Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The use of genetic sequencing technologies to determine HIV-1 viral tropism and to evaluate the effects… McGovern, Rachel Ann 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_november_mcgovern_rachel.pdf [ 5.47MB ]
Metadata
JSON: 24-1.0166738.json
JSON-LD: 24-1.0166738-ld.json
RDF/XML (Pretty): 24-1.0166738-rdf.xml
RDF/JSON: 24-1.0166738-rdf.json
Turtle: 24-1.0166738-turtle.txt
N-Triples: 24-1.0166738-rdf-ntriples.txt
Original Record: 24-1.0166738-source.json
Full Text
24-1.0166738-fulltext.txt
Citation
24-1.0166738.ris

Full Text

THE USE OF GENETIC SEQUENCING TECHNOLOGIES TO DETERMINE HIV-1 VIRAL TROPISM AND TO EVALUATE THE EFFECTS OF MARAVIROC ON PATIENT VIRAL POPULATIONS by Rachel Ann McGovern B.Sc. (Hons.), Lakehead University, 2006 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate and Postdoctoral Studies (Experimental Medicine) UNIVERSITY OF BRITISH COLUMBIA (Vancouver) September 2015 ©Rachel Ann McGovern, 2015 ii Abstract HIV-1 infection is reliant on the ability of the virus to enter target cells characterized by the expression of either the CCR5 or CXCR4 co-receptor at the cell surface.  It is now well established that the V3 loop of the HIV-1 envelope is the primary determinant of co-receptor use, and that genetic analysis of the V3 loop can be used to infer co-receptor use, or “tropism”.  This became clinically relevant with the development of the entry inhibitor, maraviroc (MVC), an anti-HIV compound designed to inhibit HIV-1 cell entry exclusively at the CCR5 co-receptor.  As such, the more pathogenic CXCR4-using variants are likely to thrive during MVC therapy.  Anti-HIV treatment guidelines now strongly suggest a tropism test be performed when considering the clinical use of MVC. There are two primary objectives discussed in this thesis, 1) the validation of a population-based sequencing tropism assay designed to predict virological response to MVC; 2) to apply a next generation sequencing tropism assay to investigate the selective pressures exerted by MVC on mixed tropic HIV-1 populations. The validation studies presented in this thesis demonstrate the reliability of a population-based sequencing assay to infer virological response to MVC in two large, multinational cohorts.  The results of these studies have promoted the worldwide expansion of this genotypic assay as a practical and affordable option for tropism inference.  In addition, a more intensive investigation into the selective pressures exerted by MVC on mixed tropic HIV-1 populations demonstrated the reproducible outgrowth of preexisting variants most genetically characteristic of CXCR4-using virus.  The results of these studies suggest MVC may be effective against a broader range of variants than previously thought.  In general, the four studies described in this thesis demonstrate the clinical utility of genetic sequencing when considering the use of MVC.  iii Preface The main body of the thesis features three peer-reviewed manuscripts published in the scientific literature, as well as a fourth manuscript not yet submitted for publication.  The three published manuscripts are reprinted in this thesis with permission granted from the respective copyright holders.  The candidate is primary author on all four manuscripts, responsible for composing the entirety of each.  In addition to reporting the findings of each study to the scientific community, the candidate actively participated in the research described in these manuscripts, including study design, laboratory generation of data and data analysis. Co-authors at the candidate’s research laboratory include senior laboratory research assistants Theresa Mo and Winnie Dong; data analysts Chanson Brumme, Art Poon, Conan Woods and Xiaoyin Zhong as well as the candidate’s graduate supervisor, Dr. Richard Harrigan.  There are a number of external collaborators associated with the studies from which samples used for the retrospective analyses presented in the thesis were taken. A version of chapter 2 has previously been published and reproduced here.  Rachel A McGovern, Thielen A, Mo T, Dong W, Woods CK, Chapman D, Lewis M, James I, Heera J, Valdez H, Harrigan PR. (2010) Population-based V3 genotypic tropism assay: a retrospective analysis using screening samples from the A4001029 and MOTIVATE studies.  AIDS 24:2517-2525.  I am the primary author of this manuscript, as such the publisher of this journal does not require copyright permission to be granted for reproduction in a doctoral dissertation.  ©Wolters Kluwer Health/Lippincott Williams & Wilkins. A version of chapter 3 has also been published in the scientific literature and reproduced here.  Rachel A McGovern, Thielen A, Portsmouth S, Mo T, Dong W, Woods CK, Zhong X, Brumme CJ, Chapman D, Lewis M, James I, Heera J, Valdez H, Harrigan PR. (2012) Population-based sequencing of the V3 loop can predict the virological response to maraviroc in treatment-naïve patients of the MERIT trial.  Journal of Acquired Immune Deficiency Syndromes 61(3):279-286.  I am the  iv primary author of this manuscript, as such the publisher of this journal does not require copyright permission to be granted for reproduction in a doctoral dissertation.  ©Lippincott Williams & Wilkins. Lastly, chapter 4 has been published and reproduced here with permission under RightsLink license number: 3532690277492.  Rachel A McGovern, Symons J, Poon AFY, Harrigan PR, van Lelyveld SFL, Hoepelman AIM, van Ham PM, Dong W, Wensing AMJ, Nijhuis M. (2013) Maraviroc treatment in non-R5-HIV-1-infected patients results in the selection of extreme CXCR4-using variants with limited effect on the total viral setpoint.  Journal of Antimicrobial Chemotherapy 68(9):2007-2014.  ©Oxford University Press. Ethical approval was granted by the Providence Health Care–University of British Columbia Research Ethics Board for the studies presented in Chapters 2 and 3, H07-01901, as well as that presented in Chapter 5, H10-00565.  v Table of Contents Abstract .......................................................................................................................................... ii Preface .......................................................................................................................................... iii Table of Contents ........................................................................................................................ v List of Tables ............................................................................................................................ viii List of Figures .............................................................................................................................. ix List of Abbreviations ................................................................................................................. xi Acknowledgements .................................................................................................................. xiv Dedication ................................................................................................................................... xv Chapter One:  A General Introduction and Thesis Objectives........................................... 1 1.1 The Human Immunodeficiency Virus .............................................................................. 1 1.1.1 The Identification of a Pandemic ................................................................................... 1 1.1.2 HIV Types, Groups & Clades ........................................................................................ 2 1.1.3 Cross-Species Transmission ........................................................................................... 4 1.1.4 Transmission among Humans ....................................................................................... 4 1.2 HIV the Virus ......................................................................................................................... 5 1.2.1 The Genome & the Virion............................................................................................... 5 1.2.2 The Replication Cycle ..................................................................................................... 9 1.2.2.1 Cellular Entry & Reverse Transcription ..................................................................... 9 1.2.2.2 Integration & Replication.......................................................................................... 10 1.2.2.3 Budding & Maturation ............................................................................................. 12 1.2.3 HIV Disease Progression & Pathogenesis .................................................................. 14 1.3 HIV Entry .............................................................................................................................. 17 1.3.1 HIV Cell Interactions..................................................................................................... 17 1.3.2 HIV Receptors & Co-Receptors ................................................................................... 20 1.3.3 Viral Structures for Cell Entry ..................................................................................... 21 1.3.4 The Entry Mechanism ................................................................................................... 23 1.3.5 The Chemokine Co-Receptors & HIV Infection ........................................................ 25 1.3.5.1 Tropism Terminology ................................................................................................ 25 1.3.5.2 Tropism Switch with Disease Progression ................................................................ 26 1.4 The Clinical Management of HIV .................................................................................... 27 1.4.1 The Concept of Antiretroviral Therapy ...................................................................... 27 1.4.2 The Antiretroviral Compounds ................................................................................... 29 1.4.3 The CCR5 Antagonists .................................................................................................. 31 1.4.3.1 Inhibition of HIV by Maraviroc ................................................................................ 32 1.4.3.2 The Clinical Trials of Maraviroc ............................................................................... 33 1.4.3.3 Maraviroc in the Presence of CXCR4-Using Virus .................................................. 35  vi 1.4.4 Maraviroc in the Clinic ................................................................................................. 36 1.4.4.1 Treatment Guidelines ................................................................................................ 36 1.4.4.2 Testing for Tropism Prior to Initiating Maraviroc ................................................... 37 1.5 Sequencing ........................................................................................................................... 39 1.5.1 Understanding HIV through Genetics ....................................................................... 39 1.5.2 Sequencing Technologies ............................................................................................. 41 1.5.2.1 Concepts in Sequencing ............................................................................................ 41 1.5.2.2 Population-based Sequencing .................................................................................... 43 1.5.2.3 Next Generation “Deep” Sequencing ....................................................................... 46 1.5.3 From Sequence to Relevance ........................................................................................ 52 1.5.4 Genotyping & Bioinformatics for Tropism Prediction ............................................. 53 1.6 Thesis Objectives & Organization ................................................................................... 56 1.6.1 Thesis Objectives ........................................................................................................... 56 1.6.2 Thesis Organization & Structure ................................................................................. 56 Chapter Two: Population-based V3 Genotypic Tropism Assay: a Retrospective Analysis Using Screening Samples from the A4001029 and MOTIVATE Studies .. 59 2.1 Introduction ........................................................................................................................ 59 2.2 Methods ................................................................................................................................. 61 2.2.1 Study Population ........................................................................................................... 61 2.2.2 Sequencing ...................................................................................................................... 62 2.2.3 Sequence Interpretation ................................................................................................ 63 2.2.4 Outcome Measures & Predictor Variables ................................................................ 64 2.2.5 Cutoff Point Optimization ............................................................................................ 65 2.3 Results ................................................................................................................................... 66 2.4 Discussion ............................................................................................................................. 79 Chapter Three: Population-based Sequencing of the V3 Loop Can Predict the Virological Response to Maraviroc in Treatment-Naïve Patients of the MERIT Trial 83 3.1 Introduction .......................................................................................................................... 83 3.2 Methods ................................................................................................................................. 85 3.2.1 Study Population ........................................................................................................... 85 3.2.2 Sequencing ...................................................................................................................... 86 3.2.3 Sequence Interpretation ................................................................................................ 87 3.2.4 Outcome Measures & Predictor Variables ................................................................. 87 3.3 Results ................................................................................................................................... 88 3.3.1 Subsequent Changes in Trofile Assay Results........................................................... 93 3.3.2 Comparison with ESTA ................................................................................................ 94 3.3.3 Assay Performance in Different HIV Clades ............................................................. 96 3.4 Discussion ............................................................................................................................. 97  vii Chapter Four: Maraviroc Treatment in Non-R5-HIV-1-Infected Patients Results in the Selection of Extreme CXCR4-using Variants with Limited Effects on the Total Viral Setpoint ..................................................................................................................................... 101 4.1 Introduction ........................................................................................................................ 101 4.2 Methods ............................................................................................................................... 103 4.2.1 Study Population ......................................................................................................... 103 4.2.2 Laboratory Methods .................................................................................................... 104 4.2.3 Sequence Processing & Tropism Interpretation ...................................................... 105 4.2.4 Fitness Analysis ........................................................................................................... 106 4.3 Results ................................................................................................................................. 106 4.4 Discussion ........................................................................................................................... 113 Chapter Five: Next Generation “Deep” Sequencing to Evaluate Viral Tropism in HIV-1 Patients Exposed to Short-term Maraviroc Add-on Therapy ....................................... 117 5.1 Introduction ........................................................................................................................ 117 5.2 Methods ............................................................................................................................... 119 5.2.1 Study Population ......................................................................................................... 119 5.2.2 Sample Preparation & “Deep” Sequencing ............................................................. 120 5.2.3 Sequence Processing ................................................................................................... 121 5.2.4 Bioinformatics Analyses ............................................................................................. 121 5.3 Results ................................................................................................................................. 123 5.4 Discussion ........................................................................................................................... 132 Chapter Six: General Discussion .......................................................................................... 138 6.1 Summary of Thesis............................................................................................................ 138 6.1.1 Validation of a Population-Based Genotypic Tropism Assay for Clinical Use .. 140 6.1.2 The Effects of Maraviroc on Non-R5 Virus Populations ........................................ 143 6.2 Contributions to the Field of HIV .................................................................................. 145 6.3 Future Directions ............................................................................................................... 147 6.4 Closing Remarks ................................................................................................................ 151 Bibliography ............................................................................................................................. 153 Appendix 1 ................................................................................................................................ 187 Chapter Three Supplementary Material ........................................................................... 187 Appendix 2 ................................................................................................................................ 196 Chapter Four Supplementary Material ............................................................................. 196 Appendix 3 ................................................................................................................................ 199 Chapter Five Supplementary Material .............................................................................. 199  viii List of Tables Table 2.1.  Percentage of patients with virological response to maraviroc at week eight. ...............69 Table 3.1.  Baseline characteristics for participants of the MERIT trial. ..............................................90 Table 3.2.  Results of rescreening the patient population of the MERIT trial for tropism. ...............91 Table 4.1.  Changes in the most prevalent sequence following exposure to maraviroc in patients screened as having non-R5 virus. ...........................................................................................................109 Table 5.1.  Baseline Characteristics Table (N=27). ................................................................................123 Table 5.2.  The most prevalent sequence at Days 0 and 7 of maraviroc therapy from patients screened as having ≥2% non-R5 virus at any time point. ...................................................................128  ix List of Figures Figure 1.1.  The HIV classification system................................................................................................ 3 Figure 1.2.  The HIV-1 genome. ................................................................................................................. 7 Figure 1.3.  The HIV-1 virion. .................................................................................................................... 8 Figure 1.4.  The HIV-1 replication cycle. .................................................................................................14 Figure 1.5.  HIV-1 target cells and latency. .............................................................................................17 Figure 1.6.  HIV-1 cell binding and membrane fusion. .........................................................................23 Figure 1.7.  Conformational changes during HIV-1 cell entry. ............................................................25 Figure 1.8.  The Sanger sequencing method. ..........................................................................................45 Figure 1.9.  Next generation 454 “deep” sequencing. ............................................................................49 Figure 2.1.  Virological response to maraviroc stratified by tropism test. ..........................................68 Figure 2.2.  Virological response to maraviroc stratified by optimized background treatment. .....70 Figure 2.3.  Virological response to maraviroc demonstrating concordance between Trofile and V3 genotype tropism assays. .....................................................................................................................72 Figure 2.4.  Virological response to maraviroc in a subset of patients screened as having non-R5 virus and randomly selected patients screened as having R5 virus. ...................................................74 Figure 2.5.  Frequency distribution of patient samples as determined by the assigned g2p score in accordance with therapeutic response categories. .................................................................................75 Figure 2.6.  Population sequencing with optimized g2p cutoff points can predict the virological response to maraviroc. ...............................................................................................................................77 Figure 2.7.  Population sequencing with optimized g2p cutoff points can effectively predict a) the probability of a tropism change and b) time to discontinuation in the study. ...................................78 Figure 3.1.  Virological response to maraviroc in MERIT trial participants rescreened for tropism using a V3 genotypic assay. ......................................................................................................................92  x Figure 3.2.  Time to change in tropism from R5 to non-R5 in MERIT trial participants rescreened for tropism using a V3 genotypic assay. ..................................................................................................94 Figure 3.3.  Concordance between tropism results in MERIT trial participants rescreened for tropism using a V3 genotypic assay and ESTA. .....................................................................................96 Figure 4.1.  Virological response to maraviroc in patients screened as having non-R5 virus. .......108 Figure 4.2.  Virological response to maraviroc in R5 and non-R5 viral populations. .....................110 Figure 4.3.  Changes in sequence prevalence by g2p FPR following maraviroc exposure.............111 Figure 4.4.  Viral fitness in the presence of maraviroc as a function of g2p false positive rate. ....112 Figure 5.1.  Virological response to short-term maraviroc exposure.................................................125 Figure 5.2.  Changes in sequence frequency stratified by g2p FPR after short-term maraviroc therapy in patients found to have ≥2% non-R5 HIV at any time point. ............................................127 Figure 5.3.  Viral fitness in the presence of maraviroc as a function of g2p FPR (FPR ≤ 20). .........129 Figure 5.4.  Virological response to maraviroc demonstrating concordance between 454 “deep” sequencing and both ESTA and TROCAI. ............................................................................................131  xi List of Abbreviations 3TC – Lamivudine A – Adenine AIDS – Acquired Immune Deficiency Syndrome ART – Antiretroviral Therapy ATP – Adenosine Triphosphate AZT – Azidothymidine (Zidovudine) b.i.d. – twice-daily dosage bp – nucleotide base pair C – Cytosine CA – Capsid Protein CCR5 – C-C-Motif Receptor 5 CD4 – Cluster of Differentiation 4 CRF – Circulating Recombinant Form CXCR4 – C-X-C-Motif Receptor 4 CYP3A – Cytochrome P450, family 3, subfamily A D/M – Dual and/or Mixed Tropism DNA – Deoxyribonucleic Acid dNTP – Deoxynucleotide Triphosphate ddNTP – Dideoxynucleotide Triphosphate dTTP – Deoxythymidine Triphosphate ddTTP – Dideoxythymidine Triphosphate EFV – Efavirenz emPCR – emulsion PCR ESTA – Enhanced Sensitivity Trofile Assay FDA – United States of America Food and Drug Administration FPR – False Positive Rate G – Guanine g2p – geno2pheno bioinformatics algorithm GALT – Gut-Associated Lymphoid Tissue Gp41 – Envelope Glycoprotein 41 Gp120 – Envelope Glycoprotein 120  xii Gp160 – Envelope Glycoprotein 160 GS-FLX – Genome Sequencer FLX (Roche/454 Life Sciences) GS Jr. – Genome Sequencer Junior (Roche/454 Life Sciences) GTR – General Time-Reversible (nucleotide substitution models) HAART – Highly Active Antiretroviral Therapy HIV – Human Immunodeficiency Virus HR – Heptad Repeat IN – Integrase INSTI – Integrase Strand Transfer Inhibitor IQR – Interquartile Range kb – kilobase pair LCB – Lower Confidence Bound LOCF – Last Observation Carried Forward LTR – Long Terminal Repeat MA – Matrix Protein MAAFT – Multiple Alignment using Fast Fourier Transform MCT – Maraviroc Clinical Test MERIT – Maraviroc versus Efavirenz in Treatment-Naïve Patients M=F – Missing Equals Failure MHC – Major Histocompatability Complex MID – Multiplex Identifier MIP – Macrophage Inflammatory Protein MOTIVATE – Maraviroc versus Optimized Therapy in Viremic Antiretroviral Treatment-Experienced Patients MT-2 – Human T-cell leukocyte cell line MSM – Men who have sex with Men MVC - Maraviroc NC – Nucleocapsid Protein Nef – Negative regulatory factor NNRTI – Non-Nucleoside Reverse Transcriptase Inhibitor NRTI – Nucleoside Reverse Transcriptase Inhibitor NSI – Non-Syncytium Inducing OBT – Optimized Background Therapy PBMC – Peripheral Blood Mononuclear Cell PBO – Placebo  xiii PCR – Polymerase Chain Reaction PI – Protease Inhibitor PIC – Pre-Integration Complex PPi – Pyrophosphate PR – Protease pr160 – Gag-Pol precursor protein 160 PSSM –Position-Specific Scoring Matrix PTP – Picotitre Plate pVL – Plasma Viral Load q.d. – once-daily dosage R5 – CCR5-Using HIV-1 RANTES – Regulated on Activation, Normal T-Cell Expressed and Secreted RAxML – Randomized Axelerated Maximum Likelihood Rev – Regulator of expression of virion proteins RNA – Ribonucleic Acid RNaseH – Ribonuclease H RT – Reverse Transcriptase RT-PCR – Reverse Transcription Polymerase Chain Reaction SDF – Stromal-Derived Factor SI – Syncytium-Inducing SIV – Simian Immunodeficiency Virus SP – Spacer Protein SVM – Support Vector Machine T – Thymine Tat – Transactivator of transcription tRNA – Transfer Ribonucleic Acid TROCAI – Tropism Coreceptor Assay Information UNAIDS – Joint United Nations Program on HIV/AIDS V3 – Third Hypervariable Loop Vif – Viral infectivity factor Vpr – Viral protein R Vpu – Viral protein U wSS – weighted Sensitivity Score X4 – CXCR4-Using HIV-1 ZDV - Zidovudine  xiv Acknowledgements First and foremost, I would like to thank my PhD supervisor, Richard Harrigan.  You are a truly patient, understanding and knowledgeable individual.  I am honoured to have been a student under your tutelage in such a productive environment.  I would also like to thank my Supervisory Committee, Thomas Kerr, Peter Phillips and Scott Tebbutt.  To all members of my supervisory committee, your support, guidance and expertise have been invaluable. I am happy to have shared the graduate journey with my fellow PhD students Chanson Brumme, Guinevere Lee, Luke Swenson and those from other labs.  Lunch dates, conference pals, commiserating companions, brainstorming comrades, thank you. The lab members, past and present, at the British Columbia Centre for Excellence in HIV/AIDS have been an incredibly supportive group.  Every individual has helped to some capacity whether with lab work, study design, problem solving, data analysis, reviewing documents, moral support and even emotional support.  I truly appreciate all they have done for me. I wish to thank study participants, coauthors and collaborators around the globe.  To the study participants, as cliché as it sounds, we wouldn’t be here without you.  Co-authors and collaborators, I am grateful for the opportunity to have met and worked with you.  The pleasure was all mine. I would also like to recognize the dedicated teachers, professors and mentors of my youth, those who never ceased to encourage me in my academic pursuits and opened my eyes to what is and what could be. And of course, I would like to express my love and utmost appreciation to my primary support group, my four pillars.  Mom, for your unconditional support, right-brain promotion and indisputable love; Grandma, for my inspiration, preschool education and appreciation of all puzzles; Rob, for your continuous encouragement, limitless support and for being my one true hero; and Charlie, for your everlasting respect, patience, partnership and for all of the adventures we have been on and have yet to embark upon.  xv Dedication I dedicate this work to my four pillars. 1 Chapter One:  A General Introduction and Thesis Objectives 1.1 The Human Immunodeficiency Virus 1.1.1 The Identification of a Pandemic In 1981 young men were being diagnosed with and succumbing to opportunistic infections that were otherwise only observed in those with weakened immune systems.  These young men happened to be homosexual, or men who have sex with men (MSM), living in New York and California.  It became clear that these previously healthy men had failing immune systems and the search for a cause was initiated.1 Nearly two years later it was announced that a retrovirus, isolated from the tissues of a patient with lymphadenopathy, might be the cause of the newly termed Acquired Immunodeficiency Syndrome (AIDS).2 This retrovirus was later to become known as Human Immunodeficiency Virus (HIV) and confirmed as the infectious agent that leads to AIDS.3–5 Over thirty years later, in 2013 it was reported by UNAIDS that 35 million people worldwide were living with HIV.6 HIV is a pandemic. The highest rates of HIV/AIDS-associated morbidity and mortality are in developing nations.  UNAIDS estimates that 70% of global HIV infections are within Africa, and in 2013, Sub-Saharan Africa accounted for an estimated 74% of AIDS-related deaths worldwide.6 Though HIV is now considered manageable with antiretroviral therapy (ART), clinical management is dependent on available resources.  With no curative strategies or working vaccines HIV continues to be a public health challenge.  Global  2 initiatives are underway in attempts to lessen the burden and the spread of HIV, with a target of reducing transmission by 50% by the end of 2015.7 In the UNAIDS Gap Report released in 2013, rates of new infection had decreased by 13% over the past three years and AIDS-related deaths had decreased by 19%.  In 2014, UNAIDS introduced the more ambitious 90-90-90 initiative.8  The 90-90-90 initiative focuses on the identification of HIV-positive individuals and ensuring they receive proper treatment.  The initiative is defined by three goals to be achieved by 2020: 1) 90% of the HIV-positive population will know their HIV status; 2) 90% of the HIV-positive population will receive antiretroviral therapy; and 3) 90% of the HIV-positive individuals receiving antiretroviral therapy will attain viral suppression.8  With initiatives to expand HIV testing, care, treatment and prevention strategies, progress is being made. 1.1.2 HIV Types, Groups & Clades There are two primary types of HIV, HIV type 1, (HIV-1) and HIV type 2 (HIV-2).  They are genetically distinct from each other and differ in many ways, including region of origin, species of origin, effects on disease progression and virulence.  HIV-1 is the cause of the global pandemic.  HIV-2 is concentrated in West Africa, and appears to be less virulent when compared to HIV-1.9,10 There are three primary groups of HIV-1 variants, and they are categorized as M (main), O (outlier) and N (non-M, non-O) based on genetic diversity and phylogeny.  Group M viruses account for the vast majority of HIV infections, estimated to be responsible for 33 million (~94%) infections worldwide.11,12 More recently, the possibility of a fourth HIV-1 group has arisen.  Currently, there are only two reported cases of HIV-1 demonstrating characteristics of a new group.  For this reason, the new group is considered “putative” and has been designated group P.13–15 Variants within HIV- 3 1 group M are further subdivided into subtypes in order to account for genetic variation, these phylogenetic clusters are called clades.16,17 Genetic variation of HIV-1 within subtypes can be up to 17%, whereas genetic variation between subtypes can be up to 35%.16 According to the Los Alamos HIV sequence database, to date there are clades (A-D, F-H, J, K) with two sub subtypes in clades A (A1 and A2) and F (F1 and F2).  In addition, there are over sixty-six circulating recombinant forms (CRFs) of HIV-1 group M viruses (Figure 1.1).18 Clades and CRFs tend to be geographically clustered, and in regions with multiple clades present may also be compartmentalized according to social and demographic characteristics.19–21 Clade C is the predominant infection in South Africa.  It can also be found in India and China, accounting for approximately 50% of global HIV-1 infections.16 Clade B is the predominant HIV-1 subtype in the western countries of Europe, the Americas and Australia but only accounts for roughly 12% of the global infection.20,11  Figure 1.1.  The HIV classification system. A diagram illustrating the four levels of classification (type, group, subtype and sub subtype) used to define HIV variants based on shared characteristics and genetic similarity.  This image was created by the author.  4 1.1.3 Cross-Species Transmission The origins of HIV have been debated.  It is now widely accepted that HIV is a human descendant of the Simian Immunodeficiency Virus (SIV)22 Using the endogenous effects of HIV as a molecular clock, it is believed that the first cross-species transmission of HIV-1 group M virus occurred at the beginning of the twentieth century, sometime between 1910 and 1930.23,24 Based on the analysis of tissue samples collected in Kinshasa, Democratic Republic of Congo, there is conclusive evidence of the presence of HIV-1 in Central Africa as early as 1959.24,25 It wasn’t until 1999 that the origin of HIV-1 was confirmed, when Gao et al. found HIV-1 virus to be phylogenetically similar to the form of SIV infecting chimpanzees, Pan troglodytes troglodytes.22,26 Keele et al. went further to differentiate between Pan troglodytes troglodytes communities inhabiting south-central and southeastern Cameroon, concluding that the two groups were the sources for group N and group M HIV-1, respectively.27 The initial route of transmission was likely via blood contact in the context of hunting chimpanzees as bushmeat.28–31 1.1.4 Transmission among Humans There are multiple routes of HIV transmission, all of which occur when bodily fluids containing virus such as the blood, or genital secretions of an infected individual come into contact with the blood, mucosal linings or damaged tissue of a new host.  A recent literature review performed by Patel et al. ranked the risk of transmission for each route based on the incidence of new infections per 10 000 exposure events.32 The risk of  5 transmission from infected blood products is extremely high, followed by vertical mother-to-child transmission.  They also report a substantial risk when engaging in anal intercourse and when sharing needles for intravenous drug use, as well as risk associated with percutaneous needlesticks and penile-vaginal intercourse.32 It is important to note that there are various prevention strategies designed to reduce the risk of transmission for all routes, though many political, social and biological factors affect the extent to which those prevention strategies are implemented, accessed and beneficial.33–38 The rates of new infections associated with the different transmission routes vary between continents.6 Based on the geographical representation of HIV incidence among risk groups published in UNAIDS Gap Report 2013, the primary route of transmission in the Americas and Western Europe is homosexual intercourse, whereas the primary route in Africa is heterosexual intercourse.  Transmission risks in Asia are greatly associated with sex work, and in Eastern Europe there are far more new infections associated with the intravenous drug use risk group. 