The role of type I interferon in the immunobiology of chikungunya

Jane Amelia Clare Wilson

B. App. Sc. (Human Biology), B. App. Sc. (Hons)

A thesis submitted for the degree of Doctor of Philosophy

at

The University of Queensland in 2015

School of Medicine & QIMR Berghofer Medical Research Institute

I Abstract

Chikungunya virus (CHIKV) is a mosquito-transmitted alphavirus that can cause explosive outbreaks of a febrile, arthritic/arthralgic disease usually lasting weeks to months, and in rare cases, more than a year. In 2004, the largest ever CHIKV outbreak began in Kenya, spreading to islands of the Indian Ocean, India, South East Asia and major outbreaks have recently occurred in the South Pacific Islands and the Caribbean. The host type I interferon (IFN) response is crucial for effective control of CHIKV infection. Herein, the dynamics, source and responses generated by the type I IFNs following CHIKV infection were investigated.

Interferon regulatory factors 3 (IRF3) and IRF7 are key transcription factors for the type I IFN response. While CHIKV infection of wild-type mice is non-lethal, infection of mice deficient in both IRF3 and IRF7 (IRF3/7-/-) resulted in mortality, illustrating that these factors are essential for protection. Using knockout mice for the adaptor molecules upstream of IRF3 and 7, IPS1 was found to be the most important for type I IFN production, with TRIF and MyD88 also contributing to the response. Mortality in IRF3/7-/- mice was also associated with type I IFN suppression of pathological levels of IFNγ and haemorrhagic shock.

Heterozygous reporter mice, in which eGFP was expressed under the control of either the IFNβ or the IFNα6 promoter on one , were employed to try and identify the cellular source of type I IFN production following CHIVK infection. However, eGFP production was found to be insufficient for detection, limiting the utility of this approach. Investigations of type I IFN production in tissues using RT-PCR revealed a strong correlation between type I IFN induction and viral tissue titres, with feet and lymph nodes producing the most robust type I IFN responses.

RNASeq was undertaken to examine, in detail, the nature of the type I IFN responses generated in the feet and inguinal lymph nodes of mice following CHIKV infection at days 2 and 7 post-infection. High quality sequencing data was generated from

II host and viral poly-adenylated RNA, permitting investigation of low level transcription from both sources, and the identification of novel /transcripts, in addition to analysis of differential gene expression.

Investigation of host transcriptional activity revealed a type I IFN dominated immune response, in spite of a low induction of type I IFN mRNA transcripts at day 2 and an absence of type I IFN mRNA induction on day 7. Many type I IFN-regulated genes, including IRF7, remained up-regulated in the feet at day 7, suggesting type I IFN- independent mechanisms for maintenance of type I IFN-regulated gene expression. Six novel IRF3 isoforms and a potentially novel coding gene were also identified.

RNA-Seq analysis of CHIKV transcriptional activity reflected the known CHIKV replication kinetics, in which virus replicates in the joints, and spreads to the draining lymph nodes, with viral titres peaking at day 2 and largely resolving at day 7. Assessment of mutations within the CHIKV genome showed an accumulation with time, with error accumulation higher in the lymph nodes than in the feet. These studies represent the first deep sequencing analysis of CHIKV genomes in tissues.

The findings in this thesis substantially add to our knowledge of the role and function of the immune response after CHIKV, illustrating for the first time that type I IFNs protect against haemorrhagic shock and that expression of type I IFN inducible genes is maintained well after type I IFN induction has waned.

III Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my research higher degree candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the policy and procedures of The University of Queensland, the thesis be made available for research and study in accordance with the Copyright Act 1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis.

IV Publications during candidature

Peer-reviewed papers:

Wilson JAC, Rudd PA,, Gardner J, Larcher T, Babarit C, Le TT, Anraku I, Kumagai Y, Loo Y-M, Gale Jr M, Akira S, Khromykh AA, Suhrbier A. 2012. Interferon Response Factors 3 and 7 Protect against Chikungunya Virus Hemorrhagic Fever and Shock. Journal of Virology 86:9888-9898.

Poo YS, Rudd PA, Gardner J, Wilson JAC, Larcher T, Colle M-A, Le TT, Nakaya HI, Warrilow D, Allcock R. 2014. Multiple Immune Factors Are Involved in Controlling Acute and Chronic Chikungunya Virus Infection. PLoS neglected tropical diseases 8:e3354.

Conference abstracts:

Wilson JAC, Ellis J, Schroder WA, Suhrbier A. 2014. RNA-Seq analysis of chikungunya virus-induced arthritis reveals significant similarity to gene expression in rheumatoid arthritis. 9th International Congress on Autoimmunity. Nice, France.

Wilson JAC, Ellis J, Schroder WA, Suhrbier A. 2013. RNASeq analysis of temporal and tissue difference in virus and host gene expression during chikungunya virus infection. Brisbane Immunology Group Meeting. Surfers Paradise, Qld.

Wilson JAC, Rudd PA,, Gardner J, Larcher T, Babarit C, Le TT, Anraku I, Kumagai Y, Loo Y-M, Gale Jr M, Akira S, Khromykh AA, Suhrbier A. 2012. Interferon Response Factors 3 and 7 Protect against Chikungunya Virus Hemorrhagic Fever and Shock. Brisbane Immunology Group 13th Annual Retreat. Kingscliff, NSW, December 2012

Wilson JAC, Rudd PA,, Gardner J, Larcher T, Babarit C, Le TT, Anraku I, Kumagai Y, Loo Y-M, Gale Jr M, Akira S, Khromykh AA, Suhrbier A. 2012. Interferon Response

V Factors 3 and 7 Protect against Chikungunya Virus Hemorrhagic Fever and Shock. 6th Australasian Virology Society Meeting. Kingscliff, NSW, December 2012.

Publications included in this thesis

No publications included.

Contributions by others to the thesis

Prof. Andreas Suhrbier supervised and conceptualised the project, assisted with interpretation of results and reviewed this thesis.

Dr Wayne Shroder provided day to day supervision, assisted with interpretation of results and reviewed this thesis.

Dr Jonathan Ellis provided support for Chapters 4-6.

Prof Thibaut Larcher performed the in situ hybridisation and assisted with interpretation and analysis of histological findings in Chapter 2.

Dr Penny Rudd contributed to non-routine technical work and data analysis in Chapter 2.

Statement of parts of the thesis submitted to qualify for the award of another degree

None.

VI Acknowledgements

I extend my sincere gratitude to my supervisors Andreas Suhrbier, Wayne Shroder and Alex Khromykh for their guidance throughout my PhD. I thank Andreas Suhrbier for allowing me the opportunity to undertake my PhD within his laboratory, and for his knowledge, advice, generosity and humour, inter alia. I thank Wayne Schroder for his dedication and thoroughness, and for his patience, humour, resolute confidence in my abilities, and genuine understanding and kindness during problematic(al) times. I thank Alex Khromykh for his encouragement, advice and outsider perspectives.

I acknowledge the University of Queensland for my Australian Postgraduate Award and QIMR Berghofer Medical Research Institute for my Student Top-Up Scholarship, and Student Travel Award.

I also wish to thank my fellow members of the Inflammation Biology Lab, past and present, for their ever present willingness to teach me and share their expertise, and for making my day to day work life (especially lunchtimes) enjoyable and often hilarious. Thank you very much Lee Major, Thuy Le, Itaru Anraku, Yee-Suan Poo, Joy Gardner, Penny Rudd, Tom Partridge, Christiana Agyei-Yeboah, Dilan Pathirana, Vinay Venugopalan and Natalie Prow - I couldn't have asked for a better group to work with!

I am also very grateful to the many excellent people I have met while completing my PhD, who I now call friends. In particular, I thank Michelle Neller, Haran Sivakumaran, Tamara Powell, Patrick Dwyer and Claire Foldi for their genuine kindness, empathy and reassuring wisdom. I also owe a very big thanks to all of my friends who encouraged me to undertake this and saw me through to the end, especially Charlie Baker, Sam Vincent and Tara Edwards.

VII Thanks also to my beautiful family Sarah, Ben, Chrissy, Jem and Noah, Pete, Janka and Emil, Si and Bron, Nick and Lucy, for their interest, encouragement and high entertainment value.

Finally, I thank my mum and dad, Clare and Michael Wilson, for encouraging a love of learning, for their relentless enthusiasm, love and generous support and for knowing me better than anyone else. This is for them.

VIII Keywords chikungunya virus, type I interferon, inflammation, innate immunity, infectious disease,

Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code: 110707, Innate Immunity, 50%

ANZSRC code: 110804, Medical Virology, 50%

Fields of Research (FoR) Classification

FoR code: 1107, Immunology 50%

FoR code: 1108, Medical Microbiology, 50%

IX Table of Contents

Chapter 1: Introduction ...... 1 1.1 Alphaviruses ...... 1 1.1.1 Virion and genome ...... 3 1.1.2 Non-structural ...... 4 1.1.3 Infection and replication ...... 5 1.1.4 Structural proteins ...... 7 1.2 Chikungunya virus ...... 8 1.2.1 Natural history ...... 8 1.2.2 The A226V mutation and the global dissemination of CHIKV ...... 9 1.2.3 Disease presentation ...... 10 1.2.3.1 Acute phase ...... 10 1.2.3.2 Chronic phase ...... 11 1.2.3.3 Atypical disease and mortality ...... 11 1.2.4 Diagnosis of CHIKV infection ...... 12 1.2.5 Treatment of CHIKV disease ...... 13 1.2.6 Vaccine development ...... 14 1.3 Immunobiology of chikungunya virus ...... 15 1.3.1 Disease models ...... 15 1.3.2 CHIKV cell and tissue tropism ...... 16 1.3.3 CHIKV pathogenesis ...... 17 1.3.4 Persistence ...... 17 1.3.5 Immune control of chikungunya virus ...... 18 1.4 The type I IFN response and RNA ...... 19 1.4.1 Toll-like receptors ...... 20 1.4.2 Cytoplasmic PRRs ...... 21 1.4.3 Signal transduction by PRRs ...... 21 1.4.3.1 MyD88-dependant signalling pathway ...... 22 1.4.3.2 TRIF-dependant signalling pathway ...... 22 1.4.3.3 IPS-1-dependant signalling pathway ...... 23 1.4.4 Activation of the IFNβ promoter ...... 23 1.4.5 Type I IFN signalling and amplification...... 24 1.4.6 Interferon regulated genes ...... 26 1.5 Alphaviruses and the type I IFN response ...... 27

X 1.5.1 Sensitivity to the antiviral effects of IFNα/β ...... 27 1.5.2 IRGs in alphavirus infection...... 27 1.5.3 Evasion of the type I IFN response ...... 28 1.6 Project aims ...... 28 Chapter 2: The role of interferon regulatory factors 3 and 7 during CHIKV infection 29 2.1 Introduction ...... 29 2.2 Materials and Methods ...... 31 2.2.1 Physical Containment Level 3 Procedures ...... 31 2.2.2 Cells ...... 31 2.2.3 Mice ...... 32 2.2.4 Blood and tissue preparation ...... 32 2.2.5 Virus and viral preparations ...... 33 2.2.6 Mouse vaccination, infection and monitoring ...... 33 2.2.7 Prophylactic IFN treatment and adoptive transfer ...... 33 2.2.8 Viraemia measurement ...... 34 2.2.9 Viral titres in tissues ...... 34 2.2.10 Foot measurements ...... 35 2.2.11 Interferon bioassay ...... 35 2.2.12 concentration by cytometric bead array ...... 35 2.2.13 Anti-CHIKV antibody ELISA ...... 36 2.2.14 Neutralisation assay ...... 36 2.2.15 Analysis of clinical parameters ...... 37 2.2.15.1 Haematocrit determination...... 37 2.2.15.2 Platelet counts ...... 37 2.2.15.3 Urine output ...... 37 2.2.15.4 Temperature measurement ...... 37 2.2.16 RNA extraction ...... 37 2.2.17 cDNA synthesis...... 38 2.2.18 Real time RT-PCR ...... 39 2.2.19 Histology ...... 39 2.2.20 In situ hybridisation ...... 39 2.2.21 Statistical analysis ...... 40 2.3 Results ...... 41 2.3.1 IRF3 and IRF7 protect against CHIKV-induced high viral burden and mortality...... 41 2.3.2 IRF3 and IRF7 protect against severe foot swelling post CHIKV infection...... 43

XI 2.3.3 Prophylactic treatment with recombinant IFNα, IFNγ or adoptive transfer of monocyte macrophages did not prevent mortality in IRF3/7-/- mice...... 45 2.3.4 IRF7 is heavily up-regulated in response to CHIKV infection in WT mice...... 47 2.3.5 An absence of IRF3 and IRF7 does not affect antibody responses to CHIKV infection ...... 49 2.3.6 Absence of IRF3 and IRF7 alters type I IFN and pro-inflammatory cytokine responses to CHIKV infection...... 51 2.3.7 CHIKV tissue tropism is altered in IRF3/7-/- mice...... 53 2.3.8 Tissue pathology and inflammatory infiltrates were altered in IRF3/7-/- mice ...... 55 2.3.9 CHIKV infection in IRF3/7-/- mice induces haemorrhage and oedema ...... 57 2.3.10 CHIKV induces clinical symptoms of hypovolemic shock in IRF 3/7 -/- mice ...... 59 2.3.11 TRIF, MyD88 and IPS-1 are involved but not critical in viraemia control, foot swelling or type I IFN production following CHIKV infection ...... 62 2.3.12 CHIKV infection of IFNAR-/- mice results in rapid, lethal hypovolemic shock...... 64 2.4 Discussion ...... 66 Chapter 3: Investigation of the source of type I interferons following CHIKV infection...... 72 3.1 Introduction ...... 72 3.2 Materials and Methods ...... 74 3.2.1 Cells ...... 74 3.2.1.1 -derived macrophages (BMMs) and DCs (BMDCs) ...... 74 3.2.1.2 Resident peritoneal macrophages (RPMs) ...... 74 3.2.1.3 Splenocytes ...... 75 3.2.2 Transfections ...... 75 3.2.3 LPS, IFN and SFV stimulation of type I IFNs ...... 75 3.2.4 Virus ...... 76 3.2.5 Mice ...... 76 3.2.6 Mouse infection, inoculation and monitoring ...... 76 3.2.7 Blood and tissue preparation ...... 77 3.2.8 Viraemia and foot measurements ...... 77 3.2.9 RNA analysis ...... 77 3.2.10 DNA isolation ...... 77 3.2.11 Genotyping PCR ...... 77 3.2.11.1 Primers ...... 77 3.2.11.2 PCR reaction mixture and cycling conditions ...... 78

XII 3.2.12 Histology ...... 79 3.2.12.1 Fluorescence imaging ...... 79 3.2.12.2 Immunohistochemistry ...... 80 3.2.13 Western Blot ...... 80 3.2.14 Statistical analysis ...... 81 3.3 Results ...... 82 3.3.1 Establishment of an accurate genotyping reaction for identification of IFNα6-GFP+/- and IFNβ-GFP+/- mice...... 82 3.3.2 CHIKV viral burden and pathogenesis was altered in IFNα6-GFP+/- and IFNβ-GFP+/- transgenic mice ...... 84 3.3.3 eGFP could not be detected in IFNα6-GFP+/- and IFNβ-GFP+/- transgenic mice following CHIKV infection mice by histological analysis ...... 86 3.3.4 eGFP could not be detected by western blot analysis of IFNα6-GFP+/- and IFNβ- GFP+/- mouse tissues following CHIKV infection...... 89 3.3.5 eGFP could not be detected in cultured cells derived from IFNα6-GFP+/- and IFNβ- GFP+/- transgenic mice following type IFN stimulation...... 90 3.3.6 IFNβ mRNA expression is reduced in IFNβ-GFP+/- mice compared to IFNβ mRNA expression in WT mice following CHIKV infection...... 91 3.3.7 Analysis of IFNα6 and IFNβ mRNA induction in WT mouse tissues following CHIKV infection ...... 93 3.4 Discussion...... 96 Chapter 4: Quality control and overview of RNA-Seq data from CHIKV infected mice100 4.1 Introduction ...... 100 4.2 Materials and Methods ...... 102 4.2.1 Mice ...... 102 4.2.2 Virus ...... 102 4.2.3 Mouse infection and tissue preparation ...... 102 4.2.4 RNA extraction ...... 102 4.2.5 Pooling and purification of RNA ...... 103 4.2.6 High throughput RNA sequencing (RNA-Seq) ...... 103 4.2.7 Data Analysis using the Tuxedo pipeline ...... 104 4.2.7.1 Read mapping and identification of splice junctions ...... 104 4.2.7.2 Transcript assembly and differential expression analysis ...... 105 4.2.8 Preparation of DEGs lists ...... 105 4.2.9 Integrated Genome Viewer ...... 106

XIII 4.2.10 Assessment of data quality ...... 106 4.3 Results ...... 107 4.3.1 Sequence quality assessment ...... 107 4.3.2 Sequence depth and mapping quality ...... 108 4.3.3 Coverage depth ...... 110 4.3.4 Coverage uniformity ...... 111 4.3.5 Global assessment of mouse gene and transcript regulation during CHIKV infection.113 4.4 Discussion ...... 116 Chapter 5: Analysis of host gene expression in response to CHIKV infection using RNA-Seq ...... 119 5.1 Introduction ...... 119 5.2 Materials and Methods ...... 121 5.2.1 High throughput RNA sequencing (RNA-Seq) data ...... 121 5.2.2 Preparation of DEGs lists ...... 121 5.2.3 Unannotated DEGs ...... 121 5.2.4 Investigation of the protein coding potential of CHIKV-induced novel transcripts ... 122 5.2.4.1 Construction of mRNA sequences ...... 122 5.2.4.2 Assessment of mRNA and amino acid sequences ...... 122 5.2.4.3 Assessment of protein homology and presence of conserved domains ...... 122 5.2.5 Interferome v2.01 analysis ...... 123 5.2.6 Ingenuity Pathway Analysis ...... 123 5.2.7 Sequence alignments and phylogenetic analysis ...... 123 5.3 Results ...... 124 5.3.1 Differentially expressed genes (DEGs) ...... 124 5.3.2 Ingenuity Pathway Analysis of upstream regulators ...... 134 5.3.3 Ingenuity pathways analysis of canonical pathways ...... 137 5.3.4 Quantitation of type I IFN regulated genes ...... 139 5.3.5 Quantitation of interferon gene expression ...... 141 5.3.6 FPKM and fold changes of IRGs in the top 50 DEG lists ...... 143 5.3.7 Up-regulated IRGs in day 7 feet are primarily regulated by type I IFNs ...... 145 5.3.8 Transcriptional activity of IRF3 and IRF7 during CHIKV infection ...... 146 5.3.9 Assessment of uncharacterised DEGs ...... 148 5.3.10 Investigation of the protein coding potential of CINT9 ...... 150 5.4 Discussion ...... 153 Chapter 6: RNA-Seq analysis of CHIKV transcription in mice ...... 158

XIV 6.1 Introduction ...... 158 6.2 Materials and Methods ...... 159 6.2.1 Alignment of CHIKV sequences ...... 159 6.2.2 Evaluation of alignments & investigation of the parental genome sequence ...... 159 6.2.3 Determination of partial CHIKV genome sequence ...... 159 6.2.4 Mutational assessment ...... 160 6.3 Results ...... 161 6.3.1 Read mapping to the LR2006_OPY1 CHIKV genome ...... 161 6.3.2 Sequencing of the parental input virion genome ...... 164 6.3.3 Investigation of mutation accrual in the CHIKV genome ...... 167 6.4 Discussion ...... 169 Chapter 7: General Discussion ...... 172 7.1 Summary of major findings ...... 172 7.2 Discussion and future directions ...... 174 7.3 Concluding remarks ...... 180 References ...... 181 Supplemental Data ...... 229

XV List of figures

Figure 1.1 Schematic diagram of alphavirus entry, replication and release from host cells. Adapted from...... 7 Figure 1.2 Schematic diagram of the type I IFN signalling pathway and positive feedback loop...... 25 Figure 2.1 CHIKV infection in WT, IRF3-/-, IRF7-/-, and IRF3/7-/- mice...... 42 Figure 2.2 Foot swelling in WT, IRF3-/-, IRF7-/-, and IRF3/7-/- mice following CHIKV infection...... 44 Figure 2.3 Prophylactic treatment of IRF3/7-/- mice and IRF7 induction...... 46 Figure 2.4 IRF3 and IRF7 mRNA induction in WT mice...... 48 Figure 2.5 Antibody production and protection in WT, IRF3-/-, IRF7-/-, and IRF3/7-/- mice following vaccination...... 50 Figure 2.6 Cytokine expression during CHIKV infection in WT, IRF3-/- IRF7-/- and IRF3/7-/- mice...... 52 Figure 2.7 Virus location by in situ hybridisation in CHIKV-infected IRF3/7-/- and WT mice...... 54 Figure 2.8 Histology in CHIKV-infected IRF3/7-/- mouse skin, joint and muscle tissues. .... 56 Figure 2.9 Histological analysis of haemorrhage and oedema in CHIKV-infected IRF3/7-/- mice...... 58 Figure 2.10 Clinical symptoms of hypovolemic shock in IRF 3/7-/- mice following CHIKV infection...... 61 Figure 2.11 CHIKV infection in WT, TRIF-/-, MyD88-/- and IPS1-/- mice...... 63 Figure 2.12 Survival, foot swelling and clinical parameters in IFNAR-/- mice following CHIKV infection...... 65 Figure 3.1. Schematic diagram of primer design and pairing for detection of transgenic and/or WT alleles...... 78 Figure 3.2 Genotyping of F1 offspring from mating of IFNα6-GFP+/- or IFNβ-GFP+/- mice with WT mice...... 83 Figure 3.3 CHIKV infection in IFNα6-GFP+/- and IFNβ-GFP+/- mice compared to WT mice...... 85 Figure 3.4 Fluorescence in fixed frozen mouse tissue sections taken from IFNα6-GFP+/- and IFNβ-GFP+/- transgenic, and WT mice following CHIKV infection...... 88 Figure 3.5 Western blot of eGFP protein expression in IFNα6-GFP+/- mice compared to WT...... 89 Figure 3.6 IFNβ and eGFP mRNA expression by RT-PCR...... 92

XVI Figure 3.7. IFNβ and IFNα6 mRNA induction in WT mice at day 2 post infection CHIKV infection...... 95 Figure 4.1 FastQC assessment of sequence quality scores...... 107 Figure 4.2 Graphical display of the percentage of total reads aligned to the mm10 or CHIKV genome...... 109 Figure 4.3 Transcription profile of host and virus gene expression at pre-infection, day and day 7 post CHIKV infection. T ...... 112 Figure 4.4 Gene and transcript expression in the feet and LN of CHIKV-infected mice at day 2 and day 7 post infection...... 115 Figure 5.1 IPA canonical pathways engaged in mouse feet and lymph nodes at day 2 and day 7 post CHIKV infection...... 138 Figure 5.2. Percentage of type I IRGs in up-regulated DEGs...... 140 Figure 5.3 Transcript abundance of the IFN subtypes in CHIKV- and mock-infected mouse feet and lymph nodes...... 142 Figure 5.4 Transcript abundance and differential expression of IRGs in CHIKV- and mock- infected mouse feet and lymph nodes...... 144 Figure 5.5 IFN-induced genes in CHIKV infected mouse feet at day 7...... 145 Figure 5.6 Gene and isoform expression of IRF3 and IRF7 in mock infected and CHIKV- infected mouse feet and lymph nodes. A & B) ...... 147 Figure 5.7. Kozak consensus sequence with nucleotide position indicated and start codon (AUG) indicated in bold, underlined text...... 149 Figure 5.8 Figure 4. Synteny analysis, and phylogenetic relationship between CINT9 and the Ly6 antigen family...... 152 Figure 6.1 Reads mapping to the CHIKV LR2006_OPY1 genome...... 163 Figure 6.2 Reads mapping to the CHIKV genome...... 166 Figure 6.3. Percentage of mutations present in RNA-Seq reads aligned to the CHIKV genome...... 168

XVII List of tables

Table 1.1 Febrile, rheumatic alphaviral disease in humans...... 2 Table 3.1 Cycling conditions for genotype of IFNα6-GFP+/- and IFNβ-GFP+/- mice ...... 79 Table 4.1 Summary of RNA-Seq reads mapped to UCSC mm10 mouse genome or CHIKV LR2006_OPY1 genome by Bowtie/Tophat...... 109 Table 4.2 Values used in coverage calculation by Lander/Waterman general equation. 110 Table 5.1 Up-regulated DEGs in foot tissues at day 2 post CHIKV infection...... 126 Table 5.2 Up-regulated DEG in foot tissues at day 7 post CHIKV infection...... 127 Table 5.3 Up-regulated DEGs in lymph nodes at day 2 post CHIKV infection...... 128 Table 5.4 Up-regulated DEGs in lymph nodes at day 7 post CHIKV infection...... 129 Table 5.5 Down-regulated DEGs in foot tissues at day 2 post CHIKV infection...... 130 Table 5.6 Down-regulated DEGs in foot tissues at day 7 post CHIKV infection...... 131 Table 5.7 Down-regulated DEGs in lymph nodes at day 2 post CHIKV infection...... 132 Table 5.8 Down-regulated DEGs in lymph nodes at day 7 post CHIKV infection...... 133 Table 5.9. The major upstream regulators of gene expression in the feet of CHIKV infected mice at day 2 and day 7 post infection...... 135 Table 5.10. The major upstream regulators of gene expression in the lymph nodes of CHIKV-infected mice at day 2 and day 7 post infection...... 136 Table 5.11 Up-regulated, unannotated genes with protein coding features ...... 149

XVIII List of Abbreviations used in this thesis x g Times gravity

˚C Degrees Celsius

µg Microgram(s)

µL Micro litre(s)

2-ME 2-mercaptoethanol

ABTS 2,2’-azino-bis(3ethylbenzthaizoline-6-sulphonic acid)

ATCC American Type Culture Collection

BLOTTO Bovo lacto transfer technique optimiser bp Base pairs

CCID50 Cell culture infectious dose 50% cDNA Complementary deoxyribonucleic acid

CHIKV Chikungunya virus

CO2 Carbon dioxide

Conc. Concentration

CPE Cytopathic effect

DHS/DSS Dengue haemorrhagic fever/dengue shock syndrome

DMEM Dulbecco’s modified Eagle’s medium

DNA Deoxyribonucleic acid dNTP(s) Deoxyribonucleic triphosphate(s)

EDTA Ethylenediaminetetraacetic acid

ELISA Enzyme-linked Immunosorbent Assay

FCS Foetal calf serum

FPKM Fragments per kilobase (of exon) per million (fragments mapped)

XIX Ft Foot g Gram(s)

GFP Green fluorescent protein

HEPES N-[2-Hydroxyethyl]piperazine-N’-[2-ethanesulphonic acid]

IFN Interferon

IFNβ-GFP+/- IFNβ-linked GFP expressing heterozygous transgenic mouse

IFNα6-GFP+/- IFNα6-linked GFP expressing heterozygous transgenic mouse

IRF Interferon regulatory factor

IRG Interferon regulated gene

LN Lymph node(s)

MEFs Mouse embryonic fibroblasts mg Milligram(s) min Minute(s) ml Millilitre(s) mM Milli Molar ng Nanogram(s)

OCT Optimal cutting temperature embedding medium

OD Optical density

ORF Open reading frame

PBS Phosphate buffer solution

PCR Polymerase chain reaction

QIMR Queensland Institute of Medical Research

RA Rheumatoid arthritis

RPMI1640 Roswell Park Memorial Institute Medium 1640

XX RIPA Radioimmunoprecipitation assay lysis buffer

RT-PCR Real time PCR s Second(s)

SDS Sodium dodecyl sulphate

SDS-PAGE Sodium dodecyl supphate polyacrylamide gel electrophoresis

TE Buffer Tris-EDTA buffer uAUG Upstream start codon

WT Wild type

XXI 1. Introduction

1.1 Alphaviruses

Alphaviruses are enveloped, positive sense, single strand ribonucleic acid (RNA) viruses (Strauss and Strauss 1994, Zuckerman, Banatvala et al. 2009). The alphavirus genus, along with genus rubivirus, comprise the Togaviridae family (Fauquet, Mayo et al. 2005). The 29 known viruses within the alphavirus genus are divided into eight antigenic complexes according to serological cross-reactivity with a species prototype – Trocara, Middleburg, Ndumu, Semliki Forest, Barmah Forest, Eastern, Western and Venezuelan Equine Encephalitis Virus complexes (Fauquet, Mayo et al. 2005). Alphaviruses are arboviruses (arthropod-borne), maintained in a zoonotic transmission cycle between an hematophagous arthropod vector (generally mosquitoes) and a vertebrate, reservoir host (Fauquet, Mayo et al. 2005, Zuckerman, Banatvala et al. 2009). Occasionally, the virus may infect an epizootic vector or incidental host, resulting in outbreaks of disease in both human and (particularly equine) populations (Strauss and Strauss 1994, Zuckerman, Banatvala et al. 2009).

Based on geographical distribution, alphaviruses are commonly referred to as ‘Old World’ and ‘New World’ viruses (Strauss and Strauss 1994). New World viruses, such as Venezuelan and Eastern Equine Encephalitis viruses are found in the Americas and target the nervous system, resulting in encephalitic disease, which can be fatal. Old World viruses are native to the African, Australian, European and Asian continents, and are more commonly associated with arthralgia/myalgia, rash and fever, and only rarely associated with mortality (Suhrbier and La Linn 2004, Zuckerman, Banatvala et al. 2009). Table 1 lists the alphaviruses associated with arthritis/arthralgia. Mayaro virus is a new world virus, associated with the febrile, rheumatic disease characteristic of the old world viruses (Tesh, Watts et al. 1999).

1 Table 1.1 Febrile, rheumatic alphaviral disease in humans.

Virus & Local Distribution Disease outbreaks Disease Symptoms Nomenclature Chikungunya virus 1 Africa, Asia, Indian Ocean Large recurrent epidemics Fever, arthralgia, myalgia, rash, Islands, Central America vomiting Barmah Forest virus2 Australia 500 – 1900 cases per annum Fever, arthralgia, myalgia, rash Ross River virus2 Australia 1500 – 8000 cases per annum Fever, arthralgia, myalgia, rash Oceania 1979/80 epidemic >60 000 cases O'nyong-nyong virus3 Africa, Asia Sporadic epidemics, 30-50 yr Fever, arthralgia, myalgia, rash Igbo Ora virus3 Uganda/Nigeria intervals 400-2 million cases Semiliki Forest virus4 Southern Africa Rare Often asymptomatic or very mild. Fever, arthralgia, headaches. Sindbis virus1,5 Africa, Asia, Australia, & Small seasonal outbreaks, larger Fever (mild), arthralgia, itching rash, Europe outbreaks every 7 years nausea Karelian Fever1, 5 Russia Rare Ockelbo virus1,5 Sweden 31 cases annual average Pogosta virus1,5 Finland 136 (1-1282) cases annual average Mayaro virus 6 Northern South America Small recurrent outbreaks (30-50 Fever (mild), headache, eye pain, cases) arthralgia, myalgia, rash, vomiting 1Suhrbier, Jaffar-Bandjee et al. (2012); 2Suhrbier and La Linn (2004); 3Powers, Brault et al. (2000), Lanciotti, Ludwig et al. (1998); 4Mathiot, Grimaud et al. (1990) 5Hubálek (2008) 6(Tesh, Watts et al. (1999), Powers, Aguilar et al. (2006)

2 1.1.1 Virion and genome

Alphaviruses are approximately 60-70 nm in diameter, spherical and comprised of three layers (Figure 1.1) (Strauss and Strauss 1994, Smith, Cheng et al. 1995, Mancini, Clarke et al. 2000). Envelope proteins E1 and E2 form 240 heterodimers that form 80 trimeric spikes and cover the virion surface. A third envelope protein (E3), is present in the virion of Semliki Forest Virus, but is not a structural component of most mature alphavirus virions (Mancini, Clarke et al. 2000, Jose, Snyder et al. 2009). Each trimeric spike is embedded in a host-derived lipid bilayer and attached, in a one to one ratio via the E2 subunits, to capsid protein monomers underneath. The capsid protein monomers form an icosahedral, shell which directs the spherical shape of the virion and encases the viral genome (Jose, Snyder et al. 2009, Leung, Ng et al. 2011). Together, the capsid protein shell and enclosed genome form the viral nucleocapsid (Strauss and Strauss 1994, Zuckerman, Banatvala et al. 2009). The 6K protein may play a role in genome release and virion assembly (Liljestrom, Lusa et al. 1991, Melton, Ewart et al. 2002, McInerney, Smit et al. 2004). However, the location or presence of the 6K protein in the virion is yet to be detected by cryo-electron microscopy (Jose, Snyder et al. 2009).

Figure 1.1 Sketch diagram of the cryoelectron microscopy image reconstruction of the CHIKV virion. A) Trimeric protein spikes of the virion surface, which are embedded in the underlying host-derived lipid bilayer. Underneath the lipid bilayer is the icosahedral nucleocapsid. B) Cross-sectional view of the virion displaying the embedded heterodimers, and the E2 subunit associating with the icosahedral monomer of the nucleocapsid underneath the lipid bilayer (adapted from Smith, Cheng et al. 1995).

3 The alphavirus genome is a positive sense, single strand RNA, approximately 11.8 kbp in length and resembles eukaryotic mRNA with a 7-methylguanylate cap at the 5’ end and a 3’ polyadenylated tail () (Strauss and Strauss 1994, Fauquet, Mayo et al. 2005). The genome has two open reading frames (ORFs), separated by a promoter for subgenomic RNA. The five structural proteins, E1-E3, capsid protein C and 6K protein are translated from 26S subgenomic RNA encoded by the 3’ ORF. The non-structural proteins 1-4 (nsp 1-4) are translated from genomic RNA encoded by the 5’ ORF and are involved in viral transcription and translation (Fauquet, Mayo et al. 2005, Leung, Ng et al. 2011).

Sub-genomic promoter 5’ Nsp1 Nsp2 Nsp3 Nsp4 6 3’ UTR C E3 E2 K E1 UTR 11.8 kbp Figure 1.2 Schematic Diagram of the CHIKV genome. The 7.6 kbp 5’ end encodes the four non-structural proteins involved in RNA synthesis and replication. The 4.2 kbp 3’ end encodes structural proteins that make up the CHIKV virion (adapted from Faragher, Meek et al. 1998).

1.1.2 Non-structural proteins

The four nsps coalesce to form membrane-associated replication complexes (RCs) but function as individual enzymes with distinct roles in viral transcription and translation (Gould, Coutard et al. 2010, Schwartz and Albert 2010).

Nsp1 contains active guanine-7-methyltransferase and guanylyltransferase domains involved in RNA methylation and capping, are required for translation initiation (Mi and Stollar 1991, Wang, Sawicki et al. 1991, Ahola and Kääriäinen 1995). An amphiphatic helix in the nsp1 structure interacts with cytoplasmic membranes, which triggers the enzymatic activity of nsp1 and also serves to anchor RCs to the cytoplasmic membranes (Ahola, Lampio et al. 1999, Spuul, Salonen et al. 2007).

The nsp2 N-terminus has nucleoside triphosphate activity thought to generate ATP required for nsp2 dsRNA helicase activity (Rikkonen, Peranen et al. 1994, Gomez de Cedrón, Ehsani et al. 1999, Gould, Coutard et al. 2010). At the C-terminus, two domains function together as the nsp2 protease, which is involved in posttranslational modification of the nsps (Russo, White et al. 2006, Pastorino, Peyrefitte et al. 2008).

4

The function of nsp3 is less well understood. While the macro domain within the N- terminal domain of nsp3 is known to have an affinity for ADP-ribose, poly-ADP-ribose and RNA, the functional significance is unclear (Malet, Coutard et al. 2009, Gould, Coutard et al. 2010). The C-terminal domain of nsp3 contains a phosphorylation site, which appears to act in association with nsp2 in the synthesis of negative strand RNA in the early stages of viral replication (Wang, Sawicki et al. 1994, De, Fata-Hartley et al. 2003).

nsp4 functions as the viral RNA-dependant RNA polymerase that is able to switch from anti-sense to sense production of genomic RNA (Lemm, Rümenapf et al. 1994, Rubach, Wasik et al. 2009). Additionally, nsp4 has been found to possess adenylyltransferase activity. Thus, nsp4 may play a role in maintenance of the poly(A) tail of the genome, which is required for initiation of viral replication (Hardy and Rice 2005, Tomar, Hardy et al. 2006).

1.1.3 Infection and replication

Virus particles enter the host cell by receptor mediated endocytosis following interaction between host cell receptor/s and the E2 subunits of the viral envelope (Jose, Snyder et al. 2009, Leung, Ng et al. 2011). The cellular receptor/s that facilitate endocytosis are yet to be clearly defined, although major histocompatibility complex, major laminin receptor (Wang, Kuhn et al. 1992), DC-SIGN and L-SIGN (Klimstra, Nangle et al. 2003), heparin sulphate (Zhang, Heil et al. 2005) and α1β1 integrin collagen receptor (La Linn, Eble et al. 2005) have been listed as potential components of such a receptor. Within the endosome, the acidic pH destabilises the E1/E2 heterodimers, causing E1 homotrimers to form (Leung, Ng et al. 2011). E1 homotrimers then insert into the endosomal membrane allowing the endosomal and viral membranes to fuse (Justman, Klimjack et al. 1993, Spyr, Käsermann et al. 1995).

Fusion of the viral and endosomal membranes releases the nucleocapsid in to the cytoplasm where it is disassembled to expose the viral genome (Figure 1.1) (Jose, Snyder et al. 2009, Leung, Ng et al. 2011). The 6K protein may facilitate membrane fusion and endosomal acidification by forming a cation-selective ion channel that allows an influx of protons into the endosome (Melton, Ewart et al. 2002). SINV has been shown to infect

5 cells without endocytosis and acidification (Paredes, Ferreira et al. 2004) and Kononchik et al., (2011) have proposed an alternative mechanism of alphavirus entry that involves direct entry at the plasma membrane. In this model, a proteinaceous pore composed of virus and host proteins allows the viral nucleocapsid to be released directly into the host cell cytoplasm (Kononchik, Hernandez et al. 2011, Kononchik, Vancini et al. 2011). Once in the cytoplasm, interaction between 60S ribosomal RNA and viral capsid proteins is thought to induce nucleocapsid disassembly (Wengler and Würkner 1992, Leung, Ng et al. 2011). Exposure to low pH following viral-endosomal membrane fusion may also assist in deconstruction of the nucleocapsid (Wengler and Gros 1996, Melton, Ewart et al. 2002).

Following nucleocapsid disassembly, two polypeptide precursors, p123 and p1234, are translated directly from the single strand genomic RNA (Jose, Snyder et al. 2009, Leung, Ng et al. 2011). Translation of the smaller p123 precursor predominates due to an opal stop codon (UAG) located after nsp3 that limits read-through translation (Jose, Snyder et al. 2009, Gould, Coutard et al. 2010). p1234 is cleaved by nsp2 to generate p123 and nsp4 (Fauquet, Mayo et al. 2005, Leung, Ng et al. 2011), which form a negative strand replication complex (-RC-). -RC synthesises a full length, negative strand RNA intermediate (Fauquet, Mayo et al. 2005, Schwartz and Albert 2010). The p123 precursor is processed to produce individual nsp1, nsp2 and nsp3. Nsps1-4 then assemble to form the positive strand RC (RC+)(Fauquet, Mayo et al. 2005, Schwartz and Albert 2010). The RC+ uses the negative strand RNA intermediate as a template for the full length 49S genome to be packaged in mature virions, and 26S subgenomic RNA that is used for translation of the structural protein polypeptide precursor (Figure 1.1) (Fauquet, Mayo et al. 2005, Schwartz and Albert 2010).

6

Figure 1.3 Schematic diagram of alphavirus entry, replication and release from host cells. Adapted from (Jose, Snyder et al. 2009).

1.1.4 Structural proteins

Translation of the 26S subgenomic RNA produces a polypeptide precursor containing the structural proteins in the sequence C-pE2-6K-E1, where C is the capsid protein and pE2 is an E2/E3 precursor (Strauss and Strauss 1994, Fauquet, Mayo et al. 2005). The C-protein is cleaved and exposes an N-terminal signal sequence on the remaining structural polypeptide precursor for translocation to the endoplasmic reticulum (ER) (Jose, Snyder et al. 2009). In the ER, pE2 and E1 undergo a series of modifications that facilitate correct formation of a pE2/E1 heterodimer (Gaedigk-Nitschko and Schlesinger 1990, Jose, Snyder et al. 2009). The 6K protein is believed to play a role in pE2/E1 formation via interaction with E2 subunit of pE2 (Gaedigk-Nitschko and Schlesinger 1990, Jose, Snyder et al. 2009). The E3 subunit of pE2 is cleaved by furin to form the E1/E2 heterodimer, which translocates to the plasma membrane via the Golgi apparatus secretory pathway (Jose, Snyder et al. 2009, Gould, Coutard et al. 2010). 7

The cleaved C-proteins form C-protein dimers (Perera, Owen et al. 2001, Solignat, Gay et al. 2009), which interact with a packaging signal present on viral RNA causing encapsidation of 49S genomic RNA within the cytoplasm and formation of the viral nucleocapsid (Tellinghuisen, Hamburger et al. 1999, Gould, Coutard et al. 2010). The newly assembled nucleocapsids then translocate to the plasma membrane where they bind the cytoplasmic domain of E2 (Strauss, Lenches et al. 2002, Jose, Snyder et al. 2009). This interaction drives the nucleocapsid across the host cell membrane where it acquires the 80 E1/E2 heterodimer trimeric spikes that form the glycoprotein envelope and the host-derived lipid bi-layer as it exits the cell as a mature virion (Figure 1.3) (Jose, Snyder et al. 2009, Gould, Coutard et al. 2010).

1.2 Chikungunya virus

1.2.1 Natural history

Chikungunya virus (CHIKV) is an alphavirus known to cause large outbreaks of febrile, rheumatic disease and is re-emerging as a potential threat to global human health (Suhrbier and Mahalingam 2009, Powers 2010, Suhrbier, Jaffar-Bandjee et al. 2012). CHIKV was first isolated from patients and mosquitoes during an outbreak of dengue-like disease in Newala, Tanzania during 1952-3 (Robinson 1955, Ross 1956). This was the first documented incidence of CHIKV and tribal elders in the region reported no knowledge of the same disease ever having occurred previously (Lumsden 1955). The name ‘chikungunya’ is derived from this initial outbreak; ‘chikungunya’ roughly translates to ‘that which bends up’ in the Makonde language and portrays the bowed posture of people suffering from joint pain that results from CHIKV infection (Robinson 1955, Ross 1956).

Outbreaks of CHIKV disease have occurred at intermittent intervals of 2-50 years and are attributed to two distinct variants that diverged from a common ancestor in the late 19th century – the Asian genotype, first isolated during a 1958 outbreak in Bangkok, and the East/Central African genotype found in Africa and India (ECSA) (Ching and Hapuarachchi 2010, Schwartz and Albert 2010). In Africa, CHIKV is naturally maintained in a zoonotic transmission cycle between non-human primate hosts and forest-dwelling Aedes furcifer or Aedes africanus mosquito vectors (Powers and Logue 2007, Zuckerman, Banatvala et al. 2009). Virus transmission by the urban dwelling Aedes aegypti mosquito is thought to allow divergence from a sylvatic, or forest dwelling, transmission cycle to an

8 urban setting, resulting in sporadic outbreaks of disease (Powers 2010, Simon, Javelle et al. 2011). Within Asia, CHIKV is predominantly transmitted by Aedes aegypti and appears to exist in an urban transmission cycle only, with no known reservoir host (Simon, Javelle et al. 2011). More recently, a new CHIKV variant, derived from the ECSA genotype has evolved to be efficiently transmitted by the Aedes albopictus mosquito (Schuffenecker, Iteman et al. 2006, Tsetsarkin, Vanlandingham et al. 2007).

1.2.2 The A226V mutation and the global dissemination of CHIKV

The largest CHIKV outbreak ever recorded began late 2004 in Kenya before spreading to over 40 countries, including Islands of the Indian Ocean, India, Sri Lanka, South East Asia (Burt, Rolph et al. 2012), the South Pacific (Dupont-Rouzeyrol, Caro et al. 2012) and the Caribbean (Van Bortel, Dorleans et al. 2014). Much of this outbreak has been attributed to the new viral variant within the ECSA strain, which is found to contain an alanine to valine mutation (A226V) within the viral E1 glycoprotein (Schuffenecker, Iteman et al. 2006, Tsetsarkin, Vanlandingham et al. 2007). The A226V mutation allowed efficient replication, dissemination and transmission of CHIKV by Aedes albopictus, previously only a minor vector for CHIKV (Tsetsarkin, Vanlandingham et al. 2007, Simon, Javelle et al. 2011). The Aedes albopictus mosquito is an aggressive mosquito with a widespread global dissemination that has been increasing over the past thirty years (Benedict, Levine et al. 2007). The increased vector competence of Aedes albopictus for the A226V-CHIKV strain has enabled CHIKV to establish autochthonous outbreaks in regions where it has not been seen before, including 205 cases in Italy during 2007, two cases in southern France in 2010 and in March 2011, New Caledonia a total of 24 cases have been confirmed (Rezza, Nicoletti et al. 2007, Gould, Gallian et al. 2010).

The most notable endogenous outbreak in CHIKV-naive regions was a particularly aggressive outbreak in the French overseas department, Reunion Island. During March 2005 and April 2006, an estimated 270 000 people, or one third of the Reunion Island population were infected with CHIKV (Schwartz and Albert 2010). A combination of an immunologically naive population and the adaptation of the A226V-CHIKV strain to the Aedes albopictus vector is thought to have contributed to the intensity of this outbreak, in which 46 000 cases of CHIKV were recorded in one week alone (Pialoux, Gaüzère et al. 2007, Renault, Solet et al. 2007, Taubitz, Cramer et al. 2007, Gould and Higgs 2009). In addition to the explosive incidence of infection, the Reunion Island epidemic was

9 associated with atypical disease manifestations and for the first time, some mortality (Pialoux, Gaüzère et al. 2007). The unusual clinical features observed in Reunion Island were subsequently observed, during the CHIKV outbreak in India. Indian government reports document 1.4 million confirmed cases of CHIKV infection, and no deaths. However, Mavalankar et al., (2007) estimate up to 6.5 million cases to have occurred and as many as 19000 deaths as a result of CHIKV disease.

While the CHIKV Indian Ocean Lineage containing the A226V mutation has facilitated increased spread of CHIKV in some parts of the globe, not all outbreaks of CHIKV are attributable to this recent strain. The recent epidemics in the Caribbean derive from Asian genotypes, while outbreaks in South America derive from the ECSA strain, prior to the A226V mutation. Additionally, outbreaks in East Asia and the subcontinent derive from both the Asian isolate and the A226V mutated Indian Ocean Lineage (Champion, Weaver et al. 2015).

1.2.3 Disease presentation

The classic indicators of CHIKV infection are fever, arthralgia/myalgia and rash. Few cases (less than 5%) are asymptomatic, and symptoms appear after a 3 -12 day incubation period (Jaffar-Bandjee, Ramful et al. 2010, Singh and Unni 2011). CHIKV disease can be divided into an acute and chronic phase.

1.2.3.1 Acute phase The acute phase comprises fever and sudden, incapacitating polyarthritis, and may also include a non-itchy rash, asthenia and headache (Sam and AbuBakar 2006, Jaffar- Bandjee, Ramful et al. 2010). Fever is severe (39-40˚C), lasting 24 -48 hours (Jaffar- Bandjee, Ramful et al. 2010). Arthralgia, like fever, has an abrupt onset and manifests largely in the hands, wrists, feet and ankles, but may also affect larger joints such as the knees and shoulders, generally in a symmetrical distribution (Sam and AbuBakar 2006, Singh and Unni 2011). Three to five days after disease onset, 40-50% of CHIKV patients develop a non-itchy, maculopapular, rash that occurs largely on the trunk and limbs and lasts no more than two days in most cases (Sam and AbuBakar 2006, Simon, Javelle et al. 2011). Manimunda et al (2010) describe the acute phase as the first month of disease. However, in uncomplicated cases, acute symptoms generally resolve within 10 days (Sam and AbuBakar 2006, Simon, Javelle et al. 2011).

10 1.2.3.2 Chronic phase Chronic disease, often marked by a second bout of fever and arthralgia, is more prevalent in adults over 40 years of age (Das, Jaffar-Bandjee et al. 2010, Singh and Unni 2011). A high viral load during the acute phase can be associated with persistent symptoms (Das, Jaffar-Bandjee et al. 2010). However, Chow, Her et al. (2011) found elevated levels of IL-6 and GM-CSF were more likely to be associated with persistent arthralgia than viral load. Chronic symptoms are fatigue and joint pain lasting several months (Manimunda, Vijayachari et al. 2010). However, arthralgia and joint effusions have been reported at 18 months (Jaffar-Bandjee, Ramful et al. 2010, Simon, Javelle et al. 2011) and long term sequale such as depression and insomnia are common (De Andrade, Jean et al. 2010, Gérardin, Fianu et al. 2011, Couturier, Guillemin et al. 2012). Manimunda (2010) reported 36% of chronic CHIKV patients met the 1987 American College of Rheumatology criteria for rheumatoid arthritis (RA) and there is some evidence that CHIKV induces RA-like illness (Bouquillard and Combe 2009). Imaging studies provide evidence of erosive arthritic lesions consistent with RA in CHIKV patients (Malvy, Ezzedine et al. 2009, Manimunda, Vijayachari et al. 2010, Mizuno, Kato et al. 2011), and some CHIKV patients respond to disease modifying anti-rheumatic drugs, such a methotrexate and sulfasalazine (Pandya 2008, Bouquillard and Combe 2009, Ganu and Ganu 2011). However, further studies are required before it can be determined whether CHIKV results in an RA-like illness, causes RA or simply brings to light an underlying condition.

1.2.3.3 Atypical disease and mortality The most recent CHIKV epidemic was associated with particularly aggressive disease in terms of the infectivity, and the nature and severity of clinical manifestations. Vertical transmission of CHIKV infection was observed (Ramful, Carbonnier et al. 2007, Jaffar-Bandjee, Ramful et al. 2010, Mangalgi, Shenoy et al. 2011) when CHIKV infection coincided with the perinatal period, prior to maternal antibody production and could not be prevented by caesarean section (Couderc and Lecuit 2009, Jaffar-Bandjee, Ramful et al. 2010). Neonatal manifestations of CHIKV infection included febrile convulsions, severe anorexia and a characteristic perioral hyperpigmentation (Gérardin, Barau et al. 2008, Valamparampil, Chirakkarot et al. 2009, Mangalgi, Shenoy et al. 2011). In older children, in whom CHIKV infection is generally mild, febrile seizures, retro-orbital pain and skin blistering that required dressing changes under general anaesthesia (Robin, Ramful et al. 2008, Sebastian, Lodha et al. 2009, Valamparampil, Chirakkarot et al. 2009). Death in

11 neonates and children was rare (Couderc and Lecuit 2009, Jaffar-Bandjee, Ramful et al. 2010).

In adults, atypical symptoms and mortality were associated with increasing age and underlying prior health conditions (Economopoulou, Dominguez et al. 2009, Das, Jaffar- Bandjee et al. 2010). A total of 254 deaths in Reunion Island and an estimated 19000 deaths in India were reported during 2005-7 (Mavalankar, Shastri et al. 2007, Pialoux, Gaüzère et al. 2007). Severe atypical manifestations included encephalitis, organ failure (Economopoulou, Dominguez et al. 2009) and transient blindness was also observed in the recent epidemic (Mahendradas, Ranganna et al. 2008). Mouth and genital ulcers, stomatitis and exfoliative dermatitis were frequent atypical manifestations (Inamadar, Palit et al. 2008). Atypical symptoms generally resolved completely, although some patients retained skin hyperpigmentation (Jaffar-Bandjee, Ramful et al. 2010).

1.2.4 Diagnosis of CHIKV infection

Clinical diagnosis of CHIKV infection is based on the ‘classic triad’ of presenting signs and symptoms: fever, arthralgia and rash, and during an epidemic, a clinical diagnosis may be adequate (Kalantri, Joshi et al. 2006). However, the recent increase in the endemicity of CHIKV, particularly to regions that harbour other pathogens that result in similar symptoms, such as malaria and dengue, or closely related viruses like Ross River virus, and in travellers returning from such regions, laboratory confirmation is required to exclude differential diagnoses and direct the appropriate treatment regime.

Enzyme–linked immunosorbance assay (ELISA) or immunofluorescence is used to detect IgM antibodies or demonstrate an increase in IgG antibodies (Kalantri, Joshi et al. 2006). IgM appears 4-5 days after disease onset and may not be present at detectable levels within the first week of disease (Taubitz, Cramer et al. 2007, Panning, Grywna et al. 2008). Demonstration of rising IgG titres requires collection of paired serum samples, taken at acute and convalescent stages of disease (Kalantri, Joshi et al. 2006). In populations where other alphaviruses are prevalent and may provide cross reactive false positives, detection of CHIKV by PCR may be more suitable (Kalantri, Joshi et al. 2006, Taubitz, Cramer et al. 2007). Recently however, an epitope-blocking ELISA, which uses a monoclonal antibody targeting an eptiope in the CHIKV E2 protein, has been developed and is not cross-reactive with the closely related Ross River virus and Barmah Forest virus

12 (Goh, Kam et al. 2015). CHIKV can be detected in patient blood from disease onset (Taubitz, Cramer et al. 2007, Panning, Grywna et al. 2008) and from tissues in patients (Simon, Javelle et al. 2011) and non-human primates (Labadie, Larcher et al. 2010) well after viraemia has passed, providing a wide timeframe in which to confirm CHIKV infection. Real time PCR is a highly sensitive technique for detection of viral RNA, with Edwards et al., (2007) reporting effective detection from as few as 20 RNA transcripts. Virus culture from tissue samples can be used to provide reference standards, but is time consuming, and within Australia, can only be performed in Biosafety Level 3 laboratories and is thus not appropriate for a clinical setting (Kalantri, Joshi et al. 2006).

1.2.5 Treatment of CHIKV disease

The treatment strategy for CHIKV is symptomatic; paracetamol or other non- salicylate analgesia and non-steroidal anti-inflammatory drugs (NSAIDs) are generally used in the acute phase of the disease (Pialoux, Gaüzère et al. 2007, W.H.O 2009). Aspirin is avoided due the potential for haemorrhage, and steroidal anti-inflammatory drugs are avoided due to their immunosuppressive effects and the potential exacerbation of symptoms (W.H.O 2009, Singh and Unni 2011). However, Padmakumar (2009) found that a treatment regimen of NSAIDs and low doses of prednisone (a corticosteroid) produced the most favourable pain relief outcomes in acute CHIKV disease patients. Chronic arthralgia is treated in the same manner as the acute phase of the disease. However, long term use of NSAIDs is not recommended (W.H.O 2009).

Several drugs have been explored for use as prophylactics, or disease modification. The antimalarial, chloroquine provided promising results in an early clinical trial (Brighton 1984). However, later studies indicated chloroquine may enhance viral replication and acute disease (Maheshwari, Srikantan et al. 1991) and was associated with a greater incidence of chronic arthritic symptoms (Lamballerie, Boisson et al. 2008). Disease modifying anti-rheumatic drugs (DMARDs) have been used in the treatment of CHIKV arthritis with varying success (Pandya, 2008, Bouquillard and Combe 2009, Ganu and Ganu, 2009, Javelle, Ribera et al., 2015) and with methotrexate treatment found to be effective in treating patients with a specific RA-like illness (Pandya 2008, Bouquillard and Combe 2009, Ganu and Ganu 2011, Javelle, Ribera et al., 2015). The antiviral drug Ribavirin works synergistically with IFNα to effectively inhibit CHIKV replication (Scheidel and Stollar 1991). However, this therapy is expensive, requires parenteral injection and is therefore not suitable for large scale use (Gould, Coutard et al. 2010). Furthermore,

13 Ribavirin treatment only provided a mild improvement in symptoms (Ravichandran and Manian 2008). Bindarit, a macrophage chemoattractant protein-1 (MCP-1) inhibitor, ameliorates inflammatory disease in mouse models of RRV and CHIKV, and may be a candidate for treatment of acute disease in humans (Rulli, Guglielmotti et al. 2009, Rulli, Rolph et al. 2011). Studies of neutralising monoclonal anti-CHIKV antibodies have demonstrated effective attenuation of viraemia and disease in knockout and neonatal mouse models of CHIKV (Couderc, Khandoudi et al. 2009, Fric, Bertin-Maghit et al. 2013, Pal, Dowd et al. 2013). This therapy may prove useful for prevention of peri-natal maternal transfer (Gould, Coutard et al. 2010, (Warter, Lee et al. 2011, Fric, Bertin-Maghit et al. 2013).

1.2.6 Vaccine development

No licensed vaccine currently exists for any alphaviruses. Early, formalin- inactivated and tween-ether treated live-attenuated vaccines, while highly immunogenic and easily prepared, were not developed further due to safety issues (Harrison, Binn et al. 1967, Eckels, Harrison et al. 1970). Later investigation of two live attenuated strains produced by serial, plaque to plaque passaging in MRC5 cells yielded the TSI-GSD-218 vaccine candidate, which progressed to a Phase II clinical trail in humans. However, this approach was abandoned after 5 of 59 trial participants developed a transient arthralgia (Edelman, Tacket et al. 2000). Muthumani et al., (2008) developed a synthetic DNA vaccine incorporating antigenic regions of the CHIKV genome, which has demonstrated protection against CHIKV disease in mice and induced protective neutralising antibody titres in non-human primates (Mallilankaraman, Shedlock et al. 2011). However, it must be delivered by concurrent intramuscular injection and electroporation.

In 2008, Wang et al. (2008) developed three chimeric vaccines comprising the CHIKV structural proteins antigens and a genetic backbone from an attenuated EEEV strain, Sindbis virus or the attenutated VEEV strain, TC-83. The TC-83 chimera produced robust antibody responses, were non-pathogenic and protective against CHIKV in type I IFN-deficient mice (Wang, Kim et al. 2011), and displayed low transmission potential in Aedes mosquitoes (Darwin, Kenney et al. 2011). However, the propensity of alphaviruses to reform into replication competent viruses hindered the development these chimeric vaccines for use in humans. Similarly, a recombinant adenovirus vector vaccine encoding the CHIKV structural proteins, has also been shown to induce protective antibody

14 responses in wild type mice (Wang, Suhrbier et al. 2011). Replacement of the subgenomic promotor of a CHIKV clone with the internal ribosomal entry site (IRES) of encephalomyocarditis virus has produced a highly attenuated vaccine that confers immunogenicity in WT and IFN-deficient mice after a single dose. The IRES-attenuated vaccine is also unable to infect or replicate in mosquito cells and is currently being tested in non-human primates as a precedent for human trials (Plante, Wang et al. 2011, Roy, Adams et al. 2014).

A virus-like particle (VLP) vaccine that elicited protective humoral immune responses in non-human primates developed in 2010 by Akahata, Yang et al. (2010) has recently been used in a Phase I clinical trial involving 25 participants. The vaccine, VRC- CHKVLP059-00-VP, was immunogenic and well tolerated with neutralising antibody titres detectable 6 months after the third and final vaccination in all dosage groups (Chang, Dowd et al. 2014). A live recombinant measles virus-based vaccine has also been found to be well tolerated in a Phase I clinical trial involving 42 participants, with neutralising antibodies detectable at day 28 post vaccination and immunogenicity unaffected by anti- measles vector immunity (Ramsauer, Schwameis et al. 2015).

1.3 Immunobiology of chikungunya virus

1.3.1 Disease models

Non-human primates provide insight in to the pathogenesis of CHIKV that is more comparable to humans, but are limited in their utility due to cost, housing and training requirements. Early studies of CHIKV pathology used cercopithecus, Rhesus (macaca mulatta) and Bonnet (macaca radiata) monkeys and also baboons (Kam, Ong et al. 2009). A rhesus model was also used to investigate the efficacy of a formalin-inactivated CHIKV vaccine against different CHIKV strains (Harrison, Binn et al. 1967). More recently, intradermal inoculation of cynomologous monkeys (macaca fascicularis), has produced a CHIKV disease model that closely resembles the viremic, clinical and pathogenic pattern of CHIKV infection in humans, including chronic disease (Labadie, Larcher et al. 2010). Tree shrews, a non-rodent primate like squirrel, have been used as a small animal alternative to non-human primates. However, these are susceptible to psychosocial stress and depression unless they are maintained in specialised housing that negates the cost and space benefit of their smaller size (Cao, Yang et al. 2003, Kam, Ong et al. 2009).

