Investigation of the role of the ubiquitin proteasome pathway in dengue virus life cycle 1

80 Figure 3.6: Proteasome inhibition decouples infectious DENV-2 production from viral RNA replication in mosquito midguts .... 101 Figure 3.17: Bortezomib decouples infectious DENV pro

                                                         Copyright by Choy Ming Ju Milly

2015

Abstract

The mosquito-borne dengue virus (DENV) is a cause of significant global health burden However, no licensed vaccine or specific antiviral treatment for dengue is available DENV interacts with host cell factors to complete its life cycle although this virus-host interplay remains to be fully elucidated Many studies have identified the ubiquitin proteasome pathway (UPP) to be important for successful DENV production, but how the UPP contributes to DENV life cycle as host factors remain ill

defined We show here that a functional UPP is critical for virus egress in both Aedes aegypti mosquitoes and human monocytic cell lines Using RNA inference studies,

we show in vivo that knockdown of ubiquitin proteasome pathway-related genes,

including proteasomal subunits, ß1, ß2 and ß5 decouples RNA replication from infectious titer production in the mosquito midgut Mechanistically, inhibition of proteasome function prevented virus egress by exacerbating endoplasmic reticulum (ER) stress through the unfolded protein response (UPR) UPR-induced translational repression reduced overall protein levels of the exocyst components needed for exocytosis This mechanism also appears to be amenable for clinical translation as inhibition of UPP in primary monocytes with the licensed proteasome inhibitor, bortezomib, inhibited DENV titers even at low nanomolar drug concentration

Furthermore, we show in vivo in a wild type mouse model that DENV replication and

spread in the mouse spleen is exquisitely sensitive to proteasome inhibition The mechanism of action suggests that such a therapeutic approach may apply to other viruses that rely on exocytosis for virus egress to complete their life cycle

Acknowledgements

I would like to express my heartfelt gratitude to Professor Duane J Gubler and

Associate Professor Ooi Eng Eong for their mentorship throughout the course of my study I thank them for being ever so generous in sharing their wealth of knowledge and experiences with me

My sincere appreciation is extended to my thesis advisory committee members, Professor Subhash Vasudevan and Assistant Professor Ashley St John for their invaluable advice and critical suggestions

Special thanks to Summer Zhang, Tan Hwee Cheng, Mah Sook Yee, Angelia Chow, Brett Ellis, Jason Tang, October Sessions and all my colleagues at Duke-NUS

Graduate Medical School for helping me in one way or another

Lastly, I would like to acknowledge my loved ones, especially my mum and husband for their encouragement all these years My elder sister, Julieanne is also instrumental

in sparking my interest in science at a very young age I am glad to be able to share

my successes and failures with them

This body of work is dedicated to my little baby, Ethan, who is the main source of