6,39,40 Differences in the geographic distribution of transmission routes may be due to a number of societal factors including laws, policies, healthcare infrastructure, and the degree of social acceptance for such behaviours as homosexual intercourse, intravenous drug use and the treatment of women. 1.2 HIV the Virus 1.2.1 The Genome & the Virion HIV is a member of the Lentivirus genus within the Retroviridae family.2,41 The virion contains two single-stranded ribonucleic acid (RNA) molecules of nearly ten kilobases (kb)  6 in length.42,43 The relatively short HIV genome is organized into nine open reading frames.  These reading frames overlap to some degree, allowing the virus to optimize the number of translated proteins (Figure 1.1).43–45 Characteristic of retroviruses, there are three large open reading frames referred to as gag, pol and env.  The gag reading frame encodes for the Gag precursor protein, pr55, which is cleaved by the viral protein, protease (PR) during virion maturation to form the structural proteins: matrix protein (MA; p17), capsid protein (CA; p24), nucleocapsid protein (NC; p7) and three smaller peptides, spacer peptide 1 (SP1), spacer peptide 2 (SP2) and protein 6 (p6).44,46–48 Similarly, the env reading frame encodes the Env precursor protein (gp160), though this precursor is later cleaved by a host cellular protease to form the two elements of the envelope surface spike, glycoprotein 120 (gp120) and glycoprotein 41 (gp41).44,49,50 The pol reading frame overlaps with gag, creating a Gag-Pol precursor protein (pr160).43 The enzymes reverse transcriptase (RT), ribonuclease H (RNaseH), integrase (IN) and protease are the products of viral proteolytic cleavage of the Gag-Pol precursor protein.44 In addition to the structural and enzymatic proteins, HIV also contains the genes for accessory proteins and regulatory proteins encoded for within the remaining six reading frames.  The role of regulatory proteins such as regulator of viral expression (Rev) and transactivating protein (Tat) is to enhance viral gene expression whereas the accessory proteins, negative factor (Nef), viral infectivity factor (Vif), viral protein r (Vpr) and viral protein u (Vpu) all act to augment viral infectivity and pathogenicity.44,51,52 The entire genome is book-ended by the 5’ and 3’ long terminal repeats (LTRs), which are placed at either end of the proviral deoxyribonucleic acid (DNA) during reverse transcription. The LTR ends contain binding sites for cellular transcription factors  7 used in the integration and transcription of viral DNA, act as a promoter for viral gene expression and participate in the post-transcription polyadenylation.53–55  Figure 1.2.  The HIV-1 genome. A linear schematic representation of the HIV-1 genome organized into the three reading frames. The gag, pol and env genes encode the structural, enzymatic and envelope proteins, respectively.  This image was created by the author. The mature infectious virion is spherical and has been estimated to be between 120 to 200 nm in diameter.56–58 The virion is composed of a lipid bilayer envelope, matrix, capsid, various viral proteins and two copies of single stranded RNA associated with nucleocapsid.  The lipid portion of the viral envelope is of cellular origin, characterized by viral proteins at the surface often referred to as spikes.  The spikes are trimeric glycoprotein complexes comprised of non-covalently linked surface gp120 protein and the transmembrane anchor gp41.59,60 The gp120-gp41 complex is necessary in the binding and fusion of the virion to the target cell.61–63 Matrix protein is associated with the inner surface of the viral envelope, and participates in the incorporation of gp120-gp41 spikes into the envelope of the new virion.64–66 Inside the mature virion is a conical capsid made of capsid protein, containing the RNA and viral proteins.  The viral proteins found within the capsid include the enzymatic proteins protease, reverse transcriptase, RNaseH, integrase and the  8 structural nucleocapsid, which forms a nucleocapsid-RNA complex.57,67 The nucleocapsid has been described as coating the RNA and thought to be associated with RNA packaging, similar to the histones of eukaryotic DNA.68,69 In the micrograph and accompanying schematic presented in Figure 1.3, panels A) and B), one can clearly see the accumulation of gag proteins at the membrane of both the budding and immature viruses, on the left and right respectively.  However, in the mature virion, pictured in the centre and featured in panel C), the formation of the capsid following proteolytic cleavage of the gag precursor protein can clearly be identified.70  Figure 1.3.  The HIV-1 virion. Panel A is a micrograph image depicting three budding stages of HIV-1, from assembly to the mature virion.  The micrograph is reproduced as a schematic in Panel B detailing the three stages, assembly of proteins at the cell membrane on the left, the released immature virion on the right and the mature virion in the centre.  Panel C is a diagrammatic representation of the mature virion highlighting structural, enzymatic and envelope proteins.  Panels A and B were taken from the open access article by Baumgärtel et al..70 The image in Panel C was created by the author.  9 1.2.2 The Replication Cycle 1.2.2.1 Cellular Entry & Reverse Transcription The first step in the HIV replication cycle is entry into the new host cell.  The circulating infectious virion binds to a target cell, releasing its components into the cytoplasm.  Target cells have two primary characteristics at their plasma membrane required for viral infection, the cluster of differentiation 4 (CD4) receptor and a chemokine co-receptor.71–75 There are two primary co-receptors used by HIV-1 to gain entry into the cell, chemokine (C-C motif) receptor 5 (CCR5) and chemokine (C-X-C motif) receptor 4 (CXCR4).73,74,76,77 In order to initiate the fusion between the virion and the cell, HIV-1 must first bind the CD4 receptor, which then exposes the viral binding site for either CCR5 or CXCR4.78,79 HIV binds to the cell surface receptors by way of the trimeric envelope glycoprotein spike, the gp120-gp41 complex.63 The conformational changes in the structure of the spike following the interactions of gp120 with both cell receptors pull the virion closer to the cell surface where gp41 can then initiate fusion of the virion with the cell membrane.  As the virion and the cell membrane fuse, the envelope becomes incorporated into the membrane and the contents of the virion are released into the cellular cytoplasm.62,63 After the capsid is released from the virion into the cytoplasm it is “uncoated”.80 Capsid molecules are partially dissociated to form a reverse transcription complex.  The details associated with the timing and location of when the capsid is uncoated have yet to be confirmed but uncoating has been shown to occur roughly one hour after viral fusion.80,81 Inside the reverse transcription complex, single stranded RNA is converted into  10 double stranded DNA through the process of reverse transcription, characteristic of retroviruses.53,82–84 First, the single-stranded viral RNA template is transcribed by the reverse transcriptase RNA-dependent DNA polymerase to synthesize a complementary DNA strand.  The reverse transcriptase RNaseH then digests the viral RNA template leaving the newly synthesized DNA strand as the template from which the reverse transcriptase DNA-dependent DNA polymerase then generates a second complementary DNA strand.  This second synthesis yields a double-stranded proviral DNA molecule.85 Reverse transcription can account for the high rate of diversity in HIV as it is a highly error prone process.  Mutations arise as newly synthesized sequences are not checked for nucleotide accuracy by the 3’-5’ exonuclease “proof-reader” characteristic of more stable genomes.86,87 Unfortunately such alterations to the viral genome and genetic diversity make the design of treatments and vaccines a challenge. 1.2.2.2 Integration & Replication When the viral genome has been transcribed into a double-stranded DNA molecule, the reverse transcription complex dissociates and the complex is reassembled to form the pre-integration complex.  It is largely accepted that the pre-integration complex contains viral DNA and the viral proteins reverse transcriptase, integrase, Vpr and matrix.88,89 The proviral DNA is ushered into the nucleus through the nuclear pore as part of the pre-integration complex.90 It is believed that Vpr helps to guide the pre-integration complex to the nuclear membrane but it is the nuclear localization signals found on matrix and integrase proteins that help guide and carry the pre-integration complex through the nuclear pore.91,92 Insertion of the newly synthesized proviral DNA recruits the host cell to begin replicating the viral genome.  Integration of the viral genome into the host DNA is  11 characteristic of retroviruses.  In most cases, the retrovirus requires a cell to be actively dividing such that the nuclear membrane is susceptible to nuclear entry and infection.  HIV however, has the ability to infect resting or non-dividing cells as it enters the nucleus through the nuclear pore as part of the pre-integration complex.90,93 The integration process is catalyzed by integrase and can be divided into two stages.90 First the 3’ ends of the proviral DNA, near the LTR, are cleaved to produce two base pair overhanging “sticky ends”.  This step is called 3’ processing and it prepares the proviral DNA for integration.94,95 Next, the host DNA is cleaved to produce a 5’ free end and a recessed 3’ end to which the viral DNA will be covalently linked.  The host DNA polymerase then completes the covalent attachment of the 5’ end of the viral DNA to the 3’ end of the host DNA referred to as strand transfer and filling in the gaps.94–96 After the viral DNA has been integrated into the host genome, the cellular replication machinery transcribes the DNA into RNA.  Replication begins in the “R” (repeat) region of the 5’ LTR by the host RNA polymerase II.  Genomic regions for transcription regulation, particularly initiation, are found within the LTRs at both the 5’ and 3’ ends of the viral DNA.44,55 The viral transcripts are processed and modified in the same way as host RNA, including the addition of a string of adenosine nucleotides forming a “polyA” tale.44 The polyA tail aids the transcripts in exiting from the nucleus, as well as preventing their enzymatic degradation and promoting their translation in the cytoplasm.44,97 Spliced and unspliced RNA are transported out of the nucleus where the mRNA molecules are translated and folded into the viral proteins, where they await virion budding at the cellular membrane with the newly transcribed RNA genome.46,99,97  12 1.2.2.3 Budding & Maturation The Gag and Gag-Pol precursor proteins are translated late in the replication cycle.  They move directly to the plasma membrane following translation and remain uncleaved until after budding.  Gag is anchored at the plasma membrane by the N-terminal end of the matrix protein.67,100 The individual proteins within the polypeptide are organized radially at the cell membrane surface: matrix, capsid, SP2, nucleocapsid, SP1 and p6.  The positions of the structural proteins in the polypeptide are relatively conserved within the mature virion (Figure 1.3).  Late proteolytic cleavage of the Gag and Gag-Pol precursor proteins ensures budding does not occur prematurely.57,100,101 Prior to budding, additional components to be incorporated into the new virions assemble at the cell membrane, including the envelope proteins, enzymatic proteins, vpr and the genomic RNA.  Host components tRNA and cyclophilin A are also collected at the membrane for packaging into the new virions.101 The former is incorporated for use in reverse transcription after the virion has entered a new host cell; the latter is believed to assist in the uncoating of the viral capsid.100,102 Prior to budding the matrix protein N-terminus is myristoylated, the myristoyl group playing a role in both transporting and attaching the matrix protein to the plasma membrane.  Budding is triggered when the matrix N-terminal myristoyl group interacts with the plasma membrane.44,102 The budding process serves to encapsulate the collected viral components to form a new virus particle.  The Gag polypeptides form a layered band along the inner surface of the immature virion, and the non-covalently linked gp120-gp41 complexes are incorporated to span the new viral envelope, forming the viral spikes that mark the surface of the virion.64,66,100,101 P6 interacts with a host trafficking system to complete budding, allowing the newly formed  13 virion to fully separate from the cell membrane.66,102 Even so, a further host restrictor called tetherin literally tethers the new virions to the surface of the cell.  The protein Vpu overcomes tetherin, completing the budding process as the virion is fully released into the extracellular matrix.100,103 However, the newly released virus particle is in its immature form at this stage. During the late stages of the HIV-1 replication cycle, protease is translated and it subsequently proteolytically cleaves itself from the Gag-Pol precursor protein.  Following budding, protease is responsible for initiating the maturation of the immature virion by cleaving the radially organized Gag and Gag-Pol precursor proteins into their structural component proteins.57,100,102,104 After cleavage, the matrix remains associated with the envelope.  The capsid proteins associate to form the capsid around the nucleocapsids, and the nucleocapsid proteins interact with the two copies of viral RNA to form the two nucleocapsid protein-RNA complexes housed within the capsid. The mature virion is now infectious (Figure 1.3c).57,101 A complete overview of the HIV-1 replication cycle is schematically represented in Figure 1.4.  14  Figure 1.4.  The HIV-1 replication cycle. A schematic diagram illustrating the HIV replication cycle, which begins with the binding and fusion of the virion to the target membrane.  Viral proteins and RNA are released into the cytoplasm where the RNA is reverse transcribed into DNA and carried into the cell nucleus via the preintegration complex (PIC).  Once in the nucleus the DNA is incorporated into the host genome and replicated using the cellular mechanisms.  The viral RNA is translated into proteins, which assemble at the cell surface alongside newly transcribed RNA.  New virions encompassing two copies of viral RNA and viral proteins bud and are released from the cell surface into the cytoplasm.  This figure has been adapted from Barré-Sinoussi et al., with the permission of Nature Publishing Group via RightsLink (License No. 3521620706908).105 ©2013 1.2.3 HIV Disease Progression & Pathogenesis The establishment and progression of an HIV infection can be subdivided into a number of phases characterized by the level of viremia and various immune factors.  Approximately 10 days following an HIV transmission event viral RNA can be detected in the blood as the first sign of an established infection.  This marks the beginning of the acute phase, which can last up to six months.106 During the acute phase increasing numbers of target cells become infected, facilitating the exponential growth of the infection and the  15 rapid spread of virus throughout the body.107 Virus can also be found in myeloid derived monocytes, as well as their macrophage and dendritic cell derivatives.  Virus continues to replicate in the lymphoid tissues to which it has spread, most often reaching a peak in viral load within roughly three to four weeks of infection.107,108 The activated immune cells of the gut-associated lymphoid tissue (GALT) are primary targets in acute infection.  This population of cells is severely depleted within the first three weeks of infection.106,109,110 At this point, flu-like symptoms including fever, lymphadenopathy, pharyngitis, skin rash, myalgia, arthralgia, fatigue and headache may emerge, indicating acute retroviral syndrome driven by the body’s natural immune response to viral infection.107,111 This symptomatic period of peak viral load is generally short-lived as the immune system begins to moderate viral replication.  In so doing, the plasma viral load (pVL) decreases markedly and remains relatively stable at what is known as the viral “set point”.107,111 Though not fixed, the viral set point foreshadows the progression of infection.  A high viral set point is most often associated with quicker progression to AIDS.112,113 Also at this time, HIV target cell populations recover incompletely, marking the end of the acute phase and the establishment of the chronic phase or the “clinical latency” period.114 Chronic infection is characterized by continuous viral replication and the infection of healthy CD4+ T-cells.  Despite persistent infection in a number of tissues and reservoirs this period of active infection is relatively asymptomatic.115,116 Without treatment, the chronic phase can last up to ten years on average, though rates of progression vary between individuals.114,117 Current anti-HIV treatment can greatly reduce the rate of disease  16 progression during this phase, but it cannot eliminate latent viral reservoirs.118–120 As depicted in Figure 1.5, as early as acute infection, a portion of infected activated CD4+ T-cells will become resting cells, forming a latent cellular compartment able to evade the immune system for prolonged periods.118 However, the resting cells are capable of reactivating and feeding the infection by producing new virus and interacting with uninfected target cells.118,121 As well, activated, infected T-cells can be marked for apoptosis, leading to the progressive depletion of the CD4+ T-cell population.116 Control over viral replication starts to wane as CD4+ T-cell populations are depleted slowly over the course of the chronic phase.122,123 Nonspecific constitutional symptoms including fever, night sweats, weight loss, and non-threatening opportunistic infections such as oral candida begin to emerge, marking the end of the chronic phase.114 At this phase of infection, known as clinical AIDS, the immune system is severely impaired, and limited in its ability to fight infection.  Pathogens that cannot overcome the immune system of a healthy host take advantage of this weakened immune state to cause opportunistic infection.  On average, when left untreated an individual with AIDS will survive approximately three years, eventually falling to opportunistic infection(s) and or tumors.  It was the identification of Pneumocystis pneumonia and Kaposi’s sarcoma in young homosexual men that lead to the discovery of AIDS over thirty years ago.1,124–127  17  Figure 1.5.  HIV-1 target cells and latency. HIV-1 target cells and their role in latent infection.  HIV-1 targets CD4+ immune cells, in particular activated CD4+ T-cells, which have the ability to become resting memory CD4+ cells, as well as resting memory CD4+ T-cells, as well as monocytes, macrophages and dendritic cells of the myeloid cell line.  The latent reservoir is namely driven by the resting memory CD4+ T-cells, which can both subsist inactive and hidden from the immune system for prolonged periods as well as become reactivated to produce new virus, transmit virus or be marked for apoptosis.  Though not depicted here, cells of the myeloid line are also capable of persisting as reservoirs.  This figure has been adapted from Deeks et al. Nature Reviews.118 ©2012 Permission to reproduce this figure has been granted by Nature Publishing Group under RightsLink license No. 3521621023311. 1.3 HIV Entry 1.3.1 HIV Cell Interactions HIV interacts with a number of immune cells, but primarily targets CD4+ T-cells and macrophages.  Following the transmission of HIV, the infectious virus particle is recognized as non-self by an antigen-presenting cell.  Antigen-presenting cells include dendritic cells and tissue macrophages; they can be found throughout the body and are  18 integral in the innate immune defense mechanism.128,129,130 Antigen-presenting cells act as sentries to the immune system.  Upon identifying as non-self, the antigen-presenting cell will engulf, digest and subsequently present an HIV epitope by displaying it on a major histocompatibility complex (MHC) found at its plasma membrane.  The presentation of epitopes to immune cells stimulates the adaptive immune response, including the activation of CD4+ T-cells.131,130 The CD4+ T-cells are lymphocytes and part of the lymphatic system; a major component of the immune system.  The lymphatic system is composed of various immune cells, lymph and lymphoid tissues like the thymus, GALT and lymph nodes.  CD4+ T-cells, or T helper cells, congregate within the lymphoid tissues, particularly the secondary lymphoid tissues like the lymph nodes, where they encounter epitope-presenting cells.  Here the CD4+ T-cells are activated and work to orchestrate the cellular immune response by releasing a variety of chemical messengers, or cytokines.132,133 Immune cells gather and interact in the lymphoid tissues exposing CD4+ cells to HIV thus hastening the spread of infection.  It is the infection and accompanied immune deficiency caused by the depletion in CD4+ T-cell populations that are used to provide a clinical definition of AIDS.  In 1990 it was shown that of all biological markers, CD4+ T cell count was the best predictor of HIV-1 disease progression.134 HIV targets tissue macrophages in addition to CD4+ T-cells.  Marcophages play a large role in HIV infection, starting with the uptake of HIV in their role as an antigen-presenting cell type.  After they have carried the virion to the lymphoid tissue, macrophages are capable of secreting chemokines as part of the immune response that  19 attract and activate CD4+ T-cells.135 Macrophages direct CD4+ T-cells to areas profuse with HIV thus providing a larger population of target cells for infection.  In addition, macrophages have the ability to transfer HIV via cell-to-cell interactions allowing the virus to evade extracellular inhibitors, including those of the host immune response or antiretroviral therapies.136 It is believed that this process is quicker and more efficient than the cell entry of freely circulating HIV virions.137 Furthermore, they can serve as a long-term reservoir to the virus throughout the body, seemingly unaffected by the virus-induced cytopathic effects, safely storing the virus from the effects of the immune system and antiretroviral therapy in treated individuals.138–140 To a lesser extent other antigen-presenting cells like blood monocytes and dendritic cells have been found to become infected and harbour HIV.  Blood monocytes are the precursor cell to both tissue macrophages and dendritic cells.141 They have been found to be somewhat resistant to HIV infection, however, this has more recently been shown to be a derivative of monocyte differentiation states.142 Though monocytes typically circulate for only a matter of days, they can act as a long-term reservoir of HIV as they do not fall victim to virus-induced cell death.143,144 Like macrophages, dendritic cells carry and present HIV to the CD4+ T-cells as part of the initial immune resonse.145 However, like monocytes they are less commonly infected with HIV.  Depending on the sub-type of dendritic cell, infectious HIV can be retained at the dendritic cell surface from which it can be transferred to CD4+ T-cells.145–147  20 1.3.2 HIV Receptors & Co-Receptors Cell surface receptors mediate the signaling events used by the cell to monitor the extracellular environment.  Receptors are integral membrane proteins, spanning the lipid membrane with extracellular, transmembrane and intracellular domains.  When a receptor is engaged with a ligand at the cell surface it opens a route of communication triggering a cellular response. The CD4 molecule is a glycoprotein receptor found on the surface of some immune cells, including the aforementioned macrophages, monocytes, dendritic cells and T helper cells.  It is characterized by four immunoglobulin domains, which interact with the MHC II molecules found on antigen presenting cells, carrying antigen.  When an antigen-carrying MHC-II complex is able to bind both a T-cell receptor and CD4, the CD4 receptor functions to augment the signals generated in this T-cell activation pathway.  Signaling in this way stimulates the CD4+ T-cell to interact with other cells and release cytokines as an immune initiative.  In the setting of HIV, CD4 is the primary receptor of the virus and is fundamental in viral entry.72,148 Though CD4 is the primary HIV receptor, the presence of a chemokine co-receptor at the surface of a target cell is necessary for viral entry.  Though other chemokine receptors have been implicated in viral entry, it is now widely accepted that CXCR4 and CCR5 are the primary co-receptors used by HIV when entering a target cell in vivo.74,149–151 The chemokine receptors are members of the G protein-coupled family of receptors, characterized by seven transmembrane loops, of which four are extracellular.150 They convey extracellular messages to the cell by binding chemokines, small, proinflammatory  21 chemotactic cytokines, used in the activation and trafficking of leukocytes.  The naturally occurring ligands for the CCR5 receptor are C-C chemokines RANTES (Regulated on Activation, Normal T-cell Expressed and Secreted), MIP-1α and MIP-1β (Macrophage Inflammatory Protein; two forms, 1α and 1β).74,152 The naturally occurring ligands for the CXCR4 receptor are C-X-C chemokines SDF-1α and SDF-1β (Stromal Cell-derived Factor; two forms, 1α and 1β).73 These naturally occurring CCR5 and CXCR4 chemokines inhibit HIV-1 replication, confirming the active role these co-receptors play in HIV infection.65,154,99 CXCR4 and CCR5 are expressed on leukocytes, though the level of expression differs between cell types.  CCR5 is found to be more prevalent on macrophages and T-memory cells, whereas CXCR4 is found to be more prevalent on naïve T-cells.  The two receptors are found on monocytes and dendritic cells relatively equally however the level of expression changes as these cell types mature.155,156 1.3.3 Viral Structures for Cell Entry HIV enters the target cell following the interaction of viral and cellular structures.  The envelope-associated spike facilitates the binding, fusion and entry of the virion.  The trimeric spike itself is composed of three copies of the heterodimer gp120 non-covalently linked to gp41.59,157 The gp120 trimer initiates contact between the virion and the target cell by binding to CD4 and a chemokine co-receptor, after which gp41 facilitates the fusion of the virion envelope and cell membrane. Envelope gp120 is the extracellular portion of the virion spike.  It is composed of five conserved regions (C1-C5) and five hypervariable regions (V1-V5).158 As the names suggest, the conserved regions have low genetic variability whereas a high level of genetic  22 diversity characterizes the variable regions.  It is believed that this high level of variability in the envelope spike allows the virus to evade immune recognition.  The core of gp120 is roughly heart-shaped with an inner and outer domain linked by a bridging sheet, the outer domain being more variable.  As a glycoprotein, gp120 is characterized by a considerable amount of N-linked glycosylation at the outer domain.50,159 This glycosylation is thought to be a mechanism through which HIV can avoid host neutralizing antibodies.160,161 Despite the modest contributions of other gp120 regions, such as V1 and V2, the third hypervariable region (V3) has been confirmed as the primary envelope region for chemokine co-receptor interaction.162–164 Within the outer domain of the gp120 core, at the apex of the inverted heart-shaped protein, V3 extends slightly as a loop.78,165–167 The V3 loop is typically composed of approximately 35 amino acid residues.  The loop structure is supported by the presence of a disulphide bridge formed between the single cysteine residues found at both the beginning and the end of the loop.  V3 can be subdivided into three basic areas, a conserved base, a variable stem and a β-hairpin tip.167,168 The positions of amino acids within the variable stem are thought to recognize different regions of the co-receptor, and amino acid polymorphisms associated with co-receptor use have been identified in this region of the V3 loop.167,169 It is the variance in the amino acid sequence of the V3 loop that determines whether the virus will use the CCR5 and or CXCR4 co-receptor. The gp41 trimer is the transmembrane portion of the gp120-gp41 complex, found at the base of gp120, in the shape of a mushroom.165,170 There are four primary domains of gp41.  At the C-terminal are the transmembrane domain and the C-terminal heptad repeat  23 (HR2); at the N-terminal, the fusion peptide and N-terminal heptad repeat (HR1).171 The HR1 and HR2, each consisting of three α-helices, are linked by short, flexible peptides to form the gp41 ectodomain.  The fusion peptide in the inactive state is found within a hydrophobic pocket of the gp120-gp41 complex.  The transmembrane domain anchors the gp120-gp41 complex to the envelope, whereas the HR1, HR2 and fusion peptide, characterized by hydrophobic residues, facilitate membrane fusion between the virion and the target cell.44,172,173 The viral structures associated with each stage of the viral entry mechanism are illustrated in Figure 1.6.  Figure 1.6.  HIV-1 cell binding and membrane fusion. A diagrammatic representation of the HIV-1 binding mechanism and subsequent fusion of the cell and viral membranes.  HIV-1 binding first requires the binding of gp120 to the CD4 receptor, the interaction of which causes a conformational change allowing the V3 loop to contact one of the two chemokine co-receptors, either CCR5 or CXCR4, and illicit the fusion mechanism.  Fusion occurs following co-receptor binding as the fusion peptide is inserted into the cell membrane and folds to pull the membranes together.  This figure was adapted from Haqqani et al., Antiviral Research.174 ©2013 with permission under the RightsLink license No. 3521621226558. 1.3.4 The Entry Mechanism Upon encountering a target cell the HIV gp120 protein first binds to the surface CD4 receptor, putting in motion a complex cascade of events from a pre-fusion closed state to an intermediate open state leading to viral entry.  The point of contact between CD4 and  24 gp120 is a recessed pocket found at the interface of the inner and outer domains at the surface of gp120.78,161 This binding interaction causes a conformational change in the gp120 protein whereby the trimer subunits rotate outwardly opening the gp120-gp41 complex, similar to the opening of a flower.  This movement flexes the CD4 molecule and draws the cell membrane closer to the bound virus.165 It also causes V1/V2 to shift, exposing the co-receptor binding site associated with V3 and the bridging sheet.79,175,176 In the CD4-bound form the gp120 bridging sheet is stabilized and the V3 loop is oriented for co-receptor binding.70,165,117 When the V3 loop contacts the co-receptor, the V3 conserved base, variable stem and hairpin loop interact with the N-terminal, extracellular loops and transmembrane helices of the co-receptor binding pocket, respectively.177–179 This co-receptor activity shifts the CD4-bound virus nearer the cell membrane and conformational changes at the gp120-gp41 interface expose gp41 to initiate fusion.180 Following HIV receptor binding, a second set of conformational changes in the gp120-gp41 complex causes the formation of a coiled coil composed of three parallel HR1 α-helices from each of the gp41 subunits of the envelope trimer.  Formation of the coiled coil orients the fusion peptide toward the cell membrane.  The fusion peptide extends through the extracellular space and into the target cell membrane, becoming a linear extension of the viral envelope and anchoring the virion to the target cell.50,172,181 Once the fusion peptide has become anchored in the cell membrane, the HR1 coiled coil folds back on itself interacting with the antiparallel HR2 helices.  The HR2 helices fit into hydrophobic grooves formed between the HR1 helices to form a six-helical bundle and hairpin loop.  The formation of the six-helical bundle draws the cell-embedded fusion peptide and the viral transmembrane domain near enough for the cell and viral membranes to fuse.62,180  25 Once fusion has occurred the virion is incorporated into the cell membrane, and the contents of the virion are released into the cytoplasm from which they can enter the replication cycle as described previously.  Figure 1.7 demonstrates the conformational changes necessary for viral entry, occurring within the gp120-gp41 complex following HIV-1 receptor binding.182  Figure 1.7.  Conformational changes during HIV-1 cell entry. A model interpretation of the conformational changes occurring during the HIV-1 entry process.  HIV-1 entry is mediated by the “envelope spike” composed of the trimeric gp120-gp41 complex.  When bound to both the CD4 and chemokine co-receptor the spike is activated and the gp41 fusion peptide extends toward the cell membrane.  The gp120 subunits rotate outward, allowing the gp41 subunit to reach the cell membrane, forming the pre-hairpin intermediate.  Once anchored in the cell membrane the pre-hairpin intermediate folds on itself creating the six-helical bundle, drawing the cell and viral membranes together, allowing contact and fusion.  This figure has been published by Tran et al., PLoS Pathogens.182 ©2012 (Open Access) 1.3.5 The Chemokine Co-Receptors & HIV Infection 1.3.5.1 Tropism Terminology HIV-1 is characterized by viral tropism, the co-receptor through which HIV enters the target cell.  When virus enters the cell via the CCR5 or the CXCR4 chemokine co-receptor it is considered “CCR5-using” (R5) or “CXCR4-using” (X4), respectively.183 There are also instances of “dual/mixed” infection, where “dual” virus is capable of using both  26 co-receptors, and “mixed” characterizes a viral population that contains both R5 and X4 virus.184,185 Collectively, X4 and dual/mixed populations are often referred to as “non-R5”. Historically, R5 virus has been referred to as macrophage-tropic, M-tropic, because R5 viral isolates are far more efficient at infecting primary macrophages when compared to T-cell lines.  Conversely, X4 virus was originally designated as T cell line-tropic.  It was called T-tropic because these viral isolates had greater infectivity in T-cell lines when compared to macrophages.  Virus capable of infecting both macrophages and T cell lines were termed dual-tropic.  These distinctions were also correlated with the formation of syncytia, which are large, irregularly shaped, multinucleated cells, in human T-cell leukocyte (MT-2) cell lines.186 T-tropic, or X4 virus was found to cause fusion of T-tropic infected cells, creating syncytia and thus termed syncytium inducing (SI).  In contrast, M-tropic, or R5 virus was not found to be associated with the formation of syncytia and was termed non-syncytium inducing (NSI).183 1.3.5.2 Tropism Switch with Disease Progression It has long been accepted that HIV-1 can be dichotomized by R5 and X4 phenotypes, and that these phenotypes differ in many aspects, and are most clinically relevant in terms of the rate of disease progression.187 Both R5 and X4 variants are capable of being transmitted however R5 virus characterizes the vast majority of primary infections regardless of transmission route.188 Whether this observation is due to the preferential transmission of R5 virus, the prevalence of target cells, biased immune pressures limiting X4 virus, or a number of other hypothesized mechanisms, has yet to be determined.189,190 However, the extreme rarity of HIV infection in individuals homozygous for a deleterious  27 mutation in the CCR5 gene, causing deficient expression of the CCR5 co-receptor, supports the observation that R5 virus is seemingly transmitted and infection established more efficiently than X4.188,191–193 There are only handful of individuals with genetically defective CCR5 expression to be identified as HIV-1 positive, and in all cases the infection is characterized solely by X4 virus.194,195 As the majority of HIV infections are initially R5, generalizations made about infection and disease progression are based on the observations of R5 virus.  However, roughly 50% of HIV-infected individuals will diverge from these generalizations and experience a tropism shift over the course of infection as increasing amounts of X4 virus emerge.196,197 This emergence of X4 virus is associated with the accelerated depletion of CD4+ T-cells and the progression to AIDS-defining illness.187,196,198–200 There is debate as to whether the emergence of X4 virus is a cause or a consequence of the immune system impairment experienced by HIV-1 positive individuals, as the mechanism driving this shift has yet to be determined.  Regardless, it is widely accepted that X4 virus serves as a prognostic marker of accelerated disease progression and decline in patient health when left untreated. 1.4 The Clinical Management of HIV 1.4.1 The Concept of Antiretroviral Therapy In the decades since HIV was confirmed as the causative agent of AIDS, a wide range of antiretroviral compounds have been approved for use in ART.  Though these compounds are not capable of curing HIV, they effectively interfere with the various stages  28 of the viral replication cycle, suppressing viral replication and limiting the amount of infectious virus in the blood.  Interfering with the HIV replication cycle is key in the fight against HIV for three reasons: 1) a reduction in plasma viral load (pVL) greatly reduces the rate of immune depletion in the infected individual and gives rise to the associated benefits; 2) reduced rates of replication translate into reduced rates of HIV mutation and the emergence of new drug resistance polymorphisms; and 3) less infectious HIV in the blood reduces the risk of HIV transmission.  