UNIVERSITÉ DE STRASBOURG

ÉCOLE DOCTORALE DES SCIENCES DE LA VIE ET DE LA SANTÉ

______________________________

THÈSE présentée par :

Hélène BOIGARD
soutenue le : 16 Décembre 2020

pour obtenir le grade de : Docteur de l’université de Strasbourg

Discipline/ Spécialité : Virologie

Vaccins à particules pseudo-virales contre la dengue et le Zika

PRESIDENTE DU JURY :

Mme SCHNEIDER-MYZONI Odile Enseignante-chercheure de l’Université de Strasbourg

AUTRES MEMBRES DU JURY :

Mme FOURNEL Sylvie Enseignante-chercheure de l’Université de Strasbourg

Mr PREVOST Gilles Enseignante-chercheure de l’Université de Strasbourg

Mme IMBERT Isabelle Enseignante-chercheure de l’Université d’Aix-Marseille

Mr BRAND Denys Enseignant-chercheur de l’Université de Tours

ZEISEL Mirjam Ingénieure de recherche de l’Université de Lyon

 Dengue and Zika viral-like particle-based vaccines
!
PART I. INTRODUCTION ............................................................................................................................... 4

CHAPTER 1. GENERAL ASPECTS .............................................................................................................................. 4

I. Virus discoveries .................................................................................................................................... 5
Discovery of dengue virus .................................................................................................................................. 5
Discovery of Zika virus ........................................................................................................................................ 6
II. Epidemiology of dengue and Zika viruses .............................................................................................. 6
Dengue ............................................................................................................................................................... 6
Zika ..................................................................................................................................................................... 7
III. Cycles of transmission ............................................................................................................................ 7
Vector-borne transmission ................................................................................................................................. 7
Non-vector borne transmission ......................................................................................................................... 8
IV. Clinical presentations ............................................................................................................................. 9
Dengue virus infection ..................................................................................................................................... 10
a) Clinical symptoms........................................................................................................................................ 10
b) Innate immune response ............................................................................................................................ 11
c) Antibody-dependent Enhancement ............................................................................................................ 12
Zika virus infection ........................................................................................................................................... 12
a) Clinical symptoms........................................................................................................................................ 12
b) Neurodegenerative syndrome and microcephaly ....................................................................................... 13
Treatments of DENV and ZIKV infections ......................................................................................................... 14
CHAPTER 2. CHARACTERISTICS OF DENGUE AND ZIKA VIRUSES ................................................................................... 15
I. Introduction ......................................................................................................................................... 15
II. Viral Genome and Structure ................................................................................................................ 15
III. Maturation of dengue and Zika viruses ............................................................................................... 17
IV. Replication and Temperature .............................................................................................................. 18
V. Humoral response, Virus neutralization and Quaternary epitopes ..................................................... 20
CHAPTER 3. DENGUE AND ZIKA VACCINE DEVELOPMENT AND CANDIDATES .................................................................. 22
I. Dengue vaccines in clinical trials ......................................................................................................... 22
Live attenuated virus vaccines ......................................................................................................................... 22
a) Attenuation through sequential passaging ................................................................................................. 22
b) Genetically engineered live-attenuated virus ............................................................................................. 22
Inactivated dengue virus vaccines .................................................................................................................... 23
DNA and RNA vaccines ..................................................................................................................................... 24
Recombinant subunit protein vaccines ............................................................................................................ 24
II. Zika vaccine in clinical trials ................................................................................................................. 24
Live attenuated virus vaccines ......................................................................................................................... 25
Inactivated Zika virus vaccines ......................................................................................................................... 25
Nucleic acid-based vaccines ............................................................................................................................. 25
a) DNA vaccines ............................................................................................................................................... 25
b) mRNA vaccine.............................................................................................................................................. 26
CHAPTER 4. VIRAL-LIKE PARTICLES ....................................................................................................................... 27
I. Introduction ......................................................................................................................................... 27
II. Current use .......................................................................................................................................... 27
III. Preclinical dengue VLP, RSP and Replicons .......................................................................................... 28

PART II. METHODS - STRATEGY FOR THE DEVELOPMENT OF VLP ................................................................ 31

I. Dengue proteins of interest ................................................................................................................. 31

The Envelope protein ....................................................................................................................................... 31
The Precursor Membrane protein .................................................................................................................... 31
 The Pre-membrane protein and envelope protein dynamic duo ..................................................................... 32
Capsid, VLP secretion, and harmony among the structural triad ..................................................................... 33
Identification and characterization of NS3 protease activity ........................................................................... 34
Glycosylation sites ............................................................................................................................................ 35
Genes of interest optimization ......................................................................................................................... 36
a) Suppression of the retention signal............................................................................................................. 36
b) Enhancement of the furin cleavage for improved maturation of VLP ......................................................... 36
c) Improvement of the NS3 catalytic site. ....................................................................................................... 37
II. Dengue VLP strategy summary ............................................................................................................ 37
III. Transition towards Zika VLP ................................................................................................................ 38
IV. Expression system and temperature .................................................................................................... 39
V. Characterization of the VLPs and evaluation as vaccine candidates. .................................................. 40
In vitro expression and formation of the VLPs ................................................................................................. 40
a) 4G2 .............................................................................................................................................................. 40
b) 3H5 .............................................................................................................................................................. 40
c) C10 .............................................................................................................................................................. 40
d) Zika antibodies ............................................................................................................................................ 41
Immunogenicity in Balb/c mice ........................................................................................................................ 41
VI. Step by step ......................................................................................................................................... 41

PART III. RESULTS .................................................................................................................................... 43

CHAPTER 1. DENGUE-2 VIRUS-LIKE PARTICLES (VLP) BASED VACCINE ELICITS THE HIGHEST TITERS OF NEUTRALIZING ANTIBODIES
WHEN PRODUCED AT REDUCED TEMPERATURE ............................................................................................................. 43
I. Article................................................................................................................................................... 43
II. Discussion ............................................................................................................................................ 53
Missing elements ............................................................................................................................................. 53
Influence of Capsid in polyprotein processing and maturation ....................................................................... 54
Structural findings – variations among serotypes. -- ........................................................................................ 54
CHAPTER 2. ZIKA VIRUS-LIKE PARTICLE (VLP) BASED VACCINE .................................................................................... 56
I. Article................................................................................................................................................... 56
II. Discussion ............................................................................................................................................ 79
a) Zika viral strains and glycosylation .............................................................................................................. 79
b) Zika virus strains and neutralizing antibodies ............................................................................................. 80
c) Antibody Dependent Enhancement and dengue virus ................................................................................ 80

PART IV. CONCLUSION ............................................................................................................................. 82

Acknowledgments

I would like to thank Dr Van Mai Cao-Lormeau for her valuable advice and guidance in th
writing of this dissertation. I am also thankful to the Doctoral School 414 for inviting me to
present my research.
 Part I. Introduction

Chapter 1. General aspects

Arthropod-borne viral diseases have a major impact on the global public health. Among these
diseases, dengue fever remains a challenge for potentially 2.5 billion people per year around
the globe. Dengue virus is transmitted by mosquitoes belonging to the Aedes genus, which
can also transmit other arthropod-borne viruses (arboviruses), such as Zika virus. While
dengue virus has been recognized as a public health threat for centuries, Zika virus generated
its first large scale outbreaks Oceania in 2013-2014 and in the Americas in 2015-2016.

Historically, dengue fever has mostly been an issue for developing countries. Despite decades
of research, in the year 2000s, there was still no vaccine nor treatment against the virus. In
order to attract more attention to unmet medical needs, the World Health Organization
(WHO) established a list of Neglected Tropical Diseases (NTDs) in which dengue fever was
included. In 2011, WHO identified five strategies to address these debilitating diseases 1. The
strategies include i) the use of preventive chemotherapy, ii) intensification of case-detection
and case management, iii) Vector and intermediate host control, iv) veterinary public health
activities at the human–animal interface, and v) provision of safe water, sanitation and
hygiene. Highlighting the impact of NTDs on the worldwide population, the WHO initiative was
supported by thirteen corporations, foundations such as the Bill and Melinda Gates
foundations, the World Bank, and some governments (UK, US, United Arab Emirates,
Bangladesh, Brazil, Mozambique, and Tanzania). By joining efforts they committed financial
contributions to the initiative in the London declaration 2,3. The five strategies identified by
WHO were compiled in an implementation roadmap with 2020 as deadline. The W.H.O. set
R&D priorities to address NTDs and the development of human vaccines is placed on the top
of the list4. The implementation of a dengue virus vaccine was one of the five technical
elements designed by the W.H.O. to reach the 2020 roadmap goal5.The first anti-dengue
vaccine candidate to have received commercialization approval is the live attenuated
tetravalent vaccine DengVaxia from Sanofi. In 2016, Dengvaxia® was authorized in eleven
countries. However, the vaccination resulted in disease enhancement in naïve children when
exposed to the wild-type virus. Amid these detrimental side-effects, the government of the
Philippines revoked its authorization in 2017. In contrast to dengue fever, Zika virus infection
has emerged recently. Therefore, vaccine opportunities and challenges were only recently
uncovered.

The body of this work presents the preclinical development of a Viral-Like Particle vaccine
candidate for dengue virus which later on served as template for the development of a VLP-
based vaccine against Zika virus. This study was carried out from March 2011 to July 2017.
I. VIRUS DISCOVERIES

Discovery of dengue virus

The earliest cases of dengue virus infection may have been reported in China, in the Chinese
medical encyclopedia from the Chin Dynasty (265–420 AD) describing an illness related to
“poison water” and “flying insects”. Later on, outbreaks of dengue-like illness were reported
in the French West Indies and Panama in 1635 and 1699 in the XVIIth century6. The first
published scientific article about dengue fever disease was published in 1852 regarding an
outbreak in 1850 in Savannah, Georgia (USA)7. In 1907 Ashburn & Craig reported the
observation of viral particles and demonstrated that dengue fever was transmitted by
mosquitoes8.

The discovery of dengue virus (DENV) was concurrent with the advances in Yellow Fever
research, another mosquito-borne disease9. In the early XXth century, the worldwide burden
of Yellow Fever post-World War I pushed the scientific community to work on a vaccine.
Through different attempts and strategies, Max Theiler managed to attenuate the Yellow
fever virus (YFV) by sequential passages in mouse brains, which increased neurotropism of the
viral strains and diminished hepatic damage as well as systemic symptoms10. The resulting
viral strain was further characterized and referred to as the 17D strain. This attenuated YFV
was eventually used as a vaccine. Theiler received the Noble Price for his discovery.

Before World War II, dengue fever outbreaks were mostly reported in North America, in the
Caribbean’s islands, East Africa, India and South East Asia11. From the 1940s, increased people
mobility led to the emergence of another epidemiological pattern: the hyperendemic
transmission of dengue virus. this pattern became more obvious in the early 1980’s, with co-
circulation of more this type of transmission co-circulation of more than 2 serotypes12. The
hyperendemic transmission is “the continuous circulation of multiple viral serotypes in an area
where a large pool of susceptible hosts and a competent vector (with or without seasonal
variation) are constantly present “13.

In 1945, vaccine pioneer Albert Sabin attempted to develop a first Dengue vaccine by
attenuation of the virus by sequential passages in mice14. Unfortunately, the technique
successfully used for the development of another flavivirus, the YFV 15, was not effective
against DENV. Sabin was able to identify the existence of two DENV serotypes using “dermal”
and “intracerebral” neutralization tests. The serotype 1 was first discovered in 1943 in Japan
follow by the serotype 2 in 1945 in Hawaii16. In 1974, the International Committee on Taxon
of viruses grouped YFV and DENV as Togaviridae. In the following decade, the committee
revised this classification and separated these viruses as Flaviruses17.

A century has passed from DENV’s first isolation to the development of a vaccine. The first and
only vaccine commercially available to this day is the Sanofi Vaccine Dengvaxia®, which uses
 the live-attenuated CYD17 YFV backbone carrying structural proteins of each of the four DENV
serotypes18.

Discovery of Zika virus

The first isolation of Zika virus (ZIKV), in 1947, was from a rhesus monkey initially used as live
sentinel for Yellow fever surveillance in the Zika forest of Uganda 19. Neutralization test
showed this mosquito-borne virus, was a different virus than YFV and DENV20. Shortly after,
ZIKV infection was described in Nigeria in 195421. Over the following fifty years, the virus only
had a minimal impact on human health with low number of mild cases whereas
seroprevalence studies indicated circulation of the virus in Africa and Asia22.

In 2007, ZIKV was detected for the first time in Oceania, on Yap Island (Federated States of
Micronesia). Six years later, in 2013, ZIKV caused its first significant outbreak in French
Polynesia23. For the first time, Van-Mai Cao-Lormeau et al. described the association between
ZIKV infection and Guillain Barré Syndrome24. In 2015, ZIKV appeared in Brazil and during the
following year this virus caused dramatic outbreaks in several countries in South and Central
America. Didier Musso published an article suggesting the virus was carried from Polynesia to
Brazil when four countries from the Pacific concurred in the Va’a World Sprint Championship
canoe race in Brazil in August 201425. Further phylogenic analyses conducted by different
groups supported the hypothesis that the ZIKV strain that founded the outbreaks in the
Americas was introduced from the Pacific islands, but they estimated the introduction of ZIKV
into Brazil rather to have occurred in late 2013 or early 201426,27.

A few months after its appearance in Brazil, ZIKV moved from the status of NTDs to be notified
as Public Health Emergency of International Concern (PHEIC), after congenital abnormalities,
including microcephaly, were reported in fetuses and newborns whose mother had been
exposed to ZIKV during pregnancy.

II. EPIDEMIOLOGY OF DENGUE AND ZIKA VIRUSES

Dengue

Dengue virus infections dramatically increased over the last fifty years. Approximately 2.5
billion people are at risk of DENV infection and there may be 390 million infections yielding 96
million symptomatic cases, globally, and every year28. Dengue endemic regions correlate with
suitable conditions for mosquito survival. The climate changes have led to the geographical
expansion of dengue vectors, particularly Aedes aegypti during the XXth century and more
 recently Aedes albopictus.29 The dengue hemorrhagic fever was described for the first time in
1954 during an epidemic episode in the Philippine islands 30. The four dengue virus serotypes
appeared independently during endemic transmission cycles between humans and arthropod
vectors31. Infection by a given serotype may induce lifelong immunity against the infecting
virus, but only short-term cross-protection against heterologous serotypes. However,
successive infections with different serotypes may increase the risk of severe dengue,
including severe hemorrhagic signs and shock syndrome.

Today, dengue virus is circulating in all the different parts of the world. In 2016 dengue
infections surged in the Americas with 2.38 million reported cases which was followed by a
significant reduction in 2017. However, the year of 2019 registered the highest number of
cases. The virus was responsible for an unprecedented outbreak in Afghanistan. In 2020,
infections increased in Bangladesh, Brazil, Cook Islands, Ecuador, India, Indonesia, Maldives,
Mauritania, Mayotte (Fr), Nepal, Singapore, Sri Lanka, Sudan, Thailand, Timor-Leste and
Yemen32.

Zika

Following the 2015 outbreak in Brazil, the virus spread in multiple countries via the arthropod
vector. According to the WHO, 87 countries and territories across the globe recorded
autochthonous cases of ZIKV infection in 2019. Data are limited in African, South East Asia and
Western Pacific however a retrospective serological study revealed seroprevalence of 9% in
children under five-year-old in Indonesia and 10% in asymptomatic adults in Laos. A cluster
outbreak occurred in India in 2018. There is to this day 61 countries have circulating Aedes
vector which could allow the spread of Zika virus 33. Because of the recent outbreak, there is
less information available regarding the global spread of Zika virus than dengue virus,
however, the continuous spread of the vector increases the areas at risk of an imminent
outbreak.

III. CYCLES OF TRANSMISSION

Vector-borne transmission

The two major mosquito specie responsible for DENV and ZIKV transmission are Aedes Aegypti
and Ae albopictus34. These hematophagous arthropods transmit the virus during blood meal.
In addition to direct transmission in urban epidemic cycle, the viruses can circulate in Enzootic
cycle which allows the virus to replicate in vertebrate host before being transmitted by the
vector (Figure 1).

Figure 1: Cycles of transmission for vector-borne infection. (extracted from Muller and Cao-Lormeau , 2018 1)

The geographical expansion of these two mosquito vectors is placing at risk over fifty percent
of the worldwide population. Ae Aegypti is mostly found in regions located between latitude
10 degrees North and latitude 10 degrees South. Ae albopictus is comparatively more present
in the northern hemisphere, including in the United States and Europe (Figure 2).

Non-vector borne transmission

Non-vector borne transmission of arboviruses has been observed. Indeed, both viruses can be
transmitted through blood transfusion. Zika virus can also be transmitted sexually. The virus
has been found in semen for over one hundred days after infection. (Error! Reference source n
ot found.). Zika virus has detected more frequently in saliva than in blood of patient
presenting symptoms of Zika fever during the 2013 French Polynesian outbreak 35. Potential
Sexual transmission was suspected during the same outbreak when the virus was identified in
semen sample from a patient previously recovered from Zika fever 8 -week prior36. In a
prospective study, semen samples were tested following early onset of Zika fever symptoms.
The Viral RNA was detected up to 141 days after infection37. The potential sexual transmission
of the virus was confirmed in 2016 when partners of individuals who just return from the
epidemic areas showed symptoms of Zika infection. The sexual transmission was recorded in
male-to-female as well as male-to-male cases38,39. Besides the sexual transmission, the vertical
transmission of the virus has had a dramatic effect on the public health. At the heart of the
 2015 outbreak in northern Brazil, serological studies concluded 35-87% of the microcephaly
in newborn infants were caused by Zika virus for a prevalence of 2-3 cases per 1000 live birth40.
The virus spread outside of the outbreak areas resulting in vertical transmission is other part
of the world. Congenital Zika Syndromes are not anecdotic with cases recorded in Vietnam
and Thailand in South East Asia, and in Portugal, imported from Angola in Africa after the 2015
outbreak 41–44

Figure 2: predicted global distribution of Aedes Aegypi (A) and Aedes Albopictus (B). Figure extracted from. 1. The
occurrence is represented from blue for 0 to 1 in red. The spatial resolution is 5km x 5km (adapted from Guzmann et al.
1).

IV. CLINICAL PRESENTATIONS

Dengue and zika viruses mostly cause asymptomatic infections. In mild cases, clinical
symptoms are similar in mild cases and distinguish themselves in severe cases. DENV infection
may induce severe hemorrhagic signs that potentially lead to shock syndrome and death. ZIKV
infection can cause neurological syndrome and defects, notably Guillain Barre Syndrome (GBS)
in adults, and microcephaly in in utero-exposed newborns. The following section describes the
clinical presentations of DENV and ZIKV infections.

Dengue virus infection

a) Clinical symptoms

Dengue virus infection is associated with a wide range of symptoms. From asymptomatic
infection to mild dengue fever to severe dengue disease – previously named as Dengue
Hemorrhagic Fever (DHF) end Dengue Shock Syndrome (DSS), until the WHO revised in 2009
the clinical classification by widening the breadth of symptoms. This “new classification” got
criticized for not defining clearly “severe dengue” clinical presentation and for not being
helpful in improving the prediction of severe symptoms45. Dengue fever is currently
characterized as a combination of two or more clinical symptoms in a febrile person who
traveled to or lives in a dengue-endemic area. The incubation time ranges from five to seven
days. The clinical presentation of the disease evolves in three consecutive phases, as shown
in Figure 3.
Figure 3: Dengue virus infection clinical phases.

b) Innate immune response

Transmission of DENV from an infected vector to the human host occurs during mosquito’s
blood fed through the injection of infectious saliva into the skin. The virus can reach the
dermis, the endodermis as well as the bloodstream. Multiple cell types are infected early on,
such as the macrophages and dendritic cells46. The aforementioned are then transferred in
the lymph nodes alongside with the virus. In the lymphatic systems, the virus interacts with
monocytes and macrophages. The virus is transported through the lymphatic system infecting
multiple cell types throughout the body47,48.

The acute illness appears about seven to ten days after the mosquito bite and it lasts about a
week (Figure 4)49. The immune reaction peaks after the 6th day. The IgM titers drop back to
the basal level starting on Day 14, while the IgG titers will drop slowly after fifty days. The
viremia falls on Day 4. Dengue virus RNA is detectable using RT-PCR until the IgM levels
plateau.
Figure 4 : Viremia and antibody kinetics after primary dengue infection. In red is represented the viral load; In dark blue,
the IgG; in light blue, the IgM (adapted from Guzman et al.)

c) Antibody-dependent Enhancement

There are four different dengue serotypes DENV-1, -2, -3, and -4. Infection by one of these
serotypes generates homotypic immunity and short term cross-protective immunity against
heterologous serotypes50. The immune serum provides long-lasting homotypic immunity by
containing neutralizing antibodies whereas the heterotypic immunity will eventually wane.
The vast majority of the antibodies is non- or weakly neutralizing. Elicitation of antibodies is
part of the humoral response, which is primordial in the defense against pathogens. In the
case of DENV, non- or weekly neutralizing antibodies play a crucial role in the exacerbation of
the immune response leading to the phenomenon called Antibody-Dependent Enhancement
of disease (ADE). The two main targets triggering the humoral response are the structural
proteins prM and E. To a lesser extent, antibodies are also generated against the only secreted
protein, NS1, as well as C, prM, NS3, NS4B, and NS551.

When weakly or non-neutralizing antibodies bind to the virus, the complex subsequently
attaches to FcyR bearing cells such as monocytes and macrophages. The virus is not
neutralized and infects the mononuclear phagocytes triggering inflammatory cytokines and
chemokines laying the ground for severe hemorrhagic dengue fever as well as shock syndrome.

Neutralization titer by circulating antibodies is one of the correlates of protection against dengue
fever. Multiple factors are influencing the clinical outcome to severe dengue. A prospective
pediatric study confirmed that a serum neutralization titer above 1:320 dilution in adults and up
to 1:1280 correlates with protection against the virus.52.

Zika virus infection

a) Clinical symptoms
It is estimated that fifty percent of the ZIKV infections are asymptomatic53. The clinical
presentation for symptomatic cases ranges from flu-like symptoms to neurodegenerative
syndromes, such as the Guillain Barre syndromes (Figure 5).

Clinical symptoms of ZIKV infection occur from three to seven days following the bite of an
infected mosquito. Mild symptoms include fever, arthralgia and myalgia, headache and
conjunctivitis, fatigue, and potentially a rash. Severe cases in adults can lead to meningitis,
encephalitis, and thrombocytopenia54–57. During pregnancy, the virus can cross the placenta.
The infection delays skull and brain development leading to microcephaly and ocular
defects58.

Systemic Nervous System

Guillain-Barré Congenital Zika

Asymptomatic Flu-like disease
Syndrome Syndrome

Figure 5: Zika virus infection and its various clinical outcomes. (adapted from Pierson and Diamond, 2018)

b) Neurodegenerative syndrome and microcephaly

Guillain Barre syndrome (GBS) was first seen associated with ZIKV infections during the French
Polynesian outbreak of 201324. In a case-control study conducted by Cao-lormeau et al.,
scientists observed correlation between the onset of GBS and the presence of anti-ZIKV IgM
in the patient sera. Indeed, they compared a GBS group with a comparable group of patients
presenting non-febrile illness and the absence of viral ZIKV RNA. The team of researchers also
compared the GBS group with a group of patients testing positive for ZIKV RNA but who did
not present neurological complications. This former group was included to assess the
potential influence of DENV infection. As a result, it was found that GBS onset begun after the
acute phase of ZIKV infection was passed (6-10 days post-infection). No correlation with prior
DENV infection and GBS was found.

The clinical presentation for GBS diagnosis is a combination of symptoms such as weaknesses
in the upper and lower extremities, a decrease in reflexes of paretic arms or legs, an increase
in protein concentration in the cerebrospinal fluid59. GBS is an immune-mediated bilateral
polyradiculoneuropathy l, which has been observed following other infectious diseases. The
symptoms beings 2 to 8 weeks after infection60–62. Patients are fully recovered three months
after their hospitalization.

This debilitating neuropathology is not the only long-term consequence of ZIKV. Indeed, ZIKV
has an affinity for the nervous tissues and can cause Congenital Zika Syndrome (CZS) in which
microcephaly in one of the most obvious consequence, by crossing the placenta in utero and
infecting the fetus. The study published in 2019 by Brady et al. showed the retrospective
analysis of microcephaly after the 2015-2017 outbreak in northern Brazil. Infection of
expecting mother increases by 17-fold the risk of developing microcephaly in the fetus.
Microcephaly is an impairment of the development of the brain resulting in smaller than
average circumference of the head. The head and the brain of the newborn may stop growing
after birth. Microcephaly can be diagnosed by ultrasound during the third trimester of
pregnancy. After delivery, the head circumference is measured and compared with the WHO
standard to determine the diagnosis. Long term defects associated to microcephaly can be
epilepsy, cerebral palsy, learning disabilities, hearing loss, and vision problems63.

Treatments of DENV and ZIKV infections

There is no specific treatment for both DENV and ZIKV infections except symptoms
management therapies. Depending on the severity of the disease, the clinical presentation of
DENV and ZIKV infection can be similar. It is essential to rule out the later before
administrating non-steroidal anti-inflammatory drugs, which can increase the risk of
hemorrhage during DENV infection64. Depending on the severity of symptoms, treatments of
DENV infection range from fluid intake, maintaining the right level of hydration to pain and
fever management with administration of non-steroid anti-inflammatory drugs. Severe case
treatments rely on fluid therapy to compensate with fluid loss via intravenous crystalloid
solution and blood transfusion. Hematocrit, blood pressure, and fluid accumulation (ascitic
fluids) are indicators of the patient’s condition and its response to symptoms management
therapy.
 Chapter 2. Characteristics of dengue and Zika viruses

The body of this work began in 2011 and was initially focused on DENV. In 2015, the outbreaks
caused by ZIKV in the Americas supported the rationale to applying the strategy initially
developed for DENV this other flavivirus. As a result, the following chapter describes mainly
the properties and characteristics of dengue viral proteins, followed by a comparison between
DENV and ZIKV structural response to temperature.

I. INTRODUCTION

Dengue and Zika viruses are part of the Flavivirus genus within the Flaviviridae Family and they
are phylogenetically close (Figure 6)65. The genetic alignment between envelop (E) gene
sequences of the four DENV serotypes presents 65-70% identity and 55-58% with E gene
sequence of ZIKV 66,67.

Figure 6: Phylogenic tree of the main mosquito-borne flaviviruses and alphavirus CHIKV based on nucleotide sequences.
(extracted from Moni and Lio’, 2017)

II. VIRAL GENOME AND STRUCTURE

The viral genome of these two enveloped viruses is composed of a single-stranded positive-
sense RNA that contains one open-reading frame (ORF) encoding from the amino-terminal
end to the carboxyterminal end, three structural proteins: capsid protein (C), the pre-
membrane protein (prM) and the envelope protein (E); and seven non-structural proteins:
NS1, NS2A/B, NS3, NS4A/B and NS5 (Figure 7a). The non-structural proteins are responsible
for the viral replication. The genome is packaged thanks to C-
Figure 7: Dengue viral genome and structure. A Schematic representation of flavivirus genome organization and processing
of the polypeptide into mature viral proteins. b, Top, primary structure of DENV E-protein ectodomain showing EDI (red),
EDII (yellow) and EDIII (blue). Bottom, ribbon and molecular structure of DENV E antiparallel dimer (top view, PDB 1UZG). c,
Surface representation of ZIKV. The asymmetrical unit is indicated by the black triangle. Blue, up to 130 Å; cyan, up to 150
Å; green, up to 190 Å; yellow, up to 230 Å; red, above 231 Å. d, Depth-based, color-coded cross-sectional view of ZIKV, with
color coding as in c. e, E- and M-protein Cα backbone structure on the mature ZIKV virion, highlighting the herringbone
organization of E dimers (color coding as in b). f, Schematic representation of the mature ZIKV, as shown in e.”. (extracted
from Campos et al. 2018)
 protein. The capsid-RNA complex is then enveloped by a lipid bilayer membrane embedded
with prM and E protein. The E protein has 3 domains (Figure 7b), and once folded, forms
antiparallel homodimers organized in herringbone patterns at the surface of the mature viral
particle. The external symmetry of the virion is organized in an icosahedral symmetry.

III. MATURATION OF DENGUE AND ZIKA VIRUSES

Dengue and Zika viruses have identical maturation processes. The viral genome is translated
in one single ORF. The resulting polyprotein undergoes several cleavages during its maturation
through the Endoplasmic Reticulum (ER) and the Golgi apparatus (Figure 8).

Figure 8: Dengue virus polyprotein and its interaction with endoplasmic reticulum membrane. (extracted from Umareddy
et al. s. d.; Leitmeyer et al. 1999).

The viral particles assemble sequentially by moving through consecutive cellular

compartments and undergoing several pH-dependent proteolytic cleavages. Once the virus is
assembled, it comprises at its surface 180 copies of the E and prM proteins at its surface. The
“pr” portion of prM covers the fusion peptide within the E protein during the maturation
process until its release in the extracellular compartment. The prME complex associates with
two other heterodimers in the Endoplasmic Reticulum forming an heterotrimers protruding
from the viral particle giving a “spiky” surface to the virion. In a single spike-like heterotrimer,
each hydrophobic fusion peptides of the E protein is covered by the loops and the Ans69
carbohydrate of the “pr” peptide providing hydrophilic properties and therefore prevent
premature fusion of the E protein (Figure 9a).

During transport in the Trans-Golgi network, vesicle’s pH drops and the heterotrimer re-
organize in 90 heterodimers, lying flat, at the surface of the virion giving a smooth aspect of
the viral particle (Figure 9b). This change of conformation renders accessible the furin cleavage
site between pr and M68. This step is critical for the maturation of viral particles69. Furin has
its optimal activity at neutral pH. However, the pr/M cleavage site is accessible only under
acidic conditions. Pr covers the fusion loop until it is released to prevent premature fusion
within the cells before the full maturation. Low pH conditions, such as the TGN, triggers E
 protein fusion. Once released in the extracellular space, the neutral pH allows dissociation of
the pr from the E-fusion loop (Figure 9c).

Figure 9: “Flavivirus maturation pathway. (a) Immature flavivirus particle at neutral pH. The virus buds into the neutral pH
environment of the endoplasmic reticulum (ER), and the immature particle contains spiky trimers of prM-E heterodimers.
One (prM-E) trimer is shown, with E colored by domains as in Figure 2; the stem and transmembrane domains are displayed
brown and translucent. The prM protein is colored cyan. It is unclear whether the region of M (denoted by a question
mark) that normally forms a membrane-proximal alpha-helix in the mature particle has this conformation in the immature
particle. (b) Immature flavivirus particle at low pH. As the virions progress through the trans-Golgi network (TGN), the
acidic environment induces the rearrangement from (prM-E) trimers to (prM-E) dimers (one of which is shown) with similar
orientation and organization as in the mature virus. This low pH-induced transition is reversible in DV (…). The transition
allows access for furin to cleave prM to pr plus M, as indicated by the red arrows. At low pH, pr remains bound to E. (c)
Mature flavivirus particle at neutral pH. The particle is secreted into the extracellular space. The pr polypeptide dissociates
from E at neutral pH, readying the virus for fusion during entry.” Extracted from
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2783195/

Particles are characterized as immature when the viral particle is released, and the Precursor
protein is not cleaved. It is estimated that 40% of the released particles are immature. A
subpopulation of partially mature particles has also been observed as well where only a
portion of the pr-M cleavage has occurred, and the surface of the virus is heterogenous 70.

The M protein acts as a link between the “pr” portion and the lipid transmembrane domains.
The immune system recognizes both “pr” and M proteins; however, the C protein remains
unseen by the immune system. The NS1 protein is the only non-structural protein released in
the extracellular compartment during an infection.

IV. REPLICATION AND TEMPERATURE

Dengue and Zika viruses can replicate in both the vector and human host. The effect of the
replication temperature on the viral structure was initially solely explored on DENV, since ZIKV
was not relevant to vaccine development prior to 2016.
Using a DENV replicon with Green Fluorescent Protein (GFP), a previous study showed a higher
expression of GFP in mammalian cells of DENV replicon at 30⁰C than 37⁰C. A much slower
decay of the GFP signal, 100h after transient gene induction, was recorded at a lower
temperature. The infectivity of replicons was then tested on highly sensitive Raji-DCSIGNR
cells. Analysis of relative infection showed a 5-fold increase in infectious Recombinant
Subviral Particles (RSP) production at 30⁰C than 37⁰C71.

In 2010, similar results were described when co-transfecting DEN non-structural proteins, the
reporter gene Luciferase, and structural proteins. The highest yields were observed when
transferring the mammalian cells, 24h after transfection cells, from 37⁰C to 30⁰C. The
authors found these conditions to be better than producing the VLP at 30⁰C all along the
process, demonstrating the importance of a slow packing phase. Infectious VLP were
successfully neutralized by 4G2 monoclonal antibody recognizing a structural epitope located
at the fusion peptide suggesting proper folding of the viral particle 72.

Additionally, production temperature is investigated at the molecular level. Electron

microscopy (EM) study was performed of DENV-2 (DV2 16681 strain) propagated in C6/36
mosquito cells at 28⁰C, purified and incubated at different temperatures: 28, 31, 33, 35, and
37⁰C. Mature virus produced at 28⁰C was smooth, whereas viruses produced at temperatures
higher than 33⁰C appeared “bumpy” with an increase of 10% of their diameter. Immature
particles remained larger than both smooth and bumpy particles. Bumpy particles are
heterogeneous and presented some exposed lipid bilayer. Heterogeneity among bumpy
particles is caused by “molecular breathing”, a phenomenon caused by the lack of contact in
between the E protein dimers. Bumpy viral particles tend to aggregate at temperatures
exceeding 35⁰C. As a result, plaque assays performed at 28⁰C showed a higher infectious rate
for bumpy viral particles than smooth ones73.

In another study, DEN2 NGC and DEN2 WRAIR S16803 were produced in C6/36 mosquito cells
at 28⁰C74. The purified viruses were incubated at 37⁰C for 30 min and at 4⁰C for 2h. The control
was maintained at 4⁰C after purification. Cryo-EM study of DENV-2 virus showed 80% of the
particles were smooth in the control sample, but they appeared bumpy when incubated at
37⁰C for both strains. Heterogeneous particles from the 37⁰C incubation were classified into
four different classes. Class I represents the relatively smooth mature particle described in
the literature. Class II and III have respectively a larger diameter and a rougher surface than
Class I. The fourth class of particle is defined by the loss of E protein icosahedral structure on
the outer layer and therefore has a smaller diameter. This study complements the structural
“breathing” phenomenon described earlier73.

Zika virus is phylogenetically closer to DENV-4, which is the dengue serotype with the highest
level of thermostability. The rise in temperature disrupts profoundly the E protein
organization at the 3-fold vertex and 5-fold vertex axis of the heterodimers composing the
icosahedral structure at the viral surface. Electron microscopy of zika virus grown in mosquito
cells and further incubated at 37°C and 40°C revealed a high level of E protein structural
 organization at the surface of the virion with a smooth surface, which is contradictory with
DEN-2 findings75. The CD-loops at the fusion peptide (amino acid 98-111) in Zika virus E protein
are spatially closer toward the 5-fold vertex than DEN-2 which might result in better stability.
In a model focusing on the effect of temperature on the vertex of E protein, it appeared Zika
virus has more inter-dimer stability76. When the CD loop is modified with single point mutation
or shortening, it reduces the viral thermostability77.

Dengue and Zika viruses have a very similar viral structure, yet, Zika virus is a lot more
thermostable than dengue virus. The lability of the dengue E protein organization at the viral
surface correlates with epitope variability and the presentation of non-neutralizing epitopes.

V. HUMORAL RESPONSE, VIRUS NEUTRALIZATION AND QUATERNARY EPITOPES

It is globally admitted that primary DENV infection confers lifelong serotype-specific immunity.
As a result of poor neutralization by heterotypic antibodies, a secondary infection with a
different serotype can generate Antibody Dependent Enhancement (ADE) of disease leading
to severe hemorrhagic signs and shock syndrome. Affinity and potency of the elicited
antibodies impact the neutralization of the virus. The different types of antibodies elicited by
DENV infection have been characterized as: group-reactive, which can recognize multiple
flaviviruses; cross-reactive, which can recognize several serotypes of one virus and, finally;
type-Specific, which can recognize only one serotype78–80. Type-specific antibodies are
generally known to strongly neutralize one single serotype. On the other hand, cross-reactive
antibodies are weakly neutralizing. There is a delicate balance between the potency of the
antibody and its type-specificity.

The antigenically dominant protein is the E protein. Type-specific antibodies mostly target
the domain III (DIII) of E protein which is involved in receptor-attachment and presents the
most antigenic variation between serotypes. Antibodies directed against this domain would
prevent DENV from entering host cells. However, a recent study addressing antibody
composition of human sera previously infected with DENV-1, DENV-2, and DENV-3 provided
evidence that strongly neutralizing antibodies do not only recognize DIII but also a cryptic
epitope covering the hinge of DI-DII and the space between the two adjacent dimers. Anti-DIII
antibodies represent a very small portion of the total antibody amount and that DENV-
immune sera can still exhibit neutralizing activity even after mutation of EDIII 81.