my inspiration

Contents

Title Page      i  

Abstract  Signature    ii  

Copyright    iii  

Abstract    iv  

Acknowledgements    v  

Table  of  Contents    vi  

List of Tables    x  

List of Figures    xi  

Chapter 1 INTRODUCTION    1  

1.1 Dengue 1

1.1.1 Dengue – a disease of global significance 1

1.1.2 Dengue structure and genome 5

1.1.3 Clinical symptoms and disease manifestations 9

1.1.4 Laboratory diagnosis of dengue 12

1.1.4.1 Immunoglobulin M (IgM) detection 12

1.1.4.2 Immunoglobulin G (IgG) detection 12

1.1.4.3 DENV isolation and propagation 12

1.1.4.4 Quantitative real time polymerase chain reaction (qRT-PCR) 14

1.1.4.5 NS1 antigen detection 15

1.2 Role of host and viral factors in dengue pathogenesis 17

1.2.1 Host factors 17

1.2.1.1 Epidemiological evidence for secondary infection and severe dengue 18

1.2.1.2 T cell response 18

1.2.1.3 Antibody-dependent enhancement (ADE) 19

1.2.2 Viral factors 23

1.2.2.1 Genotype differences 23

1.2.2.2 Glycosylation of E and NS1 proteins 24

1.2.2.3 Other viral factors 25

1.3 Animal models for dengue 26

1.3.1 Non-human primates 26

1.3.2 Laboratory mice 26

1.4 DENV infection in hosts 29

1.4.1 Life cycle of DENV 29

1.4.2 DENV infection in human 33

1.4.3 DENV infection in mosquito vector 33

1.5 Prevention and control of dengue 37

1.5.1 Ae aegypti – a public health scourge 37

1.5.2 Vector control 37

1.5.3 Vaccine development and progress 40

1.5.3.1 Live attenuated vaccines and chimeric live attenuated vaccines 40

1.5.3.2 Bivalent or trivalent vaccines – a possibility? 42

1.5.4 Antiviral drugs 44

1.5.4.1 Therapeutic antibodies 44

1.5.4.2 Viral protein inhibitors 44

1.5.4.3 Antiviral therapies for supportive management of dengue 46

1.5.4.4 Host factors inhibitors 46

1.5.4.5 Celgosivir 47

1.6 Ubiquitin proteasome pathway in viral infection 50

1.7 Specific aims 53

Chapter 2 Materials and Methods 54

2.1 Mosquitoes and cells 54

2.2 Primary monocytes isolation 54

2.3 Virus stock 55

2.4 Plaque assay 55

2.5 qRT-PCR 56

2.6 DENV-2 infection in mosquitoes 56

2.7 Gene silencing assays 57

2.8 Generation of whole-transcriptome cDNA library 58

2.9 RNAseq analysis 58

2.10 DENV-2 infection in monocytes 59

2.11 DiD labeling of DENV-2 59

2.12 Alexa Fluor labeling of DENV-2 60

2.13 MTS cell viability assay 60

2.14 Western blots 61

2.15 Immunofluorescence 61

2.16 Transmission Electron Microscopy 62

2.17 RNase treatment of β–lactone treated cells 62

2.18 Annexin V staining of primary monocytes 62

2.19 Bortezomib treatment in DENV-infected C57BL/6 mice 63

2.20 Immunohistochemistry analysis of mouse spleen 63

2.21 Measurement of hematocrit level and platelet count in whole blood 64

2.22 MCPT-1, TNF-α and IFN-γ quantification 65

2.23 Viral load quantification in mouse spleen 65

2.24 Statistical analysis 66

Chapter 3 Results    67  

3.1 Establish a mosquito infection model at Duke-NUS Graduate Medical School 67

3.1.1 Comparison of mosquito inoculation technique and qRT-PCR to measure

DENDENV concentration in mosquitoes, vertebrate and mosquito cell cultures, and h

a and human sera 67

3.2 Investigate the role of the UPP in DENV life cycle 77

3.2.1 Functional UPP is required for infectious DENV-2 production in mosquito

mmidguts 77

3.2.2 Regulation of UPP specific genes decouples infectious DENV-2 production f from viral RNA replication in mosquito midguts 83

3.2.3 Proteasome inhibition with β-lactone did not alter virus entry at non-toxic lelevels 91

3.2.4 DENV-2 egress is dependent on proteasome function 93

3.3 Proteasome inhibition exacerbates ER stress and represses translation of E

EXOC7,TC10 and EXOC1 98

3.4 Potential use of bortezomib, a proteasome inhibitor, as an anti-flaviviral drug 102

3.4.1 Bortezomib inhibits infectious DENV production in primary monocytes 102

3.4.2 Bortezomib reduced viral load and signs of dengue pathology in C57BL/6

mmice 108

3.5 Summary of results 112

Chapter 4 Discussion 113

4.1 Preface 113

4.2 Unfolded protein response during flavivirus infection 115

4.3 Repurposing proteasome inhibitors as an anti-flaviviral therapeutic 117

4.4 Learning from Ae aegypti mosquito 121

4.5 Beyond the anti-viral effects of bortezomib: Potential use as an adjuvant 122

4.6 Conclusion 123

References 124

Appendix A 147

Appendix B 223

Biography 226

List of Tables

Table 3.1: Comparative titration of C6/36 cell culture virus supernatants by qRT-PCR and mosquito inoculation 73 Table 3.2: Percentage of DENV-2-infected mosquitoes after knockdown of

proteasome subunits (p-value; Fischer’s Exact test) 82 Table 3.3: Summary of Illumina HighSeq 2000 RNA-sequencing using Partek

Genomic Suite v6.6 85 Table 4.1: Current developments of proteasome inhibitors undergoing different

phases of clinical trials 120

List of Figures

Figure 1.1: Global distribution of dengue 3

Figure 1.2: Distribution of DENVs 1-4 in (A) 1970 and (B) 2011 4

Figure 1.3: Flavivirus genome and polyprotein 7

Figure 1.4: DENV structure 8

Figure 1.5: Clinical signs and symptoms of dengue 11

Figure 1.6: Viral load and antibodies quantification in primary dengue infection 16

Figure 1.7: ADE in dengue infection 22

Figure 1.8: DENV life cycle 31

Figure 1.9: Schematic representation of the exocyst complex tethering a vesicle to plasma membrane 32

Figure 1.10: Anatomy of Ae aegypti 36

Figure 1.11: Ae aegypti distribution in the Americas during the 1930s and in 1970 and 2003 39

Figure 1.12: Dengue vaccines under development 43

Figure 1.13: Kaplan-Meier plots of viral NS1 antigen clearance 49

Figure 3.1: Replication kinetics of DENV-2 (A) PR1940 and (B) PR6913 in adult female Ae aegypti mosquitoes 70

Figure 3.2: Replication kinetics of DENV-2 derived in cell cultures measured by the mosquito inoculation technique and qRT-PCR 71