For these reasons, identifying and treating HIV positive individuals has become the keystone of new initiatives to reduce rates of HIV and AIDS related-deaths worldwide.7,201 Initially, antiretroviral compounds were administered as monotherapies due to the limited number of compounds available.  Despite the ongoing discovery of new antiretroviral compounds and classes, the development of drug resistance polymorphisms continued to be an issue in HIV treatment.202,203  In 1996 a new treatment strategy, highly active antiretroviral therapy (HAART), was introduced and accepted as the standard of care.  HAART differed from the preexisting treatment methods by dictating the co-administration of at least three antiretroviral compounds.204–206  Typically HAART consists of a protease inhibitor or an non-nucleoside reverse transcriptase inhibitor (NNRTI) alongside two nucleoside reverse transcriptase inhibitors (NRTIs).207–209 As new compounds and new drug classes are approved for clinical use they are integrated into the HAART framework, providing the continued expansion of treatment options.  29 1.4.2 The Antiretroviral Compounds In 1987 the first anti-HIV compound was introduced, zidovudine (azidothymidine; ZDV or AZT), a nucleoside reverse transcriptase inhibitor.210–212 Zidovudine was discovered by screening pharmaceutical compound collections, and was originally designed as an anti-cancer drug in the 1960s.44 Unfortunately drug resistance polymorphisms negating the effects of zidovudine soon appeared within the treated HIV positive population.213,214 The pressure to discover and develop new antiretrovirals was great.  Antiretroviral discovery quickly shifted to a target-based approach, with a number of potential targets within the HIV replication cycle including viral entry and the enzymatic proteins reverse transcriptase, integrase and protease.44 Following the introduction of zidovudine, a number of additional NRTIs were to follow including lamivudine (3TC) and emtricitabine.  NRTIs are nucleoside analogues missing the 3’hydroxyl group on the deoxyribose moiety.  Mimicking the nucleotide building blocks of DNA, NRTIs act as chain terminators during reverse transcription.  When incorporated into the new DNA strand, the missing 3’-hydroxyl group interrupts the addition of the next sequential nucleotide.  Unable to form the necessary bond between nucleotides, the extension of the new HIV DNA strand is terminated.211,85 The rapid emergence of NRTI resistance polymorphisms limited the efficacy of HIV treatment well into the 1990s.  It wasn’t until 1995 that a new class of antiretroviral compounds was introduced, the protease inhibitors.  Protease inhibitors act by blocking the active site of the protease enzyme thus preventing the proteolytic cleavage of the Gag-Pol precursor protein.215,216 A key step in the HIV replication cycle, cleavage of the Gag-Pol precursor protein is necessary for virion maturation and infectivity.104,217,218 It was later  30 discovered that protease inhibitors can affect the activity of the Cytochrome P450 (CYP3A) metabolic enzyme, which is known to quickly metabolize many antiretroviral compounds.219–221 Ritonavir has been shown to be most effective in inhibiting CYP3A, and it is now common clinical practice to use low-dose ritonavir to enhance the pharmacokinetics of a co-administered protease inhibitor, known as “boosting”.219,222–224 The non-nucleoside reverse transcriptase inhibitors debuted in 1996 with the introduction of nevirapine, and later efavirenz (EFV) in 1998.225 The NNRTIs inhibit reverse transcription, but unlike the NRTIs, they accomplish this by binding to the reverse transcriptase enzyme itself.  Despite differences in shape and structure, all NNRTIs bind to a hydrophobic pocket that is created in the p66 subunit of reverse transcriptase.  NNRTIs do not bind at the polymerase active site rather near the active site, allosterically locking the active site in an inactive state.226–228 The final HIV enzymatic protein, integrase, was targeted with the development of the Integrase Strand Transfer Inhibitors (INSTIs). As the name suggests, this class of ARV compounds inhibits the HIV replication cycle by blocking strand transfer during which the proviral DNA is incorporated into the chromosomal DNA.95 In 2007, raltegravir was the first integrase inhibitor to be approved for clinical use in United States.229–232 The current mechanism of integrase inhibition targets the Mg2+ molecule found at the active site of the integrase enzyme, competitively blocking the DNA strands from the active site.233–235 Inhibiting the entry of HIV has also become a target in anti-HIV approaches, not only preventing viral entry but the replication cycle.  The entry inhibitors can be subdivided into fusion inhibitors and CCR5 antagonists, and together with the INSTIs are  31 the most recently approved anti-HIV drug classes.  The fusion inhibitor enfuvirtide was approved for clinical use in 2003, and continues to be the only fusion inhibitor available.  However, the administration of enfuvirtide requires an ill-favoured subcutaneous injection that has limited its use.236,237 Inhibition of fusion occurs at the gp41 envelope protein.  By binding directly to the gp41 protein, enfuvirtide prevents the formation of the hairpin loop necessary for drawing the virus and cell membranes close for fusion and subsequent viral entry.238,239 In addition, viral entry can also be prevented by way of allosterically blocking the host CCR5 co-receptor, preventing HIV-co-receptor binding and thus fusion and entry.  This is accomplished by the CCR5 antagonist compounds. 1.4.3 The CCR5 Antagonists It was discovered that a 32-base pair deletion in the gene encoding the CCR5 co-receptor when homozygous offered a natural resistance to HIV infection and partial resistance when heterozygous.191,193,240–243 Truncation of the gene leaves the co-receptor dysfunctional and not expressed on the cell surface thus preventing viral entry through the CCR5 co-receptor.  As HIV co-receptor use was elucidated it became apparent that this mutation was exclusively effective against R5-virus.240,244 Inherited following classical mendelian genetics, this CCR5Δ32 mutation is found primarily in Caucasian populations particularly of Northern European decent, with an average allelic frequency of 10% in Europe.241,243–248 When it was shown that the CCR5Δ32 mutation had little effect on the health of individuals expressing it, an anti-HIV drug acting at the CCR5 co-receptor quickly became an objective, leading to the development of the CCR5 antagonist drug class.191,249  32 A number of different CCR5 antagonists have been investigated, including: aplaviroc, vicriviroc, cenicriviroc and maraviroc.  The development of aplaviroc was discontinued following a series of Phase II clinical trials in which aplaviroc was outperformed by comparator compounds, and rates of hepatotoxicity were higher than expected.250–252  Vicriviroc was discontinued following the results of the Phase III clinical trials where the addition of vicriviroc to a strong optimized background therapy did not result in a significant increase in antiviral activity.253–255 Cenicriviroc is the most recent CCR5 antagonist to be explored, and is currently in the later stages of development following the promising results of the Phase IIb studies.256,257 Despite the clinical investigations of a number of CCR5 antagonists, maraviroc (MVC) is currently the only CCR5 antagonist approved for clinical use. 1.4.3.1 Inhibition of HIV by Maraviroc Of all the antiretroviral compounds in use, maraviroc is the first to target a host protein as opposed to a viral protein.  Maraviroc is a small, non-peptidic molecule that is highly selective for the CCR5 cell receptor.  Maraviroc does not alter the expression of CCR5 on the cell surface, nor does it seem to interrupt CCR5-related intracellular signaling.249 A number of studies agree that maraviroc works by allosteric inhibition.258–260 Instead of competitively blocking the active site where the gp120 would bind to the co-receptor, maraviroc binds elsewhere on the co-receptor leaving the active site vacant.  However, this non-competitive binding induces conformational change at the active site, which prevents gp120 from recognizing and binding to the CCR5 co-receptor.  33 The N-terminal and the second extracellular loop of the CCR5 co-receptor are highly associated with CCR5 ligand binding, as well as HIV binding. 177,261–264 However, instead of engaging these regions of the co-receptor maraviroc likely binds in a pocket located deep within the transmembrane domain of the receptor, contorting the HIV-1 binding site.258,260 Some also believe that maraviroc serves to stabilize the co-receptor in an inactive state258,265 By non-competitively inhibiting gp120 at the CCR5 co-receptor, maraviroc prevents the final conformational changes associated with fusion and viral entry. 1.4.3.2 The Clinical Trials of Maraviroc In 2007 the United States Food and Drug Administration (FDA) approved MVC for clinical use in patients with treatment-experience following the successful Phase III clinical trials, MOTIVATE 1 and MOTIVATE 2 (Maraviroc versus Optimized Therapy in Viremic Antiretroviral Treatment-Experienced Patients).  It was later shown to be effective and safe for treatment-naïve HIV-positive patients, as was demonstrated in the MERIT trial (Maraviroc versus Efavirenz in Treatment-Naïve Patients). The MOTIVATE trials were multi-centre, randomized, double-blind, placebo-controlled studies run in parallel.  MOTIVATE 1 was conducted in Canada and the United States, while MOTIVATE 2 was conducted in Australia, Europe and the United States.  The studies were designed to test the efficacy and safety of maraviroc in treatment-experienced patients screened as having R5 virus compared to placebo when accompanied by an optimized background regimen.266 The optimized background regimen was designed for each patient based on safety precautions, the presence of drug resistance polymorphisms  34 and treatment history.  The viral tropism was determined using the first clinically approved tropism testing method, the original phenotypic Trofile Assay by Monogram Biosciences.  The primary end point was the mean change in plasma viral load after 48 weeks of maraviroc therapy.266,267 The results of both MOTIVATE 1 and 2 showed superior virological effects with similar safety profiles when maraviroc was compared to placebo in patients with R5 virus.266,268 Analyses of maraviroc for treatment-experienced patients conducted at 96 weeks also showed a preferential safety and efficacy profile when compared to placebo.269 As part of the maraviroc clinical trials, a study was conducted to determine the safety and efficacy of maraviroc in patients screened as having non-R5 virus by the original Trofile assay.  This MOTIVATE sister study, A4001029, enrolled highly treatment-experienced patients from Australia, Canada, Europe and the United States.270 Patients were randomized to receive either placebo or maraviroc with an optimized background regimen (OBT).  The week 24 virological response revealed that maraviroc performed similarly when compared to placebo in patients with non-R5 virus, as would be expected given the mechanism of the drug.270 The MERIT trial was designed to evaluate treatment response to MVC when compared to efavirenz in treatment-naïve patients.  The MERIT trial was a double-blind, double-dummy, multinational study enrolling only patients screened as having R5 virus by the original Trofile assay.  Each patient was assigned to one of three treatment arms, either maraviroc once-daily (q.d.), twice-daily (b.i.d.) or efavirenz, with a co-formulated zidovudine-lamivudine backbone therapy.271 Maraviroc twice-daily was found to cause  35 both better CD4+ cell recovery and fewer adverse events.  However, it was only following tropism reassessment using the improved Enhanced Sensitivity Trofile Assay (ESTA) and subsequent reanalysis that maraviroc taken twice-daily and efavirenz showed similar virological response.271–273 Rescreening using ESTA identified 15% of the enrolled study population as having dual/mixed or X4 virus partially or fully uninhibited by maraviroc, respectively.  Reanalysis excluding these patients found maraviroc twice-daily to be non-inferior to efavirenz at both of the co-primary endpoints, the number of patients with pVL <400 and <50 copies/mL at week 48.271 Following both the MOTIVATE and MERIT trials, participants were recruited to continue their maraviroc treatment and clinical observations as part of five-year, open-label observational studies to evaluate the long-term safety of maraviroc.  The results of these follow-up studies reported favourable long-term safety and tolerance profiles for maraviroc in both treatment-experienced and treatment-naïve patients.274,275 1.4.3.3 Maraviroc in the Presence of CXCR4-Using Virus As an anti-HIV drug specifically targeting the CCR5 co-receptor, maraviroc is ineffective against virus using the CXCR4 co-receptor, and as one would expect is only partially effective against dual/mixed viral populations.  A subset of patients screened as having R5 virus and enrolled in the MOTIVATE trials did not achieve viral load suppression while on maraviroc.  A retrospective analysis of these non-responders was performed using genotypic techniques.  These methods identified approximately 50% of cases to have non-R5 virus at failure, and in 70% of these cases minority non-R5 variants were found present at baseline.276 The results of this analysis indicate that a number of  36 patients experienced tropism shift from R5 to non-R5 while on maraviroc.  Similar tropism switches have been observed with the use of experimental small, non-peptide CCR5 antagonists including vicriviroc, the development of which was abandoned following the results of phase III clinical studies.255,277,278 In addition to the inefficacy of maraviroc to control viremia of non-R5 virus, the selection of X4 variants is even less desirable, as X4 virus has been associated with faster disease progression.187,196 For these reasons, clinical guidelines strongly recommend the use of a tropism-screening test prior to introducing maraviroc into an antiretroviral therapy regimen. 1.4.4 Maraviroc in the Clinic 1.4.4.1 Treatment Guidelines Based on the results of the Phase III MOTIVATE trials, maraviroc was first approved for use in treatment-experienced adults, with detectable R5-tropic viremia by the FDA and shortly after by Health Canada in 2007.  Maraviroc was later approved for use in treatment-naïve individuals with exclusively R5 virus, as first line therapy in 2010 following the results of the MERIT study.  Though approved for use in a first-line antiretroviral therapy regimen, maraviroc is rarely prescribed as such in Canada.  More often due to good safety and tolerance profiles maraviroc is used as therapy for those with extensive drug resistance or for those experiencing tolerability issues associated with other antiretroviral compounds. The dosing and treatment guidelines of maraviroc are similar between prescribing countries, such that maraviroc is prescribed as 150mg or 300mg doses taken twice-daily, and in some instances as a single 600mg dose.  To ensure safety and efficacy, the  37 prescription dose is dependent on the co-administered compounds of the antiretroviral therapy regimen and other patient characteristics such as treatment history.279 As maraviroc is primarily cleared by liver-associated CYP3A4 metabolism a 150mg dose is beneficial for patients with a CYP3A4 inhibitor, such as a ritonavir-boosted protease inhibitor, in their antiretroviral therapy regimen.  As well, prescribing maraviroc to patients with renal and hepatic impairments must be done with caution.280–282 Maraviroc is not to be prescribed when the presence of non-R5 virus has been detected at any point in the patient treatment history.  As maraviroc is only effective against R5 virus, tropism screening is highly recommended by all guidelines when considering the use of maraviroc as part of an antiretroviral therapy regimen.279,283–285 1.4.4.2 Testing for Tropism Prior to Initiating Maraviroc There are a number of validated assays to test for viral tropism using both cell-based (phenotypic) and genotypic methods.  Phenotypic methods use patient-derived virus to infect well-characterized cell lines with reporter properties such that results are dependent on successful infection of the cell lines.  For instance, the MT-2 assay co-cultures patient peripheral blood mononuclear cells (PBMCs) with MT-2 cells expressing CD4 and CXCR4.186 Tropism is determined based on the presence or absence of the formation of syncytia, which identifies patient virus as either SI (CXCR4-using) or NSI (CCR5-using).186  More recently, cell-based assays such as the Enhanced Sensitivity Trofile Assay (ESTA) make use of lab-derived viral vectors for co-transfection.267,286 First, HIV is extracted from patient plasma and the entire env gene is amplified.  The amplified patient env is then inserted into an env expression vector.  A second, non-replicative vector is also prepared where the env gene has been deleted and replaced by a luciferase indicator gene.  The env  38 expression vector and luciferase-containing vector are co-transfected in human embryonic kidney cells in order to create replication-defective recombinant virus containing both the patient env gene and the luciferase reporter gene.  These recombinant viruses are harvested and subsequently used to infect human primary glioblastoma, U87, cell lines expressing the CD4 receptor and either CXCR4 or CCR5 co-receptors.  Tropism results are indicated by the reporter luminescence of successfully infected U87 cells, and further assessed in the presence and absence of CXCR4 and CCR5 co-receptor antagonists.267,286 Genotypic methods are based on the genetic sequence of the gp120 V3 loop encoded by the env gene, and are heavily dependent on the interpretation of these sequences.  Sequence data is generated from viral RNA and interpreted using a number of different bioinformatics algorithms. 287–289 There are two primary sequencing technologies used to determine HIV-1 tropism, population-based sequencing and next generation sequencing.290–293 In the event of a highly variant DNA, as is case with HIV, population-based sequencing is only capable of identifying variants that occupy 20% or more of the viral population within a patient.  Conversely, next generation sequencing technologies claim to identify minority variants comprising <1% of the viral population.  Despite being unable to detect minority variants below the 20% threshold, population-based methods continue to be a practical tool used in the clinical detection of viral tropism, whereas next generation sequencing remains predominantly a research tool.  However, as clinical applications continue to increase and operational costs decrease with time and innovation, it is likely that next generation sequencing will quickly become a standard in clinical testing.  39 The most-widely used phenotypic and genotypic assays use viral isolates from plasma to predict tropism; however, this can become problematic when plasma viral loads are suppressed below detectable levels.  Recent studies have shown the potential of alternatives to testing plasma RNA, such as the use of viral DNA extracted from PBMCs or a stored plasma sample taken during a period of detectable viral load.292,294–299 However, despite the modest correlation between viral RNA and these alternatives, guidelines continue to recommend genotypic tropism testing be performed using plasma viral RNA collected as close to the maraviroc start date as possible.279,283 1.5 Sequencing 1.5.1 Understanding HIV through Genetics A patient with an HIV infection has millions of individual HIV variants comprising a viral population.  The average viral population is highly diverse, with only a few primary HIV variants accounting for the vast majority of virus found in the blood; the remainder of the population is composed of minority variants.  These minority variants emerge as errors occur in the viral replication cycle, and play an important role in the response to treatment and the progression of infection.300,301 Over the course of infection, a viral population becomes increasingly diverse as HIV is characterized by a high rate of replication, a high rate of mutation, is capable of recombination, and is a persistent and lifelong infection.21,86,302–305  In addition, this viral diversity is augmented by selection pressures exerted on the virus by the anti-HIV efforts of the immune system and/or antiretroviral therapy.306–309  40 The large amount of genetic variation in HIV populations can be attributed to both viral and host processes.  A principal source of HIV diversity is the low fidelity of the HIV reverse transcriptase enzyme.  Errors in nucleotide incorporation called substitutions, occur frequently during reverse transcription as reverse transcriptase lacks the help of a “proof reader” to identify and remove misincorporated nucleotides.87,310,311 In addition, single nucleotide insertions and deletions are common errors, particularly in regions of the HIV genome characterized by a string of the same nucleotide repeated, or homopolymers.87,306,311 It is estimated that on average one nucleotide error is made during the generation of each HIV genome.86,305,308 HIV has a high replication rate capable of generating up to a billion new virions in a single day, which amplifies the effects of this high rate of mutation.302,308 Furthermore, recombination in the HIV genome occurs regularly.312–314 Many infected CD4 cells harbor the viral DNA of more than one HIV variant, allowing two different HIV RNA molecules to be packaged into budding virions and the formation of heterozygous virions.314–316 During the reverse transcription of RNA from a heterozygous virion, the reverse transcriptase enzyme can jump between the two HIV RNA templates to create recombinant complementary DNA (cDNA) molecules.317–319 These viral characteristics enable the accumulation of divergent variants in the HIV population as infection progresses. The evolution of the HIV population is enabled by the aforementioned errors in viral replication, and guided by selection pressures including those exerted by the immune system.  A series of host proteins act as HIV restriction factors, including APOBEC3G, TRIM 5α and tetherin.320  Of these restriction factors, APOBEC3G exerts its antiviral effects during reverse transcription.  A cytosine deaminase expressed largely by hematopoietic  41 stem cells, APOBEC3G promotes the substitution of uracil in place of cytosine, causing guanine to adenine mutations in the growing HIV cDNA strand and instability of the new cDNA molecule.321–323 However, the HIV protein Vif has evolved to overcome these crippling effects by triggering the degradation of the APOBEC3G protein and thus preventing its inclusion into the newly forming virions.320,323  Many of the errors that occur during the replication of HIV produce non-functional virus; other errors affect viral fitness negatively and still others heighten viral fitness under pressures exerted by the immune response and/or antiretroviral treatment.  The combination of errors during replication and selection pressures, both Darwinian and purifying, lead to an extremely divergent viral population, and the potential for drug resistance or tropism-associated polymorphisms to emerge.  In order to ensure the most efficient HIV treatment experience, sequencing relevant segments of the HIV genome can reveal drug resistance polymorphisms and more recently, co-receptor use.  Furthermore, studying HIV genetically is also fundamental in the continuing ambition of the scientific community to fully understand HIV as a virus, its structure, its mechanisms of action and its interactions. 1.5.2 Sequencing Technologies 1.5.2.1 Concepts in Sequencing At the core of DNA (deoxyribonucleic acid) are four nitrogenous bases adenine (A), cytosine (C), guanine (G) and thymine (T).  Together with a deoxyribose sugar and a triphosphate group, nitrogenous bases form deoxyribonucleotide triphosphates (dNTPs), or nucleotides, the ‘N’ in dNTP representing all four bases.  The order of nucleotides in a  42 DNA molecule determines its genetic uniqueness.  DNA sequencing uses the principles of naturally occurring DNA synthesis to determine this nucleotide order, gaining information about the structure and function of proteins and other components of the virus in the sample being analyzed. The polymerase chain reaction (PCR) has become an indispensable and efficient molecular technique used to make DNA easier to detect and sequence.  Prior to the development of PCR, DNA amplification was dependent on the relatively time-consuming process of in vivo cloning.  Reliant on the biological properties of DNA synthesis and thermal cycling, PCR enables the mass replication, or amplification, of a specific primer targeted DNA sequence.324,325 The use of PCR has more recently been expanded to include the addition of reaction components for specific PCR and sequencing systems as well as tags for sample identification when sequencing multiple samples in parallel. The PCR reaction starts where the primer, a designed, short chain of nucleotides, binds to the sample DNA.  Extending from the primer at the 5’ end, the DNA polymerase enzyme incorporates dNTPs as it progresses toward the 3’ end of the newly synthesized DNA chain.324,325 The amplified DNA fragment is then used as a template for the sequencing reaction.  Based on the same principles as PCR, the placement of nucleotides during the sequencing reaction is recorded and translated into a readable DNA sequence.  Though dependent on these basic principles, sequencing techniques vary in sample preparation and the way in which nucleotides are added and detected.  43 1.5.2.2 Population-based Sequencing The viral population of an HIV infected individual is composed of millions of different HIV variants.  Population-based sequencing does not attempt to identify every individual variant, instead it takes an “average” of the sampled population to generate a consensus sequence.  This consensus sequence is generated by collapsing all of the observed sequences to produce a single sequence.  In the instance of multiple bases observed at a single position, the bases are reported together as a mixture.  In this way, the consensus does not represent a single variant but the population of variants.  Though the individual variants cannot be recognized, any clinically relevant polymorphisms present in the viral population at a prevalence of approximately 20% or more can be identified. Population-based sequencing methods were engineered using principles initially introduced in 1969 by a group investigating the use of dideoxythymidine triphosphate (ddTTP).  An analogue to the thymine nucleotide triphosphate (dTTP), ddTTP lacks a 3’-hydroxyl group necessary when forming the bonds between adjacent nucleotides during DNA synthesis.326 As ddTTP is incorporated into the newly synthesized DNA in place of dTTP, it stops DNA extension, terminating the DNA chain.326 In 1977, Frederick Sanger and his group used this and similar concepts to develop what is now known as Sanger sequencing.327,328 Sanger sequencing, or dideoxy sequencing, is based on the incorporation of dideoxynucleotide triphosphates (ddNTPs) causing chain-termination during DNA synthesis.328 In this method, four parallel reactions are run, one for each of the four ddNTPs such that each nucleotides, A, C, G, T, can be randomly incorporated into the growing  44 DNA strand as a chain terminator.  The ratio of dNTPs to ddNTPs promotes the random incorporation of ddNTPs as the sequencing reaction is repeated to generate sequence fragments of all possible lengths.  These fragments can then be sorted by size, enabling the sequence to be read sequentially based on the ddNTP incorporated into each of the consecutively sized DNA fragments.  Initially the four reactions were analyzed in parallel lanes using gel electrophoresis, one lane for each of the four ddNTPs, to reconstruct the DNA sequence.328,329 However, technical advances led to fluorescently labeling the four ddNTPs used in chain termination sequencing.  This improvement enabled the four ddNTP reactions to be combined into a single reaction and fragments separated in a single lane.330 These methods were further improved upon with the introduction of capillary electrophoresis and automated sequencing.331–335 The introduction and combination of such increasingly efficient molecular techniques has led to modern population-based sequencing.  A simplified overview of the automated dideoxy sequencing method is schematically represented in Figure 1.8. Population-based sequencing employs a single instrument to consolidate the aforementioned techniques.  Sequencers developed by Applied Biosystems, Inc. use fluorescent dideoxy chain termination to generate DNA fragments and capillary electrophoresis to determine the genetic sequence.336 Capillary electrophoresis occurs at a microscopic scale where DNA fragments are separated by size in microtubes known as capillaries.337–339 The size and automation of this process allows multiple capillaries, or arrays, to sequence multiple samples simultaneously.340,341 In order to determine the DNA sequence, separated DNA fragments pass by a laser within the capillary.  The laser is used to activate a fluorophore found on each of the chain-terminating nucleotide analogue  45 ddNTPs, and a sensor captures the fluorescent wavelength emitted by this excitation.  A computer then reconstructs the sequence based on the order of captured fluorescence and signal strength.  The assembled sequences are used to generate a consensus sequence, which can be visualized as an interpretable chromatogram.336,337,341  Figure 1.8.  The Sanger sequencing method. A simplified schematic of the Sanger sequencing method, Panel A depicts the dideoxynucleotide triphosphate (ddNTP) in comparison to the naturally occurring deoxynucleotide triphosphate (dNTP); the discrepancy is the replacement of the hydroxyl group at the 3’ carbon with a hydrogen atom.  The foundation of Sanger sequencing is the random incorporation of chain-terminating ddNTPs during DNA synthesis in place of the respective dNTPs, preventing the extension of the DNA chain as in Panel B.  Thousands of DNA fragments are then sorted by size, such that sequential nucleotides are determined by the last incorporated nucleotide, panel C.  The sequence is then depicted as a chromatogram, as in panel D.  This figure was created by the author.  Panels C and D were recreated based on a figure published by Kircher et al., Bioessays342 ©2010 (RightsLink license No. 3521630394688)   46 1.5.2.3 Next Generation “Deep” Sequencing Relatively recent advancements in technology, computing and laboratory methods have revolutionized DNA sequencing and lead to a group of new technologies often referred to as “next generation sequencing”.  Next generation sequencing technologies are massively parallel, capable of generating significantly more data than conventional sequencing methods.  As well, the ability to detect and subsequently quantify the prevalence of increasingly minority variants within a sample has made next generation sequencing an invaluable tool when studying genetically divergent samples.  These characteristics give way to the name “deep” sequencing. The first commercially successful “deep” sequencing platform, the 454 genome sequencer FLX (GS-FLX), was released by 454 Life Sciences (Roche Applied Science; Basel, Switzerland) in 2005, and the smaller 454 genome sequence junior (GS-Jr.) in 2009.343 Like population-based sequencing, “deep” sequencing sample preparation starts with the PCR amplification of a target DNA segment in order to generate thousands of copies of the fragment to enhance detection.  These primer targeted DNA amplicons are created for each sample.  Alternatively, relatively short whole genomes can be amplified using a “shotgun” approach which randomly fragments sample DNA, the sequences of which are later stitched together using computational tools.343–345 When preparing amplicon libraries for “deep” sequencing, a second PCR reaction is performed to ligate a fusion primer to each amplicon.  The fusion primer is composed of a 454 platform specific adaptor sequence, either A or B, that is linked to an oligonucleotide multiplex identifier (MID) and a sequence specific primer.  MID tagmentation provides each amplicon with a unique identifier for downstream analysis such that each sequence can be linked to a particular sample.  The  47 sequence specific primer binds to the target amplicon, linking the amplicon to the MID and adaptor sequence.  The adaptor sequence is used to manipulate the amplicons throughout sample preparation and sequencing. Amplicon libraries are generated when MID-labeled amplicons from multiple samples are combined.  Samples are pooled in equal amounts to prevent preferential amplification during the subsequent emulsion-facilitated clonal PCR.  Emulsion PCR (emPCR) for 454 “deep” sequencing utilizes oligonucleotide covered microbeads, or capture beads; these oligonucleotides are complementary to the adaptor sequences previously ligated to the sample amplicons.  Amplicons are bound to the capture beads in conditions that favour the attachment of one DNA amplicon to one bead.  A water-in-oil emulsion is then created with droplets large enough to incorporate a single DNA-bound capture bead and PCR amplification reagents.343,346–348 Within the droplet sample DNA is clonally amplified, each new DNA strand attaching to an existing bead oligonucleotide resulting in a microbead covered in millions of copies of amplicon template.343,344,347 DNA coated beads liberated from PCR emulsions are subsequently prepared for the sequencing reaction. 454 “deep” sequencing is a sequence by synthesis reaction much like population-based methods.  However, unlike population-based methods, 454 “deep” sequencing is a pyrosequencing reaction, meaning that it detects the release of a pyrophosphate group (PPi) upon nucleotide incorporation using chemiluminescence.343,344,349,350 In a fiber-optic Pico Titer Plate composed of over a million microscopic wells, individual DNA-coated beads settle into a single well.343 Surrounding the DNA bead is a three-bead system: 1)  48 packing beads added to stabilize the components of the reaction within the well; 2) enzyme beads carrying the chemiluminescent sulfurylase and luciferase enzymes; and 3) PPiase beads added to remove excess PPi and prevent signal interference.  The 454 system is designed to cycle a flow of each nucleotide sequentially over the reaction wells.343,344 As the dNTPs are incorporated into the DNA strands PPi is released and converted into adenosine triphosphate (ATP) by sulfurylase.  ATP then acts as a substrate for luciferase, generating a light response the intensity of which is proportional to the number of nucleotides incorporated during the flow and captured by an on-board camera.343,344,349 After each nucleotide flow, unincorporated dNTPs and excess ATP are degraded by apyrase, readying the reaction wells for the next nucleotide flow.343,344 Sequences are computationally reconstructed when flow cycle data is combined with image data using pixel coordinates for each well. The abundance of sequence data produced by this technology requires the use of advanced computational tools for analyses.  49  Figure 1.9.  Next generation 454 “deep” sequencing. A simplified schematic of the complete Roche 454 workflow.  Sample preparation begins with the generation of an amplicon library.  After selective amplification, amplicons are tagged with 454 adaptors that bind specifically to capture beads.  The bound amplicons are then amplified in an emulsion PCR reaction, such that one amplicon is bound per bead and incorporated into a single emulsion micelle.  The result is a “hairy” capture bead, covered in the thousands of copies of the amplicon.  The beads are loaded into picoliter wells on the 454 sequencing plate (PTP), the size of the bead permitting a single bead per well.  The pyrosequencing reaction, highlighted in the bottom panel, chemically induces a light reaction as sequential nucleotides are incorporated into the growing DNA chain.  Each light emission and its intensity is captured by a camera and interpreted computationally.  This image has been published by Mardis et al., Trends in Genetics ©2008 and reproduced here under RightsLink license No. 3522020201830.351 Following the introduction of the 454 GS-FLX sequencing platform a number of alternative next generation sequencing technologies have been, and continue to be, developed.  Biotechnology companies like Ion Torrent, Illumina and Pacific Biosciences have each developed their own next generation sequencing technology.  Each platform uses a unique combination of steps and chemistries in sample preparation and when detecting the incorporation of nucleotides in a growing DNA chain.    50 The Ion Torrent platforms (Ion Personal Genome Machine and Ion Proton) use semiconductor sequencing.  Sample library preparation is very similar to that of 454 pyrosequencing, using adaptor-liganded DNA libraries, capture beads and emulsion PCR for clonal amplification.352 These template-covered beads are distributed into millions of microscopic wells found on the ion chip, over which each nucleotide is flowed systematically.352 The primary difference between 454 and Ion Torrent sequencing is the chemistry used to detect newly incorporated nucleotides.  