These strongly neutralizing antibodies were tested for the recognition of a recombinant E
protein, which includes 80% of the E protein and lacks the two EH-1 and EH-2 helices as well
as the two transmembrane domains. As a result, recombinant E protein was recognized by
monoclonal antibodies targeting conformational epitope, such as the mAb 4G2, but not by
these strongly neutralizing antibodies. A comparison of the footprint of these neutralizing
antibodies for each of the three tested serotypes in this study shows strong evidence that the
E protein can only fold properly when expressed with both ectodomain and membrane-
associated domains82. Indeed, antibody pressure assay of human isolated neutralizing
monoclonal antibody and analysis of escape mutants showed mutations at the Hinge of DI-DII
for two different DENV-1 and DENV-3 strains82. Therefore, the sole integrity of the DIII is not
enough for viral neutralization. The proper folding of the E protein is primordial for the
recognition by strongly neutralizing antibodies.

Recent studies highlight the connection between temperature, structure, and neutralizing
epitopes display. The quaternary structure of the viral particle is paramount for vaccine
purposes. Previous crystal structure analysis of mature DENV compared to mature DENV
complexed with neutralizing antibodies showed that neutralizing epitopes are qualified as
“cryptic” as they are not present on the surface of the mature virion.

Following incubation at 37⁰C, neutralizing antibodies binding with the virus is enhanced. This
observation correlates with the EM study, which classifies the arrangement of the E protein
on the surface of the virion83. Meanwhile, temperature impacts the recognition of the virus
by neutralizing antibodies; it does not affect its infectivity. Viruses incubated at 4⁰C or 37⁰C
showed equivalent infectivity.

In another study, an anti-prM antibody has been characterized and shows recognition of a
cryptic E epitope at the EI-EII interface. This same antibody was able to restore the infectivity
of immature viral particles and bind to a recombinant soluble E protein 84. PrM antibodies are
known to be cross-reactive. This study emphasizes type-specific anti-E antibodies are not the
only players in virus neutralization.

The crystal structure analysis of four isolated human broadly neutralizing antibodies (bnAb)
revealed binding to quaternary epitopes when complexed with E protein produced in insect
cells (28°C). The epitope consists of the fusion loop made by two strands of the DII, the 150-
external loop of the adjacent DI of the second E protein of the dimer, and the two glycosylation
sites N-67 and N-153. Fusion loop and glycosylation sites are common to all four serotypes,
which relates to the broad characteristic of these four monoclonal antibodies 85.

Another strongly neutralizing serotype human monoclonal antibody (HM14c10) isolated from
an infected patient shows reactivity with a structural epitope. This epitope is discontinuous
and spans the adjacent surface of the envelope protein dimer at DI and DIII86. This structural
DENV epitope is similar to the West Nile virus (WNV) epitope recognized by the human
isolated monoclonal antibody (CR4354). This latter is also strongly neutralizing against WNV.

Evidence shows DENV neutralization does not only lie in immunity against the tertiary
structure but first and foremost within the induction of antibodies recognizing structural
quaternary epitopes. These quaternary epitopes rely on the proper folding and formation of
the structural proteins. A special focus should be attributed to the preservation of these
epitopes in vaccine development.
 Chapter 3. Dengue and Zika vaccine development and
Candidates

This chapter covers the different vaccine strategies engaged in clinical trials from the oldest
strategies, such as inactivated and live-attenuated vaccines, to the most recent ones, including
DNA and RNA-based vaccines. This chapter also describes with more details pre-clinical
strategies relying on replicons, subunits, and viral-like particle vaccines.

I. DENGUE VACCINES IN CLINICAL TRIALS

A dengue vaccine has to be safe and effective against the four serotypes in a balanced and
sustained immunity. The different strains alternate in the endemic cycles increasing the
probability of ADE. From the first attempt at a vaccine in 1945, it took nearly 70 years for the
first dengue vaccine, chimeric yellow fever 17D—tetravalent dengue vaccine (CYD-TDV),
ChimeriVax by Sanofi Pasteur, Inc. (Swiftwater, PA, USA), to reach commercialization.

Live attenuated virus vaccines

Live-attenuated vaccines are composed of replicating live viruses that have been attenuated
in order to prevent causing severe diseases in humans after inoculation. One main difficulty
with this type of vaccine is in a multivalent strategy. Some viral strains might replicate faster,
which can lead to an imbalanced immune response.

a) Attenuation through sequential passaging

The first attempt at a dengue vaccine was in 1952 by Sabin by passaging the virus multiple
times on mouse brains. However, this technique, previously used successfully to attenuate
Yellow fever virus, did not provide the desired outcome for DENV. Attenuation of DENV-2
through passages in PDK cells was conducted by the CDC, in cooperation with the company
Inviragen, Inc (Fort Collins, CO, USA)87. The resulting DEN-2 attenuated strain, called DEN-2
PDK-53, was used as backbone for a tetravalent vaccine where the original prM and E
sequences of DENV-2 were replaced by the sequences of the other serotypes88. This vaccine,
called DENVax, was eventually licensed to Takeda Pharmaceutical Company Ltd (Tokyo,
Japan). The clinical trials showed neutralizing antibody titers higher for DENV-2 compared to
the other serotypes89,90. Phase III clinical trial is still ongoing to show efficacy against the
severity of the disease. In order to round up the clinical investigation, this vaccine was also
tested with co-administration of the YFV vaccine to assess immunogenicity and safety91.

b) Genetically engineered live-attenuated virus

Genetically engineered live virus attenuation is the second most used technique for dengue
vaccine development. The NIH has licensed to four biotech companies (Biological E Ltd -
Hyderabad, India; Butantan Institute - São Paulo, brazil; Panacea Biotech Ltd - New Delhi,
India; and Vabiotech Ltd – Paris, France), a live attenuated virus with a deletion of the 3’UTR
sequence of DENV-4. The same deletion showed comparable results for DENV-1, but not for
DENV-2 and DENV-3. In order to create a tetravalent vaccine, the mutated DENV-4 strain was
used as a backbone for the other serotypes by replacing the prM and E coding sequences. The
vaccine candidate was referred to as “TV003”. Other modifications were added to the
constructs to obtain a robust and balanced immune response and showed a positive outcome
in clinical trial Phase II92–95. After further improvements, the vaccine “TV005” was
developped96. These two candidates are currently in clinical trial Phase III.

Finally, the most advanced live-attenuated dengue vaccine, Dengvaxia, is commercialized by

Sanofi Pasteur, Inc. (Swiftwater, PA, USA). This vaccine is based on the utilization of the
attenuated Yellow-Fever 17D backbone. Dengvaxia is the only one of the three live-attenuated
dengue vaccine candidates that do not comprise dengue non-structural proteins. This
chimeric attenuated virus carries the sequence and displays the structural proteins prM and E
of each of the serotypes of DENV97. The vaccine immunization schedule comprises three
different injections. During Phase II in Asia and America, the vaccine showed low and
unbalanced immunogenicity among the four strains. The efficacy was unbalanced in Phase III
clinical trial depending on the serotype, the age, and the serostatus of the patient before
immunization98,99. In Asia, the vaccine appeared to enhance disease severity in naïve patients
under nine years old with a higher incidence of hospitalization three years after
immynization100,101. This observation has changed the WHO recommendation for the vaccine
and has restricted immunization for patients from age 9 to 45 years of age. The vaccine was
licensed in Mexico, the Philippines, and Brazil in 2015, followed by El Salvador, Costa Rica,
Paraguay, Guatemala, Peru, Indonesia, Thailand, and Singapore in 2016. The adverse effects
of the vaccines have pushed the Philippines to revoke the authorization and ban the sale of
the vaccine in its country. To overcome the poor relative efficacy, the company is currently
testing the use of a fourth immunization in Singapour (Clinical trial NCT02824198) and in Latin
America (Clinical trial NCT02623725).

Inactivated dengue virus vaccines

Inactivated virus vaccine technology consists of the use of chemically inactivated virus as the
immunogen. The formalin is generally used for viral inactivation. The Walter Reed Army
Institute of Research (WRAIR, Silver Spring, MD, USA), in partnership with GlaxoSmithKline plc.
(GSK, (Brentford, UK)), developed a dengue Purified Inactivated Virus vaccine (PIV) candidate.
After a promising monovalent Phase I clinical trial102, tetravalent formulations were tested in
Phase I clinical trial103. Two different adjuvants were well tolerated. The tetravalent
formulations were able to elicit a robust response. Meanwhile, the antibody titers peaked at
day 56 post-immunization, the protection waned following the 6 months after boost
 inoculation to finally stabilize. These limited but positive results support further clinical
evaluation of the vaccine.

DNA and RNA vaccines

DNA vaccines are composed of an expression vector carrying a single or multiple genos of
interest. Once injected, these plasmids express the protein immunogen to generate an
immune reaction. DNA vaccines are generally less immunogenic than other platforms. The
first DNA candidate tested in clinical trial was a proof-of-concept study with a plasmid
expressing DENV-1 prM and E genes104 conducted by the WRAIR. In this small study of 22
healthy adults, only five patients who were attributed to a higher dose stimulated detectable
neutralizing antibodies. This study was followed up by Phase I clinical study, including a
tetravalent formulation. The formulation was enhanced with the Vaxfectin® (Vical, CA, USA)
adjuvant. Only 17.6% of the subjects in the adjuvanted high dose showed elicitation of
neutralization antibodies against 1, 3, or 4 serotypes105. The humoral response was limited.
This clinical investigation was not pushed further.

Recombinant subunit protein vaccines

Recombinant subunit protein (RSP) vaccines are also called virus-like particles (VLP).
Expression of recombinant proteins has been tested in different hosts such as baculovirus
expression vector, mammalian cells (MDCK cells, VERO cells), bacteria (Ecoli), and yeast (Pichia
Pasturis). To this date, the most advanced vaccine using RSP is developed by Merck
(Whitehouse Station, NJ, USA), which recently completed a Phase I clinical trial. The vaccine is
composed of a tetravalent formulation of truncated E proteins (80% of the amino-terminal
sequence) produced in insect cells (Drosophilia S2 cell line). The purified E protein of each
serotype is adjuvanted with the patented ISCOMATRIX™ formulation or with the Alhydrogel™.
For each formulation, there was a double amount of DENV-4 material than the other
serotypes. Despite this higher protein content, this serotype showed a lower and less durable
immune response. As a result, the vaccine was generally well tolerated but showed few
adverse events. There was no immune response in the non-adjuvanted formulation or in the
placebo. The highest dose of adjuvant resulted in the highest elicitation of neutralizing
antibodies, which was still moderate106.

II. ZIKA VACCINE IN CLINICAL TRIALS

Zika virus is an emergent virus and has been considered a major public health issue since the
major outbreak in 2015 in Brazil. The virus has only one serotype and therefore, would require
only a monovalent vaccine. However, the clinical presentation is complex, particularly due to
the risk of infection of the fetus during pregnancy via the placenta. Consequently, a live-
attenuated vaccine would not be recommended for women in childbearing age. The WHO
issued a Target Product Profile addressing all these intricate characteristics of the virus107. The
manufacturer of the only dengue virus vaccine, Sanofi, published some preclinical data using
the YFV17D virus as backbone carrying the prM and E protein from Zika108 but has not entered
clinical trial yet. There are few candidates currently in clinical trials. The vaccine technologies
are mainly focused on nucleic acid formulations (DNA and RNA vaccine) as well as inactivated
viruses. Below are presented the most advanced vaccine candidate for Zika.

Live attenuated virus vaccines

National Institute of Allergy and Infectious Diseases (NIAID, MD, USA) developed the first live-
attenuated virus vaccine candidate brought to the clinic. This candidate uses the DENV-4 live-
attenuated virus backbone with the 30 nucleotides deletion in the 3’ UTR. The prM and E
sequences are replaced with Zika’s. This candidate is currently in Phase I clinical trial and is
tested in flavivirus naïve patients (NCT03611946). The second live-attenuated virus vaccine
candidate was developed by the private company Themis Bio (wholly owned subsidiary of.
Merck & Co., Inc., Kenilworth, NJ, USA) and completed a Phase I clinical trial (NCT02996890).
This chimeric virus used the live-attenuated Measles virus vector carrying the prM and
truncated E of Zika virus109. The results have not been reported yet.

Inactivated Zika virus vaccines

Formalin inactivated virus candidates are currently being investigated in phase I clinical trial
led by NIAID and Beth Israel Deaconess Medical Center (BIDMC, Boston, MA, USA), University
of St Louis, and WRAIR Several formulations were tested in naïve patients. No serious adverse
events were registered (NCT02963909, NCT02952833, NCT02937233)110. Other inactivated
virus candidates are being tested in phase I clinical trial. These studies are sponsored by
private companies, Takeda (NCT03343626), and Valneva S.E. (Saint-Herblain, France)
(NCT03425149). The results have not been published yet.

Nucleic acid-based vaccines

a) DNA vaccines

NIAID developed the most advanced DNA vaccine. The formulation is composed of an
expression plasmid vector carrying the prM and E sequences of ZIKV with or without the prM
signal sequence and the last 98 amino acid of the carboxyterminal end of the E protein (helical
and transmembrane domains) of Japanese Encephalitis virus (JEV). In phase I, 100% of the
participants elicited neutralizing antibodies against Zika virus (NCT02840487 and
NCT02996461)1,2. Phase II clinical data have not been reported yet.

The second DNA candidate was developed by GeneOne Life Science, Inc. (Seoul, South Korea)
/ Inovio Pharmaceuticals (Plymouth Meeting, Pennsylvania) in phase I clinical trial
(NCT02809443 and NCT02887482). This formulation comprises an expression vector carrying
the prM and E sequences of ZIKV. The preliminary data reported that the vaccine elicited
neutralization antibodies in 62% of the participants111.

b) mRNA vaccine

Moderna Inc. (Cambridge, MA, USA) developed an mRNA vaccine and tested it in clinical phase
I (NCT03014089 and NCT04064905). The results of seropositive patients were published on
the company’s website with fairly successful results; however, the neutralization titers of
seronegative patients were omitted.
 Chapter 4. Viral-Like particles

Viral-like particles (VLPs) are self-assembling organized repetitive structures identical to wild-
type virus (enveloped or non-enveloped) without carrying any viral genome. These particles
are non-infectious, non-replicating, and mimic the native conformation of essential epitopes.
These particles are the results of the expression of structural genes, with or without non-
structural genes, in an expression system. The expression system can be prokaryotic to
eukaryotic, ranging from plant cells, E. coli, yeast, insect cells (mainly using baculovirus), yeast,
and mammalian cells. Each system presents its strength and weaknesses from yield,
scalability, complexity, post-translation modification, speed of production, and regulatory
limitation. This chapter describes the different expression platforms for the production of
VLPs.

I. INTRODUCTION

Viral-Like Particles are naturally formed during viral infection, they were described for the first
time in 1978 in polyoma virus112. In the following year, the first recombinant VLP were
produced using the Hepatitis B Virus (HBV)core antigen113 and HBV surface antigen114. The
first recorded electron microscopy study showed that HBV-core antigen VLPs were identical
to virions115. Viral-like particles can be generated from proteins of both enveloped and non-
enveloped viruses116. In addition to flexibility, well defined VLP platform can be used to
generate chimera and use the VLP platform as backbone where an exogenous viral proteins
sequence is added into the coding sequence of the VLP backbone resulting in a chimeric VLP.
the exogenous protein or peptide is localized at the outer surface of the VLP. For example, the
HBV surface antigen VLP mentioned above was used as backbone to generate a chimera with
dengue E protein117. A Parvovirus capsid VLP has also been used as backbone to generate a
chimera with two segment of the domain III of the dengue E protein118.

VLP can be produced in different cellular system of expression. Each system has its own
strength and weaknesses. The system is selected depending on the protein of interest and as
well as other parameters described in Table 1.

II. CURRENT USE

Viral-like particle-based vaccines are relatively recent comparing to other vaccine platforms.
There are currently three VLP vaccines approved by the U.S. Food and Drug Administration
(FDA) and one vaccine approved by the State FDA (SFDA) of China. The first VLP-based vaccine
to receive commercial license form the FDA was the HBV vaccine RecombivaxHB from Merck
& Co. in 1986. This vaccine showed safety and efficacy in infants and it contributed to the
decrease in HBV prevalence worldwide120. Later on, the U.S. FDA approved two other VLP-
 based vaccines. These vaccines, commercialized by Merck&Co from 2006 and GSK from 2009,
are targeting oncogenic-prone serotypes of Human Papilloma Virus (HPV). These vaccines are
respectively, Gardasil (originally bivalent vaccine against HPV16 and 18, which was later on
formulated as quadrivalent with serotypes 6 and 11)121,122 and Cervarix (quadrivalent vaccine
against serotypes 16, 18, 6 and 11)123. Recently, Merck and Co, Inc. received authorization for
a nonavalent vaccine, Gardasil targeting five additional serotypes: the serotypes 31, 33, 45,
52, and 58. The main difference between these two vaccines lies in the expression system.
Gardasil is produced in yeast, whereas Cervarix is produced in insect cells through baculovirus.
In 2011, Hecolin VLP-based vaccine against Hepatitis E virus was approved by the SFDA of
China. This vaccine showed 100% efficacy following the three-dose schedule regimen in phase
III clinical trial124. These four VLP-based vaccines have proven both safety and efficacy of this
platform and strategy.

Table 1 : Advantages of VLP production platforms. + low advantages,++ medium advantages, +++ high
advantages (adapted from Jeong &Seong 2017119)

Expression platform for VLP production

Cost VLP
Speed Scalability Yield
Effectiveness Complexity

Bacteria +++ +++ +++ +++ +

Yeast +++ +++ +++ ++ ++

Baculovirus
++ + ++ ++ +++
insect cells

Mammalian cells + + ++ + ++

Plants ++ ++ ++ + ++

There are other VLP-based vaccines currently being tested in clinical trials. These VLP
strategies are developed to address other viruses such as influenza (NCT02233816) and
norovirus (NCT02153112).

As I joined TechnoVax in 2011, there was still no VLP-based vaccine tested in clinical trials for
dengue or Zika viruses. Therefore, I was given the responsibility to develop a dengue VLP
vaccine. The following section covers the different approaches published before 2015 and
reflects the literature available during the development of the dengue VLP project.

III. PRECLINICAL DENGUE VLP, RSP AND REPLICONS

Different vaccine strategies relying on VLPs have been developed previously with DENV. These
strategies mostly focused on the use of prM and E proteins for the formation and secretion of
VLP.

In 2002, Konishi and Fuji established a stable CHO-K1 cell line expressing “subviral extracellular
particles” pr/M-E of DEN-2 NGC strain125. In this study, the furin n binding site between pr and
M is mutated in order to suppress the fusion activity of E protein. Polykaryocyte formation has
been previously observed in a cell line stably expressing Japanese Encephalitis Virus (JEV)
subviral extracellular particles and it seemed to be a critical impediment for particle
formation126. Sedimentation profiles of these secreted extracellular particles show lower
buoyancy than virion. After treatment with Triton X100, E protein is found to be membrane-
associated. Immunization of BALB/c mice with these extracellular particles showed no
elicitation of neutralizing antibodies following the two immunizations with these extracellular
particles. However, high neutralization titers were recorded after challenge following previous
immunization with 400ng of EPs. The same titers are observed for virus control. In this article,
the authors are referring their vaccine as “Extracellular particles” however, they did not show
electron microscopy images to ensure the formation of VLPs.

Another strategy was developed using dengue prM and E proteins in combination with
sequences of other flaviviruses producing a chimera. In this study, Zhang et al. showed
differences between serotypes and used partial sequences of JEV to generate chimeras. The
sequences of the carboxy-terminal end of the JEV E protein was used to replace dengue’s and
avoid the signal retention. Authors did not observe the secretion of VLP when using 100%
dengue sequence in 293 T cells. When replacing specific sequences with JEV, the authors
noted the secretion of VLPs. DENV-1 required replacement of the secretion signal with JEV for
efficient secretion of VLPs. However, DENV-2, DENV-3, and DENV-4 required additional
replacement of the 20% of the carboxy-terminal end of the E protein for proper secretion.
Electronic microscopy showed similar profiles between VLPs and virions with particles ranging
from 45 to 55nm. Immunization of BALB/c mice was performed intraperitoneally with 100μg
of monovalent vaccine or 25μg of each serotype for the tetravalent vaccine. Complete
Freund’s adjuvant was used for priming and Freund’s incomplete adjuvant for administration
of two boost immunization with an interval of 2 weeks. The immune response was assessed
using a recombinant DIII of E protein of all four serotypes in ELISA. Higher titers of rEIII IgG
were observed against tetravalent vaccines than tetravalent virions. Neutralization assay
using BHK-21 cells showed similar serum neutralization levels for each serotype in both
monovalent and tetravalent formulation. These titers were also similar between the VLPs and
their viral counterparts127.

Poor secretion issues of dengue VLPs were addressed in another study by using codon
optimization of the gene sequences. The prM-E coding sequences were placed downstream
of the Vesicular Stomatitis Virus-G envelope glycoprotein in the plasmid DNA vector. This
expression vector was then co-transfected in HeLa cells with a retroviral vector. Seventy
percent of the codon was replaced through a codon optimization process. This replacement
led to the 3-fold increase expression of E protein. The authors observed the RSPs went through
the ER secretory pathway as well as Golgi apparatus, simulating the natural maturation and
secretion process128. This study did not assess in vivo the RSP nor presented EM studies. There
was no evidence either of prM/E heterodimer.

A different approach to generate dengue VLPs was based on the expression of CprME in Pichia
pastoris system. The CprME or prME coding sequences were transfected into a yeast system.
Proteins of interest were detected in the cell lysate providing evidence of proper expression.
Following sucrose gradient purification, the fractions containing the E protein were used to
immunize two rabbits and assess their immunogenicity. Although the sample size was too
small to statistically compare the groups and the control, the neutralization assay showed
elicitation of neutralizing antibodies in serum 129.

These different strategies show the difficulties of tackling the production of a dengue VLP-
based vaccine. Yeast insect cell and mammalian cell systems were all tested highlighting the
specificities of the structural proteins of DENV in comparison to other flaviviruses. In order to
overcome the retention signal of E protein, the replacement, as well as the truncation of the
carboxy-terminal end, have been experimented. However, the c-terminal end of the envelope
protein has been identified as key element for the proper folding and maturation of virion. In
a similar fashion, the secretion signal of prM-E can be improved by replacing the original
dengue sequence with other secretion signals. Very little research has been done regarding
the role of the transmembrane domain of C protein, anch-C, in the maturation of the virion.
One study showed evidence the capsid protein is involved in the proper folding and secretion
of prME VLPs. The yield of released particles was higher when prM-E expression plasmid was
co-transfected with an expression vector carrying the Capsid sequence130.

Based on this first analysis of the scientific knowledge available on VLP expression platforms
at the time of the project launch, it appeared to me focusing on structural proteins and their
interactions would help in designing a novel dengue VLP-based vaccine strategy.
 Part II. Methods - Strategy for the
development of VLP

The strategy for the development of dengue and Zika VLPs relies on the mimicking of the
natural formation of the viral shell with inclusion of all the three structural proteins. In order
to maintain the natural processing of the C-prM-E, the NS2b/NS3 viral protease complex was
added to the list of proteins of interest. Firstly, the following section details the five structural
and non-structural proteins and their characteristics as well as the key features which were
selected for the formation of dengue VLP. Secondly the translation of a dengue VLP-based
vaccine strategy to a ZIKV VLP-based strategy is described. Lastly, the selection of the
expression system and the testing of the VLPs as relevant vaccine candidate is described. The
two peer-reviewed articles present the results and findings followed by a discussion for each
of these articles.

I. DENGUE PROTEINS OF INTEREST

The Envelope protein

The immunodominant structural protein is the Envelope (E) protein. This protein features
three ectodomains. From the Amino-terminal end to the carboxyl-terminal end, there is the
domain I (DI), domain II (DII), and domain III (DIII) with respective functions: molecular hinge,
dimerization domain with internal fusion peptide and Ig-like fold receptor-binding site78,131.
Domains I and II alternate with each other in the sequence of the polyprotein as domain II
folds and forms two loops. Domain I range from amino acids 1 to 52, 132 to 193, and 280 to
296. Domain II covers from amino acid 52 to 132, 193 to 280. Domain III ranges from 296 to
394132. (Figure 7). There are two glycosylation sites on the ectodomain: N-67 in DII and N-153
in DI. The ectodomain is followed by two α-helices, E-H1 and E-H2, and two transmembrane
domains, E-T1 and E-T2. The difference between the four DENV serotypes mainly lies in the
amino acid composition of the E protein DIII. The E-H1 and E-H2 are amphipathic helices. E-
H1contains a retention signal which does not exist in other flaviviruses. Helical domains of
the E protein do not have a direct role in the immune response generated because they are
hidden from the innate immune system.

The Precursor Membrane protein

The Precursor Membrane protein is composed of a “precursor” portion called “pr” in the
amino-terminal end. It is composed of 7 antiparallel β-strands, which are stabilized by three
disulfide bonds, C34-C68; C45-C80; C53-C66; and a glycosylation site at Asn69133. The
membrane or “M” follows the pr portion of the precursor membrane. It is 75 amino-acid long
and has two transmembrane domains T1 and T2, respectively, at amino acid M-40 and M-
70134.

As mentioned earlier, the host Furin is responsible for the cleavage between pr and M. The
Furin recognizes the consensus sequence: Arg-Xaa-(Lys/Arg)-Arg. The protease catalytic site is
composed of three pockets. These pockets can fit P1, P2, and P4. In all four DENV serotypes,
the P3 is an acidic residue at the opposite of all other flaviviruses. The adjacent residues of
these pockets have acidic side chain which favors basic residues at P3135. The acidic property
of the Glutamic acid residue of dengue virus in P3 might interfere with the proper cleavage of
pr/M protein and, as a result, produces immature particles136.

The Pre-membrane protein and envelope protein dynamic duo

The interaction between the two main structural proteins, prM and E, is primordial for the
expression, maturation, and formation of the viral particle. These two proteins not only
interact in their ectodomain – as mentioned earlier, the pr portion covers the E fusion peptide;
the two helical and transmembrane domains also interact with each other.

In 2012, Tsai et al., studied the role of EH2. Authors studied truncations of the E protein C-
terminal end, and their consequences in the formation, expression and structure of both prM
and E. They also investigated their interaction with each other. They showed evidence that
prM is poorly generated and unstable when it is expressed alone. Pre-membrane protein was
found to be more stable when co- transfected with EH2. On the one hand, prM expression
increases when co-transfected with an increasing amount of E. On the other hand, prM
expression decreases when co-transfected with truncations of E protein, providing the
evidence of the critical role of EH2 as well as ET1 and ET2 in the expression and stability of
prM137.

The heterodimerization of prM-E is a key element of VLP assembly138. As Zhang (2003)

described in cryo-electron microscopy studies, the two transmembrane domains of E and prM
are next to together; more precisely, E-T1 is next to MT1. The two transmembrane domains
of E protein are important for the localization of the ectodomain as they ensure proximity with
prM proteinFigure 10: illustration of the organization of E protein and localization with the membrane protein
in the lipid bilayer (green) in (a) transversal view and (b) superimposed top view. The domain I is represented in
pink, the domain II in yellow, and the domain III in blue. The M protein is represented in orange. Extracted from
10)134. Additionally, the stem region of prM aligns with the middle of
Zhang et. Al.; 2003).Figure
the DII of E protein and form an electrostatic patch involved in the heterodimer formation in
the ectodomain. Amino acids of both E and prM interact directly with each other. Indeed, the
H98 present in the middle of the M protein faces the helices αA and αB in the E protein. In the
JEV model, when the same histidine is mutated with H99A, pr/M, and E lose their
heterodimerization139.

Crystallography study revealed amino acids 1 to 81 at the N-ter of prM protein interact with
the ectodomain of the E protein and therefore influences heterodimerization. ET1 and EH2
appear to be more relevant than E ectodomain for the heterodimerization 133.

Figure 10: illustration of the organization of E protein and localization with the membrane protein in the lipid
bilayer (green) in (a) transversal view and (b) superimposed top view. The domain I is represented in pink, the
domain II in yellow, and the domain III in blue. The M protein is represented in orange. Extracted from Zhang
et. Al.; 2003).

The H244 amino acid of the E protein (DII) faces the glycosylation site in N-69 of pr protein 133.
When the pH changes from acid to neutral, they lose their affinity for each other and become
separated. The contact surface between pr and E represents respectively 16% and 4% of the
E protein133. The interaction between pr, M, and E proteins is a crucial element of the proper
folding, maturation, and, therefore, an essential role in the immune response when used as
an antigen. Group reactive antibodies, such as 4G2, cannot bind E protein when the latter is
truncated at E-H1 137. This group reactive antibody recognizes the domain II fusion peptide in
the E protein ectodomain. EH2 is essential for maintaining the proper conformation and
folding of E protein.

Capsid, VLP secretion, and harmony among the structural triad

As described by Jones et al. (2003), the capsid protein is the least conserved protein among
Flavivirus genus. There is a 40% sequence identity between most of the genus members and
only 22% identity between the mosquito-borne flaviviruses140. The capsid protein is
responsible for the encapsidation of the viral RNA in the viral envelope. The capsid protein
remains in the cytosol during expression. The carboxy-terminal end is embedded in the lipid
bilayer of the ER and is followed by the pr protein in the lumen of the ER. This transmembrane
domain is composed of four hydrophobic α-helices at its carboxy-terminal end embedded in
the membrane of the ER. This “anchor” allows the translocation of the rest polyprotein
through the membrane of the endoplasmic reticulum.

Identification and characterization of NS3 protease activity

The viral protease NS3 and its cofactor NS2b are also involved in the processing of the
polyprotein. Together they form a complex which recognizes and cleaves short sequence
motives at several protein junctions in the dengue virus polyprotein. These motives are
present at C-anchC, NS2a/NS2b, NS2b/NS3, NS3/NS4a, and NS4b/NS5 junctions. These
motives at major cleavages sites are followed by a pair of basic amino acids RR, KR, RK, or
exceptionally QR at NS2b/NS3 and preceded either by G, S, or A. The NS3 non-structural
protein has a trypsin-like serine protease within the first 180 amino acids with its catalytic
triad at D75, H51, S135, and a substrate-binding domain141. NS3 has a helicase at its Carboxy-
terminal end.

Falgout et al. showed evidence of viral serine protease activity. The authors also identified the
requirement of NS2b as cofactor141. At the minimum, NS3 requires residues 1-167 for its
protease activity. However, its optimal activity is observed with residues 1 to 183 142. The
serine protease recognizes a pair of basic amino acids in the P1 and P2 positions, which is
followed by serine or glycine residue (or Ala/Thr) in P1. Mutagenesis studies showed the
importance of the full length of NS2b in the cis-activity proteolytic activity of NS3143.

The NS3 protease is found in the soluble fraction of the cell extract when it is expressed by
itself, whereas it is present in the membrane fractions when co-expressed with NS2b
suggesting NS2b might have a role in presenting NS3 to the membrane. The hydrophilic cored
of NS2b increases NS3 activity to up to 3,3000-fold 144. More specifically, the 89EEE amino
acids of NS2b are primordial for efficient cleavage of NS2b/NS3 145. Interestingly, the flanking
transmembrane domains in NS2b are highly conserved among the flaviviruses, whereas the
core of the cofactor shows the most divergent sequence146.

The interplay between the anch-C cleavage by the viral protease and the anch-C-pr cleavage
by the host cell signalase varies among different flaviviruses. In 1993, Lobigs showed
coordination between two cleavages on each side of the ER membrane at the C-prM site of
the Murray Valley encephalitis virus (MVEV). The hydrophobic domain of the C protein (anch-
C) helps the translocation of prM through the membrane of the ER. Efficient signal peptidase
cleavage in the ER occurs only after cleavage of the C-anchC by the viral protease NS3 in the
cytosol147.
Several mutations were tested around the NS2b/NS3 cleavage site in YFV. The amino-acids
present at the cleavage site of C-AnchC have predominantly basic side chains. In contrast, the
amino acids at the cleavage sites in the non-structural regions usually have less basic side
chains. On the opposite side of the lipid bilayer of the ER membrane, when C-pr cleavage site
of the host cell signalase is uncoupled from the viral protease cleavage, the virus loses its
ability to form plaque148. The same phenomenon is observed in the context of MVEV. The P1
and P3 amino acid at the c-ter of the capsid protein are the suitable motif for the recognition
of the signalase. However, flaviviruses lack polar residues in P2, P4, and P5, and help with the
efficient recognition by the host signalase.

The mutations PQAQA at the P2, P4, and P5 of the signalase cleavage site drastically increase
the signalase cleavage of C-prM without the requirement of prior cleavage of the capsid. This
cleavage efficiency enhancement of C-prM correlates with the reduction in the growth
between wild-type and mutant. The same viral growth reduction is observed in Vero and
C6/36 cells and lower by a hundred-fold149. The membrane-anchored form of C is a poor
substrate for cleavage by the viral protease unless covalently attached to an ER luminal
polypeptide such as pr. The same dynamic is documented for YFV148. In conclusion, the PQAQA
mutation at the signalase cleavage site at Anch-C-pr is lethal for viral replication in YFV but
increases the budding of MVEV VLPs.

The two studied viruses, YFV and MVEV, show different consequences in the mutation or
uncoupling of the signalase cleavage site from the viral protease complex at the C-anchC
junction.

Glycosylation sites

Glycosylation is essential for viruses. It has multiple functions such as receptor binding,
folding, trafficking, or interaction with the immune system. As parasitic entities, viruses do not
possess their own glycosylation machinery but instead uses the host’s. During maturation,
DENV acquires its N-glycosylation in the ER-Golgi complex.

The envelope protein has two glycosylation sites, N-67 and N-143. The former is unique of
dengue virus, whereas N-143 is conserved among the flaviviruses150. Dengue protein
glycosylation is involved in host cell interaction as well as viral maturation.

Crystal structure analysis, as well as cryo-EM show the carbohydrate moiety of N-153 extends
across the DII dimer and the fusion peptide151152. Cryo-electron microscopy study shows the
interaction between lectin with the N-glycan at Asn-67 when DENV is complexed with
carbohydrate binding-domain of Dendritic Cell-specific ICAM3 grabbing non-integrin (DC-
SIGN)153 DC-SIGN is abundant in immature dendritic cells and is an alternative receptor of
DENV.
Several studies have shown the negative effect of inhibition of ER glycosidases enzyme.
Glycosidases are responsible for protein maturation and cleavage of N-glycosylation. Dengue
virus secretion is be impaired when glycosidases are inhibited. Furthermore, deficient
cleavage of the glucose residues alters protein folding by ER chaperones. As a result, proper
glycosylation of dengue structural protein is essential for virus maturation and formation.

Genes of interest optimization

The following section describes the key features, mutations and truncations that were
included in the design of the strategy to produce an immunologically relevant dengue viral-
like particle vaccine.

a) Suppression of the retention signal

As described previously the prM and E protein interacts for their dimerization. The EH2
domain plays an important role in this dimerization. This section covers the role EH1 as
retention signal. As recorded in the literature, the carboxy-terminal end of the E protein of
dengue and mainly the EH1 possess a sequence slowing down the secretion of viral particles.
Indeed, when prM and E genes are expressed using plasmid DNA vector in mammalian cells,
released VLPs demonstrate a very low yield. The replacement of 20% of the carboxy-terminal
end of the dengue envelope protein by the JEV equivalent showed efficient secretion of VLPs.
When only replacing the last 10% of the carboxy-terminal end, which corresponds to the two
transmembrane domains, the secretion was not enhanced suggesting the “retention” motif
responsible for poor VLP release is located in the two helical portion of the E protein 154. Using
the same context, the analysis of the composition of the helices EH1 and EH2, supported by
site-directed mutagenesis experiment, showed that a combination of 3 amino acid
substitutions in the EH1 domain could improve secretion by 3-fold. These amino acids are
located at the hydrophobic side of the amphipathic helix. When these amino acids are
replaced by the amino acids from the JEV EH1, i.e. I398L, M401A, M412L, it enhances the
hydrophobicity of the membrane-facing side of the helix 155. These three substitutions have
been found to be helpful for the secretion of DENV-2 and presumably DENV-1 but not DENV-
3 or DENV-4. The present goal was to maintain as much integrity as possible of the dengue
sequence therefore, the strategy was directed towards utilizing these key point mutations
targeting the troublesome amino acids inDENV-2 and establish the proof of concept.

b) Enhancement of the furin cleavage for improved maturation of VLP

The cleavage of the precursor pr from M by the furin is a key element in the maturation of the
protein. Viral particles with a cleaved pr are different in size and morphology than the mature
particles. The sequence at the cleavage site present an acidic residue at the position P3, which
 is unique to DENV compared to other flaviviruses. When this residue glutamic acid, is replaced
by alanine, which has an hydrophobic side chain, the number of released mature particles is
higher than the wild type virus70. This mutation is added to this strategy to improve the yield
of mature viral particles (Table 2).

c) Improvement of the NS3 catalytic site.