15

The first animal of model of CHIKV disease was used to isolate virus from Albino Swiss mice inoculated with sera from an infected human (Ross 1956). Older mice were resistant, while neonatal mice succumbed by day 4-6 post infection (Ross 1956). Subsequent models have used neonatal and young mice, to demonstrated tissue tropism similar to that seen in human infection, with CHIKV detected in muscle, joints, skin and brain (Couderc, Chrétien et al. 2008, Ziegler, Lu et al. 2008). Using type I IFN receptor α chain (IFNAR1) knock out mice, Couderc et al., (2008) produced an adult model of CHIKV disease that resembled the timing and tropism of CHIKV infection in humans. However, in all of these models, CHIKV infection is these mice was either very severe, or lethal. The application of a lethal knockout model, or the use of very young mice for investigation of a chronic arthralgia not normally present in children, has limited application. In 2010, Gardner, Anraku et al. (2010) developed an adult, wild type, immunologically intact, non- lethal mouse model of CHIKV disease. This model mirrors the pathogenesis and arthritic symptoms of acute CHIKV disease (Gardner, Anraku et al., 2010) and the gene expression profile of acute and chronic disease seen in humans (Wilson, Poo et al. in prep). However, this model requires virus injection into the foot, which does not mimic mosquito transmission, and is not known to induce the polyarthritis or recurrent arthritis typical of human CHIKV disease (Teo, Lum et al. 2012).

1.3.2 CHIKV cell and tissue tropism

CHIKV has been detected in skin, joints, muscle, brain, lymphoid organs and the liver where it appears to infect the connective tissues and epithelium (Couderc and Lecuit 2009). In muscle, CHIKV infects satellite cells in the basement membrane and the fascia that converges to form tendons, but not myocytes themselves (Ozden, Huerre et al. 2007, Couderc, Chrétien et al. 2008, Labadie, Larcher et al. 2010). In contrast to the neurotropic alphaviruses that infect neurons and vasculature of the brain, CHIKV infects the epithelial cells of the choroid plexus and the ependymal walls (Couderc, Chrétien et al. 2008, Couderc and Lecuit 2009, Labadie, Larcher et al. 2010). In skin, dermal fibroblasts and possibly keratinocytes are infected and in joints, CHIKV infects the joint capsule and periosteum (Couderc, Chrétien et al. 2008, Couderc and Lecuit 2009, Labadie, Larcher et al. 2010). In vitro analysis of human primary and immortalised cell lines, CHIKV was found to infect and replicate primarily fibroblasts, epithelium, endothelium (Sourisseau, Schilte et al. 2007). Haematopoietic cells, except monocyte derived-macrophages, have been reported not to be permissive to CHIKV infection (Sourisseau, Schilte et al. 2007, Solignat,

16 Gay et al. 2009). However, Her, Malleret et al. (2010) found that human peripheral blood monocytes (PBMCs) were a major target of CHIKV during the acute phase of infection, and in macaques, CHIKV was found predominantly in splenic macrophages, and the resident hepatic macrophages, Kupffer cells (Labadie, Larcher et al. 2010).

1.3.3 CHIKV pathogenesis

CHIKV enters the body following via the bite of an infected mosquito (Sourisseau, Schilte et al. 2007, Her, Malleret et al. 2010, Pakran, George et al. 2011). The virus undergoes local replication at the site of infection, then migrates to the draining lymph nodes, enters the lymph circulation and is released into the blood stream (Kam, Ong et al. 2009, Schwartz and Albert 2010). CHIKV disseminates through the blood to the liver, muscle, joints and in some cases the central nervous system, with serum titres peaking around day 6 post infection (Kam, Ong et al. 2009, Schwartz and Albert 2010). Viraemia lasts 5-7 days, but reaches very high levels, frequently over 109 virus particles/ml (Jaffar- Bandjee, Das et al. 2009). The peak viraemia at day 2-4 post infection precedes a peak in serum type I interferon (IFN) that ranges between 0.5-2 ng/ml (Ng, Chow et al. 2009, Schwartz and Albert 2010).

Acute CHIKV induces a pro-inflammatory cytokine response, predominantly involving type I IFNs, IFNγ, IL-6, IL-12, TNF-α and MCP-1 (Jaffar-Bandjee, Das et al. 2009, Ng, Chow et al. 2009, Chaaitanya, Muruganandam et al. 2011). Many of these inflammatory responses remain present in chronic disease, with elevated MCP-1, IL-6 and IL-8 detected in synovial tissues at 18 months (Hoarau, Jaffar Bandjee et al. 2010), and MCP-1, IL-6 and IL-8 and MIP-1 at 10 months post disease onset (Chaaitanya, Muruganandam et al. 2011) in synovial samples. Kelvin et al., (2011) also found increased serum IL-8 expression in patients at 12 months post disease onset, and elevated CXCL9 and IP-10 levels in these patients was correlated with raised IgG and increased severity of symptoms (Kelvin, Banner et al. 2011).

1.3.4 Persistence

CHIKV chronic arthritis/arthralgia is thought to be the result of persistent virus or viral products within joint tissues that maintain a protracted inflammatory response (Suhrbier and Mahalingam 2009, Suhrbier, Jaffar-Bandjee et al. 2012). At 18 months post disease onset, human synovial fluid was found to contain monocytes/macrophages with

17 inclusion bodies indicating viral replication (Hoarau, Jaffar Bandjee et al. 2010). In WT mice, infectious virus can be recovered up to day 14, and viral RNA can be isolated up to day 100 (Poo, Rudd et al. 2014). In macaques, infectious virus could be isolated from tissues up to 44 days post infection, while CHIKV RNA was detected up to 55 days post CHIKV infection (Labadie, Larcher et al. 2010). Macrophages have been identified as a major reservoir for this persistent virus, with CHIKV RNA detected in splenic macrophages up to 90 days post infection (Labadie, Larcher et al. 2010), and macrophage cultures have been found to harbour RRV over a period of 148 days, (Soden, Vasudevan et al. 2000, Way, Lidbury et al. 2002). Viral mRNA in these cultures was seen to subside and re- emerge to detectable levels, sporadically and via stress induction, resulting in cytopathic effect and suggesting that alphaviruses may be able to persist latently and reinfect at times when the host immune system is compromised (Way, Lidbury et al. 2002).

1.3.5 Immune control of chikungunya virus

Antibody responses and type I IFNs act to clear CHIKV viraemia usually within 5-7 days (Schwartz and Albert 2010). Type I IFN levels peak shortly after the peak in viraemia, and subside before the appearance of neutralising antibodies (Jaffar-Bandjee, Ramful et al. 2010). Virus specific IgM and IgG antibodies can be detected at 3-5 days and 5-7 days post infection, respectively (Couderc, Chrétien et al. 2008, Panning, Grywna et al. 2008, Schwartz and Albert 2010). While CHIKV infection of mice lacking the type I IFN receptor, is lethal (Couderc, Chrétien et al. 2008), these mice are capable of producing antibodies and can be protected by vaccination (Wang, Kim et al. 2011). Administration of anti-CHIKV immunoglobulins within the first 24 hours of infection is also able to protect these mice (Couderc, Khandoudi et al. 2009). However, type I IFNs have been shown to enhance humoral antiviral responses, and may be required for optimal antibody responses (Fink, Lang et al. 2006, Thompson, Whitmore et al. 2008).

CHIKV induces a profound, predominantly monocyte/macrophage infiltrate within the joints in both human patients and animal models of CHIKV disease (Couderc, Chrétien et al. 2008, Gardner, Anraku et al. 2010, Rulli, Rolph et al. 2011). Severe joint swelling in CHIKV-infected TLR3-/- (Her, Teng et al. 2014) and CCR2-/- (Poo, Nakaya et al. 2014) mice has been associated with increased neutrophil infiltrates. CD8+ T-cells have not previously been thought to have a major role in alphaviral infection (La Linn, Mateo et al. 1998, Smith-Norowitz, Sobel et al. 2000). However, recently, CD8+ T-cells have been

18 found to have a role in the control of RRV in musculoskeletal tissues (Burrack, Montgomery et al. 2015), and CHIKV in humans has been shown to activate a strong CD8+ T-cell response early in CHIKV infection (Wauquier, Becquart et al. 2011). CD4+ T- cell responses predominate later in disease (Wauquier, Becquart et al. 2011), and have been shown to contribute to CHIKV immunopathology (Nakaya, Gardner et al. 2012, Teo, Lum et al. 2013), independent of IFNγ, and have a minor role in control of CHIKV infection (Poo, Rudd et al. 2014). Neutrophils do not appear to play a major role in CHIKV disease (Alsharifi, Lobigs et al. 2006, Petitdemange, Becquart et al. 2011, Poo, Rudd et al. 2014), although severe joint swelling in CHIKV-infected TLR3-/- (Her et al. 2014) and CCR2-/- (Poo, Nakaya et al. 2014) mice has been associated with increased neutrophil infiltrates. Recently, NK cell activity during the early-acute phase of CHIKV (day 1-3 post infection) has been shown to mediate more severe CHIKV oedema and arthritis seen in the late acute phase (day 6-8 post infection; (Teo, Her et al. 2015).

1.4 The type I IFN response and RNA viruses

Type I IFNs include IFNβ, IFN-δ, IFN-ε, IFN-κ, IFN-τ and IFN-ω and the thirteen IFNα subtypes (Platanias 2005, Trinchieri 2010). IFN-δ and IFN-τ are only found in pigs and cattle/sheep, respectively (Perry, Gang et al. 2005, Platanias 2005). As the name suggests, IFNs act to ‘interfere’ with viral replication, but are secreted in response to most pathogen infections, providing one of the first lines of defence in the innate immune response, as well as activating adaptive immunity (Perry, Gang et al. 2005, Sadler and Williams 2008). The IFNα and IFNβ subtypes, are the most widely expressed and studied of the type I IFNs (Trinchieri 2010). All type I IFNs signal through a common receptor known as the type I IFN receptor (IFNAR). (Perry, Gang et al. 2005, Trinchieri 2010). IFNAR signalling promotes further production of type I IFNs, as well as IFN regulated genes (IRGs) (detailed in section 1.4.1).

Initiation of the type I IFN response to infection relies on the recognition of molecular structures, collectively termed pathogen-specific molecular patterns (PAMPs), that are not presented on healthy host cells (Kumagai, Takeuchi et al. 2008, Yoneyama and Fujita 2010). Detection of PAMPs is achieved by pattern recognition receptors (PRRs), either membrane bound toll-like receptors (TLRs), or non-TLR receptors, located within the cytoplasm (Perry, Gang et al. 2005, Kumagai, Takeuchi et al. 2008). Binding of PRRs to their respective ligands recruits adaptor molecules present within the cytoplasm,

19 and activates signalling cascades leading to transcription of type I IFNs and inflammatory mediators (Perry, Gang et al. 2005, Kumagai, Takeuchi et al. 2008).

1.4.1 Toll-like receptors

TLRs are highly conserved transmembrane glycoproteins that recognise specific PAMPs (Kawai and Akira 2010). TLRs have an extracellular leucine-rich repeat (LRR) motif involved in ligand binding, a transmembrane region and the Toll/Interleukin-1 receptor homology (TIR) domain in the cytoplasm involved in signalling (Kawai and Akira 2010).

The TLRs involved in detection of RNA viruses are TLR3, 7 and 8. TLR3 is expressed both intra- and extra-cellularly where it detects dsRNA viruses and the dsRNA replication intermediates of ssRNA viruses (Kumagai, Takeuchi et al. 2008, Jensen and Thomsen 2012). TLR7 and 8 sense ssRNA viruses, as well as small interfering RNA (siRNA) and are phylogenetically and functionally related (Bauer, Pigisch et al. 2008). However, TLR7 is expressed within pDCs and B cells, while TLR8 is expressed predominantly in conventional DCs (cDCs) and monocytes (Heil, Hemmi et al. 2004, Bauer, Pigisch et al. 2008). Both reside within the ER and translocate to endosomes or lysosomes where they detect ssRNA (Diebold, Kaisho et al. 2004, Kato, Sato et al. 2005, Wang, Liu et al. 2006, Mandl, Akondy et al. 2011, Jensen and Thomsen 2012). TLR2 and 4 are generally implicated in sensing bacterial pathogens by recognition of lipoproteins and lipopolysaccharide (LPS), respectively (Yoneyama and Fujita 2010, Kawasaki, Kawai et al. 2011). However, they have been shown to play a role in virus recognition by sensing certain viral proteins (Kurt-Jones, Popova et al. 2000, Bieback, Lien et al. 2002, Finberg and Kurt-Jones 2004, Triantafilou and Triantafilou 2004, Zhou, Halle et al. 2008).

Alphaviruses are ssRNA viruses that produce dsRNA during replication, thus TLR3, 7 and 8 are implicated in alphavirus detection (Schwartz and Albert 2010). Mice deficient in TLR3 are more susceptible to Semliki Forest virus (Schulz, Diebold et al. 2005) and CHIKV (Her, Teng et al. 2014) compared to WT mice. RRV is known to cause more severe disease in TLR7-/- mice (Neighbours, Long et al. 2012), and Sindbis virus has been shown to engage TLR7 signalling in mice and microglia with disrupted TLR7 signalling (Esen, Blakely et al. 2012). However, the absence of either TLR3 or 7 does not render the host wholly susceptible to alphavirus infection, and it is has been suggested

20 that the different TLRs may compensate for each other, and may also work in conjunction with PRRs in the cytoplasm (Martinon and Tschopp 2005, Esen, Blakely et al. 2012).

1.4.2 Cytoplasmic PRRs

In addition to TLRs, RNA viruses can be detected by PRRs in the cytoplasm. Cytoplasmic PRRs include the retinoic acid inducible gene-1 (RIG-1)-like helicases (RLRs) and NOD-Like receptors (NLRs) (Kawasaki, Kawai et al. 2011, Jensen and Thomsen 2012). NLRs cover a broad array of receptors within the cytoplasm that respond to PAMPs, non-PAMP structures and cellular stress (Wilkins and Gale Jr 2010, Jensen and Thomsen 2012). RLRs are RNA helicases and include RIG-1, Melanoma Differentiation Associated Protein 5 (MDA5) and Laboratory of Genetics and Physiology 2 (LGP2) (Yoneyama, Kikuchi et al. 2005, Wilkins and Gale Jr 2010, Jensen and Thomsen 2012). RLRs are found in the cytoplasm of most cells where they sense viral RNA and viral replication products (Yoneyama and Fujita 2010, Jensen and Thomsen 2012). All RLRs posses a C-terminal regulatory domain involved in ligand binding and a central DexD/H box helicase domain with an ATP-binding motif. However, only RIG-1 and MDA5 posses two N-terminal caspase recruitment domain (CARD)-like domains involved in signal transmission (Yoneyama, Kikuchi et al. 2005, Jensen and Thomsen 2012).

Both MDA5 and RIG-1 have also been implicated in the detection of alphaviruses; Basler and Garcia-Sastre (2007) suggest MDA5 is the primary RLR involved in alphavirus detection by sensing dsRNA replication intermediates, rather than the ssRNA genome. MDA5 has been shown to be important in the host response to salmon alphavirus in studies of rainbow trout (Chang, Collet et al. 2011). Experimentally-induced over- expression of RIG-1 was also found to prevent viral replication of VEEV in the human embryonic kidney (HEK293) cell line (Wu, Huang et al. 2008). MDA5 and RIG-1 signal through the adaptor molecule IPS-1 (see section 1.4.3.3), which has been found to play an important role in initiation of the type I IFN response to CHIKV (Schilte, Couderc et al. 2010, Schilte, Buckwalter et al. 2012, Rudd, Wilson et al. 2012), and is discussed further Chapter 2.

1.4.3 Signal transduction by PRRs

Ligand binding causes the cytoplasmic regions of PRRs to associate with adaptor molecules within the cytoplasm, which subsequently direct signalling cascades that result in production of type I IFNs and pro-inflammatory mediators (Slack, Schooley et al. 2000,

21 Kawai and Akira 2010). Four adaptor molecules play a major role in directing TLR signalling: myeloid differentiation factor-88 (MyD88), TIR-associated protein/MyD88 adaptor like (TIRAP/MAL), TIR-domain containing adaptor inducing IFNβ (TRIF/TICAM-1) and TRIF-related adaptor molecule (TRAM) (Perry, Gang et al. 2005, Kumagai, Takeuchi et al. 2008). IFNβ-promoter stimulator (IPS-1) and TNF receptor-associated death domain (TRADD) serve as adaptor molecules for MDA5/RIG-1 signalling (Kawasaki, Kawai et al. 2011, Jensen and Thomsen 2012). MyD88, TRIF, and IPS-1 are pivotal members of three major signalling pathways (Kawai and Akira 2010, Kawasaki, Kawai et al. 2011, Jensen and Thomsen 2012). Details of the three major signalling pathways, and major components involved are introduced below.

1.4.3.1 MyD88-dependant signalling pathway The MyD88 signalling pathway is used by all TLRs with the exception of TLR3, which signals exclusively through TRIF (Yamamoto, Sato et al. 2003, Kawai and Akira 2010). Activated MyD88 recruits IL-1 receptor associated kinase 4 (IRAK4), IRAK1, IRAK2 and IRAK-M (Kawai and Akira 2010). IRAK activation allows interaction with TNF- receptor associated factor 6 (TRAF6) allowing it to become ubiquinated (Kawai and Akira 2010). Ubiquinated TRAF6 then phosphorylates transforming growth factor-β-activated kinase (TAK-1) (Kawai and Akira 2010). TAK-1 activates the mitogen activated protein kinase (MAPK) pathway resulting in activation of the transcription factor activator protein-1 (AP-1) (Kawai and Akira 2010, Kawasaki, Kawai et al. 2011). Activated TAK-1 can also activate the canonical NFκB signalling pathway by recruiting NFκB essential modulator (NEMO) of the IκB kinase (IKK) complex (Doyle and O’Neill 2006, Yoneyama and Fujita 2010). This results in phosphorylation of the IκB inhibitor by the proteasome and subsequent release and translocation of NFκB to the nucleus (Doyle and O’Neill 2006, Boyd 2012). Activation of NFκB and AP-1 through the MyD88 pathway induces the production of various inflammatory mediators (Kawai and Akira 2010, Boyd 2012). In plasmacytoid dendritic cells (pDCs), MyD88 activation of TRAF6, through TLR 7, leads to recruitment of TRAF3, which activates the transcription factor IRF7 to induce production of IFNαs, specifically (Kawai, Sato et al. 2004, Ning, Pagano et al. 2011).

1.4.3.2 TRIF-dependant signalling pathway Upon TLR3 binding, TRIF is activated to employ TRADD, TRAF6 and TRAF3 (Perry, Gang et al. 2005, Kawasaki, Kawai et al. 2011). TRIF possess a C-terminal

22 receptor interacting protein (RIP) homotypic interacting domain, a TRAF6 binding domain and a toll/interleukin-1 receptor (TIR) domain that interacts with the TIR domain of TLR3/4 (Perry, Gang et al. 2005). Recruitment of TRADD allows interaction between RIP-1 and the RIP domain of TRIF, via interaction with pellino-1, an E3 ubiquitin ligase (Perry, Gang et al. 2005, Kawai and Akira 2010). Activation of RIP-1, and direct recruitment of TRAF6 by TRIF, both result in phosphorylation and activation of TAK-1 (Kawai and Akira 2010, Yoneyama and Fujita 2010). As described for MyD88 signalling, TAK-1 activation leads to either NFκB or MAPK signalling and downstream transcription of pro-inflammatory mediators by NFκB and AP-1, respectively (Yoneyama and Fujita 2010, Kawasaki, Kawai et al. 2011). TRAF3 recruitment by TRIF activates TBK1 and IKKε, which act as kinases for IRF3 and IRF7, leading to production of type I IFNs (Paz, Sun et al. 2006, Yoneyama and Fujita 2010).

1.4.3.3 IPS-1-dependant signalling pathway Both RIG-1 and MDA5 signal through the adaptor protein IPS-1, also known as mitochondrial antiviral signalling protein (MAVS), caspase activation and recruitment domain-containing adapter protein (CARDIF) and virus-induced signalling adaptor (VISA) (Wilkins and Gale Jr 2010, Jensen and Thomsen 2012). IPS-1 is a mitochondrial membrane bound protein that recruits RIG-1 and MDA5, following ligand binding, as well as recruiting the adaptor molecule, TRADD (Seth, Sun et al. 2005, Wilkins and Gale Jr 2010). IPS-1 is therefore believed to act by facilitating the interaction between RIG- 1/MDA5 and TRADD at the mitochondrial membrane (Jensen and Thomsen 2012). This interaction activates TRAF3 and consequently TBK1-IKKε, resulting in phosphorylation of IRF3 and IRF7 in the same manner described for the TRAF3 arm of the TRIF-signalling pathway (Paz, Sun et al. 2006, Kawai and Akira 2010, Yoneyama and Fujita 2010). IPS- 1/TRADD can also recruit Fas-receptor activated death domain (FADD) in a complex with RIP-1 (Yoneyama and Fujita 2010, Jensen and Thomsen 2012). This activates TRAF6 and NFκB, in the same manner described for described for TRIF signalling, leading to transcription of IFNα/β (Yoneyama and Fujita 2010, Kawasaki, Kawai et al. 2011, Jensen and Thomsen 2012).

1.4.4 Activation of the IFNβ promoter

Transcription factors, activated by PRR signalling, translocate to the nucleus (Kawasaki, Kawai et al. 2011). Here, they form transcription complexes with co-activators,

23 which then bind the virus response elements of the IFNα and IFNβ promoters to initiate IFNα/β production (Balachandran and Beg 2011, Ning, Pagano et al. 2011). IRF3, IRF7, NFκB, AP-1 and p300/CREB-binding protein form a complex – the “enhancesome” – and cooperatively bind the enhancer region upstream of the IFNβ promoter to promote the transcription of IFNβ (Balachandran and Beg 2011, Ning, Pagano et al. 2011). IRF3 and 7 can also form a ternary complex, virus activated factor (VAF), which has a weak affinity for the IFNβ and IFNα4 promoters (Ning, Pagano et al. 2011).

1.4.5 Type I IFN signalling and amplification.

Following activation of the IFNβ promoter, the resulting secreted IFNβ binds the IFNAR in an autocrine and/or paracrine manner (Platanias 2005, Yoneyama and Fujita 2010). The IFNAR is made up of two subunits, IFNAR1 and IFNAR2, the cytoplasmic domains of which are associated with Janus activated kinase1 (JAK1) and tyrosine kinase 2 (TYK2), respectively (Platanias 2005). Binding of the IFNAR1 induces auto- phosphorylation of JAKs in the cytoplasmic domains and subsequent phosphorylation of signal transducer and activator of transcription 1 (STAT1) and STAT2. STAT1/2 then forms a complex with IRF9, called IFN-stimulated gene factor 3 (ISGF3). ISGF3 translocates to the nucleus where it binds IFN-sensitive response elements (ISRE) in DNA to induce IFN regulated genes (IRG) transcription (Figure 1.2) (Platanias 2005, Yoneyama and Fujita 2010).

24

Figure 1.4 Schematic diagram of the type I IFN signalling pathway and positive feedback loop.

25 In addition to antiviral effectors (described in 1.4.6), IFNβ induced transcription TLRs, RIG-1 and MDA5 as well as IRF7 and the IFNαs. This serves to generate a positive feedback loop that greatly amplifies the type I IFN response to pathogen infection (Brierley and Fish 2002, Trinchieri 2010). The increased expression of PRRs enhances host cell capacity to detect viral infection, while secreted IFNαs signal back through the IFNAR1 to further promote ISG production (Sadler and Williams 2008, Yoneyama and Fujita 2010). Vital to the amplification of the type I IFN response is the up-regulation of IRF7 expression (Honda, Yanai et al. 2005, Ning, Pagano et al. 2011). In unstimulated cells, IRF7 is expressed only at low levels (with the exception of pDCs), and only in lymphoid cells (Ning, Pagano et al. 2011). During viral infection, increased IRF7 levels activate the IFNα promoter to vastly increase the production of IFNαs and subsequently, production of IRGs (Figure 1.2) (Honda, Yanai et al. 2005, Ning, Pagano et al. 2011).

1.4.6 Interferon regulated genes

The more well characterised IRGs include myxovirus resistance proteins (Mx), which form oligomers around viral nucleocapsids and other viral components, before degrading them (Gao, von der Malsburg et al. 2011, Haller and Kochs 2011); the 2’5’ oligoadenylate synthetases (2’5’OAS), which activate RNAseL to degrade mRNA within the cell making it detectable by RIG-1 and MDA (Malathi, Dong et al. 2007, Silverman 2007); interferon stimulated gene of 15kDa (ISG15), which is an ubiquitin homologue that exerts pleiotropic effects, including prevention of viral replication and viral degradation of IRF3, when conjugated to viral/host proteins (Minakawa, Sone et al. 2008) and double stranded RNA-activated protein kinase R (PKR), which induces host translational shutdown and thereby, cessation of viral protein synthesis (Garcia, Gil et al. 2006, Sadler and Williams 2008). PKR is also involved in stabilising type I IFN mRNA (Barry, Breakwell et al. 2009, Schulz, Pichlmair et al. 2010).

Many IRGs work cooperatively to target specific stages of viral replication and specific viruses/viral families (Zhang, Burke et al. 2007, Schoggins and Rice 2011, Karki, Li et al. 2012). The interferon-induced transmembrane proteins 1/2/3 (IFITM1/2/3) restrict viral entry by sequestering 5’-triphosphate RNA produced by enveloped ssRNA viruses, (Liu, Sanchez et al. 2011, Pichlmair, Lassnig et al. 2011). Zinc antiviral protein (ZAP), which disrupts viral RNA translation, has been shown to work synergistically with ISG20, an ssRNA-specific RNAse that disrupts RNA synthesis, to inhibit yellow fever virus more

26 effectively than if expressed alone (Zhang, Burke et al. 2007, Karki, Li et al. 2012). Schoggins and Rice (2011) suggest that expression of multiple, functionally redundant, moderately effective IRGs may serve to prevent host cytotoxicity.

1.5 Alphaviruses and the type I IFN response

1.5.1 Sensitivity to the antiviral effects of IFNα/β

Alphavirus infection efficiently induces type I IFN responses, with high concentrations detected in both animal models of disease (Gardner, Anraku et al. 2010, Rudd, Wilson et al. 2012) and CHIKV patients (Jaffar-Bandjee, Ramful et al. 2010, Schwartz and Albert 2010). In vitro, pre-treatment in an epithelial cell line with doses as low as 10 IU/ml of IFNα is known to effectively inhibit CHIKV infection (Sourisseau, Schilte et al. 2007). Pre-treatment of Swiss 3T3 fibroblasts with only 5 IU/ml murine IFNα reduced EEEV-induced cyptopathic effect by 50% with higher doses (20 - 50 IU/ ml) required in SINV and VEEV infection (Ryman and Klimstra 2008). In vivo, alphavirus infection is known to be controlled by type I IFNs. Mice lacking IFNAR1 rapidly succumb to infection with these alphaviruses, indicating that type I IFNs are a predominant contributor to protection against alphavirus infection. Similarly, IFNAR1-/- mice infected with VEEV die within 2-3 days post-infection, compared to 7-9 days in WT mice (Schoneboom, Lee et al. 2000). A live-attenuated CHIKV-strain that had displayed promising results in vaccine trials and was avirulent in adult, WT mice was found to be pathogenic in type I IFN deficient mice, again demonstrating the importance of type I IFN responses in alphavirus pathogenesis (Partidos, Weger et al. 2011).

1.5.2 IRGs in alphavirus infection.

IRGs known to be involved in alphaviral infection include PKR, which effectively limits viral replication and enhances type I IFN responses to SFV and SINV infection (Ryman, Meier et al. 2005, Ventoso, Sanz et al. 2006, Barry, Breakwell et al. 2009), the 2’5’OAS isoform, OAS3, found to exert antiviral effects against CHIKV in an RNAseL- independent manner (Bréhin, Casadémont et al. 2009), ISG15, found to be critical for survival against CHIKV in neonatal mice (Werneke, Schilte et al. 2011) …and viperin, shown to inhibit SINV replication (Wang, Hinson et al. 2007, Zhang, Burke et al. 2007, Chan, Chang et al. 2008) and limit CHIKV replication and pathology (Teng, Foo et al. 2012). ZAP has been also found to inhibit replication of several alphaviruses including 27 RRV, SFV, SINV and VEEV (Bick, Carroll et al. 2003, Zhang, Burke et al. 2007, Karki, Li et al. 2012) and over expression studies have identified C6orf150, heparanase, SLC15A3 and SLC25A28 that exert anti-alphaviral effects. However the mechanism of action of these, and many other IRGs is not well understood.

1.5.3 Evasion of the type I IFN response

Type I IFN induction by different alphaviruses in vivo has been shown to correlate with the level of viral replication, in vitro (Ryman and Klimstra 2008). While alphaviruses are shown to be highly sensitive to the antiviral effects of type I IFNs, they have also evolved to evade or impede the type I IFN response. CHIKV nsp2 has been shown to block type I IFN signalling by blocking STAT1 phosphorylation and translocation of STAT1/STAT2 to the nucleus (Fros, Liu et al. 2010). Alphavirus have also recently been found to contain secondary structures in the 5’UTR that allow alphaviruses to evade inhibition by the IFN-inducible proteins with tetratricopeptide (IFIT) family of IRGs (Hyde, Gardner et al. 2014). Alphaviruses also induce widespread host translational and transcriptional shut down in favour viral protein synthesis (Aguilar, Weaver et al. 2007, Frolov, Akhrymuk et al. 2012). The capsid protein of the encephalitic VEEV and EEEV has been found to form a complex that inhibits host gene transcription by blocking the trafficking of cellular proteins through the nuclear membrane (Garmashova, Atasheva et al. 2007, Atasheva, Fish et al. 2010). SFV and SINV have been found to inhibit host transcription via nsp2, which translocates to the nucleus and acts to degrade RNA polymerase II (Akhrymuk, Kulemzin et al. 2012). Production of dsRNA during alphavirus replication also activates PKR, and subsequently, eukaryotic initiation factor 2 α (eIF2α) leading to widespread host translational shutdown.

1.6 Project aims

The overall aim of this thesis was to examine the role of the type I IFN response to CHIKV infection in a mouse model of CHIKV infection and disease. Specifically, the thesis sought to understand the signalling pathways engaged in the induction of type I IFN responses, the primary sources of the type I IFNs during CHIKV infection, and the global gene expression profile induced following CHIKV infection, as well as gaining insights into the CHIKV genome within the tissues. The findings from these studies will further our understanding of the immunopathological inflammatory processes induced by CHIKV, which may further the development of therapeutic interventions.

28 2. The role of interferon regulatory factors 3 and 7 during CHIKV infection

2.1 Introduction CHIKV infection is primarily controlled by type I IFNs and antibodies (Gérardin, Barau et al. 2008, Suhrbier, Jaffar-Bandjee et al. 2012). The host sensor pathways implicated in detection of CHIKV include: (i) the cytoplasmic receptors, retinoic acid-inducible gene I (RIG-I) and melanoma differentiation-associated gene 5 (MDA5), which signal through the adaptor molecule interferon β promoter stimulator 1 (IPS-1), and (ii) the membrane bound toll like receptors (TLR); TLR3 via TIR domain-containing adaptor inducing interferon-β (TRIF) and TLR7 via myeloid differentiation primary response gene 88 (MyD88) (Schilte, Couderc et al. 2010, Schwartz and Albert 2010, White, Sali et al. 2011). Signalling through these pathways results in activation of IRF3 and IRF7, the key transcription factors involved in type I IFN induction (Blasius and Beutler 2010, Kawai and Akira 2011, Levy, Marié et al. 2011). The relative importance of these transcription factors in CHIKV pathogenesis and protection is unknown.

IRF3 is expressed constitutively, in an inactive state, in most cells, and its activation results in transcription of IFNβ and IFNα4 (Yoneyama and Fujita 2010, Vitour, Doceul et al. 2014). IRF7 is the primary transcription factor for the IFNαs (Honda, Yanai et al. 2005), and is only expressed constitutively in plasmacytoid dendritic cells (pDCs), also known as professional IFN producing cells (Ning, Pagano et al. 2011). Following the initial induction of IRF7 expression, subsequent rounds of pattern recognition receptor (PRRs) signalling further activate IRF3 and IRF7, resulting in transcription of the full spectrum of IFNαs and an effective type I IFN response (Yoneyama and Fujita 2010, Levy, Marié et al. 2011, Vitour, Doceul et al. 2014).

Studies of IRF3 and IRF7, including the use of IRF3 and/or IRF7 deficient mice, have revealed the importance of these transcription factors during viral infection. West Nile Virus NS1 protein, for example, interacts with IRF3 to prevent its translocation to the nucleus (Wilson, de Sessions et al. 2008). Severe fever with thrombocytopenia syndrome virus forms inclusion bodies that sequester IRF3 in the cytoplasm, again preventing IRF3

29 translocation and induction of IFNβ (Qu, Qi et al. 2012, Wu, Qi et al. 2013). IRF7 was initially discovered in studies of Epstein Barr virus infection, in which IRF7 bound a promoter region in the viral genome, thus facilitating viral latency (Ning, Pagano et al. 2011). The 3C viral protease of picornavirus enterovirus 71 effectively inhibits activity of both transcription factors, initially via cleavage of IRF7, rendering IRF7 functionally redundant. The cleaved N-terminus of IRF7 then binds IRF3, which inhibits nuclear translocation of IRF3 and subsequent induction of type I IFNs, and thereby facilitating viral replication (Lei, Xiao et al. 2013).

Given the roles of IRF3 and IRF7 in the generation of antiviral responses and the scope for their inactivation by viruses, these transcription factors were investigated in the context of CHIKV infection. Mice deficient in IRF3, IRF7 and both IRF3 and IRF7, were infected with CHIKV using the mouse model of CHIKV infection and arthritis (Gardner, Anraku et al. 2010). The relative importance of the adaptor molecules upstream of IRF3 and IRF7 was also investigated using this infection model, in MyD88-/-, TRIF-/- and IPS-1-/- mice.

30 2.2 Materials and Methods

2.2.1 Physical Containment Level 3 Procedures CHIKV is listed as a PC3 organism under Australian Standard 2243.3-2010. Chikungunya virus causes severe fever and arthritis, the National Institute for Allergy and Infectious Diseases lists CHIKV as a Category C priority pathogen due to the potential for “mass dissemination…high mortality, high morbidity and major health impact” and the virus is listed by the Australia Group (2014) as a biological agent for export control. Accordingly, all work with this virus was carried out in a PC3 facility, certified by the Office of the Gene Technology Regulator.

PC3 facilities are contained within a PC2 facility, ventilated independently, with an exhaust to the outside of the building via an absolute HEPA filter system and maintained at a negative pressure. Personal protective equipment was worn (double gloves, gown, face mask and safety glasses) before entering the facility via an antechamber. Access to PC3 facilities was restricted to registered users who had undertaken appropriate training.

2.2.2 Cells

All cell lines were maintained in a 5% CO2 humidified atmosphere. Vero cells (African green monkey kidney epithelial cells, ATCC CRL-1586) and L929 cells (murine connective tissue, ATCC CCL-1) were maintained at 37˚C. C6/36 cells (Aedes albopictus mosquito clone, ATCC CRL-1660) were maintained at 28˚C. Low endotoxin trypsin-EDTA (Gibco-Invitrogen, Carlsbad, USA) was used to passage cells at 1 in 10 dilution, twice a week, to maintain logarithmic growth. Cells were cultured in Roswell Park Memorial Institute 1640 (RPMI1640) media containing L-glutamine (Gibco-Invitrogen) and supplemented with 25 mM HEPES (Sigma-Aldrich,) Pen-Strep antibiotic (penicillin, 50U/ml; streptomycin sulphate 50 µg/ml) (Gibco-Invitrogen) and 5% (v/v) foetal calf serum (FCS) (Gibco-Invitrogen). L929 cells were also supplemented with 1% 2-mercaptoethanol (2ME) (Gibco-Invitrogen).

Primary murine embryonic fibroblasts (MEFs) derived from IRF3-/-, IRF7-/- and IRF3/7-/- transgenic C57BL/6J mice (section 2.2.3), were provided by Prof A Khryomkh (University of Queensland) and cultured in Dubelco Minimum Essential Medium (DMEM)

31 containing L-glutamine (Gibco-Invitrogen) and supplemented with 25 mM HEPES (Sigma- Aldrich), 1% 2-ME (Gibco-Invitrogen), Pen-Strep antibiotic (Gibco-Invitrogen) and 10% (v/v) FCS (Gibco-Invitrogen). Low endotoxin trypsin-EDTA (Gibco-Invitrogen) was used to passage cells at 1 in 5 dilution, twice a week to maintain logarithmic growth. MEFs were maintained at 37˚C in a 5% CO2 humidified atmosphere.

2.2.3 Mice IRF3-/- and IRF7-/- transgenic mice were generated by Dr. T. Taniguchi of Osaka University (Honda, Yanai et al. 2005). IRF3-/-, IRF7-/- and IRF3/7-/- mice (Sato, Suemori et al. 2000, Daffis, Suthar et al. 2009) were provided by Dr MS Diamond (Washington University School of Medicine, St Louis). TRIF-/- and MyD88-/- transgenic mice were generated by Dr. M. Yamamoto of Osaka University (Yamamoto, Sato et al. 2003). IPS1-/- mice (Kumar, Kawai et al. 2006) were generated by Dr. YM Loo and provided by Dr M Gale (Washington University School of Medicine, St Loius). IFNAR-/- mice (Swann, Hayakawa et al. 2007) were supplied by Dr P Hertzog (Monash University, Victoria, Australia). C57BL/6J mice (Animal Resource Centre, Canning Vale, Western Australia) were used as aged matched controls. Mice were housed in individual ventilated cages and allowed to feed ad libitum. Experiments were conducted using female, mice aged 10- 18 weeks. All animal experiments were approved by the QIMR Animal Ethics Committee and were conducted in accordance with the Australian code for the care and use of animals for scientific purposes.

2.2.4 Blood and tissue preparation Whole blood was collected by tail bleed or heart puncture, and serum was separated by centrifugation (7500 rpm, 2.5 min) using 0.8 ml MiniCollect serum tubes (Greiner Bio-One GmbH, Kremsmünster, Germany). Sera for viraemia measurement or neutralisation assays were added to the first row of a 96-well flat bottom plate in 10 µl replicates, and the plates stored at -20˚C. Sera for bioassay or cytokine quantitation were added to the first row of a 96-V-well plate and the plate stored at -80˚C.

Tissues excised for virus titration were weighed, placed in 1 ml tube containing RPMI1640 media supplemented with 10% FCS, snap frozen using dry ice and ethanol,

32 and stored at -80 ˚C. Excised tissues for PCR were placed in 1 ml of RNAlater (Invitrogen) overnight at 4˚C, then stored at -80˚C.

2.2.5 Virus and viral preparations The CHIKV isolate used (LR2006-OPY1) is a primary isolate from the 2005 Reunion Island outbreak (Parola et al., 2006). Viral stocks were prepared as per (Gardner, Anraku et al. 2010). Briefly, virus at multiplicity of infection (moi) 0.04-0.06 was passaged once in C6/36 cells. Supernatants were collected at 48 hr post infection, centrifuged (2000 rpm, 5 min) to remove cell debris then aliquoted and stored at -80˚C.

Cell culture infectious dose 50% (CCID50/ml) for each batch was determined as described for viraemia measurement, (section 2.2.8). Inactivated purified CHIKV antigen was a kind gift from Inverness Medical Innovations Australia Pty. Ltd., (Brisbane, Australia) and was used as an ELISA antigen and as a vaccine antigen. CHIKV was inactivated using binary ethylenimine and purified by polyethylene glycol precipitation as described in Gardner, Anraku et al. (2010).

2.2.6 Mouse vaccination, infection and monitoring Vaccinated mice received 10 µg inactivated CHIKV antigen, (see above) in 40 µl RPMI1640 media supplemented with 2% FCS via subcutaneous injection at the base of the tail, 6 weeks prior to CHIKV infection.

4 Mouse infection involved injection of 10 CCID50 CHIKV in 40 µl RPMI1640 supplemented with 2% FCS via subcutaneous injection into the dorso-lateral metatarsal region of each hind limb, injecting toward the ankle as described (Gardner, Anraku et al. 2010). Mock infected mice received 40 µl RPMI1640 supplemented with 2% FCS via the same route. Following infection, mice were monitored daily for signs of clinical disease.

Euthanasia was performed by CO2 asphyxiation in accordance with animal ethics guidelines.

2.2.7 Prophylactic IFN treatment and adoptive transfer IFNαA (Chemicon International, #GF023) and/or IFNγ (Chemicon International, #IF005) were administered via intravenous injection into the tail vein. CD45+/CD11b+ cells

33 for adoptive transfer were derived from the spleen of an uninfected C57BL/6J mouse by fluorescence activated cell sorting (FACS) as described in Gardener et al., (2010). Briefly, the excised spleen was passed through a fine metal mesh and resuspended in phenol red- free RMPI 1640 supplemented with 2% FCS (Gibco-Invitrogen). Red blood cells were lysed using ammonium choloride potassium (ACK) lysis buffer (Sigma) and remaining splenocytes were washed, then stained with anti-CD45-phycoerythrin (PE) (BioLegend, CA, USA) and anti-CD11b (BioLegend) labelled with fluorescein isothiocyanate (FITC). Cells were sorted using a FACSCalibur™ (Becton Dickinson, NSW, Australia) and counted. CD45+/CD11b+ cells (2.7 x 105 cells) in 100 µl RPMI supplemented with 2% FCS were adoptively transferred by intravenous injection into the tail vein 24 hours prior to CHIKV infection.

2.2.8 Viraemia measurement Serum samples were serially diluted, 10-fold, in duplicate, using RPMI1640 supplemented with 2% FCS in 96-well, flat-bottom plates. C6/36 cells (104 cells/well) in 100 µl RPMI1640 media, supplemented with 5% FCS, were then added to each well and the plates cultured for three days. A 25 µl sample of supernatant from each well was then transferred into a new parallel 96-well plate containing Vero cells (104 cells/well) in 100 µl RPMI1640 media supplemented with 5% FCS. The plates were incubated at 37˚C in a 5

% CO2 atmosphere for four days. On day four, the highest dilution at which cytopathic effect (CPE) occurred was recorded and used to calculate viral titres. CPE, observed by light microscopy (Olympus CK2, Japan) was characterised by widespread cell death, rather than plaque formation. Dead/dying cells remained adherent, but diminished in size and were clearly distinct from cell debris or detached viable cells. Viraemia was expressed as log10 CCID50/ml of serum using the Spearman-Karber method (Hamilton, Russo et al. 1977).

2.2.9 Viral titres in tissues Viral titres were determined by 5-fold serial dilutions of supernatants from disrupted tissue samples. Frozen tissue samples were thawed and soft tissues passed through a fine metal mesh using the plunger of a 5 ml syringe. Hard tissues were pulverised using individual mortar and pestles for each sample. Tissue homogenates were then centrifuged (16 000 rpm, 1 min) and 125 µl of the supernatant were diluted in triplicate, by 5-fold serial

34 dilution, in a 96-well flat bottom plate and assayed as for viraemia (section 2.2.8). Viral titres were expressed as log10 CCID50 per gram of tissue.

2.2.10 Foot measurements Area measurements of height multiplied by width of the metatarsal region were taken using digital Vernier callipers (Kincrome Pty. Ltd., Victoria, Australia). An initial, day 0 measurement was made prior to infection, followed by subsequent measurements at the indicated times post infection. Hind limb swelling was expressed as a mean percentage increase of the metatarsal area of each foot compared with pre-infection measurements.

2.2.11 Interferon bioassay Sera were thawed, placed on ice and UV inactivated (1.5 hours) using biological safety cabinet UV lamps, to inactivate live virus. Serum samples were then serially diluted (3-fold dilutions) in duplicate in RPMI supplemented with 2ME, Pen-Strep and HEPES buffer in 96-V-well plates. A 50 µl sample of diluted serum or IFNαA standard dilutions from each well were then transferred into the equivalent well of a 96-well flat bottom plate containing L929 cells (104 cells/well) in 50 µl RPMI1640 supplemented with 10% FCS.

L929 cells were cultured with the serum for 24 hours prior to the addition 400 CCID50/ml of Semliki Forest virus in 100 µl RPMI supplemented with 10% FCS). Cells and virus were cultured for 4 days, or until complete CPE was observed in the positive control wells (i.e. no interferon treatment, virus only). Supernatants were aspirated, and cells were stained with 0.01% crystal violet/10% formaldehyde. Cells were then washed twice with PBS (200 µl/well) before the addition of methanol (100 µl/well) to solubilise the crystal violet stain. The optical density of the solubilised stain was read at 595 nm on a Synergy 4 plate reader (BioTek, VT, USA). A standard curve was generated from serial dilutions of recombinant IFNαA (Chemicon International, #GF023) and used to calculate the serum IFNα concentration of each sample.

2.2.12 Cytokine concentration by cytometric bead array Serum concentrations of and chemokines was determined using a BD Biosciences Cytometric Bead Array (CBA) kit according to the manufacturer’s instructions (BD Biosciences, San Jose, USA). Briefly, 10 µl serum samples in duplicate, or cytokine standards were incubated with 1.5 µl of capture antibody bead reagent and 1.5 µl detector

35 antibody PE conjugate reagent for 2 hours in the dark, at room temperature. Samples were washed once with 100 µl wash buffer to remove unbound detector PE conjugate reagent (1000 rpm, 3 min), and a second time with formalin to inactivate CHIKV. Samples were then resuspended in 120 µl wash buffer per well. Sample data were acquired by flow cytometry using a BD FACS Array™ Bioanalyser (BD Biosciences) and analysed using FCAP Array™ v2 Software (BD Biosciences).

2.2.13 Anti-CHIKV antibody ELISA Each well of a 96-flat well microtitre ELISA plate (MaxiSorb, Nunc, Rochester, USA) was coated overnight at 4˚C with 2 µg/ml inactivated CHIKV antigen (Inverness Medical Innovations) in carbonate-bicarbonate buffer (pH 9.5). Plates were blocked with 5% skim milk powder in PBS for 1 hour at room temperature. The blocking buffer was removed and the antigen-coated wells were incubated with 3-fold serial dilutions of mouse sera for 90 min at room temperature. Plates were washed (3 x 2 min) with 0.01% Tween 20 in PBS (PBS/T) and biotin-labelled rat anti-mouse IgG1 (BD Biosciences, USA, #553441) diluted 1/3000 (~0.15 µg/ml) in PBST or rat anti-mouse IgG2c detection antibody (BD Biosciences, #553388) diluted 1/1500 (~0.30 µg/ml) in PBST, were added to each well and incubated for 90 min at room temperature. Plates were washed as above and incubated for 1 hour with streptavidin-horseradish peroxidase (HRP) (BD Biosciences, #554066) diluted 1/1000 in PBST. Plates were washed as above then developed with 100 µl 2,2’- azino-bis (3 ethylbenzthaizoline-6-sulphonic acid) (ABTS) (Sigma-Aldrich) in H2O2 for 30 min. When fully developed, the OD was read at 405 nm on a Synergy 4 plate reader (BioTek, VT, USA).

2.2.14 Neutralisation assay Serum samples were heat inactivated (60˚C, 30 min) to inactivate the virus and host complement proteins. Inactivated sera were diluted 1 in 10 in RPMI1640 before being serially diluted in RPMI1640, 3-fold down a 96-well flat-bottom plate. CHIKV (400

CCID50/ml) in 50 µl of RPMI1640 supplemented with 2% FCS was added to each well.

Sera and CHIKV were incubated at 37˚C in a 5% CO2 atmosphere for 1 hour before the addition of Vero cells (104 cells/well) in a 100 µl volume of RPMI1640 media supplemented with 5% FCS. Cells were cultured for three days and monitored for CPE. No CPE was considered indicative of complete antibody neutralisation of the virus. The lowest dilution providing complete neutralisation was used to calculate neutralising antibody titres. 36 2.2.15 Analysis of clinical parameters 2.2.15.1 Haematocrit determination. Tail vein blood was collected, in duplicate, using heparinised haematocrit tubes (Beckton Dickinson, North Ryde, Australia). Wax-sealed tubes were centrifuged (10, 000 rpm, 10 min) and the total blood volume and haematocrit measured using digital Vernier callipers (Kincrome Pty. Ltd., Victoria, Australia). Haematocrit was expressed as the percentage of red blood cells in the total blood volume (serum plus haematocrit). The average percentage increase in haematocrit of infected mice compared to mock-infected control mice was presented.

2.2.15.2 Platelet counts Platelets numbers per microliter of whole blood were determined in blood collected by tail bleed. Approximately 10 µl blood was collected in 1.5 ml tubes, and a 5 µl sample was immediately diluted 1:20 in phosphate buffered 1% ammonium oxalate solution to lyse red blood cells and prevent clotting. Lysed blood was loaded into a haemocytometer and platelets were left to settle in a humidified chamber for 20 mins before counting using a light microscope (Olympus CK2, Japan).

2.2.15.3 Urine output Urine excreted by each mouse within 1 minute of scruff restraint was collected and weighed. Sample collection was carried out at the same time each day to minimise the impact of natural, diurnal fluctuations in urine output.

2.2.15.4 Temperature measurement Body temperature was measured at least 1 hour after urine collection, and prior to blood collection. Temperatures were taken using a 2 mm thermocouple bead and a digital thermometer (Digitech, Electus Distributions Pty Ltd, Rhydalmere, Australia). The probe was positioned in the inguinal region of the mouse, immediately after restraint, for approximately 30 s, until the temperature reading stabilised.

2.2.16 RNA extraction RNA isolation was carried out using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. Briefly, excised feet stored at -80ºC in RNAlater were thawed and added to a 2 ml microcentrifuge tube containing 1.5ml TRIzol reagent and 2 x 5 mm stainless steel beads (Qiagen Pty. Ltd., Doncaster, Australia). The tubes were homogenised (3 x 2 min, 25 Htz) in a Qiagen TissueLyser II (Qiagen). Homogenates were

37 centrifuged (12 000 x g, 10 mins), 800 µl supernatant transferred to a fresh tube and 200 µl of chloroform (Sigma-Aldrich) added to each sample. The samples were shaken by hand, incubated for 2-3 mins, then centrifuged (12,000 x g, 12 min) at 4° C and 400-600 µl of the resulting RNA-containing aqueous phase transferred to a fresh tube. Samples were incubated with 500 µl isopropanol for 10 min at room temperature before centrifugation (12,000g x 10 min) at ˚C.4 The resulting precipitate was pelleted and the supernatant carefully aspirated and discarded. The RNA pellet was washed in 1 ml cold 75% ethanol, and resuspended in 20 μl RNAse/DNAase free water (Sigma). The RNA concentration of each sample was quantified using a Nanodrop ND 1000 (Nanodrop Technologies Inc.). RNA was treated for DNA contamination using RNAse-Free DNAse 1 (New England BioLabs, Ipswich, US). Briefly, 10 µg of resuspended RNA was incubated in a 100 µl final volume of 1x DNAse 1 reaction buffer and 2 units (1 µl) of DNAse 1 at ˚C37 for 10mins, before heat inactivation at 75˚C for 10 mins.

2.2.17 cDNA synthesis Synthesis of first strand total cDNA was performed in a 12 µl final volume reaction containing 0.5-2 µg RNA, 200 ng of random hexamer oligonucleotides (Sigma-Aldrich), 2.5 mM each dTTP, dCTP, dGTP, dATP (Promega) and DNAse/RNAse-free water (Sigma- Aldrich). The reaction mixture was incubated at 65˚C for 5 mins, then placed on ice and 4 µl of 5x Superscript 1st Strand Buffer (Invitrogen), 1 µl of 0.1 M dithiothreitol (DTT) (Invitrogen) and 1 µl DNAse/RNAse-free water (Sigma-Aldrich) added to each reaction. The reaction mixture was incubated at 25˚C for 1 min before the addition of 200 U of Superscript II (Invitrogen) and further incubation for 11 min at 25˚C, 50 min at 50˚C then 15 min at 70˚C.

38 2.2.18 Real time RT-PCR Quantitative real time PCR (qRT-PCR) reactions contained 1 µl of cDNA (section 0), 10 µl of Platinum SYBR Green Super mix-UDG (Invitrogen), 1x BSA, 1 µl of each 10 µM primer and DNAse/RNAse-free water (Sigma-Aldrich) to 20 µl final volume. Thermocycling and data acquisition was carried out using a Rotor-Gene Q (Qiagen). Cycling conditions were 50˚C, 2 min; 95˚C, 2 min; 35 cycles of 94˚C, 5 s; 60˚C, 10 s; 72˚C, 15 s. Expression values are calculated from a standard curve and normalised to RPL13a. Primers were (5’3’): RPL13A gag gtc ggg tgg aag tac ca (forward), tgc atc ttg gcc ttt tcc tt (reverse); IRF7 ccg aga act gga gga gtt tcg forward), gct cca gct tca cca gga tca gg (reverse); IRF3 gcc tca ctc cca gga aaa cct ac (forward), aac tcc cat tgt tcc tca gct agc (reverse); IFNβ1 ctc cag ctc caa gaa agg acg (forward), gca tct tct ccg tca tct cca (reverse); IFNα4 tca ttc tgc aat gac ctc ca (forward), tat gtc ctc aca gcc agc ag (reverse).

2.2.19 Histology Tissues were excised and placed in ≥10x volume of 10% neutral buffered formalin. Feet were fixed for 48-72 hours, and then decalcified using 15% EDTA in 0.1% phosphate buffer over ten days. Tissues were embedded in paraffin wax before being cut into ~6 µm sections and stained with haematoxylin and eosin. Stained slides were scanned using the Aperio Scan Scope XT digital slide scanner (Aperio, Vista, USA).

2.2.20 In situ hybridisation Excised tissues were fixed as for histology (section 2.2.19). In situ hybridisation using fixed tissues was carried out by C. Babarit under the supervision of T. Larcher (Ecole Nationale Vétérinaire, France). Briefly, paraffin embedded tissues were sectioned to 4 µm, deparaffinised in xylene, and rehydrated through descending graded alcohols to nuclease-free water. Tissues were permeabilised with protein kinase K (DAKO, Glostrup, Denmark) and postfixed in 4% paraformaldehyde. Sections were hybridised with a 450 bp digoxigenin-labelled CHIKV probe sequence (GenBank No.: DQ4435442) then incubated with alkaline phosphatase-conjugated anti-digoxigenin antibody (Roche). Slides were treated with nitro blue tetrazolium chloride/5-Bromo-4-chloro-3-indolyl phosphate alkaline phosphatase substrate and imaged by bright field microscopy (Nikon Eclipse 80i, Badhoevedo, Netherlands).

39 2.2.21 Statistical analysis Analyses were performed using IBM SPSS Statistics (version 19). The t test was used if the difference in the variances was <4, skewness was > -2, and kurtosis was < 2. Where the data was non-parametric and difference in variances was < 4, the Mann Whitney U test was used, if > 4 the Kolmogorov-Smirnov test was used.

40 2.3 Results 2.3.1 IRF3 and IRF7 protect against CHIKV-induced high viral burden and mortality.

IRF3 and IRF7 are key transcription factors involved in the production of type I IFNs, cytokines well known to be critical for control of CHIKV (Honda, Takaoka et al. 2006, Gérardin, Barau et al. 2008, Suhrbier, Jaffar-Bandjee et al. 2012). To assess the contribution of IRF3 and IRF7 during CHIKV infection, IRF3-/-, IRF7-/- and IRF3/7-/- mice were infected with CHIKV. Clinical signs of disease, viral burden in blood and tissues, and arthritic disease following CHIKV infection were investigated.

All WT, IRF3-/- and IRF7-/- mice survived CHIKV infection, whereas IRF3/7-/- mice became moribund, showing signs of severe clinical disease (ruffled fur, hunched posture, reduced activity), and required euthanasia, between day 4 and day 6 post infection (Figure 2.1A). Peripheral blood viraemia in all strains peaked at day 2 post infection, except in IRF3/7-/- mice, in which titres continued to increase until day 3 post infection. Viraemia in IRF3/7-/- mice was significantly higher from day 3 onwards, reaching as high as four logs greater than WT mice by day five post infection. Viraemia in IRF7-/- mice was also higher (~1 log) than WT mice from day 2 to day 4 post infection, however, this difference was only statistically significant at day 3 post infection. Viraemia in IRF3-/- mice did not differ significantly from WT mice (Figure 2.1B).

In the tissues, viral titres in muscle, spleen, liver and brain were again significantly higher in IRF3/7-/- mice compared to WT animals as early as day 2 post infection, and from day 3 onwards in the inguinal lymph nodes. As with serum viraemia, viral titres in the tissues of IRF3/7-/- mice peaked at day three post infection, 24 hours after titres peaked in WT mice. However, viral tires in the foot tissues of IRF3/7-/- mice peaked 24 hours earlier than WT mice, and remained elevated until time of death (up to day 6 post infection) (Figure 2.1C).

These results demonstrate that IRF3 and IRF7 are involved in protection against CHIKV high viral burden and absence of both transcription factors results in death.

41 A Survival B Viraemia

SEM 12

100 ± ) 10 80 8 WT 60 WT IRF3 IRF3 6 40 IRF 7

IRF 7 Log10 (CCID 50/ml 4 IRF3/7 Percent survival Percent 20 IRF3/7 titre 2

0 0 0 5 10 15 20 Serum 0 1 2 3 4 5 6 7 Days post infection Days post infection

C Tissue Titres

Feet Liver Brain 10 10 10 SEM

± 8 8 ) 8 /g WT 50 6 6 6 IRF3 (CCID 4 4 4

titre IRF 7

2 IRF3/7 2 2 Tissue

0 0 0 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

Lymph nodes Muscle 12 Spleen 12 12

10 10 10 SEM ± ) 8 8 8

6 6 6

(CCID50/g 4 4 4 titre 2 2 2

0 0

Tissue 0 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 Day post infection Day post infection Day post infection

Figure 2.1 CHIKV infection in WT, IRF3-/-, IRF7-/-, and IRF3/7-/- mice. Mice were 4 infected with 10 CCID50 via subcutaneous injection of the dorsolateral metatarsal region of the hind limb. A) Kaplan-Meier survival plot. Data were recorded when an animal was euthanased in accordance with animal ethics guidelines (n=12-20 per group). B) Virus titres in serum (n=7-12 per group). Differences in viral titres between WT and IRF3/7-/- mice reached significance at days 3, 4 and 5 (p<0.05; Mann-Whitney U). C) Virus titres in tissues (n=3-4 per group). Titres in all tissues at day 3 and 4 were significantly higher in IRF3/7-/- mice compared to WT controls (p<0.05 Mann-Whitney U).

42 2.3.2 IRF3 and IRF7 protect against severe foot swelling post CHIKV infection.

Foot swelling is an indicator of arthritic disease in CHIKV-infected WT mice, with a peak increase in foot swelling of ~60%, at day 6-7 post infection and resolution occurring by day 12-14 post infection (Gardner, Anraku et al. 2010). Infection of IRF3-/-, IRF7-/- and IRF3/7-/- mice also resulted in foot swelling. However, foot swelling in the knockout mice was earlier (day 3-5) and more pronounced than that seen in WT mice. Peak foot swelling (day 6) was slightly increased in IRF3-/- mice (~70% increase from pre-infection) more severe in IRF7-/- mice (~100% increase) and most severe in IRF3/7-/- mice (~140% increase) (Figure 2.2).

To ascertain whether foot swelling was an artefact of the virus delivery method (i.e. subcutaneous injection of the foot), WT and IRF3/7-/- mice were inoculated with heat inactivated CHIKV, sterile RPMI1640 media only, or via a different route. Injection of either heat inactivated virus or media alone did not result in foot swelling in either strain of mice (data not shown). Administration of the standard dose of CHIKV via a different route (subcutaneous injection at the base of the tail) did not induce foot swelling in WT mice. However, in IRF3/7-/- mice, subcutaneous injection at the base of the tail caused mild foot swelling and still resulted in mortality (data not shown). Together, these results confirm that foot swelling in mice following CHIKV infection is the result of replicating virus, rather than mechanical injury or local reaction to the inoculum.