Figure 3.3: Linear regression analysis between RNA copy number and MID50 72

Figure 3.4: Comparative titration of ten viremic (DENV-2) human sera by qRT-PCR and mosquito inoculation technique 76

Figure 3.5: Characterization of DENV-2 replication in the midguts and head/thoraces of Ae aegypti following ingestion of an infectious blood meal 80

Figure 3.6: Proteasome inhibition decouples infectious DENV-2 production from viral RNA replication in mosquito midguts 81

Figure 3.7: RNA-sequencing of Ae aegypti midgut 84

Figure 3.8: Differential regulation of genes belonging to UPP during DENV infection 88

Figure 3.9: Functional UPP is required for infectious DENV-2 production in mosquito

midgut 89

Figure 3.10: Functional UPP is required for infectious DENV-2 production in mosquito head/thorax 90

Figure 3.11: Proteasome inhibition with β-lactone did not alter virus entry at non-toxic levels 92

Figure 3.12: β-lactone decouples infectious DENV-2 production from viral RNA replication in THP-1 cells 94

Figure 3.13: DENV-2 egress is dependent on proteasome function 96

Figure 3.14: Increased amount of packaged DENV observed after proteasome inhibition 97

Figure 3.15: Protein levels of EXOC7, TC10 and EXOC1 decreased post β-lactone treatment, but transcript levels remained unchanged 99

Figure 3.16: UPP inhibition increases ER stress and attenuates translation of EXOC7 and TC10 101

Figure 3.17: Bortezomib decouples infectious DENV production from viral RNA replication in primary monocytes 104

Figure 3.18: Bortezomib inhibits infectious virus production for different strains of all 4 dengue serotypes and YF17D in a dose-dependent manner in primary monocytes 105

Figure 3.19: Epoxomicin reduces infectious DENV titers in a dose-dependent manner for all 4 dengue serotypes 106

Figure 3.20: Bortezomib induces apoptosis in primary monocytes at 6 hpi 107

Figure 3.21: Bortezomib inhibited DENV spread in spleen of WT mice 109

Figure 3.22: Bortezomib alleviated signs of dengue in C57BL/6 mice 110

Figure 3.23: Bortezomib alleviated inflammatory responses in DENV infected C57BL/6 mice 111

Figure 4.1: Schematic diagram of findings 114

Chapter 1: Introduction

1.1 Dengue

1.1.1 Dengue – a disease of global significance

Dengue has emerged to be the most important arthropod-borne viral disease globally, causing an estimated 390 million infections annually (Bhatt et al, 2013) The global prevalence of dengue has increased dramatically in the past 40 years and is now endemic in more than 100 countries in Africa, South-east Asia, the Americas, the Eastern Mediterranean, and the Western Pacific (Gubler, 2014) Consequently, 3 billion people that live in or travel to the tropics are at constant risk of infection with any of the four DENV serotypes (DENV-1, DENV-2, DENV-3 and DENV-4) (Bhatt

et al, 2013) (Figure 1-1)

Infection with any of the four serotypes of DENV may result in a spectrum of illness, ranging from subclinical to undifferentiated febrile illness, to classical dengue fever (DF), and in more severe cases, to dengue hemorrhagic fever (DHF)/dengue shock syndrome (DSS) (Trung & Wills, 2014) DHF/DSS is an acute and sometimes deadly illness characterized by thrombocytopenia, bleeding tendency and capillary leakage that if left untreated, could lead to hypovolemic shock Fatality rates can be as high as 20% without appropriate medical support (WHO, 2009) In addition to human

suffering, dengue results in substantial economic impact (Shepard et al, 2014)

Despite under-reporting of cases in many dengue endemic regions, it is estimated that over the next decade, dengue will collectively cost South-East Asian nations US$2.36 billion in lost productivity and medical care (Shepard et al, 2014)

The lack of effective mosquito control, environmental changes associated with rapid population growth, unplanned urbanization and increased international air travel have

collectively contributed to the geographic spread of Aedes aegypti (Ae aegypti) and

the dispersal of DENVs (Gubler, 2002; Gubler, 2004; Gubler & Meltzer, 1999; Rigau-Perez et al, 1998) (Figure 1-2) To further exacerbate the problem, there is no licensed vaccine or specific antiviral treatment for dengue The high prevalence of dengue, together with the absence of specific therapeutics and lack of effective preventive measures conspire to make dengue an even more acute global public health threat (Kyle & Harris, 2008)

Figure 1-1 Global distribution of dengue An estimated 390 million dengue

infections occur worldwide annually Dengue is now endemic in more than 100 countries in Africa, South-east Asia, the Americas, the Eastern Mediterranean, and the Western Pacific Figure adapted from (Bhatt et al, 2013)