The last key point mutation introduced into the sequence concerns the optimization of the
NS3 catalytic site. The NS2b/NS3 complex is involved in the maturation by cleaving the C from
the anch-C in the cytoplasm. This cleavage then enhances the cleavage of anch-C and pr on
the other side of the ER membrane. As described in Part II, I. 5), the NS3 protease domain of
the NS3 protein can be separated from the helicase domain in c-terminal domain without
losing its catalytic activity. The NS2b protein act as cofactor of the protease. In this strategy,
the complex NS2b with truncated NS3 is selected for the proper processing of the CprME
fragment. In addition to providing the cofactor to the protease, the catalytic site is modified
by introduction of a single amino acid mutation. A catalytic study of NS3 cleavage showed
enhancement of cleavage with L115A mutation. Most amino acid present in the S1 and S2
subsites are conserved among flaviviruses. However, the amino acid L115 of the S1 subsite of
NS3 is specific for dengue virus. When the amino-acid is mutated from Leucine to Alanine, the
cleavage efficiency of NS3 is higher than wild-type 156. This mutation was introduced into the
final VLP construct.

II. DENGUE VLP STRATEGY SUMMARY

In summary, three-point mutations were included across the five proteins and the NS3 protein
was truncated in order to solely retain the protease portion. The table below details the point
mutations that were included in the sequence of the synthesize genes to generate the VLP
(Table 2).

Table 2 : Mutations introduced in the pr,M, E and NS3 gene sequences

Mutations
Enhanced Furin cleavage site P3 Q88 →A GAA→GCA
I398→L ATC→TTG
Enhanced hydrophobicity of EH1 amphipathic helix
M401→A ATG→GCC
 M412→L ATG→TTG
Enhanced hydrophobicity of NS3 to increase Km values L115→A CTT→GCC

Dengue virus has four different serotypes. This project being a proof of concept, the most
studied and prevalent serotype was selected. The serotype 2 has been one of the most studied
(Table 3). In several studies across the globe, the serotype 2 is one of the most prevalent
serotypes. It was responsible for 25% of DENV infections in a prospective study in Indonesia157
and 57.1% in Singapour158,

Table 3 : PubMed database hits of the different dengue virus serotypes and nomenclatures. April 2020.

Database Hit
Pubmed querry "DENVx" "DENV-x" DENx "DEN-x" "dengue x" Total
Dengue serotype 1 247 792 84 269 10015 11407
Dengue serotype 2 385 1040 118 429 9344 11316
Dengue serotype 3 103 499 41 160 7719 8522
Dengue serotype 4 104 405 88 150 6829 7576

III. TRANSITION TOWARDS ZIKA VLP

In 2015-2016, the South American outbreaks and the notification of ZIKV as a Public Health
Emergency of International Concern (PHEIC) raised the opportunity to assess my dengue VLP
strategy with another Flavivirus. As presented in Part I Chapter 2 I., ZIKV is phylogenetically
close to DENV-4 and is most likely thermostable comparing to the other more thermolabile
dengue serotypes.

The objective was to translate, as simply as possible, the VLP strategy designed for DENV to
ZIKV. The amino acid sequence of the structural triad revealed ZIKV did not have the VLP
limiting features of DENV. It has a hydrophobic residue at the P3 of the furin cleavage site, and
the EH1 helix is more amphipathic and comparable to YFV. Similarly, to the dengue VLP
strategy, the NS3 protein was truncated yet, the sequence remained unmodified.

The viral sequence selected to generate the Zika VLP was obtained from the ZIKV strain
identified during the 2013 outbreak in French Polynesia. At the beginning of the project, the
MR-766 viral strain was the only strain readily available to use as control. Fortunately, it was
eventually possible to acquire a more contemporary strain, the FSS 13025 strain isolated in
2010 in Cambodia.

As a result, the Zika VLP will be formed by the concurrent expression of unmodified CprME
and NS2b/truncated NS3 proteins.
IV. EXPRESSION SYSTEM AND TEMPERATURE

The five viral proteins C, prM, E, NS2b and NS3 were selected to produce secreted VLPs for
both dengue and Zika viruses. In order to optimize the probability of co-expression of the five
proteins in the same cell and to mimic closely the natural formation of the viral particles, the
structural proteins were kept in one single polyprotein translated via the expression of a single
OFR in an expression plasmid. Since the cleavage of C-anchC is required for the polyprotein
process, the NS2b/NS3 genes were expressed under a single OFR in a separate plasmid.

The viral proteins required for the formation of VLP, the expression cell line and conditions
were determined for optimal production of the dengue VLP on the one hand and Zika VLP on
the other hand.

The HEK293 cells appeared as a suitable expression system, since it allows scalability and yield
while emphasizing on safety. These suspension cell line grows in serum-free media limiting
the presence of downstream contaminants while obtaining the best post-translation
modifications. The genes in the expression vector were codon-optimized for this particular
cell line for both dengue and Zika VLP.

The HEK293 cell line grows optimally at 37°C, however, as described earlier, DENV has two
hosts with two production temperatures which has a strong impact on the viral particle
conformation. A range of production temperature was tested in mosquito cell line C6/36 from
28°C to 37°C. The threshold of structural shift was observed as 33°C73.

In a preliminary experiment, intermediate temperatures from 28°C to 37°C were tested and
the focus was directed onto the secreted E protein yield. These results were not shown in the
resulting publication. The higher the temperature, the more E protein was secreted. The two
temperatures 31°C and 37°C were tested during the formation of the particles and probed
with the 4G2 monoclonal antibody (mAb). There was difference in the recognition of the
epitope between the products of the two different temperatures. The 4G2 mAb could only
bind the E protein when the latter was expressed at 31°C. In order to find a balance between,
conformation and yield, the cells are expanded at 37°C and transferred at 31°C right after
transfection. This temperature-dependent conformation highlights the evidence of using full
length E protein for the formation of VPL. Structural studies were performed on virus
containing native sequence. EH2 is responsible for heterodimerization, and therefore the
organization of the heterodimers at the surface of the VPLs, utilization of the full length of E
protein combined with abrogation of the secretion signal was reinforced as strategy to
produce VLPs.

In the same system, the Zika VLP showed more thermostability. The VLPs were secreted and
recognized by the 4G2 mAb when produced at either temperature, but they showed a higher
yield of expression at 37°C.
V. CHARACTERIZATION OF THE VLPS AND EVALUATION AS VACCINE CANDIDATES.

In vitro expression and formation of the VLPs

Once the VLP strategy and expression systems were set up, the VLPs were characterized by
Western Blot in the cell lysate and by Dot blot after concentration and purification of the
supernatant. Finally, the VLPs were observed by Electron Microscopy. In order to obtain the
most relevant data, the Western blot were probed with polyclonal assay and the concentrated
secreted VLPs were probed with structural monoclonal antibody.

In the 2011, the only two structural antibodies readily available against DENV were the group
reactive antibody 4G2 and the serotype-specific antibody 3H5. Few years later the C10
antibody was made available. The 4G2 is not a neutralizing antibody, on the contrary it can
promote ADE with DENV. The rationale for using this antibody in vitro was to support the
temperature-dependent change of conformation of VLPs produced by the co-expression of
the CprME and NS2bNS3 genes.

a) 4G2

The mouse 4G2 mAb was discovered in 1982159. This antibody is group reactive and can
recognize various flaviviruses including dengue and Zika viruses. It recognizes the epitope at
the fusion loop on the tip of the DII. The epitope comprised of amino acids G104, G106, L107
and W231 and plays a key role in membrane fusion78. This antibody recognizes the two strands
of this domain which is a strong indicator that the E dimer is properly formed.

b) 3H5

This mouse monoclonal antibody was discovered alongside the 4G2 antibody. However the
epitope of this antibody was characterized in 2007160,161. This antibody is type-specific for
DENV-2 and does not react with the three other serotypes. It recognizes the DIII of the E
protein and more precisely the amino acids K305 and P384. These amino acids are not
adjacent, and the antibody recognizes a conformational epitope.

c) C10

The E protein is the main immunogenic structural protein of DENV. The E protein epitopes are
classified in immunogenic group: group-reactive, type-specific, depending how broadly
antibodies recognizing these epitopes are. The monoclonal antibody [753(3)] C10 is “E dimer–
dependent epitope 1”, or EDE1, which means it recognizes an epitope at the junctions of an E
protein dimer and does not require the N-153 glycosylation site for binding85,162. The EDE
epitope overlaps with the binding site of the precursor protein. This human antibody is broadly
neutralizing and can prevent DENV infection with any of the four serotypes. This antibody
interacts with DI and DIII and interferes with the integrity of the 150-loop. In the DIII, the
paratope contact surrounds the conserved E residue K310. The side chain of this K310 amino
 acid covers W101 in the fusion loop. The latter plays an important role in the stabilization of
the E dimer contact. This antibody also interacts with the two glycosylation sites N67 and
N153. This antibody is dependent of the maturation of the viral particle. Since it recognizes
the E-dimer, it cannot bind an immature particle where prM and E are organized in trimer at
the surface of the virion.

d) Zika antibodies

Zika virus presents only one serotype and multiple clades. Therefore, the concept of broaden
neutralizing epitopes in ZIKV is not as relevant as it is for DENV. ZIKV is also much more
thermostable. Production of the VLPs at 37°C did not abrogate recognition by the 4G2 mAb.
In 2016, Zika neutralizing antibodies were identified with their epitope mapping163. The
neutralizing Z-64, Z-48 and Z-67 antibodies were obtained from mice immunized with strain
H/PF/2013. These antibodies recognize spatially distinct regions in the DIII of the E protein:
the C’-C loop and the lateral ridge. Detailed comparison with cryo-EM-derived model revealed
the Z-67 antibody recognizes the mature virion at the lateral ridge portion whereas the
antibody Z-64 recognizes a cryptic epitope at the C’-C loop. The combination of cryptic epitope
and neutralization was observed before with the anti-DENV1 DV1-E111 antibody164. The solely
Z-67 antibody was able to neutralize infection in vivo whereas Z-64 and Z-48 could moderately
neutralize the virus in vitro. These three antibodies where used to probe the secreted Zika
VLP.

Immunogenicity in Balb/c mice

Once the dengue and Zika VLPs were characterized, they were tested in vivo in Balb/c mice.
The immunogenicity was assessed with ELISA and neutralization assay. A convalescent serum
for Zika and a pool of convalescent sera for dengue were included to benchmark the potency
of the elicited antibodies. Finally, an ADE assay was developed and tested using Zika immune
sera against DENV infection. The VLP vaccines were compared to their formalin-inactivated
viral counterparts

VI. STEP BY STEP

The body of this work was initiated as proof of concept for the development of a monovalent
dengue VLP and a Zika VLP and their performance as vaccines. All the steps of the research
plan and study designs are described in Figure 11.

Figure 11 : Workflow of the dengue and Zika VLP-based vaccines development and evaluation.
 Part III. Results

Chapter 1. Dengue-2 virus-Like particles (VLP) based vaccine

elicits the highest titers of neutralizing antibodies when
produced at reduced temperature

I. ARTICLE
 Dengue-2 virus-like particle (VLP) based vaccine elicits the highest titers of
neutralizing antibodies when produced at reduced temperature
Hélène Boigard 1, Velasco Cimica 2, Jose M. Galarza !
TechnoVax, Inc., 6 Westchester Plaza, 6E, Elmsford, NY 10523, United States

article info A dengue vaccine capable of rapidly eliciting a robust and balanced immunity against the four virus serotypes after only a few
immunizations is greatly needed. We describe a new strategy to develop dengue vaccines based on the assembly of virus-like
particles (VLPs) utilizing the structural proteins CprME together with a modified complex of the NS2B/NS3 protease, which
Article history: enhances particle formation and yield. These VLPs are produced in mammalian cells and resemble native dengue virus as
Received 15 June 2018 demonstrated by negative staining and immunogold labelling electron microscopy (EM). We found that VLPs produced at lower
Received in revised form 18 October 2018 temperature (31 C) were recognized by conformational monoclonal antibodies (MAbs) 4G2, 3H5 and C10 whereas VLPs
Accepted 21 October 2018 Available
produced at higher temperature (37 C) were not recognized by these MAbs. To investigate the significance of these
online 28 October 2018
conformational discrepancies in vaccine performance, we tested the immunogenicity of VLP vaccines produced at 31 C or 37
C. Mice immunized with the VLP vaccine produced at 31 C (VLP-31 C) elicited the highest titer of neutralizing antibodies when
compared to those elicited by equivalent doses of the vaccine produced at 37 C (VLP-37 C), inactivated dengue virus vaccine
Keywords: Dengue
Vaccine or to the titer of a human anti-dengue-2 convalescence serum reference. Our results demonstrate that the conformation of the E
Viral-like particle protein displayed on the VLP vaccine plays a critical role in the induction of highly neutralizing antibodies. These findings will
Temperature guide development of a tetravalent vaccine capable of eliciting a robust and balanced neutralizing response against the four-
Neutralization dengue serotypes regardless of background immunity.
Structure
2018 TechnoVax, Inc. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
abstract (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction
Dengue is a mosquito-borne viral disease that affects humans of all ages in
tropical and subtropical regions around the world. Global expansion of the
mosquito vectors (primarily Aedes aegypti, and Aedes albopictus) has resulted
in the spread of the virus to more than 120 countries, infecting approximately
390 million people yearly [1]. This dramatic spread of the vector poses a threat
to approximately half of the world’s population and disease outbreaks impose a
heavy public health and economic burden to all affected areas of the world. Four
distinct virus serotypes (DENV14) can be transmitted by the bite of an infected
Aedes spp. mosquito causing an infection characterized by fever, headache,
myalgia, arthralgia and, depending on the severity of the infection, may progress
to dengue hemorrhagic fever/dengue shock syndrome (DHF/ DSS) [2]. Primary
infection generates long-term protection against the homologous serotype but
short-lived defense against heterologous serotypes. In fact, a secondary infection
Corresponding author.
E-mail address: jmgalarza@technovax.com (J.M. Galarza). 1 Present addresses: Applied
Biological Laboratories, 760 Parkside Avenue, Brooklyn, NY, 11226, USA. 2 Present addresses:
ATCC,10801 University Boulevard Manassas, VA 20110-2209, USA.
https://doi.org/10.1016/j.vaccine.2018.10.072
0264-410X/ 2018 TechnoVax, Inc. Published by Elsevier Ltd.
with a different serotype can trigger an antibody-dependent enhancement (ADE)
of disease that may result in DHF and DSS which are both severe life-threatening
conditions [3]. The need to elicit a robust and balanced neutralizing response
against the four-dengue serotypes, in order to prevent ADE, has hindered the
development of a safe and effective dengue vaccine.
After decades of ceaseless effort, just recently a dengue vaccine has
reached commercialization. This live-attenuated chimeric vaccine is based on
the yellow fever virus 17D vaccine strain, which provides the backbone to
carry the DENV prM-E structural proteins of each dengue serotype as
chimeric viruses. This vaccine presented serotype-dependent differences in
vaccine efficacy and lower efficacy in young children in clinical trials [1–6].
In addition to low level viremia following vaccination, this vaccine did not
evoke a sufficiently effective immunity in seronegative patients [7,8]. Other
vaccine candidates currently in clinical trials include live-attenuated virus
vaccines, purified inactivated viruses, a DNA vaccine composition and a
subunit vaccine [9–11]. The later vaccine formulation is comprised of four
recombinant truncated E proteins and is currently the only recombinant
vaccine candidate in clinical trials.
An alternative strategy for vaccine development involves the generation
of viral-like particles (VLPs) as main vaccine constituents. These
recombinant particles are self-assembled complex structures morphologically
similar to wild-type virus but are devoid of viral genetic material and thereby
unable to replicate or cause infection. Flavivirus subviral particles or VLPs
were initially detected in the supernatant of flavirius-infected cells [12] and
subsequent studies have shown that the sole expression of recombinant prM
and E were sufficient to drive the assembly and budding of subviral particles
[13–15]. These structures provide an attractive strategy for vaccine
development because the particulate nature of recombinant VLPs composed
of native proteins enhances antigen recognition, presentation and immune
stimulation [15–17]. Furthermore, the system allows for surface protein
engineering to optimize antigen configuration seeking enhancement of the
immune response. The proven efficacy and safety of licensed human VLP
based vaccines for HBV, HPV and HEV [18– 21] provide strong evidence of
the value of this approach for vaccine development.
In a previous study, we have demonstrated the assembly of Zika virus-like
particles (VLPs) using a combination of structural and non structural protein
and their ability to elicit high titers of neutralizing antibodies against Zika
virus following VLP immunization [22]. Here, we present data on the
formation of Dengue 2 VLPs using a similar strategy based on the expression
of structural and non-structural proteins and their distinct conformational and
immunological characteristics when produced at a lower temperature.

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

 Dengue viruses contain a positive sense single-stranded RNA genome and (GTX128093) and anti-NS2B (GTX124246) were purchased from GeneTex,
belong to the Flavivirus genus within the Flaviviridae family. There are four CA. The secondary antibodies anti-mouse and anti-rabbit were both purchased
distinct virus serotypes DENV-1, DENV-2, DENV-3 and DENV-4 currently from Pierce Thermo Fisher, MA (#31430 and #31460 respectively).
identified. These serotypes appear independently during endemic cycles of
transmission between humans and arthropod vectors [23]. The dengue viral
2.3. VLP production and purification
genome has a single open reading frame (ORF) encoding a polyprotein that
is cleaved co- and post-translationally by cellular and viral proteases into Production of dengue VLPs was carried out in suspension culture of
three structural proteins: capsid (C), the premembrane (prM) and the envelope Expi293TM cells transfected with the plasmids pcDNA3.4CprME (DO2)
(E) and seven non-structural proteins: NS1, NS2A/B, NS3, NS4A/B and NS5 pcDNA3.4-NS2b/NS3 (DO3) and at a 2:1 ratio and then maintained at 37 C for
(Fig. 1A). This enveloped virus is smooth and has on its surface 180 copies 4 h, at which point culture flasks were transferred to incubators set at either 37 C
of the E and (pr/M) M protein. Virus morphogenesis takes place on or 31 C. Transfected cells were harvested 72 h post-transfection and clarified via
virusinduced remodeled endoplasmic reticulum (ER) membranes leading to two successive centrifugations. The first clarification was performed at 400g for
the assembly and budding of complete but immature particles into the ER 10 min at 4⁰C followed by a second clarification at 10,000g for 10 min at 4 C.
system [24]. The spiky immature particles undergo an additional furin
Clarified supernatant fluid was concentrated by ultracentrifugation for 2 h at
protease cleavage of prM during trafficking through the ER and trans-Golgi
140,000g at 4 C. The pellet was resuspended with 1X PBSCaMg pH 7.2 (1X
network (tGN) [25] that contributes to the rearrangement of the E protein from
phosphate buffered saline supplemented with 1 mM MgCl2; and 1 mM
trimer to dimer conformations and turns the virus into smooth mature particles
CaCl2). VLPs were further purified by ultracentrifugation through a 20–60%
before they are released from infected cells. The surface E protein is the main
linear sucrose gradient in TN buffer (50 mM Tris-HCl
antigenic determinant of the virus and major target of the immune system.
Immunity directed toward the E protein is primarily mediated by neutralizing
antibody which, when present, confers protection against dengue [26]. The
conformational state of the E protein may play a significant role in the
induction of neutralizing antibody. In this work, we describe not only the
assembly of dengue 2 VLPs by solely expressing the dengue virus structural
and two non-structural proteins but also demonstrate that the conformational
properties of the particles are temperature dependent. In addition, we also
show by comparing the serum neutralizing response elicited in mice by
dengue VLP immunization, that the highest neutralizing titers were
stimulated by VLPs produced at the lower 31 C temperature.
2. Materials & methods

2.1. Gene and plasmid constructs

The structural genes of the dengue 2 strain 16681 Genbank U87411.1 were
chemically synthesized by GeneArt (Life Technology) according to a specifically
designed and codon-optimized sequence. DNA fragments were subcloned into
the plasmid vector pcDNA3.4 (Life Technologies) utilizing the NheI/NotI
restriction enzyme sites. Specific mutations were introduced in the prM gene
G88A, and in the E gene, I398L, M401A, M412L, as described by Purdy and
Chang [27,28] for optimization of VLP production.
The non-structural genes of dengue 2 strain 16681 Genbank U87411.1 NS2B
and the viral protease NS3 (NS3Pro portion of NS3) were synthesized as a single
codon optimized unit and subcloned into plasmid vector pcDNA3.4 via
XhoI/EcoRV restriction sites. A single mutation, L115A, was introduced into the
synthesized NS3 gene for enhancing its catalytic activity.
Plasmids were amplified in MAX Efficiency Stbl2TM Competent E. Coli Cells
(Life Technologies 10268-019) and purified from the bacteria utilizing an
EndoFree Plasmid Maxi Kit (Qiagen).

2.2. Virus, cells, and antibodies

Cultures of Vero cells (ATCC CCL-81TM) were maintained in VPSFM media

(Life Technologies 11681-020) supplemented with 2 mM L-Glutamine (Life
Technologies 25030-081), 2 mM GlutaMAXTM Supplement (Life Technologies
35050-061), 1X nonessential amino acids solution (Life Technologies 11140-
050), 1X ITSE (Invitria 777ITS032) and 500 ng/ml rhEGF (Life Technologies
PHG0314).
Dengue virus, DENV-2 Th-36 (ATCC VR-1810TM), was amplified in Vero
cells following virus inoculation at a low multiplicity of infection (MOI: 0.01).
Expi293TM (Life Technologies) cells were grown in Expi293 medium
(Gibco A1435101) and transfected using Expifectamine following the
manufacturer instructions (Life Technologies, A14635). The 4G2 antibody is an
in-house protein G purified monoclonal antibody from hybridoma D1-4G2-4-15
(ATCC HB-112) culture supernatant. Mouse monoclonal antibody 3H5 was
obtained through BEI Resources (NR-2556). Rabbit monoclonal anti-E antibody
[753(3) C10] was purchased from Absolute Antibody (#Ab00677-23.0). Rabbit
polyclonal antibodies, anti-C (GTX124247), anti-E (GTX127277), anti-prM
Fig. 1. The dengue virus genome contains a single open reading frame (ORF) that expresses a polyprotein comprised of both structural and non-structural proteins, which arise via several proteolytic
cleavages. (A). In early stages, the complex viral protease NS3 with its cofactor NS2B (;) self-cleaves before cleaving the capsid protein. A host cell signalase is also involved in the maturation of the
polyprotein (e). In a final step, the furin protease (D) cleaves the pr portion from the M protein to uncover E protein fusion peptide. (B). The VLP assembly strategy relies on the native properties of the
structural genes although key point mutations were introduced to improve protein processing and particle assembly. The furin cleavage site (D) was mutated at E88A to enhance cleavage at this location.
The mutations I398A, M401A, and M412L in the E protein were introduced to improve the amphipathic properties of the helical domain 1 and consequently enhance trafficking and secretion of E. The
viral NS3 protein was truncated maintaining only its N-terminal protease domain that was mutated at L115A to enhance its catalytic activity. The protease domain is kept as a single transcription unit
together with its cofactor NS2B. Purified and concentrated VLP-31 and VLPs-37 as well as dengue virus
grown at 37 C (DENV2-37) and 31 C and (DENV2-31) were tested by dot
blot using anti-E and anti-prM polyclonal antibodies as well as monoclonal
pH 7.2; 150 mM NaCl) for 4 h at 180,000g at 4 C using an SW40Ti rotor antibodies recognizing structural epitopes. Three microliters were adsorbed
(Beckman Coulter, CA). Protein content of the collected gradient fractions was onto a nitrocellulose membrane, which was then probed with primary and
analyzed by dot blot using a dengue specific antibody (Anti-E polyclonal secondary antibodies using the same protocol described in Western Blot.
antibody- Genetex #GTX127277). Selected fractions were combined, dialyzed
overnight against 1X PBS and concentrated by ultracentrifugation for 2 h at 2.6. Negative staining and immuno-gold labelling electron microscopy
140,000g at 4 C. The pellet was then suspended in 80 ml of 1XPBSCaMg and
loaded onto a second linear sucrose gradient (20–60%) in TN buffer. Fractions Gradient purified VLPs and DV2 samples were absorbed into 200-mesh
were collected, analyzed and processed in the same fashion as in the first linear carbon coated grid (EMS CF200-Cu) for 5 min. The grids were then washed
gradient described above. The total protein concentration of the purified VLP and stained with 2% uranyl acetate (EMS 224002). VLP samples for
material was estimated using the Bradford method (Fig. S1A). The E protein immunogold labelling EM examination were prepared as follows: sample
concentration of pooled fractions was determined using standard recombinant E coated carbon grids were blocked with 3% BSA in 0.1 M Sodium Cacodylate
protein (GENWAY, dengue 2 recombinant proteinGWB-BSP762) and anti-E buffer for 5 min, followed by incubation for 20 min in primary monoclonal
polyclonal antibody (Anti-E polyclonal antibody, Genetex #GTX127277) in antibody 3H5 diluted in PBS (1:100). Grids were then washed three times
quantitative Dot Blot assay (Fig. S1B). The production yield was 3 mg/L of E with 0.1 M sodium cacodylate buffer and then incubated for 20 min with
protein based on 100 ml batch. secondary goat anti-mouse antibody (1:30 dilution). After a final series of
three washes with 0.1 M sodium cacodylate buffer, grids were stained with
2.4. Western blot 2% uranyl acetate solution and examined with a Jeol-1400 electron
microscope at The Rockefeller University Imaging Center (New York, NY).
The transfected Expi293TM cell pellet was lysed with RIPA buffer (PI- 2.7. Mouse immunogenicity study
89901, Pierce Thermo Fisher, MA) and analyzed by Western blot together with
the concentrated culture supernatants. Both were loaded onto a 10–20% Tris- Ten groups of 6–8-week old BALB/c mice were inoculated twice (day 0
glycine SDS-PAGE gel and day 14) via the intramuscular (IM) route with either VLP vaccines
(EC61352BOX, Life Technologies, CA) and the proteins were separated via (TVXDO31C or TVXDO37C) or formalin inactivated DENV2 virus.
electrophoresis, and then electro-transferred from the gel onto a 0.45 mm Groups of mice received doses of either 1 mg or 5 mg of total E protein
nitrocellulose membrane (Life Technologies LC2001). The nitrocellulose content formulated alone or admixed in a 1:1 vol with a squalene-based oil-
membranes were blocked for 1 h at room temperature with blocking solution (5% in-water nano-emulsion AddaVax (InvivoGen, CA). Serum samples were
non-fat milk 1X TBS 0.1% Tween-20) followed by overnight incubation at room collected two weeks after the booster shot for immunogenicity evaluation.
temperature in primary antibody diluted in blocking solution. Membranes were We planned a study of a continuous response variable from matched pairs
washed three times for 5 min with 1X TBS 0.1% Tween-20 and then incubated of study subjects. Prior data indicate that the difference in the response of
for 1.5 h in secondary antibody (HRP conjugated anti primary) diluted in matched pairs is normally distributed with a standard deviation of 5. If the
blocking solution. Finally, membranes were washed three times with 1XTBS- true difference in the mean response of matched pairs is 18, we will need to
0.1% Tween-20 and developed using the ECL chemoluminescence system study 3 pairs of subjects to be able to reject the null hypothesis that this
(WP20005, Life Technologies, CA). The luminescent signal is detected using a response difference is zero with probability (power) 0.95. The Type I error
FluorChem imager and analyzed using Alphaview software (Protein Simple, probability associated with this test of this null hypothesis is 0.05. 2.8.
CA). Enzyme-Linked Immuno-Assay (ELISA)

2.5. Dot blot

 ELISA assays were performed in 96-well plates coated with 100 ml/well
of formalin inactivated DENV-2 virus (1 lg/ml of E protein content) and
incubated overnight at 4 C. The plates were washed three times with PBST
buffer (phosphate buffered saline plus 0.05% Tween-20) and then blocked
with 100 ll of blocking buffer (PBST plus 5% non-fat milk) for one hour at
room temperature. Mouse sera were diluted following a 4-fold serial dilution
in blocking buffer starting at 1:10; 50 ll of the diluted sera were incubated for
2 h at room temperature. After 6 washes with PBST, plates were incubated
for 2 h with 50 ll of HRP-conjugated goat anti-mouse antibody diluted 1:2000
in blocking buffer. Subsequently, plates were washed (6X) and developed by
adding 50 ll/well of ECL reagent. Light emission was measured at 425 nm
using a plate reader (Synergy, H1; BioTek, VT).
2.9. Plaque reduction neutralization test (PRNT)
The plaque reduction neutralization test was carried out in Vero cells
using dengue virus serotype 2 (DENV-2 Th-36) and sera from immunized
mice according to the method described for the evaluation of vaccine efficacy
[27,28]. Briefly, DENV-2 was amplified and titrated in Vero cells using the
same plaque visualization procedure described below. Vero cells were seeded
in 24 well plates at a density of 5 104 cells per well 24 h prior to the initiation
of the test. Control and tests sera were heat inactivated at 56 C for 30 min.
and then two-fold serially diluted in cell culture media supplemented with
penicillin and streptomycin. The same volume of diluted virus (expected to
form approximately 80 plaques per well) was added to each serum dilution.
The sera-virus mixture was then incubated for 1 h at 37 C in a 5% CO2
environment. Subsequently, each dilution was applied in duplicates into wells
of 85% confluent monolayer of Vero cells and incubated for 1 h at 37 C in a
5% CO2 environment. Thereafter, the inoculum was removed and a 1 ml
overlay of 2% carboxymethyl cellulose (CMC) in culture medium was added
to each well. Plates were then incubated for six days at which time the CMC
overlay was removed, cells were washed 1X with PBST (phosphate buffer
saline plus 0.05% Tween 20) and fixed with cold 80% acetone for 10 min at
room temperature (RT). Afterwards, plates were washed 1X with PBST and
incubated with blocking buffer (2.5% non-fat milk in PBS plus 0.5% of Triton
X-100) for 1 h at 37 C. Following one PBST wash, the primary antibody
(MAb 4G2) diluted 1/200 in blocking buffer was applied for 2 h at RT,
followed by two PBST washes and incubation for 1 h at RT with a goat anti-
mouse AP conjugated secondary antibody (1/2000) in blocking buffer.
Finally, plates were washed twice with PBST and one time with alkaline
phosphate buffer (APB: 100 mM Tris-HCl pH9.0, 150 mM NaCl, 1 mM
MgCl2). Viral plaques were detected by adding the alkaline phosphate (AP)
substrate nitro blue tetrazolium chloride (NTB) and 5-brome-4-chlore-
3indolyl phosphate (BCIP) prepared and used as previously described [22].
Plaques were counted and PRNT50s were determined using the PROBIT
method [29]. The neutralization power calculated is expressed as the
reciprocal of the highest serum dilution that neutralizes 50% of the virus.
Human pre-immune serum control and human anti-DENV-2 convalescent
serum reference were obtained from National Institute for Biological
Standards and Control (NIBSC), UK.
3. Results
3.1. Expression and processing of the structural and non-structural proteins in
the formation of dengue virus-like particles (VLPs)
We describe a new strategy to effectively assemble DENV-2 virus-like
particles (VLPs) displaying on their surface E proteins with distinct structural
conformations resulting from their production at different temperatures (31 C
versus 37 C).
Dengue virus-like particle (VLP) assembly, processing and secretion
occurred during co-expression of structural proteins CprME (plasmid DO2) and
non-structural proteins NS2B/NS3Pro (plasmid DO3) in suspension cultures of
Expi293TM human cells (Fig. 1B). Both the structural and non-structural protein
coding sequences were modified by truncation and substitution mutations to
facilitate polyprotein processing and enhance assembly and release of the VLPs.
Mutations were introduced at positions I398A, M104A and M412L within the
amphipathic domain-1 of the E protein to overcome restrictions in the processing
and maturation of the dengue polyprotein and thereby enhance VLP yield. (e.g.
retention signal in the E protein) [30]. The furin recognition sequence also carries
the mutation E88A which enhances furin cleavage. Furthermore, the viral NS3
protease/helicase was truncated retaining only the N-terminal protease domain
(NS3Pro), which contains the L115A mutation to enhance its catalytic activity.
The remaining protease cleavage sites within the CprME polypeptide, however,
were preserved in order to maintain a processing pattern analogous to the one
occurring in dengue virus infected cells (Fig. 1A and B).
Western blot analysis of transfected Expi293TM cell lysates showed
expression of both structural and non-structural proteins as well as proper
processing of the structural polyprotein. Regarding the polyprotein structural
products, the envelope E protein is detected at a molecular weight of 64 kDa and
prM and pr proteins show the predicted sizes of 30 kDa and 23 kDa respectively
as would be expected of a fully processed polyprotein (Fig. 2A and B). The
capsid protein and the M portion of prM could not be detected due to lack of
suitable specific antibody.
The molecular weight of the NS2B protein is similar to the one detected in
the DENV-2 cell lysate, suggesting that the selfcleavage between NS2B/NS3Pro
occurs efficiently at the corresponding cleavage site (Fig. 2C). The NS3Pro
portion of NS2B/ NS3Pro could not be detected due to the lack of a suitable
specific antibody. The efficient self-cleavage of NS2B/NS3Pro and the cleavage
of the capsid protein confirm that the amino acids 1–183 are sufficient for proper
protease activity as described by Li et al. [27,29]. The substitution mutation
L115A in the NS3Pro does not appear to interfere with protease activity.
To determine whether the NS2B/NS3pro have an effect on VLP production,
we analyzed by Western blot concentrated VLPs from cell culture supernatant.
DO2 (pcDNA3.4-CprME) alone showed
Fig. 2. Dengue VLPs were produced in Expi293TM cells by co-expression of CprMENS2B/NS3 or
CprME alone. Capsid protein cleavage was efficient when the NS2B/ NS3 protease was co-
expressed with CprME. The distinct cell lysates (20 mg of total protein per lane) were tested by
Western blot using specific antibodies: anti-E (A), anti-pr (B), anti-C (C) and anti-NS2B (D)
antibodies.
that a small amount of pr/M and E proteins were secreted into the culture medium
(Fig. 3A and B). However, release of the structural proteins was enhanced when
the plasmid DO2 expressing the structural genes (CprME) was co-transfected
with the plasmid DO3 that expresses the modified non-structural proteins NS2/
NS3pro. Thus, this protein combination offers the most effective approach for
VLP production (Fig. 2A–C).
3.2. Antigenic profiles of VLPs produced at different temperatures
Dengue virus is able to replicate in both mosquito and humans illustrating the
broad range of temperatures (ambient versus 37 C respectively) over which the
virus can complete its replication cycle.
 We decided to assess the antigenic profile of VLPs produced at either 31 C
or 37 C and compared them to that of wild-type virus using a dot blot assay (Fig.
4A–E). The monoclonal antibodies used for the assay recognize conformational
epitopes on the surface of the E protein. The MAb 4G2 (Fig. 4C) recognizes a
conformational epitope in the domain DII of E at the fusion loop, which is
comprised of amino acids G104, G106, L107 and W231 and plays a key role in
membrane fusion. The second MAb 3H5 (Fig. 4D) recognizes a conformation
epitope in the domain DIII of the E protein, comprised of amino acid residues
K305, P384 which are essential for binding [34,35]. The third MAb C10 (Fig.
4E) recognizes a quaternary ‘‘E dimer epitope”. This determinant includes a
strand of domain II of an E protein and the ‘‘150 loop” of Domain I of the
opposite E protein within the same dimer. The Genetex anti-E and anti-prM
polyclonal antibodies showed identical E and prM protein content in all the
samples (Fig. 4A and B respectively).
The 4G2 antibody showed the strongest reactivity with DENV237, then
DENV2-31 and finally VLP-31 (Fig. 4C). This monoclonal antibody had almost
no reactivity with VLP-37. As for the recognition of domain III of the E protein
by the 3H5 antibody, there was no difference in binding between the virus grown
at both temperatures and VLP-31, however, this antibody could barely recognize
the VLP-37 and the difference with VLP-31 was significant
(Fig. 4D).
labeling only of VLP-31, confirming the assembly and release of particles
composed of prM/E that exhibited the conformational epitopes recognized by
these MAbs (Fig. 5B). In contrast, the VLP-37 VLPs did not react with either
4G2 or 3H5 (data not shown).
3.4. Immunogenicity evaluation of VLPs in mice by ELISA
The recombinant DENV-2 VLPs showed immunological and structural
similarities with DENV-2 virus when produced at 31 C but not at 37 C. To
ascertain whether these distinctions play a significant role in the
immunogenicity of the VLPs and therefore in their effectiveness as vaccine
candidates, we performed immunization studies in BALB/c mice. Groups of
mice (n = 4) were immunized twice, two weeks apart, via the intramuscular
route with either VLP-31 or VLP-37 at the dose of 1 mg or 5 mg of total E
protein

Fig. 3. Dengue VLP formation and release is enhanced by co-expression of CprME and the viral
protease complex NS2B/NS3. Expi293TM cells were transfected with CprME alone or CprME
together with NS2B/NS3. Cell supernatant from mock transfected or CprME or CprME-
NS2B/NS3 transfected cells were purified by ultracentrifugation. Purified VLPs were analyzed
via Western blot using 20 lg of total protein per lane. Purified DENV-2 virus was used as control
(20 mg). The Western blot membranes were probed with (A) an anti-E antibody and (B) an
antiprM antibody.