43

Figure 2.2 Foot swelling in WT, IRF3-/-, IRF7-/-, and IRF3/7-/- mice following CHIKV 4 infection. Mice were infected with 10 CCID50 via subcutaneous injection of the dorsolateral metatarsal region of the hind limb. Data presented is the percentage increase in the metatarsal region from day 0 post infection (n=30 per group except IRF3/7-/- day 6 post infection, where n=6). (ns not significant, * p<0.05, ** p<0.01; Mann-Whitney U, day 4 and day 5: Kolmogov-Smirnov test).

44 2.3.3 Prophylactic treatment with recombinant IFNα, IFNγ or adoptive transfer of CD45+/CD11b+ did not prevent mortality in IRF3/7-/- mice.

IFNα is used in the treatment of a number of viral infections, namely hepatitis C (López, Molina et al. 2011, Ning, Pagano et al. 2011), and has also been shown to reduce viraemia and disease in WT mice if given prior to CHIKV infection (Gardner, Anraku et al. 2010). To asses whether prophylactic treatments could prevent the viraemia and disease observed in IRF3/7-/- mice, administration of IFNα and/or IFNγ, as well as adoptive transfer of WT cells was undertaken.

IFNα (104 IU/ml) administered 6 hours prior to CHIKV infection was only able to delay death by 24 hours in IRF3/7-/- mice. IFNγ is induced independently of IRF3 and IRF7 (Platanias 2005, Levy, Marié et al. 2011), and was shown to induce an antiviral state faster than type I IFN treatment in response to RNA viruses (Gomi, Morimoto et al. 1985, Saito 1989). Treatment of IRF3/7-/- mice with IFNγ (104 IU/ml) 24 hours post CHIKV infection was also able to delay death by 24 hours. A combined treatment of 5x103 IU/ml IFNα at 6 hours prior, and 5x103 IU/ml IFNγ 24 hours post CHIKV infection delayed death by 36 hours (Figure 2.3). Monocyte/macrophages play an important role in the pathobiology of alphaviral arthritis and can produce type I IFNs (Mateo, La Linn et al. 1999, Kumagai, Takeuchi et al. 2007, Her, Malleret et al. 2010). FACS sorted CD45+/CD11b+ cells (2.7x105) (primarily comprised of monocytes and macrophages).derived from spleens of naive WT mice were adoptively transferred to IRF3/7-/- mice, 24 hours prior to infection. Again, mortality was only delayed by 24 hours, but not prevented (Figure 2.3). This is consistent with findings by Schilte, Couderc et al. (2010) that haematopoietic cells do not produce protective levels of type IFNs during CHIKV infection.

These results demonstrate that delivery of IFNs alone is not sufficient to induce an effective antiviral response or protect IRF3/7-/- mice against CHIKV mortality.

45

Figure 2.3 Prophylactic treatment of IRF3/7-/- mice and IRF7 induction. IRF3/7-/- mice aged 8-12 weeks received either 104 IU/ml IFNα or IFNγ, 5x103 IU/ml each IFNα and IFNγ combined or 2.7x105 FACS sorted CD45+/CD11b+ splenocytes in 100 µl RPMI1640 media by intravenous injection at the base of the tail. Mice were challenged with CHIKV 6 hours post treatment. Mice treated with IFNγ were challenged 24 hours post treatment. A death event was recorded when mice were euthanased in accordance with animal ethics guidelines (n=3-4 per group).

46 2.3.4 IRF7 is induced > 150-fold in response to CHIKV infection in WT mice.

In the previous section, contrary to results seen in WT mice (Gardner, Anraku et al. 2010), administration of type I IFNs was not an effective prophylactic treatment in IRF3/7-/- mice. IRF7 is the primary transcription factor for induction of IFNαs (Honda, Yanai et al. 2005). IRF7 induction, but not IRF3 induction is known to be essential to the positive feedback loop that results in effective type I IFN responses (Marie, Durbin et al. 1998, Sato, Suemori et al. 2000).

IRF3 and IRF7 mRNA expression was analysed in the feet of WT mice infected with CHIKV. IRF3 was only up-regulated on day 1 post infection and only by ~2.5 fold. IRF7 was induced to a much greater extent, with up-regulation >100 fold by day 1 post infection, ≈150 fold increase at day 3 post infection, and expression values at day 7 still ≈50 fold higher than mock infected samples (Figure 2.4). This is consistent with the positive feedback model of type I IFN induction, in which activation of constitutively expressed IRF3 precedes the induction of IRF7 and antiviral effectors (Levy, Marié et al. 2011). IRF7 itself is induced by type I IFNs; the high level of IRF7 induction seen here, indicates that the dose of IFNα used in section 2.3.3 was not sufficient to compensate for the level of IFNαs that would ordinarily be produced in the presence of IRF7 induction. Furthermore, these results suggest that the protective role of prophylactic IFN treatment is mediated by up-regulation of IRF7, rather than induction of antiviral effectors.

47

Figure 2.4 IRF3 and IRF7 mRNA induction in WT mice. IRF3 and IRF7 mRNA expression by quantitative real-time PCR in WT mouse feet at days 0, 1, 3 and 7 post CHIKV infection. Normalised to RPL13a mRNA levels (n=6; all time points significantly greater than day 0, except IRF3 day 7, p ≤ 0.05; t-test).

48 2.3.5 An absence of IRF3 and IRF7 does not affect antibody responses to CHIKV infection

Alphaviruses, including CHIKV, are known to be controlled by host antibody responses (Gardner, Anraku et al. 2010, Holzer, Coulibaly et al. 2011, Suhrbier, Jaffar- Bandjee et al. 2012). To determine whether protective antibody responses are compromised in the absence of IRF3 and/or IRF7, IRF3-/-, IRF7-/-, IRF3/7-/- and WT mice were vaccinated with 10 µg inactivated CHIKV antigen. After four weeks, vaccinated mice were challenged with CHIKV.

At 3 weeks post vaccination, serum was collected and assayed for anti-CHIKV antibody titres by ELISA. All mouse strains were capable of mounting a robust antibody response to vaccination (Figure 2.5A). Slightly more IgG1 and less IgG2c CHIKV-specific antibody was induced in IRF3/7-/- mice compared to WT mice, indicative of a slight shift toward Th2 responses, and consistent with the role of type I IFNs in promoting Th1 responses (Bracci, La Sorsa et al. 2008, Farkas and Kemény 2011). Neutralisation assay on serum showed that all mice had produced, comparable levels of neutralising antibody titres (Figure 2.5B).

At 4 weeks post vaccination, mice were challenged with CHIKV. All strains of mice were protected against CHIKV viraemia and disease (Figure 2.5 C). Without vaccination, both IRF3/7-/- and WT mice were still able to mount neutralising antibody titres by day 4 post infection (data not shown). Nevertheless, antibody production was unable to provide protection for IRF3/7-/- mice against CHIKV infection, with mice requiring euthanasia between days 4-6.

These results indicate that antibody responses are not hindered by the absence of IRF3 and/or IRF7. However, antibody responses were either too late, or insufficient to protect IRF3/7-/- mice against cytokine dysregulation and subsequent haemorrhagic shock (see sections 2.3.6 - 2.3.11).

49

Figure 2.5 Antibody production and protection in WT, IRF3-/-, IRF7-/-, and IRF3/7-/- mice following vaccination. Mice were vaccinated with inactivated CHIKV antigen. Naïve mice received RPMI1640 instead of CHIKV antigen. At 3 weeks post vaccination, blood was collected and assayed for anti-CHIKV antibody titres and neutralisation assay. Four weeks after vaccination, mice were infected with CHIKV A) ELISA of anti-CHIKV IgG1 and IgG2c antibody titres (n=3-4 per group, 3 weeks post vaccination). B) Neutralising anti-CHIKV antibody titres (n=3-4 per group, 3 weeks post vaccination). Titre represents the serum dilution giving 100% neutralisation in CHIKV-infected Vero cell culture (n=3-4 per group, 3 weeks post infection). C) Kaplan-Meier plot of survival in vaccinated and naive IRF3/7-/- mice (n=4-6 per group).

50 2.3.6 Absence of IRF3 and IRF7 alters type I IFN and pro-inflammatory cytokine responses to CHIKV infection.

IRF3 and IRF7 are the key transcription factors for type I IFN production. To examine the extent to which IRF3 and IRF7 affect the expression of type I IFN and pro- inflammatory mediators, IRF3-/-, IRF7-/- and IRF3/7-/- mice were infected with CHIKV and their serum cytokine levels assayed.

Serum bioassay was used to determine the type I IFN levels in mice following CHIKV infection. Consistent with previous reports (Gardner, Anraku et al. 2010), high concentrations of type I IFN (peak of ~1000 IU/ml) were detected in WT mice, with levels peaking at day 2 post infection. Type I IFN levels in IRF3-/- mice were similar to WT, peaking at ~800 IU/ml on day 2 post infection. Type I IFN levels in IRF7-/- mice were markedly lower than WT mice (~2 IU/ml) and were only detectable on day 2 post infection. In IRF3/7-/- mice, type I IFN levels were below the level detectable by serum bioassay (~1 IU/ml) (Figure 2.6A). Consistent with protein expression, type I IFN mRNA expression in feet at day 2 post infection was highest in WT mice and minimal in IRF3/7-/- mice. Type I IFN mRNA was up-regulated to similar levels in IRF3-/- and IRF7-/- mice (Figure 2.6B) Circulating levels of pro-inflammatory cytokines were also measured in mice following CHIKV infection (Figure 2.6C). In IRF3/7-/- mice, peak serum levels of monocyte chemoattractant protein-1 (MCP-1), interleukin-6 (IL-6), tumour necrosis factor (TNF) and IFNγ were significantly elevated compared to WT mice; MCP-1, IL-6 and TNF were ~10 times higher, and IFNγ was almost 50 times higher in IRF3/7-/- mice compared to WT controls in response to CHIKV infection. MCP-1, IL-6 and IFNγ peaked at day 2 post infection, while TNF remained significantly elevated from day 2 until day 6 (time of death). Serum levels of these cytokines in IRF3-/- and IRF7-/- mice were not significantly different to WT mice.

These results illustrated that mortality in IRF3/7-/- mice, following CHIKV infection, is associated with very high levels of IFNγ.

51

Figure 2.6 Cytokine expression during CHIKV infection in WT, IRF3-/- IRF7-/- and IRF3/7-/- mice. Mice were infected with CHIKV. A) Serum bioassay of bioactive peripheral blood IFN levels at days 1, 2, and 3 post infection (n=3-6 per group). B) Quantitative real-time PCR of IFNα4 and IFNβ1 expression in mouse feet at day 2 post infection. Normalised to RPL13a mRNA (n=4-6 per group). C) Peripheral blood cytokine levels. (n=3-6 per group; n.s. not significant, * p<0.05, ** p<0.01; Mann-Whitney U)

52 2.3.7 CHIKV tissue tropism is altered in IRF3/7-/- mice.

The cellular tropism of CHIKV in vivo is not well characterised. Given the differences in replication kinetics, disease and immune responses seen in CHIKV infected IRF3/7-/- mice, CHIKV tropism in these mice was examined. In situ hybridisation was used to detect CHIKV RNA in the cytoplasm of cells in foot tissues of WT and IRF3/7-/- mice at day 3 and day 5 post infection. Uninfected mouse tissue, RRV anti-sense probes and CHIKV sense probes were used as negative controls.

In WT mice at day 3 post infection, staining was observed in endothelial cells (Figure 2.7A), skeletal myofibres (B), cells in the periosteum (E) and synovial membranes (D). Sporadic staining of fibroblasts and macrophages was observed within the dermis(C). In IRF3/7-/- mice at day 3 post infection, staining was observed in endothelium (F), skeletal myofibres (G) and the dermis (H). However, staining of fibroblasts and macrophage/monocytes in the dermis was considerably more apparent in IRF3/7-/- than WT mice. In contrast to WT mice, no staining was seen in the periosteum and synovial membranes of IRF3/7-/- mice, but staining was present in chondrocytes (I) of the articular cartilage, which was not seen in WT animals.

At day 5.5 post infection, staining was again observed in the endothelium (Figure 2.7J), fibroblasts and macrophages in the dermis and underlying connective tissues in WT mice (L). Positively stained macrophages could also be seen within the blood vessels (M), but staining was no longer present in the muscle (K). In IRF 3/7-/- mice at day 5.5 post infection, CHIKV RNA equivalent to WT mice was seen in the endothelium (O) and fibroblasts and macrophages of the dermis and connective tissues (Q,R,S), and also remained present in the muscle at day 5.5 post infection (P). Positive staining was also observed in the epidermis, suggesting infection of keratinocytes (S), which had not been seen at either time point in WT mice (C, N). No signal was detected in negative controls at any time point (data not shown).

The results here show that while similar to WT, the distribution and tissues targeted by CHIKV are distinct in IRF3/7-/- mice, indicating that the cellular tropism of CHIKV is altered in the absence of IRF3 and IRF7.

53

Figure 2.7 Virus location by in situ hybridisation in CHIKV-infected IRF3/7-/- and WT mice. CHIKV RNA detection by in situ hybridisation in IRF3/7-/- and WT mice at day 3 and day 5.5 post infection. Staining was observed in both strains at both time points in blood vessel walls (A, F, J, M, O), skeletal muscle (B, G, K, P) and skin (C, H, L, S). Staining was also present at day 3 in WT synovial tissue (D) and periosteum (E), IRF3/7-/- articular cartilage (I) and at day 5.5 in the dermis (L, Q). RBC: Red blood cells. Images are representative of that observed in 3 mice per group

54 2.3.8 Tissue pathology and inflammatory infiltrates were altered in IRF3/7-/- mice

To determine whether the altered tissue tropism observed in IRF3/7-/- mice might impact upon tissue pathology, histological analysis was carried out in several tissues, excised from IRF3-/-, IRF7-/-, IRF3/7-/- and WT mice at day 5 post infection.

Necrotic lesions, involving keratinocyte karyorrhexis (nuclear blebbing during cell death) were observed within the epithelial layer of skin in IRF3/7-/- mice, and were associated with bullae (small, fluid filled blisters) formation (Figure 2.8A vs B). These lesions were also present in the skin of IRF7-/- mice, but to a much lesser degree, and were not observed in IRF3-/- mice or WT mice (data not shown). Within the joints of IRF3/7-/- mice, evidence of severe exudative arthritis, characterised by the presence of fibrin and inflammatory infiltrates, could be observed (Figure 2.8C vs D). Evidence of exudative arthritis could not be seen in IRF3-/-, IRF7-/- or WT mice at day 5 post infection (data not shown); however, a similar exudative arthritis has been described previously at day 7 post infection in WT mice (Gardner, Anraku et al. 2010). In the skeletal muscle at day 5.5 post infection, a large inflammatory, predominantly macrophage infiltrate was present in IRF3-/-, IRF7-/- (Figure 2.7G & H) and WT mice (data not shown). Myofibre necrosis, likely the result of macrophage activity, was also present in these mice. In contrast, the skeletal muscle of IRF3/7-/- mice at day 5.5 post infection was characterised by minimal cellular infiltrate and endomysial oedema (Figure 2.8E vs F). The liver, spleen, kidney, gut, lung, lymph nodes and brain of IRF3/7-/- and WT at day 5.5 post infection were also examined; with the exception of lung and brain (discussed section 2.3.9), no remarkable lesions were observed in these tissues.

These differences in pathology within skin, joint and muscle tissues between IRF3/7-/- mice and WT mice thus appear to coincide with the differences with the altered tissue tropism (Figure 2.7), and indicate that CHIKV tissue pathology is altered the absence of IRF3 and IRF7.

55

Figure 2.8 Histology in CHIKV-infected IRF3/7-/- mouse skin, joint and muscle tissues. A) Skin necrosis (n) in an IRF3/7-/- mouse at day 5 post CHIKV infection. Degenerating leukocytes with pale cytoplasm and karyorrhtic nuclei (arrows) are present. B) Skin of uninfected IRF3/7-/- mice. C) Exudative arthritis in joint tissues of IRF3/7-/- mice at day 5 post infection with inflammatory infiltrate and fibrin (arrows) present in the synovial lining. Synovial space (SS) and articular cartilage (AC) are indicated. D) Articular cartilage from joint tissues of an uninfected IRF3/7-/- mouse. E) Skeletal muscle from an uninfected IRF3/7-/- mouse. F) Skeletal muscle from an IRF3/7 mouse at day 5 post infection. G) Skeletal muscle from an IRF7-/- mouse at day 5 post infection. H) Skeletal muscle from an IRF3-/- mouse at day 5 post infection. Images are representative of 3 mice per group.

56 2.3.9 CHIKV infection in IRF3/7-/- mice induces haemorrhage and oedema

In addition to changes in tissue pathology outlined in section 2.3.8, histological analysis revealed evidence of changes to the vasculature and fluid retention in IRF3/7-/- mice at day 5 post CHIKV infection.

Severe generalised oedema and multifocal haemorrhage was present within the dermis and subcutaneous tissues of feet in IRF3/7-/- mice (Figure 2.9A vs D). The vasculature of IRF3/7-/- mice was seen to contain lesions, which were not present in the vasculature of WT mice. Lesions were characterised by fibrinoid necrosis (accumulation of proteinaceous debris in necrotic tissue, resembling fibrin), involving neutrophil karyorrhexis indicative of intramural leukocytoclasis (degenerate leukocytes within vascular walls) (Figure 2.9B&C). Perivascular fibrin exudation and extravasated erythrocytes surrounding lesions were also observed (Figure 2.9C). Pulmonary consolidation, (fluid in the lungs), as well as perivascular cuffs (leukocytes surrounding the vasculature), were present in the lung tissue of IRF3/7-/- mice, and in the brain, mild pericapillary oedema was also observed, which was not present in WT mice (data not shown). Oedema and some haemorrhage was also observed in IRF7-/- mice, to a much lesser degree, and was not observed in IRF3-/- or WT mice (data not shown).

Thus, mortality in IRF3/7-/- mice following CHIKV infection was associated with severe, generalised oedema and multifocal haemorrhage.

57

Figure 2.9 Histological analysis of haemorrhage and oedema in CHIKV-infected IRF3/7-/- mice. Standard H&E staining of IRF 3/7-/- mouse tissues at day 5 post infection with uninfected IRF3/7-/- tissues or IRF3-/- and IRF7-/- mice at day 5 post infection for comparison. A) Subcutaneous foot tissue in IRF 3/7-/- mice day 5 post infection with severe, generalised oedema (asterisks) and multi-focal haemorrhage. B) Enlarged lower portion of panel A, showing oedema and haemorrhage. C) Vasculitis indicated by fibrinoid necrosis (n) and intramural degenerating leukocytes (arrows). D) Subcutaneous foot tissue in an uninfected IRF3/7-/- mouse.

58 2.3.10 CHIKV induces clinical symptoms of hypovolemic shock in IRF 3/7 -/- mice

Severe haemorrhage and/or oedema are involved in the development of hypovolemic shock, as seen during dengue haemorrhagic fever (DHS) or dengue shock syndrome (DSS) (Pang, Cardosa et al. 2007, Huy, Van Giang et al. 2013). Hypovolemic shock is a pathological drop in blood pressure due to blood and fluid loss, resulting in a failure to adequately perfuse tissues, and is often an irreversible and fatal condition. To determine whether the oedema and haemorrhage observed in IRF3/7-/- mice were causing hypovolemic shock, and ultimately death in these mice, several clinical indicators of hypovolemic shock were investigated.

Thrombocytopenia due to haemorrhage is a criterion used in the diagnosis of DSS/DHS (Ranjit, Kissoon et al. 2005, W.H.O 2009). In humans, a diagnosis of thrombocytopenia is given when circulating platelet levels drop by≥70 %. In WT mice, platelet levels at day 2 were reduced by ~20%, but returned to pre-infection levels by day 5 post infection. In IRF3/7-/- mice, platelet levels had dropped to ~55% by day 2, and at day 5, were less than 70% of pre-infection levels (Figure 2.10A).

Hypothermia, due to reduced blood circulation, is also common in hypovolemic shock, with a sudden change from fever to hypothermia a clinical indicator of the onset of DHS/DSS (W.H.O 2009). At day 2 post infection, IRF3/7-/- mice developed a fever, before a dramatic drop in body temperature at day 3 and day 4 post infection was observed. In WT mice, no significant fever was detected. There was a very mild drop in body temperature at day 3 post infection. However, body temperature stabilised to pre-infection temperatures from day 4 onwards (Figure 2.10B).

Another measure of hypovolemic shock used in DHS/DSS diagnosis is an increase in blood haematocrit levels of >20%, indicative of hemoconcentration following plasma leakage into the tissues (W.H.O 2009, Huy, Van Giang et al. 2013). At day 5 post infection, haematocrit measurements in IRF3/7-/- mice showed a mean increase of 15.8% (range 7- 24%) from pre-infection measurements, whereas haematocrit levels in infected WT mice at day 5 did not differ from pre-infection readings (Figure 2.10D).

59 Pathologically low blood pressure during hypovolemic shock can also manifest as oliguria, or reduced urine output due to poor kidney perfusion (Ranjit, Kissoon et al. 2005, W.H.O 2009). Urine output in WT and IRF3/7-/- mice was not significantly different from day 0 to day 3 post infection. However, urine output on days 4 and 5 was significantly lower in IRF3/7-/- mice compared with WT mice, with no urine output seen on day 5 post infection in IRF3/7-/- mice (Figure 2.10C).

The severe haemorrhage and generalised oedema observed in histological analysis of IRF3/7-/- mice, in conjunction with the clinical parameters observed just prior to death, provide compelling evidence that mortality in IFF3/7-/- mice following CHIKV infection was due to hypovolemic shock.

60

Figure 2.10 Clinical symptoms of hypovolemic shock in IRF 3/7-/- mice following CHIKV infection. A) Platelet levels were significantly reduced compared to WT controls on days 3, 4 and 5 post infection (p= 0.001, p<0.001 and p, 0.001 respectively; n=8-10 per group). B) IRF3/7-/- mice temperatures rose significantly from day 0 to day 2 post infection (p=0.004) and were higher than WT mice at the same time point (p=0.001). IRF3/7-/- mice temperatures were significantly lower than WT mice on day 4 and day 5 post infection (p<0.001; n=4-16 per group). C) Urine output from IRF3/7-/- mice was significantly lower than WT mice at days 4 and 5 post infection (p<0.001; n= 8-10 per group). D) Percentage increase of haematocrit levels (from uninfected control levels) in IRF3/7-/- versus WT mice at day five post infection (p<0.001; n=10-11 per group). Data pooled from four individual experiments, statistics obtained by Mann-Whitney U test).

61 2.3.11 TRIF, MyD88 and IPS-1 are involved but not critical in viraemia control, foot swelling or type I IFN production following CHIKV infection

TRIF, MyD88 and IPS-1 are used by host PRR to relay signals to IRF3 and IRF7 (Takeuchi and Akira 2010, Jensen and Thomsen 2012). The endosomal TLRs, with the exception of TLR3, transduce their signals through MyD88 (Kawai and Akira 2010); TLR3 signals exclusively via TRIF (Takeuchi and Akira 2010). The RLRs, RIG-1 and MDA5, are located in the cytoplasm, and signal through IPS-1 (Jensen and Thomsen 2012). Haematopoietic cells engage both TLRs and RLRs, while type I IFN production in fibroblasts is largely via IPS-1 (Kawai and Akira 2011). To assess the extent to which each adaptor molecules is used, and gain insight into the PRR signalling pathways used in CHIKV infection, MyD88-/-, TRIF-/-, and IPS1-/- mice were infected and monitored for viral replication, foot swelling and serum type I IFN.

CHIKV infection in MyD88-/-, TRIF-/-, and IPS1-/- mice was not fatal, but caused increased foot swelling and viraemia and lower serum type I IFNs compared to WT mice. Viraemia was highest in TRIF-/- mice, with titres ~ 2.5 and ~4 log higher than WT at day 3 (peak) and day 4, respectively. Peak viraemia in MyD88-/- and IPS-1-/- mice was earlier (day 2) and was not significantly different to WT. At day 4, viraemia in all strains was higher than WT mice. Viraemia in IPS-1-/- mice was slow to clear and remained ~2 log higher than WT on day 5 post infection (Figure 2.11A). Consistent with the prolonged viral clearance, foot swelling in IPS-1-/- mice peaked higher and later than all other strains. Swelling was also higher than WT mice in TRIF-/- mice (days 2-6 and 8) and in IPS-1 mice (days 2-8) following infection. Foot swelling in WT and MyD88-/- mice did not differ significantly (Figure 2.11B). Serum type I IFN was significantly lower in MyD88-/-, TRIF-/- and IPS-1-/- mice compared to WT mice at days 1 and 2 post infection; however in MyD88- /- mice, type I IFN was ≈100 fold higher than TRIF-/- and IPS-1-/- at day 1 post infection. By day 2 post infection, serum levels in both TRIF-/- and MyD88-/- mice were similar. Type I IFN production in IPS1-/- mice was minimal at all time points measured (Figure 2.11C).

These results demonstrate that TRIF, MyD88 and IPS-1 are involved, but not critical for protection against CHIKV infection. Consistent with results in sections 2.3.1 and 2.3.2, type I IFN production was associated with effective viraemia control and the degree of foot swelling, with IPS-1 playing the most significant role in type I IFN signalling in response to CHIKV infection. 62

Figure 2.11 CHIKV infection in WT, TRIF-/-, MyD88-/- and IPS1-/- mice. A) Serum virus titres in MyD88-/- (n=8), TRIF-/- mice (n=11) and IPS1-/- mice (n=3) compared to WT mice (n=21) Statistics by Mann-Whitney U, day 5, Kolmogrov-Smirnov test. B) Percentage increase in foot swelling in MyD88-/- (n=8-22), TRIF-/- mice (n=10-32) and IPS1-/- mice (n=3), compared to WT mice (n=14-28), statistics by Mann-whitney U test, day 4 and day 5, Kolmogrov-Smirnov test. Data derived from multiple experiments. C) Serum Type I IFN levels in MyD88-/- mice (n=5), TRIF-/- (n=5) mice and IPS1-/- (n=3) mice compared to WT mice (n=7-11). Mann-Whitney U test. (* p ≤0.05, ** p ≤0.01, n.s. not significant).

63 2.3.12 CHIKV infection of IFNAR-/- mice results in rapid, lethal hypovolemic shock.

Mice deficient in the primary type I IFN transcription factors, IRF3 and IRF7, were to shown to have impaired type I IFN production and were more susceptible to CHIKV infection, with IRF3/7-/- mice experiencing a fatal hypovolemic shock (Figure 2.10). Mice deficient in the adaptor molecules upstream of IRF3 and IRF7 - TRIF, MyD88 and IPS-1 were also seen to have impaired type I IFN production that was associated with increased disease severity (Figure 2.11). To confirm the association between increased morbidity following CHIKV infection and defective type I IFN responses, IFNAR1-/- mice, which are incapable of type I IFN signalling, were challenged with CHIKV.

CHIKV infection of IFNAR1-/- mice resulted in a severe and fatal disease phenotype, reminiscent of that seen in CHIKV infection of IRF3/7-/- mice. Mortality was rapid, with all IFNAR-/- mice experiencing extreme foot swelling (Figure 2.12B) and requiring euthanasia within 3 days of infection (Figure 2.12A). Investigations of the clinical indicators of hypovolemic shock in IFNAR-/- mice clearly demonstrated that IFNAR-/- mice were experiencing a rapid, pathological loss of blood volume prior to death. Circulating platelet levels dropped by ~90%, indicative of severe thrombocytopenia (a drop ≥of 70%) and blood loss (Figure 2.12C). The mice also displayed a fever at day 2 post infection, prior to a rapid decrease in body temperature, and by day 3 post infection (Figure 2.12D), urine output was absent, indicative of dehydration and inadequate blood pressure to maintain perfusion of the tissues (Figure 2.12E). Haematocrit readings displayed a mean ~27% increase from pre-infection levels, where an increase of 20% is considered a clinical indicator of hypovolemic shock (Figure 2.12F). No other unusual symptoms, such as neurological involvement, skin conditions etc., were observed (data not shown).

These findings clearly indicate that mortality in IFNAR1-/- mice following infection with CHIKV is the result of hypovolemic shock. Together with the findings in mice deficient in both IRF3 and IRF7, these results further substantiate the association between impaired type I IFN responses and increased susceptibility to CHIKV-induced hypovolemic shock.

64

Figure 2.12 Survival, foot swelling and clinical parameters in IFNAR-/- mice following CHIKV infection. A) Kaplan-Meier survival plot of IFNAR-/- mice and WT mice. Data were recorded when an animal was euthanased in accordance with animal ethics guidlines (n=6 per group). B) Foot swelling in IFNAR-/- mice (n=6) compared to WT mice (n=8-22). The mean percentage increase in metatarsal area is presented (n=6). C) Platelet counts in IFNAR-/- mice at day 0 versus day 3 post CHIKV infection (n=6; p=0.009, t-test). D) Body temperature in IFNAR-/- mice following CHIKV infection (n=6; p=0.04, t-test). E) Urine output in IFNAR-/- mice (n=6; p=0.002, Kolmogrov-Smirnov test). F) Haematocrit levels in IFNAR-/- mice at day 0 and day 3 post infection (n=6; p<0.001; Mann-Whitney U).

65 2.4 Discussion

Results herein demonstrate a critical role for IRF3 and IRF7 during CHIKV infection. Deficiency of either IRF3 or IRF7 in mice following CHIKV infection resulted in a more severe disease phenotype, particularly in the case of IRF7; however, a deficiency in both transcription factors resulted in mortality. CHIKV infected IRF3/7-/- mice displayed increased viral titres and serum pro-inflammatory cytokine levels, severe arthritic disease, altered tissue pathology and signs of fatal hypovolemic shock. Antibody responses were largely unaltered in IRF3/7-/- mice, with vaccination able to protect mice against CHIKV disease. However, prophylactic IFNα treatment only delayed death in IRF3/7-/- mice. Infection of IFNAR-/- mice resulted in the same lethal symptoms as those observed in IRF3/7-/- mice, confirming the protective role of IRF3 and IRF7 in CHIKV is via regulation of type I IFN responses. Infection of mice deficient in the adaptor molecules upstream of IRF3 and IRF7 demonstrated that CHIKV-induced type I IFN production is primarily via the RIG-1/MDA5/IPS-1 and TLR3/TRIFF pathways.

The disease manifestations in CHIKV infected mice were consistent with symtoms commonly observed in humans. In situ hybridisation revealed infection of mature myofibres, consistent with the myalgia commonly experienced in human disease. CHIKV has been reported to infect human muscle satellite cells (Ozden, Huerre et al. 2007) and myofibers in TLR3-/- mice (Her et al. 2014), and RRV is known to infect myofibres (Murphy, Taylor et al. 1973, Morrison, Whitmore et al. 2006). However, infection of mature myofibres following CHIKV infection of WT mice has not previously been reported. Chondrocytes were also shown to be infected by CHIKV, which has also, not previously been reported, although chondritis is commonly reported in human CHIKV (Emilie, Tiong et al. 2014) and likely contributes to rheumatic symptoms. A skin rash is also common in human CHIKV disease (Suhrbier, Jaffar-Bandjee et al. 2012), and may be caused by infection of keratinocytes (Pakran, George et al. 2011). In this study, infection of cells in the epidermis was only observed in IRF3/7-/- mice. IRF3/7-/- mice also experienced a fever and oedema, which are common in human disease (Suhrbier, Jaffar-Bandjee et al. 2012). IRF3/7-/- mice also displayed disease manifestations that have been observed in severe human disease (Economopoulou, Dominguez et al. 2009).

CHIKV infection in humans has occasionally been associated with severe, and sometimes fatal disease outcomes (Economopoulou, Dominguez et al. 2009). Severe disease manifestations including thrombocytopenia, haemorrhage, and hypovolemic shock

66 are more common in children/neonates (Gérardin, Barau et al. 2008), in whom IRF7 responses are known to be defective/underdeveloped (Levy 2007, Danis, George et al. 2008). Severe symptoms were only observed in IRF7-/-, IRF3/7-/- and IFNAR-/- mice, which were deficient type I IFN, suggesting that poor type IFN responses may predispose individuals to severe and atypical CHIKV disease manifestations.

The severe symptoms in IRF3/7-/- and IFNAR-/- mice were also analogous to dengue shock syndrome (DSS) and dengue haemorrhagic fever (DHF) involving haemoconcentration, oliguria, haemorrhage, severe oedema, fever then hypothermia (W.H.O 2009, Huy, Van Giang et al. 2013) and elevated levels of the pro-inflammatory cytokines IFNγ, IL-6, MCP-1 and TNF (Lee, Liu et al. 2006, Pang, Cardosa et al. 2007). IRF7 responses early in infection have been shown to mediate protection against the development DSS/DHF (Hoang, Lynn et al. 2010), and microarray analysis has shown that DSS patients have reduced levels of type I IFN-induced gene transcripts (Simmons, Popper et al. 2007). The paucity of type I IFNs in IRF3/7-/- and IFNAR-/- mice suggests insufficient type I IFN responses may be involved in the development of DHS/DSS.

The absence of type I IFN in IRF3/7-/- and IFNAR-/- mice likely contributes to the development of hypovolemic shock, not as a result of increased viral burden, but due to a lack of regulation of serum pro-inflammatory cytokines. TNF is a potent vasodilator that is produced by macrophages in response to IFNγ (Loppnow, Werdan et al. 2008). The very high levels of these cytokines in CHIKV infected IRF3/7-/- and IFNAR-/- mice are presumably the cause of vascular leakage and shock (Lee, Liu et al. 2006, Pang, Cardosa et al. 2007). Production of IFNγ has been shown to be inhibited by type I IFNs during Trypanosoma cruzi infection (Chessler, Caradonna et al. 2011), as well as viral infections (Nguyen, Cousens et al. 2000). Type I IFN is also known to down-regulate expression of the IFNγ-receptor, thereby suppressing the effects of IFNγ (Rayamajhi, Humann et al. 2010). Thus, in the absence of type I IFN in IRF3/7-/- and IFNAR-/- mice, excessive IFNγ production and downstream pro-inflammatory mediators cannot be effectively controlled, leading to severe vascular leakage and ultimately, hypovolemic shock. Vascular leakage and cytokine dysregulation likely also lead to the extreme foot swelling in the IRF3/7-/- mice.

67 Arthritis in WT mice is associated with a considerable monocyte/macrophage infiltrate (Gardner, Anraku et al. 2010). In IRF3/7-/- mice, however, there was a notable absence of any cellular infiltrates observed, despite a 140% increase in foot swelling (versus 50% in WT mice). Macrophage/monocyte infiltration of the tissues has been shown to be largely dependent on chemokine (C-C motif) receptor 2 (CCR2), which is the receptor for MCP-1/CCL2 (Poo, Nakaya et al. 2014). MCP-1/CCL2 was produced at very high levels in IRF3/7-/- mice. However, type I IFN signalling has been shown to be critical for migration of Ly6Chi inflammatory monocytes during chronic inflammation (Lee, Li et al. 2009). Thus, the absence of type I IFNs in IRF3/7-/- mice may have impeded the capacity for monocyte/macrophage extravasation. Furthermore, CCR2 expression in monocytes and dendritic cells have been shown to be down regulated by high levels of IFNγ (Vecchi, Massimiliano et al. 1999) and TNF (Weber, Draude et al. 1999), respectively. Foot swelling in the IRF3/7-/- mice thus appears to be the result of severe oedema, rather than arthritis.

The generation of intact type I IFN responses was seen to be more dependent on IRF7 than IRF3. In WT mice, IRF7 mRNA levels were up-regulated >100 fold following CHIKV infection. IRF7-/- mice also displayed greater disease severity and reduced serum type I IFN levels, while viral burden, foot swelling and type I IFN production in IRF3-/- mice was not significantly different to WT. These findings were consistent with studies that have shown that IRF7, but not IRF3, is required for optimal type IFN production in MEFs due to the role of IRF7 in induction of IFNαs, which then mediate amplification via a positive feedback loop (Marie, Durbin et al. 1998, Sato, Suemori et al. 2000). Hence in IRF3-/- mice, in which IRF7 could be up-regulated, IFNαs induced and the IRF7-mediated positive feedback loop engaged, disruption of the type I IFN response was minimal. This also implicates plasmacytoid dendritic cells (pDCs) in CHIKV induced type I IFN production, as pDCs can be directly stimulated by virus to produce IFNαs without prior IFNβ induction and signalling (Ning, Pagano et al. 2011). These results also suggest that an important activity of prophylactic IFNα treatment against CHIKV in WT mice is the up- regulation of IRF7, rather than the induction of antiviral effectors. Accordingly, prophylactic treatments were unable prevent mortality in IRF3/7-/- mice.

68 Given the pathology that was associated with increased IFNγ, observed in the IRF3/7-/- mice following CHIKV infection, it is perhaps not surprising that prophylactic IFNγ treatment also failed to protect mice against CHIKV. Adoptive transfer of WT CD45+/CD11b+ cells did not promote survival in the IRF3/7-/- mice, an observation consistent with the findings of Schilte, Couderc et al. (2010) that non-haematopoietic cells are the major source of protective type I IFN during CHIKV infection. In contrast to our findings, Schilte, Buckwalter et al. (2011) showed that adoptive transfer of WT haematopoietic cells into IRF3/7-/- mice was able to protect these mice from lethal CHIKV infection. In the second study by Schilte, Buckwalter et al. (2011), the CHIKV strain CHIKV-21, derived from cerebrospinal fluid of a neonate with atypical central nervous system complications (Sourisseau, Schilte et al. 2007), was used and injected intra- dermally. This study also used bone marrow cells for adoptive transfer. These differences in methodology from the study herein, which used sub-cutaneous injection of CHIKV strain LR_OPY2006 (derived from an adult with typical CHIKV disease; Parola, Lamballerie et al. 2006), and adoptive transfer of spleen-derived cells may account for the differences in the findings.

Treatment with an inactivated-CHIKV vaccine, however, was able to protect IRF3-/-, IRF7-/- and IRF3/7-/- mice against CHIKV disease and mortality, and induced high antibody and neutralising titres, indicating that antibody responses are not reliant on IRF3 and/or IRF7. IRF3/7-/- mice produced higher IgG1 and less IgG2c than WT mice, indicative of a slight Th2 bias. Type I IFNs promote Th1 responses (Bracci, La Sorsa et al. 2008, Farkas and Kemény 2011), and have also been shown to suppress Th2 responses in human CD4+ T cells (Huber, Ramos et al. 2010). It follows that in the absence of detectable levels of type I IFNs in IRF3/7-/- mice, there may be a shift toward Th2 responses. While WT mice display a predominantly Th1 response to CHIKV (Gardner, Anraku et al. 2010), the Th2-biased responses in vaccinated IRF3/7-/- mice were still capable of providing complete protection.

Investigation of the adaptor molecules upstream of IRF3 and IRF7 showed that IPS-1 was the most important for production of type I IFNs. IPS-1-/- mice developed a more protracted viraemia and increased foot swelling. This is consistent with the finding that IPS-1 is important for protection against CHIKV in non-haematopoietic cells (Schilte, Buckwalter et al. 2012), and that non-haematopoietic cells are the most critical to protective type I IFN responses during CHIKV infection (Schilte, Couderc et al. 2010). The

69 importance of IPS-1 in CHIKV infection thus implicates the cytoplasmic RLRs, RIG-1 and MDA5, as having a major role in the detection of CHIKV. Detection of alphavirus RNA by endosomal TLR3 in haematopoietic cells is well reported (Schulz, Diebold et al. 2005, Daffis, Suthar et al. 2009).While TLR3 is heavily expressed in haematopoietic cells, it is also expressed/inducible in endothelium, epithelium, fibroblasts and keratinocytes, which can all be infected by CHIKV (Schwartz and Albert 2010, Takeuchi and Akira 2010). TLR3 signals via TRIF (Takeuchi and Akira 2010), and consistent with the established role of TLR3 in detection of alphaviral RNA, CHIKV-infected TRIF mice displayed reduced serum type IFNs and increased foot swelling. Similarly, Her, Teng et al. (2014) have since demonstrated increased CHIKV viral replication in TLR3-/- and TRIF-/- MEFs, and CHIKV infection of TLR3-/- mice resulted in increased viral titres and more severe inflammation.

TLR7 also detects endosomal viral RNA (Takeuchi and Akira 2010) and is well known to recognise ssRNA viruses (Bauer, Pigisch et al. 2008, Blasius and Beutler 2010). MyD88 is the adaptor molecule employed by TLR7 for IFN signalling (Takeuchi and Akira 2010), and MyD88 deficient mice have been seen to experience increased viral burden following CHIKV infection (Schilte, Couderc et al. 2010). Although mice deficient in MyD88, did exhibit reduced type I IFN production, in this study viraemia and disease in MyD88-/- mice was largely unaltered compared to WT mice. RRV infected MyD88-/- and TRL7-/- mice also display no significant difference in early (up to day 4) viraemia and tissue titres (Neighbours, Long et al. 2012). However, RRV infection in these mice was shown to be lethal, possibly due to a germinal centre defect that results in production of low affinity antibodies with poor neutralisation capacity (Neighbours, Long et al. 2012). Importantly, the study herein was able to demonstrate the relative significance of the individual adaptor molecules involved in type I IFN responses to CHIKV. RLR/IPS-1, TLR3/TRIF and TLR7/MyD88 signalling were all shown to be involved in type I IFN responses to CHIKV, with IPS-1 having the most significant role and TLR7/MyD88 signalling relatively less important. This is consistent with recent findings by Olagnier, Scholt et al. (2014) that the antiviral response to CHIKV is dependent on a functioning RIG-1/IPS-1/TBK1/IRF3 signalling pathway.

In summary, the results presented herein demonstrate a critical role for IRF3 and IRF7 in the generation of effective type I IFN responses against CHIKV infection. IRF7 was shown to be critical for its role in priming and generation of the type I IFN positive feedback loop. These studies also reiterate the central role of type I IFNs in protection

70 against CHIKV infection; the absence of type I IFN resulted in a severe disease that was reminiscent of DHF/DSS, involving haemorrhagic shock and excessive IFNγ and TNF production. The results suggest that defective type I IFN responses may predispose to DHF/DSS and severe CHIKV disease. Induction of type I IFNs and activation of IRF3 and IRF7 by CHIKV was shown to be predominantly via the RIG-1/MDA5/IPS-1 and TLR3/TRIF signalling pathways. However, the primary source of type I IFNs remains unclear and is the basis of the ensuing chapter.

71 3. Investigation of the source of type I interferons following CHIKV infection.

3.1 Introduction

The specific cells and tissues actively engaged in production of type I IFNs following CHIKV infection in vivo remain to be fully characterised. During viral infection, haematopoietic cells engage both TLR and RLR mediated type I IFN production, while stromal cells such as fibroblasts, endothelium and epithelium, as well as dendritic cells and macrophages, generally rely on the RLR system (Akira, Uematsu et al. 2006, Kawai and Akira 2011). Plasmacytoid DCs (pDCs; murine CD11c/B220), often referred to as professional type I IFN producing cells, are reported to be the major IFNα producers (Siegal, Kadowaki et al. 1999, Ning, Pagano et al. 2011) and produce high levels of IFNαs via the TLR7/MyD88 pathway, even in the absence of IFNAR signalling and amplification (Kato, Sato et al. 2005, Ning, Pagano et al. 2011). However, the previous chapter found TLR7/MyD88 signalling to be least important for optimal type I IFN responses to CHIKV. Instead, IPS-1 was seen to have the greatest role in type I IFN responses to CHIKV, thereby implicating RLRs, and the (predominantly) stromal cells that engage RLR signalling, in CHIKV-induced type I IFN production.

Various cell types have been implicated in type I IFN production during CHIKV infection. Her, Malleret et al. (2010) found that human peripheral blood monocytes (PBMCs) were a major target of CHIKV during the acute phase of infection, and that infection of PBMCs resulted in high levels of IFNαs. In contrast, Sourisseau et al. (2007) found that haematopoietic cell cultures, except monocyte-derived macrophages, were refractory to CHIKV infection, suggesting haematopoietic cells produce type I IFNs via paracrine signalling, rather than PRR signalling. Additionally, Schilte, Couderc et al. (2010) found in vivo type I IFN production by haematopoietic cells was not vital for protection against CHIKV, with type I IFN being produced largely by IPS-1-mediated signalling in non-haematopoietic cells. In a zebra fish larvae model of CHIKV infection, IPS-1 was again found to be the major IFN signalling pathway involved. However, uninfected cells, namely neutrophils and hepatocytes were found to be the major producers of type I IFN (Palha, Guivel-Benhassine et al. 2013).

Identification of the cells producing type I IFN proteins in vivo can be difficult due to the low stability, short half-life and transient production of type I IFN mRNA and protein (McKenna, Vergilis et al. 2004). To address this issue, Kumagai, Takeuchi et al. (2007)

72 engineered mice to express enhanced green fluorescent protein (eGFP) under the control of the IFNα6 or IFNβ promoter. GFP is widely used as a fluorescent marker due to its structural stability, autocatalytic chromophore formation and low toxicity (Jackson, Craggs et al. 2006, Chudakov, Matz et al. 2010). The transgenic mice were generated by replacing the entire IFNβ or IFNα6 open reading frame (ORF) with an eGFP ORF and neomycin-resistance gene cassette (Dr Y. Kumagai, personal communication) (Kumagai, Takeuchi et al. 2007). Thus, upon type I IFN stimulation, the transgenic mice should produce a stable, non-secreted, fluorescent protein that enables in situ identification of the cells that have type I IFN promoter activation (Kumagai, Takeuchi et al. 2007).

CHIKV infection is lethal in type I IFN deficient mice (see Chapter 2) (Couderc, Chrétien et al. 2008, Rudd, Wilson et al. 2012). Thus, only mice that retained one WT allele (and therefore, the capacity to produce type I IFNs), in addition to the inserted eGFP transgene were suitable for analysing type I IFN production in vivo, following CHIKV infection. These mice were generated by mating heterozygous transgenic mice (herein referred to as IFNα6-GFP+/- and IFNβ-GFP+/- mice) with WT mice and genotyping the offspring. Heterozygous offspring were infected with CHIKV as before (Gardner, Anraku et al. 2010), with the aim of identifying the tissues and cells producing type I IFNs in response to CHIKV infection.

73 3.2 Materials and Methods

3.2.1 Cells Vero and C6/36 cell lines were maintained as described in section 2.2.2.

3.2.1.1 Bone marrow-derived macrophages (BMMs) and DCs (BMDCs) BMMs and BMDCs were derived from bone marrow in the femur. Differentiated cells were cultured in DMEM containing L-glutamine (Gibco-Invitrogen) and supplemented with 25 mM HEPES (Sigma-Aldrich), 1% 2-ME (Gibco-Invitrogen), Pen-Strep antibiotic (Gibco-Invitrogen) and 10% (v/v) FCS (Gibco-Invitrogen). Briefly, the femur was excised and the marrow was flushed from the bone with DMEM using a 5 ml syringe with 23 gauge needle. Cells were suspended in 2 ml DMEM by gently passing through the syringe needle several times. Cell suspensions were incubated with 3 ml ammonium-chloride- potassium (ACK) buffer (Life Technologies) to lyse red blood cells. After 3 min incubation at room temperature, 5 ml DMEM was added to the tube, the tube was centrifuged (200 g x 5 min), and the resulting cell pellet was resuspended in DMEM and washed (200 g x 5 min) three times. Cells were seeded (~106 cells/well) in a 24 well plate (Nunc) in 300 µl

DMEM supplemented with 10% FCS and incubated at 37˚C in a 5% CO2 atmosphere for 4 hours. Cell monolayers were washed with ~1 ml PBS per well to remove non-adherent cells and 300 µl DMEM supplemented with 10% FCS was added to each well. For generation of BMMs, cells were cultured in 30% L929-cell conditioned media (LCCM) cell culture media. DMEM (30% LCCM) supplemented with 10 ng/ml GM-CSF and 103 IU/ml IL-4 (added every 2-3 days) was used for generation of BMDCs.

3.2.1.2 Resident peritoneal macrophages (RPMs) RPMs were cultured in RPMI1640 media containing L-glutamine (Gibco-Invitrogen) and supplemented with 25 mM HEPES (Sigma-Aldrich), 1% 2-ME (Gibco-Invitrogen), Pen- Strep antibiotic (Gibco-Invitrogen) and 5% (v/v) FCS (Gibco-Invitrogen). Briefly, RPMs obtained by lavage of the peritoneal cavity using sterile, ice cold PBS, and were transferred into a 10 ml tube. Cells were centrifuged (200 g x 5 min)˚C, thenat 4 resuspended and washed (200 g x 5 min, 4˚C), before being counted a nd seeded (1-2.5 x 106 cells/well) into a 24 well plate (Nunc) in 300 µl RPMI1640 media supplemented with

5% FCS and incubated at 37˚C in a 5% CO 2 atmosphere for 3 hours prior to treatment. Non-adherent cells were washed from the plate using sterile PBS prior to treatment (only macrophages are adhered).

74 3.2.1.3 Splenocytes Splenocytes were cultured in phenol red-free RPMI1640 media containing L- glutamine (Gibco-Invitrogen) and supplemented with 25 mM HEPES, 1% 2-ME, Pen-Strep antibiotic and 10% (v/v) FCS. Whole spleens were excised and placed in ~ 1 ml cold PBS supplemented with 2% FCS in a 1.5 ml plastic tube. The contents of the tube were passed through a sterile, nylon mesh using the plunger of a 5 ml syringe, onto a sterile petrii dish on ice. The resulting cell suspension was, transferred to a 10 ml tube and centrifuged (200 g x 5 min, 4˚C). Cells were resuspended in 3 ml ACK buffer and incubated for 3 mins before the addition of 5 ml phenol red-free RPMI1640 supplemented with 5% FCS. Cells were then washed three times (200 g x 5 min,˚C) 4 using phenol red -free RPMI1640 supplemented with 5% FCS, counted and seeded (5 x 106 cells/well) in a 6 well plate in 1 ml phenol red-free RPMI1640 cell culture media.

3.2.2 Transfections Ex vivo BMM cultures (see section 3.2.1.1) were transfected with pEGFP CI mammalian expression plasmid (Clontech Laboratories Inc., CA, USA) using GeneJammer transfection reagent, as per the manufacturer’s instructions (Stratagene, La Jolla, USA). A transfection mixture was prepared by incubating 3% (v/v) transfection agent in serum-free RPMI1640 supplemented with 10%v/v Penstrep antibiotic) for 5 min at room temperature, before adding 2 µg pEGFP CI DNA construct, and incubating for a further 30 min. Cell culture medium was aspirated from BMM monolayers and replaced with 2 ml/well fresh, serum-free RPMI1640. A 100 µl volume of the prepared transfection mixture containing the pEGFP CI construct was then added to each well. Control wells received transfection reagent only (i.e. no eGFP construct). Cells and transfection mixtures were incubated at 37˚C in a 5% CO 2 humidified atmosphere. After 5 hrs, the transfection mix was replaced with RPMI1640 media supplemented with 5% FCS, and cells incubated a further 24 hours prior to use.

3.2.3 LPS, IFN and SFV stimulation of type I IFNs Ex vivo splenocyte cultures (see section 3.2.1.3) were treated with either 100 ng/ml lipopolysaccharide (LPS; E. coli 055B5) (Sigma), or 10 IU/ml IFNαA (Chemicon International, #GF023) in a final volume of 2 ml cell culture media per well. Ex vivo splenocyte, BMM, BMDC and RPM cultures (see sections 3.2.1.1-3.2.1.3) were treated with 10 IU/ml IFNαA (Chemicon International, #GF023) or 10 IU/ml IFNγ (Chemicon International, #IF005) in a final volume of 2 ml cell culture media per well. Cells for IFN

75 stimulation by SFV were incubated with 400 CCID50/ml SFV in 1 ml RPMI1640 media. After 1 hour, media and virus were aspirated and replaced with 2 ml fresh RPMI1640 supplemented with 5% FCS. Cells were monitored for eGFP expression using an EVOS® FL Cell Imaging System (Life Technologies). Prior to imaging, cell culture medium was aspirated and replaced with 1 ml PBS.

3.2.4 Virus The CHIKV isolate used (LR2006-OPY1) and preparation of viral stocks was the same as that described in section 2.2.5.

3.2.5 Mice IFNα6-GFP+/- and IFNβ-GFP+/- transgenic mice were provided by Prof. S. Akira of Osaka University (Kumagai, Takeuchi et al. 2007). C57BL/6J mice (Animal Resource Centre, Canning Vale, Western Australia) were used as aged matched controls. Transgenic mice were genotyped by PCR using DNA isolated from tail tip. Mice ≥at 21 days were anaesthetised using isoflurane (Attane™ Isoflurane, JD Medical Dist. Co. Inc., Phoenix, USA). 1 mm of the tail tip was amputated using a sterile scalpel blade and placed in individual 1 ml tubes on ice for later DNA isolation (section 3.2.10).

Mice were housed in individual ventilated cages and allowed to feed ad libitum. Experiments were conducted using female mice aged 10-18 weeks. All animal experiments were approved by the QIMR Animal Ethics Committee and were conducted in accordance with the Australian code for the care and use of animals for scientific purposes.

3.2.6 Mouse infection, inoculation and monitoring Mice were infected or mock infected with RPMI and monitored as per section 2.2.6. Naive mice were not injected. Lipopolysaccharide (LPS) (Sigma-Aldrich) was administered via intraperitoneal injection at a dose of 100 µg/ml in a 200 µl volume of sterile PBS. At 3 hours post injection, spleens were removed and processed as per section 3.2.7.

76 3.2.7 Blood and tissue preparation Blood and tissues collected for analysis were prepared and stored as described in section 2.2.4.

3.2.8 Viraemia and foot measurements Serum was assayed for viral titres as per section 2.2.8. Foot swelling was measured as described in 2.2.10.

3.2.9 RNA analysis Extraction of RNA from tissue samples, cDNA synthesis and quantitative real time PCR (RT-PCR) was carried as described in sections 2.2.16-18. Primers used were (5’3’), RPL13A gaggtcgggtggaagtacca (forward), tgcatcttggccttttcctt (reverse); IFNβ- eGFP cctacggcgtgcagtgcttcagc (forward), ccttcagctcgatgcggttcac (reverse); IFNβ ctccagctccaagaaaggacg (forward), gcatcttctccgtcatctcca (reverse); IFNα6 tggcaagactgagtgaggacg (forward) atattgctgcagcgtcttggc (reverse).

3.2.10 DNA isolation A prepGemTM Tissue DNA isolation kit (Zygem Inc., Hamilton, NZ) was used to isolate DNA from mouse tail tips (section 3.2.5) according to the manufacturer’s instructions. Scissors, decontaminated with 10% bleach and rinsed with ethanol between samples, were used to macerate each tail tip contained in a 200 µl PCR tube. prepGemTM enzyme, kit lysis buffer and DNA-free water were added to each sample and incubated (75˚C, 15 min; 95˚C for 5 min). Samples were centrifuged briefly, the supernatant transferred to a fresh tube and stored at ˚C.4 The quality and quantity of the extracted DNA was assessed using a NanoDrop 2000 (ThermoFisher Scientific). A260/280 ratios were consistently above 1.8.

3.2.11 Genotyping PCR

3.2.11.1 Primers Primer sequences were obtained from Dr Y. Kumagai of Osaka University and manufactured by GeneWorks (GeneWorks Pty Ltd., Hindmarsh, Australia). Primers specific for p107 were used as a positive control. The primers were: (5’3’) p107 gtgtatggcctgtcctagtc (forward), gcgacactaatccgagaacag (reverse); IFNα6 77 aggcaactcgaagaacccaatttgagg (forward), cagatgagtcctttgatgtgaagaggg (reverse); IFNβ cttcatttctagagtttccgactctgg (forward), gcttatagttgatggagagggctgtgg (reverse); eGFP tttacgtcgccgtccagctcg (reverse, for use with either IFNβ or IFNα6 forward primers, Figure 3.1).

The forward primer for IFNβ was redesigned. The IFNβ coding sequence was obtained using the National Centre for Biotechnology Information (NCBI) Basic Local Alignment Tool (BLAST) database. Oligomers of 24-27 nucleotide base pairs with at least 50% G/C nucleotide content were selected as potential primers. The following sequence exhibited the most favourable properties and was sent for manufacture (Geneworks Pty. Ltd., SA): (5’3’) IFNβ cactcacaacccttcccttcctgg (forward).

Figure 3.1. Schematic diagram of primer design and pairing for detection of transgenic and/or WT alleles. For each sample, two primer pairs were used to probe for DNA corresponding to the WT allele and the transgenic eGFP allele. (R) = reverse primer, (F) = forward primer.

3.2.11.2 PCR reaction mixture and cycling conditions Isolated DNA was amplified by PCR using a protocol kindly provided by Dr Y. Kumagai of Osaka University. The reaction was performed in a 50 µl volume containing 5 µl DNA (50-100 ng/µl), 2 mM each of dATP, dCTP, dGTP and dTTP, 10 µM of each primer, 1x Mg2+-free DyNAzyme™ buffer (Finnzymes Oy, Vantaa, Finland), 1 U

DyNAzyme™ II DNA Polymerase (Finnzymes) 2 mM MgCl2 and deionised water to 50 µl final volume. Cycling conditions were also modified so that the annealing temperatures were better matched to the melt temperature of the redesigned IFNβ primer. Touchdowns were introduced to cover a range of annealing temperatures. The reaction mixture was subjected to the temperature cycling conditions in Table 3.1. 1x Orange G loading dye was added to the PCR products, which were separated on a 1% agarose gel containing 0.2% ethidium bromide and visualised by UV transillumination.

78 Table 3.1 Cycling conditions for genotype of IFNα6-GFP+/- and IFNβ-GFP+/- mice

IFNα6-GFP+/- IFNβ-GFP+/-

Initial Denaturation 94˚C; 2min 94˚C; 2min x1

Denaturation 94˚C; 30sec 94˚C; 30sec

Annealing 61˚C; 1 min 62˚C; 1 min (-0.1˚C/cycle) x35

Extension 74˚C; 1 min 74˚C; 1 min

Final Extension 72˚C; 10 min 72˚C; 10 min x1

3.2.12 Histology All issues excised for histology were fixed in 10% neutral buffered formalin for 48-72 hours to kill virus, in accordance with PC3 regulations. Feet were decalcified in 15 % EDTA/0.1% PBS over 14 days at room temperature.

3.2.12.1 Fluorescence imaging Tissues for fluorescence imaging were rehydrated with 10-20% sucrose-PBS for 24 hours before being frozen at -80°C in optimal cutting temperature (OCT) cryoembedding media (Sakura Finetek,Tokyo, Japan) for cryosectioning. 7 µm frozen sections were affixed to Superfrost Plus adhesive slides (Menzel-Glaser GmBH, Braunschweig, Germany) and air dried overnight. Sections were permeabilised in 0.05% v/v Triton X100 (Life Technologies) in PBS for 2 mins, then washed (2 x 5 min) in 0.025% v/v Tween 20 in PBS (PBST). Slides were blocked with Background Sniper (Biocare Medical, CA, USA) supplemented with 1% BSA for 30 min. Slides were stained with rabbit anti-GFP antibody (Abcam, ab290 1 µg/ml or MBL JM-3999-100 50 µg/ml) in Background Sniper (Biocare Medical) supplemented with 1% BSA for 1 hour. Sections were washed in PBS (3 x 10 min) then incubated with Alexa Fluor® 488 goat anti-rabbit antibody (ab150077, Abcam) diluted 1:300 in PBS for 30 min, in the dark. Sections were washed (3 x 10 min) in PBST, then rinsed in PBS before being incubated with 4', 6-diamidino-2-phenylindole (DAPI) nuclear counterstain diluted 1: 75 000 in PBS for 10 min. Sections were washed (3 x 5 min) in PBS and mounted with ProLong Gold® Antifade mounting media (Life Technologies). Slides were visualised on an Aperio ScanScope FL (Leica Biosystems),

79 and fluorescence was quantitated using ImageScope Positive Pixel Count FL v1 Algorithm software (Aperio).

