haematologica

Journal of the European Hematology Association

Published by the Ferrata Storti Foundation

Ancient Greek The origin of a name that reflects Europe’s cultural roots.

aÂma [haima] = blood

a·matow [haimatos] = of blood
Scientific Latin
lÒgow [logos]= reasoning

Scientific Latin haematologicus (adjective) = related to blood

haematologica (adjective, plural and neuter,

Modern English used as a noun) = hematological subjects

The oldest hematology journal,

publishing the newest research results.
2016 JCR impact factor = 7.702

Haematologica, as the journal of the European Hematology Association (EHA), aims

not only to serve the scientific community, but also to promote European cultural
identify.
 haematologica
Journal of the European Hematology Association
Published by the Ferrata Storti Foundation

Editor-in-Chief
Luca Malcovati (Pavia)

Managing Director
Antonio Majocchi (Pavia)

Associate Editors
Omar I. Abdel-Wahab (New York), Hélène Cavé (Paris), Simon Mendez-Ferrer (Cambridge), Pavan Reddy (Ann
Arbor), Andreas Rosenwald (Wuerzburg), Monika Engelhardt (Freiburg), Davide Rossi (Bellinzona), Jacob Rowe
(Haifa, Jerusalem), Wyndham Wilson (Bethesda), Paul Kyrle (Vienna), Swee Lay Thein (Bethesda), Pieter Sonneveld
(Rotterdam)

Assistant Editors
Anne Freckleton (English Editor), Cristiana Pascutto (Statistical Consultant), Rachel Stenner (English Editor),
Kate O’Donohoe (English Editor), Ziggy Kennell (English Editor)

Editorial Board
Jeremy Abramson (Boston); Paolo Arosio (Brescia); Raphael Bejar (San Diego); Erik Berntorp (Malmö); Dominique
Bonnet (London); Jean-Pierre Bourquin (Zurich); Suzanne Cannegieter (Leiden); Francisco Cervantes (Barcelona);
Nicholas Chiorazzi (Manhasset); Oliver Cornely (Köln); Michel Delforge (Leuven); Ruud Delwel (Rotterdam);
Meletios A. Dimopoulos (Athens); Inderjeet Dokal (London); Hervé Dombret (Paris); Peter Dreger (Hamburg); Martin
Dreyling (München); Kieron Dunleavy (Bethesda); Dimitar Efremov (Rome); Sabine Eichinger (Vienna); Jean Feuillard
(Limoges); Carlo Gambacorti-Passerini (Monza); Guillermo Garcia Manero (Houston); Christian Geisler
(Copenhagen); Piero Giordano (Leiden); Christian Gisselbrecht (Paris); Andreas Greinacher (Greifswals); Hildegard
Greinix (Vienna); Paolo Gresele (Perugia); Thomas M. Habermann (Rochester); Claudia Haferlach (München); Oliver
Hantschel (Lausanne); Christine Harrison (Southampton); Brian Huntly (Cambridge); Ulrich Jaeger (Vienna); Elaine
Jaffe (Bethesda); Arnon Kater (Amsterdam); Gregory Kato (Pittsburg); Christoph Klein (Munich); Steven Knapper
(Cardiff); Seiji Kojima (Nagoya); John Koreth (Boston); Robert Kralovics (Vienna); Ralf Küppers (Essen); Ola Landgren
(New York); Peter Lenting (Le Kremlin-Bicetre); Per Ljungman (Stockholm); Francesco Lo Coco (Rome); Henk M.
Lokhorst (Utrecht); John Mascarenhas (New York); Maria-Victoria Mateos (Salamanca); Giampaolo Merlini (Pavia);
Anna Rita Migliaccio (New York); Mohamad Mohty (Nantes); Martina Muckenthaler (Heidelberg); Ann Mullally
(Boston); Stephen Mulligan (Sydney); German Ott (Stuttgart); Jakob Passweg (Basel); Melanie Percy (Ireland); Rob
Pieters (Utrecht); Stefano Pileri (Milan); Miguel Piris (Madrid); Andreas Reiter (Mannheim); Jose-Maria Ribera
(Barcelona); Stefano Rivella (New York); Francesco Rodeghiero (Vicenza); Richard Rosenquist (Uppsala); Simon Rule
(Plymouth); Claudia Scholl (Heidelberg); Martin Schrappe (Kiel); Radek C. Skoda (Basel); Gérard Socié (Paris); Kostas
Stamatopoulos (Thessaloniki); David P. Steensma (Rochester); Martin H. Steinberg (Boston); Ali Taher (Beirut);
Evangelos Terpos (Athens); Takanori Teshima (Sapporo); Pieter Van Vlierberghe (Gent); Alessandro M. Vannucchi
(Firenze); George Vassiliou (Cambridge); Edo Vellenga (Groningen); Umberto Vitolo (Torino); Guenter Weiss
(Innsbruck).

Editorial Office
Simona Giri (Production & Marketing Manager), Lorella Ripari (Peer Review Manager), Paola Cariati (Senior Graphic
Designer), Igor Ebuli Poletti (Senior Graphic Designer), Marta Fossati (Peer Review), Diana Serena Ravera (Peer Review)

Affiliated Scientific Societies

SIE (Italian Society of Hematology, www.siematologia.it)
SIES (Italian Society of Experimental Hematology, www.siesonline.it)
 haematologica
Journal of the European Hematology Association
Published by the Ferrata Storti Foundation

Information for readers, authors and subscribers

Haematologica (print edition, pISSN 0390-6078, eISSN 1592-8721) publishes peer-reviewed papers on all areas of experi-
mental and clinical hematology. The journal is owned by a non-profit organization, the Ferrata Storti Foundation, and
serves the scientific community following the recommendations of the World Association of Medical Editors
(www.wame.org) and the International Committee of Medical Journal Editors (www.icmje.org).

Haematologica publishes editorials, research articles, review articles, guideline articles and letters. Manuscripts should be
prepared according to our guidelines (www.haematologica.org/information-for-authors), and the Uniform Requirements
for Manuscripts Submitted to Biomedical Journals, prepared by the International Committee of Medical Journal Editors
(www.icmje.org).

Manuscripts should be submitted online at http://www.haematologica.org/.

Conflict of interests. According to the International Committee of Medical Journal Editors (http://www.icmje.org/#conflicts),
“Public trust in the peer review process and the credibility of published articles depend in part on how well conflict of
interest is handled during writing, peer review, and editorial decision making”. The ad hoc journal’s policy is reported in
detail online (www.haematologica.org/content/policies).

Transfer of Copyright and Permission to Reproduce Parts of Published Papers. Authors will grant copyright of their articles to the
Ferrata Storti Foundation. No formal permission will be required to reproduce parts (tables or illustrations) of published
papers, provided the source is quoted appropriately and reproduction has no commercial intent. Reproductions with com-
mercial intent will require written permission and payment of royalties.

Detailed information about subscriptions is available online at www.haematologica.org. Haematologica is an open access
journal. Access to the online journal is free. Use of the Haematologica App (available on the App Store and on Google
Play) is free.
For subscriptions to the printed issue of the journal, please contact: Haematologica Office, via Giuseppe Belli 4, 27100
Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, E-mail: info@haematologica.org).

Rates of the International edition for the year 2018 are as following:
Institutional Personal
Print edition Euro 600 Euro 150

Advertisements. Contact the Advertising Manager, Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone
+39.0382.27129, fax +39.0382.394705, e-mail: marketing@haematologica.org).

Disclaimer. Whilst every effort is made by the publishers and the editorial board to see that no inaccurate or misleading
data, opinion or statement appears in this journal, they wish to make it clear that the data and opinions appearing in the
articles or advertisements herein are the responsibility of the contributor or advisor concerned. Accordingly, the publisher,
the editorial board and their respective employees, officers and agents accept no liability whatsoever for the consequences
of any inaccurate or misleading data, opinion or statement. Whilst all due care is taken to ensure that drug doses and other
quantities are presented accurately, readers are advised that new methods and techniques involving drug usage, and
described within this journal, should only be followed in conjunction with the drug manufacturer’s own published liter-
ature.

Direttore responsabile: Prof. Edoardo Ascari; Autorizzazione del Tribunale di Pavia n. 63 del 5 marzo 1955.
Printing: Press Up, zona Via Cassia Km 36, 300 Zona Ind.le Settevene - 01036 Nepi (VT)
 haematologica
Journal of the European Hematology Association
Published by the Ferrata Storti Foundation

The 4th World Congress on Controversies in Multiple 23rd Congress of EHA

calendar of events

Myeloma June 14-17, 2018

Chairs: M Mohty, A Nagler, T Facon Stockholm, Sweden
May 3-5, 2018
Paris, France
EHA-SAH Hematology Tutorial on lymphoid Malignancies
and Plasma Cell Dyscrasias
EHA Hematology Tutorial on Thalassemia September 14-15, 2018
May 10-11, 2018 Buenos Aires, Argentina
Shiraz, Iran

EHA-SWG Scientific Meeting on Aging and Hematology

JACIE Inspector Training Course Chair: D Bron
EBMT / JACIE October 12-14, 2018
Chair: E McGrath Warsaw, Poland
May 24-25, 2018
Barcelona, Spain

35th International Congress of the ISBT

The International Society of Blood Transfusion (ISBT)
Chairs: K Pavenski, E van der Schoot, E Wood
June 2-6, 2018
Toronto, Canada

20th Congress of the European Society for

Haemapheresis
European Society for Haemapheresis
June 13-14, 2018
Valencia, Spain Calendar of Events updated on April 5, 2018
 haematologica
Journal of the European Hematology Association
Table of Contents Published by the Ferrata Storti Foundation
Volume 103, Issue 5: May 2018
Cover Figure
Cluster of osteoblasts with eccentric round to oval nuclei and abundant blue, mottled cytoplasm. Courtesy of Prof. Rosangela Invernizzi.

Editorials
747 (Auto-)immune signature in aplastic anemia
Antonio M. Risitano

749 Hematopoietic stem cell mobilization with plerixafor in sickle cell disease
Matthew M. Hsieh and John F. Tisdale

751 Age-related clonal hematopoiesis and monoclonal B-cell lymphocytosis / chronic lymphocytic leukemia: a new association?
Adalgisa Condoluci and Davide Rossi

Perspectives
753 Denosumab for bone lesions in multiple myeloma – what is its value?
Daniel A. Goldstein

755 Osteoprotective medication in the era of novel agents: a European perspective on values, risks and future solutions
Monika Engelhardt

Articles
Bone Marrow Failure
759 Deep sequencing and flow cytometric characterization of expanded effector memory CD8+CD57+ T cells frequently reveals T-cell receptor
Vβ oligoclonality and CDR3 homology in acquired aplastic anemia
Valentina Giudice et al.

Red Cell Biology & its Disorders

770 Safety and efficacy of plerixafor dose escalation for the mobilization of CD34+ hematopoietic progenitor cells in patients with sickle cell
disease: interim results
Farid Boulad et al.

778 Plerixafor enables safe, rapid, efficient mobilization of hematopoietic stem cells in sickle cell disease patients after exchange transfusion
Chantal Lagresle-Peyrou et al.

787 RON kinase inhibition reduces renal endothelial injury in sickle cell disease mice
Alfia Khaibullina et al.

Myeloproliferative Disorders
798 The KIT and PDGFRA switch-control inhibitor DCC-2618 blocks growth and survival of multiple neoplastic cell types in advanced
mastocytosis
Mathias Schneeweiss et al.

Acute Myeloid Leukemia

810 Distinct protein signatures of acute myeloid leukemia bone marrow-derived stromal cells are prognostic for patient survival
Steven M. Kornblau et al.

822 Clinical relevance of IDH1/2 mutant allele burden during follow-up in acute myeloid leukemia. A study by the French ALFA group
Yann Ferret et al.

Haematologica 2018; vol. 103 no. 5 - May 2018

http://www.haematologica.org/
 haematologica
Journal of the European Hematology Association
Published by the Ferrata Storti Foundation

Acute Lymphoblastic Leukemia

830 MSH6 haploinsufficiency at relapse contributes to the development of thiopurine resistance in pediatric B-lymphoblastic leukemia
Nikki A. Evensen et al.

Hodgkin Lymphoma
840 A phase II study of the oral JAK1/JAK2 inhibitor ruxolitinib in advanced relapsed/refractory Hodgkin lymphoma
Eric Van Den Neste et al.

Non-Hodgkin Lymphoma
849 A B-cell receptor-related gene signature predicts survival in mantle cell lymphoma: results from the Fondazione Italiana Linfomi
MCL-0208 trial
Riccardo Bomben et al.

857 Incidence and risk factors for relapses in HIV-associated non-Hodgkin lymphoma as observed in the German HIV-related lymphoma
cohort study
Philipp Schommer et al.

Chronic Lymphocytic Leukemia

865 Highly similar genomic landscapes in monoclonal B-cell lymphocytosis and ultra-stable chronic lymphocytic leukemia with low
frequency of driver mutations
Andreas Agathangelidis et al.

874 Toxicities and outcomes of 616 ibrutinib-treated patients in the United States: a real-world analysis
Anthony R. Mato et al.

Plasma Cell Disorders

880 A novel nano-immunoassay method for quantification of proteins from CD138-purified myeloma cells: biological and clinical utility
Irena Misiewicz-Krzeminska et al.

890 Impact of extramedullary disease in patients with newly diagnosed multiple myeloma undergoing autologous stem cell transplantation:
a study from the Chronic Malignancies Working Party of the EBMT
Nico Gagelmann et al.

Hemostasis
898 Immobilized fibrinogen activates human platelets through glycoprotein VI
Pierre H Mangin et al.

Coagulation & its Disorders

908 Circulating microRNAs as biomarkers of disease and typification of the atherothrombotic status in antiphospholipid syndrome
Carlos Pérez-Sánchez et al.

Letters to the Editor

Letters are available online only at www.haematologica.org/content/103/5.toc

e184 Iron overload in transfusion-dependent survivors of hemoglobin Bart’s hydrops fetalis.

Ali Amid et al.
http://www.haematologica.org/content/103/5/e184

Haematologica 2018; vol. 103 no. 5 - May 2018

http://www.haematologica.org/
 haematologica
Journal of the European Hematology Association
Published by the Ferrata Storti Foundation

e188 A defined culture method enabling the establishment of ring sideroblasts from induced pluripotent cells of X-linked sideroblastic anemia
Shunsuke Hatta et al.
http://www.haematologica.org/content/103/5/e188

e192 The mutational landscape of 18 investigated genes clearly separates four subtypes of myelodysplastic/myeloproliferative neoplasms
Manja Meggendorfer et al.
http://www.haematologica.org/content/103/5/e192

e196 Clonal evolution in the transition from cutaneous disease to acute leukemia suggested by liquid biopsy in blastic plasmacytoid dendritic
cell neoplasm
Eleni Ladikou et al.
http://www.haematologica.org/content/103/5/e196

e200 Heterozygous carriers of germline c.657_661del5 founder mutation in NBN gene are at risk of central nervous system relapse of B-cell
precursor acute lymphoblastic leukemia
Bartłomiej Tomasik et al.
http://www.haematologica.org/content/103/5/e200

e204 Ibrutinib for chronic lymphocytic leukemia: international experience from a named patient program
Peter Hillmen et al.
http://www.haematologica.org/content/103/5/e204

e207 IGHV segment utilization in immunoglobulin gene rearrangement differentiates patients with anti-myelin-associated glycoprotein neu-
ropathy from others immunoglobulin M-gammopathies
Jean-Sebastien Allain et al.
http://www.haematologica.org/content/103/5/e207

e211 Outcomes of patients with relapsed aggressive adult T-cell leukemia-lymphoma: clinical effectiveness of anti-CCR4 antibody and allogene-
ic hematopoietic stem cell transplantation
Shigeo Fuji et al.
http://www.haematologica.org/content/103/5/e211

e215 Sequential loss of tumor surface antigens following chimeric antigen receptor T-cell therapies in diffuse large B-cell lymphoma
Haneen Shalabi et al.
http://www.haematologica.org/content/103/5/e215

Case Reports
Case Reports are available online only at www.haematologica.org/content/103/5.toc

e219 Novel hereditary spherocytosis-associated splice site mutation in the ANK1 gene caused by parental gonosomal mosaicism
Xiong Wang, et al.
http://www.haematologica.org/content/103/5/e219

e223 Severe hemolysis and transfusion reactions after treatment with BGB-3111 and PD-1 antibody for Waldenström macroglobulinemia
Jad Othman et al.
http://www.haematologica.org/content/103/5/e223

e226 Usefulness of initial plasma dabigatran concentration to predict rebound after reversal
Nicolas Gendron et al.
http://www.haematologica.org/content/103/5/e226

Haematologica 2018; vol. 103 no. 5 - May 2018

http://www.haematologica.org/

(Auto-)immune signature in aplastic anemia
Antonio M. Risitano
Hematology, Department of Clinical Medicine and Surgery, Federico II University, Naples, Italy
E-mail: amrisita@unina.it
doi:10.3324/haematol.2018.190884
A
cquired idiopathic aplastic anemia (IAA) is a rare
hematologic disorder characterized by the failure of
hematopoiesis secondary to an immune-mediated
damage of the bone marrow. IAA is bona fide considered an
auto-immune disease with a T-cell-mediated pathophysiolo-
gy.1 In this issue of the Journal, Giudice et al.2 describe the
oligoclonal pattern of effector memory CD8+ CD57+ T cells in
IAA patients by using a combined deep sequencing and flow
cytometry approach. Indeed, Giudice et al. show that clonally
expanded T-cell populations are frequently detectable even
within the effector memory compartment, and that they tend
to correlate with disease activity. Thus, the characterization of
the T-cell receptor (TCR) repertoire by high-resolution tech-
niques may play a role in confirming the diagnosis of
immune-mediated IAA and to monitor affected patients dur-
ing their disease course.
There is widespread clinical and experimental evidence to
support the autoimmune pathophysiology of IAA.1 The
most striking is that patients with IAA may respond to T-
cell-targeted immunosuppressive therapies (IST), with rates
of hematologic responses ranging between 50% and 70%.3
Almost four decades of investigations have provided us with
a plethora of experimental data corroborating the hypothesis
of an immune-mediated pathophysiology. Increased circu-
lating activated T cells were described in IAA patients in the
‘80s.4 These T cells may suppress hematopoiesis through
the secretion of different inflammatory cytokines5,6 and/or
via cell-mediated direct killing. Among the different inhibito-
ry cytokines, interferon-γ (IFN-γ) plays a major role in sup-
pressing human hematopoietic stem cells (HSC) in vivo, as
suggested by in vitro inhibition of cell cycle progression and
induction of apoptosis of hematopoietic progenitors.7 More
recently, it has been suggested that IFN-γ may also exert its
inhibitory effect on HSC impairing the homeostatic survival
signal delivered by thrombopoietin through its cognate
receptor c-MPL.8 This inhibitory milieu is generated by
immune cells, and mostly by T cells that become activated
and proliferate in response to an antigen-driven stimulation.
While the search for these putative antigens has remained
unsuccessful, the demonstration of clonal expansion of T-cell
populations identified by their TCR has been considered
robust proof of a T-cell-mediated pathophysiology in IAA.9,10
Our growing understanding of the immune system and
the availability of powerful novel techniques has nurtured
continuous research in the field of IAA. On the one hand,
investigators have tried to further dissect the abnormalities
of the immune system in patients suffering from IAA.
Indeed, looking at specific functional T-cell subsets, recur-
rent immune derangements have been found, such as
increased T-helper type 17 cells (Th17)11 and reduced regula-
tory T cells (Treg).12,13 However, irrespective of the deep phe-
notyping of CD4+ T cells [i.e. by multiparameter mass
cytometry, termed cytometry by time-of-flight (CyTOF)],14
haematologica | 2018; 103(5)
EDITORIALS
only limited data are available about the characterization of
specific functional CD8+ T-cell subsets. On the other hand,
novel techniques of deep DNA sequencing have become
available, and their application in IAA has led to the descrip-
tion of somatic mutations in myeloid cells,15 with the subse-
quent ongoing debate about their actual meaning,16 but also
to high throughput TCR analysis.17 In their study, Giudice et
al. specifically investigated the compartment of effector
memory T cells (TEM) in IAA patients, looking for possible
clonality as assessed by flow cytometry analysis of Vβ usage
and sequencing of the hypervariable complementary deter-
mining region 3 (CDR3) of the TCR.
In agreement with previous reports,1,10 Giudice et al. con-
firm that AA patients often exhibit a skewed usage of Vβ
families, usually within the CD8+ T-cell compartment; these
oligoclonal expansions are more frequent in patients with
increased percentage of TEM (as defined by co-expression of
CD8 and CD57), which are found in approximately 70% of
AA patients. These gross abnormalities of the TCR Vβ usage
were dissected at the clonal level by deep sequencing of the
TCR, which allows the comparison of more than 107 reads
corresponding to TCRs harbored by individual T cells.
Indeed, clonal expansions were identified by repetitive use
of TCR β variable (TRBV) and joint (TRJV) genes (i.e. redun-
dant TRBV/TRBJ combinations), as well as by CDR3 size
and DJ length profiles. Immunodominant clones within dif-
ferent T-cell subsets were invariably detected in all AA
patients with increased CD8+ CD57+ TEM; however, the actu-
al magnitude of these clonal expansions was extremely het-
erogeneous in the different T-cell subsets. Indeed, these
clones remain minimally expanded (approx. 3%) within the
CD4+ compartment, while they become largely dominant
(approx. 18%) in the CD8+ compartment; the expansion
appears even larger when the CD8+ CD57+ TEM is analyzed.
This difference in behavior of TCR heterogeneity in differ-
ent T-cell subsets of AA patients was further confirmed by
analysis of Simpson’s diversity score, which shows how
TCR diversity decreased from CD4+ to CD8+ T cells, and
even more from both CD4+ and CD8+ T cells to CD8+ CD57+
TEM. In summary, the work performed by Giudice et al. con-
firms that clonal CD8+ T-cell expansions are common in AA
patients, in agreement with the well-established T-cell-medi-
ated immune pathophysiology of AA.1 But for the first time,
here the Authors provide evidence that the clonal expan-
sions are not limited to the CD8+ effector T cells, since they
can be found even in the TEM compartment. Very interesting-
ly, with the caveat of a limited sample size, Giudice et al.
show that abnormalities of the TEM compartment (i.e.
increased TEM, with possible clonal expansions) may be asso-
ciated with a dismal outcome in AA, due to refractory or
relapsed disease. This is further supported by the observa-
tion of concordance between the expansion of the immun-
odominant CD8+ CD57+ TEM clone and disease activity (sim-
747
 Editorials

ilar to what has already been shown for expansion within
the bulk CD8+ compartment10).
Memory is the hallmark of adaptive immunity; indeed,
antigen-driven clonal expansion of effector T cells may
generate antigen-specific lymphocytes that may persist
life long (reviewed by Sallusto et al.18). These cells, which
are known as memory cells, confer immediate and effec-
tive protection against pathogens upon antigen rechal-
lenging; indeed, memory immune cells are selected for
their higher affinity for cognate antigens, and rapidly acti-
vate from their resting state after antigen stimulation.18
Among T cells, three subtypes of memory cells have been
described:18 i) TEM, that carry the protective memory
because of their ability to deliver effector functions; ii)
central memory T cells (TCM), that home to secondary
lymphoid organs and exert reactive memory through
their cross-differentiation toward TEM and effector T cells
(TEFF); and iii) memory stem T cells (TSCM), which have
been described as a less differentiated subset that has bet-
ter self-renewal and the ability to differentiate into dis-
tinct subsets of memory T cells. Only limited data are
available about memory T cells in AA. In 2009, Hu et al.
reported that TEM and TEMRA (a further subset of terminally
differentiated TEM characterized by CD45RA expression
and by better effector function) are increased in CD4+ and
CD8+ T-cell subsets in AA patients, while naïve T cells
are decreased.19 More recently, the US National
Institutes of Health group has described that even TSCM
are increased in circulating CD8+ T cells of AA patients.20
However, while these observations depict the broad
immune derangement which is expected in an autoim-
mune disease such as AA, they do not provide any clues
about the pathogenic role of these cells. In contrast, the
demonstration of clonality within the TEM compartment
shown by Giudice et al. seems pathogenically more rele-
vant, according to the scenario depicted in Figure 1.
Indeed, a clonal immune response specific for (or cross-
reactive with) HSC can be elicited by unknown antigens
or triggers, eventually causing some impairment of
hematopoiesis. If these clonal TEFF do not undergo apop-
tosis (as usually occurrs in a physiological immune
response), they may lead to overt AA.1 Given the plas-
ticity of T cells, some of these clonal, antigen-specific
cells may eventually acquire a TEM or a TCM functional
phenotype, generating some skewing within the broad
TEM repertoire, as found by Giudice et al.2 The presence of
these clonal (possibly HSC-specific) TEM would account
for continuous damage to the hematopoiesis, since in the
presence of persistent antigen spread they may serve as
a reservoir for newly-generated TEFF. Thus, irrespective of
the regulatory role postulated by Giudice et al.,2 the pres-
ence of clonal TEM would represent the signature of a
deeper immune derangement, possibly associated with a
dismal clinical outcome.
In conclusion, the availability of high-throughput tech-
nologies is providing biomarkers which anticipate future

Figure 1. T-cell clonality in aplastic anemia. Unknown antigens and triggers may elicit a clonal immune response specific for (or cross-reactive with) hematopoietic
stem cells (HSC). These clonal, activated T cells tend to expand delivering their immune damage over hematopoiesis. At the same time, some of these clonally
expanded activated T cells may acquire an effector memory T cell (TEM) functional phenotype, leading to skewing of the TEM cell repertoire. While activated effector T
cells (TEFF) may undergo anergy or apoptosis (even as a consequence of immunosuppressive therapies), TEM represent a continuous reservoir for HSC-specific T cells,
which may exert their effector function upon rechallenging with the antigen. (This is very likely, given the typical antigen spread seen in autoimmune diseases.) Thus,
the presence of clonal TEM, which likely share the same antigen-specificity with large TEFF clones found in bulk CD8+ populations, represent a biomarker of a deep-root-
ed immune derangement, possibly associated with a dismal disease course.

748 haematologica | 2018; 103(5)


applications in the management of AA patients. Indeed,
deep whole exome sequencing,15 CyTOF14 and deep TCR
analysis2 all help to better describe the pathogenic events
underlying bone marrow failure syndromes. Even if none
of them translates into immediate therapeutic decisions,
they are all useful to confirm the diagnosis, to determine
the prognosis and possibly to monitor the clinical course
of AA patients. Indeed, this latter application may be use-
ful for early identification of refractory or relapsing
patients, paving the way for pre-emptive therapeutic
interventions. Moreover, the deep dissection at the clonal
and at the functional levels of the immune T-cell compart-
ment (e.g. combining CyTOF and TCR analysis) may also
answer some open questions in the field. For example,
the differential depletion of some specific T-cell subsets
might explain the different outcome seen with different
ATG preparations.3 These novel technologies may help
identify the specific T-cell subsets which are crucial to the
pathophysiology of AA (and possibly differentially
depleted by distinct ATG brands), possibly driving the
development of future targeted therapies.
References
1. Young NS. Current concepts in the pathophysiology and treatment
of aplastic anemia. Hematology Am Soc Hematol Educ Program.
2013;2013:76-81.
2. Giudice V, Feng X, Lin Z, et al. Deep sequencing and flow cytometric
characterization of expanded effector memory CD8+CD57+ T cells
frequently reveals T-cell receptor Vbeta oligoclonality and CDR3
homology in acquired aplastic anemia. Haematologica. 2018;103(5):
759-769.
3. Scheinberg P, Nunez O, Weinstein B, et al. Horse versus rabbit
antithymocyte globulin in acquired aplastic anemia. N Engl J Med.
2011;365(5):430-438.
4. Zoumbos NC, Gascón P, Djeu JY, Trost SR, Young NS. Circulating
activated suppressor T lymphocytes in aplastic anemia. N Engl J
Med. 1985;312(5):257-265.
5. Zoumbos NC, Gascón P, Djeu JY, Young NS. Interferon is a mediator
of hematopoietic suppression in aplastic anemia in vitro and possi-
bly in vivo. Proc Natl Acad Sci USA. 1985;82(1):188-192.
6. Sloand E, Kim S, Maciejewski JP, Tisdale J, Follmann D, Young NS.
Hematopoietic stem cell mobilization with plerixafor in sickle cell disease
Matthew M. Hsieh and John F. Tisdale
Cellular and Molecular Therapeutics Section, Sickle Cell Branch, National Heart, Lung, and Blood Institute (NHLBI), National
Institutes of Health (NIH), Bethesda, Maryland
E-mail: matthewhs@nhlbi.nih.gov
doi:10.3324/haematol.2018.190876
A
After more than a half-century since the molecular
basis for sickle cell disease (SCD) was described by
Linus Pauling and colleagues, we now possess the
molecular tools to contemplate a one-time cure through
genetic modification of autologous hematopoietic stem
cells (HSC). For these promising gene transfer and gene
editing strategies to become a reality, a sufficient number
of HSC of high purity must be obtained. Filgrastim, or
granulocyte colony-stimulating factor, mobilization and
haematologica | 2018; 103(5)
Editorials
Intracellular interferon-gamma in circulating and marrow T cells
detected by flow cytometry and the response to immunosuppressive
therapy in patients with aplastic anemia. Blood. 2002;100(4):1185-
1191.
7. Selleri C, Maciejewski JP, Sato T, Young NS. Interferon-gamma con-
stitutively expressed in the stromal microenvironment of human
marrow cultures mediates potent hematopoietic inhibition. Blood:
1996;87(10):4149-4157.
8. Alvarado LJ, Andreoni A, Huntsman HD, et al. Heterodimerization
of TPO and IFNγ Impairs Human Hematopoietic Stem/Progenitor
Cell Signaling and Survival in Chronic Inflammation Blood.
2017;130(Suppl 1):4.
9. Zeng W, Nakao S, Takamatsu H, et al. Characterization of T-cell
repertoire of the bone marrow in immune-mediated aplastic anemia:
evidence for the involvement of antigen-driven T-cell response in
cyclosporine-dependent aplastic anemia. Blood. 1999;93(9):3008-
3016.
10. Risitano AM, Maciejewski JP, Green S, Plasilova M, Zeng W, Young
NS. In-vivo dominant immune responses in aplastic anaemia: molec-
ular tracking of putatively pathogenetic T-cell clones by TCR beta-
CDR3 sequencing. Lancet. 2004;364(9431):355-364.
11. de Latour RP, Visconte V, Takaku T, et al. Th17 immune responses
contribute to the pathophysiology of aplastic anemia. Blood.
2010;116(20):4175-4184.
12. Solomou EE, Rezvani K, Mielke S, et al. Deficient CD4+ CD25+
FOXP3+ T regulatory cells in acquired aplastic anemia. Blood.
2007;110(5):1603-1606.
13. Kordasti S, Marsh J, Al-Khan S, et al. Functional characterization of
CD4+ T cells in aplastic anemia. Blood. 2012;119(9):2033-2043.
14. Kordasti S, Costantini B, Seidl T, et al. Deep phenotyping of Tregs
identifies an immune signature for idiopathic aplastic anemia and
predicts response to treatment. Blood. 2016;128(9):1193-1205.
15. Yoshizato T, Dumitriu B, Hosokawa K, et al. Somatic Mutations and
Clonal Hematopoiesis in Aplastic Anemia. N Engl J Med.
2015;373(1):35-47.
16. Cooper JN, Young NS. Clonality in context: hematopoietic clones in
their marrow environment. Blood. 2017;130(22):2363-2372.
17. Calis JJ, Rosenberg BR. Characterizing immune repertoires by high
throughput sequencing: strategies and applications. Trends
Immunol. 2014;35(12):581-590.
18. Sallusto F, Geginat J, Lanzavecchia A. Central memory and effector
memory T cell subsets: function, generation, and maintenance. Annu
Rev Immunol. 2004;22:745-763.
19. Hu X, Gu Y, Wang Y, Cong Y, Qu X, Xu C. Increased CD4+ and
CD8+ effector memory T cells in patients with aplastic anemia.
Haematologica. 2009;94(3):428-429.
20. Hosokawa K, Muranski P, Fenx X, et al. Memory Stem T Cells in
Autoimmune Disease: High Frequency of Circulating CD8+
Memory Stem Cells in Acquired Aplastic Anemia. J Immunol.
2016;196(4):1568-1578.
apheresis is the standard method for HSC collection in
healthy adult donors, yet this approach is associated with
high rates of adverse events requiring hospitalization in
SCD, including vaso-occlusive crises, multi-organ failure,
and even death, prompting our call for a moratorium on its
use for HSC mobilization in SCD.1 Thus, bone marrow
harvesting is the default approach, with evidence support-
ing its utility in both animal models and in vitro studies uti-
lizing patients’ material.2-4 However, bone marrow harvest-
749
 Editorials
ing employed in an ongoing HSC gene therapy trial was
recently recognized to result in suboptimal yields of high
purity HSC at the end of collection and processing, along
with substantial pain after each harvest, and most subjects
required two or three harvests to yield sufficient cell doses
for manufacturing.5,6
In this issue of the Journal, two groups of investigators
report their results using a third approach to HSC collec-
tion in SCD through mobilization with an inhibitor of the
CXCR4 chemokine receptor, plerixafor. Boulad et al. per-
formed a dose escalation study of plerixafor among a
total of 15 SCD patients at steady state.7 Ten of the
patients were receiving concomitant treatment with
hydroxyurea. Only a minority of patients in each cohort
achieved the target of ≥30 CD34+ cells/μL at 12 h after the
plerixafor injection: three out six at a dose of 80 μg/kg,
one out of three at a dose of 160 μg/kg, and two out of six
at a dose of 240 μg/kg. Two patients (15%) experienced a
vaso-occlusive crisis during the study period – one each at
80 and 240 μg/kg. None of the patients underwent leuka-
pheresis, thus attribution of these adverse events could be
narrowed to plerixafor. On the other hand, Lagresle-
Peyrou et al. reported the outcomes of three patients who
received plerixafor at a dose of 240 μg/kg.8 All three
patients received at least 2 months of red cell exchange
transfusion to target a sickle hemoglobin (HbS) near 30%
while hydroxyurea was discontinued. The peak CD34+
cell count reached >75 cells/μL at as early as 3 h after the
injection. All three patients also underwent leukapheresis
of 15 to 21 L, with a resulting total CD34+ cell yield of 4.5
to 5.8 x 106 cells/kg and a purity of 80% to 95%. No pain,
vaso-occlusive crises, or sickle-related events were
observed in these three patients.
While the number of patients is relatively small in both
studies, important lessons relevant to autologous HSC
mobilization and collection in SCD with plerixafor can be
gleaned. The first lesson regards preparation of the
patients. Specifically, stopping hydroxyurea and utilizing
red cell transfusions, simple or exchange, to target a HbS
of 30% were likely key factors in the successful mobiliza-
tion of the series reported by Lagresle-Peyrou et al.
Conversely, the absence of these measures in the study
by Boulad et al. may explain why the majority of their
patients failed to reach the target CD34+ concentration.
This is consistent with prior work demonstrating a lower
CD34+ cell content in the marrow of SCD patients on
hydroxyurea when compared to those not on the drug.3
Discontinuation of hydroxyurea combined with sched-
uled red cell transfusion to keep the HbS near 30% may
also have improved purity, which was 80% to 95% in the
study by Lagresle-Peyrou et al., while helping to mini-
mize the risk of sickle cell-related adverse events while
hydroxyurea treatment was interrupted. Secondly, leuko-
cyte and neutrophil counts increased 2- to 3-fold just
hours after a single injection of plerixafor, even at the
lowest dose of 80 μg/kg tested. Although increases of a
similar magnitude also occurred with filgrastim, the
adverse events seen with filgrastim may have been relat-
ed to the prolonged duration of 5 to 6 days from filgras-
tim that led to the high rates of adverse events in the ear-
750
lier reports. We will need more patients to ascertain the
contribution of leukocytosis alone and/or the duration of
leukocytosis in developing sickle-related complications.
Thirdly, only three patients underwent leukapheresis and
though adverse events appear acceptable, expanded
accrual could capture additional side effects. Furthermore,
if patients with SCD do not meet the goal and need addi-
tional mobilization and collection, there could be cumu-
lative side effects. Finally, the peak of mobilization of
CD34+ cells appeared to be much earlier, at 3-6 hours.
This observation is distinct from that in healthy donors,
in whom the peak is observed at 6-12 hours.9 Perhaps the
chronically hyperproliferative marrow in SCD partly
explains this early release of HSC; there could be other
factors at play. Regardless, this observation suggests that
for optimal collection, apheresis should be started within
4-6 hours of dosing.
As clinical applications of gene transfer and gene editing
strategies are being implemented in SCD, obtaining ade-
quate numbers of HSC safely from patients could be the
‘bottleneck’, preventing broad dissemination of these
exciting approaches. The early results provide optimism
that mobilization with plerixafor could be a safer and more
efficacious alternative for HSC collection to either filgras-
tim mobilization or bone marrow harvesting, and provide
general confidence for the further development of these
promising approaches to a one-time cure for SCD.
©2017 NIH (National Institutes of Health)
References
1. Fitzhugh CD, Hsieh MM, Bolan CD, Saenz C, Tisdale JF. Granulocyte
colony-stimulating factor (G-CSF) administration in individuals with
sickle cell disease: time for a moratorium? Cytotherapy.
2009;11(4):464-471.
2. Hematti P, Tuchman S, Larochelle A, Metzger ME, Donahue RE,
Tisdale JF. Comparison of retroviral transduction efficiency in CD34+
cells derived from bone marrow versus G-CSF-mobilized or G-CSF
plus stem cell factor-mobilized peripheral blood in nonhuman pri-
mates. Stem Cells. 2004;22(6):1062-1069.
3. Uchida N, Fujita A, Hsieh MM, et al. Bone marrow as a hematopoietic
stem cell source for gene therapy in sickle cell disease: evidence from
Rhesus and SCD patients. Hum Gene Ther Clin Dev. 2017;28(3):136-
144.
4. Uchida N, Bonifacino A, Krouse AE, et al. Accelerated lymphocyte
reconstitution and long-term recovery after transplantation of lentivi-
ral-transduced rhesus CD34+ cells mobilized by G-CSF and plerixafor.
Exp Hematol. 2011;39(7):795-805.
5. Kanter J, Walters MC, Hsieh MM, et al. Interim results from a phase I/II
clinical study of lenitoglobin gene therapy for severe sickle cell disease.
Blood. 2016;128(22):1176.
6. Leonard A, Bonifacino A, Dominical VM, et al. Bone marrow charac-
terization in sickle cell disease: inflammation and stress erythropoiesis
lead to suboptimal CD34 recovery compared to normal volunteer bone
marrow. Blood. 2017;130(Suppl 1):966.
7. Boulad F, Shore T, van Besien K, et al. Safety and efficacy of plerixafor
dose escalation for the mobilization of CD34+ hematopoietic progen-
itor cells in patients with sickle cell disease: interim results.
Haematologica. 2018;103(5):770-777.
8. Lagresle-Peyrou C, Lefrere F, Magrin E, et al. Plerixafor enables the safe,
rapid, efficient mobilization of haematopoietic stem cells in sickle cell
disease patients after exchange transfusion. Haematologica.
2018;103(5):778-786.
9. Pantin J, Purev E, Tian X, et al. Effect of high-dose plerixafor on
CD34(+) cell mobilization in healthy stem cell donors: results of a ran-
domized crossover trial. Haematologica. 2017;102(3):600-609.
haematologica | 2018; 103(5)

Age-related clonal hematopoiesis and monoclonal B-cell lymphocytosis / chronic
lymphocytic leukemia: a new association?
Adalgisa Condoluci and Davide Rossi
Hematology, Institute of Oncology Research, and Oncology Institute of Southern Switzerland, Bellinzona, Switzerland
E-mail: davide.rossi@eoc.ch
doi:10.3324/haematol.2018.191098
A
gathangelidis et al.1 compared the mutational land-
scape of low-count monoclonal B-cell lymphocyto-
sis (MBL), high-count MBL and highly stable chron-
ic lymphocytic leukemia (CLL) with confirmed lack of pro-
gression after a very long follow up (>10 years). The
Authors also studied the polymorphonuclear (PMN) frac-
tion and germline DNA from buccal swabs of the same
individuals. Whole genome sequencing was complement-
ed with deep sequencing of targeted genes.
The sample size, along with the low coverage imposed
by whole genome sequencing, are both limitations in
efforts to discover yet unknown variants that might actu-
ally be recurrent in these conditions. While
Agathangelidis et al. acknowledge these limitations, three
major findings characterize their manuscript.1 First, low-
count MBL, high-count MBL and highly stable CLL share
a similar genetic landscape and mutational signatures,
which include the presence of mutations in known driv-
ers associated with poor outcome, such as NOTCH1,
SF3B1, POT1,2-4 indicating that these mutations are not
sufficient to drive the aggressiveness of the disease by
themselves. Second, the mutational landscape of paired
PMN suggests that most of these patients carry a clonal
hematopoiesis that could possibly be age-related. Third,
a number of somatic mutations were found in both the
MBL/CLL cells and PMN, supporting the idea that the
MBL/CLL clone stemmed from a common ancestral
Figure 1. Hypothetical model of evolution from age-related clonal hematopoiesis to monoclonal B-cell lymphocytosis/chronic lymphocytic leukemia.
haematologica | 2018; 103(5)
Editorials
hematopoietic precursor that was able to participate in
both lymphoid and myeloid differentiation. By docu-
menting that the DNA from PMN was free from contam-
ination by MBL/CLL DNA, for example, by using molec-
ular minimal residual disease methods relying on the
individual patient’s specific immunoglobulin gene
rearrangements, the Authors have provided further evi-
dence of this important finding which, although previ-
ously reported,5,6 had remained a subject of debate.
Several hematologic malignancies, including CLL, mul-
tiple myeloma (MM) and acute myeloid leukemia (AML),
have well-defined precursor states that precede the devel-
opment of overt cancer. CLL is always preceded by a high
MBL count,7 MM is almost always preceded by mono-
clonal gammopathy of undetermined significance
(MGUS),8 and at least a quarter of all patients with
myelodysplastic syndromes (MDS) have disease that
evolves into AML.9 Deep genomic sequencing of normal
subjects revealed that during human aging, the expansion
of 1 or more hematopoietic stem and progenitor cells
(HSPC) will result in clones that will sustainably con-
tribute more than others to the production of mature
blood cells. Accordingly, age-related clonal hematopoiesis
(ARCH) is defined as the expansion of HSPC clones, har-
boring specific, disruptive, and recurrent genetic variants,
in individuals without clear diagnosis of hematologic
malignancies.10 MDS are frequently preceded by ARCH.11
751
 Editorials
Some ARCH-related mutations can increase the risk for
leukemia,12 while others possibly increase the risk for
heart disease and diabetes.13 From a pathogenetic stand-
point, the study of Agathangelidis et al.1 provides the
proof of principle that ARCH may also associate with
expansion of B-cell clones with CLL phenotype, and con-
nects ARCH with MBL and CLL in a continuum of evolu-
tion from HSCP clones to mature B-cell clones (Figure 1),
thus validating in vivo in patients the notion initially
reported from mice studies that the propensity to gener-
ate clonal B cells has already been acquired at the HSCP
stage.14 To robustly establish this association and to gain
greater insight into the pathogenetics, larger cohorts of
MBL and CLL patients should be investigated with the
rigorous approach utilized by Agathangelidis et al.1
One of the long-term complications of chemoim-
munotherapy in CLL is the development of treatment-
related MDS/AML.15 Chemoimmunotherapy poses a
strong selection bottleneck to HSCPs, and thus only the
fittest HSCPs survive and repopulate after the stress of
chemoimmunotherapy.16 HSCP fitness may be sustained
by somatic mutations in the context of a preceding
ARCH, and it is increasingly recognized as a risk factor
for therapy-related MDS/AML.17 Among elderly patients
who receive chemotherapy and develop therapy-related
MDS/AML, most have ARCH before chemotherapy.
Consistently, ARCH associates with an increased rate of
therapy-related AML/MDS.17 Following this line of evi-
dence, the study by Agathangelidis et al.1 prompts inves-
tigation into whether the finding of an ARCH in CLL
patients who receive chemoimmunotherapy is a risk fac-
tor for the development of therapy-related MDS/AML. If
this is proved to be the case, given the availability of
novel agents for the treatment of CLL that are not stress-
ful for HSCP, ARCH might become a new biomarker for
tailoring treatment in CLL.
References
1. Agathangelidis A, Ljungström V, Scarfò L, et al. Highly similar
genomic landscapes in monoclonal B-cell lymphocytosis and ultra-
stable chronic lymphocytic leukemia with low frequency of driver
752
mutations. Haematologica. 2018;103(5):865-873.
2. Rossi D, Rasi S, Fabbri G, et al. Mutations of NOTCH1 are an inde-
pendent predictor of survival in chronic lymphocytic leukemia.
Blood. 2012;119(2):521-529.
3. Wang L, Lawrence MS, Wan Y, et al. SF3B1 and other novel cancer
genes in chronic lymphocytic leukemia. N Engl J Med.
2011;365(26):2497-2506.
4. Herling CD, Klaumünzer M, Rocha CK, et al. Complex karyotypes
and KRAS and POT1 mutations impact outcome in CLL after chlo-
rambucil-based chemotherapy or chemoimmunotherapy. Blood.
2016;128(3):395-404.
5. Damm F, Mylonas E, Cosson A, et al. Acquired initiating mutations
in early hematopoietic cells of CLL patients. Cancer Discov.
2014;4(9):1088-1101.
6. Quijada-Álamo M, Hernández-Sánchez M, Robledo C, et al. Next-
generation sequencing and FISH studies reveal the appearance of
gene mutations and chromosomal abnormalities in hematopoietic
progenitors in chronic lymphocytic leukemia. J Hematol Oncol.
2017;10(1):83.
7. Landgren O, Albitar M, Ma W, et al. B-cell clones as early markers
for chronic lymphocytic leukemia. N Engl J Med. 2009;360(7):659-
667.
8. Landgren O, Kyle RA, Pfeiffer RM, et al. Monoclonal gammopathy
of undetermined significance (MGUS) consistently precedes multiple
myeloma: a prospective study. Blood. 2009;113(22):5412-5417.
9. Pfeilstöcker M, Tuechler H, Sanz G, et al. Time-dependent changes
in mortality and transformation risk in MDS. Blood.
2016;128(7):902-910.
10. Jaiswal S, Fontanillas P, Flannick J, et al. Age-related clonal
hematopoiesis associated with adverse outcomes. N Engl J Med.
2014;371(26):2488-2498.
11. Malcovati L, Gallì A, Travaglino E, et al. Clinical significance of
somatic mutation in unexplained blood cytopenia. Blood.
2017;129(25):3371-3378.
12. Genovese G, K¨ahler AK, Handsaker RE, et al. Clonal hematopoiesis
and blood-cancer risk inferred from blood DNA sequence. N Engl J
Med. 2014;371(26):2477-2487.
13. Jaiswal S, Natarajan P, Silver AJ, et al. Clonal hematopoiesis and risk
of atherosclerotic cardiovascular disease. N Engl J Med.
2017;377(2):111-121.
14. Kikushige Y, Ishikawa F, Miyamoto T, et al. Self-renewing
hematopoietic stem cell is the primary target in pathogenesis of
human chronic lymphocytic leukemia. Cancer Cell. 2011;20(2):246-
259.
15. Benjamini O, Jain P, Trinh L, et al. Second cancers in patients with
chronic lymphocytic leukemia who received frontline fludarabine,
cyclophosphamide and rituximab therapy: distribution and clinical
outcomes. Leuk Lymphoma. 2015;56(6):1643-1650.
16. Wong TN, Ramsingh G, Young AL, et al. Role of TP53 mutations in
the origin and evolution of therapy-related acute myeloid leukaemia.
Nature. 2015;518(7540):552-555.
17. Takahashi K, Wang F, Kantarjian H, et al. Preleukaemic clonal
haemopoiesis and risk of therapy-related myeloid neoplasms: a case-
control study. Lancet Oncol. 2017;18(1):100-111.
haematologica | 2018; 103(5)

Denosumab for bone lesions in multiple myeloma – what is its value?
Daniel A. Goldstein
Davidoff Cancer Center, Rabin Medical Center, Petach Tikvah, Israel
E-mail: danielagoldstein@gmail.com
doi:10.3324/haematol.2017.185264
I
n June 2017 the Food and Drug Administration (FDA)
accepted a supplemental biologics license application
seeking to expand the currently approved indication of
denosumab to patients with bone lesions from multiple
myeloma. The FDA set a prescription drug user act
(PDUFA) action date of February 3, 2018. Denosumab is an
inhibitor of receptor activator of nuclear factor κ-B ligand
(RANKL) and was previously approved for post-
menopausal women at risk of osteoporosis in addition to
patients at risk of skeletal-related events due to bone
metastases from solid tumors and giant cell tumors of the
bone. The application for use in patients with myeloma is
based on the findings of the recently presented ‘482 trial.1
This commentary seeks to understand the value of this
therapy for patients with multiple myeloma.
Denosumab is a monoclonal antibody and uses a novel
mechanism to decrease bone resorption. RANKL is a pro-
tein expressed on osteoblastic stromal cells. It binds to
receptor activator of nuclear factor-κB (RANK) and thus
mediates osteoclastic differentiation, activation, and sur-
vival. RANKL therefore controls osteoclast-mediated
bone resorption. Osteoprotegerin is a soluble RANKL
decoy receptor which binds RANKL and is the key regu-
lator of the RANKL–RANK pathway. Denosumab binds
to RANKL thus blocking the interaction of RANKL with
RANK, mimicking the endogenous effects of osteoprote-
gerin. This agent has been shown to lead to a decrease in
bone resorption, based on changes in serum and urinary
N-telopeptide, which are markers of osteoclastic bone
resorption.2
Until recently bisphosphonates had been the standard
therapy for strengthening bone in a variety of conditions
such as osteoporosis and cancer. Bisphosphonates essen-
tially bind to bone mineral and inhibit the activity of
mature osteoclasts. Non-nitrogen containing bisphospho-
nates achieve this goal by being metabolized to ATP
analogs that block osteoclast function and induce osteo-
clast apoptosis. Nitrogen-containing bisphosphonates
inhibit farnesyl pyrophosphate synthase, thus preventing
the post-translational modification of guanosine triphos-
phate binding proteins which are essential for osteoclast
function and survival.3 The essential difference between
bisphosphonates and denosumab is that bisphosphonates
inhibit mature osteoclasts while denosumab inhibits
osteoclastic precursors.
Denosumab has already gained FDA approval for mul-
tiple indications based on advanced phase clinical trials.
In postmenopausal women with low bone mineral densi-
ty, it was found to lead to a 3.0% to 6.7% increase in
bone mineral density of the lumbar spine.2 Multiple trials
have compared zoledronic acid and denosumab in
patients with solid tumors. In patients with bone metas-
tases from breast cancer, denosumab was superior to
haematologica | 2018; 103(5)
PERSPECTIVES
zoledronic acid in delaying or preventing first on-study
skeletal-related event [hazard ratio (HR)=0.82; 95% con-
fidence interval (95% CI): 0.71- 0.95; P= 0.01).4 Likewise,
denosumab was superior in terms of time to first skeletal-
related event in patients with bone metastases from
prostate cancer. The median time to first on-study skele-
tal-related event was 20.7 months (95% CI: 18.8-24.9)
with denosumab compared to 17.1 months (95% CI:
15.0-19.4) with zoledronic acid (HR=0.82, 95% CI: 0.71-
0.95; P=0.008 for superiority).5 Despite the reduction in
skeletal-related events with denosumab, there was no
associated improvement in overall survival in patients
with either breast or prostate cancer.4,5 In patients with
giant cell tumors of the bone, an open label study with
denosumab demonstrated a high level of efficacy: 96% of
patients with surgically unsalvageable giant cell tumors of
the bone did not have disease progression after a median
follow-up of 13 months.6
The ‘482 trial was an international phase 3, random-
ized, double-blind trial comparing the safety and efficacy
of monthly denosumab to monthly zoledronic acid in
patients with multiple myeloma.1 The trial enrolled 1718
patients and the primary endpoint was the time to first
on-study skeletal-related event, and was powered to
demonstrate non-inferiority. Secondary endpoints were
time to first skeletal-related event (powered to superiori-
ty), time to first and subsequent skeletal-related events
(powered to superiority), and overall survival. The study
met the primary endpoint and demonstrated that deno-
sumab was non-inferior to zoledronic acid in terms of
skeletal-related events (HR=0.98; 95% CI: 0.85-1.14;
P=0.01). The trial failed to meet the secondary endpoints
of demonstrating superiority in terms of time to first
skeletal-related event or overall survival. The authors per-
formed an unplanned exploratory analysis to evaluate
progression-free survival as an endpoint and found a pro-
longed progression-free survival in the denosumab group
(HR=0.82; 95% CI: 0.68-0.99; P=0.036). Although the
trial was well conducted with double-blind randomiza-
tion, this finding should be considered only as hypothe-
sis-generating, as it was an unplanned endpoint analysis
and such analyses are known to have a lack of statistical
reliability.7
There were no significant differences between the two
groups in terms of adverse events apart from hypocal-
cemia and renal toxicity. In patients with baseline creati-
nine clearance ≤60 mL/minute, 13% of patients in the
denosumab group developed renal toxicity, compared to
26% of patients in the zoledronic acid group (P<0.01).
The rate of creatinine doubling from baseline in the zole-
dronic acid group was nearly twice as high as in the deno-
sumab group (6.5 versus 3.3%). Conversely, there were
higher rates of hypocalcemia in patients receiving deno-
753
 Perspectives
sumab (17%) compared to those receiving zoledronic
acid (12%) (P=0.009).
In the USA we can calculate the cost of the drugs to
Medicare by using the average sales price (www.cms.gov).
This accounts for discounts and rebates and is a close esti-
mate of the cost to Medicare. The patent for zoledronic
acid expired in 2013, at which point the reimbursement
cost decreased. The average sales price for 4 mg of zole-
dronic acid is $48 and that for 120 mg of denosumab is
$2044. The annual cost is therefore $576 for zoledronic
acid, and $24,528 for denosumab – a difference of almost
$24,000. In addition to this cost there is a mark-up of
4.3% that Medicare reimburses to providers. It should be
noted that this mark-up may provide a financial incentive
to the physician to prescribe the more expensive medica-
tion, despite the higher cost to the patient and insurer.
Finally, treatment centers also charge an infusion cost of
approximately $140, billed with code 96413
(www.cms.gov). While these are the costs to Medicare,
we must also recognize that the patient often shares a sig-
nificant proportion of the cost. In 2015, the average annu-
al Medicare beneficiary cost share was $527 for deno-
sumab and $68 for zoledronic acid (www.cms.gov - 2015
Medicare drug spending data). The price of drugs is dif-
ferent in other countries around the world; however, it is
clear that everywhere in the world zoledronic acid is sig-
nificantly cheaper than denosumab. This commentary is
not intended to assess what was the most appropriate
launch price of these drugs at the very different times of
their being launched. The purpose is to discuss the most
appropriate choice of therapy in 2018, when the prices
are significantly different, due to one of the options being
available in the significantly cheaper, generic form.
There is some additional convenience from using deno-
sumab. Firstly, denosumab can be given subcutaneously
which may be preferable to the intravenous administra-
tion of zoledronic acid. Secondly, denosumab is dosed the
same for all patients, and no adjustment is needed accord-
ing to renal function, whereas dose adjustments are nec-
essary for zoledronic acid. It is doubtful however, that
this additional convenience justifies the additional annual
cost in the USA of $24,000 per patient.
Recent data for zoledronic acid demonstrate equivalent
efficacy in patients with bone metastases secondary to
breast cancer, irrespective of whether the drug is given
monthly or every 3 months.8 Could these data perhaps be
extrapolated to patients with multiple myeloma? There
are currently no good quality data regarding the use of
denosumab every 3 months in patients with neoplastic
bone disease.
In an era of financial challenges for healthcare, we, as
physicians, must be careful stewards of finite healthcare
resources. There appears to be no benefit from using
denosumab instead of zoledronic acid in terms of overall
754
survival or skeletal events. In addition, the safety profile
is very similar. There appears to be slightly more renal
toxicity with zoledronic acid; however, this is balanced
by the higher rates of hypocalcemia with denosumab.
Although there was a demonstration of benefit in terms
of progression-free survival, this finding should be treated
with caution, as it emerged from a post hoc exploratory
analysis. There are, however, significant differences in
costs – both to society and to patients. Denosumab costs
approximately $24,000 more per patient per year in the
USA. Zoledronic acid is also considerably cheaper than
denosumab in Europe as well. Perhaps the most appropri-
ate management would be for all patients to receive zole-
dronic acid, except those with a contraindication due to a
low creatinine clearance. The reason for the high cost of
new cancer drugs is complex. Without doubt, one of the
many reasons is that the cost of drug development is
high, partially related to the many regulatory require-
ments. However, while cancer is still often an incurable
disease, we must strive towards bringing forward new
therapies that provide clinically meaningful benefits to
our patients.9 In an era of medical bancruptcies and
increasing healthcare costs, we owe it to both our
patients and society to incorporate costs into clinical deci-
sion-making.
References
1. Raje NS, Roodman GD, Willenbacher W, et al. Impact of denosumab
(DMB) compared with zoledronic acid (ZA) on renal function in the
treatment of myeloma bone disease. J Clin Oncol.
2017;35(15_suppl):8005.
2. McClung MR, Lewiecki EM, Cohen SB, et al. Denosumab in post-
menopausal women with low bone mineral density. N Engl J Med.
2006;354(8):821-831.
3. Baron R, Ferrari S, Russell RG. Denosumab and bisphosphonates:
different mechanisms of action and effects. Bone. 2011;48(4):677-
692.
4. Stopeck AT, Lipton A, Body J-J, et al. Denosumab compared with
zoledronic acid for the treatment of bone metastases in patients with
advanced breast cancer: a randomized, double-blind study. J Clin
Oncol. 2010;28(35):5132-5139.
5. Fizazi K, Carducci M, Smith M, et al. Denosumab versus zoledronic
acid for treatment of bone metastases in men with castration-resis-
tant prostate cancer: a randomised, double-blind study. Lancet.
2011;377(9768):813-822.
6. Chawla S, Henshaw R, Seeger L, et al. Safety and efficacy of deno-
sumab for adults and skeletally mature adolescents with giant cell
tumour of bone: interim analysis of an open-label, parallel-group,
phase 2 study. Lancet Oncol. 2013;14(9):901-908.
7. Raghav KP, Mahajan S, Yao JC, et al. From protocols to publications:
a study in selective reporting of outcomes in randomized trials in
oncology. J Clin Oncol. 2015;33(31):3583-3590.
8. Hortobagyi GN, Van Poznak C, Harker WG, et al. Continued treat-
ment effect of zoledronic acid dosing every 12 vs 4 weeks in women
with breast cancer metastatic to bone: the OPTIMIZE-2 randomized
clinical trial. JAMA Oncol. 2017;3(7):906-912.
9. Ellis LM, Bernstein DS, Voest EE, et al. American Society of Clinical
Oncology perspective: raising the bar for clinical trials by defining
clinically meaningful outcomes. J Clin Oncol. 2014;32(12):1277-
1280.
haematologica | 2018; 103(5)

Osteoprotective medication in the era of novel agents: a European perspective on values,
risks and future solutions
Monika Engelhardt,1,2 Georg W. Herget,3 Giulia Graziani,1,2 Gabriele Ihorst,4 Heike Reinhardt,1,2 Stefanie Ajayi,1,2
Stefan Knop5 and Ralph Wasch1,2
1
Department of Medicine I, Hematology, Oncology & Stem Cell Transplantation, Medical Center, University of Freiburg;
2
Comprehensive Cancer Center Freiburg (CCCF), Medical Center, University of Freiburg; 3Department of Orthopedics and
Trauma Surgery, Medical Center, University of Freiburg; 4Clinical Trials Unit, Medical Center, University of Freiburg and
5
Hematology, Oncology, Gastroenterology, University of Würzburg, Germany
E-mail: monika.engelhardt@uniklinik-freiburg.de
doi:10.3324/haematol.2018.188516
O
steolytic bone disease is one of the most promi-
nent features of multiple myeloma (MM) and is
present in up to 80% of patients at diagnosis.1
Bone destruction leads to skeletal-related events (i.e. verte-
bral and other pathological fractures) and/or spinal cord
compression. MM is mainly due to an increase in osteo-
clastic activity which is accompanied by low osteoblastic
function.1 Bisphosphonates and other bone-targeting
agents (such as denosumab which inhibits RANKL and
osteoclast function and is not renally cleared), effective
anti-myeloma treatment, radiotherapy and surgery are the
main therapies used for the management of bone disease
in MM.1–3 Regarding the definition of MM-defining events,
there are important studies which suggest that asympto-
matic patients with more than one focal lesion detectable
by magnetic resonance imaging have a higher risk of pro-
gression to symptomatic MM (>70% within 2 years).1,4–6
These patients have been described by international
myeloma experts as having symptomatic disease.5,6
Based on phase 3 studies, the bisphosphonates,
pamidronate and zoledronic acid, have been found to
reduce skeletal-related events compared to placebo.7–9
Three randomized studies have compared the effect of
different bisphosphonates or different dosages of the
same bisphosphonate. In the first study, zoledronic acid
was as effective as pamidronate in reducing skeletal-relat-
ed events in the era of conventional chemotherapy.9,10 In
the second, two doses of intravenous pamidronate (30
versus 90 mg) showed comparable results regarding time
to skeletal-related events and survival time free of such
events.11 The limitation of this study was that it was pow-
ered to show differences in quality of life and not in
skeletal-related events.11 The third study compared intra-
venous (i.v.) zoledronic acid with oral clodronate and
showed that zoledronic acid reduced the risk of skeletal-
related events compared to clodronate in all MM patients,
irrespective of the presence of lytic lesions at diagnosis,
and improved overall survival by 10 months in patients
with lytic lesions at diagnosis.12,13 These effects continued
in patients who received zoledronic acid for >2 years.14
There was no sub-analysis according to the response sta-
tus of the patients, thus it is not clear whether the contin-
uous use of zoledronic acid produces similar results in
patients who have achieved excellent responses (≥very
good partial response). A meta-analysis was unable to
confirm superiority of zoledronic acid over pamidronate,
but revealed a survival advantage from zoledronic acid
versus placebo.15 This analysis also determined that in
haematologica | 2018; 103(5)
Perspectives
order to prevent one skeletal-related event, 6-15 MM
patients need to be treated.15
The European Myeloma Network (EMN) and
International Myeloma Working Group (IMWG) have
therefore recommended that all MM patients with ade-
quate renal function (creatinine clearance >30 mL/min)
and osteolytic disease at diagnosis should be treated with
zoledronic acid [4 mg i.v. infusion, over at least 15 min,
every 4 weeks (Q4W) or pamidronate (90 mg, in a 3-hour
infusion, Q4W], in addition to specific anti-myeloma
therapy (grade 1A; definition of evidence levels: Online
Supplementary Table S1). Symptomatic patients, without
bone disease assessed by conventional radiography, can
be treated with zoledronic acid (grade 1B). The advantage
is not clear for patients without detectable bone involve-
ment on magnetic resonance imaging or positron emis-
sion tomography/computed tomography.
Bisphosphonates are not routinely recommended in
smoldering MM (grade 1A); but in cases of osteoporosis
or vertebral fractures that are not due to the MM, bispho-
sphonates should be given at the doses given for osteo-
porosis (5 mg zoledronic acid/year). For high-risk smol-
dering MM, the treating physician should consider using
the bisphosphonate doses and schedules typically used to
treat symptomatic MM (grade 1B). Zoledronic acid
should be given continuously (grade 1B). It is debatable
whether patients who achieve a very good partial
response or better have benefits from the continuous use
of zoledronic acid. Regarding pamidronate, there are no
data to support its continuous use; thus it should be given
for 2 years and then at the physician’s discretion (grade
2C).1,2 Of note, bisphosphonates are now available as
generic drugs, whereas denosumab has just been
approved by the Food and Drug Administration for use in
MM (January 2018; likewise anticipated in Europe) and is
patent-protected. This approval was based on the results
of a large phase III study comparing denosumab with
zoledronic acid, in which the efficacy and safety of the
drugs were assessed in newly diagnosed MM. Eligible
patients were randomized 1:1 to denosumab 120 mg sub-
cutaneously Q4W or zoledronic acid 4 mg (with dose
adjustments according to renal function) i.v. Q4W along
with anti-myeloma therapy. The primary objective was
non-inferiority of denosumab to zoledronic acid with
respect to time to first on-study skeletal-related event.
Overall survival was a secondary endpoint; progression-
free survival was an exploratory endpoint. The 1718
patients enrolled were randomized into two arms, each
755
 Perspectives

with 859 participants. With regards to delaying time to
first on-study skeletal-related event, denosumab was not
inferior to zoledronic acid [P=0.01; hazard ratio
(HR)=0.98; 95% confidence interval (95% CI): 0.85-1.14].
Fewer adverse events potentially related to renal impair-
ment were reported with denosumab than with zole-
dronic acid (10.0% versus 17.1%, P<0.001). The HR for
progression-free survival was 0.82 (95% CI: 0.68-0.99;
P=0.036). The overall survival HR between denosumab
and zoledronic acid was 0.9 (95% CI: 0.70-1.16; P=0.41),
with fewer deaths in the denosumab arm (n=121; 14.1%)
than in the zoledronic acid group (n=129; 15.0%).
Therefore, denosumab showed non-inferiority to zole-
dronic acid in delaying time to first on-study skeletal-
related event. Patients on denosumab had a significantly
lower rate of renal adverse events compared to those on
zoledronic acid. The bone-specific benefits in combina-
tion with the renal function results and possible prolonga-
tion of progression-free survival with denosumab were
promising and have led to invigorating discussions about
why progression-free survival data were more favorable
with denosumab. The observations definitely need deep-

Table 1. Cost comparison of osteoprotective medications for MM: Germany versus USA.
Costs Germany Costs USA
[Monthly costs in Euro (€)] [Monthly costs in Euro (€)]
Drug (original) Dose & mode of administration Original price Generic price Original price Generic price
Denosumab (Xgeva ) ®
120 mg s.c. bolus 440 - 1890 -
Zoledronic acid (Zometa®) 4 mg i.v. over 15 min 368 279 814 49
Pamidronate 90 mg i.v. over 3 hours out of trade 251 out of trade 47
German prices: ATaxx® (Dr. Heni Software GmbH & Co.KG, Freiburg); USA prices: https://www.drugs.com/price-guide/. In Germany, the AMNOG is limiting the cost of new phar-
maceutical products; In the USA, a deflation of generic prices has been reported, due to an increasing number of competing companies entering the market
(https://www.nytimes.com/2017/08/08/health/generic-drugs-prices-falling.html).

A B

C D

Figure 1. Features of the diagnosis of multiple myeloma. (A) Time from onset of symptoms to the diagnosis of MM: patients (%) diagnosed within <3, 3-6, 6-11 or
>12 months in the retrospective versus prospective analysis. (B) First suspicion of MM: frequency (%) of patients whose MM was first suspected by different types
of physicians (n=176 patients; prospective cohort). (C-D) Prospective analysis: patients‘ satisfaction (n=176).

756 haematologica | 2018; 103(5)


er understanding and statistical evaluation: the relation-
ship between the occurrence of skeletal-related events
and progression-free survival-defining events needs to be
defined. Furthermore, an assessment of cumulative inci-
dence rates of skeletal-related events with death as a
competing event will be helpful, as the slight overall sur-
vival disadvantage in the zoledronic acid arm might have
led to fewer skeletal-related events. Nevertheless, with
full publication of the results16 and with EMA approval,
denosumab will be used in Europe in MM patients.1–3
Given these insights and major advances in the under-
standing of the disease, early MM diagnosis, especially of
symptomatic patients, has been advocated. However,
since MM is an insidiously developing malignancy and
may appear with non-specific symptoms, e.g. bone pain,
the diagnosis and therapeutic decisions can be complicat-
ed. A German study group (DSMM) and EMN project
addressed this aspect with the aim of optimizing the
prompt diagnosis and further improving the quality of
MM care.17 An initial retrospective analysis of 101 MM
patients was followed by a prospective study of 176
patients using a structured MM-specific questionnaire.
The median time from the patients' first symptoms to the
final MM diagnosis was 4 months (range, 0.5-120) in the
retrospectively studied cohort and very similar to the 6
months (range, 0.5-60) in the prospective cohort. Of inter-
est, the time from onset of symptoms to diagnosis of MM
was ≥12 months in 20% of the patients in the retrospec-
tive analysis and 35% in the prospective study (Figure
1A). The frequencies of MM-related bone fractures, renal
complications and infections occurring before the diagno-
sis of MM was made were 41%, 35% and 16%, respec-
tively. Moreover, 43% had one, 20% had two and 3%
had three of these complications. The most frequent
symptom was bone pain, which occurred in 73% of MM
patients before the final MM diagnosis was made. In 6%
of patients, MM was first suspected by orthopedists,
whereas the clinical suspicion was raised by nephrolo-
gists in 16% of cases, even though renal impairment was
less frequent (Figure 1B). Of interest, 61% of patients
were completely or fairly satisfied with the diagnostic
process, whereas 39% were less satisfied (Figure 1C).
Fifty-eight percent of the patients believed that their dis-
ease could have been diagnosed more expeditiously
(Figure 1D). Patients, who criticized the slow diagnostic
process had a much longer median time interval from
symptom onset to their final MM diagnosis compared to
those who were less critical (9 versus 3 months, respec-
tively). These results demonstrate that there is still con-
siderable latency in the diagnosis of MM. However, even
with early diagnosis and treatment with novel agents,
skeletal-related events continue to occur, in part due to
MM responses ("melting-down MM") and relapses,
reminding us that progress in MM involves understand-
ing how best to avoid skeletal-related events before the
diagnosis of the disease is made and with antimyeloma
treatment, because this substantially influences patients'
coping and their approval of our MM care.2,3,17 The notion
that treatment based on novel agents promotes bone-
healing - apart from osteoprotective supportive agents
such as bisphosphonates and denosumab - has recently
led to the demanding discussion18,19 of whether bone-
haematologica | 2018; 103(5)
Perspectives
seeking agents are currently needed. Since skeletal-related
events continue to occur in the first months of treatment
and with relapse (despite the use of novel agents and
osteoprotection1–3), effective prevention and reduction of
destructive skeletal-related events remain fundamental.1,20
Recent data from the national registry, Hospital Episode
Statistics determined fracture rates and the effect on over-
all survival in MM patients between 2001 and 2015:
expectedly, fracture rates were 18 times higher with MM
in the first year after admission than in the general popu-
lation, and remained elevated for up to 10 years. In line
with the data on early diagnosis in MM,2,3,17 the increased
fracture risk preceded the first admission with MM and
conversely the incidence of MM increased after admis-
sion with one or more fractures. Fractures were associat-
ed with poorer outcome (HR for overall survival: 1.2),
indicating the need for regular use of bone supportive
drugs despite novel agent-based treatment.21 In addition,
cost analyses in 1028 MM patients (596 with ≥1 skeletal-
related events and 432 without skeletal-related events)
demonstrated that a higher frequency of skeletal-related
events was associated with greater utilization of health-
care resources, suggesting that bone supportive drugs
need to be used diligently to avoid higher healthcare costs
due to skeletal complications and patients’ discontent.22
Since bisphosphonates in symptomatic MM have been
suggested, but beyond 2 years and with stable MM are
left to the discretion of the treating physician, a random-
ized trial assessed 170 untreated, symptomatic patients
using zolendronic acid for 4 versus 2 years.23 All patients
were treated with the same induction therapy and stem-
cell transplantation. The group treated for 4 years had
substantially fewer skeletal-related events than the group
treated for 2 years (21 versus 43%, respectively; P<0.001).
Actuarial curves at 5 years showed that progression-free
survival was 75% (95% CI: 64%-82%) and overall sur-
vival 68% (95% CI, 60%-76%) in the group treated for 4
years; these rates were not significantly different from
those of the control group treated for 2 years with zole-
dronic acid (P=0.67); but this trial was underpowered to
show differences in survival. The trial did, however, con-
firm that the continued use of zoledronic acid was useful
to reduce skeletal-related events and to preserve a better
quality of life.23
With bisphosphonates and denosumab being potent
options in MM, Goldstein, in this issue of the Journal,
comments on both costs and the fact that novel patent-
protected drugs will induce greater expenditure than
generically available alternatives.24 While there is an
unequivocal need to thoroughly evaluate and measure
"real" advances with new drugs, shortcomings of this
commentary are the "generalization" regarding patent
versus generic medications, the understatement of pro-
gression-free survival differences, convenience of subcu-
taneous versus i.v. medication, and the decreased renal
impairment and safer use of denosumab in patients with
renal impairment. Moreover, Goldstein’s conclusions
only apply to the health system in the USA, whereas
reimbursement of medication providers and financial
incentives to physicians to prescribe more expensive
drugs are different in Europe (Table 1).24 The rising cost of
patented cancer medicines in the USA is a known phe-
757
 Perspectives
nomenon: from 2000 until now, there have been 5- to 10-
fold increases in the cost of new drugs.25 Price differences
in Europe can, therefore, be substantially different from
those in the USA (Table 1). In Germany, a benefit assess-
ment of pharmaceuticals in accordance with the Act on
the Reform of the Market for Medicinal Products
(AMNOG) is limiting the cost of new pharmaceutical
products. In the USA, a deflation of generic prices has
been reported, due to an increasing number of competing
companies entering the market. As clinicians and
researchers we know that medical advances are needed,
and developmental costs for new drugs have steadily
increased with regulatory requirements. Goldstein
reminds us of the financial burden associated with new
cancer agents, which we offer to our patients with the
aim to 'never harm but always aid': nevertheless, the judge-
ment regarding bisphosphonate generics versus denosum-
ab is biased and the comparison is cumbersome. His con-
clusions that generic bisphosphonates have a novel coun-
terpart and that the financial burden with denosumab is
higher are, however, worth noting.24 Thus, the good news
prevails that treatment options for the prevention of bone
complications have increased with the introduction of
denosumab, providing a new choice for patients and
physicians. Once initiated, bisphosphonates or denosum-
ab should be continued for at least 2 years, after which a
suspension of bisphosphonate treatment may be consid-
ered in very responsive patients, although this may be
associated with skeletal risks, especially in those with
prior skeletal complications. For patients in whom bis-
phosphonates were stopped after 2 years, the drug should
be resumed on a monthly basis if the MM recurs and/or
new skeletal-related events occur, independently of the
use of novel agent-based therapies.1–3
Acknowledgment
We are deeply indebted to esteemed experts of the Deutsche
Studiengruppe Multiples Myelom, German Multiple Myeloma
Study Group, European Myeloma Network Group and
International Myeloma Working Group for their valuable dis-
cussion.
References
1. Terpos E, Kleber M, Engelhardt M, et al. European Myeloma
Network Guidelines for the management of multiple myeloma-relat-
ed complications. Haematologica. 2015;100(10):1254-1266.
2. Terpos E, Christoulas D, Gavriatopoulou M. Biology and treatment
of myeloma related bone disease. Metabolism. 2018;80:80-90.
3. Yee AJ, Raje NS. Denosumab for the treatment of bone disease in
solid tumors and multiple myeloma. Future Oncol. 2018;14(3):195-
203.
4. Hillengass J, Fechtner K, Weber M-A, et al. Prognostic significance of
focal lesions in whole-body magnetic resonance imaging in patients
with asymptomatic multiple myeloma. J Clin Oncol.
2010;28(9):1606-1610.
5. Rajkumar SV, Dimopoulos MA, Palumbo A, et al. International
Myeloma Working Group updated criteria for the diagnosis of mul-
tiple myeloma. Lancet Oncol. 2014;15(12):e538-548.
6. Caers J, Fernández de Larrea C, Leleu X, et al. The changing land-
758
scape of smoldering multiple myeloma: a European perspective.
Oncologist. 2016;21(3):333-342.
7. Berenson JR, Lichtenstein A, Porter L, et al. Efficacy of pamidronate
in reducing skeletal events in patients with advanced multiple
myeloma. Myeloma Aredia Study Group. N Engl J Med.
1996;334(8):488-493.
8. Berenson JR, Rosen LS, Howell A, et al. Zoledronic acid reduces
skeletal-related events in patients with osteolytic metastases.
Cancer. 2001;91(7):1191-1200.
9. Rosen LS, Gordon D, Kaminski M, et al. Zoledronic acid versus
pamidronate in the treatment of skeletal metastases in patients with
breast cancer or osteolytic lesions of multiple myeloma: a phase III,
double-blind, comparative trial. Cancer J. 2001;7(5): 377-387.
10. Rosen LS, Gordon D, Kaminski M, et al. Long-term efficacy and safe-
ty of zoledronic acid compared with pamidronate disodium in the
treatment of skeletal complications in patients with advanced multi-
ple myeloma or breast carcinoma: a randomized, double-blind, mul-
ticenter, comparative trial. Cancer. 2003;98(8):1735-1744.
11. Gimsing P, Carlson K, Turesson I, et al. Effect of pamidronate 30 mg
versus 90 mg on physical function in patients with newly diagnosed
multiple myeloma (Nordic Myeloma Study Group): a double-blind,
randomised controlled trial. Lancet Oncol. 2010;11(10):973-982.
12. Morgan GJ, Davies FE, Gregory WM, et al. First-line treatment with
zoledronic acid as compared with clodronic acid in multiple myelo-
ma (MRC Myeloma IX): a randomised controlled trial. Lancet.
2010;376(9757):1989-1999.
13. Morgan GJ, Child JA, Gregory WM, et al. Effects of zoledronic acid
versus clodronic acid on skeletal morbidity in patients with newly
diagnosed multiple myeloma (MRC Myeloma IX): secondary out-
comes from a randomised controlled trial. Lancet Oncol.
2011;12(8):743-752.
14. Morgan GJ, Davies FE, Gregory WM, et al. Effects of induction and
maintenance plus long-term bisphosphonates on bone disease in
patients with multiple myeloma: the Medical Research Council
Myeloma IX Trial. Blood. 2012;119(23):5374-5383.
15. Mhaskar R, Redzepovic J, Wheatley K, et al. Bisphosphonates in
multiple myeloma: a network meta-analysis. Cochrane Database
Syst Rev. 2012;(5):CD003188.
16. Raje N, Terpos E, Willenbacher W, et al. An international, ran-
domised, double-blind study of denosumab compared to zoledronic
acid in bone disease treatment of newly diagnosed multiple myelo-
ma. Lancet Hematol. 2018;19(3):370-381.
17. Graziani G, Herget G, Ihorst G, et al. Time from first symptom onset
to the final diagnosis of multiple myeloma - possible risks and future
solutions: large retrospective and conformatory prospective
“Deutsche Studiengruppe Multiples Myelom” (DSMM) analysis.
Blood. 2017;130(Suppl 1):4710.
18. Delforge M, Terpos E, Richardson PG, et al. Fewer bone disease
events, improvement in bone remodeling, and evidence of bone
healing with bortezomib plus melphalan-prednisone vs. melphalan-
prednisone in the phase III VISTA trial in multiple myeloma. Eur J
Haematol. 2011;86(5):372-384.
19. Mohty M, Malard F, Mohty B, Savani B, Moreau P, Terpos E. The
effects of bortezomib on bone disease in patients with multiple
myeloma. Cancer. 2014;120(5):618-623.
20. Engelhardt M, Terpos E, Kleber M, et al. European Myeloma
Network recommendations on the evaluation and treatment of
newly diagnosed patients with multiple myeloma. Haematologica.
2014;99(2):232-242.
21. McIlroy G, Mytton J, Evison F, et al. Increased fracture risk in plasma
cell dyscrasias is associated with poorer overall survival. Br J
Haematol. 2017;179(1):61-65.
22. Nash Smyth E, Conti I, Wooldridge JE, et al. Frequency of skeletal-
related events and associated healthcare resource use and costs in US
patients with multiple myeloma. J Med Econ. 2016;19(5):477-486.
23. Avilès A, Nambo M-J, Huerta-Guzmàn J, Cleto S, Neri N. Prolonged
use of zoledronic acid (4 Years) did not improve outcome in multiple
myeloma patients. Clin Lymphoma Myeloma Leuk. 2017;17(4):207-
210.
24. Goldstein D. Denosumab for bone lesions in multiple myeloma -
what is its value? Haematologica. 2018(5):753-754.
25. Kantarjian H, Steensma D, Rius Sanjuan J, Elshaug A, Light D. High
cancer drug prices in the United States: reasons and proposed solu-
tions. J Oncol Pract. 2014;10(4):e208-211.
haematologica | 2018; 103(5)
 Bone Marrow Failure ARTICLE

Deep sequencing and flow cytometric

characterization of expanded effector memory Ferrata Storti
Foundation
CD8+CD57+ T cells frequently reveals T-cell
receptor Vβ oligoclonality and CDR3 homology
in acquired aplastic anemia
Valentina Giudice,1 Xingmin Feng,1 Zenghua Lin,1,2 Wei Hu,3 Fanmao Zhang,3
Wangmin Qiao,3 Maria del Pilar Fernandez Ibanez,1 Olga Rios,1
and Neal S. Young1

1
Hematology Branch, National Heart, Lung, and Blood Institute (NHLBI), NIH, Bethesda,
MD, USA; 2Department of Hematology, Affiliated Hospital of Nantong University, Jiangsu,
Haematologica 2018
China and 3BGI Genomics, BGI-Shenzhen, China
Volume 103(5):759-769

ABSTRACT

O
ligoclonal expansion of CD8+CD28– lymphocytes has been con-
sidered indirect evidence for a pathogenic immune response in
acquired aplastic anemia. A subset of CD8+CD28– cells with
CD57 expression, termed effector memory cells, is expanded in several
immune-mediated diseases and may have a role in immune surveillance.
We hypothesized that effector memory CD8+CD28–CD57+ cells may
drive aberrant oligoclonal expansion in aplastic anemia. We found
CD8+CD57+ cells frequently expanded in the blood of aplastic anemia
patients, with oligoclonal characteristics by flow cytometric Vβ usage
analysis: skewing in 1-5 Vβ families and frequencies of immunodomi-
nant clones ranging from 1.98% to 66.5%. Oligoclonal characteristics
were also observed in total CD8+ cells from aplastic anemia patients
with CD8+CD57+ cell expansion by T-cell receptor deep sequencing, as Correspondence:
well as the presence of 1-3 immunodominant clones. Oligoclonality was
confirmed by T-cell receptor repertoire deep sequencing of enriched fengx2@nhlbi.nih.gov
CD8+CD57+ cells, which also showed decreased diversity compared to
total CD4+ and CD8+ cell pools. From analysis of complementarity-deter-
mining region 3 sequences in the CD8+ cell pool, a total of 29 sequences Received: July 21, 2017.
were shared between patients and controls, but these sequences were
highly expressed in aplastic anemia subjects and also present in their Accepted: December 30, 2017.
immunodominant clones. In summary, expansion of effector memory Pre-published: February 1, 2018.
CD8+ T cells is frequent in aplastic anemia and mirrors Vβ oligoclonal
expansion. Flow cytometric Vβ usage analysis combined with deep
sequencing technologies allows high resolution characterization of the doi:10.3324/haematol.2017.176701
T-cell receptor repertoire, and might represent a useful tool in the diag-
nosis and periodic evaluation of aplastic anemia patients. (Registered at Check the online version for the most updated
clinicaltrials.gov identifiers: 00001620, 01623167, 00001397, 00071045, information on this article, online supplements,
00081523, 00961064) and information on authorship & disclosures:
www.haematologica.org/content/103/5/759

Introduction ©2018 Ferrata Storti Foundation

Acquired aplastic anemia (AA) is a bone marrow (BM) failure syndrome charac-
terized by peripheral blood (PB) pancytopenia and BM hypocellularity due to
hematopoietic stem and progenitor cell destruction.1-7 There is indirect evidence to
support the hypothesis of autologous immune attack: clinical response to
immunosuppressive therapies (IST),1,2,8-10 the predominant role of activated cyto-
toxic T cells (CTLs) in BM growth inhibition,1,10-12 identification of putative
autoantigens,13,14 and oligoclonal expansion of CD8+ lymphocytes.8,14-17
Oligoclonality of T-cell populations has been defined by flow cytometry and
deep sequencing of the T-cell receptor (TCR) Vβ repertoire and by spectratyping
of complementarity-determining region 3 (CDR3) length skewing.18-21 The TCR, an
aβ or γd heterodimer, is responsible for antigen recognition and T-cell activation.22-
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.

Haematologica | 2018; 103(5) 759

 V. Giudice et al.
26
Each chain is the result of a complex gene locus
rearrangement, known as VDJ recombination.23 On a
rearranged VDJ segment, a terminal deoxynucleotidyl
transferase enzyme increases TCR variability through
insertions/deletions within a hypervariable region, the
CDR3, generating a unique potential antigen-specific
sequence.23 Early in infection, CD28+ CTLs with different
antigen affinity are selected and expanded (polyclonal
phase).27 In late stages, high antigen-affinity CD28– T cells
are in resting state as memory T cells.28-29 The expression
of CD57 on CD8+CD28– T cells identifies a subset of
memory T cells termed effector memory because of their
high antigen-affinity and ability to be rapidly activated
after antigen stimulation, as “tissue-guards”.29,30 Direct evi-
dence of their high antigen-affinity is decreased diversity
in the CD8+CD57+ TCR repertoire (oligoclonality) due to
the presence of only a few selected clones of memory
cells.27
In this work, we investigated the frequency and oligo-
clonal expansion of effector CD28–CD57+ memory cells in
CD4+ and CD8+ T cells and the TCR Vβ repertoire in AA
patients by flow cytometry and deep sequencing tech-
nologies, to provide additional evidence for the immune
hypothesis in AA pathophysiology.
Methods
Human samples
Heparinized whole PB was collected from patients and healthy
subjects after informed consent, in accordance with the
Declaration of Helsinki31 and protocols approved by the National
Heart, Lung, and Blood Institute Institutional Review Board
(National Institutes of Health, Bethesda, MD, USA) (see Online
Supplementary Table S1 for clinical characteristics). HLA haplotypes
are reported in Online Supplementary Table S2. All patients received
a diagnosis of severe AA (SAA) according to the International
Study of Aplastic Anemia and Agranulocytosis32 and to the criteria
of Camitta.33 At the time of blood sampling, none of the patients
had received therapy. Peripheral blood mononuclear cells (PBMCs)
were isolated by Ficoll-Paque density gradient centrifugation (MP
Biomedicals, LLC, Santa Ana, CA, USA) according to the manu-
facturer’s instructions.
Flow cytometry
A minimum of 4x106 PBMCs from each subject were stained for
TCR Vβ repertoire analysis (Online Supplementary Figure S1). The
manufacturer’s instructions for the IOTest Beta Mark (Beckman
Coulter, Miami, FL, USA) were optimized to avoid incorrect com-
pensation due to the existence of FITC/PE double positive anti-
bodies in the kit, as described in Online Supplementary Methods.
Expansion of CD8+CD57+ cells was defined using as a threshold
the mean frequency calculated in healthy subjects. Vβ skewing in
SAA patients was described when the frequency of one Vβ family
was higher than the mean+3Standard Deviation (SD) in healthy
subjects.16 The term “immunodominant clone” was defined as the
Vβ+ population by flow cytometry or the DNA sequence by deep
sequencing present in the highest abundance.17
TCR repertoire deep sequencing
For VDJ combination and CDR3 sequence profiling,34 DNA was
isolated from FACS- or beads- (Miltenyi Biotec Inc., San Diego,
CA, USA) sorted CD4+ and CD8+ T cells from 12 SAA patients
with CD8+CD57+ cell expansion and 9 healthy subjects
(mean: 3.2 μg of DNA; range: 0.2-27.4 μg) (Online Supplementary
760
Table S2). DNA was also isolated from beads-sorted CD8+CD57+
cells from 2 of the 12 SAA patients with enough cells for further
analysis (mean: 1.6 μg of DNA). TCR repertoire sequencing was
performed with an Illumina HiSeq 2000 sequencer (Illumina Inc.,
San Diego, CA, USA). Detailed information is provided in the
Online Supplementary Methods. Data have been deposited in the
NCBI GEO database (accession n. GSE101660).
Statistical analysis
Data were analyzed using R (RStudio, Boston, MA, USA) and
Prism (v.7.02; GraphPad software Inc., La Jolla, CA, USA). Mann-
Whitney U test, Wilcoxon signed rank sum test, pair and un-
paired t-tests, or χ2 test were used for data with abnormal distribu-
tions. Bonferroni and Dunn’s corrections were used for multiple
comparisons. P≤0.05 was considered statistically significant, after
adjustment with Bonferroni and false discovery rate (FDR).35
Linear regression was performed for correlations. Log-rank
(Mantel-Cox) test was used for progression-free survival data
analysis. Simpson's diversity index was calculated according to
the following formula:
in which ni represents the clone size as the number of copies of
each clonotype (i) or the total number of CDR3 sequences belong-
ing to each i clonotype, and n is the total number of different
clonotypes in the sample or the total number of sequences for
each sample.
Results
Effector memory CD8+CD57+ T cells frequently show
oligoclonal expansion of TCR Vβ repertoire by flow
cytometry
Immunophenotyping and flow-cytometry Vβ usage
were performed in 24 SAA patients. Clinical characteris-
tics are reported in Online Supplementary Table S1. A group
of 34 healthy subjects was studied in order to define nor-
mal ranges of T-cell populations and Vβ family expression.
SAA patients showed higher frequencies of CD8+CD57+
cells (25.6±17.3% vs. 13.3±12.6% in healthy individuals;
P=0.003), and decreased frequency of CD8+CD28+ cells
(56.8±25.7% vs. 68.8±19.1%; P=0.046). A negative corre-
lation between CD57 and CD28 expression was also pres-
ent (r2=0.601, P<0.0001). No differences were found for
CD28+ and CD57+ cells within the CD4+ subset (P=0.974
and P=0.250, respectively) (Figure 1A).
By Vβ usage, polyclonal expansion was observed in
total CD4+, CD4+CD28+, CD4+CD57+ and CD8+CD28+
cells in both healthy subjects and SAA patients (Figure 1B
and Online Supplementary Figure S2). Oligoclonal expan-
sion of CD8+CD57+ cells was present in 92% of SAA
patients with 1-3 immunodominant clones, while in total
CD8+ cells oligoclonality was reported only in 33% of
cases (Online Supplementary Figure S3). Patients did not
show expansion of a shared Vβ family, as each subject car-
ried a different TCR Vβ rearrangement in effector memo-
ry CD8+ T cells (Figure 2A). None of the 7 patients with-
out CD8+CD57+ cell expansion showed Vβ skewing in any
subgroup, with mean frequency of the immunodominant
clone of 3.8% (range: 0.21-6.01%). Conversely, all 17
patients with effector memory CD8+ cell expansion
haematologica | 2018; 103(5)
 TCR repertoire of effector memory T cells in AA

showed Vβ skewing in 1-5 Vβ subgroups, and frequencies
of the immunodominant clones ranged from 2.1% to
66.5% (mean: 9.9%).
Progression-free survival (PFS) analysis was performed
on all SAA patients, divided according to pre-treatment
frequencies of effector memory CD8+ T cells (cut-off value
13.3%) (Figure 2B). No patients with low CD8+CD57+ cell
frequency (n=7) experienced relapse (median survival not
reached; median follow-up time: 13.6 months), while 7 of
the 17 SAA subjects with expanded effector memory T
cells relapsed or died (median survival 13.2 months; medi-
an follow up: 10.4 months). However, statistical signifi-
cance between the two curves was not reached (P=0.089).
Vβ usage was analyzed in CD4+ and CD8+ T-cell subsets

**

Figure 1. Immunophenotyping and flow cytometry analysis of Vβ usage in severe aplastic anemia (SAA) patients and healthy subjects. (A) Percentages of CD28+
and CD57+ cells were calculated in both CD4+ and CD8+ compartments for healthy controls and SAA patients. Data are shown as mean±Standard Deviation (SD).
Unpaired t-test was performed. *P<0.05; **P<0.01. (B) Vβ usage was studied in T-cell compartments (by row), and percentages of each Vβ family were reported as
total CD4+ or CD8+ cell percentage. For Vβ usage in healthy subjects, data are shown as mean+SD, combining the results from all 34 healthy donors. For SAA
patients, 2 representative cases are shown.

haematologica | 2018; 103(5) 761

 V. Giudice et al.

in serial samples available for Patients 4, 22, and 34 in
order to understand the correlation of Vβ usage with clin-
ical course (Figure 2C and Online Supplementary Figure
S4A). For Patient 22, at baseline, Vβ was the immunodom-
inant clone and mostly enriched in CD8+CD57+ cell popu-
lation, rather than in CD4+CD28+, CD4+CD57+, and
CD8+CD28+ cells. After ten days of IST, the size of the
CD8+CD57+ clone was greatly reduced. The clone was
detected again at six months (3 months before clinical
relapse) and further increased at relapse (Figure 2C), sug-
gesting association of Vβ expansion with clinical status.
Patients 4 and 34 were non-responders at three months,
and non- and minimal partial-responders at the 6-month
time point. However, no significant changes in immun-
odominant clone size were observed (Online
Supplementary Figure S4A). Vβ usage was also investigated
at baseline in the BM of these 2 patients (Online
Supplementary Figure S4B), and high concordance with Vβ
usage of PB CD8+CD57+ cells was described.
Increased expansion of effector memory CD8+ T cells
with age has been reported;36 therefore, we used a pool of
age-matched healthy controls to assess the effect of age.
There was no correlation between the size of the immun-
odominant clone in CD8+CD57+ cells and age in healthy
subjects (r2=0.0003, P=0.919) or in SAA patients (r2=0.140,
P=0.079) (Online Supplementary Figure S5). However, a cor-
relation was found between CD57 expression and age in
SAA patients (r2=0.552, P<0.0001).
We then assessed the effect of transfusion history on
oligoclonal expansion of effector memory T cells, as trans-

B C

Figure 2. Vβ usage at diagnosis and during treatment. (A) Percentages of Vβ family in CD8+CD57+ cells were calculated on total CD8+ cells, and Vβ skewing in severe
aplastic anemia (SAA) patients was defined using the mean+3Standard Deviation (SD) of a given Vβ group in healthy donors. Relative expansion of each Vβ subgroup
is shown in the bar graph. Patients were divided based on the absence or presence of expanded CD8+CD57+ cells, using the mean in healthy donors (13.3%). Skewing
of one Vβ family is reported as an orange bar. (B) Progression-free survival rate of SAA patients with CD8+CD57+ cells ≤13.3% (n=7) or >13.3% (n=17) prior to treat-
ment. Log-rank (Mantel-Cox) test was performed. (C) Vβ usage was performed in Patient 22 at diagnosis, at 10 days of treatment, and at 6 and 9 months (relapse).
Perturbations during treatment and relapse are reported as percentage of positive CD8+CD57+ and Vβ2+ cells (left), or absolute lymphocyte and Vβ2 clone count
(right).

762 haematologica | 2018; 103(5)

 TCR repertoire of effector memory T cells in AA

fusions expose T cells to multiple foreign epitopes.37 A
group of 5 pure red cell aplasia (PRCA) patients, 10 sickle
cell disease (SCD) subjects, and 8 myelodysplastic (MDS)
patients who had been heavily transfused before sampling
were studied for CD8+CD57+ cell expansion and Vβ usage
(Online Supplementary Table S1). No variations were
observed in CD8+CD57+ cell frequencies when PRCA,
MDS, and SCD patients were compared to healthy sub-
jects (22.8±25.1% vs. 21.6±9.7% vs. 30.6±28.3% vs.
13.3±12.6%, respectively; P=0.216) (Online Supplementary
Figure S6A). Vβ skewing was described in 4 PRCA patients
in 1-8 subgroups, in all 8 MDS subjects in 1-4 Vβ families,
and in 7 SCD patients in 1-5 subgroups (Online
Supplementary Figure S6B). Oligoclonal expansion was
described in 91% of cases and subjects without skewing
did not present effector memory T-cell expansion.
Subsequently, we combined all subjects and divided them
according to their transfusion history and the presence of

Figure 3. Characterization of Vβ/Jβ plot, CDR3 size and DJ length profiles in healthy donors by deep sequencing. (A) T-cell receptor β variable (TRBV)/T-cell receptor
β joining (TRBJ) plots showed a “citylike” landscape for total CD4+ and CD8+ cell populations in healthy subjects (HC). (B) The size of the complementarity region 3
(CDR3) and DJ length profiles were also defined, showing similar features in CD4+ and CD8+ cells.

haematologica | 2018; 103(5) 763

 V. Giudice et al.

CD8+CD57+ T-cell expansion. By χ2 test, transfusion histo-
ry did not correlate to CD8+CD57+ expansion (P=0.255).
Total CD8+ cell TCR repertoires are polyclonal in
healthy subjects
Deep sequencing of TCR repertoire was performed in
CD4+ and CD8+ populations sorted from 9 healthy donors.
The average depth of sequencing was 10,790,646±4,050,138
in CD4+ T cells, and 10,961,961±3,879,596 in CD8+ popula-
tions. The mean frequency of the immunodominant clone
was 4.1±4.1% in CD4+ cells and 17.3±16.9% in CD8+ cells.
By plotting the number of reads of each Vβ and Jβ
matching,17 a “citylike” landscape was described for total
CD4+ and CD8+ T-cell populations (Figure 3A and Online
Supplementary Figure S7A), because of the presence of a
more homogenous distribution of TCR rearrangement fre-

Figure 4. The TCR repertoire by deep sequencing analysis from total CD8+ cells in severe aplastic anemia (SAA) patients with CD8+CD57+ cell expansion. In contrast
to healthy CD8+ profiles, most SAA patients (AA) displayed oligoclonal features in a TRBV/TRBJ rearrangement plot (A) and CDR3 size and DJ length profiles (B). CD4+
and CD8+ profiles are shown for each SAA patient. In AA1 and AA6, only CD8+ cells were sufficient for deep sequencing.

764 haematologica | 2018; 103(5)

 TCR repertoire of effector memory T cells in AA

quencies (“midtown”). The normal CDR3 size and DJ
length profiles were defined by comparing distributions
among healthy donors (Figure 3B and Online
Supplementary Figure S7B). In CD4+ and CD8+ cells, profiles
were typically distributed in a Gaussian manner with 10-
12 different size classes of 30-65 nucleotides (nt) sizes at 3
nt intervals, and they completely overlapped. Similarly, DJ
length profiles assumed an asymmetric Gaussian distribu-
tion with 2-5 predominant length classes without nt inter-
vals in CD4+ and CD8+ subsets.
Deep sequencing allows detailed characterization of
oligoclonality in SAA
Deep sequencing of TCR repertoire was performed in
CD4+ and CD8+ cells from 12 SAA patients who had
demonstrated CD8+CD57+ cell expansion by flow cytom-
etry, and also in CD8+CD57+ cells from 2 of these patients.
The average depth of sequencing was
13,861,048±4,992,677 in CD4+ T cells,
13,873,207±5,029,195 in CD8+ T cells, and
21,030,616±1,238,660 in CD8+CD57+ cells. The mean fre-
quency of the immunodominant clone was 3.3±3.4%
(range: 0.2-11.8%) in CD4+, 18.2±14.9% (range: 3.3-
54.1%) in CD8+ cells, and 59±28.9% in CD8+CD57+cells.
By plotting TRBV/TRBJ rearrangements from CD4+ cells,
the “citylike” landscape was found in 11 out of 12 patients
(Figure 4A and Online Supplementary Figure S8A). In the
CD8+ cell pool, the “citylike” landscape was found in 3
patients (AA7, AA8, and AA11), whose clones showed
very low frequencies (4.3%, 3.4%, and 3.3%, respective-
ly). The remaining patients displayed a “skyscraper” land-
scape due to the presence of 1-3 immunodominant clones.

Figure 5. The TCR repertoire by deep sequencing of enriched CD8+CD57+ cells in severe aplastic anemia (SAA) patients. (A) The enrichment of the clone in effector
memory CD8+ T cells, comparing TRBV/TRBJ rearrangement in total CD8+ cells (left) with those in CD8+CD57+ cells (right) from the same patients. (B) CDR3 size and
DJ length profiles from CD8+CD57+ cells also overlapped with those in CD4+ and CD8+ profiles.

haematologica | 2018; 103(5) 765

 V. Giudice et al.

CDR3 size and DJ length profiles were defined in total
CD4+ and CD8+ cells and compared within each patient
(Figure 4B and Online Supplementary Figure S8B). In CD4+
cells, all patients displayed CDR3 profiles with normal
Gaussian distribution, as described in healthy subjects. In
CD8+ cells, Patients AA7, AA8, and AA10 to AA12 had
CDR3 size profiles with Gaussian distributions. In the
remaining patients, CDR3 profiles showed different
shapes with 8-13 different classes and 1 or 2 predomi-
nant peaks of various nt sizes (36-57). For DJ length pro-
files in SAA patients, the asymmetric Gaussian distribu-
tion was described in all CD4+ cells, and in 6 (AA5-AA8
and AA11-AA12) CD8+ populations. Similarly,
TRBV/TRBJ rearrangement plots from the CD8+CD57+

Figure 6. Homology assessment. (A) CDR3 sequence pools were analyzed among patients (AA) and healthy subjects (HC) for the presence of homology. Shared and
immunodominant sequences were reported as a heatmap based on their relative expression: in the same row, from lowest (gray; <0.01%) to highest (red; >5%) val-
ues. (B) Structural analysis was performed comparing the sequences for common pattern, using both alignments at the N-terminal of Vβ gene (left) or at the C-ter-
minal of Jβ gene (right).

