Vox Sanguinis (2007) 93, 289–297

© 2007 The Author(s)

ORIGINAL PAPER Journal compilation © 2007 Blackwell Publishing Ltd.
DOI: 10.1111/j.1423-0410.2007.00989.x

A modified rapid monoclonal antibody-specific immobilization

Blackwell Publishing Ltd

of platelet antigen assay for the detection of human platelet

antigen (HPA) antibodies: a multicentre evaluation
K. Campbell,1 K. Rishi,2 G. Howkins,2 D. Gilby,1 R. Mushens,3 C. Ghevaert,1 P. Metcalfe,4 W. H. Ouwehand1,5 & G. Lucas3
1
National Health Service Blood and Transplant, Cambridge Centre, Long Road, Cambridge CB2 2PT, UK
2
National Health Service Blood and Transplant, Oxford Centre, John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
3
National Health Service Blood and Transplant, Bristol Centre, Southmead Road, Bristol BS10 5ND, UK
4
National Institute for Biological Standards and Control, Blanche Lane, South Mimms, Potters Bar, Hertfordshire EN6 3QG, UK
5
Department of Haematology, University of Cambridge, Long Road, Cambridge CB2 2PT, UK

Background The monoclonal antibody-specific immobilization of platelet antigens

Received: 23 April 2007,
revised 3 August 2007,
accepted 13 August 2007,
published online 7 August 2007
(MAIPA) assay is the cornerstone technique for the detection and identification of
human platelet antigen (HPA) antibodies. However, the original technique described
by Kiefel and colleagues requires approximately 8 h adding to diagnostic delay.
Moreover, proficiency exercises indicate that there are substantial variations in the
MAIPA protocol, and that these may account for interlaboratory differences in
sensitivity and specificity.
Study Design and Methods A review of current MAIPA assay protocols from six
laboratories together with performance in quality-assessment schemes identified
several key variables potentially affecting the assay results. An optimized protocol
was derived and assay time reduced to 5 h. The modified rapid MAIPA (MR-MAIPA)
assay was evaluated using 61 samples with a range of HPA antibodies typically
encountered in cases of fetomaternal alloimmune thrombocytopenia (n = 22), post-
transfusion purpura (n = 8), platelet refractoriness (n = 7) and other platelet immune
conditions (n = 24). The sensitivity of the assay was assessed using three inter-
national standards and the recombinant HPA-1a antibody CamTran007. The results
obtained were compared with the original findings obtained with the local MAIPA
assays. In addition, four different glycoprotein IIb/IIIa capture monoclonal antibodies
were evaluated for their effect on assay sensitivity.
Results Complete concordance was found between the original MAIPA results and
those obtained with the new assay when testing a selected panel of clinical samples.
The modified assay had nanogram level sensitivity for the detection of HPA-1a anti-
bodies and titration of HPA-1a and HPA-5b antibody sensitivity standards yielded
end-points equal to or greater than the mean recorded in international workshops.
Conclusion The MR-MAIPA assay offers improved turnaround for the detection of
HPA antibodies without loss of sensitivity.

Correspondence : Willem Ouwehand, MD, PhD, Department of
Haematology, University of Cambridge, Long Road, Cambridge CB2 2PT, UK
E-mail: who1000@cam.ac.uk
Introduction
Human platelet antigen (HPA) antibodies are of clinical
significance in fetomaternal alloimmune thrombocytopenia
(FMAIT), post-transfusion purpura (PTP) and platelet refrac-
toriness (PR). Rapid and sensitive detection and identification