Whereas 454 platforms use a catalyzed light reaction and camera-capture method, Ion Torrent platforms rely on the detection and measurement of the hydrogen atoms released as a result of nucleotide incorporation.352 The Illumina platforms (MiSeq, HiSeq and NextSeq) are different in their sample preparation, which relies on bridge amplification to generate clusters of clonal sequence templates on a solid surface, the flow cell.353,354  Similar to Sanger sequencing, the DNA building blocks are fluorescently labeled dNTPs that cause chain termination.  The dNTPs used in Illumina sequencing are called reversible terminators, and are characterized by a chain terminating base that inhibits the polymerization of the subsequent nucleotide.354,355 The four dNTPs are washed over the flow cell simultaneously, however, only one nucleotide will be incorporated into each growing DNA chain during each wash.356 A sequence-by-synthesis approach, the sequences are generated by capturing the fluorescent emission released when a nucleotide is incorporated into the growing DNA chain, using an on-board camera.  This is somewhat comparable to the capture of the light emission released during the pyrosequencing reaction on the 454 platform.  After each wash both the fluorescent label and chain terminator are removed from the last dNTP to enable the  51 incorporation of the next nucleotide in the sequence.357 The Illumina MiSeq platform stands out from other currently available platforms by providing the highest number of reads per run, with the lowest error rate at a reasonable cost.358,359 The SMRT (Single Molecule Real Time) DNA sequencing system developed by Pacific Biosciences does not require a clonal amplification step.  SMRT introduces fluorescent labels linked to the phosphate group of the dNTPs as opposed to the base as seen in Sanger and Illumina-platform sequencing.  This allows the fluorescent label to be released upon nucleotide incorporation, removing the need for systematic nucleotide wash cycles.360,361 As well, SMRT makes use of a chip characterized by nanophotonic, glass-bottomed wells.  The structure of these wells ensures that the laser focuses on the bottom portion of the well where a single DNA polymerase and a single DNA template are immobilized.360,361  The four fluorescently labeled nucleotides diffuse in and out of the wells simultaneously, interacting with the template DNA-bound DNA polymerase.  The nucleotides are incorporated into the growing DNA chain in real time, causing a burst of fluorescence lasting milliseconds that is used to determine the genetic sequence.360,361 This technology allows the Pacific Biosciences platform to provide the longest read lengths, in the least amount of time, at the lowest cost per run.  However, it has the highest error rate and the lowest read depth of currently available next generation sequencing platforms.358 Each platform has its advantages and disadvantages when considering practical measures like read length, read depth, run time, cost and error rate.  In all cases, an abundance of sequence data is produced by each of these platforms.358,359,362,363  For example, the 454 GS-FLX with Titanium chemistry can generate one million sequences,  52 whereas the Illumina MiSeq with updated chemistry can generate upwards of 22 million sequences.358 Regardless of the platform used, this volume of data requires the use of advanced computational tools for analyses. 1.5.3 From Sequence to Relevance Raw sequence data is relatively uninformative, and requires a stepwise series of analyses before it can be interpreted.  This process begins with ensuring the sequence is what it was intended to be by aligning and comparing each sequence to a known reference sequence.  In the case of data generated by population-based methods, sequences are representative of variants occupying at least 20% of the sample population and likely to include up to three primary variants.  The accuracy of nucleotide calls, or “base-calling”, and identified polymorphisms are reviewed for sense using manual, or more recent automated base-calling software.364–366 However, in the case of data generated by 454 “deep” sequencing, sequences are representative of variants occupying as little as <1% of the sample population and likely to include hundreds of variants for each sample.  This large amount of data requires large-scale on-board and post-sequencer processing to ensure quality and accuracy of sequence data, a step that varies as laboratories often use in-house bioinformatics processing pipelines.344 Sequence data can be evaluated in a large variety of ways using bioinformatics.  Bioinformatics is a broad term describing the computational tools used to make sense of biological data by combining computer science, statistics and mathematics.367–369 The emergence of such computational analyses as applied to sequence data coincided with the increased popularity of automated sequencing and the resultant increase in data  53 generation.367,370 Such computational tools can be designed to quickly align sequences, compare sequence databases, identify polymorphisms, illustrate genetic diversity and apply mathematical algorithms.370,371 Algorithms can be customized for specific analyses to extract detailed information from sequence data.  They are intricate sets of instructions outlining a predetermined series of calculations used for data analysis.  Regularly employed in bioinformatics, algorithms are essential in the genotypic prediction of HIV tropism. 1.5.4 Genotyping & Bioinformatics for Tropism Prediction The use of genotyping to predict HIV tropism is dependent on the relatively short gp120 V3 loop sequence.372,373 For over two decades the presence of a positive charge, basic amino acid, either arginine or lysine, at the V3 amino acid positions 11 and or 25 has been known to be strongly associated with SI, or X4 virus.287,374,375 This observation led to the longstanding 11/25 charge rule for co-receptor prediction.  Additional mutations observed within the V3 loop have since been associated with X4 virus.  However, not all mutations within the V3 loop influence the phenotype equally, which complicates this relatively simplistic rule-based approach to tropism prediction.  More recently, improvement and advancement in both sequencing technology and computing have demanded and supported the development of increasingly better predictive tools for use in tropism testing.288,289,376 There are a number of algorithms used to interpret tropism from V3 sequences including the arguably favoured Position Specific Scoring Matrix (PSSMX4/R5) and Geno2Pheno[coreceptor] (g2p) co-receptor algorithms.377–379  54 PSSMX4/R5 was introduced as a tool to predict tropism in 2003.  Based on the mathematical matrix, an array of numbers or symbols arranged in columns and rows to which operations can be applied, PSSMs are developed to identify patterns, or motifs, within a sequence.  PSSMs are trained on series of aligned sequences incorporating genetic variation, taken from virus known to express the desired traits as a baseline comparator.288,380 For instance, when applied to tropism the scoring matrix is developed using V3 sequences collected from viruses displaying characteristics associated with the R5 and X4 phenotypes.  Each amino acid position is incorporated into the scoring matrix and the potential residues are assigned a score based on the likelihood of supporting the X4 phenotype.  When the matrix is applied to generated sequences, a PSSM score is calculated and used to infer the degree of resemblance to known X4 sequence at the amino acid level, such that a higher score infers a more X4-capable sequence.288 Like PSSM, g2p is a computational method of analyzing sequence data to recognize patterns and relationships within it.  G2p is a support vector machine first applied to the estimation of HIV drug resistance over a decade ago.381–383 The algorithm has since been adapted for use in the prediction of HIV tropism and for use with data sets generated using 454 next generation sequencing.289,384,385 Somewhat more complex than a matrix, support vector machines like g2p construct a multi-dimensional plane within a multi-dimensional space.  As with PSSM, g2p is trained using a large set of reference sequences known to be associated with the traits of interest from paired genotype-phenotype data.  Sequences are mapped in space using binary classifications forming data point clusters belonging to one of two categories such as X4 and R5 when predicting tropism.  The optimal plane used for analysis delineates the two clusters of data points by the largest space possible.  This plane  55 acts as a decision boundary such that sequences are mapped on a continuous scale where sequences that are increasingly more R5-like are plotted at the greatest distance from those that are increasingly more X4-like when predicting tropism.  The g2p analysis results in a false positive rate (FPR) on a scale from 0 to 100.  When predicting HIV tropism, this value indicates the likelihood that the sequence is falsely identified as X4 therefore a lower FPR indicates a more X4-like sequence. Statistical learning methods, like PSSM and g2p, have been trained to predict HIV tropism.  Due to the continuous outputs, these methods require the application of a pre-defined cutoff point to determine the point at which sequences begin to resemble X4 variants.  There are a number of different cutoff points in use for predicting X4 variants, but an established universal cutoff point has yet to be agreed upon.  These tropism prediction tools are evaluated on their ability to rightly predict the X4 phenotype from genotypic data, as defined by the sensitivity and specificity of a test.  Sensitivity, or true positive rate, is defined as the rate of correctly identifying a sequence as X4.  The specificity, or true negative rate, is defined as the rate of correctly identifying a sequence as not being X4.  The sensitivity and specificity are calculated based on “truth” defined by the gold standard tropism testing method.  As the first commercially available tropism test, the Trofile phenotypic assay was the de facto standard at the time of the analyses described in this thesis.  Adjusting the cutoff point can influence the performance of the algorithm such that the ideal cutoff would best balance the sensitivity and specificity.  As sequencing technologies improve and algorithms are developed and optimized, genotypic tropism assays will continue to improve.  56 1.6 Thesis Objectives & Organization 1.6.1 Thesis Objectives The objectives of this thesis are to demonstrate the clinical utility of a population-based genotypic tropism assay to infer HIV-1 tropism, and to determine the necessity for such a test prior to initiating maraviroc as part of an antiretroviral treatment regimen. More specifically, this thesis seeks to describe the validation of a population-based sequencing method for inferring HIV tropism to be applied when evaluating the potential use of maraviroc in the clinical setting.  As well, this thesis highlights the necessity of tropism inference prior to initiating maraviroc by detailing the results of two retrospective studies where participants were exposed to maraviroc despite the presence of non-R5 virus. Using both population-based and next generation 454 “deep” sequencing methods, this thesis demonstrates the facility of tropism inference using genotypic methods, and illustrates the effects of maraviroc on HIV-1 variants within non-R5 viral populations. 1.6.2 Thesis Organization & Structure This thesis is divided into six chapters.  Chapter one provides a general introduction to key topics for understanding the content presented in the body of the thesis.  This introduction includes a generalized overview of the nature of HIV-1 infection and how we attempt to both control and understand it.  The body of the thesis is composed of four chapters detailing three previously published manuscripts and a fourth  57 unpublished study, all of which use sequencing technologies to determine viral tropism and the effects of maraviroc on both R5 and non-R5 viral populations.  The final chapter provides a conclusive overview and interpretation of the four studies presented in the body of the thesis and their relevance to the field of HIV-1 study. Chapters two and three discuss the validation of a “population-based,” Sanger sequencing approach to infer viral tropism from the genetic sequence of the gp120 V3 loop.  Samples were originally collected as part of the MOTIVATE, A4001029 and MERIT studies, from enrolled study participants.  As with many assays, genotyping results were compared with the original Trofile assay, the gold-standard of tropism testing at the time, and in the case of the MERIT trial they were also compared to the tropism inferences made by the Enhanced Sensitivity Trofile Assay.  The studies described in chapters two and three go one step further in adjudicating the performance of V3 genotype by comparing tropism inference to the virological response of participants of the maraviroc clinical trials.  The use of therapeutic response data to optimize a genotypic tropism assay can be considered advantageous given the clinical relevance of such a measure and population size of the studies.  Given the elusive nature of the perfect clinical assay, virological response can better represent truth when determining the effectiveness of maraviroc in viral populations. The effects of maraviroc in viral populations composed of both R5 and non-R5 variants were explored in the studies presented in chapters four and five.  Two study populations were prescribed maraviroc as an add-on therapy to a pre-existing, failing antiretroviral therapy regimen, in essence serving as short-term maraviroc monotherapy.   58 Longitudinal samples were taken to evaluate the effects of maraviroc on R5 and non-R5 variant distribution over as few as eight days.  Using next generation 454 “deep” sequencing to infer viral tropism from the gp120 V3 loop as described in other studies, the shifts in the viral population were traced.291,292 Assessment of tropism inference and virological response can provide valuable insight into the threshold of maraviroc effectiveness.  Analyses of this nature have the potential to guide the development of widely accepted and clinically relevant tests inferring tropism using genotypic methods.  59 Chapter Two: Population-based V3 Genotypic Tropism Assay: a Retrospective Analysis Using Screening Samples from the A4001029 and MOTIVATE Studies 2.1 Introduction CCR5 antagonists are a relatively recently developed class of antiretroviral drugs for the treatment of HIV-1 infection.  The activity of this drug class is based on the general requirement for HIV to use a co-receptor (CCR5 or CXCR4) to infect target CD4+ T cells.  Only viruses that use the CCR5 co-receptor are susceptible to this drug class.386 Viruses that use CCR5 are generally termed R5, whereas viruses that use CXCR4 are termed X4.  Viral populations that are able to use both CCR5 and CXCR4 receptors (dual) or contain both R5 and X4 virus (mixed) are often termed dual/mixed.76,387 A handful of CCR5 antagonists have been investigated, including: maraviroc, aplaviroc, vicriviroc and cenicriviroc.  The development of both aplaviroc and vicriviroc was abandoned following the results of Phase II and Phase III clinical trials, respectively.250–252,255 The most recent, cenicriviroc, is still in development following the promising results of the Phase II clinical trials.256,257 Currently, however, maraviroc is the only approved CCR5 antagonist, and has been approved for use in both treatment-experienced and treatment-naive patients.247,270,261 The maraviroc treatment-experienced clinical development program consisted of three studies: MOTIVATE-1, MOTIVATE-2 and A4001029.  The MOTIVATE-1 and MOTIVATE-2 studies (N=1049) have shown that maraviroc in combination with optimized  60 background therapy can effectively suppress viral load to undetectable levels in treatment-experienced patients with R5 tropic virus as determined by the Trofile assay.266 The studies were double blind and patients were randomized in a 1:2:2 ratio to receive either placebo (PBO), maraviroc once-daily (q.d.) or maraviroc twice-daily (b.i.d.) in combination with optimized background therapy.  After 48 weeks, patients receiving maraviroc showed a greater response to treatment where approximately 45% of these patients were able to suppress their HIV-1 RNA levels below detectable limits (<50 copies/ml).266 A sister safety study, A4001029 (N=186), with a similar design enrolled patients with ‘non-R5’ using virus, X4 or dual/mixed or non-phenotypable.  In these patients with non-R5 virus, maraviroc was not shown to significantly reduce viral load when compared with placebo.270 For this reason, treatment guidelines recommend screening for CCR5 use prior to initiating therapy with a CCR5 antagonist-based regimen. The original phenotypic Trofile assay (Monogram Biosciences, South San Francisco, California, USA) was used to screen patients in the MOTIVATE-1, MOTIVATE-2 and A4001029 studies.266,270 Details of the Trofile assay are described elsewhere.267 Briefly, this assay uses PCR to amplify HIV env sequence from virus in patient plasma, which is used to create a pseudovirus for characterization in a cell-based assay to determine co-receptor usage.  Samples are classified as R5, X4, or dual/mixed after the detection of a luciferase activity on CCR5+ and/or CXCR4+ expressing cell lines confirmed by inhibition with a CCR5 or CXCR4 antagonist.267 To date, Trofile has been the most widely used method for determination of tropism in both clinical practice and clinical trials.387 More recently, the original Trofile assay has been superseded by an enhanced version with increased  61 sensitivity called the Enhanced Sensitivity Trofile Assay (ESTA).286 ESTA was not used in the MOTIVATE or A4001029 studies. The genetic determinants of HIV co-receptor usage are primarily located in the third hypervariable domain (V3 loop) of the HIV envelope.373,374 This raises the possibility that a clinically useful tropism assay could be developed based on a genetic analysis of the V3 region.  Such an assay could have considerable advantages over Trofile, including cost, availability, turnaround time, and other practical considerations.  A number of different algorithms for the interpretation of V3 sequence data and prediction of tropism exist, including geno2pheno (g2p) and position-specific scoring matrices (PSSM).288,388 Earlier studies comparing genotypic methods for tropism prediction with Trofile suggested an apparently poor sensitivity for genotypic approaches, which would preclude their use.378,389 This study is the first large- scale analysis comparing the utility of V3 genotyping and the original Trofile assay to predict clinical outcomes following the use of maraviroc in treatment-experienced patients.  Such analyses linked to clinical outcomes provide a relevant parameter when comparing the predictive power of genotypic and phenotypic tropism assays. 2.2 Methods 2.2.1 Study Population Using patient populations from the MOTIVATE-1, MOTIVATE-2 and A4001029 studies, we retrospectively examined the ability of population sequencing of the HIV-1 V3  62 loop to accurately predict virological responses to maraviroc.  Patients included in the MOTIVATE-1 and MOTIVATE-2 studies were required to meet a number of criteria, including plasma RNA viral load at least 5000 copies/mL and confirmed drug resistance and/or treatment experience, before entering the study.266 A total of 3244 patients were initially screened and assessed for eligibility using the Trofile assay: 1049 were entered into the MOTIVATE-1 (Canada and the United States) and MOTIVATE-2 (Australia, Europe and the United States) studies as R5 tropic, 955 were excluded from the study following classification as X4, dual/mixed or non-phenotypable.  The first tranche of patients screened to have ‘non-R5’ virus, were entered into the A4001029 study (N=186).270 Owing to the similar baseline characteristics and identical study design of A4001029, MOTIVATE-1 and MOTIVATE-2, we were able to pool the three study populations. 2.2.2 Sequencing Viral RNA was extracted from 500 µL of frozen plasma using the automated, standard operating procedures for the NucliSENS easyMAG (bioMérieux).  Using nested RT-PCR methods, the V3 loop of gp120 from the HIV-1 env gene was amplified in triplicate from the RNA extracts.  Reverse transcriptase (RT) and first-round PCR reactions were combined into one step by employing SuperScriptTM III One-Step RT-PCR system with Platinum Taq High Fidelity enzyme (Invitrogen, Burlington, Canada) alongside forward primer SQV3F1 (5’ GAG CCA ATT CCC ATA CAT TAT TGT 3’; HXB2 6855-6878) and reverse primer CO602 (5’ GCC CAT AGT GCT TCC TGC TGC TCC CAA GAA CC 3’; 7786-7817).  Second-round PCR was performed using ExpandTM High Fidelity Enzyme  63 (Roche Diagnostics) with forward primer SQV3F2 (5’ TGT GCC CCA GCT GGT TTT GCG AT 3’; HXB2 6876-6898) and reverse primer CD4R (5 TAT AAT TCA CTT CTC CAA TTG TCC 3’; HXB2 7652-7675).  Second-round PCR amplicons were subsequently sequenced in both the 5’ and 3’ directions using the ABI 3730 automated sequencer (Applied Biosystems) using BigDye Terminator (Applied Biosystems) with forward primer V3O2F (5’ AAT GTC AGY ACA GTA CAA TGT ACA C 3’; HXB2 6945-6969) and reverse primer SQV3R1 (5’ GAA AAA TTC CCT TCC ACA ATT AAA 3’; HXB2 7350-7373). 2.2.3 Sequence Interpretation Population sequencing was used to generate a consensus sequence for each sample.  Each consensus sequence was analyzed using fully automated custom software (‘ReCall’).390,391 All potential permutations at individual base positions within the V3 loop were identified and sequences were then interpreted using the g2p bioinformatics algorithm (false-positive rate (FPR) of 5%) to predict viral tropism.  Without prior knowledge of clinical outcome, a g2p score was generated and tropism assigned using an arbitrarily chosen cutoff point of five.  In this case, a g2p score above five identified an R5-only viral population, whereas a g2p score below or equal to five identified a non-R5 viral population. For this analysis, tropism was determined using the consensus sequence from one of three PCR amplification products.  The g2p algorithm allows all possible sequence permutations within a patient sample to be analyzed and scored.  Among the triplicates, the sequence with the lowest g2p score was used to stratify patients as either R5 or non-R5 for clinical outcome analysis.  64 2.2.4 Outcome Measures & Predictor Variables The primary endpoint was described as the median change in plasma viral load after eight weeks for analyses using a continuous variable.  For responder analyses, a decrease in viral load of at least 2.0 log10 copies/mL or viral suppression of less than 50 copies/mL at week 8 defined therapeutic, or virological success.  Analysis at week 8 was performed using last observation carried forward (LOCF) in the event of missing data.  An additional endpoint, the percentage of patients with a pVL less than 50 copies/mL (percentage undetectable) at week 24, was assessed.  In the instance of incomplete data at week 24, a patient was considered to have failed treatment (‘missing equals failure’; M=F).  The change in median plasma viral load was compared between patients screened as having R5 and non-R5 virus, as well as between phenotypic and genotypic tropism screening methods. In order to predict the activity of the optimized background therapy and better assess the role of maraviroc as an antiretroviral agent, each optimized background therapy was given a weighted sensitivity score (wSS) based on phenotypic susceptibility and drug class potency.392 A weighted sensitivity score was calculated for each optimized background therapy initiated at day one, in accordance with regimen characteristics and grouped into the following three subsets: <1, 1≤wSS<2, and 2≤wSS.  If phenotypic drug resistance was exhibited, and or the regimen contained ‘inactive’ agents, a score of zero was given.  Nucleos(t)ide analogues (NRTIs) were given a score of 0.5 and both non-nucleoside analogues (NNRTIs) and protease inhibitors were given a score of one.  A treatment regimen with a weighted sensitivity score of less than one was composed of maraviroc only or maraviroc and one NRTI.  Conversely, a weighted sensitivity score of at least two was  65 given to a background therapy containing at least two NNRTIs or protease inhibitors – a regimen that in combination with maraviroc enters the realm of highly active antiretroviral therapy (HAART).  The median change in viral load for each weighted sensitivity score category and each screening method was then graphed to identify any discordance in optimized background therapy activity between the three study arms. A direct comparison of treatment response in patients with R5 and non-R5 virus was made between Trofile and genotype.  Concordance and discordance of each assay to predict therapeutic response were determined using the median change in viral load (log10 copies/ml) of R5 and non-R5 virus. In order to more accurately represent the original distribution of patient with R5 and non-R5 virus in the Trofile screening populations of MOTIVATE-1, 2, and A4001029 (approximately 29% non-R5 tropic), the entire population of individuals screened as having R5 virus, as determined genotypically, was randomly divided into four subsets.  Each subset, representing approximately 25% of the total R5-screening population, was then combined with all non-R5 tropic samples.  This sub-analysis was examined using the tropism data as determined by both phenotypic and genotypic screening methods, making the proportion of non-R5 individuals in the subset population approximately 40%. 2.2.5 Cutoff Point Optimization A follow up study to optimize potential clinically relevant g2p FPR cutoff points was conducted using the A4001029, MOTIVATE-1 and -2 sample populations.  Cutoff points to be evaluated for clinical utility were chosen based on the maraviroc-associated  66 pVL decrease, or complete viral suppression, observed during the course of the studies.  A training data set was established using 75% of the study data chosen at random from the maraviroc study arms.  The remaining 25% of the study data was subsequently used for cutoff point validation.  A second training set using the placebo data was also established.  Cutoff points were subsequently verified using bootstrapping with replacement, where 75% of the maraviroc data set was sampled at random, and repeated 1 000 times.393 The primary outcome for this analysis was the week 8 change in pVL in the combined maraviroc study arms, factoring baseline pVL, baseline CD4 count and weighted sensitivity score.  In addition, the percentage of patients with undetectable viral load, time to change tropism and time to study discontinuation were analyzed for each cutoff point.  The endpoint results as predicted by g2p were compared to those of Trofile, where X4 and dual/mixed results were collapsed into a single “X4” category.  If a patient did not complete the 48 weeks of maraviroc treatment, last observation carried forward (LOCF) or missing equals failure (M=F) were applied to fill in data gaps for both the g2p and Trofile results.393 2.3 Results Baseline characteristics in the overall screening population as well as within the MOTIVATE-1 and MOTIVATE-2 and A4001029 study populations were found to be comparable.  Data from the MOTIVATE-1, MOTIVATE-2, and A4001029 studies were combined and early virological responses to maraviroc along with optimized background therapy analyzed with regard to different tropism screening methods.  The Trofile assay  67 and V3 genotype interpreted by the g2p algorithm (5% FPR) were broadly comparable predictors of response to therapy in each of the three treatment arms of the studies (Figure 2.1).  After eight weeks of treatment, patients in the maraviroc twice-daily arm classified as having R5 virus by either Trofile (N=405) or V3 genotyping (N=394) experienced a median decrease in viral load of 2.4 log10 copies/mL, significantly greater than that of patients whose virus was classified non-R5 by the same assays (Figure 2.1a).  The maraviroc once-daily arm revealed a similar discrepancy in median response to therapy between patients with R5 and non-R5 virus (Figure 2.1b).  As expected, there was no difference in the placebo arm (Figure 2.1c). The decrease in viral load was similar regardless of the screening tropism assay method.  Similar proportions of patients achieved virological success at week 8 regardless of the assay used.  For example, of patients whose virus was classified R5 by Trofile in the maraviroc twice-daily arm, 272 patients (67%) achieved virological suppression at week 8, where patients whose virus was classified R5 by V3 genotype, 263 (67%) achieved virological suppression at week 8. (Table 2.1) Responses were consistent across screening methodologies whether the definition of virological success required a reduction of 1, 1.5, 2.5, or 3 log10 copies/mL (data not shown).  In the placebo arms, Trofile and g2p predictions paralleled one another with a median initial change of -0.75 log10 copies/mL followed by a small increase in viral load.  Similarly, screening tropism results were not strongly associated with virological response in the placebo arms.  68  Figure 2.1.  Virological response to maraviroc stratified by tropism test. Virological response in maraviroc treatment-experienced trials stratified by screening Trofile and V3 genotype results.  Combined results of patients who entered the MOTIVATE-1, MOTIVATE-2 and A4001029 studies are indicated according to screening results obtained using the original Monogram Trofile Assay (light green) and a V3 genotype assay interpreted using the g2p algorithm with a 5% FPR (dark green).  Solid lines indicate median viral load changes for patients with R5 results, whereas dashed lines indicate non-R5 results.  The three treatment arms were a) maraviroc b.i.d., b) maraviroc q.d., and c) Placebo PBO.  69 Table 2.1.  Percentage of patients with virological response to maraviroc at week eight. Virological response was defined at week 8 as achieving a viral load decrease of >2 logs or below 50 copies/mL in the twice-daily and once-daily arms.   Virological suppression was similar between the Trofile and V3 genotype screening methods regardless of the antiviral activity of the background therapy used (Figure 2.2).  Response to therapy in the three study groups was accounted for by using a weighted sensitivity score algorithm in patients receiving maraviroc.  Patients receiving a weighted sensitivity score of less than two showed similar incremental benefit of a median viral load decrease of 1 to 2 log10 copies/mL if determined by either assay to have R5 virus, with a maximum median viral load decrease of 2.5 log10 copies/mL over the first eight weeks of therapy (Figure 2.2a and 2.2b).  In contrast, a weighted sensitivity score of at least two was sufficient to confer median viral load decreases more than 2.5 log10 copies/mL regardless of screening tropism or testing methodology (Figure 2.2c).  70  Figure 2.2.  Virological response to maraviroc stratified by optimized background treatment. Virological response of maraviroc-treated individuals as predicted by screening Trofile and V3 genotype values, stratified by activity of optimized background treatment.  Combined results of patients who entered the MOTIVATE-1, MOTIVATE-2 and A4001029 studies who received maraviroc (either b.i.d. or q.d.) are indicated according to tropism screening results obtained using the Trofile (light green) and V3 genotype assays interpreted using the g2p algorithm with a 5% FPR (dark green).  Solid lines indicate median viral load changes for patients with R5 results, whereas dashed lines indicate non-R5 results.  The activity of the background regimen was assessed using a weighted sensitivity score (wSS), which reflects the sensitivity and inherent potency of the background regimen.  Scores correspond to a) wSS <1, b) wSS 1-2, or c) wSS ≥2, indicating progressively increasing activity of the background regimen. Median viral load decreases of patients screened as R5 or non-R5 by the Trofile assay and or V3 genotyping were examined as a function of tropism result concordance between assays (Figure 2.3).  Collectively, 90% concordance was observed at week 8 between Trofile and V3 genotyping results in patients receiving maraviroc, either twice daily or once daily.  Of these, a total of 737 (81%) of patients’ viruses were R5 tropic by both  71 Trofile and V3 genotype; these patients exhibited the greatest response to maraviroc treatment, with a median viral load decrease of 2.5 log10 copies/mL at week 8.  Conversely, 80 (9%) patients with non-R5 virus by both screening methods had median decreases in viral load of only 1.0 log10 copies/mL by week 8 (Figure 2.3a).  Among patients with discordant tropism screening results, 31 patients (3.5%) were identified as having non-R5 virus by Trofile but identified as having R5 virus by g2p, whereas 58 patients (6.5%) were identified as having R5 virus by Trofile but identified as having non-R5 virus by g2p with a power of 80%.  After eight weeks of treatment, patients receiving maraviroc in each of the discordant groups showed similar median viral load changes of -2.0 log10 copies/mL, regardless of the direction of discordance (Figure 2.3b).  Power estimations show that in order to detect a difference of 0.3 log10 copies/mL in median plasma viral load change with 80% power if this difference existed, assuming a standard deviation of 0.3 in viral load, 17 patients per group would be needed.  72  Figure 2.3.  Virological response to maraviroc demonstrating concordance between Trofile and V3 genotype tropism assays. Combined results of patients entered in the MOTIVATE-1, MOTIVATE-2 and A4001029 studies who received maraviroc (either b.i.d. or q.d.).  Patients were stratified according to concordance or discordance of tropism screening results obtained using the Trofile and V3 genotype (g2p FPR 5%).  Panel A) median viral load change where both assays indicated R5 virus (solid green; N=737) and non-R5 virus (dashed red; N=80).  Panel B) median viral load change where results were R5 by Trofile but non-R5 by genotype (solid orange; N=58), and non-R5 by Trofile but R5 by genotype (dashed blue; N=31). The initial screening process of MOTIVATE-1 and MOTIVATE-2 studies removed individuals with non-R5 using virus from the treatment populations; of those removed only a fraction (186/955) were enrolled into the A4001029 study270 Subsets were formulated in order to more accurately represent the initial screening population in order to explore the possible contribution of selection bias.  In the four constructed subsets, each containing the entire A4001029 study population as well as a random 25% of the MOTIVATE-1 and MOTIVATE-2 study participants all of whom received maraviroc (either once-daily or twice-daily), the Trofile and V3 genotype gave similar outcome predictions indicating the superiority of maraviroc in patients classified as having R5 virus over those classified as having non-R5 virus (Figure 2.4).  Patients whose virus screened R5 had an early median viral load decrease of 2.5 log10 copies/mL at week 8, whereas those with non-R5 virus exhibited a median change in viral load of -1.5 log10 copies/mL.  73 When week 8 virological suppression was defined as the gold standard, the Trofile assay exhibited a sensitivity of 92%, and a specificity of 20% in the twice-daily arm.  Similarly, the V3 genotypic method had a sensitivity of 89% with specificity of 24%. (Table 2.1)  In a final sensitivity analysis, the performance of the assays was compared after 24 weeks of therapy, where 46% of patients screened R5 by both Trofile and V3 genotyping had a sustained virological response (<50 copies/mL) on treatment with MVC (N=737).  If one considers the week 24 virological suppression results as the relevant benchmark for clinical utility, the Trofile assay exhibited a sensitivity of 92% and specificity of 11%, similar to the V3 genotype method, which had a sensitivity of 89% and a specificity of 14%. In order to identify clinically relevant g2p FPR cutoff points, the frequency distribution of g2p FPR values observed in the MOTIVATE and A4001029 studies was first plotted.  Based on the associated therapeutic response to maraviroc, g2p FPR scores were divided into three categories: likely to respond poorly, likely to respond moderately, and likely to respond well.  Using this distribution g2p FPR values defining the lower and upper bounds of the moderate response group were identified, 2 and 5.75, respectively (Figure 2.5).  This range of g2p FPR values was associated with partial antiviral activity of maraviroc.  74  Figure 2.4.  Virological response to maraviroc in a subset of patients screened as having non-R5 virus and randomly selected patients screened as having R5 virus. Virological response in maraviroc treatment-experienced trials stratified by screening Trofile and V3 genotype results, sampling random 25% subsets of patients screened as having R5 virus.  Combined median viral load changes from baseline in patients who entered the MOTIVATE and A4001029 studies who received maraviroc (either b.i.d. or q.d.) are indicated for four random subsets of patients screened as R5 by the Trofile assay and all patients screened as non-R5 by Trofile, in order to approximate the initial screening population, rather than the treated population.  Median viral load changes from baseline obtained using the Trofile assay (light green) and a V3 genotype assay interpreted using the g2p algorithm with a 5% FPR (dark green).  Solid lines indicate median viral load changes for patients with R5 results, whereas dashed lines indicate non-R5 results.  75  Figure 2.5.  