766 haematologica | 2018; 103(5)


cell pool of AA3 and AA4 showed the same “skyscraper”
landscape, but more enriched (Figure 5A), as well as
CDR3 size and DJ length profiles that overlapped those
in total CD8+ cells, although these were again more
enriched (Figure 5B).
TCR repertoire diversity and shared CDR3 sequences
in SAA patients and healthy subjects
Simpson’s index of diversity was calculated for each
healthy and each SAA subject. This index is used in ecol-
ogy for in-depth assessment of the degree of diversity of a
system, related to the richness (or number of species pres-
ent) and evenness (or relative abundance of each).38 Thus,
this index assesses the probability that 2 randomly select-
ed individuals from a system belong to the same species.
For a TCR repertoire, the index measures the probability
that 2 CDR3 sequences randomly selected from CD4+ or
CD8+ pools of one subject could be identical.17,38 A value
close to 0 means infinite diversity, and a value close to 1
no diversity.36 In our cohort, Simpson’s indexes were sim-
ilar in healthy controls and SAA patients for CD4+
(0.990±0.005 vs. 0.991±0.005, respectively; P=0.537) and
CD8+ cells (0.983±0.013 vs. 0.927±0.084, respectively;
P=0.060). By paired t-test, Simpson’s indexes in SAA sub-
jects were significantly different between CD4+ and CD8+
cells (P=0.047) (Online Supplementary Figure S9A). When
compared to those indices in CD8+CD57+ cells, decreased
diversity in the CD8+ effector memory compartment was
described (total CD4+ cells vs. CD8+CD57+ cells, P<0.0001;
total CD8+ cells vs. CD8+CD57+ cells, P=0.0003) (Online
Supplementary Figure S9B).
In order to test the hypothesis that an autoreactive clone
is triggered by autoantigens, CDR3 amino acid sequences
from healthy subjects and patients were screened for
homology and then compared to public and private TCR
repertoires reported in literature,39 as described in the
Online Supplementary Methods. From analysis of sequences
in the CD8+ cell pool, a total of 29 CDR3 sequences were
shared between patients and controls, but these sequences
were highly expressed in SAA patients and also present in
their immunodominant clones (Figure 6A). When we
searched for these common sequences in the whole CDR3
sequence repertoire from CD8+CD57+ cells in AA subjects,
CD8+CD57+ cell pools from AA3 and AA4 showed enrich-
ment in frequencies of the immunodominant clones (from
53.61% to 79.12% for AA3 patient and from 23.45% to
38.14% for AA4 patient). Structural analysis of shared and
immunodominant sequences did not show a pattern of
common charged residue among patients (Figure 6B and
Online Supplementary Methods). Immunodominant and
shared CDR3 sequences were compared with CDR3 β
repertoires reported in literature for infectious, autoim-
mune and malignant diseases.39 For confirmation, the
VDJdb database, a database of known antigen specific
sequences (https://vdjdb.cdr3.net), was also used. No match-
es were found for infectious diseases, while 2 sequences
present in AA9 were reported previously in AA and parox-
ysmal nocturnal hemoglobinuria (PNH).8,40 Lastly, we
sought to perform homology assessment for reported
PNH-related clonotypes on the entire TCR repertoire
without frequency restriction, as small PNH clones could
be present in healthy individuals.40 Two of the reported 12
CDR3 sequences (CATSRTGGETQYF and CATSRV-
VAGETQYF) were found in our TCR repertoires with a
haematologica | 2018; 103(5)
TCR repertoire of effector memory T cells in AA
mean frequency of 0.1±0.2% and 4.7±9.4% in SAA
patients, and 0.0003±0.0001% and 0.01±0.002% in
healthy subjects. No matches were found in the enriched
CD8+CD57+ T-cell pool (Online Supplementary Table S3).
Discussion
The character of oligoclonal expansion of CD8+CD28–
lymphocytes in AA, described by Risitano et al.,8,41 strong-
ly suggests an antigen driven mechanism of T-cell activa-
tion, ultimately leading to destruction of hematopoietic
stem and progenitor cells. In this study, we focused on a
subset of CD8+CD28–CD57+ T lymphocytes, termed effec-
tor memory cells because of their high antigen-affinity
and their ability to undergo activation after antigen stimu-
lation.27,36 Others have reported the expansion of effector
memory CD8+ T cells in the tumor microenvironment and
in PB from patients with solid tumors, hematologic malig-
nancies, chronic infections and autoimmune disorders.36,41-
44
In cancers, effector memory T cells appear to have an
important role in immune surveillance; for example, their
increase after interferon (IFN)a treatment correlates to
better prognosis in melanoma patients,43 and expanded
CD8+CD57+ T cells reach normal levels after removal of
head and neck cancer.36 Higher frequencies of CD8+CD57+
cells have also been described in SAA and MDS patients,
which decrease in responders after anti-thymocyte globu-
lin treatment.44 The expansion of effector memory cells
could also occur in older healthy subjects as a result of life-
long exposure to common pathogens, but it is related to
reduced overall immune responsiveness to novel
antigens.45,46 Consistent with previous studies, the SAA
patients in our cohort had higher frequencies of
CD8+CD57+ cells at diagnosis (25.6% in SAA patients vs.
13.3% in healthy subjects). Moreover, patients with CD8+
effector memory T-cell expansion at diagnosis experi-
enced shorter PFS. No age-effects were seen for clone size
in either patients or healthy subjects.
There is indirect evidence of immune pathophysiology
of AA, including oligoclonal expansion of effector CD8+ T
lymphocytes.8,41 Clonality of T-cell subsets can be studied
by flow cytometry analysis of Vβ usage, CDR3 size spec-
tratyping or deep sequencing of VDJ combinations, and
CDR3 nucleotide and amino acid sequences.17-20,47 In SAA
patients, expansion of at least one Vβ family in both CD4+
and CD8+ effector CD28dim cells by flow cytometry with
polyclonal features in CD4+ and oligoclonal characteristics
in CD8+ cells has been described.1,8,15,41,44 The CDR3 size
skewing by spectratyping in the CD8+ population but not
in CD4+ cells confirmed clonality in CD8+ cells.8,21,41 In our
current work, we demonstrate by flow cytometry and
deep sequencing that oligoclonal expansion occurs mainly
in effector memory CD8+ cell compartment, as the
immunodominant clones were highly enriched in
CD8+CD57+ cells and had decreased diversity on deep
sequencing. Effector memory T cells are a circulating T-
cell population that can migrate to the BM under different
types of stimulation.30 Vβ usage of peripheral CD8+CD57+
T cells may mirror that in the BM, as suggested by the
high concordance between PB and BM in our AA patients.
In our cohort, 75% of SAA patients with effector memory
cell expansion also showed oligoclonal features by deep
sequencing of the total CD8+ cell compartment:
TRBV/TRBJ rearrangement plots with a “skyscraper” pro-
767
 V. Giudice et al.

file due to the presence of 1-3 immunodominant clones,
and predominant classes in CDR3 size and DJ length pro-
files. These findings were confirmed by TCR repertoire
sequencing of CD8+CD57+ enriched cells from SAA
patients. Similar deep sequencing features have been
described in whole blood in T-cell large granular lympho-
cyte leukemia (T-LGLL).17 T-LGLL, a chronic lymphopro-
liferation of TCRaβ+CD3+CD5dimCD8+CD57+CD16+ cells
with monoclonal TCRγ-chain rearrangement and preva-
lence in the elderly, is frequently associated with autoim-
mune diseases.16,17 Thus, similarities between our SAA
cohort and T-LGLL patients suggest a common patho-
physiology of expanded autoreactive T lymphocytes.
Characterization of long-term Vβ usage has already been
proposed as a biomarker of disease progression in T-LGLL
and AA,8,16 given that immunodominant clones can
remerge during relapse.16 However, high heterogeneity in
the TCR repertoire during immunosuppressive therapies
has been reported.8,16 In our study, immunodominant
clones were longitudinally investigated in 3 AA patients.
In Patient 4, expanded clones slightly decreased during
treatment but did not disappear, as no hematologic
improvements were observed. In Patient 34, clone 17
remained stable during the course of the disease, while
clone 13.6 increased at three months and slightly
decreased at six, concomitantly with a minimal partial
response. In Patient 22, clone 2 completely disappeared at
six months of treatment and the patient achieved hemato-
logic remission; however, at the time of relapse the clone
increased again. Our data suggest that Vβ typing of the
CD8+ CD57+ T-cell population by flow cytometry might
be a useful biomarker to monitor clonal kinetics during the
course of AA, but a larger cohort of patients and a more
sensitive technique, such as deep sequencing, are needed
to validate the clinical usefulness.
Chronic antigen exposure is required to trigger T-cell acti-
vation, as in persistent viral infections or with antigen
spread during autoimmune and malignant diseases.17 For
viral infections, CDR3 homology in public sequence reper-
toires, and also cross-reactivity between viruses, is limited
by the number of possible epitopes.41,48 For autoimmune dis-
eases, including T-LGLL, clonotypes can be private to a spe-
cific disease because of the unlimited number of possible
epitopes.17,41,49 Despite the large diversity of TCR CDR3
sequences, only 29 shared CDR3 sequences were found in
our cohort; these were highly expressed in SAA patients
and enriched in the CD8+ CD57+ T-cell population. By using
sensitive techniques such as deep sequencing, we detected
sequences at very low frequencies. However, the finding
that a group of clonotypes is shared between SAA patients
and healthy subjects suggests the existence of common epi-
topes driving activation of T-cell autologous clones (as also
described by Gargiulo et al. in PNH patients40). As chronic
transfusion could be the source of antigen exposure, we
investigated its effect on effector memory T-cell expansion
in other hematologic diseases. A comparison between
transfused patients and healthy subjects showed no signifi-
cant variations in CD8+ CD57+ cell frequencies (P=0.216).
Oligoclonal expansion of effector memory CD8+ T cells
is frequent in AA and may correlate with prognosis, con-
sistent with a role of effector memory T cells in BM
destruction during active disease. Deep sequencing tech-
nologies allow in-depth characterization of the TCR reper-
toire, and flow cytometric analysis of Vβ usage may be
useful to determine diagnosis and prognosis of SAA
patients, and to monitor their clinical course.
Acknowledgments
The authors would like to thank Sachiko Kajigaya and
Keyvan Keyvanfar (Hematology Branch, NHLBI), Ying Rao
(BGI Genomics), and Swee Lay Thein (Sickle Cell Branch,
NHLBI) for assistance; Kinneret Broder (Hematology Branch,
NHLBI) for assistance in obtaining healthy volunteer samples;
and Barbara Weinstein (Hematology Branch, NHLBI) and Jim
Nichols (Sickle Cell Branch, NHLBI) for obtaining patient sam-
ples.
Funding
This research was supported by the Intramural Research
Program of the NIH, National Heart, Lung, and Blood Institute.

References
1. Young NS, Calado RT, Scheinberg P. Current
concepts in the pathophysiology and treat-
ment of aplastic anemia. Blood.
2006;108(8):2509-2519.
2. Young NS. Current concepts in the patho-
physiology and treatment of aplastic ane-
mia. Hematology Am Soc Hematol Educ
Program. 2013;2013:76-81.
3. Maciejewski JP, Selleri C, Sato T, Anderson
S, Young NS. A severe and consistent deficit
in marrow and circulating hematopoietic
cells (long-term culture-initiating cells) in
acquired aplastic anemia. Blood.
1996;88(6):1983-1991.
4. Maciejewski JP, Kim S, Sloand E, Selleri C,
Young NS. Sustained long-term hematologic
recovery despite a marked quantitative
defect in the stem cell compartment of
patients with aplastic anemia after immuno-
suppressive therapy. Am J Hematol.
2000;65(2):123-131.
5. Risitano AM, Maciejewski JP, Selleri C,
Rotoli B. Function and malfunction of
hematopoietic stem cells in primary bone
marrow failure syndromes. Curr Stem Cell
Res Ther. 2007;2(1):39-52.
6. Maciejewski JP, O’Keefe C, Gondek L, Tiu
R. Immune-mediated bone marrow failure
syndromes of progenitor and stem cells:
molecular analysis of cytotoxic T cell clones.
Folia Histochem Cytobiol. 2007;45(1):5-14.
7. Feng X, Scheinberg P, Wu CO, et al.
Cytokine signature profiles in acquired
aplastic anemia and myelodysplastic syn-
dromes. Haematologica. 2011;96(4):602-
606.
8. Risitano AM, Maciejewski JP, Green S,
Plasilova M, Zeng W, Young NS. In-vivo
dominant immune responses in aplastic
anaemia: molecular tracking of putatively
pathogenetic T-cell clones by TCR beta-
CDR3 sequencing. Lancet. 2004;
364(9431):355-364.
9. Young NS. Hematopoietic cell destruction
by immune mechanisms in aquired aplastic
anemia. Semin Hematol. 2000;37(1):3-14.
10. Sloand EM, Kim S, Maciejewski JP, et al.
Intracellular interferon- in circulating and
marrow T cells detected by flow cytometry
and the response to imunosuppressive ther-
apy in patients with aplastic anemia. Blood.
2002;100:(4)1185-1191.
11. Maciejewski JP, Selleri C, Anderson S,
Young NS. Fas antigen expression on
CD34+ human marrow cells is induced by
interferon-gamma and tumor necrosis fac-
tor-alpha and potentiates cytokine-mediat-
ed hematopoietic suppression in vitro.
Blood. 1995;85:(11)3183-3190.
12. Selleri C, Maciejewski JP, Sato T, Young NS.
Interferon-γ constitutively expressed in the
stromal microenviroment of human mar-
row cultures mediates potent hematopoietic
inhibition. Blood. 1996; 87:(10)4149-4157.
13. Hirano N, Butler MO, von Bergwelt-Baildon
MS, et al. Autoantibodies frequently detect-
ed in patients with aplastic anemia. Blood.
2003;102:(13)4567-4575.
14. Feng X, Chuhjo T, Sugimori C, et al.
Diazepam-binding inhibitor-related protein

768 haematologica | 2018; 103(5)


1: a candidate autoantigen in acquired aplas-
tic anemia patients harboring a minor popu-
latoin of paroxysmal nocturnal hemoglobin-
uria-type cells. Blood. 2004; 104:(8)2425-
2431.
15. Kook H, Risitano AM, Zeng W, et al.
Changes in T-cell receptor VB repertoire in
aplastic anemia: effects of different
immunosuppressive regimens. Blood.
2002;99(10):3668-3675.
16. Clemente MJ, Wlodarski MW, Makishima
H, et al. Clonal drift demonstrates unexpect-
ed dynamics of the T-cell repertoire in T-
large granular lymphocyte leukemia. Blood.
2011;118(16):4384-4393.
17. Clemente MJ, Przychodzen B, Jerez A, et al.
Deep sequencing of the T-cell receptor
repertoire in CD8+ T-large granular lympho-
cyte leukemia identifies signature land-
scapes. Blood. 2013;122(25):4077-4085.
18. Calis JJ, Rosenberg BR. Characterizing
immune repertoires by high throughput
sequencing: strategies and applications.
Trends Immunol. 2014;35(12):581-590.
19. Langerak AW, van Den Beemd R, Wolvers-
Tettero IL, et al. Molecular and flow cyto-
metric analysis of the Vbeta repertoire for
clonality assessment in mature
TCRalphabeta T-cell proliferations. Blood.
2001;98(1):165-173.
20. Wlodarski MW, Gondek LP, Nearman ZP, et
al. Molecular strategies for detection and
quantitation of clonal cytotoxic T-cell
responses in aplastic anemia and myelodys-
plastic syndrome. Blood. 2006;108(8):2632-
2641.
21. Schuster FR, Hubner B, Führer M, et al.
Highly skewed T-cell receptor V-beta chain
repertoire in the bone marrow is associated
with response to immunosuppressive drug
therapy in children with very severe aplastic
anemia. Blood Cancer J. 2011;1(3):e8.
22. Schatz DG, Ji Y. Recombination centres and
the orchestration of V(D)J recombination.
Nat. Rev. Immunol. 2011;11:(4)251-263.
23. Gauss GH, Lieber MR. Mechanistic con-
straints on diversity in human V(D)J recom-
bination. Mol Cell Biol. 1996;16(1):258-269.
24. Teng G, Papavasiliou FN. Immunoglobulin
somatic hypermutation. Annu Rev Genet.
2007;41:107-120.
25. Matthews AJ, Zheng S, DiMenna LJ,
Chaudhuri J. Regulation of immunoglobulin
class-switch recombination: choreography
of noncoding transcription, targeted DNA
deamination, and long-range DNA repair.
haematologica | 2018; 103(5)
TCR repertoire of effector memory T cells in AA
Adv Immunol. 2014;122:1-57.
26. Arstila TP, Casrouge A, Baron V, Even J,
Kanellopoulos J, Kourilsky P. A direct esti-
mate of the human alphabeta T cell receptor
diversity. Science. 1999;286(5441):958-961.
27. Bretschneider I, Clemente MJ, Meisel C, et
al. Discrimination of T-cell subsets and T-
cell receptor repertoire distribution.
Immunol Res. 2014;58(1):20-27.
28. Araki K, Youngblood B, Ahmed R. The role
of mTOR in memory CD8 T-cell differenti-
ation. Immunol Rev. 2010;235(1):234-243.
29. Appay V, van Lier RA, Sallusto F, Roederer
M. Phenotype and function of human T
lymphocyte subsets: consensus and issues.
Cytometry A. 2008;73(11):975-983.
30. Farber DL, Yudanin NA, Restifo NP. Human
memory T cells: generation, compartmen-
talization and homeostasis. Nat Rev
Immunol. 2014;14(1):24-35.
31. World Medical Association. World Medical
Association Declaration of Helsinki: ethical
principles for medical research involving
human subjects. JAMA. 2013;310(20):2191-
2194.
32. International Agranulocytosis and Aplastic
Anaemia Study. Incidence of aplastic ane-
mia: the relevance of diagnostic criteria. By
the International Agranulocytosis and
Aplastic Anemia Study. Blood.
1987;70(6):1718-1721.
33. Camitta BM, Thomas ED, Nathan DG, et al.
Severe aplastic anemia: a prospective study
of the effect of early marrow transplantation
on acute mortality. Blood. 1976;48(1):63-70.
34. Bolotin DA, Shugay M, Mamedov IZ, et al.
MiTCR: software for T-cell receptor
sequencing data analysis. Nat Methods.
2013;10(9):813-814.
35. Benjamini Y, Yekutieli D. The control of the
false discovery rate in multiple testing under
dependency. Ann Stat. 2001; 29:(4)1165-
1188.
36. Strioga M, Pasukoniene V, Characiejus D.
CD8+ CD28- and CD8+ CD57+ T cells and
their role in health and disease.
Immunology. 2011;134(1):17-32.
37. Nickel RS, Horan JT, Fasano RM, et al.
Immunophenotypic parameters and RBC
alloimmunization in children with sickle cell
disease on chronic transfusion. Am J
Hematol. 2015;90(12):1135-1141.
38. Hill MO. Diversity and evenness: a unifying
notation and its consequences. Ecology.
1973;54:(2)427-432.
39. Li H, Ye C, Ji G, Han J. Determinants of pub-
lic T cell responses. Cell Res. 2012;22(1):33-
42.
40. Gargiulo L, Lastraioli S, Cerruti G, et al.
Highly homologous T-cell receptor beta
sequences support a common target for
autoreactive T cells in most patients with
paroxysmal nocturnal hemoglobinuria.
Blood. 2007;109(11):5036-5042.
41. Risitano AM, Kook H, Zeng W, Chen G,
Young NS, Maciejewski JP. Oligoclonal and
polyclonal CD4 and CD8 lymphocytes in
aplastic anemia and paroxysmal nocturnal
hemoglobinuria measured by V beta CDR3
spectratyping and flow cytometry. Blood.
2002;100(1):178-183.
42. Filaci G, Fenoglio D, Fravega M et al. CD8+
CD28- T regulatory lymphocytes inhibiting
T cell proliferative and cytotoxic functions
infiltrate human cancers. J Immunol 2007;
179:(7)4323-4334.
43. Characiejus D, Pasukoniene V,
Jonusauskaite R, et al. Peripheral blood
CD8high CD57+ lymphocyte levels may
predict outcome in melanoma patients treat-
ed with adjuvant interferon-alpha.
Anticancer Res. 2008;28:(2B)1139-1142.
44. Kook H, Zeng W, Guibin C, Kirby M, Young
NS, Maciejewski JP. Increased cytotoxic T
cells with effector phenotype in aplastic ane-
mia and myelodysplasia. Exp Hematol.
2001;29:(11)1270-1277.
45. Boucher N, Dufeu-Duchesne T, Vicaut E,
Farge D, Effros RB, Schächter F. CD28
expression in T cell aging and human
longevity. Exp Gerontol. 1998;33(3):267-
282.
46. Wikby A, Ferguson F, Forsey R, et al. An
immune risk phenotype, cognitive impair-
ment, and survival in very late life: impact of
allostatic load in Swedish octogenarian and
nonagenarian humans. J Gerontol A Biol Sci
Med Sci. 2005;60(5):556-565.
47. Fozza C, Barraqueddu F, Corda G, et al.
Study of the T-cell receptor repertoire by
CDR3 spectratyping. J Immunol Methods.
2017;440:1-11.
48. Zeng W, Maciejewski JP, Chen G, Young NS.
Limited heterogeneity of T cell receptor BV
usage in aplastic anemia. J Clin Invest.
2001;108(5):765-773.
49. Le Gal FA, Ayyoub M, Dutoit V, et al.
Distinct structural TCR repertoires in natu-
rally occurring versus vaccine-induced
CD8+ T-cell responses to the tumor-specific
antigen NY-ESO-1. J Immunother.
2005;28(3):252-257.
769
ARTICLE Rec Cell Biology & its Disorders

Safety and efficacy of plerixafor dose

Ferrata Storti
Foundation
escalation for the mobilization of CD34+
hematopoietic progenitor cells in patients
with sickle cell disease: interim results
Farid Boulad,1,2 Tsiporah Shore,3 Koen van Besien,3 Caterina Minniti,4 Mihaela
Barbu-Stevanovic,5 Sylvie Wiener Fedus,6 Fabiana Perna,2 June Greenberg,7
Danielle Guarneri,7 Vijay Nandi,5 Audrey Mauguen,8 Karina Yazdanbakhsh,5
Haematologica 2018 Michel Sadelain2 and Patricia A. Shi4,5

Department of Pediatrics, BMT Service, Memorial Sloan Kettering Cancer Center, New
Volume 103(5):770-777 1

York; 2Center for Cell Engineering, Memorial Sloan Kettering Cancer Center, New York;
3
Bone Marrow and Hematopoietic Stem Cell Transplant Program, Weill Cornell
Medicine/ New York Presbyterian Hospital, New York; 4Sickle Cell Program, Division of
Hematology, Albert Einstein College of Medicine, Bronx; 5Lindsley F. Kimball Research
Institute, New York Blood Center, NY; 6Laboratory Medicine, Memorial Sloan Kettering
Cancer Center, New York; 7Division of Hematology and Oncology, Weill Cornell Medicine
/New York Presbyterian Hospital, NY and 8Department of Epidemiology and
Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY, USA

ABSTRACT

Correspondence:
pshi@nybc.org
Received: December 22, 2017.
Accepted: January 23, 2018.
Pre-published: February 1, 2018.
doi:10.3324/haematol.2017.187047
Check the online version for the most updated
information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/770
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
G
ene therapy for sickle cell disease is limited by the yield of
hematopoietic progenitor cells that can be harvested for transduc-
tion or gene editing. We therefore performed a phase I dose-esca-
lation study of the hematopoietic progenitor cell mobilizing agent pler-
ixafor to evaluate the efficacy and safety of standard dosing on periph-
eral blood CD34+ cell mobilization. Of 15 patients enrolled to date, only
one was chronically transfused and ten were on hydroxyurea. Of eight
patients who achieved a CD34+ cell concentration >30 cells/μL, six were
on hydroxyurea. There was no clear dose response to increasing plerix-
afor dosage. There was a low rate of serious adverse events; two patients
developed vaso-occlusive crises, at the doses of 80 μg/kg and 240 μg/kg.
Hydroxyurea may have contributed to the limited CD34+ mobilization
by affecting baseline peripheral blood CD34 counts, which correlated
strongly with peak peripheral blood CD34 counts. Plerixafor administra-
tion did not induce significant increases in the fraction of activated neu-
trophils, monocytes, or platelets. However, increased neutrophils posi-
tive for activated β2 integrin and Mac-1 were associated with serious
adverse events. In summary, plerixafor was well tolerated but did not
achieve consistent CD34+ cell mobilization in this cohort of patients,
most of whom were being actively treated with hydroxyurea and only
one was chronically transfused. The study will continue with escalation
of the dose of plerixafor and modification of hydroxyurea administra-
tion. Clinicaltrials.gov identifier: NCT02193191.
Introduction
Autologous gene therapy holds considerable promise for the treatment of
patients with sickle cell disease (SCD).1,2 However, its successful application
requires an adequate number of hematopoietic progenitor cells (HPC) for gene
transfer or gene editing.3 Steady-state bone marrow has been the historical source
of HPC for SCD gene therapy, but its harvest requires general anesthesia and has
been complicated in current gene therapy trials by the need for repeated bone mar-
row harvests and a high rate of adverse events.4 Granulocyte colony-stimulating
factor (G-CSF) is a standard method of mobilizing HPC but its use in SCD patients
has been associated with vaso-occlusive complications and even death.5 The
mechanism of action of G-CSF involves activation of neutrophils,6 and is also asso-
ciated with endothelial cell, platelet, and coagulation system activation,7-11 all of
which may play a crucial role in sickle cell vaso-occlusion.12
In contrast to G-CSF, plerixafor is a bicyclam reversible small molecule inhibitor

770 haematologica | 2018; 103(5)

 Safety and efficacy of plerixafor in SCD patients

Table 1. Patients’ characteristics.

Dose level Subject ID Ethnicity Age Gender Genotype Treatment Clinical complications (in addition to ACS)
yrs regimen
1 African- 33 M SS HU 16 mg/kg Avascular necrosis, retinopathy
2 American 21 F SS HU 26 mg/kg ≥ 3 vaso-occlusive crises per year
3 (1*) 29 M SS; thal(1/4) No HU ≥ 3 vaso-occlusive crises per year
80 μg/kg 4 Hispanic 32 M SS; thal (1/4) No HU Deep venous thrombosis, priapism
5 34 M SS; thal (1/4) HU 28 mg/kg Leg ulcers, retinopathy
6 African- 25 F SS; thal (1/4) HU 16 mg/kg Leg ulcers, retinal artery occlusion,
American retinopathy
7 25 F SS HU 25 mg/kg Cerebral aneurysms (2-3 mm)
160 μg/kg 8 (1*) African- 37 M SS HU 27 mg/kg Avascular necrosis, retinopathy
9 American 32 F SS Chronic transfusion Avascular necrosis, iron overload,
– No HU leg ulcers
10 46 M SS; thal (1/4) No HU Leg ulcers, priapism
12 38 M SS HU 17 mg/kg Avascular necrosis, leg ulcers, priapism
240 μg/kg 13 African- 23 F SS HU 27 mg/kg Avascular necrosis, retinopathy
14 American 27 M SS HU 23 mg/kg Avascular necrosis, priapism
3 (2*) 31 M SS- thal (1/4) No HU ≥ 3 vaso-occlusive crises per year
8 (2*) 38 M SS HU 27 mg/kg Avascular necrosis, retinopathy
*Indicates first and second enrollments for indicated subject. ACS: acute chest syndrome,; HU: hydroxyurea.

of the chemokine receptor CXCR4 and prevents binding
of its ligand CXCL12 or stromal cell derived factor-1a to
induce HPC mobilization.13 We hypothesized that plerix-
afor’s mechanism of action would lead to less marked
increases in white blood cell (WBC) counts and therefore
less cell and coagulation system activation in SCD and
supported this with data from a pre-clinical study involv-
ing a sickle cell mouse model.14 Nevertheless, the safety of
plerixafor in SCD patients remains a matter of concern
because of possible activation of WBC and neutrophils
which could still lead to vaso-occlusive complications and
the risk of early death in SCD.15-19 As CXCR4 is expressed
on most WBC and is involved in the retention of these
cells in bone marrow, a standard dose of plerixafor of 240
μg/kg increases all major WBC subsets (neutrophils, lym-
phocytes, monocytes) in normal donors about 3- to 4-
fold.20 Notably, however, in SCD patients who received G-
CSF, not all patients who had highly elevated WBC counts
experienced vaso-occlusive complications, and conversely,
not all patients who experienced vaso-occlusive complica-
tions had highly elevated WBC counts,5 suggesting that
WBC activation rather than WBC count per se may con-
tribute to vaso-occlusion in SCD.
Another issue of concern is whether enough peripheral
blood CD34+ cells can be mobilized in SCD patients with
plerixafor. The mean and median peak CD34+ counts
using plerixafor alone in normal donors are only
~25/μL.13,20 SCD patients might mobilize particularly well,
in that SCD patients might have increased circulating HPC
even at steady state, although more so during a crisis.21,22
Furthermore, SS and Sβ0 patients tend to be autosplenec-
tomized, and data from patients with thalassemia showed
that splenectomized patients mobilized about twice as
many peripheral blood CD34 cells with plerixafor alone as
non-splenectomized patients.23
Another consideration when using plerixafor is whether
to withhold hydroxyurea, the recommended standard of
care for most SCD patients. Hydroxyurea may inhibit
mobilization and withholding hydroxyurea for 2 weeks
leads to a degree of spontaneous mobilization that abets
drug-induced mobilization.23,24 However, Richard et al.
showed that two of the three SCD patients whose
hydroxyurea was withdrawn for 2 weeks developed
painful crises following the withdrawal. Given these con-
siderations, we designed a prospective phase I dose esca-
lation study of both the safety and efficacy of plerixafor in
patients with SCD in which the patients continued on
their standard outpatient treatment used for disease con-
trol. We have completed the dosing cohorts through to the
standard plerixafor dose of 240 μg/kg and report the inter-
im results here.
Methods
Study design
This study is conducted under FDA IND 122657, registered in
ClinicalTrials.gov as NCT02193191, and approved by the
Institutional Review Boards of Memorial Sloan Kettering, Weill
Cornell Medical College and the New York Blood Center. The
study design is a 3 + 3 dose escalation study with six levels of esca-
lation: doses of 80, 160, 240, 320, 400, and 480 μg/kg. There are
two primary endpoints: (i) efficacy, defined by the achievement of
a HPC mobilization level of 30 CD34+ cells/μL; and (ii) safety,
defined by the occurrence of serious adverse events (≥ grade 3)
that are at least possibly plerixafor-related (including vaso-occlu-
sive events).
At any dose level, the occurrence of at least one grade 3 serious
adverse event results in the addition of three more patients to the
initial three-patient dosing cohort. The occurrence of two grade 3
serious adverse events at a particular dose-level signifies that the
maximal tolerated dose has been exceeded and that the previous
dose-level is the maximum tolerated dose. The trial will be
stopped upon the occurrence of one grade 4 or 5 serious adverse
event at least possibly related to plerixafor. Patients are followed
for adverse events for 1 month after administration of the plerix-
afor. This design provides the following probabilities of escalation
based on the true chance of a dose-limiting toxicity at a specific
dose level:
True probability of toxicity 0.10 0.20 0.30 0.40 0.50 0.60

haematologica | 2018; 103(5) 771

 F. Boulad et al.

Figure 1. Peripheral blood white blood cell, absolute neutrophil, and CD34 cell counts. There is a trend of increasing white blood cell count (WBC, P=0.05) and
absolute neutrophil count (ANC, P=0.03), but not CD34 concentration (P=0.65) with increasing dose of plerixafor. The graphs show the peripheral blood WBC, ANC,
and CD34 concentrations of 15 patients with SCD treated whith 80 (circles), 160 (squares) and 240 (triangles) µg/Kg of plerixafor prior to administration of plerixafor
(PRE), 6-12 h and 20-24 h after plerixafor. Patients on hydroxyurea are represented by filled circles, squares and triangles, patients off hydroxyurea are represented
by open circles, squares and triangles.

Probability of escalation: 0.91 0.71 0.49 0.31 0.17 0.08
For the efficacy endpoint, the dose escalation will continue to
480 μg/kg unless all patients at a preceding dose level achieve a
peripheral blood CD34+ concentration of at least 30 cells/μL. In the
present dose-escalation phase, no leukapheresis is performed. If
and when the efficacy endpoint is safely reached, the study will
proceed to a leukapheresis phase (including preclinical transduc-
tion and editing) in three patients.
Eligible subjects are adults with SS or Sβ0 disease, normal renal
and liver function, hemoglobin concentration ≥6 g/dL, WBC count
≥3,000/μL, absolute neutrophil count (ANC) ≥1,500/μL, and
platelet count of ≥150,000/μL. Eligible subjects are admitted to the
Clinical Research Center at Weill Cornell Medical College. A sin-
gle subcutaneous injection of plerixafor (Sanofi-Genzyme) is
administered in the evening between 8-9 pm. The protocol calls
for peripheral blood sampling at three time points (baseline, 0-2 h
prior to plerixafor; peak between 6-12 h after the plerixafor dose;
at the presumed return to baseline between 20-24 h after the
dose): for reasons of feasibility and patient comfort issues, the
peak sample was consistently drawn at a mean of 12 ± 1 h after
plerixafor administration and the return to baseline sample at a
mean of 20 ± 0.29 h after the dose. Since patients have pre-existing
anemia, for reasons of safety no more than a total of 105 mL of
blood is drawn at all three time points combined.
Peripheral blood CD34 testing
Flow cytometric evaluation of the collected peripheral blood
is performed using a FACS Canto flow cytometer (Becton
Dickinson Biosciences, San Jose, CA, USA) and FACS Diva soft-
ware (BD Biosciences). Samples are stained and analyzed within
2-12 h of collection using a modification of the International
Society of Hematotherapy and Graft Engineering (ISHAGE)
method (see Online Supplementary Methods).
CD34+CD38– enumeration
Mononuclear cells are isolated from 2 mL peripheral blood by
Ficoll-Hypaque Plus density centrifugation. CD34+ cells are puri-
fied by positive selection (MidiMACS™ LS Columns, Miltenyi)
and stained with CD34 (BD PharMingen) and CD38
(Invitrogen).
Research cell and coagulation activation studies
Peripheral blood samples drawn at baseline and after 12 h are
stained within 1 h of collection for activation markers relevant
to sickle vaso-occlusion12,25-28 and assessed by flow cytometry
(BD FACSCanto™). For CD16b+ (1D3, Beckman Coulter) neu-
trophils: activated β2 integrin (clone 24, abcam), activated Mac-
1 (CBRM1/5, eBioscience), E-selectin-Fc chimera (724-ES, R&D
Systems), L-selectin (DREG-56, eBioscience), Mac-1/CD11b
(ICRF44, BD Pharmingen), and LFA-1/CD11a (HI111, BD
Pharmingen). For CD14+ (M5E2, BD Pharmingen) monocytes:
tissue factor (HTF-1, BD Pharmingen). For CD41+ (HIP2, BD
Pharmingen) platelets: CD16b (1D3, Beckman Coulter) and
CD14 (M5E2, BD Pharmingen). The percentages of positive cells
and median fluorescent intensity (MFI) are assessed for each

772 haematologica | 2018; 103(5)


A B
Figure 2. Correlation between post-plerixafor CD34+ cell counts and baseline cell counts. The CD34+ level at 12 h correlated positively with the baseline level of
CD34+ (P=0.0006) but not baseline levels of absolute neutrophil count (ANC, P=0.66) or white blood cells (WBC, P=0.49). The graphs show the association between
the value of peripheral blood CD34 concentration at 12 h after plerixafor and the baseline values of (A) CD34, (B) ANC and (C) WBC in 15 patients with SCD treated
with 80 (circles), 160 (squares) and 240 (triangles) μg/Kg of plerixafor. Patients on hydroxyurea are represented by filled circles, squares and triangles, patients off
hydroxyurea are represented by open circles, squares and triangles.
patient, with calculation of the absolute number of positive cells
by multiplying the percentage of positive cells by the relevant
number of cells obtained from a concurrent complete blood
count. For plasma studies, whole blood is centrifuged at 2500
rpm for 15 min at 4°C and plasma frozen at ≤ -80°C until batch
testing. The following coagulation system activation markers
relevant to sickle vaso-occlusion are tested:26,29 prothrombin frag-
ment 1.2 (Enzygnost F1+2, Dade-Behring) and factor VIII (STA®-
ImmunoDef VIII, Stago).
Statistical analysis
Absolute concentrations as well as the fold increases (ratio of
peak at 12 h to baseline absolute concentrations) of standard
clinical as well as research parameters were analyzed. For
patient 8(2), to avoid computing correlations on an infinite
value, we replaced the baseline value of 0 by 0.2 in the analyses
using CD34+ ratio (10/0.2 = 50). When applying non-parametric
statistics, the chosen value does not affect the results, as long as
it is close to 0. The presence of a trend in increases of CD34+,
ANC, and WBC counts (both at 12 h and fold increases) with
increasing dose was tested using the non-parametric Cuzick test
for trend. Correlations between baseline values, and values at 12
h or fold increases were estimated using the Kendall tau. A dif-
ference in the distribution of 12 h and fold increases according
to the administration of hydroxyurea was investigated using a
Wilcoxon test.
For analyses of cell activation and coagulation, Wilcoxon
signed rank paired testing was performed on the combined dose
cohorts for differences between 12 h and baseline values, where-
as the presence of significant fold differences between dose lev-
els was examined with the Kruskal-Wallis test. Correlations
between values were estimated using the Kendall tau. Two-
tailed P values <0.05 were considered statistically significant.
Results
Patients’ characteristics
Fifteen subjects have been recruited to date for the
study at the first three dose levels of 80, 160 and 240
μg/kg. Fourteen patients were enrolled from Montefiore
haematologica | 2018; 103(5)
Safety and efficacy of plerixafor in SCD patients
C
Medical Center (New York, USA) and one patient from
The Mount Sinai Hospital (New York, USA) (Table 1).
Two patients enrolled at dose levels 1 and 2 were subse-
quently re-enrolled in the study at a higher plerixafor
dose (dose level 3). All subjects had a past history of mod-
erate to severe acute chest syndrome, defined by requir-
ing treatment with simple or exchange transfusion.
Importantly, for safety and feasibility, patients were con-
tinued on their standard outpatient treatment being used
to control their disease. Ten of 15 patients were on
hydroxyurea, with a median HbF level of 12.4% (Online
Supplementary Table S1) and median baseline ANC of
4100/μL (Online Supplementary Table S2). Only one of the
15 subjects was receiving chronic transfusion therapy,
with a HbF of 1.2% and HbA of 54%; this patient was
also on deferasirox for the treatment of transfusion-relat-
ed iron overload. HbA was absent in all other patients. In
the non-transfused patients, HbF levels correlated strong-
ly with hemoglobin concentration, hematocrit, and retic-
ulocyte counts. Of nine patients for whom splenic imag-
ing was available, seven had splenic atrophy (Online
Supplementary Table S1).
Efficacy of CD34+ mobilization
Absolute WBC counts, neutrophil counts and CD34+
cell concentrations increased from baseline in all patients
(Figure 1). Absolute monocyte and lymphocyte counts
also increased from baseline (Online Supplementary Table
S2). Our target goal of mobilizing at least 30 CD34+
cells/μL was, however, reached in only 50% of patients
given the plerixafor dose of 80 μg/kg, 33% of patients
given 160 μg/kg, and 33% of patients given 240 μg/kg.
Peak ANC (P=0.03) and WBC count (P=0.05), but not
CD34+ cell count (P=0.65), increased with increasing dose
level. As previously reported in healthy donors,13,30,31 there
was a strong correlation of peak CD34+ count with base-
line CD34+ concentration (Kendall tau=0.68, P=0.0006)
but no correlation was observed with baseline ANC
(Kendall tau=0.09, P=0.66) or baseline WBC count
(Kendall tau=0.13, P=0.49) (Figure 2). There was also no
correlation, as previously reported,13 with baseline platelet
773
 F. Boulad et al.
A B
Figure 3. Association of hydroxyurea treatment with post-plerixafor cell counts. The use or not of hydroxyurea was not associated with different levels of CD34+
(P=0.66), absolute neutrophil count (ANC, P=0.66), or white blood cell count (WBC, P=0.57) 12 h after plerixafor. Peripheral blood (A) CD34 concentrations, (B) ANC,
and (C) WBC at 12 h after plerixafor in 15 patients with SCD treated with 80 (circles), 160 (squares) and 240 (triangles) μg/kg of plerixafor. Patients on hydroxyurea
are represented by filled circles, squares and triangles, patients off hydroxyurea are represented by open circles, squares and triangles.
count (Kendall tau=0.30, P=0.13) or donor age (Kendall
tau=0.14, P=0.49).
There was a significant increase in the median CD34+
fold increase with dose (P=0.01) (Online Supplementary
Figure S1). A trend was observed for ANC ratio, although
not statistically significant (1.6-, 1.8-, and 2.1-fold increas-
es, P=0.08). No trend was seen for WBC ratio (P=0.13).
With the caveat of statistical adjustment for patient 8(2)'s
baseline CD34+ count of 0/μL, there was no correlation
between CD34+ fold increase and baseline CD34+ (Kandall
tau= -0.32, P=0.11), baseline ANC (Kendall tau=0.19,
P=0.32), or baseline WBC (Kendall tau=0.22, P=0.25)
(Online Supplementary Figure S2).
Hydroxyurea was not associated with differences in
peak absolute CD34+ concentration (P=0.95), peak ANC
(P=0.59) or peak WBC (P=0.68) concentrations (Figure 3);
or with differences in the fold increases of CD34+ cell
(P=0.64), ANC (P=0.12), or WBC (P=0.36) concentrations
(Online Supplementary Figure S3).
In a subset of six patients, CD34+CD38– cells were enu-
merated (Online Supplementary Table S3), showing a medi-
an 3-fold increase in CD34+CD38– concentrations at 12 h.
Safety of plerixafor
There were no significant changes in hemoglobin con-
centration, hematocrit, or platelet counts with plerixafor
treatment (data not shown, baseline values in Online
Supplementary Table S1). Due to the occurrence of one seri-
ous adverse event at the 80 μg/kg dose and another one at
the 240 μg/kg dose, an additional three patients were
enrolled at each of these dose levels. The serious adverse
events were both pain crises, possibly related to plerixafor
(Online Supplementary Table S4), but also associated with
other possibly contributory events. Patient 13 with a seri-
ous adverse event had the second highest peak ANC (and
third highest peak WBC count) in the study, but high
WBC count and ANC were not consistently associated
with serious adverse events.
There were no significant differences between dose lev-
els for any of the activation markers of vaso-occlusion test-
ed. With the exception of tissue factor-positive (TF+) mono-
774
C
cytes at the 240 µg/kg dose, the median fold-changes in
percentage of cells at every dose cohort were ≤1.1, arguing
against generalized plerixafor-mediated cell activation
(Figure 4A). Median fold increases in absolute numbers of
activated β2 integrin-positive (aβ2+) neutrophils, activated
Mac-1-positive (aMac-1+) neutrophils, and TF+ monocytes
(160 μg/kg, 240 μg/kg) were close to 2 (Figure 4B). There
were strong correlations between the fold increase in
absolute numbers of neutrophils and fold increases in aβ2+
neutrophils (Kendall tau=0.85, P<0.001) and aMac-1+ neu-
trophils (Kendall tau=0.46, P=0.02). The absolute numbers
of aβ2+ and aMac-1+ neutrophils were significantly
increased (Online Supplementary Figure S4A,B), and the two
patients with serious adverse events (gray arrows) had rel-
atively high absolute numbers of aβ2+ and aMac-1+ neu-
trophils, albeit not the highest. There was also a significant
increase in plasma prothrombin fragment 1.2 concentra-
tions (Online Supplementary Figure S4C), but the two
patients with serious adverse events had absolute concen-
trations and fold increases at 12 h that were lower than the
median and mean for that measure. Both patients with
serious adverse events had relatively high fold-increases in
L-selectinneg neutrophils and one had a large fold increase in
TF+ monocytes (Figure 5), but their absolute numbers of L-
selectinneg neutrophils and TF+ monocytes were not partic-
ularly high (Online Supplementary Figure S4D,E). There were
significant decreases for five parameters: percentage of
aβ2+ neutrophils, MFI of aβ2 neutrophils, percentage of TF+
monocytes, and percentage and absolute number of
platelet-neutrophil aggregates (Online Supplementary Figure
S5). There were no significant changes in the MFI of aMac-
1+, Mac-1, LFA-1, or L-selectin on neutrophils (data not
shown).
Discussion
Eight of 15 patients (53%) with SCD treated with pler-
ixafor reached the peripheral blood CD34 cell target count
of at least 30 CD34+ cells/μL, including three of six
patients treated at a dose of 240 μg/kg. This is in contrast
haematologica | 2018; 103(5)

B ber of cells for TF+ mono, aβ2+ PMN, and
with the findings of a recent study by Tisdale et al., in
which mobilization was effective in seven of seven SCD
subjects (100%) at a dose of 240 μg/kg.4 It should be noted
that patients in the National Institutes of Health study
were off hydroxyurea and had been transfused for at least
2 months to achieve a HbS <20-30%4,32 while in our study,
ten of the 15 patients were on stable doses of hydroxyurea
(for at least 1 year) and only one patient was on chronic
transfusion. Although hydroxyurea, a ribonucleotide
reductase inhibitor, causes myelosuppression and was
recently found to reduce CD34 counts in peripheral blood
and bone marrow,33 there is no definitive evidence that
hydroxyurea negatively affects numbers or quality of cell
cycle-quiescent hematopoietic stem cells or immature
bone marrow progenitors as opposed to more mature
myeloid-erythroid progenitors.33-37 Indeed, in our study,
although we did not achieve consistent efficiency in CD34
cell mobilization, no correlation was found between
hydroxyurea use, and absolute or fold increases in CD34+
cells/μL.
We observed wide inter-donor variability in CD34
mobilization with plerixafor, as previously reported in
normal donors (CD34 peaks between 4-157/μL )13 and in
patients with SCD (CD34 peaks between 50-200/μL).4
However, we also observed a strong correlation between
baseline CD34+ and peak CD34+ concentrations, as previ-
ously reported with both G-CSF and plerixafor mobiliza-
tion in healthy donors (Kendall tau=0.68, P=0.0006).13,30,31
Factors contributing to baseline CD34 count remain
unclear, but our data and others’ suggest that baseline
CD34+ concentration may be affected by hydroxyurea-
haematologica | 2018; 103(5)
Safety and efficacy of plerixafor in SCD patients
Figure 4. Fold changes in the percent-
age of cells expressing activation mark-
ers, according to plerixafor dose. (A)
Other than % TF+ monocytes, median
fold changes in % cells were ≤1.1. (B)
Median fold changes in absolute num-
aMac-1+ PMN were closer to 2 than 1.
Fold changes are ratios of 12 h to base-
line values. Horizontal lines represent
medians. Filled shapes represent
hydroxyurea-treated patients, open
shapes represent patients not treated
with hydroxyurea. Gray arrows point to
the values for the two SAE patients who
had serious adverse events. TF+ mono:
tissue-factor-positive monocytes; E-Sel+
PMN: E-selectin-positive neutrophils;
aβ2+ PMN: activated β2 integrin-positive
neutrophils, aMac-1+ PMN: activated
Mac-1-positive neutrophils; L-Selneg PMN:
L-selectin-negative neutrophils; PMA:
platelet-monocyte aggregates; PNA:
platelet-neutrophil aggregates.
related myelosuppression.24,33 Patient #8, a subject re-
enrolled in the study, was particularly instructive regard-
ing this hypothesis. This patient was clinically stable on
hydroxyurea at a dose of 27 mg/kg and was enrolled twice
at an interval of 13 months. At the time of his second
treatment, however, he had a markedly lower baseline
ANC (1900/μL down from 6300/μL) and platelet count
(217,000/μL down from 400,000/μL), probably related to
oscillatory non-toxic hematopoiesis seen in SCD with
chronic and dose-intensive treatment with hydroxyurea
(ANC oscillations between 2,000-6,000/μL as determined
from review of his clinical laboratory records).38 This
myelosuppression was associated with a baseline CD34
concentration of 0/μL rather than 1/μL, possibly contribut-
ing to the relatively low 12 h CD34+ concentration of
10/μL at the 240 μg/kg dose as compared to 27/μL at the
160 μg/kg dose. In brief, because hydroxyurea can
decrease ANC and platelet count,39 hydroxyurea-related
myelosuppression may have contributed to the relatively
poor CD34+ mobilization obtained in this cohort.
However, avoiding hydroxyurea withdrawal might lower
the risk of pain crises;24 we, therefore, plan to explore tim-
ing plerixafor administration to the peak rather than nadir
of hydroxyurea-related oscillatory hematopoiesis. Finally
with regards to hydroxyurea therapy, data from the six
patients in whom we enumerated CD34+CD38– cells sug-
gest that hydroxyurea may not adversely affect HSC,
given that all patients except one (patient 10) were on
hydroxyurea and a median 3-fold increase at 12 h was
observed. Only 0.2-2.8% of CD34+ cells were CD38-neg-
ative, but this may be consistent with plerixafor’s effect in
775
 F. Boulad et al.

normal healthy donors, in that fewer HPC may be mobi-
lized by plerixafor than by G-CSF, where up to 50% of G-
CSF-mobilized CD34+ cells are CD38-negative.13,40
Because we enrolled only one patient on chronic trans-
fusion, we cannot assess any correlation between transfu-
sion and CD34 mobilization, although notably this
patient had the second highest baseline and highest peak
CD34+ cell counts in our study. Other studies of plerixafor
in SCD4,32 have initiated chronic transfusion based on the
hypothesis that the inflammatory nature of SCD affects
the bone marrow and transfusion assuages bone marrow
inflammation and stress erythropoiesis. Although replica-
tive and oxidative stress of HPC in bone marrow may
occur,41-44 there is limited evidence that HPC are damaged
in SCD.36 Five of our patients had HbF-associated increas-
es in hemoglobin concentration and hematocrit to more
than 10 g/dL and 30%, respectively (similar to values in
chronically transfused patients) but HbF levels did not cor-
relate with CD34 cell mobilization.
Based on our data, it is possible that continued dose esca-
lation could result in greater efficacy of mobilization, since
we observed a dose-related response in the median CD34+
cell fold increase (P=0.01), as also observed in healthy
donors.45 Patient 3, a repeat enrollment who had never
been on hydroxyurea and was clinically stable, is instruc-
tive in that his 12 h peak CD34+ cell count following a pler-
ixafor dose of 80 μg/kg was only 8/μL whereas at the dose
of 240 μg/kg it was 40/μL, even though his baseline CD34+
cell concentrations (1/μL and then 2/μL) were similar, sug-
gesting a dose-response to plerixafor. Notably, his two
periods in the study were separated by 19 months, sug-
gesting that, as with healthy donors,46 intra-individual
CD34+ cell counts in stable SCD patients not on hydrox-
yurea may remain stable over time. Based on these data
and given the safety and continued dose response between
240 μg/kg and 480 μg/kg observed in healthy donors,20 we
plan to continue dose escalation in SCD patients through
to the 480 μg/kg dose, barring significant adverse events.
Adding the CXCR2 agonist, GROβ, might be useful.47
Only two of 15 patients (13%) developed serious adverse
events as compared to three of seven patients (43%) in the
study of plerixafor mobilization in SCD by Tisdale et al.,4
although this must be qualified by the fact that the patients
in the study by Tisdale et al. also underwent leukapheresis.
Our low rate of serious adverse events could, however, also
be due to chronic hydroxyurea therapy and the subsequent
lower WBC and ANC peaks. As the fraction of activated
neutrophils did not increase significantly with plerixafor,
our low serious adverse event rate may be related to mod-
eration of ANC elevations by hydroxyurea, reducing the
absolute number of activated cells. Given the still uncertain
risks of morbidity, as seen with G-CSF, the use of plerixafor
in SCD requires further evaluation.
In summary, our present data suggest that, with regards
the efficacy of CD34 mobilization, red blood cell transfu-
sion may be more effective than continuing standard of
care. Whether red blood cell transfusion will remain more
effective as we escalate the plerixafor dose (as safety
allows) to 480 μg/kg, with protocol revisions for hydrox-
yurea-treated patients, is unknown. Finally, potential can-
didates for SCD gene therapy may not be able to receive
regular red blood cell transfusions (e.g. if they have red cell
alloimmunization or a history of hyperhemolysis) or may
not be willing to do so (e.g. Jehovah Witnesses), even for
the relatively short duration of 2-3 months.
This study has several limitations. Firstly, despite this
study being the largest study to date of plerixafor adminis-
tration in SCD patients, overall the number of patients
involved remains small; thus comparisons, for example,
between hydroxyurea-treated and non-hydroxyurea-treat-
ed patients, may not be representative of the actual popu-
lations. Secondly, we measured WBC, ANC and CD34
mobilization in this study only at ~12 and ~20 h after pler-
ixafor administration. It is possible that an initial peak
could have occurred at an earlier time (6-9 h) after plerix-
afor and could, therefore, have been missed. Nevertheless,
CD34 cell concentrations remain at ~70% of peak levels at
12 h.20,45,48 Our current study will be amended to include the
addition of earlier post-plerixafor assessments. Thirdly, we
determined peripheral blood CD34 cell mobilization in the
15 patients treated with plerixafor, without performing
apheresis. However, there is a well-described correlation
between peripheral blood CD34 cell concentration and the
ultimate CD34 cell dose obtained after apheresis. It is pos-
sible that technical adjustments may be required for this
equation in the context of SCD. Finally, other than enu-
merating CD34+CD38- cells, we did not further character-
ize CD34+ cells to study “stemness”, for example by deter-
mining glycophorin A positivity and CD34 dimness.4
CD34+ or CD34+CD38- enumeration is not specific for
HSC49 and it is, therefore, unclear whether patients had a
true increase in HSC, as opposed to more mature lineage-
committed CD34+ progenitors, which are either mobilized
or present in bone marrow.42 We plan to characterize the
CD34+ cells further as we move forward in the study,
which is currently enrolling at the 320 μg/kg dose level.
Despite mobilization of HSC possibly being less efficient
with plerixafor than with G-CSF, plerixafor-mobilized
HSC may have an engraftment advantage over G-CSF-
mobilized HSC with regard to better retention of CXCR4,
which facilitates homing.32
Acknowledgments
The authors would like to thank our study subjects for their
participation; the Doris Duke Charitable Foundation for a 2011
Innovation in Clinical Research Award for trial support (to PAS
and MS); Sanofi-Genzyme for provision of plerixafor; Jena
Simon for referring one study patient; W. Beau Mitchell for assis-
tance with platelet activation studies; and Henny Billett, Narla
Mohandas, and Beth Shaz for departmental support.

2. Ribeil JA, Hacein-Bey-Abina S, Payen E, et Successful Plerixafor-mediated mobilization,

References
1. Mansilla-Soto J, Riviere I, Sadelain M.
Genetic strategies for the treatment of sickle
cell anaemia. Br J Haematol. 2011;154(6):
715-727.
al. Gene therapy in a patient with sickle cell
disease. N Engl J Med. 2017;376 (9):848-855.
3. Sadelain M, Boulad F, Lisowki L, Moi P,
Riviere I. Stem cell engineering for the treat-
ment of severe hemoglobinopathies. Curr
Mol Med. 2008;8(7):690-697.
4. Tisdale JF, Pierciey Jr. JF, Kanter J, et al.
apheresis, and lentiviral vector transduction
of hematopoietic stem cells in patients with
severe sickle cell disease. Blood. 2017;130
(Suppl 1):990.
5. Fitzhugh CD, Hsieh MM, Bolan CD, Saenz
C, Tisdale JF. Granulocyte colony-stimulat-
ing factor (G-CSF) administration in individ-

776 haematologica | 2018; 103(5)


uals with sickle cell disease: time for a mora-
torium? Cytotherapy. 2009;11(4):464-471.
6. Petit I, Szyper-Kravitz M, Nagler A, et al. G-
CSF induces stem cell mobilization by
decreasing bone marrow SDF-1 and up-reg-
ulating CXCR4. Nat Immunol. 2002;3(7):
687-694.
7. Cella G, Marchetti M, Vignoli A, et al. Blood
oxidative status and selectins plasma levels
in healthy donors receiving granulocyte-
colony stimulating factor. Leukemia.
2006;20(8):1430-1434.
8. de Haas M, Kerst JM, van der Schoot CE, et
al. Granulocyte colony-stimulating factor
administration to healthy volunteers: analy-
sis of the immediate activating effects on cir-
culating neutrophils. Blood. 1994;84(11):
3885-3894.
9. Falanga A, Marchetti M, Evangelista V, et al.
Neutrophil activation and hemostatic
changes in healthy donors receiving granulo-
cyte colony-stimulating factor. Blood.
1999;93(8):2506-2514.
10. Spiel AO, Bartko J, Schwameis M, et al.
Increased platelet aggregation and in vivo
platelet activation after granulocyte colony-
stimulating factor administration. A ran-
domised controlled trial. Thromb Haemost.
2011;105(4):655-662.
11. Canales MA, Arrieta R, Gomez-Rioja R,
Diez J, Jimenez-Yuste V, Hernandez-
Navarro F. Induction of a hypercoagulability
state and endothelial cell activation by gran-
ulocyte colony-stimulating factor in periph-
eral blood stem cell donors. J Hematother
Stem Cell Res. 2002;11(4):675-681.
12. Zhang D, Xu C, Manwani D, Frenette PS.
Neutrophils, platelets, and inflammatory
pathways at the nexus of sickle cell disease
pathophysiology. Blood. 2016;127(7):801-
809.
13. Schroeder MA, Rettig MP, Lopez S, et al.
Mobilization of allogeneic peripheral blood
stem cell donors with intravenous plerixafor
mobilizes a unique graft. Blood. 2017;129
(19):2680-2692.
14. Choi E, Branch C, Cui MH, et al. No evidence
for cell activation or brain vaso-occlusion
with plerixafor mobilization in sickle cell
mice. Blood Cells Mol Dis. 2016;57:67-70.
15. Charache S. Mechanism of action of
hydroxyurea in the management of sickle
cell anemia in adults. Semin Hematol.
1997;34(3 Suppl 3):15-21.
16. Castro O, Brambilla DJ, Thorington B, et al.
The acute chest syndrome in sickle cell dis-
ease: incidence and risk factors. The
Cooperative Study of Sickle Cell Disease.
Blood. 1994;84(2):643-649.
17. Steinberg MH, Barton F, Castro O, et al.
Effect of hydroxyurea on mortality and mor-
bidity in adult sickle cell anemia: risks and
benefits up to 9 years of treatment. JAMA.
2003;289(13):1645-1651.
18. Litos M, Sarris I, Bewley S, Seed P, Okpala I,
Oteng-Ntim E. White blood cell count as a
predictor of the severity of sickle cell disease
during pregnancy. Eur J Obstet Gynecol
Reprod Biol. 2007;133(2):169-172.
19. Platt OS, Brambilla DJ, Rosse WF, et al.
Mortality in sickle cell disease. Life
expectancy and risk factors for early death.
N Engl J Med. 1994;330(23):1639-1644.
20. Pantin J, Purev E, Tian X, et al. Effect of high-
dose plerixafor on CD34(+) cell mobiliza-
tion in healthy stem cell donors: results of a
haematologica | 2018; 103(5)
Safety and efficacy of plerixafor in SCD patients
randomized crossover trial. Haematologica.
2017;102(3):600-609.
21. Lamming CE, Augustin L, Blackstad M,
Lund TC, Hebbel RP, Verfaillie CM.
Spontaneous circulation of myeloid-lym-
phoid-initiating cells and SCID-repopulating
cells in sickle cell crisis. J Clin Invest.
2003;111(6):811-819.
22. Croizat H, Ponchio L, Nicolini FE, Nagel
RL, Eaves CJ. Primitive haematopoietic
progenitors in the blood of patients with
sickle cell disease appear to be endoge-
nously mobilized. Br J Haematol. 2000;111
(2):491-497.
23. Yannaki E, Papayannopoulou T, Jonlin E, et
al. Hematopoietic stem cell mobilization for
gene therapy of adult patients with severe
beta-thalassemia: results of clinical trials
using G-CSF or plerixafor in splenectomized
and nonsplenectomized subjects. Mol Ther.
2012;20(1):230-238.
24. Richard RE, Siritanaratkul N, Jonlin E,
Skarpidi E, Heimfeld S, Blau CA. Collection
of blood stem cells from patients with sickle
cell anemia. Blood Cells Mol Dis.
2005;35(3):384-388.
25. Manwani D, Chen G, Carullo V, et al.
Single-dose intravenous gammaglobulin can
stabilize neutrophil Mac-1 activation in sick-
le cell pain crisis. Am J Hematol.
2015;90(5):381-385.
26. Jakubowski JA, Zhou C, Jurcevic S, et al. A
phase 1 study of prasugrel in patients with
sickle cell disease: effects on biomarkers of
platelet activation and coagulation. Thromb
Res. 2014;133(2):190-195.
27. Canalli AA, Proenca RF, Franco-Penteado CF,
et al. Participation of Mac-1, LFA-1 and VLA-
4 integrins in the in vitro adhesion of sickle
cell disease neutrophils to endothelial layers,
and reversal of adhesion by simvastatin.
Haematologica. 2011;96(4):526-533.
28. Wun T, Styles L, DeCastro L, et al. Phase 1
study of the E-selectin inhibitor GMI 1070 in
patients with sickle cell anemia. PLoS One.
2014;9(7):e101301.
29. Nur E, van Beers EJ, Martina S, et al. Plasma
levels of pentraxin-3, an acute phase protein,
are increased during sickle cell painful crisis.
Blood Cells Mol Dis. 2011;46(3):189-194.
30. de la Rubia J, Lorenzo JI, Torrabadella M,
Marin P, Insunza A, Sanz MA. Basal
CD34(+) cell count predicts peripheral blood
progenitor cell mobilization and collection
in healthy donors after administration of
granulocyte colony-stimulating factor.
Haematologica. 2004;89(12):1530-1532.
31. Martino M, Gori M, Pitino A, et al. Basal
CD34(+) Cell count predicts peripheral
blood stem cell mobilization in healthy
donors after administration of granulocyte
colony-stimulating factor: a longitudinal,
prospective, observational, single-center,
cohort study. Biol Blood Marrow
Transplant. 2017;23(7):1215-1220.
32. Kanter J, Walters MC, Hsieh MM, et al.
Interim results from a phase 1/2 clinical
study of lentiglobin gene therapy for severe
sickle cell disease. Blood. 2017;128
(22):1176.
33. Uchida N, Fujita A, Hsieh MM, et al. Bone
marrow as a hematopoietic stem cell source
for gene therapy in sickle cell disease: evi-
dence from Rhesus and SCD patients. Hum
Gene Ther Clin Dev. 2017;28(3):136-144.
34. Miller ST, Rey K, He J, et al. Massive acci-
dental overdose of hydroxyurea in a young
child with sickle cell anemia. Pediatr Blood
Cancer. 2012;59(1):170-172.
35. Castro O, Nouraie M, Oneal P.
Hydroxycarbamide treatment in sickle cell
disease: estimates of possible leukaemia risk
and of hospitalization survival benefit. Br J
Haematol. 2014;167(5):687-691.
36. Brunson A, Keegan THM, Bang H, Mahajan
A, Paulukonis S, Wun T. Increased risk of
leukemia among sickle cell disease patients
in California. Blood. 2017;130(13):1597-
1599.
37. Drasar ER, Jiang J, Gardner K, et al.
Leucocyte telomere length in patients with
sickle cell disease. Br J Haematol. 2014;165
(5):725-727.
38. Baird JH, Minniti CP, Lee JM, et al.
Oscillatory haematopoiesis in adults with
sickle cell disease treated with hydroxycar-
bamide. Br J Haematol. 2015;168(5):737-746.
39. Zimmerman SA, Schultz WH, Davis JS, et
al. Sustained long-term hematologic efficacy
of hydroxyurea at maximum tolerated dose
in children with sickle cell disease. Blood.
2004;103(6):2039-2045.
40. Worel N, Frank N, Frech C, Fritsch G.
Influence of plerixafor on the mobilization
of CD34+ cell subpopulations and lympho-
cyte subtypes. Transfusion. 2017;57(9):
2206-2215.
41. Javazon EH, Radhi M, Gangadharan B, Perry
J, Archer DR. Hematopoietic stem cell func-
tion in a murine model of sickle cell disease.
Anemia. 2012;2012:387385.
42. Leonard A, Bonifacino A, Dominical VM, et
al. Bone marrow characterization in sickle
cell disease: inflammation and stress ery-
thropoiesis lead to suboptimal CD34 recov-
ery compared to normal volunteer bone
marrow. Blood. 2017;130(Suppl 1):966.
43. Dallalio G, Brunson CY, Means RT, Jr.
Cytokine concentrations in bone marrow of
stable sickle cell anemia patients. J Investig
Med. 2007;55(2):69-74.
44. Colella MP, Santana BA, Conran N, et al.
Telomere length correlates with disease
severity and inflammation in sickle cell dis-
ease. Rev Bras Hematol Hemoter. 2017;39
(2):140-145.
45. Liles WC, Broxmeyer HE, Rodger E, et al.
Mobilization of hematopoietic progenitor
cells in healthy volunteers by AMD3100, a
CXCR4 antagonist. Blood. 2003;102(8):
2728-2730.
46. Eidenschink L, DiZerega G, Rodgers K,
Bartlett M, Wells DA, Loken MR. Basal lev-
els of CD34 positive cells in peripheral blood
differ between individuals and are stable for
18 months. Cytometry B Clin Cytom.
2012;82(1):18-25.
47. Hoggatt J, Singh P, Tate TA, et al. Rapid
mobilization reveals a highly engraftable
hematopoietic stem cell. Cell. 2018;172(1-
2):191-204.e10.
48. Lemery SJ, Hsieh MM, Smith A, et al. A
pilot study evaluating the safety and CD34+
cell mobilizing activity of escalating doses of
plerixafor in healthy volunteers. Br J
Haematol. 2011;153(1):66-75.
49. Notta F, Doulatov S, Laurenti E, Poeppl A,
Jurisica I, Dick JE. Isolation of single human
hematopoietic stem cells capable of long-
term multilineage engraftment. Science.
2011;333(6039):218-221.
777
ARTICLE Red Cell Biology & its Disorders

Plerixafor enables safe, rapid, efficient

Ferrata Storti
Foundation
mobilization of hematopoietic stem cells
in sickle cell disease patients after
exchange transfusion
Chantal Lagresle-Peyrou,1,2,3* François Lefrère,4* Elisa Magrin,1,4*
Jean-Antoine Ribeil,1,4* Oriana Romano,3,5,6 Leslie Weber,2,3,7 Alessandra
Magnani,1,4 Hanem Sadek,1,2,3 Clémence Plantier,1,4 Aurélie Gabrion,1,4
Brigitte Ternaux,1,4 Tristan Félix,3,5 Chloé Couzin,1,4 Aurélie Stanislas,1,4
Jean-Marc Tréluyer,8 Lionel Lamhaut,9,10 Laure Joseph,4 Marianne Delville,2,3,4
Haematologica 2018
Annarita Miccio,3,5# Isabelle André-Schmutz1,2,3# and Marina Cavazzana1,2,3,4#
Volume 103(5):778-786

1
Biotherapy Clinical Investigation Center, Groupe Hospitalier Universitaire Ouest,
Assistance Publique-Hôpitaux de Paris, INSERM CIC 1416, France; 2Laboratory of
Human Lymphohematopoiesis, INSERM UMR 1163, Imagine Institute, Paris, France:
3
Paris Descartes University – Sorbonne Paris Cité, Imagine Institute, France
4
Department of Biotherapy, Necker Children’s Hospital, Assistance Publique-Hôpitaux
de Paris, France; 5Laboratory of Chromatin and Gene Regulation during Development,
INSERM UMR1163, Imagine Institute, Paris, France; 6Department of Life Sciences,
University of Modena and Reggio Emilia, Modena, Italy; 7Paris Diderot University –
Sorbonne Paris Cité, France; 8Mère-Enfant Clinical Investigation Center, Groupe
Hospitalier Necker Cochin, Assistance Publique-Hôpitaux de Paris, France; 9Intensive
Care Unit, Anaesthesia and SAMU de Paris, Necker Hospital, Assistance Publique-
Hôpitaux de Paris, France and 10Paris Descartes University – Sorbonne Paris Cité,
France.
*CLP, FL and EM and JAR contributed equally to this work, in alphabetical order
#
AM, IAS and MC contributed equally to this work

Correspondence:
ABSTRACT
isabelle.andre-schmutz@inserm.fr

Received: November 15, 2017.
Accepted: February 13, 2018.
Pre-published: February 22, 2018.
doi:10.3324/haematol.2017.184788
Check the online version for the most updated
information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/778
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
S
ickle cell disease is characterized by chronic anemia and vaso-occlu-
sive crises, which eventually lead to multi-organ damage and pre-
mature death. Hematopoietic stem cell transplantation is the only
curative treatment but it is limited by toxicity and poor availability of
HLA-compatible donors. A gene therapy approach based on the autolo-
gous transplantation of lentiviral-corrected hematopoietic stem and pro-
genitor cells was shown to be efficacious in one patient. However, alter-
ations of the bone marrow environment and properties of the red blood
cells hamper the harvesting and immunoselection of patients’ stem cells
from bone marrow. The use of Filgrastim to mobilize large numbers of
hematopoietic stem and progenitor cells into the circulation has been
associated with severe adverse events in sickle cell patients. Thus, broad-
er application of the gene therapy approach requires the development of
alternative mobilization methods. We set up a phase I/II clinical trial
whose primary objective was to assess the safety of a single injection of
Plerixafor in sickle cell patients undergoing red blood cell exchange to
decrease the hemoglobin S level to below 30%. The secondary objective
was to measure the efficiency of mobilization and isolation of
hematopoietic stem and progenitor cells. No adverse events were
observed. Large numbers of CD34+ cells were mobilized extremely
quickly. Importantly, the mobilized cells contained high numbers of
hematopoietic stem cells, expressed high levels of stemness genes, and
engrafted very efficiently in immunodeficient mice. Thus, Plerixafor can
be safely used to mobilize hematopoietic stem cells in sickle cell
patients; this finding opens up new avenues for treatment approaches
based on gene addition and genome editing. Clinicaltrials.gov identifier:
NCT02212535.

778 haematologica | 2018; 103(5)

 Stem cell mobilization in sickle cell patients

Introduction patients,6-7 the recovery of HSPC from SCD patients’ BM

Sickle cell disease (SCD) is caused by a point mutation
in the coding region of the HBB (β-globin) gene. As a
result, an abnormal β-globin protein is incorporated into
hemoglobin tetramers. These mutant tetramers polymer-
ize when the local oxygen tension is low. The sickle
hemoglobin (HbS) polymers rigidify red blood cells,
change these cells’ shape, and are responsible for structur-
al damage to the red blood cell membrane. In turn, this
modifies the cells’ rheological properties, alters their flow
in the microcirculation, and thus causes ischemia, stroke,
multi-organ damage, severe acute and chronic pain, and
chronic hemolytic anemia. Progressive chronic organ com-
plications become the main cause of morbidity and mor-
tality in the third decade of life.1 SCD is endemic in Africa,
and the World’s Health Organization considers that 7% of
the world population carries the trait.
The only curative treatment for SCD is allogeneic
hematopoietic stem cell transplantation (HSCT) from
matched sibling donors; the disease-free survival rate 6
years after transplantation is reportedly >90%.2,3 Given
the limited availability of suitable donors and the increase
in toxicity with age, HSCT is only applied with great cau-
tion in adult SCD patients (the main target population for
curative treatment).
We recently demonstrated that gene therapy is applica-
ble to SCD patients, and that the associated toxicity and
morbidity rates seem to be lower than those for allogeneic
HSCT, at least in the first treated patient.4
As is the case with all genetic diseases, the success of
gene therapy in SCD relies on several key factors; these
include the source, quality and number of transduced
cells, the choice of the conditioning regimen, the level of
therapeutic transgene expression, and the quality of the
bone marrow (BM) microenvironment at the time of har-
vest and transplantation. It is generally acknowledged that
2 to 3x106 CD34+ hematopoietic stem and progenitor cells
(HSPC)/kg are required for a successful outcome in autol-
ogous HSCT.5 Considering the typical proportion of HSPC
that can be corrected in gene therapy clinical trials (~50%
of CD34+ HSPC) and an average recovery of 70% of
CD34+ cells post-selection, a minimum harvest of ~6x106
CD34+ cells/kg would be required. For reasons that have
not been completely elucidated, as for thalassemic
is peculiarly low (M. Cavazzana, unpublished data). In our
ongoing gene therapy trial (HGB-205, ClinicalTrials.gov
number, NCT02151526), two BM harvests (each requiring
an exchange transfusion program before general anesthe-
sia) were needed to obtain enough cells to transplant the
three SCD patients enrolled.4
Mobilization with granulocyte colony-stimulating fac-
tor (G-CSF, Filgrastim) is widely used to increase the har-
vest of HSPC (relative to that obtained in a BM aspirate).
However, attempts to mobilize HSPC with Filgrastim in
SCD have led to severe adverse events, which hamper the
cytokine’s use in this setting. The first report of a severe
adverse event following mobilization with low-dose
Filgrastim (2.5 μg/kg/day) concerned a SCD patient who
developed acute chest syndrome and an elevated white
blood cell count (63,400/mm3) as early as 3 days after the
first injection.8 The temporal relationship between
Filgrastim administration, the rapid rise in the white blood
cell count and the severe adverse event were strongly sug-
gestive of a causal link. Two other severe adverse events
(multi-organ failure and a death) were reported in 2001.9-10
Between 2003 and 2008, eight other patients requiring
autologous HSCT for a malignant hematopoietic disease
were mobilized with Filgrastim after additional precau-
tions had been taken: reduction of the HbS level to below
30% via red blood cell exchange, and carefully monitoring
of peripheral leukocytosis and the blood ion profile.11-13
Although all eight patients experienced bone pain, hyper-
tension or migraine, no severe adverse events were report-
ed - providing evidence that SCD patients can be mobi-
lized without any major complications. The median
[interquartile range] circulating CD34+ cell count was
24.4/μL [21.2-48.6].12-14
As an alternative to Filgrastim, Plerixafor (formerly
known as AMD3100) can effectively mobilize HSPC; the
CD34+ cell count in the circulation can reach 15 to 40/μL,
depending on the dose and the study.15,16 In contrast to
Filgrastim (which acts by activating monocytes and neu-
trophils in both peripheral blood and the BM), Plerixafor
directly inhibits the binding of stroma-cell-derived factor-
1a to its CXC chemokine receptor on HSPC – thus releas-
ing stem cells from the BM niches. This mechanistic dif-
ference explains the small increase in the white blood cell
count and the short time interval between Plerixafor

Table 1. Clinical parameters of patients treated with Plerixafor.

SCD Pler 1 (19 years) SCD Pler 2 (20 years) SCD Pler 3 (21 years)
Normal values D-1 D+1 D+7 D+30 D-1 D+1 D+7 D+30 D-1 D+1 D+7 D+30
HbS (HPLC G8) 0% 12·1% 12·6% 15·5% 36·0% 20·2% 15·8% ND 37·1% 6·2%*,** 13·2%** 25·2%** 36·8%**
Total bilirubin (mol/L) 0-17 112 105 109 132 32 35 ND 32 30 30 31 53
Conjugated bilirubin (mol/L) 0-5 8 7 8 9 11 13 ND 12 10 10 10 6
Lactate dehydrogenase (U/L) 125-243 399 409 361 429 566 584 ND 605 378 383 426 501
White blood cells (109/L) 4·0-10·0 9·9 20·1 11·8 13·2 11·9 11·2 ND 6·2 14·0 11·3 9·6 10·4
Neutrophils (109/L) 1·5-7 4·9 14·9 7·8 8·2 7·6 9·2 ND 2·6 9·6 9·3 6·4 6·5
Monocytes (109/L) 0·2-1 2·0 2·4 1·7 2·3 0·9* 0·9 ND 0·7 1·8* 1·3 1·5 1·8
Ferritin (μg/L) 22-275 2441 ND 2674 3111 ND 1009 1473 1962 ND 381 365 446
Liver quantification < 36 340 (±50) 120 (±30) 55 (±30)
(at inclusion period) (μmol/g)
HPLC: High-performance-liquid-chromatography; D: day. ND: not determined. * D+3. **Capillary 3 (and not G8).

haematologica | 2018; 103(5) 779

 C. Lagresle-Peyrou et al.
administration and HSPC mobilization. Furthermore,
Plerixafor synergistically augments Filgrastim-induced
mobilization of HSPC,17 and so has been widely combined
with Filgrastim (i) in Filgrastim-non-responsive patients,
(ii) as an adjunct to increase the overall recovery of CD34+
cells for autologous or allogenic HSCT, and (iii) in gene
therapy trials for β-thalassemia and Wiskott-Aldrich syn-
drome.18-20 In view of all the above data, we initiated a
phase I/II clinical trial of Plerixafor as a single mobilizing
agent (NCT02212535). Our objective was to demonstrate
that SCD patients can be safely mobilized under well-
defined, controlled clinical and biological conditions.
Moreover, this clinical trial provided an estimate of the
overall recovery of CD34+ HSPC after apheresis in adult
Plerixafor-mobilized SCD patients – thus providing a basis
for the wider use of this protocol for gene-addition or
genome-editing approaches.
780
Methods
Study design and human samples
This open-label phase I/II trial (ClinicalTrials.gov number,
NCT02212535) was sponsored by the Assistance Publique
Hôpitaux de Paris. The protocol was reviewed and approved by
the French Drug Agency (Agence Nationale de Sécurité du
Médicament) and the local independent ethics committee (Comité de
Protection des Personnes Ile-de-France II, Paris, France). The trial was
performed in accordance with the Declaration of Helsinki. Adult
SCD patients and healthy donors (for control samples) provided
their written, informed consent (Online Supplementary Table S1).
The study was designed to demonstrate the safety and efficacy of
the mobilization and harvesting of peripheral HSPC following a
single injection of 0.24 mg/kg Plerixafor in adult SCD patients,
with the inclusion criteria described in the Online Supplementary
Methods.
Figure 1. Plerixafor is highly efficient at mobilizing
hematopoietic stem and progenitor cells from sick-
le cell disease patients. (A) Changes in white blood
cell (WBC) and (B) CD34+ hematopoietic stem/prog-
enitor cell (HSPC) counts over the 66 h following
Plerixafor administration in SCD Pler 1 (red
squares), SCD Pler 2 (blue circles) and SCD Pler 3
(green triangles). Arrows indicate time and duration
of apheresis. (C) Number of hematopoietic stem
cells (HSC, black bars) and multipotent progenitors
(MPP, dotted bars) per 1,000 CD34+ cells in sam-
ples of various origins. HD: healthy donor, BM: bone
marrow, SCD: sickle cell disease, Pler: Plerixafor,
Filg: Filgrastim.
haematologica | 2018; 103(5)

The patients were carefully monitored in the Intensive Care
Unit and a few hours before the procedure started, they were
given prophylaxis for vaso-occlusive crisis [hyperhydration 60
mL/kg/day of a 0.9% saline solution and oxygen therapy (2
L/min)], as recommended by French guidelines.21 These guidelines
recommend oxygen therapy to prevent sickling phenomena in
particularly stressful conditions. As mobilization can be consid-
ered stressful, we decided to treat all patients with oxygen thera-
py. Baseline O2 saturation levels were normal and remained stable
during the procedure. The patients received a subcutaneous injec-
tion of 0.24 mg/kg of Plerixafor. We then monitored whole blood
HbS level, plasma bilirubin, conjugated bilirubin and lactate dehy-
drogenase levels, and white blood cell, neutrophil and monocyte
counts for 30 days after the Plerixafor injection. Plerixafor-mobi-
lized HSPC in peripheral blood were collected by using a COBE®
Spectra Optia apheresis system (Terumo BCT, Lakewood, CO,
USA) with modifications (Online Supplementary Material and
Methods). The apheresis lasted from 4 to 6 h. Product volumes and
total blood volumes are indicated in Table 2. Egress of the CD34+
cells was monitored hourly for at least 24 h after the Plerixafor
injection: total CD34+/kg collected and CD34+ collection efficiency
(CE1) [total CD34+ collected/[L processed x (CD34+ pre + CD34+
post)/2] were evaluated following apheresis (Table 2).
Flow cytometry analysis
Cells were stained with specific antibodies (Online
Supplementary Table S2) and analyzed on a BD FACSCanto™ II sys-
tem, with gating on viable, 7AAD-negative cells. The data were
processed using FlowJo software (version 10.2, Treestar, Ashland,
OR, USA).
RNA-Seq extraction
Total RNA was extracted from 0.1-1x106 CD34+ cells using an
RNeasy Micro kit (QIAGEN). RNA-Seq libraries were prepared
from ~10 ng of total RNA, using the Ovation Human FFPE RNA-Seq
Multiplex System kit (Nugen) after DNase treatment (ArcticZymes)
and >40 million paired-end reads/sample. Samples were generated
on a HiSeq 2500 instrument (Illumina). RNA-Seq analysis was per-
formed as described in the Online Supplementary Methods.
Transplantation into non-obese diabetic severe
combined immunodeficiency gamma mice
The non-obese diabetic severe combined immunodeficiency
gamma (NSG) mice (NOD.CgPrkdcscid Il2rgtm1Wj/SzJ, Charles
Table 2. Characteristics of apheresis and CD34+ immunoselection.
Apheresis
Total blood volumes in liters (liters
processed by the apheresis device)
Product volumes (mL)
Hematocrit value (%)*
CE1**
Total CD34+ cells collected by apheresis x106
Total CD34+ cells collected by apheresis x106/kg
of body weight
CD34+ cell product after immunoselection
Recovery
Purity post-selection
*Normal value: 2-3%; **CE1 = total CD34+ collected/ [L processed X (CD34+pre+CD34+post)/2]
haematologica | 2018; 103(5)
Stem cell mobilization in sickle cell patients
River Laboratories) were housed in a pathogen-free facility. All
experiments were performed in compliance with the French
Ministry of Agriculture’s regulations on animal experiments, and
were approved by the regional Animal Care and Use Committee
(APAFIS#2101-2015090411495178 v4). Six- to 8-week-old mice
were conditioned with intraperitoneally injected busulfan (Sigma,
25 mg/kg body weight/day) 72 h, 48 h and 24 h before transplan-
tation. CD34+ cells (300,000 cells/mouse) from Filgrastim-mobi-
lized healthy donors or Plerixafor-mobilized cells from the three
SCD patients were transplanted into the NSG mice via retro-
orbital sinus injection. Neomycin was added to the animals’ acid-
ified drinking water. Engraftment was analyzed as described in the
Online Supplementary Methods.
Statistical analysis
Statistical tests are indicated in the Figure legends. The thresh-
old for statistical significance was set at P<0.05.
Results
Patients’ characteristics
Four SCD homozygous patients followed at Necker-
Enfants Malades hospital and meeting our inclusion crite-
ria were enrolled between May 2015 and January 2017.
The first one was excluded from the study during the
enrollment stage because of elevated granulocyte counts
exceeding the limits established in the protocol (<10x109
granulocytes/L). The other three patients (P1 - P3) had
been monitored in a reference center after the diagnosis of
SCD within the first 4 years of life. All suffered from
severe SCD, with a history of acute chest syndrome and
more than two vaso-occlusive crises per year requiring
hospitalization. P1 had undergone cholecystectomy and
tonsillectomy. P2 had papillary necrosis and osteonecrosis
of both femoral heads. P3 had undergone cholecystecto-
my, and had chronic asthma and a history of osteomyelitis
events. P1 and P2 were transfused monthly because years
of hydroxyurea treatment had proven to be ineffective.
Hydroxyurea treatment was stopped in P3 3 months
before mobilization. P3 was then transfused monthly until
mobilization. In view of iron overload caused by the
transfusions (Table 1), treatment with deferasirox (P1) and
deferiprone (P2) was ongoing.
SCD Pler 1 SCD Pler 2 SCD Pler 3
3.27 (15.8) 4 (21) 4 (17.9)
278 382 322
4.8 5.8 8.2
0.24 0.30 0.29
354 412 292
4.6 (77) 5.8 (71) 4.5 (65)
82% 92% 31%
95% 94.7% 79.5%
781
 C. Lagresle-Peyrou et al.

Absence of Plerixafor-related toxicity numbers of circulating hematopoietic progenitors,22 were

Prior to Plerixafor injection, patients underwent several
erythrocytoapheresis sessions in order to decrease their
HbS levels to below 30% (Table 1). Each patient received
a single, subcutaneous injection of Plerixafor (0.24 mg/kg).
The HbS values remained low until day +1 and increased
to pre-treatment levels thereafter (Table 1). No adverse
effects were observed, other than moderate hypokalemia
related to the anticoagulant citrate infusion during the
apheresis; this was corrected by a 2-day course of potassi-
um chloride (Table 1). The white blood cell and neutrophil
counts rose to 30.4 ± 2.8x109/μL and 17.8 ± 3.5x109/L,
respectively, during the first 3 h, remained stable and then
(within 24 to 36 h of Plerixafor injection) returned to pre-
treatment values (Figure 1A and Online Supplementary
Figure S1). Bilirubin, conjugated bilirubin and lactate dehy-
drogenase values and monocyte counts remained stable
up to day +30. Serum levels of inflammatory cytokines,
including interleukin-8, which is increased in SCD
patients during acute crises and associated with higher
comparable to those of healthy controls (data not shown).
Vital parameters and blood O2 saturation remained stable
and normal during all the procedure. Monitoring of the
patients between discharge and the end of the follow-up
period (6 months after treatment) was uneventful.
Efficacy of Plerixafor in mobilizing hematopoietic stem
and progenitor cells
Baseline CD34+ cell counts were 7, 10 and 10/μL in P1,
P2 and P3, respectively. The patients exhibited a very fast,
intense increase in peripheral blood CD34+ cell count,
exceeding 80 CD34+/μL 3 h after the Plerixafor injection
(Figure 1B). Levels greater than 50 CD34+/μL were main-
tained for 6 h, then decreased and returned to normal pre-
treatment values (Figure 1B, and data not shown for day 30
and day 60). Apheresis was performed with the technical
adjustments described in the Online Supplementary
Methods. The quantities of CD34+ cells harvested by
apheresis (4.6 x 106, 5.8x106 and 4.5x106/kg body weight

A B

Figure 2. Analysis of the transcriptomic profiles of hematopoietic stem and progenitor cells from different sources (A) Hierarchical clustering analysis of HD BM,
SCD BM, SCD Plerixafor-mobilized (Pler), HD Plerixafor-mobilized and HD Filgrastim-mobilized (Filg) HSPC (cluster method: average; distance: correlation). The color
of the sample name indicates the classification. (B) Gene ontology analysis of differentially expressed genes. The most enriched biological process categories are
shown on the y-axis. The x-axis shows sample comparisons, as defined in Table 3. The orange and green color gradients correspond to the statistical significance of
the enrichment [expressed as –log10 (qvalue)] in up- and downregulated genes, respectively. The first color bar at the top indicates comparisons between HSPC from
different types of source (dark red) or the same type of source (light red). The second color bar at the top indicates comparisons between HSPC from different types
of donor (dark blue) or the same type of donor (light blue). (C) Heat map of genes involved in HSC and progenitor biology. A proportion of the HSC markers were highly
expressed in SCD Plerixafor-mobilized HSPC compared with the other samples. The row Z-score is plotted on a red-blue color scale, where red indicates high expres-
sion and blue indicates low expression. The color bar at the top indicates the sample classification. HD: healthy donor; BM: bone marrow; HSPC: hematopoietic stem
and progenitor cells; HSC: hematopoietic stem cells; SCD: sickle cell disease; Pler: Plerixafor; Filg: Filgrastim.

782 haematologica | 2018; 103(5)

 Stem cell mobilization in sickle cell patients

for P1, P2 and P3, respectively) were high in all three
patients despite a limited collection efficiency (Table 2).
This enabled the cryopreservation of 3x106 unselected
CD34+ cells/kg as a back-up for the upcoming gene thera-
py trial (used in the case of absence of engraftment) and
the immunoselection and further analyses of mobilized
CD34+ cells as detailed below. Following CD34+ selection,
CD34+ cell purity was in the normal range (Table 2).
CD34+ recovery was high in P1 and P2 (82% and 92%,
respectively) and lower in P3 (31%).
Characterization of mobilized hematopoietic stem
and progenitor cells
To determine the hematopoietic differentiation capacity
and self-renewal potential of the mobilized CD34+ cells,
we performed a number of phenotypic, transcriptomic
and functional analyses.
Hematopoietic stem cells (HSC) and their immediate
progeny (multipotent progenitors) within the CD34+ sub-
set are negative for lineage, CD38, and CD45RA markers
and positive for CD13323-25 and can, therefore, be detected
using flow cytometry. We compared the numbers of HSC
and multipotent progenitors among mobilized SCD
CD34+ cells with the values for (i) the BM of healthy
donors and SCD patients, and (ii) samples from healthy
donors mobilized with either Filgrastim or Plerixafor
(Online Supplementary Table S1, Figure 1C and Online
Supplementary Figure S2). We estimated the number of
HSC per 1000 CD34+ cells to be >25 in Plerixafor-mobi-
lized SCD samples and <5 in all other samples, suggesting
that Plerixafor mobilizes HSC with an unexpectedly high
efficacy in SCD patients.
The frequency of erythroid and granulocyte/monocyte
colony-forming cells was similar in Plerixafor-mobilized
and BM SCD samples and in the range of that observed in
Filgastrim-mobilized HSPC (Online Supplementary Figure
S3A and data not shown). Upon erythroid differentiation,
Plerixafor-mobilized as well as BM SCD HSPC gave rise to
>90% of mature GYPA+CD36lowCD71low enucleated red
blood cells (Online Supplementary Figure S3B-D).
The transcriptome of highly purified CD34+ HSPC from
the different sources was analyzed using RNA-Seq.
Unsupervised gene expression analysis showed that the
samples clustered into two major groups, based on cell
origin: the “BM” group encompassed BM samples from
SCD patients and healthy donors, whereas the “mobi-
lized” group encompassed Plerixafor-mobilized HSPC
from SCD patients and healthy donors and Filgrastim-
mobilized HSPC from healthy donors (Figure 2A).
Next, we identified differentially expressed genes
among the different populations (false discovery rate
<0.05, Table 3). In a comparison of samples from SCD
patients and healthy donors, we observed that the genes
upregulated in SCD samples are involved in inflammatory
and immune responses (e.g. defense response to other
organisms, type I interferon signaling pathway, cytokine

Figure 3. Plerixafor-mobilized CD34+ cells from sickle cell disease patients engraft to the same degree as Filgrastim-mobilized CD34+ cells from healthy donors
in NSG mice. NSG mice were sacrificed 3 to 4 months after the injection of SCD (SCD Plerixafor, n=3) or HD (HD Filgrastim, n=2) CD34+ cells. (A) Bone marrow cells
and (B) splenocytes were isolated, stained and analyzed by flow cytometry. The chimerism (defined as % human CD45+cells/total CD45+cells) and the numbers of
human B lymphocytes (CD19+IgM+), granulocytes (CD11b+CD15+), and monocytes (CD11b+CD14+) were evaluated in each group of mice (red circles and red triangles
SCD Pler1; blue circles and blue triangles SCD Pler2; green circles and green triangles SCD Pler3; the two HD Filg control are represented by gray squares/gray dia-
mond and black squares/black diamonds, respectively). Each dot represents an individual mouse. HD: healthy donor; SCD: sickle cell disease; Pler: Plerixafor; Filg:
Filgrastim.

haematologica | 2018; 103(5) 783

 C. Lagresle-Peyrou et al.

production, leukocyte migration, phagocytosis and adap- Table 3. Number of differentially expressed (up- or downregulated)
tive immune response) (Figure 2B). Major differences in genes in HSPC from different sources (false discovery rate < 0.05).
gene expression (>900 differentially expressed genes) Comparison Differentially Upregulated Downregulated
were observed when comparing mobilized and BM sam- expressed
ples (Table 3). The genes downregulated in mobilized ver- SCD BM vs. HD BM 134 52 82
sus BM HSPC are involved in cell cycle-related processes
SCD Pler vs. HD Pler 165 133 32
(e.g. DNA replication, chromosome segregation, and
nuclear division) – confirming that mobilized samples SCD Pler vs. HD Filg 265 188 77
contain more quiescent cells, presumably HSC, than pro- SCD Pler vs. HD BM 1685 761 924
genitors (Figure 2B, Online Supplementary Figure S4A and SCD Pler vs. SCD BM 1544 753 791
Online Supplementary Table S2). Interestingly, mobilized HD Pler vs. HD BM 923 315 608
populations poorly expressed genes that are typical of
committed hematopoietic progenitors, relative to BM HD Filg vs. HD BM 1893 789 1104
samples (Figure 2C, Online Supplementary Figure S4A and HD Pler vs. HD Filg 47 20 27
Online Supplementary Table S2). Conversely, a large propor- HD Filg vs. SCD BM 1683 800 883
tion of the genes involved in HSC biology were strongly HD Pler vs. SCD BM 1024 398 626
expressed in mobilized HSPC (Figure 2C, Online SCD: sickle cell disease. HD: healthy donor. BM: bone marrow. Pler: Plerixafor. Filg:
Supplementary Figure S4A and Online Supplementary Table Filgrastim.
S2). Importantly, some HSC markers (e.g. THY1, HLF and
FLT3) were more strongly expressed in Plerixafor-mobi-
lized SCD samples – confirming the latter’s high HSC con-
tent – than in BM samples and Filgrastim- or Plerixafor- dure.26 Because of the risks incurred by the patients and
mobilized HSPC from healthy donors (Figure 2C). the absence of a direct benefit, the trial was restricted to
Fewer than 300 genes were differentially expressed patients with <10x109 granulocytes/L, knowing their role
between Filgrastim- or Plerixafor-mobilized samples in vaso-occlusive crises.
(Table 3). Of note, genes encoding transcriptional regula- We decided to discontinue hydroxyurea treatment 3
tors and surface markers of plasmacytoid dendritic cell months before the mobilization in P3 and to submit all the
progenitors are upregulated in Plerixafor-mobilized sam- patients to monthly transfusions. The rationale for this
ples compared to Filgrastim-mobilized samples (Online decision was based on the following observations: (i)
Supplementary Figure S4B). FACS analyses showed the hydroxyurea has no beneficial role in CD34+ cell mobiliza-
appearance of a CD133-CD34dim cell population preferen- tion in thalassemic patients;27 (ii) hydroxyurea withdrawal
tially in Plerixafor-mobilized samples (Online is associated with an increase in the number of circulating
Supplementary Figure S2 and data not shown). This popula- CD34+ cells in SCD patients;28 and (iii) in various clinical
tion might contain plasmacytoid dendritic cell progenitors settings hydroxyurea has been associated with myelosup-
that are mobilized by Plerixafor, as recently described.16 pression29,30 suggesting BM toxicity and potential impair-
Proof of stemness was further confirmed by transplan- ment of HSC.
tation into conditioned, immunodeficient mice. Human In order to optimize the safety of the mobilization pro-
chimerism in the BM and spleen was similar in recipients cedure in SCD patients, we reduced HbS levels to below
of Filgrastim-mobilized CD34+ cells from healthy donorss 30% via erythrocyte exchanges, and we closely moni-
and Plerixafor-mobilized CD34+ cells from SCD patients tored white blood cell counts and serum levels of inflam-
(Figure 3A,B). Similar counts of lymphoid and myeloid matory cytokines. Furthermore, the use of Plerixafor
subsets were detected in both groups (Figure 3A,B). The avoided the adverse events associated with Filgrastim:
numbers of human CD34+ cells, HSC and multipotent vascular events and splenic rupture after the administra-
progenitors in bone marrow were comparable between tion of this latter have been extensively reported in clinical
the two groups (Online Supplementary Figure S5). After sec- populations and even in healthy stem cell donors.31
ondary transplantation, all recipients of Plerixafor-mobi- Although these events are usually rare, their frequency
lized SCD samples and Filgrastim-mobilized CD34+ cells may be higher in the presence of other vascular risk fac-
from healthy donors displayed engraftment - further tors (such as SCD).31 Thus the benefit/risk ratio of using a
demonstrating the presence of true HSC in HSPC mobi- hematopoietic growth factor such as Filgrastim (especially
lized with Plerixafor in SCD patients (data not shown). at the high doses required in patients without a malignant
blood disease) appears to be unacceptably low. Hence, we
gave our three patients Plerixafor at the standard dose. No
Discussion adverse events occurred, serum levels of inflammatory
cytokines were in the normal ranges (data not shown) and
Successful transplantation of autologous gene-corrected the white blood cell counts did not exceed 40x109/L (i.e. a
cells primarily depends on the collection and effective value often reported in the literature, and far from the 50
genetic modification of a sufficient number of true stem to 75x109/L often observed in healthy donors after 5 days
cells. Thus, poor harvesting of BM or mobilized peripheral of Filgrastim treatment).32
stem cells limits the success of this procedure. To over- The rapid mobilization with Plerixafor alone (compared
come the need for two or more BM aspirates, we initiated with Filgrastim) is also an important advantage. The cur-
a phase I/II clinical trial with the objective of establishing rent guidelines on mobilization in patients with a malig-
whether Plerixafor-induced stem cell mobilization in SCD nant disease recommend initiating apheresis 11 h after
patients can avoid the increased risk of vaso-occlusive Plerixafor administration;33 this contrasts with the 5 to 7
crises observed with Filgrastim mobilization and of vali- days required for collection after Filgrastim administra-
dating the efficiency of HSPC harvesting with this proce- tion. Moreover, rapid, transient stem cell mobilization by

784 haematologica | 2018; 103(5)

 Stem cell mobilization in sickle cell patients

Filgrastim+Plerixafor (especially in very poor mobilizers)
has been observed by various groups (including ours); the
rapid decrease in peripheral blood stem cells may cause
collection to fail when apheresis is initiated according to
conventional guidelines.34 We monitored the egress of
CD34+ cells into the blood every hour after Plerixafor
injection. The time course of CD34+ mobilization was
remarkably similar in our three patients. Peak counts of
over 80 CD34+/μL were achieved as early as 2 to 3 h after
Plerixafor administration; this confirms that HSPC can be
harvested immediately in SCD patients. The decrease in
CD34+ cell count observed after 6 h might be the conse-
quence of a short-term mobilization of HSC by Plerixafor
and therefore to the reduced egress of CD34+ cells from
BM combined with their return to the BM. Another possi-
bility is that the drop in CD34+ cells is due to the leuka-
pheresis procedure. Pantin et al. analyzed the kinetics of
CD34+ counts in healthy donors treated with Plerixafor
and not subjected to the apheresis procedure.35 In their
work, CD34+ cell counts peaked at 6 h at lower values
than the ones observed in our study and start to decrease
10 h after Plerixafor administration, with differences in
kinetics and range of mobilization probably related to the
clinical conditions of healthy donors versus SCD patients.
Overall, the study by Pantin et al. suggests that leuka-
pheresis per se does not cause a decrease in CD34+ cell
counts. Additionally, the limited collection efficiency
(≤30% of the circulating CD34+ cells) (Table 2) does not
support the hypothesis that the drop is due to the leuka-
pheresis procedure. Close monitoring of peripheral blood
CD34+ cell counts is therefore a crucial point for efficient
apheresis in SCD patients mobilized with Plerixafor.
The leukapheresis product contained significantly more
HSC than the other stem cell sources used as controls, i.e.
8- to 10-fold more than in BM from healthy donors or
SCD patients and in Filgrastim- or Plerixafor-mobilized
cells from healthy donors. Accordingly, HSPC from the
patients’ Plerixafor-mobilized samples showed elevated
transcription of several HSC-associated genes. We do not
have a formal explanation for this result; we can only
hypothesize that sickling cycles damage the BM stroma
and favor the mobilization of HSC.
Genes involved in inflammatory and immune responses
were upregulated in Plerixafor-mobilized SCD samples.
The inflammation-related characteristics of HSC and their
environment constitute a major obstacle to both allogene-
ic and autologous transplantation. Significant pathological
changes in hematopoiesis have been described by Weisser
et al.36 in a setting of murine and human chronic granulo-
matous disease, an inherited disease characterized by
chronic, sterile, granulomatous inflammation). In mice
and in humans with chronic granulomatous disease, BM
and Filgrastim-mobilized grafts contain a low proportion
of HSC; in transplanted mice, this feature is associated
with low reconstitution potential.36 Hence, we transplant-
ed mobilized CD34+ cells from SCD patients into
immunodeficient NSG mice, in order to establish whether
an inflammatory expression profile interfered with
engraftment and self-renewal capacity. In fact, the cells
engrafted as well as Filgrastim-mobilized samples in both
primary and secondary transplantation, demonstrating
that Plerixafor-mobilized SCD CD34+ cells contain true
stem cells that are able to reconstitute human
hematopoiesis as well as their Filgrastim-mobilized coun-
terparts. Moreover, it is possible to effectively correct
Plerixafor-mobilized SCD HSC by gene addition (manu-
script in preparation). Taken as a whole, our results show
that the inflammatory characteristics of HSC do not
impair self-renewal and engraftment.
In conclusion, the present results show that CD34+ cells
can be safely mobilized with Plerixafor in SCD patients
under well-defined clinical conditions, including a 3-
month interruption of hydroxyurea treatment, monthly
transfusions and red blood cell exchanges. The proportion
of true stem cells in the Plerixafor-mobilized CD34+ popu-
lation was significantly higher than the proportion from
any other source, although their egress must be monitored
carefully. After harvesting under specific conditions, the
cells can be successfully immunoselected. In the case that
the optimal dose of CD34+ cells required for gene therapy
(6-9x106/kg) is not reached after one apheresis and in the
absence of adverse events, a second mobilization by
Plerixafor, 24 h after the first one, will be considered. In
our small cohort, the high numbers of cells enabled us to
extend our gene-addition therapy project without having
to perform several low-yield BM harvesting steps.
Furthermore, the results removed the obstacle of collect-
ing an appropriate graft for genome-editing purposes.
Lastly, our study emphasizes the importance of consider-
ing the specific characteristics of a diseased BM; finding a
way to put the patient’s hematopoietic system into a
steady-state condition may circumvent a lack of true stem
cells in the harvested product or poor engraftment of
genetically-modified cells.
Acknowledgments
The authors would like to thank Jean-Marc Luby for excellent,
dedicated technical assistance, Valérie Jolaine, Michaela
Semeraro for the logistics of the clinical trial, Michaela Semeraro
for care of patients, and Christine Bole and Olivier Alibeu for the
RNA sequencing. We also thank Frédéric Galacteros, Pablo
Bartolucci and Susanne Matthes-Martin for revision of the clini-
cal protocol.
Funding
This work was supported by state funding from the Agence
Nationale de la Recherche as part of the Investissements d’Avenir
program (ANR-10-IAHU-01 and ANR-16-CE18-0004), the
French National Institute of Health and Medical Research
(INSERM), and Assistance Publique-Hôpitaux de Paris (AP-
HP). This work was also supported by a grant from the European
Research Council (ERC 2015-AdG, GENEFORCURE).