3.2.12.2 Immunohistochemistry Tissues for IHC were paraffin embedded and cut to 4 µm sections and affixed to Superfrost Plus adhesive slides (Menzel-Glaser GmBH) and air dried overnight. Paraffin sections were dewaxed in xylene and then rehydrated through descending graded alcohol to water. Sections were incubated in Biocare Medical Diva antigen retrieval buffer for 1 hour at 37˚C, washed (3 x 5 min) in water, treated with 0.2% pepsin in 0.1 M HCL for 10 min at 37˚C, then washed in water (2 x 5 min). Tissues were permeabilised using proteinase K in 0.05 M triphosphate buffered solution (TBS)/calcium chloride pH 8 buffer for 10 min, and then washed in water. Sections were blocked with Background Sniper (Biocare Medical) then stained with rabbit anti-GFP antibody (Abcam ab290 10 µg/ml, or MBL JM-3999-100 2.5 µg/ml) in Background Sniper (Biocare Medical) supplemented with 1% BSA for 60 min. Slides were probed with MACH1 mouse probe, then MACH2 anti- rabbit alkaline phosphatase (Biocare Medical). Signals were developed in Warp Red ® chromogen stain (Biocare Medical). Sections were counterstained in haematoxylin, and mounted using Vectasheild® Hard Set Mounting Media (Vector Labs, CA, USA).

3.2.13 Western Blot Excised tissues were placed in a 2 ml reinforced microcentrifuge tubes that contained 4 x 2.8 mm ceramic beads (MoBio Laboratories Inc., Carlsbad, USA) and 1 ml radioimmunoprecipitation assay (RIPA) buffer with Complete Protease Inhibitors Cocktail (Roche) on dry ice, then stored at -80˚C. Thawed tissue samples were homogenised (3 x 30 s, 6800 rpm) using a Precellys 24 Tissue Homogeniser (Bertin Technologies, France). Protein concentrations of the homogenates were estimated using a bicinchronic acid (BCA) protein assay kit (Pierce, Rockford, USA) according to the manufacturer’s instructions. SDS-PAGE sample buffer was added to the samples, which were then denatured by heating (98˚C, 5 min), then separated by electrophoresis on a 12% SDS - PAGE pre-cast gel (Mini-PROTEAN TGX, BioRad, Hercules, USA). Electrophoresed proteins were transferred (1 hour, 100 volts) at˚C 4 onto a nitrocellulose membrane (Hybond-C Extra, GE Healthcare, Rydalmere, Australia). Membranes were blocked for 1 hour at room temperature using bovine lacto transfer technique optimizer (BLOTTO) (1x PBS, 0.1% Tween-20 and 5% skim milk powder) on a rocking platform. Blocked membranes were washed once in 0.1% PBST then incubated with 2 µg/ml purified mouse

80 anti-GFP primary antibody (Invitrogen) in 10 ml BLOTTO overnight at˚C 4 on a rocking platform. Membranes were then washed 5 x 10 min in PBST and incubated for 1 hour at room temperature on a rocking platform with 0.25 µg/ml HRP-conjugated, pre-adsorbed rabbit anti-mouse secondary antibody (Invitrogen) in 10 ml BLOTTO. Membranes were washed 5 x 10 min in PBST, and HRP detected using Western Lightning Chemiluminesence Substrate (Perkin Elmer, Waltham, USA) exposure to x-ray film (Fujifilm, Tokyo, Japan) and/or on an ImageQuant LAS 500 scanner (GE Healthcare).

3.2.14 Statistical analysis Analyses were performed using IBM SPSS Statistics (version 19). The t test was used if the difference in the variances was <4, skewness was > -2, and kurtosis was < 2. Where the data was non-parametric and difference in variances was < 4, the Mann Whitney U test was used; if > 4 the Kolmogorov-Smirnov test was used.

81 3.3 Results

3.3.1 Establishment of an accurate genotyping reaction for identification of IFNα6-GFP+/- and IFNβ-GFP+/- mice.

IFNα6-GFP+/- and IFNβ-GFP+/- mice (Kumagai, Takeuchi et al. 2007) were a kind gift from Prof. S. Akira (Osaka University). To generate sufficient mice for experimentation these animals were mated with WT mice, with half the offspring expected to be heterozygous and therefore able to produce both type I IFNs and eGFP. All progeny were thus genotyped to identify heterozygotes for experimental use.

A lysis buffer protocol provided by Dr Y. Kumagai (Osaka University), a QIAmp DNA Mini Kit (QIAGEN) and a QIAmp Blood Mini Kit were used to extract genomic DNA from tail tip and/or blood. However, DNA yields were of low quality and yield (data not shown). Using a prepGemTM Tissue DNA isolation kit (Zygem Inc.) consistently provided DNA of adequate yield and quality for use in subsequent RT-PCR reactions (data not shown). The PCR products from the WT IFNα6 and IFNβ alleles were expected to be of a similar size (~950 bp) to the transgenic IFNα6-GFP and IFNβ-GFP alleles, respectively, and therefore could not be distinguished in the same reaction. Hence, separate reactions for the WT and the transgenic alleles were conducted for each mouse sample. Primer sequences provided by Dr Y Kumagai (Osaka University) were unsuccessful in the detection of the IFNβ-GFP allele (data not shown) and were redesigned to have a higher annealing temperature. Using the redesigned primer sets for the WT IFNβ and IFNβ-GFP alleles, a PCR product of≈ 900 bp was produced ( Figure 3.2B). The WT IFNα6 and IFNα6-GFP primer sets both produced a PCR product of≈ 1000 bp ( Figure 3.2A). The p107 and WT IFN primer sets served as positive controls for DNA isolation and the PCR reaction.

Samples in which a PCR product was detected for both a WT allele and a transgenic allele were indicative that the mouse was heterozygous (Figure 3.2A, Mouse 1 & 4; Figure 3.2B, Mouse 1 & 5). Samples in which no IFN-GFP allele could be detected indicated that the mouse was homozygous for the WT allele. Mice identified as heterozygous (IFNα6-GFP+/- and IFNβ-GFP+/- mice) were used in subsequent experiments.

82

Figure 3.2 Genotyping of F1 offspring from mating of IFNα6-GFP+/- or IFNβ-GFP+/- mice with WT mice. Genomic DNA isolated from mouse tail tips was amplified by RT- PCR using primers specific for either A) WT IFNα6 (α6) and the IFNα6-GFP transgene (∆α6) or, B) IFNβ (β) and the IFNβ-GFP transgene (∆β), in individual reactions. A mouse was deemed to be heterozygous (GFP+/-) when both the WT IFN and GFP PCR products were detected. Mice were considered homozygous for WT IFN (GFP-/-) when only WT IFN and p107 PCR products were detected. Primers for the p107 gene were used as an additional control for each sample.

83 3.3.2 CHIKV viral burden and pathogenesis was altered in IFNα6-GFP+/- and IFNβ-GFP+/- transgenic mice

To ascertain whether loss of a type I IFN allele by replacement with the eGFP reporter would affect CHIKV disease in mice, IFNα6-GFP+/- mice, IFNβ-GFP+/- mice and WT mice were infected with CHIKV. Viraemia, type I IFN production and foot swelling in transgenic mice were measured and compared with WT controls.

Viraemia in all mice peaked at similar levels on day 2, and was no longer detectable by day 6 post infection. However, viraemia in IFNβ-GFP+/- mice at day 1 post infection was ~1.5 log higher than WT and IFNα6-GFP+/- mice. At day 3 post infection, viraemia in IFNα6-GFP+/- mice was ~1 log higher compared to WT mice (Figure 3.3A). Type I IFN production peaked in all mice at day 2 post infection, and was reduced in both IFNα6- GFP+/- and IFNβ-GFP+/- mice at day 3 post infection compared to WT mice (Figure 3.3B). Foot swelling in IFNα6-GFP+/- and IFNβ-GFP+/- mice was more severe at days 3, 4 and 5 post infection compared to WT mice, with a greater increase in swelling observed in IFNβ- GFP+/- mice at days 3 and 5 (Figure 3.3C). Foot swelling in all mice peaked at comparable levels on day 6 post infection and resolved by day 11/12 post infection.

The results here indicate IFNα6-GFP+/- and IFNβ-GFP+/- mice, have a diminished type I IFN response, albeit to a small extent, following infection with CHIKV. Consequently, viral titres and foot swelling in these mice was also slightly increased. However, the overall disease phenotype in these mice was generally comparable to that seen in WT mice (Gardner, Anraku et al. 2010) suggesting these mice would be suitable for investigating the source(s) of type I IFN production during CHIKV infection.

84

Figure 3.3 CHIKV infection in IFNα6-GFP+/- and IFNβ-GFP+/- mice compared to WT mice. Mice were infected with CHIKV via s.c. injection of the feet and serum was collected at the indicated time points via tail vein bleed. Measurements of the metatarsal region were taken before and after infection. A) Virus titres in the serum of IFNα6-GFP+/- and IFNβ-GFP+/- mice (n=7 per group) compared to WT mice (n=17-28). B) Serum type I IFN levels in IFNα6-GFP+/-, IFNβ-GFP+/- and WT mice (n=4 per group). C) Foot swelling in IFNα6-GFP+/- and IFNβ-GFP+/- mice (n=7 per group) compared to WT mice (n=6-28). Data presented is the mean percentage increase from pre-infection measurements of the metatarsal region (* p<0.05, ** p<0.01; all statistics obtained by Mann-Whitney U).

85 3.3.3 eGFP could not be detected in IFNα6-GFP+/- and IFNβ-GFP+/- transgenic mice following CHIKV infection mice by histological analysis

IFNα6-GFP+/- and IFNβ-GFP+/- mice were infected with CHIKV and their tissues excised for examination by histology at day 2 post infection (peak serum type I IFN levels, Chapter 2). Fixation of all tissues was required to inactivate CHIKV, a PC3 pathogen, prior to histological processing in PC2 facilities.

Formalin fixation stabilises GFP within tissue, but is also known to enhance autofluorescence (Knight and Billinton 2001) and disrupt GFP fluorescence (Swenson, Price et al. 2007). Glycine treatment was used to quench aldehydes that produce autofluorescence. However, fixed, glycine treated cryosections of the brain, lymph node, spleen, liver, kidney, muscle, feet, gut and lung of IFNα6-GFP+/- and IFNβ-GFP+/- mice at day 2 post infection did not produce a fluorescence signal significantly different from the same tissues obtained from uninfected WT mice or uninfected IFNα6-GFP+/- and IFNβ-GFP+/- mice using an Aperio Scanscope FL immunofluorescent slide scanner (Leica Biosystems) (data not shown).

An anti-eGFP antibody (detected with an AlexaFluor-labelled secondary antibody) was used in formalin-fixed frozen sections to detect eGFP protein, as fluorescence may have been disrupted during fixation/processing (Swenson, Price et al. 2007). Again, no eGFP signal was detected above autofluorescence (i) in tissues from WT mice (Figure 3.4A and B), (ii) in tissues from mock-infected transgenic mice, or (iii) in tissues from naïve transgenic mice (data not shown). The same results were obtained using an anti-eGFP antibody from a different supplier (rabbit anti-GFP, from Abcam rather than MBL), and using both antibodies in paraffin sections (data not shown).

Ethanol and acetone also inactivate viruses and fix tissues, without enhancing autofluorescence (Ghendon and Samoilova 1968, Knight and Billinton 2001, Sharma, Gupta et al. 2012). Ethanol, or ethanol and acetone combined, were tested as alternatives to formalin fixation. Using these alternative fixation methods in conjunction with anti-eGFP antibody staining (using either anti-body), no eGFP signal above background signal in WT and naïve transgenic mouse tissues was detected (data not shown).

The results described here indicate either a lack of sensitivity in these well established assays used for detection of eGFP signals, or a lack of eGFP production by the IFNα6-GFP+/- and IFNβ-GFP+/- mice.

86

87 Figure 3.4 Fluorescence in fixed frozen mouse tissue sections taken from IFNα6- GFP+/- and IFNβ-GFP+/- transgenic, and WT mice following CHIKV infection. Mice were infected and tissues excised at day 2 post infection. A) Fluorescence micrographs of brain, lymph node and spleen in IFNα6-GFP+/- and IFNβ-GFP+/- mice compared to WT mice. Images presented are representative of each group (n=3). B) Ratio of GFP- positive pixels to DAPI positive pixels per mm2 of mouse brain, lymph node, spleen and kidney tissue.

88 3.3.4 eGFP could not be detected by western blot analysis of IFNα6-GFP+/- and IFNβ- GFP+/- mouse tissues following CHIKV infection.

Autofluorescence/background signal appeared to impede detection of eGFP fluorescence and fluorescent-labelled antibodies (section 3.3.3). Western blot analysis was therefore undertaken as an alternative method to detect eGFP protein expression in individual tissues taken from CHIKV-infected IFNα6-GFP+/- and IFNβ-GFP+/- mice.

Protein was isolated from brain, lymph node, spleen, liver, kidney, muscle, feet, gut and lung of CHIKV-infected IFNα6-GFP+/-, IFNβ-GFP+/- and WT mice for analysis by western blot using an anti-eGFP antibody. Serum type I IFNs peak at day 2 in WT mice (Chapter2) (Rudd, Wilson et al. 2012). In the experiments here, tissues were excised at 18 hours post infection with the rationale that tissue type I IFN levels are likely to peak before serum type I IFN levels, which reflect spill-over of cytokines from the tissues into the bloodstream. Despite maximal exposure of antibody stained blots, no distinct eGFP signal was detected in any tissues (Figure 3.5).

Figure 3.5 Western blot of eGFP protein expression in IFNα6-GFP+/- mice compared to WT. Western blot of protein isolated from liver and feet of CHIKV-infected IFNα6- GFP+/-, CHIKV infected WT and mock infected IFNα6-GFP+/- mouse tissues at 18 hours post infection. CHIKV infected IFNα6-GFP+/- tissue (+); mock infected IFNα6-GFP+/- tissue (-); CHIKV infected WT tissue (WT); major cross reactive bands (x); marker (M); (C) positive control.

The absence of a positive eGFP signal at 18 hours post CHIKV infection, using a highly specific and sensitive assay, again suggested low level, or absent, eGFP protein production in the IFNα6-GFP+/- and IFNβ-GFP+/- mice.

89 3.3.5 eGFP could not be detected in cultured cells derived from IFNα6-GFP+/- and IFNβ- GFP+/- transgenic mice following type IFN stimulation.

Production of eGFP by IFNα6-GFP+/- and IFNβ-GFP+/- mice in response to CHIKV could not be detected in tissues by fluorescence microscopy, IHC or western blot (see sections 3.3.3 & 3.3.4), suggesting little, or no production of eGFP. To eliminate background signal present in tissues, and establish whether cells isolated from IFNα6- GFP+/- and IFNβ-GFP+/- mice could produce eGFP, ex vivo cell cultures were stimulated with a variety of agonists, and eGFP production monitored over time.

Lipopolysaccharide (LPS) is known to activate type I IFN responses via TLR2/4, and rapidly induces type I IFN production in the spleen (Yamamoto, Sato et al. 2002, Kawai and Akira 2010). In erythrocyte-purified splenocytes derived from IFNα6-GFP+/- and IFNβ-GFP+/- mice 3 hours post i.v. injection of LPS, no positive signal for eGFP was detected by fluorescence microscopy using an EVOS® FL Cell Imaging System (Life Technologies) (data not shown). Type I IFNs themselves induce type I IFN production (Levy, Marié et al. 2011). In ex vivo splenocyte cultures treated with LPS or recombinant IFNαA for 1, 3, 6 or 9 hours, no eGFP signal was detected by fluorescence microscopy (data not shown). SFV, an alphavirus closely related to CHIKV, also induces type IFNs (Breakwell, Dosenovic et al. 2007). However, SFV is a PC2 pathogen and could thus be used to examine virus-induced type I IFN production in a PC2 laboratory, with no requirement for fixation. No eGFP signal was detected over an 18 hour time period in ex vivo splenocyte cultures following infection with SFV. Similarly, treatment of ex vivo RPM, BMM, or BMDC cultures with IFNαA, IFNγ or SFV failed to induce an eGFP signal detectable by fluorescent microscopy at 1, 3, 9 or 18 hours post infection/treatment (data not shown). In each experiment, eGFP transfected BMMs were used as positive controls, with eGFP fluorescence visible from 9 hours post transfection (data not shown).

Together with the data described in sections 3.3.3 and 3.3.4, these results indicate that IFNα6-GFP+/- and IFNβ-GFP+/- mice were incapable of producing detectable levels of eGFP protein in response to type I IFN-inducing stimuli.

90 3.3.6 IFNβ mRNA expression is reduced in IFNβ-GFP+/- mice compared to IFNβ mRNA expression in WT mice following CHIKV infection.

Investigation of eGFP expression by several methods failed to identify the presence of eGFP protein in IFNα6-GFP+/- and IFNβ-GFP+/- mouse tissues. The ability of these mice to produce eGFP mRNA was therefore analysed. Primers targeting mRNA from the eGFP allele and primers targeting the IFNβ allele were used in quantitative real time RT- PCR (RT-PCR) analysis of total RNA isolated from WT and IFNβ-GFP+/- mouse feet at day 2 post infection.

mRNA from the IFNβ allele and the eGFP allele were successfully amplified, indicating that the transgenic eGFP allele (and the IFNβ allele) were transcriptionally active and inducible in IFNβ-GFP+/- mice following CHIKV infection (Figure 3.6A). Both alleles are driven from the same promoter, and as expected, both the IFNβ allele and the eGFP allele in IFNβ-GFP+/- mice were induced to a similar degree at day 2 post infection (Figure 3.6A).

To determine whether IFNβ mRNA expression levels were different in IFNβ-GFP+/- mice compared to WT mice, qRT PCR for IFNβ mRNA in feet from IFNβ-GFP+/- and WT mice taken at day 2 post infection was undertaken. IFNβ mRNA levels in IFNβ-GFP+/- mice were shown to be ≈ 6.5 fold lower than IFNβ mRNA levels in WT mice (Figure 3.6B).

Taken together these results indicate that, despite similar induction of IFNβ and eGFP in IFNβ-GFP+/- mice, IFNβ mRNA levels were considerably reduced in IFNβ-GFP+/- mice compared with WT mice. It is possible that the IFNβ promoter was less activated in IFNβ-GFP+/- mice than in WT mice, with less than expected mRNA levels transcribed from the IFNβ and eGFP alleles as a consequence. A scarcity of mRNA available for translation likely contributed to the absence of detectable levels of eGFP protein in these mice. The reduced IFNβ responses in IFNβ-GFP+/- mice also indicates that this transgenic mouse model behaves differently with respect to type I IFN production compared with WT mice following CHIK infection and was not appropriate for investigation of CHIKV-induced type I IFN expression.

91

Figure 3.6 IFNβ and eGFP mRNA expression by RT-PCR. Total RNA was isolated from the feet of mice at day 2 post infection and analysed by RT-PCR. A) IFNβ and eGFP mRNA induction in the feet of IFNβ-GFP+/- mice (n=3). Values are normalised to RPL13a mRNA levels and are presented relative to mock-infected controls (n=3). B) IFNβ mRNA expression in the feet of IFNβ-GFP+/- versus WT mice, compared to mock infected animals of the respective genotype. Values presented are normalised to RPL13a mRNA levels (n=3 per group).

92 3.3.7 Analysis of IFNα6 and IFNβ mRNA induction in WT mouse tissues following CHIKV infection

IFNβ-GFP+/- and IFNα6-GFP+/- mice were found to be unsuitable for analysing IFN production following CHIKV infection. Visualisation of individual cells producing type I IFN via eGFP detection could therefore not be realised using this system. However, using samples from CHIKV-infected WT mice that have intact type I IFN responses, RT-PCR could be used to investigate the levels of IFNα6 and IFNβ mRNA induction.

IFNα6 and IFNβ mRNA induction was investigated in different tissues from WT mice at day 2 post CHIKV infection. IFNβ and IFNα6 mRNA induction was greatest in feet, the site of virus inoculation and arthritic disease, with induction >10 fold higher than in the other tissues analysed (Figure 3.7A). IFNβ mRNA was also substantially induced in spleen and lymph nodes, with IFNβ mRNA induction significantly higher than IFNα6 mRNA induction in these lymphoid organs (Figure 3.7A).

Simple linear regression analysis demonstrated a correlation with a high R2 value between IFNα6 and IFNβ induction in tissues and the viral tissue titres (see Chapter 2) at day 2 post infection (Figure 3.7B); kidney and gut virus titres were not analysed. The relationship between mRNA induction and viral titres was supported by Spearman’s rank coefficient analysis (Corder and Foreman 2011) (Figure 3.7C). The only tissue in which mRNA expression did not correspond to viral titres was the brain, in which induction of both IFNβ and IFNα6 mRNA was similar to induction in the liver, despite a ~2 log lower tissue viraemia.

This analysis suggests that type I IFN induction (at least on day 2) correlates with virus levels, irrespective of the tissue that is infected, with the exception of brain (see discussion), and that type I induction at day 2 post infection is heavily determined by local infection, rather than systemic.

93 94 Figure 3.7. IFNβ and IFNα6 mRNA induction in WT mice at day 2 post infection CHIKV infection. A) Total RNA was isolated from the feet of mice at day 2 post infection and analysed by RT-PCR using primers targeting IFNβ and IFNα6. Values are normalised to RPL13a mRNA levels and are relative to mock infected controls (n=3 mice per group; * = p < 0.05; t-test). B) Simple linear regression analysis of IFNβ and IFNα6 induction versus viral titres in the tissues at day 2 post infection. C) Analysis of the relationship between IFNβ and IFNα6 mRNA induction versus viral titres in the indicated tissues at day 2 post infection by Spearman’s rank correlation test. Spearman’s coefficient of correlation (rho) is provided for each tissue.

95 3.4 Discussion.

Work herein sought to determine, in vivo, the cells and tissues primarily involved in CHIKV-induced type IFN production using mice that were heterozygous for expression of eGFP under the control of the IFNα6 or IFNβ promoter (IFNα6-GFP+/- and IFNβ-GFP+/- mice, respectively). Serum type I IFN was reduced in the IFNα6-GFP+/- and IFNβ-GFP+/- mice compared to WT following CHIKV infection, and low eGFP expression, together with high autofluorescence prevented detection of eGFP fluorescence and protein. Investigation of mRNA levels showed that IFNβ mRNA was heavily reduced in IFNβ- GFP+/- mice compared to WT controls. Ultimately, the results demonstrated that the loss of one type I IFN allele by replacement with an eGFP reporter construct resulted in an altered type I IFN response, and these mice were not feasible for identification of the source of CHIKV-induced type I IFNs. However, in CHIKV infected WT mouse tissues, RT-PCR analysis showed a high level IFNβ and IFNα6 mRNA induction in feet and lymph nodes, and a strong, linear correlation between IFNβ and IFNα6 induction and viral tissue titres was observed.

A low signal to noise ratio resulted in a failure to detect an eGFP signal. Autofluorescence is a common impediment to detection of eGFP signal. Sources of autofluorescence that overlap with the eGFP spectra and were likely present in the tissues analysed, include the cofactor NAD(P)H, (Wunder, Brock et al. 2005), lipofuscin pigment (Seehafer and Pearce 2006), advanced glycation end-products (Billinton and Knight 2001), collagen (Gareau, Bargo et al. 2004), elastin (Gerson, Goldstein et al. 2009), erythrocytes (Masilamani, Al-Zhrani et al. 2004) and the ubiquitous cofactors, flavins (Billinton and Knight 2001). Furthermore, heavy formalin fixation (>24hr) required to inactive CHIKV, enhances flavin autofluorescence (Billinton and Knight 2001). Glycine treatment, employed to counter formalin-induced autofluorescence, did not improve the signal to noise ratio, nor did the use of the alternative fixatives, acetone and ethanol (Zhang, Chen et al. 2010). Formalin fixation can also impede antibody staining of tissues (Werner, Chott et al. 2000, Brazelton and Blau 2005), while antigen retrieval and decalcification of hard tissues, exposes tissue sections to harsh heat and/or pH conditions that degrade the eGFP epitope (Shi, Shi et al. 2011), resulting in low signal. The poor signal to noise ratio, despite methods employed to counter autofluorescence, was suggestive of low eGFP protein quantity.

96 A low abundance of eGFP protein was confirmed by protein detection assays that negate the problem of autofluorescence. Splenocytes, macrophages and DCs are all capable of producing type I IFNs (Akira, Uematsu et al. 2006, Kawai and Akira 2010), while LPS, IFNα and SFV are potent inducers of type I IFN (Yamamoto, Sato et al. 2002, Breakwell, Dosenovic et al. 2007, Levy, Marié et al. 2011). Yet, primary cultures of these cells derived from IFNα6-GFP+/- and IFNβ-GFP+/- mice did not produce detectable eGFP when exposed to LPS, IFNα, IFNγ and SFV. Western blot of CHIKV infected IFNα6- GFP+/- and IFNβ-GFP+/- mouse tissues, also failed to detect eGFP protein. eGFP protein folding is stable and efficient at, or below, room temperature (~20°C) (Tsien 1998). At higher temperatures, (i.e. mouse body temperature of ~36.9°C) (Rudd, Wilson et al. 2012), eGFP is prone to misfolding. Although eGFP is commonly used at 37 °C, any impediments to eGFP protein maturation may have been particularly apparent in these studies where eGFP protein expression was already low.

The method used to generate the IFNα6-GFP+/- and IFNβ-GFP+/- transgene may have hindered eGFP protein translation in these mice. The inserted eGFP fragment was flanked by a 6 bp linker sequence (Y. Kumagai, personal communication), which effectively displaced the natural IFN Kozak sequence of the IFNα6 and IFNβ gene. Kozak sequences occur immediately upstream of the translational start codon A A (GCCGCC /GCCATGG, where /G can be A or G and ATG is the start codon) in eukaryotes, and are involved in ribosomal translation initiation (Kozak 1987, Kozak 1987). Displacement or removal of the Kozak sequence heavily impairs the translational efficiency of an mRNA (Kozak 1986). Conversely, insertion of a strong Kozak sequence can be used to enhance mRNA translation (Memari, Ramanan et al. 2010). Thus, displacement of the natural Kozak sequence in the IFNα6 and IFNβ 5’ UTR may have impaired translation of the IFNα6-GFP+/- and IFNβ-GFP+/- transgene, again contributing to low eGFP signal.

The construction method used to produce IFNβ-GFP+/- mice likely also hindered transcription of IFNβ and eGFP driven by the IFNβ promoter. Early, IFNβ transcription is monoallelic (transcribed from only one allele) (Hu, Sealfon et al. 2007). Monoallelic IFNβ production up-regulates enhancesome molecules (namely IRF7) that facilitate a switch to biallelic expression, paracrine IFNβ signalling and engagement of the positive feedback loop (Apostolou and Thanos 2008, Rand, Rinas et al. 2012, Zhao, Zhang et al. 2012). Replacement of one IFNβ ORF with the eGFP ORF in this biallelic system would be

97 expected to halve the level of IFNβ mRNA transcribed, and reduce positive feedback amplification (since eGFP does not signal through the IFNAR receptor). However, IFNβ mRNA levels in the IFNβ-GFP+/- mice compared to WT mice were up to≈ 6.5 fold lower, possibly due to the stochastic nature of IFNβ expression. Not all infected cells commit to biallelic IFNβ production (Rand, Rinas et al. 2012, Zhao, Zhang et al. 2012) and this is influenced by the abundance of signalling molecules available during monoallelic IFNβ production (Apostolou and Thanos 2008, Rand, Rinas et al. 2012). Cells that produce early monoallelic eGFP, rather than IFNβ, would not have upregulated the signalling molecules required to make the switch to biallelic production, resulting not only in diminished type I IFN transcription in cells, but also a lower number of infected cells committed to type I IFN production (and therefore eGFP).

It is important to note that IFNα6-GFP+/- mice have been used to detect IFNα6 expression (work using mice in which the IFNβ allele was replaced with the eGFP ORF has not been published) (Kumagai, Takeuchi et al. 2007). Kumagai, Takeuchi et al. (2007) found eGFP+ cells were most abundant in the spleen of Newcastle disease virus (NDV)-infected IFNα6-GFP+/- mice. The percentage of eGFP+ splenocytes was ≤ 0.001%, however, with even lower levels in the liver, bone marrow and lymph node. Furthermore, these cells were identified by FACS analysis, which was not possible in PC3 facilities. In unfixed cryosections of NDV-infected spleen, eGFP+ cells were visible, albeit very sparse, supporting the notion that fixation-enhanced autofluorescence overwhelmed any eGFP signals in the studies here. Lienenklaus, Cornitescu et al. (2009) successfully monitored IFNβ expression using a similar model in which the entire IFNβ ORF was replaced with a luciferase ORF. Luciferase does not require light excitation, produces a stronger signal than eGFP (Caceres, Zhu et al. 2003, Choy, O Connor et al. 2003) and may therefore have negated the low signal to noise ratio encountered here. However, luciferase reporter assays require live cell or antemortem analysis (Caceres, Zhu et al. 2003, Choy, O Connor et al. 2003) and are not useful for in vivo detection of expression in individual cells (Caceres, Zhu et al. 2003), which was the initial aim of this study.

Investigation of type I IFN induction in WT mouse tissues by RT-PCR showed induction of IFNα6 and IFNβ was greatest in the feet, and correlated with viral titres in the tissues. Type I IFN was also heavily induced in the in the spleen and lymph node. The high levels of type I mRNA induction in the lymphoid tissues may reflect the large immune 98 cell population in lymphoid organs. However, the significant, linear correlation observed between type I mRNA induction and viral titres in the tissues, suggested that type I IFN induction is influenced by local infection levels, rather than cell/tissue type. Consistent with this, mRNA induction and viral tissue titres appeared to match the proposed route of CHIKV spread, with infection starting at the site of inoculation before moving to the draining lymph nodes, then entering the circulation (Schwartz and Albert 2010). An exception to the observed correlation was brain tissue. Defective type I IFN responses are known to increase the neurotropism and encephalitic manifestations of several viruses, including CHIKV (Abraham, Mudaliar et al. 2013, Sorgeloos, Kreit et al. 2013). The higher type I IFN induction in the brain, relative to viral titres in other tissues, may reflect increased protective measures in neural tissue to protect against neuropathology in the host.

While the cells primarily involved in type I IFN production following CHIKV infection are not well characterised, several studies (Schilte, Couderc et al. 2010, White, Sali et al. 2011, Palha, Guivel-Benhassine et al. 2013), including the previous chapter, show that CHIKV induces type I IFNs largely via RLRs, in particular RIG-1/IPS1. The IFNα6-GFP+/- and IFNβ-GFP+/- transgenic mice employed in this chapter to determine the cells producing findings proved to be unsuitable for study of the type I IFN response. However, analysis of type I IFN induction and viral titres in WT mice, suggested that detection of local viral infection, rather than a particular cell type/s, may influence the source/s of type I IFN in response to CHIKV infection.

99 4. Quality control and overview of RNA-Seq data from CHIKV infected mice

4.1 Introduction Insights into gene regulation and the transcriptome have improved vastly with the development of next generation sequencing technologies. While microarrays use hybridisation technology to examine the transcriptome, they rely on known genome sequences and are suboptimal in their detection and quantitation of genome activity (Wang, Gerstein et al. 2009, Ozsolak and Milos 2010). RNA-Seq provides comprehensive details of total genome activity via assembly of high throughput cDNA sequences (Ozsolak and Milos 2010, McGettigan 2013). Because, RNA-Seq involves multiple processes from library preparation, to computational analysis, all of which can alter the vast quantity of data generated (Nekrutenko and Taylor 2012), a high level of quality control and reporting is required for interpretation and reproducibility of results.

RNA-Seq uses next generation sequencing technology to perform massively parallel sequencing of cDNA generated from isolated RNA. cDNA is randomly fragmented and adapters are ligated to each strand before adsorption to the internal surface of the sequencer flow cell via primer-adaptor binding (Mutz, Heilkenbrinker et al. 2013). Hybridised fragments are isothermally amplified to generate clusters of approximately one thousand copies of the initial cDNA fragment. Cluster generation enhances signal detection during sequencing and enables good sequence coverage of low abundance fragments (Mutz, Heilkenbrinker et al. 2013). Sequencing can be initiated at one end only (single-end read) or both ends (paired-end reads) of a bound fragment to produce read sequences of 25 – 450 bp, with longer, paired-end protocols allowing greater precision in mapping reads that span splice junctions (McGettigan 2013, Mutz, Heilkenbrinker et al. 2013)

Read mapping and assembly algorithms are used to align the hundreds of millions of read sequences generated to a reference genome, assemble the reads into contiguous sequences, and estimate the relative abundance of individual gene and isoform transcripts (Garber, Grabherr et al. 2011). Thus, RNA-Seq can provide detailed insight into genome- wide transcriptional activity, including the abundance of known genes and their individual

100 transcripts/isoforms, as well as non-coding RNAs, and novel/unannotated gene and isoform transcripts (Garber, Grabherr et al. 2011, Mutz, Heilkenbrinker et al. 2013).

RNA-Seq of cells/tissues during infection results in amplification and sequencing of both host and pathogen RNA. Using “computational subtraction,” reads are first mapped to the host genome, and any unmapped reads are then mapped to the pathogen genome or assembled into transcripts by de novo assembly (transcript reconstruction in the absence of a reference genome) (Westermann, Gorski et al. 2012). While RNA-Seq of bacterial pathogens is well established (Westermann, Gorski et al. 2012, McClure, Balasubramanian et al. 2013), RNA-Seq of small RNA viruses from infected samples can be difficult due to the low abundance and high diversity of viral RNA, relative to host RNA (Marston, McElhinney et al. 2013).

To examine the transcriptional changes within CHIKV-infected tissues, RNA-Seq analysis was undertaken on feet and lymph node tissues - the two tissues with the highest type I IFN induction and viral titres (reported in the previous chapters). The Illumina HiSeq platform was used to generate paired-end reads of 100 bp in length, and the Tuxedo pipeline of RNA-Seq data analysis algorithms was used to align and assemble reads, and estimate the relative abundance and differential expression values. Data quality at every stage of processing was checked against the ENCODE Consortium Standards, guidelines and best practice for RNA-Seq v1.0 (ENCODE 2011).

101 4.2 Materials and Methods

4.2.1 Mice

Experiments were conducted using female, C57BL/6 mice aged 10-11 weeks old (Animal Resource Centre, Canning Vale, Western Australia). Mice were housed in individual ventilated cages and allowed to feed ad libitum. All animal experiments were approved by the QIMR Berghofer Animal Ethics Committee and were conducted in accordance with the Australian code for the care and use of animals for scientific purposes.

4.2.2 Virus

The CHIKV isolate (LR2006-OPY1) (Parola, de Lamballerie et al. 2006) and preparation of viral stocks was described previously in section 2.2.5.

4.2.3 Mouse infection and tissue preparation

4 Twenty four mice received 10 CCID50 CHIKV in 40 µl RPMI1640 media supplemented with 2% FCS via subcutaneous injection of the dorsolateral metatarsal region of each of the hind limbs, injecting toward the ankle. Another 12 mice received 40 µl RPMI1640 supplemented with 2% FCS only, and were used as mock-infected controls. Following infection, mice were monitored daily. The 12 mock-infected mice and 12 of the CHIKV-infected mice were euthanased at day 2 post infection. The remaining 12 CHIKV- infected mice were euthanased at day 7 post infection. Euthanasia was performed by CO2 asphyxiation. Feet and inguinal lymph nodes were excised. Feet were cut in half and lymph nodes were trimmed of adipose tissue, before being placed in RNAlater (Life Technologies) at 4 C. After 24 hours, excised tissues were transferred to a 2 ml screw-top microcentrifuge tube⁰ containing 4 x 2.8 mm ceramic beads (MO BIO Inc., Carlsbad, USA) and TRIzol reagent (Invitrogen) (1 ml per foot/0.5 ml per pair of inguinal lymph nodes) and stored at -80 C for later RNA isolation.

⁰

4.2.4 RNA extraction

RNA isolation was carried out using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. Briefly, the samples were homogenised in a Precellys24 Tissue Homogeniser (Bertin Technologies, Montigny-le-Bretonneux, France) (3 x 30 s,

102 6000 rpm, placed on ice for 1 min between intervals to prevent overheating). Homogenates were centrifuged (12,000g x 10 min), supernatant (~ 800 µl feet, 400 µl lymph nodes) was transferred to a fresh tube, and chloroform (Sigma-Aldrich) was added (200 µl for feet, 100 µl for lymph node). Tubes were shaken by hand and incubated at room temperature for 2 mins, then centrifuged (12,000 x g, 12 min) at 4° C. A 400-600 µl volume of the resulting aqueous phase was transferred to a fresh tube and incubated with 500 µl isopropanol for 10 min at room temperature before centrifugation (12,000g x 10 min) at 4˚C. The resulting supernatant was carefully aspirated and discarded and the remaining RNA pellet was washed in 1 ml 75% ethanol, and resuspended in 20 μl RNAse/DNAase-free water (Sigma). The RNA concentration and purity of each sample was quantified using a Nanodrop ND 1000 (NanoDrop Technologies Inc.).

4.2.5 Pooling and purification of RNA

RNA from both inguinal lymph nodes was used, and for each animal, the foot that yielded the highest quality RNA was used for DNAase treatment, pooling and RNA-Seq analysis. RNA (5 ng) from each of the 12 foot and lymph node samples were pooled to generate three replicates per time point of 20 ng each (i.e. RNA from four mice per replicate, three replicates per sample). The subsequent 20 ng, pooled sample replicates of RNA were then treated for DNA contamination and purified using an RNeasy MinElute Kit with RNAse-Free DNAse Set (QIAGEN) according to the manufacturer’s instructions. Briefly, each 20 µg pooled RNA was incubated with 10 µl RDD buffer and 7.5 Kunitz units (2.5 µl) of DNAse 1 and RNAse-free water to final volume 100 µl and incubated at room temperature for 10 mins. DNAse treated replicates were then loaded into columns for purification. RNA was precipitated by adding 700 µl buffer RLT (supplemented with 0.001% β-mercaptoethanol) and 500 µl 100% ethanol, then bound to the column membrane by centrifugation (15 s x 8000 g). Bound RNA was then washed using 500 µl buffer RW1 (15 s x 8000 g), followed by 500 µl 100% ethanol (15 s x 8000 g) then dried (5 min x 14,000 g). RNA was then eluted (1 min x 14,000 g) with RNAse/DNAse-free water (Sigma) in a 12 µl final volume. The DNAse treated, purified RNA was stored at -80 ºC.

4.2.6 High throughput RNA sequencing (RNA-Seq)

Sequencing of RNA was conducted by the Australian Genome Research Facility (AGRF) in Melbourne. This involved using a TruSeq RNA Sample Prep Kit (v2) (Illumina Inc. San Diego, USA) according to the manufacturer’s instructions. Briefly, poly-A

103 containing RNA was isolated from each 10 µg sample of total RNA using oligo-dt beads, then fragmented using divalent cations under heated conditions. First strand cDNA was reverse-transcribed from the fragmented RNA using random primers, before generation of double-stranded DNA by second-strand cDNA synthesis. Blunt-end fragments were generated by end-repair processing, before adenylation of the 3’ end of each cDNA fragment. Adenylated fragments were ligated to an adapter (to be used in cluster/library generation), and a barcode. Prepared mRNA libraries were mixed and sequenced from both ends in three lanes of an Illumina HiSeq2000 Sequencer (Illumina Inc.). The CASAVA v1.8.2 pipeline was used to separate the bar-coded sequences in each lane into their samples and extract 100 bp, paired-end read sequences in FASTQ output format.

4.2.7 Data Analysis using the Tuxedo pipeline

4.2.7.1 Read mapping and identification of splice junctions Bowtie v2.0.2 (Trapnell, Pachter et al. 2009) was used to align paired end read sequences to the UCSC Mus musculus full genome build (mm10, Dec. 2011). Default parameters were used. Reads that mapped to the genome multiple times (default: ≥10 times) were considered failed reads mapping to low complexity regions of the genome, and were discarded. Reads that contained few mismatches (default: ≤ 2 mismatches) in the high quality 5’ end of the read (default: 5’-most 28 bp) were considered successful. Bowtie output is in the form of binary alignment map (.bam) files. Reads that did not map directly to the reference genome using Bowtie were termed initially unmapped reads, while successfully mapped reads were termed accepted hits. Tophat v2.0.6 (Trapnell, Pachter et al. 2009, Trapnell, Roberts et al. 2012) identifies exons and splice junctions in (i) the accepted hits and (ii) the initially unmapped read files generated by Bowtie. Successfully aligned reads form ‘coverage islands’ of contiguous sequence, or putative exons. Tophat searches within each coverage island and any surrounding islands for canonical intron/exon donor/acceptor (GT-AG) sites for splice junctions. Sequences (default: 5 bp) upstream and downstream of potential junction sites were joined to make ‘seeds.’ Tophat searched the initially unmapped read files for sequences that contained seed sequences as evidence of a splice junction. Initially unmapped reads that were found to contain a splice junction could then be mapped to the reference genome.

Paired-end reads that did not map to the mouse reference genome using Bowtie/Tophat were then mapped to the CHIKV (LR2006-OPY1) genome (Parola, de

104 Lamballerie et al. 2006) using Bowtie. Because the CHIKV genome does not contain intron and exons, Tophat was not used to analyse splice junctions in the Bowtie output. Reads aligned to the mouse genome by Tophat were used as input for transcript assembly and analysis of differential gene expression between mock and infected samples.

4.2.7.2 Transcript assembly and differential expression analysis The Cufflinks v2.1.1 (Trapnell, Williams et al. 2010, Trapnell, Roberts et al. 2012) suite was used to assemble transcripts and estimate relative abundance. Using mapped read output from Tophat, the Cufflinks algorithm assembled mapped reads into transcripts of genes and estimated their abundance in each sample. Cufflinks takes into account the effect of exon length on the number reads/fragments that will map to a particular region, and estimates gene/transcript abundance in fragments per kilobase (of exon) per million fragments mapped (FPKM). Assembled genes and transcripts were assigned an XLOC_ (gene) or TCONs (transcript/isoform) number. Cuffmerge was then used to take transcript assemblies for each sample replicate and perform a reference annotation based transcript (RABT) assembly with the Genome Research Consortium mouse build 30 (GRCm38). RABT assembly merges transcript assemblies with the reference genome to produce a single merged annotation file (merged.gtf) that matches cufflinks generated XLOC_ identifiers with their location, gene name and accession numbers in the reference genome. Cuffdiff was used to run pair wise analyses between each of the samples in the annotated Cuffmerge output file. The fold change and q-value of differential gene and transcript expression (including novel/unannotated genes/isoforms), as well as promoter usage, coding sequence and splicing differences between each sample was calculated. The q- value is an adjusted p-value, or threshold below which the false discovery rate (proportion of false positives among the rejected null hypotheses) can be considered statistically significant (Storey 2002).

4.2.8 Preparation of DEGs lists

To generate a list of differentially expressed genes (DEGs) and their individual transcripts in the different tissues for functional analysis, output files generated in Cuffdiff were sorted and filtered using Microsoft Excel. Each file was filtered to include only the gene/isoform transcripts that displayed a statistically significant change in expression from mock samples. Transcripts for which the q-value (false discovery rate adjusted p-value) was ≤ 0.01 were selected. Of these, transcripts with a fold change of ≥ 2 from mock were 105 selected. FPKM < 1 is a heuristically determined threshold, commonly used to exclude ultra-low, (presumably) biologically insignificant changes from analysis (Hart, Komori et al. 2013, Love, Huber et al. 2014). Thus, transcripts up-regulated to < 1 FPKM, or down- regulated from < 1 FPKM were excluded from DEG lists.

4.2.9 Integrated Genome Viewer

Reads alignments generated by Bowtie and Tophat were visualised in the Integrated Genome Viewer (IGV) (Robinson, Thorvaldsdóttir et al. 2011, Thorvaldsdóttir, Robinson et al. 2012). The IGVtools count algorithm was used to compute the average density of reads aligned to the reference genome per 25 bp in each sample. The resulting transcriptional data file (.tdf) enabled the transcriptional profile across the entire genome to be viewed.

4.2.10 Assessment of data quality

FASTQC software (Andrews 2010) was used asses the quality of raw FASTQ sequences. SAMtools (Li, Handsaker et al. 2009) was used to convert .bam files generated by Bowtie/Tophat to sequence alignment/map (.sam) files. SAMtools was then used to extract read counts and quality scores (MapQ scores) from Bowtie/Tophat for assessment of mapping quality. Only reads with a MapQ ≥ 20 (i.e. uniquely mapped reads) were used for downstream analysis.

106 4.3 Results

4.3.1 Sequence quality assessment

The quality of base calling by most sequencing platforms commonly reduces as the run progresses, resulting in lower quality base calling toward the end of each read/sequence (Andrews 2010, Schmieder and Edwards 2011). FASTQC software was used to assess raw sequence data generated using the Illumina HiSeq2000 sequencing platform, and determine whether the data required trimming of poor quality bases prior to analysis.

Figure 4.1 displays the average quality scores per base along each 100 bp read in one replicate. The quality of base calling falls within the ‘very good’ (green area) quality score range of 30-40 (Figure 4.1). The average sequence quality of all samples was 36.29 ± 1.00 (range: 35.82 - 38.37). No warnings regarding duplicate sequences, overrepresented sequences, GC content or ambiguous base call numbers for any of the samples were flagged (data not shown).

FASTQC analysis of raw sequences indicated that all data was of a high quality and no reads in any samples required trimming of poor quality sequence prior to read mapping using Tophat.

Quality score across all bases (Sanger/Illumina 1.9 encoding) 40

30 20 10 0

Quality score

1 2 3 4 5 6 7 8 9 15-19 25-29 35-39 45-49 55-59 65-69 75-79 85-89 95-99 Position in read (bp) Figure 4.1 FastQC assessment of sequence quality scores. Box and whisker plot of average base calling/sequence quality across each read. Data shown is from the sample with the highest number of reads (day 7 foot) and are representative of all samples.

107 4.3.2 Sequence depth and mapping quality

Sequencing depth can impact on the type of analyses that can be done. The ENCODE Consortium Standards, Guidelines and Best Practices for RNA-Seq recommends ~30 million paired-end reads for evaluation of transcriptional profiles. For quantitation of transcripts and identification of novel transcriptional activity, a total depth of 100-200 million paired end reads ≥ 76 bp in length is recommended (ENCODE 2011). Highly expressed transcripts generate more reads, thus the percentage of mapped reads is used to indicate the depth of data in RNA-Seq experiments, rather than sequencing coverage, with > 80% generally reported for mammalian genomes (Mortazavi, Williams et al. 2008, Ruffalo, LaFramboise et al. 2011, Twine, Janitz et al. 2011, Heruth, Gibson et al. 2012). SAMtools software is a set of tools for post-processing of read alignment output and was used to extract read mapping information from Bowtie/Tophat alignment output files.

A total of 486 million, 100 bp paired-end reads were produced from the 18 samples sequenced (Foot: 3 x mock, day 2 and day 7; lymph node: 3 x mock, day 2, and day 7). Reads were aligned to the UCSC mm10 mouse genome build. Sequence fragments of which neither read mate mapped to the mouse genome were then aligned to the CHIKV (LR2006_OPY1 strain) genome. The number of reads in each sample that mapped to the genome with a MapQ ≥ 20 (i.e. the read mapped to only one position in the entire genome) are summarised in Table 4.1. The average percentage of reads that mapped uniquely to the mm10 mouse genome ranged from 84.2 – 91.2% per sample (Table 4.1 and Figure 4.1). The number of reads mapped to the CHIKV genome was 8.3% in the foot at day 2 post infection, and substantially less than 1% in all other samples (Table 4.1). In all samples, ~ 5% of reads did not map to either genome with a MapQ ≥ 20, and these reads were excluded from further analysis.

The data here indicated that the sequence depth and quality of reads mapped to the mouse genome was sufficient for quantitation of transcripts and investigation of novel transcripts. However, the percentage of total reads that mapped to the CHIKV genome was very low. This was likely due to the smaller size of the CHIKV genome, rather than poor quality read mapping.

108 Table 4.1 Summary of RNA-Seq reads mapped to UCSC mm10 mouse genome or CHIKV LR2006_OPY1 genome by Bowtie/Tophat.

Sample Total reads Reads aligned to Reads aligned to Unmapped reads (mean of 3 (SD) mm10 (% ± SD)a CHIKV (% ± SD)b (% ± SD) replicates) 55,349,451 46,620,761 4,572,386 2,654,892 Foot Day 2 (715,780) (84.231 ± 0.612) (8.260 ± 0.389) (4.796 ± 0.227) 52,331,910 47,039,395 4,094 2,790,083 LN Day 2 (2,497,885) (89.881 ± 0.293) (0.008 ± 0.003) (5.330 ± 0.439) 57,669,081 52,598,842 77,219 2,694,124 Foot Day 7 (1,359,085) (91.208 ± 0.078) (0.134 ± 0.015) (4.674 ± 0.160) 53,248,035 47,588,454 1,444 2,766,327 LN Day 7 (1,071,534) (89.373 ± 0.146) (0.003 ± 0.000) (5.196 ± 0.278) 51,011,141 46,188,873 2,510,208 Mock Foot n/a (1,550,329) (90.551 ± 0.612) (4.919 ±0.175) 54,767,209 49,093,156 2,745,703 Mock LN n/a (1,789,228) (89.638 ± 0.251) (5.018 ± 0.444) Data presented is the number of reads and their percentage of the Total reads for each Sample. a Both read mates mapped to the mm10 genome, b Neither read mate mapped to the mm10 genome.

100

n/a n/a of reads aligned ± SEM ± aligned reads of %

Figure 4.2 Graphical display of the percentage of total reads aligned to the mm10 or CHIKV genome. The mean percentage of high quality reads per sample that mapped either the CHIKV of mm10 genome, or did not map to either genome (Ft, Foot, LN: lymph node).

109 4.3.3 Coverage depth

The total number of reads and the percentage that map to the genome is a standard indicator of sequencing depth and quality for RNA-Seq experiments. However, in this experiment, reads were mapped to two genomes of different sizes. The Lander/Waterman general equation (coverage = read length x number of reads / genome length) (Lander and Waterman 1988) was thus used to estimate the coverage per lane of sequencing over the two different genomes.

The length of the mm10 mouse genome was obtained from the NCBI Assembly database and the length of the CHIKV genome was obtained from (Parola, de Lamballerie et al. 2006). The number of total reads aligned was extracted from the Tophat output files using SAMtools. Using the Lander/Waterman equation, each base in the mm10 genome and in the CHIKV genome was found to have been sequenced ~11 and ~4.6x106 times, respectively, per lane of sequencing (Table4.2) where 10x coverage is considered sufficient (Lander and Waterman 1988). Thus, while only a small percentage of total reads mapped the CHIKV genome, this was enough to provide a very high depth of data over the length of the CHIKV genome.

Table 4.2 Values used in coverage calculation by Lander/Waterman general equation.

Number of Read length Genome length Genome Coverage/ lane reads aligned (bp) (bp)

8 mm10 9.0 x 10 100 2.8 x 109 10.7

6 7 4 4.6 x 10 CHIKV 1.4 x 10 100 1.2 x 10

110 4.3.4 Coverage uniformity

To assess the uniformity of coverage, the average read density was plotted across the entire length of each genome. IGVtools ‘count’ command was used to calculate the average read density per 25 bp across the entire CHIKV and mouse genome using read mapping information contained in Bowtie/Tophat alignment output files.

Coverage was consistent across the mm10 genome, with no apparent anomalies amongst the different tissues and time points. Although two regions within 6 and 12 of the lymph nodes showed high peaks of read coverage on day 7 (but not in mock or day 2) (Figure 4.3A, red boxes), these were associated with antibody production (see Discussion). There was no indication of sequencing artefacts, such as PCR duplicates or incomplete coverage. In contrast, the transcriptional profile for CHIKV displayed a marked difference between each tissue and each time point (Figure 4.3 B). Given the quality of sequencing and read mapping (see sections 4.3.1 and 4.3.2) and the uniformity of sequence coverage in the mm10 genome, which was sequenced in the same run sample as CHIKV, these differences in coverage of the CHIKV genome are likely to reflect (i) the different levels of viral RNA in the different samples and (ii) known differences in genomic and subgenomic CHIKV RNA ratios, rather than any sequencing anomaly. The very high level of read coverage due to the smaller CHIKV genome compared to read coverage of the mouse genome can be also seen (note the difference in read depth between the two genomes (Figure 4.3 A & B; bottom right).

Together with sections 4.3.1 - 4.3.3, the results here demonstrate that (i) the total number of paired reads produced, (ii) the reads mapped and (iii) the depth of coverage per genome, were all well within the parameters outlined in the ENCODE Consortium Standards, Guidelines and Best Practices for RNA-Seq v1.0.

111

Figure 4.3 Transcription profile of host and virus gene expression at pre-infection, day and day 7 post CHIKV infection. The average read density per 25 bp interval was calculated using IGVtools count algorithm. A) Log2 average density of reads per 25 bp along the mm10 mouse genome. (Red circles highlight the peaks in read depth present on day 7 & but not day 2 or mock) B) Log2 average density of reads per 25 bp along the CHIKV (LR2006_OPY1) genome (Ft, Foot, LN: lymph node).

112 4.3.5 Global assessment of mouse gene and transcript regulation during CHIKV infection.

To assess the transcriptional activity of the host genome, Tophat alignment files were analysed using the Cufflinks suite. Cufflinks assembles aligned reads into transcripts and estimates their relative abundance in fragments per kilobase per million (FPKM). The FPKM of all transcripts associated with a gene can then be added to give the FPKM, or relative abundance, of that gene. Cuffdiff was used to generate a list of pair-wise comparisons of gene expression in all samples. The number of genes and transcripts in CHIKV infected feet and lymph nodes that were significantly different (q ≤ 0 .01) to mock infected controls, and the number of genes and transcripts significantly regulated by ≥ 2 fold was determined from these lists.

In all samples, the total number of down-regulated genes was higher than the total number of up-regulated genes (17.7% vs. 12%, 13.9% vs. 11.1% and 11.4% vs. 7.3% in day 7 feet, day 2 and day 7 lymph nodes, respectively), except day 2 feet, in which up- and down-regulation differed by < 1%. However, when considering genes with a ≥ 2 fold change, up-regulated genes outnumbered genes down-regulated in the feet at day 2 (4.6% vs. 1.8%) and day 7 (5.6% vs. 4.1%). Conversely, in the lymph nodes at day 2 post infection, genes down-regulated ≥ 2 fold out-numbered genes up-regulated ≥ 2 fold (3.8% vs. 2.3%) and at day 7, there was little difference (2.1% vs. 1.9%; Figure 4.4 A, bright red/green). Transcripts analysis showed up-regulation was more prevalent in day 2 feet (2.5% vs. 1.6%) with little difference other samples (2.9% vs. 3.2%, 2.9% vs. 2.2% and 1.4 vs. 1.5% in day 7 feet, day 2 and day 7 lymph nodes, respectively)(Figure 4.4B). Similarly, the majority of transcripts regulated ≥ 2 fold were up-regulated in day 2 feet (1.6% vs. 0.5%), with slightly more upregulation in the other samples (1.9% vs. 1.4%, 1.4% vs. 1%, 0.6% vs. 0.5% in day 7 feet, day 2 and day 7 lymph node, respectively (Figure 4.4 B bright red/green).

These results indicate a general trend toward gene down-regulation, perhaps due to virus-induced host transcriptional shut-down, cell death and/or dilution effects of infiltrating cells (Gardner, Anraku et al. 2010). A difference in the percentage of genes up- or down- regulated ≥ 2 fold (i.e. genes most likely to be of biological significance) was also observed between feet and lymph nodes. The trend toward up-regulation of transcripts reflects their relative contribution to genes up-regulated ≥ 2 fold.

113 114 Figure 4.4 Gene and transcript expression in the feet and LN of CHIKV-infected mice at day 2 and day 7 post infection. Each pie chart represents the total number of regulated genes in each sample. Differentially expressed genes, and individual gene transcripts were calculated using the Cufflinks suite. Significantly regulated (q-value ≤ 0.01) genes/transcripts are shown in green (down-regulated) or red (up-regulated). Genes/transcripts regulated by ≥2 fold are shown offset from the pie in bright green (down- regulated) and bright red (up-regulated). (Ft, Foot, LN: lymph node).

115 4.4 Discussion

The analyses presented herein demonstrates that RNA-Seq investigation of CHIKV- infected, whole mouse foot and lymph node tissues generated data of sufficient quality for assessment of both host and pathogen genome activity. Transcriptional activity across the mouse genome was largely uniform, while transcriptional activity across the CHIKV genome reflected the known viral RNA replication strategy, whereby the virus spreads from the site of infection to the draining lymph nodes and replicates in the joints (Schwartz and Albert 2010). Analysis of differential gene expression showed a high degree of down- regulation, while differentially expressed transcripts were predominantly up-regulated.

The data produced was shown to be of a standard suitable for analysis of known and novel transcriptional activity according to the ENCODE Consortium recommendations (ENCODE 2011). Over 400 million raw sequence reads of high quality were produced. A high percentage (≥ 85%) of reads in all samples mapped uniquely to the host genome, with fewer than 5% of reads generated unable to be mapped to either host or viral genome after computational subtraction. At day 2 post infection, read mapping to the host genome was lowest, while read mapping to the viral genome was the highest. In one of the few other studies of RNA viruses using high throughput sequencing methods, Marston, McElhinney et al. (2013) were able to map a maximum of 5.45% of reads isolated from brain tissue, to the genome of the highly neurovirulent rabies virus. The study used highly developed viral RNA isolation protocols and the 454 pyrosequencing platform. Herein, 8.3% and 0.13% of the total read sequences in the feet at day 2 and day 7, respectively, were seen to be derived from CHIKV (Table 4.1), indicating the utility of the RNA-Seq platform for analysing the sequence of tissue-associated viral RNA after the end of the viraemic period (Gardner, Anraku et al. 2010, Schwartz and Albert 2010).

Assessment of the mouse transcriptional profiles in each sample showed changes in genome scale transcriptional activity in the two tissues and time points. Two particularly apparent increases in the average abundance of transcripts could be seen in the lymph nodes at day 7 in chromosomes 12 (~113-116 Mb) and 6 (~68-71 Mb). These loci within the mouse genome encompass the immunoglobulin heavy chain (IgH) complex (D'Eustachio, Pravtcheva et al. 1980) and the immunoglobulin κ-light chain (Igk) complex 116 (Swan, D'eustachio et al. 1979), respectively (Pruitt, Tatusova et al. 2009). These spikes in transcriptional activity likely reflect the induction of antibody responses, known to begin at day 4 post infection in WT mice (Gardner, Anraku et al. 2010, Rudd, Wilson et al. 2012). Chromosome 16, which contains the λ-light chain genes (Pruitt, Tatusova et al. 2009), did not display a similar increase in transcript abundance. However, λ-light chains are expressed at a much lower level relative to κ-light chains in C57BL/6 mice (Woloschak and Krco 1987, Katzmann, Clark et al. 2002) due to 10-fold fewer λ-light chain genes, order of gene expression, and the inactivation of λ-light chain genes by κ-light chain expression (Takemori and Rajewsky 1981, Cherayil and Pillai 1991).

The transcriptional profile across the CHIKV genome, and between tissues and time points, was highly variable and reflected the replication kinetics of CHIKV. In all samples, sequence coverage of the structural protein ORF was higher than coverage of the non- structural protein ORF. During alphavirus replication, genomic RNA, for virion packaging, and subgenomic RNA, for production of structural proteins, are transcribed from an intermediate negative-strand RNA (Strauss and Strauss 1994). An increased ratio of subgenomic to genomic RNA, ranging from 1.3 – 16.8, is required to facilitate production of the 240 copies of E1/E2 envelope proteins that house a single genome copy per mature virion (Rümenapf, Strauss et al. 1994, Perri, Driver et al. 2000, Ng, Coppens et al. 2008).

Investigation of global gene expression revealed a high level of overall down- regulation, particularly in the foot at day 7. CHIKV infection is well known to induce host transcriptional and translational shutdown as means of delaying apoptosis and facilitating viral replication (Sudeep and Parashar 2008, Joubert, Werneke et al. 2012, Fros, Major et al. 2014). Down-regulation of host gene expression in feet at day 7 may also be an artefact of CHIKV-induced cell death, or the massive influx of inflammatory infiltrates, known to occur at day 6/7 post infection. Cell death, and/or an influx of cells at day 6/7 would effectively ‘dilute’ the RNA derived from foot tissues, causing genes expressed in mock infected feet to appear down-regulated at day 7. In support of this notion, a large number of down-regulated genes were seen to be down-regulated by < 2 fold. Despite the observed down regulation, the number of genes up-regulated by ≥ 2 fold outweighed genes down-regulated by ≥ 2 fold in the feet, and genes regulated by ≥ 2 fold in the lymph nodes were roughly equal in numbers. This indicated that those genes, with what are

117 likely to be the most biologically significant changes (herein referred to as differentially expressed genes; DEGs), were largely up-regulated.