DEN 1

DEN 2

DEN 1 DEN 2 DEN 3 DEN 4

DEN 1 DEN 2

Figure 1-2 Distribution of DENVs 1-4 in (A) 1970 and (B) 2011 Compared to

the 1970s, the increased incidence and spread of all four dengue serotypes observed in 2011 indicates the urgent need for DENV control Figure adapted from (Gubler, 1998)

DEN 1 DEN 2 DEN 3 DEN 4 DEN 1

DEN 2 DEN 3

DEN 1 DEN 2 DEN 3

DEN 2 DEN 3 DEN 4

DEN 1 DEN 2 DEN 3 DEN 4

DEN 1 DEN 2 DEN 3 DEN 4

A

B

DEN 1 DEN 2 DEN 3 DEN 4

1.1.2 Dengue structure and genome

Dengue belongs to the genus Flavivirus family Flaviviridae (Kuno et al, 1998;

Solomon & Mallewa, 2001) The Flaviviridae family includes several other viruses

that pose a threat to public health, such as the yellow fever virus, Japanese

encephalitis virus, West Nile virus and tick-borne encephalitis virus (Gould &

Solomon, 2008)

Flaviviruses are enveloped RNA viruses 45-50 nm in diameter that contain a strand, positive-sense capped RNA genome of approximately 11 kb The four DENV serotypes are immunologically distinct but antigenically related, each sharing about 65-70% homology The genome is made up of three structural and seven non-

single-structural proteins (Chambers et al, 1990), which is translated as a large precursor polypeptide molecule from a single, long open reading frame The unique open

reading frame is flanked by two untranslated regions (UTRs), which contain structural and functional elements required for viral translation and replication The N-terminal region encodes the structural proteins, capsid (C), pre-membrane (prM) and envelope (E), followed by the nonstructural proteins NS1, NS2A, NS2B, NS3, NS4A, NS4B and NS5 Individual mature viral proteins are then processed from the polyprotein by both cellular and viral proteases (Harris et al, 2006; Westaway, 1987) (Figure 1-3)

The glycoprotein E contains most of the antigenic determinants of the virus and is essential for viral attachment and entry, while M, synthesized as the precursor (prM), functions as a chaperone during maturation of the viral particle The E protein can be divided into three structural or functional domains: the central domain; the

dimerization domain, which presents a fusion peptide; and the receptor-binding domain (Figure 1-4) The highly basic C protein packages the RNA genome and is surrounded by 180 monomers of E protein that are organized into tightly packed dimers and lies flat on the surface of the viral membrane (Rey, 2003) The other seven non-structural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B and NS5) are

essential for viral RNA replication, assembly and modulation of host cell responses (Pastorino et al, 2010)

Figure 1-3 Flavivirus genome and polyprotein The 11 kb flavivirus genome is

made up of three structural and seven non-structural proteins, which is translated as

a large precursor polypeptide molecule from a single, long open reading frame The unique open reading frame is flanked by two UTRs The N-terminal region encodes the structural proteins, C, prM and E, followed by the nonstructural proteins NS1, NS2A, NS2B, NS3, NS4A, NS4B and NS5 Individual mature viral proteins are then processed from the polyprotein by both cellular and viral proteases Figure adapted from (Pastorino et al, 2010)

Figure 1-4 DENV structure DENV is a spherical and enveloped virus 50 nm in

diameter The virion contains three structural proteins C, M and E and the RNA genome The surface of the mature DENV virion is smooth with the envelope proteins aligned in pairs parallel to the virion surface The E protein can be divided into three structural or functional domains: the central domain; the dimerization domain, which presents a fusion peptide; and the receptor-binding domain Figure adapted from (Kuhn et al, 2002)

1.1.3 Clinical symptoms and disease manifestations

Infection with any one of the four serotypes of DENV may cause a wide spectrum of clinical outcome The incubation period can range from 3-14 days before symptoms develop Characteristic symptoms of classical DF include high fever accompanied by

2 or more symptoms of headache, retro-orbital pain, myalgia, arthralgia, flushing of the face, rash, anorexia, nausea and abdominal pain During fever defervescence, macular rash may be observed and most patients recover without complications about

a week after disease onset (WHO, 2009) However, a small percentage of patients may progress to severe disease during this period of fever defervescence

The most common form of severe disease, DHF is characterized by a fever lasting 2-7 days, a tendency to bleed, thrombocytopenia (platelet count ≤ 100,000/mm3), and plasma leakage When a critical volume of plasma is lost through leakage, cardiac output may become insufficient to maintain the necessary blood pressure resulting in DSS With prolonged shock, organ hypofusion can result in progressive organ

impairment, metabolic acidosis and disseminated intravascular coagulation (WHO, 1997), eventually resulting in death within 12 to 36 hours after shock onset (Figure 1-5) Currently, treatment of acute dengue is supportive, using either oral or intravenous rehydration for mild or moderate disease, and intravenous fluids and blood

transfusion for more severe cases With early diagnosis and proper management, fatality rates can be decreased to 1% or less (Kalayanarooj, 2014; Trung & Wills, 2014; WHO, 1997)