References
1. Ware RE, de Montalembert M, Tshilolo L,
Abboud MR. Sickle cell disease. Lancet.
2017;390(10091):311-323.
2. Walters MC, De Castro LM, Sullivan KM,
et al. Indications and results of HLA-identi-
cal sibling hematopoietic cell transplanta-
tion for sickle cell disease. Biol Blood
Marrow Transplant. 2016;22(2):207-211.
3. Gluckman E, Cappelli B, Bernaudin F, et al.
Sickle cell disease: an international survey
of results of HLA-identical sibling
hematopoietic stem cell transplantation.
Blood. 2017;129(11):1548-1556.
4. Ribeil JA, Hacein-Bey-Abina S, Payen E, et
al. Gene therapy in a patient with sickle

haematologica | 2018; 103(5) 785

 C. Lagresle-Peyrou et al.
cell disease. N Engl J Med. 2017;376(9):
848-855.
5. Cavazzana M, Ribeil JA, Lagresle-Peyrou
C, Andre-Schmutz I. Gene therapy with
hematopoietic stem cells: the diseased
bone marrow's point of view. Stem Cells
Dev. 2017;26(2):71-76.
6. Wang TF, Chen SH, Yang SH, Su YC, Chu
SC, Li DK. Poor harvest of peripheral
blood stem cell in donors with microcytic
red blood cells. Transfusion. 2013;53
(1):91-95.
7. Constantinou VC, Bouinta A, Karponi G, et
al. Poor stem cell harvest may not always
be related to poor mobilization: lessons
gained from a mobilization study in
patients with beta-thalassemia major.
Transfusion. 2017;57(4):1031-1039.
8. Abboud M, Laver J, Blau CA.
Granulocytosis causing sickle-cell crisis.
Lancet. 1998;351(9107):959.
9. Adler BK, Salzman DE, Carabasi MH,
Vaughan WP, Reddy VV, Prchal JT. Fatal
sickle cell crisis after granulocyte colony-
stimulating factor administration. Blood.
2001;97(10):3313-3314.
10. Grigg AP. Granulocyte colony-stimulating
factor-induced sickle cell crisis and multior-
gan dysfunction in a patient with com-
pound heterozygous sickle cell/beta+ tha-
lassemia. Blood. 2001;97(12):3998-3999.
11. Onitilo AA, Lazarchick J, Brunson CY, Frei-
Lahr D, Stuart RK. Autologous bone mar-
row transplant in a patient with sickle cell
disease and diffuse large B-cell lymphoma.
Transplant Proc. 2003;35(8):3089-3092.
12. Kamble RT, Tin UC, Carrum G. Successful
mobilization and transplantation of filgras-
tim mobilized hematopoietic stem cells in
sickle cell-hemoglobin C disease. Bone
Marrow Transplant. 2006;37(11):1065-
1066.
13. Rosenbaum C, Peace D, Rich E, Van Besien
K. Granulocyte colony-stimulating factor-
based stem cell mobilization in patients
with sickle cell disease. Biol Blood Marrow
Transplant. 2008;14(6):719-723.
14. Tormey CA, Snyder EL, Cooper DL.
Mobilization, collection, and transplanta-
tion of peripheral blood hematopoietic pro-
genitor cells in a patient with multiple
myeloma and hemoglobin SC disease.
Transfusion. 2008;48(9):1930-1933.
15. Liles WC, Broxmeyer HE, Rodger E, et al.
Mobilization of hematopoietic progenitor
cells in healthy volunteers by AMD3100, a
786
CXCR4 antagonist. Blood. 2003;102(8):
2728-2730.
16. Schroeder MA, Rettig MP, Lopez S, et al.
Mobilization of allogeneic peripheral blood
stem cell donors with intravenous plerix-
afor mobilizes a unique graft. Blood.
2017;129(19):2680-2692.
17. Broxmeyer HE, Orschell CM, Clapp DW, et
al. Rapid mobilization of murine and
human hematopoietic stem and progenitor
cells with AMD3100, a CXCR4 antagonist.
J Exp Med. 2005;201(8):1307-1318.
18. Devine SM, Flomenberg N, Vesole DH, et
al. Rapid mobilization of CD34+ cells fol-
lowing administration of the CXCR4
antagonist AMD3100 to patients with mul-
tiple myeloma and non-Hodgkin's lym-
phoma. J Clin Oncol. 2004;22(6):1095-
1102.
19. Karponi G, Psatha N, Lederer CW, et al.
Plerixafor+G-CSF-mobilized CD34+ cells
represent an optimal graft source for tha-
lassemia gene therapy. Blood. 2015;126(5):
616-619.
20. Ferrari G, Cavazzana M, Mavilio F. Gene
therapy approaches to hemoglobinopathies.
Hematol Oncol Clin North Am. 2017;31(5):
835-852.
21. Lionnet F, Arlet JB, Bartolucci P, et al.
[Guidelines for management of adult sickle
cell disease]. Rev Med Interne. 2009;30
(Suppl 3):S162-223.
22. Lamming CE, Augustin L, Blackstad M,
Lund TC, Hebbel RP, Verfaillie CM.
Spontaneous circulation of myeloid-lym-
phoid-initiating cells and SCID-repopulat-
ing cells in sickle cell crisis. J Clin Invest.
2003;111(6):811-819.
23. Doulatov S, Notta F, Eppert K, Nguyen LT,
Ohashi PS, Dick JE. Revised map of the
human progenitor hierarchy shows the ori-
gin of macrophages and dendritic cells in
early lymphoid development. Nat
Immunol. 2010;11(7):585-593.
24. Six EM, Bonhomme D, Monteiro M, et al.
A human postnatal lymphoid progenitor
capable of circulating and seeding the thy-
mus. J Exp Med. 2007;204(13):3085-3093.
25. Handgretinger R, Kuci S. CD133-positive
hematopoietic stem cells: from biology to
medicine. Adv Exp Med Biol. 2013;777:99-
111.
26. Fitzhugh CD, Hsieh MM, Bolan CD, Saenz
C, Tisdale JF. Granulocyte colony-stimulat-
ing factor (G-CSF) administration in indi-
viduals with sickle cell disease: time for a
moratorium? Cytotherapy. 2009;11(4):464-
471.
27. Yannaki E, Papayannopoulou T, Jonlin E, et
al. Hematopoietic stem cell mobilization
for gene therapy of adult patients with
severe beta-thalassemia: results of clinical
trials using G-CSF or plerixafor in splenec-
tomized and nonsplenectomized subjects.
Mol Ther. 2012;20(1):230-238.
28. Richard RE, Siritanaratkul N, Jonlin E,
Skarpidi E, Heimfeld S, Blau CA. Collection
of blood stem cells from patients with sick-
le cell anemia. Blood Cells Mol Dis.
2005;35(3):384-388.
29. Charache S, Terrin ML, Moore RD, et al.
Effect of hydroxyurea on the frequency of
painful crises in sickle cell anemia.
Investigators of the Multicenter Study of
Hydroxyurea in Sickle Cell Anemia. N Engl
J Med. 1995;332(20):1317-1322.
30. Alvarez-Larran A, Perez-Encinas M, Ferrer-
Marin F, et al. Risk of thrombosis according
to need of phlebotomies in patients with
polycythemia vera treated with hydrox-
yurea. Haematologica. 2017;102(1):103-
109.
31. Halter J, Kodera Y, Ispizua AU, et al. Severe
events in donors after allogeneic
hematopoietic stem cell donation.
Haematologica. 2009;94(1):94-101.
32. Lack NA, Green B, Dale DC, et al. A phar-
macokinetic-pharmacodynamic model for
the mobilization of CD34+ hematopoietic
progenitor cells by AMD3100. Clin
Pharmacol Ther. 2005;77(5):427-436.
33. To LB, Levesque JP, Herbert KE. How I treat
patients who mobilize hematopoietic stem
cells poorly. Blood. 2011;118(17):4530-
4540.
34. Lefrere F, Mauge L, Rea D, et al. A specific
time course for mobilization of peripheral
blood CD34+ cells after plerixafor injection
in very poor mobilizer patients: impact on
the timing of the apheresis procedure.
Transfusion. 2013;53(3):564-569.
35. Pantin J, Purev E, Tian X, et al. Effect of
high-dose plerixafor on CD34(+) cell mobi-
lization in healthy stem cell donors: results
of a randomized crossover trial.
Haematologica. 2017;102(3):600-609.
36. Weisser M, Demel UM, Stein S, et al.
Hyperinflammation in patients with chron-
ic granulomatous disease leads to impair-
ment of hematopoietic stem cell functions.
J Allergy Clin Immunol. 2016;138(1):219-
228.e219.
haematologica | 2018; 103(5)
 Red Cell Biology & its Disorders ARTICLE

RON kinase inhibition reduces renal

endothelial injury in sickle cell disease mice Ferrata Storti
Foundation

Alfia Khaibullina,1 Elena A. Adjei,1,2 Nowah Afangbedji,1 Andrey Ivanov,1

Namita Kumari,1 Luis E.F. Almeida,3 Zenaide M.N. Quezado,3 Sergei Nekhai1,4,5
and Marina Jerebtsova5

1
Center for Sickle Cell Disease, College of Medicine, Howard University, Washington, DC;
2
Departments of Genetics and Human Genetics, College of Medicine, Howard University,
Washington, DC; 3Department of Perioperative Medicine, NIH Clinical Center, National
Institutes of Health, Bethesda, MD; 4Department of Medicine, College of Medicine,
Howard University, Washington, DC and 5Department of Microbiology, College of Haematologica 2018
Medicine, Howard University, Washington, DC, USA Volume 103(5):787-798

ABSTRACT

S
ickle cell disease patients are at increased risk of developing a chron-
ic kidney disease. Endothelial dysfunction and inflammation asso-
ciated with hemolysis lead to vasculopathy and contribute to the
development of renal disease. Here we used a Townes sickle cell disease
mouse model to examine renal endothelial injury. Renal disease in
Townes mice was associated with glomerular hypertrophy, capillary
dilation and congestion, and significant endothelial injury. We also
detected substantial renal macrophage infiltration, and accumulation of
macrophage stimulating protein 1 in glomerular capillary. Treatment of
human cultured macrophages with hemin or red blood cell lysates sig-
nificantly increased expression of macrophage membrane-associated
protease that might cleave and activate circulating macrophage stimulat-
ing protein 1 precursor. Macrophage stimulating protein 1 binds to and
activates RON kinase, a cell surface receptor tyrosine kinase. In cultured
human renal glomerular endothelial cells, macrophage stimulating pro- Correspondence:
tein 1 induced RON downstream signaling, resulting in increased phos- marina.jerebtsova@howard.edu
phorylation of ERK and AKT kinases, expression of Von Willebrand fac-
tor, increased cell motility, and re-organization of F-actin. Specificity of
macrophage stimulating protein 1 function was confirmed by treatment
Received: September 15, 2017.
with RON kinase inhibitor BMS-777607 that significantly reduced
downstream signaling. Moreover, treatment of sickle cell mice with Accepted: March 1, 2018.
BMS-777607 significantly reduced glomerular hypertrophy, capillary Pre-published: March 8, 2018.
dilation and congestion, and endothelial injury. Taken together, our find-
ings demonstrated that RON kinase is involved in the induction of renal
endothelial injury in sickle cell mice. Inhibition of RON kinase activation doi:10.3324/haematol.2017.180992
may provide a novel approach for prevention of the development of
renal disease in sickle cell disease. Check the online version for the most updated
information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/787

Introduction
Sickle cell disease (SCD) is the most commonly inherited hematologic disorder
caused by a single nucleotide mutation in the β-globin gene (HBB) resulting in HbS
hemoglobin. HbS polymerization leads to sickling and hemolysis of red blood cells
(RBCs), vaso-occlusion and organ damage. SCD patients are at increased risk of
developing chronic kidney disease (CKD).1,2 Renal involvement in SCD can be pres-
ent in childhood, as evidenced in 16-28% of children with clinical manifestations
of proteinuria and microalbuminuria.3 Albuminuria and proteinuria are observed in
more than 50% of adult SCD patients, and renal failure is developed in about
30%.4,5 SCD-associated nephropathy is characterized by tubular dysfunction,
which is manifested by inability to concentrate urine, and consequent hypos-
thenuria and polyuria, and glomerular damage. Glomerular abnormalities are char-
acterized by glomerular hypertrophy, expansion of mesangium, thrombotic
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.

Haematologica | 2018; 103(5) 787

 A. Khaibullina et al.
microangiopathy, focal segmental glomerulosclerosis
(FSGS), membranoproliferative glomerulonephritis, and
albuminuria.5-8 Two major disease mechanisms of chronic
kidney disease in SCD have been proposed: 1) hemolysis-
endothelial dysfunction leading to vasculopathy; and 2)
inflammation and hyper-viscosity leading to vaso-occlu-
sion.9 Intrarenal RBC hemolysis was suggested to be a trig-
ger for both mechanisms.
Spleen, the physiological site of RBC removal from the
circulation, is abnormal in SCD patients. The functional
asplenia is likely to increase the rates of intravascular
hemolysis. Sickling of RBCs and intra-organ hemolysis
stimulate infiltration by circulating monocytes and their dif-
ferentiation into macrophages. Endocytosis of RBC lysate
products affects macrophage phenotypes.10,11 Intravascular
RBC hemolysis also releases lysate products that impair
endothelial function leading to chronic vasculopathy.
Vascular endothelium and monocytes are activated in SCD
patients, and monocyte numbers are increased.12-14
Activated macrophages express matriptase-1 (MT-SP1)
which is one of the proteases that cleavages and activates
circulating macrophage stimulating protein 1 (MSP1).15,16
MSP1 was shown to accumulate in glomeruli in the rat
model of anti-Thy1 glomerular disease; its neutralization
by antibodies reduced serum creatinine and proteinuria,
and protected rats from glomerular injury.17 MSP1 is a plas-
ma protein secreted by liver and circulated as a single-
A E
C
F by significant glomeruli hypertrophy
D D). (E and F) Quantification of
788
chain, biologically inactive pro-MSP1. It is activated by
proteolytic cleavage of Arg483-Val484 bond by either
serum proteases or proteases expressed on the cell
surface.18 Pro-MSP1 diffuses into local tissues where it is
activated by proteolytic cleavage and plays a role in the
tissue injury or repair.18 Activated MSP1 binds to and acti-
vates a cell surface receptor tyrosine kinase, Recepteur
d'Origine Nantais (RON).19 We hypothesize that endocy-
tosis of RBC lysis products by kidney-infiltrating
macrophages stimulates expression of MT-SP1, which
then locally activates circulating MSP1. We further
hypothesize that MSP1 binds to RON tyrosine kinase
receptor and activates glomerular endothelium in SCD.
We test these hypotheses using a humanized mouse
model of SCD (Townes) which recapitulates several
hematologic manifestations of human SCD, including
renal vascular occlusion, as well as vascular, tubular and
glomerular changes.20,21 Here we showed that glomerular
disease in SCD mice was associated with endothelial
injury, increase in renal macrophage infiltration, and
glomerular MSP1 accumulation. In vitro, treatment of cul-
tured human macrophages with hemin, a breakdown
product of hemoglobin, or RBC lysate significantly
increased expression of MT-SP1. In cultured human renal
glomerular endothelial cells, MSP1 treatment induced
phosphorylation of RON downstream signaling ERK and
AKT kinases, increased expression of von Willebrand fac-
tor (vWF) and cell motility, and induced re-organization of
P=4x10-17
Figure 1. Renal disease in sickle cell
disease (SCD) mice is characterized
and capillary dilation. (A-D)
Representative pictures of hematoxylin
P=5x10-9
and eosin staining (H&E) (A and B),
and periodic acid–Schiff staining (PAS)
(C and D) of renal sections. Squares
show enlarged areas (B and D). Bar
sizes on microphotographs are
100 μm (A and C) and 40 μm. (B and
glomeruli size (E) and capillary size per
glomeruli cross section (F) is per-
formed using CellSens Standard soft-
ware. Five mice per group were used
for each staining. For quantification
graphs, means are shown. Each dot
represents a value obtained from one
glomerulus cross-section. Ctrl: control.
haematologica | 2018; 103(5)
 Renal accumulation of MSP1 in sickle cell disease

F-actin. RON kinase inhibitor (RONi, BMS-777607) signif-
icantly reduced RON signaling. Moreover, injections of
RONi in young SCD mice prevented the development of
glomerular disease, substantially reducing glomerular
hypertrophy, capillary dilation and congestion, and
endothelial injury.
Taken together, these data demonstrated that renal
glomerular accumulation of MSP1 and the activation of
RON kinase were involved in the induction of renal
endothelial injury in SCD mice. Inhibition of RON kinase
activation is a novel approach to prevent CKD develop-
ment in SCD.

E F
P=5x10-16 P=8x10-9

Figure 2. Renal disease in sickle cell disease (SCD) mice is associated with significant endothelial injury. (A-D) Representative pictures of von Willebrand factor
(vWF) (A and B) and CD34 (C and D) immunostaining (red) of renal sections. Squares show enlarged areas (B and D). Non-specific primary antibodies (Abs) were
used as a negative control. Bar sizes on the microphotographs are 100 μm (A and C) and 40 μm. (B and D). (E and F) Quantification of vWF (E) and CD34 (F) expres-
sion in glomeruli cross sections is performed using ImageJ Fiji version. Five mice per group were used for each staining. For quantification graphs, means are shown.
Each dot represents a value obtained from one glomerulus cross-section. Ctrl: control.

haematologica | 2018; 103(5) 789

 A. Khaibullina et al.

Methods System. Townes mice, here referred to as SCD mice, were

obtained from the Jackson Laboratory.
Mice and RONi injections Two-month old mice were injected subcutaneously with 10
The animal protocol was approved by the Institutional Animal mg/kg of body weight of RON inhibitor (RONi, BSM-777607,
Care and Use Committee at the Children's National Health Santa Cruz Scientific) in 2% DMSO daily for 14 consecutive days.

E F
P=4x10 -5
P=3x10-9

Figure 3. Glomerular macrophages infiltration and MSP1 accumulation is increased in sickle cell disease (SCD) mice. (A-D) Representative pictures of
macrophages (F4/80) (A and B, red) and macrophage stimulating protein 1 (MSP1) (C and D, brown) immunostaining of renal sections. Squares show enlarged
areas immunostaining (B and D). Non-specific primary antibodies (Abs) were used as a negative control. Bar sizes on the microphotographs are 100 µm (A and C)
and 40 µm. (B and D). (E and F) Quantification of F4/80 positive macrophages per glomeruli cross section (E) and MSP1 accumulation in the glomeruli (F) is per-
formed using ImageJ Fiji version software. Five mice per group were used for each staining. For quantification graphs, means are shown. Each dot represents a
value obtained from one glomerulus cross-section. Ctrl: control.

790 haematologica | 2018; 103(5)


Six SCD and 4 control mice were injected with RONi. Similar
groups were injected with 2% DMSO (vehicle).
Immunohistochemistry
Paraffin embedded tissue sections were used for immunostain-
ing with rat anti-mouse F4/80 (AbD Serotec), rabbit anti-vWF
(Dako), rat anti-mouse ICAM (BioLegend), rat anti-mouse CD34
(Cedarline), and mouse anti-MSP1 (R&D Systems). AEC and DAB
kits were obtained from Vector Laboratory. Images were acquired
with an Olympus 1x51 microscope with an Olympus DP 72 cam-
era. Quantification of positive staining was performed using
ImageJ Fiji version and CellSens Standard (Olympus) software.
Cell culture
Human glomerular endothelial cell line (HGEC) was generated
as described22 and maintained in DMEM media supplemented
with 10% fetal bovine serum (FBS) and antibiotics (all from
Thermo-Fisher). THP1 human monocytic cell line was purchased
from ATCC and maintained in RPMI-1640 media supplemented
with 10% FBS, antibiotics and 2 μM β-mercaptoethanol (Sigma-
Aldrich). THP-1 cells were differentiated into macrophages by
treatment with 10 nM phorbol 12-myristate 13 acetate (PMA,
Sigma-Aldrich) for 48 hours (h).
Treatment of THP-1 cells with hemin and RBC lysate
Hemin was obtained from Frontier Scientific. Blood samples
were collected from healthy control subjects. RBCs were isolated
by centrifugation, frozen at -80°C for 15 min, and then thawed for
lysis. Cellular debris was pelleted by centrifugation, and the
hemolysates were stored at -80°C. Differentiated THP-1 cells
were treated either with different concentrations of hemin or
hemolysates for 18 h.
Immunofluorescent staining and western blots
Rabbit MT-SP1 (Calbiochem) antibody was used for immunos-
taining and western blot (WB) of THP1 cells. HGEC were treated
with 1 μM of human recombinant MSP1 (R&D Systems) with or
without RON inhibitor (200 nM) for varying times. WB analysis
was performed with rabbit anti-p44/p42 MAPK (Erk1/2), rabbit
anti-phospho-p44/p42 Erk1/2, rabbit anti-pan-Akt, and rabbit
anti-phospho-Akt antibodies (all from Cell Signaling Technology).
Mouse anti-β-actin antibodies and phalloidin-FITC conjugate
were obtained from Sigma-Aldrich.
Wounding migration assay
The HGEC monolayers were wounded by strokes across the
diameter of the well with 2.5-mm-wide pipet and medium with
MSP1 (1 μM) was added. The width of the wound was visualized
with pictures taken at pre-treatment and 5 h post treatment. A
total of 20 random measurements were analyzed for each wound
at each time point.
Isolation of mouse renal glomeruli and glomeruli
permeability assay
Mouse renal glomeruli were isolated from control mice using a
sieving technique.23 Glomerular permeability was measured by a
determination of albumin permeability, as described previously24
with slight modification (Online Supplementary Methods).
Statistical analysis
Results were expressed as mean±Standard Deviation.
Differences between two groups were compared by unpaired
parametric t-test, between multiple groups by one-way ANOVA
(GraphPad software). Differences between groups were consid-
ered significant at P<0.05.
haematologica | 2018; 103(5)
Renal accumulation of MSP1 in sickle cell disease
Results
SCD mouse kidneys have increased macrophage
infiltration and MSP1 accumulation in the glomeruli
Kidneys were collected from 4-month old SCD and con-
trol mice (n=5 per group) and renal injury was evaluated.
SCD mice develop RBC sickling, anemia, leukocytosis,
behavioral changes, and multi-organ pathology character-
istic of SCD patients.21,25,26 Glomerular abnormalities in 4-
month old SCD mice were characterized by: glomerular
hypertrophy [Figure 1A, B, and E, hematoxylin and eosin
(H&E) staining]; expansion of mesangium [Figure 1C and
D, periodic acid Schiff (PAS) staining, and Online
Supplementary Figure S1, PCNA staining]; capillary dilation,
and thickening of capillary loops and glomerular basement
membrane (Figure 1C, D and F, PAS staining). Glomerular
capillaries, peritubular cortical capillaries, and capillaries in
the medulla and papilla were markedly congested, and
sickling of RBCs was observed especially in the medulla
and papilla (Online Supplementary Figure S2, H&E staining).
Capillary congestion and dilation together with hemolysis
might induce endothelial cell injury and secretion of von
Willebrand Factor (vWF) into the surrounding blood and
sub-endothelium.27,28 Indeed, glomerular capillary expres-
sion of vWF was significantly increased in SCD mice
(Figure 2A, B and E, immunostaining, red). Murine
glomerular capillary showed positive immunostaining for
endothelial marker CD3429 that was significantly increased
in SCD mice (Figure 2C, D and F, immunostaining, red).
Because CD34 is also expressed on hematopoietic progen-
itors cells which are increased in the circulation in SCD
mice30 these cells may also be a source of increased CD34
staining in the congested capillary. We also observed a sig-
nificant renal glomerular and interstitial infiltration of acti-
vated macrophages in SCD mice (Figure 3A, B and E,
F4/80 immunostaining, arrows, and Online Supplementary
Figure S3). Glomerular infiltrated macrophages were nega-
tive for the inducible nitric oxide synthase (iNOS), a mark-
er of M1 pro-inflammatory macrophage (Online
Supplementary Figure S4). MSP1 was accumulated in SCD
renal glomerular capillary (Figure 3C, D and F, immunohis-
tostaining, brown). High levels of MSP1 accumulation
were found in 46±8% of glomeruli in SCD mice. In con-
trast, less than 20% of glomeruli demonstrated low levels
of MSP1 accumulation in control mice. Together, these
findings show that renal disease in SCD mice is associated
with significant endothelial injury, macrophage infiltra-
tion, and accumulation of MSP1 in the glomerular capillar-
ies.
Hypoxia, hemin, and red blood cell lysate stimulate
expression of matriptase 1 in human macrophages
MSP1 is produced by the liver and secreted into circula-
tion where it is activated by proteolytic cleavage.18
Macrophage membrane-bound matriptase 1 (MT-SP1) is
one of the proteases which activate MSP1 protein.16,31 We
tested whether renal hypoxia and RBC hemolysis
increased expression of MT-SP1. Meta-analysis of the
NCBI Geo database demonstrated a 2-fold increase in MT-
SP1 expression in human monocyte-derived differentiated
macrophages compared to non-differentiated monocytes
(Figure 4A). Meta analysis also showed that MT-SP1
expression was further increased in macrophages that
were cultured under hypoxia (Figure 4B). Non-differentiat-
ed human pro-monocytic cell line (THP-1 cells) expressed
791
 A. Khaibullina et al.

a low level of MT-SP1, which was increased after THP-1
differentiation into macrophages by PMA treatment
(Figure 4C and D, compare lanes 1 and 2). Treatment of
THP-1-derived macrophages with hemin or RBC lysate
further increased MT-SP1 expression (Figure 4C-F).
Immunofluorescent staining of THP-1-derived
macrophages demonstrated membrane and cytoplasmic
distribution of MT-SP1 (Figure 4G, green) which was sig-
nificantly increased upon treatment with hemin (10 μM)
or RBC lysate (Figure 4G). Collectively, hemolysis prod-
ucts (tested in this study) and hypoxia (meta-analysis data)
demonstrated increased expression of MT-SP1 in human
macrophages. Higher levels of MT-SP1 expression might
induce local activation and accumulation of circulating
MSP1.
MSP1 induces endothelial cell motility and activated
downstream RON signaling of ERK and AKT
To assess the impact of MSP1 on the endothelium, we
examined its effect on cultured human glomerular
endothelial cell line (HGEC) recently generated in our lab-
oratory.22 Treatment with 1 μM MSP1 did not induce pro-
liferation of HGEC (Figure 5A). In contrast, MSP1 treat-
ment significantly increased motility of HGEC in a 2D
wound assay (Figure 5B and C). F-actin plays a central role
in the endothelial cell motility and permeability.32
Inflammatory mediators can alter F-actin formation and
distribution.32 In confluent HGEC, F-actin formed mostly a
cortical rim with few stress fibers (Figure 5D, green).
MSP1 stimulated re-organization of F-actin and formation
of stress fibers. (Figure 5D, green). Treatment cells with
specific RON kinase inhibitor (200 nM; RONi) inhibited
MSP1-induced formation of stress fibers and vWF expres-
sion (Figure 5D, F-actin and vWF staining). MSP1 treat-
ment also increased expression of vWF in HGEC (Figure
5D, red), and this increased expression was inhibited by
RONi (Figure 5D). Next, we examined activation of RON
kinase signaling in HGEC. Binding of MSP1 to RON leads
to phosphorylation of the downstream kinases, ERK and
AKT.33 Levels of ERK and AKT phosphorylation in HGEC

A B
P=0.008 P=0.002

C D

Figure 4. Hemin, red blood

cell (RBC) count and hypoxia
treatment increase expres-
sion of matriptase MT-SP1 in
human THP1-derived
macrophages. (A and B)
Meta-analysis of membrane
type serine protease 1 (MT-
SP1) expression in human
primary monocyte-derived
macrophages using Geo
E F database NCBI (A) Data Set
GDS3203 and (B) Data Set
GDS2036. For quantification
graphs, mean and Standard
Deviation (SD) are shown. (C
and E) Western blot of MT-
SP1 in human THP-1-derived
macrophages treated with
different concentrations of
hemin (C) or RBC lysate (D).
β-actin was used for normal-
ization. (D and F)
Quantification of MT-SP1 on
Western blot. For quantifica-
tion graphs, mean and SD
G
are shown. *P<0.05. Results
are representative of three
independent experiments.
(G) Immunostaining of MT-
SP1 in THP-1-derived
macrophages treated with
hemin (10 nM) or RBC lysate
(green). DAPI was used for
nuclear staining. Bar size 40
μm. PMA: phorbol 12-myris-
tate 13-acetate.

792 haematologica | 2018; 103(5)

 Renal accumulation of MSP1 in sickle cell disease

A B C

P=0.003

E F

G H I J

Figure 5. MSP1 treatment of cultured human glomerular endothelial cell line (HGEC) induces cell motility and vWF expression, F–actin re-organization, and RON
kinase signaling in HGEC. (A) Cell growth measured by MTT assay. Five wells are used for either control or treatment with 1 μM of recombinant MSP1 in each exper-
iment. Results are representative of three independent experiments. (B) Representative picture of wound migration assay of control cells and cells treated with 1
μM of MSP1. Bar size 300 μm. (C) Quantification of wound migration assay. Three wells are used per treatment in each experiment. Results are representative of
three independent experiments. (D) Immunostaining of F-actin (green) and vWF (red) in control and treated MSP1 (1 μM) treated cells with or without RONi (200
nM). DAPI (for F-actin) and Hematoxylin (for vWF) were used for nuclear staining. Non-specific primary antibodies were used for negative control. Bar size 40 μm for
F-actin and 100 μm for vWF. (E and F) Western blots of phosphorylated and non-phosphorylated forms of ERK and AKT kinases in HGEC. (E) Cells were treated with
MSP1 (1 μM) and collected at different time points after treatment. (F) Cells were treated with MSP1 (1 μM) with or without RON inhibitor (RONi, 200 nM) for 30
min. (G-J) Quantification of pERK and pAKT on Western blot. For quantification graphs, mean and SD are shown. *P<0.05. Results are representative of three inde-
pendent experiments. β-actin used for loading normalization.

haematologica | 2018; 103(5) 793

 A. Khaibullina et al.
P=1x10-5
were significantly increased after MSP1 treatment (Figure
5E-G). Phosphorylation of ERK and AKT was significantly
reduced after treatment with RONi (Figure 5H-J). Taken
together, MSP1 activated RON receptor signaling in cul-
tured HGEC leading to the re-organization of F-actin and
increasing cell motility and vWF expression.
MSP1 increases permeability in SCD mouse glomeruli
Glomerular hyperfiltration is an early stage manifesta-
tion in the development of SCD glomerulopathy.34,35 To
test whether MSP1 increases glomeruli permeability, we
utilized a mouse whole glomeruli permeability assay.24
Renal glomerular permeability measurement was based
on the determination of albumin permeability. In the
absence of capillary filtration, the diffusional loss of albu-
min from glomeruli capillary is negligible.24 Placement of
intact glomeruli into the hypooncotic solution increases
glomeruli volume due to water accumulation inside
glomeruli according to oncotic gradient.24 Treatment of
glomeruli with permeating agents that increase glomerular
filtration significantly increases albumin loss and reduces
oncotic gradient, leading to a reduced glomerular volume
in the hypooncotic solution.24 Renal glomeruli were isolat-
ed from control mice and treated with recombinant MSP1
(1 μM) in the presence or absence of RONi (200 nM) for
30 min (see Methods for details). PBS was used as a vehicle
control. Placement of vehicle-treated glomeruli in hypoon-
cotic solution significantly increased glomerular volume
by more than 3.5-fold (Figure 6A and B). MSP1 treatment
reduces glomerular volume, whereas RONi treatment
restored the ability of glomeruli to enlarge in the hypoon-
cotic solution (Figure 6A and B). Taken together, pre-treat-
ment of mouse glomeruli with MSP1 significantly
794
Figure 6. MSP1 increases glomerular perme-
ability in whole glomeruli assay. (A)
Representative pictures of glomeruli isolated
from control mice placed in isooncotic and
P=0.008
hypooncotic solutions. Glomeruli were pre-
treated with vehicle (PBS), or with MSP1 (1
μM) with or without RON inhibitor (RONi 200
nM). Bar size 40 μm. (B) Quantification of
glomerular volume change in hypooncotic
solution. CellSens Standard software was
used for measurement of the glomeruli area
and glomeruli volume was calculated. Percent
of volume changes in isooncotic solution is
shown. For quantification, graphs means are
shown. Each dot represents a value obtained
from one glomerulus.
increased glomerular permeability for albumin, and RONi
reduced MSP1-associated glomerular permeability.
Treatment of SCD mice with RONi significantly
ameliorates glomerular endothelial injury
To test whether RONi prevents development of
glomerular endothelial injury in young SCD mice, 2-
month old SCD mice were injected with either RONi (10
mg/kg of body weight in 2% DMSO) or vehicle (2%
DMSO, n=6 per group) subcutaneously for 14 days.
Control non-SCD mice were also injected with either
RONi or vehicle (n=4 per group). We used younger (2-
month old) mice to test whether RONi prevents develop-
ment of glomerular disease, because older (4-month old)
mice had already developed profound glomerular disease
(Figures 1 and 2). Only 20% of 2-month old mice devel-
oped microalbuminuria (1 of 5 mice), and 40% of 12-week
old non-treated SCD mice developed microalbuminuria (2
of 5 mice) (data not shown). In contrast, we found that all
non-treated mice had glomerular endothelia cell injury at
12 weeks of age. Because endothelial injury was detected
before the onset of albuminuria, we focused on the role of
RON signaling in the development of endothelial injury.
Administration of RONi in SCD mice did not affect body
weight (Figure 7A), but significantly reduced kidney size
(Figure 7B and C). Moreover, administration of RONi in
SCD mice significantly reduced glomerular hypertrophy
and capillary congestion (Figure 8A and E and Online
Supplementary Figure S5, H&E staining). Capillary dilation
and thickening of capillary loops and glomerular basement
membrane were significantly reduced in SCD mice treat-
ed with RONi compared to mice with vehicle injection
(Figure 8B and F and Online Supplementary Figure S5, PAS
haematologica | 2018; 103(5)
 Renal accumulation of MSP1 in sickle cell disease

staining). RONi treatment did not affect mesangial expan- A

sion in SCD mice (Figure 8A and B, H&E and Online
Supplementary Figure S6, PAS staining). However, RONi
administration in SCD mice significantly reduced capillary
expression of vWF (Figure 8C and G and Online
Supplementary Figure S5, immunostaining, red), and
Intercellular Adhesion Molecule 1 (ICAM1) (Figure 8D
and E and Online Supplementary Figure S5, immunostain-
ing, red), which are the markers of endothelial injury.36,37
Thus, inhibiton of RON receptor in SCD mice significant-
ly ameliorated development of endothelial and glomerular
injury. B

Discussion
Renal disease in SCD patients includes a variety of
glomerular and tubular complications, but the mechanism
of their development is not fully understood. The most
important finding in our study is that MSP1 and its recep-
tor RON kinase may play a role in the activation of renal
endothelium and development of glomerular pathology in
SCD mice. Moreover, treatment of mice with RONi, an
inhibitor of RON kinase, significantly ameliorated
glomerular hypertrophy, capillary dilation and congestion,
and endothelial injury. These findings reveal a previously
unknown mechanism which contributes to glomerular
endothelial injury in SCD.
Sickle cell disease mice spontaneously develop FSGS
that may be directly associated with RBC sickling and
chronic hemolysis. The focal nature of glomerulosclerosis
in SCD mice and SCD patients apparently excludes the
effect of global factors, such as hypoxia, cell-free heme,
iron, and other circulating factors that would lead to a
global and not focal glomerulosclerosis with involvement
of only 50% of glomeruli in SCD mice. Thus, locally-pro-
duced factor(s) are more likely to contribute to the devel-
opment of FSGS.
In agreement with a previous report,21 we demonstrate
here that macrophage infiltration in renal glomeruli was
increased in SCD mice. Sickling and adhesion of RBCs,
and accumulation of RBC lysate products within the kid-
ney might stimulate renal infiltration of monocyte-derived
macrophages. Monocyte-derived renal macrophages are
present in all forms of kidney disease with inflammation,
and renal capillary macrophage infiltration is a character-
istic pathology of FSGS.38 In many human biopsy studies,
number of glomerular or interstitial macrophages correlate
with poor outcomes, suggesting their possible role in the
disease progression.39,40 However, the role of infiltrating
macrophages in the progression of renal disease is not well
understood. Phagocytosis of senescent sickled RBCs, RBC
exosomes and endocytosis of cell-free hemoglobin
increases inflammatory response in the cultured human
monocytes/macrophages.12 Activation of proteases is a
universal inflammatory response in macrophages.41 We
demonstrate here that products of RBC hemolysis signifi-
cantly increased expression of macrophage membrane-
bound protease, MT-SP1 in cultured human macrophages.
Meta-analysis data also showed that hypoxia and mono-
cyte differentiation increased MT-SP1 expression in pri-
mary human macrophages. MT-SP1 is a type II membrane
serine protease that plays important roles in cell migration
and tumor cell metastasis.42
MT-SP1 is one of the proteases that activate circulating
C P=0.04
Figure 7. Treatment of sickle cell disease (SCD) mice with RON inhibitor
reduces kidney hypertrophy. (A) Treatment of SCD mice with RON inhibitor
(RONi) does not affect body weight (BW). (B) Representative picture of kidneys
of controls and SCD mice treated either with (RONi) or vehicle (2% DMSO). (C)
Quantification of kidney weight. Kidney weights (KW) to BW ratios are shown.
MSP1.15 MSP1 expression has been found in the renal
tubular cells.43,44 In agreement with previous studies, we
did not observe MSP1 expression or accumulation in
glomeruli of control mice. In contrast, we found that
MSP1 was accumulated in approximately 46±8% of renal
glomerular capillaries in SCD mice. This accumulation
rate was similar to the percentage of injured glomeruli in
SCD mice. Glomerular accumulation of MSP1 was previ-
ously shown in the rat model of anti-Thy1 glomerular dis-
ease, and the neutralization of MSP1 by the injected anti-
bodies reduced serum creatinine and proteinuria, and pro-
tected animals from glomerular injury.17 However, the
mechanism of MSP1-associated glomerular injury was not
clarified. RON is expressed in human renal tubular cells
and glomerular mesangial cells.43,44 MSP1 treatment
induces growth, motility and collagen invasion of mesan-
gial cells.43 Expression of functional RON in endothelial
cells is unknown. MSP1 that is accumulated in glomerular
capillary of SCD mice may potentially affect endothelial
cells, or leak from capillary to affect mesangial cells or
podocytes. The increased glomeruli size associated with
dilated glomerular capillary and mesangial proliferation
was reported both in SCD patients and mouse model of
SCD.20,21,45 In our study, inhibition of RON in SCD mice
significantly reduced glomerular hypertrophy, as well as
capillary dilation and congestion without reduction of
mesangial expansion. It is possible that the short time of
treatment was not enough to produce statistically signifi-

haematologica | 2018; 103(5) 795

 A. Khaibullina et al.

cant differences in mesangial cell proliferation, or that fac-
tors other than MSP1 could stimulate mesangial growth
in SCD mice. Thus, we hypothesized that MSP1 played a
role in the activation of endothelial cells. The effect of
MSP1 on renal glomerular endothelial cells is still not
known.
To test the effect of MSP1 on endothelium, we used cul-
tured HGEC. MSP1 had previously been shown to stimu-
late motility of murine resident peritoneal macrophages
and kidney epithelial cells.18,46 We demonstrated that
MSP1 treatment increased motility of HGECs and induced
formation of F-actin stress fibers that is essential for motil-
ity and permeability of endothelial cells.32 Further studies
are needed to determine whether MSP1/RON signaling
induces permeability of cultured endothelial cells.
Interestingly, MSP1 increased vWF expression levels in
HGEC. High levels of vWF were also found in SCD
murine glomeruli. RONi treatment reduced endothelial

E F G H

P=3x10-8 P=2x10-7 P=0.002 P=0.04

Figure 8. Treatment of sickle cell disease (SCD) mice with RON inhibitor (RONi) reduces renal endothelial injury. Representative pictures of renal sections of control
and SCD mice treated with either RONi or vehicle (DMSO) are shown. (A) Hematoxylin&Eosin (H&E) staining. (B) Periodic Acid-Schiff (PAS) staining. (C) von Willebrand
factor (vWF) immunostaining (red). (D) Intercellular Adhesion Molecule (ICAM) immunostaining (red). Bar sizes 50µm. (E and F) Quantification of glomeruli size (E)
and capillary size per glomeruli cross section (F) is performed using CellSens Standard software. (G and H) Quantification of vWF (G) and ICAM (H) expression in
glomeruli cross sections is performed using ImageJ Fiji version software. Six mice per group were used for each staining. Each dot represents a value obtained from
one glomerulus cross-section. For quantification graphs, means are shown.

796 haematologica | 2018; 103(5)

 Renal accumulation of MSP1 in sickle cell disease

vWF levels in vivo and in vitro. vWF mediates adhesion of
sickle RBCs to endothelial cells, inducing oxidant stress,
and increasing expression of ICAM-1, VCAM and E-
selectin.47 Indeed, RONi treatment in SCD mice was asso-
ciated with reduced vWF levels in glomerular capillaries,
and decreased RBC adhesion and ICAM-1 levels.
Therefore, the mechanism of RON inhibition may be
associated with prevention of glomerular capillary conges-
tion, leading to improvement of renal hemodynamics. A
correlation between an alteration of renal hemodynamics
and a renal histological injury has been demonstrated in a
model of renal ischemia in SCD.48 In addition, our results
demonstrate that MSP1 treatment of glomeruli isolated
from control mice significantly increased albumin perme-
ability that was effectively prevented by RON inhibition.
Hyperfiltration is associated with glomerular hypertrophy
and glomerulosclerosis.49 We demonstrate a significant
reduction in glomerular size in SCD mice after treatment
with RON inhibitors, suggesting a possible change in
hyperfiltration. Reduction of glomerular size after 14-day
treatment with RONi is unlikely to be due to structural
remodeling or reduced proliferation of endothelial cells.
To the best of our knowledge, this is the first study
demonstrating function of MSP1/RON in the glomerular
endothelial cells. Future studies will clarify the mechanism
of MSP1-associated endothelial cell activation and its role
in the development of glomerular injury.
The present study established for the first time that
inhibition of RON kinase significantly ameliorates SCD
renal pathology in mice. Short-term inhibition of RON
kinase reduced endothelial injury in young SCD mice dur-
ing the early stage of renal disease before the onset of
albuminuria. We do not know how long this effect per-
sists after RONi withdrawal. Whether short-term inhibi-
tion of RON kinase will ameliorate already developed
renal disease in older mice is currently under investigation.
Future studies will elucidate a role of Ron kinase in renal
disease in SCD patients. These findings also highlight a
new potential therapeutic target for CKD in SCD. A
recent pre-clinical study50 and Phase 1 clinical trial (clinical-
trials.gov identifier: 01721148) of BMS-777607 RON
inhibitor for treatment of human cancers demonstrated its
safety and good tolerability, providing a proof of principal
for the potential use of this pharmacological approach.
Funding
This work was supported in part by CHaRM pilot grant
awarded to MJ through NIH grant 1P50HL118006 (to SN) and
NIH grants 1R01HL125005 (to SN), 5G12MD007597 (to
SN), and R41MD008829-01 (to ZMQ). The content is solely
the responsibility of the authors and does not necessarily represent
the official views of the National Institutes of Health.

References
1. Hamideh D, Alvarez O. Sickle cell disease
related mortality in the United States
(1999-2009). Pediatr Blood Cancer.
2013;60(9):1482-1486.
2. Elmariah H, Garrett ME, De Castro LM, et
al. Factors associated with survival in a
contemporary adult sickle cell disease
cohort. Am J Hematol. 2014;89(5):530-535.
3. Becton LJ, Kalpatthi RV, Rackoff E, et al.
Prevalence and clinical correlates of
microalbuminuria in children with sickle
cell disease. Pediatr Nephrol. 2010;25(8):
1505-1511.
4. Saraf SL, Zhang X, Kanias T, et al.
Haemoglobinuria is associated with chron-
ic kidney disease and its progression in
patients with sickle cell anaemia. Br J
Haematol. 2014;164(5):729-739.
5. Guasch A, Navarrete J, Nass K, Zayas CF.
Glomerular involvement in adults with
sickle cell hemoglobinopathies: Prevalence
and clinical correlates of progressive renal
failure. J Am Soc Nephrol. 2006;17(8):2228-
2235.
6. Sharpe CC, Thein SL. Sickle cell nephropa-
thy - a practical approach. Br J Haematol.
2011;155(3):287-297.
7. Scheinman JI. Sickle cell disease and the
kidney. Nat Clin Pract Nephrol. 2009;5(2):
78-88.
8. Tejani A, Phadke K, Adamson O, et al.
Renal lesions in sickle cell nephropathy in
children. Nephron. 1985;39(4):352-355.
9. Potoka KP, Gladwin MT. Vasculopathy and
pulmonary hypertension in sickle cell dis-
ease. Am J Physiol Lung Cell Mol Physiol.
2015;308(4):L314-324.
10. Hand WL, King-Thompson NL. Effect of
erythrocyte ingestion on macrophage anti-
bacterial function. Infect Immun. 1983;
40(3):917-923.
11. Vinchi F, Costa da Silva M, Ingoglia G, et
al. Hemopexin therapy reverts heme-
induced proinflammatory phenotypic
switching of macrophages in a mouse
model of sickle cell disease. Blood.
2016;127(4):473-486.
12. Belcher JD, Marker PH, Weber JP, Hebbel
RP, Vercellotti GM. Activated monocytes
in sickle cell disease: potential role in the
activation of vascular endothelium and
vaso-occlusion. Blood. 2000;96(7):2451-
2459.
13. Klug PP, Kaye N, Jensen WN. Endothelial
cell and vascular damage in the sickle cell
disorders. Blood Cells. 1982;8(1):175-184.
14. Mahoney DH Jr, Fernbach DJ. Monocyte
functions in sickle cell disorders. Am J
Pediatr Hematol Oncol. 1983;5(4):409-411.
15. Bhatt AS, Welm A, Farady CJ, et al.
Coordinate expression and functional pro-
filing identify an extracellular proteolytic
signaling pathway. Proc Natl Acad Sci
USA. 2007;104(14):5771-5776.
16. Qiu D, Owen K, Gray K, Bass R, Ellis V.
Roles and regulation of membrane-associ-
ated serine proteases. Biochem Soc Trans.
2007;35(Pt 3):583-587.
17. Rampino T, Soccio G, Gregorini M, et al.
Neutralization of macrophage-stimulating
protein ameliorates renal injury in anti-thy
1 glomerulonephritis. J Am Soc Nephrol.
2007;18(5):1486-1496.
18. Leonard EJ. Biological aspects of
macrophage-stimulating protein (MSP) and
its receptor. Ciba Found Symp. 1997;212:
183-191.
19. Wang MH, Julian FM, Breathnach R, et al.
Macrophage stimulating protein (MSP)
binds to its receptor via the MSP beta
chain. J Biol Chem. 1997;272(27):16999-
17004.
20. Ryan TM, Ciavatta DJ, Townes TM.
Knockout-transgenic mouse model of sick-
le cell disease. Science. 1997;278(5339):
873-876.
21. Kasztan M, Fox BM, Speed JS, et al. Long-
Term Endothelin-A Receptor Antagonism
Provides Robust Renal Protection in
Humanized Sickle Cell Disease Mice. J Am
Soc Nephrol. 2017;28(8):2443-2458.
22. Jerebtsova M, Kumari N, Obuhkov Y,
Nekhai S. Adenoviral E4 gene stimulates
secretion of pigmental epithelium derived
factor (PEDF) that maintains long-term sur-
vival of human glomerulus-derived
endothelial cells. Mol Cell Proteomics.
2012;11(11):1378-1388.
23. Misra RP. Isolation of glomeruli from mam-
malian kidneys by graded sieving. Am J
Clin Pathol. 1972;58(2):135-139.
24. Savin VJ, Sharma R, Lovell HB, Welling DJ.
Measurement of albumin reflection coeffi-
cient with isolated rat glomeruli. J Am Soc
Nephrol. 1992;3(6):1260-1269.
25. Wang L, Almeida LE, de Souza Batista CM,
et al. Cognitive and behavior deficits in
sickle cell mice are associated with pro-
found neuropathologic changes in hip-
pocampus and cerebellum. Neurobiol Dis.
2015;85:60-72.
26. Khaibullina A, Almeida LE, Wang L, et al.
Rapamycin increases fetal hemoglobin and
ameliorates the nociception phenotype in
sickle cell mice. Blood Cells Mol Dis.
2015;55(4):363-372.
27. Sporn LA, Marder VJ, Wagner DD.
Inducible secretion of large, biologically
potent von Willebrand factor multimers.
Cell. 1986;46(2):185-190.
28. Wagner DD, Bonfanti R. von Willebrand

haematologica | 2018; 103(5) 797

 A. Khaibullina et al.
factor and the endothelium. Mayo Clin
Proc. 1991;66(6):621-627.
29. Kairaitis LK, Wang Y, Gassmann M, Tay
YC, Harris DC. HIF-1alpha expression fol-
lows microvascular loss in advanced
murine adriamycin nephrosis. Am J Physiol
Renal Physiol. 2005;288(1):F198-206.
30. Blouin MJ, De Paepe ME, Trudel M.
Altered hematopoiesis in murine sickle cell
disease. Blood. 1999;94(4):1451-1459.
31. Kawaguchi M, Orikawa H, Baba T,
Fukushima T, Kataoka H. Hepatocyte
growth factor activator is a serum activator
of single-chain precursor macrophage-stim-
ulating protein. FEBS J. 2009;276(13):3481-
3490.
32. Prasain N, Stevens T. The actin cytoskele-
ton in endothelial cell phenotypes.
Microvasc Res. 2009;77(1):53-63.
33. Wagh PK, Peace BE, Waltz SE. Met-related
receptor tyrosine kinase Ron in tumor
growth and metastasis. Adv Cancer Res.
2008;100:1-33.
34. Haymann JP, Stankovic K, Levy P, et al.
Glomerular hyperfiltration in adult sickle
cell anemia: a frequent hemolysis associat-
ed feature. Clin J Am Soc Nephrol.
2010;5(5):756-761.
35. Aygun B, Mortier NA, Smeltzer MP,
Hankins JS, Ware RE. Glomerular hyperfil-
tration and albuminuria in children with
sickle cell anemia. Pediatr Nephrol.
2011;26(8):1285-1290.
36. Pober JS, Cotran RS. Cytokines and
endothelial cell biology. Physiol Rev.
798
1990;70(2):427-451.
37. Wertheimer SJ, Myers CL, Wallace RW,
Parks TP. Intercellular adhesion molecule-1
gene expression in human endothelial cells.
Differential regulation by tumor necrosis
factor-alpha and phorbol myristate acetate.
J Biol Chem. 1992;267(17):12030-12035.
38. Cao Q, Harris DC, Wang Y. Macrophages
in kidney injury, inflammation, and fibro-
sis. Physiology. 2015;30(3):183-194.
39. Hill GS, Delahousse M, Nochy D, Mandet
C, Bariety J. Proteinuria and tubulointersti-
tial lesions in lupus nephritis. Kidney Int.
2001;60(5):1893-1903.
40. Rollino C, Basolo B, Roccatello D,
Menegatti E, Piccoli G. Atypical presence of
antimyeloperoxidase antibodies in 2 trans-
planted patients. Nephron. 1993;63(4):480.
41. Tomlinson ML, Garcia-Morales C, Abu-
Elmagd M, Wheeler GN. Three matrix met-
alloproteinases are required in vivo for
macrophage migration during embryonic
development. Mech Dev. 2008;125(11-
12):1059-1070.
42. Darragh MR, Bhatt AS, Craik CS. MT-SP1
proteolysis and regulation of cell-microen-
vironment interactions. Front Biosci.
2008;13:528-539.
43. Rampino T, Collesi C, Gregorini M, et al.
Macrophage-stimulating protein is pro-
duced by tubular cells and activates mesan-
gial cells. J Am Soc Nephrol. 2002;13(3):
649-657.
44. Cantaluppi V, Biancone L, Romanazzi GM,
et al. Macrophage stimulating protein may
promote tubular regeneration after acute
injury. J Am Soc Nephrol. 2008;19(10):
1904-1918.
45. Nath KA, Hebbel RP. Sickle cell disease:
renal manifestations and mechanisms. Nat
Rev Nephrol. 2015;11(3):161-171.
46. Zhang K, Yao HP, Wang MH. Activation of
RON differentially regulates claudin
expression and localization: role of claudin-
1 in RON-mediated epithelial cell motility.
Carcinogenesis. 2008;29(3):552-559.
47. Sultana C, Shen Y, Rattan V, Johnson C,
Kalra VK. Interaction of sickle erythrocytes
with endothelial cells in the presence of
endothelial cell conditioned medium
induces oxidant stress leading to
transendothelial migration of monocytes.
Blood. 1998;92(10):3924-3935.
48. Juncos JP, Grande JP, Croatt AJ, et al. Early
and prominent alterations in hemodynam-
ics, signaling, and gene expression follow-
ing renal ischemia in sickle cell disease. Am
J Physiol Renal Physiol. 2010;298(4):F892-
899.
49. Brenner BM, Lawler EV, Mackenzie HS.
The hyperfiltration theory: a paradigm
shift in nephrology. Kidney Int.
1996;49(6):1774-1777.
50. Bieniasz M, Radhakrishnan P, Faham N, De
La O JP, Welm AL. Preclinical Efficacy of
Ron Kinase Inhibitors Alone and in
Combination with PI3K Inhibitors for
Treatment of sfRon-Expressing Breast
Cancer Patient-Derived Xenografts. Clin
Cancer Res. 2015;21(24):5588-5600.
haematologica | 2018; 103(5)
 Myeloproliferative Disorders ARTICLE

The KIT and PDGFRA switch-control inhibitor

DCC-2618 blocks growth and survival Ferrata Storti
Foundation
of multiple neoplastic cell types in
advanced mastocytosis
Mathias Schneeweiss,1,2 Barbara Peter,1,2 Siham Bibi,3 Gregor Eisenwort,1,2
Dubravka Smiljkovic,1 Katharina Blatt,1,2 Mohamad Jawhar,4 Daniela Berger,2
Gabriele Stefanzl,2 Susanne Herndlhofer,1,2 Georg Greiner,5 Gregor Hoermann,5
Emir Hadzijusufovic,1,2,6 Karoline V. Gleixner,1,2 Peter Bettelheim,7 Klaus Geissler,8
Wolfgang R. Sperr,1,2 Andreas Reiter,4 Michel Arock3 and Peter Valent1,2
Haematologica 2018
Volume 103(5):799-809
1
Ludwig Boltzmann Cluster Oncology, Medical University of Vienna, Austria; 2Department of
Internal Medicine I, Division of Hematology and Hemostaseology, Medical University of
Vienna, Austria; 3Laboratoire de Biologie et Pharmacologie Appliquee, CNRS UMR 8113,
Ecole Normale Superieure de Cachan, France; 4Department of Hematology and Oncology,
University Medical Centre Mannheim and Medical Faculty Mannheim, Heidelberg
University, Germany; 5Department of Laboratory Medicine, Medical University of Vienna,
Austria; 6Department for Companion Animals and Horses, University Clinic for Small
Animals, Internal Medicine Small Animals, University of Veterinary Medicine Vienna,
Austria; 7Elisabethinen Hospital Linz, Austria and 8Fifth Medical Department, Hospital
Hietzing, Vienna, Austria

ABSTRACT

S
ystemic mastocytosis is a complex disease defined by abnormal
growth and accumulation of neoplastic mast cells in various organs.
Most patients exhibit a D816V-mutated variant of KIT, which con-
fers resistance against imatinib. Clinical problems in systemic mastocyto-
sis arise from mediator-related symptoms and/or organ destruction
caused by malignant expansion of neoplastic mast cells and/or other
myeloid cells in various organ systems. DCC-2618 is a spectrum-selective Correspondence:
pan KIT and PDGFRA inhibitor which blocks KIT D816V and multiple peter.valent@meduniwien.ac.at
other kinase targets relevant to systemic mastocytosis. We found that
DCC-2618 inhibits the proliferation and survival of various human mast
cell lines (HMC-1, ROSA, MCPV-1) as well as primary neoplastic mast
Received: September 1, 2017.
cells obtained from patients with advanced systemic mastocytosis (IC50
<1 μM). Moreover, DCC-2618 decreased growth and survival of primary
Accepted: January 31, 2018.
Pre-published: February 8, 2018.
neoplastic eosinophils obtained from patients with systemic mastocyto-
sis or eosinophilic leukemia, leukemic monocytes obtained from patients
with chronic myelomonocytic leukemia with or without concomitant doi:10.3324/haematol.2017.179895
systemic mastocytosis, and blast cells obtained from patients with acute
myeloid leukemia. Furthermore, DCC-2618 was found to suppress the Check the online version for the most updated
proliferation of endothelial cells, suggesting additional drug effects on information on this article, online supplements,
systemic mastocytosis-related angiogenesis. Finally, DCC-2618 was and information on authorship & disclosures:
www.haematologica.org/content/103/5/799
found to downregulate IgE-mediated histamine release from basophils
and tryptase release from mast cells. Together, DCC-2618 inhibits
growth, survival and activation of multiple cell types relevant to ©2018 Ferrata Storti Foundation
advanced systemic mastocytosis. Whether DCC-2618 is effective in vivo Material published in Haematologica is covered by copyright.
in patients with advanced systemic mastocytosis is currently under inves- All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
tigation in clinical trials. conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
Introduction sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
Systemic mastocytosis (SM) is a hematopoietic neoplasm with complex biology from the publisher.
and pathology, and a variable clinical course.1-7 The disease is characterized by
abnormal expansion and accumulation of neoplastic mast cells (MC) in one or more

Haematologica | 2018; 103(5) 799

 M. Schneeweiss et al.

internal organs, including the bone marrow.1-3 Various
types of SM have been recognized by the World Health
Organization (WHO).8-11 The indolent variant of SM is
associated with ‘hematologic stability’ and thus with an
almost normal life expectancy.12-14 By contrast, the progno-
sis in patients with advanced SM, including SM with an
associated hematologic neoplasm (AHN), aggressive SM
(ASM) and MC leukemia (MCL) is unfavorable, with short
survival times and poor responses to conventional
therapy.1-5,12,13,15 Current research is, therefore, focusing on
therapeutic targets and the effects of novel antineoplastic
drugs on various cell types relevant to advanced SM.16
Since most patients with SM also suffer from mediator-
related symptoms that may sometimes be severe or even
life-threatening, such drugs are often selected based on
their dual effects on MC growth and MC activation.
Most patients with SM express the D816V-mutated vari-
ant of the stem cell factor receptor, KIT, which mediates
ligand-independent activation and autonomous growth
and differentiation of MC.17-22 The D816V KIT point muta-
tion also confers resistance against several tyrosine kinase
inhibitors, including imatinib.23-26 Novel kinase blockers act-
ing on KIT D816V have, therefore, been developed. The
highlighting example is midostaurin (PKC412).27,28
However, despite superior clinical efficacy seen in a global
phase II trial,28 patients with advanced SM often exhibit or
acquire resistance.28,29 A number of different mechanisms
may underlie resistance against midostaurin. One obvious
problem is that the drug does not suppress all clinically rel-
evant sub-clones and cell-types, especially cells lacking KIT
D816V.28,29 Such sub-clones are often seen in the context of
advanced SM. Over 50% of these patients have or develop
an AHN.30-32 Of these patients with an AHN, approximate-
ly 80-90% have an associated myeloid neoplasm, the most
frequent ones being chronic myelomonocytic leukemia
(CMML) and acute myeloid leukemia (AML).8-11,30-32 In these
patients, leukemic expansion of monocytes and/or blast
cells is typically found. In other patients, an expansion of
eosinophils, sometimes resembling chronic eosinophilic
leukemia (SM-CEL), is found. In most of these patients,
eosinophils display KIT D816V.33 By contrast, expression of
rearranged PDGFR variants is rarely seen in SM, although
in some patients with a FIP1L1/PDGFRA fusion gene, the
MC expansion has a histopathological picture indistin-
guishable from that of SM.34 Treatment of SM-AHN repre-
sents a clinical challenge because the AHN-component is
often resistant.16,32
DCC-2618 is a switch-control type II inhibitor of KIT,
which arrests KIT in an inactive state, regardless of activat-
ing mutations, such as KIT D816V.35 Moreover, several
additional oncogenic kinases, including FLT-3, PDGFRA,
PDGFRB, KDR, TIE2 and FMS are recognized by DCC-
2618.35 Recently, the first clinical trials with DCC-2618
(NCT02571036) were started in patients with kinase-dri-
ven malignancies. In addition, first preclinical studies have
shown that DCC-2618 may exert antineoplastic effects on
neoplastic MC.36
In our current study, we show that DCC-2618 is a potent
inhibitor of growth and survival of neoplastic human MC
expressing various KIT mutations. Furthermore, we show
that DCC-2618 produces growth inhibition and apoptosis
in other cell types that play a role in advanced SM. Finally,
we show that DCC-2618 inhibits IgE-dependent histamine
secretion from basophils and tryptase secretion from MC.
All in all, our data suggest that DCC-2618 is a promising,
novel drug for the treatment of advanced SM.
Methods
Reagents
The reagents used in this study are described in the Online
Supplement. DCC-2618 and its active metabolite, DP-5439, were
kindly provided by Dr. B. Smith (Deciphera Pharmaceuticals LLC,
Lawrence, KS, USA).
Isolation of primary neoplastic cells
Primary neoplastic cells were isolated from bone marrow
samples of 11 patients with SM. The bone marrow cells were
obtained during routine diagnostic investigations after written

Table 1. Characteristics of patients with systemic mastocytosis and response of neoplastic cells to DCC-2618 and DP-5439.
Patient n. Age m/f Diagnosis, SM KIT Serum tryptase BM MC % MC DCC-2618 DP-5439 PKC412
variant D816V ng/mL infiltration % in MNC IC50 IC50 IC50
#1 68 m ISM + 83.3 n.a. 1 114 nM 414 nM 56 nM
#2 56 m ISM + 103 20 5 240 nM 390 nM 35 nM
#3 49 f ISM - 22.9 10 n.a. 198 nM 554 nM n.a.
#4 66 m ISM-MPN-eo + 170 5 n.a. 394 nM 1481 nM n.a.
#5 82 f SSM + 284 50 n.a. 347 nM 1584 nM n.a.
#6 57 f ASM + 87.9 50 n.a. 331 nM 366 nM n.a.
#7 90 m ASM + 125 20 25 386 nM 679 nM 360 nM
#8 63 m ASM-AML + 33.9 15 8 393 nM 554 nM 66 nM
#9 65 m ASM-CMML + 220 50 30 460 nM 907 nM n.a.
#10 78 m ASM-MPN-eo + 45.9 15 5 83 nM 64 nM 114 nM
#11 91 m MCL + 330 20 50 321 nM 592 nM 65 nM
Diagnoses were established according to WHO criteria. Primary bone marrow cells were incubated with various concentrations of DCC-2618, DP-5439 or midostaurin (PKC412)
at 37°C for 48 h. Proliferation was then determined by measuring uptake of 3H-thymidine and IC50 values were calculated. WHO: World Health Organization; m: male; f: female;
SM: systemic mastocytosis; PB, peripheral blood; BM, bone marrow; MC: mast cells; MNC, mononuclear cells, nM, nanomolar; n.a., not available; IC50, half maximal inhibitory con-
centration; ISM: indolent SM; MPN: myeloproliferative neoplasms; SSM: smoldering SM; ASM: aggressive SM; CMML: chronic monomyelocytic leukemia; MCL: mast cell leukemia.

800 haematologica | 2018; 103(5)

 DCC-2618: a new drug against mastocytosis

informed consent had been given. Patients were classified as Results

having indolent SM (ISM; n=3), smoldering SM (SSM; n=1),
ASM (n=2), SM-AHN (n=4) and MCL (n=1) according to WHO DCC-2618 and its metabolite DP-5439 inhibit
criteria.8-11 In addition, neoplastic cells were obtained from ten proliferation of neoplastic mast cells
patients suffering from CMML, ten with AML and three with DCC-2618 and its active metabolite, DP-5439 were
hypereosinophilia. The patients’ characteristics are shown in found to inhibit 3H-thymidine uptake and thus prolifera-
Table 1 (SM patients) and Online Supplementary Table S1 (other tion in a dose-dependent manner in all MC lines tested,
hematologic disorders). Heparinized bone marrow cells were with slightly lower IC50 values obtained in HMC-1.1 cells
layered over Ficoll to isolate mononuclear cells. The study was lacking KIT D816V and ROSAKIT WT cells compared to the
approved by the ethics committee of the Medical University of KIT D816V-positive cell lines HMC-1.2 and ROSAKIT D816V
Vienna. (Figure 1A and Table 2). IC50 values obtained in HMC-1.1
cells with DCC-2618 were also lower than IC50 values
Culture of human cell lines obtained with midostaurin.25,26 In addition, DCC-2618 and
The following human MCL-like cell lines were employed in DP-5439 were found to inhibit proliferation of ROSAKIT K509I
this study: HMC-1.1 and HMC-1.2,37 three ROSA sub-clones cells with lower IC50 values (DCC-2618, IC50: 34±10 nM)
(ROSAKIT WT, ROSAKIT D816V, ROSAKIT K509I)38 and four MCPV-1 sub- compared to ROSAKIT D816V cells (Figure 1A). Unexpectedly,
clones (MCPV-1.1, MCPV-1.2, MCPV-1.3, MCPV-1.4).39 In addi- DCC-2618 and its metabolite also produced growth-inhi-
tion, we examined several AML cell lines, the CEL-related cell bition in the multi-resistant MC lines MCPV-1.1, MCPV-
line EOL-1, the microvascular endothelial cell line HMEC-1, and 1.2, MCPV-1.3 and MCPV-1.4 (Figure 1B and Table 2).
cultured human umbilical vein endothelial cells (HUVEC). A Finally, we were able to show that DCC-2618 and DP-
description of cell lines is provided in the Online Supplement. 5439 induced dose-dependent inhibition of growth of pri-
mary neoplastic bone marrow cells obtained from patients
Evaluation of growth, survival of neoplastic cells suffering from various forms of SM, including ASM and
Drug-exposed cells (cell lines or primary cells) were analyzed MCL (Figure 1C, Table 1). Interestingly, the effects of
for proliferation and survival. The bioassays employed are DCC-2618 on primary neoplastic BM cells and the related
described in the Online Supplementary Methods. IC50 values obtained in different SM variants were compa-
rable (Figure 1C, Table 1).
Western blotting
For evaluation of KIT and BTK signaling, HMC-1.1, HMC-1.2, DCC-2618 inhibits KIT, STAT5, AKT, and ERK activation in
ROSAKIT WT and ROSAKIT D816V cells were incubated in control neoplastic mast cells
medium or in DCC-2618 (0.5-5 μM) for 4 h at 37°C. Western As expected, DCC-2618 was found to suppress phos-
blotting was performed essentially as described elsewhere.26,40 phorylation of KIT in ROSAKIT WT and ROSAKIT D816V cells as
For evaluation of downstream signaling pathways of KIT, HMC- well as in both HMC-1 sub-clones (Figure 2A). In addition,
1.1, HMC-1.2, ROSAKIT WT and ROSAKIT D816V cells were first pre- DCC-2618 was found to decrease the expression of phos-
incubated overnight in Iscove modified Dulbecco medium phosphorylated (p)STAT5, pAKT and pERK1/2 in all cell
devoid of fetal calf serum and of stem cell factor. Cells (106) from lines tested (Figure 2B). As expression levels of pSTAT5 in
each line were then treated with DCC-2618 (0.001-10 μM) for
90 min at 37°C. At the end of the treatment, ROSAKIT WT cells
were stimulated with stem cell factor-containing supernatants
(10%) of Chinese hamster ovary cells transfected with the Table 2. Effects of DCC-2618 and DP-5439 on growth of various
murine scf (kl) gene (CHO-KL) at room temperature for 10 min. human cell types.
Thereafter, Western blotting was performed essentially as Cell line / cell type DCC-2618, IC50 DP-5439, IC50
described previously.26,40 Antibodies against phosphorylated
(p)Kit, STAT5, pSTAT5, ERK1/2, pERK1/2 were purchased from HMC-1.1 12.3±3.7 nM 13±4 nM
Cell Signaling (Danvers, MA, USA), antibodies against pBTK HMC-1.2 123±36 nM 96±23 nM
were bought from NovusBiologicals (Littleton, CO, USA) and ROSA KIT WT 41±5 nM 56±21 nM
antibodies against total KIT and total BTK were from Santa Cruz ROSA KIT D816V 168±65 nM 188±60 nM
Biotechnology (Santa Cruz, CA, USA).
ROSA KIT K509I 34±10 nM 28±13 nM
Measurement of mediator release MCPV-1.1 298±77 nM 357±179 nM
Drug-exposed cells (blood basophils obtained from healthy MCPV-1.2 1000±932 nM 913±378 nM
individuals and HMC-1 cells) were analyzed for histamine- and MCPV-1.3 724±511 nM 756±435 nM
tryptase release as described in the Online Supplementary MCPV-1.4 670±418 nM 697±410 nM
Methods.
MOLM-13 132±95 nM 73±31 nM
Evaluation of apoptosis in basophils MV4-11 147±88 nM 76±24 nM
Drug-exposed blood basophils obtained from healthy donors KG-1 2.7±3.7 μM 5.2±4.4 μM
by dextran sedimentation were analyzed for cell survival by U937 7.8±5.4 μM 2.5±0.3 μM
flow cytometry. Technical details are described in the Online
EOL-1 1.8±1.3 nM 1±0.8 nM
3.7±2.2 μM 2±1.3 μM
Supplementary Methods.
HMEC-1
Statistical analysis HUVEC 707±224 nM 605±222 nM
To determine the level of significance the Student t-test was Cell lines were incubated with various concentrations of DCC-2618 or DP-5439 at 37°C
for 48 h.Then, proliferation was determined by measuring uptake of 3H-thymidine and
applied. Results were considered to be statistically significantly
IC50 values were calculated. Values represent the mean±S.D. from three independent
different when P was <0.05. experiments. IC50: half maximal inhibitory concentration.

haematologica | 2018; 103(5) 801

 M. Schneeweiss et al.

HMC-1.1 cells were rather low and difficult to quantify by
Western blotting, we also performed intracellular flow
cytometry-staining experiments using an antibody against
pSTAT5. In these experiments, DCC-2618 was found to
counteract pSTAT5 expression in HMC-1.1 and HMC-1.2
cells in a dose-dependent manner (Online Supplementary
Figure S1). DCC-2618 did not inhibit phosphorylation of
BTK, another important target of tyrosine kinase inhibitors
expressed by neoplastic MC (Figure 2C).
DCC-2618 induces apoptosis in neoplastic mast cells
To explore the mechanism of drug action, we analyzed
the effects of DCC-2618 on the survival of neoplastic
MC. As assessed by light microscopy, DCC-2618 induced
apoptosis in HMC-1.1, HMC-1.2, ROSAKIT WT, ROSAKIT
D816V
and ROSA KIT K509I cells in a dose-dependent manner
(Figure 3A). The effects of DCC-2618 on survival were
more pronounced in KIT D816V-negative MC lines than
in KIT D816V-positive MC lines (Figure 3A). DCC-2618
was also found to produce apoptosis in the multi-resis-
tant MCPV-1 cell lines (Figure 3B). The apoptosis-induc-
ing effect of DCC-2618 on MC was confirmed by com-
bined annexin V/propidium-iodide staining (Figure 3A,B
and Online Supplementary Figure S2A,B). The metabolite
DP-5439 was found to be equally effective in producing
apoptosis in MC lines compared to DCC-2618 (Figure
3A,B and Online Supplementary Figure S2A,B). Together,
these data show that DCC-2618 is a novel potent anti-
neoplastic compound inducing apoptosis and growth
arrest in neoplastic MC.
DCC-2618 produces synergistic growth-inhibitory effects
with midostaurin and cladribine (2CdA) in neoplastic
mast cells
In advanced SM, drug combinations may be required to
suppress malignant cell growth. We found that DCC-2618
and midostaurin produce clear cooperative (synergistic)
growth-inhibitory effects in HMC-1.1 cells (Online
Supplementary Figure S3A,C). In HMC-1.2 cells, the drug
combination also produced cooperative antineoplastic
effects, but these effects were additive rather than syner-
gistic as defined by Calcusyn software (Online
Supplementary Figure S3A,C). In addition, we found that
DCC-2618 and 2CdA induce clear synergistic growth-
inhibitory effects on HMC-1.1 and HMC-1.2 cells (Online
Supplementary Figure S3B,D).

B C

Figures 1. DCC-2618 and its active metabolite DP-5439 inhibit proliferation of neoplastic mast cells. HMC-1, ROSA (A), MCPV-1 (B) and primary neoplastic mast
cells (C) obtained from patients with different variants of systemic mastocytosis (ISM, SSM, ASM, MCL) were incubated in control medium (0 nM) or medium con-
taining various concentrations of DCC-2618 or DP-5439, as indicated, at 37°C for 48 h. Thereafter, 3H-thymidine uptake was measured. Results in (A) and (B) are
expressed as percent of control and represent the mean±S.D. from three independent experiments. Results in (C) are expressed as percent of control and represent
mean±S.D from triplicates. Asterisk (*): P<0.05 compared to control medium.

802 haematologica | 2018; 103(5)

 DCC-2618: a new drug against mastocytosis

DCC-2618 inhibits IgE-dependent histamine release
from basophils and spontaneous tryptase release from
neoplastic mast cells
Since patients with SM often suffer from symptoms
caused by mediators released from neoplastic MC and/or
basophils, we evaluated the effect of DCC-2618 on anti-
IgE-induced histamine release. We found that DCC-2618
(0.1-1.0 μM) slightly inhibited anti-IgE mediated histamine
release from normal human blood basophils (Figure 4A).
This drug effect was found to be specific in that DCC-2618
did not inhibit C5a- or calcium ionophore-induced hista-
mine release from basophils (Online Supplementary Figure
S4A). As expected, DCC-2618 did not affect the viability of
basophils between 0.1 and 1.0 μM and did not induce his-
tamine secretion within 30 min of incubation (Online
Supplementary Figure S4B). In consecutive experiments, we
also found that DCC-2618 suppresses the spontaneous
(baseline) secretion of tryptase from HMC-1.1 and HMC-
1.2 cells during the entire incubation period (days 1
through 6) (Figure 4B).
DCC-2618 counteracts growth and survival of leukemic
monocytes and blast cells
We next explored the effects of DCC-2618 on AHN cell-
types. In a first step, we examined AML cell responses.
DCC-2618 was found to inhibit the proliferation of all
AML cell lines tested, with considerably lower IC50 values
obtained with the FLT3-mutated cell lines MOLM-13
(132±95 nM) and MV4-11 (147±88 nM) compared to KG-
1 and U937 cells (Table 2, Figure 5A). Similar effects were
seen with DP-5439 (Figure 5A). DCC-2618 was also found
to induce apoptosis in MOLM-13, MV4-11 and KG-1 cells
(Figure 5B and Online Supplementary Figure S5). Finally, we
found that DCC-2618 and DP-5439 produced dose-depen-
dent inhibition of growth in primary leukemia cells
obtained from patients with AML or CMML (Online
Supplementary Table S1 and Figure 5C). In one patient with
ASM-CMML, we isolated mononuclear cells and found
that DCC-2618 and DP-5439 induced growth inhibition of
these cells in the same way as in mononuclear cells
obtained from patients with CMML without SM (Table 1

Figures 2. DCC-2618 inhibits phosphorylation of KIT and other targets in neoplastic mast cells. (A,C) HMC-1 and ROSA cells were incubated in control medium (ROSAKIT
WT
: Iscove modified Dulbecco medium (IMDM) with stem cell factor, SCF; ROSAKIT D816V: IMDM without SCF) or medium containing various concentrations of DCC-2618, as
indicated, at 37°C for 4 h. Thereafter, cells were harvested and Western blotting was performed as described in the text using antibodies against phosphorylated (p)KIT,
total KIT, pBTK and total BTK. (B) HMC-1 and ROSA cells were first pre-incubated overnight in IMDM devoid of fetal calf serum and of SCF. Cells were then treated with
DCC-2618 (0.001-10 μM) for 90 min at 37°C. At the end of the treatment, ROSAKIT WT cells were stimulated with SCF (10% CHO-KL) at room temperature for 10 min.
Thereafter, cells were harvested and Western blotting was performed as described in the text using antibodies against pSTAT5, total STAT5, pAKT, total AKT, pERK1/2,
total ERK1/2. Western blot experiments were performed at least twice. Western blots in this figure show one representative experiment.

haematologica | 2018; 103(5) 803

 M. Schneeweiss et al.

and Figure 5C). Together, these data suggest that DCC-
2618 counteracts growth of AHN cells, including CMML
monocytes and AML blasts.
DCC-2618 inhibits the proliferation of neoplastic
eosinophils
Advanced SM is often accompanied by eosinophilia. In
addition PDGFRA is a known target of DCC-2618. We
analyzed the effects of DCC-2618 on proliferation and
survival of the FIP1L1-PDGFRA (F/P) positive EOL-1 cell
line. DCC-2618 was found to inhibit proliferation in
EOL-1 cells at low nanomolar range of concentrations
(IC50: 1.8±1.3 nM) (Online Supplementary Figure S6A).
Similar effects were seen with DP-5439 (Online
Supplementary Figure S6A). DCC-2618 also induced apop-
tosis in EOL-1 cells (Online Supplementary Figure S6C).
Next, we examined the effects of DCC-2618 on growth
of primary eosinophils. In these experiments, DCC-2618

Figures 3. DCC-2618 and DP-5439 induce apoptosis in neoplastic mast cells. HMC-1, ROSA (A) and MCPV-1 (B) were incubated in control medium (0 μM) or medium
containing various concentrations of DCC-2618 and DP-5439, as indicated, at 37°C for 48 h. Cells were then harvested and the percentage of apoptotic cells was
quantified morphologically on Wright-Giemsa-stained cytospin preparations (left panels) or by flow cytometry (determination of annexinV/PI-positive cells, right pan-
els). Results represent the mean±S.D. of three independent experiments. Asterisk (*): P<0.05 compared to control medium.

804 haematologica | 2018; 103(5)

 DCC-2618: a new drug against mastocytosis

was found to inhibit the proliferation of neoplastic bone
marrow cells obtained from a patient with ASM (Table 1
and Online Supplementary Figure S6B). In addition, DCC-
2618 was found to block the growth of bone marrow cells
obtained from patients with secondary hypereosinophilic
syndromes (Online Supplementary Table S1 and Online
Supplementary Figure S6B).
DCC-2618 inhibits growth of human endothelial cells
Increased bone marrow angiogenesis has been impli-
cated in the pathogenesis of SM.41 To investigate potential
effects of DCC-2618 on angiogenesis, we explored drug
effects on growth of HUVEC and the microvascular
endothelial cell line HMEC-1. As assessed by 3H-thymi-
dine uptake, DCC-2618 and its metabolite were found to
inhibit the proliferation of HUVEC and HMEC-1 cells in
a dose-dependent manner (Online Supplementary Figure
S7). DCC-2618 exerted stronger effects on HUVEC
(707±224 nM) than on HMEC-1 cells (3.7±2.2 µM).
Discussion
Due to the poor response to conventional drugs, treat-
ment of patients with advanced SM is still a major chal-
lenge in clinical practice. Despite the availability of new
drugs the prognosis of these patients remains poor with
short survival times.10,16,28,29 Research is, therefore, seeking
new effective drugs and novel treatment concepts. DCC-
2618 is a novel switch-control type II blocker that exerts
inhibitory effects on KIT D816V, other KIT mutants, and
several other critical target kinases, such as FLT3,
PDGFRA and KDR.35 We here describe that DCC-2618
inhibits the proliferation of nine different human MC
lines, with lower IC50 values obtained in HMC-1.1 cells
and ROSAKIT WT cells than in KIT D816V-positive HMC-1.2
and ROSAKIT D816V cells. In addition, DCC-2618 was found
to block the proliferation of primary neoplastic MC
obtained from patients with ASM or MCL. Moreover,
DCC-2618 exerted major antineoplastic effects on AHN
cells and endothelial cells, all of which may be relevant in
the pathogenesis of advanced SM. Based on these obser-
vations DCC-2618 is a novel emerging drug candidate for
advanced SM. Indeed, clinical trials with DCC-2618 have
been started recently.
The multi-kinase inhibitor midostaurin (PKC412) is
effective against the D816V-mutated variant of KIT and
has shown promising results in patients with advanced
SM in a global phase II trial, with an overall response rate
of 60%.28 In addition, midostaurin was found to suppress
mediator-related symptoms and IgE-dependent histamine
release from basophils.28,42 However, despite clinical effi-

Figures 4. Effects of DCC-2618 on anti-IgE-induced histamine release from normal

basophils. (A) Primary blood basophils from healthy donors were incubated in control medi-
um (0 μM) or in various concentrations of DCC-2618, as indicated, at 37°C for 30 min.
Thereafter, cells were incubated in control buffer or in buffer containing anti-IgE antibody
E-124.2.8 (1 μg/mL) at 37°C for 30 min. After incubation, cells were centrifuged at 4°C,
and cell-free supernatants and cell suspensions recovered and examined for histamine-
content by radioimmunoassay. Histamine release was calculated as percent of total hista-
mine and is expressed as percent of control. Results represent the mean±S.D. of four inde-
pendent experiments. Asterisk (*): P<0.05 compared to control medium. (B) HMC-1 cells
were cultured in the presence or absence of DCC-2618 (HMC-1.1: 25 nM; HMC-1.2: 500
nM) over a 6-day period. Spontaneous release of tryptase from HMC-1.1 and HMC-1.2 cells
was measured by determining tryptase concentrations in cell-free supernatants and
lysates. Tryptase release is expressed as percent of total (intra- and extra-cellular) tryptase.
Results represent the mean±S.D. of three independent experiments. Asterisk (*): P<0.05.

haematologica | 2018; 103(5) 805

 M. Schneeweiss et al.

Figures 5. Effects of DCC-2618 and DP-5439 on proliferation and survival of acute myeloid leukemia (AML) and chronic myelomonocytic leukemia (CMML). (A,C)
MOLM-13, MV4-11, KG-1, U937 and primary leukemic cells were incubated in control medium (0 μM) or medium containing various concentrations of DCC-2618 or
DP-5439, as indicated at 37°C for 48 h. Thereafter, 3H-thymidine uptake was determined. Results in (A) are expressed as percent of control and represent the
mean±S.D. from three independent experiments. Asterisk (*): P<0.05 compared to control medium. Results of (C) are expressed as percent of control and represent
the mean±S.D. from triplicates. (B) MOLM-13, MV4-11, KG-1 and U937 cells were incubated with control medium (0 nM) or various concentrations of DCC-2618 and
DP-5439, as indicated, for 48 h. Thereafter cells were harvested and the percentage of apoptotic cells was quantified morphologically on Wright-Giemsa-stained
cytospin preparations (left panels) or by flow cytometry (determination of annexinV/PI-positive cells, right panels). Results represent the mean±S.D. of three inde-
pendent experiments. Asterisk (*): P<0.05 compared to control medium.

806 haematologica | 2018; 103(5)


cacy, midostaurin is unable to produce long-lasting com-
plete remission in all patients.28 Therefore, new drugs and
drug-combinations are currently being tested in the con-
text of advanced SM. DCC-2618 might be a promising
candidate for several reasons. First, DCC-2618 exhibits a
broad target profile and is able to block growth of various
neoplastic cells.36 In the current study, DCC-2618 was
found to block growth of neoplastic cells obtained from
patients with ASM and MCL. In addition, the drug pro-
duced growth inhibition in all MCL-like cell lines tested,
including KIT-mutated cells and cell lines in which other
oncogenic pathways (such as the RAS pathway) trigger
malignant cell growth. Moreover, unlike other KIT-target-
ing drugs, DCC-2618 is able to suppress the growth and
survival of other cell types relevant to advanced SM and
AHN, including monocytes, blast cells, neoplastic
eosinophils and endothelial cells. The concentrations
required to mediate these cellular inhibitory effects are
readily achievable based on the recent report of clinical
exposure of 5 μM or higher in patients with gastrointesti-
nal stroma tumors.43
After intake, DCC-2618 is considered to be converted
to one active metabolite, DP-5439. We therefore investi-
gated whether DP-5439 is also able to counteract growth
and survival of neoplastic cells. In these experiments, we
were able to show that DP-5439 is able to suppress
growth and survival of neoplastic MC and of other
leukemic (non-MC-lineage) cells in the same way (and
with comparable IC50 values) as DCC-2618. These data
suggest that DCC-2618 treatment should be effective
even if the DP-5439 metabolite may accumulate over
time.
It is well known that about one-third of all patients
with advanced SM have an AHN at diagnosis. Of these
patients, most have a myeloid neoplasm, often in the
form of CMML or AML.2-8,13,32 The treatment of these SM-
AHN patients is a clinical challenge because the AHN is
often drug-resistant. In fact patients with SM-AHN still
have a poor prognosis with an overall survival time of
about 24 months.13,32 Because of its broad activity profile,
we asked whether DCC-2618 might be a promising agent
for patients with SM-AHN. In a first step, we found that
DCC-2618 is a potent inhibitor of proliferation and sur-
vival of the FLT3-mutated AML cell lines MOLM-13 and
MV4-11. DCC-2618 also inhibits the growth of other
AML cell lines examined (KG-1 and U937), but at IC50 val-
ues considerably higher than those for MOLM-13 or
MV4-11 cells. We also found that DCC-2618 counteracts
proliferation of primary leukemic cells obtained from
patients with SM-AHN, AML or CMML (Table 1 and
Online Supplementary Table S1). These findings suggest
that DCC-2618 may be a promising agent for SM-AHN.
In SM patients, disease progression is often accompa-
nied by expansion of neoplastic eosinophils, sometimes
even resembling (chronic) eosinophilic leukemia. In most
cases the eosinophils are of clonal origin as they express
KIT D816V.33 In rare cases, neoplastic eosinophils display
the FIP1L1/PDGFRA fusion gene.34 However, this fusion
gene is usually detectable only in eosinophilic neoplasms,
such as CEL. Since DCC-2618 is known to exert inhibito-
ry effects against PDGFRA35 we examined its effects on
EOL-1 cells harboring FIP1L1-PDGFRA. DCC-2618 was
found to exert strong anti-proliferative and apoptosis-
inducing effects in EOL-1 cells, with IC50 values in the low
nanomolar range. In addition, DCC-2618 was found to
haematologica | 2018; 103(5)
DCC-2618: a new drug against mastocytosis
inhibit growth of primary eosinophils obtained from
patients with KIT D816V-positive SM or reactive hypere-
osinophilia. Together, these data suggest that DCC-2618
inhibits multiple AHN-related cell types, which may be
relevant clinically as progression of SM is often accompa-
nied by multilineage expansion of various sub-clones,
including cells harboring or lacking KIT D816V.15,28,29
A number of different pro-oncogenic pathways and tar-
gets may be involved in KIT D816V-dependent expansion
and accumulation of MC in advanced SM.25,40,44-50 Several
of these target pathways may be sensitive to therapy
with tyrosine kinase inhibitors. We studied whether key
target pathways in neoplastic MC can be disrupted by
DCC-2618. As assessed by Western blotting, DCC-2618
was found to block the phosphorylation and thus activa-
tion of wild-type KIT and KIT D816V. In addition, we
were able to show that DCC-2618 blocks the activation
of AKT, ERK and STAT5, suggesting that multiple target
pathways are accessible to this drug. By contrast, howev-
er, the drug did not disrupt activation of BTK, another
important target displayed by neoplastic MC.46
Since the target spectrum of midostaurin (PKC412) and
DCC-2618 is not identical, we were also interested to
learn whether DCC-2618 and midostaurin can produce
synergistic antineoplastic effects on neoplastic MC.
Indeed, we found that both drugs induce cooperative or
even synergistic growth-inhibitory effects on HMC-1.1
and HMC-1.2 cells.
Specific alterations in the microenvironment, including
increased angiogenesis, are frequently detectable in
advanced bone marrow neoplasms and are often consid-
ered to play an important role in disease progression. A
typical finding in the affected bone marrow of patients
with advanced SM is increased microvessel density.41 We
found that DCC-2618 inhibits the proliferation of human
endothelial cells, including HUVEC and a microvascular
endothelial cell line, HMEC-1. These data suggest that
DCC-2618 also acts as an anti-angiogenic agent.
Interestingly, the IC50 values obtained for HMEC-1 cells
were higher than those for HUVEC, which may be
explained by the fact that HMEC-1 is a cell line, whereas
HUVEC are primary cells. An alternative explanation
would be the lack of key targets in HMEC-1 cells.
Indeed, it is well known, that KDR, a key target of DCC-
2618, is only expressed in HUVEC but not in HMEC-1
cells.
Patients with SM frequently suffer from symptoms
produced by MC-derived mediators.4-6,16 These mediators
are released on IgE-dependent activation of MC and may
cause severe problems or even lead to life-threatening
anaphylaxis. Concomitant (IgE-dependent) allergies are,
therefore, relevant comorbidities in the context of SM.
We found that DCC-2618 counteracts IgE-dependent
secretion of histamine from basophils obtained from
healthy donors. In addition, we were able to show that
DCC-2618 blocks IgE-independent, spontaneous release
of tryptase from HMC-1.1 and HMC-1.2 cells in vitro.
These results suggest that, apart from its antineoplastic
effects, DCC-2618 might also have an impact on media-
tor release (and probably on the resulting symptoms) in
patients with SM with concomitant allergies. Whether
these data can be reproduced in vivo and whether the drug
is able to suppress mediator symptoms in patients with
advanced SM or SM with concomitant allergies remains
to be determined. In fact, whereas the concentrations of
807
 M. Schneeweiss et al.
DCC-2618 required to block tryptase secretion in MC
were rather low (<1.0 μM), the concentrations required to
block IgE-dependent histamine release were rather high
(1 μM).
Collectively, our data indicate that DCC-2618 is a novel
promising agent that counteracts growth and survival of
various cell types relevant to the pathogenesis of
advanced SM. Whether DCC-2618 is able to block
References
1. Metcalfe DD. Classification and diagnosis
of mastocytosis: current status. J Invest 16.
Dermatol. 1991;96(3):2S-4S.
2. Valent P. Biology, classification and treat-
ment of human mastocytosis. Wien Klin
Wochenschr. 1996;108(13):385-397. 17.
3. Horny HP, Valent P. Diagnosis of mastocy-
tosis: general histopathological aspects,
morphological criteria, and immunohisto-
chemical findings. Leuk Res. 2001;25(7):
543-551.
4. Escribano L, Akin C, Castells M, et al.
Mastocytosis: current concepts in diagnosis 18.
and treatment. Ann Hematol. 2002;81(12):
677-690.
5. Valent P, Akin C, Sperr WR, et al. Diagnosis
and treatment of systemic mastocytosis:
state of the art. Br J Haematol. 2003;122(5):
695-717.
6. Akin C, Metcalfe DD. Systemic mastocyto- 19.
sis. Annu Rev Med. 2004;55:419-432.
7. Arock M, Valent P. Pathogenesis, classifica-
tion and treatment of mastocytosis: state of
the art in 2010 and future perspectives.
Expert Rev Hematol. 2010;3(4):497-516. 20.
8. Valent P, Horny H-P, Escribano L, et al.
Diagnostic criteria and classification of mas-
tocytosis: a consensus proposal. Leuk Res.
2001;25(27):603-625.
9. Valent P, Horny H-P, Li CY, et al. 21.
Mastocytosis (mast cell disease). World
Health Organization (WHO) Classification
of Tumours. Pathology & Genetics.
Tumours of Haematopoietic and Lymphoid
Tissues. Eds: Jaffe ES, Harris NL, Stein H, 22.
Vardiman JW 2001;1:291-302.
10. Valent P, Akin C, Metcalfe DD.
Mastocytosis: 2016 updated WHO classifi-
cation and novel emerging treatment con-
cepts. Blood. 2017;129(11):1420-1427. 23.
11. Horny HP, Akin C, Arber D, et al.
Mastocytosis. In: WHO Classification of
Tumours of Haematopoietic and Lymphoid
Tissues. Eds: Swerdlow SH, Campo E,
Harris NL et al. IARC Press Lyon, France,
vol 3, pp 62-69, 2017(2):62-69. 24.
12. Lim KH, Tefferi A, Lasho TL, et al. Systemic
mastocytosis in 342 consecutive adults: sur-
vival studies and prognostic factors. Blood.
2009;113(23):5727-5736.
13. Pardanani A, Lim KH, Lasho TL, et al. 25.
Prognostically relevant breakdown of 123
patients with systemic mastocytosis associ-
ated with other myeloid malignancies.
Blood. 2009;114(18):3769-3772.
14. Escribano L, Alvarez-Twose I, Sánchez-
Muñoz L, et al. Prognosis in adult indolent
systemic mastocytosis: a long-term study of 26.
the Spanish Network on Mastocytosis in a
series of 145 patients. J Allergy Clin
Immunol. 2009;124(3):514-521.
15. Valent P, Akin C, Sperr WR, et al. Aggressive
808
growth of neoplastic MC in patients with advanced SM is
currently being explored in a clinical trial
(NCT02571036).
Acknowledgments
This study was supported by a grant from Deciphera
Pharmaceuticals LLC. We like to thank Dr. Dan Flynn, Dr.
Bryan Smith and Dr. Oliver Rosen for helpful discussion.
systemic mastocytosis and related mast cell
disorders: current treatment options and
proposed response criteria. Leuk Res.
2003;27(7):635-641.
Arock M, Akin C, Hermine O, et al. Current
treatment options in patients with mastocy-
tosis: status in 2015 and future perspectives.
Eur J Haematol. 2015;94(6):474-490.
Furitsu T, Tsujimura T, Tono T, et al.
Identification of mutations in the coding
sequence of the proto-oncogene c-kit in a
human mast cell leukemia cell line causing
ligand-independent activation of the c-kit
product. J Clin Invest. 1993;92(4):1736-
1744.
Nagata H, Worobec AS, Oh CK, et al.
Identification of a point mutation in the cat-
alytic domain of the protooncogene c-kit in
peripheral blood mononuclear cells of
patients who have mastocytosis with an
associated hematologic disorder. Proc Natl
Acad Sci. 1995;92(23):10560-10564.
Feger F, Ribadeau Dumas A, et al. Kit and c-
kit mutations in mastocytosis: a short
overview with special reference to novel
molecular and diagnostic concepts. Int Arch
Allergy Immunol. 2002;127(2):110-114.
Longley BJ, Tyrrell L, Lu SZ, et al. Somatic c-
kit activating mutation in urticaria pigmen-
tosa and aggressive mastocytosis: establish-
ment of clonality in a human mast cell neo-
plasm. Nat Genet. 1996;12(3):312-314.
Fritsche-Polanz R, Jordan JH, Feix A, et al.
Mutation analysis of C-KIT in patients with
myelodysplastic syndromes without mas-
tocytosis and cases of systemic mastocyto-
sis. Br J Haematol. 2001;113(2):357-364.
Arock M, Sotlar K, Akin C, et al. KIT muta-
tion analysis in mast cell neoplasms: recom-
mendations of the European Competence
Network on Mastocytosis. Leukemia.
2015;29(6):1223-1232.
Shah NP, Lee FY, Luo R, et al. Dasatinib
(BMS-354825) inhibits KITD816V, an ima-
tinib-resistant activating mutation that trig-
gers neoplastic growth in most patients
with systemic mastocytosis. Blood.
2006;108(1):286-291.
Akin C, Brockow K, D'Ambrosio C, et al.
Effects of tyrosine kinase inhibitor STI571
on human mast cells bearing wild-type or
mutated c-kit. Exp Hematol.
2003;31(8):686-692.
Gleixner KV, Mayerhofer M, Sonneck K, et
al. Synergistic growth-inhibitory effects of
two tyrosine kinase inhibitors, dasatinib
and PKC412, on neoplastic mast cells
expressing the D816V-mutated oncogenic
variant of KIT. Haematologica.
2007;92(11):1451-1459.
Gleixner KV, Mayerhofer M, Aichberger KJ,
et al. PKC412 inhibits in vitro growth of
neoplastic human mast cells expressing the
D816V-mutated variant of KIT: comparison
with AMN107, imatinib, and cladribine
(2CdA) and evaluation of cooperative drug
effects. Blood. 2006;107(2):752-759.
27. Fabbro D, Ruetz S, Bodis S, et al. PKC412 —
a protein kinase inhibitor with a broad ther-
apeutic potential. Anticancer Drug Des.
2000;15(1):17-28.
28. Gotlib J, Kluin-Nelemans HC, George TI, et
al. Efficacy and safety of midostaurin in
advanced systemic mastocytosis. N Engl J
Med. 2016;374(26):2530-2541.
29. Gotlib J, Berubé C, Growney JD, et al.
Activity of the tyrosine kinase inhibitor
PKC412 in a patient with mast cell leukemia
with the D816V KIT mutation. Blood.
2005;106(8):2865-2870.
30. Sperr WR, Horny HP, Valent P. Spectrum of
associated clonal hematologic non-mast cell
lineage disorders occurring in patients with
systemic mastocytosis. Int Arch Allergy
Immunol. 2002;127(2):140-142.
31. Horny HP, Sotlar K, Sperr WR, et al.
Systemic mastocytosis with associated
clonal haematological non-mast cell lineage
diseases: a histopathological challenge. J
Clin Pathol. 2004;57(6):604-608.
32. Pardanani A. Systemic mastocytosis in
adults: 2017 update on diagnosis, risk strat-
ification and management. Am J Hematol.
2016;91(11):1146-1159.
33. Pardanani A, Reeder T, Li CY, et al.
Eosinophils are derived from the neoplastic
clone in patients with systemic mastocyto-
sis and eosinophilia. Leuk Res. 2003;27(10):
883-885.
34. Pardanani A, Ketterling RP, Brockman SR, et
al. CHIC2 deletion, a surrogate for FIP1L1-
PDGFRA fusion, occurs in systemic masto-
cytosis associated with eosinophilia and
predicts response to imatinib mesylate ther-
apy. Blood. 2003;102(9):3093-3096.
35. Smith BD, Hood MM, Wise SC, et al. DCC-
2618 is a potent inhibitor of wild-type and
mutant KIT, including refractory exon 17
D816 KIT mutations, and exhibits efficacy
in refractory GIST and AML xenograft
models. Cancer Res. 2015;75(15): abstract
2690.
36. Bai Y, Bandara G, Ching Chan E, et al,
Targeting the KIT activating switch control
pocket: a novel mechanism to inhibit neo-
plastic mast cell proliferation and mast cell
activation. Leukemia. 2013; 27(2):278-285.
37. Butterfield JH, Weiler D, Dewald G, Gleich
GJ. Establishment of an immature mast cell
line from a patient with mast cell leukemia.
Leuk Res. 1988;12(4):345-355.
38. Saleh R, Wedeh G, Herrmann H, et al., A
new human mast cell line expressing a func-
tional IgE receptor converts to factor-inde-
pendence and tumorigenicity by KIT
D816V-transfection. Blood. 2014;124(1):
111-120.
39. Hoermann G, Blatt K, Greiner G, et al.,
CD52 is a molecular target in advanced sys-
temic mastocytosis. FASEB J. 2014;28(8):
haematologica | 2018; 103(5)

3540-3551.
40. Peter B, Gleixner KV, Cerny-Reiterer S, et al.
Polo-like kinase-1 as novel target in neoplas-
tic mast cells: demonstration of growth-
inhibitory effects of siRNA and the Polo-
like kinase-1 targeting drug BI 2536.
Haematologica. 2011;96(5):672-680.
41. Wimazal F, Jordan JH, Sperr WR, et al.
Increased angiogenesis in the bone marrow
of patients with systemic mastocytosis. Am
J Pathol. 2002;160(5):1639-1645.
42. Peter B, Winter GE, Blatt K, et al. Target inter-
action profiling of midostaurin and its
metabolites in neoplastic mast cells predicts
distinct effects on activation and growth.
Leukemia. 2016;30(2):464-472.
43. Janku F, Razak ARA, Gordon MS, et al.
Pharmacokinetic-driven phase I study of
DCC-2618, a pan-KIT and PDGFRA
haematologica | 2018; 103(5)
DCC-2618: a new drug against mastocytosis
inhibitor, in patients with gastrointestinal
stroma tumor (GIST) and other solid
tumors. J Clin Oncol. 2017;35:15(S)2515.
44. Ustun C, Arock M, Kluin-Nelemans HC, et
al. Advanced systemic mastocytosis: from
molecular and genetic progress to clinical
practice. Haematologica. 2016;101(10):
1133-1143.
45. Jawhar M, Schwaab J, Schnittger S, et al.
Additional mutations in SRSF2, ASXL1
and/or RUNX1 identify a high-risk group of
patients with KIT D816V(+) advanced sys-
temic mastocytosis. Leukemia. 2016;30(1):
136-143.
46. Gleixner KV, Mayerhofer M, Cerny-Reiterer
S, et al. KIT-D816V-independent oncogenic
signaling in neoplastic cells in systemic mas-
tocytosis: role of Lyn and BTK activation
and disruption by dasatinib and bosutinib.
Blood. 2011;118(7):1885-1898.
47. Wilson TM, Maric I, Simakova O, et al.
Clonal analysis of NRAS activating muta-
tions in KIT-D816V systemic mastocytosis.
Haematologica. 2011;96(3):459-463.
48. Schwaab J, Schnittger S, Sotlar K, et al.
Comprehensive mutational profiling in
advanced systemic mastocytosis. Blood.
2013;122(14):2460-2466.
49. Bibi S, Langenfeld F, Jeanningros S, et al.
Molecular defects in mastocytosis: KIT and
beyond KIT. Immunol Allergy Clin North
Am. 2014;34(2):239-262.
50. Damaj G, Joris M, Chandesris O, et al.,
ASXL1 but not TET2 mutations adversely
impact overall survival of patients suffering
systemic mastocytosis with associated
clonal hematologic non-mast-cell diseases.
PLoS One. 2014;9(1):e85362.
809
ARTICLE Acute Myeloid Leukemia

Distinct protein signatures of acute myeloid

Ferrata Storti
Foundation leukemia bone marrow-derived stromal cells
are prognostic for patient survival
Steven M. Kornblau,1* Peter P. Ruvolo,1* Rui-Yu Wang,1 V. Lokesh Battula,1
Elizabeth J. Shpall,2 Vivian R. Ruvolo,1 Teresa McQueen,1 YiHua Qui,1
Zhihong Zeng,1 Sherry Pierce,1 Rodrigo Jacamo,1 Suk-Young Yoo,3
Phuong M. Le,1 Jeffrey Sun,1 Numsen Hail Jr,1 Marina Konopleva1 and Michael
Andreeff1

1
Section of Molecular Hematology and Therapy, Department of Leukemia; 2Department
of Stem Cell Transplantation and 3Bioinformatics and Computational Biology, The
Haematologica 2018
Volume 103(5):810-821 University of Texas M.D. Anderson Cancer Center, Houston, Texas, TX, USA
*SMK and PPR contributed equally to this work.