289
290 K. Campbell et al.
of HPA alloantibodies is imperative for good clinical care.
The monoclonal antibody (mAb)-specific immobilization of
platelet antigens (MAIPA) assay is the cornerstone technique
for HPA antibody detection and identification in many
platelet immunology laboratories [1–3]. However, quality-
assessment exercises have revealed wide interlaboratory
variation in the proficiency of HPA antibody detection by
the MAIPA assay [1–3]. Nonetheless, several laboratories
participating in the platelet immunology quality-assurance
exercises organized by the National Institute for Biological
Standards and Control (NIBSC) have shown consistently
good performance over the 10 most recent exercises. The
MAIPA assay protocols of these laboratories were analysed
and it was deduced that minor variations [e.g. sample
volume, incubation time and temperature, reagents and
glycoprotein (GP) capture mAbs] had significant conse-
quences for the sensitivity and sometimes specificity of the
assay. Although a rapid enzyme-linked immunosorbent assay
(ELISA)-based assay for the detection of HPA antibodies is
commercially available, it lacks the flexibility of the MAIPA
assay and previous reports have suggested a reduced sensi-
tivity when compared to the MAIPA assay [4,5]. In a collab-
orative project between the platelet immunology laboratories
of the National Health Service Blood and Transplant (NHSBT)
and NIBSC, a modified rapid (MR)-MAIPA assay protocol was
developed. The MR-MAIPA assay was initially validated in
the NHSBT laboratory in Cambridge and a common protocol
was defined. The sensitivity of the assay for the detection of
HPA-1a antibodies was assessed by titration of the recom-
binant HPA-1a antibody CamTran007 [6,7], the polyclonal
anti-HPA-1a potency standard 03/152 [1] and two World
Health Organization (WHO) International Reference Rea-
gents for sensitivity of HPA-1a (93/710) and HPA-5b (99/666)
[8] antibody detection. To determine interoperator variation
61 blinded samples were tested by four different operators in
three laboratories. Finally, the 42 samples with antibodies
against GPIIb/IIIa were tested with four different capture
antibodies (RFGP56, NBS-PAB-1, NBS-PAB-6 and NBS-
5B12) to determine their effect on the sensitivity of the
MR-MAIPA assay.
Finally, we analysed the results for evidence of coimmuno-
precipitation of GPIIb/IIIa with human leucocyte antigen
(HLA) class I.
Materials and methods
HPA-1a mAb
The human recombinant HPA-1a mAb CamTran007 with a
Kd of 6 × 108 M/l was used to determine the sensitivity of the
assay [9]. This antibody has been obtained from the immune
B cells of a case of FMAIT by V gene phage display techno-
logy [6] and the V genes were recombined with constant
Journal compilation © 2007 Blackwell Publishing Ltd., Vox Sanguinis (2007) 93, 289–297
domain genes to generate a recombinant immunoglobulin
G1 (IgG1) antibody [10].
Standard polyclonal HPA antibody preparations
The WHO minimum sensitivity reagents for the detection of
anti-HPA-1a (93/710) [11] and anti-HPA-5b (99/666) [8], and
the anti-HPA-1a potency standard (03/152) [1] containing a
100-international units (IU) polyclonal IgG anti-HPA-1a were
used in the initial validation studies. None of these reagents
contained HLA class I or other HPA antibodies. All were supplied
freeze-dried and were reconstituted according to the accom-
panying instructions and diluted in phosphate-buffered saline
(PBS) supplemented with 0·2% w/v bovine serum albumin (BSA).
Anti-GP mouse mAbs
The following mAbs were used in the assay, anti-GPIIb/IIIa:
RFGP56, ascites diluted 1 : 1000 (Cymbus, Southampton,
UK; a kind gift from Professor A. Goodall, University of
Leicester), NBS-PAB-1, culture supernatant (CS) diluted
1 : 10; NBS-PAB-6, CS diluted 1 : 10; NBS-5B12, CS diluted
1 : 10, NIBSC-85/661, CS, diluted 1 : 3; anti-GPIbα: NBS-
PAB-5, CS diluted 1 : 10; anti-GPIa/IIa: NBS-P16, CS diluted
1 : 10; and anti-HLA class I: W6/32, CS diluted 1 : 5.
Patient samples
Both ethylenediaminetetraacetic acid (EDTA) plasma and
serum samples from patients and normal controls were used
to validate the MR-MAIPA assay. The 61 blinded samples
were clinical referrals to NHSBT platelet immunology labo-
ratories and contained a representative range of HPA and
platelet GP specificities typically found in referrals to refer-
ence laboratories. The panel was deliberately enriched with
some samples where HPA antibodies had been difficult to
detect at initial referral. High and low titre HPA antibodies
and examples of single and multiple HPA specificities were
included, as were panreactive GPIIb/IIIa, GPIbα/IbβIX and
GPIa/IIa antibodies. All samples were coded before testing so
that operators were blinded for the antibody specificities. The
results obtained with the MR-MAIPA assay were analysed
and the results were compared with the original data
obtained by the laboratory that had submitted the sample.
Panel platelets
Samples were tested for the presence of HPA and HLA class I
antibodies against a panel of cryopreserved platelets (Table 1)
from group O, HPA and HLA class I genotyped donors. Each of
the HPA-1, -2, -3 and -5 alleles were present in homozygosity
on at least one panel cell. Panel platelets were stored until use
at the temperature of liquid nitrogen vapour phase and, after
© 2007 The Author(s)
 Rapid monoclonal antibody-specific immobilization of platelet antigen assay 291