Finally the C10 antibody showed strong reactivity with DENV237, which
decreased with DENV2-31, VLP-31 and then VLP-37. Despite the varying
reactivity with this antibody, there was no statistically significant difference
between DENV2-31 and VLP-31 and neither between VPL-37 and VLP-31
(Fig. 4E).
This evaluation showed that following the expression of identical
structural and non-structural genes and equal amount of E protein, the VLPs
secreted into the cell culture media do not exhibit equally available structural
epitopes. The maturation and conformation of the dengue VLPs are
temperature dependent. These results support the notion that the E protein
display on VLPs produced at a lower temperature adopt structures that better
exhibit conformational epitopes, which are critical for the induction of
neutralizing antibodies.

3.3. Electron microscopy (EM) examination of DEN-2 VLPs

In view of the observed conformational differences created by incubation

at either 31 C or 37 C, we chose to carry out a closer examination of purified
VLPs by negative staining and electron microscopy in order to further
evaluate the morphology, shape, size and surface composition of the secreted
virus-like particles. EM examination of VLP-31 (Fig. 5A) showed that the
particles exhibit a spherical morphology of approximately 50 nm diameter,
resembling the structural characteristics of mature dengue virus as has been
previously shown [31]. No significant differences were noticed between VLP-
37 and VLP-31 although some smaller particles and fused VLP were detected.
Immuno-gold labelling EM of dengue-2 VLPs using the conformational-
epitope recognizing MAbs 3H5 and 4G2 showed surface reactivity and
 Fig. 4. Analysis of the effect of temperature in the reactivity of the envelope (E) protein display on the surface of the dengue VLP with MAb recognizing conformational epitopes. Samples of gradient
purified virus and VLPs fractions were applied to nitrocellulose membrane and probed by dot blot with specific antibodies. Virus produced at 37 C or 31 C are shown in purple and yellow respectively
as compared to VLP produced at 37 C in red and at 31 C in green. Samples were probed with the following antibodies: A. Anti-E polyclonal antibody, B. anti-prM polyclonal antibody, C. 4G2 MAb,
D. 3H5 MAb and E. C10 MAb. VLPs produced at 31 C appear to display a better E protein folding as indicated by reactivity with 4G2 and 3H5 MAb, which also react with DENV-2 virus control.
Binding to C10, however, did not show a significant difference. Stars indicate statistically significant differences (P < 0.05).

Fig. 5. Electron microscopy study of gradient purified DENV- 2 VLP (bar represents 100 nm). (A) Negative staining with 2% Uranyl acetate. (B) Two particles observed after Immuno-gold staining
with 3H5 monoclonal antibody.
content and formulated with or without adjuvant. Two weeks after the booster neutralizing antibody titers by plaque reduction neutralization assay. Mice
immunization, we collected serum samples from vaccine and control mice and immunized with 5 mg of VLP-31 vaccine with or without adjuvant,
assessed the antibody response by measuring total IgG levels by ELISA and
formulations of VLP-37 vaccine or the inactivated DENV-2 virus control (Fig.
6). Similarly, the VLP-31 vaccine administered at the dose of 1 mg elicited
production of higher IgG levels than the VLP-37 or inactivated virus vaccine.
Although the VLP-31 vaccine induced the highest IgG levels, in a dose response
manner, it was only the 1 mg adjuvanted dose that showed a statistically
significant difference with the VLP-37 vaccine (Fig. 6).
3.5. Immunogenicity assessment by plaque reduction neutralization assay
As we have shown, both VLPs are composed of identical dengue proteins and
show similar size and shape. Observed differences between these viral-like
particles, however, seem to reside in the molecular structure of the E protein as
detected with monoclonal antibodies recognizing different conformational
epitopes. Elicitation of high titers of neutralizing antibody is paramount for
dengue protection and therefore we measured the levels of neutralizing
antibodies elicited by these vaccines using a plaque reduction neutralization test
(PRNT). This assay was performed according to the World Health Organization
(WHO) guidelines and protocols [27,29]. The PRNT50 of reciprocal dilutions
was calculated using the PROBIT methods and compared with pre-immune
mouse serum and reference controls (Fig. 7).
Both adjuvanted vaccines, at the dose of either 1 mg or 5 mg elicited higher
neutralizing titers than the non-adjuvanted preparations. The non-adjuvanted
VLP-31 vaccine at the dose of 1 mg and 5 mg stimulated PRNT50 neutralizing
antibody titers of equivalent levels. When this vaccine was admixed with
adjuvant the potency of the neutralizing response increased to a PRNT50 of 394
for the 1 mg dose and to a PRNT50 of 1517 for the 5 mg dose. In addition, the
adjuvanted formulations of both high dose and low dose of VLP-31 vaccine
presented a PRNT50 higher than the inactivated virus. Although the VLP-37
adjuvanted 5 mg dose vaccine improved neutralizing titers almost to the level of
the inactivated DENV-2, it did not reach the neutralizing power induced by the
VLP-31 vaccine (PRNT: 345 versus PRNT: 1517).
Fig. 6. Elisa Assay. Analysis of anti-DENV specific antibodies (total IgG) in mouse serum (n = 4)
following immunization with dengue VLP vaccine produced at 37 C and 31 C as well as inactivated
dengue-2 virus control (DV2) via chemiluminescence ELISA. Pre-immune sera (Pre I.) show
statistical difference (*) (P < 0.05) with the immunized groups. Stars indicate statistically significant
differences (P < 0.05).
Fig. 7. Neutralization power of vaccinated mice sera is presented as the reciprocal of serum
dilution for which 50% of the virus is neutralized (PRNT50). PRNT50 is calculated using
PROBIT method (Finney, 1952). See Material and Methods. Stars indicate statistically
significant differences (P < 0.05).
Furthermore, to better judge the neutralization power elicited by the
different vaccine doses, formulations and controls as well as the performance
of our assay, we tested a standard of human pre-immune and human
convalescence DENV-2 sera obtained from the National Institute for
Biological Standard and Control, UK (NIBSC). Because of the absence of
biological replicate for this sample, we were unable to compare statistically
with our mouse serum, however, the technical replicates of human sero-
negative control were as low as the mouse pre-immune sera. The human
convalescence DENV-2 sera control showed neutralizing titers slightly higher
than the unadjuvanted VLP vaccine compositions, however, this titer was
significantly exceeded by the neutralization response elicited by the
adjuvanted VLP-31 formulation. Thus, these results show that the VLP
vaccine produced at lower temperature, VLP-31 C and formulated with
adjuvant, elicits the strongest anti-DENV-2 neutralizing antibody response.
4. Discussion
In this study; we report a novel strategy that combines the expression of
the complete set of structural proteins (CprME) as a single polypeptide
together with the expression of a modified NS2B/NS3 protease complex for
the formation and release of dengue 2 virus-like particles (VLPs). We selected
this strategy to maintain the native processing of the polyprotein, which
retained the capsid protein upstream of prME instead of using a signal peptide
as has been previously described [32,33]. In order to completely process the
polyprotein CprME, we co-expressed a modified NS2B/NS3 complex that
provides the viral protease function. Furthermore, we have introduced specific
mutations and truncations to overcome the effects of natural features of
dengue virus such as retention signals, which may have an impact in vaccine
production.
Dengue virus can replicate in both mosquitos at ambient temperature and
in humans at 37 C. It has been reported that a lower temperature increased the
yield of dengue replicons, and also that temperature is an important factor in
the viral envelope conformation [34,35].
Molecular breathing of the E protein on the surface of dengue virus was first
described by Zhang et al. (2013) and Fibriansah et al. (2013) [36,37]. Detailed
cryo-electron microscopic analysis of dengue virus has shown that mature virus
produced at 28 C is smooth, however, when incubated at temperatures higher
than 33 C it appears bumpy, with an increase in diameter and greater
heterogeneity in shape and size [36]. The temperature changes the structure of
the virus hence affecting the antigenic conformation of the viral particle.
Although it has been reported that MAbs react differently with dengue virus of
distinct conformations [38,39], the role of dissimilar conformations of E in the
elicitation of neutralizing antibodies has not been completely elucidated
 [37,38,40].
Some broadly and highly neutralizing monoclonal antibodies have been
identified and shown to recognize the ‘‘E-dimer-depen dent-epitope” (EDE)
[38]. Another structural neutralizing antibody recognizes a conformational
epitope in the DIII and glycan loop of DI of E protein and the DII, including
the fusion peptide of the opposite E protein within the same dimer [37].
We therefore decided to investigate the effect of temperature on the
antigenic conformation of our dengue VLP using different monoclonal
antibodies that recognize conformation epitopes on the virion surface. VLPs
produced at 31 C were strongly recognized by two distinct MAbs (4G2 and
3H5) that bind conformation epitopes in domain II (4G2) and domain III
(3H5) of the E protein. In contrast, VLPs produced at 37 C were neither
recognized by 4G2 nor 3H5, revealing that the targeted epitopes were not
properly folded in two different domains of the E protein. Analysis of binding
of the C10 monoclonal antibody to the VLPs produced at either 31 C or 37 C
showed a slightly higher reactivity with VLP-31, over VLP-37 although this
difference was not significant. In contrast, the binding difference between
VLP-31 and VLP-37 when probed with 4G2 and 3H5 was pronounced and
statistically significant. These findings suggest that production temperature
has a notable effect on the folding and conformation of the E protein exhibited
on the surface of the VLPs. The difference between VLP produced at 31 C
versus 37 C seems to be subtle variation in epitope conformations, which
could not detected with the EM method used in the study. However, more
detail analysis of these structures, perhaps using cryo-electron microscopy
may reveal the basis of this distinction.
Finally we assessed the elicitation of neutralizing antibodies by both the
VLP-31 and VLP-37 in a mouse model. To provide a reference to the PRNT
results, we tested pooled human convalescent serum from patients with
clinically confirmed dengue-2 infection (NIBSC/WHO). This serum was
shown to contain lower neutralizing titers than those obtained with sera from
mice immunized with either dose of adjuvanted VLP-31 vaccine. We have
also shown that the adjuvanted formulation of VLPs produced at 31 C elicited
higher neutralizing antibody titers than that elicited by immunization with
either VLP produced at 37 C or inactivated wild-type virus. The low serum
neutralization titer elicited by the VLP37 C at high dose without adjuvant may
be due to instability and aggregation of the particles at this concentration. This
effect might have been prevented in the presence of adjuvant, which may
stabilize the VLPs and their antigens. The VLP structure produced at 31 C
appears to display a larger repertoire, arrangement or frequency of
neutralizing sites evidenced by the induction of a stronger neutralizing
response. It is even possible that these conformational attributes arise after
inoculation as the consequence of molecular breathing at the higher
temperature of the recipient host (37 C). The best neutralizing conformational
sites of the VLP-31 composition were remarkably boosted when the vaccine
was formulated with an adjuvant suggesting that the induction of a highly
neutralizing antibody response may require not only the optimal epitope
conformation but also stimulation of the most effective antigen presentation
as well as induction of robust innate and adaptive immunity.
In summary, this work not only describes a new strategy for VLP-based
dengue vaccine development but also reveals the effect of temperature on the
VLP E protein conformation and how these structural dissimilarities dictate the
neutralizing antibody response, which is higher when the VLP vaccines are
produced at lower temperature. This remarkable outcome holds great promise for
the development of a VLP based dengue vaccine.
Ethics statement
The Institutional Animal Care and Use Committee (IACAC) of the New York
Medical College approved the protocol for the mouse study. Animal work was
performed in compliance with the approved IACUC protocol (#34-2-053H) and
the Institutional Biosafety Committee policies, and performed under strict
accordance to the Office of Laboratory Animal Welfare (OLAW) guidelines, and
the Public Health Service (PHS) Policy on Humane Care and Use of Laboratory
Animals (NIH).
Conflict of interest
HB and JMG are employees of TechnoVax Inc., VC is an employee of the
ATCC BEI Resources and was employed of TechnoVax, Inc. during the
execution of this work. I have read the journal’s policy and the authors of this
manuscript have the following competing interests: HB and JMG are the
inventors on a patent related to this publication. This does not alter our adherence
to all the Vaccine Journal policies regarding sharing data and materials.
Acknowledgements
This work was supported in part by a Qualifying Therapeutic Discovery
Project Grant Reviewed by the National Institutes of Health (NIH), The
Department of Health and Human Service (HHS) for the Development of a
Multivalent Dengue Virus-Like Particle (VLP) Vaccine to TechnoVax,
Inc./J.M.G. (2010 – NIH/QTDP).
We thank Dr. George Martin and Dr. Katherine Frederick for helpful
discussions and critical review of the manuscript. The following reagents were
obtained through BEI Resources, NIAID, NIH: Dengue Virus Type 1, 228690,
NR-3796; Dengue Virus Type 3, Philippines/H87/1956, NR-80; Monoclonal
Anti-Dengue Virus Type 2 Envelope Protein, Clone 3H5-1 (produced in Vitro),
NR-2556.
Appendix A. Supplementary material
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.vaccine.2018.10.072.
References
[1] Bhatt S, Gething PW, Brady OJ, Messina JP, Farlow AW, Moyes CL, et al. The global
distribution and burden of dengue. Nature 2013;496:504–7. https://doi.
org/10.1038/nature12060.
[2] Martina BEE, Koraka P, Osterhaus ADME. Dengue virus pathogenesis: an integrated view.
Clin Microbiol Rev 2009;22:564–81. https://doi.org/10.1128/ CMR.00035-09.
[3] Guzman MG, Vazquez S. The complexity of antibody-dependent enhancement of dengue
virus infection. Viruses 2010;2:2649–62. https://doi.org/10.3390/ v2122649.
[4] Sabchareon A, Wallace D, Sirivichayakul C, Limkittikul K, Chanthavanich P, Suvannadabba
S, et al. Protective efficacy of the recombinant, live-attenuated, CYD tetravalent dengue
vaccine in Thai schoolchildren: a randomised, controlled phase 2b trial. Lancet
2012;380:1559–67. https://doi.org/10.1016/ S0140-6736(12)61428-7.
[5] Capeding MR, Tran NH, Hadinegoro SRS, Ismail HIHM, Chotpitayasunondh T, Chua MN,
et al. Clinical efficacy and safety of a novel tetravalent dengue vaccine in healthy children in
Asia: a phase 3, randomised, observer-masked, placebo-controlled trial. Lancet 2014.
https://doi.org/10.1016/S0140-6736(14) 61060-6.
[6] Villar L, Dayan GH, Arredondo-García JL, Rivera DM, Cunha R, Deseda C, et al. Efficacy
of a tetravalent dengue vaccine in children in latin America. N Engl J Med 2015;372:113–23.
https://doi.org/10.1056/NEJMoa1411037.
[7] Torresi J, Richmond PC, Heron LG, Qiao M, Marjason J, Starr-Spires L, et al. Replication
and excretion of the live attenuated tetravalent dengue vaccine CYD-TDV in a flavivirus-
naive adult population: assessment of vaccine viremia and virus shedding. J Infect Dis
2017;216:834–41. https://doi.org/10.1093/ infdis/jix314.
[8] Yang Y, Meng Y, Halloran ME, Longini IM. Dependency of vaccine efficacy on preexposure
and age: a closer look at a tetravalent dengue vaccine. Clin Infect
Dis Off Publ Infect Dis Soc Am 2018;66:178–84. https://doi.org/
10.1093/cid/cix766.
[9] Coller B-AG, Clements DE, Bett AJ, Sagar SL, ter Meulen JH. The development of
recombinant subunit envelope-based vaccines to protect against dengue virus induced
disease. Vaccine 2011;29:7267–75. https://doi.org/10.1016/
j.vaccine.2011.07.021.
[10] Weiskopf D, Angelo MA, Bangs DJ, Sidney J, Paul S, Peters B, et al. The human CD8+ T
cell responses induced by a live attenuated tetravalent dengue vaccine are directed against
highly conserved epitopes. J Virol 2014. https://doi.org/ 10.1128/JVI.02129-14.
[11] Kirkpatrick BD, Whitehead SS, Pierce KK, Tibery CM, Grier PL, Hynes NA, et al. The live
attenuated dengue vaccine TV003 elicits complete protection against dengue in a human
challenge model. Sci Transl Med 2016;8. https://doi.org/ 10.1126/scitranslmed.aaf1517.
330ra36-330ra36.
[12] Smith TJ, Brandt WE, Swanson JL, McCown JM, Buescher EL. Physical and biological
properties of dengue-2 virus and associated antigens. J Virol 1970;5:524–32.
[13] Allison SL, Tao YJ, O’Riordain G, Mandl CW, Harrison SC, Heinz FX. Two distinct size
classes of immature and mature subviral particles from tick-borne encephalitis virus. J Virol
2003;77:11357–66.
[14] Allison SL, Stadler K, Mandl CW, Kunz C, Heinz FX. Synthesis and secretion of
recombinant tick-borne encephalitis virus protein E in soluble and particulate form. J Virol
1995;69:5816–20.
[15] Lorenz IC, Kartenbeck J, Mezzacasa A, Allison SL, Heinz FX, Helenius A. Intracellular
assembly and secretion of recombinant subviral particles from tick-borne encephalitis virus.
J Virol 2003;77:4370–82. https://doi.org/ 10.1128/JVI.77.7.4370-4382.2003.
[1] Kushnir N, Streatfield SJ, Yusibov V. Virus-like particles as a highly efficient vaccine
platform: diversity of targets and production systems and advances in clinical development.
Vaccine 2012;31:58–83. https://doi.org/10.1016/ j.vaccine.2012.10.083.
[2] Frietze KM, Peabody DS, Chackerian B. Engineering virus-like particles as vaccine
platforms. Curr Opin Virol 2016;18:44–9. https://doi.org/10.1016/ j.coviro.2016.03.001.
[3] Valenzuela P, Medina A, Rutter WJ, Ammerer G, Hall BD. Synthesis and assembly of
hepatitis B virus surface antigen particles in yeast. Nature 1982;298:347–50.
[4] Keating GM, Noble S. Recombinant hepatitis B vaccine (Engerix-B): a review of its
immunogenicity and protective efficacy against hepatitis B. Drugs 2003;63:1021–51.
[5] Joura EA, Leodolter S, Hernandez-Avila M, Wheeler CM, Perez G, Koutsky LA, et al.
Efficacy of a quadrivalent prophylactic human papillomavirus (types 6, 11, 16, and 18) L1
virus-like-particle vaccine against high-grade vulval and vaginal lesions: a combined analysis
of three randomised clinical trials. Lancet Lond Engl 2007;369:1693–702.
https://doi.org/10.1016/S0140-6736(07) 60777-6.
[6] Zhang J, Zhang X-F, Huang S-J, Wu T, Hu Y-M, Wang Z-Z, et al. Long-term efficacy of a
hepatitis E vaccine. N Engl J Med 2015;372:914–22. https://doi.
org/10.1056/NEJMoa1406011.
[7] Boigard H, Alimova A, Martin GR, Katz A, Gottlieb P, Galarza JM. Zika virus-like particle
(VLP) based vaccine. PLoS Negl Trop Dis 2017;11:e0005608. https://
doi.org/10.1371/journal.pntd.0005608.
[8] Twiddy SS, Holmes EC, Rambaut A. Inferring the rate and time-scale of dengue virus
evolution. Mol Biol Evol 2003;20:122–9. https://doi.org/10.1093/ molbev/msg010.
[9] Apte-Sengupta S, Sirohi D, Kuhn RJ. Coupling of replication and assembly in flaviviruses.
Curr Opin Virol 2014;9:134–42. https://doi.org/10.1016/ j.coviro.2014.09.020.
[10] Umareddy I, Pluquet O, Wang QY, Vasudevan SG, Chevet E, Gu F. Dengue virus serotype
infection specifies the activation of the unfolded protein response. Virol J 2007;4:91.
https://doi.org/10.1186/1743-422X-4-91.
[11] Murphy BR, Whitehead SS. Immune response to dengue virus and prospects for a vaccine.
Annu Rev Immunol 2011;29:587–619. https://doi.org/10.1146/ annurev-immunol-031210-
101315.
[12] Timiryasova TM, Bonaparte MI, Luo P, Zedar R, Hu BT, Hildreth SW. Optimization and
validation of a plaque reduction neutralization test for the detection of neutralizing antibodies
to four serotypes of dengue virus used in support of dengue vaccine development. Am J Trop
Med Hyg 2013;88:962–70. https://doi.org/10.4269/ajtmh.12-0461.
[13] Roehrig JT, Hombach J, Barrett ADT. Guidelines for plaque-reduction neutralization testing
of human antibodies to dengue viruses. Viral Immunol 2008;21:123–32.
https://doi.org/10.1089/vim.2008.0007.
[14] Probit Analysis. 2nd ed. Finney DJ, editors. New York:Cambridge University Press, 1952.
[22.5 14 cm. xiv + 318 pp - 1952 – Journal of the American Pharmaceutical Association –
Wiley Online Library n.d.] https://onlinelibrary.
wiley.com/doi/abs/10.1002/jps.3030411125 [accessed April 22, 2018].
[15] Purdy DE, Chang G-JJ. Secretion of noninfectious dengue virus-like particles and
identification of amino acids in the stem region involved in intracellular retention of envelope
protein. Virology 2005;333:239–50. https://doi.org/ 10.1016/j.virol.2004.12.036.
[16] Zhang W, Chipman PR, Corver J, Johnson PR, Zhang Y, Mukhopadhyay S, et al.
Visualization of membrane protein domains by cryo-electron microscopy of dengue virus.
Nat Struct Mol Biol 2003;10:907–12. https://doi.org/10.1038/ nsb990.
[17] Zhang S, Liang M, Gu W, Li C, Miao F, Wang X, et al. Vaccination with dengue virus-like
particles induces humoral and cellular immune responses in mice. Virol J 2011;8:333.
https://doi.org/10.1186/1743-422X-8-333.
[18] Konishi E, Fujii A. Dengue type 2 virus subviral extracellular particles produced by a stably
transfected mammalian cell line and their evaluation for a subunit vaccine. Vaccine
2002;20:1058–67.
[19] Ansarah-Sobrinho C, Nelson S, Jost CA, Whitehead SS, Pierson TC. Temperature-dependent
production of pseudoinfectious dengue reporter virus particles by complementation. Virology
2008;381:67–74. https://doi. org/10.1016/j.virol.2008.08.021.
[20] Qing M, Liu W, Yuan Z, Gu F, Shi PY. A high-throughput assay using dengue-1 virus-like
particles for drug discovery. Antiviral Res 2010;86:163–71. https://
doi.org/10.1016/j.antiviral.2010.02.313.
[21] Zhang X, Sheng J, Plevka P, Kuhn RJ, Diamond MS, Rossmann MG. Dengue structure
differs at the temperatures of its human and mosquito hosts. Proc Natl Acad Sci USA
2013;110:6795–9. https://doi.org/10.1073/ pnas.1304300110.
[22] Fibriansah G, Tan JL, Smith SA, de Alwis R, Ng T-S, Kostyuchenko VA, et al. A highly
potent human antibody neutralizes dengue virus serotype 3 by binding across three surface
proteins. Nat Commun 2015;6:6341. https://doi.org/ 10.1038/ncomms7341.
[23] Rouvinski A, Guardado-Calvo P, Barba-Spaeth G, Duquerroy S, Vaney M-C, Kikuti CM, et
al. Recognition determinants of broadly neutralizing human antibodies against dengue
viruses. Nature 2015. https://doi.org/ 10.1038/nature14130.
[24] Dowd KA, DeMaso CR, Pierson TC. Genotypic differences in dengue virus neutralization
are explained by a single amino acid mutation that modulates virus breathing. MBio 2015;6.
https://doi.org/10.1128/mBio. 01559-15.
[25] de Alwis R, Smith SA, Olivarez NP, Messer WB, Huynh JP, Wahala WMPB, et al.
Identification of human neutralizing antibodies that bind to complex epitopes on dengue
virions. Proc Natl Acad Sci USA 2012;109:7439–44. https://doi.org/
10.1073/pnas.1200566109.
II. DISCUSSION

The present article described a new strategy for the development of dengue VLPs by co-
expression of the three structural proteins, CprME and two nonstructural proteins N2b and
NS3 in mammalian cells. The secreted VLP produced at 31c and 37C displayed different
structural epitopes. In an in vivo immunogenicity study in Balb/c mice, the particles produced
at lower temperature elicited more neutralizing antibodies than those produced at higher
temperature and then the inactivated virus.

Missing elements

Few years of retrospective have allowed identification of a couple of elements which would
have provided more impact to the presented findings.

In a startup environment, the priority was to gain credibility and visibility by publishing a peer-
reviewed article. While reaching this goal as fast as possible, some additional assays would
have greatly improved the attractiveness of the strategy. In retrospective the missing
elements were:

- immunogenicity of 31°C VLP transferred to 37°C before injection: it would have been very
interesting to assess immunogenicity in mice of VLP produced at 31°C and incubated at 37°C
for 30 minutes after purification as well as dot blot analysis. It would have been interesting to
find out with electron microscopy if the VLP present indeed the molecular breathing as
observed with the wild-type virus. Moreover, these particles would be more representative of
what occurs to the viral particle in the wild providing additional insights on the link between
structure and immunogenicity.

- ELISA and PRNT assay performed using virus produced at 31°C: a side-by-side comparison of
the elicitation of IgG recognizing virus amplified at both temperature as well as assessment of
the neutralization of the serum of these two viruses.

- an additional broaden neutralizing antibody: being outside of academia, it was very difficult
to obtain non-commercially available monoclonal antibodies. Cryo-EM study was not possible.
It would have been very interesting to test other broaden neutralizing antibodies recognizing
different structural epitopes.

- ADE assay: it would have been very interesting to perform an ADE assay with a different viral
serotype as well as ELISA and PRNT assays and evaluate whether the immune response to VLP
produced at 31°C is less likely to promote a subsequent ADE than the particles produced at
37°C.

Since the design of the strategy, research plan, and the publication of the article, new
information about the role of the capsid protein have been published.
 Influence of Capsid in polyprotein processing and maturation

Our strategy was innovative as it aimed at mimicking the native conformation of the viral
envelope by including the 3 structural proteins C, prM and E. The addition of NS2b/NS3 to
ensure proper processing of the polyprotein has been a novel element comparing to other
strategies. The minimal structural proteins to form and release VLPs was recorded in the
literature as prM and E. In recent years, a detailed investigation by Rana et al. of the capsid
protein of dengue and more precisely of the capsid anchor, revealed the anch-C is a
transmembrane helix and it is 14 amino acid-long in dengue virus comparing to 18-22 amino
acids in other flaviviruses165. In addition to its shorter length, the anch-C. The dengue C
transmembrane domain comprises a P110 proline which is a known helix disruptor amino acid.
This proline is positioned at P5 of the host signal peptidase cleavage site and is conserved
among all four serotypes. Rana et al. transfected WNV GFP viral replicon along with packaging
constructs in HEK293 cells and then placed onto Vero cells for infection. The packaging
constructs encoded for dengue structural proteins or for variations the latter by replacing the
C-anchor-C or anchC-pr junction sequence by heterologous sequences of ZIKV and West-Nile
virus. As a result, any replacement in the anchC produced higher amounts of infectious
particles whereas modification in the first five amino acids of pr decreased the secretion of
infectious particles. The shorten anch-c fragment in dengue results in the burial of the
NS2b/NS3 protease cleavage side. When the dengue anchC is elongated with the four amino
acids of WNV or Zika, the pseudoparticle production is improved. This study shows not only
the importance of wild-type anchC-pr junction, it also shows how c-anchC is a major driver in
VLP processing and assembly. This study highlights potential VLP secretion enhancement
strategy. The involvement of the P110 was also investigated. The mutation P110A resulted in
premature cleavage of the furin and abrogated pseudoparticle formation by preventing signal
peptidase activity resulting in the formation of AnchC-pr in both wild type and extended anchC
mutant. These findings corroborate with the MVAV mutation in MVE described in the
introduction highlighting significance of sequential cleavage and processing of the
polyprotein. Authors mentioned too that the production of wild-type dengue pseudo particles
was lower than WNV or ZIKV. These recent findings reinforce the role of the capsid in the
proper formation of the viral particle and provides an additional modification which could
potentially improve the formation and release of dengue VLP.

Structural findings – variations among serotypes. --

Cryo-EM studies generated the first evidence that temperature impacts integrity and
organization of the structural proteins. Dengue serotypes 1, 3 and 4 do not change
conformation as serotype 2166–168 when exposed to higher temperature. DENV-4 has stronger
interaction between its prME dimer contributing to the high thermostability of the serotype.

Recent amide hydrogen /deuterium exchange mass spectrometry studies on whole DENV
provided more insights on serotype-dependent conformation and compared serotypes 1 and
2169. Author tested 28°C, 37°C and 40°C. DENV-2 exhibited enlargement of the viral particle
when temperature increased from 28°C to 37°C. However, DENV-1 showed re-arrangement
when the temperature raised from 37°C to 40°C. Regarding the organization of the E protein,
the increase of temperature affects the intradimeric interface in DENV-2 whereas it affects
the interdimeric interface in DENV-1. Author reported the C-protein is also affected by the
temperature, which emphasize the interaction of C with the other structural protein and its
role in the integrity of the viral particle. Quaternary epitopes on the virion surface are both
temperature- and serotype-dependent. Dengue virus structural dynamics reinforce the use of
VLP for vaccine development above simpler vaccine strategies such as subunit vaccines.
 Chapter 2. Zika virus-like particle (VLP) based vaccine

I. ARTICLE
Introduction
Zika fever results from an infection with the Zika virus (ZIKV), which is transmitted to
humans by the bite of an infected Aedes mosquito (primarily A. aegypti). Zika virus was
isolated for the first time from a Rhesus monkey in the Zika Forest in Uganda in 1947
and later from humans in Africa in 1952 [1]. The ZIKV has been transmitted in Africa for
many years through a sylvatic cycle between mosquito vectors and nonhuman
primates, with occasional human infections [2]. In recent years, however, epidemics of
Zika have resulted from cycles of transmission between vectors and humans resulting in
the spread of disease beyond the African continent into French Polynesia and other
Pacific regions [3–5].
Since 2015, a dramatic spread of ZIKV that began in Brazil has taken place in South
America and the Caribbean Islands with the occurrence of sporadic cases in travelers
identified in the USA and Europe [6]. Currently autochthonous infections have been
reported in the continental US (Florida) [7]. Infection in most cases is asymptomatic or
produces a mild illness. However, contracting the virus during pregnancy is associated
with birth defects, primarily microcephaly (defective brain development) as well as eye
defects and hearing deficits in the infant [8]. Furthermore, an increase in cases of
Guillain-Barre syndrome has been observed following ZIKV infection [9]. The seriousness
of these disorders imposes a tremendous burden on public health. In addition to vector
transmission, ZIKV is transmitted via sexual contact [10, 11], and by body fluid [12–14].
These facts taken together with its often-asymptomatic nature makes disease control
even more difficult.
Zika virus is a member of the flavivirus genus within the Flaviviridae family. This
family consists of a large group of enveloped viruses, which includes dengue, yellow
fever, West Nile, Japanese encephalitis and others that possess a single stranded RNA
genome of positive polarity, which serves as mRNA upon the infection of susceptible
cells. The ZIKV RNA genome encodes one open reading frame (ORF, ~10,272 nt) and
translates into a single polyprotein that similarly to other flaviviruses is co- and post-
translationally cleaved by cellular and virus-encoded proteases into three structural
proteins (C, prM and E) and into seven non-structural proteins (NS1, NS2A, NS2B, NS3,
NS4A, NS4B and NS5.) that enable virus replication (Fig 1A) [15, 16]. Flavivirus
replication and morphogenesis occurs in close association with intracellular
membranes. Nascent virions are assembled and transported through the secretory
pathway and released at the cell surface. Enveloped virions are composed of a cell-
derived lipid bilayer encapsulating the C-protein wrapped viral RNA genome and
studded with 180 copies of the proteins E and M. During maturation within the
secretory pathway the precursor prM protein is cleaved by the host cell furin protease
to produce the small M protein and the fragment pr, which is released upon virus egress
from the cell. The viral surface displays the E protein as the major antigenic determinant
of the virus and mediates receptor binding and fusion during virus entry. Therefore, this
protein is a major target for vaccine development [17].
At this time there is no approved vaccine or specific treatment available to control,
combat or prevent ZIKV infection. Prophylactic vaccination represents a critical unmet
need to address disease spread and its effects globally. Vaccine strategies can rely on
inactivated, live attenuated, viral vectors or DNA based compositions, and some of
these approaches have been recently tested for Zika vaccine development [18].
 However, safety concerns may limit their potential. Here we describe a novel
strategy to assemble and produce recombinant Zika viruslike particles (VLPs) and
demonstrate their effectiveness when used as a vaccine in an animal model.
Development of a safe and efficacious vaccine candidate will provide an effective tool to
address this public health emergency.

Fig 1. Zika polyprotein processing and strategy for VLP assembly. The schematic depicts the Zika virus genome
single open reading frame (ORF) that translates into a polyprotein comprising both structural and non-structural proteins,
which arise via several proteolytic cleavages. (A) In the early stages, the complex viral protease NS3 with its cofactor
NS2B (#) self-cleaves before cleaving the capsid protein. Host cell signalase is also involved in the maturation of the
polyprotein (^). As a final step, cellular furin cleaves (5) the pr portion from the M protein to uncover E protein fusion
peptide. (B) The VLP assembly strategy relies on the co-expression of the structural protein CprME (ZO2) together with the
non-structural protein NS2B/NS3Pro (ZO3). The viral NS3 protein is truncated maintaining only its N-terminal protease
domain NS3Pro, which is kept as a single transcription unit with its cofactor NS2B. The origin of the sequence encoding
these genes and the cloning strategy are described in Material and Methods.

https://doi.org/10.1371/journal.pntd.0005608.g001

Materials and methods

Genes and plasmids construct
The sequences encoding the structural and non-structural genes of the Zika virus (ZIKV)
were chemically synthesized by GeneArt (Life Technology) according to a specifically
designed and codon-optimized GenBank sequence: KJ776791.1, derived from Zika virus
strain H/PF/2013 genome. This viral genome sequence was the most contemporary
available at the time we initiated this work. The synthesized DNA fragments were
subcloned into the plasmid vector pcDNA3.4 (Life Technologies) utilizing the XbaI/EcoRV
restriction enzyme sites.
The non-structural genes of NS2B and viral protease NS3 were synthesized as a single
codon optimized unit and subcloned into vector pcDNA3.4 via NheI/AgeI.
Plasmids were amplified in MAX Efficiency Stbl2 Competent E. Coli Cells (Life
Technologies 10268–019) and purified using an EndoFree Plasmid Maxi Kit (Qiagen,
MD).
Virus, cells, and antibodies
Cultures of Vero cells (ATCC CCL-81) were maintained in VP-SFM media (Life
Technologies, CA) supplemented with 2 mM L-Glutamine, 2 mM GlutaMAX Supplement
(Life Technologies, CA), 1X Non-essential amino acids solution (Life Technologies, CA),
1X ITSE (InVitria) and 500 ng/ml rhEGF (Life Technologies, CA).
Zika viruses. ZIKV FSS 13025, a 2013 isolate, was obtained from The University
of Texas Medical Branch (UTMB) Arbovirus Reference Collection (# TVP 22623 UTMB,
Galveston, TX) and amplified in Vero cells following a low multiplicity of virus infection
(MOI: 0.01).
ZIKV MR-766 was obtained from the American Type Culture Collection (ATCC)
(#NR50065, Manassas VA) and also amplified in Vero cells at ~MOI: 0.01 and harvested
96 h post infection. The inactivated Zika virus control (In-ZIKV) was prepared using the
MR-766 strain. Purified virus (see below) was inactivated with formalin at the final
concentration of 0.01% and incubated at 37˚C for 24 h [19]. Work with Zika virus was
conducted under biosafety level 2 (BSL2) containment and procedures.
Expi293 (Life Technologies, CA) cells were expanded in Expi293 medium (Life
Technologies,) and transfected using Expifectamine following the manufacturer
instructions (Life Technologies, CA). Monoclonal antibody 4G2 is an in-house produced
and protein G purified preparation from hybridomas D1-4G2-4-15 culture supernatant
(ATCC HB-112). Monoclonal antibodies ZV-48, ZV-64 and ZV-67 were kindly provided by
Dr. M. S. Diamond [20]. Rabbit polyclonal antibodies, anti-M (GTX133305), anti-E
(GTX133314), anti-NS2B (GTX133308) and anti-NS3 (GTX133309) were purchased from
GeneTex, CA.
Horseradish peroxidase conjugated secondary antibodies anti-mouse and anti-rabbit
were purchased from Pierce Thermo Fisher, MA (#31430 and# 31460 respectively).
Alkaline-phosphatase conjugated secondary goat anti-mouse antibody was purchased
from Pierce Thermo Fisher, MA (#31330).
The human serum was obtained from Kerafast (#EVU302) Boston, MA. The laboratory
of
James E. Crowe Jr., MD, Vanderbilt University, collected this serum from an otherwise
healthy donor who contracted laboratory confirmed Zika virus infection in July 2015
following natural exposure to mosquitoes in Brazil.