ABSTRACT

M
esenchymal stromal cells (MSC) support acute myeloid
leukemia (AML) cell survival in the bone marrow (BM)
microenvironment. Protein expression profiles of AML-derived
MSC are unknown. Reverse phase protein array analysis was performed
to compare expression of 151 proteins from AML-MSC (n=106) with
MSC from healthy donors (n=71). Protein expression differed signifi-
cantly between the two groups with 19 proteins over-expressed in
leukemia stromal cells and 9 over-expressed in normal stromal cells.
Unbiased hierarchical clustering analysis of the samples using these 28
proteins revealed three protein constellations whose variation in expres-
sion defined four MSC protein expression signatures: Class 1, Class 2,
Class 3, and Class 4. These cell populations appear to have clinical rele-
vance. Specifically, patients with Class 3 cells have longer survival and
remission duration compared to other groups. Comparison of leukemia
Correspondence: MSC at first diagnosis with those obtained at salvage (i.e. relapse/refrac-
skornblau@mdanderson.org or
tory) showed differential expression of 9 proteins reflecting a shift
mandreef@mdanderson.org toward osteogenic differentiation. Leukemia MSC are more senescent
compared to their normal counterparts, possibly due to the over-
expressed p53/p21 axis as confirmed by high β-galactosidase staining. In
Received: May 9, 2017. addition, overexpression of BCL-XL in leukemia MSC might give sur-
Accepted: February 1, 2018. vival advantage under conditions of senescence or stress and over-
expressed galectin-3 exerts profound immunosuppression. Together, our
Pre-published: March 15, 2018. findings suggest that the identification of specific populations of MSC in
AML patients may be an important determinant of therapeutic response.
doi:10.3324/haematol.2017.172429

Check the online version for the most updated Introduction

information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/810
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
There is growing evidence to support the importance of the leukemia bone mar-
row (BM) niche in the process of acute myeloid leukemia (AML) chemoresis-
tance.1,2 Hence, optimal therapeutic strategies should also address neighboring cells
in the tumor microenvironment. The critical support cells in the leukemia BM
microenvironment are mesenchymal stromal cells (MSC).3-8 Depending on the
type, MSC can act either to support or suppress tumors.4,8-15 Our group and others
have found that MSC support leukemia cell survival by diverse mechanisms that
include secretion of cytokines and chemokines, activation of survival signaling in
tumor cells, and blocking immune surveillance by suppressing natural killer (NK)
and T cells.2-5,13
Mesenchymal stromal cells are essential for human hematopoiesis, particularly
as a source of SDF-1, which regulates homing, proliferation, and differentia-
tion.6,9,10,16-18 Moreover, studies from our group and others have demonstrated that
MSC protect leukemia cells from chemotherapy.6,19-23 We have recently found that
there is reciprocal activation of NFkB signaling between MSC and AML and acute
lymphoblastic leukemia (ALL) cells that likely contribute to the effectiveness of the
microenvironment to protect malignant cells.7 Medyouf et al. recently demonstrat-

810 haematologica | 2018; 103(5)


ed that blast cells from myelodysplastic syndrome (MDS)
patients induce changes in MSC reflecting reprograming
of the stromal cells.24 MSC may also influence hematopoi-
etic precursors to promote leukemogenesis as evidenced
by the development of AML and MDS in mice where the
MSC osteo-progenitors were engineered to lack Dicer, a
key regulator of microRNA (miR) processing.2
Furthermore, a recent study from Zhao et al. reported that
p21 could be critical for inducing senescence in MSC from
MDS patients with concomitant induction of interleukin-
6 (IL6) and transforming growth factor β (TGF-β).25 This
study is consistent with findings that support the role of
Dicer in regulating MSC biology and also establish a pos-
sible mechanism of aberrant survival functions in malig-
nant MSC that may be associated with p21 and senes-
cence. The ability of malignant MSC to withstand senes-
cence may depend on the expression of the anti-apoptotic
molecule BCL-XL.26,27
The cellular composition of stromal cells in a cancer
microenvironment, such as the leukemic BM niche, is like-
ly markedly different from that of the normal BM. We,
therefore, set out to study the protein expression and acti-
vation in leukemic MSC (AML-MSC) and compared and
contracted these to normal MSC (NL-MSC) to determine
if and how they are functionally different. Reverse phase
protein array analysis (RPPA, pioneered in our laborato-
ry)28-32 was used to examine expression of 151 proteins in
MSC derived from AML BM (n=106) with those derived
from healthy donors (n=71). The results presented here
identify 28 that were differentially expressed between the
two. Importantly, the 28 proteins identified as differential-
ly expressed in the AML versus normal MSC could be
grouped into four protein constellation (PC) expression
signatures with different biological properties and clinical
implications regarding patient response to therapy.
Methods
Patients’ samples
Bone marrow was obtained from AML patients (n=106) under-
going diagnostic BM aspiration and from healthy donors (n=71)
who were undergoing BM harvest for use in allogeneic BM trans-
plantation. Samples were acquired in accordance with the regula-
tions and protocols approved by the Investigational Review Board
of MD Anderson Cancer Center. Informed consent was obtained
in accordance with the Declaration of Helsinki. Samples were ana-
lyzed under an Institutional Review Board-approved laboratory
protocol. Patients' characteristics are presented in Table 1. Details
of isolation of MSC are available in the Online Supplementary
Methods.
RPPA
Proteomic profiling was carried out on MSC samples from
patients with AML and healthy donors using RPPA. The RPPA
method and sample validation technique are described fully else-
where.28-32 Antibodies against 151 proteins were used for analysis.
(A list and the source of the antibodies and the concentrations uti-
lized is provided in the Online Supplementary Table S1). The sources
of antibodies have been been reported previously.30 An IgG sub-
type-specific secondary antibody was used to amplify the signal,
and finally a stable dye was precipitated. The stained slides were
analyzed using the Microvigene software (version 3.0, Vigene
Tech, Carlise, MA, USA) to produce quantified data. Statistical
analyses are described in the Online Supplementary Methods.
haematologica | 2018; 103(5)
Proteomic profiling reveals complexity of AML MSC
Cell senescence assessment
Microscopy assessment of β-galactosidase staining was used to
detect cell senescence using a detection kit from Cell Signal
Technology (Boston, MA, USA). Early passage cells (passage 2)
were imaged using a Nikon Coolpix 950 camera attached to a
Nikon TMS light microscope (Nikon Instruments Inc.). AML-MSC
(n=4) and NL-MSC (n=5) were lysed in kit buffer. Measurement of
β-galactosidase was performed using an in vitro fluorometric assay
with fluorescein di-β-D-galactopyranoside (FDG) as substrate.
Incubation time was 2 hours (h). Fluorescence was measured
using an Optima Fluorometer (Durham, NC, USA). Activity is pre-
sented as fluorescence units/1000 cells/minute.
Pathway analysis
String software (String 10.1; available from: http://string-db.org)33
was used to determine protein associations. Pathway analysis to
identify canonical pathways, upstream regulators, and protein net-
works was performed using Ingenuity Pathway software
(Qiagen).
Results
Proteins are differentially expressed in AML versus
healthy MSC
We have routinely utilized RPPA to analyze protein
expression from clinical samples from many hematologic
malignancies.28-32 We examined protein expression in blasts
from newly diagnosed AML patients (n=85), CD34+ cells
from normal donors (n=10), MSC from healthy donors
(n=71), and MSC from newly diagnosed AML patients
(n=54). Both normal MSC and AML-MSC expressed MSC
defining lineage markers CD73, CD90, and CD 105 as
determined by flow cytometry (Online Supplementary
Figure S1). MSC from salvage samples (i.e. relapse/refrac-
tory) were also studied (n=46). The RPPA was probed
with 151 antibodies targeting 119 different proteins (114
targeting total protein with 32 paired antibodies targeting
phosphoepitopes on 26 proteins, and 5 with only a phos-
phoepitope but not total protein epitope) covering a wide
variety of cellular functions and pathways (Online
Supplementary Table S1). Protein expression in AML-MSC,
NL-MSC, AML blasts and normal CD34+ cells was com-
pared using principle component analysis (Online
Supplementary Figure S1) and unbiased hierarchical cluster-
ing (Online Supplementary Figure S2B). NL-MSC and AML-
MSC formed a cluster distinct from AML blasts and NL-
CD34+ cells with the vast majority of the 151 proteins test-
ed showing statistically significant differential expression
(143 of 151, P=0.01; 124 of 151, P≤10-6) between MSC and
the blast/CD34+ cells. This unsurprising observation is
consistent with a previous report that gene expression
profiles are distinct between blood cells and MSC.34
Principal component analysis (PCA) also shows that pro-
tein expression in NL-CD34+ cells is distinct from that of
AML blasts. These findings were identical to those
observed when the analysis was restricted to samples
from newly diagnosed patients alone.
Next we investigated whether protein expression in
AML-MSC was different from that of NL-MSC. In PCA,
the NL-MSC occupied a distinct space from that of the
AML-MSC (Figure 1A). Unbiased hierarchical clustering
comparing AML-MSC and NL-MSC revealed differential
expression of 28 of those proteins (P<0.001; Q=0.0059).
The Q-value, a measure of the false discovery rate deter-
mined by a β-uniform mixture model highlights that these
811
 S.M. Kornblau et al.

differences are very unlikely to be random and suggest The proteins clustered into three PCs (from top to bottom
that there are significant differences in the protein expres- in Figure 1B). Nineteen of these proteins had generally
sion patterns of AML-MSC relative to those of NL-MSC. higher expression in AML-MSC, including P1 [STAT1, p-

Table 1. Patients’ demographics (see Figures 2A and 3A).

Variable Category Total Class 2 Class 1 Class 3 Class 4 P
Cohort size Number 101 6 13 18 64
Percentage 100.00% 5.90% 12.90% 17.80% 63.40%
Status New 53.50% 66.70% 53.80% 66.70% 48.40% 0.55
Salvage 45.50% 33.30% 46.20% 33.30% 50.00%
Sex M 61.40% 33.30% 30.80% 72.20% 67.20% 0.03
F 38.60% 66.70% 69.20% 27.80% 32.80%
Race Asian 6.90% 0.00% 7.70% 11.10% 6.30% 0.5
Black 7.90% 16.70% 15.40% 0.00% 7.80%
White 73.30% 50.00% 69.20% 83.30% 73.40%
Hispanic 9.90% 33.30% 7.70% 5.60% 9.40%
Unknown 3.00% 0.00% 7.70% 0.00% 3.10%
FAB M0 4.00% 0.00% 0.00% 5.60% 4.70% 0.21
M1 16.80% 16.70% 38.50% 5.60% 15.60%
M2 24.80% 16.70% 23.10% 22.20% 26.60%
M4 26.70% 66.70% 23.10% 27.80% 23.40%
M5 12.90% 0.00% 7.70% 16.70% 14.10%
M6 3.00% 0.00% 0.00% 11.10% 1.60%
RAEB-T 4.00% 0.00% 0.00% 0.00% 6.30%
Unk 7.90% 0.00% 7.70% 11.10% 7.80%
PS Asymptomatic 15.80% 16.70% 15.40% 11.10% 17.20% 0.44
Symptoms 62.40% 50.00% 46.20% 77.80% 62.50%
In bed <50 13.90% 33.30% 23.10% 5.60% 12.50%
In bed >50 2.00% 0.00% 7.70% 0.00% 1.60%
100% bedridden 1.00% 0.00% 0.00% 5.60% 0.00%
AHD N 75.20% 50.00% 84.60% 83.30% 73.40% 0.33
Y 24.80% 50.00% 15.40% 16.70% 26.60%
Cyto Favorable 2.00% 16.70% 7.70% 0.00% 0.00% 0.026
Intermediate 54.50% 83.30% 53.80% 50.00% 53.10%
Unfavorable 43.60% 0.00% 38.50% 50.00% 46.90%
FLT3 Negative 65.30% 33.30% 69.20% 66.70% 64.10% 0.0979
Positive 20.80% 16.70% 23.10% 16.70% 21.90%
Not Done 8.90% 50.00% 7.70% 16.70% 14.10%
D835 Negative 79.20% 50.00% 92.30% 77.80% 76.60% 0.005
Positive 6.90% 0.00% 0.00% 5.60% 9.40%
Not Done 8.90% 50.00% 7.70% 16.70% 14.10%
NPM1 Negative 46.50% 50.00% 69.20% 72.20% 45.30% 0.08
Positive 10.90% 0.00% 7.70% 0.00% 17.20%
Not Done 5.00% 50.00% 23.10% 27.80% 37.50%
Response CR 58.80% 66.70% 71.40% 63.60% 53.30% 0.8
PR 29.70% 0.00% 0.00% 0.00% 3.30%
HI 1.00% 0.00% 0.00% 9.10% 6.70%
No response 3.00% 33.30% 14.30% 9.10% 13.30%
FAIL 6.90% 0.00% 14.30% 18.20% 23.30%
Relapse Yes 70.00% 100.00% 80.00% 28.60% 81.30% 0.076
Vital Alive 21.60% 0.00% 28.60% 36.40% 16.70% 0.41
Dead 78.40% 100.00% 71.40% 63.60% 83.30%
M: male; F: female; FAB: French-American-British Classification; PS: propensity score; AHD: antecedent hematologic disease; N: no; Y: yes; Cyto: cytogenetic profile; CR: complete
response; PR: partial response; HI: hematologic improvement;*Statistically significant (P) sets are in bold.

812 haematologica | 2018; 103(5)

 Proteomic profiling reveals complexity of AML MSC

PDK1 (S241)], CCND1, CDKN1A (p21), ITGA2, PARP1,
PPP2R2A/B/C/D, the PP2A B regulatory subunit family
B55, BAK1, CSNK2A1, CDK4, GSK3A/B) and PC2
[STAT5A/B, BCL2L1 (BCL-XL)], DIABLO, TP53 (p53),
NOTCH 1 (cleaved 1744), SPP1, p-EGFR (Y992), and
ERBB2). Expression of 17 of the 19 proteins was validated
by immunoblot analysis (Online Supplementary Figure S3).
Although expression of EGFR and ERBB2 expression could
not be confirmed by western blot analysis, this may reflect
the enhanced sensitivity of RPPA over standard
immunoblot technology. The remaining nine proteins in
PC3 were elevated in healthy donor MSC compared to
AML-MSC: SMAD1, CREB1.p133 STMN1, SIRT1, CREB1
SMAD4, p-Foxo1/3 (S32), HSP90AA1/B1, and EIF2S1.

Figure 1. Mesenchymal stromal cell (MSC) protein expression signatures. Protein expression is distinct between normal MSC and acute myeloid leukemia (AML)-
MSC. (A) Principal component analysis (PCA) of 151 proteins examined in Class 1 (yellow), Class 2 (light blue), Class 3 (orange) and Class 4 (dark blue). (B) Unbiased
hierarchical clustering identifies 3 protein signature groups: Group 1 (11 members), Group 2 (8 members) and Group 3 (9 members) in Class 1 (yellow), Class 2
(light blue), Class 3 (orange) and Class 4 (dark blue) groups, identified in top row as “MSC protein type”. MSC derived from normal donor (light blue) or AML patient
(dark blue) is shown in the second row marked “cell type”.

haematologica | 2018; 103(5) 813

 S.M. Kornblau et al.

Table 2. List of proteins with reverse correlation to one or more other proteins in acute myeloid leukemia (AML) mesenchymal stromal cells (MSC)
versus normal MSC.
Protein AML-MSC negative AML-MSC positive
normal MSC positive normal MSC negative
AKT1S1 STAT3
AKT3 BCL2L11, ERG, SFN, SRCp416
ARC STAT3 p727 CTNNB1
ATF3 BCL2, KIT, SFN, TP53 p15
BCL2 ATF3, CTNNB1
BCL2L11 ATF3, CTNNB1
BECN1 PSMD9, YWHAE
BID GAPDH BIRC5, GAB2 p452
BIRC5 MS4A1, FN1, SRC p416 BID, EIF2S1 p51, IRS1 p1101, RPS6KB1
CAV1 PTK2
CCNB1 MAPK1.3 p202.p204
CTNNB1 BCL2, BCL2L11, FOXO1.3 p24.p32, FOXO3, KIT, ITGAL, ARC, HSP90AA1.B1, MAPK9, PTGS2, YWHAZ
PSMD9 p10, RPS6KB1 p389, SFN, SRC, SRC p416
CDK1 CREB1
CDK4 GAPDH, MAPK1.3 p202.p204, YWHAE
CREB1 CDK1, JUN p73 DIABLO, PECAM1, SFN
CREB2 p133 MAPK9 TP53
DIABLO HSPB1 CREB1
EGLN1 FOXO1.3 p24.p32, FOXO3, MS4A1, SRC p416, TCF4,
YWHAE
EIF2S1 p51 BIRC5, TGM2
ELK1 p383 SMAD1
ERBB2 p1248 TCF4
ERG AKT3
FN1 BIRC5
FOXO1.3/FOXO3 p24.p32 CTNNB1 EGLN1
FOXO3 CTNNB1, IRS1 p1101 EGLN1
GAB2 p452 BID
GAPDH BID, CDK4, GSK3A.B
GSK3A.B GAPDH, STK11
HDAC3 PSMD9
HNRNPK STAT5A.B
HSP90AA1.B1 CTNNB1
HSPB1 DIABLO
IRS1 p1101 FOXO3 BIRC5
ITGAL CTNNB1
JMJD6 STAT3 p727
JUN p73 CREB1 NRP1
KIT ATF3, CTNNB1
LEF1 PTK2
MAP2K1 XIAP
MAPK1.3 p202.p204 CCNB1, CDK4 SRC p527
MAPK9 CREB2 p133 CTNNB1
MAPK14 STAT1
MS4A1 BIRC5 EGLN1
NPM1 SMAD1
NR4A1 PECAM1
NRP1 JUN p73
PECAM1 NR4A1 CREB1
continued on the next page

814 haematologica | 2018; 103(5)

 Proteomic profiling reveals complexity of AML MSC

continued from the previous page

PPARG SMAD1
PRKAA1.2 p172 SRC p527
PSMD9 p10 CTNNB1
PSMD9 BECN1, HDAC3
PTGS2 CTNNB1
PTK2 LEF1 CAV1
RPS6KB1 p389 CTNNB1
RPS6KB1 BIRC5
SFN AKT3, ATF3, CTNNB1 CREB1, STMN1
SMAD1 ELK1 p383 NPM1, PPARG,
SRC CTNNB1
SRC p416 AKT3, BIRC5, CTNNB1, XIAP EGLN1
SRC p527 PRKAA1.2 p172, MAPK1.3 p202.p204
STAT1 MAPK14, STAT5A.B
STAT3 p727 AKT1S1
STAT3 p727 ARC, JMJD6
STAT5A.B STAT1 HNRNPK
STK11 GSK3A.B
STMN1 BCL2L11, SFN
TCF4 ERBB2 p1248 EGLN1
TGM2 EIF2S1 p51
TP53 CREB2 p133
TP53 p15 ATF3
XIAP SRC p416 MAP2K1
YWHAE BECN1, CDK4 EGLN1
YWHAZ CTNNB1

Based on various combinations of proteins expressed Clinical correlation

within these three groups, we observed that normal or
leukemic MSC samples clustered into four distinct PCs.
The first (aqua in Figure 1A; top row of annotations above
the heat map in Figure 1B), was comprised predominantly
of NL-MSC (42 of 46 members), characterized by lower
expression of proteins in constellations 1 and 2, and higher
expression of constellation 3 proteins. This was consid-
ered to represent the protein expression pattern of NL-
MSC. AML cases in this signature are hereafter called
Class 2.
A second signature (blue in Figure 1A and B) was com-
prised predominantly of AML-MSC (62 of 72 members),
and was characterized by the opposite protein expres-
sion of the NL-MSC signature, with high expression of
PC1 and 2 proteins and lower expression of PC3 pro-
teins. This signature is called Class 4. Two groups with
expression signatures between that of the Class 2 MSC
and Class 4 MSC were also recognized. One signature
(yellow in Figure 1A and top row of Figure 1B), com-
prised of 12 AML-MSC and 11 NL-MSC, mirrored the
expression of Class 2 for protein constellation 1, but that
of Class 4 for PC2 and 3. AML cases in this group are
hereafter referred to as Class 1. The final signature
(orange in Figure 1A and top row of Figure 1B), consist-
ing of 16 AML-MSC and 8 NL-MSC, mirrored the
expression of Class 4 for PC1 and 3 but had low expres-
sion of PC2, similar to Class 2. AML cases from this sig-
nature are referred to as Class 3.
Among the AML cases, there were very few clinical or
laboratory features that showed a statistically significant
correlation with the four MSC PCs. Specifically there was
no association for continuous variables: age, percent BM
or blood blasts, hemoglobin, platelet count, serum lactate
dehydrogenase, albumin, creatinine, bilirubin fibrinogen,
CD7, CD10, CD13, CD14, CD20, CD33 or CD34, or dis-
crete variables disease status (new vs. salvage), race,
French-American-British (FAB) Classification, perform-
ance status, history of an antecedent hematologic disor-
der, presence of an FLT3-ITD or presence of an NPM1
mutation (Table 1). However, a few notable associations
between MSC protein signature and clinical features were
observed. There was a statistical difference in presence of
FLT3 D835 mutation (P=0.05). While no Class 1 or Class 2
samples contained FLT3 D835 mutation, 9.4% of Class 4
MSC and 5.6% of Class 3 MSC had the mutation (Table
1). Interestingly, there was a statistical difference in MSC
population types between men and women (P=0.03)
(Table 1). Men had a higher percentage of Class 4 and
Class 3 MSC while women had a higher percentage of
Class 1 and Class 2 MSC. At present it is unclear if sex
influences MSC biology in the leukemic niche. Sex-specif-
ic effects in the microenvironment in leukemia is not
unprecedented as integrin-mediated adhesion-triggered
activation of GSK3B occurs exclusively in leukemia pro-
genitor cells derived from female AML patients.35 Total
GSK3 and ITGA2 are present in protein constellation 1

haematologica | 2018; 103(5) 815

 S.M. Kornblau et al.

and the proteins in this constellation are elevated in Class
3 and Class 4 while they are lower in Class 1 and Class 2.
Perhaps the integrin/GSK3 axis is also important in AML-
MSC.
Finally, there were statistically significant biases with
different cytogenetic groups associating with different
MSC protein signatures (P=0.026) (Table 1). Specifically,
although few patients had favorable cytogenetics in this
dataset, these were found exclusively in the Class 1 and
Class 2 signatures. Conversely, patients with unfavor-
able cytogenetics were not observed in the Class 2 sig-
nature, but they were found in relatively equal propor-
tions in the three AML specific signatures (i.e. Classes 1,
3, and 4).
Having determined that leukemic MSC divide AML
patients into discrete protein signatures, we next exam-
ined whether signature membership affected outcomes.
There was no significant difference in median overall sur-
vival between the four MSC populations, although the
median survival of 105 weeks (wk) in the Class 3 signature
was longer than the other three groups at 25, 48 and 58
wk (Figure 2A); this suggested that patients with Class 3
MSC may have a better disease prognosis. Despite the
small sample size, relapse rates among the newly diag-
nosed cases were different, with 4 of 5 Class 1 MSC
patients and 13 of 16 Class 4 MSC patients relapsing com-
pared to only 2 of 7 Class 3 MSC cases (P=0.076).
Furthermore the median remission duration was different
between patients with Class 3 MSC and the three other
signatures (101 wk for Class 3 MSC vs. 42.2 wk for Class
4 MSC; P=0.05) (Figure 2B). Other comparisons were not
made due to the small sample size. These findings suggest
that MSC can influence patient survival, although at pres-
ent the mechanism is unknown.
These findings suggest that the disease state dictates
expression of at least some of the identified proteins in
MSC.
Evidence for dysregulated signaling in AML-MSC
The observed differences in protein expression between
AML-MSC and NL-MSC suggest that pathway utilization
may be dysregulated or non-canonical in the AML-MSC.
To look for evidence of abnormal utilization, we searched
for statistically significant (R>0.2; P< 0.0001) protein-pro-
tein correlations that were reversed in the direction of cor-
relation in the AML-MSC compared to the pattern in the
NL-MSC. A representative analysis of STAT5 expression
with the other proteins in NL-MSC revealed that this tran-
scription factor is negatively correlated with hnRPK and
positively correlated with STAT1 (Online Supplementary
Figure S5B). However, in AML-MSC the correlations are
reversed (Online Supplementary Figure S5B).
A list of all protein correlations that are reversed in
AML-MSC compared to NL-MSC is provided in Table 2.
Of note, there are significant changes in protein correla-
tions involving β-catenin. As shown in Table 2, BCL2,
BCL2L11, FOXO1.3 p24.p32, FOXO3, KIT, ITGAL,
PSMD9 p10, RPS6KB1 p389, SFN, SRC, and SRC p416
are negatively correlated with β-catenin in NL-MSC but
A
P=0.26

Effect of age on protein expression in NL-MSC and P=0.6

AML-MSC
As AML is a disease of the elderly, many of the AML-
MSC samples were obtained from individuals aged over
60 years (y). Of the 106 AML-MSC samples, 50 samples
were from individuals under 60 y. Of the 71 MSC samples
from healthy donors, 68 samples were from individuals
under 60 y. To assess the impact of age on the differences
in protein expression between the AML-MSC and NL-
MSC groups, the protein expression of 24 of the differen- B
tially-expressed proteins was reassessed in age-matched
sets. The three groups were: a) under 30 y (AML n=13; NL
n=37); b) 30-49 y (AML n=17; NL n=25); and c) 50 to 59 y
(AML n=20; NL n=6). As shown in Online Supplementary
Table S2 and Online Supplementary Figure S4A, of the 19
proteins elevated in AML-MSC compared to NL-MSC,
only two (i.e. BCL-XL and PPP2R2A/B/C/D) were signifi-
cantly higher in all age groups. This suggests that
increased expression of these 2 proteins is dependent on P=0.05
the disease state and not on age. Six other proteins includ-
ing p53 and CDKN1A (p21) were also elevated in two of
the three age groups suggesting expression of these pro-
teins is also dependent on the disease state rather than on P=0.12
age (Online Supplementary Table S2).
Analysis of p53 and CDKN1A (p21) protein expression
was performed by western blot analysis on age-matched
AML and NL samples (age 40-49 y; n=3 for both groups).
Protein expression of both p53 and CDKN1A (p21) was at Figure 2. Survival and remission duration differ in patients based on mesenchy-
least 2- to 3-fold higher in AML-MSC compared to the mal stromal cell (MSC) population. Kaplan-Meir curves showing overall survival
(A) and remission duration (B). N: number.
healthy donor samples (Online Supplementary Figure 4B).

816 haematologica | 2018; 103(5)

 Proteomic profiling reveals complexity of AML MSC

are positively correlated with the protein in AML-MSC.
ARC (NOL3), HSP90AA1/B1, MAPK9, PTGS2, and
YWHAZ were positively correlated with β-catenin in
NL-MSC but are negatively correlated with the protein
in AML-MSC. Ingenuity Pathway Analysis (IPA) was
performed using software on the proteins identified as
differentially correlated with β-catenin in the MSC sets
and β-catenin itself (i.e. CTNNB1; BCL2; BCL2L11;
FOXO1; FOXO3; KIT; ITGAL; PSMD9; RPS6KB1; SFN;
SRC; NOL3; HSP90AA1; HSP90B1; MAPK9; PTGS2;
YWHAZ). IPA revealed these proteins were highly asso-
ciated with PI3K/AKT signaling (top canonical pathway;
P=7.35E-19) (Online Supplementary Figure S5). IPA also
identified the top upstream regulator of this set of pro-
teins as p53 (P=1.04E-11).
Proteins differentially expressed in AML-MSC share
interactomes
To assess the relationship among the proteins identified
in the RPPA analysis, protein association network analysis
was performed using STRING 10.533 on proteins identi-
fied as significantly different in the AML-MSC and NL-
MSC (Figure 1B). Blalock et al. used a previous version of
String software to map the nuclear interactome.36 In cases
where a family of proteins was identified (e.g. PPP2R2 set
and HSP90 set), a representative member was included in
the analysis. With the exception of PDK-1 all proteins are
interconnected at least through one association (Figure 3).
This finding suggests that there is an interconnection
between the various proteins that are distinctly expressed
between the NL-MSC and AML-MSC groups.

Activation
Inhibition
Binding
Phenotype
Catalysis
Post-translational regulation
Reaction
Transcriptional regulation

Figure 3. Proteins differentially expressed in acute myeloid leukemia (AML) and normal mesenchymal stromal cell (MSC) are highly interactive. (A) String analysis
was performed by String 10.0 using interactions based on “action”; available from: http://string-db.org. (B) Model of involvement of Group 1 and 2 proteins in AKT
signaling. Red: proteins are members of Group 1 or 2. Yellow; proteins are non-members but possible links.

haematologica | 2018; 103(5) 817

 S.M. Kornblau et al.
Pathway analysis suggests NL-MSC have prominent
adipogenic signaling while all AML-MSC populations
have prominent PI3K/AKT signaling
Pathway analysis was performed using IPA. Two sepa-
rate data sets were created: one with proteins from PC3
(elevated in normal MSC) and one with proteins from PC1
and 2 (elevated in AML-MSC). Proteins in group 3 are
associated with adipogenesis (ninth top canonical path-
way; P=2.19E-05) (Online Supplementy Figure S6). PC3 pro-
teins are elevated in MSC that were presumably normal
but reduced in AML-MSC suggesting differences in differ-
entiation potential of MSC between NL-MSC and AML-
MSC. Of the 7 proteins identified, SIRT1, FOXO1, and
SMAD1 each can activate PPARβ which is a critical regu-
lator of adipocyte differentiation.37-39 PC3 also displayed
the strongest association with AMPK signaling (P=6.84E-
07). The top upstream regulator identified in the set of
PC3 proteins was PDGFB (P=1.01E-06). PI3K/AKT path-
way was highly associated with PC1 and 2 proteins,
which are elevated in AML-MSC (top canonical path-
ways; P=7.16E-17) (Online Supplementary Figure S7). AKT
818
was identified as one of the top three upstream regulators
(P=1.22E-14), suggesting that signaling mediated by this
survival kinase is prominent in AML-MSC.
AML-MSC are senescent compared to NL-MSC
The p21 protein appears to be critical for senescence in
myelodysplastic syndrome (MDS)-MSC22 and AML-
MSC similar to MDS-MSC have elevated p21 (Figure 1B).
This finding suggests that AML-MSC might also be more
senescent than NL-MSC, as was the case in MDS-MSC.25
Senescence was observed in normal donor MSC- and
AML-derived MSC using β-galactosidase staining. AML-
MSC were more senescent than MSC derived from
healthy donors in this representative example (Figure 4A).
To account for age effects on senescence, MSC were taken
from donors under 58 y. Average age of the AML patients
(n=4) was 52 y and the average age of normal donors (n=5)
was 47 y. Also, MSC of similar cell passage (passage 2 or
3) were used, so effects of cell passage were unlikely. As
shown in Figure 4B, β-galactosidase activity was signifi-
cantly higher (almost 2-fold) in AML-MSC compared to
Figure 4. Acute myeloid leukemia (AML) mesenchymal
stromal cell (AML-MSC) are more senescent than nor-
mal MSC (NL-MSC). (A) Representative microscopy of an
AML-MSC and a normal MSC with two different slide
areas. (B). Level of β-galactosidase was assessed by
enzymatic assay in normal MSC (n=5) and AML-MSC
(n=4). Statistical significance determined by Student
t-test; *P=0.027.
haematologica | 2018; 103(5)

NL-MSC (P=0.032). These findings suggest that AML-
MSC tend toward senescence.
Therapy alters proteins expression in AML-MSC
Protein expression in AML-MSC might change between
diagnosis and relapse, perhaps as a result of relapse, or in
response to acquired changes in AML blasts. To determine
if AML-MSC protein expression was changing between
diagnosis and relapse we compared expression between
the samples obtained at first diagnosis (n=53) to those of
AML-MSC collected from primary refractory or relapsed
patients (n=54). Nine proteins are differentially expressed
between the two MSC groups (Figure 5).
Phosphorylated β-catenin, phosphorylated RPS6, and
galectin-3 are expressed at higher levels in MSC in the sal-
vage set. SMAD6, TCF4, LYN, integrin-β3, phosphorylat-
ed EIF4BP1, and phosphorylated ELK1 are higher in MSC
at first diagnosis compared to MSC from salvage AML
patients. IPA reveals that, for canonical pathways, a set
associated with osteoblast differentiation was found to be
the pathway most associated with the proteins differen-
tially expressed at diagnosis compared to salvage (Online
Supplementary Figure S8).
Discussion
This study presents the first systematic study of protein
expression differences between NL-MSC and AML-MSC.
There were several notable observations. There were clear
differences between the protein expression of MSC
(whether from healthy donor or AML patient) and AML
blast cells. This result was not surprising as one would
predict different proteins would be prominent in mes-
enchymal cells and cells of hematopoietic/myeloid line-
age. The major observation of this research was the dis-
Figure 5. A distinct set of proteins is associated with acute myeloid leukemia (AML) patient salvage status. (A) Reverse phase protein arrays (RPPA) reveals protein
expression in AML mesenchymal stromal cells (MSC) differs between diagnosis and the salvage setting for 9 proteins (P=0.05; false discovery rate=0.68).
haematologica | 2018; 103(5)
Proteomic profiling reveals complexity of AML MSC
covery that AML-MSC have significantly different protein
expression patterns compared to normal MSC, with 28 of
151 analyzed proteins being highly significantly different
(FDR<0.006). These changes assumed four signatures in
AML, 81% of which were very different from that of nor-
mal MSC, while 6% had an identical signature to NL-
MSC, and another 13% were more like the normal signa-
ture than the leukemia patterns. Signature membership
showed an association with cytogenetics, with 'favorable'
cytogenetics being limited to the more NL-MSC-like pat-
tern and 'unfavorable' cytogenetics not occurring in AML
with an NL-MSC-like pattern. There was a difference in
the distribution of the MSC population between men and
women. The significance of these differences is not clear,
but women tended to have higher percentages of Class 1
and Class 2 MSC compared to men. In leukemia progeni-
tor cells, GSK3B is activated via an integrin-mediated
mechanism in response to adhesion to a stromal cell
exclusively in women patients.35 As ITGA2 and GSK3 are
members of a protein constellation (i.e. constellation 1)
that is differentially expressed in Class 1 and Class 2
(lower levels) compared to Class 3 and Class 4 (higher lev-
els), it is tempting to speculate that integrin/GSK3 axis
may contribute to sex-specific effects in MSC.
Furthermore, these signatures influence outcomes
including response rates, remission duration and, perhaps,
survival. Patients with Class 3 MSC fare much better than
patients with Class 4 MSC as demonstrated by significant-
ly longer remission duration and a trend toward longer
OS. Changes in protein expression were often character-
ized by protein-protein correlations that were reversed
from those seen in normal MSC, providing insight into the
nature of this dysregulation and potentially providing
therapeutic targets. In NL-MSC, the signature proteins were
associated with adipocyte differentiation. That normal
MSC, but not AML-MSC, possess protein pathways impor-
819
 S.M. Kornblau et al.

tant in adipogenesis is consistent with other studies from
our group and another group that found AML-MSC are
primed toward osteoblastic differentiation and not
adipocytic differentiation.9,40,41 We reported that AML-MSC
express higher levels of osteogenic markers, including
Tissue Non-specific Alkaline Phosphatase (TNAP), RUNX2,
Osterix, and Ostepontin, compared to MSC from age-
matched heathy donors.40 In addition, in that study we
found that AML-MSC readily differentiate along the
osteogenic lineage pathway but are unable to differentiate
into adipocytes. This differentiation potential of the MSC
may be influenced by the leukemia cells themselves, as
exposure of healthy donor MSC to AML cell lines such as
OCI-AML3 induces gene expression of RUNX2, TNAP, and
other osteogenic genes, and induces osteogenic differentia-
tion of the stromal cells.40 The protein networks prevalent in
NL-MSC that we identify here are consistent with signaling
that skews toward adipocytic differentiation in the normal
cells. Diaz de la Guardia et al. found that MSC from a high-
risk AML group failed to differentiate into adipcocytes.41
The IPA data on canonical pathways in the AML-MSC
compared at first diagnosis with refractory and relapse sam-
ples suggests the importance of MSC of the osteoblastic lin-
eage in the AML niche, as many of these proteins are asso-
ciated with osteoblast survival and differentiation.
PI3K/AKT is very prevalent in group 1 proteins which
are associated with Class 4 MSC by IPA. We have previ-
ously reported that leukemia cells in co-culture with MCS
induce activation of AKT and other survival kinases.42 It is
plausible that the observed activation of AKT signaling in
AML-MSC is due to the presence of leukemia cells in the
niche or at least that the malignant cells may contribute to
activation. In MSC, AKT has been implicated in positive
regulation of Cyclin D1 (CCND1) and CDK4.43,44 The pres-
ence of the PP2A B55 a subunit in group 1 suggests that
either the protein phosphatase is not active against AKT in
those cells or that AKT is less active in the AML-Like MSC
compared to Normal-Like MSC. Despite its tumor sup-
pressor role in suppression of AKT signaling, the PP2A
subunit does support β-catenin expression by dephospho-
rylating serine and threonine residues that are required for
destruction of the transcription factor (e.g. serine 33, thre-
onine 41).45-47 It is interesting that when expression of the
151 proteins surveyed by RPPA are correlated in the nor-
mal MSC and in the AML-MSC, β-catenin exhibits the
greatest difference in correlation pattern between normal
MSC and AML-MSC compared to the other proteins. It is
plausible that the PP2A B55 a subunit may be a factor in
this phenomenon, though further investigation will be
required to verify this potential mechanism. It is interest-
ing to note that suppression of PP2A (albeit via the catalyt-
ic core subunit Ca) in MSC cell line or pre-adipocyte cells
results in differentiation to adipocytes via a mechanism
involving loss of β-catenin.48 As pathway analysis of pro-
teins associated with normal MSC (i.e. Constellation 3)
pathways identifies adipogenesis as a key pathway, it will
be important to determine if AML-MSC would be less
likely to skew toward adipocyte differentiation because of
a PP2A B55 a subunit/β-catenin axis.
BCL-XL is necessary for MSC survival during differenti-
ation.27 AML-MSC expressed higher BCL-XL and Cyclin
D1 protein. These findings may be attributed to STAT5 as
this transcription factor is a regulator of both BCL-XL and
Cyclin D1.49 We found by qRT-PCR that gene expression
of both BCL-XL and Cyclin D1 was higher in AML-MSC
(n=9) compared to normal MSC (n=10) (Online
Supplementary Figure S9). These findings suggest that ele-
vation of BCL-XL and Cyclin D1 protein in AML-MSC can
be attributed at least in part to a transcriptional mecha-
nism possibly involving STAT5.
Two prominent proteins identified as elevated in the
AML-MSC group are p21 and p53 (Figure 1B). Elevated
expression of p21 and increased senescence of MSC is
consistent with the study on MDS-MSC that showed a
role for p21 in IL-6 and TGF-β production.25 However, a
recent study from Desbourdes et al. found p21 and p53
levels were similar between AML-MSC and healthy
donor-derived MSC.50 The reason for the difference
between our results and their results is not clear. It should
be noted that the Desbourdes et al. study used less than 5
samples each of AML-derived MSC and healthy donor-
derived MSC to determine p21 and p53 levels, so perhaps
Class 1 or Class 2 MSC (which would have lower levels of
p21 and p53) are over-represented in their samples.50 In
addition, the average age of the AML patients in the
Desbourdes et al. study was 49 y while the average age of
the healthy donors was near 60 y, so perhaps the p21 and
p53 levels in the healthy donor MSC are skewed higher as
the donors are older than the patients. For p21 and p53
expression, an age match comparison of p21 and p53 in
the AML-MSC and NL-MSC shows levels of these pro-
teins are higher in the AML-MSC in at least 2 ages for each
(Online Supplementary Table S2).
In conclusion, proteomic analysis identified a distinct
set of proteins that distinguish normal MSC from AML-
MSC. Our RPPA studies identified four major signatures
of MSC in AML patients that may impact their function in
the tumor microenvironment.
Funding
This work was supported in part by the research funding from
the National Institutes of Health (P01CA49639 and MD
Anderson Cancer Center Support Grant CA016672) and the
Paul and Mary Haas Chair in Genetics (to MA).

References
1. Konopleva M, Tabe Y, Zeng Z, Andreeff M.
Therapeutic targeting of microenvironmen-
tal interactions in leukemia: mechanisms
and approaches. Drug Resist Updat.
2009;12(4-5):103-113.
2. Tabe Y, Konopleva M. Advances in under-
standing the leukaemia microenvironment.
Br J Haematol. 2014;164(6):767-778.
3. Barcellos-de-Souza P, Gori V, Bambi F,
Chiarugi P. Tumor microenvironment: bone
marrow-mesenchymal stem cells as key
players. Biochim Biophys Acta.
2013;1836(2):321-335.
4. da Silva Meirelles L, Chagastelles PC, Nardi
NB. Mesenchymal stem cells reside in virtu-
ally all post-natal organs and tissues. J Cell
Sci. 2006;119(Pt 11):2204-2213.
5. Eggenhofer E, Luk F, Dahlke MH, Hoogduijn
MJ. The life and fate of mesenchymal stem
cells. Front Immunol. 2014;5:148.
6. Greenbaum A, Hsu YM, Day RB, et al.
CXCL12 in early mesenchymal progenitors
is required for haematopoietic stem-cell
maintenance. Nature. 2013;495(7440):227-
230.
7. Jacamo R, Chen Y, Wang Z, et al. Reciprocal
leukemia-stroma VCAM-1/VLA-4-depen-
dent activation of NF-kappaB mediates

820 haematologica | 2018; 103(5)


chemoresistance. Blood. 2014;123(17):2691-
2702.
8. Studeny M, Marini FC, Dembinski JL, et al.
Mesenchymal stem cells: potential precur-
sors for tumor stroma and targeted-delivery
vehicles for anticancer agents. J Natl Cancer
Inst. 2004;96(21):1593-1603.
9. Battula VL, Chen Y, Cabreira Mda G, et al.
Connective tissue growth factor regulates
adipocyte differentiation of mesenchymal
stromal cells and facilitates leukemia bone
marrow engraftment. Blood. 2013; 122(3):
357-366.
10. Chen Y, Jacamo R, Shi YX, et al. Human
extramedullary bone marrow in mice: a
novel in vivo model of genetically controlled
hematopoietic microenvironment. Blood.
2012;119(21):4971-4980.
11. Kidd S, Spaeth E, Watson K, et al. Origins of
the tumor microenvironment: quantitative
assessment of adipose-derived and bone
marrow-derived stroma. PLoS One.
2012;7(2):e30563.
12. Ling X, Marini F, Konopleva M, et al.
Mesenchymal Stem Cells Overexpressing
IFN-beta Inhibit Breast Cancer Growth and
Metastases through Stat3 Signaling in a
Syngeneic Tumor Model. Cancer
Microenviron. 2010;3(1):83-95.
13. Poggi A, Giuliani M. Mesenchymal Stromal
Cells Can Regulate the Immune Response in
the Tumor Microenvironment. Vaccines
(Basel). 2016;4(4).
14. Spaeth E, Klopp A, Dembinski J, Andreeff
M, Marini F. Inflammation and tumor
microenvironments: defining the migratory
itinerary of mesenchymal stem cells. Gene
Ther. 2008;15(10):730-738.
15. Spaeth EL, Labaff AM, Toole BP, Klopp A,
Andreeff M, Marini FC. Mesenchymal
CD44 expression contributes to the acquisi-
tion of an activated fibroblast phenotype via
TWIST activation in the tumor microenvi-
ronment. Cancer Res. 2013; 73(17):5347-
5359.
16. Kim JA, Shim JS, Lee GY, et al.
Microenvironmental remodeling as a
parameter and prognostic factor of heteroge-
neous leukemogenesis in acute myeloge-
nous leukemia. Cancer Res. 2015; 75(11):
2222-2231.
17. Raaijmakers MH, Mukherjee S, Guo S, et al.
Bone progenitor dysfunction induces
myelodysplasia and secondary leukaemia.
Nature. 2010;464(7290):852-857.
18. Shi C. Recent progress toward understand-
ing the physiological function of bone mar-
row mesenchymal stem cells. Immunology.
2012;136(2):133-138.
19. Balakrishnan K, Burger JA, Quiroga MP, et
al. Influence of bone marrow stromal
microenvironment on forodesine-induced
responses in CLL primary cells. Blood.
2010;116(7):1083-1091.
20. Kojima K, McQueen T, Chen Y, et al. p53
activation of mesenchymal stromal cells par-
tially abrogates microenvironment-mediat-
ed resistance to FLT3 inhibition in AML
through HIF-1alpha-mediated down-regula-
tion of CXCL12. Blood. 2011; 118(16):4431-
4439.
21. Marquez ME, Hernandez-Uzcategui O,
Cornejo A, Vargas P, Da Costa O. Bone mar-
row stromal mesenchymal cells induce
down regulation of CD20 expression on B-
haematologica | 2018; 103(5)
Proteomic profiling reveals complexity of AML MSC
CLL: implications for rituximab resistance in
CLL. Br J Haematol. 2015;169(2):211-218.
22. Xia B, Tian C, Guo S, et al. c-Myc plays part
in drug resistance mediated by bone mar-
row stromal cells in acute myeloid leukemia.
Leuk Res. 2015;39(1):92-99.
23. Zeng Z, Shi YX, Samudio IJ, et al. Targeting
the leukemia microenvironment by CXCR4
inhibition overcomes resistance to kinase
inhibitors and chemotherapy in AML.
Blood. 2009;113(24):6215-6224.
24. Medyouf H, Mossner M, Jann JC, et al.
Myelodysplastic cells in patients reprogram
mesenchymal stromal cells to establish a
transplantable stem cell niche disease unit.
Cell Stem Cell. 2014;14(6):824-837.
25. Zhao Y, Wu D, Fei C, et al. Down-regulation
of Dicer1 promotes cellular senescence and
decreases the differentiation and stem cell-
supporting capacities of mesenchymal stro-
mal cells in patients with myelodysplastic
syndrome. Haematologica. 2015;
100(2):194-204.
26. Oliver L, Hue E, Priault M, Vallette FM. Basal
autophagy decreased during the differentia-
tion of human adult mesenchymal stem
cells. Stem Cells Dev. 2012; 21(15):2779-
2788.
27. Oliver L, Hue E, Rossignol J, et al. Distinct
roles of Bcl-2 and Bcl-Xl in the apoptosis of
human bone marrow mesenchymal stem
cells during differentiation. PLoS One.
2011;6(5):e19820.
28. Kornblau SM, Singh N, Qiu Y, Chen W,
Zhang N, Coombes KR. Highly phosphory-
lated FOXO3A is an adverse prognostic fac-
tor in acute myeloid leukemia. Clin Cancer
Res. 2010;16(6):1865-1874.
29. Kornblau SM, Womble M, Qiu YH, et al.
Simultaneous activation of multiple signal
transduction pathways confers poor prog-
nosis in acute myelogenous leukemia.
Blood. 2006;108(7):2358-2365.
30. Ruvolo PP, Qiu Y, Coombes KR, et al.
Phosphorylation of GSK3alpha/beta corre-
lates with activation of AKT and is prognos-
tic for poor overall survival in acute myeloid
leukemia patients. BBA Clin. 2015;4:59-68.
31. Ruvolo PP, Qui YH, Coombes KR, et al. Low
expression of PP2A regulatory subunit
B55alpha is associated with T308 phospho-
rylation of AKT and shorter complete remis-
sion duration in acute myeloid leukemia
patients. Leukemia. 2011; 25(11):1711-1717.
32. Tibes R, Qiu Y, Lu Y, et al. Reverse phase
protein array: validation of a novel proteom-
ic technology and utility for analysis of pri-
mary leukemia specimens and hematopoiet-
ic stem cells. Mol Cancer Ther.
2006;5(10):2512-2521.
33. Szklarczyk D, Franceschini A, Wyder S, et
al. STRING v10: protein-protein interaction
networks, integrated over the tree of life.
Nucleic Acids Res. 2015;43(Database
issue):D447-452.
34. Huang JC, Basu SK, Zhao X, et al.
Mesenchymal stromal cells derived from
acute myeloid leukemia bone marrow
exhibit aberrant cytogenetics and cytokine
elaboration. Blood Cancer J. 2015;5:e302.
35. Bertrand J, Despeaux M, Joly S, et al. Sex dif-
ferences in the GSK3beta-mediated survival
of adherent leukemic progenitors.
Oncogene. 2012;31(6):694-705.
36. Blalock WL, Piazzi M, Bavelloni A, et al.
Identification of the PKR nuclear interac-
tome reveals roles in ribosome biogenesis,
mRNA processing and cell division. J Cell
Physiol. 2014;229(8):1047-1060.
37. Ali AT, Hochfeld WE, Myburgh R, Pepper
MS. Adipocyte and adipogenesis. Eur J Cell
Biol. 2013;92(6-7):229-236.
38. Jin W, Takagi T, Kanesashi SN, et al.
Schnurri-2 controls BMP-dependent adipo-
genesis via interaction with Smad proteins.
Dev Cell. 2006;10(4):461-471.
39. Kauppinen A, Suuronen T, Ojala J,
Kaarniranta K, Salminen A. Antagonistic
crosstalk between NF-kappaB and SIRT1 in
the regulation of inflammation and metabol-
ic disorders. Cell Signal. 2013;25 (10):1939-
1948.
40. Battula VL, Le PM, Sun JC, et al. AML-
induced osteogenic differentiation in mes-
enchymal stromal cells supports leukemia
growth. JCI Insight. 2017;2(13).
41. Diaz de la Guardia R, Lopez-Millan B,
Lavoie JR, et al. Detailed Characterization of
Mesenchymal Stem/Stromal Cells from a
Large Cohort of AML Patients
Demonstrates a Definitive Link to
Treatment Outcomes. Stem Cell Reports.
2017;8(6):1573-1586.
42. Tabe Y, Jin L, Tsutsumi-Ishii Y, et al.
Activation of integrin-linked kinase is a crit-
ical prosurvival pathway induced in
leukemic cells by bone marrow-derived
stromal cells. Cancer Res. 2007;67(2):684-
694.
43. Gharibi B, Ghuman MS, Hughes FJ. Akt-
and Erk-mediated regulation of proliferation
and differentiation during PDGFRbeta-
induced MSC self-renewal. J Cell Mol Med.
2012;16(11):2789-2801.
44. Lai VK, Ashraf M, Jiang S, Haider K.
MicroRNA-143 is a critical regulator of cell
cycle activity in stem cells with co-overex-
pression of Akt and angiopoietin-1 via tran-
scriptional regulation of Erk5/cyclin D1 sig-
naling. Cell Cycle. 2012;11(4):767-777.
45. Hein AL, Seshacharyulu P, Rachagani S, et al.
PR55alpha Subunit of Protein Phosphatase
2A Supports the Tumorigenic and
Metastatic Potential of Pancreatic Cancer
Cells by Sustaining Hyperactive Oncogenic
Signaling. Cancer Res. 2016; 76(8):2243-
2253.
46. Stamos J, Weis W. The β-catenin destruction
complex. Cold Spring Harb Perspect Biol
2013:5: a007898.
47. Zhang W, Yang J, Liu Y, et al. PR55 alpha, a
regulatory subunit of PP2A, specifically reg-
ulates PP2A-mediated beta-catenin dephos-
phorylation. J Biol Chem. 2009; 284(34):
22649-22656.
48. Okamura H, Yang D, Yoshida K, Teramachi
J, Haneji T. Reduction of PP2A Calpha stim-
ulates adipogenesis by regulating the
Wnt/GSK-3beta/beta-catenin pathway and
PPARgamma expression. Biochim Biophys
Acta. 2014;1843(11):2376-2384.
49. Grad JM, Zeng XR, Boise LH. Regulation of
Bcl-xL: a little bit of this and a little bit of
STAT. Curr Opin Oncol. 2000;12(6):543-549.
50. Desbourdes L, Javary J, Charbonnier T, et al.
Alteration Analysis of Bone Marrow
Mesenchymal Stromal Cells from De Novo
Acute Myeloid Leukemia Patients at
Diagnosis. Stem Cells Dev. 2017;26(10):
709-722.
821
ARTICLE Acute Myeloid Leukemia

Clinical relevance of IDH1/2 mutant allele

Ferrata Storti
Foundation burden during follow-up in acute myeloid
leukemia. A study by the French ALFA group
Yann Ferret,1,2,3 Nicolas Boissel,4,5 Nathalie Helevaut,1 Jordan Madic,6 Olivier
Nibourel,1,2,3 Alice Marceau-Renaut,1,2,3 Maxime Bucci,1 Sandrine Geffroy,1 Karine
Celli-Lebras,7 Sylvie Castaigne,8 Xavier Thomas,9 Christine Terré,10 Hervé
Dombret,4,5 Claude Preudhomme1,2,3 and Aline Renneville1,2,3
1
Hematology Laboratory, Biology and Pathology Center, CHRU of Lille; 2INSERM, UMR-S
1172, Cancer Research Institute of Lille; 3University of Lille, F-59000; 4Hematology
Haematologica 2018
Department, Saint-Louis Hospital, AP-HP, Paris; 5EA3518, Institut Universitaire
Volume 103(5):822-829
d’Hématologie (IUH), University 7 Paris Diderot; 6Circulating Biomarkers Laboratory, Curie
Institute, Paris; 7Acute Leukemia French Association (ALFA) coordination, Saint-Louis
Hospital, AP-HP, Paris; 8Hematology Department, Versailles Hospital, Le Chesnay;
9
Hematology Department, Lyon Sud Hospital, Pierre Benite and 10Cytogenetic Laboratory,
Versailles Hospital, Le Chesnay, France

ABSTRACT

A
ssessment of minimal residual disease has emerged as a powerful
prognostic factor in acute myeloid leukemia. In this study, we
investigated the potential of IDH1/2 mutations as targets for min-
imal residual disease assessment in acute myeloid leukemia, since these
mutations collectively occur in 15-20% of cases of acute myeloid
leukemia and now represent druggable targets. We employed droplet
digital polymerase chain reaction assays to quantify IDH1R132,
IDH2R140, and IDH2R172 mutations on genomic DNA in 322 samples
from 103 adult patients with primary IDH1/2 mutant acute myeloid
leukemia and enrolled on Acute Leukemia French Association (ALFA) -
Correspondence: 0701 or -0702 clinical trials. The median IDH1/2 mutant allele fraction in
aline.renneville@gmail.com bone marrow samples was 42.3% (range, 8.2 - 49.9%) at diagnosis of
acute myeloid leukemia, and below the detection limit of 0.2% (range,
<0.2 - 39.3%) in complete remission after induction therapy. In univariate
Received: October 29, 2017. analysis, the presence of a normal karyotype, a NPM1 mutation, and an
Accepted: February 16, 2018. IDH1/2 mutant allele fraction <0.2% in bone marrow after induction
Pre-published: February 22, 2018.
therapy were statistically significant predictors of longer disease-free sur-
vival. In multivariate analysis, these three variables remained significantly
predictive of disease-free survival. In 7/103 (7%) patients, IDH1/2 muta-
doi:10.3324/haematol.2017.183525 tions persisted at high levels in complete remission, consistent with the
presence of an IDH1/2 mutation in pre-leukemic hematopoietic stem
Check the online version for the most updated cells. Five out of these seven patients subsequently relapsed or progressed
information on this article, online supplements, toward myelodysplastic syndrome, suggesting that patients carrying the
and information on authorship & disclosures:
www.haematologica.org/content/103/5/822 IDH1/2 mutation in a pre-leukemic clone may be at high risk of hemato-
logic evolution.

©2018 Ferrata Storti Foundation

Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
Introduction
Assessment of minimal residual disease (MRD) has emerged as a powerful prog-
nostic factor in acute myeloid leukemia (AML).1–6 Many studies have shown that
MRD detection using multiparameter flow cytometry or real-time quantitative
polymerase chain reaction (qPCR) provides powerful independent prognostic infor-
mation in AML.7,8 Chimeric fusion genes, such as PML-RARA, CBFB-MYH11 or
RUNX1-RUNX1T1, NPM1 mutations, as well as WT1 expression are well-estab-
lished molecular markers for MRD monitoring in AML. However, sensitive and
leukemia-specific MRD markers are lacking in approximately 40% of AML
patients. This prompted us to investigate the potential of other recurrent molecular

822 haematologica | 2018; 103(5)


abnormalities as targets for MRD assessment in AML, such
as mutations in isocitrate dehydrogenase (IDH) 1 and 2.
IDH1/2 mutations affecting IDH1R132, IDH2R140, and
IDH2R172 residues are single-nucleotide mutations that
collectively occur in 15-20% of AML and represent driver
mutations in leukemogenesis.9 Mutant IDH1/2 enzymes
have neomorphic activity and catalyze the reduction of a-
ketoglutarate to an oncometabolite, the R-enantiomer of
2-hydroxyglutarate (2-HG), which promotes DNA and
histone hypermethylation, altered gene expression, and
impaired hematopoietic differentiation.10,11 Quantification
of single-nucleotide mutations by qPCR can be challeng-
ing because of problems with background amplification
from the wild-type allele. Recently, the development of
digital PCR has enabled absolute quantification of various
genomic targets with high precision and sensitivity and
has, therefore, turned out to be a promising technique for
MRD monitoring, especially for gene mutations.7,12
The clinical significance of residual IDH1/2 mutations in
bone marrow in complete remission after chemotherapy
is currently unknown. In this study, we employed digital
PCR assays to quantify IDH1/2 mutant allele fraction at
AML diagnosis and during follow-up in a large cohort of
AML patients intensively treated in the Acute French
Leukemia Association (ALFA) trials to investigate whether
IDH1/2 mutations are suitable MRD markers that could
predict clinical outcome in AML patients and provide fur-
ther information for risk-adapted therapy.
Methods
Patients and treatment
This study was performed in 103 adult patients (18-70 years)
with previously untreated primary IDH1/2 mutated AML and
enrolled on the prospective ALFA-0701 (Eudra-CT 2007-002933-
haematologica | 2018; 103(5)
IDH1/2 mutant allele burden in AML
36; ClinicalTrials.gov NCT00927498 or ALFA-0702 (Eudra-CT
2008-000668-18; ClinicalTrials.gov NCT00932412) trials.
Treatment schemes have been previously reported for both tri-
als.13,14 These studies were approved by the ethics committee of
Saint-Germain en Laye and Sud Est IV, France, respectively, and
the institutional review board of the French Regulatory Agency.
Bone marrow or peripheral blood samples collected at the time of
diagnosis of AML and during follow-up were obtained from the
tissue bank Tumorothèque du Centre de Référence Régional en
Cancérologie de Lille (CRRC)” and approval for this study was
obtained from the institutional review board of CHRU of Lille
(CSTMT089). All patients provided written informed consent to
both treatment and genetic analysis before inclusion in the study,
in accordance with the declaration of Helsinki. Among all patients
included in the ALFA-0701 (n=278) or ALFA-0702 (n=704) trials,
we selected patients meeting the following criteria: (i) the pres-
ence of an IDH1R132 or an IDH2R140/R172 mutation at AML
diagnosis (n=160), (ii) achievement of complete remission after
induction therapy (n=130), and (iii) one or more bone marrow fol-
low-up sample available for IDH1/2 variant allele fraction
(IDH1/2-VAF) assessment (n=103) (Figure 1).
Molecular analysis
Droplet DigitalTM PCR (ddPCR) assays were used to quantify
the IDH1/2 mutant allele and its wild-type counterpart in diagnos-
tic and follow-up samples. During complete remission, only bone
marrow samples were analyzed for IDH1/2-VAF assessment.
IDH1/2-VAF was quantified on genomic DNA using Bio-RadTM
reagents, primers and probes (HEX-labeled wild-type allele; FAM-
labeled mutant alleles). All samples were tested in duplicate wells,
using 90 ng of DNA per well. The PCR product from each well
was then subjected to the QX100 droplet reader (Bio-RadTM),
which measures the fluorescence of each droplet individually
using a two-color detection system. Raw data were analyzed
using QuantaSoft software, version 1.7.4.0917 (Bio-RadTM).
Representative two-dimensional plots of droplet fluorescence for
Figure 1. Patient flow chart. VAF,
variant allele fraction.
823
 Y. Ferret et al.

IDH1/2 wild-type controls and IDH1/2 mutant samples are shown mutations in AML.9,17 In our cohort, IDH2R172K muta-
in Online Supplementary Figure S1. The mutant allele frequency was tions were mutually exclusive with NPM1 and FLT3 muta-
then estimated using a Poisson distribution model as the fraction tions, but co-occurred with DNMT3A mutations. An iso-
of positive droplets divided by total droplets containing a target. lated trisomy 11 was identified in 5/21 (24%) patients
The limit of detection was defined for each mutation as the mean with the IDH2R172K mutation, while this cytogenetic
value of IDH1/2 wild-type controls plus three standard deviations abnormality was not found in any patient with other
(Online Supplementary Table S1). The upper detection limit of these types of IDH1/2 mutations (24% versus 0%; P<0.001)
ddPCR assays (rounded to 0.2% of mutant allele frequency) was (Figure 2). In the subgroup of IDH2R172K mutant AML
further considered as the threshold for statistical analysis. An (n=21), single-nucleotide polymorphism array analysis
IDH1/2-VAF level below 0.2% was hereafter considered as nega- revealed an additional genomic lesion involving chromo-
tive MRD. Gene mutation analysis and next-generation sequenc- some 11, consisting of a 11p11.2-q12.1 uniparental dis-
ing assays are described in the Online Supplementary Methods and omy, in one patient with normal karyotype AML. No MLL
Online Supplementary Tables S2-S4. partial tandem duplication, known to be strongly associat-
ed with trisomy 11,18 was found by reverse transcriptase
Statistical analysis
Group comparison for categorical and continuous variables was
performed with the Fisher exact and Mann-Whitney test, respec-
Table 1. Baseline characteristics of the patients and acute myeloid
tively. Overall survival was calculated from the date of AML diag- leukemias.
nosis to the last follow-up date by censoring patients alive at that
date. Disease-free survival was calculated from the date of com- Number of patients (%)
plete remission to the date of relapse or death, censoring patients
ALFA-0701 ALFA-0702 Total
alive without an event at the last follow-up date. In some analyses, Gender
data were censored at the time of allogeneic stem cell transplanta- Male 10 38 48 (47)
tion. Univariate and multivariate analyses assessing the impact of Female 16 39 55 (53)
categorical and continuous variables were performed with a Cox Median age (range), years 62 (51-70) 50 (22-60) 54 (22-70)
model.15 The proportional-hazards assumption was checked before Median white blood cell 18 (1-157) 5 (1-377) 7 (1-377)
conducting multivariate analyses.16 Covariates with a P-value <0.1 count (range), x 109/L
in univariate analysis were included in the multivariable models.
Cytogenetics
Statistical analyses were performed with STATA software (STATA
Normal 21 50 71 (69)
12.0 Corporation, College Station, TX, USA). P-values were two-
Abnormal 4 23 27 (26)
sided, with P<0.05 denoting statistical significance. Failure 1 4 5 (5)
IDH1/2 mutation
IDH1 p.R132H/C/G 10 26 36 (35)
Results IDH2 p.R140Q 10 36 46 (45)
IDH2 p.R172K 6 15 21 (20)
Baseline characteristics of the patients and acute
myeloid leukemias Other gene mutations
The patients’ median age was 54 years (range, 22-70). NPM1 mutation 16 34 50 (48)
FLT3 internal tandem duplication 3 16 19 (19)
The median follow-up was 2.7 years (95% CI: 2.3-3.0).
FLT3-tyrosine kinase domain mutation2 7 9 (9)
Results of conventional cytogenetic studies were available DNMT3A mutation 6 23 29 (35)
for 98/103 (95%) patients, of whom 72% had normal TET2 mutation 2 2 4 (4)
karyotype AML. A concomitant NPM1 mutation was CEBPA mutation 1 (1 sm) 3 (2 sm, 1 dm) 4 (4)
found in 50/103 (48%) patients. Only 4/103 (4%) patients
European LeukemiaNet 2008 risk-group
harbored a concomitant TET2 mutation (Table 1), in accor-
Favorable 13 20 33 (32)
dance with the fact that IDH1/2 and TET2 mutations tend Non-favorable 12 52 64 (62)
to be mutually exclusive.10 As opposed to IDH1R132 and Not defined 1 5 6 (6)
IDH2R140 mutations, IDH2R172 mutations are less likely
sm: single mutation; dm: double mutation.
to be accompanied by additional frequently recurring

Figure 2. Barcoding representing the co-occurrence of gene mutations and cytogenetic alterations in our cohort of 103 patients with IDH1/2 mutant acute myeloid
leukemia. ITD: internal tandem duplication; TKD: tyrosine kinase domain.

824 haematologica | 2018; 103(5)


PCR in this subgroup (data not shown). The association
between IDH2R172 mutation and trisomy 11 observed in
our cohort is consistent with results from a previous
study,19 and suggests a potential cooperation between
these two genetic alterations in leukemogenesis.
IDH1/2 mutation level at diagnosis of acute myeloid
leukemia and during follow-up
At AML diagnosis, IDH1/2-VAF could be assessed by
next-generation sequencing in 80/103 patients (Online
Supplementary Table S5). The median IDH1/2-VAF value
was 41% (range, 16-53%) in bone marrow and 39.5%
(range, 6-50%) in peripheral blood samples. In the subset
of NPM1-mutated AML, IDH1/2-VAF was systematically
higher than NPM1-VAF, except in one patient with similar
VAF for both mutations [n=34 comparisons; median dif-
ference IDH1/2-VAF - NPM1-VAF, 10.5% (range, 0-25%);
P<0.001] (Online Supplementary Figure S2). This finding
supports the notion that IDH1/2 mutations were present
in pre-existing clones that subsequently acquired NPM1
mutations.
We also performed ddPCR assays in diagnostic and fol-
low-up samples to quantify the IDH1/2-VAF. A total of
322 samples from 103 patients with IDH1/2 mutations
were analyzed by ddPCR at diagnosis (n=97, of which 69
were bone marrow and 28 peripheral blood samples), dur-
ing hematologic remission (n=211 bone marrow samples),
and at relapse (n=14 bone marrow samples). At AML diag-
nosis, the median IDH1/2-VAF assessed by ddPCR was
42.3% (range, 8.2-49.9%) in bone marrow and 40.6%
(range, 5.5-53%) in peripheral blood samples, consistent
with our next-generation sequencing data. After induction
therapy, the IDH1/2 mutant allele fraction in bone marrow
samples decreased significantly compared to the pretreat-
ment levels (P<0.001) with a median value below 0.2%
(range, <0.2-39.3%). At AML relapse, the median IDH1/2-
VAF was 21.3% (range, 0.2-38.5%). Among the 14
patients for whom a bone marrow sample was available
for molecular analysis at AML relapse, only one lost the
mutation during disease evolution (Figure 3A).
Persistent clonal hematopoiesis with IDH1/2
mutations
IDH1/2 mutations persisted at high levels during hema-
tologic remission in 7/103 (7%) patients, including four
A B
Figure 3. IDH1/2 mutant allele fraction assessed by droplet digital polymerase chain reaction at diagnosis of acute myeloid leukemia and during follow-up (A)
in the whole cohort and (B) for the seven patients with persistent clonal hematopoiesis with IDH1/2 mutations. The plain lines in the dot plot indicate the median
values.
haematologica | 2018; 103(5)
IDH1/2 mutant allele burden in AML
with an IDH2R140 mutation and three with an IDH1R132
mutation, but none with an IDH2R172 mutation. The
main characteristics of these seven patients are summa-
rized in Table 2 and their IDH1/2-VAF profiles are shown
in Figure 3B. The only common characteristic identified in
these patients was age over 50 years. In this subgroup, the
median IDH1/2-VAF was 8% (range, 0.8-28.5%) after
induction and 40% (range, 26-43.5%) after consolidation
therapy. Of these seven patients, only one is still alive in
first complete remission, one died from transplant-related
mortality, three relapsed, and two developed overt
myelodysplastic syndrome. Altogether, 5/7 (71%) patients
with persistent clonal hematopoiesis with IDH1/2 muta-
tions relapsed or progressed toward myelodysplastic syn-
drome within 1 to 4 years after AML diagnosis.
Univariate and multivariate prognostic analyses
The prognostic impact of IDH1/2 mutations in AML
remains controversial.9 In the present cohort composed
exclusively of IDH1/2-mutated AML, the presence of an
IDH2R172 mutation was associated with a shorter dis-
ease-free survival compared to other IDH1/2 mutation
types, but without the difference reaching statistical sig-
nificance (P=0.088). No difference according to the type of
IDH1/2 mutation was observed regarding overall survival
(Table 3; Figure 4).
The prognostic impact of IDH1/2-VAF was evaluated in
complete remission after induction therapy in a subset of
95 patients for whom a post-induction bone marrow sam-
ple was available for IDH1/2-VAF assessment (Figure 1).
We were not able to perform statistical analysis at later
follow-up time-points, such as post-consolidation,
because of the lack of available DNA samples for many
patients. Variables considered for univariate and multivari-
ate analyses were age, white blood cell count, cytogenet-
ics, mutational status of five genes, and IDH1/2-VAF after
induction therapy. In univariate analysis for disease-free
survival, the presence of a normal karyotype, a NPM1
mutation, and a IDH1/2-VAF <0.2% were significantly
associated with a longer disease-free survival. In multivari-
ate analysis, these three variables remained significantly
predictive of disease-free survival. Factors significantly
associated with overall survival were age, the presence of
a normal karyotype, the presence of a NPM1 mutation or
a TET2 mutation. Other molecular abnormalities studied,
825
 Y. Ferret et al.

as well as IDH1/2-VAF, had no impact on overall survival
(Table 3; Figure 5).
Discussion
In this study including 103 adult patients with primary
IDH1/2 mutant AML who were intensively treated, we
showed the feasibility of IDH1/2-VAF monitoring using
ddPCR and its prognostic relevance after induction thera-
py, independently of pretreatment risk factors. Our find-
ings also suggest that patients with persistent IDH1/2-
mutated clonal hematopoiesis may be at high risk of dis-
mal hematologic evolution.
The prognostic value of IDH1/2 mutations is still a mat-
ter of debate9 and may be influenced by the type of muta-
tions, as we previously reported,20,21 or the profile of con-
comitant mutations, such as NPM1 or DNMT3A muta-
tions.17,22 The present study, which only included patients
with IDH1/2 mutations, was not designed to explore the
prognostic significance of IDH1/2 mutations.
The role of MRD in the management of AML patients is
growing. Because of the marked heterogeneity of AML,
no single MRD marker can be applied to all patients.
Additionally, the optimal method for measuring clearance
of leukemia cells after chemotherapy remains to be deter-
mined. Here, we focused on IDH1/2 mutations because
they are recurrent genetic events in AML, mostly in nor-
mal karyotype AML, and now represent druggable targets.
The digital PCR technique had been previously shown to
allow absolute quantification of a nucleic acid target with
high precision and sensitivity.12 Our data provide evidence
that measurement of IDH1/2-VAF by ddPCR is feasible.

Table 2. Clinical and biological characteristics of the seven patients with persistent clonal hematopoiesis with IDH1/2 mutations.
UPN 1 UPN 2 UPN 3 UPN 4 UPN 5 UPN 6 UPN 7
Age (years) 50 55 55 50 68 60 63
Gender F M M M F F F
9
WBC count, x 10 /L 28 2.4 4.7 43 34 100 3.2
Cytogenetics Normal Trisomy 8 Normal Normal Failure Normal Normal
NPM1 mutation Pos. Neg. Pos. Neg. Pos. Pos. Neg.
FLT3-ITD Pos. Neg. Neg. Pos. Pos. Neg. Neg.
FLT3-TKD mutation Pos. Neg. Neg. Neg. Neg. Neg. Neg.
CEBPA mutation Neg. NA Neg. Neg. Neg. Neg. Neg.
DNMT3A mutation NA p.R882H (VAF 26%) NA p.R882H (VAF 48%) Neg. Neg. Neg.
TET2 mutation NA Neg. NA Neg. Neg. Neg. Neg.
IDH1/2 mutation IDH2 p.R140Q IDH2 p.R140Q IDH2 p.R140Q IDH2 p.R140Q IDH1 p.R132G IDH1 p.R132C IDH1 p.R132C
(VAF at diagnosis) (44%) (43%) (39%) (47%) (44%) (48%) (43%)
IDH1/2-VAF in CR 28.5% 0.76% 4.87% 28.1% NA 8.2% NA
after induction
IDH1/2-VAF in CR 28.1% 7.2% 42.9% 39.9% 43.5% NA 27.1%
after consolidation
Clinical outcome Alive in CR1 2 years Relapse 4 years MDS 1 year after Relapse 1.5 year Death after Relapse 1.5 year MDS 2.5 years
after AML after AML AML diagnosis after AML allo-SCT after AML after AML
diagnosis diagnosis diagnosis diagnosis diagnosis
UPN: unique patient number; F: female; M: male; WBC: white blood cell; Pos.: positive; Neg.: negative; ITD: internal tandem duplication; TKD: tyrosine kinase domain; NA: not available;
VAF: variant allele fraction; CR: complete remission; AML: acute myeloid leukemia; MDS: myelodysplastic syndrome; allo-SCT: allogeneic stem cell transplantation.

Table 3. Prognostic analysis for disease-free survival and overall survival.

Disease-free survival Overall survival
Univariate Multivariate Univariate
Variable HR 95% CI P HR 95% CI P HR 95% CI P
Age* 1.04 0.98 1.10 0.202 - - - - 1.13 1.00 1.27 0.047
Log10 (white blood cell count)* 1.00 0.99 1.01 0.537 - - - - 1.00 0.99 1.01 0.698
NPM1 mutation 0.23 0.11 0.50 <0.001 0.32 0.12 0.88 0.027 0.19 0.05 0.72 0.014
Normal karyotype 0.26 0.12 0.59 0.001 0.41 0.17 0.99 0.046 0.24 0.07 0.76 0.016
FLT3 internal tandem duplication 1.11 0.33 3.70 0.865 - - - - 1.00 0.12 8.08 1.000
FLT3 tyrosine kinase domain mutation 0.20 0.03 1.48 0.115 - - - - - - - 0.078
DNMT3A mutation 1.42 0.61 3.31 0.413 - - - - 2.42 0.64 9.15 0.192
TET2 mutation 2.66 0.58 12.30 0.209 - - - - 12.59 1.68 94.62 0.014
IDH2 p.R172K mutation 2.04 0.90 4.61 0.088 0.85 0.31 2.32 0.751 1.44 0.39 5.34 0.586
IDH1/2-VAF after induction <0.2% 0.32 0.15 0.69 0.004 0.40 0.18 0.90 0.026 0.46 0.14 1.54 0.208
HR: hazard ratio; CI: confidence interval; VAF: variant allele fraction; *: continuous variable.

826 haematologica | 2018; 103(5)

 IDH1/2 mutant allele burden in AML

However, despite technical optimizations, we were not
able to reach the 0.01% or even 0.1% threshold that we
would expect as the quantitative detection limit with the
specific ddPCR assays. This problem was due to a relative-
ly high background observed in negative controls, which
always consisted of double-positive (actually false-posi-
tive) droplets. Polymerase errors occurring during the PCR
amplification step seem to be responsible for the genera-
tion of these false-positive signals.
The present study is the first to quantify IDH1/2 muta-
tion levels in a large cohort of AML patients. Previous
studies using Sanger sequencing,23 qPCR,24 or next-genera-
tion sequencing technology25 suggested that the presence
or the level of IDH1/2 mutations was correlated to disease
status in most patients with AML, but the small number
of IDH1/2-mutated patients included in these studies pre-
cluded statistical analysis. Our study revealed that a posi-
tive IDH1/2-VAF after induction chemotherapy was asso-
ciated with a shorter disease-free survival. Whether
patients with residual IDH1/2 mutations in complete
remission may benefit from allogeneic stem cell transplan-
tation remains to be addressed by future studies. In clinical
practice, IDH1/2-VAF assessment during and after treat-
ment could be especially valuable in AML patients with-
out recurrent fusion genes or NPM1 mutations, which are
both leukemia-specific and more sensitive MRD markers.
Keeping in mind the caveat that IDH1/2 mutations can be
present in the pre-leukemic clone in some cases, one could
argue that these mutations could be good MRD markers
for those patients in whom MRD becomes undetectable
after induction or at early follow-up time-points.
However, sequential monitoring of IDH1/2-VAF after con-
solidation therapy or allogeneic stem cell transplantation
could still help to detect disease persistence and guide pre-
emptive therapy to prevent hematologic relapse, as sug-
gested in a recent study.26 An alternative approach to MRD
monitoring in IDH1/2-mutated patients is to quantify the
oncometabolite 2-HG.27,28 A previous study from the ALFA
group showed that total 2-HG serum levels <2 μmol/L
after induction were associated with better disease-free
survival and overall survival.29 We were not able to corre-
late IDH1/2-VAF and 2-HG levels in this study because of
the lack of serum samples.
We found that 7/103 (7%) patients had an IDH1/2

A B

Figure 4. Kaplan-Meier estimates of (A) disease-free survival and (B) overall survival according to the type of IDH1/2 mutation.

A B

Figure 5. Prognostic analysis according to post-induction IDH1/2 mutant allele fraction. Kaplan-Meier estimates of (A) disease-free survival and (B) overall survival
according to IDH1/2 variant allele fraction (IDH1/2-VAF). MRD+ denotes IDH1/2-VAF ≥0.2% and MRD- denotes IDH1/2-VAF <0.2%.

haematologica | 2018; 103(5) 827

 Y. Ferret et al.

mutation that persisted at high levels in hematologic
remission, consistent with the presence of this mutation
in pre-leukemic hematopoietic stem cells. Unlike AML
blasts, these hematopoietic stem cells survive chemother-
apy and persist in remission bone marrow, providing a
potential reservoir for leukemic progression.30 In our study,
5/7 (71%) patients with persistent clonal hematopoiesis
with IDH1/2 mutations relapsed or progressed toward
myelodysplastic syndrome, suggesting that these patients
may be at high risk of hematologic evolution and should
probably be monitored more closely. Klco et al. showed
that initiating mutations, such as DNMT3A, TET2, and
IDH1/2 mutations, are less likely to be cleared after
chemotherapy than cooperating mutations,31 in accor-
dance with our own and previous data.25,32 Furthermore,
the prognostic value of persisting somatic mutations in
complete remission appears to vary depending on the
gene involved. Recent studies suggested that the presence
of persistent mutations in DNMT3A, TET2 or ASXL1 lacks
prognostic impact in terms of AML relapse or survival,33,34
in contrast with what we observed for IDH1/2 mutations.
Patients with IDH1/2 mutations are candidates for tar-
geted therapies. Small-molecule inhibitors of mutant IDH1
such as ivosidenib or IDH2 such as the recently approved
enasidenib are currently under clinical investigation and,
when used as single agents, have shown promising results
in patients with AML or myelodysplastic syndrome as a
first-line treatment or in relapsed or refractory diseases.9
These molecules have been shown to induce differentia-
tion of primary leukemic cells in vitro35,36 and in vivo37 to pro-
mote clinical responses. Future studies should determine
whether patients with high levels of IDH1/2-VAF after
induction therapy could benefit from a consolidation or
maintenance therapy including IDH1/2 inhibitors.
Ultimately, one could imagine that the use of these small
molecules might also be considered in patients with per-
sistence of clonal hematopoiesis with IDH1/2 mutations,
although clearance of the clone carrying the drug targets
seems to occur only in a small subset of treated patients,
even with the most potent inhibitors.38 Additionally, pre-
clinical and clinical data indicate that IDH1/2 mutations
may identify patients likely to respond to pharmacological
BCL-2 inhibition.39,40 The use of IDH1/2-VAF monitoring in
patients treated with an IDH1/2 or BCL-2 inhibitor, such as
venetoclax, could therefore contribute to the evaluation of
treatment efficacy.
In conclusion, our study is the first to show that IDH1/2
mutant allele fraction in complete remission after induc-
tion therapy significantly correlates with disease-free sur-
vival, independently of pretreatment prognostic factors.
However, this difference did not translate into distinct
overall survival rates in our cohort. Our data provide evi-
dence that IDH1/2 mutant allele fraction has the potential
to become a useful tool for the management of AML
patients as a biomarker of treatment response, in addition
to being a molecular predictor of response to targeted ther-
apies. Further studies based on larger cohorts of patients
are required to confirm and extend our findings, and to
address the question of whether the residual level of
IDH1/2 mutation may help to refine the assignment into
distinct risk groups and guide the decision of whether to
perform allogeneic stem cell transplantation or give target-
ed therapies.
Acknowledgments
This work was supported by the Association Laurette Fugain,
the Ligue Contre le Cancer (North Center), the SIRIC
ONCOLille, the North-West Canceropole (GIRCI AAP-AE
2015_53), and the Institut National du Cancer - Direction
Generale de l’Offre de Soin (INCA-DGOS_9967).

References
1. Schnittger S, Weisser M, Schoch C,
Hiddemann W, Haferlach T, Kern W. New
score predicting for prognosis in PML-
RARA+, AML1-ETO+, or CBFBMYH11+
acute myeloid leukemia based on quantifi-
cation of fusion transcripts. Blood.
2003;102(8):2746–2755.
2. Stentoft J, Hokland P, Ostergaard M, Hasle
H, Nyvold CG. Minimal residual core bind-
ing factor AMLs by real time quantitative
PCR--initial response to chemotherapy pre-
dicts event free survival and close monitor-
ing of peripheral blood unravels the kinet-
ics of relapse. Leuk Res. 2006;30(4):389–
395.
3. Cilloni D, Renneville A, Hermitte F, et al.
Real-time quantitative polymerase chain
reaction detection of minimal residual dis-
ease by standardized WT1 assay to
enhance risk stratification in acute myeloid
leukemia: a European LeukemiaNet study. J
Clin Oncol. 2009;27(31):5195–5201.
4. Freeman SD, Virgo P, Couzens S, et al.
Prognostic relevance of treatment response
measured by flow cytometric residual dis-
ease detection in older patients with acute
myeloid leukemia. J Clin Oncol.
2013;31(32):4123–4131.
5. Terwijn M, van Putten WLJ, Kelder A, et al.
High prognostic impact of flow cytometric
minimal residual disease detection in acute
myeloid leukemia: data from the
HOVON/SAKK AML 42A study. J Clin
Oncol. 2013;31(31):3889–3897.
6. Krönke J, Schlenk RF, Jensen K-O, et al.
Monitoring of minimal residual disease in
NPM1-mutated acute myeloid leukemia: a
study from the German-Austrian acute
myeloid leukemia study group. J Clin
Oncol. 2011;29(19):2709–2716.
7. Grimwade D, Ivey A, Huntly BJP.
Molecular landscape of acute myeloid
leukemia in younger adults and its clinical
relevance. Blood. 2016;127(1):29–41.
8. Ommen HB. Monitoring minimal residual
disease in acute myeloid leukaemia: a
review of the current evolving strategies.
Ther Adv Hematol. 2016;7(1):3–16.
9. Medeiros BC, Fathi AT, DiNardo CD,
Pollyea DA, Chan SM, Swords R. Isocitrate
dehydrogenase mutations in myeloid malig-
nancies. Leukemia. 2017;31(2):272–281.
10. Figueroa ME, Abdel-Wahab O, Lu C, et al.
Leukemic IDH1 and IDH2 mutations result
in a hypermethylation phenotype, disrupt
TET2 function, and impair hematopoietic
differentiation. Cancer Cell. 2010;18(6):
553–567.
11. Losman J-A, Looper RE, Koivunen P, et al.
(R)-2-hydroxyglutarate is sufficient to pro-
mote leukemogenesis and its effects are
reversible. Science. 2013;339(6127):1621–
1625.
12. Hindson BJ, Ness KD, Masquelier DA, et al.
High-throughput droplet digital PCR sys-
tem for absolute quantitation of DNA copy
number. Anal Chem. 2011;83(22):8604–
8610.
13. Castaigne S, Pautas C, Terré C, et al. Effect
of gemtuzumab ozogamicin on survival of
adult patients with de-novo acute myeloid
leukaemia (ALFA-0701): a randomised,
open-label, phase 3 study. Lancet.
2012;379(9825):1508–1516.
14. Thomas X, de Botton S, Chevret S, et al.
Randomized phase II study of clofarabine-
based consolidation for younger adults
with acute myeloid leukemia in first remis-
sion. J Clin Oncol. 2017;35(11):1223-1230.
15. Cox D. Regression models and life tables. J
R Stat Soc B. 1972;34:187-220.
16. Grambsch PM, Therneau TM. Proportional
hazards tests and diagnostics based on
weighted residuals. Biometrika. 1994;81(3):
515–526.
17. Papaemmanuil E, Gerstung M, Bullinger L,
et al. Genomic classification and prognosis
in acute myeloid leukemia. N Engl J Med.
2016;374(23):2209–2221.
18. Caligiuri MA, Strout MP, Schichman SA, et
al. Partial tandem duplication of ALL1 as a
recurrent molecular defect in acute myeloid
leukemia with trisomy 11. Cancer Res.
1996;56(6):1418–1425.
19. Eisfeld A-K, Kohlschmidt J, Mrózek K, et al.
Adult acute myeloid leukemia with tri-
somy 11 as the sole abnormality is charac-
terized by the presence of five distinct gene
mutations: MLL-PTD, DNMT3A, U2AF1,
FLT3-ITD and IDH2. Leukemia.

828 haematologica | 2018; 103(5)


2016;30(11):2254–2258.
20. Boissel N, Nibourel O, Renneville A, et al.
Prognostic impact of isocitrate dehydroge-
nase enzyme isoforms 1 and 2 mutations in
acute myeloid leukemia: a study by the
Acute Leukemia French Association group.
J Clin Oncol. 2010;28(23):3717–3723.
21. Boissel N, Nibourel O, Renneville A,
Huchette P, Dombret H, Preudhomme C.
Differential prognosis impact of IDH2
mutations in cytogenetically normal acute
myeloid leukemia. Blood. 2011;117(13):
3696–3697.
22. Patel JP, Gönen M, Figueroa ME, et al.
Prognostic relevance of integrated genetic
profiling in acute myeloid leukemia. N Engl
J Med. 2012;366(12):1079–1089.
23. Chou W-C, Peng K-Y, Lei W-C, et al.
Persistence of mutant isocitrate dehydroge-
nase in patients with acute myeloid
leukemia in remission. Leukemia.
2012;26(3):527–529.
24. Jeziskova I, Razga F, Toskova M, et al.
Quantitative detection of IDH2 mutation for
minimal residual disease monitoring in
patients with acute myeloid leukemia and
its comparison with mutations in NPM1
gene. Leuk Lymphoma. 2013;54(4):867–870.
25. Debarri H, Lebon D, Roumier C, et al.
IDH1/2 but not DNMT3A mutations are
suitable targets for minimal residual disease
monitoring in acute myeloid leukemia
patients: a study by the Acute Leukemia
French Association. Oncotarget. 2015;6
(39):42345–42353.
26. Brambati C, Galbiati S, Xue E, et al. Droplet
digital polymerase chain reaction for
DNMT3A and IDH1/2 mutations to
haematologica | 2018; 103(5)
improve early detection of acute myeloid
leukemia relapse after allogeneic
hematopoietic stem cell transplantation.
Haematologica. 2016;101(4):e157-161.
27. Ward PS, Patel J, Wise DR, et al. The com-
mon feature of leukemia-associated IDH1
and IDH2 mutations is a neomorphic
enzyme activity converting a-ketoglutarate
to 2-hydroxyglutarate. Cancer Cell.
2010;17(3):225–234.
28. Pollyea DA, Kohrt HE, Zhang B, et al. 2-
hydroxyglutarate in IDH mutant acute
myeloid leukemia: predicting patient
responses, minimal residual disease and
correlations with methylcytosine and
hydroxymethylcytosine levels. Leuk
Lymphoma. 2013;54(2):408–410.
29. Janin M, Mylonas E, Saada V, et al. Serum
2-hydroxyglutarate production in IDH1-
and IDH2-mutated de novo acute myeloid
leukemia: a study by the Acute Leukemia
French Association group. J Clin Oncol.
2014;32(4):297–305.
30. Shlush LI, Zandi S, Mitchell A, et al.
Identification of pre-leukaemic
haematopoietic stem cells in acute
leukaemia. Nature. 2014;506(7488):328–
333.
31. Klco JM, Miller CA, Griffith M, et al.
Association between mutation clearance
after induction therapy and outcomes in
acute myeloid leukemia. JAMA.
2015;314(8):811–822.
32. Pløen GG, Nederby L, Guldberg P, et al.
Persistence of DNMT3A mutations at long-
term remission in adult patients with AML.
Br J Haematol. 2014;167(4):478–486.
33. Gaidzik VI, Weber D, Paschka P, et al.
IDH1/2 mutant allele burden in AML
DNMT3A mutant transcript levels persist
in remission and do not predict outcome in
patients with acute myeloid leukemia.
Leukemia. 2018;32(1):30-37.
34. Jongen-Lavrencic M, Grob T, Kavelaars FG,
et al. Prospective molecular MRD detection
by NGS: a powerful independent predictor
for relapse and survival in adults with
newly diagnosed AML. Blood. 2017;130
(Suppl 1):LBA-5.
35. Chaturvedi A, Araujo Cruz MM, Jyotsana
N, et al. Mutant IDH1 promotes leukemo-
genesis in vivo and can be specifically tar-
geted in human AML. Blood. 2013;122(16):
2877–2887.
36. Wang F, Travins J, DeLaBarre B, et al.
Targeted inhibition of mutant IDH2 in
leukemia cells induces cellular differentia-
tion. Science. 2013;340(6132):622–626.
37. Amatangelo MD, Quek L, Shih A, et al.
Enasidenib induces acute myeloid leukemia
cell differentiation to promote clinical
response. Blood. 2017;130(6):732–741.
38. Perl AE. The role of targeted therapy in the
management of patients with AML.
Hematol Am Soc Hematol Educ Program.
2017;2017(1):54–65.
39. Chan SM, Thomas D, Corces-Zimmerman
MR, et al. Isocitrate dehydrogenase 1 and 2
mutations induce BCL-2 dependence in
acute myeloid leukemia. Nat Med.
2015;21(2):178–184.
40. Konopleva M, Pollyea DA, Potluri J, et al.
Efficacy and biological correlates of
response in a phase II study of venetoclax
monotherapy in patients with acute myel-
ogenous leukemia. Cancer Discov. 2016;6
(10):1106–1117.
829
ARTICLE Acute Lymphoblastic Leukemia

MSH6 haploinsufficiency at relapse

Ferrata Storti
Foundation contributes to the development of thiopurine
resistance in pediatric B-lymphoblastic
leukemia
Nikki A. Evensen,1 P. Pallavi Madhusoodhan,1 Julia Meyer,2 Jason Saliba,1
Ashfiyah Chowdhury,1 David J. Araten,3 Jacob Nersting,4 Teena Bhatla,1
Tiffaney L. Vincent,5 David Teachey,5 Stephen P. Hunger,5 Jun Yang,6
Haematologica 2018 Kjeld Schmiegelow4 and William L. Carroll1

Departments of Pediatrics and Pathology, Perlmutter Cancer Center, NYU-Langone

Volume 103(5):830-839
1

Medical Center, New York, NY, USA; 2Huntsman Cancer Institute, University of Utah
Medical Center, Salt Lake City, USA; 3Department of Medicine, Perlmutter Cancer
Center, NYU-Langone Medical Center, New York NY, USA; 4Department of Pediatrics and
Adolescent Medicine, The University Hospital Rigshospitalet, Copenhagen, Denmark;
5
Department of Pediatrics and the Center for Childhood Cancer Research, Children’s
Hospital of Philadelphia and The Perelman School of Medicine at The University of
Pennsylvania, Philadelphia, PA, USA and 6St. Jude Children’s Research Hospital,
Memphis, TN, USA
ABSTRACT

Correspondence:
william.carroll@nyumc.org
Received: July 13, 2017.
Accepted: February 7, 2018.
Pre-published: February 15, 2018.
doi:10.3324/haematol.2017.176362
Check the online version for the most updated
information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/830
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
S
urvival of children with relapsed acute lymphoblastic leukemia is
poor, and understanding mechanisms underlying resistance is
essential to developing new therapy. Relapse-specific heterozygous
deletions in MSH6, a crucial part of DNA mismatch repair, are frequent-
ly detected. Our aim was to determine whether MSH6 deletion results
in a hypermutator phenotype associated with generation of secondary
mutations involved in drug resistance, or if it leads to a failure to initiate
apoptosis directly in response to chemotherapeutic agents. We knocked
down MSH6 in mismatch repair proficient cell lines (697 and UOCB1)
and showed significant increases in IC50s to 6-thioguanine and 6-mer-
captopurine (697: 26- and 9-fold; UOCB1: 5- and 8-fold) in vitro, as well
as increased resistance to 6-mercaptopurine treatment in vivo. No shift in
IC50 was observed in deficient cells (Reh and RS4;11). 697 MSH6 knock-
down resulted in increased DNA thioguanine nucleotide levels com-
pared to non-targeted cells (3070 vs. 1722 fmol/μg DNA) with no differ-
ence observed in mismatch repair deficient cells. Loss of MSH6 did not
give rise to microsatellite instability in cell lines or clinical samples, nor
did it significantly increase mutation rate, but rather resulted in a defect
in cell cycle arrest upon thiopurine exposure. MSH6 knockdown cells
showed minimal activation of checkpoint regulator CHK1, γH2AX
(DNA damage marker) and p53 levels upon treatment with thiopurines,
consistent with intrinsic chemoresistance due to failure to recognize
thioguanine nucleotide mismatching and initiate mismatch repair.
Aberrant MSH6 adds to the list of alterations/mutations associated with
acquired resistance to purine analogs emphasizing the importance of
thiopurine therapy.
Introduction
Relapsed B-precursor acute lymphoblastic leukemia (B-ALL) is a leading cause of
cancer mortality amongst children. Development of chemoresistance is a crucial
factor contributing to relapse, therefore understanding the biological mechanisms
underlying this resistance is imperative for discovering innovative treatment strate-
gies.1 Recent work has begun to highlight the direct role of relapse
specific/enriched genetic alterations in the emergence of clones that have gained a
selective advantage under the pressure of specific chemotherapeutics, such as
NT5C2, TBL1XR1, PRPS1, and CREBBP.1-5 Many of these mutations cause resist-
ance specifically to thiopurines, which are the backbone of maintenance therapy

830 haematologica | 2018; 103(5)


and have proven vital for achieving cures.6 Our analysis of
copy number alterations (CNAs) in diagnosis/relapse pairs
revealed a relapse specific hemizygous deletion on chro-
mosome 2p16.3 involving MSH6 in 4-10% of patients.7,8
MutS homolog 6 (MSH6) is a major component of the
mismatch repair (MMR) system, which is a highly con-
served biological process that recognizes and repairs
errors in nascent DNA strands during replication to main-
tain genomic integrity. Initial recognition of replicative
errors is carried out by protein heterodimers consisting of
either MSH6 and MSH2 (hMutSa), or MSH3 and MSH2
(hMutSβ). Upon recognition of a mismatch, hMutSa
recruits MutLa (MLH1-PMS2) which engages down-
stream proteins and enzymes involved in DNA repair.9,10
Constitutional defects in MMR, including monoallelic
mutations in Lynch syndrome and biallelic loss in consti-
tutional mismatch repair deficiency (CMMRD), are
strongly linked to carcinogenesis, where loss of MMR
functionality causes increased mutability and predisposi-
tion to malignancy.11-15 Previous work has linked defects
in MMR to drug resistance, including thiopurines, in var-
ious cancers.16-19 However, it is uncertain if resistance in
MMR defective clones occurs through the acquisition of
secondary mutations as a consequence of mutagenic ther-
apy, or the outgrowth of clones that have intrinsic drug
resistance. Our lab previously demonstrated that lower
expression of MSH6 in patient samples was associated
with increased ex vivo resistance to 6-mercaptopurine and
prednisone,7 highlighting the clinical importance of under-
standing the role of this genetic alteration in B-ALL.
The mechanism of action of thiopurines is based upon
the insertion of a false nucleotide, namely a thioguanine
(TGN), into DNA that when thiomethylated pairs with a
thymine instead of a cytosine.18 Cytotoxicity is thought to
be dependent on the MMR machinery recognizing the
mismatch and attempting to match the TGN on the
parental strand with an appropriate base on the daughter
strand.19,20 Whether the DNA damage induced by the
repetitive, futile cycles of DNA excision and repair, or
simply the recognition of mismatches by hMutSa is
enough to initiate a signaling cascade culminating in cell
cycle arrest and apoptosis is not entirely understood.
We sought to delineate whether reduced expression of
MSH6 could give rise to chemoresistance in B-precursor
ALL and elucidate the mechanism responsible for the
resistance. Our data here support the view that reduced
MSH6 directly results in an increased tolerance to incor-
porated TGN and subsequent mismatches through a fail-
ure to initiate MMR, thus allowing cells to proliferate and
survive under thiopurine treatment both in vitro and in
vivo. We demonstrate that ALL cell lines with a functional
MMR trigger a CHK1-mediated cell cycle arrest in
response to thiopurines that is followed by DNA damage
and apoptosis. In contrast, upon reduction of MSH6, the
MMR signaling cascade is not fully activated and cells do
not undergo apoptosis.
Methods
Cells and reagents
The B-lineage leukemia cell lines RS4;11 (ATCC, Manassas, VA,
USA), Reh (ATCC), 697 (DSMZ, Braunschweig, Germany), and
UOCB1 (a kind gift from Dr. Terzah Horton at Texas Children’s
Cancer Center/Baylor College of Medicine) were grown in
haematologica | 2018; 103(5)
MSH6 haploinsufficiency contributes to resistance
RPMI1640 medium. HEK293T (ATCC) cells were grown in
DMEM medium. All media were supplemented with 10% FBS,
1% penicillin/streptomycin under 5% CO2 at 37°C.
Drug preparation, viral preparation, immunoblotting,
apoptosis assays, and cell cycle
Standard protocols were followed and have been previously
described.3,21 More detailed information is provided in the Online
Supplementary Appendix.
Patients’ samples
Cryopreserved pediatric B-ALL specimens were obtained from
the Children's Oncology Group (COG) ALL cell bank. All patients
were treated on COG protocols for newly diagnosed ALL. All sub-
jects provided consent for banking and future research use of these
specimens in accordance with the regulations of the institutional
review boards of all participating institutions.
Microsatellite instability analysis
Microsatellite instability (MSI) analysis was performed using
MSI Analysis System, v.1.2 (Promega, Madison, WI, USA) follow-
ing the manufacturer’s protocol. Detailed information is provided
in the Online Supplementary Appendix.
Measurement of mutation rate
Spontaneous mutation rate was measured using a flow cytom-
etry assay previously described by Araten et al.22 that detects the
presence of numerous glycosylphosphatidylinositol-linked (GPI)
membrane proteins (see Online Supplementary Appendix). Briefly,
GPI(+) isolated clones from the NT and MSH6-KD cell lines were
expanded either untreated or treated with 6-TG (0.040 μg/mL and
0.100 μg/mL, respectively, based on IC50 values determined for
clones). Cells were then stained for GPI-dependent markers
including FLAER-Alexa 488 (Pinewood), CD48, CD52, and CD59
(Serotec),23 and analyzed by flow cytometry. The mutant frequen-
cy (f) was calculated as the number of GPI(-) events divided by the
total number of live events, and mutation rate (μ) was calculated
as f divided by cell divisions.22
Thioguanine quantification assay
Cells were treated with 6-thioguanine (6TG) and collected every
day for four days. DNA was extracted using Puregene Core Kit A
(QIAGEN). DNA TGN levels were measured using liquid chro-
matography-tandem mass spectrometry as described previously.24
In vivo mouse model of chemoresistance
All experiments were conducted on protocols approved by
the Institutional Animal Care and Use Committee and
Institutional Review Board of the Children's Hospital of
Philadelphia. Briefly, 1 million UOCB1 NT GFP-CBG or MSH6
shRNA1 GFP-CBR cells were injected into NSG mice via tail vein
on day 0 (total 20 mice; 10 per cell line). On day 6 leukemic bur-
den was confirmed via bioluminescence imaging (BLI) (IVIS
Spectrum imaging system, Perkin Elmer) and animals were ran-
domized to treatment groups [PBS vehicle or Purixan (50 mg/kg)
diluted in PBS]. Mice were treated on day 7 by gavage (0.2
mL/mouse). For BLI, 3 mg of luciferin was injected intraperi-
toneally and mice were imaged ten minutes post injection.
Quantification of total flux was determined by analyzing the BLI
images using Living Image Software (Perkin Elmer) (see Online
Supplementary Appendix).
Statistical analysis
Statistical significance was calculated using unpaired t-test for
IC50s, paired t-test for mutation rates, one-way ANOVA for cell
831
 N.A. Evensen et al.

Figure 1. Knockdown of MSH6 in mismatch repair (MMR) proficient cells lead to decreased sensitivity to thiopurines. (A-C, left) Western blot analysis of whole cell
lysates from 697 (A), UOCB1 (B), and Reh and RS4;11 (C). (A-C, right) Apoptotic cells measured by Annexin V and 7AAD staining followed by flow cytometry after 5
days of treatment. Graphs represent 3 experiments each performed with duplicates. Bars indicate mean+Standard Deviation.