Many dengue infections present as viral syndrome and may be missed or mistaken for

influenza or another viral infection A definitive diagnosis of dengue infection can be made only in the laboratory and relies on detecting specific antibodies in the patient’s serum, isolating the virus, or detecting viral antigen or RNA in serum or tissues

Figure 1-5 Clinical signs and symptoms of dengue WHO case definition of (A)

DF, (B) DHF, and (C) DSS (D) The incubation period before signs of symptoms develop can range from day 3 to day 14, average 4-7 Figure adapted from (Whitehead et al, 2007)

1.1.4 Laboratory diagnosis of dengue

1.1.4.1 Immunoglobulin M (IgM) detection

The efficacy of clinical diagnosis, surveillance, prevention and control of dengue has been limited by the lack of low-cost, easy to use and sensitive diagnostic tests

Currently, laboratory diagnosis in most dengue endemic countries relies on detecting immunoglobulin M (IgM) antibody in acute serum samples IgM antibodies can be detected as early as 4 days after the onset of fever IgM peaks at approximately two weeks following fever onset and then decline to undetectable levels over the next two months (Guzman et al, 2010) (Figure 1-6) IgM antibody capture ELISA (MAC-ELISA) is based on capturing human IgM antibodies on a microtiter plate using anti-human-IgM antibody, followed by the addition of DENV specific antigen containing viral E protein One limitation of this assay is that serologic diagnosis of dengue may

be confounded if the patient had recently been infected or vaccinated with an

antigenically related flavivirus due to the cross-reactive nature of the antibody

1.1.4.2 Immunoglobulin G (IgG) detection

During primary infection, immunoglobulin G (IgG) can be detected 7- 10 days after illness onset, making it less useful for early diagnosis (Figure 1-6) However, the rapid increase of IgG levels during secondary infection (as early as day 4 from illness onset) can be suggestive of dengue when the ratio of IgM and IgG is used

1.1.4.3 DENV isolation and propagation

Besides serological tests, isolation of virus from patient samples during the febrile phase is routinely performed in the laboratory as a diagnostic test The added

advantage is the availability of viral isolates, which is essential for characterizing

virus strain differences, critical for viral surveillance and pathogenesis studies

DENV has been among the most difficult viruses to isolate and propagate because the virus does not infect or replicate well in most laboratory animals and mammalian cell cultures Studies on dengue in the 1940s to 1960s used suckling mice and various tissue culture systems for isolation and assay of DENV (Halstead et al, 1964;

Hammon et al, 1960; Igarashi, 1978; Sukhavachana et al, 1966; Yuill et al, 1968) Although the use of mammalian and insect cell culture systems improved sensitivity (Igarashi, 1978; Sukhavachana et al, 1966; Yuill et al, 1968), which is influenced by both the cell culture system and the strain of DENV, many unpassaged DENV do not produce cytopathic effects (CPE) when grown in these cells The lack of a sensitive isolation and assay system that could be used for unpassaged wild type viruses,

prevented rapid advancement in dengue research

In large urban centers of the tropics, DENVs are usually maintained naturally in an

Ae aegypti-human-Ae aegypti cycle with periodic epidemics (Gubler, 1988) The

development of the mosquito inoculation technique in the early 1970s provided a highly sensitive method for the isolation, propagation and quantitation of DENV (Rosen & Gubler, 1974) The use of a natural host proved to be 10 to 1000 times more sensitive in detecting dengue viruses than the commonly used plaque assay,

depending on the serotype or strain of virus (Choy et al, 2013) As DENV can be frequently recovered from serum in mosquitoes when the same serum tested negative

in both mammalian and insect cell culture, the use of mosquitoes is especially

important when attempting to propagate viruses with low replication efficiency from mildly symptomatic illness, and to isolate viruses from sera, tissues of naturally infected humans, wild animals or field caught mosquitoes

The quantitation of virus in sera, human tissues, animals, mosquitoes, and cell culture

is calculated using the method of Reed and Muench (Reed & Muench, 1938) and expressed as the dose that infects 50% of the mosquitoes inoculated (MID50) Using this method, many studies in the South Pacific islands, Indonesia, Thailand, Sri Lanka and Puerto Rico showed a high rate of isolation of DENV from primary clinical samples Equally important was the demonstration of considerable variation in

viremia levels of different DENV strains and serotypes in patients showing a

correlation with disease severity (Gubler et al, 1978; Gubler et al, 1981; Gubler et al, 1979; Kuberski et al, 1977) This technique has also been used as an assay for other arboviruses