VLP production and purification

Expi293 cells in suspension culture were transfected with a 1:2 ratio of pcDNA3.4-
NS2b/NS3
(ZO3) and pcDNA3.4-CprME (ZO2) at 37˚C in a 5% CO2 environment and agitated at
150 rpm. Transfected cells were harvested 72 h post-transfection and clarified via two
successive centrifugations. The first clarification was performed at 400g for 10 min at
4˚C followed by a second clarification at 10,000g for 10 min at 4˚C. The remaining
proteins in the supernatant were precipitated using 8% (w/v) of polyethylene glycol
8000 and incubated overnight at 4˚C. A protein pellet was collected after centrifugation
at 14,000g for 30 min at 4˚C and loaded onto a 20% w/v sucrose cushion in TNE buffer
(10 mM Tris-HCl, pH 8.0, 120 mM NaCl and 1 mM EDTA) and spun by ultracentrifugation
at 150,000g for 2 h at 4˚C. The pellet was resuspended in TNE buffer. Viruses and VLP
were then purified by ultracentrifugation through a linear potassium tartrate 10–35%
(w/v) / glycerol 7–30% (v/v) density gradient at 180,000g for 4 h at 4˚C. Fractions were
collected and analyzed by dot blot using 4G2 monoclonal mouse antibody. The 4G2
MAb reacts with dengue, Zika and other flavivirus [21] and has been shown to recognize
a conformational epitope within domain II of the dengue E protein [17] and presumably
recognizes a similar epitope in the Zika virus. Selected fractions were further purified
and concentrated using Amicon Ultra Centrifugal Filter Unit (Millipore, MA).

Dot blot, Western blot and Coomassie blue stain

The cell protein content was analyzed after clarification of transfected Expi293 cells. The
cell pellets were lysed with RIPA buffer (Pierce Thermo Fisher, MA) according to the
vendor protocol. For Western blotting, cell lysates and concentrated culture
supernatants were loaded onto a 4–12% Bis-Tris SDS-polyacrylamide gel (Life
Technologies, CA). After electrophoretic separation, proteins were electro-transferred
from the gel onto a 0.45 μm nitrocellulose membrane (Life Technologies LC2001). For
dot blot analysis, 3 μl of sample was applied to a 0.45 μm nitrocellulose membrane and
allowed to dry for 5 min. The nitrocellulose membranes were then blocked with 5%
non-fat milk in TBST (10 mM Tris-HCl, pH 7.4, 130 mM NaCl, 2.7 mM KCl and 0.1%
Tween-20) for 1 h at room temperature followed by overnight incubation at room
temperature in primary antibody diluted with blocking buffer. Membranes were
washed 3 times with 1X TBST and then incubated for 2 h with secondary antibody
diluted in blocking solution. Finally, membranes were washed 3 times with 1X TBST and
developed with ECL system (Life Technologies, CA).
The total protein concentration was determined using the Bradford method. The E
protein content in the purified VLPs and inactivated ZIKV (In-ZIKV) was determined using
densitometry analysis of Coomassie blue stained SDS-PAGE gels. Different
concentrations of a BSA standard were loaded onto the same gel. The stained gel image
was acquired with a FluorChem M imager instrument (Protein Simple, CA). The optical
density of the bands was analyzed using Alphaview software. The BSA concentrations
were used to establish standard curve and parameters of the linear regression curve
were used to determine E protein concentration of each VLP vaccine and In-ZIKV. Dot
blot were performed with purified and quantified VLP or ZIKV and the amount of sample
applied is specified in the corresponding figure legend.

Negative staining and immuno-gold labeling TEM

VLP or Zika virus samples were prepared for TEM examination as follow: Samples for
immunogold labeling TEM (2.5 μl) were loaded onto CF200-CU carbon film 200 mesh
copper grid (product of Electron Microscopy Science) and incubated for 5 min at room
temperature, then washed with PBS, blocked in 1% BSA in PBS for 5 min and incubated
on the surface of a primary antibody drop for 30 min at room temperature. We use two
antibodies, the mouse MAb 4G2 diluted 1:500 in PBS and a polyclonal human serum
from a Zika virus infected patient diluted 1:20 dilution in PBS. The grids were then
washed 6 times with PBS and incubated on the surface of a drop of secondary antibody
conjugated with gold beads for 30 min. A goat anti-mouse antibody conjugated with 10
nm gold beads was used with the mouse MAb 4G2 and a goat anti-human antibody
conjugated with 6 nm gold beads was used with the human serum. Grids were finally
washed 6 times with PBS, and then fixed for 15 min with 4% paraformaldehyde in PBS,
washed with PBS and subsequently washed twice with 0.2M Sodium cacodylate buffer
and finally stained with 1% uranyl acetate solution.
Mouse immunogenicity and efficacy study
The aim of the study was to determine the immunogenicity and efficacy of a VLP-based
Zika vaccine and its capacity to elicit a strong neutralizing serum protective immune
response. Experimental groups comprised eight mice (n = 8) in order to study pairs of
subjects while being able to reject a null hypothesis with statistical power of >99% and a
standard deviation of 19. The alpha (!) error probability associated with this test is 0.01.
Nine groups of 6 to 8-week old female BALB/c mice (n = 8) were inoculated twice
(day 0 and day 24) via the intramuscular (IM) route with either VLP vaccine, formalin
inactivated ZIKV MR-766 (In-ZIKV) control or PBS plus adjuvant negative control (Neg.
Ctr.). Mice received doses of either 1 μg or 4 μg of total E protein content formulated
alone or admixed in a 1:1 volume ratio with a squalene-based oil-in-water nano-
emulsion AddaVax (InvivoGen, CA). Serum samples were collected from all animals at
day 42.

Evaluation of serum antibody levels by ELISA

Serum IgG titers against Zika viruses were determined by ELISA. Briefly, assays were
performed in 96-well plates coated with 50 μl/well of purified and inactivated Zika virus
MR-766 or FSS-13025 (2 μg/ml total protein concentration) at 4˚C overnight.
Subsequently, plates were washed 3 times with PBS-T (0.05% Tween-20 in phosphate
buffered saline) and then blocked with 100 μl of blocking solution (5% non-fat milk in
PBS-T) for 1 h at room temperature (RT). Vaccine and control sera were diluted in
fourfold series in blocking buffer and applied to each well in triplicate and incubated for
2 h at RT. After 6 washes with PBS-T, plates were incubated for 2 h at RT with 50 μl of
HRP-conjugated goat anti-mouse IgG (H+L) secondary antibody diluted (1:5000) in
blocking buffer). Subsequently, plates were washed 6 times with PBS-T and developed
by adding 50 μl per well of Ultra TMB solution (Thermo Scientific, MA) for 20 min of
incubation and stopped with 100 μl of stop solution (2 M HCL solution). ELISA end-point
titer for each group was calculated as the reciprocal of the highest dilution with OD
value 2σ above the mean of the negative control wells.

Plaque reduction neutralization test (PRNT)

The plaque reduction neutralization test was carried out in Vero cells using sera from
immunized mice and either MR-766, or FSS-13025 Zika viruses following the method
described for the evaluation of flavivirus vaccine efficacy [22, 23]. Briefly, ZIKV was
amplified and titrated in Vero cells using the same plaque visualization procedure
described below. Vero cells were seeded in 24-well plates at a density of 6x104 cells per
well 24 h prior to the initiation of the test. Control and test sera were heat inactivated
at 56˚C for 30 min. Starting at a 1:20 dilution, the sera were two-fold serially diluted
with cell culture media supplemented with penicillin and streptomycin. Equal volumes
of diluted virus to form ~60 plaques per well were added to each serum dilution. The
sera-virus mixtures were incubated for 1 h at 37˚C in a 5% CO2 environment.
Afterwards, each dilution was applied in duplicated to wells of 85% confluent
monolayer of Vero cells and incubated for 90 min at 37˚C in a 5% CO 2 incubator.
Thereafter, the inocula were removed and a 1 ml overlay of 1.8% carboxymethyl
cellulose (CMC) in culture medium was added to each well. Plates were then incubated
for 3 days at which time the CMC overlay was removed, cells washed with PBST
(phosphate buffer saline plus 0.05% Tween 20) and fixed with cold 80% acetone for 10
min at -20˚C. Subsequently, plates were washed with PBST and then incubated with
blocking buffer (2.5% non-fat milk in 0.5% of Triton X-100 PBS solution) for 1 h in a 37˚C
incubator. After one wash, the primary antibody (MAb 4G2) diluted 1:500 in blocking
buffer was applied for 2 h at RT, followed by two PBST washes and incubation for 1 h at
RT with a goat anti-mouse AP conjugated secondary antibody diluted (1:2000) in
blocking buffer. Finally, plates were washed twice with PBST and once with alkaline
phosphate buffer (APB: 100 mM Tris-HCl pH9.0, 150 mM NaCl, 1 mM MgCl2). Viral
plaques were detected by adding the alkaline phosphate (AP) substrate nitro blue
tetrazolium chloride (NTB) and 5-brome-4-chlore-3-indolyl phosphate (BCIP) prepared
and used as follows: Combine 33 μl of NTB (50 mg/ml in 70% dimethylformamide) with
5 ml of APB mix well and then add 16.5 μl of BCIP (50 mg/ml in 100%
dimethylformamide) and the mixture should be used within 1 h of preparation. Plaques
were counted and PRNT50s were determined using the PROBIT method [24]. The
neutralization power calculated is expressed as the reciprocal of the highest serum
dilution that neutralizes 50% of the virus and the histograms represent the average
neutralization 50% per group and their standard deviations.

Antibody-dependent enhancement (ADE) assay of DENV-2

infection
Suspension cultures of U937 cells grown in RPMI supplemented with 10% of FBS were
used in the ADE assay, which was performed according to Diamond et. at. [25]. Prior to
the initiation of the experiment (72h), the U397 cells were differentiated toward the
granulocyte or macrophage lineage by the addition of dimethyl sulfoxide (DMSO;
1.25%) [25]. DENV-2 virus (50μl) was mixed with diluted serum samples (four for each
vaccine group) and controls (virus alone or 4G2 MAb-200ng) and then added to 2.5x105
cells resulting in a multiplicity of infection of 3 (MOI: 3). The resulting mixture of cells,
virus and serum was incubated for 2 h at 37˚C.
Subsequently, cells were washed four times to remove free virus, suspended in culture
medium and then incubated for 96 h at 37˚C. Thereafter, virus titers in the culture
supernatants were determined by plaque assay as described above in the PRNT assay.

Statistics
All statistical calculations were carried out using GraphPad Prism 4 software. Tests
between two groups used two-tailed Student’s t-test. The P-value was calculated for
<0.05 level of significance.

Ethical statement
Mouse studies were carried out in the Department of Comparative Medicine, Animal
Facility of the New York Medical College, Valhalla, NY. All animals were cared for in
compliance with the Guide for the Care and Use of Laboratory Animals and all
experiments were approved by the New York Medical College’s IACUC. Mice were
housed in an AAALAC-accredited facility.
Results
Strategy for the assembly of Zika virus-like particles (VLPs)
Early studies on flavivirus replication have demonstrated that together with complete
infectious virions, particles are also produced that lack the viral RNA genome, which
have been termed small, non infectious subviral particles (SVPs) or virus-like particles
(VLPs) [26]. Assembly of these non-infectious particles using recombinant methods has
been utilized to study protein function, morphogenesis and structure [27–29] and to
generate flaviviruses vaccines, which have proven to be immunogenic [30]. We have
found a new and effective strategy to assemble ZIKA VLPs and utilize them for vaccine
development. Formation of VLP is accomplished by the co-expression of the Zika virus
structural proteins CprME together with a truncated form of the protease NS3Pro linked
to its cofactor NS2B constituting the viral NS2B/NS3Pro protease complex (Fig 1B).
Transfection of suspension cultures of Expi-293 cells with plasmids expressing CprME
and NS2B/NS3Pro produces the structural polypeptide CprME, which undergoes
proteolytic cleavages mediated by cellular and the co-expressed NS2B/NS3Pro
proteases (Fig 1B). Expression of NS2B/NS3Pro and processing of the polyprotein CprME
was evaluated by Western blot analysis using lysates of cells transfected with a single
plasmid expressing CprME (ZO2) or a single plasmid expressing NS2B/NS3Pro (ZO3) or a
combination of both plasmids. This analysis showed that the protease NS2B/NS3Pro
was expressed and self-cleaved rendering NS2B (~14KDa) (Fig 2D) and NS3Pro (~19KDa)
(Fig 2C) when cells were transfected with the plasmid ZO3 alone (expressing
NS2B/NS3Pro) or together with ZO2 (expressing CprME) but not in cells only transfected
with the plasmid ZO2 (CprME) (Fig 2C and 2D). Analysis of a ZIKV infected Vero cell
lysate revealed the expression of the full-length NS3 protein (containing both the
protease and the helicase domains), which was absent from the lysate of uninfected
(mock) Vero cells (Fig 2C). Similarly, Western blot studies demonstrated that the
polyprotein CprME was expressed and processed yielding the expected proteins E and
prM, as well as intermediate cleavage products including prM and CprM (Fig 2A and 2B).
Release of prM and C from ZO2 alone seems to occur at a lower frequency in the
absence of the virus protease complex NS2B/NS3Pro. Therefore to detect M and pr in
ZO2 alone requires higher amount of loaded material than the one used (Fig 2B). We
were unable to detect the C protein in neither the ZIKV infected cell lysate nor the VLP
transfected lysates using the currently available anti-C antibody. However, the
detection of E and
Fig 2. Analysis of protein expression and processing in lysates of transfected cells by
Western blot. Cell lysates (30μg total protein) of Expi-HEK293 transfected with plasmids ZO2
(CprME), or ZO3 (NS2B/ NS3Pro) or both (ZO2/ZO3) were probed with anti-Zika protein specific
polyclonal antibodies (See Material and Methods). Zika virus (ZIKV) infected and uninfected (mock)
Vero cell were controls. (A) Anti-E specific antibody detected E protein expression and cleavage in
ZO2/ZO3 and ZO2 (lesser amount) cell lysates but not in that of ZO3. E was also detected in ZIKV
infected cells but not in the mock control. (B) The prM protein (~32KDa) was detected in ZO2/ZO3
transfected cells as well as in ZIKV infected cells, which also showed a small amount of M (~ 8KDa).
(C) NS3Pro (~19KDa) was detected in ZO3 and ZO2/ZO3 transfected cells whereas the full-length
protease/helicase NS3 (~69KDa) was detected in ZIKV infected cells. Neither was detected in ZO2
transfected cells or the mock control. (D) NS2B (~14KDa) was detected in ZO2 and ZO2/ZO3
transfected cells and ZIKV infected cells whereas ZO2 and mock transfected cells were negative.

https://doi.org/10.1371/journal.pntd.0005608.g002

prM suggests that C is cleaved as well. This analysis demonstrates protein expression as
well as the activity of NS2B/NS3Pro, which is able to self-cleave uncoupling NS2B and
NS3Pro before subsequently cutting of C from the polyprotein (Figs 1 and 2A–2D).
Further cleavage processing steps mediated by cellular signalase proteases separate E
from M and C from pr. The released E protein shows slightly greater size than the E of
Zika MR-766 virus infected Vero cells (Fig 2A), due to the lack in this ZIKV strain of four
amino acids in the E protein (153– 156) including the glycosylation site asparagine (N-
154), which are present in the E protein of ZIKV H/PF/2013 used for VLP production as
well as other ZIKV strains such as the more contemporary strain FSS-13025 used as
control [31]. At the time of immunization, only the viral strain MR-766 was available.

Production and characterization of Zika virus-like particles (VLPs)

The recombinant production of Zika VLPs was carried out in suspension cultures of
ExpiHEK293 cells following co-transfection of the plasmids ZO2 (CprME) and ZO3
(NS2B/NS3
Pro). The VLPs were harvested from the culture supernatant and purified as described in
Material and Methods. The Zika virus grown in Vero cells was subjected to an analogous
purification scheme as the one described for the VLPs. The purity and identity of the
protein component of the purified VLP and Zika virus were assessed by Coomassie blue
staining and

Fig 3. Analysis of purified VLPs and ZIKV by Coomassie blue stain and Western blot. A total of
1.8 μg of purified VLPs or ZIKV were resolved by SDS-PAGE and protein content and identity
analyzed by (A) Coomassie stain and (B) Western blot using Zika protein specific antibodies, anti-E
upper panel and anti-M lower panel (C) Comparison of the size of the E protein of ZIKV MR-766;
FSS-13025 and VLPs by Western blot.

https://doi.org/10.1371/journal.pntd.0005608.g003

Western blot analysis. As shown in Fig 3A, the E protein was readily detected as the
predominant component of both the ZIKV and the VLP preparations. The identity of the
E protein was verified by Western blot analysis and exactly correlated with the E protein
visualized in the stained gel (Fig 3B). The migration difference of the E proteins detected
in the cell lysates was also seen between the E protein of the purified ZIKV-MR-766 and
VLPs (Fig 3A and 3B). This disparity, as noted above, is due to the lack of four amino
acids including the glycosylation site N-154 in ZIKV-MR-766. These residues are present
in ZIKV-FSS-13025 and H/PF/ 2013 and sequences of this strain were used to produce
the VLPs. Examination by Western blot of the migration patterns of the E protein from
ZIKV-MR766, FSS-13025 and VLPs (genes derived from H/PF/2013) showed that the MR-
766 E protein migrated slightly faster than the E protein from ZIKV FSS-13025 and VLPs,
which exhibit similar size (Fig 3C). Coomassie staining also showed minor unidentified
proteins in the VLP preparation that were also present in the ZIKV albeit in lesser
amounts (Fig 3A). The M protein, a major structural component of the Zika virus, was
clearly identified by Western blot analysis in both the ZIKV and VLP preparations but it
was not clearly seen in the stained gel, possibly due to its small size and low
concentration, which precluded retention of sufficient amounts of bound dye for clear
detection (Fig 3A). Both, E and M the major structural proteins were clearly identified by
Western blot (Fig 3B).
Furthermore, to corroborate whether conformational neutralizing epitopes of the E
protein were exhibited on the VLPs, we probed by dot blot purified VLPs and ZIKVs with
the domain specific neutralizing MAbs, ZV-48, ZV-64 and ZV-67 [20] in addition to the
 mouse MAb 4G2. This examination showed that the VLPs were indeed recognized by
these MAbs indicating that these neutralizing epitopes were properly displayed in the
VLPs as they also were in the ZIKV MR-766 and FSS-13025 (Fig 4). We have also
observed that the reactivity of Zika viruses with these MAbs is mostly abrogated after
the viruses are inactivated with formaldehyde (Fig 4). The VLP sample remained
untreated.

Fig 4. Examination of reactivity of native VLPs and native and inactivated ZIKV with
monoclonal antibodies recognizing structural epitopes. VLPs were produced and purified as
described in material and methods. ZIKV MR-766 and FSS-13025 were grown in Vero cells, PEG-
8000 precipitated and concentrated by centrifugation through a 20% sucrose cushion. Half of the
virus sample was inactivated with formalin and then treated and untreated (native) virus samples
were further purified via density gradient centrifugation. Dot blot analysis of purified VLPs and ZIKVs
(0.2μg/μl, 3μl per dot of E protein content) were probed the MAbs, 4G2, ZV-48, ZV-64 and ZV-67,
which recognize conformational epitopes in distinct domains of the E protein, showed that VLPs and
native ZIKVs reacted with the MAbs, whereas the reactivity of formalin inactivated ZIKVs virus was
greatly reduced as compared to native VLPs. This outcome suggests that formalin has a deleterious
effect on the ZIKV conformational epitope recognized by these MAbs.

https://doi.org/10.1371/journal.pntd.0005608.g004

Negative staining and immunogold labeling electron microscopy

To further characterize the structure and surface composition of the VLPs, we examined
purified material by transmission electron microscopy (TEM). Negative staining studies
revealed that the VLPs are fairly homogeneous spherical structures with an average
diameter of 60 nm. The VLP surfaces appeared smooth without noticeable projection or
rough features. Comparative analysis with purified ZIKV showed that both VLP and virus
particles are similar in size, morphology and surface appearance (Fig 5A, 5B and 5E).
This is in agreement with recent reports of the Zika virus structure [31, 32]. To better
define the surface composition of the VLPs, we probed the recombinant particles by
immunogold labeling with two distinct antibodies, 4G2, a MAb that recognizes a
conformational loop in domain II of the E protein shared by some members of the
flavivirus family including Zika [21], and a polyclonal human serum from a Zika patient.
Subsequently these preparations were examined by negative staining TEM. These
studies demonstrated the reactivity of the VLPs with both antibodies (Fig 5C and 5D)
indicating not only that the E protein is displayed on the surface of the particles but that
it also maintains its native conformation based on reactivity with the MAb 4G2 (Fig 5C).

Fig 5. Electron micrographs of negative staining and immunogold labeling of VLPs and ZIKV.
Purified VLPs and ZIKV were examined by transmission electron microscopy (TEM). Panels A and B
show uranyl acetate negatively stained VLPs which exhibit particle sizes from 50nm to 65nm in
diameter (mean 60nm) and a structure that resembles the morphology and surface appearance of
wild type Zika virus which is shown in Panel E. Immunogold labeling with two antibodies, mouse anti-
E protein MAb 4G2 and a human serum from a Zika patient served as primary and counterstained
with anti-mouse or anti-human secondary antibody conjugated with gold beads of 6 nm and 10nm in
diameter, respectively. Both antibodies, 4G2 panel C and human serum panel D, bind to the particle
surfaces as revealed by the presence of gold beads. Panel F shows wild-type ZIKV probed with the
4G2 antibody, which also binds the virus surface as revealed by the detection of gold beads; in this
case goat anti-mouse secondary antibody conjugated with 10 nm gold bead was applied. These
studies demonstrate that a specific anti-E MAb and an anti-Zika polyclonal antibody reacts with the
VLP surfaces indicating that the major Zika surface antigen, the E glycoprotein, is indeed displayed
on the VLP surfaces.

https://doi.org/10.1371/journal.pntd.0005608.g005

Examination of purified ZIKV also showed equivalent reactivity with the 4G2 antibody
(Fig 5F) confirming that both the VLP and live ZIKV exhibit on their surfaces the
conformational site recognized by 4G2 (Fig 5C and 5F). We further probed VLPs with the
serum from a Zika patient and found that it also binds to the particles’ surface providing
addition evidence that the ZIKV major surface glycoprotein E is indeed present on the
VLP surface (Fig 5D).
VLP based Zika vaccine evaluation
We sought to assemble Zika VLPs with the main objective of developing a safe and
effective vaccine to stem the rapidly spreading Zika epidemic. To establish
immunogenicity and invitro efficacy of a VLP-based Zika vaccine, we designed a mouse
study to assess the performance of a VLP vaccine and to compare the outcome to an
equivalent formulation and dose of an inactivated Zika virus (In-ZIKV) control. Given the
urgency of advancing a Zika vaccine and the lack, at the onset of the study, of well-
established Zika animal models, we selected a mouse model to carry out this vaccine
evaluation. Two groups of 6 to 8- week old female Balb/ c mice (n = 8 each) were
immunized via the intramuscular route (IM) twice, three weeks apart, with two
difference doses, 1 μg and 4 μg of total E protein of VLP vaccine formulated with or
without adjuvant (AddaVax, InVivoGen) a squalene-oil-in water nano-emulsion that
stimulates a balanced immune response and has a formulation similar to MF59, a
licensed adjuvant in flu vaccines in Europe [33, 34]. Similar groups of mice (n = 8 each)
were immunized via the same route and schedule with a formalin inactivated Zika virus
(In-ZIKV) control at the doses of either 1 μg or 4 μg of total E protein content also
formulated with or without adjuvant. A negative control (Neg. Ctr.) group (n = 8)
received phosphate buffered saline (PBS) plus adjuvant following the same
immunization regimen as used for the vaccine groups. Three weeks after the booster
immunization, animals in all groups were terminally bled and sera samples prepared for
immunogenicity and efficacy evaluation.

Evaluation of immunogenicity by ELISA

We assessed the level of the total serum IgG response in all vaccinated animals by ELISA
using as antigens two Zika virus strains, one isolated in 1947 (Zika virus MR-766) and a
second more current strain isolated in Cambodia in 2013 (Zika virus, FSS 13025). The
ELISA results demonstrated that mice vaccinated with the low dose (1 μg) or high dose
(4 μg) of VLP vaccine stimulated the production of high levels of serum antibodies
against both Zika virus strains (Fig 6 Upper and Lower panels). This response was
enhanced when the VLP vaccines were
Fig 6. Evaluation of serum IgG response elicited by VLP and control vaccinations. Serum antibody (total
IgG) elicited by the Zika VLP vaccine (VLP), inactivated Zika virus (In-ZIKV) and negative control (Neg. Ctr.)
were measured by ELISA. Groups of BALB/c mice (n = 8) were immunized on days 0 and 21 with a VLP-based
Zika vaccine (VLP), or an inactivated Zika virus vaccine (In-ZIKV) at doses of 1ug or 4ug with or without
adjuvant. Blood samples were collected three weeks after the booster immunization and total serum IgG was
measured via ELISA using as antigens two different Zika viruses MR766 (Upper panel) and FSS-13025 (Lower
panel). Mice that received low dose (1μg alone or plus adjuvant) or high dose (4μg alone or plus adjuvant) of
either vaccine showed a strong antibody response to the MR-766 virus compared to the placebo group (Upper
panel). Mice showed high titers of IgG antibodies against FSS-13025 Zika virus as well (Lower panel). The
values shown represent geometric means of the reciprocal dilutions of each mouse serum and data point
standard deviations.

https://doi.org/10.1371/journal.pntd.0005608.g006

formulated with adjuvant. The 1 μg VLP vaccine plus adjuvant elicited a markedly
increased level of serum IgG as compared to VLP alone. This enhancement was even
more significant using the 4 μg VLP vaccine dose formulated with adjuvant. Similar
results were obtained with the two Zika virus antigens used in the ELISA studies (Fig 6
Upper and Lower panels). Mice vaccinated with the inactivated Zika virus (In-ZIKV)
produced a serum IgG response that was quite comparable to the one triggered by the
VLP vaccines.
Clearly, both the VLP and the inactivated Zika virus control compositions were
capable of eliciting a high antibody response against two Zika viruses. In contrast, the
negative control group did not demonstrate a specific IgG response against Zika.

Assessment of VLP vaccine neutralization efficacy

Elicitation of high titers of specific neutralizing antibodies has been found to correlate
with protection against several flaviviruses including yellow fever (YF) tick-borne
 encephalitis, dengue and Japanese encephalitis [35–37]. Although this has not yet been
established for Zika it is very likely that the presence of high titers of neutralizing
antibodies confer protection to infection. Therefore, we assessed the level of
neutralizing antibodies elicited by the VLP vaccine and controls using a plaque reduction
neutralization test (PRNT) with the two Zika viruses previously described. This
evaluation with the ZIKV MR-766 showed that the VLP vaccine at the lower dose (1 μg)
elicited neutralizing antibody titers significantly higher than the negative control group,
which corresponds to a PRNT50: 1085 for the vaccine versus PRNT50: <25 for the
negative control (Fig 7 Left panel and Table 1). This response was enhanced when
adjuvant was part of the formulation raising the PRNT50: 1301. In contrast, the
inactivated Zika virus

Fig 7. Serum neutralizing antibody responses by plaque reduction neutralization test (PRNT) against two Zika
viruses. The serum neutralizing activity was plotted as PRNT50. (Left Panel) PRNT with MR-766 showed that the
VLP vaccines elicited significantly higher neutralizing titers than equivalent composition of In-ZIKV vaccine. The VLP 1
μg plus adjuvant elicits higher titers than the corresponding In-ZIKV vaccine but the difference was not significant. The
VLP (4μg+ adj) showed titers 5 fold greater than the human serum and the 4 μg alone was within this range, however
statistical analysis was not feasible with only one human sample. (Right Panel) Similarly, PRNT with FSS-13025
showed that VLP vaccines exhibit stronger neutralization than the corresponding In-ZIKV vaccine. Neutralizing titers
elicited by the In-ZIKV vaccine against FSS-13025 were much lower than those seen with the MR-766 and this may be
due to the fact the In-ZIK vaccine was prepared with the MR-766 virus, which elicited a higher neutralizing titer toward
the analogous virus (MR-766) than to a more distant one (FSS-13025). Similarly, the human serum showed a higher
titer against FSS-13025 than MR-766, presumably because it resulted from a human infection in Brazil with a virus
antigenically related to the FSS-13025. P-Values indicates t-tests.

https://doi.org/10.1371/journal.pntd.0005608.g007

Table 1. Comparison of the neutralizing antibody responses (PRNT50) elicited by the VLP vaccine and controls. A summary of geometric
mean titers of PRNT50 GMT for comparison between the two ZIKV strains, MR-766 and FSS 13025 for both vaccine and virus control groups.
Neutralizing activity elicited by the VLP vaccines against the two viruses were comparable, except the VLP for 1μg alone, which had greater activity
against MR766 than that of FSS (a p=0.0033). The In-ZIK virus control 4 μg plus adjuvant also showed a difference between the two viruses
(bp=0.026). The statistical analysis was performed using t-test at confidence intervals of 95% with GraphPad Prism 4 software. Human serum
isolated from a patient who recovered from a Zika infection was used to validate PRNT technique. However, no statistical analysis can be
performed to compare the neutralization with two different Zika viruses due to the low sample size.
 Neutralizing Antibody Titers in VLP Vaccinated and Control Groups

Vaccine Dose (μg) Adjuvant PRNT50 MR766 PRNT50 FSS13025

Neg. Ctr. - PBS + ADJ 11 5

a
VLP Vaccine 1 No ADJ 1,085 357 a

1 ADJ 1,301 605

4 No ADJ 2,978 2,476

4 ADJ 20,854 20,826

In-ZIKV 1 No ADJ 79 57

1 ADJ 794 79

4 No ADJ 430 160

4 ADJ 1,408 b 177 b

Human serum from Zika patient, Brazil 2013 4,267 17,067

https://doi.org/10.1371/journal.pntd.0005608.t001

(In-ZIKV) control at equivalent dose and formulation stimulated much lower neutralizing
titers than the VLP vaccine reaching PRNT50: 79 for the 1 μg dose and PRNT50: 794 for
the 1 μg plus adjuvant formulation. Increasing the VLP vaccine dose to 4 μg with or
without adjuvant significantly raised the neutralizing antibody titers to PRNT50: 2978
and PRNT50: 20854, respectively. Including the adjuvant in the 4 μg VLP vaccine
formulation heightened neutralization titers seven fold. The In-ZIKV high dose (4 μg)
control showed improved titers when adjuvant was added PRNT50: 1408 versus
PRNT50: 430 for In-ZIKV alone, but neither formulation gives rise to the neutralization
activity attained with the VLP vaccine. All VLP vaccine formulations elicited statistically
significant higher neutralizing antibodies titers than the equivalent compositions of the
inactivated Zika virus vaccine with the exception of the 1 μg plus adjuvant dose, where
the VLP induced higher titers but the difference was not significant. Neutralization
analysis with the more contemporaneous ZIKV FSS-13025 resulted in a neutralization
pattern that resembled the one seen with ZIKV MR-766. The VLP vaccine showed
significantly higher neutralizing titers against ZIKV FSS-13025 than the In-ZIKV vaccine
and this increased with the rise in dose and the incorporation of adjuvant (Fig 7 Right
panel and Table 1). Furthermore, we used as a reference in the PRNT50 a human serum
from patient who had recovered from a Zika infection. This revealed high titers of
neutralizing antibody against the two Zika viruses tested; although the neutralizing
activity was 4-fold higher for ZIKV-FSS virus (PRNT50: 17,067 for FSS versus and PRNT50:
4267 for MR-766) (Fig 7 Left and Right panels and Table 1). Since the human serum was
obtained from a recently infected patient, it is likely that it better neutralized the more
contemporaneous ZIKV–FSS suggesting that differences in neutralizing epitopes may
exist between these two viruses. Given the fact that until now we have had access to
only one Zika infected human serum sample, we could not perform a more
comprehensive comparison with the VLP vaccine. Nonetheless, these results showed
that the VLP vaccine was superior in eliciting neutralizing antibody responses than
equivalent compositions of an In-ZIKV control (Fig 7 Left and Right panels and Table 1).
Considering that flavivirus infections may induce antibody-dependent enhancement of
infection (ADE) with heterologous/co-circulating viruses, we tested whether the
 antibody response elicited by the VLP or controls vaccination against ZIKV may play a
role in

Fig 8. Antibody dependent enhancement (ADE) of dengue virus infection by the anti-ZIKV
response elicited by vaccination. Serum from four VLP immunized mice (4ug +adj); inactivated ZIK
(4ug+adj) and negative control (PBS+adj) diluted 1/500 were mixed with DENV-2 virus and added to
U397 cells. The MAb 4G2 and virus alone were used as controls. After 96h incubation, DENV-2 virus
titers were measured on the culture supernatant by plaque assay. Error bars indicate standard
deviation of the four samples performed in duplicates. Asterisk (*) indicates that difference between
4G2 (positive control) and remaining samples is statistically significant (p< 0.05).

https://doi.org/10.1371/journal.pntd.0005608.g008

augmenting dengue virus infection. Here, we utilized an in vitro ADE assay [25] and
measured whether DENV-2 infection was enhanced in the presence of anti-ZIKV
antibodies. We used the MAb 4G2, which enhances dengue virus infections as positive
control and DENV-2 alone as reference. This study showed that the antibody response
to ZIKV elicited by VLP vaccination did not enhance DENV-2 infection (Fig 8). Nor did the
sera from the In-ZIKV and the negative control vaccination when compared to DENV-2
with 4G2 MAb, which serves as the positive control and significantly enhanced DENV-2
infection (Fig 8). These data suggest that the antibody response elicited by these ZIKV
vaccines at the dilution tested in the assay did not induce ADE of dengue 2-virus
infection.

Discussion
The ongoing Zika epidemic has already resulted in over 1500 cases of microcephaly in
Brazil and 15 cases in the continental United States born to mothers who had traveled
to affected areas [38, 39]. Many other countries in South America and the Caribbean
are also experiencing the effects of the Zika epidemics and some autochthonous
infections have been reported in the continental US in Florida. In addition to the
mosquito vector direct infection there appear to be other forms of transmission
including via sexual contact [11, 13] potentially increasing the risk of viral dissemination.
Developing a safe and effective vaccine to control and combat the spread of the Zika
virus is a high priority. Our work describes a new strategy to generate Zika virus-like
particles (VLPs), providing a potentially safe and effective platform for the rapid
production of a candidate for clinical development of a prophylactic Zika vaccine. The
single coexpression of the structural polyprotein CprME together with the non-
structural NS2B/ NS3Pro suffices for the processing of CprME leading to the self-
assembly and release of VLPs devoid of viral RNA. Purified VLPs resemble the wild type
Zika virus in size, morphology and antigenic composition offering a suitable strategy for
vaccine development. Several highly effective viral vaccines are based on the VLP
strategy [40, 41] providing rationale for the use of Zika VLP for vaccine development.
VLP vaccine production can be attained in suspension cultures of mammalian cells using
standard fermentation technology. Considering that VLPs are not infectious or able to
replicate, inactivation is not required better preserving protein structure and
conformational epitopes. Our results show that all VLP vaccine and In-ZIKV control
formulations elicited high titers of serum IgG antibodies. However these high levels of
Zika specific antibody did not correlate, in the In-ZIKV vaccine control, with induction of
high titers of neutralizing antibodies, which are the benchmark of protection. Most of
our VLP vaccine formulations stimulated neutralizing antibody titers that were
significantly higher than those induced by the In-ZIKV control. The discrepancy between
high ELISA titers and low neutralization response in the In-ZIKV control, as compared to
the VLP vaccine, may be attributed to the distortion of critical neutralizing epitopes as a
result of ZIKV inactivation, which may reduce the repertoire of neutralizing sites. We
have seen some of these effects where formalin inactivation of the Zika viruses
abrogated most of the reactivity with the MAbs 4G2, ZV-48, ZV-64 and ZV-67, which
recognize conformational epitopes in the fusion loop of domain II and domain III of the
E protein (Fig 4). Therefore, the display of native and properly configured proteins on
the surfaces of the VLPs appears to exhibit to the immune system a better antigenic
array, which promotes the stimulation of a more diverse neutralizing response as we
have seen with these VLP vaccine formulations.
Suboptimal neutralizing antibody levels may contribute to antibody-enhancement of
disease in dengue, which by phylogenetic analysis is closer to Zika than other
flaviviruses. In fact, one study has shown that combining dengue cross-reacting
antibodies with Zika virus enhanced Zika infection in a cell culture assay [42]. In recent
years, quaternary epitopes identified on the surface of dengue have been recognized as
strong neutralizing sites [43–46]. Furthermore, a recent study has shown that
monoclonal antibodies recognizing these highly neutralizing epitopes of dengue have
also shown protection against Zika highlighting the significance of these types of
epitopes in providing protection against Zika virus infection [47]. Although the
correlates of protection for a Zika vaccine remain to be established, the elicitation of a
strong and diverse neutralizing response may greatly contribute to protection against
Zika infection. As shown in this study, the VLP based Zika vaccine elicited a robust
neutralizing antibody response suggesting that critical conformational epitopes were
indeed displayed for effective immune stimulation. In contrast to the response elicited
by our VLPs, the In-ZIKV produced a much-reduced neutralizing response, even though
the ELISA titers were as high as those induced by the VLP vaccine, suggesting that
important neutralizing epitopes were not properly displayed in the In-ZIKV composition.
 To investigate the possibility that the anti-ZIKV antibodies elicited by VLP vaccination
may partially bind to dengue virus and contribute to antibody-dependent enhancement
(ADE) of dengue infection, we tested diluted serum samples of VLP and control
immunized animals in an ADE in vitro assay. This evaluation demonstrated lack of
enhancement of DENV-2 infection when either the VLP vaccine or control
immunizations stimulated antibodies were tested. In contrast, the 4G2 MAb significantly
enhanced dengue infection, a property well established for this antibody [42]. This
outcome provides some indication that ZIKV VLP vaccination does not appear to
mediate ADE at least with DENV-2.
This study describes a safe, effective and straightforward strategy to rapidly produce
a Zika vaccine. VLPs are produced in mammalian suspension cell cultures offering a
suitable system for rapid scale up of manufacturing, without the risk of working with an
infectious agent. The lack of infectivity of the product eliminates the need for chemical
inactivation, which may compromise vaccine efficacy and safety. The particulate nature
of the vaccine and the preservation of a variety of conformational antigenic sites may
even render this vaccine efficacious in humans without using adjuvant if its
incorporation into a vaccine raises safety concerns. In summary, the Zika VLP platform
puts forward a vaccine composition and production system ready for clinical
development of a safe and effective prophylactic Zika vaccine, which is greatly needed
to meet the challenges imposed by the spread of the Zika epidemic.