832 haematologica | 2018; 103(5)


cycle analysis, and two-way ANOVA for mutation rates with and
without treatment as well as in vivo studies.
Results
Previously we noted relapse-specific heterozygous dele-
tions in MSH6 in 4 out of 76 patients that were near iden-
tical and deleted MSH6 only for 3 patients while one har-
bored a larger deletion involving more genes within the
region (Online Supplementary Figure S1). To begin to eluci-
date the impact of MSH6 deletion on the development of
relapsed disease, we knocked down expression of MSH6
using shRNA in 697 cells, a B-ALL cell line that expresses
all four MMR proteins (Figure 1A) and is MMR profi-
cient,25 and tested for changes in chemosensitivity. We
observed approximately 80-90% (shRNA1) and 50-60%
(shRNA2) knockdown of MSH6 expression, as well as
decreased expression of MSH2, compared to non-target-
ing (NT) control cells (Figure 1A), which is consistent with
literature on the loss of protein stability of MSH2 and
MSH6 when not dimerized.17,26 Knockdown of MSH6
with both shRNA1 and shRNA2 leads to a significant
decrease in apoptotic cells when treated with thiopurines
for five days (Figure 1A). A 26-fold increase in IC50 with
6-TG (NT: 0.027 vs. shRNA1: 0.716 μg/mL; P=0.007) and
8.5-fold increase for 6-MP (NT: 0.340 vs. shRNA1: 2.89
μg/mL; P=0.006) was observed for shRNA1 (Online
Supplementary Figure S2). A 1.7-fold (NT: 0.015 vs.
shRNA2: 0.025 μg/mL; P=0.015) and a 2.6-fold (NT: 0.143
vs. shRNA2: 0.373 μg/mL; P=0.032) increase in IC50 for 6-
TG and 6-MP, respectively, were observed for shRNA2
cells compared to NT cells (Online Supplementary Figure
A B
haematologica | 2018; 103(5)
MSH6 haploinsufficiency contributes to resistance
S2). However, no significant differences were observed
when cells were treated with prednisolone (Pred), doxoru-
bicin (Doxo), cytarabine (Ara-c), or methotrexate (MTX)
(Online Supplementary Figure S3A). Interestingly, we found
that knockdown of MSH6 also resulted in decreased sen-
sitivity to temozolomide (TMZ), an alkylating agent used
to treat glioblastomas, as reported previously (Online
Supplementary Figure S3B).27,28
To further support the role of MSH6 in chemoresis-
tance, we knocked down expression in UOCB1 cells,
another B-ALL cell line that expresses all four MMR pro-
teins (Figure 1B). Similar to the effect observed in 697 cells,
depletion of MSH6 with either shRNA significantly
reduced the induction of apoptosis upon treatment with
thiopurines (Figure 1B) [fold increase in IC50 as compared
to NT with 6TG: 4.8 for shRNA1 (P=0.007) and 3 for
shRNA2 (P<0.001); 6MP: 8.3 for shRNA1 (P<0.001) and
9.2 shRNA2 (P<0.001)] (Online Supplementary Figure S4).
Additionally, a similar impact on TMZ resistance was
observed with UOCB1 MSH6 shRNA1 expressing cells
compared to NT control cells, although shRNA2 did not
show the same effect, possibly due to less depletion by
shRNA2 (Online Supplementary Figure S3B).
To determine the specificity of the phenotype
observed for MSH6 depletion versus defects in other
MMR proteins, we assessed the effect of MSH6 knock-
down in MMR deficient B-ALL cell lines Reh and
RS4;11.25,29 Both Reh and RS4;11 have minimal to no
expression of MLH1 and PMS2 (Figure 1C). Knockdown
of MSH6 expression had no effect on the sensitivity of
either Reh or RS4;11 to 6-TG or 6-MP (Figure 1C and
Online Supplementary Figure S5).
To begin to elucidate the mechanism of resistance, we
Figure 2. Knockdown of MSH6 leads to increased
tolerance of incorporated thioguanine (TGN). (A and
B) TGN incorporation into DNA was measured over
time after treatment with 0.1 μg/mL of 6-thiogua-
nine (6-TG) in 697 (A) and Reh (B) cells using liquid
chromatography-tandem mass spectrometry. A rep-
resentative graph from 3 independent experiments
is shown. (C) Combined ratio of TGN fmole/μg of DNA
in MSH6 shRNA1 knockdown (KD) cells compared to
non-targeting (NT) cells from 3 experiments. Bars
indicate mean+Standard Deviation.
833
 N.A. Evensen et al.

measured the level of TGN incorporation into DNA upon
treatment with 6-TG. 697 MSH6 shRNA1 cells accumulat-
ed more TGN/μg DNA over time than NT cells (NT 1722
and KD 3070 fmol/μg DNA) (Figure 2A and C). In con-
trast, no difference in TGN levels was observed in Reh
cells (Figure 2B and C). Additionally, Reh cells had approx-
imately 10-fold higher TGN levels compared to 697 cells
(Figure 2A and B), highlighting the difference between
MMR deficient and proficient cells in their ability to
respond to and survive thiopurine exposure. Thus, MMR
proficient cells with high TGN succumb to the damage
and therefore display less TGN/μg DNA over time, mean-
while deficient cells tolerate higher levels of TGN.
We next tested whether or not a change occurs in cell
cycle progression upon treatment. 697 NT cells slowed
their growth and had a significantly higher proportion of
cells in S phase and less cells in G1 beginning at 96 hours
(h) (6-TG, P=0.014; 6-MP, P=0.051) and progressing

Figure 3. Thiopurine treatment resulted in an S phase arrest, which was abrogated upon knockdown of MSH6. 697 (A) and UOCB1 (B) NT and MSH6 shRNA1 and
2 expressing cells were treated with indicated drug for 5 days. Cells were fixed with 70% ethanol, treated with RNAse, and then stained with propidium iodide. DNA
content was analyzed by flow cytometry. Representative images from 3 individual experiments are shown. A one-way ANOVA was performed to determine statistical
significance of the increase in % of cells in S phase at each time point.

834 haematologica | 2018; 103(5)

 MSH6 haploinsufficiency contributes to resistance

through 120 h of thiopurine treatment (6-TG, P=0.013; 6-
MP, P=0.001) compared to the MSH6 shRNA lines by one-
way ANOVA (Figure 3A and Online Supplementary Figure
S6A). MSH6 shRNA1 cells had only a modest decrease in
growth with no clear S phase arrest (6-TG, P=0.011, and
6-MP, P=0.001, for percent of cells in S phase compared to
NT at 120 h using Tukey’s multiple comparison test), even
at higher concentrations of 6-TG (Figure 3A and Online
Supplementary Figure S6A and B). MSH6 shRNA2 cells had
a more moderate accumulation of cells in S phase and
drop of cells in G1 (Figure 3B and Online Supplementary
Figure S6A), which is consistent with the modest levels of
knockdown and apoptosis. Similar trends were observed
with UOCB1 cells (6-TG, P=0.31; and 6-MP, P=0.34 at 120
h) (Figure 3B). This more moderate effect observed with
the UOCB1 cells is consistent with the degree of impact
MSH6 knockdown had on chemoresistance compared to
the 697 cells. Neither NT nor MSH6 shRNA1 Reh cells
showed alterations in cell cycle upon exposure (Online
Supplementary Figure S6B).
To gain a more complete understanding of the mecha-
nism leading to apoptosis following TGN incorporation,
we analyzed downstream pathways in 697 NT and MSH6
shRNA1 cells after treatment with 6-TG. Based on the
observed S phase arrest and previous research demonstrat-
ing activation of the ataxia telangiectasia and Rad3-related
(ATR)-Chk1 pathway downstream of MMR,30,31 we first
assessed the level of activation of Chk1 by probing for
phosphorylation of serine 317 (pChk1). 697 NT cells had
a low level of pChk1 at 48 h with a significant increase
through 96 h of exposure. 697 MSH6 shRNA1 cells had
minimal to low levels of pChk1 at 72 h with minimal
increase over time (Figure 4A). The 697 cells expressing
shRNA2 had a similar pattern but slightly lower levels of
pChk1 compared to NT cells, which is consistent with the
cell cycle data. We next assessed the level of phosphory-
lated H2AX (γH2AX), a marker of DNA damage that is
phosphorylated downstream of the ATR/ATM pathways
following drug treatment,32 as well as levels of the apopto-
sis marker p53. There was a very modest level of γH2AX
starting at 72 h that increased to a higher level at 96 h after
treatment in 697 NT cells compared to very modest levels
in the MSH6 shRNA1-2 cells (Figure 4A), suggesting that
the functional MMR system in the NT cells was attempt-
ing to repair the DNA, leading to nicks. Additionally, the
levels of phosphorylated and total p53 were higher in NT
cells at 72 and 96 h compared to MSH6 shRNA1-2 cells
(Figure 4A) and the shRNA2 cells had higher levels than
the shRNA1 cells. Similar results were found in UOCB1
cells (Figure 4B).
We next examined the impact of MSH6 knockdown on
mutation rate by performing two assays that measure
genomic instability and mutation burden. Microsatellite
instability (MSI) is a marker for genomic instability and
has been observed in cases where expression of MLH1 or
MSH2 is lost.25,33 We investigated MSI on 2 patient sample
pairs that we previously found to have deletions of MSH6
at relapse, as well as on 697 NT and MSH6 shRNA1 cells
treated with 6-TG for 120 h. No MSI was observed in the
patient samples comparing diagnosis to relapse or in the
697 cells comparing either untreated to 6-TG treated or
NT to MSH6 shRNA1 cells (Figure 5A). These data are
consistent with previous literature that found alterations
in MSH6 expression alone do not lead to high MSI.34 To
investigate the effect of MSH6 disruption on the rate of
spontaneous mutations in PIG-A, which is required for
expression of GPI, we used a flow cytometry-based assay
that measures surface expression of several GPI-depen-
dent markers (CD48, CD52, and CD59).22,35 Although
there was a trend to suggest that 697 MSH6 shRNA1 cells
had a slightly higher mutation rate, statistical significance
was not achieved (Figure 5B). Furthermore, treatment of
the clones from each cell line with 6-TG did not lead to an
increased mutation rate (Figure 5B).
To investigate the clinical relevance of reduced MSH6
expression and drug resistance, we utilized an in vivo
mouse model. We injected mice with either UOCB1 NT
or UOCB1 MSH6 shRNA1 cell lines (knockdown con-

A B

Figure 4. Thiopurine treatment leads to activation of cell cycle regulator Chk1 and DNA repair that ultimately resulted in DNA damage and cell death. Western
blot analysis of whole cell lysates from 697 (A) and UOCB1 (B) non-targeting (NT), MSH6 shRNA1, and shRNA2 cells after treatment with 6-thioguanine (0.1 μg/mL,
and 0.025 μg/mL, respectively). (C) Untreated cells; numbers are hours after treatment. Blots were probed for Chk1 activation, γH2AX for DNA damage, and apop-
tosis marker p53. Total Chk1, actin, and total H2AX were used as loading controls. Images are representative of 3 individual experiments.

haematologica | 2018; 103(5) 835

 N.A. Evensen et al.

firmed day of injection; Figure 6C) and, following confir-
mation of leukemic burden on day 6, the mice were treat-
ed with PBS (control) or purixan (an oral suspension form
of 6-MP). Following the 10-day course of purixan treat-
ment, the leukemic burden was significantly diminished
in the mice harboring NT cells compared to that observed
in the NT PBS treated mice (P=0.0001), suggesting that
these cells were unable to survive and expand under the
selective pressure of the purixan (Figure 6A and B). In con-
trast, the MSH6 shRNA1 mice treated with purixan were
not significantly different from the NT PBS group
(P=0.828). Although purixan also had a statistically signif-
icant impact on MSH6 shRNA1 cells compared to MSH6
shRNA1 PBS control (P=0.0005), these cells were able to
continue proliferating under the selective pressure, unlike
the NT cells (Figure 6A and B). Finally, a comparison
between PBS MSH6 shRNA1 and PBS control NT cells at
day 17 showed that MSH6 depleted cells also had a
growth advantage in vivo (P<0.0001).
Discussion
In recent years, there has been an abundance of evi-
dence demonstrating the outgrowth of clones at relapse in
ALL that are associated with unique or enriched relapse
specific mutations that confer drug resistance. Some of the
most common relapse specific mutations found thus far
occur in NT5C2 and PRPS1 and lead to the outgrowth of
thiopurine resistant clones.2,4 Our data presented here
demonstrate that reduction of MSH6 in ALL also leads to
decreased sensitivity to purine analogs due to a failure to
initiate the apoptotic cascade directly in response to
nucleotide mismatches. Even with only 50-60% reduced
expression, which potentially mimics levels in patients
with heterozygous loss, we demonstrate a significant
decrease in sensitivity to thiopurines. Our data are consis-
tent with the recent work of Diouf et al. who showed that
lower levels of MSH2 in cell lines were associated with
resistance to 6-TG and 6-MP. They found 11% of ALL
samples showed decreased protein levels of MSH2
through copy number loss of genes controlling MSH2
degradation.17 Thus defects in MMR, including heterozy-
gous deletion of MSH6, can be added to the list of genetic
alterations that result in the development of resistance to
purine analogs, the foundation of maintenance therapy.
The variety of mutations that lead to selective outgrowth
of such clones in a substantial number of patients under-
scores the selective pressure of thiopurines on tumor cells.
The outgrowth of MSH6 deleted/mutated clones not
found at diagnosis has been observed at relapse in malig-
nant gliomas following treatment with temozolo-

A B

Figure 5. Knockdown of MSH6 did not lead to a mutator phenotype or increased mutation rate. (A) Microsatellite instability (MSI) was measured in diagnosis/relapse
pairs that had relapse specific, heterozygous MSH6 deletions and in 697 non-targeting (NT) and MSH6 shRNA1 cells left untreated or treated with 0.05 μg/mL of
6-thioguanine (6-TG) for 5 days. (B) Mutation rate in the PIC-A gene was measured in 697 NT and MSH6 shRNA1 clones that were expanded for 2-3 weeks with or
without 6-TG. The cells were analyzed for loss of GPI-dependent cell surface markers, including FLAER, CD48, CD52, and CD59 using flow cytometry. (Left) Individual
mutation rates/cell divisions for each clone; the line represents mean+Standard Deviation. (Right) Mutation rates/cell divisions for three clones with and without 6-
TG treatment.

836 haematologica | 2018; 103(5)

 MSH6 haploinsufficiency contributes to resistance

mide,28,36,37 which produces DNA O6-methyguanine, a
lesion structurally similar to 6-TG.38 Our data demonstrat-
ing decreased sensitivity of MSH6 knockdown cells to
temozolomide support the hypothesis that MMR defi-
cient clones gain an advantage under this selective pres-
sure leading to resistant recurrences.27,39 The difference in
TMZ sensitivity between the two UOCB1 shRNA knock-
down cell lines could be due to the interplay between
MSH6 and SETD2 protein levels since UOCB1 cells have
a copy number loss of SETD2 (NA Evensen et al., 2018,
unpublished data) and there is a greater reduction of MSH6
with shRNA1 compared to shRNA2. SETD2, the gene that
codes for the methyltransferase responsible for the
trimethylation of H3K36 that serves as the docking site for
MSH6,40 is among the epigenetic regulators commonly
found mutated in relapse patients.41 Ongoing studies in our
lab are focused on identifying the relationship of epigenet-
ic readers, writers, and erasers, such as SETD2, MSH6,
and WHSC1 in chemoresistance.
Mechanistically, our in vitro and in vivo data support the
hypothesis that the delayed cytotoxic response to thiop-
urines is due to the MMR system recognizing a mismatch
and initiating futile, damaging DNA repair that ultimately
leads to apoptosis.20,42,43 This pathway is not fully activated
in cells with reduced MSH6 because the mismatch goes
undetected, allowing these cells to tolerate excess TGN
mismatches and, ultimately, to continue to survive and
proliferate while under treatment. Our data provide evi-
dence that, upon recognition of mismatches, NT cells
slow their progression through S phase by activating
Chk1 as they begin to repair their DNA. Due to the mis-
match being on the daughter strand, the excision/repair
process is unsuccessful, and over time nicks build up in the
DNA, demonstrated by increased levels of γH2AX.
Eventually, the damage becomes overwhelming and cells
initiate apoptosis, as shown by increased p53. MSH6
shRNA1 cells exhibited minimal to no change in cell cycle,
activation of Chk1, or increased γH2AX and p53. The
moderate changes observed with the MSH6 shRNA2 cells
highlight the idea that even a more modest reduction in
MSH6 expression could lead to subtle changes that have a
significant impact on chemoresistance. The MMR defi-
cient Reh cells also had no alteration in their cell cycle,
suggesting that recognition of mismatches by MutSa is
not sufficient for full activation of this cascade, but rather
damage induced by the repair, which is orchestrated by

B C

Figure 6. Knockdown of MSH6 leads to decreased sensitivity to purixan in vivo. (A) Bioluminescence imaging (BLI) of mice injected with UOCB1 non-targeting (NT)
or MSH6 shRNA1 cells. Six days after injection mice were imaged and then randomized to treatment. Treatment was started on day 7 and images were taken again
on days 13 and 17. C: PBS control treatment; T: purixan treatment. (B) Quantification of total flux was determined by analyzing the BLI images using Living Image
software. (C) Western blot to confirm knockdown of MSH6 in cells used to inject mice. Actin was used as loading control.

haematologica | 2018; 103(5) 837

 N.A. Evensen et al.

MutLa, is what initiates apoptosis. Our data demonstrat-
ing the involvement of the ATR-Chk1-H2AX signaling
cascade is supported by the work of Eich et al. which
demonstrated activation of this pathway upon treatment
with temozolomide.32 Interestingly, understanding how
MMR deficient cells respond to thiopurines in terms of
TGN incorporation could prove essential given the emerg-
ing idea of measuring these parameters in patients on
maintenance therapy.44
The data presented here do not support the hypothesis
of increased mutation burden, genomic instability, or MSI
when MSH6 is reduced. Our inability to demonstrate
MSI-high, which is considered a standard method for clin-
ical testing of MMR deficiencies in tumors,45 in MSH6
depleted cell lines and clinical samples is consistent with
the lack of MSI in glioma samples with MSH6 deletions or
mutations.46,47 Haploinsufficiency of MSH6 or compensa-
tion by MSH2/MSH3 may account for this observation.25,48
In addition, ALL clonal evolution from diagnosis to relapse
is not associated with increased mutation burden support-
ing our mutation rate analysis, although Ma et al. reported
a subset of hypermutated relapse cases.49 Of these, one
had a bialleic mutation of PMS2, another had multiple
damaging MSH6 mutations as well as an MLH1 splice site
mutation, while the others harbored no MMR muta-
tions.49 Furthermore, one case demonstrated that a het-
erozygous deletion of MSH6 at diagnosis was not suffi-
cient to cause a hypermutator phenotype, but the acquisi-
tion of a second hit in the WT allele at relapse was.49
Likewise, the majority of hypermutated gliomas at relapse
show defects in multiple MMR genes or loss of heterozy-
gosity.50 Thus, our work supports a model whereby hap-
loinsufficiency of MSH6 results in TGN tolerance and
resistance directly rather than by generation of secondary
mutations. However, it does not rule out the possibility
that haploinsufficiency, along with other defects in the
MMR pathway, may result in a mutator phenotype.
Overall, it has become increasingly evident that the
genetic and epigenetic landscape of cancer cells is vital to
the overall effectiveness of treatment. These studies illus-
trate yet another example of a mutation/deletion found at
relapse that directly influences the response to a therapeu-
tic agent that is currently heavily relied on. Continuous
efforts to elucidate the potential functions and mecha-
nisms of genes found mutated at relapse will help lead us
to novel treatment strategies.
Funding
This work was supported by the Leukemia and Lymphoma
Society SCOR grant: 7010-14 (WLC, JY, DT, SPH), the US
National Institutes of Health (NIH) funded grant RO1
CA140729 (WLC), and the Perlmutter Cancer Center Support
Grant: P30 DA016087.
Acknowledgments
We gratefully acknowledge the Children’s Oncology Group
(COG) Specimen Bank for samples. Support for flow cytometry
was provided by NYU School of Medicine’s Cytometry and Cell
Sorting Laboratory, which is supported in part by grant
P30CA016087 from the NIH/NCI, and the CHOP Flow
Cytometry Core. We acknowledge the VA-Mertid award
1I01BX-000670, which helped support this work.

References
1. Pui CH, Evans WE. Treatment of acute
lymphoblastic leukemia. N Engl J Med.
2006;354(2):166-178.
2. Meyer JA, Wang J, Hogan LE, et al. Relapse-
specific mutations in NT5C2 in childhood
acute lymphoblastic leukemia. Nat Genet.
2013;45(3):290-294.
3. Jones CL, Bhatla T, Blum R, et al. Loss of
TBL1XR1 disrupts glucocorticoid receptor
recruitment to chromatin and results in glu-
cocorticoid resistance in a B-lymphoblastic
leukemia model. J Biol Chem. 2014;
289(30):20502-20515.
4. Li B, Li H, Bai Y, et al. Negative feedback-
defective PRPS1 mutants drive thiopurine
resistance in relapsed childhood ALL. Nat
Med. 2015;21(6):563-571.
5. Mullighan CG, Zhang J, Kasper LH, et al.
CREBBP mutations in relapsed acute lym-
phoblastic leukaemia. Nature. 2011;
471(7337):235-239.
6. Nielsen SN, Grell K, Nersting J, et al. DNA-
thioguanine nucleotide concentration and
relapse-free survival during maintenance
therapy of childhood acute lymphoblastic
leukaemia (NOPHO ALL2008): a prospec-
tive substudy of a phase 3 trial. Lancet
Oncol. 2017;18(4):515-524.
7. Yang JJ, Bhojwani D, Yang W, et al.
Genome-wide copy number profiling
reveals molecular evolution from diagnosis
to relapse in childhood acute lymphoblastic
leukemia. Blood. 2008;112(10):4178-4183.
8. Hogan LE, Meyer JA, Yang J, et al.
Integrated genomic analysis of relapsed
childhood acute lymphoblastic leukemia
reveals therapeutic strategies. Blood.
2011;118(19):5218-5226.
9. Li GM. Mechanisms and functions of
DNA mismatch repair. Cell Res. 2008;18
(1):85-98.
10. Edelbrock MA, Kaliyaperumal S, Williams
KJ. Structural, molecular and cellular func-
tions of MSH2 and MSH6 during DNA
mismatch repair, damage signaling and
other noncanonical activities. Mutat Res.
2013;743-744:53-66.
11. Dunlop MG, Farrington SM, Carothers AD,
et al. Cancer risk associated with germline
DNA mismatch repair gene mutations.
Hum Mol Genet. 1997;6(1):105-110.
12. Tiwari AK, Roy HK, Lynch HT. Lynch syn-
drome in the 21st century: clinical perspec-
tives. QJM. 2016;109(3):151-158.
13. Goecke T, Schulmann K, Engel C, et al.
Genotype-phenotype comparison of
German MLH1 and MSH2 mutation carri-
ers clinically affected with Lynch syn-
drome: a report by the German HNPCC
Consortium. J Clin Oncol. 2006;
24(26):4285-4292.
14. Ripperger T, Schlegelberger B. Acute lym-
phoblastic leukemia and lymphoma in the
context of constitutional mismatch repair
deficiency syndrome. Eur J Med Genet.
2016;59(3):133-142.
15. Ripperger T, Beger C, Rahner N, et al.
Constitutional mismatch repair deficiency
and childhood leukemia/lymphoma--
report on a novel biallelic MSH6 mutation.
Haematologica. 2010;95(5):841-844.
16. Fink D, Aebi S, Howell SB. The role of
DNA mismatch repair in drug resistance.
Clin Cancer Res. 1998;4(1):1-6.
17. Diouf B, Cheng Q, Krynetskaia NF, et al.
Somatic deletions of genes regulating
MSH2 protein stability cause DNA mis-
match repair deficiency and drug resistance
in human leukemia cells. Nat Med.
2011;17(10):1298-1303.
18. Swann PF, Waters TR, Moulton DC, et al.
Role of postreplicative DNA mismatch
repair in the cytotoxic action of thiogua-
nine. Science. 1996;273(5278):1109-1111.
19. Waters TR, Swann PF. Cytotoxic mecha-
nism of 6-thioguanine: hMutSalpha, the
human mismatch binding heterodimer,
binds to DNA containing S6-methylth-
ioguanine. Biochemistry. 1997;36(9):2501-
2506.
20. Karran P, Attard N. Thiopurines in current
medical practice: molecular mechanisms
and contributions to therapy-related cancer.
Nat Rev Cancer. 2008;8(1):24-36.
21. Morrison DJ, Hogan LE, Condos G, et al.
Endogenous knockdown of survivin
improves chemotherapeutic response in
ALL models. Leukemia. 2012;26(2):271-
279.
22. Araten DJ, Golde DW, Zhang RH, et al. A
quantitative measurement of the human
somatic mutation rate. Cancer Res.
2005;65(18):8111-8117.
23. Araten DJ, Sanders KJ, Anscher D,
Zamechek L, Hunger SP, Ibrahim S.
Leukemic blasts with the paroxysmal noc-
turnal hemoglobinuria phenotype in chil-
dren with acute lymphoblastic leukemia.
Am J Pathol. 2012;181(5):1862-1869.
24. Jacobsen JH, Schmiegelow K, Nersting J.
Liquid chromatography-tandem mass spec-
trometry quantification of 6-thioguanine in

838 haematologica | 2018; 103(5)


DNA using endogenous guanine as internal
standard. J Chromatogr B Analyt Technol
Biomed Life Sci. 2012;881-882:115-118.
25. Gu L, Cline-Brown B, Zhang F, Qiu L, Li
GM. Mismatch repair deficiency in hema-
tological malignancies with microsatellite
instability. Oncogene. 2002;21(37):5758-
5764.
26. Marra G, Iaccarino I, Lettieri T, Roscilli G,
Delmastro P, Jiricny J. Mismatch repair defi-
ciency associated with overexpression of
the MSH3 gene. Proc Natl Acad Sci USA.
1998;95(15):8568-8573.
27. McFaline-Figueroa JL, Braun CJ, Stanciu M,
et al. Minor Changes in Expression of the
Mismatch Repair Protein MSH2 Exert a
Major Impact on Glioblastoma Response to
Temozolomide. Cancer Res. 2015;
75(15):3127-3138.
28. Xie C, Sheng H, Zhang N, Li S, Wei X,
Zheng X. Association of MSH6 mutation
with glioma susceptibility, drug resistance
and progression. Mol Clin Oncol. 2016;
5(2):236-240.
29. Matheson EC, Hall AG. Assessment of mis-
match repair function in leukaemic cell
lines and blasts from children with acute
lymphoblastic leukaemia. Carcinogenesis.
2003;24(1):31-38.
30. Yoshioka K, Yoshioka Y, Hsieh P. ATR
kinase activation mediated by MutSalpha
and MutLalpha in response to cytotoxic
O6-methylguanine adducts. Mol Cell.
2006;22(4):501-510.
31. Yan T, Desai AB, Jacobberger JW,
Sramkoski RM, Loh T, Kinsella TJ. CHK1
and CHK2 are differentially involved in
mismatch repair-mediated 6-thioguanine-
induced cell cycle checkpoint responses.
Mol Cancer Ther. 2004;3(9):1147-1157.
32. Eich M, Roos WP, Nikolova T, Kaina B.
Contribution of ATM and ATR to the
resistance of glioblastoma and malignant
melanoma cells to the methylating anti-
cancer drug temozolomide. Mol Cancer
Ther. 2013;12(11):2529-2540.
haematologica | 2018; 103(5)
MSH6 haploinsufficiency contributes to resistance
33. Verma L, Kane MF, Brassett C, et al.
Mononucleotide microsatellite instability
and germline MSH6 mutation analysis in
early onset colorectal cancer. J Med Genet.
1999;36(9):678-682.
34. Bodo S, Colas C, Buhard O, et al. Diagnosis
of Constitutional Mismatch Repair-
Deficiency Syndrome Based on
Microsatellite Instability and Lymphocyte
Tolerance to Methylating Agents.
Gastroenterology. 2015;149(4):1017-1029.
e1013.
35. Araten DJ, Krejci O, Ditata K, et al. The rate
of spontaneous mutations in human
myeloid cells. Mutat Res. 2013;749(1-2):49-
57.
36. Cahill DP, Codd PJ, Batchelor TT, Curry
WT, Louis DN. MSH6 inactivation and
emergent temozolomide resistance in
human glioblastomas. Clin Neurosurg.
2008;55:165-171.
37. Cahill DP, Levine KK, Betensky RA, et al.
Loss of the mismatch repair protein MSH6
in human glioblastomas is associated with
tumor progression during temozolomide
treatment. Clin Cancer Res. 2007;
13(7):2038-2045.
38. Zhang J, Stevens MF, Laughton CA,
Madhusudan S, Bradshaw TD. Acquired
resistance to temozolomide in glioma cell
lines: molecular mechanisms and potential
translational applications. Oncology. 2010;
78(2):103-114.
39. Hunter C, Smith R, Cahill DP, et al. A
hypermutation phenotype and somatic
MSH6 mutations in recurrent human
malignant gliomas after alkylator
chemotherapy. Cancer Res. 2006;
66(8):3987-3991.
40. Li F, Mao G, Tong D, et al. The histone
mark H3K36me3 regulates human DNA
mismatch repair through its interaction
with MutSalpha. Cell. 2013;153(3):590-600.
41. Mar BG, Bullinger LB, McLean KM, et al.
Mutations in epigenetic regulators includ-
ing SETD2 are gained during relapse in pae-
diatric acute lymphoblastic leukaemia. Nat
Commun. 2014;5:3469.
42. Yan T, Berry SE, Desai AB, Kinsella TJ.
DNA mismatch repair (MMR) mediates 6-
thioguanine genotoxicity by introducing
single-strand breaks to signal a G2-M arrest
in MMR-proficient RKO cells. Clin Cancer
Res. 2003;9(6):2327-2334.
43. Tidd DM, Paterson AR. Distinction
between inhibition of purine nucleotide
synthesis and the delayed cytotoxic reac-
tion of 6-mercaptopurine. Cancer Res.
1974;34(4):733-737.
44. Nielsen SN, Grell K, Nersting J, Frandsen
TL, Hjalgrim LL, Schmiegelow K. Measures
of 6-mercaptopurine and methotrexate
maintenance therapy intensity in child-
hood acute lymphoblastic leukemia.
Cancer Chemother Pharmacol. 2016;
78(5):983-994.
45. Buza N, Ziai J, Hui P. Mismatch repair defi-
ciency testing in clinical practice. Expert
Rev Mol Diagn. 2016;16(5):591-604.
46. Yip S, Miao J, Cahill DP, et al. MSH6 muta-
tions arise in glioblastomas during temo-
zolomide therapy and mediate temozolo-
mide resistance. Clin Cancer Res.
2009;15(14):4622-4629.
47. Maxwell JA, Johnson SP, McLendon RE, et
al. Mismatch repair deficiency does not
mediate clinical resistance to temozolo-
mide in malignant glioma. Clin Cancer Res.
2008;14(15):4859-4868.
48. Umar A, Risinger JI, Glaab WE, Tindall KR,
Barrett JC, Kunkel TA. Functional overlap
in mismatch repair by human MSH3 and
MSH6. Genetics. 1998;148(4):1637-1646.
49. Ma X, Edmonson M, Yergeau D, et al. Rise
and fall of subclones from diagnosis to
relapse in pediatric B-acute lymphoblastic
leukaemia. Nat Commun. 2015;6:6604.
50. van Thuijl HF, Mazor T, Johnson BE, et al.
Evolution of DNA repair defects during
malignant progression of low-grade
gliomas after temozolomide treatment.
Acta Neuropathol. 2015;129(4):597-607.
839
ARTICLE
Ferrata Storti
Foundation
Haematologica 2018
Volume 103(5):840-848
Preliminary results were presented at the 58th
Annual Meeting of the American Society of
Hematology, held on December 3 – 6, 2016,
in San Diego, USA.
Correspondence:
franck.morschhauser@chru-lille.fr
Received: September 12, 2017.
Accepted: January 10, 2018.
Pre-published: January 19, 2018.
doi:10.3324/haematol.2017.180554
Check the online version for the most updated
information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/840
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
840
Hodgkin Lymphoma
A phase II study of the oral JAK1/JAK2
inhibitor ruxolitinib in advanced
relapsed/refractory Hodgkin lymphoma
Eric Van Den Neste,1 Marc André,2 Thomas Gastinne,3 Aspasia Stamatoullas,4
Corinne Haioun,5 Amine Belhabri,6 Oumedaly Reman, 7Olivier Casasnovas,8
Hervé Ghesquieres,9 Gregor Verhoef,10 Marie-José Claessen,11 Hélène A.
Poirel,12 Marie-Christine Copin,13 Romain Dubois,13 Peter Vandenberghe,14
Ioanna-Andrea Stoian,15 Anne S. Cottereau,16 Sarah Bailly,1 Laurent Knoops17
and Franck Morschhauser18
Department of Hematology, Cliniques Universitaires Saint-Luc, UCL Brussels, Belgium;
1
2
Hematology Department, CHU UCL Namur, Yvoir, Belgium; 3Hematology, CHU Nantes,
France; 4Clinical Hematology, Centre Henri Becquerel, Rouen, France; 5Lymphoid
Malignancies Unit, AP-HP, Groupe Hospitalier Mondor, Créteil, France; 6Onco-hematology,
Centre Leon Berard, University Claude Bernard Lyon 1, France; 7Hematology, Centre
Hospitalier Universitaire, Caen, France; 8Hematology Department, Hopital Le Bocage, CHU
Dijon, France; 9Hospices Civils de Lyon, Université Claude Bernard, Centre Hospitalier Lyon-
Sud, Pierre Bénite, France; 10Department of Hematology, University Hospitals Leuven,
Belgium; 11Erasmus MC, Rotterdam, the Netherlands; 12Center for Human Genetics,
Cliniques universitaires Saint-Luc, Université Catholique de Louvain, Brussels, Belgium;
13
CHRU de Lille, France; 14Center for Human Genetics, Katholieke Universiteit - Leuven,
Belgium; 15Nuclear Medicine, Cliniques Universitaires Saint-Luc, UCL Brussels, Belgium;
16
Nuclear Medicine, Hôpital Tenon, Paris, France; 17Cliniques Universitaires Saint-Luc and
de Duve Institute, Université Catholique de Louvain, Brussels, Belgium and 18CHU Lille,
Hematology Department, and Université de Lille, GRITA, France
ABSTRACT
J
AK2 constitutive activation/overexpression is common in classical
Hodgkin lymphoma, and several cytokines stimulate Hodgkin lym-
phoma cells by recognizing JAK1-/JAK2-bound receptors. JAK block-
ade may thus be therapeutically beneficial in Hodgkin lymphoma. In this
phase II study we assessed the safety and efficacy of ruxolitinib, an oral
JAK1/2 inhibitor, in patients with relapsed/refractory Hodgkin lym-
phoma. The primary objective was overall response rate according to the
International Harmonization Project 2007 criteria. Thirty-three patients
with advanced disease (median number of prior lines of treatment: 5;
refractory: 82%) were included; nine (27.3%) received at least six cycles
of ruxolitinib and six (18.2%) received more than six cycles. The overall
response rate after six cycles was 9.4% (3/32 patients). All three respon-
ders had partial responses; another 11 patients had transient stable dis-
ease. Best overall response rate was 18.8% (6/32 patients). Rapid allevia-
tion of B-symptoms was common. The median duration of response was
7.7 months, median progression-free survival 3.5 months (95% CI: 1.9-
4.6), and the median overall survival 27.1 months (95% CI: 14.4-27.1).
Forty adverse events were reported in 14/33 patients (42.4%). One event
led to treatment discontinuation, while 87.5% of patients recovered
without sequelae. Twenty-five adverse events were grade 3 or higher.
These events were mostly anemia (n=11), all considered related to ruxoli-
tinib. Other main causes of grade 3 or higher adverse events included
lymphopenia and infections. Of note, no cases of grade 4 neutropenia or
thrombocytopenia were observed. Ruxolitinib shows signs of activity,
albeit short-lived, beyond a simple anti-inflammatory effect. Its limited
toxicity suggests that it has the potential to be combined with other ther-
apeutic modalities. ClinicalTrials.gov: NCT01877005
haematologica | 2018; 103(5)

Introduction
Hodgkin lymphoma (HL) is regarded as a curable malig-
nancy in most cases, yet treatment failure still occurs in
about 10% of patients with early-stage disease.1 In
advanced-stage disease, up to 10% of cases do not reach
complete remission and are thus considered to have pri-
mary refractory HL,2 while 20-30% of primary responders
eventually relapse following first-line treatment.3
For most patients with relapsed or refractory HL (R/R
HL), the standard of care consists of high-dose salvage
chemotherapy followed by autologous stem-cell transplan-
tation (SCT). For patients who develop R/R HL within 1
year of autologous SCT, the prognosis proves extremely
poor, since they have a median survival of 1.2 years.4 For
patients in whom all classical approaches have failed, new
strategies, including checkpoint inhibitors targeting PD-1
and antibody-drug conjugates targeting CD30, have
become part of the therapeutic armamentarium against
R/R HL.5-8 However, patients with multiple relapses or
those who develop refractory disease remain in medical
need, especially those in whom treatment with brentux-
imab-vedotin (BV) and PD-1 blockers fails.
Classical HL is characterized by the presence of
Hodgkin and Reed-Sternberg (HRS) cells and their vari-
ants.9 HRS cells were demonstrated to shape their envi-
ronment by secreting immunosuppressive cytokines and
chemokines.10 With this in mind, the Janus kinase (JAK) –
signal transducer and activator of transcription (STAT)
pathway appears to be a relevant cytokine-induced signal
transduction pathway that has been shown to transfer sig-
nals directly from cell surface cytokine receptors to the cell
nucleus. Given that enhanced JAK-mediated signaling has
been demonstrated in a significant number of HL
patients,11 this signaling pathway has become a focus for
developing novel therapeutic agents for the disease. Van
Roosbroeck et al. reported the translocation of JAK2 in
several cases of HL,12 and JAK inhibition was shown to
decrease the proliferation of cell lines. Although such
translocations are relatively rare, 9p24.1 genomic amplifi-
cation including the JAK2 locus appears common in HL,
along with increased protein expression and activity,
resulting in the constitutive activation of STAT6, an essen-
tial messenger of tumor cell growth.13-15 In corollary, JAK
1/2 inhibition may be suitable to target the constitutive
activation caused by either JAK2 translocation or JAK2
amplification and to modify the reactive microenviron-
ment which contributes to HL growth via aberrant
cytokine production.16
Ruxolitinib is the first potent, selective, and oral inhibitor
of JAK1/2 being developed for clinical use.17 Its major
effects include inhibition of proliferation, induction of
apoptosis, and reduction in cytokine plasma levels, all
mediated by the drug's ability to inhibit JAK-induced phos-
phorylation of STAT.18 Used in the treatment of myelofi-
brosis, ruxolitinib had durable efficacy in reducing
splenomegaly and alleviating constitutional symptoms, the
patients gained weight and their general physical condition
improved.19 The dose-limiting toxicity was thrombocy-
topenia, which was fairly well managed by dose reduc-
tions or brief interruptions of treatment. In the present
phase II study, we sought to investigate the safety and effi-
cacy of ruxolitinib in patients with R/R HL. Exploratory
biomarker analyses pertaining to plasma cytokine profiles
and aberrations of JAK2 were also carried out.
haematologica | 2018; 103(5)
Ruxolitinib in advanced relapsed/refractory HL
Methods
Patients’ eligibility
Patients aged 18 years or older with a diagnosis of R/R HL for
whom no treatment with proven efficacy was available were eli-
gible to enter the trial after having receiving at least one prior
therapy provided that they had measurable nodal disease at
baseline (≥1 cm in the longest transverse diameter, clearly meas-
urable in at least two perpendicular dimensions) on computed
tomography or magnetic resonance imaging, as well as an
Eastern Cooperative Oncology Group performance score of ≤3.
Additional inclusion criteria were an absolute neutrophil count
≥1.0 x 109/L, platelet count ≥75 x 109/L, serum creatinine ≤1.5 x
upper limit of normal, serum bilirubin ≤1.5 x upper limit of nor-
mal, and ALT and AST levels ≤2.5 or ≤5.0 x upper limit of nor-
mal in the event the transaminase increase was due to HL-relat-
ed liver disease. Pregnant or lactating patients were not allowed
to enter the trial, and men and women of childbearing potential
had to agree to employ an adequate contraceptive method dur-
ing the study treatment. Patients were permitted to have
received an undefined number of prior lines of therapy, and a
previous allogeneic SCT was likewise allowed provided that
patients had not received any immunosuppressive therapy with-
in the 90 days prior to starting the screening procedures. Patients
were required to have a life expectancy of ≥3 months.
Study design and treatment
This multicenter, open-label, phase II study (HIJAK,
NCT01877005) was conducted at ten LYSA centers in France and
Belgium, with patients recruited from July 2013 through
December 2014. Its primary efficacy endpoint was overall
response rate (ORR), defined as the proportion of patients with a
complete response or partial response at 6 months of treatment by
investigator assessment based on the revised 2007 International
Harmonization Project response criteria for malignant lym-
phoma.20 Secondary objectives included relief of B symptoms, best
ORR (occurring at any time during study), duration of response,
progression-free survival, overall survival, as well as the incidence
and severity of adverse events.
The study was carried out in line with the ethical principles of
the Helsinki Declaration and in compliance with the International
Conference on Harmonization Guideline for Good Clinical
Practice. The protocol was approved by the institutional review
board of each study site and written informed consent was
obtained from all patients.
The starting dose of ruxolitinib was 20 mg given twice daily dur-
ing six 28-day cycles for the induction period if the platelet count
was >200 x 109/L. The ruxolitinib dose was decreased to 15 mg
twice daily in patients with platelet counts between 75 x 109/L and
200 x 109/L. Patients who achieved at least stable disease at the end
of cycle 6 and who had, in the investigator's opinion, a clinical ben-
efit were eligible to continue ruxolitinib (15 mg or 20 mg), which
was defined as “maintenance” therapy. Treatment could be contin-
ued for up to 2 years or until progressive disease, intolerability, or
as long as the investigator thought that there was clinical benefit.
Administration of the study drug could be stopped for any grade
≥3 non-hematologic toxicity, with the exception of deep venous
thrombosis and alopecia. Following event resolution to grade ≤1,
ruxolitinib could be resumed, with a 5 mg dose reduction and a
maximum delay of 4 weeks. Mandatory dose decreases or inter-
ruptions for hematologic toxicity as well as the rules for perma-
nent discontinuation are detailed in the Online Supplementary
Appendix. Growth factors were allowed as per American Society of
Clinical Oncology guidelines and infectious prophylaxis as per the
guidelines of Heine et al.21
841
 E. Van Den Neste et al.

Study assessments safety analysis comprised all patients who received at least one
Baseline assessments comprised documentation of disease- dose of the study drug. All statistical analyses were performed
related symptoms, physical examination, laboratory tests, and using SAS software, version 9.2. P-values <0.05 were consid-
imaging studies of the neck, chest, abdomen, and pelvis, using ered statistically significant. All available data were included in
computed tomography or magnetic resonance imaging. Biopsy data listings and tabulations, with no imputations of values for
prior to inclusion was recommended, but not mandatory. Tumors missing data. An interim analysis was neither planned nor per-
were measured at baseline, at the end of every two cycles of rux- formed.
olitinib, and following the six-cycle induction, as well as during
maintenance therapy. Given the exploratory nature of the study,
there was no centralized review of computed tomography Results
response. However, positron emission tomographic images of the
responders were all centrally reviewed by a nuclear physician Patients’ disposition and characteristics
(ASC) to confirm partial or complete metabolic response based on The patients’ characteristics are listed in Table 1. From
the Deauville five-point scale. The evaluable study population for July 2013 to December 2014, a total of 33 patients with
efficacy was restricted to patients who had received at least 28 R/R HL were recruited. Their median age was 37 years
days of the study drug. (range, 19-80). Most of the patients had advanced HL
Safety was monitored for up to 1 month after treatment. (stage III/IV) and had been heavily pretreated, with a
Adverse events were summarized by means of the Medical median number of five prior regimens including autolo-
Dictionary for Regulatory Activities, and graded using the gous SCT (54%), allogeneic SCT(15%), and BV (82%).
National Cancer Institute’s Common Terminology Criteria for Of the 33 patients recruited, 27 (82%) had refractory HL
Adverse Events (NCI-CTCAE), version 3.0. Laboratory abnormal- and 22 had biopsy-confirmed relapse of HL. Among the
ities were assessed according to NCI-CTCAE version 4.0. Only six patients displaying a response, a biopsy was per-
grade 3 or 4 toxicities and grade 2 infections were to be reported. formed in five of them at relapse [8 days, 12 days, 6
All patients were included in the toxicity analysis. weeks (n=2) and 14 months prior to inclusion in the
study].
Exploratory biomarker analysis
Blood samples (5 mL) were taken at baseline prior to drug
administration and on day 1 of cycle 2 for the measurement of Table 1. Patient’s demographics and characteristics.
27 cytokines related to the immune system using bead-based Patients’ demographics and characteristics All patients (n=33)
immunoassays. JAK2 gains, amplifications, and gene rearrange-
Gender, n (%)
ments were also investigated using fluorescent in situ hybridiza-
Male 21 (63.6%)
tion with two tri-color sets of probes associating JAK2/9p24 Female 12 (36.4%)
break-apart probes with a control centromeric probe
Age in years, median (range) 37.0 (19.0-80.0)
(CEP9/9q21): the already prepared probes from Empire
genomics on the one hand, and the association of the JAK2 B/A ECOG score
probe from Kreatech with the CEP9 probe from Vysis on the 0 11 (33.3%)
other hand. The CD274/PDL1 and PDCD1LG2/PDL2 loci at 1 15 (45.5%)
9p24 were studied with home-made prepared bacterial artificial 2 5 (15.2%)
chromosome probes purchased from the Chori BACPAC
3 2 (6.1%)
Resources Center (Oakland, CA, USA). Extraction, labeling and Ann Arbor stage
hybridization were performed on paraffin-embedded tissue, as I 1 (3.0%)
previously reported.22 II 7 (21.2%)
III 3 (9.1%)
Statistical methods IV 22 (66.7%)
The sample size for this phase II study was calculated using B symptoms
an exact single-stage phase II design.23 A two-stage design with Yes 16 (48.5%)
interim analysis for activity or toxicity was not planned given No 17 (51.5%)
the very advanced stage of the patients, the relative paucity of Extranodal involvement
alternative options, and the potential toxicity of ruxolitinib that Bone 13 (39.4%)
was expected to be in the low range, based on myelofibrosis Liver 6 (18.2%)
data. The treatment was considered ineffective if the ORR was Lung 12 (36.4%)
≤15%, and effective if the ORR was ≥35%. Under the assump- Soft tissues 4 (12.1%)
tion of an alpha first-order risk error set at 5% and beta at 20% Time since initial diagnosis in months, median (range) 55.4 (8.7 – 216.1)
with a one-sided test, it was deemed necessary to include a Prior therapies
total of 28 evaluable patients with a cut-off number of eight. If Prior lines, median (range) 5 (1 – 16)
at least eight patients had a response, the hypothesis of an ORR Chemotherapy 33 (100%)
≤15% was rejected with both a target error rate and an actual Radiotherapy 18 (54.5%)
error rate of 0.05. If seven or fewer patients had a response, the Brentuximab vedotin 27 (82%)
hypothesis of an ORR ≥35% was rejected with a target error Autologous SCT 18 (54.5%)
rate of 0.2 and an actual error rate of 0.187. The ORR estimate Allogeneic SCT 5 (15.2%)
and its 90% confidence intervals (CI) were calculated for all Interval since last treatment in months, median (range) 6 (1.1 – 75.0)
patients who completed at least one cycle of the study drug. Disease status at inclusion
The Kaplan-Meier method was employed to estimate the Relapse 6
median value and its 95% CI for time to response, duration of Refractory 27 (81.8%)
response, progression-free survival and overall survival. The SCT: stem cell transplantation; ECOG: Eastern Cooperative Oncology Group.

842 haematologica | 2018; 103(5)

 Ruxolitinib in advanced relapsed/refractory HL

Patients’ exposure to the study drug
The median number of ruxolitinib cycles administered
was four (range, 1 to 12) (Table 2). Nine patients received
all six of the planned cycles of ruxolitinib and six of these
patients continued on maintenance therapy with the JAK
inhibitor. The remainder discontinued ruxolitinib therapy,
most because of progressive disease and in one case due to
adverse events.
Responses and outcomes
The patients’ disposition through the study is illustrated
in Figure 1. Among the 33 HL patients included in the trial,
one patient did not complete the first cycle of treatment
because of progressive disease and was not, therefore,
included in the efficacy analysis. At the end of the ruxoli-
tinib induction period (6 months) three of 32 patients had
a response, for an ORR of 9.4% (90% CI: 2.6-22.5%); the
response in all three was partial. At some point during
induction six of the 32 patients had a response, which
was, in all six cases a partial response, for a best ORR of
18.8% (95% CI: 7.2-36.4%). A detailed analysis of the
responders’ characteristics is provided in Table 3. Figures 2
and 3 illustrate metabolic evolution in two patients.
Interestingly, UPN 611001, who had achieved a partial
response after six cycles of treatment, eventually entered
complete remission during the follow-up, beyond the six
cycles. Achievement of complete metabolic response was
confirmed by central review. At the time of writing, two
patients (UPN 611001 and 881001) are still taking ruxoli-
tinib. Figure 4 illustrates changes in target tumor measure-
ments in individual patients. The best reduction, if any, at
any time throughout treatment is shown.
In addition, during the 6-month induction, transient sta-
ble disease was recorded in 11 patients, albeit of limited
duration. Overall, the disease control rate (including stable
disease with complete and partial responses) was 53.1%
(17/32 patients) (95% CI: 34.7-70.9%) with a median
duration of 1.9 months.
The alleviating effect on systemic symptoms, such as
pruritus, fever, and sweating, was noteworthy, starting
within the first month of drug administration and com-
monly lasting. The impact was most remarkable on the

Table 2. Treatment exposure and modifications.

Treatment exposure and modifications All patients (n=33)
Cycles given, number, median (range) 4 (1-22)
Received full induction (6 cycles), % 9 (27.3%)
Received maintenance (> 6 cycles), % 6 (18.2%)
Percentage of planned dosea
<75% 3 (9.1%)
75%-90% 4 (12.1%)
90%-110% 25 (75.8%)
110%-125% 1 (3.0%)
Dose modification
Yes 20 (60.6%)
Type of modificationb
Dose reduction 2 (10.0%)
Dose interruption 18 (90.0%)
Dose increase 3 (15.0%)
Number of days of interruption if any, 4 (1 – 28)
median (range)
a
Defined as follows: (total number of tablets taken/total expected number of tablets)

Figure 1. Patients’ disposition. *N months on maintenance therapy: 4, 6, 6, 21,

*100, taking into account protocol-defined dose reduction. bThe total sum of the per-

16, 22; PD: progressive disease.

centages for the type and the reasons of modification may be greater than 100.0% as
a patient may have had several types of modification and reasons for treatment mod-
ification.

Table 3. Characteristics of responders (best response achieved during the 6-month ruxolitinib induction).
UPN Prior treatment Extranodal involvement Response (Cheson 2007)20
N Type
611001 9 ABVD, BEACOPP, MINE, IGEV, GVD, CAELYX, GVD, RT, BV Liver PR*,1
211004 8 ABVD, RT, IVA, transplantation, MINE, GVD, BV, ASHAP Breast PR
601001 5 BEACOPP, DHAP, IGEV, transplantation-RT, BV Liver, bone, lung PR
601004 5 ABVD, DHAP, RT, RT, BV None PR
641001 1 ABVD2 None PR
881001 5 ABVD, transplantation, MINE, BV, GEMOX Lung PR1
*Patient eventually achieved a complete response during maintenance therapy. 1Patients still under treatment with ruxolitinib at the time of writing; 2Patient with morbid obesity
not eligible for standard approaches with chemo/immunotherapy. UPN:unique patient number; RT: radiotherapy; BV: brentuximab vedotin; PR: partial response.

haematologica | 2018; 103(5) 843

 E. Van Den Neste et al.
control of pruritus, which affected 35.5% of patients prior
to initiating therapy but only 6.6% after the first cycle of
ruxolitinib. Sweating, which was present in 32.2% of the
patients at inclusion, was reduced, with 20% still having
this symptom after one cycle of treatment. Fever was
abolished in 3/4 patients with this symptom at inclusion.
The median follow-up was 17.5 months. As illustrated
in Figure 5, the median progression-free survival was 3.5
months (95% CI: 1.9-4.6). The median duration of
response was 7.7 months (95% CI: 1.8-NA) for the six
patients who eventually a achieved response (data not
shown). Overall, 30 patients had progressive disease, with
97% at the initial site and/or 60% at new sites. Following
ruxolitinib discontinuation, 25 (83.3%) patients were
given further treatments, consisting of chemotherapy in
19, and immunotherapy in nine, the latter comprising rit-
uximab in four, BV in three, and nivolumab in two.
Transplantation was eventually carried out in five
patients, and was allogeneic in four cases and autologous
in the remaining one. Among the 25 patients prescribed
further therapy, the observed complete and partial
response rates were 10% and 15%, respectively. Overall,
12 patients died on account of lymphoma progression
(83.3%), toxicity of other treatments (8.3%), or other rea-
sons (8.3%). The median overall survival was 27.1 months
(95% CI: 14.4-27.1).
A monary embolism. Of the eight serious adverse events, six
Figure 2. Response after ruxolitinib. Illustrative patient (UPN 601004). (A)
Positron emission tomography (PET)-computer tomography (CT) frontal view. (B)
PET-CT sagittal view. Partial response with allievation of B symptoms and blood
inflammation was achieved 2 months after starting ruxolitinib. At month 6, the
patient had slowly progressive disease but refused to stop ruxolitinib. CRP: C-
reactive protein.
844
Safety
All patients enrolled in the study received at least one
dose of study medication and were, therefore, included in
the safety analysis. Overall, 40 adverse events were
observed in 14/33 patients (42.4%). In six patients, the
adverse events were related to the ruxolitinib therapy. In
eight of them, the events were of grade ≥3 (Table 4A).
Among the 40 adverse events recorded, 30 (75%) occurred
during induction and 18 (45%) were related to ruxolitinib.
No drug-related deaths were recorded. One adverse event
resulted in permanent drug discontinuation, while 87.5%
of the adverse events resolved without sequelae. The
characteristics and grade of the adverse events, listed by
system organ class and preferred terms, are displayed in
Table 4B. Twenty-five (62.5%) were of grade ≥3. These
were mostly anemia (n=11), all considered related to rux-
olitinib. Other main causes of grade ≥3 adverse events
included lymphopenia (n=4), infections (n=3) and miscel-
laneous causes. Of note, no cases of grade 4 neutropenia
or thrombocytopenia were observed.
Eight serious adverse events were reported in four
patients, during induction (n=5), maintenance (n=1) or
after the end of treatment (n=2). These serious adverse
events consisted of infection in three patients (device-
related sepsis, gastroenteritis, and lung infection). The
other serious adverse events were anemia, diarrhea, sub-
dural hematoma, bone pain, and pulmonary embolism.
Two serious adverse events (anemia, lung infection) were
deemed drug-related and six were considered grade ≥3:
infection (n=3), anemia, subdural hematoma, and pul-
resolved without sequelae, while the device-related sepsis
and pulmonary embolism, observed in the same patient,
persisted until the patient died due to progressive disease
and were thus not considered as the cause of death.
Among the 33 patients, one second primary malignancy
was observed (adenocarcinoma of the colon in an 80-year
old male patient).
Biomarker analysis
Using bead-based immunoassays, plasma levels of 27
cytokines related to the immune system were measured at
Figure 3. Response after ruxolitinib. Illustrative patient (UPN 601001).
Comparison of frontal positron emission tomography (PET)-scan prior to inclu-
sion and after 2 months of ruxolitinib. There was a rapid improvement of con-
stitutional symptoms after a few days on ruxolitinib. PET after 2 months showed
metabolic partial response with a total volume reduction of tumor lung lesions
of 64%.
haematologica | 2018; 103(5)

baseline and after the first cycle of treatment. At baseline,
there was no difference in cytokine levels between
responders and non-responders. In responders, the only
cytokine that decreased significantly was CX-CL10
(P=0.01). In patients presenting with pruritus (n=11), the
levels of platelet-derived growth factor-BB (PDGF-BB)
(Online Supplementary Appendix), interleukin (IL)-5, IL-10,
IL-12, IL-13, IL-17, eotaxin, fibroblast growth factor basic
(FGF basic), macrophage inflammatory protein 1b
(MIP1b), regulated on activation, normal T-cell expressed
and secreted (RANTES), and vascular endothelial growth
factor (VEGF) were significantly increased. In the latter
patients, ruxolitinib treatment significantly decreased the
levels of PDGF-BB, IL-10, IL-12, IL-13, IL-17, FGF basic
and VEGF. Among the patients who could be analyzed for
JAK2 amplification in HRS cells (n=12), polysomy (sug-
gesting hyperdiploidy) was detected in all of them, and
specific JAK2 amplification in only one. This latter patient
haematologica | 2018; 103(5)
Ruxolitinib in advanced relapsed/refractory HL
achieved a partial response as determined by computed
tomography criteria and also a positron emission tomog-
raphy-determined response lasting 4 months. It is note-
worthy that the PDL1 and PDL2 loci (which are in the
vicinity of the JAK2 locus at 9p24), analyzed by fluores-
cent in situ hybridization with bacterial artificial chromo-
some probes, showed the same pattern of gains as for the
JAK2 locus.
Discussion
JAK/STAT activation, driven by an aberrant network of
cytokines and chemokines in the HL microenvironment,
is critical for the proliferation and survival of neoplastic
HRS cells.24,25 The JAK/STAT pathway also plays a role in
immune evasion by HL cells via the secretion of
chemokines leading to Th2 homing or via the regulation
Figure 4. Waterfall plot demonstrating percent
change from baseline in target tumor dimensions
(best response, n=27). Of note, among the 32
patients evaluable for disease response, five had no
end-of-treatment SPD measurements by computer
tomography (CT) scan as planned by protocol
because there were obvious signs of disease progres-
sion. *Persisting positive positron emission tomogra-
phy scan, considered as partial response.
Figure 5. Kaplan-Meier estimate of progression-free
survival in 32 evaluable patients with Hodgkin lym-
phoma receiving ruxolitinib.
845
 E. Van Den Neste et al.

of PD-L1/L2 expression, which confers an immune privi- Table 4. Treatment-emergent adverse events.
lege to HRS cells. Chromosome 9p24.1/PD-L1/PD-L2 A. Patients with an adverse event.
alterations increase the abundance of the PD-1 ligands, Treatment-emergent adverse events1 All patients (n=33)
PD-L1 and PD-L2, and their further induction through
JAK/STAT signaling.26-28 This complex crosstalk between Patients with > 1 AE 14 (42.4%)
malignant HRS cells and the reactive microenvironment N. of AE/patient, median (range) 2 (1-11)
N. of patients with AE > grade 3 8 (24.2%)
could be targeted to overcome chemoresistance. Based on
this rationale, we explored JAK 1/2 inhibition in a phase Patients with AE related to ruxolitinib 6 (18.2%)
II study of fixed dose ruxolitinib in patients with Patients with AE leading to drug discontinuation 1 (3%)
advanced HL patients before the onset of the era of PD-1 Patients with AE leading to death 0 (0%)
blockers. With an ORR of 9.4% at the end of the 6-month AE: adverse event. 1Total number of AE, 40.
induction period, this study did not reach its primary effi-
cacy goal. Nevertheless, when including transient B. Characteristics of the adverse event (N = 40) by system organ class
responses seen before the 6-month evaluation, the ORR and preferred terms
was 18.8% in some heavily pretreated patients, most of
whom were refractory and had failed treatment with BV.
Adverse event Any grade Grade 2 Grades > 3
These responses were sometimes durable (median=7.7 Any adverse event 40 15 25
months). Some other patients had disease control, but Infections and infestations 13 10 31
with uncertain clinical benefit. A notable finding to be Anemia 11 0 11
highlighted was the relief of B symptoms and pruritus,
which was quick and long-lasting, resulting in a number Lymphopenia 4 0 4
of patients being reluctant to discontinue the compound, Thrombocytopenia 2 0 2
despite progressive disease. The latter effect should not Weight decrease 1 0 1
be interpreted as a proven surrogate of anti-lymphoma Respiratory and thoracic disorders 3 2 1
activity.
Diarrhea 1 1 0
These results tend to lend some support to the concept
of JAK1/2 inhibition as a potential therapeutic means in Infuenza-like illness 1 1 0
HL. There are presently only scarce data available on the Subdural hematoma 1 0 1
use of ruxolitinib in HL. In a preliminary report of an Bone pain 1 1 0
ongoing study, Kim et al. described rapid achievement of Epilepsy 2 0 2
disease control (1 complete response, 5 partial responses, 1
Implantable device infection, gastro-enteritis, lung infection.
1 stable disease) in 13 patients with advanced HL treated
with ruxolitinib at a dose of 20 mg bid.29 Younes et al.
reported changes in tumor measurements in HL patients
treated in a phase I study with SB1518, an inhibitor of contains the JAK2 gene, are more frequent in advanced
JAK2 and FLT-3.30 In vitro, AZD1480, an inhibitor of JAK1 disease.28 Surprisingly, in our patients, a low incidence of
and JAK2, could regulate proliferation in HL cell lines.27 JAK2 amplification was seen, suggesting a low proportion
The multikinase inhibitor lestaurtinib also inhibited of patients harboring the target of ruxolitinib, although
growth and increased apoptosis of HL cell lines and HL this inference should be considered with caution since
cells from lymph nodes.31 Finally, a clinical grade JAK2 not all patients could be analyzed.
inhibitor, fedratinib, inhibited the proliferation of classi- With respect to safety, ruxolitinib was by and large
cal HL cell lines in a JAK2 copy number-dependent man- well-tolerated, with no drug-related mortality reported.
ner implying decreased phosphorylation of STAT and The most prominent toxicities included drug-related ane-
expression of downstream targets including PD-L1 show- mia and manageable infectious events with no specific
ing immunomodulation by JAK inhibitors.32 pattern. The relative lack of hematologic toxicity suggests
If JAK2 is actually an appropriate target, questions arise that it could be feasible to combine ruxolitinib treatment
as to why the study outcome was not more convincing. with genotoxic compounds. For patients who discontin-
Could the drug's limited activity be attributed to insuffi- ued ruxolitinib therapy, a switch to chemotherapy and/or
cient dosage? Given that we observed unambiguous immunotherapy was feasible, suggesting that the com-
cytokine profile changes and frequent improvements in B pound does not jeopardize further treatment.
symptoms, it would seem that the dosage of 20 mg twice The question now remains as to how this compound
daily, a dosage at which target inhibition occurs in can best be utilized in the future. The exploratory nature
myelofibrosis,33 was appropriate. Another factor possibly of our study did not allow identification of the best can-
influencing the outcome was our patients’ disease stage, didates on the basis of clinical stage or biomarkers. The
represented by a high percentage of refractoriness. At this cytokine profile showed some changes in patients with
late stage, the genetic changes would be so complex that pruritus, but these changes were not correlated with clin-
selective inhibition of JAK is insufficient in cells depend- ical response. Although JAK2 status was explored in a
ent on other signaling pathways to promote their sur- minority of patients, the only patient with JAK2 amplifi-
vival, thus further curbing the study's potential. It is cation achieved a response. It will be important to focus
known that genomic aberrations, such as chromosome on biomarker results in ongoing studies of JAK inhibition
breakpoints, are more numerous in later clinical stages of in HL. Given ruxolitinib’s limited benefits as monothera-
HL.34 Mechanisms of resistance to JAK/STAT inhibition py, use in combination with other drugs may possibly
have been reported such as a feedback loop of paradoxi- enhance its therapeutic potential. Ruxolitinib, which has
cally activated extracellular signal-regulated kinases 1 and no overlapping toxicity with chemotherapy, has been
2 (ERK1/2).27 Aberrations of the 9p24.1 amplicon, which combined with hypomethylating agents, lenalidomide,

846 haematologica | 2018; 103(5)

 Ruxolitinib in advanced relapsed/refractory HL

and even intensive chemotherapy.35-38 In vitro data have
shown that ruxolitinib could restore the sensitivity of cis-
platin-resistant cell lines with higher Jak2 expression.39
Interestingly, the combination of BV with ruxolitinib
resulted in additive and synergistic killing in a xenograft
mouse model of HL through a mechanism involving
mitochondrial control of apoptosis.40 Another means to
boost ruxolitinib’s potential would be to combine it with
agents blocking other signaling pathways. Interestingly,
the combination of ruxolitinib with a Bcl2/Bcl-xL
inhibitor displayed dramatic synergy in an adult T-cell
leukemia cell line via a mechanism implying BAX activa-
tion.41 Finally, the effect of combining chemical JAK
blockade and an anti-PD1/L1 strategy should be analyzed
in HL, keeping in mind, however, that a potential antag-
onism may be encountered due to these two drugs acting
on the same target, given that PD1-L1 expression is
dependent on JAK2 activity.
In conclusion, based on a strong biological rationale for
clinical evaluation of JAK2 blockade in HL, we initiated a
phase II study of ruxolitinib in R/R HL patients. The
study failed to fulfill the efficacy criteria for further devel-
opment of the drug as monotherapy. Nonetheless, in
patients with very advanced disease ruxolitinib showed
hints of activity that surpassed solely an anti-inflammato-
ry effect. This may suggest that further improvements
will come from a more complete inhibition of signaling
pathways involved in HRS cell survival or from combina-
tion with chemotherapy, such as BV.
Acknowledgments
The authors would like to thank the patients, their families and
their caregivers who made this study possible. We also thank all
the study investigators and study staff at each of the clinical sites.
For the LYSARC, we acknowledge the project manager and all
members of the data monitoring committee. We express our
thanks to Loïc Chartier and Sami Boussetta, biostatisticians at
the LYSARC, who contributed to the statistical design and analy-
sis of the study. Editorial assistance was provided by Cremer
Consulting.

References Blood. 2009;114(14):2945-2951. ing PRDM16 in myeloid and lymphoid

1. Meyer RM, Hoppe RT. Point/counterpoint:
early-stage Hodgkin lymphoma and the role
of radiation therapy. Hematology Am Soc
Hematol Educ Program. 2012;2012:313-321.
2. Horning SJ. Primary refractory Hodgkin's
disease. Ann Oncol. 1998; 9 (Suppl 5): S97-
101.
3. Kuruvilla J, Keating A, Crump M. How I
treat relapsed and refractory Hodgkin lym-
phoma. Blood. 2011;117(16):4208-4217.
4. Arai S, Fanale M, DeVos S, et al. Defining a
Hodgkin lymphoma population for novel
therapeutics after relapse from autologous
hematopoietic cell transplant. Leuk
Lymphoma. 2013;54(11):2531-2533.
5. Ansell SM. Hodgkin lymphoma: MOPP
chemotherapy to PD-1 blockade and
beyond. Am J Hematol. 2016;91(1):109-112.
6. Armand P, Shipp MA, Ribrag V, et al.
Programmed death-1 blockade with pem-
brolizumab in patients with classical
hodgkin lymphoma after brentuximab
vedotin failure. J Clin Oncol. 2016 Jun 27.
[Epub ahead of print].
7. Younes A, Bartlett NL, Leonard JP, et al.
Brentuximab vedotin (SGN-35) for relapsed
CD30-positive lymphomas. N Engl J Med.
2010;363(19):1812-1821.
8. Moskowitz C. Novel agents and strategies
in transplant-eligible patients with relapsed
and refractory Hodgkin lymphoma.
Hematology Am Soc Hematol Educ
Program. 2016;2016(1):331-338.
9. Slovak ML, Bedell V, Hsu YH, et al.
Molecular karyotypes of Hodgkin and Reed-
Sternberg cells at disease onset reveal dis-
tinct copy number alterations in chemosen-
sitive versus refractory Hodgkin lymphoma.
Clin Cancer Res. 2011;17(10):3443-3454.
10. Aldinucci D, Gloghini A, Pinto A, De Filippi
R, Carbone A. The classical Hodgkin's lym-
phoma microenvironment and its role in
promoting tumour growth and immune
escape. J Pathol. 2010;221(3):248-263.
11. Navarro A, Diaz T, Martinez A, et al.
Regulation of JAK2 by miR-135a: prognostic
impact in classic Hodgkin lymphoma.
12. Van Roosbroeck K, Cox L, Tousseyn T, et al.
JAK2 rearrangements, including the novel
SEC31A-JAK2 fusion, are recurrent in classi-
cal Hodgkin lymphoma. Blood. 2011;117
(15):4056-4064.
13. Meier C, Hoeller S, Bourgau C, et al.
Recurrent numerical aberrations of JAK2
and deregulation of the JAK2-STAT cascade
in lymphomas. Mod Pathol. 2009;22(3):476-
487.
14. Hartmann S, Martin-Subero JI, Gesk S, et al.
Detection of genomic imbalances in
microdissected Hodgkin and Reed-
Sternberg cells of classical Hodgkin's lym-
phoma by array-based comparative genom-
ic hybridization. Haematologica. 2008;93(9):
1318-1326.
15. Joos S, Granzow M, Holtgreve-Grez H, et al.
Hodgkin's lymphoma cell lines are charac-
terized by frequent aberrations on chromo-
somes 2p and 9p including REL and JAK2.
Int J Cancer. 2003;103(4):489-495.
16. Aldinucci D, Celegato M, Casagrande N.
Microenvironmental interactions in classical
Hodgkin lymphoma and their role in pro-
moting tumor growth, immune escape and
drug resistance. Cancer Lett. 2016;380(1):
243-252.
17. Assi R, Verstovsek S, Daver N. 'JAK-ing' up
the treatment of primary myelofibrosis:
building better combination strategies. Curr
Opin Hematol. 2017;24(2):115-124.
18. Vannucchi AM, Harrison CN. Emerging
treatments for classical myeloproliferative
neoplasms. Blood. 2017;129(6):693-703.
19. Massaro F, Molica M, Breccia M. How rux-
olitinib modified the outcome in myelofi-
brosis: focus on overall survival, allele bur-
den reduction and fibrosis changes. Expert
Rev Hematol. 2017;10(2):155-159.
20. Cheson BD, Pfistner B, Juweid ME, et al.
Revised response criteria for malignant lym-
phoma. J Clin Oncol. 2007;25(5):579-586.
21. Heine A, Brossart P, Wolf D. Ruxolitinib is a
potent immunosuppressive compound: is it
time for anti-infective prophylaxis? Blood.
2013;122(23):3843-3844.
22. Duhoux FP, Ameye G, Lambot V, et al.
Refinement of 1p36 alterations not involv-
malignancies. PLoS One. 2011;6(10):e26311.
23. A'Hern RP. Sample size tables for exact sin-
gle-stage phase II designs. Stat Med.
2001;20(6):859-866.
24. Aldinucci D, Pinto A, Gloghini A, Carbone
A. Chemokine receptors as therapeutic tools
in Hodgkin lymphoma: CCR4 and beyond.
Blood. 2010;115(3):746-747; author reply
748.
25. Carbone A, Gloghini A, Castagna L, Santoro
A, Carlo-Stella C. Primary refractory and
early-relapsed Hodgkin's lymphoma: strate-
gies for therapeutic targeting based on the
tumour microenvironment. J Pathol.
2015;237(1):4-13.
26. Green MR, Monti S, Rodig SJ, et al.
Integrative analysis reveals selective 9p24.1
amplification, increased PD-1 ligand expres-
sion, and further induction via JAK2 in
nodular sclerosing Hodgkin lymphoma and
primary mediastinal large B-cell lymphoma.
Blood. 2010;116(17):3268-3277.
27. Derenzini E, Lemoine M, Buglio D, et al.
The JAK inhibitor AZD1480 regulates prolif-
eration and immunity in Hodgkin lym-
phoma. Blood Cancer J. 2011;1(12):e46.
28. Roemer MG, Advani RH, Ligon AH, et al.
PD-L1 and PD-L2 genetic alterations define
classical Hodgkin lymphoma and predict
outcome. J Clin Oncol. 2016;34(23):2690-
2697.
29. Kim SJ, Kang HJ, Dong-Yeop S, et al. The
efficacy of JAK2 inhibitor in heavily pre-
treated classical Hodgkin lymphoma: a
prospective pilot study of ruxolitinib in
relapsed or refractory classical Hodgkin
lymphoma and primary mediastinal large
B-cell lymphoma. Blood. 2016;128(22):
1820.
30. Younes A, Romaguera J, Fanale M, et al.
Phase I study of a novel oral Janus kinase 2
inhibitor, SB1518, in patients with relapsed
lymphoma: evidence of clinical and biologic
activity in multiple lymphoma subtypes. J
Clin Oncol. 2012;30(33):4161-4167.
31. Diaz T, Navarro A, Ferrer G, et al.
Lestaurtinib inhibition of the Jak/STAT sig-
naling pathway in Hodgkin lymphoma
inhibits proliferation and induces apoptosis.

haematologica | 2018; 103(5) 847

 E. Van Den Neste et al.
PLoS One. 2011;6(4):e18856.
32. Hao Y, Chapuy B, Monti S, Sun HH, Rodig
SJ, Shipp MA. Selective JAK2 inhibition
specifically decreases Hodgkin lymphoma
and mediastinal large B-cell lymphoma
growth in vitro and in vivo. Clin Cancer Res.
2014;20(10):2674-2683.
33. Cervantes F, Vannucchi AM, Kiladjian JJ, et
al. Three-year efficacy, safety, and survival
findings from COMFORT-II, a phase 3
study comparing ruxolitinib with best avail-
able therapy for myelofibrosis. Blood.
2013;122(25):4047-4053.
34. Falzetti D, Crescenzi B, Matteuci C, et al.
Genomic instability and recurrent break-
points are main cytogenetic findings in
Hodgkin's disease. Haematologica. 1999;84
(4):298-305.
35. Naqvi K, Daver N, Pemmaraju N, et al.
848
Clinical use of ruxolitinib in an academic
medical center in unselected patients with
myeloproliferative neoplasms not on clini-
cal study. Leuk Lymphoma. 2017;58(4):
866-871.
36. Daver N, Cortes J, Newberry K, et al.
Ruxolitinib in combination with lenalido-
mide as therapy for patients with myelofi-
brosis. Haematologica. 2015;100(8):1058-
1063.
37. Devillier R, Raffoux E, Rey J, et al.
Combination therapy with ruxolitinib plus
intensive treatment strategy is feasible in
patients with blast-phase myeloproliferative
neoplasms. Br J Haematol. 2016;172(4):628-
630.
38. Mayfield JR, Czuchlewski DR, Gale JM, et
al. Integration of ruxolitinib into dose-inten-
sified therapy targeted against a novel JAK2
F694L mutation in B-precursor acute lym-
phoblastic leukemia. Pediatr Blood Cancer.
2017;64(5).
39. Hu Y, Hong Y, Xu Y, Liu P, Guo DH, Chen Y.
Inhibition of the JAK/STAT pathway with
ruxolitinib overcomes cisplatin resistance in
non-small-cell lung cancer NSCLC.
Apoptosis. 2014;19(11):1627-1636.
40. Ju W, Zhang M, Wilson KM, et al.
Augmented efficacy of brentuximab vedotin
combined with ruxolitinib and/or
Navitoclax in a murine model of human
Hodgkin's lymphoma. Proc Natl Acad Sci
USA. 2016;113(6):1624-1629.
41. Zhang M, Mathews Griner LA, Ju W, et al.
Selective targeting of JAK/STAT signaling is
potentiated by Bcl-xL blockade in IL-2-
dependent adult T-cell leukemia. Proc Natl
Acad Sci USA. 2015;112(40):12480-12485.
haematologica | 2018; 103(5)
 Non-Hodgkin Lymphoma ARTICLE

A B-cell receptor-related gene signature

predicts survival in mantle cell lymphoma: Ferrata Storti
Foundation
results from the Fondazione Italiana Linfomi
MCL-0208 trial
Riccardo Bomben,1* Simone Ferrero,2,3* Tiziana D'Agaro,1 Michele Dal Bo,1
Alessandro Re,4 Andrea Evangelista,5 Angelo Michele Carella,6 Alberto Zamò,7
Umberto Vitolo,8 Paola Omedè,3 Chiara Rusconi,9 Luca Arcaini,10
Luigi Rigacci,11 Stefano Luminari,12,13 Andrea Piccin,14 Delong Liu,15
Adrian Wiestner,15 Gianluca Gaidano,16 Sergio Cortelazzo,17 Marco Ladetto2,18
and Valter Gattei1**
Haematologica 2018
Volume 103(5):849-856
1
Clinical and Experimental Onco-Hematology Unit, Centro di Riferimento Oncologico,
IRCCS, Aviano (PN), Italy; 2Department of Molecular Biotechnologies and Health
Sciences, Hematology Division 1, University of Torino, Italy; 3Hematology Division 1, AOU
“Città della Salute e della Scienza di Torino” University-Hospital, Italy; 4Hematology, AO
“Spedali Civili di Brescia”, Italy; 5Unit of Cancer Epidemiology, AOU “Città della Salute e
della Scienza di Torino” University-Hospital, Italy; 6Hematology Division 1, IRCCS AOU
San Martino IST, Genova, Italy; 7Department of Diagnostics and Public Health, University
of Verona, Italy; 8Hematology Division 1, AOU “Città della Salute e della Scienza di
Torino” University-Hospital, Italy; 9Hematology Division, “Niguarda Ca’ Granda” Hospital,
Milano, Italy; 10Hematology Division, Department of Molecular Medicine, IRCCS
Fondazione San Matteo, Pavia, Italy; 11Hematology Division, AOU “Careggi”, University of
Firenze, Italy; 12Hematology, Azienda Sanitaria Locale IRCCS, Reggio Emilia, Italy;
13
Department of Diagnostic, Clinical and Public Health Medicine, University of Modena
and Reggio Emilia, Reggio Emilia, Italy; 14Department of Hematology and BMT Unit,
Bolzano/Bozen Regional Hospital, Italy; 15Hematology Branch, National Heart, Lung, and
Blood Institute, NIH, Bethesda, MD, USA; 16Division of Haematology, Department of
Translational Medicine –University of Eastern Piedmont, Novara, Italy; 17Hematology,
Medical Oncology and Hematology Division, “Istituto Clinico Humanitas Gavazzeni”,
Bergamo, Italy and 18SC Ematologia Azienda Ospedaliera Nazionale SS. Antonio e Biagio
e Cesare Arrigo, Alessandria, Italy
*RB and SF contributed equally to this work as first authors: **ML and VG contributed equally to
this work as senior authors

Correspondence:
ABSTRACT rbomben@cro.it

M
antle cell lymphoma patients have variable clinical courses,
ranging from indolent cases that do not require immediate treat-
ment to aggressive, rapidly progressing diseases. Thus, diagnos- Received: November 10, 2017.
tic tools capable of stratifying patients according to their risk of relapse Accepted: February 14, 2018.
and death are needed. This study included 83 samples from the
Fondazione Italiana Linfomi MCL-0208 clinical trial. Through gene Pre-published: February 22, 2018.
expression profiling and quantitative real-time PCR we analyzed 46
peripheral blood and 43 formalin-fixed paraffin-embedded lymph node doi:10.3324/haematol.2017.184325
samples. A prediction model to classify patients was developed. By ana-
lyzing the transcriptome of 27 peripheral blood samples, two subgroups Check the online version for the most updated
characterized by a differential expression of genes from the B-cell recep- information on this article, online supplements,
tor pathway (B-cell receptorlow and B-cell receptorhigh) were identified. and information on authorship & disclosures:
The prediction model based on the quantitative real-time PCR values of www.haematologica.org/content/103/5/849
six representative genes (AKT3, BCL2, BTK, CD79B, PIK3CD, and SYK),
was used to classify the 83 cases (43 B-cell receptorlow and 40 B-cell
receptorhigh). The B-cell receptorhigh signature associated with shorter pro- ©2018 Ferrata Storti Foundation
gression-free survival (P=0.0074), selected the mantle cell lymphoma Material published in Haematologica is covered by copyright.
subgroup with the shortest progression-free survival and overall survival All rights are reserved to the Ferrata Storti Foundation. Use of
(P=0.0014 and P=0.029, respectively) in combination with high (>30%) published material is allowed under the following terms and
Ki-67 staining, and was an independent predictor of short progression- conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
free survival along with the Mantle Cell Lymphoma International Copies of published material are allowed for personal or inter-
Prognostic Index-combined score. Moreover, the clinical impact of the 6- nal use. Sharing published material for non-commercial pur-
gene signature related to the B-cell receptor pathway identified a mantle poses is subject to the following conditions:
cell lymphoma subset with shorter progression-free survival intervals https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
also in an external independent mantle cell lymphoma cohort homoge- mercial purposes is not allowed without permission in writing
nously treated with different schedules. In conclusion, this 6-gene signa- from the publisher.
ture associates with a poor clinical response in the context of the MCL-
0208 clinical trial. (clinicaltrials.gov identifier: 02354313).

Haematologica | 2018; 103(5) 849

 R. Bomben et al.
Introduction
Mantle cell lymphoma (MCL) is a distinctive B-cell
malignancy accounting for 5-10% of all lymphomas,1-3
whose molecular hallmark and initiating oncogenic event,
the t(11;14)(q13;q32) translocation, leads to constitutive
overexpression of the proto-oncogene cyclin D1
(CCND1).2,4
Once considered as uniformly characterized by a poor
prognosis, MCL has been demonstrated to have unexpect-
edly variable clinical courses, ranging from indolent cases
that do not require immediate treatment to aggressive,
rapidly progressing disease.2,5-10 Even among patients
requiring treatment, prognosis is highly heterogeneous,
with patients experiencing prolonged remissions and oth-
ers rapidly relapsing even after cytarabine-containing
induction regimens followed by autologous transplanta-
tion. Thus, diagnostic tools capable of stratifying MCL
patients in different risk classes are warranted in order to
direct treatment strategies.11 For this reason, many
attempts have been made to identify clinical, histological,
and molecular markers that can stratify patients according
to their risk of relapse and death.12-25
In addition to the clinical MCL prognostic score (MCL-
International Prognostic Index, MIPI)12,14 capable of strati-
fying patients into risk groups with different overall sur-
vival (OS),14 the Ki-67 proliferation index has been pro-
posed as one of the most powerful and independent pre-
dictors of survival in MCL even in the context of prospec-
tive trials and modern therapies,5,13,26 and for these reasons
has been integrated into the so-called MIPI-combined
(MIPI-c) score.13,26 Moreover, effective prognostic discrimi-
nation is achieved by post-treatment response monitoring
by positron emission tomography (PET)-scan and minimal
residual disease (MRD). Furthermore, a seminal study
identified a specific signature associated with proliferation
Table 1. Characteristics of 83 mantle cell lymphoma cases entering
the study.
Number of cases
Median age, years (range)
Ratio male/female (%)
Abnormal LDH (%)
Median WBC (x109/L)
Typical morphology
Blastoid morphology
Median Proliferation Index (Ki-67 staining), %
MIPI-c class
Low
Low/intermediate
High/intermediate
High
na
Median survival, months (range) 34.7 (1.4-73.4)
Median progression-free survival, months (range) 31.3 (1.4-73.4)
LDH: lactate dehydrogenase; WBC: white blood count; MIPI-c: MCL:-International
Prognostic Index-combined; na: not available.
850
as the strongest predictor of OS in a large MCL series.20 In
this context, a cohort of 20 proliferation-associated genes
constructed on the basis of gene expression analysis was
demonstrated to be superior to other molecular markers.20
Since approaches based on microarray technology have
not yet been incorporated into routine clinical practice, a
PCR-based surrogate method investigating expression of
five genes has been proposed and applied to paraffin-
embedded tissues.18
Recent evidence suggests that the B-cell receptor (BCR)
pathway may contribute to the pathogenesis of several
histological types of B-cell non-Hodgkin lymphomas,
including MCL.27-30 The importance of BCR signaling path-
way in B-cell malignancy pathogenesis has driven interest
in the use of small-molecule inhibitors of BCR-associated
kinases, potentially preventing the activation of one or
more of the distal BCR signaling pathway proteins.28,31
In the present study, we developed a survival predictive
model for younger patients with advanced MCL treated in
the context of the Fondazione Italiana Linfomi (FIL) MCL-
0208 Phase III randomized clinical trial. This model is
based upon the quantitative evaluation of six genes, most-
ly from the BCR pathway, selected from a gene expression
profile (GEP) of peripheral blood (PB) MCL cells and was
applied to formalin-fixed paraffin-embedded (FFPE) tissue
specimens. Notably, the model predicts poor response in
the context of the FIL-MCL-0208 trial.
Methods
Primary MCL cases
The study included 83 out of 300 samples of adult patients
under 66 years of age with advanced stage MCL, enrolled in the
FIL-MCL-0208 prospective, multicenter, Phase III randomized
clinical trial (clinicaltrials.gov identifier: 02354313),32 divided as fol-
lows: i) a panel of 27 PB samples utilized for GEP upon positive
sorting of the clonal CD5+/CD19+ MCL cells; ii) an additional
panel of 19 PB samples utilized for quantitative real-time PCR
(qRT-PCR) of the identified gene signature in the purified MCL
cell component; iii) a panel of 43 lymph node (LN) samples uti-
lized for qRT-PCR of the identified gene signature (in this LN
Samples
panel 6 samples had a matched PB sample). The clinical and histo-
83 pathological details of the 83 MCL cases used in this study are
56 (28-65) reported in Table 1. No significant differences were found
between the 83 cases entering the study versus the 217 remaining
57/26 (68)
cases enrolled in the clinical trial in terms of median age, MIPI
45/29 (39) score, Ki-67 index and PFS intervals (Online Supplementary Table S1
13.7 and Online Supplementary Figure S1). No differences in clinical and
74 biological parameters were observed between PB and LN MCL
5 samples (data not shown). All patients were treated according to the
FIL-MCL-0208 clinical trial, as reported in Online Supplementary
20.0 (0-99)
Figure S2.
Mantle cell lymphoma diagnosis was prospectively confirmed
38 (46%) by centralized histological review according to the World Health
23 (28%) Organization (WHO) 2008 criteria.3,33 All patients provided
11 (13%) informed consent in accordance with Institutional Review Board
requirements (0016331-BZ 09/02/2010) and the Declaration of
6 (7%)
Helsinki, and protocol consent included use of MRD sample left-
5 (6%) overs for the study.
All the procedures employed for RNA extraction, GEP and
downstream analyses, qRT-PCR, analyses and qRT-PCR valida-
tions were carried out according to standard protocols, as report-
ed previously.34-37 (See the Online Supplementary Appendix for
details). Microarray data are available in Gene Expression
haematologica | 2018; 103(5)

Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under accession num-
ber GSE89447. Cases used for these procedures are reported in
Online Supplementary Table S2.
Validation procedures
The 6-gene signature was tested in the MCL cohort described
by Saba et al.,30 enrolled in another clinical trial (clinicaltrials.gov
identifier: 00114738), by using the sum of the array gene expression
values, as reported.30 Gene signatures related to MCL outcome
were retrieved from previous papers,30,38 and imported in the
GeneSpring GX and tested in the present cohort with GEP data
available.
Statistical analysis
Overall survival was computed from trial registration to death
as a result of any cause, censored at the latest follow up in patients
who were still alive. Progression-free survival (PFS) was computed
from trial registration to progression or death as a result of any
cause, censored at the latest tumor assessment if no progression
was observed. Clinical correlations, performed with the MedCalc
v.9.5 software, were made using Kaplan-Meier plots and log-rank
test. The Cox proportional model was chosen for multivariable
analysis. Clinical outcome results were up-dated as of January
A B
Figure 1. Gene Expression Profile (GEP) analysis of 27 mantle cell lym-
phoma samples. (A) Principal Component Analysis (PCA) scores represent-
ed in a 3D scatter plot. One point per array/sample is shown. Black line indi-
cates separation between PCA classes. (B) Hierarchical clustering of 14
group-1 cases and 13 group-2 cases, using 50,739 probes. (C) Hierarchical
clustering of 14 group1 cases and 13 group2 cases, using the 922 differ-
entially expressed probes. Color codes for gene expression values refer to
mean centered log-ratio values.
haematologica | 2018; 103(5)
B-cell receptor signature in mantle cell lymphoma
2017.32 Investigators are still blinded to the investigation arm as
the primary study end point has still not been met.
Results
GEP identifies MCL patients with distinct expression of
genes belonging to the BCR pathway
Global GEP was performed in purified MCL cells from
27 PB samples. An unsupervised analysis performed by
principal component analysis (PCA) divided the cohort
into two groups of 14 cases and 13 cases, respectively
(Figure 1A). Consistently, a hierarchical clustering, which
was run with all the GEP features, split MCL cases into
two major groups perfectly resembling the PCA groups
(Figure 1B).
Supervised analysis according to the PCA classification
defined a gene expression signature composed of 922
probes, 713 up-regulated and 209 down-regulated in
group-2 versus group-1 samples (Figure 1C and Online
Supplementary Table S3).
Pathway analysis revealed that “Antigen processing and
presentation” and “B-cell receptor signaling pathway”
were among the top ranked pathways enriched in the
851
 R. Bomben et al.

group-2 category (Online Supplementary Table S4). Similar
results were obtained by GSEA which highlighted a con-
stitutive overexpression of genes related to the BCR sig-
naling pathways in the context of group-2 patients (Figure
2A and Online Supplementary Table S5). Therefore, here-
after the two PCA groups were identified as BCRlow
(group-1) and BCRhigh (group-2).
A 6-gene signature identifies BCRlow and BCRhigh MCL
samples
Having identified two different groups of MCL patients
at diagnosis with a different expression of genes related
to the BCR pathway, we overlapped the genes included in
the gene sets related to the BCR pathway (115 probes)
and the differentially expressed genes (922 probes) to cre-

A B

D E

Figure 2. 6-gene signature and Decision Tree (DT) prediction model. (A) Gene Expression Profile data of BCRlow and BCRhigh MCL samples were tested using Gene set
enrichment analysis (GSEA). Reported are the significant gene sets differentially expressed and related to the B-Cell Receptor (BCR) pathway. (B) Venn diagram
derived by merging the differentially expressed probes and the genes belonging to the BCR related gene sets. In bold genes selected as the 6-gene signature. (C)
Hierarchical clustering of 14 BCRlow cases and 13 BCRhigh cases, using the six gene values. Color codes for gene expression values refer to mean centered log-ratio
values. (D) Hierarchical clustering of 8 BCRlow cases and 9 BCRhigh cases belonging to the training set of DT prediction model, using the six gene qRT-PCR values. (E)
Hierarchical clustering of 6 BCRlow cases and 4 BCRhigh cases belonging to the validation set of DT prediction model, using the six gene qRT-PCR values. Bar under the
heat-map refers to prediction generated by the DT prediction model. Color codes for gene expression values refer to mean centered log-ratio values.

852 haematologica | 2018; 103(5)

 B-cell receptor signature in mantle cell lymphoma

ate a reduced signature (Figure 2B). In this way, 18 probes
corresponding to 15 genes, all over-expressed in BCRhigh
cases were identified (Figure 2B). Among these genes, a
subgroup of six genes (AKT3, BCL2, BTK, CD79B,
PIK3CD, and SYK) was selected for further validations
due to their direct involvement in the BCR pathway
and/or the existence of drugs targeting the related pro-
teins. A hierarchical cluster using only these six genes was
able to discriminate patients belonging to the BCRlow or
BCRhigh groups (Figure 2C).
Development of a qRT-PCR-based predictor for BCRlow
and BCRhigh in MCL samples
By analyzing the expression levels of the selected six
genes in the same 27 MCL PB samples by qRT-PCR
approach, a strict correlation with GEP data was found
(Online Supplementary Figure S3B). Moreover, the 27 MCL
cases were randomly divided into a training set (17 cases;
8 BCRlow and 9 BCRhigh samples) and a validation set (10
cases; 6 BCRlow and 4 BCRhigh samples) to develop and test
a decision tree (DT) model based on qRT-PCR data capa-
ble of categorizing patients into one of the two categories.
The DT model based on qRT-PCR data correctly classified
16 of 17 cases belonging to the training set and 10 of 10
cases of the validation cohort, and allowed the classifica-
tion of 19 additional PB samples screened with qRT-PCR
(9 BCRlow and 10 BCRhigh (Figure 2D and E and Online
Supplementary Table S2).
(median PFS: 21.6 months vs. not reached; P=0.0375)
(Figure 3).
Application of the 6-gene signature to LN samples
from MCL patients
To evaluate the capability of the 6-gene signature to
identify different subgroups also in the context of MCL
LN cases, we tested our qRT-PCR approach in a series of
43 LN samples preserved as FFPE LN specimens. Thirty-
five (81%) out of 43 samples were amplifiable for all six
genes, and using a DT model based on qRT-PCR values
from FFPE, 23 cases were classified as BCRlow and 20 clas-
sified as BCRhigh (Online Supplementary Table S2). Notably,
for 6 out of 43 LN samples, a PB matched sample was
available, and by comparing qRT-PCR results performed
on PB samples and LN FFPE samples from these cases, a
good concordance was overall observed, although FFPE
samples generally amplified at higher Ct values (Online
Supplementary Figure S5). Of note, 5 out of 6 these MCL
cases were consistently classified. The misclassified case

Association between BCR categories and biological

and clinical parameters
Collectively, the 6-gene signature was re-evaluated by
setting up a validated qRT-PCR approach (see Online
Supplementary Appendix and Online Supplementary Table S6)
in PB samples from 46 MCL cases, 23 were identified as
BCRlow and 23 as BCRhigh. By correlating the BCR groups
with the available biological parameters, no association
was found between the 6-gene signature and IGHV gene
status (P=0.93) (Online Supplementary Table S2 and Online
Supplementary Figure S4A), Ki-67 expression, white blood
cells, hemoglobin, lymphocytes, platelets, and neutrophil
count (data not shown). The only significant difference was
between the BCR classification and lactate dehydrogenase
Figure 3. BCRhigh mantle cell lymphoma (MCL) group is associated with a worse
(LDH) levels; BCRhigh cases showed a higher level of LDH
clinical outcome. Kaplan-Meier curves obtained by comparing progression-free
with respect to BCRlow MCL (416.6±191.6 vs. 292.2±127.4;
P=0.023) (Online Supplementary Figure S4B). survival intervals of 23 BCRlow MCL cases with 23 BCRhigh MCL cases. The number
Clinically, MCL patients classified as BCRhigh experi- of patients in each group is reported under relative categories; P refers to log-
rank test.
enced shorter PFS with respect to BCRlow MCL cases

Table 2. Cox regression analysis on mantle cell lymphoma cases.

Univariable Multivariable
Variable HR (95%CI) P HR (95%CI) P
BCR signature
BCR high 2.81 (1.28-6.19) 0.01 3.48 (1.47-8.25) 0.005
MIPI-c
Low/intermediate 1.5 (0.6-3.73) 0.384
High/intermediate 1.41 (0.45-4.43) 0.557
High 3.46 (1.1-10.91) 0.034 4.17 (1.3-13.34) 0.016
Multivariable Cox regression analysis of progression-free survival was performed by including the 6-gene BCR categorization and the mantle cell lymphoma International
Prognostic Index-combined (MIPI-c) score as defined by Hoster et al.13 HR: Hazard Ratio; CI: Confidence Interval; BCR: B-cell receptor.

haematologica | 2018; 103(5) 853

 R. Bomben et al.

was considered as BCRlow according to GEP data. Also in
the context of LN samples, no correlation was found
between the different biological parameters and BCR
groups (data not shown).
By merging the MCL cases analyzed either in PB or in
LN, a total of 83 cases were collected, 43 BCRlow and 40
BCRhigh. BCRhigh patients had a shorter PFS with respect to
BCRlow patients (median PFS: 42.1 months vs. not reached;
P=0.0074) (Figure 4A). Since Ki-67 is a well-known prog-
nosticator in MCL,26 we combined the BCR groups with
the prognostic groups defined by Ki-67 score. Cases with
high Ki-67 (≥30% of Ki-67 expressing cells) and classified
in the BCRhigh group experienced the shortest PFS, while
cases classified as BCRlow had similar longer PFS intervals
irrespective of the high or low Ki-67 score (median PFS:
20.5 months vs. not reached for all the other combinations;
P=0.0014) (Figure 4B). Consistently, multivariable analysis
carried out by including the BCR signature and the MIPI-c
categories selected the BCRhigh and the high risk MIPI-c cat-
egory as independent predictors of PFS (Table 2). Regarding
OS, while the BCR readout failed to identify groups with
different OS intervals, possibly due to the low rate of
events and short follow up (Figure 4C), the combination of
high Ki-67 score and a BCRhigh 6-gene signature was able
again to select the MCL subgroup with the shortest OS
(46.7 vs. not reached; P=0.029) (Figure 4D).

A B

Figure 4. BCRhigh mantle cell

lymphoma (MCL) group is
associated with a worse clini-
cal outcome (overall series).
C (A) Kaplan-Meier curves
obtained by comparing pro-
D
gression-free survival (PFS)
intervals of 43 BCRlow MCL
cases with 40 BCRhigh MCL
cases. (B) Kaplan-Meier
curves obtained by compar-
ing PFS intervals of 19 BCRlow
and low Ki-67 MCL cases,
with 20 BCRlow and high Ki-67
MCL cases, with 21 BCRhigh
and Ki-67 low MCL cases,
with 10 BCRhigh and Ki-67 high
MCL cases. (C) Kaplan-Meier
curves obtained by compar-
ing overall survival (OS) inter-
vals of 43 BCRlow MCL cases
with 40 BCRhigh MCL cases.
(D) Kaplan-Meier curves
obtained by comparing OS
intervals of 19 BCRlow and low
Ki-67 MCL cases, with 20
BCRlow and high Ki-67 MCL
cases, with 21 BCRhigh and Ki-
67 low MCL cases, with 10
BCRhigh and Ki-67 high MCL
cases. The number of
patients in each group is
reported under relative cate-
gories; P refers to log-rank
test.