1.1.4.4 Quantitative real time polymerase chain reaction (qRT-PCR)

For DENV detection and quantitation, quantitative real time polymerase chain

reaction (qRT-PCR) has become the method of choice in the past 20 years This method is generally more sensitive and efficient than isolation assays, and can provide

a rapid serotype-specific diagnosis Moreover, DENVs can be identified and

quantified directly from clinical samples A number of methods involving primers from different locations in the genome and different approaches to detect the qRT-PCR products have been developed over the past several years (Guzman & Kouri,

1996; Lanciotti et al, 1992)

1.1.4.5 NS1 antigen detection

Dengue NS1 is a highly conserved glycoprotein essential for DENV viability and is secreted from infected cells as a soluble hexamer (Flamand et al, 1999) It can be found in the peripheral blood circulation for up to 9 days from illness onset (Alcon et

al, 2002) (Figure 1-6) Currently, the NS1 ELISA based antigen assay is

commercially available for DENV, and has been shown to be useful for differential diagnostics between flaviviruses because of its specificity Anti-NS1 antibodies can also be used in immunohistochemistry to detect infection in postmortem tissues More recently, commercial tests combining NS1 antigen and IgM antibody detection have become increasingly popular (Blacksell et al, 2011)

Figure 1-6 Viral load and antibodies quantification in primary dengue infection NS1 is detectable up to 9 days from illness onset As viremia subsides,

IgM antibodies antibodies can be detected as early as 4 days after the onset of fever and declines thereafter IgG antibodies are produced at the later stage of infection Figure adapted from (Guzman et al, 2010)

1.2 Role of host and viral factors in dengue pathogenesis

1.2.1 Host factors

The pathogenesis of DHF/DSS is a product of host and viral factors Host factors such

as the immune status, sex, age, innate immune system and individual/species level genetics in combination appear to contribute to the pathogenesis of severe dengue disease (de la et al, 2007; Sangkawibha et al, 1984; Thein et al, 1997)

For dengue, the immune status of the host may modify the disease in two directions – toward infections accompanied by mild or no disease because of partial or complete protection, or in the direction of increased disease severity (Halstead, 1988; Sabin, 1952) This could be attributed to the following reasons Firstly, as the four serotypes

of DENV have evolved from a common ancestor, they share many common antigens but at the same time, have sufficient differences to allow sequential infection with different types (Wang et al, 2000; Weaver & Vasilakis, 2009) Although infection with one dengue serotype provides lifelong immunity to that virus, there is no cross protective immunity to the other serotypes (Gubler, 1998) Secondly, altered T cell responses during secondary infections with heterologous serotypes have been

postulated to contribute to cytokine storm and immunopathogenesis of DHF/DSS

(discussed in 1.2.1.2) Lastly, cells of the mononuclear cell lineage are capable of

supporting dengue virus infection via an antibody-dependent enhancement (ADE)

process (discussed in 1.2.1.3), which can increase disease severity (Halstead, 1988;

Halstead & O'Rourke, 1977a; Halstead & O'Rourke, 1977b; Kliks et al, 1989)

Supporting evidence for each of these possibilities are as follows:

1.2.1.1 Epidemiological evidence for secondary infection and severe dengue

The first evidence that a prior dengue infection can somehow modify a subsequent heterologous dengue infection from mild to severe was inferred from a highly

significant association found between the occurrence of secondary type antibody responses and severe illness (Halstead et al, 1967) In a large-scale Bangkok study, Halstead et al., 1967, found that approximately 85% of dengue-infected patients with DHF were experiencing a secondary infection with a different serotype of DENV Clinical studies performed on Rhesus monkeys also showed a significant correlation between the occurrence of thrombocytopenia and secondary infection with DENV-2

in the monkeys (Halstead et al, 1973)

Furthermore, the pathogenic role of secondary infections is evident in three epidemics

in Cuba where DHF/DSS occurred in patients who were infected with DENV-1 in

1977 and then infected with DENV-2 in 1981 or 1997 (Guzman et al, 2002) For the

1997 outbreak, silent infections occurred for people with primary infection of DENV

In the same outbreak, a very high proportion of DENV-1-immune adults (from 1977 epidemic) infected with DENV-2 developed either DF or DHS/DSS (Guzman et al, 1990; Kouri et al, 1989)

1.2.1.2 T cell response

A possible explanation offered for the more severe disease outcome observed in secondary infection is the original antigenic sin hypothesis (Rothman, 2011) T cell responses play a major role in cell-mediated immunity T cells can recognize viral

epitopes presented on infected cells by major histocompatibility complex (MHC) molecules, and direct release of pro-inflammatory cytokines It is proposed that during secondary infection, the preferential expansion of lower avidity cross-reactive memory T cells from the previous infection may out-compete the higher avidity specific naive T cells (Mongkolsapaya et al, 2003) This can delay viral clearance and result in production of pro-inflammatory cytokines which may lead to vascular

leakage (Rothman, 2011)