Frequency distribution of patient samples as determined by the assigned g2p score in accordance with therapeutic response categories. The frequency distribution of g2p false positive rate (FPR) values observed in the genotypic tropism rescreening of the MOTIVATE and A4001029 studies.  G2p FPR values were divided into therapeutic response categories as observed in the MOTIVATE and A4001029 studies, such that red indicates patients likely to respond poorly to maraviroc, yellow predicts compromised maraviroc efficacy, and green indicates patients likely to respond well to maraviroc.  Note two scales along the x-axis, g2p FPR values <10 are divided by increments of 0.5 units and g2p FPR values >10 are divided by increments of 10 units. Following the identification of cutoff points defining the likelihood of poor and moderate therapeutic response to maraviroc, their potential use as clinically relevant cutoff points was evaluated using the MOTIVATE and A4001029 study data.  Five analyses were conducted, each with a different sample set, 1) training data set (75% of the data from the maraviroc study arms); 2) placebo training set (placebo study arms); 3) study validation set (25% of data from the maraviroc study arms); 4) bootstrapping (75% of the entire maraviroc data set sampled at random 1 000 times with replacement); 5) all maraviroc data (entire maraviroc data set).  Data for each sample set were stratified into three bins based on the optimized g2p FPR values (2 < FPR, 2 ≤ FPR ≤ 5.75, FPR > 5.75).  Results were consistent  76 regardless of end point being analyzed, change in pVL or percent undetectable.  In all maraviroc sample sets, a g2p FPR < 2 predicted a poor response to maraviroc, whereas a g2p FPR greater than 5.75 predicted virological success following maraviroc exposure.  G2p FPR values between 2 and 5.75 inferred a slightly diminished response to maraviroc.  When compared to the virological response as predicted by the original Trofile assay, the g2p FPR cutoff point of 2 appeared to better predict virological failure, whereas the g2p FPR cutoff point of 5.75 performed broadly similar to the Trofile assay in predicting virological success (Figure 2.6). To further validate the use of the optimized g2p FPR cutoff points, survival analyses were performed using time to change tropism and time to study discontinuation as endpoints.  The optimized cutoff points were able to predict the probability of tropism change from R5 to non-R5 following the start of maraviroc and whether a patient would discontinue participation in the maraviroc studies.  A g2p FPR < 2 clearly predicted a change in tropism from R5 to non-R5 in this prescreened R5 study population, suggesting the maraviroc-selected outgrowth of preexisting but unidentified non-R5 virus.  G2p FPR values > 5.75 were associated with little probability of tropism change.  When analyzing the g2p FPR predictions of study discontinuation, the distinction between outcomes based on g2p FPR bin was less noteworthy when compared to the other analyses.  This may be an artifact of non-maraviroc-associated factors causing discontinuation.  The optimized cutoff points performed similarly to the original Trofile assay in both analyses (Figure 2.7).  77  Figure 2.6.  Population sequencing with optimized g2p cutoff points can predict the virological response to maraviroc. Using viral load change and percent of patients suppressed below 50 HIV RNA copies/mL as endpoints, the optimized cutoff points of 2 and 5.75 were demonstrated to predict response to maraviroc.  Outcomes for the five sample sets: training data set (75%), the placebo arm, validation set (25%), bootstrapping and all samples from the maraviroc dataset are shown above.  Coloured lines represent the three g2p FPR bins (2 < FPR, 2 ≤ FPR ≤ 5.75, FPR > 5.75); black lines represent R5 (solid) or non-R5 (dotted) samples as determined by the original Monogram Trofile assay.  A g2p FPR value < 2 inferred a poor response to maraviroc, a score of > 5.75 inferred therapeutic success.  G2p FPR value 5.75 indicated the point at which maraviroc efficacy starts to become compromised.  Note that more patients were identified as having g2p FPR ≤ 5.75 than identified by the original Trofile assay.  78  Figure 2.7.  Population sequencing with optimized g2p cutoff points can effectively predict a) the probability of a tropism change and b) time to discontinuation in the study. The optimized g2p FPR cutoff points of 2 and 5.75 were demonstrated to predict response to maraviroc using a) the time in which tropism may shift from R5 to non-R5 following initiation of maraviroc and b) the time required to discontinue participation in the study as endpoints.  Outcomes for three maraviroc sample sets: training data set (75%), validation set (25%), and all samples are shown above.  Coloured lines represent the three g2p FPR bins (2 < FPR, 2 ≤ FPR ≤ 5.75, FPR > 5.75); black lines represent R5 (solid) or non-R5 (dotted) samples as determined by the original Monogram Trofile assay.  Those with g2p FPR < 2 were more likely to switch tropism over the course of the study than those with g2p > 5.75 and less likely to remain in the study for the 48 week duration following the initiation of maraviroc.  79 2.4 Discussion Sequence determination of the HIV envelope V3 loop can be used to predict HIV co-receptor tropism, but discordance with results obtained by the original Trofile assay can be observed.  To date, this has limited the application of genotyping for clinical screening of individuals who could potentially benefit from maraviroc-containing therapy.  This study represents the largest attempt to demonstrate the potential clinical utility of genotypic HIV tropism determination based on virological response to a CCR5 antagonist rather than concordance with the phenotypic Trofile assay.  Here, employing the same patient population used to demonstrate the efficacy of maraviroc for its subsequent regulatory approval, we have found that determination of the HIV envelope V3 sequence provides broadly similar early virological outcomes compared with the original Trofile assay.  These results were consistent across a variety of subset analyses and responses were maintained over 24 weeks of observation. The population studied here offers several unique advantages for the evaluation of tropism methodologies.  The combination of several randomized, multinational registrational trials allows the investigation of large numbers of very well characterized patients.  In particular, the inclusion of patients screened as having non-R5 virus from the parallel A4001029 safety study, which had the same design as the MOTIVATE studies.  An additional critical factor was that the patient population was extensively pretreated, with demonstrated resistance to three drug classes, and the trial design excluded the use of the then-experimental agents raltegravir, darunavir, and etravirine.  As a result, large numbers of individuals had virological success and large numbers failed to respond to therapy, such that if important differences in assay performance existed they could be detected.  80 There is published data comparing Trofile and V3 genotype-based tropism screening assays.  In 2007, Low et al. analyzed the ability of V3 genotyping with various genotypic algorithms, including g2p, to identify X4 variants as determined by Trofile, genotyping was found to exhibit lower sensitivity when compared with Trofile.  At this time, owing to low observed sensitivity, it was found that genotypic algorithms were not suitable for predicting HIV X4 variants in clinical samples378 In 2009, however, Poveda et al. compared V3 genotyping to Trofile using various algorithms including g2p and PSSMX4R5/SINSI.  Overall, at least 72% concordance was found between Trofile and genotypic methods with a range of sensitivity from 31–76%.  Adjustments to the X4 cutoff points of PSSM algorithms led to an increase in sensitivity to 80%.  The results of this study led to the conclusion that the use of the genotypic algorithms may prove to be a useful tool in determining the appropriate use of CCR5 antagonists in antiretroviral regimens.389 Adjustments to algorithm matrices and cutoff points can be made to optimize the ability of V3 genotyping to identify R5 and non-R5 viral populations for use in the prediction of CCR5 antagonist clinical outcomes.  And, as illustrated here, cutoff point optimization can be used to define various levels of maraviroc activity as has been done for other antiretroviral compounds.  The study presented in this chapter is based on the ability of V3 genotyping to predict viral tropism in comparison to Trofile where results were generated in reference to clinical outcome/virological success. A limitation of the approach of using population-based V3 genotyping is that the reliable detection of minority species within a viral population may be difficult.  The detection limit for population-based sequencing is approximately 15–30%, though this is highly dependent on the input copy number of the virus, a varying factor which limits all  81 tropism assays.  In an attempt to partially overcome this limitation, we performed RT-PCR in triplicate to increase the likelihood of identifying minority species.  Swenson et al. performed 454 “deep” sequencing of the same MOTIVATE and A4001029 study populations.394 This method has a lower detection limit than that for population-based sequencing.  When compared with the results presented here, “deep” sequencing showed an increased ability to detect low abundance minority non-R5 species, though only a modest improvement in terms of predicting virological response following the start of maraviroc was demonstrated.  Our analysis did not allow comparison with the current ESTA, which has a claimed sensitivity of 0.3% of a viral population compared with 10% for the original assay based on in-vitro mixing experiments.  This is because the ESTA has never been tested on the patients in the MOTIVATE or A4001029 studies.  However, the lack of comparison between these two assays may not be a significant failing, as the specificity of this assay is unknown.  In the ACTG5211 study of treatment-experienced individuals receiving a vicriviroc-containing regimen, the use of the ESTA was associated with only marginal improvement in prediction of virological outcomes at 24 weeks compared with the original Trofile assay.395 Furthermore, if increased sensitivity to detect X4 virus is desired, the g2p FPR (or other genotypic algorithms, such as PSSM) cutoff point used to assign a tropism to V3 sequences can simply be adjusted to increase sensitivity.  Indeed, Lengauer demonstrated that, as ESTA is a more sensitive adaptation of the original Trofile, changing the genotypic algorithm cutoff points allows similar agreement of V3 genotyping with the ESTA as with the original Trofile.396 Additional research is currently underway to determine the optimum cutoff points for clinical application of the different genotypic algorithms, but cutoff points will be optimized to improve associations with virological outcomes, rather than the ESTA assay results.  Future comparisons of both the  82 sensitivity and specificity of V3 genotyping and the ESTA for the prediction of virological response to CCR5 antagonists are clearly warranted. Genotypic assays have considerable practical advantages, including more rapid turnaround time, broader availability and lower cost.  In direct comparisons neither approach has been shown to be consistently better than the other.  It is likely that both approaches have their place.  Finally, the results presented here support further analysis of V3 loop sequencing using newer technologies such as “deep” next generation sequencing, which is currently being explored in studies of HIV drug resistance and tropism. This large study showed that V3 genotyping and the original Trofile assay were comparable in predicting early antiviral responses to maraviroc in treatment-experienced patients.  Furthermore, this observation was consistent across a variety of sub-analyses and maintained out to 24 weeks.  The data suggest that analysis of HIV envelope V3 genotype is an attractive option for predicting co-receptor tropism of HIV in patient plasma and, perhaps more importantly, for identification of treatment-experienced patients who could benefit from CCR5 antagonists such as maraviroc.  83 Chapter Three: Population-based Sequencing of the V3 Loop Can Predict the Virological Response to Maraviroc in Treatment-Naïve Patients of the MERIT Trial 3.1 Introduction CCR5 antagonists are a relatively new class of antiretroviral drugs for the treatment of HIV-1.  Currently, maraviroc (MVC) (ViiV Healthcare) is the only clinically approved CCR5 antagonist.397 HIV-1 cellular entry is dependent on interactions of two host cell receptors, the CD4 receptor in conjunction with one of two chemokine co-receptors, CCR5 or CXCR4.  Viral populations may consist of CCR5- using virus, “R5”; CXCR4-using virus, “X4”; virus able to use both CCR5 and CXCR4 or a mixture of both CCR5-using and CXCR4-using virus, “dual/mixed”.76,387 R5 viruses are responsible for the majority of new infections, whereas the presence of X4 viruses, usually in mixture with R5 virus, is generally associated with late-stage disease and poor prognosis.398,399 As the name suggests, CCR5 antagonists target CCR5-using virus exclusively by binding to the CCR5 co-receptor thus inhibiting the CCR5-gp120 interaction.249,400 As such, it is highly recommended that a tropism screening assay be performed before the initiation of a CCR5 antagonist as CCR5 antagonists are ineffective against X4 or dual/mixed variants.  The original Monogram Biosciences Trofile Assay and Enhanced Sensitivity Trofile Assay (ESTA) have been the most widely used phenotypic tropism assays.267,286,386  84 The MOTIVATE trials demonstrated that maraviroc, when used with an optimized background therapy, can effectively reduce plasma viral load in treatment-experienced patients infected with R5 virus.266,268 The MERIT trial was a multicenter, randomized, double-blind comparative trial in drug-naive individuals consisting of three arms: maraviroc once-daily, maraviroc twice-daily and efavirenz (EFV), all in combination with Combivir (zidovudine (ZDV)/lamivudine (3TC)).271 Only patients with R5 virus determined by the original phenotypic Trofile assay were enrolled in the trial.  At week 16, the Data Safety Monitoring Board discontinued the maraviroc once-daily study arm due to failure to meet pre-specified non-inferiority criteria of maraviroc once-daily versus efavirenz.271 In the primary 48-week analysis, maraviroc twice-daily was non-inferior to the efavirenz arm at the <400 copies per milliliter endpoint but not at the <50 copies per milliliter endpoint (65% vs. 69%) using a threshold of -10%.  However, when retrospectively rescreened using the ESTA assay, approximately, 15% of patients were shown to have non-R5 virus, and the non-inferiority margin was met for the co-primary endpoint of <50 copies per milliliter, suggesting the enrollment of patients with non-R5 virus in MERIT who were unlikely to respond to maraviroc.401 Genotypic methods have also been developed to determine tropism to reduce the costs and practical limitations of phenotypic tropism testing.  Numerous genetic determinants of tropism have been identified primarily within the V3 loop of the gp120 protein encoded by the HIV-1 env gene.373,374 From RNA extracted from plasma samples, the V3 loop can be sequenced using population-based or “deep” sequencing technologies. Resultant sequences can then be interpreted using a bioinformatics algorithm, such as geno2pheno[co-receptor] (g2p) or the position-specific scoring matrix, from which tropism may  85 be inferred.288,402 Previously, our group retrospectively analyzed the virological response to maraviroc in treatment-experienced patients of the MOTIVATE trial, where patient samples were rescreened using a population-based sequencing method.290,403 Here we retrospectively analyzed maraviroc treatment response in treatment-naive individuals from the MERIT trial where tropism was predicted using a population-based V3 genotypic tropism assay. 3.2 Methods 3.2.1 Study Population Treatment-naive patients enrolled in the MERIT trial receiving at least one dose of study medication, with screening tropism results available using the original Trofile, ESTA and population-based sequencing assays (N=876) were included in this analysis.  Patient viral loads were ≥2000 copies per milliliter for purposes of successfully performing the Trofile assay.  Of the original screening population, only those using the CCR5 co-receptor exclusively, as determined by the original Trofile assay, were included in the MERIT trial.271,401 Patients were enrolled in a ratio of 1:1:1 to the three treatment arms: efavirenz (600 mg taken once-daily), maraviroc once-daily (300 mg q.d.), or maraviroc twice-daily (300 mg b.i.d.).  All patients were also assigned a fixed-dose background therapy of Combivir (zidovudine (ZDV)/lamivudine (3TC)) taken twice daily.  Enrollment in MERIT required patients to harbor no genotypic resistance mutations to the trial drugs, including efavirenz, lamivudine, and zidovudine.271,401  86 3.2.2 Sequencing The nucleic acid from stored frozen plasma samples taken during the screening visit of patients enrolled in the MERIT trial was extracted using the standard procedures for the NucliSENS easyMAG (bioMérieux, St Laurent, Canada).  Staff was blinded to virological outcome data and ESTA results.  The V3 loop of gp120 from the HIV-1 env gene was amplified in triplicate from the RNA extracts using nested reverse transcriptase–polymerase chain reaction (RT-PCR).  Reverse transcription and first-round polymerase chain reactions (PCRs) were combined into a single step with the Super- Script III One-Step RT-PCR system with Platinum Taq High Fidelity enzyme (Invitrogen, Burlington, Canada).  One-step RT-PCR was performed using forward primer SQV3F1 (5’ GAG CCA ATT CCC ATA CAT TAT TGT 3’; HXB2 6855-6878) and reverse primer CO602 (5’ GCC CAT AGT GCT TCC TGC TGC TCC CAA GAA CC 3’; HXB2 7786-7817).  Expand high-fidelity enzyme (Roche, Laval, Canada) with forward primer SQV3F2 (5’ TGT GCC CCA GCT GGT TTT GCG AT 3’; HXB2 6876-6898) and reverse primer CD4R (5’ TAT AAT TCA CTT CTC CAA TTG TCC 3’; HXB2 7652-7675) were used in the second-round PCR reaction.  The resultant PCR amplicons were sequenced in both the 5’ and 3’ directions using the ABI 3730 automated sequencer (Applied Biosystems, Burlington, Canada) using BigDye Terminator (Applied Biosystems).  The sequencing reactions employed forward primer V3O2F (5’ AAT GTC AGY ACA GTA CAA TGT ACA C 3’; HXB2 6945-6969) and reverse primer SQV3R1 (5’ GAA AAA TTC CCT TCC ACA ATT AAA 3’; HXB2 7350-7373).  87 3.2.3 Sequence Interpretation Population-based sequencing methods were used to generate a consensus sequence of the V3 loop (~105bp) for each replicate.  Fully automated base calling was performed using custom software “RECall”.390,391 Mixtures were determined if the area under the secondary peak exceeded 12.5% of the area under the primary peak.  Sequences were processed with the clonal geno2pheno (g2p) algorithm with no additional parameters added.  The previously optimized false-positive rate (FPR) cutoff of 5.75% was used, where samples below this FPR value were inferred as being non-R5.  This clinical cutoff was based on analyses of the combined MOTIVATE/A4001029 trials.393 The synthesis of viral cDNA, amplification and sequencing were performed in triplicate for each sample, and the sequence with the lowest g2p FPR value was used to assign the viral tropism.  Of the 885 samples attempted, we were unable to sequence nine (1%); four patients included in the analyses were amplified successfully in duplicate, of which three were R5 and one was non-R5.  To investigate the effect of different cutoff points, sequences were also processed using three additional FPR cutoff points, 3.5%, 10%, and 20%, and the effect of triplicate analyses were explored (see Appendix I). 3.2.4 Outcome Measures & Predictor Variables The predefined study endpoints included the percentage of the population able to suppress their viral load below detectable levels at week 48 (<50 copies/mL, where missing data were considered failure (M=F)).  A secondary analysis investigated the median change in plasma viral load from baseline at all time points.  In the event of missing plasma viral load data, the last observation was carried forward (LOCF).  A total of 244 patients were  88 missing plasma viral load data at week 48.  A survival analysis was performed based upon the time elapsed between baseline and the first sample with non-R5 virus determined using the original Trofile assay.  A comparative analysis of the rate of concordance and discordance in tropism calls between the V3 genotypic and ESTA methods was also performed.  The treatment response, change in plasma viral load and percent of patients with viral load <50 copies per milliliter associated with each was analyzed to week 48. The maraviroc once-daily arm served as a second independent investigation of the V3 genotype assay used here.  This trial arm was discontinued in January 2006 due to failure to achieve predetermined non-inferiority criteria of maraviroc once-daily.271 At this point, the once-daily arm was unblinded and patients were given the option to switch to the maraviroc twice- daily arm.  In this study, a censored analysis consisting of only those patients who remained on maraviroc once-daily for the duration of the study was conducted as well as an uncensored analysis consisting of all recipients of maraviroc once-daily, including those who went on to receive maraviroc twice-daily.  Note that not all maraviroc once-daily samples were rescreened using ESTA therefore, these comparisons could not be made for this arm. 3.3 Results Upon rescreening the MERIT efavirenz and maraviroc twice-daily populations (N=705) using a population-based V3 genotype assay, 58 patients (8%) were identified as having non-R5 virus based on their g2p score despite having been screened as having R5 virus by the original Trofile assay.  Baseline characteristics were broadly comparable  89 between the R5 and non-R5 groups and treatment arms (Table 3.1).  R5 and non-R5 genotype categories were used in the analyses, including the percentage of the population able to suppress their viral load to undetectable levels at 48 weeks (<50 HIV RNA copies/mL) after the start of treatment (Table 3.2). Patients screened as having R5 virus by V3 genotype had similar responses at all weeks in the maraviroc twice-daily arm and efavirenz arm (Figure 3.1A), whereas those screened with non-R5 virus in the maraviroc twice-daily arm did not (Figure 3.1B).  We performed the same analysis at three additional commonly used g2p FPR cutoff points (3.5%, 10%, and 20%) which demonstrated an improved response to maraviroc twice-daily in patients screened as having non-R5 virus when the g2p FPR cutoff is increased.  This suggests decreased sensitivity in detecting X4 virus as the g2p FPR cutoff increases (see Supplemental Figures 3.1a and 3.1b in Appendix I).  At the predefined primary endpoint of 48 weeks, maraviroc twice-daily was non-inferior when compared with that of efavirenz in the same group of individuals (67% vs. 68% with plasma viral load <50 RNA copies/mL) (Figure 3.1A; Table 3.2).   90 Table 3.1.  Baseline characteristics for participants of the MERIT trial. Baseline characteristics for patient enrolled in the maraviroc b.i.d. and efavirenz q.d. arms of the MERIT study and rescreened for tropism using both V3 genotyping and the ESTA.  Based on the re-screening results, patients were stratified into R5 (g2p>5.75) and non-R5 (g2p≤5.75) categories.   91 Table 3.2.  Results of rescreening the patient population of the MERIT trial for tropism. Patients enrolled in either the maraviroc twice-daily or the efavirenz once-daily treatment arms of the MERIT study were initially screened as having R5 virus by the original Trofile assay and rescreened for tropism using the ESTA and V3 genotyping (g2p FPR >5.75, R5).  For each screening assay, the number of patients achieving treatment success, defined as viral load of <50 copies/mL at week 48 after maraviroc or efavirenz treatment, are shown.  Non-inferiority of maraviroc to efavirenz was predefined at -10%, with a 97.5% LCB stratified on geographic region (hemisphere) and screening plasma viral load.  The absolute difference represents the percentage of patients achieving success in the efavirenz once-daily arm from the maraviroc twice-daily arm. * LCB – lower confidence bound Median plasma viral load decline in the maraviroc arms was roughly 1 log10 copies/mL less in the population identified as having non-R5 virus by genotype when compared with the R5 population.  For example, when screened as having R5 virus by the V3 genotypic assay, patients assigned maraviroc twice-daily had a reduction in their plasma HIV RNA viral load of approximately 3 log10 copies/mL after 16 weeks on treatment similar to the efavirenz arm (Figure 3.1C).  In comparison, a viral load reduction of just over 2 log10 copies/mL was observed after 16 weeks on maraviroc for patients screened as having non-R5 virus by V3 genotype (Figure 3.1D).  The median viral load in the non-R5 group on efavirenz decreased roughly 3 log10 copies/mL (Figure 3.1D).  The analysis was repeated using three additional g2p FPR cutoff points (see Supplemental Figures 3.2a and 3.2b in Appendix I).  92  Figure 3.1.  Virological response to maraviroc in MERIT trial participants rescreened for tropism using a V3 genotypic assay. All patients were originally screened as having R5 virus by the Original Trofile Assay and retrospectively screened by V3 genotype using g2p with a 5.75 FPR cutoff point.  Patients were re-stratified as having R5 or non-R5 virus by genotype and the response to therapy evaluated.  At week 48, primary endpoint, maraviroc twice-daily was non-inferior to efavirenz in those patients screened as having R5 virus by V3 genotype.  A) patients screened as having R5 virus comparing virological response to maraviroc twice-daily (N=323) and efavirenz (N=324).  B) patients screened as having non-R5 virus comparing virological response to maraviroc.  93 To further investigate the ability of V3 genotype to determine tropism and predict clinical outcome, patients initially assigned to the maraviroc once-daily arm were also investigated regardless of whether or not they subsequently switched to a maraviroc twice-daily regimen.  Patients assigned to the maraviroc once-daily arm had greater and more sustained suppression of viremia if rescreened as having R5 virus (Figure 3.1E) than if rescreened as having non-R5 virus (Figure 3.1F).  Similar patterns were observed in a censored analysis, which included only those patients who remained on the once-daily regimen, though the number of patients becomes very small beyond week 16 (data not shown).  As with MVC twice-daily, the analysis was repeated using three additional g2p FPR cutoff points (see Supplemental Figures 3.3a and 3.3b in Appendix I). 3.3.1 Subsequent Changes in Trofile Assay Results A longitudinal analysis was performed using the time to a change of tropism from R5 to non-R5 as determined by the original Trofile assay to investigate whether changes in on-therapy tropism were predicted by screening V3 genotype.  Patients screened as having non-R5 virus by genotype had a greater likelihood of changing tropism than patients with R5 virus whether initially assigned to the maraviroc twice-daily (Figure 3.2A) or maraviroc once- daily (Figure 3.2B) arms.  After two weeks on maraviroc twice-daily, 10% of patients screened as having non-R5 virus by genotype had shifted to non-R5 by Trofile.  By week 48, 35% of patients screened as having non-R5 virus by genotype had shifted to non-R5 by Trofile, compared with the shift occurring in only 10% of subjects screened as R5 by V3 genotype.  To further investigate the use of different g2p FPR cutoff points, we analyzed the data in terms of the time to change tropism from R5 to non-R5 at 3.5%, 10%, and 20%,  94 again showing decreased sensitivity as the g2p FPR cutoff increases (see Supplemental Figures 3.4a and 3.4b in Appendix I).  Figure 3.2.  Time to change in tropism from R5 to non-R5 in MERIT trial participants rescreened for tropism using a V3 genotypic assay. The original Trofile assay was used to determine tropism at screening and only those determined as having R5 virus were enrolled in the MERIT study, the assay was performed at subsequent visits where possible.  Screening plasma samples were rescreened using the V3 genotypic assay and re-stratified as being R5 (green line) or non-R5 (red line) for patients enrolled in the maraviroc twice-daily arm (A) or patients enrolled in the maraviroc once-daily arm (B). 3.3.2 Comparison with ESTA The maraviroc twice-daily baseline samples were also rescreened using ESTA.  There were 283 concordant R5 calls (81%) and nine concordant non-R5 calls (2%) between ESTA and V3 genotype.  Of the 59 discordant cases, there were 39 (11%) where ESTA identified non-R5 virus but V3 genotype identified R5-using virus, while the converse occurred in 20 cases (6%).  Where both assays determined an R5 viral population, patients on maraviroc twice-daily had a viral load reduction of approximately 3 log10 copies/mL compared with approximately 2 log10 copies/mL where both assays identified a non-R5  95 population (Fig. 3.3A).  Of interest, in the instance of discordance between the two assays, patients showed an intermediate response after the start of maraviroc twice-daily, with a viral load reduction of roughly 2.5 log10 copies/mL regardless of the direction of discordance (Figure 3.3A).  Similar results were obtained if concordance and discordance between the two assays were examined in terms of the percentage of patients with undetectable plasma viral loads (Figure 3.3B).  Patients identified as having non-R5 virus by both assays respond poorly to maraviroc twice-daily with an average of only 15% achieving an undetectable viral load at week 48.  In the instance of concordant R5 or where assay results were discordant, roughly 60%–70% of patients achieved undetectable viral loads by week 16 and sustained these to week 48.  From this data, it is difficult to determine whether one assay outperformed the other in the case of discordance.  Using triplicate testing resulted in the interpretation of 29 patients as having non-R5 virus, compared with only 20 if only single amplification testing was used (see Supplemental Figure 3.5, in Appendix I).  96  Figure 3.3.  Concordance between tropism results in MERIT trial participants rescreened for tropism using a V3 genotypic assay and ESTA. Screening samples from the MERIT study were rescreened for tropism by both ESTA and V3 genotype, and re-stratified as having R5 or non-R5 virus as indicated.  Using the maraviroc twice-daily population, concordance was evaluated using A) the change in pVL after the start of maraviroc, and B) the percentage of the population able to suppress viral load <50 copies/mL.  Green represents patients screened R5 by both assays (N=283); red represents patients screened non-R5 by both assays (N=9). Black lines indicate discordance between the two assays (N=59). 3.3.3 Assay Performance in Different HIV Clades The MERIT trial was performed at multiple centers worldwide, enabling the comparison of assay performance in B and non-B clades of HIV.  Most non-B clade virus was clade C, originating in South Africa.271,404 Regardless of the screening assay used, patients with clade B virus had a greater response to maraviroc twice-daily than non-B virus.  At week 48, the percentage of the population whose virus was suppressed below 50 copies/mL was comparable between those rescreened as R5 by ESTA (74%) and V3 genotype (70%) who were clade B (N=176; 194) and also similar between ESTA and V3 genotype among those with clade C virus (63% and 62%, respectively) rescreened as R5 (N=105; 96).404 Fewer than 36 patients infected with HIV other than clades B or C received  97 maraviroc, resulting in limited ability to distinguish differences in the response to maraviroc in other subtypes. 3.4 Discussion This study was designed to investigate the ability of a V3 genotypic tropism assay to predict the treatment response to CCR5 antagonists in the MERIT trial which compared maraviroc versus efavirenz in treatment-naive individuals determined to have exclusively R5 virus at screening by the original Trofile assay.  Within this population, V3 genotyping was able to identify patients who were less likely to respond well to maraviroc treatment and more likely to change tropism as defined by the original Trofile assay after exposure to maraviroc, but not to efavirenz, similar to ESTA. The MERIT trial has particular advantages when comparing tropism-screening assays in comparison to studies without treatment outcomes as endpoints.  The trial was a large, worldwide, multicenter, double blind, randomized controlled trial.  Patients enrolled were treatment naive and all received the same nucleoside background therapy, reducing confounding factors.  The efavirenz control arm was particularly useful in highlighting the difference in response for those rescreened non-R5.  In addition, the maraviroc treatment arms (twice-daily and once-daily) provided two independent populations for further investigation of the genotypic method in treatment response prediction. Numerous studies have compared V3 genotypic prediction methods and algorithms with Trofile and ESTA, the most widely used tropism assays.  Initially, Trofile was found to have an apparently greater sensitivity in identifying non-R5 variants.378 In  98 2007, however, Skrabal et al. compared the prediction of viral tropism using two phenotypic methods, including original Trofile assay, and a genotypic method consisting of population-based sequencing paired with a bioinformatics algorithm based on a support vector machine.405 In this study, they were able to identify 86.5% concordance in tropism call between the original Trofile assay and the genotypic method used.405 Few of these studies, however, have been able to directly investigate the relative ability of genotypic algorithms to determine tropism using actual treatment outcomes in well-powered studies as the gold standard.  The V3 genotypic tropism assay described here was also employed to rescreen the MOTIVATE population, where, despite discordant results genotype was broadly comparable to the original Trofile assay in predicting treatment response after the start of maraviroc.403 It is interesting to note that a considerable proportion of patients achieve virological success despite having non-R5 virus by either genotype or ESTA.  This is likely due to the activity of the nucleoside reverse transcriptase inhibitor background coupled with additional partial suppression of R5 variants by maraviroc in what are likely mixed tropic viral populations. It is not clear why maraviroc response rates were higher in patients infected with clade B versus clade C HIV, but there exist a very large number of possible confounders including differences in race, gender, and regional differences in health care management.  From the viewpoint of a screening assay, however, there was no obvious evidence of clade-specific decreases in performance relative to the ESTA assay between clade B and clade C. There are a number of inherent limitations of this study, including the retrospective nature of rescreening and the resulting fact that only patients with R5 results by the  99 original Trofile assay could be examined for treatment outcomes.  However, a similar analysis was performed with the MOTIVATE trial populations where patients with non-R5 virus by the original Trofile assay were enrolled in a sister safety study, A4001029.270,403 The study was also restricted to patient screening samples.  Follow-up data would have allowed for ongoing comparison in tropism assays as the study progressed and viral populations changed.  In addition, the majority of patients enrolled in the study were infected with HIV-1 clades B or C, leaving only limited power to investigate other non-B clades.  It should also be noted that the Combivir nucleoside reverse transcriptase inhibitor background used at the time that the MERIT trial was designed is much less commonly used now.  A further limitation of this study is the sequencing technology employed.  Population-based sequencing, generally, is only able to identify minority species that exist at a prevalence of approximately 20%.  Therefore, samples were amplified in triplicate to increase the sensitivity for detecting non-R5 virus.  The 5.75% g2p FPR cutoff point used here was optimized using data from the MOTIVATE trial and may not be ideal as the study populations of MOTIVATE and MERIT were different (treatment experienced vs. treatment naive, respectively).  Although the MERIT population may not be ideal for determining an optimal g2p FPR cutoff point (both because it is a smaller population and has been completely pre-screened, unlike MOTIVATE/A4001029), the data are consistent with the suggestion that g2p FPR cutoff points of 5.75 to 10% combined with triplicate sequence analysis are quite reasonable values for clinical interpretation of population-based V3 sequencing data (see Appendix I).  Similar analyses using “deep” sequencing, with a detection limit for minority species of approximately 2%, yield similar results and showed only a minor incremental benefit of this increased sensitivity.406  100 In summary, when patients enrolled in the MERIT trial were retrospectively rescreened using a population-based V3 genotypic tropism assay, maraviroc was found to be non-inferior to efavirenz in the population identified as R5 by genotype, but not in the non-R5 populations.  