854 haematologica | 2018; 103(5)


Validations of BCR signature
To verify whether the BCR signature maintained its
prognostic impact in an independent set of patients, we
used the gene expression data of MCL LN biopsies report-
ed by Saba et al.30 Also in this different setting, a high
expression of the 6-gene signature, as in the context of
BCRhigh cases, identified an MCL patient subset with infe-
rior PFS (P=0.049) (Online Supplementary Figure S6).
In another set of analyses, by taking advantage of our 27
MCL cases with GEP data available, we correlated our
BCR signature with other MCL signatures with proven
clinical impact.30,38 As reported in Online Supplementary
Figure S7A, the BCR signature reported in Saba et al.30
divided MCL cases in two groups that corresponded
exactly to our BCR definition (Online Supplementary Figure
S7B).30 Similarly, the 17 genes of the proliferation signa-
ture reported by Scott et al.38 split our MCL cases in 3 dif-
ferent groups resembling the 3 different groups originally
defined (Online Supplementary Figure S8A). In this context,
the shortest PFS and OS intervals were observed in the
third group characterized by a higher expression of genes
related to proliferation and a BCRhigh phenotype in keeping
with our findings (Online Supplementary Figure S8A-C).
Discussion
In this study, we demonstrated that a BCR-derived sig-
nature based on the differential expression of six genes
correlated with shorter PFS intervals in the context of a
Phase III prospective clinical trial (FIL-MCL-0208) for
younger MCL patients receiving R-CHOP induction, fol-
lowed by high-dose cytarabine and autologous stem cell
transplantation (clinicaltrials.gov identifier: 023541313).32
Notably, the BCR-related 6-gene signature reported here
was able to identify an MCL subset with shorter PFS inter-
vals also in the context of an external independent MCL
cohort homogenously treated with different schedules.30
On the other hand, when the signature described by Saba
et al.30 and Scott et al.38 was applied to our MCL cases, the
patient subsets with the worse prognosis turned out to be
particularly enriched in BCRhigh cases, even though these
signatures did not include any gene from our signature.
Therefore, although composed of genes located upstream
of the BCR machinery, our signature was able to identify
cases with an active BCR pathway as defined by other sig-
natures.30 In this regard, however, experiments with pri-
mary MCL cases and/or MCL cell lines combining BCR
stimulation with the use of specific BCR inhibitors should
be performed to investigate the contribution of the 6-gene
signatures described here to the actual activation of the
BCR pathway.
Again in agreement with this line of reasoning, BCRhigh
samples presented a significant upregulation of PAX5 (see
GEP data in Online Supplementary Table S3), a gene whose
product is known to prevent plasma cell differentiation
thus preserving the capacity to respond to antigen-
induced activation and proliferation.39 Taken together
these data corroborate recent findings of ongoing active
BCR signaling in MCL cell in vivo,29,30 and further underline
the role of antigen stimulation in the ontogeny of MCL, as
suggested by the skewed IGVH gene repertoire found in
MCL cells.40
In order to discriminate between BCRlow and BCRhigh
MCL samples, we developed a DT model based on the
haematologica | 2018; 103(5)
B-cell receptor signature in mantle cell lymphoma
expression of the selected six genes.28 This DT model was
applied in an independent cohort of PB samples and then
to a further series of FFPE LN samples, thus demonstrating
that two MCL subsets with different expression levels of
BCR-related genes could also be recognized in the LN
compartment, mirroring PB. Taken together, by combin-
ing data from the PB and LN compartments, MCL cases
classified as BCRhigh showed higher LDH levels and shorter
PFS with respect to BCRlow patients, suggesting that activa-
tion of BCR signaling drives tumor proliferation and deter-
mines clinical outcome of MCL patients, which is in keep-
ing with recent findings.30
By combining the predictive capacity of the 6-gene BCR
signature with the Ki-67 index, we identified a particularly
unfavorable category (BCRhigh and high Ki-67) with a sub-
stantially shorter PFS and OS than the other groups.
Consistently, the BCRhigh signature turned out to be an
independent prognosticator along with the high-risk
MIPI-c category for short PFS by multivariate analysis.
There is no indication that the validity of the model may
be affected by the different recruitment site (PB vs. LN), or
by different sample storage (frozen vs. FFPE) because the
main clinical parameters were equally distributed
between the different series (PB/frozen vs. LN/FFPE) (R
Bomben et al., 2018, unpublished observation). In this regard,
an important feature of this model/assay is its applicabili-
ty to both PB and LN FFPE samples, having, therefore, the
chance to combine results of qRT-PCR with Ki-67 staining
in all the cases.
Our data underscore the increasing importance of BCR-
related genes in the pathogenesis and development of
MCL, further underlined by the clinical significance of
drugs specifically targeting genes belonging to this path-
way. In particular, therapeutic targeting of BTK41 can be
rationally exploited in lymphoid malignancies that have
been proved to be dependent on an antigen-dependent
BCR-mediated active signaling. However, despite the rel-
atively high response rate to single agent ibrutinib in
relapsed/refractory MCL, it remained unclear as to why
some patients showed clear responses, while others
received little therapeutic benefit.31,42 The BCR-related sig-
nature described here may provide insights into molecular
factors that explain the divergent responses of MCL
patients to ibrutinib, although other causes of primary
resistance might be related to gene mutations in the other
pathways, e.g. NF-κB pathway and epigenetic modifiers,
as recently reported.43,44
In conclusion, in the present study we developed a sur-
vival model for patients with MCL composed of six genes
(AKT3, BTK, CD79B, PIK3CD, SYK, BCL2) whose expres-
sion can easily be investigated by qRT-PCR and also in
FFPE specimens. The signature was associated with a poor
clinical response in the context of a high-dose chemo-
immunotherapy regimen, and might, therefore, be con-
sidered for validation and application in future clinical tri-
als.
Acknowledgments
The authors would like to thank Progetto Giovani Ricercatori
GR-2011-02347441, GR-2009-1475467, and GR-2011-
02351370, Ministero della Salute, Rome, Italy; Progetto Ricerca
Finalizzata RF-2009-1469205, and RF-2010-2307262,
Ministero della Salute, Rome, Italy; Associazione Italiana contro
le Leucemie, linfomi e mielomi (AIL), Venezia Section,
Pramaggiore Group, Italy; Associazione Italiana Ricerca Cancro
855
 R. Bomben et al.
(AIRC), Investigator Grant IG-2015 (17622); “5x1000
Intramural Program”, Centro di Riferimento Oncologico, Aviano,
Italy; Provincia Autonoma di Bolzano/Bozen, Italy; A.O. S.
Maurizio, Bolzano/Bozen, Italy; Progetto di Rilevante Interesse
Nazionale (PRIN2009) 7.07.02.60 AE01, Ministero Italiano
dell'Università e della Ricerca (MIUR), Roma, Italy; Fondi di
References
1. Dreyling M, Ferrero S, Hermine O. How to
manage mantle cell lymphoma. Leukemia.
2014;28(11):2117-2130.
2. Cheah CY, Seymour JF, Wang ML. Mantle
Cell Lymphoma. J Clin Oncol. 2016;
34(11):1256-1269.
3. Campo E, Swerdlow SH, Harris NL, Pileri S,
Stein H, Jaffe ES. The 2008 WHO classifica-
tion of lymphoid neoplasms and beyond:
evolving concepts and practical applications.
Blood. 2011;117(19):5019-5032.
4. Jares P, Colomer D, Campo E. Molecular
pathogenesis of mantle cell lymphoma. J
Clin Invest. 2012;122(10):3416-3423.
5. Dreyling M, Ferrero S, Vogt N, Klapper W.
New paradigms in mantle cell lymphoma: is
it time to risk-stratify treatment based on
the proliferative signature? Clin Cancer Res.
2014;20(20):5194-5206.
6. Ghielmini M, Zucca E. How I treat mantle
cell lymphoma. Blood. 2009;114(8):1469-
1476.
7. Herrmann A, Hoster E, Zwingers T, et al.
Improvement of overall survival in advanced
stage mantle cell lymphoma. J Clin Oncol.
2009;27(4):511-518.
8. Zucca E, Roggero E, Pinotti G, et al. Patterns
of survival in mantle cell lymphoma. Ann
Oncol. 1995;6(3):257-262.
9. Barista I, Romaguera JE, Cabanillas F.
Mantle-cell lymphoma. Lancet Oncol.
2001;2(3):141-148.
10. Martin P, Chadburn A, Christos P, et al.
Outcome of deferred initial therapy in man-
tle-cell lymphoma. J Clin Oncol. 2009;
27(8):1209-1213.
11. Dreyling M, Ferrero S. The role of targeted
treatment in mantle cell lymphoma: is trans-
plant dead or alive? Haematologica.
2016;101(2):104-114.
12. Geisler CH, Kolstad A, Laurell A, et al. The
Mantle Cell Lymphoma International
Prognostic Index (MIPI) is superior to the
International Prognostic Index (IPI) in pre-
dicting survival following intensive first-line
immunochemotherapy and autologous
stem cell transplantation (ASCT). Blood.
2010;115(8):1530-1533.
13. Hoster E, Rosenwald A, Berger F, et al.
Prognostic Value of Ki-67 Index, Cytology,
and Growth Pattern in Mantle-Cell
Lymphoma: Results From Randomized
Trials of the European Mantle Cell
Lymphoma Network. J Clin Oncol. 2016;
34(12):1386-1396.
14. Hoster E, Dreyling M, Klapper W, et al. A
new prognostic index (MIPI) for patients
with advanced-stage mantle cell lymphoma.
Blood. 2008;111(2):558-565.
15. Bea S, Valdes-Mas R, Navarro A, et al.
Landscape of somatic mutations and clonal
evolution in mantle cell lymphoma. Proc
Natl Acad Sci USA. 2013;110(45):18250-
18255.
16. Zhang J, Jima D, Moffitt AB, et al. The
genomic landscape of mantle cell lymphoma
is related to the epigenetically determined
chromatin state of normal B cells. Blood.
2014;123(19):2988-2996.
856
Ricerca Locale, Università degli Studi di Torino, Italy;
Fondazione Neoplasie Del Sangue (Fo.Ne.Sa), Torino, Italy;
CRT 2015.1044, Fondazione CRT, Torino, Italy. We are grate-
ful to all the Clinical Investigators, to the Pathologists, to Luigia
Monitillo, Daniela Barbero, Marina Ruggeri, Paola Ghione, and
Gian Maria Zaccaria.
17. Katzenberger T, Petzoldt C, Holler S, et al.
The Ki67 proliferation index is a quantita-
tive indicator of clinical risk in mantle cell
lymphoma. Blood. 2006;107(8):3407.
18. Hartmann E, Fernandez V, Moreno V, et al.
Five-gene model to predict survival in man-
tle-cell lymphoma using frozen or formalin-
fixed, paraffin-embedded tissue. J Clin
Oncol. 2008;26(30):4966-4972.
19. Ek S, Bjorck E, Porwit-MacDonald A,
Nordenskjold M, Borrebaeck CA. Increased
expression of Ki-67 in mantle cell lym-
phoma is associated with de-regulation of
several cell cycle regulatory components, as
identified by global gene expression analy-
sis. Haematologica. 2004;89(6):686-695.
20. Rosenwald A, Wright G, Wiestner A, et al.
The proliferation gene expression signature
is a quantitative integrator of oncogenic
events that predicts survival in mantle cell
lymphoma. Cancer Cell. 2003;3(2):185-197.
21. Mozos A, Royo C, Hartmann E, et al.
SOX11 expression is highly specific for
mantle cell lymphoma and identifies the
cyclin D1-negative subtype. Haematologica.
2009;94(11):1555-1562.
22. Navarro A, Clot G, Royo C, et al. Molecular
subsets of mantle cell lymphoma defined by
the IGHV mutational status and SOX11
expression have distinct biologic and clinical
features. Cancer Res. 2012;72(20):5307-
5316.
23. Majlis A, Pugh WC, Rodriguez MA,
Benedict WF, Cabanillas F. Mantle cell lym-
phoma: correlation of clinical outcome and
biologic features with three histologic vari-
ants. J Clin Oncol. 1997;15(4):1664-1671.
24. Wiestner A, Tehrani M, Chiorazzi M, et al.
Point mutations and genomic deletions in
CCND1 create stable truncated cyclin D1
mRNAs that are associated with increased
proliferation rate and shorter survival.
Blood. 2007;109(11):4599-4606.
25. Jares P, Colomer D, Campo E. Genetic and
molecular pathogenesis of mantle cell lym-
phoma: perspectives for new targeted thera-
peutics. Nat Rev Cancer. 2007;7(10):750-
762.
26. Determann O, Hoster E, Ott G, et al. Ki-67
predicts outcome in advanced-stage mantle
cell lymphoma patients treated with anti-
CD20 immunochemotherapy: results from
randomized trials of the European MCL
Network and the German Low Grade
Lymphoma Study Group. Blood. 2008;
111(4):2385-2387.
27. Perez-Galan P, Dreyling M, Wiestner A.
Mantle cell lymphoma: biology, pathogene-
sis, and the molecular basis of treatment in
the genomic era. Blood. 2011;117(1):26-38.
28. Young RM, Staudt LM. Targeting pathologi-
cal B cell receptor signalling in lymphoid
malignancies. Nat Rev Drug Discov. 2013;
12(3):229-243.
29. Akhter A, Street L, Ghosh S, et al.
Concomitant high expression of Toll-like
receptor (TLR) and B-cell receptor (BCR) sig-
nalling molecules has clinical implications in
mantle cell lymphoma. Hematol Oncol.
2015;35(1):79-86.
30. Saba NS, Liu D, Herman SE, et al.
Pathogenic role of B-cell receptor signaling
and canonical NF-kappaB activation in man-
tle cell lymphoma. Blood. 2016; 128(1):82-
92.
31. Wang ML, Rule S, Martin P, et al. Targeting
BTK with ibrutinib in relapsed or refractory
mantle-cell lymphoma. N Engl J Med.
2013;369(6):507-516.
32. Cortellazzo S, Martelli M, Ladetto M,
Ferrero S, Ciccone G, Evangelista A, et al.
High dose sequential chemotherapy with
rituximab and ASCT as first line therapy in
adult MCL patients: clinical and molecular
response of the MCL-0208 trial, a FIL study.
Haematologica. 2015;100(s1):3.
33. Swerdlow SH, Campo E, Pileri SA, et al. The
2016 revision of the World Health
Organization (WHO) classification of lym-
phoid neoplasms. Blood. 2016;
127(20):2375-2390.
34. Bomben R, Gobessi S, Dal BM, et al. The
miR-17-92 family regulates the response to
Toll-like receptor 9 triggering of CLL cells
with unmutated IGHV genes. Leukemia.
2012;26(7):1584-1593.
35. Dal Bo M, D'Agaro T, Gobessi S, et al. The
SIRT1/TP53 axis is activated upon B-cell
receptor triggering via miR-132 up-regula-
tion in chronic lymphocytic leukemia cells.
Oncotarget. 2015;6(22):19102-19117.
36. Draghici S, Khatri P, Martins RP, Ostermeier
GC, Krawetz SA. Global functional profiling
of gene expression. Genomics.
2003;81(2):98-104.
37. Subramanian A, Tamayo P, Mootha VK, et
al. Gene set enrichment analysis: a knowl-
edge-based approach for interpreting
genome-wide expression profiles. Proc Natl
Acad Sci USA. 2005;102(43):15545-15550.
38. Scott DW, Abrisqueta P, Wright GW et al.
New Molecular Assay for the Proliferation
Signature in Mantle Cell Lymphoma
Applicable to Formalin-Fixed Paraffin-
Embedded Biopsies. J Clin Oncol. 2017;
35(15):1668-1677.
39. Nera KP, Kohonen P, Narvi E, et al. Loss of
Pax5 promotes plasma cell differentiation.
Immunity. 2006;24(3):283-293.
40. Hadzidimitriou A, Agathangelidis A,
Darzentas N, et al. Is there a role for antigen
selection in mantle cell lymphoma?
Immunogenetic support from a series of 807
cases. Blood. 2011;118(11):3088-3095.
41. Rickert RC. New insights into pre-BCR and
BCR signalling with relevance to B cell
malignancies. Nat Rev Immunol. 2013;
13(8):578-591.
42. Wang ML, Blum KA, Martin P, et al. Long-
term follow-up of MCL patients treated
with single-agent ibrutinib: updated safety
and efficacy results. Blood. 2015; 126(6):739-
745.
43. Rahal R, Frick M, Romero R, et al.
Pharmacological and genomic profiling iden-
tifies NF-kappaB-targeted treatment strate-
gies for mantle cell lymphoma. Nat Med.
2014;20(1):87-92.
44. Lenz G, Balasubramanian S, Goldberg J,
Rizo A, Schaffer M, Phelps C, et al.
Sequence variants in patients with primary
and acquired resistance to ibrutinib in the
phase 3 MCL3001 (RAY) trial.
Haematologica. 2016;101(s1):155.
haematologica | 2018; 103(5)
 Non-Hodgkin Lymphoma ARTICLE

Incidence and risk factors for relapses in

HIV-associated non-Hodgkin lymphoma as Ferrata Storti
Foundation
observed in the German HIV-related lymphoma
cohort study
Philipp Schommers,1,2,* Daniel Gillor,1,* Marcus Hentrich,3 Christoph Wyen,1,4
Timo Wolf,5 Mark Oette,6 Alexander Zoufaly,7 Jan-Christian Wasmuth,8
Johannes R. Bogner,9 Markus Müller,10 Stefan Esser,11 Alisa Schleicher,12
Björn Jensen,13 Albrecht Stoehr,14 Georg Behrens,15,16 Alexander Schultze,17
Haematologica 2018
Jan Siehl,18 Jan Thoden,19 Ninon Taylor20 and Christian Hoffmann12,21
Volume 103(5):857-864

1
Department I of Internal Medicine, University Hospital Cologne, Germany; 2German
Center for Infection Research (DZIF), Partner Site Bonn-Cologne, Germany; 3Department
of Medicine III, Red Cross Hospital Munich, Germany; 4Praxis am Ebertplatz, Cologne,
Germany; 5Department of Medicine II, University of Frankfurt, Germany; 6Department of
General Medicine, Gastroenterology and Infectious Diseases, Augustinerinnen Hospital,
Cologne, Germany; 7Department of Medicine IV, Kaiser Franz Josef Hospital, Vienna,
Austria; 8Department of Internal Medicine I, University of Bonn, Germany; 9Department
of Medicine IV, University of Munich, Munich, Germany; 10Department of Infectious
Diseases, Vivantes Auguste-Viktoria- Hospital, Berlin, Germany; 11Department of
Dermatology, University Hospital Essen, Germany; 12University of Schleswig Holstein,
Campus Kiel, Kiel, Germany; 13Department of Gastroenterology, Hepatology and
Infectious Diseases, Düsseldorf University Hospital, Germany; 14Ifi-Institute for
Interdisciplinary Medicine, Hamburg, Germany; 15Department of Clinical Immunology
and Rheumatology, Hannover Medical School, Germany; 16German Center for Infection
Research (DZIF), Hannover, Germany; 17Department of Emergency Medicine, University
Medical Center Hamburg-Eppendorf, Hamburg, Germany; 18Ärzteforum Seestraße,
Berlin, Germany; 19Medical Group Practice for Internal Medicine and Rheumatology,
Freiburg, Germany; 20Department of Internal Medicine III with Hematology, Medical
Oncology, Hemostaseology, Infectious Diseases, Rheumatology, Oncologic Center,
Laboratory of Immunological and Molecular Cancer Research, Paracelsus Medical
University Salzburg, Austria and 21IPM Study Center, Hamburg, Germany
*PS and DG contributed equally to this work.
Correspondence:
philipp.schommers@uk-koeln.de
ABSTRACT

O
utcome of HIV-infected patients with AIDS-related lymphomas
has improved during recent years. However, data on incidence,
risk factors, and outcome of relapses in AIDS-related lym-
phomas after achieving complete remission are still limited. This
prospective observational multicenter study includes HIV-infected
patients with biopsy- or cytology-proven malignant lymphomas since
2005. Data on HIV infection and lymphoma characteristics, treatment
and outcome were recorded. For this analysis, AIDS-related lymphomas
patients in complete remission were analyzed in terms of their relapse-
free survival and potential risk factors for relapses. In total, 254 of 399
(63.7%) patients with AIDS-related lymphomas reached a complete
remission with their first-line chemotherapy. After a median follow up
of 4.6 years, 5-year overall survival of the 254 patients was 87.8%
(Standard Error 3.1%). Twenty-nine patients relapsed (11.4%). Several
factors were independently associated with a higher relapse rate, includ-
ing an unclassifiable histology, a stage III or IV according to the Ann
Arbor Staging System, no concomitant combined antiretroviral therapy
during chemotherapy and R-CHOP-based compared to more intensive
chemotherapy regimens in Burkitt lymphomas. In conclusion, complete
remission and relapse rates observed in our study are similar to those
reported in HIV-negative non-Hodgkin lymphomas. These data provide
further evidence for the use of concomitant combined antiretroviral ther-
apy during chemotherapy and a benefit from more intensive chemother-
apy regimens in Burkitt lymphomas. Modifications to the chemotherapy
regimen appear to have only a limited impact on relapse rate.
Received: September 17, 2017.
Accepted: February 2, 2018.
Pre-published: February 8, 2018.
doi:10.3324/haematol.2017.180893
Check the online version for the most updated
information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/857
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.

Haematologica | 2018; 103(5) 857

 P. Schommers et al.
Introduction
Over the last two decades, the incidence of AIDS-relat-
ed lymphomas (ARL) has markedly declined due to the
introduction of combination antiretroviral therapy
(cART). However, ARL remain a major cause of morbidity
and mortality, and represent the highest proportion of all
AIDS-related deaths.1 Patients with ARL are usually treat-
ed with the same chemotherapy protocols established in
the HIV-negative setting,2 and the rates of complete remis-
sion (CR) achieved are comparable to those reported in
their HIV-negative counterparts.3,4 However, available data
on the incidence and potential risk factors of recurrent dis-
ease in ARL are scarce, and treatment of disease relapse
remains challenging.4,5 In recent studies on HIV-negative
patients with diffuse large B-cell lymphoma (DLBCL), R-
CHOP-based regimens (rituximab, cyclophosphamide,
adriamycin, vincristine, and prednisone) resulted in CR
rates of around 65-80%. Differences in response rates
largely depend on the pre-treatment International
Prognostic Index (IPI) for aggressive lymphomas.6,7 In
patients who had achieved a CR, relapse rates ranged
from 6-10%.6,8 The second most common ARL are Burkitt
or Burkitt-like lymphomas (BL).9,10 In the HIV-negative set-
ting, CR and overall survival (OS) rates of around 80-90%
were reported by different groups.10-12 In a large prospec-
tive trial on short-intensive chemotherapy combined with
rituximab for patients with BL, the relapse rate was 12%.10
Although this approach also proved feasible in HIV-relat-
ed BL,9 it remains unclear whether relapse rates reported
in HIV-negative DLBCL and BL are different to those in
ARL. Thus, we investigated the risk factors and incidence
of relapse in a large cohort of ARL patients who had
achieved a CR after first-line treatment.
Methods
Study design
The German HIV Lymphoma Cohort is an ongoing, prospective
observational multicenter study including all adult HIV infected
patients who are diagnosed with biopsy- or cytology-proven
malignant lymphoma in 33 participating centers since January
2005. Data on HIV-infection and lymphoma characteristics, treat-
ment and outcome are recorded. From the time of lymphoma
diagnosis, patients are followed every six months. Ethics approval
was obtained from the ethics committees of the University of
Cologne (IRC Cologne: 05-174), Germany, and written informed
consent was given by each participating patient.
The present analysis includes only patients with aggressive B-
cell lymphoma in first CR. Lymphomas were grouped in DLBCL,
BL, plasmablastic lymphoma (PBL) and ARL, not further classifi-
able, the latter group representing aggressive B-cell non-Hodgkin
lymphomas (B-NHL) that could not be classified into any subtype.
To study the impact of chemotherapy dose intensity on the risk of
relapse in patients treated with either R-CHOP-based regimens or
the short intensive GMALL protocol,10 we performed an analysis
of dose reductions and delays in chemotherapy cycles.
(Information about the GMALL and R-CHOP protocol can be
found in the Online Supplementary Tables S1 and S2, respectively).
A full-intensity treatment was considered to consist of six cycles
of chemotherapy according to the R-CHOP or GMALL protocol
administered at 3-week intervals without dose reductions and
within a period of 120 days (5x21 days for 6 cycles plus a maxi-
mum of 3 days delay per cycle). Less than 6 cycles of chemother-
858
apy were classified as “cycle reduction” and the treatment dura-
tion was calculated according to the number of cycles given. If the
dose of any chemotherapy drug was reduced by 20% or more,
treatment intensity was considered to be reduced.
As positron emission tomography (PET) scans were not routine-
ly performed, the 1999 standardized response criteria for non-
Hodgkin’s lymphomas13 were used rather than the 2007 criteria.14
CR was defined as the disappearance of all disease manifestations
for at least three months. This definition also includes uncertain
complete remission (CRu) that implied a residual mass of 1.5 cm
or smaller that remained unchanged over at least three months.
Statistical analysis
Statistical analyses were performed using IBM SPSS Statistics
software (IBM, Armonk, NY, USA), v.24.0. Univariate statistics
were performed using Pearson’s χ², Fisher’s exact one-way
Analysis of Variance (ANOVA) with Bonferroni-corrected post-
hoc test, or Kruskal-Wallis test depending on data. For the multi-
variate Cox regression analysis, continuous clinically meaningful
breakpoints that showed P-values below 0.1 in the univariate
analysis were considered. Kaplan-Meier curves were used to illus-
trate the relapse-free survival (RFS) and overall survival.
Differences between subgroups were assessed with the log-rank
test. RFS was defined as the period between first diagnosis and
any lymphoma relapse according to the STEEP criteria.15 OS was
defined as the period between first diagnosis and death from any
cause. All-cause deaths as well as “lost to follow up” were cen-
sored. All P-values were two-sided. P<0.05 was considered statis-
tically significant.
Results
Patients' characteristics and outcome
Numbers and characteristics of patients included in the
present analysis are depicted in Figure 1. In total, 254 of
399 (63.7%) patients with high-grade NHL of B-cell origin
(classified as ARL) reached a CR with their first-line
chemotherapy. Of those, 127 had DLBCL, 91 BL, 29 PBL,
and 7 ARL, not further classified. ARL was CD20-negative
in 24 of 254 cases (9.5%), among them 22 PBL and 2
DLBCL cases. Among 22 PBL cases with information on
Epstein-Barr-Virus (EBV) status, EBV was present in 15
(68%). Overall, 86.2% of patients with CD20+ lym-
phomas received rituximab. Notably, patients diagnosed
before 2010 were less frequently treated with rituximab
than those diagnosed from 2010 onwards (79% vs. 96%,
P<0.001). Further, 73% of patients with CD4 cell counts
less than 50/µl received rituximab compared to 88% with
CD4 counts 50/mL or over (P=0.096). Patients' character-
istics with respect to treatment outcomes are listed in
Table 1. OS of patients who achieved CR with first-line
therapy was significantly better than that of patients in
other response groups (Figure 2). After a median follow up
of 4.6 years, 5-year OS of the entire group of all 262
patients in first CR was 87.1% [standard error (SE) 2.3%]
(Figure 3A) with differences between lymphoma sub-
types: 87.8% (SE 3.1%) in DLBCL, 87.6% (SE 3.7%) in BL,
79.6% (SE 11.3%) in PBL, and 83.3% (SE 15.2%) in ARL,
not further classified (P=0.994) (Figure 3B).
Incidence of recurrent disease in ARL
After a median follow up of 4.6 years, a relapse of the
ARL had occurred in 29 of 254 patients (11.4%). Relapses
were observed in 14 patients with DLBCL (11.0%), 9 with
haematologica | 2018; 103(5)

BL (9.9 %), 3 with PBL (11.5%) and 3 with ARL, not fur-
ther classified (42.9%), after a median follow up of 5.0,
4.6, 3.5 and 6.0 years, respectively. Isolated central nerv-
ous system (CNS) relapses were observed in 3 of 29
patients (DLBCL: n=2; BL: n=1). RFS depicted by Kaplan
Meier curves is shown in Figure 3C and D. Five-year RFS
(5yRFS) was 88.4% (SE 2.9%) in DLBCL, 88.9% (SE 3.5%)
in BL, and 88.6% (SE 6.2%) in PBL. By contrast, 5yRFS
was lower in patients with ARL, not further classified
[57.1% (SE 18.7%); P=0.057) (Figure 3D).
Among patients who achieved CR with first-line R-
CHOP-based protocols, 5yRFS was 87.8% (SE 3.1%) and
84.4% (SE 8.3%) in DLBCL and PBL, respectively, as com-
pared to 65.5% (SE 12.6) in BL and 40.0% in ARL, not fur-
ther classified (SE 21.9%; P=0.005) (Figure 3E). No signifi-
cant differences in 5yRFS between ARL subtypes were
observed in patients treated with the GMALL protocol
haematologica | 2018; 103(5)
Relapses in AIDS-related lymphomas
(P=0.884) (Figure 3F), although the number of patients
with subtypes other than BL was very small in this analy-
sis. Of note, patients with BL who received the GMALL
protocol had a significantly better 5yRFS than those
receiving R-CHOP-based protocols [94.2% (SE 2.8%) vs.
65.5% (SE 12.6%); P=0.001].
Risk factors for recurrent disease in ARL
Univariate analysis identified several factors associated
with a lower risk for ARL relapse such as a low IPI, stage
I or II according to the Ann Arbor Staging System, cART
given during chemotherapy, CD4 T-cell counts
>200x109/L, pathology other than ARL, not further classi-
fied, and chemotherapy according to the GMALL-proto-
col (Table 2). These factors were analyzed in a multivari-
ate Cox proportional hazards model, with backward step-
wise elimination based on a Wald statistic with P≤0.1.
Figure 1. Flow chart of patients included in the present analysis. NHL: non-
Hodgkin lymphoma; T-NHL: T-cell non-Hodgkin lymphoma; CR: complete remis-
sion; BL: Burkitt lymphoma; DLBCL: diffuse large B-cell lymphoma; PBL: plas-
mablastic lymphoma; ARL: AIDS-related lymphoma.
859
 P. Schommers et al.

After two elimination steps, histology [BL: Hazard ratio
(HR) 2.60 95% Confidence Interval (95%CI): 0.92 – 7.4,
PBL: HR 1.28 95% CI 0.36-4.57, ARL, not further classi-
fied: HR 5.08 95% CI 1.13 – 22.90, indicator = DLBCL),
stage III or IV according to the Ann Arbor Staging System
(HR 4.85 95%CI 1.44 – 16.34), no concomitant cART (HR
4.28 95%CI 1.19 – 15.39) and use of R-CHOP (HR: 7.59
95%CI 1.87 – 30.81) remained in the model (Table 2). A
higher IPI was no longer predictive anymore in the multi-
variate model.
Dose intensity of chemotherapy
Since chemotherapy regimen (R-CHOP or GMALL)
seems to be critical for RFS, we investigated how many
patients of all aggressive B-NHL had any kind of reduc-
tion (either in the number of chemotherapy cycles or in
the treatment intensity) or a delay during their treatment.
Results of dose intensity analysis are shown in Online
Supplementary Table S3. Overall, 32.7% of the patients
had a treatment delay, 13.8% had dose reductions, and
16.5% had reduced numbers of chemotherapy cycles

Table 1. Patients’ characteristics based on their treatment outcome.

CR after CR after Progressive Partial On Treatment No Total P
first-line further lines disease remission chemotherapy related chemotherapy (N=387)
chemotherapy chemotherapy (n=45) (n=22) (n=15) deaths given
(n=254) (n=21) (n=23) (n=7)
Median age (years) 44 48 45 44 50 48 46 45 0.104b
Male 230 (91%) 20 (95%) 44 (98%) 21 (96%) 15 (100%) 21 (91%) 6 (86%) 357 (92%) 0.514a
Median viral load 19031 26790 40208 7159 11387 2390 105000 18557 0.784b
(copies/mL)
HIV-RNA below limit of detection 74 (30%) 5 (24%) 11 (25%) 8 ((36%) 4 (29%) 8 (35%) 2 (29%) 112 (30%) 0.903a
Median CD4+ 248 190 111 186 153 157 58 212 0.006b
T cells (x109/L)
CD20+ lymphoma 213 (90%) 20 (95%) 34 (81%) 18 (90%) 10 (83%) 13 (68%) 4 (67%) 312 (87%) 0.044a
BM involvement 47 (20%) 7 (33%) 15 (39%) 6 (29%) 3 (21%) 6 (21%) 1(17%) 85 (23%) 0.190a
CNS involvement 17 (8%) 2 (10%) 8 (21%) 2 (11%) 1 (9%) 5 (33%) 0 (0%) 36 (11%) 0.023
IPI score
Low 100 (42%) 5 (24%) 8 (18%) 2 (11%) 5 (39%) 2 (9%) 0 (0%) 122 (36%)
Intermediate 104 (44%) 11 (52%) 20 (46%) 13 (68%) 6 (42%) 11 (50%) 3 (43%) 168 (46%)
High 34 (14%) 5 (24%) 16 (36%) 4 (21%) 2 (15%) 9 (41%) 4 (57%) 74 (20%) <0.001a
Lymphoma subtype
DLBCL 127 (50%) 7 (33%) 24 (53%) 13 (59%) 7 (47%) 7 (30%) 4 (57%) 189 (49%)
BL 91 (36%) 11 (52%) 12 (27%) 5 (23%) 4 (27%) 9 (39%) 1 (14%) 133 (34%)
PBL 29 (11%) 3 (14%) 7 (16%) 3 (14%) 3 (20%) 7 (30%) 2 (29%) 54 (14%)
ARL, not further classifiable 7 (3%) 0 (0%) 2 (4%) 1 (5%) 1 (7%) 0 (0%) 0 (0%) 11 (3%) 0.420a
Median follow up (years) 4.64 5.4 0.5 0.7 0.1 0.2 0 2.4 <0.001b
BL: Burkitt lymphoma; DLBCL: diffuse large B-cell lymphoma; PBL: plasmablastic lymphoma; IPI: International Prognostic Index; BM: bone marrow; CNS: central nervous sys-
tem. aTwo-sided Pearson’s χ2. bKruskal-Wallis test.

Figure 2. Kaplan-Meier estimates for overall survival (OS) of the differ-

ent observed treatment outcomes of aggressive non-Hodgkin lym-
phomas. CR: complete remission; Prog. Disease: progressive disease;
Part. Remission: partial remission. Dotted line indicates 3 months.

860 haematologica | 2018; 103(5)

 Relapses in AIDS-related lymphomas

given. Only 37.0% of patients received their full planned
course of therapy.
Patients who experienced treatment delays and/or dose
reductions or received a reduced number of chemotherapy
cycles were significantly older (42 vs. 45 years; P=0.042)
and had received the GMALL-protocol significantly more
often than patients who completed the full planned course
of therapy (P=<0.001) (Online Supplementary Table S3).
Overall, 86.2% of patients with CD20+ lymphomas
received rituximab. There was no difference in the relapse
rate between patients with or without administration of
rituximab (P=0.75), and our results remained consistent
when patients without rituximab were excluded (data not
shown).
Factors influencing the RFS
To investigate the influence of different factors, includ-
ing treatment reduction or delay on 5yRFS, we investigat-

Table 2. Risk factors for 5-year relapse-free survival (including all aggressive non-Hodgkin lymphoma in first complete remission; n=254).
Aggressive
NHL (N=254)
5-year relapse-free P P
survival % (univariate) (multivariate)
Sex Male (n=230) 87
Female (n=24) 96 0.229
Age >60Y (n=24) 77
<60Y (n=228) 89 0.178
CNS involvement Yes (n=17) 82
No (n=204) 90 0.305
BM involvement Yes (n=47) 81
No (n=193) 90 0.101
Bulky Disease Yes (n=44) 86
No (n=137) 88 0.637
CD4+ T cells <50x109/l Yes (n=37) 86
No (n=202) 88 0.573
Prior AIDS-defining illness Yes (n=56) 87
No (n=193) 88 0797
IPI score Low (n=100) 95 Indicator
Intermediate (n=104) 84 0.760
High (n=34) 82 0.039 0.853
Ann Arbor stage I/II (n=93) 95
III/IV (n=156) 83 0.005 0.011
Extranodal involvement Yes (n=71) 87
No (n=181) 88 0.936
ECOG score 0-1 (n=159) 89
2-5 (n=78) 86 0.635
Elevated LDH Yes (n=146) 86
No (n=96) 92 0.143
Antiretroviral Treatment Viral load b.d. (n=74) 88
Naive (n=134) 88
Therapy failure (n=39) 87 0.986
cART during Chemotherapy Yes (n=234) 89
No (n=9) 67 0.033 0.026
CD20 positive lymphoma Positive (n=213) 88
Negative (n=24) 87 0.888
Lymphoma subtype DLBCL (n=127) 88 Indicator
BL (n=91) 89 0.072
PBL (n=29) 89 0.703
ARL, not further classifiable (n=7) 57 0.064 0.034
Chemotherapy CHOP (n=163) 84
GMALL (n=87) 95 0.013 0.005
Univariate statistics: Log rank test. Multivariate statistics: Cox regression. Viral load b.d.: Viral load below limit of detection; cART: combination antiretroviral therapy; BL: Burkitt
lymphoma; DLBCL: diffuse large B-cell lymphoma; PBL: plasmablastic lymphoma; IPI: International Prognostic Index; BM: bone marrow; CNS: central nervous system; ECOG:
Eastern Cooperative Oncology Group scale.

haematologica | 2018; 103(5) 861

 P. Schommers et al.

ed two selected groups: patients with DLBCL receiving R- Discussion

CHOP-based regimens and patients with BL receiving the
GMALL-regimen (see also Figure 1). In patients with
DLBCL who received a R-CHOP-based treatment, a high
IPI, an elevated LDH, stage III or IV according to the Ann
Arbor Staging System, and bone marrow involvement at
first diagnosis were associated with a significantly
increased risk of relapse (Online Supplementary Table S4).
However, none of these parameters turned out to be an
independent risk factor in the Cox proportional hazards
model. Of note, chemotherapy dose reductions, treatment
delays or a reduced number of R-CHOP-cycles given did
not adversely affect 5yRFS (Online Supplementary Table S4).
By univariate analysis, patients with BL who underwent
GMALL-chemotherapy had a significantly increased risk
of relapse if the following factors were present: diagnosis
of another AIDS-defining disease prior to BL diagnosis,
CNS-involvement, failure of cART defined as measurable
viral loads despite concomitant cART, concomitant cART
during chemotherapy, and reduced numbers of
chemotherapy cycles administered (Online Supplementary
Table S4). However, none of these factors remained signif-
icant in the Cox proportional hazards model.
In this large prospective cohort study of 254 ARL
patients who had achieved a CR with first-line
chemotherapy (64% of all cases), the total relapse rate was
11% after a median follow up of 4.6 years. Patients with
DLBCL who were mainly treated with R-CHOP-based
regimens had a relapse rate of 11%. These rates are in line
with the 6-10% relapse rates reported in HIV-negative
DLBCL in first CR.6,8 Notably, the 10% relapse rate of
patients with BL who were mainly treated with the
GMALL protocol compares favorably with the 12%
relapse rate reported in the HIV-negative setting.10
Outcome of patients with ARL in which histology was
not further classifiable was poor with a 5yRFS of only
57%. There was no difference in type and intensity of
chemotherapy to that used in DLBCL (data not shown),
therefore these cases may represent a subgroup of highly
aggressive lymphomas that may benefit from intensive
chemotherapy regimens.
Even though the overall survival of patients with PBL
was shown to be significantly worse than that in DLBCL
and BL,16,17 there was no difference in relapse rates

A B

C D

E F

Figure 3. Kaplan-Meier estimates for aggressive non-Hodgkin lymphoma (NHL) that achieved complete remission (CR) after first-line chemotherapy. (A) Overall
survival of all AIDS-related lymphomas (ARL) and of (B) different subtypes (Log rank test: P=0.982). (C) Relapse-free survival of all ARL and of (D) different subtypes
(P=0.064). (E) Relapse-free survival of different subtypes treated with R-CHOP-based regimens (rituximab, cyclophosphamide, adriamycin, vincristine, and pred-
nisone) and (F) GMALL-based chemotherapeutic regimens (P=0.006 and P=0.79, respectively). DLBCL: diffuse-large B-cell lymphoma; BL: Burkitt-lymphoma; PBL:
plasmablastic lymphoma. Dotted line indicates 3 months.

862 haematologica | 2018; 103(5)

 Relapses in AIDS-related lymphomas

between patients who have achieved a CR (12%) to those
observed in DLBCL and BL. The inferior OS of the small
patient group of PBL may at least in part be explained by
3 late deaths 4-7 years after first diagnosis that were unre-
lated to lymphoma: one case of sepsis due to pneumonia
and 2 cases of secondary malignancies (lung cancer and
oral cavity cancer).
AIDS-related lymphoma patients with an intermediate
and high IPI had higher relapse rates and a lower 5yRFS
than those with a low IPI in univariate analysis. The lack
of significance in the multivariate analysis was somewhat
surprising as previous studies have demonstrated strong
prognostic relevance of the IPI in ARL.17,18 Whether
patients with HIV-related DLBCL and intermediate or
high IPI may benefit from more intensive treatments such
as the CHOEP regimen, as has been shown in the HIV-
negative setting, remains to be seen.19,20
Previous studies have shown that concomitant cART
was associated with improved CR rates and a trend
toward improved OS.21 Our results also support a concur-
rent use of cART as it was associated with better 5yRFS.
Several cART regimens with a good safety and tolerability
profile and low interaction potential are now available,
strongly arguing for a simultaneous cART during ARL
chemotherapy.4
The use of R-CHOP-based regimens showed signifi-
cantly less treatment delays and reductions, as compared
to the GMALL protocol. However, the majority of
patients with BL (84%) received chemotherapy according
to the GMALL-protocol which resulted in significantly
lower relapse rates compared to R-CHOP-based regimens
(Figure 3E and F). Notably, treatment delays and a reduced
chemotherapy intensity appeared to have no impact on
the relapse-rate in GMALL-treated BL, while, at least in
the univariate analysis, a reduced number of chemothera-
py cycles was associated with lower 5yRFS. Thus, our
data indicate that HIV-infected patients with BL should be
treated with the planned number of intensive chemother-
apy cycles.3,8 By contrast, reduced relative dose intensity
did not negatively impact 5yRFS in patients treated with
R-CHOP-based regimens for DLBCL. This finding does
not correspond to data reported in HIV-negative DLBCL
and warrants further investigation.22,23
It is important to note that this analysis focuses on
patients in first CR, and that factors that predict RFS were
not necessarily associated with initial treatment response.
Our study has several limitations. First, given the uncon-
trolled design selection biases cannot be ruled out. Second,
the analysis of risk factors associated with outcome is
more exploratory in nature. Given the relatively low num-
ber of patients in some of the selected subgroups, the sta-
tistical power of the analysis is limited and does not allow
any firm conclusions to be drawn. Notably, data on poten-
tial risk factors for lymphoma relapse such as adherence to
ART or cumulative viremia between CR and relapse are
not available.24 Nevertheless, if a CR has been reached, the
relapse rate was low regardless of whether the CR was
achieved with or without dose reduction and whether rit-
uximab was used or not. Fourth, CRs were not generally
confirmed by negative positron emission tomography
(PET) scans as recommended by current guidelines for
HIV-negative lymphomas.14,25 However, the role of PET-
scanning in HIV-lymphoma remains controversial as the
rate of false positive results appears to be higher than in
the HIV-negative setting.26,27 Finally, the number of
patients with ARL, not further classified, may have been
lowered by reference pathology services which, in turn,
may have slightly altered our findings.
In conclusion, both CR rates and relapse rates observed
in the German HIV-related Lymphoma Cohort Study are
similar to those reported in HIV-negative NHL. These data
add to the growing body of evidence showing that treat-
ment outcomes compare favorably with those in patients
with NHL and no HIV infection.

6. Pfreundschuh M, Kuhnt E, Trumper L, et al. multicenter trial. Blood. 2014;124(26):3870-

References
1. Lewden C, May T, Rosenthal E, et al.
Changes in causes of death among adults
infected by HIV between 2000 and 2005:
The "Mortalite 2000 and 2005" surveys
(ANRS EN19 and Mortavic). J Acquir
Immune Defic Syndr. 2008;48(5):590-598.
2. Barta SK, Dunleavy K, Mounier N. Diffuse
large B-cell lymphoma. Hentrich M, Barta
SK (eds.). HIV-associated Hematological
Malignancies. Springer International
Publishing; 2016.
3. Dunleavy K, Wilson WH. How I treat HIV-
associated lymphoma. Blood. 2012;
119(14):3245-3255.
4. Brunnberg U, Hentrich M, Hoffmann C,
Wolf T, Hubel K. HIV-Associated
Malignant Lymphoma. Oncol Res Treat.
2017;40(3):82-87.
5. Re A, Michieli M, Casari S, et al. High-dose
therapy and autologous peripheral blood
stem cell transplantation as salvage treat-
ment for AIDS-related lymphoma: long-
term results of the Italian Cooperative
Group on AIDS and Tumors (GICAT) study
with analysis of prognostic factors. Blood.
2009;114(7):1306-1313.
CHOP-like chemotherapy with or without
rituximab in young patients with good-
prognosis diffuse large-B-cell lymphoma: 6-
year results of an open-label randomised
study of the MabThera International Trial
(MInT) Group. Lancet Oncol. 2011;
12(11):1013-1022.
7. Feugier P, Van Hoof A, Sebban C, et al.
Long-term results of the R-CHOP study in
the treatment of elderly patients with dif-
fuse large B-cell lymphoma: a study by the
Groupe d'Etude des Lymphomes de
l'Adulte. J Clin Oncol. 2005;23(18):4117-
4126.
8. Rovira J, Valera A, Colomo L, et al.
Prognosis of patients with diffuse large B
cell lymphoma not reaching complete
response or relapsing after frontline
chemotherapy or immunochemotherapy.
Ann Hematol. 2015;94(5):803-812.
9. Montoto S, Noy A, Ribera JM. Burkitt lym-
phoma. Hentrich M, Barta SK (eds): HIV-
associated Hematological Malignancies.
Springer International Publishing; 2016.
10. Hoelzer D, Walewski J, Dohner H, et al.
Improved outcome of adult Burkitt lym-
phoma/leukemia with rituximab and
chemotherapy: report of a large prospective
3879.
11. Dunleavy K, Pittaluga S, Shovlin M, et al.
Low-intensity therapy in adults with
Burkitt's lymphoma. N Engl J Med. 2013;
369(20):1915-1925.
12. Ribrag V, Koscielny S, Bosq J, et al.
Rituximab and dose-dense chemotherapy
for adults with Burkitt's lymphoma: a ran-
domised, controlled, open-label, phase 3
trial. Lancet. 2016;387(10036):2402-2411.
13. Cheson BD, Horning SJ, Coiffier B, et al.
Report of an international workshop to
standardize response criteria for non-
Hodgkin's lymphomas. NCI Sponsored
International Working Group. J Clin Oncol.
1999;17(4):1244.
14. Cheson BD, Pfistner B, Juweid ME, et al.
Revised response criteria for malignant
lymphoma. J Clin Oncol. 2007;25(5):579-
586.
15. Hudis CA, Barlow WE, Costantino JP, et al.
Proposal for standardized definitions for
efficacy end points in adjuvant breast can-
cer trials: the STEEP system. J Clin Oncol.
2007;25(15):2127-2132.
16. Schommers P, Wyen C, Hentrich M, et al.
Poor outcome of HIV-infected patients
with plasmablastic lymphoma: results from

haematologica | 2018; 103(5) 863

 P. Schommers et al.
the German AIDS-related lymphoma
cohort study. Aids. 2013;27(5):842-845.
17. Schommers P, Hentrich M, Hoffmann C, et
al. Survival of AIDS-related diffuse large B-
cell lymphoma, Burkitt lymphoma, and
plasmablastic lymphoma in the German
HIV Lymphoma Cohort. Br J Haematol.
2015;168(6):806-810.
18. Barta SK, Xue X, Wang D, et al. A new
prognostic score for AIDS-related lym-
phomas in the rituximab-era.
Haematologica. 2014;99(11):1731-1737.
19. Hentrich M, Hoffmann C, Mosthaf F, et al.
Therapy of HIV-associated lymphoma-rec-
ommendations of the oncology working
group of the German Study Group of
Physicians in Private Practice Treating HIV-
Infected Patients (DAGNA), in cooperation
with the German AIDS Society (DAIG).
Ann Hematol. 2014;93(6):913-921.
20. Schmitz N, Nickelsen M, Ziepert M, et al.
Conventional chemotherapy (CHOEP-14)
864
with rituximab or high-dose chemotherapy
(MegaCHOEP) with rituximab for young,
high-risk patients with aggressive B-cell
lymphoma: an open-label, randomised,
phase 3 trial (DSHNHL 2002-1). Lancet
Oncol. 2012;13(12):1250-1259.
21. Barta SK, Xue X, Wang D, et al. Treatment
factors affecting outcomes in HIV-associat-
ed non-Hodgkin lymphomas: a pooled
analysis of 1546 patients. Blood. 2013;
122(19):3251-3262.
22. Bosly A, Bron D, Van Hoof A, et al.
Achievement of optimal average relative
dose intensity and correlation with survival
in diffuse large B-cell lymphoma patients
treated with CHOP. Ann Hematol. 2008;
87(4):277-283.
23. Hirakawa T, Yamaguchi H, Yokose N, Gomi
S, Inokuchi K, Dan K. Importance of main-
taining the relative dose intensity of CHOP-
like regimens combined with rituximab in
patients with diffuse large B-cell lymphoma.
Ann Hematol. 2010;89(9):897-904.
24. Zoufaly A, Stellbrink HJ, Heiden MA, et al.
Cumulative HIV viremia during highly
active antiretroviral therapy is a strong pre-
dictor of AIDS-related lymphoma. J Infect
Dis. 2009;200(1):79-87.
25. Younes A, Hilden P, Coiffier B, et al.
International Working Group consensus
response evaluation criteria in lymphoma
(RECIL 2017). Ann Oncol. 2017;28(7):1436-
1447
26. Mhlanga JC, Durand D, Tsai HL, et al.
Differentiation of HIV-associated lym-
phoma from HIV-associated reactive
adenopathy using quantitative FDG PET
and symmetry. Eur J Nucl Med Mol
Imaging. 2014;41(4):596-604.
27. Sathekge M. Differentiation of HIV-associ-
ated lymphoma from HIV-reactive
adenopathy using quantitative FDG-PET
and symmetry. Eur J Nucl Med Mol
Imaging. 2014;41(4):593-595.
haematologica | 2018; 103(5)
 Chronic Lymphocytic Leukemia ARTICLE

Highly similar genomic landscapes in

monoclonal B-cell lymphocytosis and Ferrata Storti
Foundation
ultra-stable chronic lymphocytic leukemia with
low frequency of driver mutations
Andreas Agathangelidis,1* Viktor Ljungström,2* Lydia Scarfò,1 Claudia Fazi,1
Maria Gounari,1,3 Tatjana Pandzic,2 Lesley-Ann Sutton,2,4
Kostas Stamatopoulos,3 Giovanni Tonon,5 Richard Rosenquist2,4**
and Paolo Ghia1**
Haematologica 2018
1
Strategic Research Program on CLL and B-cell Neoplasia Unit, Division of Experimental Volume 103(5):865-873
Oncology, Università Vita-Salute San Raffaele and IRCCS Istituto Scientifico San
Raffaele, Milan, Italy; 2Department of Immunology, Genetics and Pathology, Science for
Life Laboratory, Uppsala University, Sweden; 3Institute of Applied Biosciences, Center for
Research and Technology Hellas, Thessaloniki, Greece; 4Department of Molecular
Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden and 5Functional
Genomics of Cancer Unit, Division of Experimental Oncology, IRCCS Istituto Scientifico
San Raffaele, Milan, Italy
*
AA and VL contributed equally as first authors. **RR and PG contributed equally as last authors

ABSTRACT

D
espite the recent discovery of recurrent driver mutations in chron-
ic lymphocytic leukemia, the genetic factors involved in disease
onset remain largely unknown. To address this issue, we per-
formed whole-genome sequencing in 11 individuals with monoclonal B-
cell lymphocytosis, both of the low-count and high-count subtypes, and
5 patients with ultra-stable chronic lymphocytic leukemia (>10 years
without progression from initial diagnosis). All three entities were indis-
tinguishable at the genomic level exhibiting low genomic complexity
and similar types of somatic mutations. Exonic mutations were not fre-
quently identified in putative chronic lymphocytic leukemia driver genes Correspondence:
in all settings, including low-count monoclonal B-cell lymphocytosis. To ghia.paolo@hsr.it
corroborate these findings, we also performed deep sequencing in 11
known frequently mutated genes in an extended cohort of 28 monoclon-
al B-cell lymphocytosis/chronic lymphocytic leukemia cases.
Received: August 8, 2017.
Interestingly, shared mutations were detected between clonal B cells and
paired polymorphonuclear cells, strengthening the notion that at least a Accepted: February 7, 2018.
fraction of somatic mutations may occur before disease onset, likely at Pre-published: February 15, 2018.
the hematopoietic stem cell level. Finally, we identified previously unre-
ported non-coding variants targeting pathways relevant to B-cell and
chronic lymphocytic leukemia development, likely associated with the doi:10.3324/haematol.2017.177212
acquisition of the characteristic neoplastic phenotype typical of both
monoclonal B-cell lymphocytosis and chronic lymphocytic leukemia. Check the online version for the most updated
information on this article, online supplements,
and information on authorship & disclosures:
Introduction www.haematologica.org/content/103/5/865

Chronic lymphocytic leukemia (CLL), the most common adult leukemia in the
West, is a clinically heterogeneous disease.1 At one end of the spectrum, CLL
patients present with an indolent disease that does not require therapy for
decades. At the other end of the spectrum, patients experience a rapidly progres-
sive disease, need early treatment, and frequently relapse.2,3
High-throughput studies14,15 have established that, though displaying a markedly
lower mutational burden compared to solid tumors,16 CLL is characterized by a
diverse genetic landscape with driver gene mutations in pathways considered cen-
tral for disease pathogenesis, e.g. NOTCH and NF-κB signaling.7,9,17 The frequency
of most driver gene mutations in CLL tends to increase in aggressive/refractory
cases supporting their involvement mainly in disease progression.18-20
Chronic lymphocytic leukemia is preceded by a condition termed monoclonal
B-cell lymphocytosis (MBL) that is characterized by the presence of circulating
monoclonal B cells with a CLL phenotype, however, at a lower concentration than
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.

Haematologica | 2018; 103(5) 865

 A. Agathangelidis et al.
required for a clinical diagnosis of CLL (≥5x109/L).21-24
MBL, found in otherwise healthy individuals, is divided
into 2 subtypes based on the number of circulating cells:
‘high-count MBL’ (HC-MBL: 0.5-5x109/L) that evolves into
CLL requiring therapy at a rate of 1%/year,25 and ‘low-
count MBL’ (LC-MBL: <0.5x109/L) that has not been
observed to progress into a clinical disease,26 yet persists
over time.26,27 Several typical CLL driver gene mutations
have been reported in HC-MBL9,28,29 even years before the
transition to CLL,30 and these correlate with adverse dis-
ease course.31 Such mutations have been reported in mul-
tipotent hematopoietic progenitor CD34+ cells from
patients with CLL,32 suggesting that such aberrations may
also be implicated in CLL onset.
Here, we aimed to gain insight into the genetic lesions
that may be involved in the transformation from MBL to
CLL, analyzing LC-MBL cases for the first time. To this
end, we used whole-genome sequencing (WGS) and tar-
geted re-sequencing to profile LC-MBL, HC-MBL and a
particularly indolent subset of CLL, i.e. patients with ultra-
stable disease for more than ten years, thus, clinically anal-
ogous to MBL. Moreover, in order to explore the possible
origin of genetic lesions at the hematopoietic progenitor
cell level, we analyzed polymorphonuclear (PMN) cells
from the study participants.
We report that the genomic profiles of ultra-stable CLL
patients are very similar to individuals with LC-MBL and
HC-MBL, characterized by infrequent CLL driver gene
mutations that, however, were not associated with dis-
ease progression. Furthermore, we observed non-coding
variants (NCVs) that target key pathways/cellular process-
es relevant to normal and neoplastic B-cell development,
thus, potentially contributing to the leukemic transforma-
tion. We also found shared somatic mutations between
MBL/CLL and PMN cells, strengthening the notion that at
least a proportion of somatic mutations may occur before
the onset of CLL.
Methods
The research protocol was approved by the Institutional Ethics
Committee and all participants gave written informed consent in
accordance with the Declaration of Helsinki.
Study population
The study cohort comprised 9 subjects with LC-MBL, 13 sub-
jects with HC-MBL, and 7 patients with Rai stage 0 CLL, herein
called ‘ultra-stable’ CLL. Detailed information about the study
cohort is provided in the Online Supplementary Appendix.
Cell samples
Chronic lymphocytic leukemia cells were stained with anti-
CD19, anti-CD5 and anti-CD20 antibodies. CD19+CD5+CD20dim
cells were sorted using a High Speed FACS Sorter MoFLo
(Beckman Coulter) according to previously published methods.26
PMN cells were sorted based on physical parameters. Buccal cells
were collected with the use of appropriate buccal swabs
(Epicentre, Madison, USA).
DNA extraction
The NucleoSpin® Tissue XS kit (Macherey-Nagel, Germany)
was used for DNA extraction in samples with less than 5x104 cells
and the QIAamp DNA Micro kit (Qiagen, Germany) in samples
with cell numbers ranging between 5x104 and 1x106. The QIAamp
866
DNA Blood Mini kit (Qiagen, Germany) was used for samples
with more than 1x106 cells as well as for the buccal samples. DNA
quantity and quality were assayed using the Qubit dsDNA HS
Assay Kit (Life Technologies, USA).
WGS: library preparation
The Nextera technology was utilized for the library construc-
tion (Nextera™ DNA Sample Prep Kit, Illumina, USA) as it
requires low input material whilst maintaining library complexity.
Fifty ng of genomic DNA were used for the construction of
libraries that were sequenced in paired-end mode 2x100bp on a
HiSeq 2000 (Illumina, USA).
A variant allele frequency (VAF) of 10% was used as threshold
for variant calling. More detailed information regarding the bioin-
formatics analysis is given in the Online Supplementary Appendix.
Targeted re-sequencing: library preparation
Probes targeting all coding exons or hotspot regions of 11
known or postulated CLL driver genes (ATM, BIRC3, MYD88,
NOTCH1, SF3B1, TP53, EGR2, POT1, NFKBIE, XPO1, FBXW7)
(Online Supplementary Table S1) were designed using Agilent’s
SureDesign service (https://earray.chem.agilent.com/suredesign/home.
htm). The target regions were captured using the HaloPlex HS tar-
geting enrichment kit (Agilent Technologies, USA). Paired-end
sequencing (150 bp reads) was performed on the NextSeq instru-
ment with the use of the 500/550 High Output Kit (Illumina,
USA).
Gene enrichment analysis
The identification of genes/gene pathways (gene enrichment
analysis, GEA) enriched within the targets of NCVs and motif-
breaking events caused by NCVs was performed with Enrichr31
using the KEGG 2016 gene database.
Results
WGS reveals highly similar mutational profiles in MBL
and ultra-stable CLL
Whole-genome sequencing was performed on 6 individ-
uals with LC-MBL, 5 individuals with HC-MBL, and 5
patients with ultra-stable CLL. For each individual/patient,
samples from MBL/CLL cells and PMN cells were evaluat-
ed against buccal (control) cells resulting in a total of 48
samples sequenced with an average autosomal coverage
of 32X (Online Supplementary Table S2). Basic demographic
and biological characteristics of the MBL/CLL cases
included in the WGS analysis are provided in Online
Supplementary Table S3. Overall, 37,033 somatic variants
were detected in MBL/CLL samples with an average of
2040 somatic variants in LC-MBL (range: 298-2871), 2558
in HC-MBL (range: 1428-3483), and 2400 in CLL (range:
1650-3176), respectively. Notably, 2792 variants were
identified in the 15 PMN control samples compared with
buccal DNA, with an average of 186 variants/sample (the
PMN sample from case CLL_3 was excluded from the
analysis due to tumor cell contamination) (Figure 1A).
Highly analogous mutation rates were observed in HC-
MBL and CLL (0.79 and 0.74 mutations per Mb, respec-
tively), while a slightly lower rate was seen in LC-MBL
(0.63 mutations per Mb); this latter finding was due to
sample LC-MBL_1 (excluding this sample, the average
mutation rate for LC-MBL would have been 0.74 muta-
tions per Mb) (Figure 1B). The ratio of single-nucleotide
variants (SNVs)/small indels was again almost identical in
haematologica | 2018; 103(5)
 WGS in MBL and ultra-stable CLL

the three entities (ranging from 11.4 to 12.6), whereas it
was significantly lower in the PMN samples (3.9 for LC-
MBL and CLL and 4.1 for HC-MBL, respectively) (P<0.005
for all cases) (Figure 1C).
The transition to transversion (Ti/Tv) ratio ranged from
0.99 to 1.13 in the MBL/CLL samples, while it was slightly
lower in the PMN samples (0.94) (Figure 2A). No clear dif-
ferences were observed between MBL/CLL and PMN
samples when the distribution of mutations among the six
types was examined (Figure 2B). We then evaluated the
sequence context of each mutation by incorporating infor-
mation on the bases immediately upstream and down-
stream of the mutated base, hence leading to 96 possible
mutation types in this classification16 (Online
Supplementary Table S4). Almost all major differences were
identified between MBL/CLL samples and PMN samples
with the former group exhibiting mainly C>T mutations
at NpCpG trinucleotides (P<0.05 for ultra-stable CLL and
HC-MBL).
The MutationalPatterns package33 was used to delineate
the mutational signatures in our cohort. Mutational pat-
terns identified in the MBL/CLL samples resembled those
reported by Puente et al.;9 this finding was corroborated by
calculating the pairwise similarity with the 30 previously
published signatures, where signature 9 and, to some
extent, signature 1 where the main contributors (Figure
2C). The same analysis in the PMN samples gave different
results, with a strong impact of mutational signatures 3
and 5 (Figure 2D). Signature 3 had previously been identi-
fied in solid tumors and is associated with failure of DNA
double-strand break-repair by homologous recombina-
tion.16 Signature 5 exhibits transcriptional strand bias for
C>T and T>C mutations at ApTpN context and displays
a correlation between smoking history and mutation con-
tribution.16
MBL and ultra-stable CLL display a paucity of
mutations in putative CLL driver genes
Whole-genome sequencing identified 186 non-synony-
mous exonic variants amongst MBL/CLL samples and 15
amongst PMN samples. The average number was 8.9 for
LC-MBL (range: 1-16), 14.8 for HC-MBL (range: 9-27),
11.6 for ultra-stable CLL (range: 7-19), and 0.9 for the
PMN samples (range: 0-6), respectively (Figure 3A). In
MBL/CLL samples, the vast majority of non-synonymous
mutations were missense [LC-MBL: 47 of 53 (88.7%); HC-

B C

Figure 1. Somatic mutational analysis of ultra-stable chronic lymphocytic leukemia (CLL), high-count monoclonal B-cell lymphocytosis (HC-MBL), low-count mon-
oclonal B-cell lymphocytosis (LC-MBL) and control polymorphonuclear (PMN) cell samples. (A) Total number of somatic mutations identified by whole-genome
sequencing (WGS) in CLL cell samples from MBL, CLL and the respective PMN samples. All samples carried similar mutational loads with the exception of a single
LC-MBL sample (LC-MBL_1) that displayed a very low number of mutation events; as can be seen, the corresponding PMN sample had a mutation load similar to
the other PMN samples, where comparison of mutation profiles between the MBL and PMN sample showed few common hits, thus excluding the likelihood of con-
tamination. Concerning PMN control samples, they were also characterized by high homogeneity regarding the mutational load. There was a single sample with a
very high mutational load; detailed comparison against its respective CLL sample showed a high overlap of mutations indicating potential tumor cell contamination,
hence this sample was removed from downstream analysis. (B) Average mutation rates ± Standard Deviation (SD) for LC-MBL, HC-MBL and CLL. Highly analogous
mutation rates were observed in the HC-MBL (0.79 mutations per Mb) and CLL (0.74 mutations per Mb) samples, while LC-MBL samples had a slightly lower ratio
(0.63 mutations per Mb). (C) Average SNV to small indels ratio ± SD for all sample groups. All 3 entities displayed similar ratios in clear contrast to the PMN samples
where the ratio was much lower.

haematologica | 2018; 103(5) 867

 A. Agathangelidis et al.

MBL: 61 of 73 (83.7%); ultra-stable CLL: 50 of 59
(84.7%)], whereas the remainder concerned either non-
sense mutations or frameshift indels (Figure 3B and Online
Supplementary Table S5). Forty-nine of the 186 mutations
(26.3%) had a VAF more than 50%. Concerning the 7
PMN samples harboring mutations, only a single mutation
(6.7%) had a VAF more than 50% (Figure 3C) (Online
Supplementary Table S6). The most commonly mutated
gene was IGLL5, in accordance with a recently reported
study,6 carrying mutations in 5 different samples (2 LC-
MBL, 2 HC-MBL and 1 CLL samples), likely introduced by
the somatic hypermutation (SHM) process. Only 6 of 186
mutations (1.6%) detected in the MBL/CLL samples con-
cerned putative CLL driver genes, according to 2 recently
reported lists.7,9 In detail, 3 were identified in individuals
with HC-MBL: i) a single NOTCH1 p.P2514Rfs*4 deletion
(VAF 20%), a known hotspot mutation in CLL10,28,34-36 in
HC-MBL_4; ii) a single FBXW7 p.W307S mutation (VAF
26%) in HC-MBL_2; and iii) a single KIAA0947 p.L2093X
(VAF 43%) in HC-MBL_5. Two mutations concerned indi-
viduals with LC-MBL: i) a KLHL6 p.A91D mutation (VAF
45%) in LC-MBL_5; and ii) a single CD79A p.E200G
mutation (VAF 53%) in LC-MBL_6. Finally, a CD79B
p.N68S mutation (VAF 41%) was identified in a single
CLL sample (CLL_5). Although most of these exact muta-
tions have not previously been reported in CLL, functional
prediction using Polyphen-2 classified all but the CD79B
mutation as probably damaging. No CLL driver gene
mutations were found in the PMN samples.
To assess whether the non-synonymous mutations
identified here might be potentially relevant to CLL, we
compared our findings to the variants reported by Puente
et al.9 and the International Cancer Genome Consortium
(ICGC) database.37 Overall, the vast majority of genes car-
rying mutations in our series were also reported as mutat-
ed in either or both datasets: 94% in LC-MBL, 89% in
HC-MBL, and 97% in CLL.
We extended our analysis by performing targeted re-
sequencing of 11 putative CLL driver genes in 8 LC-MBL,
13 HC-MBL and 7 ultra-stable CLL samples as well as 24
corresponding PMN samples. All but one LC-MBL case
(LC-MBL_4) subjected to WGS were included in this
analysis (Online Supplementary Table S7). In total, 5 variants
were detected in 3 different HC-MBL samples, including 4
missense variants and 1 frameshift deletion. Two variants
(targeting the NOTCH1 and FBXW7 genes) had been
already identified by WGS, whereas the remaining 3 con-
cerned the POT1 (n=2) and SF3B1 (n=1) genes. In detail, a
single HC-MBL case (HC-MBL_5) harbored an SF3B1
mutation (p.K700E; VAF 1.1%), a known hotspot muta-
tion in CLL,18,29,38,39 and a POT1 mutation (p.M1V; 4.3%),
while the other POT1 mutation (p.S38R; 6.7%) was found

A B

C D

Figure 2. Detailed analysis of mutation types. (A) Transition to transversion (Ti/Tv) ratios were comparable in all monoclonal B-cell lymphocytosis (MBL)/chronic lym-
phocytic leukemia (CLL) samples and somewhat lower in the polymorphonuclear (PMN) cell samples. (B) Similar distribution of mutations among the 6 mutation
classes for each MBL/CLL entity and PMN samples (average values ± Standard Deviation). Similar profiles were evident for all entities with the G>A mutation pre-
dominating in all cases. (C) Mutational signatures that contribute to the somatic mutations observed in the MBL/CLL samples. (D) Mutational signatures that dom-
inate in the PMN samples.

868 haematologica | 2018; 103(5)


in HC-MBL_2, which also harbored the FBXW7 mutation.
Two somatic non-synonymous variants were identified in
2 different PMN samples: an ATM variant (p.R337C; VAF
20%) detected in an LC-MBL case, frequently reported
and probably with a low functional impact, and a TP53
mutation (p.G245A, VAF 3%) found in a HC-MBL case,
previously reported in several human cancers and lym-
phomas40 (Online Supplementary Table S8).
Non-coding variants in MBL and ultra-stable CLL target
genes in pathways relevant to CLL pathogenesis
Coding and non-coding regions enriched for mutations
were detected using Fishhook29 in 10 kilobase windows
across the genome and with compensation for replication
timing. In line with previous findings,9 the analysis
revealed highest mutational enrichment in the IG loci and
within sites known to be recurrently affected by off-target
somatic hypermutation (e.g. BTG2, BCL6 and TCL1A)
(Online Supplementary Figure S1).
Funseq2,41 a bioinformatics tool investigating the link-
age between NCVs and target genes using integrated
bisulfite sequencing, ChIP-Sequencing, and RNA-sequenc-
ing data from the Roadmap Epigenomics Project, was
used for the examination of the NCVs. This analysis
revealed a total of 1517 variants in the MBL/CLL samples
and 39 in the PMN samples. After stringent filtering, 106
NCVs of potential relevance to MBL and CLL were iden-
tified (Online Supplementary Table S9): 29 in LC-MBL (aver-
age 4.8), 45 in HC-MBL (average 9), and 32 in CLL samples
(average 6.4), respectively; only 4 NCVs were found in 2
PMN samples. Since we intentionally selected for NCVs in
transcription factor (TF) highly occupied regions (see
Online Supplementary Appendix), not unexpectedly most
variants were located in gene promoter sites (Figure 4A).
Twenty-nine variants (26.4%) concerned 16 cancer-associ-
ated genes and were evenly distributed amongst the 3
entities: 9 in LC-MBL samples, 11 in HC-MBL, and 9 in
ultra-stable CLL samples. Three of these cancer-associated
genes were recurrently targeted: 9 variants concerned the
BTG2 gene in 4 samples (2 CLL, 1 HC-MBL and 1 LC-
MBL), 5 variants involved the BCL6 gene in 2 samples (1
HC-MBL and 1 LC-MBL), and 2 variants targeted the
BIRC3 gene in 2 samples (1 CLL and 1 HC-MBL). We also
identified 6 variants concerning the ST6GAL1 gene in 3
CLL samples and the same NKIRAS1-related variant in 2
CLL samples (Figure 4B). Moreover, pathway analysis
with Enrichr31 showed that 30 of 110 (27.3%) of the vari-
ants targeted genes were implicated in key CLL pathways
and cellular processes, such as the PI3K-AKT pathway
(TCL1A, CCND1, BCL2, PKN1, DDIT4 and SGK3)
(P<0.05), the NF-κB pathway (BIRC3, BCL2 and PLAU)
(P<0.05) and the spliceosome machinery (DDX46 and
HSPA2). In most of these cases (22 of 30, 73.3%), the vari-
ants were located at promoter sites. Comparison to the
series by Puente et al.9 identified 4 common gene targets:
BTG2, BCL6, BACH2 and TCL1A; none of our samples
carried variants affecting either the 3’ UTR of the
NOTCH1 gene or the PAX5 gene enhancer.
We also analyzed the predicted impact of the NCVs on
TF binding and found that 72 of 110 (65.5%) of the vari-
ants could result in a motif-breaking event (LC-MBL:
n=21, HC-MBL: n=33, ultra-stable CLL: n=18) (Online
Supplementary Table S10). We subsequently investigated
genes and gene pathways that may be affected by such TF
motif breaks. In 55 of 72 (76.4%) cases, variants disrupted
haematologica | 2018; 103(5)
WGS in MBL and ultra-stable CLL
a DNase I hypersensitive site, while enrichment analysis
of the implicated target genes using Enrichr31 led to the
identification of genes participating in pathways relevant
to CLL pathogenesis, such as the MAPK, WNT and AP-1
pathways (P<0.0005) (Online Supplementary Table S11).
Moreover, we examined the potential relation of the
NCVs that affect TF binding to AID activity by checking
if they occurred in the known hotspots (WRCY, RGYW,
WA, TW). According to our findings, 21 of 72 (29.2%)
NCVs were located at AID hotspots. Gene enrichment
analysis of the remaining 51 target genes revealed similar
pathways as in the original analysis (namely AP-1 and
DNA damage response pathways).
Shared mutations between CLL and polymorphonuclear
cells indicate that somatic variants can arise before
CLL onset
Shared mutations between MBL/CLL samples and their
respective PMN samples were identified in all samples
irrespective of origin. Regarding exonic mutations, the
same synonymous GSE1 mutation was found in an HC-
MBL case and its paired PMN sample with comparable
VAF (28% vs. 26%). In addition, a LC-MBL sample and its
paired PMN sample carried an identical mutation within
the ncRNA gene LOC339874, though with different VAF
(16% vs. 31%). In the case of non-exonic mutations, 179
shared NCVs were identified between MBL/CLL and
PMN samples (Online Supplementary Table S12); the aver-
age number per sample was 15.8 for LC-MBL, 8.2 for HC-
MBL, and 9 for ultra-stable CLL (range: 2-34), respectively.
Most of these mutations were intergenic (128 of 179,
71.5%) (Figure 4C). Interestingly, 6 NCVs were recurrent-
ly found in more than one MBL/CLL-PMN sample pair: 3
were intergenic and the other 3 were intronic (Online
Supplementary Table S13). Finally, we also examined the
mutational signatures for shared mutations between
MBL/CLL and PMN samples but did not observe clear cor-
relations with any signature (data not shown).
In order to exclude the possibility of contamination of
the PMN cell fraction by MBL/CLL DNA, we designed
allele-specific primers (Online Supplementary Table S14) and
performed PCR amplification of the clonotypic IGH gene
rearrangement in both the MBL/CLL and the respective
PMN samples in 11 of 16 cases with available material.
We identified the clonotypic rearrangement in all 11
MBL/CLL samples but in none of the corresponding PMN
samples examined, effectively ruling out the possibility
that the observed results were due to contamination.
Somatic copy-number analysis
sCNA analysis was performed in 3 samples from each
entity and their respective PMN control samples, as well
as in 4 additional PMN samples from 1 HC-MBL and 3
LC-MBL cases. In total, 16 sCNAs were identified in the
MBL/CLL samples (average: 1.8, range: 1-6): 7 in LC-MBL,
4 in HC-MBL, and 5 in the ultra-stable CLL samples, all
but one concerning deletion events (Online Supplementary
Table S15). Of the recurrent cytogenetic aberrations
included in the Döhner hierarchical model,42 del(17p),
del(11q) and trisomy 12 were not identified in any of the
samples, whereas del(13q) was detected in 7 of 9
MBL/CLL cases (2 LC-MBL, 3 HC-MBL and 2 CLL cases).
FISH analysis gave concordant results in 5 of 7 cases
where data from both techniques were available; in the
remaining 2 cases del(13q) was detected with a single
869
 A. Agathangelidis et al.
A B
technique each (Online Supplementary Table S16). All other
sCNAs represented unique events. In terms of distribution
across the chromosomes, 7 of the 32 (21.9%) sCNAs were
found in the vicinity of centromeres, whereas 11 of 32
(34.4%) were located close to a telomere (distance
<10x107 bp). None of the PMN samples demonstrated
sCNAs typical of CLL. Only one MBL case showed a
shared del(8)(p11.22) between the LC-MBL sample and its
paired PMN sample.
Discussion
Limited information is available concerning the genomic
landscape at the very early or indolent phases of CLL. To
this end, we compared the genomes of ultra-stable CLL
cases, defined as those cases stable for more than ten years
after diagnosis, to genomes from individuals with: i) LC-
MBL, a condition that does not progress into a clinically
relevant leukemia;26 and, ii) HC-MBL, a clinically identifi-
able pre-leukemic state.25
Both types of MBL and ultra-stable CLL exhibited the
same low level of genomic complexity, similar genome-
wide mutation rates, and average number of exonic muta-
tions, which were distinct from those of the control sam-
ples. Reflecting this similarity, analysis relating to pub-
lished mutational signatures revealed similar patterns in
samples from all 3 entities. In more detail, signature 9 that
predominated in the MBL/CLL cohort has been previous-
ly identified in CLL and B-cell lymphomas and is attrib-
uted to polymerase η that is involved in AID-induced
870
Figure 3. Exonic mutations in our monoclonal B-
cell lymphocytosis (MBL)/chronic lymphocytic
leukemia (CLL) cohort and polymorphonuclear
(PMN) cell samples. (A) Average numbers of
exonic non-synonymous mutations in MBL/CLL
entities and PMN samples. (B) The vast majority
of non-synonymous mutations were missense in
all 3 MBL/CLL entities. PMN samples failed to
show such predominance. (C) Average numbers
of mutations with VAF≥50% for all 3 MBL/CLL
entities were comparable. A single PMN sample
carried a clonal mutation.
somatic hypermutation.16 The second ranking signature 1
is an age-related signature stemming from spontaneous
deamination of 5-methylcytosine that has been detected
in many cancer types.16 Analogies between MBL and ultra-
stable CLL extended also to sCNAs in that all samples,
irrespective of origin, carried very few sCNA. Del(13q)
predominated in all three entities, as shown in previous
studies.26 Most of the sCNAs were located in close prox-
imity to either centromeres or telomeres, in keeping with
previous findings reporting significant over-representa-
tions in these regions due to duplication rates.43 Thus,
most of the sCNAs identified here may not be directly
related to the MBL/CLL phenotype.
Interestingly, PMN cells harbored a significantly higher
load of mutations compared to buccal cells. Mutations
detected in the PMN samples were characterized by the
dominance of distinct mutational signatures compared to
the MBL/CLL cohort. However, these samples carried
shared somatic mutations with the respective MBL/CLL
cell samples in all analyzed cases. Most shared mutations
concerned intergenic regions, yet we also identified a sin-
gle shared exonic mutation. This finding supports the
notion that some mutations present in the CLL clone
could be acquired prior to disease onset, as previously sug-
gested.32
Almost all genes that were found mutated in HC-MBL
and/or LC-MBL had been previously described as recur-
rently mutated in CLL.9,37 In contrast to our recent WES
study on relapsing CLL,8 where the great majority of cases
carried at least one CLL driver mutation, such mutations
were relatively scarce in our cohort. Most importantly, the
haematologica | 2018; 103(5)
 WGS in MBL and ultra-stable CLL

A B

Figure 4. Analysis of non-coding variants and shared mutations between monoclonal B-cell lymphocytosis (MBL)/chronic lymphocytic leukemia (CLL) and poly-
morphonuclear (PMN) cell samples in the present cohort. (A) Topology of the 106 relevant non-coding variants identified in the present study. The majority con-
cerned gene promoter sites. (B) Recurrent non-coding variants in genes relevant to CLL. The BIRC3, BCL6 and BTG2 genes are known to be associated with various
types of cancer. (C) Topology of shared mutations between MBL/CLL and PMN samples. Intergenic mutations predominated, followed by intronic mutations.

lack of any obvious impact of the identified mutations on
disease progression after a prolonged follow up highlights
the fact that the mere presence of a given driver mutation
does not axiomatically equate with disease progression, as
previously reported.31,44 Additional studies are required in
order to clarify this phenomenon.
Chronic lymphocytic leukemia cases have been shown
to harbor detrimental gene mutations in subclones not vis-
ible at diagnosis that are progressively selected, e.g. fol-
lowing the use of chemotherapy.45,46 Such mutations were
identified by targeted re-sequencing, yet only in a minor
fraction of the present cohort. In this context, it has been
recently proposed that sequence depths greater than
4000X will be essential in order to robustly identify all
subclonal mutations and predict aggressive cases.12
Arguably, the absence of such driver mutations when
applying highly sensitive methods may potentially help to
identify individuals with very indolent disease for whom
less frequent, if any follow up will be warranted.
Whole-genome sequencing revealed that the non-cod-
ing mutome commonly targets gene pathways and cellu-
lar processes involved in CLL pathobiology. A sizeable
proportion of variants affected the promoter sites of genes
previously associated with cancer, e.g. BTG2, BCL6,
BACH2 and TCL1A. Interestingly, all 4 genes have been
recently reported as non-IG targets of the SHM process in
patients with lymphomas,47 implicating this otherwise
normal process in the emergence of CLL-like clones.
Additional studies need to be performed to address the
relevance of such mutations; however, since AID-related
mutations are common in CLL,16 they may indeed be rel-
evant to disease pathogenesis. We did not identify any
“poor-prognostic” 3’ UTR NOTCH1 mutations;9 howev-
er, we did discover two NCVs targeting indirectly the
BIRC3 gene. We also found a number of variants within
the promoter sites of genes implicated in pathways rele-
vant to CLL biology, including the PI3K-AKT and NF-κB
pathways as well as the spliceosome machinery.
Furthermore, we identified TF binding motif breaking
events that may arise due to NCVs; most concerned the

haematologica | 2018; 103(5) 871

 A. Agathangelidis et al.

MAPK, WNT and AP-1 signaling pathways. In this con-
text, preliminary results (data not shown) from our ongoing
high-throughput study on aggressive CLL cases showed a
great degree of consistency in the targeting of NCVs: the
same “CLL-relevant” gene pathways were again among
the most common targets of NCVs, further corroborating
our present findings. Having said that, some of these vari-
ants could represent bystander SHM targets of unknown
significance or minor contributors to disease pathogenesis,
therefore requiring further studies before definitive con-
clusions can be drawn regarding their actual significance.
It is important to note that a recent study on the epige-
netic profile of CLL48 reported a novel pathogenic role of
TF dysregulation in CLL, with increased activity of EGR
and NFAT as well as loss of EBF and AP-1, causing imbal-
ances in the normal B-cell epigenetic program.
Interestingly, certain members of these networks (e.g.
EBF1, JUN and FOS) were among the most commonly
affected TFs across all sample types tested. Collectively,
our findings support the notion that gene pathways could
be indirectly targeted by NCVs with the targets being
either the genes themselves or other interacting genes, e.g.
TFs.
Limitations of the present work involve the relatively
small size of the cohort, mainly due to the rarity of sam-
ples meeting the selection criteria. In particular, CLL
patients had to have stable disease after a prolonged fol-
low up, whereas all individuals with MBL had to have a
persistent monoclonal B-cell population. Concerning LC-
MBL, low CLL cell number was an additional challenging
factor. Furthermore, although our targeted re-sequencing
approach covered almost 50% of reported mutations in
putative driver genes (as reported by Puente et al.),9 by def-
inition this approach is not exhaustive.
In summary, we report that MBL and ultra-stable CLL
are virtually indistinguishable at the genomic level. While
this may be reflective of a passive and slow accumulation
of mutations, we identified both exonic and NCV-targeted
pathways central for B-cell biology and CLL development,
likely linked to the acquisition of the MBL/CLL pheno-
type. Importantly, ultra-stable CLL cases carried few
known driver gene mutations, even after ten years of fol-
low up, perhaps reflecting the central role of microenvi-
ronmental signals rather than cell-intrinsic defects in shap-
ing clonal behavior. In other words, cell-extrinsic trigger-
ing, specifically mediated through the B-cell receptor,
might represent the major driving force in the early stages
of CLL, whereas disease progression will require acquisi-
tion of genetic driver mutations.
Funding
This research project was supported by the Associazione
Italiana per la Ricerca sul Cancro, AIRC (Investigator Grant
#15189 to PG and Special Program Molecular Clinical
Oncology – 5 per mille #9965), Milano, Italy, Ricerca
Finalizzata 2010 (#2318823 to PG); Swedish Cancer Society,
the Swedish Research Council, Uppsala University, Uppsala
University Hospital, Lion’s Cancer Research Foundation, and
Selander’s Foundation, Uppsala; and, H2020 “AEGLE, An
analytics framework for integrated and personalized healthcare
services in Europe”, by the European Union; “MEDGENET,
Medical Genomics and Epigenomics Network” (No.692298) by
the European Union; “GCH-CLL” funded by the General
Secretariat for Research and Technology (GSRT) of Greece and
the Italian Ministry of Health (MoH); and IMI2 “HARMO-
NY”, funded by the European Union. AA is a fellow of
Associazione Italiana per la Ricerca sul Cancro AIRC (Triennial
fellowship “Guglielmina Lucatello é Gino Mazzega”).

8. Ljungstrom V, Cortese D, Young E, et al. lymphocytic leukemia. N Engl J Med.

References
1. Chiorazzi N, Rai KR, Ferrarini M. Chronic
lymphocytic leukemia. N Engl J Med.
2005;352(8):804-815.
2. Parikh SA, Shanafelt TD. Prognostic factors
and risk stratification in chronic lympho-
cytic leukemia. Semin Oncol. 2016;
43(2):233-240.
3. Scarfo L, Ferreri AJ, Ghia P. Chronic lym-
phocytic leukaemia. Crit Rev Oncol
Hematol. 2016;104:169-182.
4. Fabbri G, Khiabanian H, Holmes AB, et al.
Genetic lesions associated with chronic
lymphocytic leukemia transformation to
Richter syndrome. J Exp Med.
2013;210(11):2273-2288.
5. Fabbri G, Rasi S, Rossi D, et al. Analysis of
the chronic lymphocytic leukemia coding
genome: role of NOTCH1 mutational acti-
vation. J Exp Med. 2011;208(7):1389-1401.
6. Kasar S, Kim J, Improgo R, et al. Whole-
genome sequencing reveals activation-
induced cytidine deaminase signatures dur-
ing indolent chronic lymphocytic
leukaemia evolution. Nat Commun. 2016;
6(8866).
7. Landau DA, Tausch E, Taylor-Weiner AN,
et al. Mutations driving CLL and their evo-
lution in progression and relapse. Nature.
2015;526(7574):525-530.
Whole-exome sequencing in relapsing
chronic lymphocytic leukemia: clinical
impact of recurrent RPS15 mutations.
Blood. 2016;127(8):1007-1016.
9. Puente XS, Bea S, Valdes-Mas R, et al. Non-
coding recurrent mutations in chronic lym-
phocytic leukaemia. Nature. 2015;
526(7574):519-524.
10. Puente XS, Pinyol M, Quesada V, et al.
Whole-genome sequencing identifies recur-
rent mutations in chronic lymphocytic
leukaemia. Nature. 2011;475(7354):101-
105.
11. Quesada V, Conde L, Villamor N, et al.
Exome sequencing identifies recurrent
mutations of the splicing factor SF3B1 gene
in chronic lymphocytic leukemia. Nat
Genet. 2012;44(1):47-52.
12. Rose-Zerilli MJ, Gibson J, Wang J, et al.
Longitudinal copy number, whole exome
and targeted deep sequencing of 'good risk'
IGHV-mutated CLL patients with progres-
sive disease. Leukemia. 2016;30(6):1301-
1310.
13. Schuh A, Becq J, Humphray S, et al.
Monitoring chronic lymphocytic leukemia
progression by whole genome sequencing
reveals heterogeneous clonal evolution pat-
terns. Blood. 2012;120(20):4191-4196.
14. Wang L, Lawrence MS, Wan Y, et al. SF3B1
and other novel cancer genes in chronic
2011;365(26):2497-2506.
15. Burns A, Alsolami R, Becq J, et al. Whole-
genome sequencing of chronic lymphocyt-
ic leukaemia reveals distinct differences in
the mutational landscape between
IgHV(mut) and IgHV(unmut) subgroups.
Leukemia. 2018;32(2):332-342.
16. Alexandrov LB, Nik-Zainal S, Wedge DC,
et al. Signatures of mutational processes in
human cancer. Nature. 2013;
500(7463):415-421.
17. Landau DA, Carter SL, Stojanov P, et al.
Evolution and impact of subclonal muta-
tions in chronic lymphocytic leukemia.
Cell. 2013;152(4):714-726.
18. Rossi D, Bruscaggin A, Spina V, et al.
Mutations of the SF3B1 splicing factor in
chronic lymphocytic leukemia: association
with progression and fludarabine-refrac-
toriness. Blood. 2011;118(26):6904-6908.
19. Rossi D, Fangazio M, Rasi S, et al.
Disruption of BIRC3 associates with flu-
darabine chemorefractoriness in TP53
wild-type chronic lymphocytic leukemia.
Blood. 2012;119(12):2854-2862.
20. Schnaiter A, Paschka P, Rossi M, et al.
NOTCH1, SF3B1, and TP53 mutations in
fludarabine-refractory CLL patients treated
with alemtuzumab: results from the
CLL2H trial of the GCLLSG. Blood.
2013;122(7):1266-1270.

872 haematologica | 2018; 103(5)


21. Landgren O, Albitar M, Ma W, et al. B-cell
clones as early markers for chronic lym-
phocytic leukemia. N Engl J Med. 2009;
360(7):659-667.
22. Marti GE, Rawstron AC, Ghia P, et al.
Diagnostic criteria for monoclonal B-cell
lymphocytosis. Br J Haematol. 2005;
130(3):325-332.
23. Dagklis A, Fazi C, Scarfo L, et al.
Monoclonal B lymphocytosis in the general
population. Leuk Lymphoma. 2009;
50(3):490-492.
24. Ghia P, Prato G, Scielzo C, et al.
Monoclonal CD5+ and CD5- B-lympho-
cyte expansions are frequent in the periph-
eral blood of the elderly. Blood. 2004;
103(6):2337-2342.
25. Rawstron AC, Bennett FL, O'Connor SJ, et
al. Monoclonal B-cell lymphocytosis and
chronic lymphocytic leukemia. N Engl J
Med. 2008;359(6):575-583.
26. Fazi C, Scarfo L, Pecciarini L, et al. General
population low-count CLL-like MBL per-
sists over time without clinical progression,
although carrying the same cytogenetic
abnormalities of CLL. Blood. 2011;
118(25):6618-6625.
27. Matos DM, Furtado FM, Falcao RP.
Monoclonal B-cell lymphocytosis in indi-
viduals from sporadic (non-familial) chron-
ic lymphocytic leukemia families persists
over time, but does not progress to chronic
B-cell lymphoproliferative diseases. Rev
Bras Hematol Hemoter. 2015;37(5):292-
295.
28. Rasi S, Monti S, Spina V, et al. Analysis of
NOTCH1 mutations in monoclonal B-cell
lymphocytosis. Haematologica. 2012;
97(1):153-154.
29. Imielinski M, Guo G, Meyerson M.
Insertions and Deletions Target Lineage-
Defining Genes in Human Cancers. Cell.
2017;168(3):460-472 e414.
30. Kreutzer DA, Essigmann JM. Oxidized,
haematologica | 2018; 103(5)
deaminated cytosines are a source of C -->
T transitions in vivo. Proc Natl Acad Sci
USA. 1998;95(7):3578-3582.
31. Kuleshov MV, Jones MR, Rouillard AD, et
al. Enrichr: a comprehensive gene set
enrichment analysis web server 2016
update. Nucleic Acids Res. 2016;
44(W1):W90-97.
32. Damm F, Mylonas E, Cosson A, et al.
Acquired initiating mutations in early
hematopoietic cells of CLL patients. Cancer
Discov. 2014;4(9):1088-1101.
33. Blokzijl FJ, R.; van Boxtel, R.; Cuppen, E.
MutationalPatterns: comprehensive
genome-wide analysis of mutational
processes. bioRxiv 071761; doi: https://
doi.org/10.1101/071761.
34. Baliakas P, Hadzidimitriou A, Sutton LA, et
al. Recurrent mutations refine prognosis in
chronic lymphocytic leukemia. Leukemia.
2015;29(2):329-336.
35. Sutton LA, Young E, Baliakas P, et al.
Different spectra of recurrent gene muta-
tions in subsets of chronic lymphocytic
leukemia harboring stereotyped B-cell
receptors. Haematologica. 2016;101(8):959-
967.
36. Lionetti M, Fabris S, Cutrona G, et al. High-
throughput sequencing for the identifica-
tion of NOTCH1 mutations in early stage
chronic lymphocytic leukaemia: biological
and clinical implications. Br J Haematol.
2014;165(5):629-639.
37. Zhang J, Baran J, Cros A, et al. International
Cancer Genome Consortium Data Portal--a
one-stop shop for cancer genomics data.
Database (Oxford). 2011;2011:bar026.
38. Strefford JC, Sutton LA, Baliakas P, et al.
Distinct patterns of novel gene mutations
in poor-prognostic stereotyped subsets of
chronic lymphocytic leukemia: the case of
SF3B1 and subset #2. Leukemia. 2013;
27(11):2196-2199.
39. Jeromin S, Weissmann S, Haferlach C, et al.
WGS in MBL and ultra-stable CLL
SF3B1 mutations correlated to cytogenetics
and mutations in NOTCH1, FBXW7,
MYD88, XPO1 and TP53 in 1160 untreated
CLL patients. Leukemia. 2014;28(1):108-
117.
40. Wilda M, Bruch J, Harder L, et al.
Inactivation of the ARF-MDM-2-p53 path-
way in sporadic Burkitt's lymphoma in
children. Leukemia. 2004;18(3):584-588.
41. Fu Y, Liu Z, Lou S, et al. FunSeq2: a frame-
work for prioritizing noncoding regulatory
variants in cancer. Genome Biol. 2014;
15(10):480.
42. Dohner H, Stilgenbauer S, Benner A, et al.
Genomic aberrations and survival in chron-
ic lymphocytic leukemia. N Engl J Med.
2000;343(26):1910-1916.
43. Nguyen DQ, Webber C, Ponting CP. Bias of
selection on human copy-number variants.
PLoS Genet. 2006;2(2):e20.
44. Hurtado AM, Chen-Liang TH,
Przychodzen B, et al. Prognostic signature
and clonality pattern of recurrently mutat-
ed genes in inactive chronic lymphocytic
leukemia. Blood Cancer J. 2015;5:e342.
45. Ojha J, Ayres J, Secreto C, et al. Deep
sequencing identifies genetic heterogeneity
and recurrent convergent evolution in
chronic lymphocytic leukemia. Blood.
2015;125(3):492-498.
46. Rossi D, Khiabanian H, Spina V, et al.
Clinical impact of small TP53 mutated sub-
clones in chronic lymphocytic leukemia.
Blood. 2014;123(14):2139-2147.
47. Khodabakhshi AH, Morin RD, Fejes AP, et
al. Recurrent targets of aberrant somatic
hypermutation in lymphoma. Oncotarget.
2012;3(11):1308-1319.
48. Oakes CC, Seifert M, Assenov Y, et al.
DNA methylation dynamics during B cell
maturation underlie a continuum of disease
phenotypes in chronic lymphocytic
leukemia. Nature Genetics. 2016;48(3):253-
264.
873
ARTICLE Chronic Lymphocytic Leukemia

Toxicities and outcomes of 616

Ferrata Storti
Foundation ibrutinib-treated patients in the United States:
a real-world analysis
Anthony R. Mato,1 Chadhi Nabhan,2 Meghan C. Thompson,1 Nicole Lamanna,3
Danielle M. Brander,4 Brian Hill,5 Christina Howlett,6,7 Alan Skarbnik,7 Bruce D.
Cheson,8 Clive Zent,9 Jeffrey Pu,10 Pavel Kiselev,11 Andre Goy,7 David Claxton,10
Krista Isaac,12 Kaitlin H. Kennard,1 Colleen Timlin,1 Daniel Landsburg,1 Allison
Winter,5 Sunita D. Nasta,1 Spencer H. Bachow,3 Stephen J. Schuster,1 Colleen
Dorsey,1 Jakub Svoboda,1 Paul Barr13* and Chaitra S. Ujjani8*
Haematologica 2018
Hematology and Oncology, University of Pennsylvania, Philadelphia, PA; 2Cardinal Health,
Volume 103(5):874-879
1

Dublin, OH 3Hematology/Oncology, Presbyterian/Columbia University Medical Center, New

York, NY; 4Hematology/Oncology, Duke University, Durham, NC; 5Hematology and Medical
Oncology, Cleveland Clinic, OH; 6Pharmacy, Ernest Mario School of Pharmacy, New
Brunswick, NY; 7Hematology/Oncology, John Theurer Cancer Center, Hackensack, NY;
8
Hematology/Oncology, Georgetown University Hospital, Washington DC;
9
Hematology/Oncology, University of Rochester Medical Center, NY; 10Hematology/Oncology,
Penn State Milton S Hershey Medical Center, PA; 11Lymphoma, Celgene Corp, Summit, NY;
12
Hematology/Oncology, Lankenau Hospital, Wynnewood, PA and 13Division of Hematology
and Oncology, University of Rochester, NY, USA
*Indicates shared last authorship

ABSTRACT

Correspondence:
amato@mskcc.org
Received: October 20, 2017.
Accepted: January 26, 2018.
Pre-published: February 1, 2018.
doi:10.3324/haematol.2017.182907
Check the online version for the most updated
information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/874
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
C
linical trials that led to ibrutinib’s approval for the treatment of
chronic lymphocytic leukemia showed that its side effects differ
from those of traditional chemotherapy. Reasons for discontinua-
tion in clinical practice have not been adequately studied. We conducted a
retrospective analysis of chronic lymphocytic leukemia patients treated
with ibrutinib either commercially or on clinical trials. We aimed to com-
pare the type and frequency of toxicities reported in either setting, assess
discontinuation rates, and evaluate outcomes. This multicenter, retrospec-
tive analysis included ibrutinib-treated chronic lymphocytic leukemia
patients at nine United States cancer centers or from the Connect®
Chronic Lymphocytic Leukemia Registry. We examined demographics,
dosing, discontinuation rates and reasons, toxicities, and outcomes. The
primary endpoint was progression-free survival. Six hundred sixteen ibru-
tinib-treated patients were identified. A total of 546 (88%) patients were
treated with the commercial drug. Clinical trial patients were younger
(mean age 58 versus 61 years, P=0.01) and had a similar time from diagnosis
to treatment with ibrutinib (mean 85 versus 87 months, P=0.8). With a
median follow-up of 17 months, an estimated 41% of patients discontin-
ued ibrutinib (median time to ibrutinib discontinuation was 7 months).
Notably, ibrutinib toxicity was the most common reason for discontinua-
tion in all settings. The median progression-free survival and overall sur-
vival for the entire cohort were 35 months and not reached (median fol-
low-up 17 months), respectively. In the largest reported series on ibrutinib-
treated chronic lymphocytic leukemia patients, we show that 41% of
patients discontinued ibrutinib. Intolerance as opposed to chronic lympho-
cytic leukemia progression was the most common reason for discontinua-
tion. Outcomes remain excellent and were not affected by line of therapy
or whether patients were treated on clinical studies or commercially. These
data strongly argue in favor of finding strategies to minimize ibrutinib
intolerance so that efficacy can be further maximized. Future clinical trials
should consider time-limited therapy approaches, particularly in patients
achieving a complete response, in order to minimize ibrutinib exposure.

874 haematologica | 2018; 103(5)


Introduction
Ibrutinib is an orally bioavailable, irreversible inhibitor
of Bruton tyrosine kinase (BTK). It is a standard of care in
the treatment of chronic lymphocytic leukemia (CLL) in
the relapsed/refractory1-3 as well as front-line4 settings.
The toxicities of ibrutinib and reasons for its discontin-
uation were initially defined through several landmark
studies comparing ibrutinib to chlorambucil (front-line,
RESONATE 2),4 ofatumumab (RESONATE)2 and ben-
damustine and rituximab +/- ibrutinib versus placebo
(HELIOS).5 Higher proportions of patients discontinuing
therapy (ranging from 25% to 51%) have been found with
long-term follow-up (median follow-ups ranging between
20 months and approximately 5 years) of these and other
ibrutinib clinical trials. Although the percentage of discon-
tinuations due to progression of disease (21% to 45%)
remained similar, there were high proportions of discon-
tinuations due to adverse events, ranging from 12% to
32%.6-10
Woyach et al. suggested that discontinuation due to CLL
progression (distinguished from disease transformation)
occurred later in the disease course (cumulative incidence
of 7.3% at 2 years, 19.1% at 4 years) while other reasons
for discontinuation, such as Richter transformation,
peaked early and reached a plateau by 3 years (18.7% at 2
years, 23.9% at 3 years).8 In clinical practice, adverse
events were found to be the most common cause of ibru-
tinib discontinuation among 143 patients, approaching
50%.11 Atrial fibrillation, infectious complications, and
cytopenias were the most commonly described adverse
events.11 High percentages of patients discontinuing thera-
py, ranging from 19-41%, were encountered in additional
studies conducted in the USA and outside of the USA;
however, these studies were limited by relatively small
sample sizes and short follow-up periods.12-15
We, therefore, aimed to characterize patterns of care
among ibrutinib-treated CLL patients in clinical practice in
the USA focusing on rates and reasons for discontinuation,
and how these affect outcomes. To our knowledge, this
series is the largest report on ibrutinib-treated CLL
patients.
Methods
We conducted a multicenter, retrospective cohort study of CLL
patients at nine USA cancer centers and from the Connect® CLL
Registry (199 USA centers, 80% community sites) who were
treated with ibrutinib either as part of a clinical trial or with com-
mercially available drug.16 The institutional review board of each
participating institution approved this study. Investigators at each
institution were asked to utilize chart review, institutional elec-
tronic medical records and clinical/pathological databases to
obtain required information for all CLL patients treated with ibru-
tinib. Data collected included: patients’ demographics, genetic
characteristics, number of prior therapies, dosing and dose adjust-
ments, discontinuation rates and reasons, toxicities and outcomes.
The period of enrollment was January 2014 to August 2016.
The primary study endpoint was progression-free survival,
which was defined as time from ibrutinib treatment to progres-
sion or death from any cause as per the Kaplan Meier method.17
Patients were otherwise censored, regardless of progression status,
at the time of last follow-up and at the time of next therapy. When
interpreting medical records, investigators were advised to use the
haematologica | 2018; 103(5)
Ibrutinib: treatment toxicities and outcomes
International Workshop on Chronic Lymphocytic Leukemia
(IWCLL) criteria to define response and progression of disease.18,19
Patients were stratified by line of therapy (front-line versus
relapsed/refractory), reason for discontinuation (intolerance versus
progressive disease), clinical trial participation versus commercial
use, and depth of response (complete response versus partial
response and partial response with lymphocytosis). Secondary
endpoints included overall survival and reasons for ibrutinib dis-
continuation. Overall survival was defined as the time in months
from initiation of ibrutinib to death.
Reasons for ibrutinib discontinuation were categorized as fol-
lows: toxicity, progressive disease, Richter transformation to
either diffuse large B-cell lymphoma or Hodgkin lymphoma,
planned cellular therapy (allogeneic hematopoietic stem cell trans-
plantation or chimeric antigen receptor genetically modified T-cell
therapy), secondary malignancies, physician’s or patient’s prefer-
ence, financial concerns, and other/unrelated death. Toxicities
leading to discontinuation were categorized as: hematologic toxi-
city, infection, atrial fibrillation, congestive heart failure, drug-
induced pneumonitis, drug-induced colitis, transaminitis, bleed-
ing, arthralgia/myalgia, dermatological, neurotoxicity, other, or
unknown.
Survival data were compared using the log-rank test.20
Univariate Cox regression analyses were used to estimate hazard
ratios.21 All other comparison analyses were descriptive. All tests
were two-sided at the 5% level. Statistical analyses were per-
formed using STATA 10.1 (Stata Statistical Software: Release 10.
2007; StataCorp LP, College Station, TX, USA).
Results
Patients
We identified 616 patients who received ibrutinib,
including 536 relapsed-refractory and 80 previously
untreated patients. Data from the nine contributing aca-
demic centers were collected retrospectively and included
information for 399 patients treated with ibrutinib; data
from the Connect CLL registry were collected prospec-
tively and included information on 217 patients largely
collected from community sites (80% community). The
patients’ baseline characteristics stratified by line of thera-
py are available in Table 1. A total of 546 (88%) patients
were treated with commercially available drug/off study.
Clinical trial patients were younger (mean age 58 versus 61
years, P=0.01), had a similar time from diagnosis to treat-
ment with ibrutinib (mean 85 versus 87 months, P=0.8)
and were more consistently initiated at a dose of 420 mg
daily (100% versus 89%).
Reasons for discontinuation, toxicities and timing
of events
At a median follow-up of 17 months (range, 1-60
months), 41% of patients discontinued ibrutinib. The
median time to ibrutinib discontinuation was 7 months
(range, 0.1–41), with a median time of 6 months for
patients who discontinued due to intolerance and 10
months for those who discontinued due to progression of
disease (Online Supplementary Figure S1). Among patients
on ibrutinib monotherapy, 41% discontinued therapy
while the percentage of discontinuation among patients
receiving ibrutinib-based combination therapy was
43.9%. Ibrutinib starting dose (420 mg daily versus <420
mg daily) did not correlate with the proportion of patients
who discontinued ibrutinib due to toxicity (51% versus
875
 A.R. Mato et al.

50%) or disease progression (19.6% versus 21.4%).
Reasons for ibrutinib discontinuation are listed in Table
2. Percentages listed indicate the proportion of discontin-
uations due to each category. Toxicity was the most com-
mon reason for discontinuation in all settings, accounting
for 63.1% of discontinuations in front-line use (n=12/80
front-line patients) and 50.2% of discontinuations in
relapsed/refractory use (n=116/536 relapsed/refractory
patients). Toxicity was the most common reason for dis-
continuation in several settings including: commercial use
and clinical trial use (50% of discontinuations in front-line
commercial use, 77.7% of discontinuations in front-line
clinical trial use, 52.5% of discontinuations in
relapsed/refractory commercial use, and 39.7% of discon-
tinuations in relapsed/refractory trial use). Notably, the
proportion of discontinuations due to progressive disease
was lower: 15.8% in the front-line setting and 20.9% in
relapsed/refractory use. Richter transformation to diffuse
large B-cell lymphoma or Hodgkin lymphoma accounted
for 5.3% of the discontinuations in the front-line setting
and 5.0% in the relapsed/refractory setting.
Among the patients treated front-line with ibrutinib, the
three most common toxicities leading to discontinuation
were arthralgia (41.6%), atrial fibrillation (25%), and rash
(16.7%). In the relapsed/refractory population, the most
common toxicities leading to discontinuation were atrial
fibrillation (12.3%), infection (10.7%), pneumonitis
(9.9%), bleeding (9%) and diarrhea (6.6%).
The median time to ibrutinib discontinuation varied by
toxicity: bleeding (8 months), diarrhea (7.5 months), atrial
fibrillation (7 months), infection (6 months), arthralgia (5
months), pneumonitis (4.5 months) and rash (3.5 months).
Outcomes
At a median follow-up of 17 months, the median pro-
gression-free and overall survival for the entire cohort
were 35 months and not reached, respectively (Figure
1A,B). Overall survival from the start of ibrutinib therapy,

Table 1. Baseline characteristics.