Finally, a major benefit of RNA-Seq over other next generation sequencing technologies is the ability to estimate, not only the overall expression of a gene, but also the relative abundance of the individual transcripts derived from that gene (Ozsolak and Milos 2010, Trapnell, Roberts et al. 2012). Consistent with the DEGs, there were a greater number of up-regulated transcripts compared to down-regulated transcripts, in all samples. However, the total number of significantly regulated transcripts was lower than the number of significantly regulated genes in all samples. While more transcripts exist relative to genes, changes in individual transcript expression may not reach significance, but contribute in summation with other transcripts, to a significant change in expression of the whole gene. Transcripts regulated ≥ 2 fold are likely those genes/isoforms engaged in, or actively inhibited, during the response to CHIKV infection, and were used for later analysis of differential gene and pathway expression.

The results presented here demonstrated that the data generated by RNA-Seq of CHIKV infected tissues was of a standard of quality suitable for in-depth analysis of transcriptional activity. Global analysis of gene expression showed a high level of transcriptional activity in response to CHIKV, which was investigated in further detail in the Chapter 6.

118 5. Analysis of host gene expression in response to CHIKV infection using RNA-Seq

5.1 Introduction Microarray technology has been used to study gene expression in the mouse model of CHIKV (Nakaya, Gardner et al. 2012). However, such technologies allow investigation of known/characterised genes only, whereas RNA-Seq technology provides insight into the full spectrum of genome regulation during infection, including coding and non-coding gene expression, low-level gene expression, novel genes and individual isoform expression (Wang, Gerstein et al. 2009, Trapnell, Williams et al. 2010, McGettigan 2013).

To date, the only study of the mammalian response to an alphavirus using RNA- Seq, examined salmonoid alphavirus in fish cell lines (Xu, Evensen et al. 2015). Other studies of alphaviruses using RNA-Seq have examined infected invertebrate vectors (Morazzani, Wiley et al. 2012, Akbari, Antoshechkin et al. 2013, Poelchau, Huang et al. 2014, Shrinet, Jain et al. 2014), or viral transcription only (Chen, Kam et al. 2013, Donaszi- Ivanov, Mohorianu et al. 2013). Microarray by Nakaya, Gardner et al. (2012) identified a significant type I IFN response, with overlap between inflammatory processes in RA and acute CHIKV-arthritis. Atasheva, Akhrymuk et al. (2012) found that a different set of IFN regulated genes (IRGs) were expressed during VEEV replication versus viral clearance. Ryman, Meier et al. (2005) identified 44 genes induced during SINV infection, five of which (ISG20, ISG15, ISG56, viperin/rsad2 and ZAP) were confirmed to inhibit SINV translation in mice. Briolat, Jouneau et al. (2014) also observed a dominant type I IFN response in zebrafish larvae responses to CHIKV and this study also demonstrated tissue specific differences in IRG production.

Although all nucleated cells have the capacity to respond to or produce type I IFNs (MacMicking 2012), tissue differences in type I IFN and IRG expression exist. Pulverer, Rand et al. (2010), using a luciferase-Myxovirus resistance (Mx) reporter gene as an indicator of type I IFN responses, found the responses to IFNβ or Thogoto virus were most prominent in the liver, spleen and kidneys. Zebrafish models of CHIKV found the liver and spleen to be major producers of IRG mRNAs (Briolat, Jouneau et al. 2014) and the liver a major producer of type I IFN protein (Palha, Guivel-Benhassine et al. 2013). Results in Chapter 3 found that feet, lymph nodes, spleen and liver (ranked highest to lowest) produced the highest levels of type I IFN mRNA following CHIKV infection. 119

What drives these differences and what impact they have on the downstream gene expression is still little understood. Herein, RNA-Seq of CHIKV-infected mouse feet and lymph nodes at day 2 and day 7, as well as mock-infected tissues, was used to investigate the gene expression profiles in 2 tissues and at 2 time points following CHIKV infection.

120 5.2 Materials and Methods

5.2.1 High throughput RNA sequencing (RNA-Seq) data The RNA-Seq data analysed herein is that described in Chapter 4, and was generated as per sections 4.2.1-4.2.9. Briefly, RNA samples from feet and lymph nodes of CHIKV-infected mice at day 2 and day 7 and from mock-infected animals were sequenced on an Illumina HiSeq 2000 sequencer. The 100 bp paired-end read sequences produced were aligned to the UCSC mm10 Refseq build mouse genome the Tophat v2.0 read mapping software (Trapnell, Pachter et al. 2009). Tophat output files were analysed using the Cufflinks suite (Trapnell, Roberts et al. 2012). Cufflinks was used to estimate the relative abundance of individual gene transcripts in each sample. Relative abundance is measured in fragments per kilobase (of exon) per million fragments mapped (FPKM) and is based on the number of reads that align to a particular genomic region. Cuffmerge/Cuffdiff were used to run pair wise comparisons of FPKM values for each gene in each sample.

5.2.2 Preparation of DEGs lists To generate a list of differentially expressed genes (DEGs) in the different tissues for functional analysis, output files generated in Cuffdiff were sorted and filtered using Microsoft Excel. Each file was filtered to include only the genes with a statistically significant change in FPKM from mock values. FPKM values for which the q-value (p- value adjusted for false discovery rate) was≤ 0.01 were selected. Of these, transcripts with a fold change of≥ 2 difference from mock values were selected. FPKM < 1 is a heuristically determined threshold, commonly used to exclude genes with ultra-low, (presumably) biologically insignificant expression from analysis (Hart, Komori et al. 2013, Love, Huber et al. 2014). Thus, transcripts up-regulated to < 1 FPKM, or down-regulated from < 1 FPKM were excluded from DEG lists.

5.2.3 Unannotated DEGs Cufflinks assigns every expressed region of the genome an X_LOC (gene) and T_CONS (transcript) identifier, which was subsequently linked to, or annotated against the UCSC mm10 Refseq reference genome. Regions of the genome that were expressed but were not annotated (i.e. do not exist in the reference genome) appear as ‘-‘in the gene ID column of cufflinks output files, but can be tracked by their X_LOC/T_CONS identifiers. The genome co-ordinates of each unannotated gene in the top 50 up-regulated DEGs was entered into the NCBI Assembly database, and the corresponding nucleotide sequence

121 downloaded in FASTA format from the Genome Reference Consortium mouse build 38 (GRCm38) C57BL/6J mouse genome assembly. The BLAST algorithm was used to search for the unannotated gene mRNA in the NCBI nucleotide database. Where a match was found, the DEGs list was updated. The remaining unannotated, up-regulated DEGs were termed CHIKV induced novel transcripts (CINTs)

5.2.4 Investigation of the protein coding potential of CHIKV-induced novel transcripts CINTs in each top 50 DEGs list were investigated for their potential to encode a protein.

5.2.4.1 Construction of mRNA sequences The XLOC_ identifier of each CINT was used to find its corresponding transcripts (TCONS_) and exon coordinates in the merged.gtf annotation file. Exon sequences were extracted from the NCBI Assemblies database using the co-ordinates provided in Cufflinks. Downloaded exon sequences were concatenated to form the full mRNA sequence of the individual transcripts/isoforms for each unannotated transcript. The mRNA sequence of each gene isoform was subsequently identified using the nomenclature system chikungunya induced novel transcript number.variant number (eg. CINT1.1, CINT1.2 etc.).

5.2.4.2 Assessment of mRNA and amino acid sequences ExPasy Translate tool was used to translate the mRNA sequence of each CINT isoform into its amino acid sequence and identify the open reading frames (ORFs) present in all six possible frames. The translated sequences were assessed based on criteria set out by the NIH for identification of novel genes with protein coding potential. Sequences in which the longest open reading frame was i) < 30 aa or ii) not in the same strand direction reported by Cufflinks (i.e. 5’ 3’, or 3’5’) iii) or had a 5’ UTR < 20 aa or iv) contained many (>10) and/or long (> 30 aa) upstream ORFs (uORFs) or v) had upstream start site (uAUGs) within 9 bp of the longest ORF start site were excluded from further analysis. Remaining CINTs were then assessed for the presence and strength of the Kozak sequence in their mRNA sequences.

5.2.4.3 Assessment of protein homology and presence of conserved domains The mRNA sequences of CINTs that exhibited protein coding gene features were investigated for homology to other known proteins. The BLASTx algorithm was used to search for regions of similarity between known proteins in the NCBI protein database and amino acid sequences in each of the six possible reading frames that can be translated

122 from a nucleotide sequence. BLASTx was also used to identify the presence of conserved protein domains in translated CUNT isoforms. CINT sequences found to have 100% identity with a protein in the NCBI database were renamed accordingly.

5.2.5 Interferome v2.01 analysis The percentage of differentially expressed genes in each tissue and time point that were regulated by type I IFNs was determined using the Interferome v2.01 database of interferon regulated genes (Rusinova, Forster et al. 2013). Parameters were set to identify murine genes that are regulated ≥ 2 fold, either in vitro or in vivo, following treatment with type I IFNs.

5.2.6 Ingenuity Pathway Analysis Ingenuity Pathway Analysis (IPA) software (QIAGEN) was used to identify the canonical signalling pathways and upstream regulators employed in each tissue/time point based on the gene expression profile of each data set. Up- and down-regulated DEGs lists were uploaded into IPA and analysis was performed using the “direct and indirect relationships” setting. Upstream regulator analysis was used to predict the genes/molecules regulating gene expression seen in each sample. Upstream regulator output was ranked by p-value (significance of overlap with the IPA database) then Z-score (extent and direction of up- or down-regulation). Canonical pathways analysis was used to identify the pathways being engaged in each sample based on expression changes in each DEGs list. Canonical pathways output was filtered to exclude “disease specific pathways,” and ranked by p-value.

5.2.7 Sequence alignments and phylogenetic analysis MEGA v6.0 software (Tamura, Dudley et al. 2007) was used to perform multiple sequence alignments and phylogenetic analysis of amino acid sequences. Sequence alignments were rendered using consensus–colouring (CHROMA) (Goodstadt and Ponting 2001).

123 5.3 Results

5.3.1 Differentially expressed genes (DEGs)

A global assessment of DEGs was presented in Chapter 4 (section 4.3.5), which showed a greater number of genes were down-regulated in tissues post CHIKV infection than up-regulated. Herein, DEGs lists were used to assess the biological processes engaged following CHIKV infection. The 50 most up- and down- regulated genes are shown in Tables 1-8. Down-regulated, unannotated DEGs are indicated by “-” in the gene symbol list. Up-regulated, unannotated DEGs were termed CHIKV induced novel transcripts (CINTs) and were further investigated in section 5.3.9.

There was a 30% overlap in DEGs between day 2 and 7 in the feet, a 32% overlap of DEGs at day 2 in feet and lymph node, and 20% of DEGs in the top 50 most up- regulated were common to day 2 feet, day7 feet and day2 lymph nodes. The DEG lists included several known anti-CHIKV effectors such as Mx proteins (Schoggins, Wilson et al. 2011), OAS genes (Bréhin, Casadémont et al. 2009), rsad2/viperin (Teng, Foo et al. 2012), ISG15 (Werneke, Schilte et al. 2011) and ISG20 (Schoggins, Wilson et al. 2011, Karki, Li et al. 2012), as well as the three IFIT genes (Hyde, Gardner et al. 2014). Granzymes A and B were heavily induced, particularly in day 7 feet (≈119 and ≈630 fold increase, respectively), as were several other inflammatory cell markers and chemoattractants. A broad array of GTPases, in particular the p47 IFN-inducible GTPases and the p65 guanylate binding proteins (GBPs) were also up-regulated (see discussion). Type I IFNs were induced at day 2, and largely absent on day 7. Interestingly, IRF7 remained strongly up-regulated in day 7 feet, despite the lack of type I IFN mRNAs (see section 5.3.5).

DEGs in day 7 lymph node differed from the other samples. 20.7% of all up- regulated genes, and 60% of the top 50 most up-regulated genes were immunoglobulin heavy- and light- chain variable region genes. To consider non-immunoglobulin mediated process, unnamed immunoglobulin genes were removed from DEGs lists; a singular entry was retained for immunoglobulin heavy (IgH-V) and light (Igκ-V) chains variable regions (Table 5.4). Day 7 lymph node also revealed a high proportion of CINTs in the top 50

124 (18% versus 10%, 6% and 6% in day 2 lymph node, day 2 and day 7 feet, respectively) (see 5.3.9).

Generally, the top 50 down-regulated genes had lower fold change than up- regulated genes in the top 50. 16% of down-regulated DEGs were common to all samples. All CD209 isoforms and their homologue CD207 were seen to be down- regulated in all tissues and time points, as were other lectins/carbohydrate-binding molecules. Keratins, keratin-associated genes, epithelial development genes and genes associated with striated muscle function and energy metabolism were also down-regulated in the feet. This may, in part, reflect cell death in infected cells including epithelial cells and muscle cells (Nakaya, Gardner et al. 2012, Rudd, Wilson et al. 2012). A number of M2 (anti-inflammatory) markers (Retnla, F13a1 Mrc1, CD163) were also down-regulated in each sample. The most heavily down-regulated gene in lymph nodes at both time points was the microRNA, mirlet7h (see Discussion).

Thus, gene regulation during CHIKV infection shows both tissue and time specific difference in regulation.

125 Table 5.1 Up-regulated DEGs in foot tissues at day 2 post CHIKV infection.

Gene Mock Day 2 Log2 Official Gene Name Symbol FPKM FPKM (FC) Ifnb1 Interferon beta 1 0.00 6.79 21.00 CINT6 Chikungunya induced novel transcript 6 0.00 4.65 21.00 Ifna4 Interferon alpha 4 0.00 4.40 21.00 Gm15056* Beta defensin 52 0.00 3.18 21.00 Ifna5 Interferon alpha 5 0.00 3.04 21.00 Gcg Glucagon 0.00 2.00 21.00 Ifna6 Interferon alpha 6 0.00 1.33 21.00 Ifna2 Interferon alpha 2 0.00 1.15 21.00 CINT5 Chikungunya induced novel transcript 5 0.00 1.13 21.00 Cxcl10 Chemokine (C-X-C motif) ligand 10 1.69 1473.04 9.76 Gm4841*† Interferon gamma-inducible gtpase alpha 3 0.44 295.51 9.42 Gm14446* Interferon-induced protein with tetratricopeptide repeats 1c 0.29 151.64 9.04 Cxcl9 Chemokine (C-X-C motif) ligand 9 0.56 220.40 8.61 Mx1 Myxovirus (influenza virus) resistance 1 1.32 375.53 8.16 Gm12185* Interferon gamma-inducible gtpase beta 7 0.08 20.30 8.07 CINT7 Chikungunya induced novel transcript 7 0.03 8.48 8.01 Iigp1† Interferon inducible gtpase 1 0.54 132.28 7.97 Isg15 Interferon stimulated gene 15 (ubiquitin-like modifier) 5.34 1211.61 7.83 Rsad2 Radical s-adenosyl methionine domain containing 2 (viperin) 1.52 337.12 7.79 Ifi44 Interferon-induced protein 44 0.51 110.50 7.75 Ifit1 Interferon-induced protein with tetratricopeptide repeats 1/ISG56 2.24 457.72 7.68 Oasl1 2',5'-oligoadenylate synthetase-like 1 0.88 134.28 7.25 Gm5431*† Interferon gamma-inducible gtpase beta 8 0.02 2.40 7.06 Gm12250*† Interferon regulated gtpase 10 0.33 43.34 7.03 Ifit3 Interferon-induced protein with tetratricopeptide repeats 3/ISG60 8.59 1075.31 6.97 Cxcl11 Chemokine (C-X-C motif) ligand 11 0.12 14.46 6.87 Irf7 Interferon regulatory factor 7 8.43 979.32 6.86 Mx2 Myxovirus (influenza virus) resistance 2 0.86 95.59 6.80 Prm1 Protamine 1 0.02 2.19 6.48 Pydc4 Pyrin domain containing 4 0.32 28.15 6.48 Irg1 Immunoresponsive 1 homolog 0.36 31.89 6.47 Slfn4 Schlafen 4 2.42 212.88 6.46 Ttll9 Tubulin tyrosine ligase-like family, member 9 0.01 1.20 6.36 Oas1g 2'-5' oligoadenylate synthetase 1g 0.24 18.96 6.33 Gbp5 Guanylate binding protein 5 1.60 126.27 6.30 Tgtp1† T-cell specific gtpase 1 1.62 123.67 6.25 Ifit2 Interferon-induced protein with tetratricopeptide repeats 2 3.66 276.80 6.24 Apol9b Apolipoprotein L 9b 0.79 54.05 6.09 Gzmb Granzyme B 0.16 11.06 6.08 Ms4a4c Membrane-spanning 4-domains, subfamily A, member 4C 1.67 110.11 6.04 Pydc3 Pyrin domain containing 3 0.45 29.69 6.04 Igtp† Interferon gamma induced gtpase 7.95 510.75 6.00 Gbp2 Guanylate binding protein 2, interferon-inducible 9.01 572.19 5.99 Ccl5 Chemokine (C-C motif) ligand 5 5.12 321.80 5.97 Usp18 Ubiquitinspecific peptidase 18 3.25 201.24 5.95 Batf2 Basic leucine zipper transcription factor, atf-like 2 0.41 25.14 5.95 Apol9a Apolipoprotein L 9a 0.86 51.00 5.90 Cmpk2 Cytidine monophosphate (UMP-CMP) kinase 2, mitochondrial 2.16 125.75 5.86 Phf11c PHD finger protein 11C 5.28 299.45 5.83 Ccl4 Chemokine (C-C motif) ligand 4 0.99 55.89 5.82

* Gene name/symbol assigned after manual search of gene coordinates; bold: Interferome designated IRG; italics: literature validated IRG; † p47 IFN inducible GTPase;

126

Table 5.2 Up-regulated DEG in foot tissues at day 7 post CHIKV infection.

Gene Mock Day 7 Log2 Official Gene Name Symbol FPKM FPKM (FC) CINT1 CHIKV-induced novel transcript 1 0.000 1.044 21.000 CINT4 CHIKV-induced novel transcript 4 0.000 1.045 21.000 Gm15056* Beta defensin 52 0.000 11.584 21.000 Tcrb* receptor beta chain 0.000 1.080 21.000 Gzmb granzymeB 0.164 102.165 9.284 Gm4841† Interferon gamma-inducible GTPase alpha 3 0.443 143.131 8.335 Cxcl9 chemokine (C-X-C motif) ligand 9 0.563 181.529 8.333 Gbp1 guanylate binding protein 1, interferon-inducible 0.082 24.961 8.248 Fcgr4 Fc receptor, IgG, low affinity IV 0.801 174.742 7.770 Cd300e CD300e molecule 0.071 14.610 7.688 Cxcl10 chemokine (C-X-C motif) ligand 10 1.695 313.526 7.531 Gm12185*† interferon gamma inducible GTPase beta 7 0.075 13.855 7.521 Gm12250*† interferon-gamma-inducible p47 GTPase 0.331 42.716 7.011 Gzma granzymeA 0.612 76.408 6.963 Iigp1† Interferon inducible GTPase 1 0.542 59.376 6.775 Irg1 immunoresponsive 1 homolog (mouse) 0.360 33.923 6.559 Tgtp1† T cell specific GTPase 1 1.620 142.555 6.460 Gm5431*† interferon gamma inducible GTPase beta 8 0.018 1.500 6.385 Slamf8 SLAM family member 8 0.366 29.947 6.354 Batf2 basic leucine zipper transcription factor, ATF-like 2 0.406 31.637 6.283 Klra2 killer cell lectin-like receptor, subfamily A, member 2 0.353 25.736 6.186 Ms4a4c membrane-spanning 4-domains, subfamily A, member 4C 1.673 112.111 6.067 Tigit T cell immunoreceptor with Ig and ITIM domains 0.094 5.670 5.919 Fam26f family with sequence similarity 26, member F 1.104 65.832 5.898 Zbp1 Z-DNA binding protein 1 4.313 254.285 5.882 Pydc4 Pyrin domain containing 4 0.316 18.522 5.872 CINT2 CHIKV-induced noel transcript 2 0.079 4.554 5.841 Gbp2 guanylate binding protein 2, interferon-inducible 9.010 482.328 5.742 Gm5150* SIRP beta 1 like 1 protein 0.273 14.590 5.741 Gbp8 guanylate binding protein 8, interferon-inducible 0.690 36.166 5.713 Gm14446* Interferon-induced protein with tetratricopeptide repeats 1c 0.288 14.614 5.663 Fcrl5 Fc receptor-like 5 0.033 1.675 5.658 Cd8b1 CD8b molecule 0.161 8.140 5.658 Prm1 protamine 1 0.024 1.227 5.647 Il12rb1 interleukin 12 receptor, beta 1 0.162 7.890 5.605 Slfn1 Schlafen 1 1.217 57.859 5.572 Ubd ubiquitin D 0.495 23.193 5.549 Ly6c2 antigen 6 complex, locus C2 10.216 465.144 5.509 Igtp† interferon gamma induced GTPase 7.955 361.793 5.507 Ifi44 interferon-induced protein 44 0.513 22.977 5.486 Isg15 Interferon Stimulated Gene 15 ubiquitin-like modifier 5.341 238.298 5.479 Gbp5 guanylate binding protein 5 1.597 70.955 5.473 Klra3 killer cell lectin-like receptor, subfamily A, member 2 0.078 3.348 5.418 Ccl5 chemokine (C-C motif) ligand 5 5.123 211.614 5.368 Ms4a4b membrane-spanning 4-domains, subfamily A, member 4b 1.153 45.347 5.298 Gm4951*† Interferon gamma inducible GTPase 2 0.450 17.659 5.294 Irf7 interferon regulatory factor 7 8.430 327.235 5.279 Nos2 nitric oxide synthase 2, inducible 0.220 8.337 5.241 Ccl4 chemokine (C-C motif) ligand 4 0.987 36.929 5.225 Oasl1 2',5'-oligoadenylate synthetase-like 1 0.883 32.760 5.214

* Gene name/symbol assigned after manual search of gene coordinates; bold: Interferome designated IRG; italics: literature validated IRG; † p47 IFN inducible GTPase;

127 Table 5.3 Up-regulated DEGs in lymph nodes at day 2 post CHIKV infection.

Gene Mock Day 2 Log2 Official Gene Name Symbol FPKM FPKM (FC) CINT8 Chikv-upregulated novel transcript 8 0.000 1.774 21.000 Ifna5 Interferon, alpha 5 0.000 1.136 21.000 Ifnb1 Interferon, beta 1, fibroblast 0.000 1.121 21.000 CINT9 Chikv-upregulated novel transcript 9 0.116 24.319 7.712 Cxcl11 Chemokine (C-X-C motif) ligand 11 0.122 15.319 6.968 Sct Secretin 0.642 70.782 6.785 Slfn4 Schlafen 4 1.982 116.073 5.872 Irg1 Immunoresponsive 1 homolog (mouse) 0.150 7.693 5.677 CINT10 Chikv-upregulated novel transcript 10 0.037 1.744 5.571 Gbp10 Guanylate-binding protein 10 1.252 58.399 5.409 Isg15 Interferon stimulated gene 15 ubiquitin-like modifier 25.186 798.073 4.986 Gzmb Granzyme B 2.967 87.551 4.883 Ifit3 Interferon-induced protein with tetratricopeptide repeats 3/ISG60 22.617 646.874 4.838 Rsad2 Radical s-adenosyl methionine domain containing 2 (viperin) 3.952 112.745 4.834 Cdkl4 Cyclin-dependent kinase-like 4 0.123 3.356 4.775 Nts Neurotensin 0.658 16.680 4.663 CINT11 Chikv-upregulated novel transcript 11 1.581 38.183 4.594 Fgf23 Fibroblast growth factor 23 0.082 1.941 4.566 Tuba8 Tubulin, alpha 8 0.095 2.112 4.475 Ttll9 Tubulin tyrosine ligase-like family, member 9 0.166 3.670 4.469 Ifit2 Interferon-induced protein with tetratricopeptide repeats 2/ISG54 10.000 221.063 4.466 Oasl2 2'-5' oligoadenylate synthetase-like 2 13.828 304.946 4.463 Ifit1 Interferon-induced protein with tetratricopeptide repeats 1/ISG56 9.259 201.926 4.447 Tmem171 Transmembrane protein 171 0.170 3.636 4.417 Gm5431*† Interferon gamma inducible gtpase beta 8 0.864 17.969 4.379 Irf7 Interferon regulatory factor 7 49.905 1003.810 4.330 Nmes1* Normal mucosa of esophagus-specific gene 1 protein 0.695 13.928 4.325 Oas2 2'-5' oligoadenylate synthetase 1a 18.265 363.633 4.315 Oasl1 2'-5' oligoadenylate synthetase-like 1 8.188 160.097 4.289 Mx2 Myxovirus (influenza virus) resistance 2 (mouse) 3.847 74.201 4.270 Il1rn Interleukin 1 receptor antagonist 1.968 36.516 4.214 Ifi44 interferon-induced protein 44 4.532 79.906 4.140 Gm12185*† interferon gamma inducible GTPase beta 7 1.275 21.753 4.093 Oas1g 2'-5' oligoadenylate synthetase 1g 0.851 14.413 4.081 Ptgs2 prostaglandin-endoperoxide synthase 2 0.134 2.187 4.032 Gpr31 G protein-coupled receptor 31 0.134 2.179 4.026 Cxcl10 chemokine (C-X-C motif) ligand 10 20.777 331.740 3.997 Gm14446*† Interferon-induced protein with tetratricopeptide repeats 1c 15.998 246.750 3.947 Prm1 protamine 1 0.148 2.271 3.939 Mx1 myxovirus (influenza virus) resistance 1 16.661 252.306 3.921 Isg20 interferon stimulated exonuclease gene 20kDa 13.592 199.121 3.873 Zbp1 Z-DNA binding protein 1 36.520 529.494 3.858 Il6 interleukin 6 (interferon, beta 2) 0.113 1.626 3.845 Usp18 ubiquitin specific peptidase 18 20.441 289.266 3.823 Klra2 killer cell lectin-like receptor, subfamily A, member 2 1.008 13.857 3.780 Cmpk2 cytidine monophosphate (UMP-CMP) kinase 2, mitochondrial 6.825 92.285 3.757 Oas1a 2'-5' oligoadenylate synthetase 1a 14.164 190.364 3.748 Pira6 paired-Ig-like receptor A6 0.955 12.825 3.747 CINT22 CHIKV-upregulated novel transcript 22 0.372 4.988 3.744 Perm1 PGC-1 and ERR-induced regulator in muscle 1 4.825 60.997 3.660

* Gene name/symbol assigned after manual search of gene coordinates; bold: Interferome designated IRG; italics: literature validated IRG; † p47 IFN inducible GTPase;

128 Table 5.4 Up-regulated DEGs in lymph nodes at day 7 post CHIKV infection.

Gene Mock Day 7 Log2 Official Gene Name Symbol FPKM FPKM (FC) CINT12 Chikv-induced novel transcript 12 0.000 1.777 21.000 Igh-V* Immunoglobulin heavy chain variable region 42.904 3716.430 6.437 Gm5535 - 0.115 8.582 6.225 Igk-V* Immunoglobulin light chain variable region 9.687 610.436 5.978 Gm6455* EG623849 pseudogene 0.126 4.904 5.279 Gnat3 Guanine nucleotide binding protein, alpha transducing 3 0.289 9.229 4.996 Slpi Secretory leukocyte peptidase inhibitor 15.915 461.330 4.857 Tmed6 Transmembrane emp24 protein transport domain containing 6 0.130 3.699 4.830 Eaf2 ELL associated factor 2 2.086 46.594 4.482 CINT13 Chikv-induced novel transcript 13 0.066 1.400 4.396 Ly6B* Model protein coding gene containing LU domain 1.091 19.685 4.173 Derl3 Derlin 3 4.291 76.115 4.149 Aicda Activation-induced cytidine deaminase 0.784 13.470 4.103 CINT15 Chikv-induced novel transcript 12 0.215 3.483 4.015 Igj Immunoglobulin J polypeptide 115.486 1793.620 3.957 Bhlha15 Basic helix-loop-helix family, member a15 0.940 14.465 3.943 Oosp1 Oocyte secreted protein 1 0.762 10.706 3.812 Nuggc Nuclear gtpase, germinal center associated 0.702 9.831 3.809 Arhgef10 Rho guanine nucleotide exchange factor (GEF) 10 0.145 2.009 3.792 Il21 Interleukin 21 0.137 1.802 3.722 CINT16 Chikv-induced novel transcript 16 0.137 1.686 3.622 CINT17 Chikv-induced novel transcript 17 0.121 1.430 3.558 Mzb1 Marginal zone B and B1 cell-specific protein 56.721 613.061 3.434 Cacna1h Calcium channel, voltage-dependent, T type, alpha 1H subunit 0.667 6.977 3.388 CINT18 Chikv-induced novel transcript 12 0.170 1.733 3.348 CINT19 Chikv-induced novel transcript 12 0.102 1.023 3.329 Gzmk Granzyme K (granzyme 3; tryptase II) 0.443 4.352 3.297 Fkbp11 FK506 binding protein 11, 19 kda 5.103 45.429 3.154 CINT20 Chikv-induced novel transcript 20 0.178 1.351 2.920 5830418P13Rik Non-coding RNA 0.522 3.923 2.910 Mybl1 V-myb avian myeloblastosis viral oncogene homolog-like 1 1.153 8.106 2.814 Ankrd55 Ankyrin repeat domain 55 0.699 4.646 2.733 Cryba4 Crystallin, beta A4 0.912 5.410 2.568 CINT21 Chikv-induced novel transcript 21 0.184 1.071 2.539 Endou Endonuclease, polyu-specific 1.593 9.174 2.525 Tnfrsf17 Tumor necrosis factor receptor superfamily, member 17 2.025 11.607 2.519 Rrm2 Ribonucleotide reductase M2 12.413 66.768 2.427 Creld2 Cysteine-rich with egf-like domains 2 22.299 118.987 2.416 Fam46c Family with sequence similarity 46, member C 8.812 46.654 2.404 Mir148a Micro rna 148a 0.299 1.563 2.388 Pif1 PIF1 5'-to-3' DNA helicase 0.984 5.016 2.350 Cdc20 Cell division cycle 20 7.294 37.152 2.349 Ccnb1 CyclinB1 5.214 26.387 2.339 Nek2 Nima-related kinase 2 2.322 11.715 2.335 Cdc25c Cell division cycle 25C 1.014 5.071 2.322 Gzmb Granzyme B 2.967 14.613 2.300 Cdca3 Cell division cycle associated 3 7.447 36.513 2.294 Sapcd2 Suppressor APC domain containing 2 0.851 4.148 2.285 Ly6c2 Lymphocyte antigen 6 complex, locus C2 121.659 591.624 2.282 Plk1 Polo-like kinase 1 5.364 25.822 2.267

Genes in this list are the top 50 most upregulated genes after immunoglobulin variable region genes (Ig-V) have been removed. Ig-V genes represented by a single entry. *Gene name/symbol assigned after manual search of gene coordinates. Bold; Interferome designated IRG; Italics: literature validated IRG; † p47 IFN-inducible GTPase.

129 Table 5.5 Down-regulated DEGs in foot tissues at day 2 post CHIKV infection.

Gene Mock Day 7 Log2 Official Gene Name Symbol FPKM FPKM (FC) - - 1.03 0.00 -21.00 Cd209g Cluster of differentiation 209g 7.97 0.57 -3.81 - - 2.93 0.30 -3.27 Cd209f Cluster of differentiation 209f 17.71 1.86 -3.26 Mrgprg Mas-related GPR, member G 2.74 0.32 -3.09 Vsig4 V-set and immunoglobulin domain containing 4 17.80 2.37 -2.91 Inmt Indolethylamine n-methyltransferase 51.52 7.31 -2.82 Gkn3 Gastrokine 3 5.66 0.83 -2.77 Vstm2b V-set and transmembrane domain containing 2B 2.41 0.36 -2.75 Pck1 Phosphoenolpyruvate carboxykinase 1 (soluble) 7.70 1.15 -2.74 Gpr34 G protein-coupled receptor 34 12.82 1.97 -2.71 Cd163 CD163 molecule 20.57 3.35 -2.62 Retnla Resistinlike alpha 145.84 25.88 -2.49 Fcrls Fc receptor-like S, scavenger receptor 3.47 0.62 -2.49 Tmem179 Transmembrane protein 179 1.67 0.30 -2.48 Abca9 Atp-binding cassette, sub-family A (ABC1), member 9 19.02 3.53 -2.43 Ccl24 Chemokine (C-C motif) ligand 24 15.23 2.84 -2.43 Abca6 Atp-binding cassette, sub-family A (ABC1), member 6 2.97 0.59 -2.33 Cd209d Cluster of differentiation 209 28.45 5.73 -2.31 Hepacam2 HEPACAM family member 2 1.61 0.33 -2.28 2610016A17Rik Ncrna 1.93 0.41 -2.24 Islr2 Immunoglobulin superfamily containing leucine-rich repeat 2 3.63 0.79 -2.21 Cd209b Cluster of differentiation 209b 6.29 1.41 -2.16 Pkhd1l1 Polycystic kidney & hepatic disease 1 (autosomal recessive)-like 1 1.70 0.39 -2.14 Aldh1a3 Aldehyde dehydrogenase 1 family, member A3 10.18 2.34 -2.12 Krt6b Keratin 6B 19.57 4.55 -2.10 Cd209a Cluster of differentiation 209a 7.37 1.73 -2.09 Wnt2 Wingless-type MMTV integration site family member 2 3.54 0.84 -2.08 Mrc1 Mannose receptor, C type 1 45.39 10.75 -2.08 Nnat Neuronatin 29.51 7.08 -2.06 Nrac* Nutritionally-regulated adipose and cardiac-enriched protein 2.24 0.55 -2.04 Siglech Sialic acid binding ig-like lectinH 1.62 0.40 -2.02 Cd209c Cluster of differentiation 209c 4.83 1.20 -2.01 Slc6a7 Solute carrier family 6 (neurotransmitter transporter), member 7 3.19 0.80 -2.00 Cd33 CD33 molecule 6.05 1.52 -2.00 Gpr1 G protein-coupled receptor 1 7.08 1.79 -1.98 LOC102633590 Non-coding RNA 3.69 0.93 -1.98 Gsta1 Glutathione s-transferase alpha 1 1.17 0.30 -1.98 Tnfrsf17 Tumor necrosis factor receptor superfamily, member 17 1.80 0.46 -1.97 Gsc Goosecoid homeobox 1.41 0.36 -1.96 Fcna FicolinA 9.39 2.42 -1.96 Gdf10 Growth differentiation factor 10 40.12 10.34 -1.96 Cyp2e1 Cytochrome P450, family 2, subfamily E, polypeptide 1 24.42 6.34 -1.95 Stmn4 Stathmin-like 4 1.18 0.31 -1.93 Cx3cr1 Chemokine (C-X3-C motif) receptor 1 2.49 0.65 -1.93 Megf10 Multiple egf-like-domains 10 1.89 0.50 -1.91 F13a1 Coagulation factor XIII, A1 polypeptide 88.71 24.10 -1.88 Ednrb Endothelinreceptor type B 7.25 1.98 -1.87 Odf3l2 Outer dense fiber of sperm tails 3-like 2 9.19 2.51 -1.87 - - 1.31 0.36 -1.86

130 Table 5.6 Down-regulated DEGs in foot tissues at day 7 post CHIKV infection.

Gene Official gene name Mock Day 7 Log 2 Symbol FPKM FPKM (FC)

Plac9b Placenta specific 9b 5.29 0.00 -21 - - 1.27 0.00 -21 Egfbp2 Epidermal growth factor binding protein type b 6.75 0.63 -3.42 Krt6b Keratin 6B 19.57 1.88 -3.38 Ep1 Epithelial progenitor 1 2.91 0.31 -3.24 Smco1 Single-pass membrane protein with coiled-coil domains 1 6.22 0.73 -3.08 Mir3965 Micro rna 3965 1.04 0.12 -3.07 Gm13490 Predicted gene 13490 1.39 0.18 -2.95 - - 1.31 0.17 -2.95 Prss55 Protease, serine, 55 4.09 0.55 -2.91 - - 1.81 0.25 -2.86 Ccl24 Chemokine (c-c motif) ligand 24 15.23 2.13 -2.84 - - 1.09 0.16 -2.76 Gm4861 Predicted gene 4861 2.29 0.34 -2.74 - - 3.92 0.64 -2.61 Mylk4 Myosin light chain kinase family, member 4 11.85 2.03 -2.54 - - 1.15 0.20 -2.53 - - 1.89 0.33 -2.51 Retnla Resistinlike alpha 145.84 25.75 -2.50 Agbl1 ATP/GTP binding protein-like 1 2.64 0.47 -2.48 - - 1.31 0.23 -2.48 2310065F04Rik - 3.35 0.61 -2.46 Mss51 MSS51 mitochondrial translational activator 73.10 13.68 -2.42 Gdap1 Ganglioside induced differentiation associated protein 1 1.09 0.21 -2.39 Gm4861 Predicted gene 1.11 0.21 -2.39 Cox7a1 Cytochrome c oxidase subunit viia polypeptide 1 (muscle) 163.38 31.22 -2.39 Ckmt2 Creatine kinase, mitochondrial 2 (sarcomeric) 46.40 8.87 -2.39 Cd209b Cluster of differentiation 209a 6.29 1.20 -2.39 - - 6.89 1.32 -2.38 D830013O20Rik - 1.40 0.28 -2.32 Crnn Cornulin 1.84 0.37 -2.32 Mucl1 Mucin-like 1 7.75 1.56 -2.31 Krtap15 Keratin associated protein 15 8.82 1.78 -2.31 Dupd1 Dual specificity phosphatase & pro isomerase domain containing 1 1.48 0.30 -2.29 - - 1.12 0.23 -2.29 Mrgprg Mas-related gpr, member g 2.74 0.56 -2.28 Dhrs7c Dehydrogenase/reductase (SDR family) member 7C 17.23 3.56 -2.27 - - 2.93 0.61 -2.27 Vsig4 V-set and immunoglobulin domain containing 4 17.80 3.71 -2.26 Mpp3 Membrane protein, palmitoylated 3 (MAGUK p55 subfamily 3) 4.96 1.06 -2.23 Col9a1 Collagen, type IX, alpha 1 3.32 0.71 -2.22 Myoz3 Myozenin3 11.74 2.56 -2.20 2310050B05Rik Ncrna 27.83 6.15 -2.18 Defb14 Defensin beta 14 9.74 2.18 -2.16 Sprr2h Small proline-rich protein 2H 5.03 1.13 -2.16 Krt25 Keratin 25 31.30 7.09 -2.14 Hhatl Hedgehog acyltransferase-like 18.52 4.22 -2.13 Myoz1 Myozenin1 173.32 39.55 -2.13 Krt27 Keratin 27 14.60 3.33 -2.13 Phkg1 Phosphorylase kinase, gamma 1 (muscle) 21.28 4.91 -2.12

131 Table 5.7 Down-regulated DEGs in lymph nodes at day 2 post CHIKV infection.

Mock Day 7 Log2 Gene Symbol Official gene name FPKM FPKM (FC) Mirlet7h MicroRNAlet7h 229.71 3.45 -6.06 Bpifb6 BPI fold containing family B, member 6 1.01 0.03 -5.03 Bpifb3 BPI fold containing family B, member 3 11.84 0.46 -4.70 Bpifb2 BPI fold containing family B, member 2 7.05 0.29 -4.59 Cd207 CD207 molecule, langerin 2.82 0.21 -3.75 Cd209e Cluster of differentiation 209e 5.49 0.41 -3.74 Igh-V* Immunoglobulin heavy chain variable region 2.34 0.23 -3.34 Clec4g C-type lectin domain family 4, member G 3.27 0.33 -3.32 Cd209d Cluster of differentiation 209d 52.93 6.07 -3.12 Rasd2 RASD family, member 2 2.51 0.29 -3.12 Syndig1 Synapse differentiation inducing 1 3.65 0.43 -3.08 Cd209a Synapse differentiation inducing 1 10.19 1.21 -3.08 - 1.79 0.23 -2.98 Rgr Retinal g protein coupled receptor 1.43 0.20 -2.83 Necab2 Ral guanine nucleotide dissociation stimulator-like 4 1.01 0.15 -2.79 Calca N-terminal ef-hand calcium binding protein 2 3.21 0.49 -2.71 Cd209f Calcitonin-related polypeptide alpha 13.21 2.21 -2.58 Il22ra2 Interleukin 22 receptor, alpha 2 31.05 5.57 -2.48 1300017J02Rik* Inhibitorof carbonic anhydrase 1.74 0.32 -2.43 Cbr2 Carbonyl reductase 2 34.73 6.51 -2.42 Wasf1 WAS protein family, member 1 2.68 0.51 -2.41 Cd209b Cluster of differentiation 209b 214.60 40.46 -2.41 Cd209g Cluster of differentiation 209g 4.89 0.92 -2.40 Ccl24 Chemokine (C-C motif) ligand 24 2.19 0.41 -2.40 Mrc1 Mannose receptor, C type 1 18.04 3.43 -2.40 F8 Coagulation factor VIII, procoagulant component 11.58 2.26 -2.36 Abca9 Atp-binding cassette, sub-family A (ABC1), member 9 8.62 1.69 -2.35 Sept3 Septin 3 2.58 0.52 -2.30 Gcsam Germinal center-associated, signaling and motility 3.48 0.71 -2.29 Epor Erythropoietin receptor 3.50 0.72 -2.28 Inmt Indolethylamine n-methyltransferase 46.71 9.77 -2.26 Nxph3 Neurexophilin 3 1.29 0.27 -2.24 Adra1b Adrenoceptor alpha 1B 1.22 0.26 -2.23 - 4.16 0.89 -2.23 Ces2e Carboxylesterase 2e 3.52 0.78 -2.18 Il13ra2 Interleukin 13 receptor, alpha 2 3.94 0.88 -2.17 Marco Macrophage receptor with collagenous structure 131.67 29.43 -2.16 Hpgd Hydroxyprostaglandin dehydrogenase 15-(NAD) 20.12 4.52 -2.15 Stab2 Stabilin 2 11.84 2.67 -2.15 Slc40a1 Solute carrier family 40 (iron-regulated transporter), member 1 7.76 1.76 -2.14 Ak8 Adenylate kinase 8 2.46 0.56 -2.13 Crisp3 Cysteine-rich secretory protein 3 6.39 1.46 -2.13 Gchfr GTP cyclohydrolase I feedback regulator 2.94 0.68 -2.11 Chp2 Calcineurin-like ef-hand protein 2 8.72 2.04 -2.10 Srpx2 Sushi-repeat containing protein, x-linked 2 5.13 1.20 -2.09 Gria3 Glutamate receptor, ionotropic, AMPA 3 4.38 1.03 -2.09 Fam65c Family with sequence similarity 65, member C 7.21 1.70 -2.08 Mogat1 Monoacylglycerol o-acyltransferase 1 2.98 0.71 -2.08 Slc16a12 Solute carrier family 16, member 12 1.34 0.32 -2.07 A530099J19Rik* G protein coupled receptor like 5.94 1.43 -2.05

132 Table 5.8 Down-regulated DEGs in lymph nodes at day 7 post CHIKV infection.

Mock Day 7 Log2 Gene Symbol Official gene name FPKM FPKM (FC) Mirlet7h MicroRNA let 7h 229.71 2.52 -6.51 Calca Calcitonin-related polypeptide alpha 3.21 0.19 -4.08 Crabp1 Cellular retinoic acid binding protein 1 1.66 0.25 -2.73 Gm2083* major urinary protein LOC100048885 2.96 0.54 -2.46 Krt15 Keratin 15 3.20 0.58 -2.45 Bpifb6 BPI fold containing family B, member 6 1.01 0.23 -2.12 Pck1 Phosphoenolpyruvate carboxykinase 1 (soluble) 126.14 31.26 -2.01 Wfdc18 Wap-fourdisulphide core domain18 16.09 4.08 -1.98 Cd209e Cluster of differntiation 209e 5.49 1.40 -1.98 Bpifb2 BPI fold containing family B, member 2 7.05 1.89 -1.90 Cldn8 Claudin 8 1.83 0.52 -1.82 Nxph3 Neurexophilin 3 1.29 0.37 -1.81 Muc15 Mucin 15, cell surface associated 2.30 0.66 -1.80 Fcgbp Fc fragment of igg binding protein 3.84 1.13 -1.76 A530053G22Rik Non-coding RNA 3.32 0.98 -1.76 Mup3 Major urinary protein 3 1.13 0.34 -1.73 4931408D14Rik Non-coding RNA 2.10 0.63 -1.73 Slc44a4 Solute carrier family 44, member 4 1.72 0.52 -1.73 Pdk4 Pyruvate dehydrogenase kinase, isozyme4 10.72 3.26 -1.72 Tshr Thyroid stimulating hormone receptor 2.46 0.75 -1.71 Saa3 Serum amyloid A3 pseudogene 4.05 1.24 -1.71 Il17b Interleukin 17B 1.86 0.57 -1.70 Wwc1 WW and C2 domain containing 1 1.39 0.43 -1.70 S100a14 S100 calcium binding protein A14, 2.23 0.71 -1.65 Mmp12 Matrix metallopeptidase 12 (macrophage elastase) 1.12 0.36 -1.65 Ccdc129 Coiled-coil domain containing 129 1.07 0.35 -1.63 B3galt2 Udp-gal:betaglcnac beta 1,3-galactosyltransferase, polypeptide 2 6.17 1.99 -1.63 Tph1 Tryptophan hydroxylase 1 1.27 0.41 -1.63 Defb1 Defensin, beta 1 4.09 1.33 -1.62 Cbr2 Carbonyl reductase 2 34.73 11.36 -1.61 Arhgap40 Rho gtpaseactivating protein 40 1.60 0.52 -1.61 Tfap2b Transcription factor activating enhancer binding protein 2 beta 2.03 0.67 -1.60 LOC102638835 Non-coding RNA 1.61 0.53 -1.60 Sox10 SRY (sex determining region y)-box 10 3.16 1.06 -1.57 Mogat1 Monoacylglycerol o-acyltransferase 1 2.98 1.00 -1.57 Ptgds Prostaglandin D2 synthase 21kda (brain) 4.94 1.67 -1.56 Alb Albumin 3.14 1.07 -1.56 Elf5 Fas (TNFRSF6) binding factor 1, e74-like factor 5 (ets domain TF) 2.11 0.71 -1.56 Mmd2 Monocyte to macrophage differentiation-associated 2 4.46 1.51 -1.56 Plekhb1 Pleckstrinhomology domain containing, family B (evectins) 1 4.59 1.56 -1.55 Wif1 WNT inhibitory factor 1 1.77 0.61 -1.55 Col5a3 Collagen, type V, alpha 3 7.63 2.62 -1.54 Col6a5 Collagen, type VI, alpha 5 2.71 0.95 -1.51 Ttpa Tocopherol (alpha) transfer protein 1.30 0.46 -1.51 Orm1 Orosomucoid 1 44.29 15.58 -1.51 plet1 Placenta expressed transcript 1 3.86 1.36 -1.51 Spon2 Spondin2, extracellular matrix protein 3.30 1.17 -1.50 Enpep Glutamyl aminopeptidase (aminopeptidase A) 5.84 2.07 -1.49 Adra1a Adrenoceptor alpha 1A 1.08 0.38 -1.49 Areg Amphiregulin 17.60 6.25 -1.49

133 5.3.2 Ingenuity Pathway Analysis of upstream regulators

To gain further insight into gene regulation following CHIKV infection, up- and down-regulated DEGs were analysed using Ingenuity Pathways Analysis (IPA) of up- stream regulators.

IPA up-stream regulator analysis predicts the transcription factors and signalling molecules mediating transcription based on gene expression changes within a data set (i.e. a DEGs list). The majority of predicted up-stream regulators at day 2 in both tissues and in the feet at day 7 were components of type I IFN signalling (Table 5.9 & Table 5.10, bolded entries). IFNγ and IL-12 were also predicted upstream regulators, consistent with CHIKV-induced Th1 responses (Gardner, Anraku et al. 2010) and the up-regulation of multiple IRGs (see sections 5.3.4 & 5.3.5). Consistent with activation of apoptosis during CHIKV (Krejbich-Trotot, Denizot et al. 2011), activity down stream of TRIM24 (negative regulator of p53-dependant apoptosis) (Allton, Jain et al. 2009) and the microRNA mir-21 (suppressor of apoptosis in activated T-cells) (Meisgen, Xu et al. 2012) were down- regulated in samples at day 2 and day 7 feet.

While STAT 1 and IFNγ were more prominent upstream regulators, STAT3 and IL- 4, involved in anti-inflammatory and wound repair processes, were also predicted regulators at day 7 in the feet. In the lymph node at day 7, predicted upstream regulators also reflected the production of humoral immune responses; CSF2, EP400, FOxM1 are drivers of cell cycle progression (Costa 2005, Kim, Woo et al. 2010) with, Myc, CD38 and IL-5 associated with proliferation and differentiation (Pasqualucci, Neumeister et al. 2001) (Table 5.10). Consistent with increased cellular proliferation, the activity of the tumour suppressors SMARCB1 (Modena, Lualdi et al. 2005), RB1 and RBL2 (Villanueva 2014) and pro-apoptosis gene BNIPL3 (Diwan, Koesters et al. 2008) was predicted to be inhibited. The activity of PPARA and PPARG, key regulators of fatty acid metabolism (AlSaleh, Sanders et al. 2012) and transcriptional regulation (Salter and Tarling 2007) and the T-cell marker, CD3 as well as IRGM, an up-regulated gene in day 2 samples and day 7 feet, were also predicted to be inhibited in lymph nodes at day 7.

These analysis have identified a type I IFN mediated pathway that may play a role in gene regulation during CHIKV infection.

134 Table 5.9. The major upstream regulators of gene expression in the feet of CHIKV infected mice at day 2 and day 7 post infection.

P-value indicates the likelihood of a signalling pathway downstream of the listed upstream regulator is involved. Z-score indicates the degree to which a downstream pathway is activated or inhibited. Bold: gene has previously been shown to modulate type I IFN signalling.

135 Table 5.10. The major upstream regulators of gene expression in the lymph nodes of CHIKV-infected mice at day 2 and day 7 post infection.

P-value indicates the likelihood of a signalling pathway downstream of the listed upstream regulator is involved. Z-score indicates the degree to which a downstream pathway is activated or inhibited. Bold: gene has previously been shown to modulate type I IFN signalling.

136 5.3.3 Ingenuity pathways analysis of canonical pathways

Up- and down-regulated DEGs were also analysed using the IPA canonical pathway analysis tool. Activated pathways were similar between feet at day 2 and day 7, and lymph nodes at day 2, with communication, activation, maturation and migration of immune cells common to each sample. The liver X receptor and retinoic acid receptor (LxR/RxR) Activation was predicted to be engaged in the lymph nodes at day 2 but not at day 7 (Figure 5.1). LxR/RxR form heterodimers with PPARA and PPARG (predicted to be inhibited upstream regulators at day 7 in the lymph nodes; Table 5.10), that can regulate gene transcription and chromatin unwinding (Salter and Tarling 2007). Consistent with the absence of type I IFNs and waning viral titres at day 7 (Chapter 2), Interferon signalling and Activation of IRF by cytosolic PRRs pathways were not activated in the feet, nor lymph node at day 7. The pathways predicted to be engaged in the lymph nodes at day 7 were dominated by cell cycle control, DNA damage repair and homologous recombination. This is indicative of the generation of antibody responses that begin as early as day 4 post CHIKV infection (Kam, Lum et al. 2012), and is consistent with the predicted upstream regulators engaged in B cell proliferation (previous section 5.3.2) and the abundance of immunoglobulin variable regions up-regulated in day 7 lymph node (Table 5.4).

The results here, together with results of section 5.3.2, are consistent with the known immunobiology of the host response to CHIKV and are also consistent with the previous microarray studies (Nakaya, Gardner et al. 2012).

137

Figure 5.1 IPA canonical pathways engaged in mouse feet and lymph nodes at day 2 and day 7 post CHIKV infection. Up- or down- regulated DEGs that are part of an IPA canonical pathway are presented as the percentage of total genes in that pathway. Pathways are ranked by p-value.

138 5.3.4 Quantitation of type I IFN regulated genes

Given the critical role of type I IFNs in CHIKV infection (Chapter 2) (Couderc, Chrétien et al. 2008), the abundance of type I IFN regulated genes (IRGs) in the DEGs lists was assessed using the Interferome v2.01 database (Rusinova, Forster et al. 2013).

In feet at day 2 and day 7, and lymph nodes at day 2,≥ 65% of the top 50 up - regulated DEGs identified by Interferome were IRGs (Figure 5.2, grey bars). In all samples, there was a higher percentage of IRGs in the top 50 up-regulated DEGs compared to total DEGs. In both tissues, the percentage of type I IRGs was greatest at day 2 (Figure 5.2, black bars), consistent with the peak production of type I IFN at day 2 post infection seen in Chapter 2. However, the percentage of type I IRGs remained elevated at day 7 in the feet. In lymph nodes at day 7, the percentage of type I IRGs was substantially lower than the other samples.

A number of DEGs, were not recognised by, or listed in Interferome, such as predicted genes and the type I IFNs themselves. A literature search of unrecognised genes brought the percentage of IRGs in the top 50 up-regulated DEGs lists to 86%, 82% and 72% in the feet at day 2 and day 7 and in the lymph nodes at day 2, respectively (Figure 5.2, white bars, Tables 5.1-8, italised text). No additional type I IRGs were found in the top 50 for lymph nodes at day 7 beyond those identified by the Interferome analysis.

These results show that the majority of genes induced following CHIKV infection are regulated by the type I IFNs, particularly at day 2, with sustained up-regulation of type I IRGs in the feet at day 7.

139 Percentage of type I IFN regulated genes

100 80 60 40 20 0 Day 2 Ft Day7 Ft Day 2 LN Day 7 LN Total Top 50 Ammended top 50

Figure 5.2. Percentage of type I IRGs in up-regulated DEGs. The percentage of IRGs in the total (black) and the top 50 (grey) up-regulated DEG lists as determined by Interferome v2.01. White bars represent the percentage of type I IRGs in the top 50 up- regulated DEGs identified by Interferome, plus type I IRGs identified by literature search. Antibody variable regions were excluded from day 7 lymph node DEGs lists. LN: lymphnode; Ft: foot.

140 5.3.5 Quantitation of interferon gene expression

The previous section showed continued up-regulation of IRGs in feet at day 7. Eighteen known murine type I IFNs exist, including thirteen IFNα subtypes with often confounding sequence homology (Fung, Sia et al. 2004, Woelk, Frost et al. 2007). RNA- Seq (unlike microarray technologies) enables identification of the individual IFNα subtypes, despite their similarity. To elucidate the individual IFNs involved in gene regulation following CHIKV infection, transcript levels of all type I IFN subtypes, as well as type II IFN (IFNγ), the two type III IFNs (λ2 & λ3) were examined through out the course of infection.

The type I IFNs had a surprisingly low relative abundance of transcripts (Figure 5.3). In both tissues, type I IFNs were highest at day 2, and returned to mock levels by day 7. IFNβ was considerably higher in the feet compared to lymph nodes at day 2, as were IFNα4 and IFNα5, with a small but significant increase in IFNα1, IFNα2 and IFNα6. Conversely, IFNγ was significantly higher in the lymph node compared to feet at day 2. However, IFNγ returned to mock infected levels in the lymph nodes at day 7, while in the feet, IFNγ expression had increased and was significantly elevated compared to both mock and day 2 levels. Type III IFNs did not differ significantly from mock in any tissues or time points. The type I IFN, IFNε is specific to the female reproductive tract (Fung, Mangan et al. 2013) and was not detected in any tissues or time points analysed.

The results here demonstrate the remarkably low level of type I IFN transcripts required to generate the strong type I IFN response induced following CHIKV infection. These results also reconfirm the involvement of IFNγ in the immune response to CHIKV (Gardner, Anraku et al. 2010), and provide a better understanding of the induction of the individual IFNα subtypes.

141

Figure 5.3 Transcript abundance of the IFN subtypes in CHIKV- and mock-infected mouse feet and lymph nodes. FPKM vales of IFN transcripts in the lymph nodes and feet of mock infected mice and at day 2 and day 7 post infection, as determined by Cufflinks. Fragments per kilobase (of exon) per million fragments mapped (FPKM) is the estimated number of gene transcripts based on the number of reads that align to a particular genomic region. LN: lymphnode. Statistics obtained by t-test.

142 5.3.6 FPKM and fold changes of IRGs in the top 50 DEG lists

The previous section showed a distinction between in the expression of type I and type II IFNs in the feet and lymph nodes at the two time points. To assess how IFN expression influenced IRG up-regulation, the expression of a set of IRGs common to day 2 samples and day 7 feet in the top 50 DEGs was investigated.

In mock infected tissues, transcript FPKM levels of the selected IRGs were consistently higher in the lymph nodes than in feet (Figure 5.4). However, upon infection, these genes were up-regulated by an equal or significantly higher fold change in the feet compared to lymph nodes. Interestingly, while many IRGs in the lymph nodes were less up-regulated at day 7 than at day 2, at day 7 all IRGs in the feet remained significantly up- regulated. Gbp1 and Gzmb were higher at day 7 than day 2 in the feet. In the lymph node at day 7, all IRGs except ISG20 and Gzmb had returned to mock-infected levels.

These results suggest continued up-regulation of IRGs in the feet of CHIKV infected mice on day 7. These IRGs emerge to be regulated by both type I and type II IFNs (Interferome). The continued presence of IFNγ in day 7 feet is established in this model of CHIKV (Gardner, Anraku et al. 2010), likely explains the up-regulation of these IRGs.

143

Figure 5.4 Transcript abundance and differential expression of IRGs in CHIKV- and mock-infected mouse feet and lymph nodes. Mock FPKM values of IRGs in the lymph nodes and feet and their Log2 (fold change) at day 2 and day 7 post infection, as determined by Cufflinks. Fragments per kilobase (of exon) per million fragments mapped (FPKM) is the estimated number of gene transcripts, calculated by Cufflinks based on the number of reads that align to a particular genomic region. LN: lymphnode; Ft: foot. Statistics obtained by t-test.

144 5.3.7 Up-regulated IRGs in day 7 feet are primarily regulated by type I IFNs

The observations in section 5.3.6 suggested that type I IRGs may remain up- regulated despite the absence of type I RNAs in day 7 feet. To investigate this phenomenon further, all up-regulated genes in the feet at day 7 were analysed using Interferome v2.0 to quantify gene induction by the different IFN analysed subtypes. The majority of IRGs in day 7 feet emerged to be regulated by type I IFN, with some overlap with IFNγ regulated genes and no apparent type III IFN involvement. Thus in the feet at day 7 feet, type I IRGs were shown to be widely induced, despite the absence of type I IFN transcripts.

Figure 5.5 IFN-induced genes in CHIKV infected mouse feet at day 7. Venn diagram of all up-regulated IRGs in the feet at day 7 post infection detailing the number of IRGs induced by each IFN subtype.

145 5.3.8 Transcriptional activity of IRF3 and IRF7 during CHIKV infection

IRF3 and IRF7 are key type I IFN transcription factors. Chapter 2 demonstrated the importance of IRF3 and IRF7 in preventing severe CHIKV disease. To further understand their function during CHIKV infection, the expression of IRF3 and IRF7 alternatively spliced isoforms were compared in all samples.

Whole gene expression (i.e. all transcriptional isoforms) of IRF3 did not differ between mock, day 2 and day 7 time points in either tissue (Figure 5.6A). However, the number of IRF3 transcripts was significantly higher in the lymph nodes compared to feet at each time point. The transcript abundance of IRF7 was significantly elevated at day 2 post infection in feet and lymph nodes (Figure 5.6 B) consistent with RT PCR data (Chapter 2). At day 7, IRF7 gene expression in the feet had reduced from day 2 levels, but was still significantly higher than in mock infected feet. In contrast, in the lymph nodes at day 7, IRF7 gene expression had returned to mock infected levels.

Seven novel IRF3 transcript isoforms were identified by Cufflinks, in addition to the two known IRF3 mRNA sequences (Karpova, Ronco et al. 2001) (Figure 5.6C). Alignment of the novel isoforms to the mouse and the human IRF3 amino acid sequences indicated that five of the novel isoforms were lacking the activation domain (AD) (Supplementary data, Figure S1). There were no significant differences in expression of the IRF3 isoforms at the different time points. Although there was a substantially higher proportion of IRF3 isoforms lacking the AD in seen in the lymph nodes (Figure 5.6C), this did not reach statistical significance. The three known IRF7 isoforms were also identified. However, no novel isoforms were reported by Cufflinks and there was no change in isoform transcript abundance in the different tissues and time points (Figure 5.6D).