Table 1 HPA and HLA antigen profile of cryopreserved platelet panel BSA). The resuspended platelets were incubated with 25 μl of
either test or control serum/plasma for 30 min at 37 °C in a
HPA -1a -1b -2a -2b -3a -3b -5a -5b HLA-A HLA-B well of a round bottom microtitre plate (Dynex Technologies,
Worthing, UK). Serum/plasma was removed by two washings
PL No.
with 200 μl TBS/BSA and the pelleted platelets were
1 0 + + 0 0 + + 0 2,11 40,51
resuspended in 50 μl TBS/BSA before the addition of appro-
2 + 0 + 0 + 0 0 + 1,3 8
3 + 0 0 + + 0 + 0 2,29 40,44 priately diluted GP-specific mAb. Following a further
4 + 0 + 0 0 + + 0 2,32 62,27 incubation for 30 min at 37 °C, the platelets were washed
5 0 + + 0 + 0 + 0 2,24 35,44 three times with TBS/BSA, followed by solubilization of the
6 0 + + 0 0 + 0 + 2,23 7,44 platelets with 0·5% Triton-X-100 (BDH Laboratory Supplies,
7 + 0 0 + 0 + + + 2 8,44 Poole, UK) for 15 min at 20 °C. The microtitre plates were
centrifuged at 1400 g for 15 min to precipitate unsolubilized
PL, platelet donor; No., rank number; HPA, human platelet antigen; HLA,
platelet debris. Of the supernatant, 100 μl was transferred to
human leucocyte antigen; 0, antigen absent; +, antigen present.
a flat bottom well microtitre plate (Nunc Maxisorp, Roskilde,
Denmark) that had been previously coated with goat anti-
removal of cryoprotectant by washing with PBS/EDTA/BSA, mouse Ig diluted 1 : 500 (Jackson Immunoresearch Labora-
a suspension 1 × 108 platelets/ml was prepared. Recovered tories, West Grove, PA, USA) and washed and blocked with
panel platelets were stored at 4 °C in PBS/EDTA/BSA for not wash buffer (TBS, Nonidet P40 substitute 0·5% v/v, Tween
more than 3 days. The detection of HPA-15 antibodies was 0·01% v/v, CaCl2 0·01% w/v, BSA 0·1% w/v). The plates were
not evaluated because CD109 is not preserved after cryop- incubated for 30 min at 37 °C to allow effective capture of
reservation of panel platelets [12]. the trimolecular complex of antigen, murine and human
antibodies. The lysate was discarded, and the plate was
washed six times with 125 μl of wash buffer. Bound human
Historical testing
IgG was detected by the addition of 100 μl of a 1 : 5000 dilu-
The 61 blinded samples were investigated at the time of referral tion of horseradish peroxidase-conjugated goat anti-human
for HPA antibodies by the indirect platelet immunofluorescence IgG (Jackson Immunoresearch Laboratories) in wash buffer.
test (PIFT), the local MAIPA assay, and in one sample a whole- After 60-min incubation at 20 °C, the plates were washed a
platelet solid-phase ELISA. The local MAIPA protocols were further six times and the colour was developed by addition
based on the original protocol [13] and subsequent modifi- of 100 μl o-phenylenediamine (OPD) solution (DakoCytoma-
cations [14,15] but none included shortened incubation times tion, Ely, UK; OPD tablets, 2 mg, code no. S2045) and 5 μl of
at higher temperature or a standardized panel of mAbs. 30% w/v hydrogen peroxide (H2O2) in reverse osmosis water.
The enzymatic reaction was stopped by the addition of 100 μl
of 0·5 M H2SO4 to each well. The absorbance was read at
MR-MAIPA assay
490 nm with a reference filter of 630 nm. The major differ-
Panel platelets (100 μl) were centrifuged, the supernatant was ences between the MAIPA assay as described by Kiefel and
decanted and the platelet pellet was resuspended in 50 μl of colleagues [13–15] and the MR-MAIPA are summarized in
TBS/BSA (Tris-buffered saline supplemented with 0·2% w/v Table 2.

Table 2 Comparison of modified original and MR-MAIPA protocol

Step/reagent Modified original MAIPA [13–15] MR-MAIPA

Buffer used in serum and mAb sensitization phases PBS TBS

Platelets/test (×107/l) 2 1
Volume serum used per test (μl) 5–200 25
Solubilization buffer Nonidet P40 Triton-X-100
Centrifugation of lysate 15 000 g for 10 min at 4 °C 1400 g for 15 mins at RT
Incubation of trimolecular complex with capture antibody 90 min at 4 °C 30 min at 37 °C
Incubation with goat anti-human IgG horseradish peroxidase 90 min at 4 °C 60 min at 20 °C
Total assay time ~8 h ~5 h

mAb, monoclonal antibody; PBS, phosphate buffered saline; TBS, Tris-buffered saline; RT, room temperature; trimolecular complex: platelet glycoprotein,
murine monoclonal antibody and human IgG.

© 2007 The Author(s)