Baseline characteristics
Ibrutinib in front line Ibrutinib in relapse
Total number 80 536
Median age at diagnosis, years (range) 62 (37-88), n=78 60 (22-95), n=532
Median time from diagnosis to ibrutinib start, months (range) 25.5 (1-232), n=80 78.3 (1-366), n=536
del17p(+) 37%, n=76 26%, n=368
del11q(+) 19%, n=75 35%, n=367
Tp53mut(+) 12%, n=42 13%, n=142
Complex karyotype (+) (>3) 40%, n=60 34%, n=214
Median time diagnosis to ibrutinib, months (range) 26 (1-232) 78 (1-660)
Median ibrutinib starting dose 420 mg 420 mg
Ibrutinib administered as monotherapy 68%, n=80 89%, n=536
Ibrutinib suspension required 30%, n=79 37%, n=310
Ibrutinib dose adjusted 15%, n=79 20%, n=309
Median follow-up 17 months
del17p(+) (deletion 17p mutation positive); del11q(+) (deletion 11q mutation positive); Tp53mut(+) (Tp53 mutation positive); Complex karyotype (+)(>3) (complex karyotype
positive defined as three or more abnormalities).

A B

Figure 1. Outcomes for the entire cohort. Kaplan Meier curves at a median follow-up of 17 months showing (A) progression-free survival (PFS) for the entire cohort
and (B) overall survival (OS) for the entire cohort.

876 haematologica | 2018; 103(5)


stratified by whether the drug was being used in the
front-line versus relapsed/refractory setting, is shown in
Online Supplementary Figure S2A,B. Notably, there was no
significant difference in progression-free survival by
front-line versus relapsed/refractory use (P=0.27, log-rank
test) (Figure 2A) or at first, second or third relapse
(P=0.45) (Online Supplementary Figure S3). Progression-free
survival was similar when stratified by ibrutinib use in
the clinical practice setting as compared to the clinical
trial setting (P=0.14, log-rank test) (Figure 2B). Patients
who discontinued due to toxicity had significantly longer
progression-free survival and overall survival than those
who discontinued due to disease progression (P=0.01 and
P=0.02, respectively, log-rank test) (Figure 2C,D).
A B
C D
haematologica | 2018; 103(5)
Ibrutinib: treatment toxicities and outcomes
Investigator-assessed depth of response (complete
response versus partial response versus partial response
with lymphocytosis versus stable disease versus progres-
sive disease) appeared to correlate with a longer progres-
sion-free survival (Figure 2E). We also stratified progres-
sion-free survival by deletion 17p status and complex
karyotype status (≥3 abnormalities) in CLL patients treat-
ed in the relapsed/refractory setting. Progression-free sur-
vival was not significantly different in patients with dele-
tion 17p (P=0.70), but was significantly shorter in patients
with a complex karyotype (hazard ratio=1.8, 95% confi-
dence interval: 1.1-3.0, P=0.01). The Kaplan Meier curves
for these analyses are shown in Online Supplementary
Figure S4A-C.
Figure 2. Outcomes stratified by line of therapy, clinical trial participation,
reason for discontinuation and depth of response. Kaplan Meier curves
showing outcomes stratified by (A) line of therapy (progression-free survival),
(B) clinical trial participation (progression-free survival), (C) reason for dis-
continuation (progression-free survival), (D) reason for discontinuation (over-
all survival), and (E) depth of response. CR: complete response; PR: partial
response; PR-L: partial response with lymphocytosis; SD: stable disease; PD:
progressive disease.
877
 A.R. Mato et al.
Discussion
In the largest reported series of ibrutinib-treated CLL
patients so far, we found that the median progression-free
survival was 35 months. Interestingly, this outcome was
comparable between previously untreated and
relapsed/refractory patients. Our observed progression-
free survival of 35 months is shorter than that previously
described in clinical trials in which the median progres-
sion-free survival was 52 months in relapsed/refractory
disease.9 These observations suggest that outcomes vary
when comparing clinical trial patients and those treated in
clinical practice, underscoring the need to better under-
stand outcomes and toxicities in a real-world setting.
Patient-specific factors, including molecular prognostic
markers, performance status and prior therapies, may par-
tially account for these discrepancies. Our front-line
cohort of patients had a greater number of molecular
abnormalities than typically seen in the treatment-naïve
population; 37% of patients had del17p and 40% had a
complex karyotype. Both features are associated with an
inferior progression-free survival.22 For the relapsed
cohort, patients may have received prior idelalisib which
could have affected their subsequent response to ibrutinib;
notably, this was prohibited in the earliest clinical trials of
ibrutinib.11
At a median follow-up of 17 months, overall discontin-
uation in our study was high at 41%, suggesting that ibru-
tinib discontinuation is an emerging issue in clinical prac-
tice. This mirrors the high overall discontinuation patterns
seen in longer follow-up studies of clinical trial patients,6-9
particularly the 51% estimated discontinuations seen at a
median follow-up of 3.4 years by Woyach et al.8 In addi-
tion, 15% of front-line and 20% of relapsed ibrutinib-
treated patients required a dose reduction, which is signif-
icantly higher than the 4% of patients requiring dose
reduction due to adverse events in the RESONATE trial
due to adverse events (4%).2
Early ibrutinib trials suggested that progressive disease
was the cause of the majority of cases of discontinuation,2
a pattern that has persisted in many,6,8,9 but not all,5,7,10 stud-
ies including longer follow-up periods. Surprisingly, our
results demonstrate that intolerance (50.2% in the
relapsed/refractory setting, 63% in the front-line setting),
rather than progressive disease (21% in the
relapsed/refractory setting, 16% in the front-line setting),
accounts for the majority of cases of discontinuation.
Similar findings were previously noted in a smaller series
reported by our group.11 With respect to adverse events
leading to discontinuation, these values are greater than
those reported in a pooled analysis from the phase III
relapsed RESONATE and front-line RESONATE-2 studies
in which the discontinuation rate was 10%.23 Because at
least 50% of discontinuations are due to intolerance rather
than progression of disease, it is unlikely that these
patients harbor ibrutinib resistance mutations and, there-
fore, should likely be sensitive to alternate kinase
inhibitors with different side effect profiles. Two clinical
trials are being conducted at this time studying umbralisib
(NCT02742090) and acalabrutinib (NCT02717611) in
kinase inhibitor-intolerant patients.
For the first time we have established a timeline associ-
ated with time to discontinuation due to specific ibrutinib-
related toxicities. Similar to the experience with idelalisib
where specific toxicities appear to occur after different
878
Table 2. Reasons for Ibrutinib discontinuation.
Reason for ibrutinib Ibrutinib in Ibrutinib in relapse
discontinuation front-line (n=19) (n=231)
Toxicity 63.1% (n=12) 50.2% (n=116)
CLL progression 15.8% (n=3) 20.9% (n=49)
Other/unrelated death 5.3% (n=1) 12.1% (n=28)
Physician’s or patient’s preference 10.5% (n=2) 6.7% (n=15)
RT DLBCL 5.3% (n=1) 4.6% (n=10)
Stem cell transplantation/CAR T-cell 0 3.3% (n=8)
Financial concerns 0 0.8% (n=2)
Secondary malignancy 0 0.8% (n=2)
RT Hodgkin lymphoma 0 0.4% (n=1)
CLL: chronic lymphocytic leukemia; RT DLBCL: Richter transformation to diffuse large
B-cell lymphoma; CAR T-cell: chimeric antigen receptor T-cell); RT: Richter transforma-
tion.
periods of time on treatment,24 these data may be of value
in developing monitoring strategies for specific toxicities
and educational material related to ibrutinib toxicities.
Reasons for discontinuation also varied by line of thera-
py. Our analysis revealed a strikingly higher number of
discontinuations due to toxicity in the front-line setting;
63% compared to the previously reported 9% in RES-
ONATE-2.4 Similar findings, although not as marked,
were noted in the relapsed setting; 50% of patients in our
analysis discontinued therapy due to intolerance com-
pared to 12% discontinuing due to adverse events in the
4-year follow-up of RESONATE data.6 In line with prior
reports, we found that atrial fibrillation and pulmonary
complications were common reasons for discontinuation.
Arthralgias and rash were frequently noted as well, partic-
ularly in treatment-naïve patients. The higher discontinu-
ation rate for toxicity may reflect lack of physicians’ com-
fort in toxicity management, a higher incidence of toxicity
in clinical practice, differences in patients’ comorbidities
and age, or a lower threshold for discontinuation given an
increasing number of available alternative treatment
choices. This may be in contrast to the limited number of
therapies available to ibrutinib-treated patients in early
clinical trials. In a recent series by Lampson et al. of treat-
ment-naïve patients treated with idelalisib plus ofatu-
mumab adverse events (particularly liver toxicity) were
the most common reasons for discontinuation.25 This
study suggested that younger patients’ age and intact
immunity may lead to autoimmune treatment-related tox-
icities in treatment-naïve patients.25
Progression of disease was the second most common
indication for ibrutinib discontinuation. Sixteen percent of
previously untreated patients experienced progression
compared to 10% progressions/deaths at an 18-month fol-
low-up in the front-line RESONATE-2 study.4 This differ-
ence may be related to the exclusion of patients with
del17p from the RESONATE-2 study.4 Rates of discontin-
uation due to progression in the relapsed population in our
cohort were more comparable with the 27% found in the
4-year follow-up of the RESONATE study.6 Outcomes of
patients in our series who discontinued therapy due to
toxicity were superior to those of patients who discontin-
ued due to progressive disease.
We also considered limitations in the study design.
Conducted by physicians and research coordinators from
several institutions in a retrospective manner, data collec-
haematologica | 2018; 103(5)
 Ibrutinib: treatment toxicities and outcomes

tion may have been affected by inconsistencies in chart
interpretation, as well as clinical experience and practice
style. There were a disproportionate number of
relapsed/refractory patients compared to front-line
patients.
It is vital to develop strategies to mitigate ibrutinib intol-
erance so that efficacy can be further maximized.
Examples include the creation of guidelines for the evalu-
ation and management of problematic side effects such as
atrial fibrillation, rash, and arthralgias. An educational
forum focused on oncologists, physician educators, and
nurses should be implemented. In addition, the design of
future clinical trials should allow for cessation of therapy
in order to minimize ibrutinib exposure, particularly in the
small subset of patients who achieve complete remission.
This strategy has been successfully demonstrated in
patients receiving venetoclax who were able to achieve
minimal residual disease negativity.26 For example, the
incorporation of BCL-2 inhibitors and/or anti-CD20 mon-
oclonal antibody therapies in combination with ibrutinib
may enable patients to experience minimal residual dis-
ease-negative responses that may translate into shorter
durations of treatment.27,28
Acknowledgments
The authors thank Joseph and Cindy Riggs for their ongoing
support of this work. They also thank the Center for CLL,
University of Pensylvania.

References
1. Byrd J, Furman R, Coutre S, et al. Targeting
BTK with ibrutinib in relapsed chronic lym-
phocytic leukemia. N Engl J Med.
2013;369(1):32-34.
2. Byrd J, Brown J, O'Brien S, et al. Ibrutinib
versus ofatumumab in previously treated
chronic lymphoid leukemia. N Engl J Med.
2014;371(3):213-223.
3. O'Brien S, Jones J, Coutre S, et al. Ibrutinib
for patients with relapsed or refractory
chronic lymphocytic leukaemia with 17p
deletion (RESONATE-17): a phase 2, open-
label, multicentre study. Lancet Oncol.
2016;17(10):1409-1418.
4. Burger J, Tedeschi A, Barr P, et al. Ibrutinib as
initial therapy for patients with chronic lym-
phocytic leukemia N Engl J Med.
2015;373(25):2425-2437.
5. Chanan-Khan A, Cramer P, Demirkan F, et
al. Ibrutinib combined with bendamustine
and rituximab compared with placebo, ben-
damustine, and rituximab for previously
treated chronic lymphocytic leukaemia or
small lymphocytic lymphoma (HELIOS): a
randomised, double-blind, phase 3 study.
Lancet Oncol. 2016;17(2):200-211.
6. Byrd J, Hillmen P, O’Brien S, et al. Long-term
efficacy and safety with ibrutinib (ibr) in
previously treated chronic lymphocytic
leukemia (CLL): up to four years follow-up
of the RESONATE study. J Clin Oncol.
2017;35(suppl; abstr 7510).
7. Maddocks K, Ruppert A, Lozanski G, et al.
Etiology of ibrutinib therapy discontinua-
tion and outcomes in patients with chronic
lymphocytic leukemia. JAMA Oncol. 2015;
1(1):80-87.
8. Woyach J, Ruppert A, Guinn D, et al.
BTKC481S-mediated resistance to ibrutinib
in chronic lymphocytic leukemia. J Clin
Oncol. 2017;35(13):1437-1443.
9. O’Brien S, Furman R, Coutre S, et al. Five-
year experience with single-agent ibrutinib
in patients with previously untreated and
relapsed/refractory chronic lymphocytic
leukemia/small Lymphocytic Leukemia.
Blood. 2016; 128:233.
10. Jain P, Thompson P, Keating M, et al. Long-
term outcomes for patients with chronic
lymphocytic leukemia who discontinue
ibrutinib. Cancer. 2017;123(12): 2268-2273.
11. Mato A, Nabhan C, Barr P, et al. Outcomes
of CLL patients treated with sequential
kinase inhibitor therapy: a real world expe-
rience. Blood. 2016;128(18):2199-2205.
12. Parikh S, Chaffee K, Call T, et al. Ibrutinib
therapy for chronic lymphocytic leukemia
(CLL): an analysis of a large cohort of
patients treated in routine clinical practice.
Blood. 2015;126(23):2935.
13. Sandoval-Sus J, Chavez J, Dalia S, et al.
Outcomes of patients with relapsed/refrac-
tory chronic lymphocytic leukemia after
ibrutinib discontinuation outside clinical tri-
als: a single institution experience. Blood.
2015;126:2945.
14. Winqvist M, Asklid A, Andersson P, et al.
Real-world results of ibrutinib in patients
with relapsed or refractory chronic lympho-
cytic leukemia: data from 95 consecutive
patients treated in a compassionate use pro-
gram. Haematologica. 2016;101(12): 1573-
1580.
15. UK CLL Forum. Ibrutinib for
relapsed/refractory chronic lymphocytic
leukemia: a UK and Ireland analysis of out-
comes in 315 patients. Haematologica.
2016;101(12):1563-1572.
16. Mato A, Nabhan C, Kay N, et al. Real-world
clinical experience in the Connect chronic
lymphocytic leukaemia registry: a prospec-
tive cohort study of 1494 patients across 199
US centres. Br J Haematol. 2016;175(5):892-
903.
17. Bland J, Altman D. Survival probabilities
(the Kaplan-Meier method). BMJ.
1998;317(7172):1572.
18. Hallek M, Cheson B, Catovsky D, et al.
Guidelines for the diagnosis and treatment
of chronic lymphocytic leukemia: a report
from the International Workshop on
Chronic Lymphocytic Leukemia updating
the National Cancer Institute-Working
Group 1996 guidelines. 2008;111(12):5445-
5456.
19. Cheson B, Byrd J, Rai K, et al. Novel targeted
agents and the need to refine clinical end
points in chronic lymphocytic leukemia. J
Clin Oncol. 2012;30(23):2820-2822.
20. Matthews D, Farewell V. 7 The log-rank or
Mantel-Haenszel test for the comparison of
survival curves in: Basel S, Karger A, editors.
Using and Understanding Medical Statistics
2007:67-75.
21. Anderson P, Gill R. Cox’s regression model
for counting processes: a large sample study.
Ann Statist. 1982;10(4):1100-1120.
22. Thompson PA, O'Brien SM, Wierda WG, et
al. Complex karyotype is a stronger predic-
tor than del(17p) for an inferior outcome in
relapsed or refractory chronic lymphocytic
leukemia patients treated with ibrutinib-
based regimens. Cancer. 2015;121(20): 3612-
3621.
23. O'Brien SM, Byrd JC, Hillmen P, et al.
Outcomes with ibrutinib by line of therapy
in patients with CLL: analyses from phase III
data. J Clin Oncol. 2016;34(15_suppl): 7520.
24. Coutré SE, Barrientos JC, Brown JR, et al.
Management of adverse events associated
with idelalisib treatment: expert panel opin-
ion. Leuk Lymphoma. 2015;56(10): 2779-
2786.
25. Lampson B, Kasar S, Matos T, et al. Idelalisib
given front-line for treatment of chronic
lymphocytic leukemia causes frequent
immune-mediated hepatotoxicity. Blood
2016;128(12):195-203.
26. Seymour JF, Ma S, Brander DM, et al.
Venetoclax plus rituximab in relapsed or
refractory chronic lymphocytic leukaemia: a
phase 1b study. Lancet Oncol. 1017;18(2):
230-240.
27. Sharman J, Brander D, Mato A, et al.
Ublituximab and ibrutinib for previously
treated genetically high-risk chronic lym-
phocytic leukemia: results of the GENUINE
phase 3 study. J Clin Oncol.
2017;35(Supplemental, ASCO abstract
7504).
28. Hillmen P, Rawstron A, Munir T, et al. The
initial report of bloodwise tap clarity study
combining ibrutinib and venetoclax in
relapsed, refractory CLL shows acceptable
safety and promising early indications EHA
Oral Presentation June 22-25, Madrid Spain.
2017.

haematologica | 2018; 103(5) 879

ARTICLE Plasma Cell Disorders

A novel nano-immunoassay method for

Ferrata Storti
Foundation quantification of proteins from CD138-purified
myeloma cells: biological and clinical utility
Irena Misiewicz-Krzeminska,1,2,3 Luis Antonio Corchete,1,2 Elizabeta A. Rojas,1,2
Joaquín Martínez-López,4 Ramón García-Sanz,1,2,5 Albert Oriol,7 Joan Bladé,8
Juan-José Lahuerta,6 Jesús San Miguel,9 María-Victoria Mateos1,2,5
and Norma C. Gutiérrez1,2,5
Haematologica 2018 1
Cancer Research Center-IBMCC (USAL-CSIC), Salamanca, Spain, 2Institute of Biomedical
Research of Salamanca (IBSAL), Spain; 3National Medicines Institute, Warsaw, Poland;
Hematology Department, Hospital 12 de Octubre, CNIO, Complutense University,
Volume 103(5):880-889 4

CIBERONC, Madrid, Spain; 5Hospital Universitario de Salamanca, CIBERONC, Spain;

6
Hospital 12 de Octubre, Madrid, Spain; 7Hospital Germans Trias i Pujol, Barcelona, Spain;
8
Hospital Clinic, Barcelona, Spain and 9Clínica Universidad de Navarra, CIMA, IDISNA,
CIBERONC, Pamplona, Spain

ABSTRACT

P
rotein analysis in bone marrow samples from patients with multiple
myeloma has been limited by the low concentration of proteins
obtained after CD138+ cell selection. A novel approach based on
capillary nano-immunoassay could make it possible to quantify dozens
of proteins from each myeloma sample in an automated manner. Here
we present a method for the accurate and robust quantification of the
expression of multiple proteins extracted from CD138-purified multiple
myeloma samples frozen in RLT Plus buffer, which is commonly used for
nucleic acid preservation and isolation. Additionally, the biological and
clinical value of this analysis for a panel of 12 proteins essential to the
pathogenesis of multiple myeloma was evaluated in 63 patients with
Correspondence: newly diagnosed multiple myeloma. The analysis of the prognostic
normagu@usal.es impact of CRBN/Cereblon and IKZF1/Ikaros mRNA/protein showed
that only the protein levels were able to predict progression-free survival
of patients; mRNA levels were not associated with prognosis.
Received: September 27, 2017. Interestingly, high levels of Cereblon and Ikaros proteins were associated
Accepted: January 31, 2018. with longer progression-free survival only in patients who received
Pre-published: March 15, 2018. immunomodulatory drugs and not in those treated with other drugs. In
conclusion, the capillary nano-immunoassay platform provides a novel
opportunity for automated quantification of the expression of more than
doi:10.3324/haematol.2017.181628 20 proteins in CD138+ primary multiple myeloma samples.

Check the online version for the most updated

information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/880 Introduction

©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
Genomics has come to dominate biomedical research in recent years. For exam-
ple, high-throughput genomic technologies have been used for the comprehensive
analysis of multiple myeloma (MM). In particular, gene expression profiling has
enabled the molecular classification of MM, which is widely used in biological
research.1 However, a knowledge of protein expression is essential for identifying
therapeutic targets, since proteins are the molecules through which most new
drugs achieve their efficacy. The limited amount of sample remaining after plasma
cell purification means that messenger RNA (mRNA) quantification is still used as
an indirect measure of protein expression in most cases. However, several studies
have shown that protein levels cannot be predicted from mRNA measurements.2
Immunohistochemistry and flow cytometry have been used to analyze expres-
sion at the protein level, although to a limited extent. These methods are of great
value and are of proven clinical utility, but they have some limitations that make
them less useful for studying intracellular protein levels. They mostly use directly

880 haematologica | 2018; 103(5)

 Nano-scale protein quantification in multiple myeloma

marked antibodies that reduce the sensitivity of detection,
and fewer antibodies are available for these techniques,
even when used in an indirect assay.3
Immunohistochemistry allows only semiquantitative
analysis of protein expression, and requires a well-trained
pathologist to interpret the results. Moreover, neither
technique is able to identify non-specific antibody binding
to other proteins.3
Western blotting (WB) remains the “gold standard” tech-
nique for protein characterization in most laboratories.
However, WB consumes large quantities of reagents, has a
low throughput, and requires a great deal of time and
effort involving many laborious manual processing steps.
Moreover, WB only yields semiquantitative data of poor
repeatability, making it a challenge to go beyond using the
assay in discovery research to apply it reliably in the clin-
ical setting.4–6 A further drawback is that it is not always
possible to obtain the quantity of protein extract required
for WB from primary cancer samples.
MM is a clear prototype of a bone marrow-infiltrating
tumor for which a relatively small quantity of sample is
available after the diagnostic procedure, which involves
morphological evaluation, immunophenotypic characteri-
zation by flow cytometry, and CD138+ plasma cell separa-
tion for routine fluorescence in situ hybridization analysis.
The recent development of a method based on the combi-
nation of capillary nano-electrophoresis with immunoas-
say (CNIA), also known as ‘simple western’, requires only
very small amounts of sample to be able to measure pro-
tein expression.3,7 This technical advance makes it possible
to analyze the expression of 50-100 proteins in a single
MM sample. Here we present the results of a pilot study
using this platform in MM patients. The main goal was to
quantify accurately and robustly the proteins extracted
from CD138-purified MM samples frozen in RLT Plus
buffer, which is commonly used as a method for RNA and
DNA preservation. Additionally, we attempted to estab-
lish the clinical value of this analysis using a panel of pro-
teins essential to MM pathogenesis, comparing it with
that of the corresponding RNA expression.
extracted separately. After overnight incubation at -20ºC, the pro-
teins were centrifuged at 13,000 x g for 30 min at 4ºC, and washed
twice with 70% ice-cold ethanol followed by centrifugation for a
further 10 min. The protein precipitate was dried at 39ºC and dis-
solved with 50 μL 0.2 M NaOH for 10 min at room temperature
and 4x WB sample buffer for at least 15 min at room temperature.
Samples were then denatured at 95ºC for 5 min, cooled to room
temperature and stored in aliquots at -80ºC. Before any assay,
samples were heated to room temperature, then kept at 37ºC for
30 min in order to re-dissolve any protein that had precipitated
during freezing.
Capillary electrophoresis immunoassay
Capillary electrophoresis immunoassay or simple western
analysis was performed using the WESTM machine (ProteinSimple,
San Jose, CA, USA) in accordance with the manufacturer’s proto-
cols. The Total Protein Assay (ProteinSimple) was used to quantify
the protein concentration. In brief, 5 μL of proteins were loaded on
the plate, separated by size, labeled with a biotin reagent and
detected by chemiluminescence using streptavidin-horseradish
peroxidase. At the end of the run, the proportion of the protein of
interest in the total protein in the sample was measured, in com-
parison to a standard curve previously generated using JJN3 cell
line extracts of known protein concentrations.
Primary antibodies used in the study and the optimized condi-
tions for each antibody are presented in Table 1. Data were ana-
lyzed using CompassTM software. Each protein peak was meas-
ured automatically and normalized with respect to the GAPDH
median area under the peak. Expression of each protein is present-
ed as its abundance relative to GAPDH.

Methods
For more specific information see the Online Supplementary File.

Patients and multiple myeloma cell lines

Sixty-three samples from patients diagnosed with MM
between October 2013 and November 2015 were included in the
study (Online Supplementary Table S1). Forty-three had been
enrolled in two Spanish Myeloma Group clinical trials: GEM2010
[bortezomib/melphalan/prednisone and lenalidomide/dexam-
ethasone in a sequential or alternating manner; (n=24)] and
BenVelPres [bendamustine/bortezomib/prednisone; (n=19)]. The
other 20 patients were not treated as part of a clinical trial.
The impact of RNA and protein expression on patients’ survival
was evaluated only in the group of patients that took part in the
clinical trials (Online Supplementary Table S1). The scheme of the
study is presented in Figure 1.

Protein extraction from RLT Plus buffer

Proteins were extracted by ice-cold acetone precipitation from
RNA-column flow-through liquid. To increase the rate of protein
Figure 1. Scheme of the study.
precipitation 10 mM NaCl was added to the acetone at 80% (v/v).
For technical reasons, each sample was divided into two tubes and

haematologica | 2018; 103(5) 881

 I. Misiewicz-Krzeminska et al.

DNA/RNA extraction and quantitative real-time
polymerase chain reaction analysis
mRNA expression was evaluated by Taqman assay quantitative
real-time polymerase chain reaction (qRT-PCR) analysis using the
respective GAPDH Taqman assay as a control, by the 2-ΔCt method.
Statistical analysis
Spearman correlations were calculated. Progression-free sur-
vival (PFS) was calculated for each gene and protein. Survival
curves were plotted using the Kaplan–Meier method and statisti-
cal significance was evaluated with the log-rank test.
Results
Protein extraction from RLT Plus buffer results
in optimal quality and quantity
Firstly, we evaluated the amount and quality of the pro-
tein extracted with our protocol. The data generated by
the WESTM system were visualized as virtual blots (Figure
2A) or peaks (Figure 2B) that were quantified as the area
under the curve using the inbuilt algorithm of the
CompassTM software. Using JJN3 myeloma cell line lysates
of known concentration, we generated a protein standard
curve that proved to be linear over the evaluated range of
concentrations (Figure 2C). The amount of protein
obtained from each sample ranged between 0.00 and 0.36
mg protein, with a median quantity of 0.06 mg per sam-
ple. Three of the 63 samples had insufficient material to
analyze protein expression (Figure 2D). We compared the
expression of the various proteins extracted from MM
cells stored in RLT Plus buffer with that obtained using the
standard RIPA protocol and found the signals to be similar
for the two protocols (Figure 2E,F).
Optimization of protein quantification by capillary
nano-electrophoresis with immunoassay
For each analyzed protein, we first searched in the
ProteinSimple antibody database for the optimized condi-
tions (http://www.proteinsimple.com/antibody/antibod-
ies.html). If the antibody was present, we re-evaluated it in
our system, using the antibody at the indicated concentra-
tion and at double and half the indicated concentration. In
the event that the protein evaluated was not present in the
database, we performed a full optimization, which consist-
ed of running the assays in the cell line samples at two con-
centrations (0.1 mg/mL and 0.2 mg/mL) with at least five
antibody dilutions in order to determine the optimal con-
centration at which the antigen-antibody binding was sat-
urated and no change in antibody concentration influenced
the result. The optimized concentrations for each anti-
body, the molecular weights at which the peaks were
observed, and the coefficients of variation arising from the
validation of each protein are shown in Table 1.
Standard curves were produced for each protein to eval-
uate the range of linearity over which the expression of
each protein could be quantified. Briefly, each capillary
contained the sample at a different dilution, and the pro-
tein detection was visualized as virtual blots, as exempli-
fied by the use of Aiolos in Figure 3A. The peaks obtained
for each dilution, which were obtained automatically by
the program, have a distinct height and width, depending
on the sample dilution (Figure 3B). Once they had been
quantified the standard curve was generated (Figure 3C).
After protein quantification, we compared the value
obtained for each sample and each protein with the
respective standard curve to ensure correct measurement.
The limit of quantitation was set as signal-to-noise ratio of
10:1 in accordance with the guidelines from the European
Directorate for the Quality of Medicine set out in the
European Union Pharmacopoeia.8
The results of Aiolos quantification in six samples are
shown in Figure 3D, in which virtual blots for both Aiolos
and GAPDH are presented.
Analysis of mRNA and protein expression
We analyzed the expression of 12 genes and their
encoded proteins, together with GAPDH as a control
(Figure 1). We decided to select proteins involved in MM
or cancer pathogenesis: Cyclin D1 and Cyclin D2, whose
overexpression is a unifying event for most MM;9–11 c-myc,
which is consistently found to be involved in the transfor-
mation of monoclonal gammopathy of undetermined sig-
nificance into MM;12,13 HSP90, which is upregulated in

Table 1. Summary of proteins and antibodies used in the study.

Target protein Company Cat number Species Clonality Dilution Ab MW peak Intra-assay CV (%)
1 Aiolos Cell Signaling 12720 Rabbit Polyclonal 1:100 85 10.9
2 Calnexin Enzolifesciences ADI-SPA-860 Rabbit Polyclonal 1:250 119 4.9
3 Cereblon NovusBio NBP1-91810 Rabbit Polyclonal 1:80 59 9.7
4 c-myc Cell Signalling 5605 Rabbit Monoclonal 1:50 75 8.3
5 Cyclin D1 Abcam Ab134175 Rabbit Monoclonal 1:50 40 9.9
6 Cyclin D2 Cell Signaling 3741 Rabbit Monoclonal 1:100 38 7.8
7 DDX21 Abcam Ab182156 Rabbit Monoclonal 1:100 110 11.4
8 HSP90 Cell Signaling 4877 Rabbit Monoclonal 1:50 95 7.0
9 Ikaros Santa Cruz Sc-13039 Rabbit Polyclonal 1:50 70 10.8
10 PSME1 NovusBio NBP1-83121 Rabbit Polyclonal 1:50 37 10.2
11 RIPK1 Cell Signaling 3493 Rabbit Monoclonal 1:50 79 5.8
12 XAF1 Cell Signaling 13805 Rabbit Monoclonal 1:25 46, 110 15.5
13 GAPDH Cell Signaling 2118 Rabbit Monoclonal 1:50 42 8.6
Ab: antibody; MW: molecular weight; CV: coefficient of variation.

882 haematologica | 2018; 103(5)

 Nano-scale protein quantification in multiple myeloma

many solid and hematologic malignancies, including
MM;14 Calnexin, which forms endoplasmic reticulum and
is upregulated in MM relative to normal plasma cells in
genetically identical twin samples;15 and DDX21 or RIPK1,
with known involvement in several tumors.16–18 In addi-
tion, proteins involved in the mechanism of action of
antimyeloma drugs were included: Cereblon, Ikaros,
Aiolos for immunomodulatory drugs; XAF1 for melpha-
lan; and PSME1 for bortezomib.19–22
At the protein level, PSME1 and Calnexin showed the
highest median level of expression, while HSP90 was the
most strongly expressed mRNA (Figure 4A,B). Conversely,
proteins Cyclin D1, Cyclin D2 and c-myc had the lowest
median level of expression, and CRBN, RIPK1 and XAF1
were the least expressed mRNA.
In general, the expression of mRNA was more homoge-
nous than that of proteins, as indicated by the higher coef-
ficients of variation for the latter (Figure 4C). In fact, the
coefficients of variation were significantly lower than
those for c-myc, DDX21, HSP90, IKZF1 and PSME1
mRNA than for their respective encoded proteins. The
highest variability in expression, both at the mRNA and
protein levels, was observed for CCND2/Cyclin D2 and
CCND1/Cyclin D1, as well as for c-myc protein.
Next, we analyzed the correlation between the two lev-
els of gene expression, mRNA and protein. Interestingly,

A B D

E F

Figure 2. Optimization of protein extraction from RLT Plus samples. Due to the various additives in the sample buffer there is marked incompatibility with most of
the normally used protein quantification methods. The Total Protein assay was therefore used, as it is insensitive to high SDS concentrations. The standard curve was
generated using JJN3 cell line extracts at 0.25 mg/mL concentration, and serial dilutions thereof. Each capillary contained one sample of a known concentration.
Results were visualized as virtual gels (A) and the numbers correspond to the areas under the curves of the peaks (B). A standard curve was generated, plotting the
result for each capillary quantification (C). Amount of protein obtained from each sample (D). Comparison of results from Calnexin, Cyclin D2, GAPDH and Ikaros quan-
tification in the U266 cell line, from which protein extracts were obtained by RLT Plus and RIPA extraction, and visualized as virtual blots or two distinct dilutions of
the sample (E), and one dilution extracted with both protocols visualized as peaks (F).

haematologica | 2018; 103(5) 883

 I. Misiewicz-Krzeminska et al.

only Cyclin D1 and Cyclin D2 protein levels were strongly
correlated with the respective CCND1 and CCND2 mRNA
levels (Figure 4D). We observed a modest correlation for
Aiolos, Calnexin and DDX21 proteins with their respective
mRNA.
Although the number of proteins analyzed was limited,
we examined the correlation between the levels of the dif-
ferent proteins. A positive correlation was observed
between most of the protein pairs (Online Supplementary
Figure S2). We confirmed the previously described relation-
ship between Ikaros/Aiolos and c-myc.19 Additionally, c-
myc protein expression was positively correlated with
Cereblon, Calnexin, and RIPK1, and negatively correlated
with DDX21 (Online Supplementary Figure S2). We found
that protein levels of Cereblon, Ikaros and Aiolos, all of
which are required for the activity of immunomodulatory
drugs, were correlated with each other (Online
Supplementary Figure S2).
The potential association between the expression of pro-
teins and mRNA tested in the study and chromosomal
abnormalities was also explored. We confirmed the well-
established pattern of CCND1/Cyclin D1 and
CCND2/Cyclin D2 expression in t(11;14) and t(4;14)
(Online Supplementary Figure S3). A lower level of expression
of PSME1 and RIPK1 proteins in MM with 1q gains, and a
higher level of IKZF1 mRNA expression in MM with
t(11;14) were also observed.
Influence of mRNA and protein levels on survival
of myeloma patients
Since clinical data were available for 43 MM patients, 24
enrolled in GEM 2010 and 19 in BenVelPres clinical trials,
we also performed survival analysis for proteins and
mRNA using PFS as the endpoint (Table 2). Cereblon and
Ikaros were the only proteins able to predict PFS.
Interestingly, mRNA levels of CRBN and IKZF1 were not
associated with prognosis (Figure 5). Accordingly, patients
with a high level of Cereblon protein had a longer PFS
than those with a low level (50.4 versus 16.3 months,
P<0.001). Similarly, high levels of Ikaros protein were
associated with longer PFS (45.1 versus 17.8 months,
P<0.01). The levels of two mRNA were associated with
longer PFS: a high level of PSME1 (50.4 versus 23.5
months, P<0.05) and a low level of XAF1 (20.3 versus 45.1
months, P<0.05).
Since Cereblon, Ikaros and Aiolos are involved in the
mechanism of action of immunomodulatory drugs, and
only GEM2010 patients were treated with lenalidomide,
we examined whether the prognostic value of these pro-
teins was influenced by the type of treatment. Indeed,
high levels of Cereblon and Ikaros were both associated
with longer PFS only in patients who received
immunomodulatory drugs and not in those treated with
other drugs (Figure 6).

A B D

Figure 3. Optimization of protein expression quantification. For each protein the standard curve was generated using the sample with the strongest signal to prepare
the serial dilutions. Each dilution was run in a separate capillary. The sample result for Aiolos is visualized as a virtual blot (A) or as peaks (B). The standard curve
established the linear range for each protein (C). The sample result for Aiolos together with the respective GAPDH for each sample was run in separate capillaries
(D). Each peak area was quantified and Aiolos was normalized with respective to GAPDH, as shown in the example.

884 haematologica | 2018; 103(5)

 Nano-scale protein quantification in multiple myeloma

Discussion ples after CD138+ separation are usually stored in buffers

MM has been comprehensively studied at the DNA and
RNA levels using high-throughput technologies such as
microarrays for detecting copy number abnormalities, and
gene expression profiling, and, more recently, next-gener-
ation sequencing for DNA mutation analysis. MM sam-
such as TRIZOL or RLT Plus, which preserve nucleic acids
for subsequent use in genomic studies. While it is techni-
cally possible to extract proteins from these buffers,23–25 the
quantity of protein would not be sufficient for multiple
WB to be carried out. In the classic WB, the amount of the
purified plasma cells, even if all of it were available for the

A C

Figure 4. Two levels of analysis of each gene RNA and protein. mRNA expression of each gene was assessed by qRT-PCR and normalized relative to GAPDH and
expressed as 2-ΔCt (A). Abundance of each protein was assessed by CNIA and normalized relative to GAPDH abundance in each case (B). The Y axis of graphs (A) and
(B) are expressed on a log scale. The variability of each mRNA and protein measurement in the analyzed population of patients with MM, measured as percentage
coeffcient of variation (CV%). The threshold of statistical significance (*P<0.05) was determined as described in the Methods section (C). Spearman correlation coef-
ficient for each mRNA/protein pair ranked by increasing P value (D).

haematologica | 2018; 103(5) 885

 I. Misiewicz-Krzeminska et al.

Table 2. Univariate analysis of progression-free survival.

mRNA HR Group n Median PFS (months) P-value Protein HR Group n Median PFS (months) P-value
(CI) (CI)
IKZF3 1.84 H 10 21.7 0.19 Aiolos 0.51 H 33 45.1 0.16
(0.74-4.6) L 33 45.1 (0.19-1.34) L 10 23.5
CANX1 0.44 H 21 50.4 0.074 Calnexin 1.82 H 22 30 0.2
(0.18-1.11) L 22 23.5 (0.72-4.6) L 18 43.8
CRBN 0.49 H 10 50.4 0.24 Cereblon 0.23 H 31 50.4 0.00034
(0.14-1.67) L 33 29.8 (0.09-0.55) L 12 16.3
CCND1 2.55 H 27 29.8 0.066 CyclinD1 1.55 H 23 30 0.33
(0.91-7.14) L 16 NR (0.64-3.78) L 20 NR
CCND2 0.38 H 17 NR 0.053 CyclinD2 0.24 H 11 NR 0.051
(0.13-1.05) L 26 29.8 (0.05-1.03) L 32 29.8
MYC 1.54 H 33 30 0.44 c-myc 0.63 H 27 45.1 0.31
(0.51-4.64) L 10 45.1 (0.26-1.56) L 13 29.8
DDX21 0.49 H 29 43.8 0.1 Ddx21 0.47 H 14 50.4 0.18
(0.2-1.17) L 14 23.5 (0.15-1.46) L 23 30
HSP90 0.52 H 24 50.4 0.14 Hsp90 1.5 H 12 29.8 0.44
(0.22-1.25) L 19 23.5 (0.53-4.27) L 25 50.4
IKZF1 1.93 H 25 29.8 0.15 Ikaros 0.31 H 32 45.1 0.0075
(0.77-4.83) L 18 43.8 (0.12-0.77) L 11 17.8
PSME1 0.33 H 18 50.4 0.018 Psme1 0.48 H 30 45.1 0.13
(0.12-0.86) L 25 23.5 (0.19-1.26) L 10 16.3
RIPK1 0.52 H 28 50.4 0.13 Ripk1 1.41 H 11 30 0.49
(0.22-1.24) L 15 23.5 (0.53-3.73) L 29 45.1
XAF1 2.39 H 11 20.3 0.048 Xaf1 1.56 H 15 30 0.34
(0.98-5.8) L 32 45.1 (0.63-3.85) L 25 45.1
HR: hazard ratio; CI: confidence interval; PFS: progression-free survival; H: high level; L: low level.

protein studies, would have allowed at most six proteins
to be evaluated, the median amount of sample being suf-
ficient to analyze two proteins.
Here we present for the first time a method for quanti-
fying the expression of multiple proteins from myeloma
cells stored in the buffer commonly used for nucleic acid
preservation and isolation. We report a protocol for pro-
tein extraction from MM cells stored in RLT Plus buffer
based on a well-known acetone precipitation proce-
dure.26,27 We decided to add NaCl to the sample before pre-
cipitation, since a greater inorganic salt content is known
to improve protein yield.28 After testing several types of
salts and concentrations, we chose the one with the best
performance. We also assessed several methods of protein
pellet dissolution, finally settling for a 0.2 M NaOH and 4x
WB sample buffer, since slight changes in the pH of the
environment change protein solubility.26,29
In contrast to the classic WB, which provides only semi-
quantitative (blot-based) results, CNIA quantifies the area
under the curve of the signal in each capillary, enabling
expression relative to the control protein to be calculat-
ed.3,30 To determine whether the CNIA method is suitable
for evaluating the expression of multiple proteins, using
various antibodies, we tested the performance of each
protein in the WESTM system. We first optimized the con-
centration of the antibody to be employed using RIPA-
extracted proteins from MM cell lines, as suggested by the
system provider. The antibody dilution used has to be the
one that saturates the epitope-antibody binding, so that
the additional increase in antibody concentration would
not have caused the increase in the signal. Although the
concentration of antibodies used by the CNIA platform is
higher than that used in WB, lower amounts of antibody
are required because of the small volume of antibody.
Comparing the signal detected by each antibody when
the sample was extracted by the RIPA method with that
obtained using our protocol revealed no significant differ-
ences for any of the proteins evaluated, which supports
the suitability of the present protocol for extracting pro-
teins from RLT Plus buffer. We observed differences
between the predicted molecular weight and that detected
by the CNIA system for some proteins, regardless of the
extraction method. The most probable explanation for
this phenomenon is that migration depends on the mobil-
ity in the matrix.31 In fact, each system provides a unique
molecular weight value that depends on the particular
interactions between the matrix and protein, and the true
molecular weight can only be determined by mass spec-
trometry or sequencing.32 In our assay, we analyzed only
the data obtained from the antibodies that detected peaks
whose signal intensity was linear in the serial dilutions,
enabling the frequent biases in WB to be eliminated.
Although such standard curves are not usually calculated
as part of the WB method, this is highly recommendable.33
Even though, at first glance, the optimization step
required for each antibody seems laborious, one CNIA run
allows 24 samples to be analyzed, so it would be possible
to optimize, for example, four antibodies in 3 hours.
Therefore, bearing in mind the subsequent possibility of
quantifying protein abundances, it is not such a time-con-
suming process.
After demonstrating that our approach accurately quan-

886 haematologica | 2018; 103(5)


A B
tified the proteins extracted at the same time as the DNA
and RNA from the RLT Plus buffer, we investigated the
applicability of the method to the analysis of the expres-
sion of key proteins in MM biology, such as D Cyclins, c-
myc, Cereblon, Ikaros, and Aiolos, among others. We also
wanted to compare protein expression with the corre-
sponding mRNA level, since many basic studies have
revealed that only 30-40% of protein abundance can be
explained by the mRNA level.34 Our results showed a
moderate or low correlation between mRNA and protein
levels of expression, and are consistent with the general
observation that ~60% of the variation in protein concen-
tration cannot be explained by measuring mRNA alone.34
We also observed that the mRNA level was less variable
than protein expression among MM patients for all the
mRNA and protein pairs analyzed. Indeed, protein abun-
dance is regulated by a variety of complex mechanisms,
such as post-transcriptional and post-translational modifi-
cations, and protein degradation regulation.34,35 By measur-
ing mRNA abundance, only the early steps in a long chain
of regulatory events are considered.36 However, the
mRNA level is still often employed as a proxy for protein
abundance, mostly because of the lack of appropriate
technology to quantify proteins quickly and efficiently.
Our results reproduce the well-known pattern of
CCND1/Cyclin D1 and CCND2/Cyclin D2 expression in
MM with t(11;14) and t(4;14).9 We also found a correla-
tion between c-myc and Ikaros and Aiolos levels, analyz-
ing either mRNA or protein expression, consistent with
the previously demonstrated regulation of c-myc by
Ikaros and Aiolos in MM.19 Interestingly, the correlation
between Ikaros and Aiolos levels was stronger for the
protein than for the mRNA. To our knowledge, this is the
first time that the protein levels of c-myc, Ikaros and
haematologica | 2018; 103(5)
Nano-scale protein quantification in multiple myeloma
Figure 5. Progression-free sur-
vival according to levels of
MRNA and protein expression.
Progression-free survival in
patients with low and high lev-
els of mRNA (A) and protein (B)
expression. The log-rank test
was performed for each gene
and protein and Kaplan-Meier
curves represent the PFS of
MM patients depending on
mRNA and protein status.
Cutoff Finder software
(http://molpath.charite.de/cut
off) was used to obtain the
optimal cutoff, which was
defined as that producing the
most significant split that dis-
criminates between good and
poor survival by examining all
the possible cutoffs using the
log-rank test.
Aiolos have been quantified and the relationship between
their expressions analyzed in MM. In T-cell acute lym-
phoblastic leukemia, for example, the levels of mRNA
encoding Ikaros and Aiolos were weakly, but significantly
correlated.37
Among the proteins included in our study, we observed
a significant association between protein level and PFS for
Cereblon and Ikaros, while this association was not
observed for the respective mRNA levels. Cereblon forms
an E3 ubiquitin ligase complex together with the damaged
DNA binding protein 1 (DDB1), Cullin4A (CUL4) and
Roc1. Immunomodulatory drugs, such as lenalidomide or
pomalidomide, bind to Cereblon in a region located at the
C-terminus of this protein.38,39 Our results did not demon-
strate a correlation between Cereblon protein and mRNA
level, and showed that only high levels of Cereblon pro-
tein were associated with a good prognosis in MM. These
findings are concordant with those of previous studies and
support the usage of protein expression to evaluate
Cereblon levels.40
Several independent groups have identified Ikaros and
Aiolos as the downstream targets of Cereblon after
immunomodulatory drug activation.41–43 The role of the
level of Ikaros in MM survival is controversial. When
Ikaros expression was investigated at the RNA level, a low
level of mRNA IKZF1 expression was associated with bet-
ter prognosis in newly diagnosed patients treated with
immunomodulatory drugs.44 On the other hand, low
IKZF1 levels were found to predict a lack of responsive-
ness to immunomodulatory drugs and a shorter overall
survival in refractory MM patients.45 We observed that a
high level of Ikaros protein was associated with longer
PFS, while no significant impact on prognosis was
observed when PFS was estimated from mRNA levels.
887
 I. Misiewicz-Krzeminska et al.

Figure 6. Progression-free sur-

vival in patients with low and
high Cereblon, Aiolos and
Ikaros protein levels, depend-
ing on the treatment scheme
(only patients treated accord-
ing to GEM2010 trial received
lenalidomide). The log-rank
test was performed for each
protein and Kaplan-Meier
curves represent progression-
free survival of MM patients
depending on protein status.
Cutoff Finder software
(http://molpath.charite.de/cut
off) was used to obtain the
optimal cutoff, which was
defined as that producing the
most significant split that dis-
criminates between good and
poor survival by examining all
the possible cutoffs using the
log-rank test.

These results are consistent with the longer survival dis-
played by relapsed/refractory MM patients treated with
lenalidomide who expressed high levels of IKZF1/3 pro-
tein, as evaluated by immunohistochemical staining.46
Although the number of patients analyzed in this
study is relatively small, the survival analysis carried out
dividing patients according to drug therapy showed that
high levels of Cereblon and Ikaros proteins are associat-
ed with a longer PFS only in patients who receive
immunomodulatory drugs and not in those who are
treated with other drugs. Other studies reached the same
conclusion that the level of Cereblon can predict the out-
come of patients with MM mainly in those treated with
immunomodulatory drugs.47–50 By contrast, in the present
series of MM patients, the level of Aiolos did not influ-
ence the outcome of the patients treated with
immunomodulatory drugs.
In summary, we present the implementation of a novel
technique based on capillary nano-immunoassay for
quantifying protein expression in MM samples in the clin-
ical setting. The requirement for only a relatively small
amount of material means that, for the first time, more
than 20 proteins can be analyzed using the same sample
frozen for DNA and RNA analysis. This makes the CNIA
platform a fast, effective and accurate tool for exploring
the impact of different proteins on the survival of patients
with MM and for investigating new protein biomarkers
that could help to predict the response to new drugs that
directly target specific proteins. These encouraging results
require further validation in a larger cohort of patients
with MM or other hematologic malignancies.
Acknowledgments
The authors thank Isabel Isidro, Teresa Prieto and Vanesa
Gutierrez for their technical assistance.
Funding
This work was funded by a grant from the International
Myeloma Foundation's Black Swan Research Initiative® and
“Gerencia Regional de Salud, Junta de Castilla y León”
(BIO/SA35/14). WES™ platform was acquired thanks to
INNOCAMPUS Program (CEI10-1-0010).

888 haematologica | 2018; 103(5)


References
1. Broyl A, Hose D, Lokhorst H, et al. Gene
expression profiling for molecular classifica-
tion of multiple myeloma in newly diag-
nosed patients. Blood. 2010;116(14):2543–
2553.
2. Tian Q, Stepaniants SB, Mao M, et al.
Integrated genomic and proteomic analyses
of gene expression in mammalian cells. Mol
Cell Proteomics. 2004;3(10):960–969.
3. Chen J-Q, Wakefield LM, Goldstein DJ.
Capillary nano-immunoassays: advancing
quantitative proteomics analysis, biomarker
assessment, and molecular diagnostics. J
Transl Med. 2015;13:182.
4. Chen J-Q, Heldman MR, Herrmann MA, et
al. Absolute quantitation of endogenous
proteins with precision and accuracy using a
capillary western system. Anal Biochem.
2013;442(1):97–103.
5. Ghosh R, Gilda JE, Gomes AV. The necessity
of and strategies for improving confidence in
the accuracy of western blots. Expert Rev
Proteomics. 2014;11(5):549–560.
6. Gassmann M, Grenacher B, Rohde B, Vogel
J. Quantifying western blots: pitfalls of den-
sitometry. Electrophoresis. 2009;30(11):
1845–1855.
7. Fan AC, Deb-Basu D, Orban MW, et al.
Nanofluidic proteomic assay for serial analy-
sis of oncoprotein activation in clinical spec-
imens. Nat Med. 2009;15(5):566–571.
8. Technical guide for the eleboration of mono-
graphs.
9. Bergsagel PL, Kuehl WM, Zhan F, Sawyer J,
Barlogie B, Shaughnessy J. Cyclin D dysreg-
ulation: an early and unifying pathogenic
event in multiple myeloma. Blood.
2005;106(1):296–303.
10. Chesi M, Bergsagel PL. Molecular pathogen-
esis of multiple myeloma: basic and clinical
updates. Int J Hematol. 2013;97(3):313–323.
11. Misiewicz-Krzeminska I, Sarasquete ME,
Vicente-Dueñas C, et al. Post-transcriptional
modifications contribute to the upregulation
of Cyclin D2 in multiple myeloma. Clin
Cancer Res. 2016;22(1):207–217.
12. Chiecchio L, Dagrada GP, Protheroe RKM,
et al. Loss of 1p and rearrangement of MYC
are associated with progression of smoulder-
ing myeloma to myeloma: sequential analy-
sis of a single case. Haematologica.
2009;94(7):1024–1028.
13. Chng W-J, Huang GF, Chung TH, et al.
Clinical and biological implications of MYC
activation: a common difference between
MGUS and newly diagnosed multiple
myeloma. Leukemia 2011;25(6):1026–1035.
14. Zhang L, Fok JHL, Davies FE. Heat shock
proteins in multiple myeloma. Oncotarget.
2014;5(5):1132–1148.
15. Munshi NC, Hideshima T, Carrasco D, et al.
Identification of genes modulated in multi-
ple myeloma using genetically identical twin
samples. Blood. 2004;103(5):1799–1806.
16. Cimino D, Fuso L, Sfiligoi C, et al.
Identification of new genes associated with
breast cancer progression by gene expres-
sion analysis of predefined sets of neoplastic
tissues. Int J Cancer. 2008;123(6):1327–1338.
17. Schneider AT, Gautheron J, Feoktistova M,
et al. RIPK1 suppresses a TRAF2-dependent
pathway to liver cancer. Cancer Cell.
2017;31(1):94–109.
18. Yoo JY, Jaime-Ramirez AC, Bolyard C, et al.
haematologica | 2018; 103(5)
Nano-scale protein quantification in multiple myeloma
Bortezomib treatment sensitizes oncolytic
HSV-1-treated tumors to NK cell
immunotherapy. Clin Cancer Res. 2016;22
(21):5265–5276.
19. Bjorklund CC, Lu L, Kang J, et al. Rate of
CRL4(CRBN) substrate Ikaros and Aiolos
degradation underlies differential activity of
lenalidomide and pomalidomide in multiple
myeloma cells by regulation of c-Myc and
IRF4. Blood Cancer J. 2015;5e354.
20. Kortüm KM, Zhu YX, Shi CX, Jedlowski P,
Stewart AK. Cereblon binding molecules in
multiple myeloma. Blood Rev.
2015;29(5):329–334.
21. Park GB, Kim Y-S, Kim D, et al. Melphalan-
induced apoptosis of EBV-transformed B
cells through upregulation of TAp73 and
XAF1 and nuclear import of XPA. J
Immunol. 2013;191(12):6281–6291.
22. Dytfeld D, Luczak M, Wrobel T, et al.
Comparative proteomic profiling of refrac-
tory/relapsed multiple myeloma reveals bio-
markers involved in resistance to borte-
zomib-based therapy. Oncotarget.
2016;7(35):56726-56736.
23. Chomczynski P. A reagent for the single-
step simultaneous isolation of RNA, DNA
and proteins from cell and tissue samples.
BioTechniques. 1993;15(3):532–534, 536–
537.
24. Vorreiter F, Richter S, Peter M, Baumann S,
von Bergen M, Tomm JM. Comparison and
optimization of methods for the simultane-
ous extraction of DNA, RNA, proteins, and
metabolites. Anal Biochem. 2016;50825–33.
25. Mathieson W, Thomas GA. Simultaneously
extracting DNA, RNA, and protein using
kits: is sample quantity or quality preju-
diced? Anal Biochem. 2013;433(1):10–18.
26. Pace CN, Treviño S, Prabhakaran E, Scholtz
JM. Protein structure, stability and solubility
in water and other solvents. Philos Trans R
Soc Lond B Biol Sci. 2004;359(1448):1225-
1234; discussion 1234-1235.
27. Race J. The determination of blood-proteins
by acid-acetone. Biochem J.
1932;26(5):1571–1584.
28. Crowell AMJ, Wall MJ, Doucette AA.
Maximizing recovery of water-soluble pro-
teins through acetone precipitation. Anal
Chim Acta. 2013;79648–54.
29. Cutler P, editor. Protein Purification
Protocols. 2nd. Totowa, NJ: Humana Press;
2004. 484 p.
30. Rustandi RR, Loughney JW, Hamm M, et al.
Qualitative and quantitative evaluation of
SimonTM, a new CE-based automated
Western blot system as applied to vaccine
development. Electrophoresis. 2012;33(17):
2790–2797.
31. Copeland RA. Methods for Protein Analysis.
Boston, MA: Springer US.
32. Strupat K. Molecular weight determination
of peptides and proteins by ESI and MALDI.
Meth Enzymol. 2005;405:1–36.
33. McDonough AA, Veiras LC, Minas JN, Ralph
DL. Considerations when quantitating pro-
tein abundance by immunoblot. Am J Physiol
Cell Physiol. 2015;308(6):C426-433.
34. Vogel C, Marcotte EM. Insights into the reg-
ulation of protein abundance from proteom-
ic and transcriptomic analyses. Nat Rev
Genet. 2012;13(4):227–232.
35. Liu Y, Beyer A, Aebersold R. On the
Dependency of cellular protein levels on
mRNA abundance. Cell. 2016;165(3):535–
550.
36. Payne SH. The utility of protein and mRNA
correlation. Trends Biochem Sci. 2015;40
(1):1–3.
37. Mitchell JL, Yankee TM. Variations in
mRNA and protein levels of Ikaros family
members in pediatric T cell acute lym-
phoblastic leukemia. Ann Transl Med.
2016;4(19):363–363.
38. Ito T, Ando H, Suzuki T, et al. Identification
of a primary target of thalidomide terato-
genicity. Science. 2010;327(5971):1345–
1350.
39. Ito T, Ando H, Handa H. Discovery of the
target for immunomodulatory drugs
(IMiDs). Rinsho Ketsueki. 2016;57(5):556–
562.
40. Gandhi AK, Mendy D, Waldman M, et al.
Measuring cereblon as a biomarker of
response or resistance to lenalidomide and
pomalidomide requires use of standardized
reagents and understanding of gene com-
plexity. Br J Haematol. 2014;164(2):233–244.
41. Lu G, Middleton RE, Sun H, et al. The
myeloma drug lenalidomide promotes the
cereblon-dependent destruction of Ikaros
proteins. Science. 2014;343(6168):305–309.
42. Krönke J, Udeshi ND, Narla A, et al.
Lenalidomide causes selective degradation
of IKZF1 and IKZF3 in multiple myeloma
cells. Science. 2014;343(6168):301–305.
43. Zhu YX, Braggio E, Shi C-X, et al. Cereblon
expression is required for the antimyeloma
activity of lenalidomide and pomalidomide.
Blood. 2011;118(18):4771–4779.
44. Krönke J, Kuchenbauer F, Kull M, et al.
IKZF1 expression is a prognostic marker in
newly diagnosed standard-risk multiple
myeloma treated with lenalidomide and
intensive chemotherapy: a study of the
German Myeloma Study Group (DSMM).
Leukemia. 2017;31(6):1363-1367.
45. Zhu YX, Braggio E, Shi C-X, et al.
Identification of cereblon-binding proteins
and relationship with response and survival
after IMiDs in multiple myeloma. Blood.
2014;124(4):536–545.
46. Pourabdollah M, Bahmanyar M, Atenafu
EG, Reece D, Hou J, Chang H. High IKZF1/3
protein expression is a favorable prognostic
factor for survival of relapsed/refractory
multiple myeloma patients treated with
lenalidomide. J Hematol Oncol. 2016;9
(1):123.
47. Bila J, Sretenovic A, Jelicic J, et al. Prognostic
significance of cereblon expression in
patients with multiple myeloma. Clin
Lymphoma Myeloma Leuk. 2016;16(11):
610–615.
48. Jung S-H, Choi H-J, Shin M-G, et al.
Thalidomide-based induction regimens are
as effective as bortezomib-based regimens
in elderly patients with multiple myeloma
with cereblon expression. Ann Hematol.
2016;95(10):1645–1651.
49. Huang S-Y, Lin C-W, Lin H-H, et al.
Expression of cereblon protein assessed by
immunohistochemicalstaining in myeloma
cells is associated with superior response of
thalidomide- and lenalidomide-based treat-
ment, but not bortezomib-based treatment,
in patients with multiple myeloma. Ann
Hematol. 2014;93(8):1371–1380.
50. Broyl A, Kuiper R, van Duin M, et al. High
cereblon expression is associated with better
survival in patients with newly diagnosed
multiple myeloma treated with thalidomide
maintenance. Blood. 2013;121(4):624–627.
889
ARTICLE Plasma Cell Disorders

Impact of extramedullary disease in patients

Ferrata Storti
Foundation with newly diagnosed multiple myeloma
undergoing autologous stem cell
transplantation: a study from the Chronic
Malignancies Working Party of the EBMT
Nico Gagelmann,1 Diderik-Jan Eikema,2 Simona Iacobelli,3 Linda Koster,2
Hareth Nahi,4 Anne-Marie Stoppa,5 Tamás Masszi,6 Denis Caillot,7
Haematologica 2018 Stig Lenhoff,8 Miklos Udvardy,9 Charles Crawley,10 William Arcese,11
Clara Mariette,12 Ann Hunter,13 Xavier Leleu,14 Martin Schipperus,15
Michel Delforge,16 Pietro Pioltelli,17 John A. Snowden,18 Maija Itälä-Remes,19
Volume 103(5):890-897
Maurizio Musso,20 Anja van Biezen,2 Laurent Garderet21 and Nicolaus Kröger1

1
Department of Stem Cell Transplantation, University Medical Center Hamburg-
Eppendorf, Hamburg, Germany; 2EBMT Data Office, Leiden, the Netherlands;
3
Dipartimento di Biologia, Università degli Study di Roma “Tor Vergata”, Italy; 4Center for
Hematology and Regenerative Medicine, Karolinska Institutet, Stockholm, Sweden;
5
Institut Paoli Calmettes, Marseille, France; 6St. István and St. László Hospital,
Budapest, Hungary; 7Hématologie Clinique, Dijon University Hospital, Dijon, France;
8
Department of Hematology, Skane University Hospital Lund, Sweden; 9Department of
Hematology, Bone Marrow Transplant Unit, Debrecen Medical University, Hungary;
10
Department of Haematology, Cambridge University Hospitals, UK; 11University Tor
Vergata, Roma, Italy; 12Department of Hematology, Grenoble University Hospital, France;
13
Leicester Royal Infirmary, Leicester; 14Hematology, Hôpital La Mileterie, Poitiers,
France; 15Haga Teaching Hospital, the Hague, the Netherlands; 16Department of
Hematology, UZ Leuven, Belgium; 17Hematology, Ospedale San Gerardo ASST Monza-
Università degli Studi di Milano Bicocca, Monza, Italy; 18Department of Haematology,
Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK; 19Turku University
Hospital, Finland; 20Division of Hematology, Casa di Cura "La Maddalena", Palermo, Italy
and 21Hopital St Antoine, Paris, France

ABSTRACT
Correspondence:

nkroeger@uke.uni-hamburg.de
Received: August 15, 2017.
Accepted: January 26, 2018.
Pre-published: February 1, 2018.
doi:10.3324/haematol.2017.178434
Check the online version for the most updated
information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5890
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
W
e investigated extramedullary disease in newly diagnosed mul-
tiple myeloma patients and its impact on outcome following
first-line autologous stem cell transplantation. We identified
3744 adult myeloma patients who received up-front single (n=3391) or
tandem transplantation (n=353) between 2005 and 2014 with available
data on extramedullary involvement at diagnosis. The overall incidence
of extramedullary disease was 18.2% (n=682) and increased per year
from 6.5% (2005) to 23.7% (2014). Paraskeletal involvement was found
in 543 (14.5%) and extramedullary organ involvement in 139 (3.7%).
More patients with extramedullary organ involvement had multiple
involved sites (≥2; P<0.001). In a comparison of patients with single sites
with patients without the disease, up-front transplantation resulted in at
least similar 3-year progression-free survival (paraskeletal: P=0.86, and
extramedullary organ: P=0.88). In single paraskeletal involvement, this
translated less clearly into worse 3-year overall survival (P=0.07) while
single organ involvement was significantly worse (P=0.001). Multiple
organ sites were associated with worse outcome (P<0.001 and P=0.01).
First-line treatment with tandem compared with single transplantation
resulted in similar survival in patients with extramedullary disease at
diagnosis (P=0.13 for both).
Introduction
Multiple myeloma (MM) accounts for approximately 2% of all new cancer cases
and 13% of hematologic cancers with an age-adjusted incidence of 6 per 100,000
per year in the USA and Europe.1 Autologous stem cell transplantation (ASCT) and
the development of new agents have considerably increased the median survival of
MM patients.2 The disease is characterized by a clonal proliferation of malignant
plasma cells with a strong dependence on the bone marrow (BM) microenviron-
ment.3

890 haematologica | 2018; 103(5)


However, in some MM patients, myeloma cells escape
the BM, resulting in extramedullary disease (EMD),
which can be further characterized by two different types
of involvement: 1) paraskeletal (PS), consisting of masses
that arose from bone lesions; and 2) extramedullary organ
involvement (EM), resulting from hematogenous spread
into different organs, skin and lymph nodes.4,5 At the time
of MM diagnosis, the incidence of EM involvement in
observational studies ranges from 1.7% to 4.5 using a
baseline staging that includes whole-body magnetic reso-
nance imaging (MRI) or positron emission tomography-
computed tomography (PET-CT).6 Paraskeletal involve-
ment is more frequent and varies from 7% to 34.2% due
to different definitions and access of sensitive imaging
techniques.7-10 Rates are also considered to be higher at
relapse or after surgery.11,12 Several studies reported that
EMD was associated with shorter survival rates, and thus
considered EMD as a high-risk feature. However, the evi-
dence of the effect of EMD at diagnosis is limited due to
small populations, heterogenous patient or intervention
selection, and relapse settings.13-16 Therefore, very limited
data are available to assess the role of EMD at diagnosis
of MM patients after up-front ASCT. This lack of evi-
dence is striking, since ASCT is standard therapy in first-
line therapy in eligible patients.17,18
Therefore, the objective of this study was to determine
the demographic and clinical characteristics of EMD in
MM patients at diagnosis and to evaluate its impact on
outcome after up-front ASCT as first-line therapy. For
this purpose, we analyzed 3744 patients with or without
EMD at diagnosis after up-front single or tandem ASCT
who had been reported to the European Society for Blood
and Marrow Transplantation (EBMT) registry between
2005 and 2014.
Methods
Study design and data collection
We included adult patients with MM who had available data
on extramedullary involvement at time of diagnosis who
received an up-front single ASCT within 12 months of diagnosis
or a tandem ASCT within six months from first ASCT as first-
line therapy and who had been reported to the EBMT registry
between January 2005 and December 2014. Patients were con-
sidered eligible for analysis if there were full data available on
extramedullary involvement (yes or no) at time of diagnosis, its
location, and the number of sites. This study was performed in
accordance with the principles of the Declaration of Helsinki
and was approved by the Chronic Malignancies Working Party
of the EBMT. The EBMT is a non-profit, scientific society repre-
senting more than 600 transplant centers, mainly in Europe.
Data are entered, managed, and maintained in a central database
with internet access. Audits are routinely performed to deter-
mine the accuracy of the data. Data on extramedullary involve-
ment were extracted from the database using Med-B forms.
Patients whose transplant data are reported provided informed
consent to use the information for research purposes and data
are anonymized.
Definitions and statistical analysis
The primary end point was 3-year progression-free survival
(PFS), which was defined as the time from ASCT to disease pro-
gression or death from any cause. The secondary end points
were 3-year overall survival (OS), non-relapse mortality (NRM)
haematologica | 2018; 103(5)
Autologous SCT for extramedullary myeloma
and response. Overall survival was defined as the time from
ASCT to death from any cause or last follow up. Non-relapse
mortality was defined as death without evidence of relapse or
progression, with relapse or progression as competing events.
Remission, progression and relapse were defined according to
standard EBMT criteria.19
On the basis of type of extramedullary involvement, we
defined three groups of myeloma patients: 1) without EMD
(MM group); 2) with paraskeletal (PS group); and 3)
extramedullary organ involvement (EM group). In addition, we
determined and analyzed the impact of the number of involved
sites as one or multiple (≥ 2) sites. Disease stage at diagnosis was
determined according to the International Staging System (ISS; I-
III),20 Salmon and Durie stages I, II or III, and also according to
renal function A or B.21 Performance status at ASCT was
assessed with the Karnofsky score (≤80 indicating poor and >80
good status).22 Categorical variables were compared with the use
of the Fisher’s exact test or the χ² test. Continuous variables
were analyzed using the Kruskal-Wallis test for independent
samples.
Survival probabilities were estimated by the Kaplan-Meier
method,23 and the Log-Rank test was used for univariate com-
parison. Median follow up was calculated according to the
reverse Kaplan-Meier method.24 Outcomes were artificially cen-
sored at three years. We used cumulative incidence analysis to
assess NRM, and labeled death from relapse as a competing
event.25,26 The proportional hazards assumption was verified
using graphical methods. Scaled Schoenfeld27 residuals and
graphical checks proposed by Klein and Moeschberger28 were
performed to find evidence of violations. To minimize the effect
of selection bias, we used a landmark analysis at six months
whenever single and tandem ASCT were compared.
To assess the multivariate effect of factors on each end point,
we used the Cox proportional hazards model to estimate hazard
ratios (HR).29 Only complete cases were included in the analysis.
All tests were two-sided, with the type I error rate fixed at
a=0.05. All analyses were performed using the statistical soft-
ware R, v.3.1.0 (R Foundation for Statistical Computing, Vienna,
Austria) and SPSS Statistics 23 (SPSS, IBM Corp, Armonk, NY,
USA).
Results
Incidence and sites
Among the 3744 patients identified in the registry,
14.5% (n=543) had paraskeletal involvement (PS group)
and 3.7% (n=139) extramedullary organ involvement (EM
group), while 81.8% (n=3062) had no EMD (MM group).
Between 2005 and 2014, the EMD incidence per year
increased from 6.5% to 23.7%.
Within the EM group, the involved sites were: kidney
(27.3%, n=38), skin (23.0%, n=32), lymph nodes (17.3%,
n=24), central nervous system (CNS; 10.1%, n=14), lung
and respiratory tract (6.5%, n=9), gastrointestinal tract
(GI) and liver (5.8%, n=8), pleura and heart (5.0%, n=7),
and spleen, ovaries and testes (5.0%, n=7). Most patients
with EMD (93.5%, n=639) presented with one involved
site (PS1 and EM1), 5.7% (n=36) had two sites, 0.7%
(n=5) had three sites, while four and five sites were pres-
ent in 0.1% (n=1) of patients, respectively. Notably, with-
in the PS group, all 19 patients with multiple (≥2) sites
had only additional paraskeletal involvement (PS2), while
further involvement in all 24 EM patients was also
restricted to other organs (EM2).
891
 N. Gagelmann et al.

Table 1. Patients’, disease and transplantation characteristics.

Patients without EMD Patients with EMD
Characteristic MM group PS group EM group Total P
N. of patients (%) 3062 (81.8) 543 (14.5) 139 (3.7) 3744
Sex, n. (%)
Female 1279 (41.8) 240 (44.2) 57 (41.0) 1576 (42.1) 0.55
Male 1783 (58.2) 303 (55.8) 82 (59.0) 2168 (57.9)
Age at diagnosis in years
Median 59.8 59.8 59.0 0.59
Range 27.4 to 77.7 26.8 to 76.8 31.8 to 72.8
ISS, n. (%)
I 781 (36.9) 158 (38.6) 29 (30.5) 968 (36.9)
II 759 (35.8) 148 (36.2) 29 (30.5) 936 (35.7) 0.11
III 578 (27.3) 103 (25.2) 37 (38.9) 781 (27.4)
Unknown 944 134 44 1122
Renal function, n. (%)
A 2188 (82.7) 410 (83.2) 85 (65.9) 2683 (82.1) <0.001
B 458 (17.3) 83 (16.8) 44 (34.1) 585 (17.9)
Unknown 416 50 10 476
Karnofsky score, n. (%)
Good 1877 (67.3) 344 (67.7) 81 (62.8) 2302 (67.2)
Poor 914 (32.7) 164 (32.3) 48 (37.2) 1126 (32.8) 0.55
Unknown 271 35 10 316
Status at ASCT, n. (%)
CR 580 (19.1) 115 (21.5) 16 (11.7) 711 (19.2)
PR 2262 (74.7) 389 (72.6) 109 (79.6) 2760 (74.6) 0.10
< PR 187 (6.2) 32 (6.0) 12 (8.8) 231 (6.2)
Unknown 33 7 2 42
Type of myeloma, n. (%)
Light chain only 672 (22.1) 122 (22.5) 39 (28.3) 833 (22.4)
Non-secretory 74 (2.4) 29 (5.4) 3 (2.2) 106 (2.9) 0.002
Heavy and light chain 2292 (75.4) 389 (72.0) 96 (69.6) 2777 (74.7)
Unknown 24 3 1 28
Ig-type, n. (%)
G 1648 (70.8) 282 (72.1) 72 (73.5) 2002 (71.1)
A 634 (27.2) 101 (25.8) 22 (22.4) 757 (26.9) 0.51
D/E/M 45 (1.9) 8 (2.0) 4 (4.1) 57 (2.0)
Unknown 735 152 41 928
Light chain type, n. (%)
Kappa 1853 (63.7) 327 (65.1) 73 (55.7) 2253 (63.6) 0.13
Lambda 1054 (36.3) 175 (34.9) 58 (44.3) 1287 (36.4)
Unknown 155 41 8 204
Years of ASCT, n. (%)
0 3062 (100) 3062 (81.8) <0.001
1 524 (96.5) 115 (82.7) 639 (17.1)
≥2 19 (3.5) 24 (17.3) 43 (1.1)
Years of ASCT, n. (%)
< 2009 518 (16.9) 82 (15.1) 25 (18.0) 625 (16.7) 0.001
2009-11 1400 (45.7) 204 (37.6) 62 (44.6) 1666 (44.5)
> 2011 1144 (37.4) 257 (47.3) 52 (37.4) 1453 (38.8)
Time to 1st ASCT in months
Median 6.2 6.2 6.1 0.81
Range 1.1 to 11.9 2.1 to 11.9 3.9 to 11.9
Type of ASCT, n.
Tandem 249 89 15 353
Single 2813 454 124 3391
EMD: extramedullary disease; MM: patients without extramedullary disease; PS: paraskeletal involvement; EM: extramedullary organ involvement; N.: number; ISS: International
Staging System; ASCT: autologous stem cell transplantation; CR: complete remission; PR: partial remission.

892 haematologica | 2018; 103(5)


Figure 1. Progression-free survival (A) and overall survival (B) with numbers at risk of myeloma patients following up-front autologous stem cell transplantation
according to presence of involvement. MM: no extramedullary disease; PS: paraskeletal involvement; EM: extramedullary organ involvement; N: number.
Patients’ and disease characteristics
Median age at diagnosis was 59.8 years in both MM and
PS, and 59.0 years in EM patients (P=0.59). In all groups
there were more males (57.9%) than females (42.1%).
More EM patients (34.1%) had worse renal function (stage
B) in comparison to PS (16.8%) and MM patients (17.3%;
P<0.001). Patients with EM involvement (28.3%) were
more likely to have light chain disease compared to PS
(22.5%) and MM patients (22.1%; P=0.002). More detailed
patients’ characteristics are listed in Table 1.
Transplantation characteristics and responses
The source of stem cells for all patients was peripheral
blood. A total of 3391 patients underwent up-front single
ASCT and 353 patients up-front tandem ASCT; there was
no difference in time to first ASCT between the two
groups (P=0.81). Complete remission (CR) before the first
ASCT was reported in 21.5% PS, 11.7% EM and 19.1%
MM patients; partial remission (PR) was achieved by
72.6% PS, 79.6% EM and 74.7% MM patients (P=0.1)
(Table 1). After ASCT, complete response was achieved by
41.6% PS, 36.1% EM and 43.9% MM patients, while
54.0% PS, 51.9% EM and 49.8% MM patients showed
partial response (P=0.001).
EMD and survival
Median follow up was 36.3 months (range: 1-118.9
months) after the date of ASCT. In the univariate analysis,
the MM and PS groups showed similar 3-year PFS of
47.9% (95%CI: 45.8-50.1) versus 50.0% (95%CI: 44.6-55.3;
P=0.78) and similar 3-year OS of 80.1% (95%CI: 78.4-81.8)
versus 77.7% (95%CI: 73.3-82.1; P=0.09) (Figure 1A and B).
In contrast, EM patients had a significantly worse 3-year
PFS of 39.9% (95%CI: 30.3-49.5) in comparison to MM
(P=0.001) and PS patients (P=0.007), and a significantly
worse 3-year OS of 58.0% (95%CI: 48.1-67.9) compared
to MM and PS patients (P<0.001, respectively). Within the
EM group, 3-year PFS differed according to involved
organs: kidney (59.5%), skin (20.1%), lymph nodes
(37.6%), CNS (47.9%), lung/respiratory tract (44.4%),
haematologica | 2018; 103(5)
Autologous SCT for extramedullary myeloma
GI/liver (22.5%), and spleen, ovaries and testes (60.0%)
(Table 2).
Comparing the MM group without EMD to those with
EMD, one involved site resulted in a similar 3-year PFS of
49.4% (95%CI: 44.6-54.3; P=0.36) while multiple involved
sites showed a worse PFS of 22.7% (95%CI: 5.2-40.2;
P=0.001) (Figure 2A). Both one and multiple involved sites
showed worse 3-year OS rates of 73.5% (95%CI: 69.2-
77.7; P<0.001) and 71.4% (95%CI: 55.1-87.7; P=0.05) in
comparison to patients without EMD (80.1%) (Figure 2B).
After stratification of EMD groups according to one
versus multiple involved sites (PS1 vs. PS2, and EM1 vs.
EM2), PS patients showed no significant difference in 3-
year PFS of 50.5% (95%CI: 45.0-55.9%) versus 36.0%
(95%CI: 5.2-66.8%; P=0.71), and OS of 77.2% (95%CI:
72.7-81.7%) versus 91.7% (95%CI: 76.0-100; P=0.27). In
EM patients, this comparison resulted in a significantly
worse 3-year PFS of multiple sites in the univariate analy-
sis: 44.7% (95%CI: 34.1-55.3%) versus 13.9% (95%CI: 0-
35.5%; P=0.03) (Figure 3A). In contrast, 3-year OS was
58.7% (95%CI: 47.9-69.5%) for EM1 versus 57.5%
(95%CI: 34.2-80.8%; P=0.51) (Figure 3B).
Tandem transplantation and survival
A landmark analysis was used to compare tandem and
single ASCT, considering a total of 3139 patients who were
alive at six months. In patients without EMD, the compar-
ison of tandem versus single ASCT resulted in similar 3-
year PFS, 53.8% (95%CI: 46.7-60.9) versus 51.3% (95%CI:
48.9-53.7; P=0.37), and similar 3-year OS: 84.7% (95%CI:
79.6-89.8) versus 81.6% (95%CI: 79.8-83.4; P=0.26).
Patients with EMD showed a 3-year PFS of 59.0%
(95%CI: 46.3-71.8) after tandem versus 53.0% (95%CI:
47.5-58.6) after single (P=0.43) ASCT, while 3-year OS was
77.0% (95%CI: 66.1-87.9) versus 76.9% (95%CI: 72.4-81.4;
P=0.91).
Within each EMD group, PS patients showed a similar 3-
year PFS of 59.4% (95%CI: 45.3-73.6) after tandem
versus 54.3% (95%CI: 48.0-60.5; P=0.44) after single ASCT
and similar 3-year OS of 82.6% (95%CI: 72.3-92.8) versus
893
 N. Gagelmann et al.

80.3% (95%CI: 75.6-85.1; P=0.88). Patients with EM
involvement showed no significant difference in both 3-
year PFS and OS after tandem versus single transplantation:
56.2% (95%CI: 27.2-85.3) versus 48.3% (95%CI: 36.6-60.1;
P=0.98), and 52.0% (95%CI: 20.0-84.0) versus 64.9%
(95%CI: 54.2-75.7; P=0.39).
Role of other factors on survival and causes of death
All patients in CR before first ASCT showed a signifi-
cantly better 3-year PFS of 59.8% (95%CI: 55.3-64.3) com-
pared to 30.7% (95%CI: 28.2-33.2) in PR and 24.7%
(95%CI: 17.6-31.8; P<0.001) in less than PR. There was
also a significant difference in 3-year OS, with patients in
CR showing 83.6% (95%CI: 80.2-87.0) compared to
78.8% (95%CI: 76.9-80.6) in patients with PR and 27.8%
(95%CI: 20.8-34.9) in patients with less than PR (P<0.001).
Other factors associated with worse PFS in patients with
EMD were: older age (P=0.04), transplantation before 2011
(P=0.01), higher disease stage according to ISS (P=0.01) and
Salmon and Durie (P=0.02), and lower remission status at
transplantation (P<0.001). Factors associated with worse
OS in EMD patients were: transplantation before 2011
(P=0.02), higher disease stage according to ISS (P=0.002)
and Salmon and Durie (P=0.02), and lower remission status
at transplantation (P<0.001).
Non-relapse mortality at three years occurred in 3.0%
(95%CI: 2.0-4.0) of MM, 3.0% (95%CI: 2.0-5.0) of PS
patients, and 7.0% (95%CI: 2.0-12.0) of EM patients
(P=0.05). Main causes of death were relapse or progression
(86.3%), infection (7.1%), secondary malignancy or post-
transplant lymphoproliferative disorder (3.6%), organ
damage or failure (1.8%), toxicity (0.4%), and unknown in
83 patients.
Multivariate analyses
A multivariable model was constructed to examine the
effect of EMD on 3-year PFS and OS after adjusting for
possible prognostic factors. All factors and covariates
including corresponding references are listed in Table 3. To
avoid linearly dependent covariates, we merged the dis-
ease group and the new variable of the number of involved
sites into a 5-level variable consisting of patients without
EMD (MM group) and patients with EMD according to
number of involved sites (PS1, PS2, EM1 and EM2). Cox
proportional hazards regression considering independent
factors for worse PFS yielded significant results for EM2

Table 2. Involved sites in extramedullary organ involvement (EM) group and survival after autologous stem cell transplantation (ASCT).
Site N. of N. of deaths 3-year PFS 3-year OS
patients (%) in % (95% CI) in % (95% CI)
Kidney 38 (27.3) 7 59.5 (41.1 to 77.9) 75.3 (59.0 to 91.7)
CNS 14 (10.1) 4 47.9 (18.3 to 77.4) 64.3 (35.5 to 93.1)
Lung / respiratory tract 9 (6.5) 3 44.4 (7.4 to 81.5) 41.7 (0 to 85.1)
GI tract / liver 8 (5.8) 3 22.5 (0 to 58.8) 58.3 (22.0 to 94.7)
Pleura / heart 7 (5.0) 5 NE NE
Spleen / ovaries / testes 7 (5.0) 2 60.0 (17.1 to 100) 60.0 (17.1 to 100)
Skin 32 (23.0) 10 20.1 (3.4 to 36.7) 53.3 (30.5 to 76.0)
Lymph nodes 24 (17.3) 10 37.6 (16.4 to 58.7) 48.2 (25.1 to 71.3)
PFS: progression-free survival; OS: overall survival; N.: number; CI: Confidence Interval; CNS: central nervous system; GI: gastrointestinal; NE: not estimable.

Figure 2. Progression-free survival (A) and overall survival (B) with numbers at risk of myeloma patients following up-front autologous stem cell transplantation
according to number of involvements: 0, 1 and ≥ 2. N: number.

894 haematologica | 2018; 103(5)

 Autologous SCT for extramedullary myeloma

with an HR of 3.40 (95%CI: 1.74-6.61; P<0.001). Discussion

Interestingly, there was no difference in PFS in EM1 com-
pared to MM with an HR of 1.03 (95%CI: 0.66-1.62;
P=0.88). Comparison of PFS between PS and MM showed
no difference for PS1, with an HR of 1.02 (95%CI: 0.83-
1.27; P=0.86), and a less clear HR of 2.46 (95%CI: 0.92-
6.62; P=0.07) for PS2.
In the OS analysis, EM1 and EM2 were associated with
worse outcome, showing HRs of 2.30 (95%CI: 1.43-3.70;
P=0.001) and 3.64 (95%CI: 1.48-8.94; P=0.01). The differ-
ence between patients with one site of PS involvement and
those without EMD was less clear, with an HR of 1.33
(95%CI: 0.98-1.83; P=0.07), while PS2 resulted in a similar
outcome with an HR of 0.74 (95%CI: 0.10-5.32; P=0.77).
Tandem ASCT showed similar 3-year PFS and OS com-
pared to single ASCT, with HRs of 0.83 (95%CI: 0.66-1.06;
P=0.13) and 0.74 (95%CI: 0.51-1.09; P=0.13). However,
other factors made a significant contribution to an
increased risk of worse outcome (Table 3). For PFS, these
factors were: ISS stage II and III, PR and less than PR at
ASCT. OS was significantly influenced by ISS stage II and
III, male sex, PR and less than PR at ASCT, and the pres-
ence of heavy and light chains.
Extramedullary disease in patients with MM is consid-
ered a poor prognostic factor. This EBMT registry study
including 682 EMD patients identified an increase per year
of EMD incidence at diagnosis from 2005 to 2014. We
demonstrated that first-line ASCT resulted in at least simi-
lar 3-year PFS in patients with single sites of EMD com-
pared to patients without EMD. Another finding, even
though this was less clear, was that this translated into
worse 3-year OS in single PS involvement while single sites
of EM were significantly associated with worse outcome,
which worsened still further when multiple sites of organs
were involved. As far as treatment options for EMD at
diagnosis are concerned, we found both first-line tandem
and single ASCT resulted in similar 3-year PFS and OS.
Evidence on the role of EMD at diagnosis after first-line
ASCT is still limited. A retrospective single center study30
of 27 patients concluded that ASCT might overcome poor
prognosis at onset compared to patients without EMD,
while another study showed extramedullary organ
involvement in only 4 patients and that its impact on out-
come could be under-estimated.5 A prospective study31 of

Table 3. Multivariate analysis.

3-year PFS 3-year OS
Factors – reference Hazard ratio (95% CI) P Hazard ratio (95% CI) P
Group – MM without EMD < 0.001 < 0.001
PS1 1.02 (0.82 to 1.27) 0.86 1.33 (0.98 to 1.83) 0.07
PS2 2.46 (0.92 to 6.62) 0.07 0.74 (0.10 to 5.32) 0.77
EM1 1.03 (0.66 to 1.62) 0.88 2.30 (1.43 to 3.70) 0.001
EM2 3.40 (1.74 to 6.61) < 0.001 3.64 (1.48 to 8.94) 0.01
Sex – male
Female 0.86 (0.74 to 1.01) 0.06 0.71 (0.56 to 0.91) 0.01
Age in years, > 60 0.22 0.45
< 50 0.81 (0.74 to 1.03) 0.08 1.04 (0.73 to 1.48) 0.85
50 to 60 0.95 (0.80 to 1.11) 0.50 1.17 (0.91 to 1.50) 0.21
ISS – I < 0.001 < 0.001
II 1.48 (1.23 to 1.77) < 0.001 1.75 (1.29 to 2.37) < 0.001
III 1.81 (1.46 to 2.24) < 0.001 2.68 (1.92 to 3.74) < 0.001
Renal function – A
B 0.99 (0.79 to 1.24) 0.93 1.25 (0.92 to 1.69) 0.16
Status at ASCT – CR < 0.001 0.02
PR 1.58 (1.26 to 1.97) < 0.001 1.48 (1.05 to 2.08) 0.03
< PR 2.18 (1.54 to 3.10) < 0.001 2.08 (1.22 to 3.54) 0.01
Type of myeloma - light chain 0.10 0.09
Non-secretory 0.77 (0.42 to 1.43) 0.41 1.70 (0.80 to 3.61) 0.17
Heavy and light 1.19 (0.98 to 1.46) 0.08 1.38 (1.01 to 1.88) 0.04
Year of ASCT - > 2011 0.21 0.91
< 2009 1.22 (0.97 to 1.53) 0.09 1.07 (0.75 to 1.53) 0.71
2009 to 2011 1.05 (0.88 to 1.26) 0.61 1.00 (0.76 to 1.33) 0.98
Type of ASCT - single
Tandem 0.83 (0.66 to 1.06) 0.13 0.74 (0.51 to 1.09) 0.13
PFS: progression-free survival; OS: overall survival; CI: Confidence Interval; MM: patients without extramedullary disease; PS: patients with paraskeletal involvement arising from
bone lesions; PS1: patients with paraskeletal involvement having one involved site; PS2: patients with paraskeletal involvement and multiple involved sites; EM1: patients with
extramedullary organ involvement having one involved site; EM2: patients with extramedullary organ involvement and multiple involved sites; ISS: International Staging System;
CR: complete remission; PR: partial remission; ASCT: autologous stem cell transplantation.

haematologica | 2018; 103(5) 895

 N. Gagelmann et al.
Figure 3. Progression-free survival (A) and overall survival (B) with numbers at risk of myeloma patients with extramedullary organ involvement following up-front
autologous stem cell transplantation according to number of involvements: 1 and ≥ 2. EM: patients with extramedullary organ involvement; EM1: patients with one
site of EM; EM2: patients with two or more sites of EM; N: number.
patients in relapse with either soft-tissue or bone-related
involvement at a single institution found that bone-related
relapses were associated with better OS. However, treat-
ments before diagnosis of extramedullary relapse signifi-
cantly differed between groups. Since different types of
involvement were reported, this variable was examined
closely.
In our study, especially EM involvement in 139 MM
patients was associated with lower rate of CR before and
after ASCT, a higher frequency of ISS stage III, and worse
renal function. Importantly, the impact of the number of
involved sites on outcome in EMD at diagnosis had not
previously been described. We found 20% of all EMD
patients having multiple sites of involvement, which is in
line with previous reports (16%).13 Notably, the location
of further involvement was only paraskeletal in the PS
group and was also restricted to other organs in the EM
group.32
The use of radiation therapy might contribute to the dif-
ference in PFS and OS of patients with single sites of EMD
compared to patients without EMD, because it is consid-
ered effective in reducing progression in patients with soli-
tary osseous and extraosseous involvement,33,34 in particu-
lar because reports about the efficacy of novel agents in
these cohorts at diagnosis are very limited. Some results
would suggest an induction bortezomib-based regimen
followed by high-dose melphalan/ASCT for patients with
paraskeletal rather than extramedullary involvement.14,35-37
In a retrospective study38 investigating carfilzomib alone or
in combination as salvage therapy in relapse, presence of
extramedullary involvement resulted in shorter duration of
response compared to absent EMD, suggesting limited
treatment effect. Smaller reports on the possible impact of
immunomodulatory drugs showed partial efficacy regard-
ing response rates in EMD patients.10,39,40
Retrospective studies highlighted an extremely poor
prognosis for CNS involvement with a median OS of less
than six months.41,42 However, in addition to systemic anti-
MM therapy, CNS irradiation and the use of novel combi-
nation therapies have been shown to improve the duration
896
of response.42 With regard to these analyses outside trans-
plantation settings, we investigated survival according to
involved sites in EM patients, finding most of the patients
had kidney, skin or lymph node involvement. After up-
front ASCT, best outcomes were found in kidney and CNS
involvement while skin and lymph node involvement
showed worse outcome. Interestingly, our CNS cohort
showed higher rates of OS compared to previous reports,
which might be due to the selection of patients with CNS
involvement at diagnosis, while most reports evaluated
patients at later phases of the disease.41,42
A pooled analysis of prospective studies regarding trans-
plantation strategies suggested the superiority of tandem
ASCT in patients with poor prognostic features at diagno-
sis.4,43 Our landmark analyses of EMD patients who
received either tandem or single ASCT as first-line therapy
found no difference in PFS and OS. However, this analysis
was conducted with the use of retrospective data and is,
therefore, subject to the attendant limitations. Regression
modeling and landmark analyses were performed as a
means of controlling for differences between the patients,
but such adjustment cannot account for all discrepancies in
clinical and diagnostic characteristics between groups. The
increasing incidence of EMD might be caused by a more
frequent use of whole-body MRI or PET-CT in recent
years. However, although recent evidence promotes the
use of more sensitive imaging techniques,44 data are not
routinely documented, and they are still not part of routine
diagnostics and were thus not available in our study.45,46 A
randomized trial is the only way to overcome these chal-
lenges and to assess the definite impact of EMD in newly
diagnosed MM patients after ASCT.
In conclusion, this EBMT study identified an increase in
incidence per year of EMD in newly diagnosed MM
patients from 2005 to 2014. We revealed that first-line
ASCT in patients with single sites of EMD (PS or EM)
resulted in at least similar 3-year PFS compared to patients
without EMD. Nevertheless, single EM involvement was
associated with worse 3-year OS, which worsened still fur-
ther when multiple sites of organs were involved.
haematologica | 2018; 103(5)

References
1. National Cancer Institute. Surveillance epi-
demiology and end results program. SEER
stat fact sheets: Myeloma. 2013;
http://seer.cancer.gov/statfacts/html/mulmy
.html [Last accessed 14 April 2017].
2. Kumar SK, Rajkumar SV, Dispenzieri A, et
al. Improved survival in multiple myeloma
and the impact of novel therapies. Blood.
2008;111(5):2516-2520.
3. Palumbo A, Anderson K. Multiple myeloma.
N Engl J Med. 2011; 364(11):1046-1060.
4. Touzeau C, Moreau P. How I treat
extramedullary myeloma. Blood. 2016;
127(8):971-976.
5. Bladé J, Fernández de Larrea C, Rosiñol L,
Cibeira MT, Jiménez R, Powles R. Soft-tis-
sue plasmacytomas in multiple myeloma:
incidence, mechanisms of extramedullary
spread, and treatment approach. J Clin
Oncol. 2011;29(28):3805-3812.
6. Weinstock M, Aljawai Y, Morgan EA, et al.
Incidence and clinical features of
extramedullary multiple myeloma in
patients who underwent stem cell transplan-
tation. Br J Haematol. 2015;169(6):851-858.
7. Cavo M, Terpos E, Nanni C, et al. Role of
18F-FDG PET/CT in the diagnosis and man-
agement of multiple myeloma and other
plasma cell disorders: a consensus statement
by the International Myeloma Working
Group. Lancet Oncol. 2017;18(4):e206-e217.
8. Bartel TB, Haessler J, Brown TL, et al. F18-
fluorodeoxyglucose positron emission
tomography in the context of other imaging
techniques and prognostic factors in multi-
ple myeloma. Blood. 2009; 114(10):2068-
2076.
9. Usmani SZ, Heuck C, Mitchell A, et al.
Extramedullary disease portends poor prog-
nosis in multiple myeloma and is overrepre-
sented in high-risk disease even in the era of
novel agents. Haematologica.
2012;97(11):4761-4767.
10. Short KD1, Rajkumar SV, Larson D, et al.
Incidence of extramedullary disease in
patients with multiple myeloma in the era of
novel therapy and the activity of pomalido-
mide in extramedullary myeloma.
Leukemia. 2011;25(6):906-908.
11. Rosiñol L, Fernández de Larrea C, Bladé J.
Extramedullary myeloma spread triggered
by surgical procedures: an emerging entity?
Acta Haematol. 2014;132(1):36-38.
12. Pérez-Simón JA, Sureda A, Fernández-Aviles
F, et al. Reduced-intensity conditioning allo-
geneic transplantation is associated with a
high incidence of extramedullary relapses in
multiple myeloma patients. Leukemia.
2006;20(3):542-545.
13. Varettoni M1, Corso A, Pica G,
Mangiacavalli S, Pascutto C, Lazzarino M.
Incidence, presenting features and outcome
of extramedullary disease in multiple myelo-
ma: a longitudinal study on 1003 consecu-
tive patients. Ann Oncol. 2010;21(2):325-
330.
14. Rosiñol L, Oriol A, Teruel AI, et al.
Superiority of bortezomib, thalidomide, and
dexamethasone (VTD) as induction pre-
transplantation therapy in multiple myelo-
ma: a randomized phase 3 364.
PETHEMA/GEMstudy. Blood.
2012;120(8):1589-1596.
15. Usmani SZ, Rodriguez-Otero P, Bhutani M,
Mateos MV, Miguel JS. Defining and treat-
ing high-risk myeloma. Leukemia.
2015;29(11):2119-2125.
16. Sonneveld P, Avet-Loiseau H, Lonial S, et al.
Treatment of multiple myeloma with high-
haematologica | 2018; 103(5)
Autologous SCT for extramedullary myeloma
risk cytogenetics: a consensus of the
International Myeloma Working Group.
Blood. 2016;127(24):2955-2962.
17. Cavo M, Rajkumar SV, Palumbo A, et al.
International Myeloma Working Group con-
sensus approach to the treatment of multi-
ple myeloma patients who are candidates
for autologous stem cell transplantation.
Blood. 2011;117(23):6063-6073.
18. Bianchi G, Richardson PG, and Anderson
KC. Best Treatment Strategies in High-Risk
Multiple Myeloma: Navigating a Gray Area.
J Clin Oncol. 2014;32(20):2125-2132.
19. Bladé J, Samson D, Reece D, et al. Criteria
for evaluating disease response and progres-
sion in patients with multiple myeloma
treated by high-dose therapy and
haemopoietic stem cell transplantation.
Myeloma Subcommittee of the EBMT.
European Group for Blood and Marrow
Transplant. Br J Haematol. 1998;
102(5):1115-1123.
20. Greipp PR, San Miguel J, Durie BG, et al.
International staging system for multiple
myeloma. J Clin Oncol. 2005;23(15):3412-
3420.
21. Durie BG, Salmon SE. A clinical staging sys-
tem for multiple myeloma. Correlation of
measured myeloma cell mass with present-
ing clinical features, response to treatment,
and survival. Cancer. 1975;36(3):842-854.
22. Karnofsky DA, Abelmann WH, Craver LF,
Burchenal JH. The Use of the Nitrogen
Mustards in the Palliative Treatment of
Carcinoma – with Particular Reference to
Bronchogenic Carcinoma. Cancer.
1948;1(4):634-56.
23. Kaplan E, Meier P. Nonparametric estima-
tion from incomplete observations. J Am
Stat Assoc. 1958;53(282):457-481.
24. Schemper M, Smith TL. A note on quantify-
ing follow-up in studies of failure time.
Control Clin Trials. 1996;17(4):343-346.
25. Fine JP, Gray RJ. A proportional hazards
model for the subdistribution of a compet-
ing risk. J Am Stat Assoc. 1999;94(446):496-
509.
26. Gray RJ. A class of k-sample tests for com-
paring the cumulative incidence of a com-
peting risk. Ann Stat. 1988;16(3):1141-1154.
27. Iacobelli S, EBMT Statistical Committee.
Suggestions on the use of statistical method-
ologies in studies of the European Group for
Blood and Marrow Transplantation. Bone
Marrow Transplant. 2013;48(Suppl 1):S1-37.
28. Klein JP, Moeschberger ML (eds): Survival
Analysis: Techniques for Censored and
Truncated Data, 2nd edn. Springer-Verlag:
New York, USA, 2003.
29. Cox DR: Regression models and life tables. J
R Stat Soc B. 1972;34(2):187-220.
30. Lee SE, Kim JH, Jeon YW, et al. Impact of
extramedullary plasmacytomas on out-
comes according to treatment approach in
newly diagnosed symptomatic multiple
myeloma. Ann Hematol. 2015; 94(3):445-
452.
31. Pour L, Sevcikova S, Greslikova H, et al.
Soft-tissue extramedullary multiple myelo-
ma prognosis is significantly worse in com-
parison to bone-related extramedullary
relapse. Haematologica. 2014;99(2):360-
32. Ghobrial IM. Myeloma as a model for the
process of metastasis: implications for ther-
apy. Blood. 2012;120(1):20-30.
33. Hu K, Yahalom J. Radiotherapy in the man-
agement of plasma cell tumors. Oncology
(Williston Park). 2000;14(1):101-108.
34. Terpos E, Morgan G, Dimopoulos MA, et
al. International Myeloma Working Group
Recommendations for the Treatment of
Multiple Myeloma–Related Bone Disease. J
Clin Oncol. 2013;31(18):2347-2357.
35. Paubelle E, Coppo P, Garderet L, et al.
Complete remission with bortezomib on
plasmocytomas in an end-stage patient
with refractory multiple myeloma who
failed all other therapies including
hematopoietic stem cell transplantation:
possible enhancement of graft-vs-tumor
effect. Leukemia. 2005;19(9):1702-1704.
36. Sonneveld P, Goldschmidt H, Rosiñol L, et
al. Bortezomib-based versus nonborte-
zomib-based induction treatment before
autologous stem-cell transplantation in
patients with previously untreated multiple
myeloma: a meta-analysis of phase III ran-
domized, controlled trials. J Clin Oncol.
2013;31(26):3279-3287.
37. Mateos MV, Oriol A, Matínez-López J, et al.
Bortezomib, melphalan, and prednisone
versus bortezomib, thalidomide, and pred-
nisone as induction therapy followed by
maintenance treatment with bortezomib
and thalidomide versus bortezomib and
prednisone in elderly patients with untreat-
ed multiple myeloma: a randomised trial.
Lancet Oncol. 2010;11(10):934-941.
38. Muchtar E, Gatt ME, Rouvio O, et al.
Efficacy and safety of salvage therapy using
Carfilzomib for relapsed or refractory mul-
tiple myeloma patients: a multicentre retro-
spective observational study. Br J Haematol.
2016;172(1):89-96.
39. Calvo-Villas JM, Alegre A, Calle C, et al.
Lenalidomide is effective for
extramedullary disease in relapsed or refrac-
tory multiple myeloma. Eur J Haematol.
2011;87(3):281-284.
40. Rosiñol L, Cibeira MT, Bladé J, et al.
Extramedullary multiple myeloma escapes
the effect of thalidomide. Haematologica.
2004;89(7):832-836.
41. Nieuwenhuizen L, Biesma DH. Central
nervous system myelomatosis: review of
the literature. Eur J Haematol. 2008;80(1):1-
9.
42. Chen CI, Masih-Khan E, Jiang H, et al.
Central nervous system involvement with
multiple myeloma: long term survival can
be achieved with radiation, intrathecal
chemotherapy, and immunomodulatory
agents. Br J Haematol. 2013;162(4):483-488.
43. Cavo M, Salwender H, Rosiñol L, et al.
Double vs single autologous stem cell trans-
plantation after bortezomib-based induc-
tion regimens for multiple myeloma: an
integrated analysis of patient-level data
from phase European III studies. Blood.
2013;122:767.
44. Moreau P, Attal M, Caillot D, et al.
Prospective evaluation of magnetic reso-
nance imaging and [18F]
Fluorodeoxyglucose positron emission
tomography-computed tomography at
diagnosis and before Maintenance Therapy
in Symptomatic Patients With Multiple
Myeloma Included in the IFM/DFCI 2009
Trial: Results of the IMAJEM Study. J Clin
Oncol. 2017; 35(25):2911-2918.
45. Kumar S, Paiva B, Anderson KC, et al.
International Myeloma Working Group con-
sensus criteria for response and minimal
residual disease assessment in multiple
myeloma. Lancet Oncol. 2016;17(8):e328-
346.
46. Pawlyn C, Fowkes L, Otero S, et al. Whole-
body diffusion-weighted MRI: a new gold
standard for assessing disease burden in
patients with multiple myeloma?
Leukemia. 2016;30(6):1446-1448.
897
ARTICLE Hemostasis

Immobilized fibrinogen activates human

Ferrata Storti
Foundation platelets through glycoprotein VI
Pierre H Mangin,1 Marie-Blanche Onselaer,2 Nicolas Receveur,1 Nicolas Le
Lay,3 Alexander T Hardy,2 Clare Wilson,4 Ximena Sanchez,5 Stéphane Loyau,3
Arnaud Dupuis,1 Amir K Babar,6 Jeanette LC Miller,6 Helen Philippou,4 Craig E
Hughes,2,8 Andrew B Herr,6 Robert AS Ariëns,4 Diego Mezzano,5 Martine
Jandrot-Perrus,3,7 Christian Gachet1 and Steve P. Watson2,9
1
Université de Strasbourg, INSERM, EFS Grand-Est, BPPS UMR-S 1255, FMTS, France;
Haematologica 2018 2
Institute of Cardiovascular Sciences, IBR Building, College of Medical and Dental
Sciences, University of Birmingham, UK; 3Université de Paris Diderot, INSERM
UMR_S1148, Hôpital Bichat, Paris, France ; 4Thrombosis and Tissue Repair Group, Institute
Volume 103(5):898-907
of Cardiovascular and Metabolic Medicine, University of Leeds, UK; 5Laboratorio de
Hemostasia, Pontificia Universidad Catolica de Chile, Santiago, Chile; 6Division of
Immunobiology, Center for Systems Immunology & Division of Infectious Diseases,
Cincinnati, OH, USA; 7Acticor Biotech, Hôpital Bichat, INSERM, UMR-S 1148, Paris, France;
8
Institute for Cardiovascular and Metabolic Research, Harborne Building, University of
Reading, UK and 9Centre of Membrane Proteins and Receptors (COMPARE), Universities of
Birmingham and Nottingham, Midlands, UK

ABSTRACT

G
lycoprotein VI, a major platelet activation receptor for collagen
and fibrin, is considered a particularly promising, safe antithrom-
botic target. In this study, we show that human glycoprotein VI
signals upon platelet adhesion to fibrinogen. Full spreading of human
platelets on fibrinogen was abolished in platelets from glycoprotein VI-
deficient patients suggesting that fibrinogen activates platelets through
glycoprotein VI. While mouse platelets failed to spread on fibrinogen,
human-glycoprotein VI-transgenic mouse platelets showed full spreading
Correspondence: and increased Ca2+ signaling through the tyrosine kinase Syk. Direct bind-
pierre.mangin@efs.sante.fr or ing of fibrinogen to human glycoprotein VI was shown by surface plas-
s.p.watson@bham.ac.uk mon resonance and by increased adhesion to fibrinogen of human glyco-
protein VI-transfected RBL-2H3 cells relative to mock-transfected cells.
Received: October 20, 2017. Blockade of human glycoprotein VI with the Fab of the monoclonal anti-
Accepted: February 13, 2018.
body 9O12 impaired platelet aggregation on preformed platelet aggre-
gates in flowing blood independent of collagen and fibrin exposure.
Pre-published: February 22, 2018.
These results demonstrate that human glycoprotein VI binds to immobi-
lized fibrinogen and show that this contributes to platelet spreading and
doi:10.3324/haematol.2017.182972 platelet aggregation under flow.