These results confirmed earlier results obtained when the same population was rescreened using ESTA.  As well, the performance characteristics of ESTA and genotype are broadly similar.  When population-based sequencing results were discordant with ESTA, neither assay outperformed the other in terms of treatment response.  These results have also been obtained using “deep” sequencing technology.   With the reduced cost and broader worldwide availability, population-based genotyping is an attractive option as a routine screening tool for tropism assessment.  101 Chapter Four: Maraviroc Treatment in Non-R5-HIV-1-Infected Patients Results in the Selection of Extreme CXCR4-using Variants with Limited Effects on the Total Viral Setpoint 4.1 Introduction HIV-1 infection is dependent on viral entry into a susceptible host cell. Entry is mediated by the binding of virus to the host CD4 receptor and one of two chemokine co-receptors present on the cell surface—either chemokine (C-C motif) receptor 5 (CCR5) or chemokine (C-X-C motif) receptor 4 (CXCR4).76,77,153 The majority of new infections appear to be due to CCR5-using ‘R5’ virus, which continues to predominate during the course of the infection.187,399 As time progresses however, ~50% of patients with HIV subtype B have a shift in their viral population to include a growing amount of CXCR4-using ‘X4’ virus, which is strongly correlated with rapid disease progression.183,187,196 Many patients with X4 virus have a viral population that contains both X4 and R5 virus (mixed infection). In some cases, the virus may also be able to use both co-receptors (dual infection).184,407 For the purposes of this paper, both X4 and dual/mixed populations will be referred to as ‘non-R5’. The co-receptor that HIV uses to enter the host cell (‘tropism’) is largely but not entirely dependent on the characteristics of the V3 loop region within the gp120 protein encoded by the env gene.373,374 The V3 loop region is usually 35 amino acid residues in size and bookended by cysteine residues between which a disulphide bridge forms, resulting in  102 the loop structure.167 Particular mutations have been associated with CXCR4 co-receptor use, including the substitution of basic residues at V3 loop positions 11 and/or 25.169,375 This observation formed the basis of the simpler rule-based co-receptor prediction methods.375 Multiple interpretive algorithms have since been designed to predict co-receptor use based on the sequence of relevant regions of the viral genome, in particular the envelope’s V3 loop region.  The use of machine-learning techniques has allowed prediction methods to become more complex, resulting in the improved sensitivity of genotypic methods.378,402 Maraviroc is the first clinically available CCR5 antagonist.249 Patients screened as having R5 virus in the clinical trials of maraviroc had a significantly greater virological response to maraviroc than those with non-R5 virus populations, demonstrating the need to identify HIV tropism prior to therapeutic initiation of maraviroc.266,270,408 A number of both phenotypic and genotypic screening methods have been developed in order to determine HIV-1 tropism in patient samples. Initially, the most widely used method of identifying tropism was the phenotypic original Trofile assay (Monogram Biosciences).267 Since its introduction, adjustments have been made to increase its sensitivity and it is now referred to as the Enhanced Sensitivity Trofile Assay (ESTA).286 More recently, a population-based genotypic tropism assay has been developed in which the V3 loop is sequenced and interpreted using a bioinformatics algorithm such as geno2pheno[coreceptor] (g2p).383,402,403 The same concept can be achieved with enhanced sensitivity using next-generation sequencing technology.291,409 Next generation sequencing can be used to quantify viral variants accounting for <2%–3% of the viral population, in comparison with  103 the approximate 20% limit of standard sequencing technologies.291,409,410 This allows a more thorough evaluation of the non-R5 and R5 variants present within a sample. Using 454 “deep” sequencing methods to identify tropism, we longitudinally investigated and quantified the selective pressure of maraviroc on the dynamics and composition of the viral population in four patients with non-R5 virus who had experienced maraviroc treatment failure. 4.2 Methods 4.2.1 Study Population The four patients were participants in the AIDS Therapy Evaluation in the Netherlands (ATHENA) observational cohort, which has been approved by local and national institutional review boards.411  Three patients harboured non-R5 HIV populations by ESTA.  The remaining patient harboured R5 virus, but was predicted to harbour non-R5 virus by 454 “deep” sequencing.  Despite the presence of non-R5 virus, the potential value of maraviroc as an addition to an ongoing failing therapy regimen was investigated.  No other drugs were added, to preserve future treatment options in these heavily pre-treated individuals. Patient A started treatment in 1995 and was treated sequentially with two duo nucleoside reverse transcriptase inhibitor (NRTI) regimens, followed by four protease inhibitor (PI)-based regimens and one non- nucleoside reverse transcriptase inhibitor (NNRTI)-based regimen.  Patient A discontinued all highly active antiretroviral therapy  104 components (Combivir, tenofovir and ritonavir-boosted atazanavir) on his own initiative prior to starting maraviroc.  Patient B started treatment in 1995 with mono and duo NRTI-based and PI-based regimens followed by one PI+NNRTI-based regimen.  At the time of starting maraviroc, Patient B was being treated with Combivir, tenofovir and ritonavir- boosted tipranavir.  Patient C had started triple drug therapy in 1997 with a PI-based regimen followed by two NNRTI-based regimens.  Patient C was being treated with tenofovir, emtricitabine and nevirapine when maraviroc was introduced; however, during the course of maraviroc treatment, nevirapine was substituted with ritonavir-boosted lopinavir as part of the background regimen.  Patient D started therapy in 1995 and was treated with mono and duo NRTI regimens, three PI-based regimens and one NNRTI-based regimen.  Patient D was being treated with tenofovir and ritonavir-boosted tipranavir when maraviroc was added. Four longitudinal samples were collected from each patient over a median of 53 days (IQR 27 days).  The median length of maraviroc add-on therapy was three weeks (range 13 – 26 days).  The period between the start of maraviroc and the first sample analyzed ranged from six to 26 days (median 11 days).  Three of the four patients had two post- maraviroc samples sequenced. 4.2.2 Laboratory Methods A total of 17 samples were processed for 454 “deep” sequencing.  Viral RNA was extracted from 500µL of frozen plasma.  Of the extraction eluate, 4µL were used in a RT-PCR reaction using the Superscript III One-Step RT-PCR system (Invitrogen) to reverse transcribe and amplify the V3 loop of the gp120 protein encoded by the HIV-1 env gene.   105 This reaction was performed in triplicate.291,292 Nested second-round PCR was performed on the triplicate samples with primers incorporating unique sample-specific tags, or multiplex identifiers (MIDs), for sample identification (Appendix 2, Supplemental Table 4.1).  In preparation for clonal amplification using emulsion PCR, the resulting triplicate PCR amplicons were quantified and combined in equal proportions such that 2x1012 DNA amplicons from each triplicate sample were included in the amplicon libraries.  An amplicon library contained up to 12 samples, each identified by a specific MID.  Following this combination step, amplicon libraries were then purified and re-quantified.291,292 Amplicon libraries were diluted to 1×107 molecules/mL, such that five molecules of DNA were present for every DNA capture bead.  Approximately 790 000 beads were loaded into each of the four regions of the 454 pyrosequencing picotitre plate.  Samples were sequenced in both the forward and reverse directions using the 454 GS-FLX platform (Roche, 454 Life Sciences). 4.2.3 Sequence Processing & Tropism Interpretation A median of 14 000 sequence reads per sample was obtained.  After processing for sequence quality, a median of 12 500 sequence reads per sample were included in the analyses.412 Viral tropism was interpreted using the g2p algorithm with a false-positive rate (FPR) cutoff point of 3.5% for each variant sequence.  As previously reported for “deep” sequence analyses, sequences below this g2p FPR value were inferred as being non-R5.291,409 Samples were determined to be non-R5 when the proportion of non-R5 variants exceeded the 2% cutoff point previously described.291 The percentage of non-R5 variants within each viral population was determined from the total useable read counts for each sample.  Having calculated this, the viral load of non-R5 and R5 populations could be determined  106 separately.  The non-R5 plasma viral load (pVL) was calculated by simply multiplying the percentage of non-R5 by the total pVL, with the R5 pVL as the remainder. 4.2.4 Fitness Analysis Following g2p scoring, sequences were stratified into five bins based on their g2p FPR (FPR, 2%, 2% ≤FPR, 3.5%, 3.5%≤ FPR, 10%, 10%≤ FPR, 20%, 20%≤ FPR).  The observed frequency of the sequences within each bin was calculated and plotted over time for each time point and patient (allele frequencies).  Each bin represents a different ‘level’ of tropism and level of response to maraviroc. Relative viral fitness in the presence of maraviroc was calculated according to Wright’s recursive formula for allele frequency evolution under haploid selection, with the generation time assumed to be one day.413 We fit a four-parameter fitness curve based on the flexible gamma cumulative distribution function to the longitudinal 454 “deep” sequence data, given this model.  The expected allele frequency of evolution was calculated from the initial observed allele frequencies at the earliest time point given the four model parameters of the fitness function.  Here, a relative fitness of 1.0 is equal to the most-fit allele in the patient viral population. 4.3 Results All four patients were confirmed as having non-R5 virus at baseline by 454 “deep” sequencing, the baseline being defined as the most recent sampling time point prior to the start of maraviroc treatment.  The viral populations were composed of between <1%  107 and 58% non-R5 virus at baseline (median 24%, IQR 6%–44%).  Even at the first time point following the start of maraviroc treatment, it was apparent that non-R5 virus was not inhibited by maraviroc as the percentage of non-R5 variants increased in all patients (Figure 4.1).  In fact, the proportion of non-R5 populations increased by a median of 71% (IQR 49% – 91%) after a median of 11 days of maraviroc exposure (Figure 4.1).  We note an increase in the percentage of non-R5 virus in Patient A several days before the documented start of maraviroc; we could not find an explanation for this observation.  When tested using ESTA, Patient C was determined to have R5 virus only at baseline (<1% non-R5 virus using 454 “deep” sequencing).  The result from the previous time point, 49 days prior to baseline, was determined as 11% non-R5 virus using 454 “deep” sequencing.  This non-R5 population subsequently increased to 99% after 21 days of maraviroc exposure.  In all four patients, the predominant non-R5 variant contained a basic residue (Lysine or Arginine) at V3 loop positions 11 or 25 (Table 4.1). The results here demonstrate a rapid replacement of the R5 population by non-R5 variants following the start of maraviroc treatment (Figure 4.2).  The overall pVL increased, with a median of 0.36 log10 copies/mL by the final time point (Figure 4.1).  Viral loads in patients A, B and D increased (0.27, 0.73 and 0.45 log10 copies/mL, respectively) after 26, 21 and 13 days of exposure, respectively.  Patient C showed a decrease in viral load, with 1.32 log10 copies/mL after 21 days of maraviroc.  When separated into R5 and non-R5 viral loads, it is evident that changes in the overall pVL are due to a large decrease in the R5 populations (median -1.64 log10 copies/mL, IQR -1.85 to -1.27) and the outgrowth of the non-R5 populations, which expanded to comprise a median of 98% (IQR 99%–96%) by the last time point (Figure 4.1).  108  Figure 4.1.  Virological response to maraviroc in patients screened as having non-R5 virus. The change in plasma viral load (log10 copies/mL) and percent non-R5 following exposure to maraviroc in four patients experiencing treatment failure and screened as having dual/mixed HIV-1.  Viral tropism was determined from the V3 sequences generated using “deep” sequencing and interpreted using g2p with a 3.5 FPR cutoff point.  The black line represents the change in total pVL, the green line represents the change in R5 only pVL, and the dashed red line represents the change in %non-R5 on the secondary vertical axis.  The vertical dashed black line indicates the date at which the patient started maraviroc add-on therapy.  109 Table 4.1.  Changes in the most prevalent sequence following exposure to maraviroc in patients screened as having non-R5 virus. A summary of the change in prevalence and FPR for the most common HIV-1 variants from each of the four patients with dual/mixed HIV-1, before and after exposure to maraviroc.  Amino acid substitutions between the most prevalent sequence at the first and last time points are noted, dashes represent no change at the position.  Sequence prevalence, as a percentage of the number of useable reads, was calculated for each sequence as found at the pre-therapy and post-therapy time points.  A variant with a g2p FPR ≤ 3.5% was considered to be non-R5 (red).  Amino acid positions marked by a square are associated with mutations characteristic of CXCR4-using virus.  ** Unusually, the majority of sequences in the forward direction from the first two time points for patients A and C contain a frameshift mutation and were omitted from the analyses.  Where blank, the sequence was not detected at the time point.  110  Figure 4.2.  Virological response to maraviroc in R5 and non-R5 viral populations. The change in median plasma viral load of non-R5 and R5 variants prior to and following the start of maraviroc in the viral populations of four patients with dual/mixed HIV-1 experiencing treatment failure.  The red bars represent the change in the non-R5 population; the green bars represent the change in the R5 population.  Error bars are shown for the interquartile range.  Following maraviroc exposure, all four individuals showed decreases of the R5-viral load of at least 1 log, the R5-population was replaced by non-R5 variants (g2p FPR<3.5). Variants were divided into five strata, or “bins”, (FPR<2%, 2% ≤FPR< 3.5%, 3.5% ≤FPR< 10%, 10% ≤FPR< 20%, 20% ≤FPR) according to their g2p FPR in order to estimate relative viral fitness in vivo in the presence of maraviroc (Figure 4.3).  The frequency of the sequences was plotted for each bin at each time point, showing the change in g2p FPR over time.  The non-R5 variants with extremely low g2p FPR values (<2%) had the most substantial increase in prevalence.  R5 variants (FPR≥ 3.5%) were represented in three bins, the higher g2p FPR indicating variants increasingly characteristic of R5 virus and most likely to respond to maraviroc.  There was no corresponding increase in the proportion of virus with a g2p FPR between 3.5% and 20%, thus implying the relative inhibition of these variants (Figure 4.3).  111  Figure 4.3.  Changes in sequence prevalence by g2p FPR following maraviroc exposure. Viral sequences were stratified based on FPR, the frequency of sequences within each stratification was then plotted against time from baseline.  Sequences were generated using 454 “deep” sequencing.  The gray shaded area of each graph represents the time prior to the start of maraviroc; white representing time after the start of maraviroc.  The circle marker (o) represents non-R5 variants with extremely low FPR values (FPR<2), the triangle marker (Δ) and plus sign marker (+) represents variants with low FPR bordering the non-R5 or R5 call (2≤FPR≤3.5; 3.5≤FPR≤10, respectively).  The x mark (x) and asterisk (*) markers represent R5 variants (10≤FPR≤20), the asterisk (*) with the highest FPR values and most likely to respond to MVC (20≤FPR). Based on the data collected, we employed a population genetic model to measure the relative fitness advantage of predicted X4 viral variants in the presence of maraviroc in vivo under haploid selection (generation time of one day).  The relative frequencies of these allelic classes over time were fitted to Wright’s haploid selection model to estimate four parameters of a flexible monotonic function relating relative fitness to g2p FPR (Figure 4.4).   112 The viral fitness of variants with an g2p FPR> 5% has little selective advantage in the presence of maraviroc, remaining relatively constant at a relative fitness of 0.2, where 1.0 is equal to the most-fit allele in the patient viral population.  However, when variants had a g2p FPR< 5% the relative fitness began to increase markedly.  At a g2p FPR between 2% and 0%, the relative fitness was shown to increase from ~0.65 to 1.0, representing an increase in fitness of ~0.2 units for every FPR unit (Figure 4.4).  Figure 4.4.  Viral fitness in the presence of maraviroc as a function of g2p false positive rate. Data from four individuals with non-R5 virus and exposed to maraviroc was pooled to analyse viral fitness as a function of g2p FPR in the presence of maraviroc.  Relative fitness in the presence of maraviroc was calculated according to the Wright recursive formula for allele frequency evolution under haploid selection where generation time was assumed to be one day.   A four-parameter fitness curve was fit based on the gamma cumulative distribution function to the longitudinal 454 “deep” sequence data.  Sequences were categorized into bins based on g2p FPR cutoff points (0, 2, 3.5, 10, and 20) and observed allele frequencies calculated across bins for each time point and patient.  Expected allele frequencies were calculated from the initial observed allele frequencies (earliest time point) given the model parameters (four parameters of the fitness function).  Here, a relative fitness of 1.0 is equal to the most-fit allele in the patient viral population.  Red circles indicate commonly used false-positive cutoff values.  113 The most common sequences from the pre-therapy and post-therapy samples tested were compared for each patient (Table 4.1 and Supplemental Table 4.2 in Appendix II).  The prevalence of the most common sequences as a percentage of the total useable sequences obtained for the sample was calculated.  For each patient, the most common sequence was different between the pre-therapy and post-therapy time points; however, the most common sequence at the post- therapy time point was always present at a pre-therapy time point (Table 4.1 and Supplemental Table 4.2 in Appendix II).  In the case of Patient C, the most common sequence at the post-therapy time point was not found at the first pre-therapy time point but was found at baseline.  With the exception of Patient B, the most common pre-therapy sequence was an R5 variant, whereas the post- therapy sequence was a non-R5 variant.  The most common sequences for Patient B, pre- and post-therapy, were both non-R5 variants.  Notably, the non-R5 variants that became the most common sequence all had extremely low g2p FPR values (between 1.1% and 1.8%).  For each patient, the most common sequence at the post-therapy time point had a greater prevalence than the most common sequence at the pre-therapy time- point, showing a decrease in sequence variation and a selective advantage for those sequences with a low g2p FPR in the presence of maraviroc (Table 4.1). 4.4 Discussion Using 454 “deep” sequencing, viral tropism was examined in plasma samples from patients with non-R5 HIV-1 who were exposed to maraviroc add-on therapy as an attempt to obtain a potential clinical benefit.  The R5 viral population in all patients responded to maraviroc as indicated by the observed decrease in R5 pVL.  However, consistent with the  114 A4001029 study, the total viral population did not have a significant virological response to maraviroc.270 In fact, alongside the suppression of R5 virus, the non-R5 populations modestly expanded such that the total viral load remained nearly constant.  Interestingly, the most successful non-R5 variants were those with an extremely low FPR using g2p, which suggests that maraviroc may inhibit HIV with a lower g2p FPR than previously thought. It has previously been suggested that the emergence of non-R5 variants is due not to a switch in co-receptor use by variants but to the selective outgrowth of a pre-existing non-R5 reservoir in the presence of maraviroc or vicriviroc.277,414 This is consistent with the more recent data generated by applying the same 454 “deep” sequencing screening method to a group of 24 individuals from Spain with non-R5 HIV populations who had been exposed to maraviroc for 8 days.415 Unlike most previous analyses using individuals participating in clinical trials of maraviroc and vicriviroc, patients here were given maraviroc despite having been screened as having non-R5 HIV by ESTA.  In heavily pre-treated patients experiencing drug-related toxicity, such as those studied here, clinicians have to apply individually tailored empirical therapy regimens.  Although this is not recommended, all four patients received maraviroc to assess its virological or immunological efficacy.  This study was a retrospective analysis of four of these cases. After a median of 21 days of maraviroc add-on therapy, patients with non-R5 virus at baseline had a dramatic increase in their non-R5 population, such that despite a minimum decrease of 1 log10 copies/mL in the median R5 pVL, the overall pVL remained nearly constant.  In each patient from the study presented here, the most prevalent  115 sequence after roughly three weeks of exposure to maraviroc was non-R5, and in agreement with the findings of Westby et al. these sequences had already been observed at a pre-therapy time point.414 Given the short amount of time that elapsed between sampling, it is unlikely that the mutations conferring non-R5 usage that had been observed in the pre-maraviroc samples had disappeared only to re-emerge de novo after the start of maraviroc therapy.  The more parsimonious interpretation of our results is that non-R5 variants that had been present before maraviroc therapy continued to persist in the population until they gained a selective advantage at the start of maraviroc therapy. There is an ongoing debate regarding the appropriate g2p FPR cutoff point to apply in order to effectively predict a patient’s viral response to maraviroc.  Genotypic tropism testing in British Columbia, Canada, uses a cutoff point of 5.75% in population-based sequencing methods and 3.5% in next generation sequencing methods, both of which have been retrospectively validated using populations from clinical trials of maraviroc.291,416 However, the European guide- lines recommend a more conservative g2p FPR cutoff point of 10%.283 Similarly, the German–Austrian guidelines recommend a g2p FPR cutoff range of 5%–15% depending on the treatment options (www.daignet.de).  Longitudinal analysis of the independent CCR5- and non-CCR5-using HIV populations shows that maraviroc selects for viruses with an extremely low g2p FPR (between 0% and 2%), suggesting that the g2p FPR cutoff points as high as 10% may exclude individuals who could respond to maraviroc. This analysis provides another way in which to look at and interpret g2p FPR cutoff points clinically.  By composing a fitness map, measuring the shape of a fitness function  116 over the range of co-receptor-tropism predictions from g2p, there is a possibility of identifying results that are overlooked when making inferences based on outcome.  The data are more consistent with the lower g2p cutoff points of the newer German–Austrian guidelines (as low as 5%) compared with the values of up to 20% suggested in older guidelines.  Based on the results of this study and a retrospective analysis of clinical trials, a cutoff point of 5% may be more applicable and should be explored further.  However, due to our small sample size, this finding needs to be approached with caution, and oversampling is a concern as patients, with the exception of Patient C, generally had a high pVL following the start of maraviroc treatment.  Unfortunately, we were unable to use the primer ID method more recently described by Jabara et al., and therefore the viral input copy number as well as the effects of PCR and sequencing error are unknown.417 In conclusion, when patients with dual/mixed virus were exposed to maraviroc add-on therapy, selection for non-R5 populations caused them to expand to fill the space once occupied by R5 variants.  Selection favoured non-R5 variants with an extremely low g2p FPR, which may indicate that the antiretroviral activity of maraviroc may be effective in a broader-range of HIV variants than currently suspected.  117 Chapter Five: Next Generation “Deep” Sequencing to Evaluate Viral Tropism in HIV-1 Patients Exposed to Short-term Maraviroc Add-on Therapy 5.1 Introduction CCR5 antagonists are one of the more recently approved antiretroviral drug classes with maraviroc (MVC) first to be approved for clinical use.397 Viral entry into the host cell is contingent on the binding of HIV to the CD4 receptor and to one of the chemokine co-receptors present on the target cell surface, either CXCR4 or CCR5.76,77,150 Maraviroc impedes HIV by allosterically inducing a conformational change in the structure of the CCR5 receptor extracellular loops, making the receptor unrecognizable to the virus and thereby inhibiting entry into the cell.258,259 Maraviroc is, however, ineffective against CXCR4-using (X4) virus.187,196 An individual’s viral population may be composed of virus using CCR5, CXCR4, or a mixture of both co-receptors, referred to as dual and or mixed (dual/mixed).  In order to ensure virological response when introducing a CCR5 antagonist as part of a therapy regimen, it is highly recommended that the co-receptor(s) used by the patient’s HIV population, or tropism, first be identified.279,283,284 Genotypic and phenotypic assays are used to determine tropism, each with their own set of limitations including: cost, availability, turnaround time and sensitivity.  The Enhanced Sensitivity Trofile Assay (ESTA) is the most widely known phenotypic assay at present.273 The original Trofile assay, the ESTA predecessor, was used for the enrollment  118 screening used in the maraviroc clinical trials.266,267,270,271 A number of other research groups have developed phenotypic tropism tests such as the Toulouse Tropism Test in France or Tropism Coreceptor Assay Information (TROCAI) in Spain.418,419 Genotypic assays using both population-based and next generation “deep” sequencing technologies have been developed using the study populations from the clinical trials of maraviroc, the MOTIVATE 1 and 2, A401029 and MERIT.291,292,403,409,416 The population-based genotypic assay is comparatively cost efficient and accessible, however, “deep” sequencing methods offer increased sensitivity.283 A third approach to determine the potential effectiveness of treatment with maraviroc was explored in the trial of the Maraviroc Clinical Test (MCT).  As a supplement to tropism testing, the Maraviroc Clinical Test is used to evaluate virological response to eight days of true or “functional” maraviroc monotherapy as an in vivo test to determine drug sensitivity.  Here, “functional” maraviroc monotherapy is the addition of maraviroc to a failing antiretroviral therapy regimen.420,421  A small number of studies exposing patients to short-term CCR5 antagonist monotherapy have been conducted.422,423 For instance, as part of the Phase II clinical trials of maraviroc, patients were exposed to maraviroc monotherapy for 10 days in order to evaluate efficacy.  At all twice-daily doses patients had a mean plasma viral load (pVL) decrease of ≥1.6 log10 copies/mL at 10 to 15 days post exposure.422 Similar results were observed when the activity and safety of vicriviroc was assessed in patients exposed to vicriviroc monotherapy for 14 days.423 However, the patient populations in these studies had been screened for tropism, enrolling those with R5 virus only.  A small study conducted in the Netherlands observed the effects of maraviroc given to four individuals with dual/mixed virus failing antiretroviral therapy, over a median of 21 days.  Here, the  119 R5 viral populations decreased in prevalence and the non-R5 viral populations flourished such that the overall plasma viral load remained relatively constant.424 The results of this study demonstrate the maraviroc-associated selection of preexisting non-R5 variants. In order to further investigate the effects of maraviroc on pre-existing non-R5 variants and the utility of genotypic tropism inference, we retrospectively investigated the selective pressures of maraviroc exposure on the viral populations of 27 individuals enrolled in the Maraviroc Clinical Test trial.  The study population included both individuals screened as having R5 and non-R5 HIV.  Patients received either true or “functional” maraviroc monotherapy for eight days, with plasma samples drawn at multiple time points.  Using the 454 “deep” sequencing platform we traced the effects of short-term maraviroc monotherapy on both R5 and non-R5 viral populations. 5.2 Methods 5.2.1 Study Population As part of the prospective Maraviroc Clinical Test trial, patients received maraviroc for eight days (Day 0 to Day 7) as either a true monotherapy or added to a failing antiretroviral regimen as a “functional monotherapy”.420,425 The Maraviroc Clinical Test trial enrolled patients at the Virgen del Rocio University Hospital in Seville, Spain between 1 July 2008 and 1 February 2011.  All patients had a detectable pVL (>50 copies/mL) for six months prior to enrollment and had no prior experience with co-receptor antagonists.420,425 Plasma was drawn and plasma viral load determined at Days 0 (pre-therapy), 2, 4 and 7.  Tropism had been determined at Day 0 using ESTA and TROCAI for comparison with the  120 Maraviroc Clinical Test results.419 Nine and eight individuals were found to have dual/mixed virus at Day 0 by Trofile and TROCAI, respectively.  We received the stored plasma samples from the first 27 consecutive patients enrolled in the Maraviroc Clinical Test trial.  A total of 83 samples were tested, a median of three time points for each of the 27 individuals.  All samples were tested for tropism using the 454 “deep” sequencing assay previously described by Swenson et al..291,292,412 The original Maraviroc Clinical Test trial conducted in Spain was approved by Virgen del Rocío University Hospital Ethical Committee.420,421 The Providence Health Care and University of British Columbia Research Ethics Board approved the retrospective genotypic study described here. 5.2.2 Sample Preparation & “Deep” Sequencing Viral RNA was extracted from 500µL of frozen plasma using the NucliSENS EasyMAG (bioMérieux).  The V3 loop was amplified in triplicate using 4µL of extract in a one-step RT-PCR reaction (Superscript III One-Step RT-PCR System, Invitrogen).  A nested second-round PCR reaction was performed using primers containing system-specific adaptors, multiplex identifiers (MIDs) and target-specific sequence (Appendix 3, Supplemental Table 5.1) Tagged triplicate PCR products were quantified using the Quanti-iT Picogreen dsDNA Assay Kit (Invitrogen) and combined in equal proportions (2 x1012 molecules/µL) to generate amplicon libraries.  Each amplicon library was composed of 12 samples tagged with a unique MID to be sequenced in either the forward or reverse direction for a total of 24 samples, each in triplicate.  These libraries were purified and diluted to 2x106 DNA molecules/µL.  Libraries were clonally amplified using emulsion  121 PCR and pyrosequencing was subsequently performed using 454 titanium chemistry, following the manufacturer’s protocol, and run on a 454 GS-FLX platform (Roche/454 Life Sciences) as previously described.291,292 5.2.3 Sequence Processing A median of 6998 (IQR 5712–8412) usable sequences were collected for each sample.  Sequences passing quality filters at various points in the pipeline were also required to meet three basic criteria:  ≥96 and ≤118 nucleotides, the number of nucleotides composing the V3 sequence must be divisible by three, and the V3 loop must start and end with cysteine.412 Sequences that did not meet these expectations were removed from the analysis.  Gaps placed in the sequences during alignment were removed and sequences were then interpreted using the geno2phenocoreceptor algorithm (g2p) and viral tropism inferred using a false positive rate (FPR) cutoff point of 3.5, as previously described.291,409,412 Sequences that were unable to be scored by g2p for reasons including stop codons were removed from the analysis as well.  Sequences with a g2p FPR ≤3.5 were considered non-R5; a patient was screened as having non-R5 virus if their viral population was ≥2% non-R5.291 5.2.4 Bioinformatics Analyses As an exploratory analysis of cutoff points, sequences were grouped into five categorical bins according to their FPR: 0≤FPR≤2, 2<FPR≤3.5, 3.5<FPR≤10, 10<FPR≤20, FPR>20.  These bins were then plotted by frequency for each of the four time points.  To visualize intra-patient variation, we constructed phylogenetic trees for each patient using maximum likelihood and sequences from Day 0 and Day 7.  Sequences with a count of 15  122 or less were removed in order to remove any potential minor cross contamination.  The remaining sequences were aligned with Multiple Alignment using Fast Fourier Transform (MAFFT) using the L-INS-I strategy.426 The trees were built using Randomized Axelerated Maximum Likelihood (RAxML) with the GTR+Gamma model (of the General Time-Reversible (GTR) family of models) of molecular evolution.427 The measured in vivo fitness of a variant as a function of g2p FPR was also determined.  The relative fitness for each variant in the presence of maraviroc was calculated in accordance with the Wright recursive formula for allele frequency evolution under haploid selection.  Here the generation time was assumed to be one day.  A four-parameter fitness curve based on the gamma cumulative distribution function was fit to the longitudinal 454 “deep” sequence data.  Again, variants were categorized into bins based on FPR (0, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 5, 10, 25, 50, 100), incorporating additional FPR bins in order to better define the point at which viral fitness began to increase in the presence of maraviroc.  The observed allele frequencies were calculated across bins for each time point and each patient.  Expected allele frequencies were calculated from the initial observed allele frequencies (the earliest time point) given the four parameters of the fitness function.  A variant was equally fit to the most-fit allele in the patient viral population if the relative fitness was calculated to be 1.0.  It is assumed that the most-fit allele is an X4 variant with a FPR of 0, and the least-fit allele is an R5 variant with a FPR of 100.  123 5.3 Results A summary of the patient baseline characteristics (N=27) is presented in Table 5.1.  Of the 83 samples tested, 81 samples were sequenced successfully.  We were unable to sequence one sample; the second sample failed due to sample contamination and was removed from the analyses.  Ten of the 27 patients (37%) tested were found to have a non-R5 infection (≥2% non-R5 virus) according to 454 “deep” sequencing of the baseline (Day 0) samples.  The median percentage of non-R5 sequences in these baseline samples was 24.5% (IQR 15.3 – 35.0).  In the same set of samples, by the final time point (Day 7) the median percentage of non-R5 sequences had increased to 70.2% (IQR 8.30 – 95.4) (Figure 5.1).  There were two instances where patients screened as having non-R5 virus at Day 0 (14.6 and 26.0% non-R5) seemingly cleared their non-R5 virus by Day 7 of maraviroc (0 and 1% non-R5, respectively).  Conversely, there were two instances of patients being screened as having a R5 infection at Day 0 (0.4 and 0.5% non-R5) that had an outgrowth of non-R5 virus by Day 7 (70.9 and 23.3% non-R5, respectively). Table 5.1.  Baseline Characteristics Table (N=27).  *MVC monotherapy here indicates patients on true monotherapy, not those on “functional MVC monotherapy”  124 The median pVL at Day 0 of those screened as having non-R5 was 4.85 log10 copies/mL (IQR 4.54 - 5.14), slightly greater than the pVL of those screened as having R5 virus, 4.58 log10 copies/mL (IQR 4.15 – 4.89).  After short-term exposure to maraviroc there was a median decrease of 1.37 log10 copies/mL (IQR 1.08 – 1.53) in the pVL of those screened as having R5 virus, compared to a median decrease of 0.2 log10 copies/mL (IQR -0.13 to 1.24) in those screened as having non-R5 virus (Figure 5.1).  When the overall pVL was subdivided into separate R5 and non-R5 pVL based on the 3.5 g2p FPR cutoff, the median R5 pVL decreased by 1.37 and 1.27 log10 copies/mL in patients screened as having R5 and non-R5 virus, respectively.  Conversely, in patients screened as having non-R5 virus the non-R5 pVL increased slightly by a median of 0.06 log10 copies/mL (IQR -1.23 to 0.36).  Differences in the virological response at all time points were not found to be statistically significant between individuals screened as having R5 or non-R5 virus using the Wilcoxon-Mann-Whitney test (Figure 5.1).  125  Figure 5.1.  