Journal compilation © 2007 Blackwell Publishing Ltd., Vox Sanguinis (2007) 93, 289–297
292 K. Campbell et al.
Study design
The initial approach was to analyse the details of the
MAIPA protocols in consistently ‘good performer’ laboratories
participating in the NIBSC proficiency scheme and to identify
factors that had a negative or positive effect on sensitivity.
Based on this analysis, a protocol was derived (the MR-MAIPA
assay) that was extensively evaluated in a step-wise process.
The first step was to use four well-characterized HPA anti-
body reference reagents: the human recombinant IgG1
HPA-1a antibody CamTran007 and the three aforementioned
polyclonal anti-HPA standards. The characterization of all
four reagents has been extensive and has been reported
elsewhere [9–11,16]. The validation phase of the study was
preceded by a workshop in which staff from NHSBT platelet
immunology laboratories were trained to the MR-MAIPA
standard operating procedure. The 61 blinded samples were
tested twice by a single operator (K. Campbell, Cambridge)
and then by three different operators in NHSBT laboratories
in Bristol, Cambridge (different operators) and Oxford in
batches of 30, 16 and 15 samples, respectively, using NBS-
PAB-1, NBS-P16 and NBS-PAB-5 and the standard set of
panel platelets (Table 1). In addition all 42 samples with
GPIIb/IIIa antibodies were tested with four capture mAbs:
NBS-PAB-1, NBS-PAB-6, NBS-5B12 and RFGP56 by a single
operator in the NHSBT Cambridge laboratory.
Results
Sensitivity of the MR-MAIPA assay
The sensitivity for the detection of HPA-1a antibodies was
assessed with the human recombinant HPA-1a antibody
CamTran007 and with the two polyclonal NIBSC reference
reagents. Antibody CamTran007 was tested over a concen-
tration range from 10 μg/ml to 0·015 ng/ml using the GPIIb/
IIIa capture antibodies NBS-PAB-1, NBS-PAB-6 and NBS-
5B12 (Fig. 1). The results indicate that the sensitivity for the
detection of this intermediate affinity HPA-1a antibody is in
Journal compilation © 2007 Blackwell Publishing Ltd., Vox Sanguinis (2007) 93, 289–297
the nanogram range and that this was independent of the
capture antibody.
Serial dilutions of the anti-HPA-1a potency standard
NIBSC 03/152 in PBS/BSA were tested using the same three
capture mAbs. The potency standard was detected at a titre
of 1 in 512 with NBS-PAB-1 and NBS-5B12, and at a titre of
1 in 256 with NBS-PAB-6 (Fig. 2a). Dilutions of the anti-
HPA-1a sensitivity standard NIBSC 93/710 showed reactivity
at a titre of 64 with all three capture antibodies (Fig. 2b),
while titres ranging from 1 in 4 to 1 in 64 had been observed
in international proficiency exercises [2]. The first stage of
the assessment of the MR-MAIPA assay was completed by
testing the HPA-5b sensitivity antibody preparation NIBSC
99/666 and a titre of 1 in 32 was consistently observed with
capture antibody NBS-P16 (Fig. 2c), which compares well
with the titres observed in previous exercises ranging from
undiluted to 1 : 64 (median 1 : 4) [2].
Clinical samples
Clinical samples typically referred to a platelet immunology
laboratory were used in the second phase of the assessment
of the MR-MAIPA. The characteristics of the 61 blinded sam-
ples are presented in Table 3, with 49 samples containing
HPA antibodies, a unique sample with anti-Moua [17], five
with panreactive GP antibodies, one that reacted in the PIFT
only, two with HLA antibodies only (no platelet-specific anti-
bodies) and three inert samples. Testing of the 61 blinded
samples by a single operator (KC) in the Cambridge laboratory
on two occasions with two different but partially overlapping
sets of capture mAbs (RFGP56, NBS-P16, CLB-MB45 and W6/
32 vs. NBS-PAB-1, NBS-PAB-6, NBS-5B12, NBS-P16, NBS-
PAB-5 and W6/32) showed complete concordance between
the two tests with positive results for all but two of the 58
samples with platelet antibodies and no significant signal
with the three inert samples (data not shown). The anti-Moua
that was previously only detected in a solid-phase platelet
ELISA but not in PIFT or MAIPA, was also non-reactive in the
modified MAIPA. Similarly, the ‘PIFT only’ sample was still
Fig. 1 Titration of anti-HPA-1a CamTran007 by
MAIPA with three capture antibodies. Sensitivity
of the MR-MAIPA to detect recombinant human
immunoglobulin G1 (IgG1) anti-HPA1a
CamTran007. Doubling dilutions of the antibody
were tested by MAIPA with an HPA-1a1a
homozygous platelets. The capture mAbs
NBS-PAB-1, NBS-PAB-6 and NBS-5B12 were
used. x-axis, dilution of CamTran007 in μg/ml.
y-axis, absorbance at 490 nm.
© 2007 The Author(s)
 Rapid monoclonal antibody-specific immobilization of platelet antigen assay 293

Fig. 2 Titration of three NIBSC anti-HPA

standards by MAIPA with three capture
antibodies. Doubling dilutions of three NIBSC
standards with HPA antibodies were tested by
MR-MAIPA. Samples of HPA-1a homozygous
platelets were sensitized with the samples
prepared from the samples. Dilution of the
standard is on the x-axis and absorbance at
490 nm on the y-axis. (a) Anti-HPA-1a potency
standard 03/152 was tested with three GPIIb/IIIIa
capture antibodies, NBS-PAB-1, NBS-PAB-6 and
NBS-5B12; (b) anti-HPA-1a sensitivity standard
93/710 was tested as per (a) and (c); anti-HPA-5b
sensitivity standard 99/666 was tested with the
GPIa/IIa capture antibody P16.

non-reactive in the modified MAIPA. The identified specifi-
cities were identical to specificities previously determined by
the indirect immunofluorescence test and the local MAIPA.
HLA class I antibodies were detected in 32 (52%) samples,
and in 28 samples in combination with HPA antibodies. GP-
reactive antibodies without allelic-restricted reactivity were
detected in all five (8%) samples known to contain these
antibodies (three against GPIIb/IIIa, one against GPIbα/IbβIX
and one against GPIIb/IIIa, GPIbα/IbβIX and GPIa/IIa).
In a parallel exercise the same 61 blinded samples were
tested in batches of 30, 16 and 15 by three different operators
in the platelet immunology laboratories situated in NBS Bristol,
Cambridge (different operators) and Oxford laboratories,
respectively. Again the anti-Moua and the ‘PIFT only’ samples
were non-reactive and the expected results were obtained
with all but one of the remaining samples, one weak anti-
HPA-1a, sample OX-01 that was only positive in the original
local MAIPA only when the test serum volume was tripled.
This sample gave positive results in the modified MAIPA with
a normal serum volume when tested with two panels of mAbs
(above), but was negative in the modified MAIPA when
retested by the originating laboratory. However, after break-
ing of the code and upon retesting with a double volume of
serum a positive signal was obtained.

© 2007 The Author(s)

Journal compilation © 2007 Blackwell Publishing Ltd., Vox Sanguinis (2007) 93, 289–297
294 K. Campbell et al.