Check the online version for the most updated

information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/898 Introduction

The immunoglobulin receptor glycoprotein (GP) VI is expressed on megakary-

©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
ocytes and platelets. GPVI associates with the Fc receptor (FcR) γ-chain in the
membrane, and with the Src family kinases (SFK) Lyn and Fyn through its cytosolic
tail.1 Ligand binding clusters GPVI at the platelet surface promoting phosphoryla-
tion of the immunoreceptor tyrosine-based motif (ITAM) of the FcR γ-chain by
SFK.2–4 This results in the recruitment of Syk and formation of a LAT-based signalo-
some that activates PLCγ2 leading to an increase in Ca2+, activation of integrin and
secretion of granules.5
GPVI is widely known as a platelet activation receptor for fibrillar collagen.5
However, in recent years, GPVI has been shown to bind to additional ligands,
including subendothelial and plasma adhesive proteins such as laminins and fib-
rin,6–8 the hormone adiponectin and the transmembrane protein emmprin.9,10
Several of these interactions are relatively weak and of unclear significance. For
example, GPVI supports adhesion and efficient activation of platelets to collagen

898 haematologica | 2018; 103(5)


and fibrin but is only involved in post-adhesive events on
laminin.6,11 Determining the importance of the interaction
of GPVI with each ligand in mediating platelet activation
in vivo will require development of selective inhibitors.
GPVI is involved in arterial thrombosis and in several of
the more recently discovered roles of platelets, including
maintenance of vascular integrity at sites of inflammatory
challenge.12 We and others have reported that the absence
of GPVI reduces experimental thrombosis in mouse mod-
els of atherosclerotic plaque rupture13,14 and abolishes
occlusive thrombus formation following FeCl3 injury.15 In
contrast, the absence or blockade of GPVI has a relatively
minor impact on hemostasis in mice16 and patients defi-
cient in GPVI have a relatively mild bleeding diathesis.17–19
These results highlight that GPVI is a promising anti-
thrombotic target with inhibitors predicted to have a
minor effect on hemostasis.20
Following ligand binding, GPVI stimulates signals that
convert integrin aIIbβ3 from a low to a high affinity state
for fibrinogen and other physiological ligands.21 Ligand
engagement of integrin aIIbβ3 has been reported to gen-
erate outside-in signals that are similar to those of GPVI,
including activation of Src and Syk kinases, PLCγ2 and
Ca2+ mobilization.22–24 Paradoxically, however, human but
not mouse platelets generate extensive lamellipodial
sheets and stress fibers on fibrinogen whereas on colla-
gen, which stimulates similar signals, full spreading of
platelets is seen in both species.25 One explanation for
this difference is the presence of the low affinity immune
receptor, FcγRIIA, in the human but not the rodent
genome, as FcγRIIA-transfected transgenic mouse
platelets exhibit increased spreading and Syk activation
upon adhesion to fibrinogen, although the increase in
spreading is only partial.26–28 Outside-in signaling by
aIIbβ3 is also mediated by two conserved tyrosines pres-
ent in a NxxY motif in the integrin β3 intracytoplasmic
domain independent of Src and Syk activation. Mutation
of these two tyrosine residues to phenylalanine leads to
a re-bleeding diathesis that has been attributed to a
defect in clot retraction.29 This shows that engagement of
integrin aIIbβ3 leads to activation of multiple signaling
pathways.
In the present study, we showed that full spreading of
human platelets on fibrinogen is abolished in patients
deficient in GPVI and that transgenic mouse platelets
expressing human GPVI, in contrast to wild-type
platelets, spread fully on fibrinogen. Direct binding of fib-
rinogen to human GPVI was demonstrated using human
GPVI-transfected cell lines and recombinant GPVI.
Inhibiting the binding of fibrinogen to GPVI limited
platelet aggregation under conditions that excluded
involvement of collagen and fibrin.
Methods
Patients
Family 1 and family 2 are two families who are heterozygous or
homozygous for an adenine insertion in exon 6 of GP6 that gen-
erates a premature stop codon in position 242 of the protein.17
They have been described previously.30 Patient 3 is a 10-year old
boy suffering from an autoimmune disease with anti-GPVI anti-
bodies. The platelets of this patient do not aggregate to collagen
and GPVI is not detected at the platelet surface using flow cytom-
etry and western blot (data not shown).
haematologica | 2018; 103(5)
Fibrinogen, GPVI and platelet activation
Mice
Wild-type mice were generated from breeding of heterozy-
gotes or purchased from Harlan Laboratories (Hillcrest, UK) or
Charles River (Lyon, France). GPVI-/- mice were provided by Dr
Jerry Ware.31 Mouse platelets expressing human GPVI have
been described and characterized previously.32 Syk chimera
mice have been described previously.27 Ethical approval for ani-
mal experimentation was obtained from the French Ministry of
Research and UK Home Office in accordance with the
European Union guidelines, the Guide for the Care and Use of
Laboratory Animals.
Reagents
PRT-060318 was obtained from Caltag Medsystems
(Buckingham, UK), REOPRO from E. Lilly (Indianapolis, IN,
USA). Recombinant GPVI was made as described elsewhere.33
Fibrinogen was from Kabi (Bad Homburg, Germany) or from
ERL (South Bend, IN, USA). RAM.1 (anti-GPIbβ) was generated
in U949.34 The blocking Fab fragment of monoclonal antibody
directed against human GPVI, 9O12.2, and its humanized ver-
sion are referred to as 9O12 in this manuscript.35,36 All other
reagents were from previously described sources.11,37 The anti-
fibrin antibody 59D8 was obtained from CT Esmon (Oklahoma
Medical Research Foundation, OK, USA).38
Generation and characterization of RBL-2H3 cells
The cDNA of WT human GPVI33 was inserted in pSRαNeo
between the 5’-XhoI and 3’-BamHI restriction sites. Rat
basophilic leukemia cells, RBL-2H3, were cultured in Dulbecco
modified Eagle medium supplemented with 10% fetal bovine
serum albumin and stably transfected with 1 μg of DNA corre-
sponding to the empty vector or WThGPVI vector mixed with
FuGENE6 (Roche, Boulogne-Billancourt, France) and selected in
growth medium containing G418 0.7%; 1 mg/mL geneticin
(GibcoBRL, Invitrogen, Cergy Pontoise). Cell surface expression
of recombinant GPVI and constitutively expressed integrin
aIIbβ3 was confirmed by flow cytometry and immunoblot
(data not shown).
Cell adhesion to fibrinogen
LAB TEK 4 wells were coated with 400 μL/well of collagen
(50 µg/mL) or fibrinogen (100 μg/mL) overnight at 37°C. Wells
were saturated with human serum albumin (10 mg/mL) for 1 h
at 37°C. Trypsinised RBL cells (3x105 cells/mL) were incubated
with phosphate-buffered saline or 9O12 (50 μg/mL) and/or
REOPRO (40 μg/mL) for 15 min at 37°C. Subsequently, 100,000
cells were added to the wells (300 μL) for 1 h at 37°C. After
three washing steps, cells were fixed with 400 μL paraformalde-
hyde 4% for 20 min. Pictures were taken with an EVOS optic
microscope (x10). Actin was stained with Alexa-488-phalloidin
and the nucleus with DAPI.
Washed platelets
Human blood was taken from patients or from healthy
donors using 3.8% (v/v) sodium citrate (1:9) as the anticoagu-
lant. Human and mouse washed platelets were obtained by
centrifugation using prostacyclin (2.8 μmol/L) and resuspended
in modified Tyrode-HEPES buffer as described elsewhere.37,39
Platelet spreading
Platelet adhesion to immobilized fibrinogen was achieved as
described previously.37 Images of the platelets were obtained
with a Zeiss Axiovert 200 mol/L microscope or with a Leica
DMI400 microscope. Platelet surface area was analyzed using
ImageJ software (NIH, Bethesda, USA).
899
 P.H. Mangin et al.
Western blotting
For stimulation on fibrinogen, washed platelets were pre-treat-
ed with 10 μmol/L indomethacin and 2 U/mL apyrase. Platelets
(1.5 mL containing 5x108/mL) were allowed to adhere to 10 cm
dishes coated with 100 μg/mL fibrinogen or heat-inactivated
bovine serum albumin for 45 min at 37°C. Non-adherent platelets
were removed and lysed by addition of 2X lysis buffer (150
mmol/L NaCl, 10 mmol/L Tris, 1 mmol/L EGTA, 1 mmol/L EDTA,
1% NP-40; pH 7.4, plus 1.25 mmol/L Na3VO4, 50 μg/mL AEBSF,
2.5 μg/mL leupeptin, 2.5 μg/mL aprotinin and 0.25 μg/mL pep-
statin). Adherent platelets were washed twice with Tyrode buffer
then lysed with 1X lysis buffer on ice for 15 min before scraping.
Proteins were immunoprecipitated with a-Syk antibody and pro-
tein A-sepharose beads for 2 h. The beads were washed, proteins
eluted in sodium dodecyl sulfate (SDS) sample buffer, separated
by SDS-polyacrylamide gel electrophoresis (PAGE), electro-trans-
ferred, and western blotted with the stated antibodies. For whole
platelet lysates, washed platelets (5x108/mL) were lysed directly
with an equal volume of 2xSDS sample buffer, separated by SDS-
PAGE, electro-transferred, and western blotted with the stated
antibodies.
Ca2+ assay and in vitro perfusion assay
Intraplatelet Ca2+ concentrations following platelet adhesion to
fibrinogen were measured using a dual-dye ratiometric method
and hirudinated blood perfusion was performed as previously
described.40 Three-dimensionsal reconstructed images were
obtained using the 3D module of Leica LAS X software.
Solid-based binding assay
Binding studies were performed with the recombinant proteins,
GPVI-Fc fusion (dimer) and GPVI-His tagged (monomer). Cover
slips were coated with collagen or fibrinogen overnight at 4°C.
The plates were blocked with 3% bovine serum albumin – phos-
phate-buffered saline for 1 h and washed prior to addition of
monomeric or dimeric GPVI at a concentration of 100 nmol/L for
1 h. After washing, 4 μg/mL of secondary antibodies, horseradish
peroxidase (HRP)-conjugated goat anti-human IgG Fc or HRP-con-
jugated anti-His Tag, were added for 1 h. GPVI binding was
detected using 3,3′,5,5′-tetramethylbenzidine. The reaction was
stopped with H2SO4 (2 mol/L) and absorbance was measured at
450 nm with a spectrofluorometer.
Surface plasmon resonance
Surface plasmon resonance was performed on a Pioneer plat-
form from PALL® FortéBio® (Portsmouth, UK). IF-1 purified fib-
rinogen was diluted to 100 μg/mL using 10 mmol/L NaAc pH 5.0.
IF-1 fibrinogen was adsorbed to the chip surface via amine cou-
pling to a level of 3825 resonance units (RU) at flow-cell 1 and
3423 RU on flow-cell 3. Flow-cell 2 was activated using amine
coupling and blocked using 1 mol/L ethanolamine and was the
designated reference channel. GPVI analytes were dialysed and
diluted to 1 µmol/L using the same batch of running buffer as used
for the blanks. Analytes were injected using the OneStep® titration
function at a flow rate of 30 μL/min with a 100% loop-inject and
400 s dissociation. The chip surface was regenerated by flushing
with 1 mol/L NaCl at 60 μL/min for 10 s, followed by a further 400
s dissociation. Qdat data analysis software (PALL® FortéBio®, UK)
was used to analyze the data. Binding data were fitted using a one
site KA/KD model and analyte aggregation parameters adjusted per
binding curve according to goodness of fit and curve type.
Statistics
The statistical analyses were performed using the GraphPad
Prism program, version 5.0 (Prism, GraphPad, LaJolla, CA, USA).
900
The values are indicated as mean ± standard error of the mean
(SEM). The statistical analysis is described in the Figure legends.
Results
Abolition of spreading on fibrinogen in glycoprotein
VI-deficient human platelets
Human platelets undergo robust spreading on immobi-
lized fibrinogen, generating lamellipodial sheets and stress
fibers.41 This is illustrated in Figure 1A with over 90% of
platelets from a control donor undergoing full spreading
on fibrinogen over 30 min; the small number of partially-
spread platelets most likely represent newly adhered cells.
In 2013, Matus et al. described four unrelated families with
index cases who are homozygous for an adenine insertion
in exon 6 of human GP6, which leads to a premature ‘stop
codon’ in position 242 prior to the transmembrane
domain.17 All four homozygous patients lack expression of
GPVI on their platelets and heterozygous relatives express
approximately 50% of the receptor. Since then, two fur-
ther unrelated families with the same mutation have been
identified by the same group and also been shown to lack
surface expression of GPVI with absent platelet aggrega-
tion to collagen.30 Unexpectedly, in studying platelets from
two unrelated index cases in these families, we observed
reduced adhesion on immobilized fibrinogen and a failure
to form lamellipodial sheets and stress fibers (Figure 1A).
The absence of GPVI was confirmed by flow cytometry
and by abolition of aggregation to collagen but not to
other agonists in both cases30 (data not shown). In contrast,
spreading and adhesion of platelets from heterozygote
carriers from each family and platelets from a control were
similar (Figure 1A). The same result was also seen in a
patient with an auto-immune thrombocytopenia associat-
ed with the absence of GPVI expression (Figure 1Bi).
Adhesion of platelets was blocked by the aIIbβ3 receptor
antagonist, REOPRO (Figure 1B), as previously shown in
controls. These results demonstrate that adhesion of
human platelets on fibrinogen is critically dependent on
integrin aIIbβ3 with a minor contribution from GPVI, but
that full spreading requires GPVI.
Mouse platelets expressing human glycoprotein VI
undergo full spreading on fibrinogen
Mouse platelets adhere and undergo limited spreading
on human or mouse fibrinogen, forming filopodia and
limited lamellipodia but not stress fibers (Figure 2A). A
similar response is seen in platelets deficient in GPVI
(Figure 2A), whereas adhesion is abolished in the absence
of the integrin β3-subunit (Figure 2B). A similar level of
adhesion is seen in human GPVI transgenic mouse
platelets but is associated with the formation of lamellipo-
dial sheets and stress fibers (Figure 2C). These results
demonstrate that full spreading but not adhesion of mouse
platelets is dependent on human GPVI and not mouse
GPVI. One potential explanation for these results is that
human but not mouse GPVI is able to bind to fibrinogen
and mediate platelet activation.
Fibrinogen binds to monomeric human glycoprotein VI
To test whether fibrinogen is able to bind to GPVI,
increasing concentrations of recombinant soluble GPVI
extracellular domain, expressed either as a monomer
(GPVI-ex) or dimer (GPVI-Fc), was flowed over immobi-
haematologica | 2018; 103(5)

lized fibrinogen and binding monitored by surface plas-
mon resonance. As shown in Figure 3A, clear binding of
monomeric GPVI (ka = 1.17 ± 0.01 x 104 M-1s-1) was
observed with a kd of 3.94 ± 0.01 x 10-3 s-1. Binding was fit-
ted to a single site with an equilibrium dissociation con-
stant (KD) of 336 ± 1 nmol/L. In contrast, binding of dimer-
ic GPVI to fibrinogen was not detected at concentrations
up to 1 μmol/L (Figure 3Ai). In a second approach, fibrino-
gen was immobilized on a plastic surface and a solid
phase binding assay was performed. There was increased
binding of monomeric GPVI, but not dimeric GPVI, which
was inhibited by D-dimer (Figure 3Aii) where the binding
motif in fibrin resides.30 To further investigate the ability
of GPVI to bind to fibrinogen, we transfected rat RBL-2H3
basophilic cells, which constitutively express integrin
aIIbβ3 at low levels, with human GPVI and studied adhe-
sion to immobilized fibrinogen. We observed a 3-fold
increase in adhesion of GPVI-transfected cells to fibrino-
gen and to collagen relative to the adhesion of mock-trans-
fected control cells (Figure 3Bi, ii). RBL-2H3 cells express-
ing human GPVI also formed stress fibers upon adhesion
to fibrinogen. The human GPVI-blocking monoclonal
B GPVI on platelets (Patient 3) were
haematologica | 2018; 103(5)
Fibrinogen, GPVI and platelet activation
antibody 9012 Fab blocked the increase in adhesion.
Blocking the integrin aIIbβ3 with REOPRO reduced cell
adhesion to immobilized fibrinogen to the same level as
9O12 Fab, with no further inhibition in the presence of
both inhibitors (data not shown), indicating the presence of
additional binding proteins for fibrinogen in the adherent
cell line although binding to these was not sufficient to
induce spreading (Figure 3Biii and not shown). These
results demonstrate that GPVI binds to immobilized fib-
rinogen and is able to contribute to cell adhesion.
Spreading of human platelets but not mouse platelets
is dependent on Syk
The formation of lamellipodial sheets and stress fibers
in human platelets on fibrinogen and collagen is blocked
by the inhibitors of Src and Syk tyrosine kinases, PP2 and
PRT060318, respectively (Figure 4Ai & ii). Adhesion of
human platelets to fibrinogen induces phosphorylation of
Syk which co-precipitates with the phosphorylated FcR γ-
chain (Figure 4Aiii). These results provide further evidence
of GPVI activation in human platelets by immobilized fib-
rinogen. In contrast, the morphological modifications of
Figure 1. Glycoprotein VI supports
platelet adhesion and spreading on
immobilized fibrinogen. (A, B) Washed
human platelets were allowed to
adhere to immobilized fibrinogen (10
μg/mL) for 30 min. (A)(i).
Representative epifluorescence
images of fibrinogen-adherent
platelets from healthy donors (Control)
or members of a family with a mutation
in the GP6 gene (Heterozygotes: Family
1 +/-; homozygotes: Family 1 -/-). Scale
bars represent 10 μm. (A)(ii). Bar graph
representing the percentage of
platelets spreading on fibrinogen.
Spreading is expressed as the
mean±SEM in five random fields, in
two separate experiments (one-way
ANOVA, Kruskal-Wallis post-hoc test,
***P<0.0002; ****P<0.0001).
(A)(iii). Bar graph representing the
number of platelets adhering to immo-
bilized fibrinogen per mm² of a control
(Control) and two families with a muta-
tion in the GP6 gene. Adhesion is
expressed as mean±SEM in five ran-
dom fields, in two separate experi-
ments (one-way ANOVA, Kruskal-Wallis
post-hoc test, **P<0.002;
****P<0.0001). (B). Washed platelets
from a control or a patient with an
immune thrombocytopenic purpura
presenting with undetectable levels of
allowed to adhere to fibrinogen (100
μg/mL). (B)(i). Representative epifluo-
rescence images of washed platelets
from patient 3 adhering to immobilized
fibrinogen for 30 min, in the presence
or absence of REOPRO (40 μg/mL).
Scale bars represent 10 μm. (B)(ii). Bar
graph represents the number of
platelets adhering to fibrinogen per
mm2. Adhesion is expressed as
mean±SEM in eight random fields, in
two separate experiments.
901
 P.H. Mangin et al.
mouse platelets on fibrinogen is blocked by the Src kinase
inhibitor PP2 but not by the Syk kinase inhibitor
PRT060318 (Figure 4B). Morphological changes of mouse
platelets on fibrinogen are also not altered in platelets
from irradiated mice transplanted with Syk-deficient fetal
liver (Figure 4B) or from PF4.Cre-Sykfl/fl mice (Online
Supplementary Figure S1). Western blotting for Syk con-
firmed lack of expression of the tyrosine kinase in the two
transgenic models (Figure 4B and not shown). Thrombin
stimulated full spreading of wild-type and Syk-deficient
platelets (Figure 4B). Formation of lamellipodia and stress
A
B fibrinogen per mm². Adhesion is
C fibrinogen per mm². Adhesion is
902
fibers in human GPVI transgenic mouse platelets was
blocked by PRT060318 (Figure 4C). The ability of Src and
Syk inhibitors to block spreading of human platelets and
human GPVI transgenic mouse platelets on fibrinogen is
consistent with platelet activation by GPVI. This is sup-
ported by demonstration of phosphorylation of the FcR γ-
chain. The limited spreading of mouse platelets on fibrino-
gen is mediated through a Src-dependent but Syk-inde-
pendent pathway. Together, these results support a model
in which immobilized fibrinogen activates human but not
mouse platelets through GPVI.
Figure 2. Human but not mouse glyco-
protein VI supports platelet adhesion
and spreading on immobilized fibrino-
gen. (A). Washed platelets from wild-
type mice (WT mice), GPVI-deficient
mice (GPVI-/- mice) or from healthy
donors (Human) were allowed to adhere
to human or mouse fibrinogen (FGN) for
30 or 45 min, respectively, and fixed
with PFA and stained with Alex-488-
phalloidin (4 μg/mL). (A)(i).
Representative epifluorescence images
of washed platelets adhering to fibrino-
gen. Scale bars represent 10 µm. (A)(ii).
Bar graph representing the number of
platelets adhering to immobilized fib-
rinogen per mm². Adhesion is expressed
as mean±SEM in five random fields, in
three separate experiments (two-way
ANOVA, Bonferroni post-hoc test:
P>0.05). (A)(iii). Bar graph representing
the percentage of platelets spreading
on immobilized fibrinogen. Spreading is
expressed as the mean±SEM in five ran-
dom fields, in six separate experiments.
Significance was attained using a two-
way ANOVA, Bonferroni post-hoc test:
****P<0.001. (B). Washed control (WT)
or β3-deficient (β3-/-) platelets were
allowed to adhere to fibrinogen for 60
min, fixed with PFA and stained with
TRITC-phalloidin (2 μg/mL). (B)(i).
Representative epifluorescence images
of washed mouse platelets adhering to
fibrinogen. Scale bars represent 10 μm.
(B)(ii). Bar graph representing the num-
ber of platelets adhering to immobilized
expressed as the mean±SEM in eight
random fields, in four separate experi-
ments (Mann-Whitney test,
**P<0.001). (C). Washed platelets from
wild-type mice (WT mice) or mice
expressing human GPVI (hGPVI mice)
were allowed to adhere to fibrinogen for
60 min, fixed with PFA and stained with
TRITC-phalloidin (2 μg/mL). (C)(i).
Representative epifluorescence images
of washed platelets adhering to fibrino-
gen. Scale bars represent 10 μm. (C)(ii).
Bar graph (left) representing the num-
ber of platelets adhering to immobilized
expressed as the mean±SEM in eight
random fields, in four separate experi-
ments (Mann-Whitney test, P>0.05).
Bar graph (right) representing the per-
centage of platelets spreading on immo-
bilized fibrinogen. Spreading is
expressed as the mean±SEM in eight
random fields, in four separate experi-
ments (Mann-Whitney test,
**P<0.001).
haematologica | 2018; 103(5)

Fibrinogen stimulates an increase of Ca2+ in human
glycoprotein VI-transgenic mouse platelets
The observation that platelets expressing human but
not mouse GPVI undergo full spreading suggests that sig-
nals from human GPVI are of significance. To investigate
this, a dual-dye Ca2+ assay was used to monitor cytoplas-
mic Ca2+ levels as a marker of PLCγ2 activation. Analysis
of single platelet Ca2+ profiles by confocal microscopy
highlighted that signals generated on fibrinogen are com-
posed of Ca2+ spikes (Figure 5A). The number of Ca2+
spikes in mouse platelets expressing human GPVI was sig-
nificantly increased relative to the number in wild-type
haematologica | 2018; 103(5)
Fibrinogen, GPVI and platelet activation
platelets and was blocked in the presence of PRT-060318
(Figure 5Ai-ii) highlighting the critical role of Syk in Ca2+
mobilization. These results demonstrate that human GPVI
stimulates Ca2+ signaling in fibrinogen-adherent mouse
platelets.
Fab 9O12 blocks aggregate growth of humanized
glycoprotein VI mouse platelets
Fibrinogen plays a critical role in hemostasis and arterial
thrombosis through crosslinking of platelets in the grow-
ing thrombus. In addition, we now show that fibrinogen
induces platlet activation by GPVI. To establish whether
Figure 3. Monomeric but not dimeric GPVI binds to
immobilized fibrinogen and supports cellular adhe-
sion. (A)(i). IF-1 purified fibrinogen was immobilised
to a COOH-V chip surface using amine coupling
covalent linkage, to a level of 3,825 RU. Monomeric
or dimeric GPVI was titrated over three to four orders
of magnitude across the chip surface to a maximum
of 1 µmol/L, generating a single binding curve (1 RU
represents the binding of approximately 1 pg/mm2
protein). The graphs shown are representative of
three repeats and the KD is expressed as mean ±
SEM. (A)(ii). Solid-phase binding assays were per-
formed in Nunc maxisorb 96-well plates coated
overnight with BSA and fibrinogen (FGN).
Monomeric or dimeric GPVI (100 nmol/L) was incu-
bated as described. Bound GPVI was detected using
HRP-coupled to an anti-6×His monoclonal antibody
for monomeric GPVI or an anti-human IgG for dimer-
ic GPVI. The histogram (mean ± SD) shows the
results from five independent experiments.
Significance was determined using a one-way
ANOVA, Dunnett post-hoc test : *P<0.05. (B). RBL-
2H3 cells (3x105 cells/mL; 300 μL) transduced with
an empty vector (RBL-2H3 Ctrl) or with human full-
length GP6 cDNA (RBL-2H3 huGPVI) were pre-incu-
bated with PBS, REOPRO (20 μg/mL), and 9O12 (50
μg/mL) and allowed to adhere to immobilized fib-
rinogen for 1 h at 37°C, in 5% CO2. After three gentle
washing steps, cells were permeabilised with TRI-
TON x100 0.2% and stained with Alexa-568 phal-
loidin (1.5 U/mL). (B)(i). Representative epifluores-
cence images of RBL cells adhering to immobilized
fibrinogen. (B)(ii). and (B)(iii). Cells were manually
counted in six to eight different experiments
(*P<0.05 Mann Whitney t-test in (ii) and one-way
ANOVA was followed by the Bonferroni multiple com-
parison test).
903
 P.H. Mangin et al.
activation of GPVI by platelet-bound fibrinogen partici-
pates in platelet aggregation we performed an in vitro flow
adhesion assay under conditions that prevent activation of
GPVI by collagen and by fibrin. To achieve this, we gener-
ated a platelet aggregate over type I fibrillar collagen using
hirudin-treated blood to prevent formation of fibrin. We
B
C body to confirm equal protein loading. Blots
904
then perfused additional blood from the same donor over
the aggregate at a wall shear rate of 300 s-1 in the presence
or absence of the Fab fragment of the GPVI blocking mon-
oclonal antibody 9O12. As expected, we were unable to
detect the presence of fibrin in the aggregate using a spe-
cific antibody (data not shown). The aggregate continued to
Figure 4. Syk promotes platelet spreading
on fibrinogen downstream of glycoprotein
VI but not in integrin aIIbβ3 outside-in sig-
naling. Washed human platelets, or
washed platelets from wild-type mice (WT
mice), Syk chimera mice (Syk chimera mice)
and mice expressing human GPVI (hGPVI
mice) were allowed to adhere to fibrinogen
or to collagen in the presence of either the
Src inhibitor PP2 (20 μmol/L), the Syk
inhibitor (5 μmol/L), thrombin (0.1 U/mL),
or vehicle control, for 30 min (human
platelets) or 45 min (mouse platelets) at
37°C followed by fixation with PFA. (A)(i).
Representative DIC images of human
platelets adhering to fibrinogen or collagen.
Scale bars represent 5 μm. (A)(ii). Bar graph
representing the surface area of platelets
spreading on fibrinogen. Spreading is
expressed as the mean±SEM in five or more
random fields, in three separate experi-
ments. Significance was determined by one-
way ANOVA and Bonferroni multiple compar-
ison test: *P<0.01. (A)(iii) Representative
western blot from human platelets following
adhesion to fibrinogen. Following immuno-
precipitation of Syk, proteins were separat-
ed by sodium dodecyl sulfate-polyacry-
lamide gel electrophoresis and western blot
for phosphotyrosine. Membranes were then
stripped and reprobed for Syk to confirm
equal protein loading. Blots are representa-
tive of three separate experiments. (B)(i).
Representative DIC images of mouse
platelets adhering to fibrinogen. Scale bars
represent 5 μm. (B)(ii). Bar graph represent-
ing the surface area of platelets spreading
on fibrinogen. Spreading is expressed as
the mean±SEM in five or more random
fields, in three separate experiments.
Significance was determined using one-way
ANOVA, with the Bonferroni post-hoc test:
*P<0.05, **P<0.001. (B)(iii). Expression of
Syk was measured by western blotting
platelet whole-cell lysates with an a-Syk
antibody. Membranes were then stripped
and reprobed with an anti-a-tubulin anti-
are representative of three separate experi-
ments. (C)(i). Representative epifluores-
cence images of mouse platelets adhering
to fibrinogen. Scale bars represent 10 µm.
(C)(ii). Bar graph representing the surface
area of platelets spreading on fibrinogen.
Spreading is expressed as the mean±SEM
in eight random fields, in five separate
experiments (Mann-Whitney test,
**P<0.01).
haematologica | 2018; 103(5)
 Fibrinogen, GPVI and platelet activation

grow in the presence of a Fab control but was dramatically gen to bind GPVI in suspension due to conformational dif-
inhibited in the presence of Fab 9O12 (Figure 6A and ferences between circulating and immobilized fibrinogen.
Online Supplementary Figure S2). In contrast, and as previ- Alternatively it may be due to the inability of the dimeric
ously shown, blockade of GPVI did not impair aggrega- fibrinogen to cluster GPVI on the platelet surface in sus-
tion measured by light transmission aggregometry in pension or because of a dependency on binding to integrin
response to ADP, U46619 and thrombin.35 These results aIIbβ3. While the affinity of fibrinogen for GPVI is in the
demonstrate a critical role for GPVI in aggregate growth range of that for collagen for GPVI,44,45 we have shown that
under flow when the roles of collagen and fibrin are negat- fibrinogen (and fibrin) bind selectively to monomeric
ed. In contrast, fibrinogen does not induce activation of GPVI and this would not be sufficient to induce activation
platelets in suspension either because the interaction is because of the absence of crosslinking. The reason why
dependent on activation of integrin aIIbβ3 or because it human platelets, but not mouse platelets, spread on fib-
cannot crosslink GPVI. rinogen is unclear. Based on the fact that human and
mouse GPVI share 64% homology,33 one could imagine
that only human GPVI binds to fibrinogen or that both
Discussion bind to this adhesive protein but only human GPVI is able
to promote activation.
In this study we show that human GPVI binds to immo- GPVI is primarily known as the major signaling receptor
bilized fibrinogen and that this leads to intracellular sig- for collagen. However, in recent years, GPVI has been
nals, which drive the formation of lamellipodial sheets shown to bind to other ligands including laminin, the
and stress fibers in human platelets and in human GPVI- transmembrane protein emmprin, adiponectin, histones
expressing mouse platelets. This explains the previously and fibrin.6-8,10,47 The physiological significance of many of
paradoxical observation that only human platelets form these interactions is uncertain, in part because of their low
lamellipodial sheets and stress fibers on a fibrinogen sur- affinity or because of whether they acutally occur in vivo.
face, despite mouse platelets being able to form both actin The interaction that has received the greatest attention is
structures in the presence of G protein-coupled receptor that with fibrin which lies at the interface of the core and
agonists such as thrombin. We also show that the interac- shell of the growing platelet aggregate.7,8 This interaction
tion of fibrinogen with GPVI is important for aggregate
growth providing a new understanding of hemostasis and
thrombosis.
We recently identified fibrin as a novel ligand for
GPVI7,8,42 and have shown that binding resides in the D- A
dimer region.30 The observation that fibrinogen also acti-
vates GPVI should not, therefore, be a surprise.
Nevertheless, this was unexpected and came from the
observation that human platelets deficient in GPVI adhere
to but do not spread on fibrinogen. This raises the ques-
tion as to why this has been previously overlooked. One
reason is that mouse platelets do not spread on fibrinogen
and thus there is no defect in the absence of GPVI. A sec-
ond reason is the low level of phosphorylation of the FcR
γ-chain induced by fibrinogen in human platelets relative
to that by collagen and other GPVI-agonists. This may
reflect the extent to which each ligand is able to cluster
GPVI and, in the case, of fibrinogen, the dependency on
the interaction with integrin aIIbβ3. A third reason is that
fibrinogen binds selectively to monomeric GPVI whereas
the original binding studies were performed with dimeric
GPVI.7,8 It is worth noting that we have reported a reduced
number of dimers on immobilized fibrinogen relative to
collagen.42
Adhesion of human platelets to fibrinogen is dependent
on integrin aIIbβ3. At present, it is not known whether
binding to integrin aIIbβ3 is critical for activation of GPVI
or simply to promote adhesion such that activation of
GPVI can occur. As a dimer, fibrinogen should be able to
Figure 5. Human glycoprotein VI, but not integrin aIIbβ3 plays a major role in
the regulation of the calcium signaling after platelet adhesion to fibrinogen.
bind two GPVI monomers, but alternatively the interac-
tion with integrin aIIbβ3 may be required to support acti- Washed platelets from wild-type (WT) or mice expressing human GPVI (hGPVI)
vation of monomeric GPVI. A similar role for an integrin were loaded with Oregon-green Bapta-1-AM and Calcein red orange and
in the activation of an ITAM receptor has been reported in deposited on immobilized fibrinogen (100 μg/mL). Modifications in fluores-
cence of individual adherent platelets were monitored for 7 min by confocal
microscopy and the Ca2+ concentrations were determined as detailed in the
other hematopoietic cells with the postulate that the inte-
grin and the ITAM receptor would be associated via a link- Methods section. (A)(i). Typical time-course Ca2+ profile of one representative
er protein.43 platelet adhering to fibrinogen. (A)(ii). Dot plot representing the number of cal-
cium spikes over a period of 5 min. Each point represents an individual
platelet. The results are presented as the mean±SEM of five independent
Fibrinogen is present in whole blood at a concentration
of 2 - 4 mg/mL but does not induce platelet activation. experiments (one-way ANOVA, Bonferroni post-hoc test, ***P<0.001).
This may be explained by an inability of soluble fibrino-

haematologica | 2018; 103(5) 905

 P.H. Mangin et al.

Figure 6. Blockade of human glycoprotein VI limits platelet accumulation to a growing aggregate. (A)(i). Hirudinated human whole blood labeled with DIOC6 (1
μmol/L) was perfused over immobilized collagen (200 μg/mL) to preform aggregates for 2 min 30 s, before perfusing hirudinated blood from the same donor in the
presence of the Alexa Fluor 647-conjugated monoclonal antibody against GPIbβ (5 μg/mL) and with a Fab control (Control) or the blocking anti-GPVI antibody 9O12
(50 μg/mL). (A)(i). Representative 3D reconstructions from confocal images of aggregates obtained after 7 min 30 s of blood perfusion at 300 s-1. Preformed aggre-
gates are represented in gray, aggregates formed in the presence of a Fab control are represented in red and aggregates formed in the presence of the Fab 9O12
are depicted in orange. (A)(ii). Bar graph representing the volume of the platelet aggregates (mean±SEM) in eight random fields, in six separate experiments per-
formed with different blood donors (Mann-Whitney test, ***P<0.001). The gray, red and orange colors represent the volume of the preformed aggregates, the aggre-
gates formed in the presence of a Fab control and the aggregates formed in the presence of the Fab 9O12, respectively.

takes place at a critical checkpoint in aggregate consolida-
tion and aggregate growth. Thus, GPVI has the potential
to both initiate and propagate thrombus formation
through interactions with collagen and fibrin and, it
appears now, also with fibrinogen. Moreover, while colla-
gen and fibrin are localized at the base of the thrombus
and in the core, respectively, fibrinogen is found through-
out the aggregate. This suggests a model in which throm-
bus growth could be sustained by GPVI/fibrinogen, poten-
tially in association with other adhesive proteins. Indeed,
in addition to fibrinogen other adhesive proteins such as
von Willebrand factor and fibronectin have been shown to
support thrombus growth.47–50 Whether these proteins par-
ticipate in GPVI-mediated platelet aggregation is unclear
since von Willebrand factor is not known to be a ligand of
GPVI and fibronectin does not directly promote platelet
adhesion and activation through GPVI.51 Selective inhibi-
tion of the interaction of GPVI with collagen, fibrinogen
and fibrin is required to establish their respective contribu-
tions to platelet activation in hemostasis and thrombosis.
The discovery that GPVI initiates and propagates
platelet aggregation at sites of vessel injury suggests a
major role in hemostasis and thrombosis. Paradoxically,
however, mice and humans deficient in GPVI only have
at most a mild bleeding diathesis, which in the case of
humans may be due to additional confounders such as a
low platelet count as seen in patients with immune-
induced thrombocytopenia caused by antibodies to
GPVI. The relatively minor role of GPVI in hemostasis
can be explained by redundancy in pathways of platelet
activation, with the GPIb-von Willebrand factor axis ini-
tiating hemostasis, and ADP, thromboxane and thrombin
inducing powerful activation. Additionally, the reactive
fibrillar type I and III collagen present in deeper layers of
the vessels would limit the role of GPVI in the hemosta-
sic response following superficial injury. On the other
hand, the discovery that fibrin and immobilized fibrino-
gen activate GPVI may be of significance at sites of fib-
rinogen deposition or fibrin formation in diseased vessels
following inflammation or loss of vascular integrity. The
ability of fibrin and immobilized fibrinogen to activate
GPVI may also reflect yet-to-be-discovered new roles for
GPVI.
In conclusion, in the present study, we have identified
immobilized fibrinogen as a novel activator of human but
not mouse GPVI and have shown that this interaction
supports platelet aggregation under flow. This further
emphasizes the contribution of GPVI to platelet activation
in thrombosis.
Acknowledgments
This work was supported by the British Heart Foundation
(RG/13/18/30563); SPW holds a BHF Chair (CH03/003) and
ATH holds a BHF Studentship (FS/15/71/31677). NLL’s con-
tract was funded by the Agence Nationale pour la Recherche
ANR-14-CE35-0022-02. The authors would like to thank Victor
Tybulewicz and Edina Schweighoffer for providing critical
reagents.

2. Ezumi Y, Shindoh K, Tsuji M, Takayama H. chain downstream of glycoprotein VI in

References
1. Suzuki-Inoue K, Tulasne D, Shen Y, et al.
Association of Fyn and Lyn with the proline-
rich domain of glycoprotein VI regulates
intracellular signaling. J Biol Chem.
2002;277(24):21561-21566.
Physical and functional association of the Src
family kinases Fyn and Lyn with the colla-
gen receptor glycoprotein VI-Fc receptor γ
chain complex on human platelets. J Exp
Med. 1998;188(2):267-276.
3. Quek LS, Pasquet JM, Hers I, et al. Fyn and
Lyn phosphorylate the Fc receptor gamma
murine platelets, and Lyn regulates a novel
feedback pathway. Blood. 2000;96(13):4246-
4253.
4. Severin S, Nash CA, Mori J, et al. Distinct
and overlapping functional roles of Src fam-
ily kinases in mouse platelets. J Thromb
Haemost. 2012;10(8):1631-1645.

906 haematologica | 2018; 103(5)


5. Nieswandt B, Watson SP. Platelet-collagen
interaction: is GPVI the central receptor?
Blood. 2003;102(2):449-461.
6. Inoue O, Suzuki-Inoue K, McCarty OJ, et al.
Laminin stimulates spreading of platelets
through integrin alpha6beta1-dependent
activation of GPVI. Blood. 2006;107(4):1405-
1412.
7. Mammadova-bach E, Ollivier V, Loyau S, et
al. Platelet glycoprotein VI binds to polymer-
ized fibrin and promotes thrombin genera-
tion. Blood. 2016;126(5):683-692.
8. Alshehri OM, Hughes CE, Montague S, et
al. Fibrin activates GPVI in human and
mouse platelets. Blood. 2015;126(13):1601-
1608.
9. Riba R, Hughes CE, Graham A, Watson SP,
Naseem KM. Globular adiponectin induces
platelet activation through the collagen
receptor GPVI-Fc receptor chain complex. J
Thromb Haemost. 2008;6(6):1012-1020.
10. Seizer P, Borst O, Langer HF, et al. EMM-
PRIN (CD147) is a novel receptor for platelet
GPVI and mediates platelet rolling via GPVI-
EMMPRIN interaction. Thromb Haemost.
2009;101(4):682-686.
11. Schaff M, Tang CJ, Maurer E, et al. Integrin
6 1 is the main receptor for vascular
laminins and plays a role in platelet adhe-
sion, activation, and arterial thrombosis.
Circulation. 2013;128(5):541-552.
12. Boulaftali Y, Hess PR, Getz TM, et al.
Platelet ITAM signaling is critical for vascu-
lar integrity in inflammation. J Clin Invest.
2013;123(2):908-916.
13. Kuijpers MJE, Gilio K, Reitsma S, et al.
Complementary roles of platelets and coag-
ulation in thrombus formation on plaques
acutely ruptured by targeted ultrasound
treatment: a novel intravital model. J
Thromb Haemost. 2009;7(1):152-161.
14. Hechler B, Gachet C. Comparison of two
murine models of thrombosis induced by
atherosclerotic plaque injury. Thromb
Haemost. 2011;105 (Suppl):S3-12.
15. Bender M, Hagedorn I, Nieswandt B.
Genetic and antibody-induced glycoprotein
VI deficiency equally protects mice from
mechanically and FeCl3-induced thrombo-
sis. J Thromb Haemost. 2011;9(7):1423-
1426.
16. Nieswandt B, Schulte V, Bergmeier W, et al.
Long-term antithrombotic protection by in
vivo depletion of platelet glycoprotein VI in
mice. J Exp Med. 2001;193(4):459-469.
17. Matus V, Valenzuela G, Sáez CG, et al. An
adenine insertion in exon 6 of human GP6
generates a truncated protein associated
with a bleeding disorder in four Chilean
families. J Thromb Haemost.
2013;11(9):1751-1759.
18. Dumont B, Lasne D, Rothschild C, et al.
Absence of collagen-induced platelet activa-
tion caused by compound heterozygous
GPVI mutations. Blood. 2009;114(9):1900-
1903.
19. Hermans C, Wittevrongel C, Thys C,
Smethurst PA, Van Geet C, Freson K. A com-
pound heterozygous mutation in glycopro-
tein VI in a patient with a bleeding disorder.
J Thromb Haemost. 2009;7(8):1356-1363.
20. Zahid M, Mangin P, Loyau S, et al. The
future of glycoprotein VI as an antithrom-
botic target. J Thromb Haemost.
2012;10(12):2418-2427.
21. Bennett JS, Berger BW, Billings PC. The
structure and function of platelet integrins. J
Thromb Haemost. 2009;7(Suppl 1):200-205.
haematologica | 2018; 103(5)
Fibrinogen, GPVI and platelet activation
22. Obergfell A, Judd BA, Del Pozo MA,
Schwartz MA, Koretzky GA, Shattil SJ. The
molecular adapter SLP-76 relays signals from
platelet integrin aIIbβ3 to the actin
cytoskeleton. J Biol Chem. 2001;276(8):
5916-5923.
23. Goncalves I, Hughan SC, Schoenwaelder
SM, Yap CL, Yuan Y, Jackson SP. Integrin
aIIbβ3-dependent calcium signals regulate
platelet-fibrinogen interactions under flow:
Involvement of phospholipase Cγ2. J Biol
Chem. 2003;278(37):34812-34822.
24. Wonerow P, Pearce AC, Vaux DJ, Watson SP.
A critical role for phospholipase Cγ2 in
aIIbβ3-mediated platelet spreading. J Biol
Chem. 2003;278(39):37520-37529.
25. Watson SP, Auger JM, McCarty OJ, Pearce
AC. GPVI and integrin aIIbβ3 signaling in
platelets. J Thromb Haemost. 2005;3(8):
1752-1762.
26. Boylan B, Gao C, Rathore V, Gill JC,
Newman DK, Newman PJ. Identification of
FcgammaRIIa as the ITAM-bearing receptor
mediating alphaIIbbeta3 outside-in integrin
signaling in human platelets. Blood.
2008;112(7):2780-2786.
27. Hughes CE, Finney BA, Koentgen F, Lowe
KL, Watson SP. The N-terminal SH2 domain
of Syk is required for (hem) ITAM, but not
integrin, signaling in mouse platelets. Blood.
2015;125(1):144-155.
28. Zhi H, Rauova L, Hayes V, et al. Cooperative
integrin/ITAM signaling in platelets
enhances thrombus formation in vitro and
in vivo. Thromb Haemost. 2013;121(10):
1858-1867.
29. Law DA, DeGuzman FR, Heiser P,
Ministri-Madrid K, Killeen N, Phillips DR.
Integrin cytoplasmic tyrosine motif is
required for outside-in aIIbβ3 signalling
and platelet function. Nature. 1999;401
(6755):808-811.
30. Onselaer M-B, Hardy AT, Wilson C, et al.
Fibrin and D-dimer bind to monomeric
GPVI. Blood Adv. 2017; 1(19):1495-1504.
31. Kato K, Kanaji T, Russell S, et al. The contri-
bution of glycoprotein VI to stable platelet
adhesion and thrombus formation illustrat-
ed by targeted gene deletion. Blood.
2003;102(5):1701-1707.
32. Mangin PH, Tang C, Bourdon C, et al. A
humanized glycoprotein VI (GPVI) mouse
model to assess the antithrombotic effica-
cies of anti-GPVI agents. J Pharmacol Exp
Ther. 2012;341(1):156-163.
33. Jandrot-Perrus M, Busfield S, Lagrue AH, et
al. Cloning, characterization, and functional
studies of human and mouse glycoprotein
VI: a platelet-specific collagen receptor from
the immunoglobulin superfamily. Blood.
2000;96(5):1798-1807.
34. Perrault C, Moog S, Rubinstein E, et al. A
novel monoclonal antibody against the
extracellular domain of GPIbbeta modulates
vWF mediated platelet adhesion. Thromb
Haemost. 2001;86(5):1238-1248.
35. Lecut C, Feeney LA, Kingsbury G, et al.
Human platelet glycoprotein VI function is
antagonized by monoclonal antibody-
derived Fab fragments. J Thromb Haemost.
2003;1(12):2653-2662.
36. Lebozec K, Jandrot-Perrus M, Avenard G,
Favre-Bulle O, Billiald P. Design, develop-
ment and characterization of ACT017, a
humanized Fab that blocks platelet’s glyco-
protein VI function without causing bleed-
ing risks. MAbs. 2017:9(6):945-958.
37. Hughes CE, Sinha U, Pandey A, Eble JA,
O’Callaghan CA, Watson SP. Critical role for
an acidic amino acid region in platelet signal-
ing by the HemITAM (hemi-immunorecep-
tor tyrosine-based activation motif) contain-
ing receptor CLEC-2 (C-type lectin receptor-
2). J Biol Chem. 2013;288(7):5127-5135.
38. Hui KY, Haber E, Matsueda GP,Monoclonal
antibodies to a synthetic fibrin-like peptide
bind to human fibrin but not fibrinogen.
Science 1983,222(4628):1129-1132.
39. Cazenave J-P, Ohlmann P, Cassel D, Eckly A,
Hechler B, Gachet C. Preparation of washed
platelet suspensions from human and rodent
blood. Methods Mol Biol. 2004;272:13-28.
40. Schaff M, Receveur N, Bourdon C, et al.
Novel function of tenascin-C, a matrix pro-
tein relevant to atherosclerosis, in platelet
recruitment and activation under flow.
Arterioscler Thromb Vasc Biol. 2011;31(1):
117-124.
41. Haimovich B, Lipfert L, Brugge JS, Shattil SJ.
Tyrosine phosphorylation and cytoskeletal
reorganization in platelets are triggered by
interaction of integrin receptors with their
immobilized ligands. J Biol Chem.
1993;268(21):15868-15877.
42. Poulter NS, Pollitt AY, Owen DM, et al.
Clustering of glycoprotein VI (GPVI) dimers
upon adhesion to collagen as a mechanism
to regulate GPVI signaling in platelets. J
Thromb Haemost. 2017;15(3):549-564.
43. Evans R, Patzak I, Svensson L, et al. Integrins
in immunity. J Cell Sci. 2009;122(2):215-225.
44. Miura Y, Takahashi T, Jung SM, Moroi M.
Analysis of the interaction of platelet colla-
gen receptor glycoprotein VI (GPVI) with
collagen: a dimeric form of GPVI, but not
the monomeric form, shows affinity to
fibrous collagen. J Biol Chem. 2002;277
(48):46197-46204.
45. Loyau S, Dumont B, Ollivier V, et al. Platelet
glycoprotein VI dimerization, an active
process inducing receptor competence, is an
indicator of platelet reactivity. Arterioscler
Thromb Vasc Biol. 2012;32(3):778-785.
46. Aslan JE, Phillips KG, Healy LD, et al.
Histone deacetylase 6-mediated deacetyla-
tion of a-tubulin coordinates cytoskeletal
and signaling events during platelet activa-
tion. AJP Cell Physiol. 2013;305(12):C1211-
C1213.
47. Kulkarni S, Dopheide SM, Yap CL, et al. A
revised model of platelet aggregation. J Clin
Invest. 2000;105(6):783-791.
48. Yang H, Reheman A, Chen P, Zhu G.
Fibrinogen and von Willebrand factor-inde-
pendent platelet aggregation in vitro and in
vivo. J Thromb Haemost. 2006;4(10):2230-
2237.
49. Reheman A, Yang H, Zhu G, et al. Plasma
fibronectin depletion enhances platelet
aggregation and thrombus formation in
mice lacking fibrinogen and von Willebrand
factor plasma fibronectin depletion
enhances platelet aggregation and thrombus
formation in mice lacking fibrinogen and
von Willebrand factor. 2012;113(8):1809-
1817.
50. Cho J, Mosher DF. Impact of fibronectin
assembly on platelet thrombus formation in
response to type I collagen and von
Willebrand factor. Blood. 2006;108(7):2229-
2236.
51. Maurer E, Schaff M, Receveur N, Bourdon
C, Mercier L, Nieswandt B. Fibrillar cellular
fibronectin supports efficient platelet func-
tion and procoagulant activity. Thromb
Haemost. 2015;114(6):1175-1188.
907
ARTICLE
Ferrata Storti
Foundation
Haematologica 2018
Volume 103(5):908-918
Correspondence:
rosario.lopez.exts@juntadeandalucia.es
Received: November 10, 2017.
Accepted: February 22, 2018.
Pre-published: March 15, 2018.
doi:10.3324/haematol.2017.184416
Check the online version for the most updated
information on this article, online supplements,
and information on authorship & disclosures:
www.haematologica.org/content/103/5/908
©2018 Ferrata Storti Foundation
Material published in Haematologica is covered by copyright.
All rights are reserved to the Ferrata Storti Foundation. Use of
published material is allowed under the following terms and
conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode.
Copies of published material are allowed for personal or inter-
nal use. Sharing published material for non-commercial pur-
poses is subject to the following conditions:
https://creativecommons.org/licenses/by-nc/4.0/legalcode,
sect. 3. Reproducing and sharing published material for com-
mercial purposes is not allowed without permission in writing
from the publisher.
908
Coagulation & its Disorders
Circulating microRNAs as biomarkers
of disease and typification of the
atherothrombotic status in antiphospholipid
syndrome
Carlos Pérez-Sánchez,1* Iván Arias-de la Rosa,1* María Ángeles Aguirre,1,2
María Luque-Tévar,1 Patricia Ruiz-Limón,1 Nuria Barbarroja,1
Yolanda Jiménez-Gómez,1 María Carmen Ábalos-Aguilera,1
Eduardo Collantes-Estévez,1,2,3 Pedro Segui,1,4 Francisco Velasco,5
María Teresa Herranz,6 Jesús Lozano-Herrero,6 María Julia Hernandez-Vidal,6
Constantino Martínez,7 Rocío González-Conejero,7 Massimo Radin,8
Savino Sciascia,8 Irene Cecchi,8 María José Cuadrado9** and Chary López-
Pedrera1,2**
1
Instituto Maimónides de Investigación Biomédica de Córdoba (IMIBIC), Spain; 2Unidad
de Gestión Clínica Reumatología, Hospital Universitario Reina Sofía, Córdoba, Spain;
3
Departamento de Medicina (Medicina, Dermatología y Otorrinolaringología),
Universidad de Córdoba, Spain; 4Unidad de Gestión Clínica Radiología, Hospital
Universitario Reina Sofía, Córdoba, Spain; 5Unidad de Gestión Clínica Hematología,
Hospital Universitario Reina Sofía, Córdoba, Spain; 6Servicio de Medicina Interna,
Hospital Morales Meseguer, Murcia, Spain; 7Centro Regional de Hemodonación,
Universidad de Murcia, IMIB-Arrixaca, Spain; 8Department of Clinical and Biological
Sciences, Center of Research of Immunopathology and Rare Diseases, Torino, Italy and
9
Lupus Research Unit and St. Thomas’ Hospital, London, UK
*CP-S and IA-R shared first authorship and contributed equally to this work. **MJC and CL-P shared last author-
ship and contributed equally to this work.
ABSTRACT
W
e aimed to identify the plasma miRNA profile of antiphospho-
lipid syndrome (APS) patients and to investigate the potential
role of specific circulating miRNAs as non-invasive disease bio-
markers. Ninety APS patients and 42 healthy donors were recruited.
Profiling of miRNAs by PCR-array in plasma of APS patients identified
a set of miRNAs differentially expressed and collectively involved in
clinical features. Logistic regression and ROC analysis identified a signa-
ture of 10 miRNA ratios as biomarkers of disease. In addition, miRNA
signature was related to fetal loss, atherosclerosis, and type of thrombo-
sis, and correlated with parameters linked to inflammation, thrombosis,
and autoimmunity. Hard clustering analysis differentiated 3 clusters rep-
resenting different thrombotic risk profile groups. Significant differences
between groups for several miRNA ratios were found. Moreover,
miRNA signature remained stable over time, demonstrated by their
analysis three months after the first sample collection. Parallel analysis
in two additional cohorts of patients, including thrombosis without
autoimmune disease, and systemic lupus erythematosus without
antiphospholipid antibodies, each displayed specific miRNA profiles
that were distinct from those of APS patients. In vitro, antiphospholipid
antibodies of IgG isotype promoted deregulation in selected miRNAs
and their potential atherothrombotic protein targets in monocytes and
endothelial cells. Taken together, differentially expressed circulating
miRNAs in APS patients, modulated at least partially by antiphospho-
lipid antibodies of IgG isotype, might have the potential to serve as
novel biomarkers of disease features and to typify patients’
atherothrombotic status, thus constituting a useful tool in the manage-
ment of the disease.
Introduction
Antiphospholipid syndrome (APS) is a clinical disorder characterized by the
occurrence of thrombosis and/or pregnancy morbidity associated with the persist-
ent presence of antiphospholipid antibodies (aPL), including anti-cardiolipin anti-
haematologica | 2018; 103(5)
 MicroRNAs as antiphospholipid syndrome biomarkers

Table 1. Clinical and laboratory parameters of the antiphospholipid syndrome (APS) patients and the healthy donors (HDs).
APS HDs P
(total n. 90) (total n. 42)
Females/males 48/42 22/20
Age, years 51.2 ± 13.1 46.2 ± 13.4 NS
Arterial thrombosis 35/90 0/42
Venous thrombosis 55/90 0/42
Recurrences 37/90 0/42
Pregnancy morbidity 23/90 0/42
Pathological CIMT 24/90 6/42 0.00
ABI – left* 1.3 ± 0.11 1.2 ± 0.09 0.02
ABI – right* 1.27 ± 0.11 1.2 ± 0.09 0.02
LA positivity 85/90 0/42 0.00
aCL-IgG# GPL 23.4 (0.5-448) 1.3 (0.5-5) 0.00
aCL-IgM# MPL 21.8 (0-354) 4.9 (0.8-17) 0.00
Anti-β2GPI IgG# SGU 23.9 (0-361) 1 (1-2) 0.02
Anti-β2GPI IgM# SMU 14.8 (0-289) 1.2 (1-2.6) 0.02
Antiplatelet agents† 30/90 0/42
Anticoagulant agents‡ 62/90 0/42
Total cholesterol level,* mg/dL 191.7 ± 32.06 190 ± 41.8 NS
Cholesterol HDL level,* mg/dL 51.09 ± 12.4 55.4 ± 13.1 NS
Cholesterol LDL level,* mg/dL 112.1 ± 33.2 118.4 ± 26.9 NS
Triglycerides level,* mg/dL 155.1 ± 163.2 88.3 ± 50.07 NS
ESR,* mm/h 13.5 ± 13.6 6.6 ± 4 0.05
n: number; NS: not significant; CIMT: carotid intima-media thickness; ABI: ankle brachial index; LA: lupus anticoagulant; aCL: anti-cardiolipin antibodies; GPL: IgG phospholipid
units; MPL: IgM phospholipid units; anti-β2GPI: anti-β2 glycoprotein 1 antibodies; SGU: stantard IgG units; SMU: standard IgM units; HDL: high-density lipoprotein; LDL: low-density
lipoprotein; ESR: erythrocyte sedimentation rate. *Results expressed as mean±Standard Deviation. #Results expressed in mean and values range. †Antiplatelet agents include
acetylsalicylic acid and clopidogrel. ‡Anticoagulant agents indicate vitamin K antagonists, including warfarin and acenocumarol.

bodies (aCL), anti-β2-glycoprotein 1 antibodies (anti-
β2GPI) and/or lupus anticoagulant (LA). Cardiac, cerebral
and vascular strokes in these patients are responsible for a
significant reduction in life expectancy.1 The course of car-
diovascular disease (CVD) in APS patients may rapidly
change from asymptomatic to severe life-threatening
manifestations that are difficult to deal with. Timely diag-
nosis and accurate monitoring of the course of APS are
essential to improve the quality of therapy and avoid
approaches based on medical empirical protocols. In the
same way, like many other autoimmune diseases, APS is
a heterogeneous entity, and this has a dramatic impact on
diagnosis and treatment.2
Understanding of the pathophysiological mechanisms
explaining how atherosclerosis and CVD are associated to
APS has been greatly broadened with the application of
genomic technologies.3 One emerging and important
mechanism controlling gene expression is epigenetics,
which regulates gene packaging and independent expres-
sion of alterations in the DNA sequence. Epigenetics,
which comprises DNA methylation, histone modifica-
tions, and microRNAs (miRNAs) activity, is providing
new directions linking genomics and environmental fac-
tors.4 miRNAs are small, non-coding RNAs that, depend-
ing upon base pairing to messenger RNA (mRNA), medi-
ate mRNA cleavage, translational repression or mRNA
destabilization. miRNAs are known to be involved in cru-
cial cellular processes and their dysregulation has been
described in many cell types and fluids in a broad range of
diseases.5-7
In the setting of APS, a previous study by our group
recognized that aPL modulate the expression of 2
miRNAs in monocytes (miR-19b and miR-20a) that con-
trol the expression of key proteins involved in the
pathology of the disease, such as tissue factor (TF).8
Moreover, we recently demonstrated that both aPL and
the anti-double stranded DNA antibodies (anti-dsDNA)
promote specific changes in the expression of proteins
related to the biogenesis of miRNAs in leukocytes of APS
and systemic lupus erythematosus (SLE) patients, which
are translated in the altered expression of the miRNAs
profile and that of their protein targets (related to CVD)
in these disorders.9
Extensive analyses have shown that miRNAs are
released into the circulation where they are present in con-
centration levels that differ between healthy subjects and
patients. Although little is known about the origin and
function of such circulating miRNAs, these molecules are
increasingly recognized as non-invasive and readily acces-
sible biomarkers for risk stratification, diagnosis and prog-
nosis of multiple forms of CVD.10
Specific profiles of circulating miRNAs are also associat-
ed to the pathophysiology of different systemic autoim-
mune diseases, including SLE, systemic sclerosis, and
rheumatoid arthritis (RA), and some of them appear to be
of diagnostic and, possibly, of prognostic value.11 To date,
in the context of APS, no study has analyzed the potential
role of the circulating miRNAs as biomarkers of the dis-
ease. Therefore, the present study was designed to deter-
mine the plasma miRNA specific profile of APS patients,

haematologica | 2018; 103(5) 909

 C. Perez-Sánchez et al.

A B

Figure 1. Antiphospholipid syndrome (APS) patients showed a specific circulating miRNAs profile related to clinical features of this autoimmune disorder. (A) To
identify the changes that occurred in the expression levels of microRNAs (miR) in plasma from antiphospholipid syndrome versus controls, Human Serum & Plasma
miRNA PCR-array (Qiagen) was performed in the study cohort. Expression levels of 19 miRNAs were found up-regulated in antiphospholipid syndrome, while 20
miRNAs were down-regulated. (B) Ingenuity Pathway Analysis (IPA) uncovered the main enriched biological functions and pathways in which these microRNAs are
involved. The analysis included only the functions and pathways with average IPA score >2 [indicated as -log (P value)]. (C) Validation of selected miRNAs by RT-PCR
in the whole cohort of APS patients and healthy donors. *P<0.05.

their modulation by autoantibodies, and their potential ical parameters, B-Mode Ultrasound IMT and Ankle Brachial
role as non-invasive biomarkers of disease features. Index measurements see the Online Supplementary Appendix.

Isolation of miRNAs and analysis of miRNAs expression

Methods
Patients
Ninety patients with primary APS and 42 healthy donors (HDs)
were included in this study over a period of 24 months (see Online
Supplementary Appendix). All experimental protocols were
approved by the ethics committee of the Reina Sofia Hospital,
Cordoba, Spain, and written informed consent was obtained. The
characteristics of patients and HDs are shown in Table 1.
The adjusted global anti-phospholipid syndrome score
(aGAPSS) was calculated for each APS patient, as previously
described.12 Briefly, aGAPSS was calculated by adding the points
corresponding to both the cardiovascular and thrombotic risk fac-
tors, based on a linear transformation derived from the β regres-
sion coefficient as follows: 3 for hyperlipidemia, 1 for arterial
hypertension, 5 for aCL IgG/IgM, 4 for anti-β2GPI IgG/IgM, and 4
for LA.
Two additional cohorts of patients were further analyzed as dis-
ease control, including 23 patients with thrombosis in the absence
of an associated autoimmune disease [12 non-pregnant women
and 11 men; mean age 44 (range: 21-73 years), including patients
with objectively verified thrombotic events: 14 deep venous
thrombosis and 9 thrombosis in intra-cerebral vessels], and 25 SLE
patients without aPLs (Online Supplementary Table S1).
For details of blood sample collection and assessment of biolog-
profiling
Total RNA, including the miRNA fraction, was extracted from
both plasma and supernatants obtained from in vitro studies by
using the QIAzol miRNeasy kit (Qiagen, Valencia, CA, USA) fol-
lowing the manufacturer's instructions13 (Online Supplementary
Appendix). To identify the changes that occurred in the expression
levels of miRNAs in plasma from APS patients and HDs, a Human
Serum & Plasma miRNA PCR-array (Qiagen) was performed
(Online Supplementary Appendix) on an exploratory cohort (Online
Supplementary Table S2).
Quantitative real-time PCR
A fixed volume of 3 μl of RNA solution from the 14 μl-eluate
from RNA isolation of 200 μl plasma sample was used as input
into the reverse transcription. Input RNA was reverse transcribed
using the TaqMan miRNA Reverse Transcription kit and miRNA-
specific stem-loop primers (Life Technologies, Madrid, Spain)
(Online Supplementary Appendix). The expression levels of miRNAs
were calculated by using 2-Ct and reciprocal ratios were performed
[Ratio miR-A/miR-B = log2 (2-Ct miR-A/2-Ct miR-B)], as previously
described.14-19 Reciprocal ratios analysis is an approach that
bypasses the controversial issue of data normalization of miRNAs
in plasma (self-normalization). Furthermore, miRNAs whose con-
centrations are changed because of a pathology in opposite direc-
tions can be effective in differentiating investigated populations.

910 haematologica | 2018; 103(5)

 MicroRNAs as antiphospholipid syndrome biomarkers

Figure 2. Interaction network of microRNAs identified potential mRNA targets involved in clinical features of antiphospholipid syndrome. Using microRNA Target
Filter of QIAGEN’s Ingenuity Pathway Analysis (IPA, QIAGEN Redwood City, CA, USA, www.qiagen.com/ingenuity), the software generated a network including the
selected microRNAs (miRNAs or miR) and their mRNA targets, filtered by coronary artery disease, thrombosis, abortion and cerebrovascular dysfunction. Only targets
experimentally observed and predicted with high confidence are shown and related by direct interactions to their specific miRNA regulators.

Target gene prediction and integrated analysis by
Ingenuity Pathway Analysis
The altered miRNAs were further analyzed to obtain informa-
tion about biological functions, pathways and networks by using
the web-based bioinformatics tool QIAGEN’s Ingenuity Pathway
Analysis (IPA; Ingenuity Systems, http://www.INGENUITY.com).
For this purpose, all differentially regulated miRNAs and fold
changes were imported into IPA20 (Online Supplementary Appendix).
Details of purification of IgG, in vitro exposure of monocytes and
endothelial cells to aPL antibodies, and the statistical analysis are
available in the Online Supplementary Appendix.
Results
Differentially expressed miRNAs in the plasma of APS
patients and HDs
In the discovery phase (exploratory cohort), we identi-
fied 39 miRNAs that were differentially expressed
between APS patients and HDs (cut off: 1.7-fold change),
including 19 up-regulated and 20 down-regulated (Figure
1A). Using the IPA software, the functional analysis of the
altered miRNAs in APS patients showed that a large num-
ber of them had validated and putative target mRNAs
mainly involved in connective tissue disorders, inflamma-
tory response, reproductive system disease, CVD or skele-
tal and muscular disorders (Figure 1C).
Bioinformatic identification and analysis of
deregulated miRNAs related to the pathophysiology of
APS and analysis of potential protein targets
In silico studies were performed to identify the altered
miRNAs that might have as potential targets a number of
genes/proteins involved in the development of clinical
manifestations related to APS, such as coronary artery dis-
ease, thrombosis, abortion, and cerebrovascular dysfunc-
tion. IPA identified 11 altered miRNAs as the main regula-
tors of proteins involved in the pathology of APS, includ-
ing miRNA 34a-5p, 15a-5p, 145a-5p, 133b-3p, 124-3p,
206, 20a-5p, 19b-3p, 210-3p, 296-5p and 374a-5p. This set
of 11 miRNAs included, among others, the top 5 up-regu-
lated miRNAs and 3 out of the top 5 down-regulated
miRNAs in the PCR-array. The expression levels of the 11
selected miRNAs were analyzed in all study subjects by
RT-PCR (Figure 1B). MiR-124 and miR-34a were found
increased in APS patients in relation to healthy donors,
while miR-20a, miR-19b and miR145a were found
reduced. The remaining microRNAs were also found to be
altered, showing a trend to either increase or reduction as
observed in the discovery phase, thus validating the data
obtained by PCR-array.
We further developed a network that defined the inter-
action between miRNA-mRNA targets (Figure 2). Key pro-
teins involved in the pathophysiology of APS, and identi-

haematologica | 2018; 103(5) 911

 C. Perez-Sánchez et al.

Figure 3. A circulating miRNA signature in antiphospholipid syndrome (APS) might have potential value as biomarkers of disease. (A) Selected microRNAs (miRNAs
or miR) were analyzed in the whole cohort, including 90 APS patients and 42 healthy donors, and reciprocal ratios were performed. Beeswarm plot of each differen-
tially expressed miR ratio is shown, along with mean, Standard Deviation, and P-value. For statistical analysis, after normality and equality of variance tests, compar-
isons were made by paired Student t-test or a non-parametric test (Mann-Whitney rank sum test). (B) A combination of the 10 miRNA ratios as a panel was carried
out by using logistic regression on the data set. ROC curve of miRNA panel and cut off were generated based on the predicted probability (P) for each subject as a
single score. The equation used in our model was: “Combined miRNA-ratio panel [Logit(p)] = - 0.64 + 0.034x(miR-19b/miR-34a) + 1.061x(miR-19b/miR-15a) +
0.248x(miR-19b/miR-124) – 1.704x(miR-19b/miR-145) + 2.34x(miR-20a/miR-145) – 0.729x(miR-20a/miR-374a) – 0.624x(miR-20a/miR-210) + 0.088x(miR-
20a/miR-133b) + 0.166x(miR-206/miR-34a) + 0.056x(mir-124/miR-296)”. The area under the curve (AUC), sensitivity and specificity are displayed, and a cut-off
value with higher specificity was selected.

fied as potential mRNA targets of those miRNAs, were
quantified in the plasma of APS patients and HDs. As pre-
viously reported,20-23 APS patients showed significantly
increased plasma levels of TF, PAI-1, MCP-1, VEGF-A and
VEGFR-1 (Online Supplementary Figure S1).
Circulating miRNA signature as potential biomarkers
of disease in APS
It has been shown that the combination of miRNAs
improves their predictive potential to differentiate two
pathological conditions.14-19 Thus, to assess the potential of
specific circulating miRNAs in APS patients as biomarkers
of disease features, reciprocal ratios of the miRNAs ana-
lyzed were performed by using statistical tools. By this
approach, we identified 10 miRNA ratios, integrated by
the 11 selected miRNAs, and differentially expressed in
plasma of APS patients in comparison with HDs, includ-
ing miR-19b/miR-34a, miR-19b/miR-15a, miR19b/miR-
124, miR-19b/miR-145, miR-20a/miR-145, miR-20a/miR-
374a, miR-20a/miR-210, miR-20a/miR-133b, miR-
206/miR-34a and miR-124/miR-296 (Figure 3A). To fur-
ther explore the efficiency of these biomarkers to identify
APS patients, a combination of the 10 miRNA ratios as a
panel was carried out by using a logistic regression on the
data set, as previously described.24 Thus, all miRNA-ratios
were integrated into a single model or equation, which
provided a single ‘score’ that allowed us to perform the
ROC analysis and establish the cut off for prediction. The
ROC curve for the 10 miRNA ratios signature revealed a
marked accuracy, evidenced by an AUC of 0.81. At the
optimal cut-off value of 0.6, the sensitivity and specificity
of the combined miRNA panel for APS identification were
78% and 80%, respectively (Figure 3B).
Stability of miRNA expression profile over time in APS
Plasma from 21 APS patients included in the study was
evaluated again three months after the first blood sample
collection to analyze the stability of the circulating
miRNA profile. Results demonstrated that miRNA
expression in the second sample collection did not
change in relation to the first analysis (Online
Supplementary Figure S2A). Moreover, the levels of
miRNA ratios at baseline correlated significantly with
the levels of these ratios three months later (Online

912 haematologica | 2018; 103(5)

 MicroRNAs as antiphospholipid syndrome biomarkers

Figure 4. Antiphospholipid syndrome (APS) patients show a specific miRNA profile distinct from both non-autoimmune patients with previous thrombotic events
and aPL-negative systemic lupus erythematosus (SLE) patients. Twenty-three thrombotic non-antiphospholipid syndrome patients (non-APS) and 25 aPL-negative
SLE patients were included, and the circulating microRNA (miRNA or miR) signatures of APS were compared. One-way ANOVA was used for statistical comparisons.
A Bonferroni correction was applied for multiple testing. P<0.05 was considered statistically significant. Beeswarm plot of each differentially expressed miRNA ratio
is shown along with mean, Standard Deviation and P-value. n.s.: no significant statistical differences.

Supplementary Figure S2B). Thus, our data support the
theory that there is a specific circulating miRNA signa-
ture in APS which remains stable over time.
APS patients show a specific miRNA profile different
from both non-autoimmune patients with previous
thrombotic events and aPL-negative SLE patients
To assess the specificity of the miRNA signature found
in APS patients, and in order to analyze whether the
altered expression of the circulating miRNAs evaluated
was linked to their thrombophilic status, an additional dis-
ease group including 23 patients with thrombosis in the
absence of an associated autoimmune disease was evalu-
ated. In these patients, the ratios formed by the expression
levels of the 11 selected miRNAs were significantly differ-
ent from those described in APS patients, except for the
ratios miR-19b/miR-15a and miR-19b/miR-145 which
exhibited non-significant differences (Figure 4). To evalu-
ate if the altered expression of the miRNA signature was
a sign of an autoimmune status, an additional disease
group, including 25 SLE patients negative for aPL, was
analyzed. In this SLE cohort, the ratios produced by the
selected circulating miRNAs were significantly different
from those found in APS patients, except for the ratios
miR-19b/miR-34a, miR-20a/miR-374a and miR-124/miR-
296 which exhibited non-significant differences (Figure 4).
Potential influence of standard therapy on the profile
of circulating miRNAs in APS
APS patients were classified into two groups based on
the treatment received, including 30 primary APS patients
treated with antiplatelet agents and 62 primary APS
patients treated with anticoagulant drugs. The statistical
comparison between patients treated with antiplatelet
and anticoagulant agents showed no significant differ-
ences in miRNA signature, except for the ratio miR-
20a/374 (Online Supplementary Figure S3).
Circulating miRNAs are associated with clinical
features of APS and show potential as biomarkers for
the development of atherosclerosis
The levels of some circulating miRNA ratios that inte-
grate the signature in APS were associated with the ocur-
rence of fetal losses in these patients, including elevated
levels of miR-19b/miR-124 and miR-20a/miR-374, and
reduced levels of miR-124/miR-296 (Figure 5A).
Associations between miRNA ratios and the type of
thrombosis suffered by APS patients were also identified.
Thus, elevated levels of ratios miR-20a/miR-145 and miR-
20a/miR-374 were significantly associated with the ocur-
rence of arterial thrombosis in APS patients (Figure 5B).
Furthermore, the ratios of miR-19b/miR-124 and miR-
124/miR-296 were also found to be associated with the
presence of a pathological CIMT in these patients (Figure
5C). To accurately evaluate their relevance as biomarkers
of early atherosclerosis, we conducted combined ROC
analyses of these miRNA ratios. The combination of both
circulating miRNA ratios as a panel showed an evident
accuracy, with an AUC of 0.76 at a sensitivity of 67% and
specificity of 78% from a cut-off value of 0.41 (Figure 5D).
Cluster analysis
Hard clustering analysis in the APS cohort differentiated
3 clusters representing different thrombotic risk profile
groups. Clinical and laboratory parameters of each cluster
are resumed (Figure 6A). Briefly, cluster 1 (50% of the clus-
tered cohort) was characterized by lower prevalence of
cardiovascular risk factors and aPL multiple positivity.
Conversely, cluster 1 shows a higher rate of venous
thrombotic event when compared to the other clusters.

haematologica | 2018; 103(5) 913

 C. Perez-Sánchez et al.

A B

C D

Figure 5. Circulating miRNAs are related to clinical features of antiphospholipid syndrome (APS) and show potential as biomarkers for the development of ather-
osclerosis. Association studies of altered circulating microRNA (miRNA or miR) ratios and the occurrence of previous fetal loss (A), the type of thrombosis suffered
(B) and the presence of a pathological carotid intima-media thickness (CIMT) (C). Beeswarm plot of each miR ratio is shown, along with mean, Standard Deviation,
and P-value. (D) A combination as a panel of the 2 miRNA ratios associated to the pathological CIMT was carried out by using logistic regression on the data set and
receiver operator characteristics (ROC) curve analyses were performed. ROC curve of miRNA panel and cut off were generated based on the predicted probability
(P) for each patient as a single score. The equation used was: “Combined miRNA-ratio panel [Logit(p)] = 0.599 – 0.133x(miR-19b/miR-124) + 0.007x(miR-124/miR-
296)”. The area under the curve (AUC), sensitivity and specificity are shown, and a cut-off value with higher specificity was selected.

Cluster 2 (17.6% of the clustered cohort) was character-
ized by a higher rate of cardiovascular risk factors, arterial
thrombotic events, recurrences and a low prevalence of
multiple aPL positivity. Cluster 3 (32.4% of the clustered
cohort) was represented by a higher rate of multiple aPL
positivity, arterial thrombotic events, and lower rate of
cardiovascular risk factors. When evaluating different
miRNA ratio expression among clusters, we found a sta-
tistically significant difference between groups for the fol-
lowing miRNA ratios: miR-19b/miR-124 (P<0.001,
ANOVA), miR-20a/miR-374 (P<0.05, ANOVA), miR-
20a/miR-210 (P<0.001, ANOVA) and miR-124/miR-296
(P<0.05, ANOVA). miRNA ratio expression in the differ-
ent clusters are summarized in Figure 6C.
When comparing the aGAPSS values among the differ-
ent clusters, we found a significant difference (P=0.008,
t-test) between cluster 1 [mean aGAPSS 5.38;
1.628±Standard Deviation (SD)] and Cluster 2 (mean
aGAPSS 8,67; 3.67±SD). Similarly, we found a significant
difference (P<0.001, t-test) between cluster 1 and cluster 3
(mean aGAPSS 10.82; 2.316±SD). aGAPSS values stratify-
ing for clusters are represented in Figure 6B.
Circulating miRNAs correlate with clinical and
serological parameters in APS
The miRNA ratios that integrate the signature in APS
were linked with clinical parameters, such as ABI, pres-
ence of elevated titers of aPL, particularly aCL and anti-
β2GPI antibodies, and erythrocyte sedimentation rate
(Online Supplementary Table S3). Correlation analyses with
serological markers related to atherothrombosis further
showed significant positive correlations with the expres-
sion levels of various miRNA ratios and with levels of TF,
PAI-1, VEGF-A, VEGF-R1 and MCP-1 (Online
Supplementary Table S3). Some of these correlations were
also found among various miRNA ratios in plasma of APS
patients.
Antiphospholipid antibodies modulate the expression
of both the circulating miRNAs that integrate the
signature in APS and their potential protein targets
The expression of the 11 selected miRNAs was signifi-
cantly altered in the supernatant of HUVECs treated with
aPL-IgG in relation to those treated with a non-immune-
IgG (Figure 7A), except for the miR-124 and miR-206.
Accordingly, this treatment promoted in HUVECs the
secretion of atherothrombotic proteins, such as TF, PAI-1
and VEGF-R1 (Figure 7B), potential targets of the miRNAs
analyzed. On the other hand, the expression levels of sev-
eral miRNAs were deregulated in the supernatant of
monocytes treated with aPL-IgG, including miR-19b, miR-
20a, miR-145, miR-210 and miR-296 (Figure 7C).
Concomitantly, aPL-IgG treatment promoted in mono-
cytes an increase in the secretion of TF, PAI-1 and MCP-1
(Figure 7D).
Discussion
The present study identifies, for the first time, a specific
signature of circulating miRNAs in APS patients that
might serve as potential biomarkers of clinical features of

914 haematologica | 2018; 103(5)


A B
this autoimmune disorder. Moreover, this signature could
represent a useful tool to typify and stratify patients based
on their thrombotic status and cardiovascular risk profile
(Online Supplementary Figure S3).
Circulating miRNAs were firstly described in peripheral
blood as promising specific biomarkers for a wide range of
diseases, such as cancer and other inflammatory patholo-
gies.25,26 Thereafter, several studies revealed the altered
expression of numerous miRNAs in plasma, blood cells,
and tissues of systemic autoimmune conditions, such as
RA and SLE, which were directly associated to disease
activity, making them potential useful biomarkers for clin-
ical features and follow up.9,26-29 However, to date, the spe-
cific profile of circulating miRNAs in APS patients has not
been evaluated. In the present study, the profiling of
miRNAs by PCR-array in plasma of APS patients has
helped to identify a set of miRNAs differentially expressed
and collectively associated to clinical features of the dis-
ease, such as inflammatory response, reproductive system
disease, and CVD, among others. Using logistic regres-
sion, we further developed a model that identified 10
miRNA ratios, differentially expressed, that showed great
potential as biomarkers of disease of APS patients.
Recent studies support the evidence that an miRNAs
signature has a higher diagnostic value than individual
miRNAs.14-19 The use of ratios is a feasible approach that
overcomes the controversial question of normalizing plas-
ma levels of miRNAs, given the lack of a reliable normal-
izer for circulating miRNAs. In addition, the establishment
haematologica | 2018; 103(5)
MicroRNAs as antiphospholipid syndrome biomarkers
Figure 6. Specific miRNA signatures might
identify subgroups of antiphospholipid syn-
drome (APS) patients showing different
thrombotic risk profiles: cluster analysis. (A)
Clinical and laboratory parameters of the 3
clusters. (B) Comparison of the adjusted
global anti-phospholipid syndrome score
(aGAPSS) values among the different clus-
ters. (C) Evaluation of different microRNA
(miR) ratios expression among clusters.
HTA: arterial hypertension; LA: lupus antico-
agulant; aCL: anti-cardiolipin IgG/IgM; anti-
β2GPI: anti-β2 glycoprotein 1 IgG/IgM.
of these ratios allows the identification of a combination
of expression profiles closer to reality in vivo in patients,
where the interactions between miRNAs and their specific
potential targets never occur in a unique or individualized
way. In fact, it is likely that, in some cases, various
miRNAs, whose concentrations are shifted in opposite
directions in a particular pathology, contribute together
and specifically to certain clinical profiles.
The signatures of circulating miRNAs identified in APS
patients integrated miRNAs previously described to be
altered in other autoimmune and CVD. Thus, miR-19b
and miR-20a have been shown to be essential modulators
of TF expression in APS and SLE patients,8 so that reduced
expression of such miRNAs contributes to the overexpres-
sion of TF in monocytes, which is directly associated with
the occurrence of thrombotic events in APS.21 On the
other hand, miR-124, found altered in APS, SLE and RA
patients at both cellular and plasma levels, modulates the
overexpression of MCP-1, a key chemokine directly
involved in CVD associated to these autoimmune condi-
tions.30-33 Likewise, miR-133b and miR-145 have been
identified as the most promising biomarkers of the patho-
genesis of CVD. Both miRNAs participate in the differen-
tiation of vascular smooth muscle cells. In addition, miR-
133b regulates angiogenesis and endothelial function,
while miR-145 participates in the stabilization of athero-
matous plaque.34 The miR-34a is highly expressed in
endothelial cells, and elevated circulating levels of this
miRNA have been associated to myocardial infarction.35
915
 C. Perez-Sánchez et al.

A B

C D

Figure 7. Antiphospholipid antibodies modulate the expression of both the circulating miRNAs that integrate the signature in antiphospholipid syndrome (APS)
and their putative protein targets. Human umbilical vein endothelial cells (HUVECs) were treated with antiphospholipid antibodies and secreted selected microRNAs
(miRNAs) (A) and putative target protein (B) levels were determined in the supernatant. Monocytes were also treated with antiphospholipid antibodies and secreted
selected miRNAs (C) and putative target proteins (D) levels were evaluated in the supernatant of culture. Differences were analyzed by Student t-test. Values are the
means and Standard Error of Mean of 4 independent experiments performed in triplicate. P<0.05 was considered statistically significant. TF: tissue factor; PAI-1:
plasminogen activator inhibitor-1; VEGF-A: vascular endothelial growth factor A; VEGF-R1: VEGF-Receptor-1; MCP-1: monocyte chemotactic protein.

Moreover, the main target of miR-34a is VEGF-A, a key
inflammatory protein involved in numerous cardiovascu-
lar and autoimmune pathologies, including APS.23,36 In the
same way, miR-374 has been described as regulator of
maintenance of vascular integrity.37 The remaining
miRNAs members of the signature, including miR-296,
miR-210, miR-206 and miRNA-15, have been found
altered in severe pre-eclampsia, one of the leading causes
of maternal mortality and neonatal morbidity world-
wide.38-40 Thus, all the processes regulated by these
miRNAs seem to orchestrate distinct aspects of APS
pathogenesis.
To assess the specificity of the circulating miRNA signa-
ture in APS we evaluated the miRNA profile in an addi-
tional cohort of patients characterized by the presence of
previous thrombotic events in the absence of an associat-
ed autoimmune disease. The miRNAs analysis revealed a
differential pattern of expression between these two
cohorts. Those results substantiate previous studies that
evidenced the presence of a distinct miRNA profile in
monocytes and neutrophils of thrombotic non-autoim-
mune patients compared to APS patients.9 This could
reflect a differential mode of regulation and activity of
miRNAs in thrombotic patients compared to APS patients,
on which the role of autoantibodies might be crucial.
Moreover, the analysis of a parallel autoimmune popula-
tion (SLE patients) negative for aPL, also identified an
miRNA signature distinct from that of APS, thus underly-
ing the potential role of aPLs as regulators of thrombosis-
related miRNAs in APS, and pointing to the presence of a
specific miRNA profile relative to the pathogenesis of
each disease.
Antiphospholipid syndrome patients recruited in this
study were mainly treated with anticoagulant and/or
antiplatelet agents. All of them have been shown to influ-
ence miRNAs expression, an epigenetic process that might
help to delineate the mechanisms underlying their
effects.9,41,42 Thus, we evaluated the potential effect of
these treatments on the circulating miRNA expression
profile. No significant differences were observed in our
cohort of APS patients between those who received
antiplatelet and those treated with anticoagulant agents,
suggesting that the prothrombotic status induced by
effects of aPLs, and the consequently deregulated
miRNAs, were not differentially modulated among these
drugs.
In order to understand the clinical relevance of the
altered circulating miRNA signature, association and cor-
relation studies were perfomed. Altered expression of var-
ious miRNA ratios was associated with the presence of
previous fetal losses. In line with these findings, several
studies have shown that the misregulation of circulating
placental miRNAs in maternal blood might lead to preg-
nancy complications, thus acting as non-invasive diagnos-
tic and prognostic biomarkers for pregnancy monitoring.42-
44
Association studies further established a significantly
increased expression of 2 miRNA ratios in APS patients
that had suffered arterial thrombosis in comparison with
those who experienced venous thrombotic events.
Interestingly, both miRNA ratios were integrated by the
miR-20a, previously reported to be the main regulator of
TF, whose expression levels have been found to be related
to the development of arterial thrombosis in the setting of
APS.8,45 Finally, we identified 2 miRNA ratios as clinical rel-
evant biomarkers related to early atherosclerosis develop-
ment in APS patients, which were integrated by the miR-

916 haematologica | 2018; 103(5)

 MicroRNAs as antiphospholipid syndrome biomarkers

19b and miR-124, both of them critical players in the
expression of proteins related to inflammation and throm-
bosis in APS and SLE.8,9
Correlation studies revealed that the altered circulating
miRNA signature in APS is linked to parameters related to
increased risk of peripheral artery disease such as ABI.
Moreover, correlations between circulating miRNA levels
and numerous altered parameters related to inflammation
and thrombosis were also identified. These correlations
support the relationship observed in the in silico study
between the selected miRNAs and potential target pro-
teins involved in various clinical features of APS. The
influence of the autoimmunity in the circulating profile of
miRNAs in APS was also revealed by the significant corre-
lation between high titers of aPL-IgG and the altered
expression of several miRNAs integrating the signature.
These relationships further sustain those previously iden-
tified among the altered profile of miRNAs in APS and SLE
at cellular level and the autoimmune and inflammatory
profile of both autoimmune conditions.9 Therefore, our
data suggest that the altered plasma profile of miRNAs is
an important mechanisin that might contribute to the reg-
ulation of the pro-atherothrombotic status of APS
patients, on which aPL seem to play a key role.
Our in vitro studies further confirmed this hypothesis,
demonstrating that aPL-IgG antibodies promoted a signif-
icant deregulation in the expression levels of both the
selected miRNAs and their potential protein targets in the
supernatant of cultured monocytes and HUVECs, the
main drivers of the CVD in the setting of APS. These
results also confirm and complement previous studies
which showed the in vitro effects of aPL-IgG in the induc-
tion of prothrombotic/inflammatory mediators3,31 and the
modulation of specific cellular miRNAs involved in their
modulation.8,9 Nevertheless, although our data show spe-
cific effects of aPL-IgG on the secretion of several circulat-
ing microRNAs related to CVD, the contribution of other
components of the vascular and immune system to the
altered profile of circulating miRNAs still has to be
defined. In addition, since we did not perform a complete
plasma human microarray analysis, we cannot exclude the
complementary role of other circulating miRNAs in the
physiopathology of APS.
Interestingly, our analysis supports a clinical role for the
use of miRNA ratios when stratifying patients for their
thrombotic risk. While studying the miRNA expression
profile has widened the understanding of APS pathogene-
sis,9 its clinical utility is still a question of debate. Our data
support the view that specific miRNA signatures could
identify subgroups of APS patients showing different clin-
ical profiles (in terms of site of thrombosis and risk of
recurrences), potentially paving the way for their use as
useful biomarkers that will increase the specificity and
sensitivity of thrombotic risk assessment.
Taken together, our data suggest that differentially
expressed miRNAs in the plasma of APS patients, modu-
lated at least partially by aPL-IgG antibodies, might have
the potential to serve as novel biomarkers of disease fea-
tures and could help typify the atherothrombotic status of
patients, thus constituting a useful tool in the manage-
ment of this disease.
Acknowledgments
We thank all patients for their participation in this study.
Funding
This study was supported by grants from the Junta de
Andalucia (CTS-7940), the Instituto de Salud Carlos III (ref. n.
PI15/01333), Cofinanciado por el Fondo Europeo de Desarrollo
Regional de la Unión Europea 'Una manera de hacer Europa',
Spain, and the Spanish Inflammatory and Rheumatic Diseases
Network (RIER), Instituto de Salud Carlos III
(RD16/0012/0015). CL-P was supported by a contract from the
Spanish Junta de Andalucía. YJ-G was supported by a contract
from the University of Cordoba (Co-financing of the Research
Plan of the University of Cordoba and the Operating Program of
the European Regional Development Funds -ERDF- for
Andalusia).

immune regulatory functions. Clin Dev and extracellular communicators in cardio-

References
1. Young SP, Kapoor SR, Viant MR, et al. The
impact of inflammation on metabolomic
profiles in patients with arthritis. Arthritis
Rheum. 2013;65(8):2015-2023.
2. Palisi A, Grimaldi M, Sabatini P, et al. A
serum nuclear magnetic resonance-based
metabolomic signature of antiphospholipid
syndrome. J Pharm Biomed Anal.
2017;133:90-95.
3. Perez-Sanchez C, Barbarroja N, Messineo
S, et al. Gene profiling reveals specific
molecular pathways in the pathogenesis of
atherosclerosis and cardiovascular disease
in antiphospholipid syndrome, systemic
lupus erythematosus and antiphospholipid
syndrome with lupus. Ann Rheum Dis.
2015;74(7):1441-1449.
4. López-Pedrera C, Pérez-Sánchez C,
Ramos-Casals M, Santos-Gonzalez M,
Rodriguez-Ariza A, Cuadrado MJ.
Cardiovascular risk in systemic autoim-
mune diseases: epigenetic mechanisms of
Immunol. 2012;2012:9746-9748.
5. Huang Y, Shen XJ, Zou Q, Wang SP, Tang
SM, Zhang GZ. Biological functions of
microRNAs: a review. J Physiol Biochem.
2011;67(1):129-139.
6. Xiao C, Rajewsky K. MicroRNA control in
the immune system: basic principles. Cell.
2009;136(1):26-36.
7. Alevizos I, Illei GG. MicroRNAs as bio-
markers in rheumatic diseases. Nat Rev
Rheumatol. 2010;6(7):391-398.
8. Teruel R, Pérez-Sánchez C, Corral J, et al.
Identification of miRNAs as potential mod-
ulators of tissue factor expression in
patients with systemic lupus erythemato-
sus and antiphospholipid syndrome. J
Thromb Haemost. 2011;9(10):1985-1992.
9. Pérez-Sánchez C, Aguirre MA, Ruiz-Limón
P, et al. Atherothrombosis-associated
microRNAs in Antiphospholipid syndrome
and Systemic Lupus Erythematosus
patients. Sci Rep. 2016;6:31375.
10. Creemers EE, Tijsen AJ, Pinto YM.
Circulating microRNAs: novel biomarkers
vascular disease? Circ Res. 2012;110(3):
483-495.
11. Heegaard NHH, Carlsen AL, Skovgaard K,
Heegaard PMH. Circulating Extracellular
microRNA in Systemic Autoimmunity.
EXS. 2015;106:171-195.
12. Sciascia S, Sanna G, Murru V, Roccatello D,
Khamashta MA, Bertolaccini ML. GAPSS:
The global anti-phospholipid syndrome
score. Rheumatology (Oxford).
2015;54(1):134-138.
13. Castro-Villegas C, Pérez-Sánchez C,
Escudero A, et al. Circulating miRNAs as
potential biomarkers of therapy effective-
ness in rheumatoid arthritis patients treat-
ed with anti-TNFa. Arthritis Res Ther.
2015;17:49.
14. Boeri M, Verri C, Conte D, et al. MicroRNA
signatures in tissues and plasma predict
develoment and prognosis of computed
tomography detected lung cancer. Proc
Natl Acad Sci USA. 2011;108(9):3713-3718.
15. Hennessey PT, Sanford T, Choudhary A, et
al. Serum microRNA biomarkers for detec-

haematologica | 2018; 103(5) 917

 C. Perez-Sánchez et al.
tion of non-samll cell lung cancer. PLoS
One. 2012;7(2):e32307.
16. Sheinerman KS, Tsivinsky VG, Abdullah L,
Crawford F, Umansky SR. Plasma
microRNA biomarkers for detection of mild
cognitive impairment: biomarker validation
study. Aging (Albany NY). 2013;5(12):925-
938.
17. Matthaei H, Wylie D, Lloyd MB, et al.
miRNA biomarkers in cyst fluid augment
the diagnosis and management of pancreatic
cysts. Clin Cancer Res. 2012;18(17):4713-
4724.
18. Fortunato O, Boeri M, Verri C, et al.
Assessment of circulating microRNAs in
plasma of lung cancer patients. Molecules.
2014;19(3):3038-3054.
19. Sharova E, Grassi A, Marcer A, et al. A circu-
lating miRNA assay as a first-line test for
prostate cancer screening. Br J Cancer.
2016;114(12):1362-1366.
20. Mayburd AL, Martlínez A, Sackett D, et al.
Ingenuity network-assisted transcription
profiling: Identification of a new pharmaco-
logic mechanism for MK886. Clin Cancer
Res. 2006;12(6):1820-1827.
21. Cuadrado MJ, López-Pedrera C, Khamashta
MA, et al. Thrombosis in primary antiphos-
pholipid syndrome: a pivotal role for mono-
cyte tissue factor expression. Arthritis
Rheum. 1997;40(5):834-841.
22. López-Pedrera C, Aguirre MA, Buendía P, et
al. Differential expression of protease acti-
vated receptors in monocytes from patients
with primary Antiphospholipid syndrome.
Arthritis Rheum. 2010; 62(3):869-877.
23. Cuadrado MJ, Buendía P, Velasco F, et al.
Vascular endothelial growth factor expres-
sion in monocytes from patients with pri-
mary antiphospholipid syndrome. J Thromb
Haemost. 2006;4(11):2461-2469.
24. Xiao B, Wang Y, Li W, et al. Plasma
microRNA signature as a noninvasive bio-
marker for acute graft-versus-host disease.
Blood. 2013;122(19):3365-3375.
25. Mitchell PS, Parkin RK, Kroh EM, et al.
Circulating microRNAs as stable blood-
based markers for cancer detection. Proc
918
Natl Acad Sci USA. 2008;105(30):10513-
10518.
26. Chen X, Ba Y, Ma L, et al. Characterization
of microRNAs in serum: a novel class of bio-
markers for diagnosis of cancer and other
diseases. Cell Res. 2008;18(10):997-1006.
27. Dai Y, Huang YS, Tang M, et al. Microarray
analysis of microRNA expression in
peripheral blood cells of systemic lupus
erythematosus patients. Lupus. 2007;16:
939-946.
28. Rodriguez A, Vigorito E, Clare S, et al.
Requirement of bic/microRNA-155 for nor-
mal immune function. Science.
2007;316(5824):608-611.
29. Wang J, Chen J, Sen S. MicroRNA as
Biomarkers and Diagnostics. J Cell Physiol.
2016;231(1):25-30.
30. Nakamachi Y, Kawano S, Takenokuchi M.
MicroRNA-124a is a key regulator of prolif-
eration and monocyte chemoattractant pro-
tein 1 secretion in fibroblast-like synovio-
cytes from patients with rheumatoid arthri-
tis. Arthritis Rheum. 2009;60(5):1294-1304.
31. Pérez-Sánchez C, Ruiz-Limón P, Aguirre
MA, et al. Mitochondrial dysfunction in
antiphospholipid síndrome: implications in
the pathogenesis of the disease and effects
of coenzyme Q10 treatment. Blood. 2012;
119(24):5859-5870.
32. Ruiz-Limón P, Barbarroja N, Pérez-Sánchez
C, et al. Atherosclerosis and cardiovascular
disease in systemic lupus erythematosus:
effects of in vivo statin treatment. Ann
Rheum Dis. 2015;74(7):1450-1458.
33. Barbarroja N, Pérez-Sanchez C, Ruiz-Limon
P, et al. Anticyclic citrullinated protein anti-
bodies are implicated in the development of
cardiovascular disease in rheumatoid arthri-
tis. Arterioscler Thromb Vasc Biol.
2014;34(12):2706-2716.
34. Navickas R, Gal D, Laucevičius A,
Taparauskaitė A, Zdanytė M, Holvoet P.
Identifying circulating microRNAs as bio-
markers of cardiovascular disease: a system-
atic review. Cardiovasc Res.
2016;111(4):322-337.
35. Huang Y, Qi Y, Du JQ, Zhang DF.
MicroRNA-34a regulates cardiac fibrosis
after myocardial infarction by targeting
Smad4. Expert Opin Ther Targets. 2014;
18(12):1355-1365.
36. Li N, Wang K, Li PF. MicroRNA-34 Family
and Its Role in Cardiovascular Disease. Crit
Rev Eukaryot Gene Expr. 2015;25(4):293-
297.
37. Licholai S, Blaż M, Kapelak B, Sanak M.
Unbiased Profile of MicroRNA Expression
in Ascending Aortic Aneurysm Tissue
Appoints Molecular Pathways Contributing
to the Pathology. Ann Thorac Surg.
2016;102(4):1245-1252
38. Choi SY, Yun J, Lee OJ, et al. MicroRNA
expression profiles in placenta with severe
preeclampsia using a PNA-based microarray.
Placenta. 2013;34(9):799-804.
39. Sheikh AM, Small HY, Currie G, Delles C.
Systematic Review of Micro-RNA
Expression in Pre-Eclampsia Identifies a
Number of Common Pathways Associated
with the Disease. PLoS One. 2016;11(8):
e0160808.
40. Akehurst C, Small HY, Sharafetdinova L, et
al. Differential expression of microRNA-206
and its target genes in preeclampsia. J
Hypertens. 2015;33(10):2068-2074.
41. Chan ES, Cronstein BN. Methotrexate--how
does it really work? Nat Rev Rheumatol.
2010;6(3):175-178.
42. Tsochandaridis M, Nasca L, Toga C, Levy-
Mozziconacci A. Circulating microRNAs as
clinical biomarkers in the predictions of
pregnancy complications. Biomed Res Int.
2015;2015:294954.
43. Zhao Z, Moley KH, and Gronowski AM.
Diagnostic potential for miRNAs as bio-
markers for pregnancy-specific diseases.
Clinical Biochemistry. 2013;46(10-11):953-
960.
44. Wu L, Zhou H, Lin H, et al. Circulating
microRNAs are elevated in plasma from
severe preeclamptic pregnancies.
Reproduction. 2012;143(3):389-397.
45. Tatsumi K, Mackman N. Tissue Factor and
Atherothrombosis. J Atheroscler Thromb.
2015;22(6):543-549
haematologica | 2018; 103(5)