However, a recent study demonstrated that during heterologous infections 2/DENV-3 and DENV-3/DENV-2 infections), the recognition of conserved or cross-reactive epitopes was either constant or expanded compared to that in homologous infections (Weiskopf & Sette, 2014), indicating that antigenic sin does not impair the quality of T cell responses significantly Also, a temporal mismatch with the

(DENV-production of CD8+ T cells to the time of onset of vascular leakage in children with DHF suggested that the mechanism that triggers vascular leakage in children may be independent of CD8+ T cell responses (Dung et al, 2010) While the role of T cells in the pathogenesis of dengue remains to be fully elucidated, T cells alone cannot fully explain the epidemiological trends observed with DHF/DSS

1.2.1.3 Antibody-dependent enhancement (ADE)

To provide another molecular explanation of severe disease during a heterologous secondary infection, it has been hypothesized that prior infection of patients can enhance the infection and replication of dengue viruses in cells of the mononuclear cell lineage via fragment crystallisable receptors (FcγR) This hypothesis is referred to

as ADE of virus infection DENV can form immune complex with heterotypic

antibodies from previous infection or maternal transfer The virus-antibody immune complex binds to FcγR bearing cells, facilitating DENV uptake It is thought that increased infection of these cells produce greater amounts of vasoactive mediators in response to dengue infection that in turn causes increased vascular permeability leading to hypovolemia and shock (Halstead, 1970; Halstead & O'Rourke, 1977b; Kliks et al, 1989) (Figure 1-7) ADE is also believed to cause DHF in infants with primary dengue infection Maternal DENV antibodies acquired passively from infants are shown to be capable of enhancing dengue when added to monocytes in the

presence of dengue virus (Kliks et al, 1988; Simmons et al, 2007)

However, in a recent nested case-control study on DHF in infants (Libraty et al, 2009), no significant association was found between the severity of dengue and the DENV3 ADE activity due to the presence of maternal antibodies in infants at illness onset What is unique about this study is the focus on DENV3, while DENV2

predominates in the earlier studies At the same time, studies have shown that not all secondary infections result in DHF epidemics and that the disease severity of dengue could not be explained by the secondary infection alone Indeed, while DHF/DSS immediately accompanied an Asian genotype DENV-2 in Cuba in 1981 in DENV-1 immune individuals, similar sequential infection of DENV-1 in 1990 and an

American genotype DENV-2 in 1995 did not result in DHF/DSS in Peru Genomic studies on DENV-2 strains found a total of six amino acid charge differences between the American (non-DHF-related) and Asian (DHF-related) viruses, suggesting that DENV strain differences is also be a determinant of disease outcome

Epidemiological observations have also suggested that DENV-2 and 4 cause more

symptomatic disease in secondary infection compared to DENV-1 and -3 (Henchal et

al, 1982) On the other hand, severe DENV-1 and DENV-3 infections have been observed to be more likely to occur in DENV-nạve individuals than either DENV-2

or DENV-4 infections (Balmaseda et al, 2006; Harris et al, 2000; Nisalak et al, 2003; Vaughn et al, 2000) Collectively, the literature indicates that while outbreaks of DF and DHF are influenced by multiple host factors (Halstead, 2007; Halstead, 2008), secondary infections are not always associated with increased viremia Viral factors also appear to play a critical role (Gubler & Trent, 1993)

Figure 1-7 ADE in dengue infection DENVs form immune complexes with

heterotypic antibodies from previous infection or from maternal transfer, which binds to FcγRs on monocytes, resulting in increased infection and viral load, increasing severity of disease outcome Figure adapted from (Murphy & Whitehead, 2011)

et al, 2007)

While the association between genetic changes in the dengue viral genome and

epidemic transmission has been repeatedly observed epidemiologically (Bennett et al, 2006; Messer et al, 2003; Rico-Hesse et al, 1997), the mechanism underlying how these genetic changes influence epidemic transmission is unclear (Rico-Hesse, 2010) Phenotypic expression of genetic changes in the virus genome may include increased virus replication and viremia, disease severity, and epidemic potential

1.2.2.2 Glycosylation of E and NS1 proteins

Addition of carbohydrates to Asn67 of the E protein was observed to be crucial for virus particle production in mammalian and mosquito cell lines (Bryant et al, 2007) This carbohydrate group has also been implicated to mediate binding of DENV to dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN) on dendritic cells (DCs) (Pokidysheva et al, 2006) In addition to the E protein, several studies have suggested that glycosylation of the NS1 protein is important for DENV production (Pryor et al, 1998; Tajima et al, 2008) While removal of just one

of the two glycosylation site allowed replication of the virus albeit with a less virulent phenotype; showing reduced growth in cell culture and attenuation of neurovirulence when inoculated into 3-day-old mice, complete removal of DENV2 NS1

glycosylation resulted in an unstable virus with numerous mutations (Pryor et al, 1998) Moreover, NS1 protein glycosylation is required for efficient secretion of the protein from infected cells (Flamand et al, 1999) All these, coupled with the

observation that patients experiencing DHF have high levels of NS1 protein in the blood, suggest that NS1 protein glycosylation may be important in disease

pathogenesis (Avirutnan et al, 2006; Libraty et al, 2002)