Table 3 Characteristics of the 61 blinded samples PAB-6 and NBS-5B12 and two with RFGP56. The results of
another unusual sample (OX-13) are included in Table 4. This
Anti-HPA/GP FMAIT PTP PR Other Total sample had been selected on basis that discrepant results had
been reported in two quality assurance exercises [Sample 4
1a 10 5 6 21
(01/448) in workshop 2001-A and Sample 3 (03/138) in
1a+5b 2 2
workshop 2003-A]. In both exercises the detection of this
1a+2b+3a 1 1
HPA-1a antibody varied between laboratories with clear
1b 2 2 1 5
1b+2b 1 1 positive results in 17 but negative results in the remaining
1b+3b 1 1 nine laboratories. The failure to detect this antibody in some
1b+3b+5b+15b 1 1 laboratories may be multifactorial, but results from this
3a 1 2 3 study showed that robust positive signals were obtained with
3a+5a 1 1 all four capture antibodies (Table 4).
3b 1 1 Overall, the absorbance signal produced by NBS-PAB-1
5a 1 1 was slightly higher, but not significantly different from the
5b 6 4 10 NBS-PAB-6 signal (2·61 ± 1·25 vs. 2·44 ± 1·30, P = 0·068,
5b+2b 1 1
t-test for paired samples; average ± 1 × SD on panel cell 1 with
Moua 1 1
24 samples containing anti-HPA-1a).
GPIIb/IIIa 3 3
There is a potential risk with MAIPA assays that incomplete
GPIb/IX 1 1
GPIIb/IIIa, Ibα/IbβIX, Ia/IIa 1 1 solubilization of the platelet membrane may lead to coimmu-
PIFT only 1 1 noprecipitation of platelet-specific GP complexes with HLA
HLA only 2 2 class I. This is thought to be one of the important causes of false-
Inert 3 3 positive results for HPA antibodies with panel cells that carry
Total 22 8 6 25 61 the cognate HLA antigen. The results were analysed for pos-
sible evidence of coimmunoprecipitation (Fig. 3). All samples
HPA, human platelet antigen; GP, glycoprotein; FMAIT, fetomaternal
were tested for reactivity against HLA class I using mAb W6/
alloimmune thrombocytopenia; PTP, post-transfusion purpura; PR,
32. In 18 of the 24 samples with anti-HPA-1a, reactivity on
platelet refractoriness.
one or more panel cells with HLA class I was observed. In 14
of the 18 samples the cognate HLA class I antigen was present
The effect of capture mAb on assay sensitivity was inves- on the one or more of the HPA-1b1b cells. Of the 14 samples
tigated in detail with the 42 samples with anti-GPIIb/IIIa with HPA-1a antibodies, five produced positive reactions
reactivity. These samples were tested in the MR-MAIPA assay with the captured GPIIb/IIIa of the HPA-1b1b genotype
with four panel cells (cells 1, 2, 3 and 6) and capture clones (Fig. 3; average ± 1 SD, 0·288 ± 0·071, range 0·194–0·385).
RFGP56, NBS-PAB6, NBS-5B12 in addition to PAB-1. The
correlation of signal intensities between capture antibodies
Discussion
was generally good, but interestingly, in five samples
(Table 4) false-negative results were obtained with one of the The detection of HPA antibodies is of clinical relevance for
four capture antibodies, one each with NBS-PAB-1, NBS- the management of patients with FMAIT, PTP and PR. The

Table 4 Selected samples with GPIIb/IIIa antibodies that produced false-negative results in the MR-MAIPA with one of the four capture mAbs

Capture mAb

Sample code Condition NBS-PAB-1 NBS-PAB-6 NBS-5B12 RFGP56 Specificity anti-HPA

OX-01 PTP 0·118 0·087 0·254 0·107 1a+5b

OX-06 ITP 0·015 0·380 0·160 0·701 1a+5b
CAM-16 PR 0·141 0·187 0·140 0·074 1b
OX-09 PR 0·179 0·216 0·549 0·052 1b
CAM-04 NAIT 0·169 0·140 0·097 0·121 3a
OX-13 Blood donor 0·914 0·865 0·981 0·609 1a

HPA, human platelet antigen; sample OX-13 is positive with all four capture antibodies but was included because it produced a clear positive–negative split
in previous NIBSC proficiency schemes. False-negative results are in bold and italic.
PTP, post-transfusion purpura; ITP, immune thrombocytopenic purpura; FMAIT, fetomaternal alloimmune thrombocytopenia.

© 2007 The Author(s)