Virological response to short-term maraviroc exposure. Box-and-whisker plot illustrating the change in plasma viral load over the course of short-term maraviroc treatment, Day 0 to Day 7.  Patients were screened by 454 “deep” sequencing as having R5 or non-R5 virus at Day 0 (g2p FPR 3.5%), where populations comprised of 2% non-R5 variants were considered to be non-R5. In order to identify changes in variant frequency as a function of g2p FPR in the viral populations during short-term exposure to maraviroc, HIV sequences were stratified into five FPR bins ranging from FPR 0 to >20 based on the g2p interpretation.  The frequency of each bin was analyzed at each time point for each patient.  The results of this analysis are summarized in two figures, patients screened as having non-R5 virus and those experiencing a tropism switch from R5 to non-R5 are summarized in Figure 5.2, whereas patients screened as having R5 virus are summarized in Supplemental Figure 5.1 of Appendix III.  There were three basic patterns in g2p FPR frequency.  First, those screened as having R5 virus with a g2p FPR >20 at baseline had very little change in the frequency of variants based on g2p FPR.  In general, if non-R5 variants with a g2p FPR ≤2 were present at Day 0 these variants were likely to increase markedly by Day 7.  This was  126 accompanied by a decrease of similar magnitude in the R5 variants with a g2p FPR >20 as was observed clearly in patients 11, 22, 23, 24 and 25 in Figure 5.2.  However, in the instance of variants with 3.5<FPR≤20 the frequency changes based on g2p FPR were somewhat unpredictable.  There were a few instances where in the absence of non-R5 virus, variants with a g2p FPR>20 decreased in frequency as variants with lower, yet R5, g2p FPR increased in frequency (see Supplemental Figure 5.1 in Appendix III).  In the instances where non-R5 variants with g2p FPR≤3.5 become the predominant sequence at Day 7, these variants were detected at Day 0.  In two cases these non-R5 variants were found at a prevalence of <1% of the Day 0 viral population. (Table 5.2) Similar to the frequency as a function of g2p FPR plots, the phylogenetic analysis also revealed basic patterns of change in the viral populations after short-term maraviroc exposure.  In a majority of patients the R5 population became less diverse during treatment with maraviroc, with many R5 variants present at Day 0 becoming undetectable by Day 7.  The instances of shifting dominance in the viral populations by non-R5 virus are also clearly visible, where R5 variants become less diverse non-R5 variants become more diverse (see Supplemental Figure 5.2 in Appendix III).  For patients that experienced a tropism switch from non-R5 to R5, the non-R5 variants treed closely with the R5 variants, however, by Day 7 these variants were not detected (Supplemental Figure 5.3a in Appendix III).  When tropism switch was observed from R5 to non-R5, the non-R5 variants present at Day 7 seem to extend away from the cluster of R5 variants and become quite diverse (Supplemental Figure 5.3b in Appendix III).  127  Figure 5.2.  Changes in sequence frequency stratified by g2p FPR after short-term maraviroc therapy in patients found to have ≥2% non-R5 HIV at any time point. Patients screened as having non-R5 virus by 454 “deep” sequencing and two patients experiencing a tropism switch from R5 to non-R5 (p13 and P17).  FPR stratifications were based on the g2p interpretation of viral sequences generated by “deep” sequencing.  Variants were placed into bins according to their calculated g2p FPR and plotted over time.  The red line represents non-R5 variants with extremely low g2p FPR values (0≤FPR≤2), the yellow line represents variants just below the currently accepted 3.5 g2p FPR cutoff (2<FPR≤3.5), and the green line represents variants with low g2p FPR bordering the non-R5 or R5 call (3.5<FPR≤10).  The blue (10<FPR≤20) and black (20<FPR) lines represent R5 variants, black with the highest FPR values and most likely to respond to maraviroc.  128 Table 5.2.  The most prevalent sequence at Days 0 and 7 of maraviroc therapy from patients screened as having ≥2% non-R5 virus at any time point. Nucleotide sequences were translated into amino acids for patients screened as having non-R5 virus at Day 0 and those switching from having R5 to non-R5 virus by Day 7.  Amino acid differences between time points are indicated; a dash represents no change at a position.  Prevalence at pre- and post-maraviroc time points was calculated as a percentage of the total useable reads collected for the sample.  A g2p FPR of ≤3.5 is considered to be non-R5 (red); g2p FPR >3.5 is considered to be R5 (green).  Day 0 screening results by the phenotypic ESTA and TROCAI assays are also indicated. * indicates a tropism switch from non-R5 to R5; ** indicates a tropism switch from R5 to non-R5          N/A: tropism test wasn’t performed; “---“: no tropism result available; “D/M”: dual/mixed tropic 129 When the relative fitness of a viral variant was calculated as a function of g2p FPR, the results revealed a marked increase in viral fitness starting at a g2p FPR of approximately two.  Viral fitness continued to increase as g2p FPR approached zero (Figure 5.3).  This is in concordance with the observation demonstrated in Figure 5.2 where variants with g2p FPR≤2 expanded in the presence of maraviroc.  Conversely, no selective advantage was observed for variants with g2p FPR>2 in the presence of maraviroc.  The increase in fitness at extremely low g2p FPR (≤2) increased from approximately 0.61 to 1.0, which translates to an increase in viral fitness of 0.2 for every FPR unit in this range.  Figure 5.3.  Viral fitness in the presence of maraviroc as a function of g2p FPR (FPR ≤ 20). Relative fitness in the presence of maraviroc was calculated according to the Wright recursive formula for allele frequency evolution under haploid selection where generation time was assumed to be one day.  We fit a four-parameter fitness curve based on the gamma cumulative distribution function to 454 “deep” sequence data.  Sequences were categorized into bins based on g2p FPR cutoffs (0, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 5, 10, 25, 50, 100) and observed allele frequencies were calculated across bins for each time point and patient.  Expected allele frequencies were calculated from the initial observed allele frequencies (earliest time point) given the model parameters (four parameters of the fitness function).  The vertical dashed red line indicates the 3.5 g2p FPR cutoff point.  Here, a relative fitness of 1.0 is equal to the most-fit allele in the patient viral population.  130 Tropism testing of Day 0 samples was performed using three assays, 454 “deep” sequencing as well as the phenotypic ESTA and TROCAI assays.  Results were generated by 454 “deep” sequencing for all 27 patients at multiple time points.  Phenotypic testing was only performed at Day 0 and results were generated for 25 and 19 samples by ESTA and TROCAI, respectively.  The ability to predict virological response was compared between “deep” sequencing and each of the two phenotypic assays.  When concordant, patients screened as having R5 virus had a median pVL decrease of >1 log10 copies/mL after eight days of maraviroc exposure.  Conversely, median pVL remained relatively unchanged in patients screened as having non-R5 virus.  When discordant (ESTA N=6; TROCAI N=4), based on virological response the phenotypic assays performed better when predicting response to short-term maraviroc exposure (Figure 5.4).  Discordance between genotype and phenotype was observed for each of the four cases of tropism switch as identified by “deep” sequencing.  131  Figure 5.4.  Virological response to maraviroc demonstrating concordance between 454 “deep” sequencing and both ESTA and TROCAI. Ability to predict virological response of short-term maraviroc exposure by phenotypic ESTA (N=25) and TROCAI (N=19) was compared with that of genotypic 454 “deep” sequencing (g2p FPR 3.5%).  Comparisons of “deep” sequencing with ESTA and TROCAI are reported separately.  Patients with concordant tropism results have been combined and reported using the median pVL.  Instances of discordance (ESTA N=6; TROCAI N=4) were reported for each patient, however, pVL was only reported for Days 0 and 7 due to low N-values, and coloured based on the direction of discordance.  132 5.4 Discussion In this study the use of 454 “deep” sequencing technology facilitated the identification and approximate quantification of non-R5 minority species within the viral populations of individuals enrolled in the trial of the Maraviroc Clinical Test.  Regardless of co-receptor use, these patients received maraviroc as a true monotherapy or as an add-on therapy to a failing antiretroviral regimen, “functional monotherapy”.  Strictly R5 populations responded well to maraviroc, as expected.  However, in viral populations characterized with ≥2% non-R5 virus, the number and diversity of R5 variants present decreased.  This reduction of the R5 population provided non-R5 variants room to expand as demonstrated by the percentage of non-R5 virus present and pVL measures of these viral populations.  HIV sequence data also demonstrated a strong selection of non-R5 variants having extremely low g2p FPR (0 ≤ FPR ≤ 2), those most characteristic of X4 virus, in as few as eight days after the addition of maraviroc when viral populations were determined to be non-R5 at Day 0. Based on previously published work, it is not surprising that after eight days of maraviroc exposure the median R5 pVL decreased by at least 1 log10 copies/mL in both patients screened as having R5 and non-R5 viral populations.  In the Phase II clinical trials of maraviroc, patients screened as having R5 virus were exposed to maraviroc for 10 to 15 days, during which time the pVL had decreased by more than 1.6 log10 copies/mL.422 In the study reported here, those found to have R5 virus only at screening had a decrease in R5 pVL that mirrored the decrease in the overall pVL.  However, in the instances of non-R5 virus at screening the overall pVL stayed relatively constant with a modest median pVL decrease of approximately 0.2 log10 copies/mL.  When the R5 populations were isolated in  133 these samples, the R5 variants had decreased by a median of 1.27 log10 copies/mL whereas the median percentage of non-R5 virus had nearly doubled.  These results suggest that as the R5 variants were inhibited in the presence of maraviroc, opportunistic outgrowth allowed the expansion of the non-R5 population such that the median pVL of the entire viral population remained relatively constant.  This observation is in agreement with the findings of a similar study conducted by our group where four individuals shown to have non-R5 virus exhibited an increase of 71% in the median percentage of non-R5 virus after 11 days of maraviroc exposure.424 The lesser degree of dominance of the non-R5 variants at Day 7 in the population observed here may be due to the difference in the duration of maraviroc exposure, or any potential antiviral activity in the failing background regimen. Under maraviroc “monotherapy”, the non-R5 variants identified as the predominate variant in viral populations at Day 7 were also found to be present at Day 0, though as minority species.  This observation demonstrates that the tropism shifts in viral populations are due to the outgrowth of preexisting non-R5 variants.  These findings have previously been reported in various studies that have shown preexisting non-R5 minority variants expand to predominate a viral population in as few as six days.276,414,424 The non-R5 variants predominating at Day 7 in each of the non-R5 populations had extremely low g2p FPR values (FPR<2), similar to the findings described by our group previously.424 According to our results, the viral fitness in the presence of maraviroc begins to increase quite dramatically at a g2p FPR of approximately two.  In fact, viral fitness increased roughly 20% with each FPR unit decrease such that variants with a g2p FPR of zero were the most fit in the viral population in the presence of maraviroc.  134 There were four instances of tropism switch occurring within this study population after eight days of maraviroc exposure.  In two cases a tropism switch from R5 to non-R5 was observed and conversely, two instances from non-R5 to R5.  When a tropism switch from non-R5 to R5 was observed, the most prevalent variants for each viral population were R5.  However, the cumulative prevalence of non-R5 variants at Day 0 was greater than the 2% non-R5 cutoff point used to infer the tropism of the viral population.  These tropism switches may suggest that maraviroc activity may extend to lower g2p FPR than previously assumed.  Or perhaps indicate that despite an observed increase in viral fitness as g2p FPR decreases, there may still be non-R5 variants with a low fitness either in the presence of maraviroc or not.  In each of the patients experiencing tropism switch, the phenotypic assays correctly predicted the direction of virological response to maraviroc.  However, the sample size of this study is too small to draw any conclusions regarding assay performance.  Interestingly, in both cases of tropism switch from non-R5 to R5, the prevalence of the variant eliciting the non-R5 result would have been below the detection threshold of population-based sequencing and the sample would likely have been called R5. In each of the two cases of tropism switch from R5 to non-R5, the non-R5 viral population accounted for <1% at Day 0, below the 2% cutoff point used to infer a non-R5 population.  Yet, after eight days of maraviroc add-on therapy these small non-R5 viral populations shifted to account for >20% of the viral population.  Interestingly, these were not the only viral populations to contain non-R5 variants at <1%, yet they were the only instances of extreme minority virus (<1%) with a g2p FPR <1.0.  This may indicate that, in concordance with the observation that viral fitness increases markedly as g2p FPR  135 decreases below two, despite extremely low prevalence these non-R5 variants (g2p FPR<1.0) were at a selective advantage in the presence of maraviroc.424 However, Knapp et al. previously described the observation of inconsistent detection and increased variability of minority non-R5 variants, particularly when variants fall below a prevalence of 2-3%.412 Using 47 replicates of a non-homogeneous sample sequenced on the 454 GX-FLX, it was found that extreme minority non-R5 variants were inconsistently observed, and when detected there was a high amount of variability in the non-R5 variants present between sample replicates.  These inconsistencies were inversely correlated with both sample pVL and prevalence.412  The tropism switches observed in this study could be the result of incomplete sampling of the viral population leading to unidentified and unaccounted for non-R5 variants at baseline.  Similarly, it may be due to the inconsistent detection of extreme minority variants using the 454 platform, and current shortcomings in the ability to distinguish between true minority variants at extremely low prevalence and sequencing artifacts.  There are a number of limitations inherent with this study, including the small sample size, the short duration of maraviroc exposure and the potential of incomplete sampling or amplification errors using the current laboratory methods for both genotypic and phenotypic assays discussed.  In order to account for sampling and PCR error we extracted and amplified viral RNA in triplicate.  A further limitation is the 454 pyrosequencing reaction itself, which is prone to insertion and deletion errors, particularly at homopolymeric regions.  Nucleotide incorporation during the sequencing reaction is detected as light intensity, and in the event of nucleotide repetition this light intensity is cumulative.  However, the ability to distinguish the number of times a nucleotide is  136 incorporated using this accumulated light intensity becomes increasingly difficult, often leading to one more or one less nucleotide in the homopolymeric string.  The presence of these homopolymer errors make it difficult to differentiate between a sequencing artifact and a true extreme minority variant.428,429 In order to attempt to overcome this limitation and any potential cross-contamination between wells, sequences were removed from the analysis if they were observed fewer than fifteen times.  As well, sequences were removed if the sequence nucleotide length was aberrant, which may indicate single base insertions or deletions.  As well, in attempt to reduce the number of false positives caused by sequencing artifacts, tropism inference was based on a cutoff of 2% where a non-R5 viral population must be composed of at least 2% non-R5 variants (FPR ≤3.5).  As tropism is determined by the algorithmic interpretation of the V3 loop sequence with the application of a cutoff to a generated score, it could be possible that the cutoff point is more conservative or more lenient than may be necessary.  The results of this study support the idea that with improved understanding of variants with low g2p FPR and their response to maraviroc, cutoff points may yet be further optimized.  It may become increasingly important to better understand and identify the dual/mixed variants hovering between the non-R5 and R5 cutoff point from a clinical perspective.  In addition, some patients in this study had remained on a failing antiretroviral regimen at the time of introducing maraviroc.  Though they were previously failing therapy with high pVL, there may have been some remnant antiretroviral activity, however, there was no control for such background effects in this study.  137 In conclusion, this study provides further support that maraviroc should not be prescribed as part of an anti-HIV regimen for those found to have non-R5 tropic virus, as the inhibition of R5 variants promotes the selection of pre-existing minority non-R5 variants.  For this reason, it is important to pay attention to minority species that are at the limits of detection of the current tropism detection methods.  As well, the currently used g2p FPR cutoff point for non-R5 variants may more appropriately be a dual/mixed variant cutoff point suggesting that maraviroc may have a more potent effect on variants with a lower g2p FPR than previously thought.  This may also indicate that an improved understanding of low g2p FPRs and response to maraviroc will help to further optimize the cutoff points for genotypic tropism assays as newer sequencing methods become available. 138 Chapter Six: General Discussion 6.1 Summary of Thesis This thesis focuses on viral response to the CCR5-antagonist, maraviroc and two genotypic methods used to identify HIV-1 tropism in order to predict maraviroc-associated virological response.  In doing so it has also provided additional insight into the intricacies of R5 and non-R5 HIV-1 variants under the selective pressures of maraviroc, and how we can better differentiate between them for clinical use. With the approval of the CCR5 antagonist maraviroc for clinical use, treatment guidelines now state that individuals considering the addition of maraviroc to their antiretroviral regimen require a viral tropism test so as to infer the clinical utility of the drug on a case-by-case basis.  The ideal tropism test is one that can provide clinically reliable results in a timely manner at an affordable cost.  At the time of the maraviroc clinical trials highlighted in Chapters Two and Three, the only clinically available tropism assay was the original phenotypic Trofile assay.  In fact, Trofile was the tropism screening assay used for enrollment into the clinical trials of maraviroc: MOTIVATE 1, MOTIVATE 2, A4001029 and MERIT.  The phenotypic Trofile Assay, now replaced by the optimized Enhanced Sensitivity Trofile Assay (ESTA), continues to be a useful standard in tropism testing despite limitations including: turnaround time, required sample volume, location and cost.  At the onset of this thesis genotypic approaches had been found to be inadequate in predicting HIV-1 tropism, demonstrating poor sensitivity and specificity.  However  139 partly as a result of the work outlined in this thesis, genotypic tropism testing approaches have become more common. The studies presented in Chapters Two and Three detail the validation of a new population-based sequencing method coupled with the geno2pheno tropism bioinformatics algorithm (g2p) for inferring viral tropism from V3 loop sequences.  Using the well-characterized study populations from the maraviroc Phase III clinical trials, genotyping was shown to be broadly similar to the original and Enhanced Trofile assays in predicting virological response to maraviroc.  The results of retrospectively rescreening the maraviroc clinical trials study populations for viral tropism using V3 genotyping provided sufficient validation of its clinical utility.  As standard “Sanger” sequencing is a widely available technology, the validation of the described population-based tropism assay has provided an affordable and reliable method of predicting virological response to maraviroc. The development of new “next generation” sequencing technologies such as 454 “deep” sequencing, have made it possible to examine sequence variation in patient viral populations by enabling the generation of thousands of sequence reads for an individual sample.  This encompassing snapshot of the viral population makes possible the identification of species existing at a minority within a population.  When studying HIV the ability to identify minority variants, the majority of which go undetected by standard sequencing methods, can be very useful particularly when studying characteristics like drug resistance polymorphisms and viral tropism within a viral population.  In Chapters Four and Five, retrospective tropism analyses using next generation 454 “deep” sequencing  140 are described, where patient populations were given maraviroc in addition to a failing antiretroviral regimen despite predicted viral tropism.  These study populations are of particular value because they, in essence, represent maraviroc mono-therapy treatment groups. The application of next generation 454 “deep” sequencing to study HIV-1 patient samples characterized by maraviroc add-on therapy and minority non-R5 variants provided a means of examining viral population dynamics as directed by viral tropism and maraviroc exposure.  As described in Chapters Four and Five, the presence of pretreatment non-R5 variants in a viral population being treated with maraviroc causes a shift in viral tropism such that non-inhibited non-R5 variants thrive under maraviroc treatment conditions.  When viral sequences were stratified according to g2p FPR, it was clear that viruses with an extremely low FPR were more fit in the presence of maraviroc. 6.1.1 Validation of a Population-Based Genotypic Tropism Assay for Clinical Use Chapters Two and Three discuss two studies designed to assess the potential utility of a population-based sequencing assay to infer viral tropism and predict the virological response to maraviroc.  These studies remain the largest attempts to prove the clinical utility of a population-based genotypic tropism assay including patients with both non-R5 and R5 viruses treated with a CCR5 antagonist.  Using the patient populations from the four large-scale, multi-national, randomized controlled Phase III clinical trials of maraviroc, samples were retrospectively tested for viral tropism using the aforementioned population-based methods.  Enrollment into the trials was based on the results of the original Trofile  141 assay, such that only individuals found to have R5-tropic HIV at screening were enrolled in the MOTIVATE and MERIT trails, and a subset of those found to have non-R5 virus enrolled in the sister safety study, A4001029.  Retrospectively, V3 loop sequences were generated, processed and interpreted using the g2p tropism bioinformatics algorithm.  In order to address the limit of detection of population-based sequencing and broaden the sampling of variants within a population, samples were extracted and processed in triplicate.  The genotypic rescreening tropism results were compared to the clinical outcome results from the respective trials. The V3 population-based sequencing assay was shown to predict virological response to maraviroc in treatment-experienced patients, performing comparably to Trofile exhibiting 90% concordance between the two assays (Chapter 2).  The study confirmed that those found to have R5 virus as determined by genotypic methods tended to respond well to maraviroc whereas those found to have non-R5 virus generally experience a poor virological response.  It is important to note that a non-R5 result does not characterize the viral population as purely X4; it can include dual and mixed virus populations, which affect the virological response to maraviroc.  In those with discordant tropism results between testing methods, the virological response to maraviroc was intermediate.  This observation may indicate viral populations composed of a mixture of R5, X4 and/or dual variants being sampled differently between the two assays. The samples from the MERIT trial were rescreened using both the V3 population-based sequencing assay as well as the Enhanced Sensitivity Trofile Assay (Chapter 3).  According to the retrospective genotypic tropism results of the MERIT screening samples,  142 roughly 8% of the participants of the MERIT study would have been excluded from study enrollment based on the presence of non-R5 virus.  When considering the virological response of these patients, those found to harbour non-R5 HIV-1 sustained little response to maraviroc yet efavirenz elicited a favourable response, confirming the presence of non-responsive non-R5 virus.  When the rescreening results of genotyping and ESTA were compared, it was evident that the discordances observed were associated with populations likely composed of both R5 and non-R5 variants.  As with the MOTIVATE rescreening results, these viral populations were likely characterized by non-R5 variants near the detection threshold of either assay.  This reasoning is supported by the results of a subsequent study designed to elucidate the mechanism(s) of failure in the MOTIVATE trials, which concluded that reduced virological response to maraviroc may in part be due to the outgrowth of preexisting minority X4 variants.276 Performing broadly similar to both Trofile and ESTA, the results of these studies confirmed the practical utility of population-based V3 sequencing to infer viral tropism when considering maraviroc in an antiretroviral regimen. Based on a reliable technology, V3 population-based genotyping is comparatively inexpensive, requires little time to process samples and most importantly, it can be performed in any laboratory equipped with a standard sequencing platform.  Comparable to the Trofile Assays in tropism inference, and having been released into the public domain based on the results of the studies presented in Chapters Two and Three, the V3 population-based assay has been adopted as a routine tropism assay in many parts of the world.  143 6.1.2 The Effects of Maraviroc on Non-R5 Virus Populations Chapters Four and Five discuss the results of two in-depth investigations exploring the selective pressures evoked by maraviroc on non-R5 HIV-1 populations.  Patient samples were obtained from a small subset of the ATHENA cohort in the Netherlands and a second subset from the Maraviroc Clinical Test trial in Spain.  In both studies, patients had been exposed to short-term maraviroc add-on therapy during a period of treatment failure, in essence creating a surrogate maraviroc mono-therapy.  In order to monitor the virological response to maraviroc under the conditions of a failing regimen, samples were collected in a longitudinal fashion at multiple time points.  Retrospectively, samples collected at each time point were used to generate V3 loop sequences for tropism inference, and the effects of maraviroc evaluated as a function of the depression of the R5 populations and the expansion of the non-R5 populations. As maraviroc successfully inhibits the targeted R5 HIV population there is less competition within the physical space of the viral population, allowing the expansion of the non-inhibited non-R5 variants, leading to a shift in viral tropism.  As was shown with the Maraviroc Clinical Test study population described in Chapter Five, this tropism shift can occur in as few as eight days.  The four patients from the ATHENA cohort were exposed to maraviroc for a median of 53 days, however, by day 21 a shift in the viral population was evident such that virus most genetically associated with X4 accounted for more than 90% of sample virus.  Ten patients from the Spanish Maraviroc Clinical Test trial were identified as having non-R5 virus.  After a brief eight days of maraviroc exposure the median percent non-R5 virus had increased 54% in the non-R5 viral populations.  Variants with an extremely low FPR (< 2%) were seemingly unaffected by maraviroc, becoming the  144 dominant viral strains.  This was not surprising as this group of viruses represents the variants that are most genotypically characteristic of X4 variants.  However, it was perhaps surprising that selection of non-R5 viruses in the presence of maraviroc was so clearly delineated at extremely low g2p FPR (<3%).  Viral fitness markedly improved be below g2p FPR 3%, such that viral fitness in the presence of maraviroc doubled at g2p FPR 3% and tripled at g2p FPR 2%.  These observed patterns were consistent between the two patient populations despite the difference in the length of maraviroc exposure.  Despite the antiviral efforts of maraviroc, the plasma viral load remained relatively constant during the course of maraviroc add-on therapy in both study populations.  This lack of virological response was caused by a relatively equivalent rate of non-R5 expansion and R5 suppression. Though neither study population was particularly large, it is evident that the selective pressures of maraviroc in the absence of a fully functioning background therapy promote the outgrowth of variants whose sequence is most characteristic of X4 virus.  The detection limit of next generation 454 “deep” sequencing allowed us to identify extreme X4 minority variants, including those below the currently applied g2p FPR cutoff points.  In the majority of patients found to have non-R5 virus at baseline, a population inversion was observed such that an expanding non-R5 variant population overtook the once dominant R5 population.  This inversion was clearly visible in 60% of individuals screened as having non-R5 virus from the Maraviroc Clinical Test trial sample set.  The non-R5 variants present at the highest prevalence at the final time point were observed at baseline, in some cases as extreme minority variants occupying <1% of the viral population.  By exposing patients to maraviroc under the conditions of therapy failure, the effects of maraviroc on  145 the viral population can be evaluated without the confounding factors associated with a treatment backbone. The studies described in Chapters Four and Five complement each other, strengthening the observation that maraviroc, when in the presence of non-R5 virus, will select for viruses with extremely low g2p FPR.  Whether these results imply potential antiretroviral activity of maraviroc against variants falling to the obscure “dual-mixed” class or whether the delineation between X4 and R5 virus is at a lower g2p FPR than currently thought is uncertain.  However, these studies do reinforce the significance of tropism testing prior to initiating maraviroc as part of an antiretroviral treatment regimen and suggest that the antiretroviral activity of maraviroc may extend to a broader range of HIV-1 variants than previously suspected. 6.2 Contributions to the Field of HIV Population-based sequencing has been the primary technology used in molecular studies for decades and remains a favoured and reliable tool despite the introduction of next generation sequencing methods.  It has a longstanding, proven utility in the HIV clinical setting, including screening for drug resistance polymorphisms and host genetic factors influencing treatment responses.  The availability and wide distribution as well as the relatively inexpensive and simple processing methods make the use of population-based sequencing an appealing option for determining viral tropism clinically. A functional, clinical genotypic tropism assay designed to predict virological response to maraviroc was developed and validated as described in Chapters Two and  146 Three.  The validation of this population-based tropism assay led to its release in the public domain allowing laboratories worldwide to access the methodology.430  This assay was shown capable of inferring co-receptor use and predicting virological response to maraviroc and maraviroc-induced tropism changes.  In fact, it performed similarly to other tropism testing methods, but provides several advantages including: smaller sample volume requirements, shorter turnaround time, decreased cost and accessibility.  Subsequent optimization of the population-based assay identified two primary cutoff points that seemingly indicate a range of g2p FPR values associated with diminished maraviroc activity and a minimum g2p FPR value associated with maraviroc-induced virological suppression.  Identification of these clinically relevant g2p FPR values has improved the definition of non-R5 and R5 HIV variants and their clinical significance in maraviroc treatment. The studies conducted in Chapters Four and Five are further examples of retrospectively rescreening study samples using genotypic methods, and evaluating the performance of the genotypic method using maraviroc response data.  Here, patients with non-R5 virus were exposed to maraviroc under therapy failure conditions, providing a unique and informative study population.  Creating surrogate conditions of maraviroc mono-therapy, these studies describe the evaluation and interpretation of the antiretroviral effects of maraviroc on HIV-1 variants without the influence of any potential confounding factors associated with background therapy.  The next generation 454 “deep” sequencing data collected in these studies is invaluable, allowing for the analysis of selective pressures exerted exclusively by maraviroc on individual HIV-1 variants ranging from “extremely R5” to “extremely X4”.  Such studies better our understanding of the in vivo effects of  147 maraviroc on mixed viral populations and how to better interpret genotypic tropism results and better integrate the CCR5 antagonist drug class in the treatment of HIV-1 6.3 Future Directions This thesis focused on the development of screening assays to determine the clinical utility of CCR5 antagonists.  Viral entry is an appealing antiretroviral target, preventing the infection of new cells, and stopping the viral life cycle before it can begin.  Maraviroc was successfully introduced as the first approved CCR5-antagonist with demonstrated efficiency and good tolerability in the maraviroc clinical trials.  However, maraviroc is not widely used at present.  This is likely due to the range and availability of other antiretroviral drugs and drug classes, and possibly due to the necessity for tropism screening prior to prescription. Few new CCR5 antagonists appear to be in development, with the exception of cenicriviroc, which is in Phase IIb clinical trials.  More excitingly, because R5 viruses cause the majority of new HIV-1 infections, maraviroc has become a focus for the improvement of preventative strategies including pre-exposure prophylaxis, microbicides and vaginal rings.431–433 There are a number of reasons to exploit CCR5 antagonists, like maraviroc, for this purpose including the mechanism of action as well as good tolerability and safety profiles making this drug class a more desirable option for regular preventative use. Treatment and prevention strategies continue to act against HIV by introducing a physical means to inhibit the activity of the virus.  However, curative and gene therapy strategies focusing on the expression of the human CCR5 co-receptor are being explored.   148 To date, the greatest HIV success story has been that of the “Berlin Patient” who underwent a full blood transfusion for treatment of acute myeloid leukemia in 2007.434 Tested as having an R5-dominant HIV-1 population (<3% non-R5 by next generation sequencing), doctors took this opportunity to attempt a cure by identifying a blood donor match homozygous for the Δ32 mutation.434,435 Having survived this high-risk procedure the Berlin Patient continues to suppress HIV replication in the absence of antiretroviral therapy.436,437 At present such a cure seems unlikely in clinical practice, as it is associated with a high mortality rate, and there is a limited donor pool in addition to other logistical barriers including cost.  Attempts to recreate this case have been made, however, patients did not survive longer than one year.438 Perhaps in the future this may become a clinically viable option at which point the identification of pre-transplantation non-R5 variants will require a sensitive tropism assay to ensure success.  In the short-term, a more practical approach may be the use of gene therapy strategies to prevent infection. The use of zinc finger nucleases is currently the most relevant form of gene therapy in HIV medicine.  Zinc finger nucleases are artificial restriction enzymes that can be designed to target specific DNA sequences and subsequently remove the sequence fragment from the DNA strand.  