The results here reveal the presence of seven, previously undescribed IRF3 isoforms that are expressed at different ratios in the feet and lymph nodes. IRF3 gene and isoform expression and IRF7 isoform expression were stable following CHIKV infection. In contrast, IRF7 was substantially up-regulated in both tissues by day 2, and remained elevated in the feet at day 7.

146

Figure 5.6 Gene and isoform expression of IRF3 and IRF7 in mock infected and CHIKV-infected mouse feet and lymph nodes. A & B) FPKM values of total gene transcripts in mock infected feet and lymph nodes and at day 2 and day 7 post infection. C & D) The proportion of individual isoform transcripts as a percentage of total gene transcripts. IRF3 isoforms that contain the activation domain are coloured green. v: variant/isoform. LN: lymphnode; Ft: foot. Statistics obtained by t-test.

147 5.3.9 Assessment of uncharacterised DEGs

Many DEGS induced following CHIKV infection were unannotated, or uncharacterised. Unannotated/uncharacterised genes in the top 50 up-regulated DEGs lists were termed CHIKV-induced novel transcripts (CINTs) and were investigated for their validity and protein coding potential (described in material and methods section 5.2.4). Three CINTs, from a total of 22, were identified as having the potential to encode a novel protein (Table 5.11) (full list in supplementary data, Table S1) and their RNA sequences were investigated for the strength of their Kozak sequence. A Kozak sequence is required for ribosomal recognition of the translational start site and was considered strong if the RNA sequence matched the consensus sequence (Figure 5.7) at position -3, or +4, adequate if position -6, -3 or +4 matched, and weak if no nucleotides in the CINT RNA sequence matched the Kozak consensus (Kozak 1986). Only CINT9 was found to have a strong Kozak sequence (Supplementary data, Figure S2). Thus, CINT11.2 and 15.2 were considered potential long non-coding RNAs, and were not investigated further for protein coding potential.

148 Table 5.11 Up-regulated, unannotated genes with protein coding features

Protein coding gene CINT 9 CINT 11.2 CINT 15.2 feature

log2 (FC) 7.71 4.59 3.24

FPKM 24.34 38.19 3.48

ORF ≥ 30 aa Y Y Y

uORF 0 0 0

uAUG 0 4 2

In frame stop codon N N N

5’UTR ≥ 20aa Y Y Y

Protein homology Ly6-I TRIM30a mCG14456

Conserved domains LU Domain SPRY HMG box

Kozak sequence Strong None Very weak

Figure 5.7. Kozak consensus sequence with nucleotide position indicated and start codon (AUG) indicated in bold, underlined text.

149 5.3.10 Investigation of the protein coding potential of CINT9

The previous section showed that one CHIKV induced novel transcript (CINT), CINT9, identified/predicted by Cufflinks, exhibited protein coding potential. The possible function of the CINT9 gene was investigated further.

CINT9 was expressed in the lymph nodes at day 2 post infection, and was up- regulated 208 fold. Using the BLASTx algorithm to search the NCBI protein database, significant homology (e-value = 3 x 10-46) was found between lymphocyte 6 antigen B (Ly6B) and CINT9 sequences. The CINT9 locus was located within the Ly6 antigen cluster on chromosome 15, between Ly6C1 and Ly6C2 (Figure 5.8A). An amino acid sequence alignment of all mouse Ly6 antigens confirmed that the ly6/uPA receptor-like (LU) protein domain, conserved in all Ly6 antigens, was also present in CINT9 (Figure 5.8B, full alignment in supplementary data, Figure S3). Evolutionary analysis based on the amino acid sequence alignment of all murine Ly6 antigens further confirmed the relatedness of CINT9 with the Ly6 antigen gene family. However, synteny analysis showed that CINT9 does not have a human orthologue (Figure 5.8A) and is located in a divergent gene cluster that is not found in humans (Lee, Wang et al. 2013) (Figure 5.8 A and C).

These results indicate the existence of potentially novel protein coding Ly6 antigen gene, which is up-regulated in the lymph nodes following CHIKV infection.

150

151

Figure 5.8 Figure 4. Synteny analysis, sequence homology and phylogenetic relationship between CINT9 and the Ly6 antigen family. A) Diagram of the Ly6 antigen gene cluster on murine chromosome 15 and corresponding region on human chromosome 8. Orthologous genes are connected by lines. Adapted from Lee, Wang et al. (2013). B) Alignment of Ly6 antigen amino acid sequences and the predicted protein sequence of CINT9 showing the conserved LU domain, containing 10 cysteine residues, present in all sequences. Conservation threshold was 60%. C) Phylogenetic analysis of CINT9 and Ly6 antigens in the Ly6 gene cluster conducted using MEGA v4.0 (Tamura, Dudley et al. 2007). The percentage of replicate trees in which The bootstrap test (500 replicates) percentages are shown next to the branches. The evolutionary distances were computed using the p-distance method and are in the units of the number of amino acid differences per site.

152 5.4 Discussion

The results here represent the first deep sequencing study of the in vivo mammalian host response to an alphavirus. The findings demonstrate the nature of the genes expressed in mouse feet and lymph nodes following infection with CHIKV. Early infection (day 2) was dominated by the type I IFN responses, with the exception of gene up-regulation in day 7 lymph nodes. This was despite a relatively low level of type I IFN mRNAs. IRGs remained substantially elevated in the feet at day 7 post infection, despite an absence of type I IFN mRNA. In the lymph nodes, up-regulated genes were dominated by immunoglobulin variable region genes on day 7, with IFNγ heavily up-regulated on day 2 and dropping to mock-infected levels by day 7. Several uncharacterised genes were up- regulated, including one potentially novel Ly6 antigen. Six novel IRF3 isoforms were also identified.

The dominance of type I IFN responses. CHIKV infection induces serum type I IFNs by day 2 post infection, and up to 80% of genes in feet and lymph nodes at this time were type I IFN-regulated, yet the FPKM of type I IFN mRNA transcripts was very low (≤ 8 for all subtypes). This may have been to the low stability of the type I IFNs are subject to tight post-transcriptional control; AU-rich repeats in type I IFN mRNA can be stabilised, but also targeted for rapid for degradation and prevention of detrimental inflammatory responses (Hao and Baltimore 2009, Anderson 2010). Type I IFN mRNA, particularly IFNβ mRNA, expression was higher in feet compared to lymph nodes, consistent with the notion that type I IFNs are predominantly produced by non-haematopoietic cells (Schilte, Couderc et al. 2010). Furthermore, IFNβ mRNA levels were relatively low in the lymph nodes whereas IFNγ levels were high. This is perhaps also consistent with findings that IFNβ can, in certain settings, suppress IFNγ (Nguyen, Cousens et al. 2000, Rayamajhi, Humann et al. 2010, Chessler, Caradonna et al. 2011).

IRF7 and IRF7 activities. All IFN mRNAs had returned to mock-infected levels by day 7. Despite this, IRF7, a key transcription factor for IFNαs (Rudd, Wilson et al. 2012), as well as many other IRGs (Ning, Pagano et al. 2011, Bidwell, Slaney et al. 2012) remained strongly up-regulated in the feet at day 7. IRF7 is itself heavily up-regulated by type I IFNs, and can also be activated in a type I IFN-independent manner when the IRF7

153 protein binds to the virus activated element (VAF) in the IRF7 promoter (Ning, Huye et al. 2005). IRF7 may also be up-regulated by IFNγ (Dölken, Ruzsics et al. 2008, O’Donnell, Conway et al. 2012), which remained up-regulated by ≈208 fold in the feet at day 7. However, IRF7 mRNA translation can be inhibited via binding of the 3’UTR by OASL (Lee, Kim et al. 2013), which was highly expressed at day 7. Guanylate binding protein (GBP4), which was also heavily up-regulated in the feet at day 7 is known to inhibit IRF7-mediated gene induction via interaction with the upstream adaptor molecule, TRAF6 (Hu, Wang et al. 2011).

Persistent up-regulation of IRGs. With no type I IFN mRNA, continued up- regulation of many of the top 50 most up-regulated IRGs in the feet at day 7 appeared to be associated with elevated IFNγ mRNA expression (in the lymph nodes, IFNγ and IRG mRNA expression had largely returned to mock levels by day 7). IFNγ has been shown to work synergistically with BRCA1 to up-regulate the IRGs IFIT2, Mx1 and IRF7 as well as apoptotic responses (Andrews, Mullan et al. 2002, Buckley, Hosey et al. 2007). Maloney, Thackray et al. (2012) found IFNγ-regulated IRF1 and STAT1 induction of gene expression was required for clearance of viral infection in macrophages, and viperin can be induced directly by IFNγ-induced IRF1 (Stirnweiss, Ksienzyk et al. 2010). However, Interferome analysis of all up-regulated genes in feet at day 7 suggests that about 50% of IRGs were likely to be regulated by type I exclusively, despite the absence of type I IFN transcripts. Ongoing stimulation of IRGs by viral RNA, independent of type I IFNs, has been reported and may involve IRF9 (Pulit-Penaloza, Scherbik et al. 2012), unphosphorylated ISGF3 (Cheon, Holvey‐Bates et al. 2013) and/or IRF1 (Stirnweiss, Ksienzyk et al. 2010, Harman, Lai et al. 2011). A concept of viral stress-inducible genes has also been proposed where type I IRGs are stimulated directly by stress responses, rather than type I IFNs (Sarkar and Sen 2004). CHIKV RNA is also known to persist for an extended period (Poo, Rudd et al. 2014) and may be responsible for type I IFN- independent stimulation of type I IRGs.

Granzyme A, B and K. The up-regulation of these granzymes, predominantly in the lymph nodes at day 7, may reflect activation of NK and CD8+ T-cells (Bratke, Klug et al. 2008, Gardner, Anraku et al. 2010). However, neither NK cells (Alsharifi, Lobigs et al. 2006, Poo, Rudd et al. 2014) nor CD8+ T-cells (La Linn, Mateo et al. 1998, Smith-

154 Norowitz, Sobel et al. 2000) appear to play important roles in controlling alphaviral infections. Although recently, CD8+ T cells have recently been seen to contribute to control of RRV in muscle cells (Burrack, Montgomery et al. 2015). Granzymes are believed to be involved in inducing apoptosis of infected cells as well as other anti-viral activities (Domselaar and Bovenschen 2011). However, they have recently also been shown to regulate pro-inflammatory cytokine responses (Wensink, Hack et al. 2015). Their presence in CHIKV infected tissues might play a role in CHIKV immunopathology, more so than anti-viral activities.

GTPases and their roles in viral infection. The IFNγ-induced p47 immunity regulated GTPases and the p65 guanylate binding proteins (GBPs) constituted 20 and 24% of the top 50 most up-regulated DEGs in the feet at day 2 and 7, respectively. The p47 GTPases and GBPs, induced by IFNγ (and to a lesser extent by type I IFNs) (Taylor, Feng et al. 2004), associate with pathogen-containing vacuoles (PVs) to activate the PV’s pathogen killing activities (Kim, Shenoy et al. 2011, Haldar, Saka et al. 2013). The p47 GTPases disrupt PV membranes to promote autophagy and phagolysosome formation (Taylor, Feng et al. 2004, Kim, Shenoy et al. 2011), while GBPs facilitate trafficking of antimicrobial substrates, including NADPH oxidase (NOX2) and autophagy effectors into PVs (Haldar, Saka et al. 2013). Up-regulation of GTPases in response to an alphavirus has been reported previously (Ryman, White et al. 2002). While p47 GTPases are known to be critical for protection against several bacteria and protozoa (Taylor, Feng et al. 2004, Saha, Prasanna et al. 2010), little is known about their role in viral infections . GBPs may have a role in viral infection with over-expression of GBP2 shown to inhibit the single- stranded RNA viruses vesicular stomatitis virus (VSV) and encephalomyocarditis virus (EMCV) in mice (Carter, Gorbacheva et al. 2005), and GBP1 and 3 capable of inhibiting influenza, in vitro (Nordmann, Wixler et al. 2012). CHIKV is known to engage autophagy processes to facilitate replication (Krejbich-Trotot, Gay et al. 2011, Joubert, Werneke et al. 2012). Thus it is plausible that CHIKV might induce and exploit the GTPases to facilitate replication, or that the GTPases represent a class of antiviral effectors not previously described for alphaviral infections.

IRF3 isoforms. Seve novel IRF3 isoforms were identified. Two IRF3 isoforms are known to exist, with one isoform a non-coding RNA that is targeted for non-sense-

155 mediated RNA decay (Zhai, Gao et al. 2008). While all seven novel isoforms identified here encoded the highly conserved DNA binding domain, five isoforms were shown to be missing all, or part of the activation domain (Lin, Mamane et al. 1999), indicating that these isoforms may act as antagonists of full length IRF3 (Karpova, Ronco et al. 2001, Li, Ma et al. 2011). IRF3 protein is expressed constitutively and only becomes activated following TLR stimulation (Jensen and Thomsen 2012). Accordingly, overall IRF3 expression was not seen to differ greatly between different time points. However, expression of the different isoforms was seen to differ between the two tissues. Further research will be required to understand the activities of these isoforms; however, it is possible that isoforms without activation domains act as inhibitors of IRF3 transcriptional activities (Karpova, Ronco et al. 2001, Zhai, Gao et al. 2008).

Protein coding potential of CINT9 CINT9 was found to meet the criteria for being a protein coding gene according to the Havana Annotation Guidelines (Wilming, Frankish et al. 2012) and the NCBI CCDS Curation Guidelines (Pruitt, Harrow et al. 2009). High homology with Ly6B, including the presence of the conserved LU domain and was located within the Ly6 antigen locus on mouse chromosome 15. Ly6B and Ly6C2 were up- regulated in samples at day 2 and in the feet at day 7. The ly6 antigens are cysteine-rich, generally GPI-anchored cell surface proteins , commonly expressed on monocytes, DCs and neutrophils (Bamezai 2004, Lee, Wang et al. 2013) and have roles in cytokine production, cell migration and differentiation (Leuschner, Dutta et al. 2011, Ji, Liu et al. 2012). However, the CINT9 gene is located in a region of the chromosome that is present only in mice (Lee, Wang et al. 2013), and is unlikely to have a human orthologue.

Non coding RNAs. As understanding of the genome progresses, it is apparent that gene expression is not limited to protein coding genes, but also diverse RNA species that can modulate gene and protein expression (Rinn and Chang 2012). Non-coding genes are increasingly emerging to have a role in immune regulation, with microRNAs able to inhibit transcription or target mRNA for degradation (Heward and Lindsay 2014), long non- coding RNAs (lncRNA) able to exert epigenetic regulation of gene expression (Lee 2012, Rinn and Chang 2012), and recently a class of IFN-regulated lncRNAs has been identified (Josset, Tchitchek et al. 2014). CINT11 was found to have several characteristics of a long non-coding RNA: it is >200 bp long, exhibited characteristics of a protein coding

156 gene, was immediately upstream of TRIM30α, a gene with which CINT11 shared considerable homology, but was unlikely to be protein coding (Poliseno 2012, Qureshi and Mehler 2012, Rinn and Chang 2012). TRIM30α down-regulates inflammasome activation (Hu, Mao et al. 2010) and TLR signalling (Shi, Deng et al. 2008), and was identified as a gene involved in viral clearance in later stages of encephalitic alphavirus, Venezuelan equine encephalitis virus (Atasheva, Akhrymuk et al. 2012). LncRNAs are often the product of gene duplication and function to block the activity of neighbouring homologues (Poliseno 2012, Rinn and Chang 2012). It is possible that expression of CINT11 at day 2 following CHIKV had a role in blocking TRIM30α, thus promoting TLR signalling.

Down-regulation of CD209 isoforms. All CD209 isoforms and their homologue CD207 (Langerhin) were seen to be heavily down-regulated in all tissues and time points, as were other lectins/carbohydrate-binding molecules. Down-regulation of these genes likely reflects stimulation of antigen presenting cells, which commonly down-regulate cell surface receptors, including CD209, once they have taken up an antigen (Relloso, Puig- Kröger et al. 2002). The down-regulation of a number of other M2 (anti-inflammatory) markers (Retnla, F13a1 Mrc1, CD163) likely reflects active suppression of homeostatic processes during an inflammatory response

The role of the microRNA, mirlet7. In the lymph nodes, the most heavily down- regulated gene at both day 2 and day 7 was the microRNA, mirlet7h. The different mirlet7 subtypes are known to be involved in a broad range of functions, such as apoptosis cell cycle regulation and tumour suppression (Ambros 2004, Rozen-Gagnon, Stapleford et al. 2014), where they function as negative regulators in post-transcriptional control. Down- regulation of mirlet7 is known to occur in stimulated macrophages and results in increased production of IL-6 and IL-10 (Liu, Chen et al. 2011, Schulte, Eulalio et al. 2011). Consistent with down-regulation of the suppressive role of mirlet7 genes, a number of mirlet7 targets involved in apoptosis (Bcl proteins, Casp3) (Shimizu, Takehara et al. 2010, Wang, Tang et al. 2011) and cell cycle regulation (cyclin dependant kinases and cell division cycle genes, AIM1, E2f and CCNa) (Johnson, Esquela-Kerscher et al. 2007, Lee and Dutta 2007, Peter 2009) were up-regulated in the lymph nodes.

157 6. RNA-Seq analysis of CHIKV transcription in mice

6.1 Introduction

CHIKV enters the body by mosquito bite and is thought to spread from the site of infection to the draining lymph nodes (Schwartz and Albert 2010). CHIKV causes acute and chronic, often debilitating polyarthritis that can last for months, to over a year, after viraemia has been cleared (Suhrbier, Jaffar-Bandjee et al. 2012). In mice, infectious virus can be recovered from arthritic feet up to 14 day post infection, while viral RNA can persist for up to 100 days post infection (Poo, Rudd et al. 2014). Viral RNA can also be detected in joint tissues for extended periods after the end of the viraemic period in humans (Hoarau, Jaffar Bandjee et al. 2010) and monkeys (Labadie, Larcher et al. 2010), with persisting viral RNA thought to be responsible for chronic inflammatory disease (Suhrbier and Mahalingam 2009).

The alphaviral genome is a ~12kb single-strand, positive sense RNA with two open reading frames (ORFs) encoding four non-structural proteins (nsps) and five structural proteins (Strauss and Strauss 1994). The genomic RNA resembles messenger RNA with 5’ and 3’UTRs that facilitate viral RNA transcription and translation (Hyde, Chen et al. 2015). Like all RNA viruses, alphaviruses encode their own RNA polymerase, which is believed to have a high error rate (~10-4 errors per base) (Crotty, Cameron et al. 2001, Coffey, Beeharry et al. 2011). Upon infection, the non-structural proteins are translated and form replication complexes (Strauss and Strauss 1994). Replication complexes then interact with a conserved sequence element (CSE) in the 3’UTR, which serves as the promoter for negative strand RNA synthesis (Kuhn, Hong et al. 1990, Frolov, Hardy et al. 2001). Negative strand RNA is used a template for transcription of (i) full length, positive strand genomic RNA for virion packaging and translation, and (ii) 26S subgenomic RNA (encoding structural proteins), via the subgenomic promoter (Strauss and Strauss 1994, Schwartz and Albert 2010).

Herein, RNA-Seq was used to gain insights into the evolution of CHIKV RNA in tissues, given that this viral RNA is likely responsible for acute and chronic disease. Total RNA was isolated from the feet and lymph nodes of CHIKV infected mice (described in Chapter 4) and used for RNA-Seq analysis of polyadenylated RNA, which includes both positive strand genomic, and subgenomic CHIKV RNA.

158 6.2 Materials and Methods

6.2.1 Alignment of CHIKV sequences The data analysed in this chapter is from the experiments described in Chapter 4. Specifically, the data files from which read sequences mapping to the mouse genome had been removed (section 4.2.7) were aligned to two CHIKV genomes. Bowtie v2.0.2 (Trapnell, Pachter et al. 2009) (default parameters) was used to align non-mouse reads to (i) the LR2006_OPY1 CHIKV reference genome (Parola, de Lamballerie et al. 2006) (accession: DQ443544.2, described in section 4.2.7), and (ii) a parental/input virus genome generated by sequencing of input virion RNA (Poo, Nakaya et al. 2014) and further sequencing of the 3’ UTR.

6.2.2 Evaluation of alignments & investigation of the parental genome sequence Reads alignments generated by Bowtie were visualised in the Integrated Genome Viewer (IGV) (Robinson, Thorvaldsdóttir et al. 2011, Thorvaldsdóttir, Robinson et al. 2012). Where only one ‘mate’ in a paired-end read mapped to the CHIKV genome, IGV was used to extract the sequence of its unmapped mate. The unmapped sequences were used as queries in homology searches using the basic local alignment search tool (BLAST, NCBI). Unmapped read sequences that aligned to the CHIKV genome were used to create a consensus sequence. The consensus sequence was incorporated into the LR2006_OPY1 genome to create a reference genome for the input/parental virus.

6.2.3 Determination of partial CHIKV genome sequence The inferred sequence of the parental genome was confirmed by RT-PCR amplification of the input virus using the reaction and cycling conditions described in Chapter 2 (section 3.3.1) and the following primer sequences (5’3’): cgggaagctgagatagaagttg (forward), cataccgaactcttccacgatt (reverse). The virus was also sequenced using BigDye terminator v3.1 (Life Technologies), according to the manufacturer’s instructions.

159 6.2.4 Mutational assessment SAMtools count algorithm was used to obtain read mapping counts from alignment data generated by Bowtie v2.0.2. The total number of reads mapped to the genome as well as the number of reads mapped with a mapQ >20 (i.e. uniquely mapped to the genome), which also contained at least 1 mismatched base or insertion/deletion mutation, were derived and used to calculated the percentage of mutated reads per sample. The IGV allele frequency threshold parameter was set to 5% such that positions in the genome at which 5% or more called nucleotides varied from the reference genome were highlighted and available for quantitation of potential mutations. The Ilumina HiSeq200 sequencing error rate is 0.4 – 2.0% depending on read length (Glenn 2011, Minoche, Dohm et al. 2011, Wall, Tang et al. 2014).

160 6.3 Results

6.3.1 Read mapping to the LR2006_OPY1 CHIKV genome RNA-Seq data described in Chapter 4, with mouse genome sequences removed were mapped to the LR2006_OPY1 CHIKV genome and alignment files loaded into IGV. The frequency allele threshold was set to 5% (sequencing error rate is estimated to be 0.4 - 2%), to enable identification of sites in the sequenced virus that differed from the LR2006_OPY1 genome in ≥ 5% of the nucleotides mapped to that site.

Deviations of the mouse-derived CHIKV RNA sequences from the LR2006_OPY1 genome of ≥ 5% are indicated by coloured vertical bars (Figure 6.1A, upper panels). Some deviations were present in all samples, with ≥97% frequency at each base position (Figure 6.1A, red boxes). A single nucleotide deletion at position 9952 bp was also present in ~7.6% of reads in all samples, except day 7 lymph node (Figure 6.1A, blue box). A region within the 3’UTR of the genome was heavily divergent from the reference genome, with few, heavily mismatched (> 5 mismatches) or no reads mapping to the region 11695 - 11740 bp (Figure 5.1B). Perplexingly, a high number of reads that mapped uniquely to the genome (mapQ ≥20) either side of this aberrant region had a mate pair that could not be mapped to the LR2006_OPY1 genome (Figure 6.1B, red horizontal reads), suggesting a deletion

The presence of unmapped mate-pairs, and consistent deviations from the LR2006_OPY1 sequence, present in multiple samples from different mice, indicated that the input virus was distinct from LR2006_OPY1, rather than representing genuine mutations. Sequencing of the parental input virus was deemed necessary.

161

162 Figure 6.1 Reads mapping to the CHIKV LR2006_OPY1 genome. A) Schematic diagram of reads aligned to the CHIKV genome. A bar chart of the log10 transformed read coverage (i.e. the number of reads aligned at a position in the genome, upper panels) and a condensed image of the individual 100 bp reads mapped to the genome (grey horizontal bars, lower panel) are shown. Coloured, vertical bars in the upper coverage panels denote positions in the genome at which a nucleotide differs from the reference genome in ≥ 5% of quality-weighted aligned reads. Red boxes indicate positions at which ≥ 97% of nucleotides differ from the reference. Blue boxes indicate an insertion in ~7.6% of reads. Base positions with < 20 reads mapped were excluded from analysis. Images are representative of the three replicates for each sample. B) Enlarged image of the poorly mapped region in the 3’UTR of day 2 foot showing the reads with unmapped mate-pairs either side. Red horizontal bar: read with unmapped mate, purple: single nucleotide insertion, black: single nucleotide deletion. Images representative of 3 replicates.

163 6.3.2 Sequencing of the parental input virion genome Section 6.3.1 suggested differences between the input virus used to infect mice and the LR2006_OPY1 genome. The parental virion genome sequence was established by Ion Torrent sequencing as described by Poo, Rudd et al. (2014). However, the 3’ UTR was not fully sequenced using this technique. PCR primers were designed that flanked the aberrant 3’ UTR region and the region was amplified by RT-PCR. RT-PCR products were then sequenced confirming a deletion in the 3’UTR, which was 44 bp in length, and new parental reference CHIKV genome, which incorporated the site changes seen by Poo, Rudd et al. (2014), as well as the 44 bp deletion in the 3’UTR, was generated

5.3.3. Alignment to the new parental CHIKV genome

When non-mouse reads were re-mapped to the new parental virus genome (5.3.2), the site differences and the aberrant 3’ UTR region, observed in Figure 6.1, were no longer present (Figure 6.2A & B), and general features of the sequence analysis could now be described.

A reduction of ≈ 10 fold in the number of viral transcripts from day 2 to day 7 post infection in both tissues was apparent. This is consistent with the pattern of viral replication seen in Chapter 2, with peak tissue titres and serum viraemia seen at day 2 and undetectable viraemia at day 7. The number of transcripts in the feet was significantly higher than in lymph nodes at both time points (Mean reads mapped: foot day 2: 4,572,386 and day 7: 77,219 versus LN day 2: 4097 and day 7: 1444) (Figure 6.2A). The lower number of transcripts in the lymph node compared to feet at each time point reflects the lower viral tissue titres seen in the lymph nodes compared to feet, as reported previously (Chapter 2) (Rudd, Wilson et al. 2012). In all samples, a higher number of reads were seen to align to the structural protein ORF (non-structural: structural, foot day 2: 1:95±14, day 7: 1:57±14; lymph node day 2 1:13±4, day 7 1:18±2), consistent with the high ratio of subgenomic to genomic RNA found in alphavirus infected cells (Perri, Driver et al. 2000, Ng, Coppens et al. 2008). A feature of the alignments not previously reported in CHIKV replication was a high number of viral transcripts seen to map to the 5’UTR and Nsp1 at day 7 in the foot.

164 165 Figure 6.2 Reads mapping to the CHIKV genome. A) Schematic diagram of reads aligned to the CHIKV genome. A bar chart of the log10 transformed read coverage (i.e. number of reads aligned at a position in the genome, upper panels) and a condensed image of the individual 100 bp reads mapped to the genome (grey horizontal bars, lower panel) are shown. Coloured, vertical bars in the upper coverage panels denote positions in the genome at which a nucleotide differs from the reference genome in ≥ 5% of aligned reads. Base positions with < 20 reads mapped were excluded from analysis. Images are representative of the three replicates for each sample. B) Enlarged image of the 3’UTR of day 2 foot. Red horizontal bar: read with unmapped mate, black: single nucleotide deletion. Images representative of 3 replicates.

166 6.3.3 Investigation of mutation accrual in the CHIKV genome

Having established the sequence of the parental input virus, the accrual of mutations could be analysed. The percentage of mutated, 100 bp reads in each sample, and the number of positions in the viral genome at which 5% or more of called bases differed from the reference were quantified (Fig. 5.3A and B).

Figure 6.3A shows the percentage of mapped reads in each sample that mapped uniquely to the genome (i.e. MapQ >20), but also contained at least one mismatched base and/or insertion/deletion mutation. In the feet, ~20% of mapped reads contained 1 or more mutations at day 2 and day 7. The percentage of mutated reads was slightly but significantly higher in the lymph nodes (~23%) than in feet (~20%) on day 2. Over 40% of reads at day 7 in LN were found to contain at least one mutation, which was significantly higher than both day 2 LN (23%) and day 7 foot (20%). The percentage of mutated reads in each of the 3 replicate samples was remarkably consistent, as indicated by the very small error bars.

Figure 6.3B presents the percentage of sites within the genome where 5% or more of called bases differed from the reference genome. Only positions in the genome that had a coverage level of at least 20 reads were considered (e.g. 1 read in 20 with a mutation in a given site would reach the 5% threshold at that site). The percentage of sites mutated increased significantly in both feet and lymph nodes from day 2 to day 7 post infection. Consistent with Figure 5.3A, the percentage of mutated sites was significantly higher in lymph nodes at both time points, and was particularly evident on day 7.

The results clearly show that there is an increase in mutations in the CHIKV genome over time (from day 2 to day 7). Furthermore, these results also show that mutations are more prevalent in lymph nodes than in feet.

167

Figure 6.3. Percentage of mutations present in RNA-Seq reads aligned to the CHIKV genome. A) The number of reads that had a Map Q>20, plus at least one mismatched base and/or and insertion/deletion mutation as a percentage of total reads aligned to the modified CHIKV genome in each tissue and time point. B) The number of positions in the CHIKV genome at which ≥ 5% of aligned nucleotides differed from the reference genome, as a percentage of total base pairs in the modified CHIKV genome. *: p ≤ 0.05, **: p ≤ 0.01, n.s.: not significant by t-test. n = 3 replicates per sample.

168 6.4 Discussion

This Chapter represents the first deep sequencing analysis of CHIKV genomes in tissues. The data presented here reflected the known CHIKV replication kinetics, in which virus spreads from the site of infection (feet) to the draining lymph nodes, via antigen presenting cells (Schwartz and Albert 2010), with viral titres peaking at day 2 and largely resolved at day 7 (Chapter 2) (Rudd, Wilson et al. 2012). A 44 bp deletion was found to be present in the parental input virus. When mapped to a reference genome that was consistent with the input/parental virus, mutations within the CHIKV genome appeared to accumulate with time, particularly in the nsp ORF, and in the lymph nodes more so than in feet.

The CHIKV genome was seen to accumulate mutations over time and the rate of mutation accumulation was higher in lymph nodes than in feet. The error rate of RNA polymerases is estimated to be 10-4 mutations per nucleotide copied and is the likely cause of mutation accumulation (Holland, Spindler et al. 1982, Steinhauer and Holland 1987, Coffey, Beeharry et al. 2011). The accumulation of errors eventually results in error catastrophe and loss of replication capacity (Crotty, Cameron et al. 2001). This is may account for the inability to recover infectious virus from feet of CHIKV infected mice after day 14 post infection, despite the presence of CHIKV RNA (Poo, Rudd et al. 2014). Why the mutation rate was higher in the lymph nodes than in feet is unclear. Viral RNA in the lymph nodes may be subject to greater immune pressure, or alternatively, the higher temperature of the lymph nodes compared with peripheral limbs (i.e. feet) may increase the mutation rate. This increased rate of mutation accrual may also account for the lack of persistent infectious virus and viral RNA in lymphoid organs such as spleen and lymph nodes (Poo, Nakaya et al. 2014).

As expected, fewer reads mapped to the nsp ORF than the structural ORF, in all samples. As described in Chapter 4, an increased ratio of subgenomic to genomic RNA, driven by the subgenomic promoter, is required during alphavirus replication to facilitate production of 240 structural E1/E2 envelope proteins that house a single genome copy per mature virion (Rümenapf, Strauss et al. 1994, Perri, Driver et al. 2000, Ng, Coppens et al. 2008). The increased number of reads aligned to the genome at day 2 compared to day 7, is consistent with previous findings, including those in Chapter 2, in which tissue titres

169 peaked at day 2, and almost resolved by day 7 (Gardner, Anraku et al. 2010, Rudd, Wilson et al. 2012). The increased number of reads aligned to the genome in feet compared to lymph nodes is consistent with the higher titres in feet compared to lymph nodes (Chapter 2) (Rudd, Wilson et al. 2012). Feet were also the site of inoculation, and contain abundant synovial tissue, where alphaviruses are known to replicate (Strauss and Strauss 1994, Schwartz and Albert 2010). Virus in the lymph node had to spread from the site of inoculation, so titres might be expected to be lower in this tissue due to delay.

Interestingly, a greater number of reads aligned to the 5’UTR/nsp1 (compared to nsps 2 - 4) in feet at day7. Nsp1 facilitates capping of the 5’UTR (Ahola and Kääriäinen 1995) and the 5’UTR also forms secondary structures that provide mRNA stability (Nickens and Hardy 2008), as well as containing promoter elements required for initiation of transcription (Gorchakov, Hardy et al. 2004). The higher number of reads mapped to the 5’UTR/nsp1 in feet at day 7 may reflect increased RNA stability of degradation products or aborted transcripts.

A frame shift mutation in the 6K protein involving a T insertion was identified in around ~ 7.6% of reads. With the exception of day 7 lymph node (due to the paucity of mapped reads) this frame shift mutation was seen in all samples, and possibly represents a subset of viral species in the inoculum. However, Forrester, Guerbois et al. (2011), in a study of intra-host variation of viral genomes, have previously reported repeated deletions in the 6K protein of Venezualean Equine Encephalitis virus, resulting defective interfering (DI) RNAs. DI RNAs can antagonise viral proteins, and compete with parent virus for host transcription/translation factors, effectively enhancing viral fitness by slowing replication (Bangham and Kirkwood 1990, Thompson, Rempala et al. 2009). DI RNA can also be packaged into DI particles which cannot replicate independently, but can still infect host cells (Thompson, Rempala et al. 2009, Firth, Atasheva et al. 2011). DI particles have been associated with alphaviral persistence, in vivo (Atkinson, Barrett et al. 1986, Firth, Atasheva et al. 2011) and may contribute to the development of chronic arthritis following CHIKV infection, despite the absence infectious virus.

170 RNA-Seq of CHIKV RNA in tissues prompted a re-sequencing of RNA from parental input virions (Poo, Rudd et al. 2014), which confirmed several deviations from the LR2006_OPY1 genome. The LR2006_OPY1 genome sequence was derived from human isolates that were passaged once in mammalian (Vero) cells prior to sequencing (Parola, de Lamballerie et al. 2006). The parental virus used herein was from the same stocks sequenced by Parola, de Lamballerie et al. (2006), with two additional rounds of passaging in C6/36 mosquito cells (Gardner, Anraku et al. 2010). RNA-Seq of the tissues also revealed a 44 bp deletion in the 3’UTR of the parental virus. The 3’UTR of CHIKV varies widely between different strains due to direct sequence repeats that are duplicated, inverted or deleted in different CHIKV strains (Chen, Wang et al. 2013). These variations are thought to facilitate viral fitness in different hosts/vectors (Chen, Wang et al. 2013, Hyde, Chen et al. 2015). The 44 bp deletion detected in the parental virus sequence corresponds to part of a direct sequence repeat that occurs twice in the Asian lineage, and once in the ECSA lineage, of which LR2006_OPY1 derives (Chen, Wang et al. 2013). The parental virus used herein thus likely acquired the site changes and the deletion in the 3’UTR during the two passages in C6/36 mosquito cells.

The results presented here highlight the simple, yet effectual nature of alphavirus replication. CHIKV quickly replicates to high titres, can mutate to evade host immune pressures and rapidly adapt to different hosts. Despite encoding only eight genes in 2 ORFs, CHIKV operates as a highly refined and effective pathogen.

171 7. General Discussion

7.1 Summary of major findings

Chapter 2: IRF3/7 and adaptor molecules. This chapter demonstrated a critical role for IRF3 and IRF7 in effective type I IFN responses and protection against CHIKV infection and mortality. IRF7, which was substantially up-regulated by day 1 of CHIKV infection, was essential for the type I IFN positive feedback loop. IRF3/7 deficient mice developed severe disease that was associated with excessive IFNγ and TNF production and ultimately resulted in a lethal haemorrhagic shock, reminiscent of dengue haemorrhagic shock/dengue shock syndrome (DHS/DSS). The findings implicated the type I IFN response in suppression of pathological levels of IFNγ, and in preventing the development of severe CHIKV disease. Using mice deficient in the adaptor molecules upstream of IRF3 and IRF7, multiple pathways were able to induce type I IFN production. However, IPS1 was found to be the most important for type I IFN production, then TRIF, with MyD88-/- mice capable of producing type I IFN, at only slightly reduced levels compared to WT mice. Type I IFN induction following CHIKV infection was thereby shown to be predominantly via the RIG-1/MDA5/IPS-1 and TLR3/TRIF signalling pathways.

Chapter 3: GFP reporter mice. This chapter sought to identify the cellular source of type I IFN production following CHIKV infection. Initially, reporter mice, t heterozygous for expression of eGFP under the control of either the IFNβ or the IFNα6 promoter, were infected with CHIKV. However, these mice emerged not to be informative, as eGFP production was too low for detection under all conditions tested. This was presumably due to disruption of positive feedback loops required to enhance the intrinsically low activity of the type I IFN promoter.

Analysis of tissue-specific type I IFN induction by RT-PCR and comparison with viral tissue titres in infected WT mice revealed a strong correlation between local type I IFN induction and viral tissue titres, with feet and lymph nodes producing the most robust type I IFN response. This suggested that detection of local viral infection directs the production of type I IFN in responses to CHIKV infection

172 Chapter 4: Assessment of RNA-Seq data. Comprehensive analysis of RNA isolated from the feet and inguinal lymph nodes of mice following CHIKV infection was undertaken using RNA-Seq. Quality assessment of the RNA-Seq data was undertaken in this chapter, which demonstrated that the data was of a standard suitable for analysis of transcriptional activity of both the host and virus. The RNA-Seq approach was sensitive enough to detect viral genomes, even at day 7 post infection, 1-2 days after the end of the viraemia period. Read depth, coverage and mapping was of high quality, and was also sufficient to enable host differential gene expression, reliable quantitation of low level gene transcription and identification of novel genes/transcripts, according to the ENCODE Consortium Standards, Guidelines and Best Practices for RNA-Seq v1.0 (ENCODE 2011). Global analysis of host gene expression showed a general trend toward down-regulation, with more genes down- regulated than up-regulated in all samples except day 2 feet. However genes up- regulated >2 fold out-numbered genes down-regulated >2 fold in all samples.

Chapter 5: Analysis of host gene expression following CHIKV infection. In this chapter, induction of host transcriptional activity was shown to be dominated by type I IFN immune responses, with the majority of up-regulated genes being IRGs. Relatively low levels of type I IFN mRNA transcripts were sufficient for this potent induction of IRGs at day 2. At day 7 in the feet, IRGs, including IRF7, remained up-regulated despite the return of type I IFN mRNA transcripts to basal levels. This suggested a type I IFN-independent mechanism of maintaining IRG expression in the feet, following initial type I IFN responses. An exception to this was day 7 lymph node, where genes associated with B- cell activation were the major class.

RNA-Seq analysis also revealed the specific classes of IRGs induced by CHIKV infection, with major up-regulation of granzymes A, B and K and the p47 and p65 GTPases – two classes of genes that have not previously be assessed for their role in the host response to CHIKV infection. Seven novel IRF3 isoforms and a set of novel RNAs were identified. The majority of these novel RNAs were assessed as likely to be non- coding. However one of these, CINT9, was determined to potentially encode a novel protein with a high degree of similarity to Ly6B.

Chapter 6: Analysis of CHIKV RNA in host tissues. The studies in this chapter represent the first deep sequencing analysis of CHIKV RNA in tissues. RNA-Seq analysis 173 of CHIKV RNA was shown to reflect the known kinetics of CHIKV replication, in which virus replicates in the joints (i.e. the feet), and spreads to the draining lymph nodes, with viral titres peaking at day 2 and largely cleared by day 7. More transcripts were produced from the structural region of the genome than the non-structural region. Mutations in the CHIKV genomes accumulated with time, and the accrual of mutations was higher in the lymph nodes than in the feet. The observed difference in mutation accrual may have implications chronic disease, whereby less mutated viral RNA is able to persist long term in the feet, while heavily mutated viral RNA in the lymph nodes accumulates too many mutations, and is thus unable to persist.

7.2 Discussion and future directions

Type I IFN deficiency and DHS/DSS. Mice deficient in type I IFN responses (IRF3/7-/- and IFNAR-/- mice, Chapter 2) developed fatal hypovolemic shock, characterised by severe oedema, oliguria, hypothermia and increased haematocrit, as well as haemorrhagic manifestations; all criteria defined by WHO for hypovolemic shock (W.H.O 2009). This pathology closely resembled the characteristic symptoms of DHF/DSS in humans (Rajapakse 2011, Huy, Van Giang et al. 2013), implicating defective type I IFN responses in the development DHF/DSS and severe CHIKV disease. Infants, known to have immature IRF7 responses (Levy 2007, Danis, George et al. 2008), are more prone to development of DHF/DSS (Guzmán, Kouri et al. 2002), and a study of children and young adults, found IRF7 transcriptional activity was associated with protection against development of DHS/DSS (Hoang, Lynn et al. 2010). Furthermore, DHS/DSS patients have been found to have reduced type I IFN transcript levels (Simmons, Popper et al. 2007). The findings in this thesis provide impetus for investigation of the type I IFNs as a potential indicator for risk of developing DHS/DSS, and as therapeutic agent for DHS/DSS, which affects up 500,000 people annually and results in up to 22,000 deaths, mainly children, each year (W.H.O 2015).

Regulation of IFNγ by type I IFNs. Mortality in type I IFN deficient mice was associated with excessive IFNγ production, as well as IL-6, MCP-1 and TNF (Chapter 2). IFNγ induces production by macrophages of the 3 latter cytokines (Loppnow, Werdan et al. 2008), which promote vasodilation and inflammation, and were likely the drivers of 174 vascular leakage leading to hypovolemic shock in the type I IFN deficient mice (Lee, Liu et al. 2006, Pang, Cardosa et al. 2007). Many of the top 50 up-regulated genes at day 7 in the feet (peak arthritis) were IFNγ-regulated (Chapter 5, Table 5.2), …and furthermore, CHIKV infection of IFNγ-/- mice in our hands does not result in more severe arthritis (Nakaya, Gardner et al. 2012; N. Prow, personal communication). However, it is important to note that Lum, Teo et al. (2013) found foot swelling was increased in IFNγ-/- mice infected with the Asian SGP11 CHIKV strain, perhaps indicating a difference in the role of IFNγ in CHIKV pathobiology between different viral strains. Type I IFNs can inhibit IFNγ responses during infection in other settings (Nguyen, Cousens et al. 2000, Chessler, Caradonna et al. 2011), and type I IFNs are also known to down regulate the IFNγ receptor (Rayamajhi, Humann et al. 2010). These findings suggest anti-IFNγ agents may have utility in the treatment of both DHF/DSS and CHIKV arthritis. Anti-IFNγ antibody was recently shown by our collaborators to ameliorate disease in a mouse model of the related alphavirus, Ross River virus (Mahalingam, 2015, unpublished data, see below).

Identification of the cells producing type I IFNs. Attempts to identify the cellular source of type I IFNs using eGFP reporter mice proved to be unsuccessful (Chapter 3). The lack of eGFP produced by these mice was likely the result of low type I IFN promoter activity. This was later illustrated in Chapter 5, with the finding that only very low levels of type I IFN transcripts were produced in response to CHIKV infection, even at the time of peak serum type I IFN (day 2) in WT mice. In heterozygous IFNα6-GFP+/- and IFNβ- GFP+/- mice, the number of eGFP transcripts would likely be even lower. Despite this, these mice were successfully used for detection of IFNα6 production in response to NDV using unfixed tissue histology and flow cytometry (Kumagai, Takeuchi et al. 2007). The IFNβ-GFP+/- mice have never been reported in a publication (Y.A. Kumagai, personal communication). Identification of the cell types producing type I IFN in response to CHIKV using these mice might be achieved by FACS-based isolation of specific cell types from the individual organs followed by RT-PCR and/or RNA-Seq analysis of type I IFN mRNA

175 levels. However, CHIKV is a PC3 pathogen, and this approach would require access to PC3 flow cytometry facilities, of which only two have recently come into existence in Australia.

Continued IRG production. RNA-Seq analysis of gene expression showed that all type I IFN subtypes had returned to mock infected levels at day 7 in both the feet and lymph nodes (Chapter 5, Figure 5.3). However, analysis of DEGs using the Interferome database, showed almost a quarter of all differentially up-regulated genes in the feet at day 7 post infection are regarded as being regulated by type I IFNs, but not IFNγ or type III IFNs. This suggested a type I IFN-independent mechanism of type I IRG induction. Alternatively, the idea that these genes are exclusively regulated by type I interferon should be reconsidered.

The continued up-regulation of pro-inflammatory IRGs has potential to play a role in the chronic inflammatory arthritis caused by CHIKV. To gain insights into the mechanism/s driving gene induction, promoter reporter constructs could be used to test the responsiveness of the promoter of these up-regulated IRGs to IFNγ. In addition, bioinformatics analysis could be undertaken to identify a common transcription factor binding site/s in the promoters of these genes. Gene silencing and/or knock out mouse models of candidate transcription factors could be used to block potential type I- independent mechanisms of IRG production. RT-PCR assessment of target IRGs downstream of the blocked pathways could be used to confirm the type I IFN-independent mechanism of IRG-induction following CHIKV infection. RNA-Seq analysis during the chronic phase of the disease (i.e. day 30 post infection) would be very useful to assess which IRGs are persistently up-regulated and potentially involved in maintaining chronic disease. In particular, Li, Ding et al. (2013) have identified an IFNβ-induced protein that appears to attenuate the type I IFN response in an IRF3/7 dependant mechanism. Silencing of this gene, activating signal co-integrator complex 3 (Acc3), in HeLa cells resulted in a 20-fold reduction in CHIKV titres due to reduced ISG expression. In light of the role of IRF3/7 in CHIKV pathobiology and the continued ISG production observed in these studies, as well as the persistent nature of CHIKV-induced arthritis/arthralgia, future studies of Ascc3 in CHIKV are warranted.

176 Increasing CHIKV mutation rates as a potential therapy. Site mutations, identified by RNA-Seq, in virus isolated from the tissues accumulated over time and the number of mutations was significantly higher in the lymph nodes than in feet (Chapter 6, Figure 6.4). Analysis of viral titres in Chapter 2 showed that no cytopathic virus could be detected in lymph nodes at day 7 post infection. In contrast, viral titres in the feet at day 7 were still about 5 log10CCID50/ml. Poo, Rudd et al. (2014) also reported recovery of infectious virus from WT mice feet up to day 14 post infection, with viral RNA persisting up to day 100. Together, these findings suggested that CHIKV in the feet may persist due to a lower rate of mutation accrual compared to virus in the lymph nodes. Ribavirin is a nucleoside inhibitor that induces mutations into RNA during RNA virus replication, and can induce lethal mutagenesis, or virus extinction via error catastrophe (Crotty, Cameron et al. 2001, Ortega-Prieto, Sheldon et al. 2013). Persistent viral RNA is thought to contribute to the chronic nature of CHIKV disease (Labadie, Larcher et al. 2010, Suhrbier, Jaffar-Bandjee et al. 2012), therefore Ribavirin might be suggested as an agent for treatment of chronic CHIKV.

The CHIKV genome. While RNA-Seq of CHIKV RNA isolated from the tissues was useful for identification of single nucleotide mutations, there are some limitations to this method of viral genome analysis. Alphaviruses are prone to production of defective interfering (DI) RNA, which contain a range of deletions within the genome (Bangham and Kirkwood 1990, White, Thomson et al. 1998, Thompson, Rempala et al. 2009). RNA isolated from tissues for RNA-Seq is fragmented prior to library preparation and cDNA synthesis. Although the mean length of fragments is around 200 bp, a large range of fragment lengths is produced. This makes it difficult to determine whether 100 bp paired- reads mapping to distant parts of the genome derive from (i) a large cDNA library fragment, or (ii) a short cDNA library fragment from a DI RNA genome in which a large portion of the genome has been deleted.

To investigate deletions in individual CHIKV genomes, and the length of CHIKV genomes produced during replication in the tissues, CHIKV RNA genomes could be isolated using oligo pull-down technologies, cloned and then sequenced. A key question for the future will be to understand why infectious virus cannot be isolated post day 14, despite the continued presence of CHIKV RNA up to day 100 post infection (Poo, Rudd et al. 2014).

177

Novel IRF3 isoforms and CINT9. RNA-Seq analysis identified several novel RNA transcripts up-regulated in response to CHIKV infection, including seven IRF3 isoforms and a potential protein-coding gene related to the Ly6 antigens. Chapter 2 highlighted the important role of IRF3 in CHIKV infection, with IRF3-/- mice developing more severe disease, and IRF3/7-/- mice succumbing to the virus. Both human (Li, Ma et al. 2011) and mouse IRF3 (Zhai, Gao et al. 2008) genes encode ubiquitously expressed splicing variants that inhibit virus-induced, IRF3-mediated type I IFN production. Five of the seven novel variants observed here were also predicted to have an inhibitory effect on IRF3 signalling, due to partial deletion of the IRF3 activation domain. The inhibitory activity of these novel IRF-3 isoforms could be tested by cloning them and recombinantly expressing them in cells, both in isolation and together with full-length IRF3 in different ratios. The transduced cells could then be treated with type I interferon or infected with CHIKV and the induction of key IRGs analysed by RT-PCR.

CINT9 had significant homology with the Ly6 antigens, in particular Ly6B, which was upregulated ≈18 fold in the lymph nodes at day 7 and is expressed on nascent inflammatory macrophages (Rosas, Thomas et al. 2010). CINT9 was located upstream of Ly6C2, which was heavily up-regulated at day 7 in the feet and lymph nodes. Inflammatory monocyte/macrophages that express high levels of Ly6C, also express high levels of CCR2 (Geissmann, Manz et al. 2010), which targets these cells to sites of inflammation (Leuschner, Dutta et al. 2011).

To validate the existence of full-length CINT9, RT-PCR would need to be performed. The product could be then be cloned, transfected into cells to look for biological activity, as for the IRF3 isoforms above. Recombinant protein could also be expressed by bacterial expression systems and used for the production of antibodies. Evidence of CINT9 protein production could then be sought by western blotting of CHIKV infected tissues.

Granzymes A, B & K in CHIKV infection. RNA-Seq analysis in Chapter 5 showed considerable induction of granzymes A and B in the feet, and B and K in the lymph nodes at day 7 post infection. Granzymes are typically associated with induction of apoptosis in infected cells, and are secreted by CD8+ T-cells and NK cells (Domselaar and Bovenschen 2011). These cell types have not previously been shown to play a major role 178 in alphaviral infection (La Linn, Mateo et al. 1998, Smith-Norowitz, Sobel et al. 2000, Alsharifi, Lobigs et al. 2006, Poo, Rudd et al. 2014), thus major up-regulation of granzymes was unexpected. However, CD8+ T-cells have recently been shown to be important for control of Ross River virus in mouse muscle (Burrack, Montgomery et al. 2015), and in humans, strong CD8+ T-cell responses have been reported early in infection (Wauquier, Becquart et al. 2011). Additionally, granzymes have recently been proposed to have extracellular, immuno-modulatory roles in inflammation, possibly by induction and cleavage of secreted cytokines (Wensink, Hack et al. 2015).

Further investigation of granzymes A, B and K in CHIKV infection would first involve confirmation of protein up-regulation in serum and/or tissue extracts by ELISA or western blot. Analysis of protein levels in CHIKV-infected human serum samples would also be useful to establish the relevance of mouse studies to human disease. Mice deficient in granzymes A, B and K are available, and preliminary data indicate a role for these genes in the development of CHIKV-induced arthritic disease.

The p47 and p65 GTPases in CHIKV p47 and p65 GTPases associate with pathogen-containing vacuoles during infection (Shenoy, Kim et al. 2008). The p47 GTPases promote autophagy (Shenoy, Kim et al. 2008, Haldar, Saka et al. 2013), which CHIKV is thought to exploit to delay apoptosis and facilitate replication (Joubert, Werneke et al. 2012). The p47 GTPases also promote acidification of lysosomes (Tiwari and MacMicking 2008), which is required for release of alphavirus genomes from the nucleocapsid (Jose, Snyder et al. 2009, Leung, Ng et al. 2011). Thus p47 GTPases may facilitate viral un-coating and replication. The function of the GBPs is less well understood (Shenoy, Kim et al. 2008), but these proteins are also recruited to PVs, possibly via p47 GTPases (Haldar, Saka et al. 2013), and facilitate delivery of microbicidal peptides (Kim, Shenoy et al. 2011).

While up-regulation of the p65 and p47 GTPases in response to CHIKV has been reported in previous microarray studies (Ryman, White et al. 2002, Xu, Evensen et al. 2015), their function in alphaviral infection has not been investigated. However, several studies have implicated a role for GBPs in viral inhibition in both mice (Carter, Gorbacheva et al. 2005) and humans (Anderson, Carton et al. 1999). In contrast, mice deficient p47 GTPases genes do not alter in their resistance to MCMV (Collazo, Yap et al. 2001) and Ebola viruses (Taylor, Collazo et al. 2000). However, these studies have only looked at

179 single gene knock outs. Given several genes in these classes were expressed together it may be possible that they do not exert activity in isolation. To asses the potential role of these proteins in CHIKV, their presence at the protein level would first need to be confirmed, either by ELISA or western blot. To assess the function of these GTPases in CHIKV infection, mice deficient in relevant proteins could then be examined, with double or triple gene knockouts likely required to produce significant phenotypes.

7.3 Concluding remarks The findings presented in this thesis have further elucidated role of the type I IFNs in the immunobiology of CHIKV. The findings have defined the contribution of key components of type I IFN signalling pathways for induction of the type I IFN response to CHIKV. It was illustrated that type I IFNs protect against lethal haemorrhagic shock; this has implications for our understanding of DHF/DSS. This was the first deep sequencing analysis of the host response to CHIKV infection. Several genes and gene classes not previously associated with CHIKV were identified. These findings have also provided insights into CHIKV genome changes within the host tissues, namely reduced mutational rates in the joint tissues, which may contribute to persistence of viral RNA and arthritic symptoms. Thus, this thesis advances our understanding of the inflammatory processes induced by CHIKV.

180 8. References Abraham, R., P. Mudaliar, A. Padmanabhan and E. Sreekumar (2013). "Induction of cytopathogenicity in human glioblastoma cells by chikungunya virus." PloS One 8(9): e75854.

Aguilar, P. V., S. C. Weaver and C. F. Basler (2007). "Capsid protein of eastern equine encephalitis virus inhibits host cell gene expression." Journal of Virology 81(8): 3866-3876.

Ahola, T. and L. Kääriäinen (1995). "Reaction in alphavirus mRNA capping: formation of a covalent complex of nonstructural protein nsP1 with 7-methyl-GMP." Proceedings of the National Academy of Sciences 92(2): 507-511.

Ahola, T., A. Lampio, P. Auvinen and L. Kääriäinen (1999). "Semliki Forest virus mRNA capping enzyme requires association with anionic membrane phospholipids for activity." The EMBO Journal 18(11): 3164-3172.

Akahata, W., Z. Y. Yang, H. Andersen, S. Sun, H. A. Holdaway, W. P. Kong, M. G. Lewis, S. Higgs, M. G. Rossmann and S. Rao (2010). "A virus-like particle vaccine for epidemic Chikungunya virus protects nonhuman primates against infection." Nature Medicine 16(3): 334-338.

Akbari, O. S., I. Antoshechkin, H. Amrhein, B. Williams, R. Diloreto, J. Sandler and B. A. Hay (2013). "The developmental transcriptome of the mosquito Aedes aegypti, an invasive species and major arbovirus vector." G3: Genes| Genomes| Genetics 3(9): 1493-1509.

Akhrymuk, I., S. V. Kulemzin and E. I. Frolova (2012). "Evasion of the Innate Immune Response: the Old World Alphavirus nsP2 Protein Induces Rapid Degradation of Rpb1, a Catalytic Subunit of RNA Polymerase II." Journal of Virology 86(13): 7180-7191.

Akira, S., S. Uematsu and O. Takeuchi (2006). "Pathogen recognition and innate immunity." Cell 124(4): 783-801.

Allton, K., A. K. Jain, H.-M. Herz, W.-W. Tsai, S. Y. Jung, J. Qin, A. Bergmann, R. L. Johnson and M. C. Barton (2009). "Trim24 targets endogenous p53 for degradation." Proceedings of the National Academy of Sciences 106(28): 11612-11616.

AlSaleh, A., T. A. Sanders and S. D. O'Dell (2012). "Effect of interaction between PPARG, PPARA and ADIPOQ gene variants and dietary fatty acids on plasma lipid profile and adiponectin concentration in a large intervention study." Proceedings of the Nutrition Society 71(01): 141-153. 181 Alsharifi, M., M. Lobigs, M. M. Simon, A. Kersten, K. Müller, A. Koskinen, E. Lee and A. Müllbacher (2006). "NK cell-mediated immunopathology during an acute viral infection of the CNS." European Journal of Immunology 36(4): 887-896.

Ambros, V. (2004). "The functions of animal microRNAs." Nature 431(7006): 350-355.

Anderson, P. (2010). "Post-transcriptional regulons coordinate the initiation and resolution of inflammation." Nature Reviews Immunology 10(1): 24-35.

Anderson, S. L., J. M. Carton, J. Lou, L. Xing and B. Y. Rubin (1999). "Interferon-induced guanylate binding protein-1 (GBP-1) mediates an antiviral effect against vesicular stomatitis virus and encephalomyocarditis virus." Virology 256(1): 8-14.

Andrews, H. N., P. B. Mullan, S. McWilliams, S. Sebelova, J. E. Quinn, P. M. Gilmore, N. McCabe, A. Pace, B. Koller and P. G. Johnston (2002). "BRCA1 regulates the interferon γ- mediated apoptotic response." Journal of Biological Chemistry 277(29): 26225-26232.

Andrews, S. (2010). "FastQC: A quality control tool for high throughput sequence data." http://www.bioinformatics.babraham.ac.uk/projects/fastqc/.

Apostolou, E. and D. Thanos (2008). "Virus Infection Induces NF-kappaB-dependent interchromosomal associations mediating monoallelic IFN-beta gene expression." Cell 134(1): 85-96.

Atasheva, S., M. Akhrymuk, E. I. Frolova and I. Frolov (2012). "New PARP gene with an anti-alphavirus function." Journal of Virology 86(15): 8147-8160.

Atasheva, S., A. Fish, M. Fornerod and E. I. Frolova (2010). "Venezuelan equine encephalitis virus capsid protein forms a tetrameric complex with CRM1 and importin α/β that obstructs nuclear pore complex function." Journal of Virology 84(9): 4158-4171.

Atkinson, T., A. Barrett, A. Mackenzie and N. Dimmock (1986). "Persistence of virulent Semliki Forest virus in mouse brain following co-inoculation with defective interfering particles." Journal of General Virology 67(6): 1189-1194.

Balachandran, S. and A. A. Beg (2011). "Defining Emerging Roles for NF-κB in Antivirus Responses: Revisiting the Interferon-β Enhanceosome Paradigm." PLoS Pathogens 7(10): e1002165.

182 Bamezai, A. (2004). "Mouse Ly-6 proteins and their extended family: markers of cell differentiation and regulators of cell signaling." ARCHIVUM IMMUNOLOGIAE ET THERAPIAE EXPERIMENTALIS-ENGLISH EDITION- 52(4): 255-266.

Bangham, C. M. and T. B. Kirkwood (1990). "Defective interfering particles: effects in modulating virus growth and persistence." Virology 179(2): 821-826.

Barry, G., L. Breakwell, R. Fragkoudis, G. Attarzadeh-Yazdi, J. Rodriguez-Andres, A. Kohl and J. K. Fazakerley (2009). "PKR acts early in infection to suppress Semliki Forest virus production and strongly enhances the type I interferon response." Journal of General Virology 90(6): 1382-1391.

Basler, C. F. and A. García-Sastre (2007). "Sensing RNA virus infections." Nature Chemical Biology 3(1): 20-21.

Bauer, S., S. j. Pigisch, D. Hangel, A. Kaufmann and S. Hamm (2008). "Recognition of nucleic acid and nucleic acid analogs by Toll-like receptors 7, 8 and 9." Immunobiology 213(3-4): 315-328.