Journal compilation © 2007 Blackwell Publishing Ltd., Vox Sanguinis (2007) 93, 289–297
 Rapid monoclonal antibody-specific immobilization of platelet antigen assay
Fig. 3 Evidence for coimmunprecipitation of
HLA class I with GPIIb/IIIa. Five of the 14 samples
with HPA-1a antibodies and concurrent HLA class
I antibodies also produced a significant reading
with the GPIIb/IIIa derived from the HPA-1b1b
platelets. The results by MR-MAIPA with capture
mAbs NBS-PAB-1 and W6/32 on the HPA-1a1a
(first set of two columns) and HPA-1b1b panel
platelets (second set of two columns). Absorbance
at 490 nm is presented on the y-axis. Sample
BR-04 is an FMAIT case and the four others are
transfusion purpura cases.
definition of the molecular basis of HPA antigens and the
introduction of DNA-based genotyping assays has greatly
improved the accuracy of typing of donors and patients
[18–20]. In contrast, the detection and identification of HPA
antibodies remains cumbersome and proficiency of HPA
antibody detection varies considerably between laboratories
that participate in quality-assurance exercises [2,3]. There are
a number of reasons for poor performance in HPA antibody
detection. First, sensitivity for antibody detection is tech-
nique dependent [2,3]. Second, within a single technique,
interoperator variation may occur and even in this study in
which all operators were trained to a standard procedure,
discrepancies were observed between two laboratories with
one example of anti-HPA-1a. Finally, the composition of the
platelet panel cells and particularly the need for homozygo-
sity for all HPA alleles is important.
Half the laboratories participating in the biannual Inter-
national Society of Blood Transfusion platelet immunology
exercise use the MAIPA assay for antibody detection and
identification [3]. This technique originally described by
Kiefel and colleagues [13–15] has the advantage that the
HPAs are in the native configuration during the sensitization
phase. Therefore, the assay has remained the cornerstone
technique in many platelet immunology laboratories in spite
of the availability of a commercial ELISA for the detection of
HPA antibodies. Since the original description of the assay,
modifications of the MAIPA principle have been reported and
variables have been identified that may influence sensitivity
[14,15].
Using the recombinant HPA-1a antibody CamTran007
that has a Kd of 6 × 10–8 M/l, it was demonstrated that the
sensitivity of the MR-MAIPA is in the upper nanogram range
(Fig. 1). This level of sensitivity compares poorly with the
sensitivities of the antiglobulin and autoanalyser techniques
for anti-D detection that are in the lower nanogram range
[21]. There are several possible reasons for the lower sensitiv-
ity of the MAIPA assay. First, the affinity of HPA antibodies
plays a critical role. The affinity of two recombinant HPA-1a
© 2007 The Author(s)
Journal compilation © 2007 Blackwell Publishing Ltd., Vox Sanguinis (2007) 93, 289–297
295
antibodies, CamTran007 and 19-7 (Kd 6·0 and 3·97 × 10–8 M/l,
respectively [9]), is substantially lower than most monoclonal
RhD antibodies, with some having affinities in the lower
nanomolar range [22]. Second, GPIIb/IIIa is known to
undergo a dramatic shape change upon platelet activation
with a major repositioning of the plexn–semaphorin–integrin
(PSI) domain that carries the HPA-1 antigens [23,24]. It is
possible that after membrane solubilization, GPIIb/IIIa may
undergo conformational change that may lead to the displace-
ment of HPA antibodies. If so, then one could speculate that
certain capture antibodies may, depending on their epitope
and affinity, constrain this gymnastic tour de force of
GPIIb/IIIa.
These results indicate that the sensitivity of the MR-MAIPA
may be greater than current MAIPA assays but that question
cannot be unequivocally answered by this study. However, there
is evidence that the MR-MAIPA is at least as sensitive as the
original assay. First, the end-point titres of the recombinant
HPA-1a antibody and of the three polyclonal NIBSC refer-
ence reagents were similar or superior to those obtained in
both the NIBSC proficiency schemes [2] and the International
Society of Blood Transfusion platelet immunology workshops
[25] (Figs 1 and 2). Second, there was complete concordance
between the original results and the MR-MAIPA results for
the panel of 61 blinded samples that had been deliberately
enriched for samples with HPA antibodies that had been dif-
ficult to detect at initial referral. Third, the HPA-3a antibodies
in sample CAM-04 from a clinically significant case of
FMAIT were reactive in the MR-MAIPA but not in our previ-
ously used MAIPA assays. In addition, it has been shown that
the sensitivity of the MR-MAIPA is maintained when the
assay is transferred to other operators in the same or other
laboratories provided this process is supported by a standard
operating procedure and training.
Our study confirmed the observation made by several
groups that the type of capture mAb influences the sensitivity
of the MAIPA (Table 4). It demonstrates that no single capture
antibody is ideal. The observed differences between them
296 K. Campbell et al.
could be explained by epitope overlap between capture anti-
body and the HPA antibody, or variation in the ability of
capture antibodies to constrain the GPIIb/IIIa conformation