Acknowledgments
We thank Walter Orenstein, MD and Katherine Frederick, MD for their encouragement
and critical reading of the manuscript. We also thank Sulli Popilskis, DVM, Director of
the New York Medical College Animal Facility and Kellie Elson and their team for their
support with the animal study. We thank Dr. Diamond for generously providing the
monoclonal antibodies FV-48, FV-64 and FV-67.
 References
1. Dick GW, Kitchen SF, Haddow AJ. Zika virus. I. Isolations and serological specificity. Trans R
Soc Trop Med Hyg. 1952; 46(5):509–20. PMID: 12995440
2. Duffy MR, Chen TH, Hancock WT, Powers AM, Kool JL, Lanciotti RS, et al. Zika virus outbreak
on Yap Island, Federated States of Micronesia. The New England journal of medicine. 2009;
360(24):2536–43. https://doi.org/10.1056/NEJMoa0805715 PMID: 19516034
3. Lanciotti RS, Kosoy OL, Laven JJ, Velez JO, Lambert AJ, Johnson AJ, et al. Genetic and
serologic properties of Zika virus associated with an epidemic, Yap State, Micronesia, 2007.
Emerging infectious diseases. 2008; 14(8):1232–9. PubMed Central PMCID:
PMCPMC2600394. https://doi.org/10.3201/ eid1408.080287 PMID: 18680646
4. Haddow AD, Schuh AJ, Yasuda CY, Kasper MR, Heang V, Huy R, et al. Genetic
characterization of
Zika virus strains: geographic expansion of the Asian lineage. PLoS neglected tropical diseases. 2012;
6(2):e1477. PubMed Central PMCID: PMCPMC3289602.
https://doi.org/10.1371/journal.pntd.0001477 PMID: 22389730
5. Faye O, Freire CC, Iamarino A, Faye O, de Oliveira JV, Diallo M, et al. Molecular evolution of
Zika virus during its emergence in the 20(th) century. PLoS neglected tropical diseases. 2014;
8(1):e2636. PubMed Central PMCID: PMCPMC3888466.
https://doi.org/10.1371/journal.pntd.0002636 PMID: 24421913
6. Zanluca C, de Melo VC, Mosimann AL, Dos Santos GI, Dos Santos CN, Luz K. First report of
autochthonous transmission of Zika virus in Brazil. Mem Inst Oswaldo Cruz. 2015; 110(4):569–
72. PubMed Central PMCID: PMCPMC4501423. https://doi.org/10.1590/0074-02760150192
PMID: 26061233
7. CDC. Zika virus disease in the United States, 2015–2016. 2016.
8. Valentine G, Marquez L, Pammi M. Zika Virus-Associated Microcephaly and Eye Lesions in the
Newborn. J Pediatric Infect Dis Soc. 2016.
9. Cao-Lormeau VM, Blake A, Mons S, Lastere S, Roche C, Vanhomwegen J, et al. Guillain-Barre
Syndrome outbreak associated with Zika virus infection in French Polynesia: a case-control
study. Lancet.
2016; 387(10027):1531–9. https://doi.org/10.1016/S0140-6736(16)00562-6 PMID: 26948433
10. Foy BD, Kobylinski KC, Chilson Foy JL, Blitvich BJ, Travassos da Rosa A, Haddow AD, et al.
Probable non-vector-borne transmission of Zika virus, Colorado, USA. Emerging infectious
diseases. 2011; 17
(5):880–2. PubMed Central PMCID: PMCPMC3321795.
https://doi.org/10.3201/eid1705.101939 PMID: 21529401
11. D Ortenzio E, Matheron S, Yazdanpanah Y, de Lamballerie X, Hubert B, Piorkowski G, et al.
Evidence of Sexual Transmission of Zika Virus. The New England journal of medicine. 2016;
374(22):2195–8. https://doi.org/10.1056/NEJMc1604449 PMID: 27074370
12. Musso D, Roche C, Nhan TX, Robin E, Teissier A, Cao-Lormeau VM. Detection of Zika virus in
saliva. J Clin Virol. 2015; 68:53–5. https://doi.org/10.1016/j.jcv.2015.04.021 PMID: 26071336
13. Fourcade C, Mansuy JM, Dutertre M, Delpech M, Marchou B, Delobel P, et al. Viral load
kinetics of Zika virus in plasma, urine and saliva in a couple returning from Martinique, French
West Indies. J Clin Virol.
2016; 82:1–4. https://doi.org/10.1016/j.jcv.2016.06.011 PMID: 27389909
14. Gourinat AC, O’Connor O, Calvez E, Goarant C, Dupont-Rouzeyrol M. Detection of Zika virus
in urine. Emerging infectious diseases. 2015; 21(1):84–6. PubMed Central PMCID:
PMCPMC4285245. https:// doi.org/10.3201/eid2101.140894 PMID: 25530324
15. Lindenbach BD, Murray C.L.,Thiel H-J., Rice C.M. Flaviviridae. Fields Virology, Sixth Edition,
Lippincott Williams & Wilkins. 2013;Vol. I, Chapter 25:713–46.
16. Baronti C, Piorkowski G, Charrel RN, Boubis L, Leparc-Goffart I, de Lamballerie X. Complete
coding sequence of zika virus from a French polynesia outbreak in 2013. Genome Announc.
2014; 2(3). PubMed Central PMCID: PMCPMC4047448.
17. Crill WD, Chang GJ. Localization and characterization of flavivirus envelope glycoprotein cross-
reactive epitopes. Journal of virology. 2004; 78(24):13975–86. PubMed Central PMCID:
PMCPMC533943. https://doi.org/10.1128/JVI.78.24.13975-13986.2004 PMID: 15564505
18. Larocca RA, Abbink P, Peron JP, Zanotto PM, Iampietro MJ, Badamchi-Zadeh A, et al. Vaccine
protection against Zika virus from Brazil. Nature. 2016.
19. Eckels KH, Putnak R. Formalin-inactivated whole virus and recombinant subunit flavivirus
vaccines. Advances in virus research. 2003; 61:395–418. PMID: 14714438
20. Zhao H, Fernandez E, Dowd KA, Speer SD, Platt DJ, Gorman MJ, et al. Structural Basis of
Zika VirusSpecific Antibody Protection. Cell. 2016; 166(4):1016–27. PubMed Central PMCID:
PMCPMC4983199. https://doi.org/10.1016/j.cell.2016.07.020 PMID: 27475895
21. Gentry MK, Henchal EA, McCown JM, Brandt WE, Dalrymple JM. Identification of distinct
antigenic determinants on dengue-2 virus using monoclonal antibodies. The American journal
of tropical medicine and hygiene. 1982; 31(3 Pt 1):548–55.
22. Timiryasova TM, Bonaparte MI, Luo P, Zedar R, Hu BT, Hildreth SW. Optimization and
validation of a plaque reduction neutralization test for the detection of neutralizing antibodies to
four serotypes of dengue virus used in support of dengue vaccine development. The American
journal of tropical medicine and hygiene. 2013; 88(5):962–70. PubMed Central PMCID:
PMCPMC3752766. https://doi.org/10. 4269/ajtmh.12-0461 PMID: 23458954
23. WHO. Guidelines for Plaque-Reduction Neutralization Testing of human antibodies to Dengue
viruses. 2007.
24. Finney DJ. Probit Analysis. Ed, Cambridge, England, Cambridge University Press. 1952.
25. Diamond MS, Edgil D, Roberts TG, Lu B, Harris E. Infection of human cells by dengue virus is
modulated by different cell types and viral strains. Journal of virology. 2000; 74(17):7814–23.
PubMed Central PMCID: PMCPMC112311. PMID: 10933688
26. Smith TJ, Brandt WE, Swanson JL, McCown JM, Buescher EL. Physical and biological
properties of dengue-2 virus and associated antigens. Journal of virology. 1970; 5(4):524–32.
PubMed Central PMCID: PMCPMC376035. PMID: 4195055
27. Schalich J, Allison SL, Stiasny K, Mandl CW, Kunz C, Heinz FX. Recombinant subviral particles
from tick-borne encephalitis virus are fusogenic and provide a model system for studying
flavivirus envelope glycoprotein functions. Journal of virology. 1996; 70(7):4549–57. PubMed
Central PMCID: PMCPMC190391. PMID: 8676481
28. Allison SL, Stadler K, Mandl CW, Kunz C, Heinz FX. Synthesis and secretion of recombinant
tick-borne encephalitis virus protein E in soluble and particulate form. Journal of virology. 1995;
69(9):5816–20. PubMed Central PMCID: PMCPMC189449. PMID: 7637027
29. Allison SL, Tao YJ, O’Riordain G, Mandl CW, Harrison SC, Heinz FX. Two distinct size classes
of immature and mature subviral particles from tick-borne encephalitis virus. Journal of virology.
2003; 77 (21):11357–66. PubMed Central PMCID: PMC229348.
https://doi.org/10.1128/JVI.77.21.1135711366.2003 PMID: 14557621
30. Suphatrakul A, Yasanga T, Keelapang P, Sriburi R, Roytrakul T, Pulmanausahakul R, et al.
Generation and preclinical immunogenicity study of dengue type 2 virus-like particles derived
from stably transfected mosquito cells. Vaccine. 2015; 33(42):5613–22.
https://doi.org/10.1016/j.vaccine.2015.08.090 PMID: 26382602
31. Sirohi D, Chen Z, Sun L, Klose T, Pierson TC, Rossmann MG, et al. The 3.8 A resolution cryo-
EM structure of Zika virus. Science. 2016; 352(6284):467–70. PubMed Central PMCID:
PMCPMC4845755. https://doi.org/10.1126/science.aaf5316 PMID: 27033547
32. Kostyuchenko VA, Lim EX, Zhang S, Fibriansah G, Ng TS, Ooi JS, et al. Structure of the
thermally stable Zika virus. Nature. 2016; 533(7603):425–8.
33. Khurana S, Verma N, Yewdell JW, Hilbert AK, Castellino F, Lattanzi M, et al. MF59 adjuvant
enhances diversity and affinity of antibody-mediated immune response to pandemic influenza
vaccines. Sci Transl Med. 2011; 3(85):85ra48. PubMed Central PMCID: PMCPMC3501657.
https://doi.org/10.1126/ scitranslmed.3002336 PMID: 21632986
34. O Hagan DT, Ott GS, Nest GV, Rappuoli R, Giudice GD. The history of MF59((R)) adjuvant: a
phoenix that arose from the ashes. Expert review of vaccines. 2013; 12(1):13–30.
https://doi.org/10.1586/erv. 12.140 PMID: 23256736
35. Belmusto-Worn VE, Sanchez JL, McCarthy K, Nichols R, Bautista CT, Magill AJ, et al.
Randomized, double-blind, phase III, pivotal field trial of the comparative immunogenicity,
safety, and tolerability of two yellow fever 17D vaccines (Arilvax and YF-VAX) in healthy infants
and children in Peru. The American journal of tropical medicine and hygiene. 2005; 72(2):189–
97. PMID: 15741556
36. Heinz FX, Holzmann H, Essl A, Kundi M. Field effectiveness of vaccination against tick-borne
encephalitis. Vaccine. 2007; 25(43):7559–67. https://doi.org/10.1016/j.vaccine.2007.08.024
PMID: 17869389
37. Katzelnick LC, Montoya M, Gresh L, Balmaseda A, Harris E. Neutralizing antibody titers against
dengue virus correlate with protection from symptomatic infection in a longitudinal cohort.
Proceedings of the National Academy of Sciences of the United States of America. 2016;
113(3):728–33. PubMed Central
38. PMCID: PMCPMC4725482. https://doi.org/10.1073/pnas.1522136113 PMID: 26729879
39. CDC. Outcomes of Pregnancies with Laboratory Evidence of Possible Zika Virus Infection in
the United States, 2016. 2016.
40. PAHO. Regional Zika Epidemiology Update (Americas) July, 29, 2016. 2016.
41. Joura EA, Leodolter S, Hernandez-Avila M, Wheeler CM, Perez G, Koutsky LA, et al.
Efficacy of a quadrivalent prophylactic human papillomavirus (types 6, 11, 16, and 18) L1 virus-
like-particle vaccine against high-grade vulval and vaginal lesions: a combined analysis of three
randomised clinical trials. Lancet. 2007; 369(9574):1693–702. https://doi.org/10.1016/S0140-
6736(07)60777-6 PMID: 17512854
42. Zhang J, Zhang XF, Huang SJ, Wu T, Hu YM, Wang ZZ, et al. Long-term efficacy of a
hepatitis E vaccine. The New England journal of medicine. 2015; 372(10):914–22.
https://doi.org/10.1056/ NEJMoa1406011 PMID: 25738667
43. Dejnirattisai W, Supasa P, Wongwiwat W, Rouvinski A, Barba-Spaeth G, Duangchinda
T, et al. Dengue virus sero-cross-reactivity drives antibody-dependent enhancement of infection
with zika virus. Nature immunology. 2016.
44. de Alwis R, Smith SA, Olivarez NP, Messer WB, Huynh JP, Wahala WM, et al.
Identification of human neutralizing antibodies that bind to complex epitopes on dengue virions.
Proceedings of the National Academy of Sciences of the United States of America. 2012;
109(19):7439–44. PubMed Central
PMCID: PMCPMC3358852. https://doi.org/10.1073/pnas.1200566109 PMID: 22499787
45. Rouvinski A, Guardado-Calvo P, Barba-Spaeth G, Duquerroy S, Vaney MC, Kikuti
CM, et al. Recognition determinants of broadly neutralizing human antibodies against dengue
viruses. Nature. 2015; 520
(7545):109–13. https://doi.org/10.1038/nature14130 PMID: 25581790
46. Fibriansah G, Tan JL, Smith SA, de Alwis R, Ng TS, Kostyuchenko VA, et al. A highly
potent human antibody neutralizes dengue virus serotype 3 by binding across three surface
proteins. Nat Commun.
2015; 6:6341. PubMed Central PMCID: PMCPMC4346626.
https://doi.org/10.1038/ncomms7341 PMID: 25698059
47. Gallichotte EN, Widman DG, Yount BL, Wahala WM, Durbin A, Whitehead S, et al. A
new quaternary structure epitope on dengue virus serotype 2 is the target of durable type-
specific neutralizing antibodies. MBio. 2015; 6(5):e01461–15. PubMed Central PMCID:
PMCPMC4620467. https://doi.org/10.1128/ mBio.01461-15 PMID: 26463165
48. Swanstrom JA, Plante JA, Plante KS, Young EF, McGowan E, Gallichotte EN, et al.
Dengue Virus Envelope Dimer Epitope Monoclonal Antibodies Isolated from Dengue Patients
Are Protective against Zika Virus. MBio. 2016; 7(4). PubMed Central PMCID:
PMCPMC4958264.
II. DISCUSSION

As this Zika VLP research project occurred concurrently to the outbreak, there was very limited
resources from biologicals to literature. The strategy was a straightforward translation of the
gene combination used for the dengue VLP project, but without point mutations. It was
expected that Zika VLP would be less complex in terms of assembly and secretion than dengue
VLP. Dengue virus was identified previously as peculiar from other flaviviruses with non-
optimal cleavage site sequence of the furin and retention signal in the EH-1 domain. Zika virus
also has one single serotype and was expected to be thermostable. This study supports this
expectation. The Zika VLP secretion yield was much higher than for dengue VLP. The
ultracentrifugation gradient was improved from sucrose to potassium tartrate and glycerol.
After concentration and purification, the particles per electron microscopy fields were far
more abundant in for Zika VLPs than dengue VLPs. The yield of production was greater for Zika
VLP and Dengue VLP as well.

a) Zika viral strains and glycosylation

In order to conduct the experiments, the sole ZIKV strain available at the beginning of the
project was the MR-766 isolate from 1947. This strain does not carry the N-glycosylation at N-
154 and has a four amino acid deletion in the 150-loop region. This characteristic is unique for
viruses belonging to the African lineage.

Viral E protein glycosylation plays an important role in infectivity with receptor attachment,
protein folding and assembly. Glycosylation also plays a role in immune evasion of the virus170.
Immune evasion is a strategy allowing the virus to evade or interfere with the host immune
system170. The N-154 glycosylation loop protects the fusion loop and contributes in preventing
premature fusion. Throughout its evolution, the African genotype of Zika virus underwent
recombination and lost its glycosylation as a potential adaptive response to Aedes
dalzieli vector171. The N-154 glycosylation has been identified to correlates with infectivity in
vivo in mice172. In Non-Human Primates (NHP), the virus lacking the N-154 glycosylation
reverted the deletion173 emphasizing the important of glycosylation in the fitness of the virus.
Zika virus of Asian lineage reverted to restoration of the four amino acids including the N-154
glycosylation site.

Because of time constraint, we conducted mice immunization using the African MR-766 strain.
Additional mice immunization experiments using the more recent FSSZIKV FSS 13025 strain
isolated in 2010 in Cambodia that presents the glycosylation would certainly have provide
complementary findings. Moreover, the Cambodian 2010 strain is closely related to the viral
strain responsible for the 2013 outbreak in French Polynesia and in 2015 in Brazil.
 b) Zika virus strains and neutralizing antibodies

VLP, like inactivated virus and subunit vaccine are all vaccine strategies presenting assembled
viral proteins to the immune system. Although viral inactivation is cost- and time efficient, it
is not a suitable option for viruses yielding a poor immune response laying the ground for re-
infection. This was tragically observed with RSV, when infants died during clinical trial of
enhanced infection174. Formalin inactivation causes crosslinking of viral protein and alters the
conformation. This altered conformation does not mimic adequately viral epitopes presented
during natural infection yielding non-neutralizing antibodies175. In our study, dot blot analysis
confirms inactivated ZIKV does not present structural epitopes while VLPs do. The
neutralization titer is lesser for inactivated virus comparing to VLP even though they
stimulated similar levels of IgG.

c) Antibody Dependent Enhancement and dengue virus

As described in the introduction Zika and dengue viruses circulate in overlapping endemic
areas since the 2015 outbreak. Dengue virus has been circulating for over a century in these
regions. Local populations have high level of seroconversion against DENV however they did
not have pre-existing immunity against ZIKV. The scenario of dengue Antibody-Dependant
Enhancement (ADE) caused by pre-existing ZIKV infection would only occur when children
born from Zika infected mother are exposed to DENV. The other scenario would be from a
traveler having received Zika vaccine prior to travel in dengue endemic area.

At the time of our study it was unknown whether cross-reactivity between DENV and ZIKV
could increase the incidence of ADE. Zika virus infection has decreased since its peak in 2015.
Antibody-Enhancement of disease occurs in DENV after a secondary infection with a different
serotype and it causes by low level of neutralizing antibodies. There is to this day no clinical
record of Zika-induced ADE.

In this present article, reviewers asked that we addressed the potential dengue ADE following
Zika immunization. The mouse sera containing high titers of neutralization antibodies were
added with DENV in cell culture. The monoclonal 4G2 antibody was used as positive control
as it is known to induce ADE in vitro. There was no ADE recorded. These results are in
contradiction with an in vivo study with Balb/c and C57BL/6Stat2-/- mice assessing the
immunogenicity of a nanoparticle vaccine with AG129 and C57BL/6Stat2-/- mice for ADE using
sub-lethal inoculation176. In these two mice models, neither DENV-2 nor ZIKV induced
enhancement of disease. In another study testing another particle-based ZIK vaccine, the
DENV-2 induced ADE was assessed in vitro in THP-1 cells. the vaccine elicited neutralizing titers
against ZIKV and did not trigger ADE in DENV infection177. Lastly, another published study
assessed DENV-2 ADE in vitro after Zika subunit vaccine immunization in Balb/c mice178. In this
vaccine preparation, author identified E monomer and E dimer population. The E-dimer
vaccine did not induce ADE with DEN-2 whereas E-monomer population did trigger ADE. This
last study is in alignment with our results in Balb/c mice and emphasize the importance of
tertiary and quaternary structure of the E protein immunogen.
 Part IV. Conclusion

This manuscript presented the preclinical evaluation of a dengue serotype 2 VLP-based

vaccine and a Zika VLP-based vaccine with in vitro and in vivo data. These two vaccines have
been developed by generating a respective Viral-Like Particle using a novel combination of
structural and non-structural protein genes. From the inception of the projects to the
finalization (Figure 12), the priority was to generate two VLP as structurally close as possible
to the wild-type virus.

Figure 12 : Strengths-Weaknesses-Opportunities-Threats (S.W.O.T.) analysis of the future of dengue and Zika VLP-based
vaccines

The future of a dengue and a Zika VLP-based vaccine relies on internal and external factors
which can benefit or impede their development:
Strengths

From 2011 and onwards, the findings in structural vaccinology, on the thermolability of viral
proteins, provide new insights and evidence of the relevance of antigen quaternary structure
in vaccine design. A recent example is the development of an RSV subunit vaccine based on
the pre-fusion F protein. A natural infection with RSV does not elicit lasting protection and
individuals can be re-infected multiple times through life. After 50 years of vaccine research
with multiple strategies, a promising candidate is finally reaching Phase III clinical trial. This
vaccine developed by Pfizer is structurally designed to maintain the F protein in pre-fusion
state. This vaccine is safer than the post-fusion protein, which is widely prevalent naturally
occurring state of the protein. This structural vaccine is so effective, it will be tested without
adjuvant in the last phase of the clinical trial.

Structural strategies have attracted a lot of attention. Indeed, a research team from the
university of Texas published after us a similar approach for the development of Zika VLP. They
also used a combination of genes expression CprME and NS2bNS3 and HEK293 cells179. In their
study mice were immunized with both prME-derived and CprME-derived VLPs. The VLPs
comprising the capsid protein induced statistically significantly higher neutralization titers
than capsid-free particles. The authors also used NS2b/NS3 to process the CprME. Following
this first publication, the same research team published a follow-up work where they describe
the establishment of a stable cell line expressing ZIKV CprME via transduction of a bicistronic
plasmid carrying CprME-ires-NS2b-NS3 showing elicitation of neutralization antibodies in mice
after immunization with the resulting VLPs180. Finally, the authors translated their platform to
two other flaviviruses JEV and YFV and one alphavirus CHIKV181. The tetravalent formulation
of the four VLPs induced high neutralization titers against the four viruses. This publication
confirms the use of CprME-NS2b/NS3 as platforms for a flavivirus VLP platform as I patented.
It is important to note, the authors of this study did not publish anything regarding VLP in
Dengue virus, which is very likely due to the complexity of the formation and secretion of the
VLPs. The strategy presented by the author would require more optimization to address the
low yield.

Opportunities

Zika and Dengue virus co-circulates in endemic areas. The potential ADE resulting from a
primary dengue infection with a secondary Zika infection, or vice-versa, has been clarified as
has to be addressed for vaccine purposes. A pentavalent protecting against the four dengue
serotypes and Zika virus would be ideal.

There has been contradictory evidence between animal models and clinical data. Balb/c mice
were used as animal model to generate preliminary data for these two articles. These mice
are not permissive to dengue nor Zika infection; however, the serum neutralization titers after
immunization with a vaccine candidate allows a preliminary evaluation of the candidate. Over
the past five years, there is contradictory evidence regarding the onset of ADE with a different
combination of primary and secondary infections with both dengue and Zika virus. Results
vary among the different models and samples.

The intricacies of the homotypic- heterotypic protection and enhancement of infections are
an on-going conversation within the scientific community. Several published studies have
explored dengue immunity effect in Zika virus infection. In a mice study using human sera, the
animal model does not correlate with the clinical observation182. Authors argue that in
humans, previous dengue infection can be protective against Zika symptomatic infection and
congenital Zika syndrome. In mice, however, ADE is observed. The timing between infection
and sampling of the human sera is also an influencing factor of the protection or enhancement
evaluation of heterotypic serum. Fully immunocompetent mice with existing immunity against
dengue virus presented ADE during pregnancy with Zika infection and with higher impact in
congenital zika syndrome in vertical transmission183. The opposite was observed in pregnant
immuno-deficient Stat2-/- mice, transfused with dengue immune serum and challenged with
Zika virus. Transfused human immune sera provided protection against infection. It decreased
the pathogenicity, the viral replication, and the inflammation in Zika virus and fetus 184.

Bridging animal and clinical data has been a challenge. Focusing on Zika virus infection, the in
vitro analysis of human immune samples from either symptomatic or asymptomatic cases and
revealed no generation of ADE when the serum was sampled within the next 18-days post
infection185. However, it generated ADE after the initial innate immune response. The
enhancement of diseases was recorded even 297 days after infection. The same sera were
used to infuse immune-deficient Ifnar1-/- C57BL/6 pregnant mice. When challenging the mice
with a lethal dose of Zika virus, the onset of ADE, and eventually the survival of the animals,
was dependent on the amount of perfused human sera.

The in vitro data using human tissues provides another avenue for preclinical testing.
Inoculation of human skin explants with human immune and naïve sera and challenged with
a virus load representing natural infection dose, the viral load was assessed as marker of ADE.
DEN-2 and DEN-3 serotypes were tested186. The DEN-3 infection was self-limiting however the
infection with DEN-2 induced recruitment of macrophages and dendritic cells in immune sera-
infused explants. This in vitro model has provided interesting insights showed heterotypic
immunity can enhance migration of dendritic cells and macrophage and enhance viral
replication for both Zika and dengue viruses. This model does not address a clinical outcome
but correlated with other in vitro and mouse studies.

In one Non-human primates study, ADE was induced in vitro of a ZIK infection in the presence
of DEN immune sera; however it does not translate with in vivo observation where no increase
of viral load was recorded187. In another study, the ADE was not induced in vitro and pre-
existing immunity against DEN was found to be actually protective and limit a ZIK infection188.

Exploration of clinical evidence leans towards the induction of ADE in DEN infection in
combination with pre-existing dengue immunity. Hospitalization records and viral load
analysis of Puerto Rico human sera showed that a primary DEN infection followed by a Zika
secondary infection did not result in ADE. In contrast, it did result in ADE with a secondary
infection with DEN-2 or DEN-3189.

In a Brazilian study, previous dengue or Zika infection can drive either ADE or neutralization
of other flaviviruses. These findings were obtained comparing clinical sample human sera
isolate with either Dengue seroconversion, Zika seroconversion, or an active Zika infection
combined with Dengue seroconversion190. The ADE assay was performed in vitro using
different strains as well as YFV17D chimeras. This study reports also variability between the
dengue serotypes. Serotype 1 and 4 immunity are cross-protective against ZIKV, whereas
serotype 2 and 3 are enhancing in vitro. However, another study carried out in Brazil have
shown no impact on Zika infection with pre-existing dengue immunity191. A third study
assessed the different factors of vertical transmission and birth defects in pregnant Brazilian
women in 2015 and 2016 in Brazil. No correlation was identified between pre-existing dengue
immunity and the severity of congenital Zika syndrome in vertical transmission 192. Finally, in
French Polynesia, preexisting dengue immunity does not correlate with the enhancement of
Zika infection either193.

In vivo data differ from pre-clinical mouse studies and in vitro assays. One consensus
observation was an increase in cytokine release in DEN-immune serum undergoing Zika
infection in both NHP and human. However, this increase was not statistically significant. A
detailed analysis of T-cell response could help determine a significant effect. These clinical
data are very encouraging as no ADE has been linked between sequential infections of the two
viruses.

Weaknesses

The main challenge for a dengue vaccine is to induce a well-balanced protection against all
the four serotypes. In order to achieve this goal, a vaccine must be able to generate a highly
neutralizing immunity in children and adults. These challenges have been faced by the
vaccines currently in clinical trials or commercialized.

In 2020, there are three dengue vaccines, currently in Phase III clinical trial or completed the
Phase III. These candidates are all LAV vaccines. Viruses were attenuated through different
processes and while they carry the same prM and E proteins, their backbone is either DEN-2
for Takeda’s candidate(DENVax), Yellow Fever for Sanofi’s candidate and finally either DEN1,
-3 or -4 for the Butantan/NIAID’s candidate(TV003,-)5. As discussed previously, Sanofi’s
candidate failed to elicit a robust and balanced immunity and actually increase serious adverse
events in naïve immunized children. The preclinical data already hinted poor results. Takeda’s
vaccine failed to provide protection to NHP since viremia was recorded in a challenge study
with DEN-1, and viral RNA was detected with DEN-1 and DEN-3. This vaccine was well
tolerated in Phase I clinical trial but could the preclinical data forecasts, once again, a failed
immunity after vaccination? Thankfully a schedule of prime-boost resulted in the elicitation of
broad neutralizing antibodies comparing to three doses, as seen in Denvaxia. The only sole
serotype providing complete protection was the serotype 2, which is the serotype of the
attenuated virus backbone. The NIAID vaccine, however, carries the non-structural proteins
of three serotypes. During a challenge study, volunteers were inoculated with the
heterologous DEN-2 and were protected. As mentioned earlier, interference can be a major
issue for a balanced immunity in dengue virus and live-attenuated vaccine. The NIAID
candidate has been showing the best results so far, raising questions regarding the importance
of non-structural proteins. Unfortunately, aside from live-attenuated virus and inactivated
virus technologies, there are only three candidates currently in clinical trial, which rely on
different technologies; the DNA vaccine from USAMRMC in Phase I and the Merk and Hawai
Biotech subunit vaccines both in Phase I. in order to limit the damages and obtain valuable
results in clinical studies, other vaccinal strategies have integrated a prime-boost vaccination
schedule with two different vaccines. The USAMRMC conducted a Phase I clinical trial using
an inactivated vaccine for prime immunization and was followed by a LAV vaccine for boost
immunization. None of the aforementioned strategies addressed the structural aspect of the
immunogen.

The balanced immunity is not of concern in the case of Zika virus vaccine since there is only
one serotype. During the early phase of vaccine development, the priority was given to
innovative strategies such as RNA vaccine that has been deployed over live attenuate one’s.
Moderna presented on their website the early results of the Phase I with a seroconversion
rate of 94.4% in seronegative patients and a 4-fold of the neutralization titer in seropositive
patients. The lack of transparency regarding the neutralization titers of the seronegative group
hint that the lack of pre-existing immunity might be a problem for Zika as well. The sole
collection of additional in vivo data and comparison with the different Zika vaccine candidates
will allow determining if the elicitation of vaccine robust protection is mutually exclusive from
patient immune status at the time of immunization.

A 2020 study highlighted the decrease of neutralizing titers in patients previously infected
over the course of seven years in French Polynesia and four years in Fiji analyzing pre- and
post-outbreak neutralizing titers. The seroprevalence for Zika virus decreased 1.5 years after
the outbreak; however, it did not decrease for Dengue virus. Neutralization assay from Fiji
sample showed a 2-log increase in their neutralization titers against Zika at the time of the
outbreak in 2013-2014 and a decrease of the same amplitude over 2015-2017. Some
previously seropositive individuals actually reverted. From 2015 to 2017, however, the
neutralization titers did not decrease over the three years following the concurrent outbreak
of DEN-3 in 2014. It is important to notice that the neutralization titer of Zika virus at the time
of infection was higher than of dengue, meaning even though Zika virus infection results in
high level neutralizing antibody titers, it does not generate a lasting immunity. The vaccine
design for Zika virus should account for a lasting protection.

Threats
The presented in vivo data of the two VLP-based vaccines are very encouraging. However, the
Dengvaxia experience has taught the scientific community that the preclinical and clinical
evaluation of a dengue vaccine, and potentially Zika vaccine, can be very complex. Retracing
the Dengvaxia development steps could help understand in vitro and preclinical requirements
to properly evaluate a dengue vaccine.

Firstly, the chimeric DEN-2 monovalent vaccine was tested in NHP194. The vaccine provided
elicitation of neutralizing titers at day 14 for the highest dose and at day 30 for all the dose
tested in NHP. The virus also provided protection in a challenge test even at low dose. This
study was of low statistical power and the FRNT50 test did not report the actual titer, but
instead the highest dilution step providing a 50% reduction. The tetravalent formulation was
then tested in NHP as well195. This time, the vaccine-induced neutralization titers were
compared to the titers induced by wild-type dengue virus. This live-attenuated virus did not
induce a balanced immunity against the four serotypes, but instead showed higher titers for
DEN-4 and lower titer for DEN-2. These early stages of development already rose concerns
about balancing the immune response in a tetravalent formulation as the vaccine relies on
live-attenuated virus. In a more powered study using inoculation of the dengue chimeras in
the cerebral cortex frontal lobe, the neutralization titers were imbalanced 196. The serotype
three generated low titers comparing to the other serotypes even though all the animals had
seroconverted. DEN-4 serotype produced the highest titers. These NHP data were
encouraging to follow a single dose immunization schedule which, later on, had to be modified
to a 3-dose schedule. At this early stage, author questioned this imbalance of the immune
response and hypothesized that viral interference was responsible for the imbalance197.

The monovalent DEN-2 live-attenuated vaccine was pushed to Phase I clinical trial testing198.
The vaccine was established to be safe and resulted in seroconversion of subjects. A follow-
up Phase I199 was conducted in the US with the tetravalent formulation which eventually
reached commercialization. This tetravalent Phase I should have raised concerns regarding
the immunogenicity and the efficacy of this vaccine. The 3-dose regimen was effective in NHP
to elicit high levels of neutralization titers whereas there was no improvement after the
second dose in human. The PRNT50 GMT titers for DEN-1,-2, -3, and -4 were respectively 67,
538, 122, and 154, revealing a very imbalanced response 4 weeks after the third
immunization200. Authors of the study suggested than three to four months might not be
optimal to induce a robust immune response in humans. The adjusted immunization schedule
was later used in Phase II and Phase III trial. A second and third Phase I trials were conducted
in a non -endemic country, Mexico201, and in an endemic country, the Philippines202. It was
the first time the vaccine was tested in children. In these two trials, the neutralization GMT
were low and imbalanced after the three immunization in children. Overall, the data published
does not address potential ADE and focuses on the seroconversion by mentioning a PRNT50
titer above ten but elude to address the low titers and imbalanced immune response, yet, the
vaccine was pushed forward with Phase II and III. These phases focused on efficacy in school
children.
The results of Phase II203, Phase IIb99 and Phase III98 showed an unbalanced response (Table 4)
with a lower efficacy for DEN-2.

Table 4 : Dengvaxia vaccine efficacy in percentage of each of the serotypes in clinical trial Phases II, IIb and III.

DEN-1 DEN-2 DEN-3 DEN-4

Phase II 55.6 9.2 75.3 100
Phase IIb 61.2 3.5 81.9 89.9
Phase III 50 35 78.4 75.3

As observed in preclinical studies, a low titer of neutralizing antibodies generates ADE in vitro
and in vivo. In the following years after Phase II-III clinical trials, seronegative vaccinated
children had a higher hospitalization rate during acute dengue infection. Some children
presented ADE with vascular leakage. Dr Halstead, a highly renowned scientist and doctor,
warned the scientific community about these clinical data and the long term impact of the
vaccine204. After refuting the concerns in 2016205, the vaccine manufacturer eventually
changed their recommendation and in 2017 to only administer the vaccine to seropositive
patients, which requires immunological screening prior vaccination.

This Live-attenuated virus-based vaccine-induced sensitization to ADE as well as an

unbalanced immune response. Sensitization to ADE occurs when neutralization titers are low.
The reported titers of NHP and clinical study are very low comparing to titers of a natural
infection, which should be used as minimal requirements since a primary infection with
dengue virus naturally sensitize to ADE. In the present study, I used the WHO standard in order
to assess the strength of my neutralization assay and also compare it with mouse titers. The
WHO standard had a PRNT50 titer of 228 and mouse sera ranged from 245 to 1517 at the
highest dose which is very promising considering Chimerivax was evaluated based on the
seroconversion discrimination, which corresponded to a PRNT50 titer above 10.