Benedict, M. Q., R. S. Levine, W. A. Hawley and L. P. Lounibos (2007). "Spread of the tiger: global risk of invasion by the mosquito Aedes albopictus." Vector-borne and Zoonotic Diseases 7(1): 76-85.

Bick, M. J., J. W. N. Carroll, G. Gao, S. P. Goff, C. M. Rice and M. R. MacDonald (2003). "Expression of the zinc-finger antiviral protein inhibits alphavirus replication." Journal of Virology 77(21): 11555-11562.

Bidwell, B. N., C. Y. Slaney, N. P. Withana, S. Forster, Y. Cao, S. Loi, D. Andrews, T. Mikeska, N. E. Mangan and S. A. Samarajiwa (2012). "Silencing of Irf7 pathways in breast cancer cells promotes bone metastasis through immune escape." Nature Medicine 18(8): 1224-1231.

Bieback, K., E. Lien, I. M. Klagge, E. Avota, J. Schneider-Schaulies, W. P. Duprex, H. Wagner, C. J. Kirschning, V. Ter Meulen and S. Schneider-Schaulies (2002). "Hemagglutinin protein of wild-type measles virus activates toll-like receptor 2 signaling." Journal of Virology 76(17): 8729-8736.

Billinton, N. and A. W. Knight (2001). "Seeing the wood through the trees: A review of techniques for distinguishing green flourescent protein from endogenous autofluoresence " Analytical Biochemistry 291: 715-197.

183 Blasius, A. L. and B. Beutler (2010). "Intracellular toll-like receptors." Immunity 32(3): 305- 315.

Bouquillard, É. and B. Combe (2009). "A report of 21 cases of rheumatoid arthritis following Chikungunya fever. A mean follow-up of two years." Joint Bone Spine 76(6): 654- 657.

Boyd, J. H. (2012). "Toll-Like Receptors and Opportunities for New Sepsis Therapeutics." Current Infectious Disease Reports 14(3): 1-7.

Bracci, L., V. La Sorsa, F. Bellardelli and E. Proietti (2008). "Type I interferons as vaccine adjuvants against infectious diseases and cancer." Expert Review of Infectious Diseases and cancer 7: 373-381.

Bratke, K., A. Klug, P. Julius, M. Kuepper, M. Lommatzsch, G. Sparmann, W. Luttmann and J. C. Virchow (2008). "Granzyme K: a novel mediator in acute airway inflammation." Thorax 63(11): 1006-1011.

Brazelton, T. R. and H. M. Blau (2005). "Optimizing techniques for tracking transplanted stem cells in vivo." Stem Cells 23(9): 1251-1265.

Breakwell, L., P. Dosenovic, G. B. Karlsson Hedestam, M. D'Amato, P. Liljeström, J. Fazakerley and G. M. McInerney (2007). "Semliki Forest virus nonstructural protein 2 is involved in suppression of the type I interferon response." Journal of Virology 81(16): 8677-8684.

Bréhin, A. C., I. Casadémont, M. P. Frenkiel, C. Julier, A. Sakuntabhai and P. Desprès (2009). "The large form of human 2', 5'-Oligoadenylate Synthetase (OAS3) exerts antiviral effect against Chikungunya virus." Virology 384(1): 216-222.

Brierley, M. M. and E. N. Fish (2002). "Review: IFN-α/β receptor interactions to biologic outcomes: understanding the circuitry." Journal of Interferon & Cytokine Research 22(8): 835-845.

Brighton, S. (1984). "Chloroquine phosphate treatment of chronic chikungunya arthritis." South Africa Medicine Journal 11: 217-218.

Briolat, V., L. Jouneau, R. Carvalho, N. Palha, C. Langevin, P. Herbomel, O. Schwartz, H. P. Spaink, J.-P. Levraud and P. Boudinot (2014). "Contrasted innate responses to two

184 viruses in zebrafish: insights into the ancestral repertoire of vertebrate IFN-stimulated genes." The Journal of Immunology 192(9): 4328-4341.

Buckley, N. E., A. M. Hosey, J. J. Gorski, J. W. Purcell, J. M. Mulligan, D. P. Harkin and P. B. Mullan (2007). "BRCA1 regulates IFN-γ signaling through a mechanism involving the type I IFNs." Molecular Cancer Research 5(3): 261-270.

Burrack, K. S., S. A. Montgomery, D. Homann and T. E. Morrison (2015). "CD8+ T Cells Control Ross River Virus Infection in Musculoskeletal Tissues of Infected Mice." The Journal of Immunology 194(2): 678-689.

Burt, F. J., M. S. Rolph, N. E. Rulli, S. Mahalingam and M. T. Heise (2012). "Chikungunya: a re-emerging virus." The Lancet 379(9816): 662-671.

Caceres, G., X. Y. Zhu, J.-a. Jiao, R. Zankina, A. Aller and P. Andreotti (2003). "Imaging of luciferase and GFP-transfected human tumours in nude mice." Luminescence 18(4): 218- 223.

Cao, J., E. B. Yang, J. J. Su, Y. Li and P. Chow (2003). "The tree shrews: adjuncts and alternatives to primates as models for biomedical research." Journal of Medical Primatology 32(3): 123-130.

Carter, C., V. Gorbacheva and D. Vestal (2005). "Inhibition of VSV and EMCV replication by the interferon-induced GTPase, mGBP-2: differential requirement for wild-type GTP binding domain." Archives of Virology 150(6): 1213-1220.

Chaaitanya, I. K., N. Muruganandam, S. G. Sundaram, O. Kawalekar, A. P. Sugunan, S. P. Manimunda, S. R. Ghosal, K. Muthumani and P. Vijayachari (2011). "Role of Proinflammatory Cytokines and Chemokines in Chronic Arthropathy in CHIKV Infection." Viral Immunology 24(4): 265-271.

Campion, E. W., S. C. Weaver and M. Lecuit (2015). "Chikungunya virus and the global spread of a mosquito-borne disease." New England Journal of Medicine 372(13): 1231- 1239.

Chan, Y. L., T. H. Chang, C. L. Liao and Y. L. Lin (2008). "The cellular antiviral protein viperin is attenuated by proteasome-mediated protein degradation in Japanese encephalitis virus-infected cells." Journal of Virology 82(21): 10455-10464.

185 Chang, M., B. Collet, P. Nie, K. Lester, S. Campbell, C. J. Secombes and J. Zou (2011). "Expression and functional characterization of the RIG-I-like receptors MDA5 and LGP2 in rainbow trout (Oncorhynchus mykiss)." Journal of Virology 85(16): 8403-8412.

Chang, L.-J., K. A. Dowd, F. H. Mendoza, J. G. Saunders, S. Sitar, S. H. Plummer, G. Yamshchikov, U. N. Sarwar, Z. Hu and M. E. Enama (2014). "Safety and tolerability of chikungunya virus-like particle vaccine in healthy adults: a phase 1 dose-escalation trial." The Lancet 384(9959): 2046-2052.

Chen, K. C., Y.-W. Kam, R. Lin, M. Ng, L. Ng and J. J. H. Chu (2013). "Comparative analysis of the genome sequences and replication profiles of chikungunya virus isolates within the East, Central and South African (ECSA) lineage." Virology Journal 10: 169.

Chen, R., E. Wang, K. A. Tsetsarkin and S. C. Weaver (2013). "Chikungunya virus 3 ′ untranslated region: adaptation to mosquitoes and a population bottleneck as major evolutionary forces." PLoS Pathogens 9(8): e1003591.

Cheon, H., E. G. Holvey‐Bates, J. W. Schoggins, S. Forster, P. Hertzog, N. Imanaka, C. M. Rice, M. W. Jackson, D. J. Junk and G. R. Stark (2013). "IFNβ-dependent increases in STAT1, STAT2, and IRF9 mediate resistance to viruses and DNA damage." The EMBO Journal 32(20): 2751-2763.

Cherayil, B. J. and S. Pillai (1991). "The omega/lambda 5 surrogate immunoglobulin light chain is expressed on the surface of transitional B in murine bone marrow." The Journal of Experimental Medicine 173(1): 111-116.

Chessler, A.-D. C., K. L. Caradonna, A. Da'dara and B. A. Burleigh (2011). "Type I interferons increase host susceptibility to Trypanosoma cruzi infection." Infection and Immunity 79(5): 2112-2119.

Ching, N. L. and H. C. Hapuarachchi (2010). "Tracing the path of Chikungunya virus- Evolution and adaptation." Infection, Genetics and Evolution 10: 876-885.

Chow, A., Z. Her, E. K. Ong, J.-m. Chen, F. Dimatatac, D. J. Kwek, T. Barkham, H. Yang, L. Rénia and Y.-S. Leo (2011). "Persistent arthralgia induced by Chikungunya virus infection is associated with interleukin-6 and granulocyte macrophage colony-stimulating factor." Journal of Infectious Diseases 203(2): 149-157.

186 Choy, G., S. O Connor, F. E. Diehn, N. Costouros, H. R. Alexander, P. Choyke and S. K. Libutti (2003). "Comparison of noninvasive fluorescent and bioluminescent small animal optical imaging." Biotechniques 35(5): 1022-1031.

Chudakov, D. M., M. V. Matz, S. Lukyanov and K. A. Lukyanov (2010). "Fluorescent proteins and their applications in imaging living cells and tissues." Physiological Reviews 90(3): 1103-1163.

Coffey, L. L., Y. Beeharry, A. V. Bordería, H. Blanc and M. Vignuzzi (2011). "Arbovirus high fidelity variant loses fitness in mosquitoes and mice." Proceedings of the National Academy of Sciences 108(38): 16038-16043.

Collazo, C. M., G. S. Yap, G. D. Sempowski, K. C. Lusby, L. Tessarollo, G. F. V. Woude, A. Sher and G. A. Taylor (2001). "Inactivation of LRG-47 and IRG-47 reveals a family of interferon γ–inducible genes with essential, pathogen-specific roles in resistance to infection." The Journal of Experimental Medicine 194(2): 181-188.

Corder, G. W. and D. I. Foreman (2011). Nonparametric Statistics for Non-Statisticians: A Step-by-Step Approach, John Wiley & Sons.

Costa, R. H. (2005). "FoxM1 dances with mitosis." Nature Cell Biology 7(2): 108-110.

Couderc, T., F. Chrétien, C. Schilte, O. Disson, M. Brigitte, F. Guivel-Benhassine, Y. Touret, G. Barau, N. Cayet, I. Schuffenecker, P. Desprès, F. Arenzana-Seisdedos, A. Michault, M. L. Albert and M. Lecuit (2008). "A Mouse Model for Chikungunya: Young Age and Inefficient Type-I Interferon Signaling Are Risk Factors for Severe Disease." PLoS Pathogens 4(2): e29.

Couderc, T., N. Khandoudi, M. Grandadam, C. Visse, N. Gangneux, S. Bagot, J. F. Prost and M. Lecuit (2009). "Prophylaxis and therapy for Chikungunya virus infection." Journal of Infectious Diseases 200(4): 516-523.

Couderc, T. and M. Lecuit (2009). "Focus on Chikungunya pathophysiology in human and animal models." Microbes and Infection 11(14-15): 1197-1205.

Couturier, E., F. Guillemin, M. Mura, L. Léon, J. M. Virion, M. J. Letort, H. De Valk, F. Simon and V. Vaillant (2012). "Impaired quality of life after chikungunya virus infection: a 2-year follow-up study." Rheumatology 51(7): 1315-1322.

187 Crotty, S., C. E. Cameron and R. Andino (2001). "RNA virus error catastrophe: direct molecular test by using ribavirin." Proceedings of the National Academy of Sciences 98(12): 6895-6900.

D'Eustachio, P., D. Pravtcheva, K. Marcu and F. Ruddle (1980). "Chromosomal location of the structural gene cluster encoding murine immunoglobulin heavy chains." The Journal of Experimental Medicine 151(6): 1545-1550.

Daffis, S., M. S. Suthar, K. J. Szretter, M. Gale and M. S. Diamond (2009). "Induction of IFN-β and the innate antiviral response in myeloid cells occurs through an IPS-1- dependent signal that does not require IRF-3 and IRF-7." PLoS Pathogens 5(10): e1000607.

Danis, B., T. C. George, S. Goriely, B. Dutta, J. Renneson, L. Gatto, P. Fitzgerald- Bocarsly, A. Marchant, M. Goldman and F. Willems (2008). "Interferon regulatory factor 7- mediated responses are defective in cord blood plasmacytoid dendritic cells." European Journal of Immunology 38(2): 507-517.

Darwin, J. R., J. L. Kenney and S. C. Weaver (2011). "Transmission Potential of Two Chimeric Chikungunya Vaccine Candidates in the Urban Mosquito Vectors, Aedes aegypti and Ae. albopictus." The American Journal of Tropical Medicine and Hygiene 84(6): 1012- 1015.

Das, T., M. C. Jaffar-Bandjee, J. J. Hoarau, P. Krejbich Trotot, M. Denizot, G. Lee-Pat- Yuen, R. Sahoo, P. Guiraud, D. Ramful and S. Robin (2010). "Chikungunya fever: CNS infection and pathologies of a re-emerging arbovirus." Progress in Neurobiology 91(2): 121-129.

De Andrade, D., S. Jean, P. Clavelou, R. Dallel and D. Bouhassira (2010). "Chronic pain associated with the Chikungunya Fever: long lasting burden of an acute illness." BMC Infectious Diseases 10(1): 31.

De, I., C. Fata-Hartley, S. G. Sawicki and D. L. Sawicki (2003). "Functional analysis of nsP3 phosphoprotein mutants of Sindbis virus." Journal of Virology 77(24): 13106-13116.

Diebold, S. S., T. Kaisho, H. Hemmi, S. Akira and C. Reis e Sousa (2004). "Innate antiviral responses by means of TLR7-mediated recognition of single-stranded RNA." Science's STKE 303(5663): 1529-1531.

188 Diwan, A., A. G. Koesters, D. Capella, H. Geiger, T. A. Kalfa and G. W. Dorn II (2008). "Targeting erythroblast-specific apoptosis in experimental anemia." Apoptosis 13(8): 1022- 1030.

Dölken, L., Z. Ruzsics, B. Rädle, C. C. Friedel, R. Zimmer, J. Mages, R. Hoffmann, P. Dickinson, T. Forster and P. Ghazal (2008). "High-resolution gene expression profiling for simultaneous kinetic parameter analysis of RNA synthesis and decay." Rna 14(9): 1959- 1972.

Domselaar, R. and N. Bovenschen (2011). "Cell death-independent functions of granzymes: hit viruses where it hurts." Reviews in Medical Virology 21(5): 301-314.

Donaszi-Ivanov, A., I. Mohorianu, T. Dalmay and P. P. Powell (2013). "Small RNA analysis in Sindbis virus infected human HEK293 cells." PloS One 8(12): e84070.

Doyle, S. L. and L. A. J. O’Neill (2006). "Toll-like receptors: from the discovery of NFκB to new insights into transcriptional regulations in innate immunity." Biochemical Pharmacology 72(9): 1102-1113.

Dupont-Rouzeyrol, M., V. Caro, L. Guillaumot, M. Vazeille, E. D'Ortenzio, J.-M. Thiberge, N. Baroux, A.-C. Gourinat, M. Grandadam and A.-B. Failloux (2012). "Chikungunya virus and the mosquito vector Aedes aegypti in New Caledonia (South Pacific Region)." Vector- Borne and Zoonotic Diseases 12(12): 1036-1041.

Eckels, K. H., V. R. Harrison and F. M. Hetrick (1970). "Chikungunya virus vaccine prepared by Tween-ether extraction." Applied Microbiology 19(2): 321-325.

Economopoulou, A., M. Dominguez, B. Helynck, D. Sissoko, O. Wichmann, P. Quenel, P. Germonneau and I. Quatresous (2009). "Atypical Chikungunya virus infections: clinical manifestations, mortality and risk factors for severe disease during the 2005–2006 outbreak on Reunion." Epidemiology and Infection 137(04): 534-541.

Edelman, R., C. Tacket, S. Wasserman, S. Bodison, J. Perry and J. Mangiafico (2000). "Phase II safety and immunogenicity study of live chikungunya virus vaccine TSI-GSD- 218." The American journal of tropical medicine and hygiene 62(6): 681-685.

Edwards, C. J., S. R. Welch, J. Chamberlain, R. Hewson, H. Tolley, P. A. Cane and G. Lloyd (2007). "Molecular diagnosis and analysis of Chikungunya virus." Journal of Clinical Virology 39(4): 271-275.

189 Emilie, J., T. H. Tiong, L.-G. Isabelle, S. Hélène and S. Fabrice (2014). "Inflammation of the external ear in acute chikungunya infection: Experience from the outbreak in Johor Bahru, Malaysia, 2008." Journal of Clinical Virology.

ENCODE (2011). Standards, Guidelines and Best Practices for RNA-Seq v1. 0. https://genome.ucsc.edu/ENCODE/protocols/dataStandards/ENCODE_RNAseq_Standard s_V1.0.pdf, National Research Institute.

Esen, N., P. K. Blakely, E. K. Rainey-Barger and D. N. Irani (2012). "Complexity of the microglial activation pathways that drive innate host responses during lethal alphavirus encephalitis in mice." ASN Neuro 4(4): AN20120016.

Faragher, S. G., A. D. J. Meek, C. M. Rice and L. Dalgarno (1988). "Genome SequenceGenome Sequence of a Mouse-Avirulent and a Mouse Virulent Strain of Ross River Virus of a Mouse-Avirulent and a Mouse Virulent Strain of Ross River Virus." Virology 163: 509-526.

Farkas, Á. and L. Kemény (2011). "Interferon-α in the generation of monocyte-derived dendritic cells: recent advances and implications for dermatology." British Journal of Dermatology 165(2): 247-254.

Fauquet, C., M. A. Mayo, J. Maniloff, U. Desselberger and L. A. Ball, Eds. (2005). Virus taxonomy: classification and nomenclature of viruses: eighth report of the International Committee on the Taxonomy of Viruses. Amsterdam, Elsevier Academic Press.

Finberg, R. W. and E. A. Kurt-Jones (2004). "Viruses and Toll-like receptors." Microbes and Infection 6(15): 1356-1360.

Fink, K., K. S. Lang, N. Manjarrez‐Orduno, T. Junt, B. M. Senn, M. Holdener, S. Akira, R. M. Zinkernagel and H. Hengartner (2006). "Early type I interferon-mediated signals on B cells specifically enhance antiviral humoral responses." European Journal of Immunology 36(8): 2094-2105.

Firth, A. E., S. Atasheva, E. I. Frolova and I. Frolov (2011). "Conservation of a packaging signal and the viral genome RNA packaging mechanism in alphavirus evolution." Journal of Virology 85(16): 8022-8036.

Forrester, N., M. Guerbois, A. Adams, X. Liang and S. Weaver (2011). "Analysis of intrahost variation in Venezuelan equine encephalitis virus reveals repeated deletions in the 6-kilodalton protein gene." Journal of Virology 85(17): 8709-8717. 190 Fric, J., S. Bertin-Maghit, C.-I. Wang, A. Nardin and L. Warter (2013). "Use of human monoclonal antibodies to treat Chikungunya virus infection." Journal of Infectious Diseases 207(2): 319-322.

Frolov, I., M. Akhrymuk, I. Akhrymuk, S. Atasheva and E. I. Frolova (2012). "Early events in alphavirus replication determine the outcome of infection." Journal of Virology 86(9): 5055-5066.

Frolov, I., R. Hardy and C. M. Rice (2001). "Cis-acting RNA elements at the 5'end of Sindbis virus genome RNA regulate minus-and plus-strand RNA synthesis." RNA 7(11): 1638-1651.

Fros, J. J., W. J. Liu, N. A. Prow, C. Geertsema, M. Ligtenberg, D. L. Vanlandingham, E. Schnettler, J. M. Vlak, A. Suhrbier and A. A. Khromykh (2010). "Chikungunya virus nonstructural protein 2 inhibits type I/II interferon-stimulated JAK-STAT signaling." Journal of Virology 84(20): 10877-10887.

Fros, J. J., L. D. Major, F. E. Scholte, J. Gardner, M. J. van Hemert, A. Suhrbier and G. P. Pijlman (2014). "Chikungunya virus nsP2-mediated host shut-off disables the unfolded protein response." Journal of General Virology: vir. 0.071845-071840.

Fung, K. Y., N. E. Mangan, H. Cumming, J. C. Horvat, J. R. Mayall, S. A. Stifter, N. De Weerd, L. C. Roisman, J. Rossjohn and S. A. Robertson (2013). "Interferon-ε protects the female reproductive tract from viral and bacterial infection." Science 339(6123): 1088- 1092.

Fung, M., S. Sia, K. Leung and N. Mak (2004). "Detection of differential expression of mouse interferon-alpha subtypes by polymerase chain reaction using specific primers." Journal of Immunological Methods 284(1): 177-186.

Gaedigk-Nitschko, K. and M. J. Schlesinger (1990). "The Sindbis virus 6K protein can be detected in virions and is acylated with fatty acids." Virology 175(1): 274-281.

Ganu, M. A. and A. Ganu (2011). "Post-Chikungunya Chronic Arthritis–Our Experience with DMARDs Over Two Year Follow Up." JAPI 59: 83-86.

Gao, S., A. von der Malsburg, A. Dick, K. Faelber, G. F. Schröder, O. Haller, G. Kochs and O. Daumke (2011). "Structure of myxovirus resistance protein a reveals intra-and intermolecular domain interactions required for the antiviral function." Immunity 35(4): 514- 525.

191 Garber, M., M. G. Grabherr, M. Guttman and C. Trapnell (2011). "Computational methods for transcriptome annotation and quantification using RNA-seq." Nature Methods 8(6): 469-477.

Garcia, M., J. Gil, I. Ventoso, S. Guerra, E. Domingo, C. Rivas and M. Esteban (2006). "Impact of protein kinase PKR in cell biology: from antiviral to antiproliferative action." Microbiology and Molecular Biology Reviews 70(4): 1032-1060.

Gardner, J., I. Anraku, T. T. Le, T. Larcher, L. Major, P. Roques, W. A. Schroder, S. Higgs and A. Suhrbier (2010). "Chikungunya virus arthritis in adult wild-type mice." Journal of Virology 84(16): 8021.

Gareau, D. S., P. R. Bargo, W. A. Horton and S. L. Jacques (2004). "Confocal fluorescence spectroscopy of subcutaneous cartilage expressing green fluorescent protein versus cutaneous collagen autofluorescence." Journal of Biomedical Optics 9(2): 254-258.

Garmashova, N., S. Atasheva, W. Kang, S. C. Weaver, E. Frolova and I. Frolov (2007). "Analysis of Venezuelan equine encephalitis virus capsid protein function in the inhibition of cellular transcription." Journal of Virology 81(24): 13552-13565.

Geissmann, F., M. G. Manz, S. Jung, M. H. Sieweke, M. Merad and K. Ley (2010). "Development of monocytes, macrophages, and dendritic cells." Science 327(5966): 656- 661.

Gérardin, P., G. Barau, A. Michault, M. Bintner, H. Randrianaivo, G. Choker, Y. Lenglet, Y. Touret, A. Bouveret, P. Grivard, K. L. Roux, S. Blanc, I. Schuffenecker, T. Couderc, F. Arenzana-Seisdedos, M. Lecuit and P.-Y. Robillard (2008). "Multidisciplinary Prospective Study of Mother-to-Child Chikungunya Virus Infections on the Island of La Réunion." PLoS Medicine 5(3): 0413-0423.

Gérardin, P., A. Fianu, D. Malvy, C. Mussard, K. Boussaïd, O. Rollot, A. Michault, B. A. Gauzere, G. Bréart and F. Favier (2011). "Perceived morbidity and community burden after a Chikungunya outbreak: the TELECHIK survey, a population-based cohort study." BMC Medicine 9(1): 5-16.

Gerson, C. J., S. Goldstein and A. E. Heacox (2009). "Retained structural integrity of collagen and elastin within cryopreserved human heart valve tissue as detected by two- photon laser scanning confocal microscopy." Cryobiology 59(2): 171-179.

192 Ghendon, Y. and G. Samoilova (1968). "Antiviral effect of acetone." Journal of General Virology 3(2): 271-273.

Glenn, T. C. (2011). "Field guide to next-generation DNA sequencers." Molecular Ecology Resources 11(5): 759-769.

Goh, L. Y., Y.-W. Kam, S. W. Metz, J. Hobson-Peters, N. A. Prow, S. McCarthy, D. W. Smith, G. P. Pijlman, L. F. Ng and R. A. Hall (2015). "A sensitive epitope-blocking ELISA for the detection of Chikungunya virus-specific antibodies in patients." Journal of Virological Methods 222: 55-61.

Gomez de Cedrón, M., N. Ehsani, M. L. Mikkola, J. A. García and L. Kääriäinen (1999). "RNA helicase activity of Semliki Forest virus replicase protein NSP2." FEBS Letters 448(1): 19-22.

Gomi, K., M. Morimoto and N. Nakamizo (1985). "Characteristics of antiviral and anticellular activities of human recombinant interferon-gamma." Japanese Journal of Cancer Research: Gann 76(3): 224-234.

Goodstadt, L. and C. P. Ponting (2001). "CHROMA: consensus-based colouring of multiple alignments for publication." Bioinformatics 17(9): 845-846.

Gorchakov, R., R. Hardy, C. M. Rice and I. Frolov (2004). "Selection of functional 5′ cis- acting elements promoting efficient Sindbis virus genome replication." Journal of Virology 78(1): 61-75.

Gould, E., B. Coutard, H. Malet, B. Morin, S. Jamal, S. Weaver, A. Gorbalenya, G. Moureau, C. Baronti and I. Delogu (2010). "Understanding the alphaviruses: Recent research on important emerging pathogens and progress towards their control." Antiviral Research 87(2): 111-124.

Gould, E. A., P. Gallian, X. De Lamballerie and R. N. Charrel (2010). "First cases of autochthonous dengue fever and chikungunya fever in France: from bad dream to reality!" Clinical Microbiology and Infection 16(12): 1702-1704.

Gould, E. A. and S. Higgs (2009). "Impact of climate change and other factors on emerging arbovirus diseases." Transactions of the Royal Society of Tropical Medicine and Hygiene 103(2): 109-121.

193 Guzmán, M. G., G. Kouri, J. Bravo, L. Valdes, V. Susana and S. B. Halstead (2002). "Effect of age on outcome of secondary dengue 2 infections." International Journal of Infectious Diseases 6(2): 118-124.

Haldar, A. K., H. A. Saka, A. S. Piro, J. D. Dunn, S. C. Henry, G. A. Taylor, E. M. Frickel, R. H. Valdivia and J. Coers (2013). "IRG and GBP host resistance factors target aberrant,“non-self” vacuoles characterized by the missing of “self” IRGM proteins." PLoS Pathogens 9(6): e1003414.

Haller, O. and G. Kochs (2011). "Human MxA protein: an interferon-induced dynamin-like GTPase with broad antiviral activity." Journal of interferon & cytokine research 31(1): 79- 87.

Hamilton, M. A., R. C. Russo and R. V. Thurston (1977). "Trimmed Spearman-Karber method for estimating median lethal concentrations in toxicity bioassays." Environmental Science & Technology 11(7): 714-719.

Hao, S. and D. Baltimore (2009). "The stability of mRNA influences the temporal order of the induction of genes encoding inflammatory molecules." Nature Immunology 10(3): 281- 288.

Hardy, R. W. and C. M. Rice (2005). "Requirements at the 3'end of the sindbis virus genome for efficient synthesis of minus-strand RNA." Journal of Virology 79(8): 4630- 4639.

Harman, A. N., J. Lai, S. Turville, S. Samarajiwa, L. Gray, V. Marsden, S. K. Mercier, K. Jones, N. Nasr and A. Rustagi (2011). "HIV infection of dendritic cells subverts the IFN induction pathway via IRF-1 and inhibits type 1 IFN production." Blood 118(2): 298-308.

Harrison, V., L. Binn and R. Randall (1967). "Comparative immunogenicities of chikungunya vaccines prepared in avian and mammalian tissues." The American Journal of Tropical Medicine and Hygiene 16(6): 786-791.

Hart, T., H. K. Komori, S. LaMere, K. Podshivalova and D. R. Salomon (2013). "Finding the active genes in deep RNA-seq gene expression studies." BMC Genomics 14(1): 778.

Heil, F., H. Hemmi, H. Hochrein, F. Ampenberger, C. Kirschning, S. Akira, G. Lipford, H. Wagner and S. Bauer (2004). "Species-specific recognition of single-stranded RNA via toll- like receptor 7 and 8." Science's STKE 303(5663): 1526-1529.

194 Her, Z., B. Malleret, M. Chan, E. K. S. Ong, S. C. Wong, D. J. C. Kwek, H. Tolou, R. T. P. Lin, P. A. Tambyah and L. Rénia (2010). "Active infection of human blood monocytes by Chikungunya virus triggers an innate immune response." The Journal of Immunology 184(10): 5903-5913.

Her, Z., T. S. Teng, J. J. Tan, T. H. Teo, Y. W. Kam, F. M. Lum, W. W. Lee, C. Gabriel, R. Melchiotti and A. K. Andiappan (2014). "Loss of TLR3 aggravates CHIKV replication and pathology due to an altered virus‐specific neutralizing antibody response." EMBO Molecular Medicine: e201404459.

Heruth, D. P., M. Gibson, D. N. Grigoryev, L. Q. Zhang and S. Q. Ye (2012). "RNA-seq analysis of synovial fibroblasts brings new insights into rheumatoid arthritis." Cell and Bioscience 2(1): 43.

Heward, J. A. and M. A. Lindsay (2014). "Long non-coding RNAs in the regulation of the immune response." Trends in Immunology 35(9): 408-419.

Hoang, L. T., D. J. Lynn, M. Henn, B. W. Birren, N. J. Lennon, P. T. Le, K. T. H. Duong, T. T. H. Nguyen, L. N. Mai and J. J. Farrar (2010). "The early whole-blood transcriptional signature of dengue virus and features associated with progression to dengue shock syndrome in Vietnamese children and young adults." Journal of Virology 84(24): 12982- 12994.

Hoarau, J.-J., M.-C. Jaffar Bandjee, P. Krejbich Trotot, T. Das, G. Li-Pat-Yuen, B. Dassa, M. Denizot, E. Guichard, A. Ribera, T. Henni, F. Tallet, M. P. Moiton, B. A. Gauzère, S. Bruniquet, Z. Jaffar Bandjee, P. Morbidelli, G. Martigny, M. Jolivet, F. Gay, M. Grandadam, H. Tolou, V. Vieillard, P. Debré, B. Autran and P. Gasque (2010). "Persistent Chronic Inflammation and Infection by Chikungunya Arthritogenic Alphavirus in Spite of a Robust Host Immune Response." The Journal of Immunology 184(10): 5914-5927.

Holland, J., K. Spindler, F. Horodyski, E. Grabau, S. Nichol and S. VandePol (1982). "Rapid evolution of RNA genomes." Science 215(4540): 1577-1585.

Holzer, G. W., S. Coulibaly, G. Aichinger, H. Savidis-Dacho, J. Mayrhofer, S. Brunner, K. Schmid, O. Kistner, J. G. Aaskov and F. G. Falkner (2011). "Evaluation of an inactivated Ross River virus vaccine in active and passive mouse immunization models and establishment of a correlate of protection." Vaccine 29(24): 4132-4141.

195 Honda, K., A. Takaoka and T. Taniguchi (2006). "Type I inteferon gene induction by the interferon regulatory factor family of transcription factors." Immunity 25(3): 349-360.

Honda, K., H. Yanai, H. Negishi, M. Asagiri, M. Sato, T. Mizutani, N. Shimada, Y. Ohba, A. Takaoka and N. Yoshida (2005). "IRF-7 is the master regulator of type-I interferon- dependent immune responses." Nature 434(7034): 772-777.

Hu, J., S. C. Sealfon, F. Hayot, C. Jayaprakash, M. Kumar, A. C. Pendleton, A. Ganee, A. Fernandez-Sesma, T. M. Moran and J. G. Wetmur (2007). "Chromosome-specific and noisy IFNB1 transcription in individual virus-infected human primary dendritic cells." Nucleic Acids Research 35(15): 5232-5241.

Hu, Y., K. Mao, Y. Zeng, S. Chen, Z. Tao, C. Yang, S. Sun, X. Wu, G. Meng and B. Sun (2010). "Tripartite-motif protein 30 negatively regulates NLRP3 inflammasome activation by modulating reactive oxygen species production." The Journal of Immunology 185(12): 7699-7705.

Hu, Y., J. Wang, B. Yang, N. Zheng, M. Qin, Y. Ji, G. Lin, L. Tian, X. Wu and L. Wu (2011). "Guanylate binding protein 4 negatively regulates virus-induced type I IFN and antiviral response by targeting IFN regulatory factor 7." The Journal of Immunology 187(12): 6456-6462.

Hubálek, Z. (2008). "Mosquito-borne viruses in Europe." Parasitology Research 103(1): 29-43.

Huber, J. P., H. J. Ramos, M. A. Gill and J. D. Farrar (2010). "Cutting edge: Type I IFN reverses human Th2 commitment and stability by suppressing GATA3." The Journal of Immunology 185(2): 813-817.

Huy, N. T., T. Van Giang, D. H. D. Thuy, M. Kikuchi, T. T. Hien, J. Zamora and K. Hirayama (2013). "Factors associated with dengue shock syndrome: A systematic review and meta-analysis." PLoS Neglected Tropical Diseases 7(9): e2412.

Hyde, J. L., R. Chen, D. W. Trobaugh, M. S. Diamond, S. C. Weaver, W. B. Klimstra and J. Wilusz (2015). "The 5′ and 3′ ends of alphavirus RNAs–Non-coding is not non- functional." Virus Research. 206:99-107.

Hyde, J. L., C. L. Gardner, T. Kimura, J. P. White, G. Liu, D. W. Trobaugh, C. Huang, M. Tonelli, S. Paessler and K. Takeda (2014). "A viral RNA structural element alters host recognition of nonself RNA." Science 343(6172): 783-787.

196 Inamadar, A. C., A. Palit, V. Sampagavi, S. Raghunath and N. Deshmukh (2008). "Cutaneous manifestations of chikungunya fever: observations made during a recent outbreak in south India." International Journal of Dermatology 47(2): 154-159.

Jackson, S. E., T. D. Craggs and J.-r. Huang (2006). "Understanding the folding of GFP using biophysical techniques." Expert Reviews of Proteomics 3(5): 545-559.

Jaffar-Bandjee, M. C., T. Das, J. J. Hoarau, P. Krejbich Trotot, M. Denizot, A. Ribera, P. Roques and P. Gasque (2009). "Chikungunya virus takes centre stage in virally induced arthritis: possible cellular and molecular mechanisms to pathogenesis." Microbes and Infection 11(14-15): 1206-1218.

Jaffar-Bandjee, M. C., D. Ramful, B. A. Gauzere, J. J. Hoarau, P. Krejbich-Trotot, S. Robin, A. Ribera, J. Selambarom and P. Gasque (2010). "Emergence and clinical insights into the pathology of Chikungunya virus infection." Expert Review of Anti-infective Therapy 8(9): 987-996.

Javelle, E., A. Ribera, I. Degasne, B.-A. Gaüzère, C. Marimoutou and F. Simon (2015). "Specific Management of Post-Chikungunya Rheumatic Disorders: A Retrospective Study of 159 Cases in Reunion Island from 2006-2012." PLoS Neglected Tropical Diseases 9(3): e0003603-e0003603.

Jensen, S. and A. R. Thomsen (2012). "Sensing of RNA Viruses: a Review of Innate Immune Receptors Involved in Recognizing RNA Virus Invasion." Journal of Virology 86(6): 2900-2910.

Ji, D., P. Liu, F. Wang, S. Zhang and H. Li (2012). "Identification and expression of a novel member of Ly-6 superfamily in zebrafish Denio rerio." Development Genes and Evolution 222(2): 119-124.

Johnson, C. D., A. Esquela-Kerscher, G. Stefani, M. Byrom, K. Kelnar, D. Ovcharenko, M. Wilson, X. Wang, J. Shelton and J. Shingara (2007). "The let-7 microRNA represses cell proliferation pathways in human cells." Cancer Research 67(16): 7713-7722.

Jose, J., J. E. Snyder and R. J. Kuhn (2009). "A structural and functional perspective of alphavirus replication and assembly." Future Microbiology 4(7): 837-856.

Josset, L., N. Tchitchek, L. E. Gralinski, M. T. Ferris, A. J. Eisfeld, R. R. Green, M. J. Thomas, J. Tisoncik-Go, G. P. Schroth and Y. Kawaoka (2014). "Annotation of long non- coding RNAs expressed in Collaborative Cross founder mice in response to respiratory

197 virus infection reveals a new class of interferon-stimulated transcripts." RNA Biology 11(7): 875-890.

Joubert, P.-E., S. W. Werneke, C. de la Calle, F. Guivel-Benhassine, A. Giodini, L. Peduto, B. Levine, O. Schwartz, D. J. Lenschow and M. L. Albert (2012). "Chikungunya virus– induced autophagy delays caspase-dependent cell death." The Journal of Experimental Medicine 209(5): 1029-1047.

Justman, J., M. R. Klimjack and M. Kielian (1993). "Role of spike protein conformational changes in fusion of Semliki Forest virus." Journal of Virology 67(12): 7597-7607.

Kalantri, S., R. Joshi and L. W. Riley (2006). "Chikungunya epidemic: an Indian perspective." Nationall Medical Journal of India 19(6): 315-322.

Kam, Y.-W., F.-M. Lum, T.-H. Teo, W. W. Lee, D. Simarmata, S. Harjanto, C.-L. Chua, Y.- F. Chan, J.-K. Wee and A. Chow (2012). "Early neutralizing IgG response to Chikungunya virus in infected patients targets a dominant linear epitope on the E2 glycoprotein." EMBO Molecular Medicine 4(4): 330-343.

Kam, Y. W., E. K. S. Ong, L. Rénia, J. C. Tong and L. F. P. Ng (2009). "Immuno-biology of Chikungunya and implications for disease intervention." Microbes and Infection 11(14-15): 1186-1196.

Karki, S., M. M. H. Li, J. W. Schoggins, S. Tian, C. M. Rice and M. R. MacDonald (2012). "Multiple Interferon Stimulated Genes Synergize with the Zinc Finger Antiviral Protein to Mediate Anti-Alphavirus Activity." PLoS One 7(5): e37398.

Karpova, A. Y., L. V. Ronco and P. M. Howley (2001). "Functional characterization of interferon regulatory factor 3a (IRF-3a), an alternative splice isoform of IRF-3." Molecular and Cellular Biology 21(13): 4169-4176.

Kato, H., S. Sato, M. Yoneyama, M. Yamamoto, S. Uematsu, K. Matsui, T. Tsujimura, K. Takeda, T. Fujita and O. Takeuchi (2005). "Cell type-specific involvement of RIG-I in antiviral response." Immunity 23(1): 19-28.

Katzmann, J. A., R. J. Clark, R. S. Abraham, S. Bryant, J. F. Lymp, A. R. Bradwell and R. A. Kyle (2002). "Serum reference intervals and diagnostic ranges for free κ and free λ immunoglobulin light chains: relative sensitivity for detection of monoclonal light chains." Clinical Chemistry 48(9): 1437-1444.

198 Kawai, T. and S. Akira (2010). "The role of pattern-recognition receptors in innate immunity: update on Toll-like receptors." Nature Immunology 11(5): 373-384.

Kawai, T. and S. Akira (2011). "Toll-like receptors and their crosstalk with other innate receptors in infection and immunity." Immunity 34(5): 637-650.

Kawai, T., S. Sato, K. J. Ishii, C. Coban, H. Hemmi, M. Yamamoto, K. Terai, M. Matsuda, J. Inoue and S. Uematsu (2004). "Interferon-α induction through Toll-like receptors involves a direct interaction of IRF7 with MyD88 and TRAF6." Nature Immunology 5(10): 1061-1068.

Kawasaki, T., T. Kawai and S. Akira (2011). "Recognition of nucleic acids by pattern‐recognition receptors and its relevance in autoimmunity." Immunological Reviews 243(1): 61-73.

Kelvin, A. A., D. Banner, G. Silvi, M. L. Moro, N. Spataro, P. Gaibani, F. Cavrini, A. Pierro, G. Rossini and M. J. Cameron (2011). "Inflammatory Cytokine Expression Is Associated with Chikungunya Virus Resolution and Symptom Severity." PLoS Neglected Tropical Diseases 5(8): e1279.

Kim, B.-H., A. R. Shenoy, P. Kumar, R. Das, S. Tiwari and J. D. MacMicking (2011). "A family of IFN-γ–inducible 65-kD GTPases protects against bacterial infection." Science 332(6030): 717-721.

Kim, J., A. J. Woo, J. Chu, J. W. Snow, Y. Fujiwara, C. G. Kim, A. B. Cantor and S. H. Orkin (2010). "A Myc network accounts for similarities between embryonic stem and cancer cell transcription programs." Cell 143(2): 313-324.

Klimstra, W. B., E. M. Nangle, M. S. Smith, A. D. Yurochko and K. D. Ryman (2003). "DC- SIGN and L-SIGN can act as attachment receptors for alphaviruses and distinguish between mosquito cell-and mammalian cell-derived viruses." Journal of Virology 77(22): 12022-102232.

Knight, A. W. and N. Billinton (2001). "Distinguishing GFP from cellular autofluorescence." Biophotonics International 8(7): 42-51.

Kononchik, J., R. Hernandez and D. Brown (2011). "An Alternative Pathway for Alphavirus Entry." Virology Journal 8(1): 304-314.

199 Kononchik, J. P., R. Vancini and D. T. Brown (2011). "Alphavirus adsorption to mosquito cells as viewed by freeze fracture immunolabeling." Virology 415(2): 132-140.

Kozak, M. (1986). "Point mutations define a sequence flanking the AUG initiator codon that modulates translation by eukaryotic ribosomes." Cell 44(2): 283-292.

Kozak, M. (1987). "An analysis of 5'-noncoding sequences from 699 vertebrate messenger RNAs." Nucleic Acids Research 15(20): 8125-8148.

Kozak, M. (1987). "At least six nucleotides preceding the AUG initiator codon enhance translation in mammalian cells." Journal of Molecular Biology 196(4): 947-950.

Krejbich-Trotot, P., M. Denizot, J. J. Hoarau, M. C. Jaffar-Bandjee, T. Das and P. Gasque (2011). "Chikungunya virus mobilizes the apoptotic machinery to invade host cell defenses." The FASEB Journal 25(1): 314-325.

Krejbich-Trotot, P., B. Gay, G. Li-Pat-Yuen, J.-J. Hoarau, M.-C. Jaffar-Bandjee, L. Briant, P. Gasque and M. Denizot (2011). "Chikungunya triggers an autophagic process which promotes viral replication." Virology Journal 8(1): 432.

Kuhn, R. J., Z. Hong and J. H. Strauss (1990). "Mutagenesis of the 3'nontranslated region of Sindbis virus RNA." Journal of Virology 64(4): 1465-1476.

Kumagai, Y., O. Takeuchi and S. Akira (2008). "Pathogen recognition by innate receptors." Journal of Infection and Chemotherapy 14(2): 86-92.

Kumagai, Y., O. Takeuchi, H. Kato, H. Kumar, K. Matsui, E. Morii, K. Aozasa, T. Kawai and S. Akira (2007). "Alveolar Macrophages Are the Primary Interferon-[alpha] Producer in Pulmonary Infection with RNA Viruses." Immunity 27(2): 240-252.

Kumar, H., T. Kawai, H. Kato, S. Sato, K. Takahashi, C. Coban, M. Yamamoto, S. Uematsu, K. J. Ishii and O. Takeuchi (2006). "Essential role of IPS-1 in innate immune responses against RNA viruses." The Journal of Experimental Medicine 203(7): 1795- 1803.

Kurt-Jones, E. A., L. Popova, L. Kwinn, L. M. Haynes, L. P. Jones, R. A. Tripp, E. E. Walsh, M. W. Freeman, D. T. Golenbock and L. J. Anderson (2000). "Pattern recognition receptors TLR4 and CD14 mediate response to respiratory syncytial virus." Nature Immunology 1(5): 398-401.

200 La Linn, M., J. A. Eble, C. Lubken, R. W. Slade, J. Heino, J. Davies and A. Suhrbier (2005). "An arthritogenic alphavirus uses the [alpha] 1 [beta] 1 integrin collagen receptor." Virology 336(2): 229-239.

La Linn, M., L. Mateo, J. Gardner and A. Suhrbier (1998). "Alphavirus-specific cytotoxic T lymphocytes recognize a cross-reactive epitope from the capsid protein and can eliminate virus from persistently infected macrophages." Journal of Virology 72(6): 5146-5153.

Labadie, K., T. Larcher, C. Joubert, A. Mannioui, B. Delache, P. Brochard, L. Guigand, L. Dubreil, P. Lebon and B. Verrier (2010). "Chikungunya disease in nonhuman primates involves long-term viral persistence in macrophages." The Journal of Clinical Investigation 120(3): 894-906.

Lamballerie, X. D., V. Boisson, J. C. Reynier, S. Enault, R. N. Charrel, A. Flahault, P. Roques and R. L. Grand (2008). "On chikungunya acute infection and chloroquine treatment." Vector-borne and Zoonotic Diseases 8(6): 837-840.

Lanciotti, R. S., M. L. Ludwig, E. B. Rwaguma, J. J. Lutwama, T. M. Kram, N. Karabatsos, B. C. Cropp and B. R. Miller (1998). "Emergence of epidemic O'nyong-nyong fever in Uganda after a 35-year absence: genetic characterization of the virus." Virology 252(1): 258-268.

Lander, E. S. and M. S. Waterman (1988). "Genomic mapping by fingerprinting random clones: a mathematical analysis." Genomics 2(3): 231-239.

Lee, J. T. (2012). "Epigenetic regulation by long noncoding RNAs." Science 338(6113): 1435-1439.

Lee, M. S., B. Kim, G. T. Oh and Y.-J. Kim (2013). "OASL1 inhibits translation of the type I interferon-regulating transcription factor IRF7." Nature Immunology 14(4): 346-355.

Lee, P. Y., Y. Li, Y. Kumagai, Y. Xu, J. S. Weinstein, E. S. Kellner, D. C. Nacionales, E. J. Butfiloski, N. van Rooijen and S. Akira (2009). "Type I interferon modulates monocyte recruitment and maturation in chronic inflammation." The American Journal of Pathology 175(5): 2023-2033.

Lee, P. Y., J.-X. Wang, E. Parisini, C. C. Dascher and P. A. Nigrovic (2013). "Ly6 family proteins in neutrophil biology." Journal of Leukocyte Biology 94(4): 585-594.

201 Lee, Y.-R., M.-T. Liu, H.-Y. Lei, C.-C. Liu, J.-M. Wu, Y.-C. Tung, Y.-S. Lin, T.-M. Yeh, S.-H. Chen and H.-S. Liu (2006). "MCP-1, a highly expressed chemokine in dengue haemorrhagic fever/dengue shock syndrome patients, may cause permeability change, possibly through reduced tight junctions of vascular endothelium cells." Journal of General Virology 87(12): 3623-3630.

Lee, Y. S. and A. Dutta (2007). "The tumor suppressor microRNA let-7 represses the HMGA2 oncogene." Genes & Development 21(9): 1025-1030.

Lei, X., X. Xiao, Q. Xue, Q. Jin, B. He and J. Wang (2013). "Cleavage of interferon regulatory factor 7 by enterovirus 71 3C suppresses cellular responses." Journal of Virology 87(3): 1690-1698.

Lemm, J. A., T. Rümenapf, E. G. Strauss, J. H. Strauss and C. Rice (1994). "Polypeptide requirements for assembly of functional Sindbis virus replication complexes: a model for the temporal regulation of minus-and plus-strand RNA synthesis." The EMBO Journal 13(12): 2925-2934.

Leung, J. Y.-S., M. M.-L. Ng and J. J. H. Chu (2011). "Replication of Alphaviruses: A Review on the Entry Process of Alphaviruses into Cells." Advances in Virology 2011. 9 pages. doi:10.1155/2011/249640.

Leuschner, F., P. Dutta, R. Gorbatov, T. I. Novobrantseva, J. S. Donahoe, G. Courties, K. M. Lee, J. I. Kim, J. F. Markmann and B. Marinelli (2011). "Therapeutic siRNA silencing in inflammatory monocytes in mice." Nature Biotechnology 29(11): 1005-1010.

Levy, D. E., I. J. Marié and J. E. Durbin (2011). "Induction and function of type I and III interferon in response to viral infection." Current Opinion in Virology 1(6): 476-486.

Levy, O. (2007). "Innate immunity of the newborn: basic mechanisms and clinical correlates." Nature Reviews Immunology 7(5): 379-390.

Li, C., L. Ma and X. Chen (2011). "Interferon regulatory factor 3-CL, an isoform of IRF3, antagonizes activity of IRF3." Cellular & Molecular Immunology 8(1): 67-74.

Li, J., S. C. Ding, H. Cho, B. C. Chung, M. Gale, S. K. Chanda and M. S. Diamond (2013). "A short hairpin RNA screen of interferon-stimulated genes identifies a novel negative regulator of the cellular antiviral response." MBio 4(3): e00385-00313.

202 Li, H., B. Handsaker, A. Wysoker, T. Fennell, J. Ruan, N. Homer, G. Marth, G. Abecasis and R. Durbin (2009). "The sequence alignment/map format and SAMtools." Bioinformatics 25(16): 2078-2079.

Lienenklaus, S., M. Cornitescu, N. Ziętara, M. Łyszkiewicz, N. Gekara, J. Jabłońska, F. Edenhofer, K. Rajewsky, D. Bruder and M. Hafner (2009). "Novel reporter mouse reveals constitutive and inflammatory expression of IFN-β in vivo." The Journal of Immunology 183(5): 3229-3236.

Liljestrom, P., S. Lusa, D. Huylebroeck and H. Garoff (1991). "In vitro mutagenesis of a full-length cDNA clone of Semliki Forest virus: the small 6,000-molecular-weight membrane protein modulates virus release." Journal of Virology 65(8): 4107-4113.

Lin, R., Y. Mamane and J. Hiscott (1999). "Structural and functional analysis of interferon regulatory factor 3: localization of the transactivation and autoinhibitory domains." Molecular and Cellular Biology 19(4): 2465-2474.

Liu, S. Y., D. J. Sanchez and G. Cheng (2011). "New developments in the induction and antiviral effectors of type I interferon." Current Opinion in Immunology 23(1): 57-64.

Liu, Y., Q. Chen, Y. Song, L. Lai, J. Wang, H. Yu, X. Cao and Q. Wang (2011). "MicroRNA-98 negatively regulates IL-10 production and endotoxin tolerance in macrophages after LPS stimulation." FEBS letters 585(12): 1963-1968.

López, V., I. Molina, C. Monteagudo and E. Jordá (2011). "Cutaneous sarcoidosis developing after treatment with pegylated interferon and ribavirin: a new case and review of the literature." International Journal of Dermatology 50(3): 287-291.

Loppnow, H., K. Werdan and M. Buerke (2008). "Invited review: Vascular cells contribute to atherosclerosis by cytokine-and innate-immunity-related inflammatory mechanisms." Innate Immunity 14(2): 63-87.

Love, M. I., W. Huber and S. Anders (2014). "Moderated estimation of fold change and dispersion for RNA-Seq data with DESeq2." BioRxiv.

Lumsden, W. H. R. (1955). "An epidemic of virus disease in Southern Province, Tanganyika territory, in 1952-1953 II. General description and epidemiology." Transactions of the Royal Society of Tropical Medicine and Hygiene 49(1): 33-57.

203 MacMicking, J. D. (2012). "Interferon-inducible effector mechanisms in cell-autonomous immunity." Nature Reviews Immunology 12(5): 367-382.

Mahendradas, P., S. K. Ranganna, R. Shetty, R. Balu, K. M. Narayana, R. B. Babu and B. K. Shetty (2008). "Ocular manifestations associated with chikungunya." Ophthalmology 115(2): 287-291.

Maheshwari, R. K., V. Srikantan and D. Bhartiya (1991). "Chloroquine enhances replication of Semliki Forest virus and encephalomyocarditis virus in mice." Journal of Virology 65(2): 992—995.

Malathi, K., B. Dong, M. Gale and R. H. Silverman (2007). "Small self-RNA generated by RNase L amplifies antiviral innate immunity." Nature 448(7155): 816-819.

Malet, H., B. Coutard, S. Jamal, H. Dutartre, N. Papageorgiou, M. Neuvonen, T. Ahola, N. Forrester, E. A. Gould and D. Lafitte (2009). "The crystal structures of Chikungunya and Venezuelan equine encephalitis virus nsP3 macro domains define a conserved adenosine binding pocket." Journal of Virology 83(13): 6534-6545.

Mallilankaraman, K., D. J. Shedlock, H. Bao, O. U. Kawalekar, P. Fagone, A. A. Ramanathan, B. Ferraro, J. Stabenow, P. Vijayachari and S. G. Sundaram (2011). "A DNA Vaccine against Chikungunya Virus Is Protective in Mice and Induces Neutralizing Antibodies in Mice and Nonhuman Primates." PLoS Neglected Tropical Diseases 5(1): e928-940.

Maloney, N. S., L. B. Thackray, G. Goel, S. Hwang, E. Duan, P. Vachharajani, R. Xavier and H. W. Virgin (2012). "Essential cell-autonomous role for interferon (IFN) regulatory factor 1 in IFN-γ-mediated inhibition of norovirus replication in macrophages." Journal of Virology 86(23): 12655-12664.

Malvy, D., K. Ezzedine, M. Mamani-Matsuda, B. Autran, H. Tolou, M.-C. Receveur, T. Pistone, J. Rambert, D. Moynet and D. Mossalayi (2009). "Destructive arthritis in a patient with chikungunya virus infection with persistent specific IgM antibodies." BMC Infectious Diseases 9(1): 200-207.

Mancini, E. J., M. Clarke, B. E. Gowen, T. Rutten and S. D. Fuller (2000). "Cryo-electron microscopy reveals the functional organization of an enveloped virus, Semliki Forest virus." Molecular Cell 5(2): 255-266.

204 Mandl, J. N., R. Akondy, B. Lawson, N. Kozyr, S. I. Staprans, R. Ahmed and M. B. Feinberg (2011). "Distinctive TLR7 Signaling, Type I IFN Production, and Attenuated Innate and Adaptive Immune Responses to Yellow Fever Virus in a Primate Reservoir Host." The Journal of Immunology 186(11): 6406-6416.

Mangalgi, S. M., S. Shenoy, P. G. Maralusiddappa and I. V. Aprameya (2011). "Neonatal Chikungunya–A Case Series." Journal of Pediatric Sciences 3(2): e74-78.

Manimunda, S. P., P. Vijayachari, R. Uppoor, A. P. Sugunan, S. S. Singh, S. K. Rai, A. B. Sudeep, N. Muruganandam, I. K. Chaitanya and D. R. Guruprasad (2010). "Clinical progression of chikungunya fever during acute and chronic arthritic stages and the changes in joint morphology as revealed by imaging." Transactions of the Royal Society of Tropical Medicine and Hygiene 104(6): 392-399.

Marie, I., J. E. Durbin and D. E. Levy (1998). "Differential viral induction of distinct interferon‐α genes by positive feedback through interferon regulatory factor 7." The EMBO Journal 17(22): 6660-6669.

Marston, D. A., L. M. McElhinney, R. J. Ellis, D. L. Horton, E. L. Wise, S. L. Leech, D. David, X. de Lamballerie and A. R. Fooks (2013). "Next generation sequencing of viral RNA genomes." BMC genomics 14(1): 444.

Martinon, F. and J. Tschopp (2005). "NLRs join TLRs as innate sensors of pathogens." Trends in immunology 26(8): 447-454

Masilamani, V., K. Al-Zhrani, M. Al-Salhi, A. Al-Diab and M. Al-Ageily (2004). "Cancer diagnosis by autofluorescence of blood components." Journal of Luminescence 109(3): 143-154.

Mateo, L., M. La Linn, S. R. McColl, S. Cross, J. Gardner and A. Suhrbier (1999). "An arthrogenic alphavirus induces monocyte chemoattractant protein-1 and interleukin-8." Intervirology 43(1): 55-60.

Mathiot, C. C., G. Grimaud, P. Garry, J. C. Bouquety, A. Mada, A. M. Daguisy and A. J. Georges (1990). "An outbreak of human Semliki Forest virus infections in Central African Republic." The American Journal of Tropical Medicine and Hygiene 42(4): 386-393.

Mavalankar, D., P. Shastri and P. Raman (2007). "Chikungunya epidemic in India: a major public-health disaster." The Lancet 7(5): 306-308.

205 McClure, R., D. Balasubramanian, Y. Sun, M. Bobrovskyy, P. Sumby, C. A. Genco, C. K. Vanderpool and B. Tjaden (2013). "Computational analysis of bacterial RNA-Seq data." Nucleic Acids Research 41(14): e140-e140.

McGettigan, P. A. (2013). "Transcriptomics in the RNA-seq era." Current Opinion Chemical Biology 17(1): 4-11.

McInerney, G. M., J. M. Smit, P. Liljestrom and J. Wilschut (2004). "Semliki Forest virus produced in the absence of the 6K protein has an altered spike structure as revealed by decreased membrane fusion capacity." Virology 325(2): 200-206.

McKenna, S. D., K. Vergilis, A. R. N. Arulanadam, W. Y. Weiser, R. Nabiollin and M. A. Tepper (2004). "Formation of Human IFN-β Complex with the Soluble Type I Interferon Receptor IFNAR-2 Leads to Enhanced IFN Stability, Pharmacokinetics, and Antitumor Activity in Xenografted SCID Mice." Journal of Interferon & Cytokine Research 24(2): 119- 129.

Meisgen, F., N. Xu, T. Wei, P. C. Janson, S. Obad, O. Broom, N. Nagy, S. Kauppinen, L. Kemény and M. Ståhle (2012). "MiR-21 is up-regulated in psoriasis and suppresses T cell apoptosis." Experimental Dermatology 21(4): 312-314.

Melton, J. V., G. D. Ewart, R. C. Weir, P. G. Board, E. Lee and P. W. Gage (2002). "Alphavirus 6K proteins form ion channels." Journal of Biological Chemistry 277(49): 46923-46931.

Memari, H. R., R. N. Ramanan and A. B. Ariff (2010). "Comparison of expression systems for the production of human interferon-α2b." Central European Journal of Biology 5(4): 446-455.

Mi, S. and V. Stollar (1991). "Expression of Sindbis virus nsP1 and methyltransferase activity inEscherichia coli." Virology 184(1): 423-427.

Minakawa, M., T. Sone, T. Takeuchi and H. Yokosawa (2008). "Regulation of the nuclear factor (NF)-kappaB pathway by ISGylation." Biological & Pharmaceutical Bulletin 31(12): 2223-2227.

Minoche, A. E., J. C. Dohm and H. Himmelbauer (2011). "Evaluation of genomic high- throughput sequencing data generated on Illumina HiSeq and genome analyzer systems." Genome Biology 12(11): R112.

206 Mizuno, Y., Y. Kato, N. Takeshita, M. Ujiie, T. Kobayashi, S. Kanagawa, K. Kudo, C.-K. Lim and T. Takasaki (2011). "Clinical and radiological features of imported chikungunya fever in Japan: a study of six cases at the National Center for Global Health and Medicine." Journal of Infection and Chemotherapy 17(3): 419-423.

Modena, P., E. Lualdi, F. Facchinetti, L. Galli, M. R. Teixeira, S. Pilotti and G. Sozzi (2005). "SMARCB1/INI1 tumor suppressor gene is frequently inactivated in epithelioid sarcomas." Cancer Research 65(10): 4012-4019.

Morazzani, E. M., M. R. Wiley, M. G. Murreddu, Z. N. Adelman and K. M. Myles (2012). "Production of virus-derived ping-pong-dependent piRNA-like small RNAs in the mosquito soma." PLoS Pathogens 8(1): e1002470.

Morrison, T. E., A. C. Whitmore, R. S. Shabman, B. A. Lidbury, S. Mahalingam and M. T. Heise (2006). "Characterization of Ross River virus tropism and virus-induced inflammation in a mouse model of viral arthritis and myositis." Journal of Virology 80(2): 737-749.

Mortazavi, A., B. A. Williams, K. McCue, L. Schaeffer and B. Wold (2008). "Mapping and quantifying mammalian transcriptomes by RNA-Seq." Nature Methods 5(7): 621-628.