1.2.2.3 Other viral factors

Other viral protein, NS2A, NS4A and NS4B are also implicated to play a role in disease severity, possibly via interference of the interferon (IFN) signaling pathway,

an important antiviral mechanism at play during dengue infection (Jones et al, 2005; Munoz-Jordan et al, 2005; Munoz-Jordan et al, 2003)

Besides viral proteins, the non-coding untranslated region of the DENV genome can also play a critical role in viral replication and pathogenicity A non-coding RNA approximately 0.5 kb in size, derived from incomplete degradation of the viral 3′UTR

by the cellular 5′-3′ exonuclease is found to be produced abundantly by all

flaviviruses This non-coding RNA, termed subgenomic flaviviral RNA (sfRNA), was reported to be required for viral pathogenicity in a mouse model of the attenuated Kunjin strain of West Nile virus (Pijlman et al, 2008) Follow-up studies determined that Kunjin virus sfRNA counteracted IFN antiviral activity (Pijlman et al, 2008; Schuessler et al, 2012) Similarly for DENV, DENV-2 sfRNA binds to RNA binding proteins, and inhibits their antiviral activity, leading to profound inhibition of IFN-stimulated genes (ISGs) mRNA translation (Bidet et al, 2014) Strikingly, while the majority of genomes synthesized during infection were processed into sfRNA, sfRNA

is dispensable for RNA replication in IFN-incompetent cells, arguing for its role in antagonizing immune responses of the host

1.3 Animal models for dengue

1.3.1 Non-human primates

The degree in which each of the above factors contributes to dengue pathogenesis

could be understood through in vivo experimentation Unfortunately, an animal model

that reliably reproduces the clinical features observed in human cases is lacking Although primates and mouse models may display some signs of human dengue, each have their inherent limitations (Buck et al, 2014; Clark et al, 2013; St John et al, 2013; Zellweger et al, 2010) Common marmosets have been shown to display high levels

of viremia and exhibit clinical signs of dengue infection, including fever and

thrombocytopenia during primary infection with clinical isolates of DENV (Omatsu

et al, 2011) The viremia and antibody responses exhibited in marmosets were also similar to that in humans after secondary infection with heterotypic serotypes of DENV in that infection with one serotype did not confer protection to others (Moi et

al, 2014) Similarly, high doses of DENV inoculated intravenously in Rhesus

macaques can also induce similar hemorrhagic manifestations observed in humans with DHF (Onlamoon et al, 2010) However, due to ethical reasons, limited facilities and high costs, the large-scale use of non-human primates remains a challenge

1.3.2 Laboratory mice

Mouse models serve as a cheaper alternative but most of them do not reliably mimic the disease observed in humans and display limited viremia Another major stumbling block is their reliance on mouse-adapted DENVs for infection, as human clinical isolates do not replicate well in these mice (Zellweger & Shresta, 2014) The AG129

mouse model, deficient in receptors for types I and II IFN, is increasingly used for the investigation of virus tropism, vascular leakage, vaccine and drug testing (Plummer & Shresta, 2014) Although the mice may show some dengue-like characteristics such as intestinal hemorrhage, vascular leakage and viremia, they have limited usefulness for investigation into the immune responses against DENV infection or vaccination, as these mice are severely immunocompromised The lack of IFNα/β and γ signaling might draw some limitations and calls for caution in the interpretation of findings derived from this model, especially since most human dengue cases are

immunocompetent One recent example is the testing of a α-glucosidase inhibitor, celgosivir, as a treatment for acute DF Celgosivir was found to fully protect AG129 mice from lethal infection with a mouse adapted DENV (Rathore et al, 2011;

Watanabe et al, 2012) However, in a phase 1b clinical trial, celgosivir did not reduce viremia or duration of fever in patients with dengue compared to those treated with placebo (Low et al, 2014)

Although wild type (WT) mice have been shown to be less susceptible to infection than mice lacking the capacity of innate immune activation (Shresta et al, 2004), studies have observed that WT mice can sustain replicating DENV infection

(Atrasheuskaya et al, 2003; Boonpucknavig et al, 1981; Chen et al, 2007; Huang et al, 2000; St John et al, 2011) DENV-induced vascular hemorrhaging or pathology has also been examined in these mice (Assuncao-Miranda et al, 2010; Chen et al, 2007)

St John and colleagues optimized a system of infecting WT mice to generate systemic DENV infection, by injecting 1 × 106 plaque-forming units (PFU) of DENV, intra-peritoneally (i.p.), to bypass the stages of natural, peripheral infection which WT mice can clear virus quickly (St John et al, 2013) Although replication can be detected in