Secondly, the live-attenuated virus strategy is limited by viral interference when multivalent
formulations are required as they are in dengue. The proposed VLP approach is expected to
be a lot more balance for dengue as it has been observed for the nonavalent formulations of
the Gardasil vaccine.

In conclusion, dengue and Zika virus are very similar in their replication, organization,
transmission vector, and structure, yet the small differences among them lead to
neurotropism and sexual and vertical transmission for Zika, and ADE and hemorrhagic fever
for dengue virus.

There is currently no Zika virus approved for commercialization and the first and only
commercialized dengue virus vaccine is a live-attenuated virus and has seen its authorization
retracted in certain countries after recording an increase of hospitalization in seronegative
children at the time of the vaccination. The seropositivity of the patient is relevant in order to
be eligible for vaccination according to the WHO adjusted recommendations. This entails the
vaccine is not suitable to prevent primary infections. The most advanced Zika virus vaccine
utilizes a very different technology and is a modRNA vaccine. The publicly available data are
mentioning a seroconversion in naïve patients. An elicitation of high neutralizing titers was
seen in seropositive patients. The similarities between these two clinical outcomes indicate
there is still a lot of room for improvement.

An ideal vaccine for dengue virus would take into consideration the thermolability of the virus
and engage in a structural approach for the design of the immunogen. Regarding Zika, an ideal
vaccine would target a lasting and sustainable protection. For both of these vaccines, the
neutralization titer should the benchmark for efficacy over the seroconversion.
REFERENCES

1. WHO EMRO | Accelerating work to overcome the global impact of neglected tropical diseases:
a roadmap for implementation | NTD-infocus | Neglected tropical diseases.
http://www.emro.who.int/neglected-tropical-diseases/ntd-infocus/ntd-roadmap.html.
2. WHO | WHO roadmap inspires unprecedented support to defeat neglected tropical diseases.
WHO https://www.who.int/neglected_diseases/London_meeting_follow_up/en/.
3. Research-based pharma pledges on neglected tropical diseases.
https://www.thepharmaletter.com/article/research-based-pharma-pledges-on-neglected-
tropical-diseases.
4. WHO | Research & development for neglected tropical diseases. WHO
http://www.who.int/research-observatory/analyses/neglected_diseases/en/.
5. WHO | Global Strategy for dengue prevention and control, 2012–2020. WHO
https://www.who.int/denguecontrol/9789241504034/en/.
6. Howe, G. M. A world geography of human diseases. World Geogr. Hum. Dis. (1977).
7. Arnold, R. D. On the Dengue or Break-Bone Fever, as It Appeared at Savannah in the Summer
and Autumn of 1850. Edinb. Med. Surg. J. 78, 304–319 (1852).
8. Ashburn, P. M. & Craig, C. F. Experimental investigations regarding the etiology of dengue
fever. 1907. J. Infect. Dis. 189, 1747–1783; discussion 1744-1746 (2004).
9. Reed, W. & Agramonte, A. Landmark article. Feb 16, 1901: The etiology of yellow fever. An
additional note. By Walter Reed, Jas. Carroll and Aristides Agramonte. JAMA 250, 649–658
(1983).
10. Theiler, M. Studies on the Action of Yellow Fever Virus in Mice. Ann. Trop. Med. Parasitol. 24,
249–272 (1930).
11. Henchal, E. A. & Putnak, J. R. The dengue viruses. Clin. Microbiol. Rev. 3, 376–396 (1990).
12. Messina, J. P. et al. Global spread of dengue virus types: mapping the 70 year history. Trends
Microbiol. 22, 138–146 (2014).
13. Dengue: Practice Essentials, Background, Pathophysiology. (2019).
14. Sabin, A. B. & Schlesinger, R. W. Production of Immunity to Dengue with Virus Modified by
Propagation in Mice. Science 101, 640–642 (1945).
15. Frierson, J. G. The Yellow Fever Vaccine: A History. Yale J. Biol. Med. 83, 77–85 (2010).
16. Hotta, S. Experimental Studies on DengueI. Isolation, Identification and Modification of the
Virus. J. Infect. Dis. 90, 1–9 (1952).
17. Westaway, E. G. et al. Flaviviridae. Intervirology 24, 183–192 (1985).
18. Thomas, S. J. & Yoon, I.-K. A review of Dengvaxia®: development to deployment. Hum. Vaccines
Immunother. 15, 2295–2314 (2019).
19. Dick, G. W. A., Kitchen, S. F. & Haddow, A. J. Zika virus. I. Isolations and serological specificity.
Trans. R. Soc. Trop. Med. Hyg. 46, 509–520 (1952).
20. Kerr, J. A. Studies on certain viruses isolated in the tropics of Africa and South America;
immunological reactions as determined by cross complement-fixation tests. J. Immunol. Baltim.
Md 1950 68, 461–472 (1952).
21. Macnamara, F. N. Zika virus: a report on three cases of human infection during an epidemic of
jaundice in Nigeria. Trans. R. Soc. Trop. Med. Hyg. 48, 139–145 (1954).
22. WHO | Zika: the origin and spread of a mosquito-borne virus. WHO
http://www.who.int/bulletin/volumes/94/9/16-171082/en/.
23. aubry, A. et al. Serosurvey of dengue, Zika and other mosquito-borne viruses in French
Polynesia. in vol. 765 (2015).
24. Cao-Lormeau, V.-M. et al. Guillain-Barré Syndrome outbreak associated with Zika virus infection
in French Polynesia: a case-control study. Lancet Lond. Engl. 387, 1531–1539 (2016).
25. Musso, D. Zika Virus Transmission from French Polynesia to Brazil. Emerg. Infect. Dis. 21, 1887
(2015).
26. Metsky, H. C. et al. Zika virus evolution and spread in the Americas. Nature 546, 411–415
(2017).
27. Faria, N. R. et al. Zika virus in the Americas: Early epidemiological and genetic findings. Science
352, 345–349 (2016).
28. Bhatt, S. et al. The global distribution and burden of dengue. Nature 496, 504–507 (2013).
29. Shope, R. Global climate change and infectious diseases. Environ. Health Perspect. 96, 171–174
(1991).
30. Halstead, S. B. Mosquito-borne haemorrhagic fevers of South and South-East Asia. Bull. World
Health Organ. 35, 3–15 (1966).
31. Twiddy, S. S., Holmes, E. C. & Rambaut, A. Inferring the Rate and Time-Scale of Dengue Virus
Evolution. Mol. Biol. Evol. 20, 122–129 (2003).
32. WHO. Dengue and severe dengue. WHO
http://www.who.int/mediacentre/factsheets/fs117/en/.
33. W.H.O. Countries and territories with current or previous Zika virus transmission.
https://www.who.int/emergencies/diseases/zika/countries-with-zika-and-vectors-table.pdf
(2019).
34. Malavige, G., Fernando, S., Fernando, D. & Seneviratne, S. Dengue viral infections. Postgrad.
Med. J. 80, 588–601 (2004).
35. Musso, D. et al. Detection of Zika virus in saliva. J. Clin. Virol. Off. Publ. Pan Am. Soc. Clin. Virol.
68, 53–55 (2015).
36. Musso, D. et al. Potential Sexual Transmission of Zika Virus. Emerg. Infect. Dis. 21, 359–361
(2015).
37. Mansuy, J. M. et al. Zika virus in semen and spermatozoa. Lancet Infect. Dis. 16, 1106–1107
(2016).
38. Russell, K. et al. Male-to-Female Sexual Transmission of Zika Virus—United States, January–
April 2016. Clin. Infect. Dis. 64, 211–213 (2017).
39. Deckard, D. T. Male-to-Male Sexual Transmission of Zika Virus — Texas, January 2016. MMWR
Morb. Mortal. Wkly. Rep. 65, (2016).
40. Krow-Lucal, E. R. et al. Association and birth prevalence of microcephaly attributable to Zika
virus infection among infants in Paraíba, Brazil, in 2015–16: a case-control study. Lancet Child
Adolesc. Health 2, 205–213 (2018).
41. Moi, M. L. et al. Zika virus infection and microcephaly in Vietnam. Lancet Infect. Dis. 17, 805–
806 (2017).
42. Bui, T. C. et al. Zika virus outbreaks and treatment in pregnant women in Ho Chi Minh City,
Vietnam. Int. J. Gynecol. Obstet. 141, 390–391 (2018).
43. Wongsurawat, T. et al. Case of Microcephaly after Congenital Infection with Asian Lineage Zika
Virus, Thailand. Emerg. Infect. Dis. 24, 1758–1761 (2018).
44. Sassetti, M. et al. First case of confirmed congenital Zika syndrome in continental Africa. Trans.
R. Soc. Trop. Med. Hyg. 112, 458–462 (2018).
45. Clinical Presentation | Dengue | CDC. https://www.cdc.gov/dengue/healthcare-
providers/clinical-presentation.html (2019).
46. Mueller, C. G. & Cao-Lormeau, V.-M. Chapter 8 - Insect-Borne Viruses and Host Skin Interface.
in Skin and Arthropod Vectors (ed. Boulanger, N.) 275–292 (Academic Press, 2018).
doi:10.1016/B978-0-12-811436-0.00008-3.
47. Johnston, L. J., King, N. J. C. & Halliday, G. M. Langerhans Cells Migrate to Local Lymph Nodes
Following Cutaneous Infection with an Arbovirus. J. Invest. Dermatol. 114, 560–568 (2000).
48. Marovich, M. et al. Human dendritic cells as targets of dengue virus infection. J. Investig.
Dermatol. Symp. Proc. 6, 219–224 (2001).
49. Guzman, M. G. et al. Dengue: a continuing global threat. Nat. Rev. Microbiol. 8, S7–S16 (2010).
50. Rico-Hesse, R. Microevolution and virulence of dengue viruses. in Advances in Virus Research
vol. 59 315–341 (Academic Press, 2003).
51. Campos, J. L. S., Mongkolsapaya, J. & Screaton, G. R. The immune response against flaviviruses.
Nat. Immunol. 19, 1189–1198 (2018).
52. Katzelnick, L. C. et al. Antibody-dependent enhancement of severe dengue disease in humans.
Science 358, 929–932 (2017).
53. Aubry, M. et al. Zika Virus Seroprevalence, French Polynesia, 2014–2015. Emerg. Infect. Dis. 23,
669–672 (2017).
54. Karimi, O. et al. Thrombocytopenia and subcutaneous bleedings in a patient with Zika virus
infection. The Lancet 387, 939–940 (2016).
55. Carteaux, G. et al. Zika Virus Associated with Meningoencephalitis. N. Engl. J. Med. 374, 1595–
1596 (2016).
56. Swaminathan, S., Schlaberg, R., Lewis, J., Hanson, K. E. & Couturier, M. R. Fatal Zika Virus
Infection with Secondary Nonsexual Transmission. N. Engl. J. Med. 375, 1907–1909 (2016).
57. Pierson, T. C. & Diamond, M. S. The emergence of Zika virus and its new clinical syndromes.
Nature 560, 573–581 (2018).
58. Satterfield-Nash, A. Health and Development at Age 19–24 Months of 19 Children Who Were
Born with Microcephaly and Laboratory Evidence of Congenital Zika Virus Infection During the
2015 Zika Virus Outbreak — Brazil, 2017. MMWR Morb. Mortal. Wkly. Rep. 66, (2017).
59. Fokke, C. et al. Diagnosis of Guillain-Barré syndrome and validation of Brighton criteria. Brain J.
Neurol. 137, 33–43 (2014).
60. Tam, C. C. et al. Guillain-Barré Syndrome and Preceding Infection with Campylobacter,
Influenza and Epstein-Barr Virus in the General Practice Research Database. PLoS ONE 2,
(2007).
61. Lehmann, H. C., Hartung, H.-P., Kieseier, B. C. & Hughes, R. A. C. Guillain-Barré syndrome after
exposure to influenza virus. Lancet Infect. Dis. 10, 643–651 (2010).
62. Grimaldi-Bensouda, L. et al. Guillain-Barre syndrome, influenzalike illnesses, and influenza
vaccination during seasons with and without circulating A/H1N1 viruses. Am. J. Epidemiol. 174,
326–335 (2011).
63. Microcephaly. https://www.who.int/news-room/fact-sheets/detail/microcephaly.
64. W.H.O. Zika virus. https://www.who.int/news-room/fact-sheets/detail/zika-virus.
65. Moni, M. A. & Lio’, P. Genetic Profiling and Comorbidities of Zika Infection. J. Infect. Dis. 216,
703–712 (2017).
66. Flipse, J. & Smit, J. M. The Complexity of a Dengue Vaccine: A Review of the Human Antibody
Response. PLoS Negl. Trop. Dis. 9, e0003749 (2015).
67. Dejnirattisai, W. et al. Dengue virus sero-cross-reactivity drives antibody-dependent
enhancement of infection with zika virus. Nat. Immunol. 17, 1102–1108 (2016).
68. Yu, I.-M. et al. Structure of the Immature Dengue Virus at Low pH Primes Proteolytic
Maturation. Science 319, 1834–1837 (2008).
69. Perera, R. & Kuhn, R. J. Structural Proteomics of Dengue Virus. Curr. Opin. Microbiol. 11, 369–
377 (2008).
70. Junjhon, J. et al. Influence of pr-M Cleavage on the Heterogeneity of Extracellular Dengue Virus
Particles. J. Virol. 84, 8353–8358 (2010).
71. Ansarah-Sobrinho, C., Nelson, S., Jost, C. A., Whitehead, S. S. & Pierson, T. C. Temperature-
dependent production of pseudoinfectious dengue reporter virus particles by
complementation. Virology 381, 67–74 (2008).
72. Qing, M., Liu, W., Yuan, Z., Gu, F. & Shi, P. Y. A high-throughput assay using dengue-1 virus-like
particles for drug discovery. Antiviral Res. 86, 163–171 (2010).
73. Zhang, X. et al. Dengue structure differs at the temperatures of its human and mosquito hosts.
Proc. Natl. Acad. Sci. U. S. A. 110, 6795–6799 (2013).
74. Fibriansah, G. et al. Structural Changes in Dengue Virus When Exposed to a Temperature of
37?C. J. Virol. 87, 7585–7592 (2013).
75. Kostyuchenko, V. A. et al. Structure of the thermally stable Zika virus. Nature 533, 425–428
(2016).
76. Pindi, C. et al. Molecular Basis of Differential Stability and Temperature Sensitivity of ZIKA
versus Dengue Virus Protein Shells. Sci. Rep. 10, (2020).
77. Gallichotte, E. N. et al. CD-loop Extension in Zika Virus Envelope Protein Key for Stability and
Pathogenesis. J. Infect. Dis. 216, 1196–1204 (2017).
78. Crill, W. D. & Chang, G.-J. J. Localization and Characterization of Flavivirus Envelope
Glycoprotein Cross-Reactive Epitopes. J. Virol. 78, 13975–13986 (2004).
79. Gromowski, G. D., Barrett, N. D. & Barrett, A. D. T. Characterization of Dengue Virus Complex-
Specific Neutralizing Epitopes on Envelope Protein Domain III of Dengue 2 Virus. J. Virol. 82,
8828–8837 (2008).
80. Lin, H.-E. et al. Analysis of Epitopes on Dengue Virus Envelope Protein Recognized by
Monoclonal Antibodies and Polyclonal Human Sera by a High Throughput Assay. PLoS Negl.
Trop. Dis. 6, (2012).
81. Wahala, W. M. P. B., Huang, C., Butrapet, S., White, L. J. & de Silva, A. M. Recombinant dengue
type 2 viruses with altered e protein domain III epitopes are efficiently neutralized by human
immune sera. J. Virol. 86, 4019–4023 (2012).
82. de Alwis, R. et al. Identification of human neutralizing antibodies that bind to complex epitopes
on dengue virions. Proc. Natl. Acad. Sci. U. S. A. 109, 7439–7444 (2012).
83. Lok, S.-M. et al. Binding of a neutralizing antibody to dengue virus alters the arrangement of
surface glycoproteins. Nat. Struct. Mol. Biol. 15, 312–317 (2008).
84. Chan, A. H. Y. et al. A Human PrM Antibody That Recognizes a Novel Cryptic Epitope on Dengue
E Glycoprotein. PloS One 7, e33451 (2012).
85. Rouvinski, A. et al. Recognition determinants of broadly neutralizing human antibodies against
dengue viruses. Nature (2015) doi:10.1038/nature14130.
86. Teoh, E. P. et al. The structural basis for serotype-specific neutralization of dengue virus by a
human antibody. Sci. Transl. Med. 4, 139ra83 (2012).
87. Kanesa-thasan, N. et al. Safety and immunogenicity of attenuated dengue virus vaccines
(Aventis Pasteur) in human volunteers. Vaccine 19, 3179–3188 (2001).
88. Osorio, J. E., Huang, C. Y.-H., Kinney, R. M. & Stinchcomb, D. T. Development of DENVax: a
chimeric dengue-2 PDK-53-based tetravalent vaccine for protection against dengue fever.
Vaccine 29, 7251–7260 (2011).
89. Rupp, R. et al. Safety and immunogenicity of different doses and schedules of a live attenuated
tetravalent dengue vaccine (TDV) in healthy adults: A Phase 1b randomized study. Vaccine 33,
6351–6359 (2015).
90. Sirivichayakul, C. et al. Safety and Immunogenicity of a Tetravalent Dengue Vaccine Candidate
in Healthy Children and Adults in Dengue-Endemic Regions: A Randomized, Placebo-Controlled
Phase 2 Study. J. Infect. Dis. 213, 1562–1572 (2016).
91. Immunogenicity and Safety of Tetravalent Dengue Vaccine (TDV) Administered With a Yellow
Fever Vaccine in Adults ClinicalTrials.gov. https://clinicaltrials.gov/ct2/show/NCT03342898
(2017).
92. Kallas, E. G. et al. Safety and immunogenicity of the tetravalent, live-attenuated dengue vaccine
Butantan-DV in adults in Brazil: a two-step, double-blind, randomised placebo-controlled phase
2 trial. Lancet Infect. Dis. (2020) doi:10.1016/S1473-3099(20)30023-2.
93. Whitehead, S. S. et al. In a randomized trial, the live attenuated tetravalent dengue vaccine
TV003 is well-tolerated and highly immunogenic in subjects with flavivirus exposure prior to
vaccination. PLoS Negl. Trop. Dis. 11, e0005584 (2017).
94. Durbin, A. P. et al. A 12-Month-Interval Dosing Study in Adults Indicates That a Single Dose of
the National Institute of Allergy and Infectious Diseases Tetravalent Dengue Vaccine Induces a
Robust Neutralizing Antibody Response. J. Infect. Dis. 214, 832–835 (2016).
95. Kirkpatrick, B. D. et al. The live attenuated dengue vaccine TV003 elicits complete protection
against dengue in a human challenge model. Sci. Transl. Med. 8, 330ra36 (2016).
96. Kirkpatrick, B. D. et al. Robust and Balanced Immune Responses to All 4 Dengue Virus Serotypes
Following Administration of a Single Dose of a Live Attenuated Tetravalent Dengue Vaccine to
Healthy, Flavivirus-Naive Adults. J. Infect. Dis. 212, 702–710 (2015).
97. Guy, B. et al. Preclinical and clinical development of YFV 17D-based chimeric vaccines against
dengue, West Nile and Japanese encephalitis viruses. Vaccine 28, 632–649 (2010).
98. Capeding, M. R. et al. Clinical efficacy and safety of a novel tetravalent dengue vaccine in
healthy children in Asia: a phase 3, randomised, observer-masked, placebo-controlled trial. The
Lancet (2014) doi:10.1016/S0140-6736(14)61060-6.
99. Sabchareon, A. et al. Protective efficacy of the recombinant, live-attenuated, CYD tetravalent
dengue vaccine in Thai schoolchildren: a randomised, controlled phase 2b trial. The Lancet 380,
1559–1567 (2012).
100. Hadinegoro, S. R. et al. Efficacy and Long-Term Safety of a Dengue Vaccine in Regions of
Endemic Disease. N. Engl. J. Med. 373, 1195–1206 (2015).
101. Arredondo-García, J. L. et al. Four-year safety follow-up of the tetravalent dengue vaccine
efficacy randomized controlled trials in Asia and Latin America. Clin. Microbiol. Infect. Off. Publ.
Eur. Soc. Clin. Microbiol. Infect. Dis. 24, 755–763 (2018).
102. Martinez, L. J. et al. Safety and Immunogenicity of a Dengue Virus Serotype-1 Purified-
Inactivated Vaccine: Results of a Phase 1 Clinical Trial. Am. J. Trop. Med. Hyg. 93, 454–460
(2015).
103. Schmidt, A. C. et al. Phase 1 Randomized Study of a Tetravalent Dengue Purified Inactivated
Vaccine in Healthy Adults in the United States. Am. J. Trop. Med. Hyg. 96, 1325–1337 (2017).
104. Beckett, C. G. et al. Evaluation of a prototype dengue-1 DNA vaccine in a Phase 1 clinical trial.
Vaccine 29, 960–968 (2011).
105. Danko, J. R. et al. Safety and Immunogenicity of a Tetravalent Dengue DNA Vaccine
Administered with a Cationic Lipid-Based Adjuvant in a Phase 1 Clinical Trial. Am. J. Trop. Med.
Hyg. 98, 849–856 (2018).
106. Manoff, S. B. et al. Immunogenicity and safety of an investigational tetravalent recombinant
subunit vaccine for dengue: results of a Phase I randomized clinical trial in flavivirus-naïve
adults. Hum. Vaccines Immunother. 15, 2195–2204 (2019).
107. WHO | Zika virus vaccine product development. WHO
http://www.who.int/immunization/research/development/zika/en/.
108. Giel-Moloney, M. et al. Chimeric yellow fever 17D-Zika virus (ChimeriVax-Zika) as a live-
attenuated Zika virus vaccine. Sci. Rep. 8, 1–11 (2018).
109. Nürnberger, C., Bodmer, B. S., Fiedler, A. H., Gabriel, G. & Mühlebach, M. D. A Measles Virus-
Based Vaccine Candidate Mediates Protection against Zika Virus in an Allogeneic Mouse
Pregnancy Model. J. Virol. 93, (2019).
110. Modjarrad, K. et al. Preliminary aggregate safety and immunogenicity results from three trials
of a purified inactivated Zika virus vaccine candidate: phase 1, randomised, double-blind,
placebo-controlled clinical trials. Lancet Lond. Engl. 391, 563–571 (2018).
111. Tebas, P. et al. Safety and Immunogenicity of an Anti-Zika Virus DNA Vaccine - Preliminary
Report. N. Engl. J. Med. (2017) doi:10.1056/NEJMoa1708120.
112. Brady, J. N. & Consigli, R. A. Chromatographic separation of the polyoma virus proteins and
renaturation of the isolated VP1 major capsid protein. J. Virol. 27, 436–442 (1978).
113. Burrell, C. J., Mackay, P., Greenaway, P. J., Hofschneider, P. H. & Murray, K. Expression in
Escherichia coli of hepatitis B virus DNA sequences cloned in plasmid pBR322. Nature 279, 43–
47 (1979).
114. Valenzuela, P., Medina, A., Rutter, W. J., Ammerer, G. & Hall, B. D. Synthesis and assembly of
hepatitis B virus surface antigen particles in yeast. Nature 298, 347–350 (1982).
115. Cohen, B. J. & Richmond, J. E. Electron microscopy of hepatitis B core antigen synthesized in E.
coli. Nature 296, 677–679 (1982).
116. Lin, S.-C. et al. Heterologous prime-boost immunization regimens using adenovirus vector and
virus-like particles induce broadly neutralizing antibodies against H5N1 avian influenza viruses.
Biotechnol. J. 8, 1315–1322 (2013).
117. Bisht, H., Chugh, D. A., Raje, M., Swaminathan, S. S. & Khanna, N. Recombinant dengue virus
type 2 envelope/hepatitis B surface antigen hybrid protein expressed in Pichia pastoris can
function as a bivalent immunogen. J. Biotechnol. 99, 97–110 (2002).
118. Amexis, G. & Young, N. S. Parvovirus B19 Empty Capsids as Antigen Carriers for Presentation of
Antigenic Determinants of Dengue 2 Virus. J. Infect. Dis. 194, 790–794 (2006).
119. Jeong, H. & Seong, B. L. Exploiting virus-like particles as innovative vaccines against emerging
viral infections. J. Microbiol. Seoul Korea 55, 220–230 (2017).
120. Ott, J. J., Stevens, G. A., Groeger, J. & Wiersma, S. T. Global epidemiology of hepatitis B virus
infection: New estimates of age-specific HBsAg seroprevalence and endemicity. Vaccine 30,
2212–2219 (2012).
121. Ault, K. A. Effect of prophylactic human papillomavirus L1 virus-like-particle vaccine on risk of
cervical intraepithelial neoplasia grade 2, grade 3, and adenocarcinoma in situ: a combined
analysis of four randomised clinical trials. The Lancet 369, 1861–1868 (2007).
122. Muñoz, N. et al. Impact of Human Papillomavirus (HPV)-6/11/16/18 Vaccine on All HPV-
Associated Genital Diseases in Young Women. JNCI J. Natl. Cancer Inst. 102, 325–339 (2010).
123. Paavonen, J. et al. Efficacy of human papillomavirus (HPV)-16/18 AS04-adjuvanted vaccine
against cervical infection and precancer caused by oncogenic HPV types (PATRICIA): final
analysis of a double-blind, randomised study in young women. The Lancet 374, 301–314 (2009).
124. Zhu, F.-C. et al. Efficacy and safety of a recombinant hepatitis E vaccine in healthy adults: a
large-scale, randomised, double-blind placebo-controlled, phase 3 trial. The Lancet 376, 895–
902 (2010).
125. Konishi, E. & Fujii, A. Dengue type 2 virus subviral extracellular particles produced by a stably
transfected mammalian cell line and their evaluation for a subunit vaccine. Vaccine 20, 1058–
1067 (2002).
126. Konishi, E., Fujii, A. & Mason, P. W. Generation and Characterization of a Mammalian Cell Line
Continuously Expressing Japanese Encephalitis Virus Subviral Particles. J. Virol. 75, 2204–2212
(2001).
127. Zhang, S. et al. Vaccination with dengue virus-like particles induces humoral and cellular
immune responses in mice. Virol. J. 8, 333 (2011).
128. Wang, P.-G. et al. Efficient Assembly and Secretion of Recombinant Subviral Particles of the
Four Dengue Serotypes Using Native prM and E Proteins. PLoS ONE 4, e8325 (2009).
129. Sugrue, R. J., Fu, J., Howe, J. & Chan, Y. C. Expression of the dengue virus structural proteins in
Pichia pastoris leads to the generation of virus-like particles. J. Gen. Virol. 78 ( Pt 8), 1861–1866
(1997).
130. Hsieh, S.-C., Liu, I.-J., King, C.-C., Chang, G.-J. & Wang, W.-K. A strong endoplasmic reticulum
retention signal in the stem-anchor region of envelope glycoprotein of dengue virus type 2
affects the production of virus-like particles. Virology 374, 338–350 (2008).
131. Christian, E. A. et al. Atomic-level functional model of dengue virus Envelope protein infectivity.
Proc. Natl. Acad. Sci. 110, 18662–18667 (2013).
132. Modis, Y., Ogata, S., Clements, D. & Harrison, S. C. Structure of the dengue virus envelope
protein after membrane fusion. Nature 427, 313–319 (2004).
133. Li, L. et al. The flavivirus precursor membrane-envelope protein complex: structure and
maturation. Science 319, 1830–1834 (2008).
134. Zhang, W. et al. Visualization of membrane protein domains by cryo-electron microscopy of
dengue virus. Nat. Struct. Mol. Biol. 10, 907–912 (2003).
135. Henrich, S. et al. The crystal structure of the proprotein processing proteinase furin explains its
stringent specificity. Nat. Struct. Biol. 10, 520–526 (2003).
136. Junjhon, J. et al. Differential Modulation of prM Cleavage, Extracellular Particle Distribution,
and Virus Infectivity by Conserved Residues at Nonfurin Consensus Positions of the Dengue
Virus pr-M Junction. J. Virol. 82, 10776–10791 (2008).
137. Tsai, W.-Y. et al. C-Terminal Helical Domains of Dengue Virus Type 4 E Protein Affect the
Expression/Stability of prM Protein and Conformation of prM and E Proteins. PLoS ONE 7,
(2012).
138. Kuhn, R. J. et al. Structure of dengue virus: implications for flavivirus organization, maturation,
and fusion. Cell 108, 717–725 (2002).
139. Lin, Y.-J. & Wu, S.-C. Histidine at Residue 99 and the Transmembrane Region of the Precursor
Membrane prM Protein Are Important for the prM-E Heterodimeric Complex Formation of
Japanese Encephalitis Virus. J. Virol. 79, 8535–8544 (2005).
140. Jones, C. T. et al. Flavivirus Capsid Is a Dimeric Alpha-Helical Protein. J. Virol. 77, 7143–7149
(2003).
141. Falgout, B., Pethel, M., Zhang, Y. M. & Lai, C. J. Both nonstructural proteins NS2B and NS3 are
required for the proteolytic processing of dengue virus nonstructural proteins. J. Virol. 65,
2467–2475 (1991).
142. Li, H., Clum, S., You, S., Ebner, K. E. & Padmanabhan, R. The Serine Protease and RNA-
Stimulated Nucleoside Triphosphatase and RNA Helicase Functional Domains of Dengue Virus
Type 2 NS3 Converge within a Region of 20 Amino Acids. J. Virol. 73, 3108–3116 (1999).
143. Falgout, B., Miller, R. H. & Lai, C. J. Deletion analysis of dengue virus type 4 nonstructural
protein NS2B: identification of a domain required for NS2B-NS3 protease activity. J. Virol. 67,
2034–2042 (1993).
144. Yusof, R., Clum, S., Wetzel, M., Murthy, H. M. & Padmanabhan, R. Purified NS2B/NS3 serine
protease of dengue virus type 2 exhibits cofactor NS2B dependence for cleavage of substrates
with dibasic amino acids in vitro. J. Biol. Chem. 275, 9963–9969 (2000).
145. Wu, H.-C. et al. Identification of a dengue virus type 2 (DEN-2) serotype-specific B-cell epitope
and detection of DEN-2-immunized animal serum samples using an epitope-based peptide
antigen. J. Gen. Virol. 84, 2771–2779 (2003).
146. Chandramouli, S. et al. Serotype-specific structural differences in the protease-cofactor
complexes of the dengue virus family. J. Virol. 84, 3059–3067 (2010).
147. Lobigs, M. Flavivirus premembrane protein cleavage and spike heterodimer secretion require
the function of the viral proteinase NS3. Proc. Natl. Acad. Sci. U. S. A. 90, 6218–6222 (1993).
148. Amberg, S. M. & Rice, C. M. Mutagenesis of the NS2B-NS3-mediated cleavage site in the
flavivirus capsid protein demonstrates a requirement for coordinated processing. J. Virol. 73,
8083–8094 (1999).
149. Lobigs, M. & Lee, E. Inefficient Signalase Cleavage Promotes Efficient Nucleocapsid
Incorporation into Budding Flavivirus Membranes. J Virol 78, 178–186 (2004).
150. Heinz, F. X. & Allison, S. L. Flavivirus Structure and Membrane Fusion. in Advances in Virus
Research vol. Volume 59 63–97 (Academic Press, 2003).
151. Modis, Y., Ogata, S., Clements, D. & Harrison, S. C. A ligand-binding pocket in the dengue virus
envelope glycoprotein. Proc. Natl. Acad. Sci. U. S. A. 100, 6986–6991 (2003).
152. Mondotte, J. A., Lozach, P.-Y., Amara, A. & Gamarnik, A. V. Essential role of dengue virus
envelope protein N glycosylation at asparagine-67 during viral propagation. J. Virol. 81, 7136–
7148 (2007).
153. Pokidysheva, E. et al. Cryo-EM reconstruction of dengue virus in complex with the carbohydrate
recognition domain of DC-SIGN. Cell 124, 485–493 (2006).
154. Chang, G.-J. J. et al. Enhancing biosynthesis and secretion of premembrane and envelope
proteins by the chimeric plasmid of dengue virus type 2 and Japanese encephalitis virus.
Virology 306, 170–180 (2003).
155. Purdy, D. E. & Chang, G.-J. J. Secretion of noninfectious dengue virus-like particles and
identification of amino acids in the stem region involved in intracellular retention of envelope
protein. Virology 333, 239–250 (2005).
156. Salaemae, W., Junaid, M., Angsuthanasombat, C. & Katzenmeier, G. Structure-guided
mutagenesis of active site residues in the dengue virus two-component protease NS2B-NS3. J.
Biomed. Sci. 17, 68 (2010).
157. Rahayu, A. et al. Prevalence and Distribution of Dengue Virus in Aedes aegypti in Yogyakarta
City before Deployment of Wolbachia Infected Aedes aegypti. Int. J. Environ. Res. Public. Health
16, (2019).
158. Yung, C.-F. et al. Dengue Serotype-Specific Differences in Clinical Manifestation, Laboratory
Parameters and Risk of Severe Disease in Adults, Singapore. Am. J. Trop. Med. Hyg. 92, 999–
1005 (2015).
159. Henchal, E. A., Gentry, M. K., McCown, J. M. & Brandt, W. E. Dengue Virus-Specific and
Flavivirus Group Determinants Identified with Monoclonal Antibodies by Indirect
Immunofluorescence. Am. J. Trop. Med. Hyg. 31, 830–836 (1982).
160. Gromowski, G. D. & Barrett, A. D. T. Characterization of an antigenic site that contains a
dominant, type-specific neutralization determinant on the envelope protein domain III (ED3) of
dengue 2 virus. Virology 366, 349–360 (2007).
161. Sukupolvi-Petty, S. et al. Type- and subcomplex-specific neutralizing antibodies against domain
III of dengue virus type 2 envelope protein recognize adjacent epitopes. J. Virol. 81, 12816–
12826 (2007).
162. Dejnirattisai, W. et al. A new class of highly potent, broadly neutralizing antibodies isolated
from viremic patients infected with dengue virus. Nat. Immunol. 16, 170–177 (2015).
163. Zhao, H. et al. Structural Basis of Zika Virus-Specific Antibody Protection. Cell 166, 1016–1027
(2016).
164. Austin, S. K. et al. Structural Basis of Differential Neutralization of DENV-1 Genotypes by an
Antibody that Recognizes a Cryptic Epitope. PLOS Pathog. 8, e1002930 (2012).
165. Rana, J., Slon Campos, J. L., Poggianella, M. & Burrone, O. R. Dengue virus capsid anchor
modulates the efficiency of polyprotein processing and assembly of viral particles. J. Gen. Virol.
100, 1663–1673 (2019).
166. Fibriansah, G. et al. A highly potent human antibody neutralizes dengue virus serotype 3 by
binding across three surface proteins. Nat. Commun. 6, 6341 (2015).
167. Kostyuchenko, V. A., Chew, P. L., Ng, T.-S. & Lok, S.-M. Near-Atomic Resolution Cryo-Electron
Microscopic Structure of Dengue Serotype 4 Virus. J. Virol. 88, 477–482 (2014).
168. Fibriansah, G. et al. A potent anti-dengue human antibody preferentially recognizes the
conformation of E protein monomers assembled on the virus surface. EMBO Mol. Med. 6, 358–
371 (2014).
169. Lim, X.-X. et al. Conformational changes in intact dengue virus reveal serotype-specific
expansion. Nat. Commun. 8, 1–13 (2017).
170. Watanabe, Y., Bowden, T. A., Wilson, I. A. & Crispin, M. Exploitation of glycosylation in
enveloped virus pathobiology. Biochim. Biophys. Acta Gen. Subj. 1863, 1480–1497 (2019).
171. Faye, O. et al. Molecular Evolution of Zika Virus during Its Emergence in the 20th Century. PLoS
Negl. Trop. Dis. 8, (2014).
172. Carbaugh, D. L., Baric, R. S. & Lazear, H. M. Envelope Protein Glycosylation Mediates Zika Virus
Pathogenesis. J. Virol. 93, (2019).
173. Aliota, M. T. et al. Heterologous Protection against Asian Zika Virus Challenge in Rhesus
Macaques. PLoS Negl. Trop. Dis. 10, e0005168 (2016).
174. Kim, H. W. et al. RESPIRATORY SYNCYTIAL VIRUS DISEASE IN INFANTS DESPITE PRIOR
ADMINISTRATION OF ANTIGENIC INACTIVATED VACCINE. Am. J. Epidemiol. 89, 422–434 (1969).
175. Fan, Y.-C., Chiu, H.-C., Chen, L.-K., Chang, G.-J. J. & Chiou, S.-S. Formalin Inactivation of Japanese
Encephalitis Virus Vaccine Alters the Antigenicity and Immunogenicity of a Neutralization
Epitope in Envelope Protein Domain III. PLoS Negl. Trop. Dis. 9, (2015).
176. Shukla, R. et al. Zika virus envelope nanoparticle antibodies protect mice without risk of disease
enhancement. EBioMedicine 54, (2020).
177. Cabral-Miranda, G. et al. Zika Virus-Derived E-DIII Protein Displayed on Immunologically
Optimized VLPs Induces Neutralizing Antibodies without Causing Enhancement of Dengue Virus
Infection. Vaccines 7, (2019).
178. Slon-Campos, J. L. et al. A protective Zika virus E-dimer-based subunit vaccine engineered to
abrogate antibody-dependent enhancement of dengue infection. Nat. Immunol. 20, 1291–1298
(2019).
179. Garg, H., Sedano, M., Plata, G., Punke, E. B. & Joshi, A. Development of Virus-Like-Particle
Vaccine and Reporter Assay for Zika Virus. J. Virol. 91, (2017).
180. Garg, H., Mehmetoglu-Gurbuz, T., Ruddy, G. M. & Joshi, A. Capsid containing virus like particle
vaccine against Zika virus made from a stable cell line. Vaccine 37, 7123–7131 (2019).
181. Garg, H., Mehmetoglu-Gurbuz, T. & Joshi, A. Virus Like Particles (VLP) as multivalent vaccine
candidate against Chikungunya, Japanese Encephalitis, Yellow Fever and Zika Virus. Sci. Rep. 10,
(2020).
182. Marques, E. T. A. & Drexler, J. F. Complex Scenario of Homotypic and Heterotypic Zika Virus
Immune Enhancement. mBio 10, (2019).
183. Rathore, A. P. S., Saron, W. A. A., Lim, T., Jahan, N. & St. John, A. L. Maternal immunity and
antibodies to dengue virus promote infection and Zika virus–induced microcephaly in fetuses.
Sci. Adv. 5, (2019).
184. Brown, J. A. et al. Dengue Virus Immunity Increases Zika Virus-Induced Damage during
Pregnancy. Immunity 50, 751-762.e5 (2019).
185. Shim, B.-S. et al. Zika Virus-Immune Plasmas from Symptomatic and Asymptomatic Individuals
Enhance Zika Pathogenesis in Adult and Pregnant Mice. mBio 10, (2019).
186. Castanha, P. M. S. et al. Reciprocal immune enhancement of dengue and Zika virus infection in
human skin. JCI Insight 5, (2020).
187. McCracken, M. K. et al. Impact of prior flavivirus immunity on Zika virus infection in rhesus
macaques. PLoS Pathog. 13, (2017).
188. Pantoja, P. et al. Zika virus pathogenesis in rhesus macaques is unaffected by pre-existing
immunity to dengue virus. Nat. Commun. 8, (2017).
189. Santiago, G. A. et al. Prior Dengue Virus Infection Is Associated With Increased Viral Load in
Patients Infected With Dengue but Not Zika Virus. Open Forum Infect. Dis. 6, (2019).
190. Oliveira, R. A. et al. Previous dengue or Zika virus exposure can drive to infection enhancement
or neutralisation of other flaviviruses. Mem. Inst. Oswaldo Cruz 114, (2019).
191. Terzian, A. C. B. et al. Viral Load and Cytokine Response Profile Does Not Support Antibody-
Dependent Enhancement in Dengue-Primed Zika Virus–Infected Patients. Clin. Infect. Dis. Off.
Publ. Infect. Dis. Soc. Am. 65, 1260 (2017).
192. Halai, U.-A. et al. Maternal Zika Virus Disease Severity, Virus Load, Prior Dengue Antibodies, and
Their Relationship to Birth Outcomes. Clin. Infect. Dis. Off. Publ. Infect. Dis. Soc. Am. 65, 877
(2017).
193. Cao-Lormeau, V. et al. Guillain-Barré Syndrome outbreak caused by ZIKA virus infection in
French Polynesia. Lancet Lond. Engl. 387, 1531–1539 (2016).
194. Guirakhoo, F. et al. Recombinant Chimeric Yellow Fever-Dengue Type 2 Virus Is Immunogenic
and Protective in Nonhuman Primates. J. Virol. 74, 5477–5485 (2000).
195. Guirakhoo, F. et al. Viremia and Immunogenicity in Nonhuman Primates of a Tetravalent Yellow
Fever–Dengue Chimeric Vaccine: Genetic Reconstructions, Dose Adjustment, and Antibody
Responses against Wild-type Dengue Virus Isolates. Virology 298, 146–159 (2002).
196. Guirakhoo, F. et al. Safety and Efficacy of Chimeric Yellow Fever-Dengue Virus Tetravalent
Vaccine Formulations in Nonhuman Primates. J. Virol. 78, 4761–4775 (2004).
197. Guy, B. et al. Evaluation of Interferences between Dengue Vaccine Serotypes in a Monkey
Model. Am. J. Trop. Med. Hyg. 80, 302–311 (2009).
198. Guirakhoo, F. et al. Live Attenuated Chimeric Yellow Fever Dengue Type 2 (ChimeriVaxTM-DEN2)
Vaccine: Phase I Clinical Trial for Safety and Immunogenicity: Effect of Yellow Fever Pre-
 immunity in Induction of Cross Neutralizing Antibody Responses to All. Hum. Vaccin. 2, 60–67
(2006).
199. Morrison, D. et al. A novel tetravalent dengue vaccine is well tolerated and immunogenic
against all 4 serotypes in flavivirus-naive adults. J. Infect. Dis. 201, 370–377 (2010).
200. Guy, B. et al. From research to phase III: Preclinical, industrial and clinical development of the
Sanofi Pasteur tetravalent dengue vaccine. Vaccine 29, 7229–7241 (2011).
201. Poo, J. et al. Live-attenuated Tetravalent Dengue Vaccine in Dengue-naïve Children,
Adolescents, and Adults in Mexico City: Randomized Controlled Phase 1 Trial of Safety and
Immunogenicity. Pediatr. Infect. Dis. J. 30, e9 (2011).
202. Capeding, R. Z. et al. Live-attenuated, tetravalent dengue vaccine in children, adolescents and
adults in a dengue endemic country: Randomized controlled phase I trial in the Philippines.
Vaccine 29, 3863–3872 (2011).
203. Villar, L. et al. Efficacy of a Tetravalent Dengue Vaccine in Children in Latin America. N. Engl. J.
Med. 372, 113–123 (2015).
204. Halstead, S. B. & Russell, P. K. Protective and immunological behavior of chimeric yellow fever
dengue vaccine. Vaccine 34, 1643–1647 (2016).
205. Hadinegoro, S. R. S. et al. Answer to the review from Halstead and Russell “Protective and
immunological behavior of chimeric yellow fever dengue vaccine” (DOI
10.1016/j.vaccine.2016.02.004). Vaccine 34, 4273–4274 (2016).
 RÉSUMÉ
Introduction
Chapitre 1 : Aspects généraux
La dengue menace 2.5 milliards d’individus au niveau mondial. Le virus de la dengue (DENV)
est transmis par les moustiques du genre Aedes. Ces moustiques sont aussi capables de
transmettre d’autres maladies virales dont la fièvre Zika. La dengue est connue pour son
impact sur la santé publique depuis près d’un siècle. En revanche, le virus du Zika (ZIKV) s’est
révélé seulement récemment avec les épidémies d’Océanie en 2013 et d’Amérique du Sud de
2015 à 2016.
En 2016, le premier vaccin contre la dengue, Dengvaxia de Sanofi, a reçu l’autorisation de mise
sur le marché dans 11 pays. Malheureusement ce vaccin engendre la « facilitation de
l'infection par des anticorps » (ADE) chez les enfants immunologiquement naïfs.
Ce manuscrit présente le développement préclinique réalisé entre 2011 et 2016 d’un vaccin à
particules pseudo-virales contre la dengue. Ce travail a servi par la suite de matrice pour le
développement d’un vaccin contre le Zika entre 2016 et 2017.
I. Epidémiologie de la dengue et du Zika
La dengue : Le nombre annuel d’infection est estimé à 390 millions comprenant 96 millions de
cas symptomatiques. Les régions endémiques correspondent aux zones tropicales favorables
à la survie du moustique Aedes. Le changement climatique a largement contribué à
l’expansion du vecteur. Il y a quatre sérotypes différents qui circulent en parallèle les uns des
autres.
Le Zika : A la suite de l’épidémie de 2015 au Brésil, le virus s’est répandu dans plusieurs pays
grâce aux vecteurs. D’après le WHO, le virus est présent dans 87 pays et territoires autour du
monde.
II. Cycles de transmission
Les deux espèces de moustiques responsables de la transmission de la dengue et du Zika sont
les moustiques Aedes Aegypti et Aedes Albopictus. La répartition géographique du moustique
représente un risque pour 50% de la population mondiale. Les transmissions non-vectorielles
sont aussi possibles par transfusions sanguines, sexuelles et verticales dans le cas du Zika.
III. Présentations cliniques
Les infections à la dengue et au Zika sont généralement asymptomatiques. Les cas
symptomatiques de dengue varient de relativement bénins (céphalées, douleurs oculaires et
articulaires, éruption cutanée) à des cas sévères avec une fièvre hémorragique. Après une
première infection, la protection homotypique est durable. En revanche lors d’une infection
hétérotypique secondaire, les anticorps non-neutralisants favorisent l’amplification du virus
dans les cellules immunitaires. Les deux cibles principales de la réponse humorale sont les
protéines prM et E. Lorsque des anticorps non- ou peu neutralisants s’attachent au virus, ils
sont ensuite intégrés par des cellules ayant le récepteur Fc-y tels que les monocytes et
macrophages.
Pour les cas les moins sérieux, les infections au Zika engendrent des symptômes très similaires
à la dengue. Les cas les plus sérieux peuvent engendrer méningite, encéphalite et
thrombopénie. En revanche, il peut y avoir des conséquences neurologiques de long terme
notamment par le syndrome Guillain-Barré chez les adultes et de retard de développement
cérébral chez les nouveau-nés tel que la microcéphalie.
Il n’existe aucun traitement contre DENV ni pour ZIKV mise à part la prise en charge des
Chapitre 2 : Caractéristiques des virus de la dengue and du Zika
Le corps principal de ce travail a commencé en 2011 et était focalisé uniquement sur le virus
de la dengue. En 2015, l’épidémie de Zika en Amérique du Sud fut l’opportunité d’appliquer
la stratégie à un autre Flavivirus. La dengue et le Zika, du genre Flavivirus dans la famille des
Flaviviridae., sont phylogénétiquement proches avec une homologie de la protéine
d’Enveloppe de 65-70% entre les quatre sérotypes de la dengue et 55-58% entre la dengue et
le Zika. En conséquent, ce chapitre se focalise sur les propriétés et caractéristiques des
protéines virales de dengue et se termine par une comparaison entre les effets de la
température sur les protéines structurales de la dengue et du Zika.
Le génome viral de ces deux virus est un RNA double brin à sens positif comportant un seul
cadre de lecture ouvert codant pour une seule polyprotéine composée des protéines
structurales à l’extrémité N-terminale : Capside, pré-membrane, enveloppes et les protéines
non-structurales à l’extrémité C-terminale: NS1 NS2A/B, NS3, NS4A/B and NS5. Les protéines
non-structurales sont responsables de la réplication virale tandis que les protéines
structurales forment le virion par l’encapside l’ARN viral dans une enveloppe composée d’une
bicouche lipidique incrustée de protéines E et prM. Ce dimère prME est organisées en chevron
d’homodimères donnant une structure icosaèdre à la particule virale.
Pour le virus de la dengue comme pour celui du Zika, une fois la polyprotéine exprimée, elle
est clivée dans le Réticulum Endoplasmique (ER) par la protéase virale et par des protéases
cellulaires avant d’être transportée dans l’appareil de Golgi. Le virus assemblé comporte 180
copies de la protéine E et de la protéine prM à sa surface. La partie « pr » couvre le peptide
de fusion de la protéine E prévenant ainsi la fusion prématurée de la particule virale jusqu’à
la sécrétion dans le milieu extracellulaire Les complexes prM-E s’associent deux par deux
formant des hétérodimères qui dépassent de la surface virale donnant un aspect rugueux au
virion. Lors du transport dans le réseau transgolgique, le pH de la vésicule décroît poussant la
réorganisation des dimères en 90 trimères qui sont aplatis sur le virion et changent la surface
en surface lisse. Le réarrangement des protéines rend le site de clivage de la furine accessible
séparant la portion pr de la portion M. Une fois la protéine virale secrétée, le changement de
pH vers un pH neutre permet à la portion pr de se détacher de l’Enveloppe La partie M de la
protéine prM sert de lien entre la partie pr et la membrane bi-lipidique.
La réplication des virus de la dengue et du Zika se réalise dans le vecteur à 28°C et chez
l’humain à 37°C. Les effets de la température sur la réplication et la structure virale ont été
initialement abordés uniquement avec le virus DENV-2, avant l’épidémie de Zika en 2016.
D’un point de vue moléculaire, des études de microscopie électronique sur des virions
produits en cellule C6/36 à 28°C et ensuite incubés entre 28°C et 37⁰C ont différents aspects
de surface. Les particules incubées à basse température ont un aspect lisse tandis que les
particules incubées au-dessus de 33°C étaient bosselées avec un diamètre supérieur de 10%,
dû au manque de contact entre les protéines d’enveloppe. La surface bosselée est différente
de la surface rugueuse des particules immatures. Ce changement de structure induit par
changement de température est appelé « respiration moléculaire ». La séroneutralisation par
réduction des plages de lyse (PRNT) révèle que les particules bosselées ont un taux d’infection
supérieur.
En contraste avec la particule virale de DENV-2, l’étude de microscopie électronique de
particule de Zika produite à 28°C et ensuite incubée à 37°C et 40°C montre une forte stabilité
structurelle avec conservation de l’aspect lisse de la surface lorsque la température augmente.
La labilité thermique de la dengue avec la respiration moléculaire des protéines structurelles
et principalement de l’Enveloppe met en évidence la difficulté de présenter un antigène
vaccinal pertinent pour stimuler une immunité protectrice. Des anticorps sont utilisés pour
mettre en avant les différences entre ces particules.
Les anticorps reconnaissant la dengue sont séparés en trois catégories :
- les anticorps réactifs au groupe des Flavivirus, qui peuvent reconnaitre des Flavivirus
différents,
- les anticorps réactifs croisés qui peuvent reconnaitre plusieurs sérotypes de la dengue
- et enfin les anticorps spécifiques de type qui ne reconnaissent qu’un des quatre sérotypes.
Les anticorps spécifiques de type sont généralement fortement neutralisants. A l’inverse, les
anticorps réactifs croisés sont généralement faiblement neutralisants. Il existe un équilibre
fragile entre l’efficacité de l’anticorps et sa spécificité.
Une étude récente a permis de mettre en évidence que les anticorps fortement neutralisants
ne reconnaissent pas uniquement le domaine III mais aussi un épitope cryptique couvrant la
charnière entre le domaine I et II ainsi que l’espace entre les deux dimères adjacents.
Ces anticorps fortement neutralisants sont capables de reconnaitre la protéine E uniquement
lorsque cette dernière est proprement repliée. Les deux hélices amphipathiques et les deux
domaines transmembranaires sont très importants pour le repliement de l’éctodomaine.
L’intégrité du domaine III de la protéine E n’est pas suffisant pour la neutralisation du virus
par des anticorps, le repliement global de l’enveloppe est primordial.
Le maintien de la structure quaternaire de la particule virale est important pour la vaccination.
Des anticorps neutralisants peuvent reconnaitre des épitopes qualifiés de cryptiques, c’est-à-
dire qu’ils ne sont pas présents à la surface du virion. En revanche, la reconnaissance est
possible après une incubation à 37°C. Il est important de noter que la température impacte la
reconnaissance du virus par des anticorps neutralisants mais n’affecte pas l’infectivité.
Un anticorps anti-prM est capable de reconnaitre un épitope cryptique à l’interface entre EI
et EIII. Le même anticorps fut capable de restaurer l’infectivité de particules immatures
confirmant que les anticorps spécifiques de type et par conséquent que le domaine III n’est
pas le seul épitope important pour la neutralisation virale.
Les « anticorps à réactivité croisée et fortement neutralisants » (bnAb) reconnaissent des
épitopes quaternaires lorsqu’ils sont complexés avec une protéine recombinante E produite à
28°C. L’épitope en question consiste en la boucle de fusion qui est composée de deux brins
du domaine II ainsi que de la boucle externe à l’acide amine 150 du domaine I adjacent de
l’autre protéine E composant le dimère et enfin des sites de glycosylation N153 et N-67. La
boucle de fusion et les sites de glycosylation sont communs aux quatre sérotypes.
D’autre études sont en accord avec cette dernière confirmant que la neutralisation du virus
dépend d’anticorps reconnaissants des épitopes structuraux quaternaires. Ces épitopes
dépendent du bon repliement et de la formation des protéines structurales. Une attention
particulière doit être attribuée à ces épitopes dans le cadre de développement de vaccins.
Chapitre : Vaccins en cours de développement
1) Vaccin contre la dengue couramment en essai clinique
Après 70 ans de recherche, le premier vaccin contre la dengue fut autorisé à la mise sur le
marché en 2016. Ce vaccin est un vaccin vivant atténué chimérique qui contient l’ensemble
de protéines et régions non-traduites de la fièvre jaune (YFV) à l’exception des protéines prM
and E qui appartiennent à la dengue. Il existe deux autres vaccins vivants atténués en cours
d’essai cliniques, ainsi que des vaccins vivants inactivés et des vaccins à ADN et ARN.
Les vaccins de sous-unités recombinantes, aussi appelés Particule Pseudo-Virale (VLP)
reposent sur l’immunisation par des protéines recombinantes issues de la transfection de
gènes d’intérêt dans un système d’expression. Le vaccin de ce type le plus avancé est le vaccin
développé par Merck and Co qui a finalisé la Phase I d’essai clinique.
2) Vaccins contre le Zika en essais cliniques
Il y a deux vaccins vivants atténués en essai clinique, trois vaccins vivants inactivés, deux
vaccins à ADN et un vaccin à RNA
Chapitre : Les particules pseudo-virales
Les particules pseudo-virales sont composées de structures protéiques répétitives qui
s’assemblent entre elles, donc la structure générale est identique au virus de type sauvage ne
transportant pas de matériel génétique viral. Ces particules sont non-infectieuses, ne peuvent
se répliquer et imitent la conformation native d’épitopes essentiels. Ces particules résultent
de l’expression de gènes codant pour les protéines structurelles en combinaison ou non, de
gènes codant pour des protéines non-structurelles.
Ces protéines peuvent être exprimée dans plusieurs type de cellules. Chaque système a ses
propres avantages et inconvénients variant du rendement, de la complexité de la mise en
place, des modifications post-traductionnelles, la rapidité de production et enfin les limites
règlementaires. Ces systèmes peuvent utiliser des cellules bactériennes, des levures, des
insectes, des mammifères et des végétaux.
Les particules pseudo-virales ont été décrites pour la première fois en 1978 par l’expression
de deux protéines du HBV dans des cellules de levures. L’étude de microscopie électronique a
révélé que les particules formées sont identiques au virus.
I. Vaccin à VLP en essai clinique
En 2020, il y a trois vaccins à particules pseudo-virales qui ont reçu une autorisation de mise
sur le marché. Le premier vaccin qui a reçu sa licence est le vaccin Recombivax de Merck contre
l’hépatite B. Il y a eu ensuite deux vaccins à particules pseudo-virales contre le papilloma virus
humain. Il s’agit des vaccins Gardasil et Cervarix respectivement développés par Merck & Co
en 2006 et GlaxoSmithKline en 2009. Il n’y a aucun vaccin VLP contre la dengue et le Zika en
essai clinique.
II. Etudes précliniques de VLPs de dengue
Plusieurs stratégies utilisant des VLPs ont été développées précédemment au niveau
préclinique. Ces stratégies se focalisent principalement sur l’utilisation des protéines prM et E
pour la formation et la sécrétion de VLPs.
Un signal de rétention dans la partie associée à la membrane de l’enveloppe entrave la
secretions de VLPs. Ce signal de rétention est propre à la dengue. Lorsque cette partie est
échangée avec la séquence homologue de JEV, des particules pseudo-virales sont secrétées à
la suite de l’expression des gènes prM et de la protéine chimérique E dans des cellules 293T.
Ces particules produites stimulent des anticorps neutralisants.
Une approche différente pour l’expression de VLP a utilisé des cellules de levures pour
l’expression des protéines CprME. Les sérums immuns de lapins présentent des anticorps
neutralisants après immunisation de ces derniers.
Un autre facteur important dans la maturation de la particule virale est le rôle du domaine
transmembranaire de la capside (anch-C). Une étude a mis en évidence le rôle de la capside
dans le repliement et la sécrétion de VLPs de prM et E. Le rendement de sécrétion était
supérieur lorsque les protéines prM et E étaient coexprimées avec un vecteur exprimant la
capside.
Une autre étude a révélé que l’optimisation du codon du plasmide d’expression permet une
augmentation de sécrétion trois fois plus importante qu’en utilisant la séquence génique de
type sauvage.
Une analyse de paysage scientifique contemporain au lancement du projet a permis d’obtenir
une première vision de stratégie de développement pour des VLP de dengue.
Method - Stratégie pour le développement de VLP
La stratégie utilisée pour le développement de VLP de dengue et de Zika repose sur la
formation de l’enveloppe virale avec la co-expression des trois protéines structurelles C-prM-
E ainsi que la protéase virale NS2b/NS3 en conservant le plus possible le processus et la
séquence d’acides aminés du type sauvage
I. Les protéines d’intérêt pour les VLPs de dengue
1) L’Enveloppe
L’éctodomaine de l’enveloppe est composé de 3 domaines : DI, DII et DIII. Ils ont pour
fonctions respectives la charnière moléculaire, le domaine de dimérisation avec le peptide
fusogène et enfin le site d’attachement au récepteur.
Les domaines I et II sont discontinus et s’alternent à l’extrémité N-terminale. La protéine
comporte deux sites de glycosylation : N-67 sur le domaine II et N-153 sur le domaine I. les
domaines associés à la membrane sont composés de deux hélices amphipathiques E-H1 et E-
H2 ainsi que deux domaines transmembranaires E-T1 et E-T2. Le signal de rétention unique à
la dengue est localisé dans l’E-H1.
2) Le Précurseur de Membrane
La protéine Précurseur de Membrane est composée du précurseur à l’extrémité N-terminal,
qui lui-même est composé de 7 feuillets antiparallèles β qui sont stabilisés par 3 ponts
disulfures ainsi qu’un site de glycosylation N-69. Le domaine de la membrane, M, est long de
75 acides aminés et est suivi par deux domaines transmembranaires T1 et T2. La furine est
responsable du clivage entre pr et M. La séquence consensus est Arg-Xaa-(Lys/Arg)-Arg. Le
site catalytique de la protéase contient 3 cavités qui reconnaissent les acides aminés
positionnés en P1, P2 et P4. Pour les quatre sérotypes de la dengue, l’acide amine P3 du site
de clivage a des propriétés physico-chimiques différentes des autres Flavivirus, il est à l’origine
de la production de particules immatures puisque le site de clivage n’est pas optimal.
3) Le duo prM et E
L’interaction entre les deux protéines structurelles prM et E est primordiale pour l’expression,
la maturation et la formation de la particule virale. En 2012, Tsai et al. ont démontré que prM
est faiblement secrétée lorsque la protéine est exprimée seule tandis que son expression est
stable lorsqu’elle est co-transfectée avec E-H2. Enfin, l’expression de prM augmente
lorsqu’elle est co-transfectée avec des quantités croissantes d’enveloppe et décroît lorsqu’elle
est co-transfectée avec l’éctodomaine de la protéine E.
L’hétérodimérisation de prM-E est un élément dans l’assemblage des VLPs. Une étude de
microscopie électronique révèle que E-T1 est proche de M-T1. Cette proximité entre les
domaines transmembranaires, permet d’assurer l’interaction entre les deux éctodomaines.
L’acide aminé H98 de la protéine M de prM s’aligne avec les hélices αA et αB du domaine II de
la protéine E formant un patch électrostatique assurant l’hétérodimérisation. Une étude
cristallographique confirme que ET1 et EH2 sont plus importants pour l’hétérodimérisation
avec prM que l’éctodomaine de E. L’acide amine H244 de la protéine E fait face au site de
glycosylation N-69 de prM. Lorsque le pH acide du compartiment cellulaire transitionne au pH
neutre du milieu extracellulaire, ces deux éléments perdent leur affinité. L’hétérodimérisation
est aussi importante pour la reconnaissance du virion par des anticorps réactifs de groupe. Par
exemple, l’anticorps 4G2 ne peut reconnaitre la protéine E lorsqu’elle est tronquée à
l’éctodomaine bien que 4G2 reconnaisse le domaine II et non un domaine associé à la
membrane.
La Capside est la protéine la moins conservée parmi les Flavivirus. Il existe 40% d’homologie
entre les séquences de virus appartenant au même genre et 22% entre les Flavivirus transmis
par un vecteur arthropode. La capside reste cytosol tout au long de la maturation virale.
L’extrémité C-terminale est intégrée dans la membrane du réticulum endoplasmique et se
termine par le site de clivage par la protéase cellulaire dans la partie interne du E.R. Le
domaine transmembranaire est composé de 4 hélices hydrophobiques et est appelé « ancre »
(anch-c). La coordination entre le clivage de l’anch-C et le clivage C-prM est important pour
maintenir l’infectivité virale.
Il existe des différences de. Dans le cas du Murray Valley Encephalitis Virus (MVEV), les clivages
de chaque côté de la membrane du ER sont séquentiels. En effet, C-prM ne peut être clivé que
lorsque C-anchC est précédemment clivé. Dans le cas de YFV, lorsque le clivage de C-AnchC
est découplé du clivage du clivage C-prM, alors le virus est non-infectieux car incapable de
former des plaques de lyse in vitro.