Murphy, F., W. Taylor, C. Mims and I. Marshall (1973). "Pathogenesis of Ross River virus infection in mice. II. Muscle, heart, and brown fat lesions." Journal of Infectious Diseases 127(2): 129-138.

Muthumani, K., K. M. Lankaraman, D. J. Laddy, S. G. Sundaram, C. W. Chung, E. Sako, L. Wu, A. Khan, N. Sardesai and J. J. Kim (2008). "Immunogenicity of novel consensus- based DNA vaccines against Chikungunya virus." Vaccine 26(40): 5128-5134.

Mutz, K.-O., A. Heilkenbrinker, M. Lönne, J.-G. Walter and F. Stahl (2013). "Transcriptome analysis using next-generation sequencing." Current Opinion in Biotechnology 24(1): 22- 30.

Nakaya, H. I., J. Gardner, Y. S. Poo, L. Major, B. Pulendran and A. Suhrbier (2012). "Gene profiling of Chikungunya virus arthritis in a mouse model reveals significant overlap with rheumatoid arthritis." Arthritis & Rheumatism 64(11): 3553-3563.

Neighbours, L. M., K. Long, A. C. Whitmore and M. T. Heise (2012). "Myd88-dependent toll-like receptor 7 signaling mediates protection from severe Ross River virus-induced disease in mice." Journal of virology 86(19): 10675-10685.

207 Nekrutenko, A. and J. Taylor (2012). "Next-generation sequencing data interpretation: enhancing reproducibility and accessibility." Nature Reviews Genetics 13(9): 667-672.

Ng, C. G., I. Coppens, D. Govindarajan, J. Pisciotta, V. Shulaev and D. E. Griffin (2008). "Effect of host cell lipid metabolism on alphavirus replication, virion morphogenesis, and infectivity." Proceedings of the National Academy of Sciences 105(42): 16326-16331.

Ng, L. F. P., A. Chow, Y. J. Sun, D. J. C. Kwek, P. L. Lim, F. Dimatatac, L. C. Ng, E. E. Ooi, K. H. Choo and Z. Her (2009). "IL-1β, IL-6, and RANTES as biomarkers of Chikungunya severity." PLoS One 4(1): e4261.

Nguyen, K. B., L. P. Cousens, L. A. Doughty, G. C. Pien, J. E. Durbin and C. A. Biron (2000). "Interferon α/β-mediated inhibition and promotion of interferon γ: STAT1 resolves a paradox." Nature Immunology 1(1): 70-76.

Nickens, D. G. and R. W. Hardy (2008). "Structural and functional analyses of stem–loop 1 of the Sindbis virus genome." Virology 370(1): 158-172.

Ning, S., L. E. Huye and J. S. Pagano (2005). "Regulation of the transcriptional activity of the IRF7 promoter by a pathway independent of interferon signaling." Journal of Biological Chemistry 280(13): 12262-12270.

Ning, S., J. Pagano and G. Barber (2011). "IRF7: activation, regulation, modification and function." Genes and Immunity 12(6): 399-414.

Nordmann, A., L. Wixler, Y. Boergeling, V. Wixler and S. Ludwig (2012). "A new splice variant of the human guanylate-binding protein 3 mediates anti-influenza activity through inhibition of viral transcription and replication." The FASEB Journal 26(3): 1290-1300.

O’Donnell, L. A., S. Conway, R. W. Rose, E. Nicolas, M. Slifker, S. Balachandran and G. F. Rall (2012). "STAT1-independent control of a neurotropic measles virus challenge in primary neurons and infected mice." The Journal of Immunology 188(4): 1915-1923.

Olagnier, D., F. E. Scholte, C. Chiang, I. C. Albulescu, C. Nichols, Z. He, R. Lin, E. J. Snijder, M. J. van Hemert and J. Hiscott (2014). "Inhibition of dengue and chikungunya virus infections by RIG-I-mediated type I interferon-independent stimulation of the innate antiviral response." Journal of Virology 88(8): 4180-4194.

208 Ortega-Prieto, A. M., J. Sheldon, A. Grande-Pérez, H. Tejero, J. Gregori, J. Quer, J. I. Esteban, E. Domingo and C. Perales (2013). "Extinction of hepatitis C virus by ribavirin in hepatoma cells involves lethal mutagenesis." PloS One 8(8): e71039.

Ozden, S., M. Huerre, J. P. Riviere, L. L. Coffey, P. V. Afonso, V. Mouly, J. De Monredon, J. C. Roger, M. El Amrani and J. L. Yvin (2007). "Human muscle satellite cells as targets of Chikungunya virus infection." PLoS One 2(6): e527.

Ozsolak, F. and P. M. Milos (2010). "RNA sequencing: advances, challenges and opportunities." Nature Reviews Genetics 12(2): 87-98.

Padmakumar, B., J. B. Jayan, R. M. R. Menon, B. Krishnankutty, R. Payippallil and R. Nisha (2009). "Comparative evaluation of four therapeutic regimes in chikungunya arthritis: a prospective randomized parallel-group study." Indian Journal of Rheumatology 4(3): 94- 101.

Pakran, J., M. George, N. Riyaz, R. Arakkal, S. George, U. Rajan, A. Khader, S. Thomas, R. Abdurahman and S. Sasidharanpillai (2011). "Purpuric macules with vesiculobullous lesions: a novel manifestation of Chikungunya." International journal of dermatology 50(1): 61-69.

Pal, P., K. A. Dowd, J. D. Brien, M. A. Edeling, S. Gorlatov, S. Johnson, I. Lee, W. Akahata, G. J. Nabel and M. Richter (2013). "Development of a highly protective combination monoclonal antibody therapy against Chikungunya virus." PLoS Pathogens 9(4): e1003312.

Palha, N., F. Guivel-Benhassine, V. Briolat, G. Lutfalla, M. Sourisseau, F. Ellett, C.-H. Wang, G. J. Lieschke, P. Herbomel and O. Schwartz (2013). "Real-time whole-body visualization of Chikungunya virus infection and host interferon response in zebrafish." PLoS Pathogens 9(9): e1003619.

Pandya, S. (2008). "Methotrexate and hydroxychloroquine combination therapy in chronic chikungunya arthritis: a 16 week study." Indian Journal of Rheumatology 3(3): 93-97.

Pang, T., M. J. Cardosa and M. G. Guzman (2007). "Of cascades and perfect storms: the immunopathogenesis of dengue haemorrhagic fever-dengue shock syndrome (DHF/DSS)." Immunology and Cell Biology 85(1): 43-45.

209 Panning, M., K. Grywna, M. Van Esbroeck, P. Emmerich and C. Drosten (2008). "Chikungunya fever in travelers returning to Europe from the Indian Ocean region, 2006." Emerging Infectious Diseases 14(3): 416-412.

Paredes, A. M., D. Ferreira, M. Horton, A. Saad, H. Tsuruta, R. Johnston, W. Klimstra, K. Ryman, R. Hernandez, W. Chiu and D. T. Brown (2004). "Conformational changes in Sindbis virions resulting from exposure to low pH and interactions with cells suggest that cell penetration may occur at the cell surface in the absence of membrane fusion." Virology 324(2): 373-386.

Parola, P., X. de Lamballerie, J. Jourdan, C. Rovery, V. Vaillant, P. Minodier, P. Brouqui, A. Flahault, D. Raoult and R. Charrel (2006). "Novel chikungunya virus variant in travelers returning from Indian Ocean islands." Emerging Infectious Diseases 12(10): 1493-1499.

Partidos, C. D., J. Weger, J. Brewoo, R. Seymour, E. M. Borland, J. P. Ledermann, A. M. Powers, S. C. Weaver, D. T. Stinchcomb and J. E. Osorio (2011). "Probing the attenuation and protective efficacy of a candidate chikungunya virus vaccine in mice with compromised interferon (IFN) signaling." Vaccine 29(16): 3067-3073

Pasqualucci, L., P. Neumeister, T. Goossens, G. Nanjangud, R. Chaganti, R. Küppers and R. Dalla-Favera (2001). "Hypermutation of multiple proto-oncogenes in B-cell diffuse large- cell lymphomas." Nature 412(6844): 341-346.

Pastorino, B. A. M. w., C. N. Peyrefitte, L. Almeras, M. Grandadam, D. Rolland, H. J. Tolou and M. Bessaud (2008). "Expression and biochemical characterization of nsP2 cysteine protease of Chikungunya virus." Virus Research 131(2): 293-298.

Paz, S., Q. Sun, P. Nakhaei, R. Romieu-Mourez, D. Goubau, I. Julkunen, R. Lin and J. Hiscott (2006). "Induction of IRF-3 and IRF-7 phosphorylation following activation of the RIG-I pathway." Cellular and molecular biology (Noisy-le-Grand, France) 52(1): 17-28.

Perera, R., K. E. Owen, T. L. Tellinghuisen, A. E. Gorbalenya and R. J. Kuhn (2001). "Alphavirus nucleocapsid protein contains a putative coiled coil α-helix important for core assembly." Journal of Virology 75(1): 1-10.

Perri, S., D. A. Driver, J. P. Gardner, S. Sherrill, B. A. Belli, T. W. Dubensky and J. M. Polo (2000). "Replicon vectors derived from Sindbis virus and Semliki forest virus that establish persistent replication in host cells." Journal of Virology 74(20): 9802-9807.

210 Perry, A. K., C. Gang, D. Zheng, T. Hong and G. Cheng (2005). "The host type I interferon response to viral and bacterial infections." Cell Research 15(6): 407-422.

Peter, M. E. (2009). "Let-7 and miR-200 microRNAs: guardians against pluripotency and cancer progression." Cell Cycle 8(6): 843-852.

Petitdemange, C., P. Becquart, N. Wauquier, V. Béziat, P. Debré, E. M. Leroy and V. Vieillard (2011). "Unconventional repertoire profile is imprinted during acute chikungunya infection for natural killer cells polarization toward cytotoxicity." PLoS pathogens 7(9): e1002268.

Pialoux, G., B. A. Gaüzère, S. Jauréguiberry and M. Strobel (2007). "Chikungunya, an epidemic arbovirosis." The Lancet infectious diseases 7(5): 319-327.

Pichlmair, A., C. Lassnig, C. A. Eberle, M. W. Górna, C. L. Baumann, T. R. Burkard, T. Bürckstümmer, A. Stefanovic, S. Krieger and K. L. Bennett (2011). "IFIT1 is an antiviral protein that recognizes 5 [prime]-triphosphate RNA." Nature immunology 12(7): 624-630.

Platanias, L. C. (2005). "Mechanisms of type-I-and type-II-interferon-mediated signalling." Nature Reviews Immunology 5(5): 375-386.

Plante, K., E. Wang, C. D. Partidos, J. Weger, R. Gorchakov, K. Tsetsarkin, E. M. Borland, A. M. Powers, R. Seymour and D. T. Stinchcomb (2011). "Novel chikungunya vaccine candidate with an IRES-based attenuation and host range alteration mechanism." PLoS Pathog 7(7): e1002142-e1002142

Poelchau, M. F., X. Huang, A. Goff, J. Reynolds and P. Armbruster (2014). "An Experimental and Bioinformatics Protocol for RNA-seq Analyses of Photoperiodic Diapause in the Asian Tiger Mosquito, Aedes albopictus." Journal of Visualized Experiments(93): e51961-e51961.

Poliseno, L. (2012). "Pseudogenes: newly discovered players in human cancer." Science signaling(5): re5.

Poo, Y. S., H. Nakaya, J. Gardner, T. Larcher, W. A. Schroder, T. T. Le, L. D. Major and A. Suhrbier (2014). "CCR2 deficiency promotes exacerbated chronic erosive neutrophil- dominated chikungunya virus arthritis." Journal of Virology 88(12): 6862-6872.

Poo, Y. S., P. A. Rudd, J. Gardner, J. A. Wilson, T. Larcher, M.-A. Colle, T. T. Le, H. I. Nakaya, D. Warrilow and R. Allcock (2014). "Multiple Immune Factors Are Involved in

211 Controlling Acute and Chronic Chikungunya Virus Infection." PLoS neglected tropical diseases 8(12): e3354.

Powers, A. M. (2010). "Chikungunya." Clinics in laboratory medicine 30(1): 209-219.

Powers, A. M., P. V. Aguilar, L. J. Chandler, A. C. Brault, T. A. Meakins, D. Watts, K. L. Russell, J. Olson, P. F. C. Vasconcelos and R. DA (2006). "Genetic relationships among Mayaro and Una viruses suggest distinct patterns of transmission." The American journal of tropical medicine and hygiene 75(3): 461-469.

Powers, A. M., A. C. Brault, R. B. Tesh and S. C. Weaver (2000). "Re-emergence of Chikungunya and O’nyong-nyong viruses: evidence for distinct geographical lineages and distant evolutionary relationships." Journal of General Virology 81(2): 471-479.

Powers, A. M. and C. H. Logue (2007). "Changing patterns of chikungunya virus: re- emergence of a zoonotic arbovirus." Journal of General Virology 88(9): 2363-2377.

Pruitt, K. D., J. Harrow, R. A. Harte, C. Wallin, M. Diekhans, D. R. Maglott, S. Searle, C. M. Farrell, J. E. Loveland and B. J. Ruef (2009). "The consensus coding sequence (CCDS) project: Identifying a common protein-coding gene set for the human and mouse genomes." Genome Research 19(7): 1316-1323.

Pruitt, K. D., T. Tatusova, W. Klimke and D. R. Maglott (2009). "NCBI Reference Sequences: current status, policy and new initiatives." Nucleic Acids Research 37(suppl 1): D32-D36.

Pulit-Penaloza, J. A., S. V. Scherbik and M. A. Brinton (2012). "Type 1 IFN-independent activation of a subset of interferon stimulated genes in West Nile virus Eg101-infected mouse cells." Virology 425(2): 82-94.

Pulverer, J. E., U. Rand, S. Lienenklaus, D. Kugel, N. Ziętara, G. Kochs, R. Naumann, S. Weiss, P. Staeheli and H. Hauser (2010). "Temporal and spatial resolution of type I and III interferon responses in vivo." Journal of Virology 84(17): 8626-8638.

Qu, B., X. Qi, X. Wu, M. Liang, C. Li, C. J. Cardona, W. Xu, F. Tang, Z. Li and B. Wu (2012). "Suppression of the interferon and NF-κB responses by severe fever with thrombocytopenia syndrome virus." Journal of Virology 86(16): 8388-8401.

212 Qureshi, I. A. and M. F. Mehler (2012). "Emerging roles of non-coding RNAs in brain evolution, development, plasticity and disease." Nature Reviews Neuroscience 13(8): 528- 541.

Rajapakse, S. (2011). "Dengue shock." Journal of Emergencies, Trauma and Shock 4(1): 120.

Ramful, D., M. Carbonnier, M. Pasquet, B. Bouhmani, J. Ghazouani, T. Noormahomed, G. Beullier, T. Attali, S. Samperiz and A. Fourmaintraux (2007). "Mother-to-child transmission of Chikungunya virus infection." The Pediatric Infectious Disease Journal 26(9): 811-815.

Ramsauer, K., M. Schwameis, C. Firbas, M. Müllner, R. J. Putnak, S. J. Thomas, P. Desprès, E. Tauber, B. Jilma and F. Tangy (2015). "Immunogenicity, safety, and tolerability of a recombinant measles-virus-based chikungunya vaccine: a randomised, double-blind, placebo-controlled, active-comparator, first-in-man trial." The Lancet infectious diseases 15(5): 519-527

Rand, U., M. Rinas, J. Schwerk, G. Nöhren, M. Linnes, A. Kröger, M. Flossdorf, K. Kály- Kullai, H. Hauser and T. Höfer (2012). "Multi-layered stochasticity and paracrine signal propagation shape the type-I interferon response." Molecular Systems Biology 8(1).

Ranjit, S., N. Kissoon and I. Jayakumar (2005). "Aggressive management of dengue shock syndrome may decrease mortality rate: a suggested protocol*." Pediatric Critical Care Medicine 6(4): 412-419.

Ravichandran, R. and M. Manian (2008). "Ribavirin therapy for Chikungunya arthritis." The Journal of Infection in Developing Countries 2(02): 140-142.

Rayamajhi, M., J. Humann, K. Penheiter, K. Andreasen and L. L. Lenz (2010). "Induction of IFN-αβ enables Listeria monocytogenes to suppress macrophage activation by IFN-γ." The Journal of Experimental Medicine 207(2): 327-337.

Relloso, M., A. Puig-Kröger, O. M. Pello, J. L. Rodríguez-Fernández, G. de la Rosa, N. Longo, J. Navarro, M. A. Muñoz-Fernández, P. Sánchez-Mateos and A. L. Corbí (2002). "DC-SIGN (CD209) expression is IL-4 dependent and is negatively regulated by IFN, TGF- β, and anti-inflammatory agents." The Journal of Immunology 168(6): 2634-2643.

Renault, P., J. L. Solet, D. Sissoko, E. Balleydier, S. Larrieu, L. Filleul, C. Lassalle, J. Thiria, E. Rachou and H. de Valk (2007). "A major epidemic of chikungunya virus infection

213 on Reunion Island, France, 2005–2006." The American Journal of Tropical Medicine and Hygiene 77(4): 727-731.

Rezza, G., L. Nicoletti, R. Angelini, R. Romi, A. Finarelli, M. Panning, P. Cordioli, C. Fortuna, S. Boros and F. Magurano (2007). "Infection with chikungunya virus in Italy: an outbreak in a temperate region." The Lancet 370(9602): 1840-1846.

Rikkonen, M., J. Peranen and L. Kaariainen (1994). "ATPase and GTPase activities associated with Semliki Forest virus nonstructural protein nsP2." Journal of Virology 68(9): 5804-5810.

Rinn, J. L. and H. Y. Chang (2012). "Genome regulation by long noncoding RNAs." Annual Review of Biochemistry 81.

Robin, S., D. Ramful, F. Le Seach, M. C. Jaffar-Bandjee, G. Rigou and J. L. Alessandri (2008). "Neurologic manifestations of pediatric chikungunya infection." Journal of Child Neurology 23(9): 1028-1035.

Robinson, J. T., H. Thorvaldsdóttir, W. Winckler, M. Guttman, E. S. Lander, G. Getz and J. P. Mesirov (2011). "Integrative genomics viewer." Nature Biotechnology 29(1): 24-26.

Robinson, M. C. (1955). "An epidemic of virus disease in Southern Province, Tanganyika territory, in 1952-1953." Transactions of the Royal Society of Tropical Medicine and Hygiene 49(1): 28-32.

Rosas, M., B. Thomas, M. Stacey, S. Gordon and P. R. Taylor (2010). "The myeloid 7/4- antigen defines recently generated inflammatory macrophages and is synonymous with Ly-6B." Journal of leukocyte biology 88(1): 169-180.

Ross, R. W. (1956). "The Newala Epidemic. III. The Virus: Isolation, Pathogenic Properties and Relationship to the Epidemic." The Journal of Hygiene 54(2): 177-191.

Roy, C. J., A. P. Adams, E. Wang, K. Plante, R. Gorchakov, R. L. Seymour, H. Vinet- Oliphant and S. C. Weaver (2014). "Chikungunya vaccine candidate is highly attenuated and protects nonhuman primates against telemetrically-monitored disease following a single dose." Journal of Infectious Diseases: jiu014

Rozen-Gagnon, K., K. A. Stapleford, V. Mongelli, H. Blanc, A.-B. Failloux, M.-C. Saleh and M. Vignuzzi (2014). "Alphavirus mutator variants present host-specific defects and attenuation in mammalian and insect models." PLoS pathogens 10(1): e1003877.

214 Rubach, J. K., B. R. Wasik, J. C. Rupp, R. J. Kuhn, R. W. Hardy and J. L. Smith (2009). "Characterization of purified Sindbis virus nsP4 RNA-dependent RNA polymerase activity in vitro." Virology 384(1): 201-208.

Rudd, P. A., J. Wilson, J. Gardner, T. Larcher, C. Babarit, T. T. Le, I. Anraku, Y. Kumagai, Y.-M. Loo and M. Gale Jr (2012). "Interferon Response Factors 3 and 7 Protect against Chikungunya Virus Hemorrhagic Fever and Shock." Journal of Virology 86(18): 9888- 9898.

Ruffalo, M., T. LaFramboise and M. Koyutürk (2011). "Comparative analysis of algorithms for next-generation sequencing read alignment." Bioinformatics 27(20): 2790-2796.

Rulli, N. E., A. Guglielmotti, G. Mangano, M. S. Rolph, C. Apicella, A. Zaid, A. Suhrbier and S. Mahalingam (2009). "Amelioration of alphavirus‐induced arthritis and myositis in a mouse model by treatment with bindarit, an inhibitor of monocyte chemotactic proteins." Arthritis & Rheumatism 60(8): 2513-2523.

Rulli, N. E., M. S. Rolph, A. Srikiatkhachorn, S. Anantapreecha, A. Guglielmotti and S. Mahalingam (2011). "Protection From Arthritis and Myositis in a Mouse Model of Acute Chikungunya Virus Disease by Bindarit, an Inhibitor of Monocyte Chemotactic Protein-1 Synthesis." Journal of Infectious Diseases 204(7): 1026-1030.

Rümenapf, T., E. G. Strauss and J. H. Strauss (1994). "Subgenomic mRNA of Aura alphavirus is packaged into virions." Journal of Virology 68(1): 56-62.

Rusinova, I., S. Forster, S. Yu, A. Kannan, M. Masse, H. Cumming, R. Chapman and P. J. Hertzog (2013). "Interferome v2. 0: an updated database of annotated interferon-regulated genes." Nucleic Acids Research 41(D1): D1040-D1046.

Russo, A. T., M. A. White and S. J. Watowich (2006). "The crystal structure of the Venezuelan equine encephalitis alphavirus nsP2 protease." Structure 14(9): 1449-1458.

Ryman, K. D. and W. B. Klimstra (2008). "Host responses to alphavirus infection." Immunological reviews 225(1): 27-45.

Ryman, K. D., K. Meier, E. M. Nangle, S. Ragsdale, N. Korneeva, R. Rhoads, M. R. MacDonald and W. B. Klimstra (2005). "Sindbis virus translation is inhibited by a PKR/RNase L-independant effector induced by alpha/beta intereferon priming of dendritic cells." Journal of Virology 79(3): 1487-1499.

215 Ryman, K. D., L. J. White, R. E. Johnston and W. B. Klimstra (2002). "Effects of PKR/RNase L-dependent and alternative antiviral pathways on alphavirus replication and pathogenesis." Viral Immunology 15(1): 53-76.

Sadler, A. J. and B. R. G. Williams (2008). "Interferon-inducible antiviral effectors." Nature Reviews Immunology 8(7): 559-568.

Saha, B., S. J. Prasanna, B. Chandrasekar and D. Nandi (2010). "Gene modulation and immunoregulatory roles of Interferonγ." Cytokine 50(1): 1-14.

Saito, S. (1989). "Possible involvement of virus-induced protein kinase in the antiviral state induced with interferon-gamma against Sindbis virus." Journal of Cytokine Research 9(1): 23-24.

Salter, A. and E. Tarling (2007). "Regulation of gene transcription by fatty acids."

Sam, I. C. and S. AbuBakar (2006). "Chikungunya virus infection." Medical Journal of Malaysia 61(2): 264-270.

Sarkar, S. N. and G. C. Sen (2004). "Novel functions of proteins encoded by viral stress- inducible genes." Pharmacology & Therapeutics 103(3): 245-259.

Sato, M., H. Suemori, N. Hata, M. Asagiri, K. Ogasawara, K. Nakao, T. Nakaya, M. Katsuki, S. Noguchi and N. Tanaka (2000). "Distinct and Essential Roles of Transcription Factors IRF-3 and IRF-7 in Response to Viruses for IFN-α/β Gene Induction." Immunity 13(4): 539-548.

Scheidel, L. M. and V. Stollar (1991). "Mutations that confer resistance to mycophenolic acid and ribavirin on Sindbis virus map to the nonstructural protein nsP1." Virology 181(2): 490-499.

Schilte, C., M. R. Buckwalter, M. E. Laird, M. S. Diamond, O. Schwartz and M. L. Albert (2012). "Cutting edge: independent roles for IRF-3 and IRF-7 in hematopoietic and nonhematopoietic cells during host response to Chikungunya infection." Journal of Immunology 188(7): 2967-2971.

Schilte, C., T. Couderc, F. Chretien, M. Sourisseau, N. Gangneux, F. Guivel-Benhassine, A. Kraxner, J. Tschopp, S. Higgs and A. Michault (2010). "Type I IFN controls chikungunya virus via its action on nonhematopoietic cells." The Journal of Experimental Medicine 207(2): 429-442.

216 Schmieder, R. and R. Edwards (2011). "Quality control and preprocessing of metagenomic datasets." Bioinformatics 27(6): 863-864.

Schoggins, J. W. and C. M. Rice (2011). "Interferon-stimulated genes and their antiviral effector functions." Current Opinion in Virology 1: 519-525.

Schoggins, J. W., S. J. Wilson, M. Panis, M. Y. Murphy, C. T. Jones, P. Bieniasz and C. M. Rice (2011). "A diverse range of gene products are effectors of the type I interferon antiviral response." Nature 472(7344): 481-485.

Schoneboom, B. A., J. S. Lee and F. B. Grieder (2000). "Early expression of IFN- alpha/beta and iNOS in the brains of Venezuelan equine encephalitis virus-infected mice." Journal of interferon & cytokine research 20(2): 205-216.

Schuffenecker, I., I. Iteman, A. Michault, S. Murri, L. Frangeul, M. C. Vaney, R. Lavenir, N. Pardigon, J. M. Reynes and F. Pettinelli (2006). "Genome microevolution of chikungunya viruses causing the Indian Ocean outbreak." PLoS medicine 3(7): 1058-1070.

Schulte, L. N., A. Eulalio, H. J. Mollenkopf, R. Reinhardt and J. Vogel (2011). "Analysis of the host microRNA response to Salmonella uncovers the control of major cytokines by the let-7 family." The EMBO journal 30(10): 1977-1989.

Schulz, O., S. S. Diebold, M. Chen, T. I. Näslund, M. A. Nolte, L. Alexopoulou, Y. T. Azuma, R. A. Flavell, P. Liljeström and C. R. e Sousa (2005). "Toll-like receptor 3 promotes cross-priming to virus-infected cells." Nature 433(7028): 887-892.

Schulz, O., A. Pichlmair, J. Rehwinkel, N. C. Rogers, D. Scheuner, H. Kato, O. Takeuchi, S. Akira, R. J. Kaufman and C. Reis e Sousa (2010). "Protein kinase R contributes to immunity against specific viruses by regulating interferon mRNA integrity." Cell Host & Microbe 7(5): 354-361.

Schwartz, O. and M. L. Albert (2010). "Biology and pathogenesis of chikungunya virus." Nature Reviews Microbiology 8(7): 491-500.

Sebastian, M. R., R. Lodha and S. Kabra (2009). "Chikungunya infection in children." Indian Journal of Pediatrics 76(2): 185-189.

Seehafer, S. S. and D. A. Pearce (2006). "You say lipofuscin, we say ceroid: defining autofluorescent storage material." Neurobiology of Aging 27(4): 576-588.

217 Seth, R. B., L. Sun, C. K. Ea and Z. J. Chen (2005). "Identification and characterization of MAVS, a mitochondrial antiviral signaling protein that activates NF-[kappa] B and IRF3." Cell 122(5): 669-682.

Sharma, A., P. Gupta and R. K. Maheshwari (2012). "Inactivation of Chikungunya virus by 1, 5 iodonapthyl azide." Virology Journal 9(1): 301.

Shenoy, A. R., B.-H. Kim, H.-P. Choi, T. Matsuzawa, S. Tiwari and J. D. MacMicking (2008). "Emerging themes in IFN-γ-induced macrophage immunity by the p47 and p65 GTPase families." Immunobiology 212(9): 771-784.

Shi, M., W. Deng, E. Bi, K. Mao, Y. Ji, G. Lin, X. Wu, Z. Tao, Z. Li and X. Cai (2008). "TRIM30α negatively regulates TLR-mediated NF-κB activation by targeting TAB2 and TAB3 for degradation." Nature Immunology 9(4): 369-377.

Shi, S.-R., Y. Shi and C. R. Taylor (2011). "Antigen retrieval immunohistochemistry review and future prospects in research and diagnosis over two decades." Journal of Histochemistry & Cytochemistry 59(1): 13-32.

Shimizu, S., T. Takehara, H. Hikita, T. Kodama, T. Miyagi, A. Hosui, T. Tatsumi, H. Ishida, T. Noda and H. Nagano (2010). "The let-7 family of microRNAs inhibits Bcl-xL expression and potentiates sorafenib-induced apoptosis in human hepatocellular carcinoma." Journal of hepatology 52(5): 698-704.

Shrinet, J., S. Jain, J. Jain, R. K. Bhatnagar and S. Sunil (2014). "Next generation sequencing reveals regulation of distinct Aedes microRNAs during chikungunya virus development." PLoS neglected tropical diseases 8(1): e2616.

Siegal, F. P., N. Kadowaki, M. Shodell, P. A. Fitzgerald-Bocarsly, K. Shah, S. Ho, S. Antonenko and Y. J. Liu (1999). "The nature of the principal type 1 interferon-producing cells in human blood." Science 284(5421): 1835-1837.

Silverman, R. H. (2007). "Viral encounters with 2′, 5′-oligoadenylate synthetase and RNase L during the interferon antiviral response." Journal of Virology 81(23): 12720-12729.

Simmons, C. P., S. Popper, C. Dolocek, T. N. B. Chau, M. Griffiths, N. T. P. Dung, T. H. Long, D. M. Hoang, N. V. Chau and L. T. T. Thao (2007). "Patterns of Host Genome— Wide Gene Transcript Abundance in the Peripheral Blood of Patients with Acute Dengue Hemorrhagic Fever." Journal of Infectious Diseases 195(8): 1097-1107.

218 Simon, F., E. Javelle, M. Oliver, I. Leparc-Goffart and C. Marimoutou (2011). "Chikungunya Virus Infection." Current Infectious Disease Reports 13: 218-228.

Singh, S. K. and S. K. Unni (2011). "Chikungunya virus: host pathogen interaction." Reviews in Medical Virology 21: 78-88.

Slack, J. L., K. Schooley, T. P. Bonnert, J. L. Mitcham, E. E. Qwarnstrom, J. E. Sims and S. K. Dower (2000). "Identification of two major sites in the type I interleukin-1 receptor cytoplasmic region responsible for coupling to pro-inflammatory signaling pathways." Journal of Biological Chemistry 275(7): 4670-4678.

Smith-Norowitz, T. A., R. A. Sobel and F. Mokhtarian (2000). "B cells and antibodies in the pathogenesis of myelin injury in Semliki Forest virus encephalomyelitis." Cellular Immunology 200(1): 27-35.

Smith, T. J., R. H. Cheng, N. H. Olson, P. Peterson, E. Chase, R. J. Kuhn and T. S. Baker (1995). "Putative receptor binding sites on alphaviruses as visualized by cryoelectron microscopy." Proceedings of the National Academy of Sciences 92(23): 10648-10652.

Soden, M., H. Vasudevan, B. Roberts, R. Coelen, G. Hamlin, S. Vasudevan and J. La Brooy (2000). "Detection of viral ribonucleic acid and histologic analysis of inflamed synovium in Ross River virus infection." Arthritis & Rheumatism 43(2): 365-369.

Solignat, M., B. Gay, S. Higgs, L. Briant and C. Devaux (2009). "Replication cycle of chikungunya: a re-emerging arbovirus." Virology 393(2): 183-197.

Sorgeloos, F., M. Kreit, P. Hermant, C. Lardinois and T. Michiels (2013). "Antiviral type I and type III interferon responses in the central nervous system." Viruses 5(3): 834-857.

Sourisseau, M., C. Schilte, N. Casartelli, C. Trouillet, F. Guivel-Benhassine, D. Rudnicka, N. Sol-Foulon, K. Le Roux, M. C. Prevost and H. Fsihi (2007). "Characterization of reemerging chikungunya virus." PLoS Pathogens 3(6): e89.

Spuul, P., A. Salonen, A. Merits, E. Jokitalo, L. Kaariainen and T. Ahola (2007). "Role of the amphipathic peptide of Semliki forest virus replicase protein nsP1 in membrane association and virus replication." Journal of Virology 81(2): 872-883.

Spyr, C. A., F. Käsermann and C. Kempf (1995). "Identification of the pore forming element of Semliki Forest virus spikes." FEBS Letters 375(1-2): 134-136.

219 Steinhauer, D. and J. Holland (1987). "Rapid evolution of RNA viruses." Annual Reviews in Microbiology 41(1): 409-431.

Stirnweiss, A., A. Ksienzyk, K. Klages, U. Rand, M. Grashoff, H. Hauser and A. Kröger (2010). "IFN regulatory factor-1 bypasses IFN-mediated antiviral effects through viperin gene induction." The Journal of Immunology 184(9): 5179-5185.

Storey, J. D. (2002). "A direct approach to false discovery rates." Journal of the Royal Statistical Society: Series B (Statistical Methodology) 64(3): 479-498.

Strauss, E. G., E. M. Lenches and J. H. Strauss (2002). "Molecular genetic evidence that the hydrophobic anchors of glycoproteins E2 and E1 interact during assembly of alphaviruses." Journal of Virology 76(20): 10188-10194.

Strauss, J. H. and E. G. Strauss (1994). "The alphaviruses: gene expression, replication, and evolution." Microbiological Reviews 58(3): 491-562.

Sudeep, A. and D. Parashar (2008). "Chikungunya: an overview." Journal of biosciences 33(4): 443-449.

Suhrbier, A., M. C. Jaffar-Bandjee and P. Gasque (2012). "Arthritogenic alphaviruses—an overview." Nature Reviews Rheumatology 8: 420-429.

Suhrbier, A. and M. La Linn (2004). "Clinical and pathologic aspects of arthritis due to Ross River virus and other alphaviruses." Current Opinion in Rheumatology 16(4): 374- 379.

Suhrbier, A. and S. Mahalingam (2009). "The immunobiology of viral arthritides." Pharmacology & Therapeutics 124(3): 301-308.

Swan, D., P. D'eustachio, L. Leinwand, J. Seidman, D. Keithley and F. Ruddle (1979). "Chromosomal assignment of the mouse kappa light chain genes." Proceedings of the National Academy of Sciences 76(6): 2735-2739.

Swann, J. B., Y. Hayakawa, N. Zerafa, K. C. Sheehan, B. Scott, R. D. Schreiber, P. Hertzog and M. J. Smyth (2007). "Type I IFN contributes to NK cell , activation, and antitumor function." The Journal of Immunology 178(12): 7540-7549.

Swenson, E. S., J. G. Price, T. Brazelton and D. S. Krause (2007). "Limitations of green fluorescent protein as a cell lineage marker." Stem Cells 25(10): 2593-2600.

220 Takemori, T. and K. Rajewsky (1981). "Lambda chain expression at different stages of ontogeny in C57BL/6, BALB/c and SJL mice." European Journal of Immunology 11(8): 618-625.

Takeuchi, O. and S. Akira (2010). "Pattern recognition receptors and inflammation." Cell 140(6): 805-820.

Tamura, K., J. Dudley, M. Nei and S. Kumar (2007). "MEGA4: molecular evolutionary genetics analysis (MEGA) software version 4.0." Molecular biology and evolution 24(8): 1596-1599.

Taubitz, W., J. P. Cramer, A. Kapaun, M. Pfeffer, C. Drosten, G. Dobler, G. D. Burchard and T. Löscher (2007). "Chikungunya fever in travelers: clinical presentation and course." Clinical infectious diseases 45(1): e1-4.

Taylor, G. A., C. M. Collazo, G. S. Yap, K. Nguyen, T. A. Gregorio, L. S. Taylor, B. Eagleson, L. Secrest, E. A. Southon and S. W. Reid (2000). "Pathogen-specific loss of host resistance in mice lacking the IFN-γ-inducible gene IGTP." Proceedings of the National Academy of Sciences 97(2): 751-755.

Taylor, G. A., C. G. Feng and A. Sher (2004). "p47 GTPases: regulators of immunity to intracellular pathogens." Nature Reviews Immunology 4(2): 100-109.

Tellinghuisen, T. L., A. E. Hamburger, B. R. Fisher, R. Ostendorp and R. J. Kuhn (1999). "In vitro assembly of alphavirus cores by using nucleocapsid protein expressed in Escherichia coli." Journal of Virology 73(7): 5309-5319.

Teng, T.-S., S.-S. Foo, D. Simamarta, F.-M. Lum, T.-H. Teo, A. Lulla, N. K. Yeo, E. G. Koh, A. Chow and Y.-S. Leo (2012). "Viperin restricts chikungunya virus replication and pathology." The Journal of clinical investigation 122(12): 4447-4460.

Teo, T.-H., Z. Her, J. J. Tan, F.-M. Lum, W. W. Lee, Y.-H. Chan, R.-Y. Ong, Y.-W. Kam, I. Leparc-Goffart and P. Gallian (2015). "Caribbean and La Réunion Chikungunya virus isolates differ in their capacity to induce pro-inflammatory Th1 and NK cell responses and acute joint pathology." Journal of virology: JVI. 00909-00915.

Teo, T.-H., F.-M. Lum, C. Claser, V. Lulla, A. Lulla, A. Merits, L. Rénia and L. F. Ng (2013). "A pathogenic role for CD4+ T cells during Chikungunya virus infection in mice." The Journal of Immunology 190(1): 259-269.

221 Tesh, R. B., D. M. Watts, K. L. Russell, C. Damodaran, C. Calampa, C. Cabezas, G. Ramirez, B. Vasquez, C. G. Hayes and C. A. Rossi (1999). "Mayaro virus disease: an emerging mosquito-borne zoonosis in tropical South America." Clinical infectious diseases 28(1): 67-73.

The Australia Group (2014). "Common control list handbook." 2: 259.

Thompson, J. M., A. C. Whitmore, H. F. Staats and R. Johnston (2008). "The contribution of type I interferon signaling to immunity induced by alphavirus replicon vaccines." Vaccine 26(39): 4998-5003.

Thompson, K. A. S., G. A. Rempala and J. Yin (2009). "Multiple-hit inhibition of infection by defective interfering particles." Journal of General Virology 90(4): 888-899.

Thorvaldsdóttir, H., J. T. Robinson and J. P. Mesirov (2012). "Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration." Briefings in Bioinformatics: bbs017.

Tiwari, S. and J. D. MacMicking (2008). Bacterial Phagosome Acidification Within IFN-γ- Activated Macrophages: Role of Host p47 Immunity-Related GTPases IRGs). Autophagosome and Phagosome, Springer: 407-415.

Tomar, S., R. W. Hardy, J. L. Smith and R. J. Kuhn (2006). "Catalytic core of alphavirus nonstructural protein nsP4 possesses terminal adenylyltransferase activity." Journal of Virology 80(20): 9962-9969.

Trapnell, C., L. Pachter and S. L. Salzberg (2009). "TopHat: discovering splice junctions with RNA-Seq." Bioinformatics 25(9): 1105-1111.

Trapnell, C., A. Roberts, L. Goff, G. Pertea, D. Kim, D. R. Kelley, H. Pimentel, S. L. Salzberg, J. L. Rinn and L. Pachter (2012). "Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks." Nature Protocols 7(3): 562- 578.

Trapnell, C., B. A. Williams, G. Pertea, A. Mortazavi, G. Kwan, M. J. van Baren, S. L. Salzberg, B. J. Wold and L. Pachter (2010). "Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation." Nature Biotechnology 28(5): 511-515.

222 Triantafilou, K. and M. Triantafilou (2004). "Coxsackievirus B4-induced cytokine production in pancreatic cells is mediated through toll-like receptor 4." Journal of Virology 78(20): 11313-11320.

Trinchieri, G. (2010). "Type I interferon: friend or foe?" The Journal of Experimental Medicine 207(10): 2053-2063.

Tsetsarkin, K. A., D. L. Vanlandingham, C. E. McGee and S. Higgs (2007). "A single mutation in chikungunya virus affects vector specificity and epidemic potential." PLoS Pathogens 3(12): 1895-1906.

Tsien, R. Y. (1998). "The green fluorescent protein." Annual review of biochemistry 67(1): 509-544.

Twine, N. A., K. Janitz, M. R. Wilkins and M. Janitz (2011). "Whole transcriptome sequencing reveals gene expression and splicing differences in brain regions affected by Alzheimer's disease." PloS One 6(1): e16266.

Valamparampil, J. J., S. Chirakkarot, S. Letha, C. Jayakumar and K. Gopinathan (2009). "Clinical profile of Chikungunya in infants." Indian journal of pediatrics 76(2): 151-155.

Van Bortel, W., F. Dorleans, J. Rosine, A. Blateau, D. Rousset, S. Matheus, I. Leparc- Goffart, O. Flusin, C. Prat and R. Cesaire (2014). "Chikungunya outbreak in the Caribbean region, December 2013 to March 2014, and the significance for Europe." Euro Surveill 19(13): 20759.

Vecchi, A., L. Massimiliano, S. Ramponi, W. Luini, S. Bernasconi, R. Bonecchi, P. Allavena, M. Parmentier, A. Mantovani and S. Sozzani (1999). "Differential responsiveness to constitutive vs. inducible chemokines of immature and mature mouse dendritic cells." Journal of leukocyte biology 66(3): 489-494.

Ventoso, I., M. A. Sanz, S. Molina, J. J. Berlanga, L. Carrasco and M. Esteban (2006). "Translational resistance of late alphavirus mRNA to eIF2α phosphorylation: a strategy to overcome the antiviral effect of protein kinase PKR." Genes & development 20(1): 87-100.

Villanueva, M. T. (2014). "Tumorigenesis: Establishing the origin of retinoblastoma." Nature Reviews Cancer 14(11): 706-707.

223 Vitour, D., V. Doceul, S. Ruscanu, E. Chauveau, I. Schwartz-Cornil and S. Zientara (2014). "Induction and control of the type I interferon pathway by Bluetongue virus." Virus research 182: 59-70.

W.H.O (2009). Dengue: guidelines for diagnosis, treatment, prevention and control - New edition. Geneva, Switzerland, WHO Library Cataloguing-in-Publication Data.

W.H.O (2009). "World Health Organisation. Guidelines on clinical management of chikungunya fever." Regional Office for South-East Asia 2009 Edition: 1-19.

W.H.O. (2015). "Global Alert and Response (GAR): Impact of Dengue." from http://www.who.int/csr/disease/dengue/impact/en/.

Wall, J. D., L. F. Tang, B. Zerbe, M. N. Kvale, P.-Y. Kwok, C. Schaefer and N. Risch (2014). "Estimating genotype error rates from high-coverage next-generation sequence data." Genome research: gr. 168393.168113.

Wang, D., A. Suhrbier, A. Penn-Nicholson, J. Woraratanadharm, J. Gardner, M. Luo, T. T. Le, I. Anraku, M. Sakalian, D. Einfeld and J. Y. Dong (2011). "A complex adenovirus vaccine against chikungunya virus provides complete protection against viraemia and arthritis." Vaccine 29(15): 2803-2809.

Wang, E., D. Y. Kim, S. C. Weaver and I. Frolov (2011). "Chimeric Chikungunya Viruses are Nonpathogenic in Highly Sensitive Mouse Models, but Efficiently Induce a Protective Immune Response." Journal of Virology 85(17): 9249-9252.

Wang, E., E. Volkova, A. P. Adams, N. Forrester, S. Y. Xiao, I. Frolov and S. C. Weaver (2008). "Chimeric alphavirus vaccine candidates for chikungunya." Vaccine 26(39): 5030- 5039.

Wang, J. P., P. Liu, E. Latz, D. T. Golenbock, R. W. Finberg and D. H. Libraty (2006). "Flavivirus activation of plasmacytoid dendritic cells delineates key elements of TLR7 signaling beyond endosomal recognition." The Journal of Immunology 177(10): 7114- 7121.

Wang, K. S., R. Kuhn, E. Strauss, S. Ou and J. Strauss (1992). "High-affinity laminin receptor is a receptor for Sindbis virus in mammalian cells." Journal of Virology 66(8): 4992-5001.

224 Wang, S., Y. Tang, H. Cui, X. Zhao, X. Luo, W. Pan, X. Huang and N. Shen (2011). "Let- 7/miR-98 regulate Fas and Fas-mediated apoptosis." Genes and Immunity 12(2): 149-154.

Wang, X., E. R. Hinson and P. Cresswell (2007). "The interferon-inducible protein viperin inhibits influenza virus release by perturbing lipid rafts." Cell Host & Microbe 2(2): 96-105.

Wang, Y., S. G. Sawicki and D. L. Sawicki (1991). "Sindbis virus nsP1 functions in negative-strand RNA synthesis." Journal of Virology 65(2): 985-988.

Wang, Y. F., S. G. Sawicki and D. L. Sawicki (1994). "Alphavirus nsP3 functions to form replication complexes transcribing negative-strand RNA." Journal of Virology 68(10): 6466- 6475.

Wang, Z., M. Gerstein and M. Snyder (2009). "RNA-Seq: a revolutionary tool for transcriptomics." Nature Reviews Genetics 10(1): 57-63.

Warter, L., C. Y. Lee, R. Thiagarajan, M. Grandadam, S. Lebecque, R. T. Lin, S. Bertin- Maghit, L. F. Ng, J.-P. Abastado and P. Desprès (2011). "Chikungunya virus envelope- specific human monoclonal antibodies with broad neutralization potency." The Journal of Immunology 186(5): 3258-3264.

Wauquier, N., P. Becquart, D. Nkoghe, C. Padilla, A. Ndjoyi-Mbiguino and E. M. Leroy (2011). "The acute phase of chikungunya virus infection in humans is associated with strong innate immunity and T CD8 cell activation." Journal of Infectious Diseases 204(1): 115-123.

Way, S. J. R., B. A. Lidbury and J. L. Banyer (2002). "Persistent Ross River virus infection of murine macrophages: an in vitro model for the study of viral relapse and immune modulation during long-term infection." Virology 301(2): 281-292.

Weber, C., G. Draude, K. S. Weber, J. Wübert, R. L. Lorenz and P. C. Weber (1999). "Downregulation by tumor necrosis factor-α of monocyte CCR2 expression and monocyte chemotactic protein-1-induced transendothelial migration is antagonized by oxidized low- density lipoprotein: A potential mechanism of monocyte retention in atherosclerotic lesions." Atherosclerosis 145(1): 115-123.

Wengler, G. and C. Gros (1996). "Analyses of the role of structural changes in the regulation of uncoating and assembly of alphavirus cores." Virology 222(1): 123-132.

225 Wengler, G. and D. Würkner (1992). "Identification of a sequence element in the alphavirus core protein which mediates interaction of cores with ribosomes and the disassembly of cores." Virology 191(2): 880-888.

Wensink, A. C., C. E. Hack and N. Bovenschen (2015). "Granzymes Regulate Proinflammatory Cytokine Responses." The Journal of Immunology 194(2): 491-497.

Werneke, S. W., C. Schilte, A. Rohatgi, K. J. Monte, A. Michault, F. Arenzana-Seisdedos, D. L. Vanlandingham, S. Higgs, A. Fontanet and M. L. Albert (2011). "ISG15 is critical in the control of Chikungunya virus infection independent of UbE1L mediated conjugation." PLoS Pathogens 7(10): e1002322.

Werner, M., A. Chott, A. Fabiano and H. Battifora (2000). "Effect of formalin tissue fixation and processing on immunohistochemistry." The American Journal of Surgical Pathology 24(7): 1016-1019.

Westermann, A. J., S. A. Gorski and J. Vogel (2012). "Dual RNA-seq of pathogen and host." Nature Reviews Microbiology 10(9): 618-630.

White, C. L., M. Thomson and N. J. Dimmock (1998). "Deletion analysis of a defective interfering Semliki Forest virus RNA genome defines a region in the nsP2 sequence that is required for efficient packaging of the genome into virus particles." Journal of Virology 72(5): 4320-4326.

White, L. K., T. Sali, D. Alvarado, E. Gatti, P. Pierre, D. Streblow and V. R. DeFilippis (2011). "Chikungunya virus induces IPS-1-dependent innate immune activation and protein kinase R-independent translational shutoff." Journal of Virology 85(1): 606-620.

Wilkins, C. and M. Gale Jr (2010). "Recognition of viruses by cytoplasmic sensors." Current Opinion in Immunology 22(1): 41-47.

Wilming, L., et al. (2008). "The vertebrate genome annotation (Vega) database." Nucleic Acids Research 36: D753-D760.

Wilson, J. R., P. F. de Sessions, M. A. Leon and F. Scholle (2008). "West Nile virus nonstructural protein 1 inhibits TLR3 signal transduction." Journal of Virology 82(17): 8262- 8271.

Woelk, C. H., S. D. Frost, D. D. Richman, P. E. Higley and S. L. K. Pond (2007). "Evolution of the interferon alpha gene family in eutherian mammals." Gene 397(1): 38-50.

226 Woloschak, G. E. and C. J. Krco (1987). "Regulation of kappa/lambda immunoglobulin light chain expression in normal murine lymphocytes." Molecular Immunology 24(7): 751- 757.

Wu, J. Q. H., D. Huang, E. Chan-Wong and D. Chau (2008). "Inhibition of the replication of Venezuelan equine encephalitis virus by retinoic acid-inducible gene-I." Defence Research & Developement Canada: Technical Memorandum. 213(213):396-402.

Wu, X., et al. (2013). "Evasion of antiviral immunity through sequestering TBK1/IKKB/IRF3 into viral inclusion bodies." Journal of Virology 88(6): 3067-3076.

Wunder, C., R. W. Brock, A. Krug, N. Roewer and O. Eichelbrönner (2005). "A remission spectroscopy system for in vivo monitoring of hemoglobin oxygen saturation in murine hepatic sinusoids, in early systemic inflammation." Comparative Hepatology 4(1): 1476- 5926.

Xu, C., Ø. Evensen and H. M. Munang’andu (2015). "De novo assembly and transcriptome analysis of Atlantic salmon macrophage/dendritic-like TO cells following type I IFN treatment and Salmonid alphavirus subtype-3 infection." BMC genomics 16(1): 96.

Yamamoto, M., S. Sato, H. Hemmi, K. Hoshino, T. Kaisho, H. Sanjo, O. Takeuchi, M. Sugiyama, M. Okabe and K. Takeda (2003). "Role of adaptor TRIF in the MyD88- independent toll-like receptor signaling pathway." Science's STKE 301(5633): 640-643.

Yamamoto, M., S. Sato, H. Hemmi, H. Sanjo, S. Uematsu, T. Kaisho, K. Hoshino, O. Takeuchi, M. Kobayashi and T. Fujita (2002). "Essential role for TIRAP in activation of the signalling cascade shared by TLR2 and TLR4." Nature 420(6913): 324-329.

Yoneyama, M. and T. Fujita (2010). "Recognition of viral nucleic acids in innate immunity." Reviews in Medical Virology 20(1): 4-22.

Yoneyama, M., M. Kikuchi, K. Matsumoto, T. Imaizumi, M. Miyagishi, K. Taira, E. Foy, Y. M. Loo, M. Gale Jr and S. Akira (2005). "Shared and unique functions of the DExD/H-box helicases RIG-I, MDA5, and LGP2 in antiviral innate immunity." The Journal of Immunology 175(5): 2851-2858.

Zhai, J., D. Gao, W. Liu, R. Hong, Y. Qin, H. Ouyang, Y. Kong, Y. Wang, Y. Xie and J. Liu (2008). "Characterization of a novel isoform of murine interferon regulatory factor 3." Biochemical and Biophysical Research Communications 377(2): 384-388.

227 Zhang, G.-f., X.-g. Chen, P.-q. Shuai and L. Lin (2010). "Effect of Cell Fixative on GFP Luminous Characteristics." Chemistry & Bioengineering 12: 014.

Zhang, W., M. Heil, R. J. Kuhn and T. S. Baker (2005). "Heparin binding sites on Ross River virus revealed by electron cryo-microscopy." Virology 332(2): 511-518.

Zhang, Y., C. W. Burke, K. D. Ryman and W. B. Klimstra (2007). "Identification and characterization of interferon-induced proteins that inhibit alphavirus replication." Journal of Virology 81(20): 11246-11255.

Zhao, M., J. Zhang, H. Phatnani, S. Scheu and T. Maniatis (2012). "Stochastic expression of the interferon-β gene." PLoS Biology 10(1): e1001249.

Zhou, S., A. Halle, E. A. Kurt-Jones, A. M. Cerny, E. Porpiglia, M. Rogers, D. T. Golenbock and R. W. Finberg (2008). "Lymphocytic choriomeningitis virus (LCMV) infection of CNS glial cells results in TLR2-MyD88/Mal-dependent inflammatory responses." Journal of Neuroimmunology 194(1): 70-82.

Ziegler, S. A., L. Lu, A. da Rosa, T. PA, S. Y. Xiao and R. B. Tesh (2008). "An animal model for studying the pathogenesis of chikungunya virus infection." The American Journal of Tropical Medicine and Hygiene 79(1): 133-139.

Zuckerman, A. J., J. E. Banatvala, B. D. Schoub, P. Griffiths and P. Mortimer, Eds. (2009). Principles and Practice of Clinical Virology. Alphaviruses. Oxford, Wiley-Blackwell

228 9. Supplemental Data

229

Figure S1. IRF3 amino acid sequence alignment. Amino acid sequences of the human, mouse and novel IRF3 isoforms were aligned using MEGA v6.0 software. Conservation threshold was 60%. Sequence alignments were rendered using consensus– colouring of multiple alignments (CHROMA). Red arrows indicate the DNA-binding domain (D), the nuclear export signal (NES) and the activation domain (AD).

230 Table S1. Protein coding features in up-regulated, unannotated genes from top 50 DEG lists.

Log2 (fold Transcript Longest Top BLASTx Foot D7 v uORFs uAUGs e value BLASTx description Similarity (%) Other observations change) length (bp) ORF (aa) accession PREDICTED: hypothetical protein CINT1 v1 Inf 694 78 1 3 XP_001924410.2 5.00E-10 76 ID; 25 Cov LOC100156886 [Sus scrofa] L-lactate dehydrogenase A chain CINT2 v1 5.84 502 33 0 13 ERE78392.1 1.40E+00 64 ID; 4 cov [Cricetulus griseus] v2 776 22 5 12 nothing v3 762 33 1 6 nothing v4 1006 22 2 8 nothing v5 1938 107 2 7 nothing v6 2104 107 5 19 nothing v7 388 22 1 3 nothing v8 1389 33 2 8 nothing v9 1276 46 2 7 nothing CINT3 v1 Inf 572 53 0 7 EDL16837.1 2.00E-15 mCG147594 [Mus musculus] 100 ID; 19 cov WP_017731516. hypothetical protein [Nafulsella contains stop codons, BLAST hit CINT4 v1 Inf 389 56 0 3 3.80E+00 32 ID; 50 cov 1 turpanensis] in wrong frame

Foot D2

CINT5 v1 Inf 586 42 1 2 nothing CINT6 v1 Inf 319 60 0 0 nothing no 5'UTR CINT7 v1 8.01 79 2 7 EDL79940.1 9.00E-09 rCG26676 [Rattus norvegicus] 54 ID; 3 cov PREDICTED: Hhex-like protein [Cricetulus CINT8 v1 Inf 2053 89 1 2 XP_003515706.1 5.00E-27 87 ID; 10 cov contains 2 stop codons griseus] LND2 PREDICTED: uncharacterized protein CINT9 v1 7.71 737 165 0 0 XP_914999.1 3.00E-46 84 ID, 47 cov contains LU domain LOC546643 [Mus musculus] PREDICTED: lymphocyte antigen 6I-like v2 750 165 0 0 XP_914999.1 5.00E-46 84 ID; 36 cov [Mus musculus] CINT10 v1 5.57 4435 108 15 42 EDL87013.1 1.60E+00 rCG63130 [Rattus norvegicus] 42 ID; 5 cov BLAST hit in wrong frame hypothetical protein [Plasmodium berghei v2 2283 138 6 15 XP_673000.1 4.00E-28 78 ID; 9 cov BLAST hit in wrong frame strain ANKA] PREDICTED: tripartite motif-containing Contains SPRY superfamily CINT11 v1 4.59 4147 125 3 14 XP_485980.5 1.00E-138 100 ID; 15 cov protein 30A [Mus musculus] domain PREDICTED: tripartite motif-containing Contains SPRY superfamily v2 2232 224 0 0 XP_485980.5 0.00E+00 100 ID; 34 cov protein 30A [Mus musculus] domain mCG140494, isoform CRA_a [Mus CINT22 v1 3.74 1867 32 5 9 EDL28658.1 6.00E-40 100 ID, 13 cov musculus]

231 Table S1. Protein coding features in up-regulated, unannotated genes from top 50 DEG lists (cont.)

Log2 (fold Transcript Longest Top BLASTx LND7 v uORFs uAUGs e value BLASTx description Similarity (%) Other observations change) length (bp) ORF (aa) accession CINT12 v1 Inf 336 36 0 0 WP_022020858. 5.90E+00 ftsK/SpoIIIE family protein [Bacteroides sp. 29 ID, 88 cov contains stop codons 1 CAG 661] CINT13 v1 4.40 2163 132 0 0 AAI51696.1 2.00E-139 RPLP0 protein [Bos taurus] 78 ID; 37 cov BLAST hit in wrong frame PREDICTED: uncharacterized protein 100 ID, 100 Contains LU superfamily CINT14 v1 4.17 804 134 0 0 XP_914999.1 6.00E-81 LOC546643 [Mus musculus] cov domain CINT15 v1 1543 93 0 3 EDL05858.1 6.00E-51 mCG144566 [Mus musculus] 64 ID; 35 cov Contains HMG box domain, v2 969 93 0 2 v3 2834 69 12 29 EDL08408.1 3.00E-14 mCG147230 [Mus musculus] 80 ID; 4 cov BLAST hit in wrong frame v4 1647 85 6 11 AAH31435.1 5.00E-14 Chpt1 protein [Mus musculus] 86 ID; 7 cov BLAST hit in wrong frame E3 ubiquitin-protein ligase NEDD4 BLAST hit in wrong frame, v5 742 85 0 2 EGW01477.1 4.30E+00 55 ID; 16 cov [Cricetulus griseus] contains stop codons BLAST hit in wrong frame, CINT16 v1 4.01 5751 145 8 28 ERE74287.1 2.00E-07 caltractin-like protein [Cricetulus griseus] 76 ID; 2 cov contains stop codons, no 5'UTR

CINT17 v1 3.62 3975 105 7 29 nothing v2 642 35 0 0 nothing CINT18 v1 3.56 2676 81 0 3 ERE74288.1 5.00E-09 E3 ubiquitin-protein ligase [Cricetulus 63 ID; 3 cov BLAST hit in wrong frame E3 ubiquitin-protein ligase [Cricetulus v2 867 89 0 3 ERE74288.1 2.00E-09 63id, 6 cov BLAST hit in wrong frame griseus] v3 5501 104 4 18 EDL09413.1 3.00E-26 mCG147326 [Mus musculus] 82 ID; 4 cov BLAST hit in wrong frame CINT19 v1 3.35 4776 107 5 12 NP_001041357.1 3.00E-17 uncharacterized protein LOC317165 52 ID; 5 cov [Rattus norvegicus] CINT20 v1 3.33 2226 168 3 14 EDK97113.1 0.00E+00 mCG145673, isoform CRA_a [Mus 100 ID; 13 cov BLAST hit in wrong frame musculus] >gb|AFJ59921.1| stem loop binding BLAST hit in wrong frame, no 5' v2 1274 80 2 NP_001257345.1 1.00E-55 57 ID; 44 cov protein 2 [Rattus norvegicus] UTR CINT21 v1 2.92 887 110 1 1 nothing V: variant/isoforms; LN: lymph node; uORF: upstream open reading frame; uAUG: upstream start codon; CINT: CHIKV-induced novel transcript.

232

Figure S3. CINT9 isoform mRNA and protein sequences. A) The full mRNA sequences of the two CINT9 variants are shown. The open reading frame is indicated by bold, uppercase font. The presence of the Kozak sequence is indicated by underlined text. B) The translated amino acid sequence encoded by the two CINT9 mRNA isoforms. mRNA sequences were translate using ExPASY Translate Tool.

233

FigureS2. Alignment of full Ly6 antigen amino acid sequences and the full predicted protein sequence of CINT9. The conserved LU domain is indicated (red arrows). Alignment constructed using MEGA v6.0 software. Conservation threshold was 60%. Sequence alignments were rendered using consensus–colouring of multiple alignments (CHROMA).

234