after membrane solubilization. The three different capture
antibodies did not affect the sensitivity of the MR-MAIPA for
the detection of recombinant anti-HPA-1a CamTran007 (Fig. 1).
This observation is compatible with the notion that monoclonal
or polyclonal HPA-1a antibodies of good affinity are detected
irrespective of the capture antibody, as long as there is no
epitope overlap. Interestingly, the signal produced with the
anti-HPA-1a potency standard 03/152 was less good with
clone NBS-PAB-6 when compared with NBS-PAB-1 (Fig. 2a).
Recent studies from our laboratory have shown that NBS-
PAB-1 and NBS-PAB-6 target overlapping epitopes [26],
suggesting that it may be a difference in affinity, which
explains the poorer performance of NBS-PAB-6.
Two samples (Table 4) that produced equivocal results
between our laboratories or in quality exercises were partic-
ularly informative. Sample OX-13 with anti-HPA-1a showed
a positive–negative split between laboratories participating
in the NIBSC proficiency schemes. This sample produced
robust positive results in the MR-MAIPA with all four capture
mAbs. The HPA-3a antibodies in sample CAM-04 was asso-
ciated with widespread petechiae and a platelet count of
12 × 109/l in a term neonate after an uneventful delivery.
This antibody was not consistently detected by the MAIPA
protocols that were previously operational in our laboratories
but a good signal was observed in the MR-MAIPA with three
out of four mAbs.
The issue of coimmunoprecipitation remains a concern
when using the MAIPA assay for HPA antibody detection.
Incomplete solubilization may cause immunoprecipitation of
the HLA class I molecule by anti-integrin antibodies. In 14 of
the 24 samples with anti-HPA-1a and anti-HLA, HLA anti-
bodies were against HLA class I antigens present on the HPA-
1b1b panel cell. Nine of the 14 samples were non-reactive on
the HPA-1b1b panel cell with GPIIb/IIIa capture antibodies.
The remaining five samples produced some signal (Fig. 4) but
four of these were from PTP patients and the reactivity with
the HPA-1b1b platelets is most likely caused by autoanti-
bodies against GPIIb/IIIa [27,28]. The single FMAIT sample
(BR-04) with HPA-1a antibodies and positive results with the
HPA-1b1b platelets (OD 0·385) contained strong HLA class I
antibodies that may have caused the weak positive reaction
on the antithetical HPA-1b1b panel cell. There were insuffi-
cient samples containing HPA-2 or HPA-5 reactive antibodies
to answer the question of possible coimmunoprecipitation of
HLA class I with GPIbα/IbβIX and GPIa/IIa, respectively.
In conclusion, this study describes the extensive validation
of a modified and rapid MAIPA assay protocol using a large
number of samples with HPA antibodies and several capture
mAbs. The reduction of the duration of the assay from 8 to
5 h without loss of sensitivity is of key importance to diagnostic
Journal compilation © 2007 Blackwell Publishing Ltd., Vox Sanguinis (2007) 93, 289–297
laboratories. In addition, the sensitivity of the MR-MAIPA
assay for the detection of HPA-1a antibodies is in the higher
nanogram range and appears better than the standard MAIPA
assay. Improved assay sensitivity has clinical relevance as it
has been shown that HPA-1a negative, HLA DRB3*0101 positive
mothers with thrombocytopenic neonates may have a negative
HPA-1a antibody screen during pregnancy with antibodies only
becoming detectable in the postpartum period [29]. In addi-
tion, severe FMAIT cases with cerebral bleeds occur in women
with low-titre HPA-1a antibodies. This study demonstrates
that the detection of some antibodies (possibly low titre) by
MAIPA assay can be significantly influenced by the choice
of capture antibody. The use of recombinant HPA antigens [30]
and alternative immunoassay platforms like Luminex® beads
may be a possible avenue to further increase the sensitivity
of HPA-1a antibody detection. Experiments are currently
proceeding in our laboratories to investigate this [31].
Acknowledgements
We thank the staff of the NHSBT platelet immunology labora-
tories at Bristol, Cambridge and Oxford for their expertise in
platelet immunology. We gratefully acknowledge Stephen
Garner, Dr Nick Watkins, Graham Smith, Dave Allen, Rizwan
Yusuf and Gerry Campbell for their assistance and valuable
comments during the manuscript preparation.
References
1 Allen D, Rigsby P, Bessos H, Berry J, Wilson D, Ouwehand WH,
Urbaniak S, Metcalfe P: Collaborative study to establish the first
international standard for quantitation of anti-HPA-1a. Vox
Sang 2005; 89:100–104
2 Metcalfe P, Allen D, Chapman J, Ouwehand WH: Interlaboratory
variation in the detection of clinically significant alloantibodies
against human platelet alloantigens. Br J Haematol 1997;
97:204–207
3 Bessos H, Wilson DW, Metcalfe P, Allen D, Urbaniak SJ: Report
on the 12th International Society of Blood Transfusion platelet
immunology workshop. Vox Sang 2005; 89:105–113
4 Lucas GF, Rogers SE: Evaluation of an enzyme-linked immuno-
sorbent assay kit [GTI PakPlus(R)] for the detection of antibodies
against human platelet antigens. Transfus Med 1999; 9:385–386
5 Lubenko A, Savage J: Antigen capture ELISA for platelet anti-
body detection: choice of conjugate influences assay result.
Transfus Med 2000; 10:213–218
6 Griffin HM, Ouwehand WH: A human monoclonal antibody
specific for the leucine-33 (P1A1, HPA-1a) form of platelet