4) Identification and caractérisation de la protéase NS3

La protéase NS3 et son cofacteur NS2b sont responsables des clivages C-anchC, NS2a/NS2b,
NS2b/NS3, NS3/NS4a, NS4b/NS5. NS3 est une protéase à serine dans la partie N-terminale
jusqu’à l’acide aminé 180 avec la triade catalytique en D75, H51, S135 et un domaine
d’attachement du substrat. L’extrémité C-terminale de la protéine est une hélicase. Le
cofacteur NS2b est nécessaire à la protéase NS3. Lorsque NS3 est exprimé seule, la protéase
est soluble tandis que lorsqu’elle est co-exprimée avec son cofacteur S2b, elle reste associée
à la membrane favorisant ainsi le clivage de la polyprotéine qui est incrustée dans la
membrane de l’ER et voit son activité protéolytique multipliée de 3,300 fois. La séquence
entière du cofacteur est nécessaire à la bonne interaction avec la protéase. Les domaines
transmembranaires aux deux extrémités du cofacteur sont très conservés parmi les Flavivirus
tandis que le centre a le plus de divergence.
La glycosylation est essentielle pour le repliement des protéines virales, l’attachement au
récepteur et l’interaction du virus avec le système immunitaire. La glycosylation virale est
réalisée par les mécanismes cellulaires de l’hôte. La glycosylation des protéines virales est
conserve dans cette stratégie.
Afin de produire des VLP imitant au plus proche le virus, des optimisations sont apportées aux
protéines d’intérêt.
1) Suppression du signal de rétention de l’E-H1 par amélioration des propriétés
amphipatiques de l’hélice : mutations I398L, M401A, M412L
2) Amélioration du site de clivage de la furine par mutation de l’acide aminé positionné
en P3 du site : mutation Q88A
3) Optimisation du site catalytique de la protéine NS3 et troncation de cette dernière
pour conserver uniquement la protéase : mutation L115A
II. Transition vers les VLP de Zika
En 2015, l’épidémie de Zika virus m’a permis d’appliquer la stratégie développée pour la
dengue à ce virus. Zika est phylogénétiquement proche de DENV-4 qui est le sérotype le plus
stable lors de changement de température. La triade de protéines structurelles ne possède
pas les caractéristiques physico-chimiques limitantes de celles de la dengue. Au début du
projet seule la souche virale MR-766 isolée en 1947 était disponible. La souche plus
contemporaine FSS 13025 isolée en 2010 fut disponible lorsque le projet était plus avancé. Les
VLPs de Zika ont été formés par la co-expression des protéines CprME et NS2b/NS3.
III. Système d ’expression et température
Les protéines structurelles sont exprimées par un plasmide d’expression différent de celui des
protéines non-structurelles. Le codon des gènes est optimisé pour l’expression des protéines
dans les cellules Hek293. Cette lignée cellulaire a été choisie puisqu’elle tolère une densité
cellulaire importante, est en suspension et est facilement transrectale avec un bon rendement
de production de protéines recombinantes. Cette lignée cellulaire permet aussi d’apporter les
bonnes modifications post traductionnelles et peut facilement s’amplifier dans un milieu de
culture ne contenant pas de sérum, ce qui est un avantage pour la sureté et la purification du
vaccin final.
La température optimale pour cette lignée cellulaire est 37°C. les deux virus, la dengue et le
Zika, peuvent tous les deux se répliquer à 28°C dans le vecteur et à 37°C lors d’une infection
humaine. La température impacte la structure des protéines prM et E à la surface du virion et
la transition est observée à 33°C.
Dans une étude préliminaire, la réduction de la température fut testée pour améliorer le
rendement de production et plus particulièrement 31°C afin d’éviter un ralentissement trop
important, tout en s’approchant de la température de transition. Lors de cette étude, il est
apparu que l’anticorps 4G2 ne reconnaissait que les particules produites à 31°C et non à 37°C.
Ce phénome fut unique aux VLP de dengue. Les VLP de Zika produites à 31C et 37°C ont pu
être reconnues par l’anticorps 4G2.
IV. Caractérisation des VLPs et évaluation des vaccins
1) Expression in vitro et formation des VLPs
Une fois la stratégie mise en place, les VLPs ont été caractérisées dans un premier temps par
Western blot pour vérifier l’expression intracellulaire des protéines d’intérêt en utilisant des
anticorps polyclonaux. Dans un second temps, la formation et sécrétion des VLP sont
confirmées par microscopie électronique et dot blot en utilisant des anticorps monoclonaux
reconnaissant des épitopes structurels. Pour la dengue, il a s’agit des anticorps :
-4G2 : anticorps réactif au groupe qui reconnait un épitope discontinu autour du peptide
fusogène sur le DII. Les acides aminés impliqués dans l’épitope sur le virus de la dengue sont
G104, G106, L107 and W231. Cet anticorps reconnait les deux brins du domaine II, ce qui
indique une bonne formation de l’éctodomaine.
- 3H5 : cet anticorps est spécifique de type et reconnait les acides aminés non-adjacents K305
et P384 du domaine III de la protéine E
- C10 : cet anticorps reconnait un épitope quaternaire. Cet épitope est caractérisé d’épitope
« dépendant de la dimérisation de la protéine E » EDE1 puisqu’il implique domaine I et l’acide
amine K310 du III. La chaine latérale de ce dernier couvre l’acide amine W101 de la boucle
fusogène. Cette boucle est primordiale pour la stabilisation du dimère E. Cet anticorps
interagit aussi avec les deux sites de glycosylation N-67 et N-153 et affecte l’intégrité de la
boucle 150 mais ne peut reconnaitre que la particule mature puisque son site de
reconnaissance de la protéine E se superpose avec le site d’interaction de pr avec E.
- Anticorps anti-Zika : le virus du Zika a un sérotype unique. La production de VLPs à 37°C ne
prévient pas la reconnaissance de la particule par l’anticorps 4G2. Les anticorps utilisés, Z-64,
Z-48 and Z-67, sont fortement neutralisants. Ces anticorps reconnaissent des régions
spatialement distinctes sur le domaine III de la protéine E.
2) Etude in vivo
Une fois que les VLPS de dengue et de Zika ont été caractérisées, elles ont été utilisées pour
l’immunisation de souris Balb/c. L’immunogénicité de ces particules est évaluée par ELISA et
essai de neutralisation. Un sérum convalescent de Zika et un pool de sérum convalescent de
dengue sont utilisés comme référence. Les vaccins VLP sont comparés avec le virus homologue
inactivé par formaline.
V. Etapes par Etapes
Les étapes du plan de recherche sont décrites ci-dessous :
 Résultats
Chapitre 1: Dengue-2 virus-Like particles (VLP) based vaccine elicits the highest titers of
neutralizing antibodies when produced at reduced temperature
Discussion :
Quelques années après la réalisation du projet, le recul et la maturité me permettent
d’identifier à ce jour des éléments qui auraient pu être apportés pour donner plus d’impact à
l’article publié :
- tester par dot blot, microscopie électronique et immunisation, des VLPs produites à 31°C et
incubées à 37°C pendant 30min
- réaliser un test ELISA et une séroneutralisation par réduction des plages de lyse avec le virus
produit à 31°C pour comparer la stimulation d’anticorps neutralisants pour les deux
températures.
- évaluer in vitro la facilitation de l’infection par anticorps en utilisant un virus de sérotype
différent.
Depuis la publication de cet article, il y a maintenant plus d’informations disponibles
concernant le rôle de la capside et ses caractéristiques. En effet, l’hélice transmembranaire
est composée de 14 acides aminés tandis que celle des autres Flavivirus est en général de 18-
22 acides aminés.
De plus cette hélice comprend une proline en position 110. Cette proline est localisée en P5
du site de clivage C-pr. Elle est aussi un acide aminé qui a des propriétés physico-chimiques
perturbatrices d’hélice et est commune aux cartes sérotypes de la dengue. Une élongation de
l’hélice de la capside ainsi que la modification de la P110 pourraient être des éléments
additionnels contribuant à l’optimisation de la formation et la sécrétion de VLPs. Des études
récemment publiées soutiennent l’utilisation de ces modifications.
Enfin, les effets de la température sur les particules virales révèlent que les quatre sérotypes
réagissent différemment, renforçant l’utilisation de VLP dont la composition se rapproche le
plus possible du type sauvage.
Chapitre 2: Zika virus-like particle (VLP) based vaccine
Discussion :
Malgré le peu de connaissance du virus du Zika au début de l’épidémie, il était attendu que la
formation de VLP de Zika serait moins laborieuse que celle de la dengue. Le rendement
d’expression de VLP était supérieur à celui de VLP de dengue et la température de production
n’a pas eu d’impact sur l’organisation structurelle générale de la particule virale. Il n’existe
qu’un seul sérotype de ce virus mais plusieurs génotypes.
Au cours de l’évolution du virus, les génotypes africains ont une délétion au niveau du
domaine de glycosylation N-154 et ont quatre acides aminés manquants autour de la boucle
150. Cette boucle de glycosylation présente dans le génotype asiatique, responsable des
épidémies contemporaines, est importante pour la pathogénicité du virus et plus
particulièrement pour l’attachement du virus au récepteur cellulaire.
Dengue et Zika circulent grâce au même vecteur dans des zone géographiques identiques. La
question de la facilitation de l’infection par anticorps se pose puisque ces deux virus sont très
proches. Dans notre étude, il a été demandé d’explorer un potentiel ADE par les anticorps
stimulés par la vaccination de VLP de Zika lors d’une infection avec la dengue. Les résultats
présentés dans la publication montrent qu’il n’y a pas d’ADE, ce qui est en accord avec la
littérature récente.
Conclusion

Le futur des vaccins VLP contre la Dengue et le Zika

Interne

Atouts:
-Approche structurelle
- Stratégie utilisée par
une autre équipe de
recherche
Faiblesse:
- Dengue: immunité
équilibrée parmi les 4
sérotypes
- Zika: immunité durable

Positif Négatif

Opportunités:
- Formulation
pentavalente composée
des 4 sérotypes de
dengue et du génotype
asiatique Zika
Menaces:
- Evaluation malavisée
des données
précliniques et cliniques
- Manque d’intérêt et de
financement

Externe
 UNIVERSITÉ DE STRASBOURG

La dengue et le Zika sont deux virus transmis par le moustique Aedes et ont un impact significatif sur la population
mondiale. La création de vaccins sûrs et efficaces répondrait à un besoin médical. Les particules pseudo-virales
(VLP) sont composées de structures protéiques répétitives qui s’assemblent entre elles, donc la structure
générale est identique au virus de type sauvage ne transportant pas de matériel génétique viral. Ces particules
sont non-infectieuses, ne peuvent se répliquer et imitent la conformation native d’épitopes essentiels. Le travail
présenté décrit le développement de VLP de dengue et de Zika en utilisant la co-expression des protéines
structurales CprME et de la protéase virale NS2b/NS3. Ces VLPs ont stimulé des anticorps neutralisants in vivo.
Dans le cas de la dengue, les VLP produites à basse température (31°C) ont stimulé des anticorps plus neutralisant
que les VLP produites à 37°C.

Mots-clés : Particules pseudo-virales, Dengue, Zika, Vaccin, température, épitope quaternaire, structure

Dengue and Zika virus are transmitted the Aedes mosquito and have a high impact on the global public health.
The development of safe and efficacious vaccine would respond to an unmet medical need. Viral-like particles
(VLPs) are self-assembling organized repetitive structures identical to wild-type virus (enveloped or non-
enveloped) without carrying any viral genome. These particles are non-infectious, non-replicating, and mimic the
native conformation of essential epitopes. The body of this work describes the development of dengue and Zika
VLP by using the co-expression of the three structural proteins CprME with the NS2b/NS3 protease complex.
these VLPs elicited neutralizing antibodies in vivo. In the context of dengue virus, VLP produced at low
temperature stimulated more neutralizing antibodies then VLPs produced at 37°C.

Keywords : Viral-like particles, Dengue, Zika, Vaccine, Temperature, quaternary epitopes