glycoprotein IIIa from a V gene phage display library. Blood
1995; 86:4430–4436
7 Watkins NA, Schaffner-Reckinger E, Allen DL, Howkins GJ,
Brons NHC, Smith GA, Metcalfe P, Murphy MF, Kieffer N,
Ouwehand WH: HPA-1a phenotype–genotype discrepancy reveals
a naturally occurring Arg93Gln substitution in the platelet β3 integrin
that disrupts the HPA-1a epitope. Blood 2002; 99:1833–1839
© 2007 The Author(s)
 Rapid monoclonal antibody-specific immobilization of platelet antigen assay
8 Metcalfe P, Ouwehand WH, Sands D, Barrowcliffe TW: Collabora-
tive studies to establish the first WHO Reference Reagent for
detection of human antibody against human platelet antigen-
5b. Vox Sang 2003; 84:237–240
9 Santoso S, Kroll H, Andrei-Selmer CL, Socher I, Rankin A,
Kretzschmar E, Watkins NA, Ouwehand WH: A naturally occur-
ring LeuVal mutation in β3-integrin impairs the HPA-1a epitope:
the third allele of HPA-1. Transfusion 2006; 46:790–799
10 Armour KL, Clark MR, Hadley AG, Williamson LM: Recombinant
human IgG molecules lacking Fcγ receptor I binding and mono-
cyte triggering activities. Eur J Immunol 1999; 29:2613–2624
11 WHO: WHO Expert Committee on Biological Standardisation,
48th report. WHO Technical Report Series 1999; 889:17
12 Berry JE, Murphy CM, Smith GA, Ranasinghe E, Finberg R,
Walton J, Brown J, Navarrete C, Metcalfe P, Ouwehand WH:
Detection of Gov system antibodies by MAIPA reveals an
immunogenicity similar to the HPA-5 alloantigens. Br J Haematol
2000; 110:735–742
13 Kiefel V, Santoso S, Weisheit M, Mueller-Eckhardt C: Mono-
clonal antibody – specific immobilization of platelet antigens
(MAIPA): a new tool for the identification of platelet-reactive
antibodies. Blood 1987; 70:1722–1726
14 Kiefel V: The MAIPA assay and its applications in immunohae-
matology. Transfus Med 1992; 2:181–188
15 Kiefel V, Scheld S, Freitag E, Kroll H, Mueller-Eckhardt C: [Two
modifications of MAIPA assays for the demonstration of platelet
antibodies]. Beitr Infusionsther Transfusionsmed 1994; 32:237–
239
16 Watkins NA, Armour KL, Smethurst PA, Metcalfe P, Scott ML,
Hughes DL, Smith GA, Williamson LM, Clark MR, Ouwehand
WH: Rapid phenotyping of HPA-1a using either diabody-based
hemagglutination or recombinant IgG1-based assays. Transfusion
1999; 39:781–789
17 Masters R, Taaning E: Three cases of platelet alloimmunisation
associated with the presence of a novel platelet-specific anti-
body. Vox Sang 1998; 75:242–246
18 Bugert P, McBride S, Smith G, Dugrillon A, Klüter H, Ouwehand
WH, Metcalfe P: Microarray-based genotyping for blood groups:
comparison of gene array and 5′-nuclease assay techniques with
human platelet antigen as a model. Transfusion 2005; 45:654–
659
19 Hurd CM, Cavanagh G, Schuh A, Ouwehand WH, Metcalfe P:
Genotyping for platelet-specific antigens: techniques for the
detection of single nucleotide polymorphisms. Vox Sang 2002;
83:1–12
© 2007 The Author(s)
Journal compilation © 2007 Blackwell Publishing Ltd., Vox Sanguinis (2007) 93, 289–297
297
20 Metcalfe P, Cavanagh G, Hurd C, Ouwehand WH: HPA genotyp-
ing by PCR-SSP: report of 4 exercises. Vox Sang 1999; 77:40–43
21 Anstee D, Klein H: Mollison’s Blood Transfusion in Clinical
Medicine, 11th edn. Oxford, UK, Blackwell Publishing, 2005
22 Bye JM, Carter C, Cui Y, Gorick BD, Songsivilai S, Winter G,
Hughes-Jones NC, Marks JD: Germline variable region gene seg-
ment derivation of human monoclonal anti-Rh(D) antibodies.
Evidence for affinity maturation by somatic hypermutation and
repertoire shift. J Clin Invest 1992; 90:2481–2490
23 Xiao T, Takagi J, Coller BS, Wang JH, Springer TA: Structural
basis for allostery in integrins and binding to fibrinogen-
mimetic therapeutics. Nature 2004; 432:59–67
24 Xiong JP, Stehle T, Diefenbach B, Zhang R, Dunker R, Scott DL,
Joachimiak A, Goodman SL, Arnaout MA: Crystal structure of
the extracellular segment of integrin α Vβ3. Science 2001;
294:339–345
25 Metcalfe P: Platelet antigens and antibody detection. Vox Sang
2004; 87:82–86
26 Jennings NS, Harmer IJ, Campbell K, Stafford P, Smith GA,
Metcalfe P, Benton MA, Marsh JC, Ouwehand WH: Molecular
characterization of the variable domains of an αIIbβ3-specific
immunoglobulin M κ platelet cold agglutinin in a follicular
lymphoma patient with treatment refractory autoimmune
thrombocytopenia: idiotypic overlap between αIIbβ3 integrin
antibodies. Transfusion 2007; 47:499–510
27 Watkins NA, Smethurst PA, Allen D, Smith GA, Ouwehand WH:
Platelet αIIbβ3 recombinant autoantibodies from the B-cell
repertoire of a post-transfusion purpura patient. Br J Haematol
2002; 116:677–685
28 Taaning E, Tonnesen F: Pan-reactive platelet antibodies in post-
transfusion purpura. Vox Sang 1999; 76:120–123
29 Williamson LM, Hackett G, Rennie J, Palmer CR, Maciver C,
Hadfield R, Hughes D, Jobson S, Ouwehand WH: The natural
history of fetomaternal alloimmunization to the platelet-specific
antigen HPA-1a (PlA1, Zwa) as determined by antenatal screen-
ing. Blood 1998; 92:2280–2287
30 Li CQ, Garner SF, Davies J, Smethurst PA, Wardell MR,
Ouwehand WH: Threonine-145/methionine-145 variants of
baculovirus produced recombinant ligand binding domain of
GPIbα express HPA-2 epitopes and show equal binding of von
Willebrand factor. Blood 2000; 95:205–211
31 Stafford P, Ghevaert C, Campbell K, Smith G, Williamson L,
Watkins NA, Ouwehand WH: Recombinant mini-β3 integrin
fragments for the detection of HPA-1 antibodies. Blood 2005,
106:Abstract 218