This technique is being used experimentally to engineer homozygous Δ32 cells, devoid of CCR5 co-receptor expression at the cell surface.439–441 In essence, zinc finger nucleases can be used to create cells incapable of infection with R5 virus by mimicking the naturally occurring CCR5 Δ32 mutation.  More recently they have explored the possibility of removing a segment from the gene encoding CXCR4 to inhibit X4 virus, however the effects of preventing the expression of the CXCR4 co-receptor is less understood.441 As new treatment and prevention strategies exploit entry inhibition at the  149 CCR5 co-receptor, tropism testing is becoming increasingly important.  In order to ensure the success of entry inhibition strategies tropism testing will remain a valuable tool for the long-term both clinically and in the development of new anti-HIV strategies. The results of the two studies presented in Chapters Four and Five indicate that maraviroc is effective against a broader range of HIV variants than previously thought.  In both studies, maraviroc selected for “extremely” X4 variants with a g2p FPR <2%.  This is below all currently applied cutoff points for genotypic tropism inference.  However, due to small sample sizes the relevance of this observation has yet to be determined, thus leaving room for future studies to further explore and optimize g2p FPR cutoff points if deemed necessary.  Such studies may better characterize HIV tropism and help to better define the elusive division between R5, X4 and dual HIV variants.  The continued application of CCR5-targeting anti-HIV strategies will continue to rely on tropism inference.  For this reason and the development of new genotyping technologies, algorithm cutoff points will continually need to be optimized in order to improve genotypic tropism inference tools for clinical application. During the course of this thesis, DNA sequencing technologies have changed dramatically.  As sequencing platforms become increasingly sensitive, with simpler methods and optimized chemistry, genetic clinical tools have become more widely applied.  Population-based sequencing, as used in Chapter One and Two, will remain a dominant technique in molecular virology laboratories in the short term.  It has superior sequence read length and a simple design making it a valuable tool for small laboratories, small-scale experiments, or for assays when sequencing difficult regions of the genome.  As an  150 affordable and reliable option, its use has the potential to flourish in resource-poor settings.  With the increasing availability of antiretroviral compounds globally, the demand for clinical monitoring in such settings has increased. However, in North America, population-based sequencing may be phased out of clinical use as drug resistance and tropism assays are designed and validated for “deep” sequencing platforms.  At present large HIV clinical laboratories like Quest Diagnostics in the United States and the British Columbia Centre for Excellence in HIV/AIDS in Canada, use “deep” sequencing in their clinical diagnostics for HIV tropism.  But the current rate of development in the area of sequencing technology and chemistry is rapid.  Over the course of approximately ten years 454 “deep” sequencing has gone from being revolutionary to being abandoned as newer, more efficient platforms are introduced.  The discontinuation of 454 products by mid-2016 has led to the development of HIV clinical assays using other “deep” sequencing platforms. With a broad range of applications including the detection of drug resistance polymorphisms and viral tropism, genotyping options are becoming increasingly affordable, applicable and accessible, supporting the push for global equality in HIV treatment.  In addition, efforts being made to develop and improve next generation sequencing platforms can and will continue to offer new opportunities in genetic studies, clinical monitoring and HIV treatment.  The read depth of these technologies can provide valuable insight into patient viral populations by identifying individual virus variants and their susceptibility to treatment, perhaps somewhat augmenting what we consider to be “personalized medicine” in HIV treatment.  151 In addition to the clinical applications described above, research using sequencing technologies continues to illuminate aspects of the HIV-1 genome, envelope proteins and entry mechanism as well as the dynamics within the diverse viral populations.  Such research may help to improve our ability to delineate between variants that will and those that will not respond to maraviroc, or may provide a better understanding of the selective pressures elicited by maraviroc.  It is also fundamental in the development of new antiretrovirals and potential curative strategies.  In any case, further research using sequencing techniques to elucidate the mechanisms of HIV-1 infection will easily translate into clinical application as new treatment strategies and subsequent curative strategies.  The continuous progression in the fields of genetics and bioinformatics betters our understanding of HIV-1 cell entry and how we can exploit it to effectively control infection. 6.4 Closing Remarks Though the population-based tropism assay described in Chapters Two and Three of this thesis is not as specific, nor as sensitive as next generation sequencing technologies, based on the results of studies presented here it remains a reliable and affordable tool, capable of predicting the virological response to maraviroc.  However, the use of next generation “deep” sequencing has provided valuable insight into the population dynamics of mixed HIV-1 under maraviroc-induced selection pressures.  The results presented in Chapters Four and Five, confirm that tropism inference prior to initiating maraviroc is necessary though these results may insinuate maraviroc is effective against a broader range of HIV-1 variants than previously thought.  The field of clinical genotyping is developing rapidly as current technologies and chemistries are being optimized, and new ones being  152 explored.  As genotyping continues to improve in the HIV clinical setting, so too does our ability to detect and interpret HIV-1 tropism and the clinical utility of maraviroc. 153 Bibliography 1. Gottlieb MS, Schanker HM, Fan PT, Saxon A, Weisman JD, Pozalski I. Pneumocystis pneumonia --- Los Angeles. MMWR. Morb. Mortal. Wkly. Report. 30, 250–2 (1981). 2. Barré-Sinoussi, F. et al. Isolation of a T-lymphotropic retrovirus from a patient at risk for acquired immune deficiency syndrome (AIDS). Science 220, 868–71 (1983). 3. Broder, S. & Gallo, R. C. A pathogenic retrovirus (HTLV-III) linked to AIDS. N. Engl. J. Med. 311, 1292–7 (1984). 4. Gallo, R. C. & Montagnier, L. The chronology of AIDS research. Nature 326, 435–6 (1987). 5. Blattner, W., Gallo, R. C. & Temin, H. M. HIV causes AIDS. Science 241, 515–6 (1988). 6. UNAIDS. The Gap Report. (2013). 7. UNAIDS. Global Report: UNAIDS Report on the Global AIDS Epidemic 2013. (2013). 8. UNAIDS. 90-90-90: An Ambitious Treatment Target to Help End the AIDS Epidemic. (2014). Accessed September 24, 2015, at: <http://www.unaids.org/sites/default/files/media_asset/90-90-90_en_0.pdf> 9. Marlink, R. et al. Reduced rate of disease development after HIV-2 infection as compared to HIV-1. Science 265, 1587–90 (1994). 10. Popper, S. J. et al. Lower human immunodeficiency virus (HIV) type 2 viral load reflects the difference in pathogenicity of HIV-1 and HIV-2. J. Infect. Dis. 180, 1116–21 (1999). 11. Hemelaar, J. The origin and diversity of the HIV-1 pandemic. Trends Mol. Med. 18, 182–92 (2012). 12. Ward, M. J., Lycett, S. J., Kalish, M. L., Rambaut, A. & Leigh Brown, A. J. Estimating the rate of intersubtype recombination in early HIV-1 group M strains. J. Virol. 87, 1967–73 (2013). 13. Plantier, J.-C. et al. A new human immunodeficiency virus derived from gorillas. Nat. Med. 15, 871–2 (2009).  154 14. Vallari, A. et al. Confirmation of putative HIV-1 group P in Cameroon. J. Virol. 85, 1403–7 (2011). 15. Robertson, D. L. et al. HIV-1 nomenclature proposal. Science 288, 55–6 (2000). 16. Hemelaar, J., Gouws, E., Ghys, P. D. & Osmanov, S. Global trends in molecular epidemiology of HIV-1 during 2000-2007. AIDS 25, 679–89 (2011). 17. Ariën, K. K. et al. The replicative fitness of primary human immunodeficiency virus type 1 (HIV-1) group M, HIV-1 group O, and HIV-2 isolates. J. Virol. 79, 8979–90 (2005). 18. Los Alamos National Laboratory. HIV Sequence Database. Accessed September 24, 2015, at: <http://www.hiv.lanl.gov/> 19. Abecasis, A. B. et al. HIV-1 subtype distribution and its demographic determinants in newly diagnosed patients in Europe suggest highly compartmentalized epidemics. Retrovirology 10, 7–19 (2013). 20. Geretti, A. M. HIV-1 subtypes: epidemiology and significance for HIV management. Curr. Opin. Infect. Dis. 19, 1–7 (2006). 21. Hemelaar, J. Implications of HIV diversity for the HIV-1 pandemic. J. Infect. 66, 391–400 (2013). 22. Huet, T., Cheynier, R., Meyerhans, A., Roelants, G. & Wain-Hobson, S. Genetic organization of a chimpanzee lentivirus related to HIV-1. Nature 345, 356–9 (1990). 23. Korber, B. et al. Timing the ancestor of the HIV-1 pandemic strains. Science 288, 1789–96 (2000). 24. Worobey, M. et al. Direct evidence of extensive diversity of HIV-1 in Kinshasa by 1960. Nature 455, 661–4 (2008). 25. Zhu, T. et al. An African HIV-1 sequence from 1959 and implications for the origin of the epidemic. Nature 391, 594–7 (1998). 26. Gao, F. et al. Origin of HIV-1 in the chimpanzee Pan troglodytes troglodytes. Nature 397, 436–41 (1999). 27. Keele, B. F. et al. Chimpanzee reservoirs of pandemic and nonpandemic HIV-1. Science 313, 523–6 (2006). 28. Hahn, B. H., Shaw, G. M., De Cock, K. M. & Sharp, P. M. AIDS as a zoonosis: scientific and public health implications. Science 287, 607–14 (2000).  155 29. Peeters, M. et al. Risk to human health from a plethora of simian immunodeficiency viruses in primate bushmeat. Emerg. Infect. Dis. 8, 451–7 (2002). 30. Aghokeng, A. F. et al. Extensive survey on the prevalence and genetic diversity of SIVs in primate bushmeat provides insights into risks for potential new cross-species transmissions. Infect. Genet. Evol. 10, 386–96 (2010). 31. Wolfe, N. D. et al. Naturally acquired simian retrovirus infections in central African hunters. Lancet 363, 932–7 (2004). 32. Patel, P. et al. Estimating per-act HIV transmission risk: a systematic review. AIDS 28, 1509–19 (2014). 33. Chudy, M. et al. Blood screening nucleic acid amplification tests for human immunodeficiency virus Type 1 may require two different amplification targets. Transfusion 52, 431–9 (2012). 34. Voronin, Y. et al. HIV monoclonal antibodies: a new opportunity to further reduce mother-to-child HIV transmission. PLoS Med. 11, e1001616 (2014). 35. Degenhardt, L. et al. Prevention of HIV infection for people who inject drugs: why individual, structural, and combination approaches are needed. Lancet 376, 285–301 (2010). 36. Aung, T., Paw, E., Aye, N. M. & McFarland, W. Coverage of HIV prevention services for female sex workers in seven cities of Myanmar. AIDS Behav. 18 Suppl 1, S37–41 (2014). 37. Joseph, N. M., Elan, S., Vadivu, S. & Kanungo, R. Needlestick injuries among healthcare workers of a tertiary care hospital in South India. Infect. Control Hosp. Epidemiol. 35, 103–5 (2014). 38. Yan, H. et al. The Increased Effectiveness of HIV Preventive Intervention among Men Who Have Sex with Men and of Follow-Up Care for People Living with HIV after ‘Task-Shifting’ to Community-Based Organizations: A ‘Cash on Service Delivery’ Model in China. PLoS One 9, e103146 (2014). 39. Nguyen, A. T. et al. Intravenous drug use among street-based sex workers: a high-risk behavior for HIV transmission. Sex. Transm. Dis. 31, 15–9 (2004). 40. Strathdee, S. A. et al. Correlates of injection drug use among female sex workers in two Mexico-U.S. border cities. Drug Alcohol Depend. 92, 132–40 (2008). 41. Narayan, O. & Clements, J. E. Biology and pathogenesis of lentiviruses. J. Gen. Virol. 70, 1617–39 (1989).  156 42. Ratner, L. et al. Complete nucleotide sequence of the AIDS virus, HTLV-III. Nature 313, 277–84 (1985). 43. Muesing, M. A. et al. Nucleic acid structure and expression of the human AIDS/lymphadenopathy retrovirus. Nature 313, 450–8 (1985). 44. Richman DD, Kinchington D, Fitch N, Griffin S, Persaud D, Siliciano R, Goldthorpe Stein SE, Gazzard BG, Powderly WG, Tebas P, Race E, Rizzardi GP, Pantaleo G, Razavi B, Friedland G, Pawinski R, Govender M, G. T. Human Immunodeficiency Virus. (International Medical Press, 2003). 45. Watts, J. M. et al. Architecture and secondary structure of an entire HIV-1 RNA genome. Nature 460, 711–6 (2009). 46. Mervis, R. J. et al. The gag gene products of human immunodeficiency virus type 1: alignment within the gag open reading frame, identification of posttranslational modifications, and evidence for alternative gag precursors. J. Virol. 62, 3993–4002 (1988). 47. Turner, B. G. & Summers, M. F. Structural biology of HIV. J. Mol. Biol. 285, 1–32 (1999). 48. Ganser-Pornillos, B. K., Yeager, M. & Sundquist, W. I. The structural biology of HIV assembly. Curr. Opin. Struct. Biol. 18, 203–17 (2008). 49. Moulard, M. & Decroly, E. Maturation of HIV envelope glycoprotein precursors by cellular endoproteases. Biochim. Biophys. Acta 1469, 121–32 (2000). 50. Wyatt, R. & Sodroski, J. The HIV-1 envelope glycoproteins: fusogens, antigens, and immunogens. Science 280, 1884–8 (1998). 51. Emerman, M. & Malim, M. H. HIV-1 regulatory/accessory genes: keys to unraveling viral and host cell biology. Science 280, 1880–4 (1998). 52. Seelamgari, A. et al. Role of viral regulatory and accessory proteins in HIV-1 replication. Front. Biosci. 9, 2388–413 (2004). 53. Piekna-Przybylska, D. & Bambara, R. A. Human Immunodeficiency Virus Reverse Transcriptase. (Springer New York, 2013). doi:10.1007/978-1-4614-7291-9 54. Sierra, S., Kupfer, B. & Kaiser, R. Basics of the virology of HIV-1 and its replication. J. Clin. Virol. 34, 233–44 (2005). 55. Krebs, F. C., Hogan, T. H., Quiterio, S., Gartner, S. & Wigdahl, B. in HIV Sequence Compendium 2001 (eds. Kuiken, C. et al.) 29–70 (Los Alamos, NM: Theoretical Biology and Biophysics Group, Los Alamos National Laboratory, 2001). Accessed September 24, 2015, at:  157 <http://www.hiv.lanl.gov/content/sequence/HIV/COMPENDIUM/2001/partI/Wigdahl.pdf> 56. Benjamin, J., Ganser-Pornillos, B. K., Tivol, W. F., Sundquist, W. I. & Jensen, G. J. Three-dimensional structure of HIV-1 virus-like particles by electron cryotomography. J. Mol. Biol. 346, 577–88 (2005). 57. Briggs, J. A. G., Wilk, T., Welker, R., Kräusslich, H.-G. & Fuller, S. D. Structural organization of authentic, mature HIV-1 virions and cores. EMBO J. 22, 1707–15 (2003). 58. Zhu, P. et al. Electron tomography analysis of envelope glycoprotein trimers on HIV and simian immunodeficiency virus virions. Proc. Natl. Acad. Sci. U. S. A. 100, 15812–7 (2003). 59. Zhu, P. et al. Distribution and three-dimensional structure of AIDS virus envelope spikes. Nature 441, 847–52 (2006). 60. Roux, K. H. & Taylor, K. A. AIDS virus envelope spike structure. Curr. Opin. Struct. Biol. 17, 244–52 (2007). 61. Landau, N. R., Warton, M. & Littman, D. R. The envelope glycoprotein of the human immunodeficiency virus binds to the immunoglobulin-like domain of CD4. Nature 334, 159–62 (1988). 62. Blumenthal, R., Durell, S. & Viard, M. HIV entry and envelope glycoprotein-mediated fusion. J. Biol. Chem. 287, 40841–9 (2012). 63. Gallo, S. A. et al. The HIV Env-mediated fusion reaction. Biochim. Biophys. Acta 1614, 36–50 (2003). 64. Yu, X., Yuan, X., Matsuda, Z., Lee, T. H. & Essex, M. The matrix protein of human immunodeficiency virus type 1 is required for incorporation of viral envelope protein into mature virions. J. Virol. 66, 4966–71 (1992). 65. Dorfman, T., Mammano, F., Haseltine, W. A. & Göttlinger, H. G. Role of the matrix protein in the virion association of the human immunodeficiency virus type 1 envelope glycoprotein. J. Virol. 68, 1689–96 (1994). 66. Tedbury, P. R. & Freed, E. O. The role of matrix in HIV-1 envelope glycoprotein incorporation. Trends Microbiol. 22, 372–8 (2014). 67. Lee, S.-K., Potempa, M. & Swanstrom, R. The choreography of HIV-1 proteolytic processing and virion assembly. J. Biol. Chem. 287, 40867–74 (2012). 68. Feng, Y. X. et al. HIV-1 nucleocapsid protein induces ‘maturation’ of dimeric retroviral RNA in vitro. Proc. Natl. Acad. Sci. U. S. A. 93, 7577–81 (1996).  158 69. Zhang, Y. & Barklis, E. Nucleocapsid protein effects on the specificity of retrovirus RNA encapsidation [published erratum appears in J Virol 1997 Jul;71(7):5712]. J. Virol. 69, 5716–22 (1995). 70. Baumgärtel, V., Müller, B. & Lamb, D. C. Quantitative live-cell imaging of human immunodeficiency virus (HIV-1) assembly. Viruses 4, 777–99 (2012). 71. Singer, I. I. et al. CCR5, CXCR4, and CD4 are clustered and closely apposed on microvilli of human macrophages and T cells. J. Virol. 75, 3779–90 (2001). 72. Dalgleish, A. G. et al. The CD4 (T4) antigen is an essential component of the receptor for the AIDS retrovirus. Nature 312, 763–7 (1984). 73. Bleul, C. C. et al. The lymphocyte chemoattractant SDF-1 is a ligand for LESTR/fusin and blocks HIV-1 entry. Nature 382, 829–33 (1996). 74. Alkhatib, G. et al. CC CKR5: a RANTES, MIP-1alpha, MIP-1beta receptor as a fusion cofactor for macrophage-tropic HIV-1. Science 272, 1955–8 (1996). 75. Björndal, A. et al. Coreceptor usage of primary human immunodeficiency virus type 1 isolates varies according to biological phenotype. J. Virol. 71, 7478–87 (1997). 76. Berger, E. A., Murphy, P. M. & Farber, J. M. Chemokine receptors as HIV-1 coreceptors: roles in viral entry, tropism, and disease. Annu. Rev. Immunol. 17, 657–700 (1999). 77. Deng, H. et al. Identification of a major co-receptor for primary isolates of HIV-1. Nature 381, 661–6 (1996). 78. Kwong, P. D. et al. Structure of an HIV gp120 envelope glycoprotein in complex with the CD4 receptor and a neutralizing human antibody. Nature 393, 648–59 (1998). 79. Sattentau, Q. J. & Moore, J. P. Conformational changes induced in the human immunodeficiency virus envelope glycoprotein by soluble CD4 binding. J. Exp. Med. 174, 407–15 (1991). 80. Hulme, A. E., Perez, O. & Hope, T. J. Complementary assays reveal a relationship between HIV-1 uncoating and reverse transcription. Proc. Natl. Acad. Sci. U. S. A. 108, 9975–80 (2011). 81. Arhel, N. Revisiting HIV-1 uncoating. Retrovirology 7, 96–105 (2010). 82. Fassati, A. & Goff, S. P. Characterization of intracellular reverse transcription complexes of human immunodeficiency virus type 1. J. Virol. 75, 3626–35 (2001).  159 83. Lanchy, J.-M. et al. Dynamics of the HIV-1 Reverse Transcription Complex during Initiation of DNA Synthesis. J. Biol. Chem. 275, 12306–12 (2000). 84. Warrilow, D., Tachedjian, G. & Harrich, D. Maturation of the HIV reverse transcription complex: putting the jigsaw together. Rev. Med. Virol. 19, 324–37 (2009). 85. Sarafianos, S. G. et al. Structure and function of HIV-1 reverse transcriptase: molecular mechanisms of polymerization and inhibition. J. Mol. Biol. 385, 693–713 (2009). 86. Preston, B., Poiesz, B. & Loeb, L. Fidelity of HIV-1 reverse transcriptase. Science 242, 1168–71 (1988). 87. Roberts, J., Bebenek, K. & Kunkel, T. The accuracy of reverse transcriptase from HIV-1. Science 242, 1171–73 (1988). 88. Miller, M. D., Farnet, C. M. & Bushman, F. D. Human immunodeficiency virus type 1 preintegration complexes: studies of organization and composition. J. Virol. 71, 5382–90 (1997). 89. Haffar, O. K. et al. Two nuclear localization signals in the HIV-1 matrix protein regulate nuclear import of the HIV-1 pre-integration complex. J. Mol. Biol. 299, 359–68 (2000). 90. Bukrinsky, M. I. et al. Active nuclear import of human immunodeficiency virus type 1 preintegration complexes. Proc. Natl. Acad. Sci. U. S. A. 89, 6580–4 (1992). 91. Heinzinger, N. K. et al. The Vpr protein of human immunodeficiency virus type 1 influences nuclear localization of viral nucleic acids in nondividing host cells. Proc. Natl. Acad. Sci. U. S. A. 91, 7311–7315 (1994). 92. Popov, S., Rexach, M., Ratner, L., Blobel, G. & Bukrinsky, M. Viral Protein R Regulates Docking of the HIV-1 Preintegration Complex to the Nuclear Pore Complex. J. Biol. Chem. 273, 13347–13352 (1998). 93. Von Schwedler, U., Kornbluth, R. S. & Trono, D. The nuclear localization signal of the matrix protein of human immunodeficiency virus type 1 allows the establishment of infection in macrophages and quiescent T lymphocytes. Proc. Natl. Acad. Sci. U. S. A. 91, 6992–6996 (1994). 94. Delelis, O., Carayon, K., Saïb, A., Deprez, E. & Mouscadet, J.-F. Integrase and integration: biochemical activities of HIV-1 integrase. Retrovirology 5, 114–126 (2008). 95. Pommier, Y., Johnson, A. A. & Marchand, C. Integrase inhibitors to treat HIV/AIDS. Nat. Rev. Drug Discov. 4, 236–48 (2005).  160 96. Yoder, K. E. & Bushman, F. D. Repair of gaps in retroviral DNA integration intermediates. J. Virol. 74, 11191–200 (2000). 97. Ashe, M. P., Griffin, P., James, W. & Proudfoot, N. J. Poly(A) site selection in the HIV-1 provirus: inhibition of promoter-proximal polyadenylation by the downstream major splice donor site. Genes Dev. 9, 3008–25 (1995). 98. Cullen, B. R. Nuclear mRNA export: insights from virology. Trends Biochem. Sci. 28, 419–24 (2003). 99. Fischer, U. et al. Evidence that HIV-1 Rev directly promotes the nuclear export of unspliced RNA. EMBO J. 13, 4105–12 (1994). 100. Sundquist, W. I. & Kräusslich, H.-G. HIV-1 assembly, budding, and maturation. Cold Spring Harb. Perspect. Med. 2, a006924 (2012). 101. Briggs, J. A. G. & Kräusslich, H.-G. The molecular architecture of HIV. J. Mol. Biol. 410, 491–500 (2011). 102. Bell, N. M. & Lever, A. M. L. HIV Gag polyprotein: processing and early viral particle assembly. Trends Microbiol. 21, 136–44 (2013). 103. Martin-Serrano, J. & Neil, S. J. D. Host factors involved in retroviral budding and release. Nat. Rev. Microbiol. 9, 519–31 (2011). 104. Ross, E. K. et al. Maturation of Human Immunodeficiency Virus Particles Assembled from the gag Precursor Protein Requires In Situ Processing by gag-pol Protease. AIDS Res. Hum. Retroviruses 7, 475–83 (1991). 105. Barré-Sinoussi, F., Ross, A. L. & Delfraissy, J.-F. Past, present and future: 30 years of HIV research. Nat. Rev. Microbiol. 11, 877–83 (2013). 106. McMichael, A. J., Borrow, P., Tomaras, G. D., Goonetilleke, N. & Haynes, B. F. The immune response during acute HIV-1 infection: clues for vaccine development. Nat. Rev. Immunol. 10, 11–23 (2010). 107. Kahn, J. O. & Walker, B. D. Acute human immunodeficiency virus type 1 infection. N. Engl. J. Med. 339, 33–9 (1998). 108. Ribeiro, R. M. et al. Estimation of the initial viral growth rate and basic reproductive number during acute HIV-1 infection. J. Virol. 84, 6096–102 (2010). 109. Veazey, R. S. Gastrointestinal Tract as a Major Site of CD4+ T Cell Depletion and Viral Replication in SIV Infection. Science 280, 427–431 (1998). 110. Brenchley, J. M. et al. CD4+ T cell depletion during all stages of HIV disease occurs predominantly in the gastrointestinal tract. J. Exp. Med. 200, 749–59 (2004).  161 111. Kinloch-de Loës, S. et al. Symptomatic primary infection due to human immunodeficiency virus type 1: review of 31 cases. Clin. Infect. Dis. 17, 59–65 (1993). 112. Deeks, S. G. et al. Immune activation set point during early HIV infection predicts subsequent CD4+ T-cell changes independent of viral load. Blood 104, 942–7 (2004). 113. Lavreys, L. et al. Higher set point plasma viral load and more-severe acute HIV type 1 (HIV-1) illness predict mortality among high-risk HIV-1-infected African women. Clin. Infect. Dis. 42, 1333–9 (2006). 114. Pantaleo, G., Graziosi, C. & Fauci, A. S. New concepts in the immunopathogenesis of human immunodeficiency virus infection. N. Engl. J. Med. 328, 327–35 (1993). 115. Pantaleo, G. et al. HIV infection is active and progressive in lymphoid tissue during the clinically latent stage of disease. Nature 362, 355–8 (1993). 116. Douek, D. C., Picker, L. J. & Koup, R. A. T cell dynamics in HIV-1 infection. Annu. Rev. Immunol. 21, 265–304 (2003). 117. Fauci, A. S. Host factors and the pathogenesis of HIV-induced disease. Nature 384, 529–34 (1996). 118. Deeks, S. G. et al. Towards an HIV cure: a global scientific strategy. Nat. Rev. Immunol. 12, 607–14 (2012). 119. Deeks, S. G., Lewin, S. R. & Havlir, D. V. The end of AIDS: HIV infection as a chronic disease. Lancet 382, 1525–33 (2013). 120. Nasi, M., Pinti, M., Mussini, C. & Cossarizza, A. Persistent inflammation in HIV infection: Established concepts, new perspectives. Immunol. Lett. 161, 184–8 (2014). 121. Battistini, A. & Sgarbanti, M. HIV-1 latency: an update of molecular mechanisms and therapeutic strategies. Viruses 6, 1715–58 (2014). 122. Bangsberg, D. R. et al. Non-adherence to highly active antiretroviral therapy predicts progression to AIDS. AIDS 15, 1181–3 (2001). 123. Hazenberg, M. D. et al. Persistent immune activation in HIV-1 infection is associated with progression to AIDS. AIDS 17, 1881–8 (2003). 124. Friedman-Kien A; Laubenstein L; Marmor M; Hymes K; Green J; Ragaz A; Gottleib J; Muggia F; Demopoulos R; Weintraub M; Kaposi’s sarcoma and Pneumocystis pneumonia among homosexual men--New York City and California. MMWR. Morb. Mortal. Wkly. Rep. 30, 305–8 (1981).  162 125. Fannin S; Gottlieb MS; Weisman JD; Rogolsky E; Prendergast T; Chin J; Friedman-Kien AE; Friedman S; Rothenberg R. A cluster of Kaposi’s sarcoma and Pneumocystis carinii pneumonia among homosexual male residents of Los Angeles and Orange Counties, California. MMWR. Morb. Mortal. Wkly. Rep. 31, 305–7 (1982). 126. Chang, Y. et al. Identification of herpesvirus-like DNA sequences in AIDS-associated Kaposi’s sarcoma. Science 266, 1865–1869 (1994). 127. Antman, K. & Chang, Y. Kaposi’s sarcoma. N. Engl. J. Med. 342, 1027–38 (2000). 128. Lanzavecchia, A., Roosnek, E., Gregory, T., Berman, P. & Abrignani, S. T cells can present antigens such as HIV gp120 targeted to their own surface molecules. Nature 334, 530–2 (1988). 129. Hamilos, D. L. Antigen presenting cells. Immunol. Res. 8, 98–117 (1989). 130. Chougnet, C., Shearer, G. M. & Landay, A. L. The Role of Antigen-presenting Cells in HIV Pathogenesis. Curr. Infect. Dis. Rep. 4, 266–271 (2002). 131. Moris, A. et al. Dendritic cells and HIV-specific CD4+ T cells: HIV antigen presentation, T-cell activation, and viral transfer. Blood 108, 1643–51 (2006). 132. Itano, A. A. & Jenkins, M. K. Antigen presentation to naive CD4 T cells in the lymph node. Nat. Immunol. 4, 733–9 (2003). 133. Zeng, M., Haase, A. T. & Schacker, T. W. Lymphoid tissue structure and HIV-1 infection: life or death for T cells. Trends Immunol. 33, 306–14 (2012). 134. Fahey, J. L. et al. The prognostic value of cellular and serologic markers in infection with human immunodeficiency virus type 1. N. Engl. J. Med. 322, 166–72 (1990). 135. Swingler, S. et al. HIV-1 Nef mediates lymphocyte chemotaxis and activation by infected macrophages. Nat. Med. 5, 997–1003 (1999). 136. Duncan, C. J. A., Russell, R. A. & Sattentau, Q. J. High multiplicity HIV-1 cell-to-cell transmission from macrophages to CD4+ T cells limits antiretroviral efficacy. AIDS 27, 2201–6 (2013). 137. Carr, J. M., Hocking, H., Li, P. & Burrell, C. J. Rapid and efficient cell-to-cell transmission of human immunodeficiency virus infection from monocyte-derived macrophages to peripheral blood lymphocytes. Virology 265, 319–29 (1999). 138. Lavi, E. et al. CXCR-4 (Fusin), a co-receptor for the type 1 human immunodeficiency virus (HIV-1), is expressed in the human brain in a variety of cell types, including microglia and neurons. Am. J. Pathol. 151, 1035–42 (1997).  163 139. Sinclair, E., Gray, F., Ciardi, A. & Scaravilli, F. Immunohistochemical changes and PCR detection of HIV provirus DNA in brains of asymptomatic HIV-positive patients. J. Neuropathol. Exp. Neurol. 53, 43–50 (1994). 140. Jambo, K. C. et al. Small alveolar macrophages are infected preferentially by HIV and exhibit impaired phagocytic function. Mucosal Immunol. 7, 1116–26 (2014). 141. Palucka, K. A., Taquet, N., Sanchez-Chapuis, F. & Gluckman, J. C. Dendritic cells as the terminal stage of monocyte differentiation. J. Immunol. 160, 4587–95 (1998). 142. Ellery, P. J. et al. The CD16+ monocyte subset is more permissive to infection and preferentially harbors HIV-1 in vivo. J. Immunol. 178, 6581–9 (2007). 143. Auffray, C., Sieweke, M. H. & Geissmann, F. Blood monocytes: development, heterogeneity, and relationship with dendritic cells. Annu. Rev. Immunol. 27, 669–92 (2009). 144. Coleman, C. M. & Wu, L. HIV interactions with monocytes and dendritic cells: viral latency and reservoirs. Retrovirology 6, 51–62 (2009). 145. Geijtenbeek, T. B. et al. DC-SIGN, a dendritic cell-specific HIV-1-binding protein that enhances trans-infection of T cells. Cell 100, 587–97 (2000). 146. Heath, S. L., Tew, J. G., Szakal, A. K. & Burton, G. F. Follicular dendritic cells and human immunodeficiency virus infectivity. Nature 377, 740–4 (1995). 147. Kwon, D. S., Gregorio, G., Bitton, N., Hendrickson, W. A. & Littman, D. R. DC-SIGN-mediated internalization of HIV is required for trans-enhancement of T cell infection. Immunity 16, 135–44 (2002). 148. Sattentau, Q. J. & Weiss, R. A. The CD4 antigen: physiological ligand and HIV receptor. Cell 52, 631–3 (1988). 149. Lapham, C. K. et al. Evidence for cell-surface association between fusin and the CD4-gp120 complex in human cell lines. Science 274, 602–5 (1996). 150. Feng, Y., Broder, C. C., Kennedy, P. E. & Berger, E. a. HIV-1 entry cofactor: functional cDNA cloning of a seven-transmembrane, G protein-coupled receptor. Science 272, 872–7 (1996). 151. Dragic, T. et al. HIV-1 entry into CD4+ cells is mediated by the chemokine receptor CC-CKR-5. Nature 381, 667–73 (1996). 152. Cocchi, F. et al. Identification of RANTES, MIP-1 alpha, and MIP-1 beta as the major HIV-suppressive factors produced by CD8+ T cells. Science 270, 1811–5 (1995).  164 153. Feng, Y., Broder, C. C., Kennedy, P. E. & Berger, E. A. HIV-1 entry cofactor: functional cDNA cloning of a seven-transmembrane, G protein-coupled receptor. Science 272, 872–7 (1996). 154. Oberlin, E. et al. The CXC chemokine SDF-1 is the ligand for LESTR/fusin and prevents infection by T-cell-line-adapted HIV-1. Nature 382, 833–5 (1996). 155. Lee, B., Sharron, M., Montaner, L. J., Weissman, D. & Doms, R. W. Quantification of CD4, CCR5, and CXCR4 levels on lymphocyte subsets, dendritic cells, and differentially conditioned monocyte-derived macrophages. Proc. Natl. Acad. Sci. U. S. A. 96, 5215–20 (1999). 156. Tuttle, D. L., Harrison, J. K., Anders, C., Sleasman, J. W. & Goodenow, M. M. Expression of CCR5 increases during monocyte differentiation and directly mediates macrophage susceptibility to infection by human immunodeficiency virus type 1. J. Virol. 72, 4962–9 (1998). 157. Mao, Y. et al. Subunit organization of the membrane-bound HIV-1 envelope glycoprotein trimer. Nat. Struct. Mol. Biol. 19, 893–9 (2012). 158. Starcich, B. R. et al. Identification and characterization of conserved and variable regions in the envelope gene of HTLV-III/LAV, the retrovirus of AIDS. Cell 45, 637–48 (1986). 159. Raska, M. et al. Glycosylation patterns of HIV-1 gp120 depend on the type of expressing cells and affect antibody recognition. J. Biol. Chem. 285, 20860–9 (2010). 160. Wei, X. et al. Antibody neutralization and escape by HIV-1. Nature 422, 307–12 (2003). 161. Wyatt, R. et al. The antigenic structure of the HIV gp120 envelope glycoprotein. Nature 393, 705–11 (1998). 162. Chesebro, B., Wehrly, K., Nishio, J. & Perryman, S. Mapping of independent V3 envelope determinants of human immunodeficiency virus type 1 macrophage tropism and syncytium formation in lymphocytes. J. Virol. 70, 9055–9 (1996). 163. Cheng-Mayer, C., Quiroga, M., Tung, J. W., Dina, D. & Levy, J. A. Viral determinants of human immunodeficiency virus type 1 T-cell or macrophage tropism, cytopathogenicity, and CD4 antigen modulation. J. Virol. 64, 4390–8 (1990). 164. Hwang, S. S., Boyle, T. J., Lyerly, H. K. & Cullen, B. R. Identification of envelope V3 loop as the major determinant of CD4 neutralization sensitivity of HIV-1. Science 257, 535–7 (1992).  165 165. Liu, J., Bartesaghi, A., Borgnia, M. J., Sapiro, G. & Subramaniam, S. Molecular architecture of native HIV-1 gp120 trimers. Nature 455, 109–13 (2008). 166. Chen, B. et al. Structure of an unliganded simian immunodeficiency virus gp120 core. Nature 433, 834–41 (2005). 167. Huang, C. et al. Structure of a V3-containing HIV-1 gp120 core. Science 310, 1025–8 (2005). 168. Jiang, X. et al. Conserved structural elements in the V3 crown of HIV-1 gp120. Nat. Struct. Mol. Biol. 17, 955–61 (2010). 169. Resch, W., Hoffman, N. & Swanstrom, R. Improved success of phenotype prediction of the human immunodeficiency virus type 1 from envelope variable loop 3 sequence using neural networks. Virology 288, 51–62 (2001). 170. Veronese, F. D. et al. Characterization of gp41 as the transmembrane protein coded by the HTLV-III/LAV envelope gene. Science 229, 1402–5 (1985). 171. Tan, K., Liu, J., Wang, J., Shen, S. & Lu, M. Atomic structure of a thermostable subdomain of HIV-1 gp41. Proc. Natl. Acad. Sci. U. S. A. 94, 12303–8 (1997). 172. Weissenhorn, W., Dessen, A., Harrison, S. C., Skehel, J. J. & Wiley, D. C. Atomic structure of the ectodomain from HIV-1 gp41. Nature 387, 426–30 (1997). 173. Epand, R. M. Fusion peptides and the mechanism of viral fusion. Biochim. Biophys. Acta 1614, 116–21 (2003). 174. Haqqani, A. A. & Tilton, J. C. Entry inhibitors and their use in the treatment of HIV-1 infection. Antiviral Res. 98, 158–70 (2013). 175. Wu, L. et al. CD4-induced interaction of primary HIV-1 gp120 glycoproteins with the chemokine receptor CCR-5. Nature 384, 179–83 (1996). 176. Rizzuto, C. D. et al. A conserved HIV gp120 glycoprotein structure involved in chemokine receptor binding. Science 280, 1949–53 (1998). 177. Cormier, E. G. & Dragic, T. The crown and stem of the V3 loop play distinct roles in human immunodeficiency virus type 1 envelope glycoprotein interactions with the CCR5 coreceptor. J. Virol. 76, 8953–7 (2002). 178. Tamamis, P. & Floudas, C. A. Molecular recognition of CXCR4 by a dual tropic HIV-1 gp120 V3 loop. Biophys. J. 105, 1502–14 (2013). 179. Tamamis, P. & Floudas, C. A. Molecular recognition of CCR5 by an HIV-1 gp120 V3 loop. PLoS One 9, e95767 (2014).  166 180. Pancera, M. et al. Structure and immune recognition of trimeric pre-fusion HIV-1 Env. Nature 514, 455–461 (2014). 181. Chan, D. C., Fass, D., Berger, J. M. & Kim, P. S. Core structure of gp41 from the HIV envelope glycoprotein. Cell 89, 263–73 (1997). 182. Tran, E. E. H. et al. Structural mechanism of trimeric HIV-1 envelope glycoprotein activation. PLoS Pathog. 8, e1002797 (2012). 183. Berger, E., Doms, R. & Fenyö, E. A new classification for HIV-1. Nature 391, 4408 (1998). 184. Doranz, B. J. et al. A dual-tropic primary HIV-1 isolate that uses fusin and the beta-chemokine receptors CKR-5, CKR-3, and CKR-2b as fusion cofactors. Cell 85, 1149–58 (1996). 185. Trouplin, V. et al. Determination of coreceptor usage of human immunodeficiency virus type 1 from patient plasma samples by using a recombinant phenotypic assay. J. Virol. 75, 251–9 (2001). 186. Koot, M. et al. HIV-1 biological phenotype in long-term infected individuals evaluated with an MT-2 cocultivation assay. AIDS 6, 49–54 (1992). 187. Schuitemaker, H. et al. Biological phenotype of human immunodeficiency virus type 1 clones at different stages of infection: progression of disease is associated with a shift from monocytotropic to T-cell-tropic virus population. J. Virol. 66, 1354–60 (1992). 188. van’t Wout, A. B. et al. Macrophage-tropic variants initiate human immunodeficiency virus type 1 infection after sexual, parenteral, and vertical transmission. J. Clin. Invest. 94, 2060–7 (1994). 189. Moore, J. P., Kitchen, S. G., Pugach, P. & Zack, J. A. The CCR5 and CXCR4 coreceptors--central to understanding the transmission and pathogenesis of human immunodeficiency virus type 1 infection. AIDS Res. Hum. Retroviruses 20, 111–26 (2004). 190. Regoes, R. R. & Bonhoeffer, S. The HIV coreceptor switch: a population dynamical perspective. Trends Microbiol. 13, 269–77 (2005). 191. Liu, R. et al. Homozygous defect in HIV-1 coreceptor accounts for resistance of some multiply-exposed individuals to HIV-1 infection. Cell 86, 367–77 (1996). 192. Dean, M. et al. Genetic restriction of HIV-1 infection and progression to AIDS by a deletion allele of the CKR5 structural gene. Hemophilia Growth and Development Study, Multicenter AIDS Cohort Study, Multicenter Hemophilia Cohort Study, San Francisco City Cohort, ALIVE. Science 273, 1856–62 (1996).  167 193. Meyer, L. et al. Early protective effect of CCR-5 delta 32 heterozygosity on HIV-1 disease progression: relationship with viral load. The SEROCO Study Group. AIDS 11, F73–8 (1997). 194. Sheppard, H. W. et al. HIV-1 infection in individuals with the CCR5-Delta32/Delta32 genotype: acquisition of syncytium-inducing virus at seroconversion. J. Acquir. Immune Defic. Syndr. 29, 307–13 (2002). 195. Oh, D.-Y. et al. CCR5Delta32 genotypes in a German HIV-1 seroconverter cohort and report of HIV-1 infection in a CCR5Delta32 homozygous individual. PLoS One 3, e2747 (2008). 196. Richman, D. D. & Bozzette, S. A. The impact of the syncytium-inducing phenotype of human immunodeficiency virus on disease progression. J. Infect. Dis. 169, 968–74 (1994). 197. De Roda Husman, A.-M. Association between CCR5 Genotype and the Clinical Course of HIV-1 Infection. Ann. Intern. Med. 127, 882–90 (1997). 198. Koot, M. et al. Prognostic value of HIV-1 syncytium-inducing phenotype for rate of CD4+ cell depletion and progression to AIDS. Ann. Intern. Med. 118, 681–8 (1993). 199. Karlsson, A., Parsmyr, K., Sandström, E., Fenyö, E. M. & Albert, J. MT-2 cell tropism as prognostic marker for disease progression in human immunodeficiency virus type 1 infection. J. Clin. Microbiol. 32, 364–70 (1994). 200. Maas, J. J. et al. Strong association between failure of T cell homeostasis and the syncytium-inducing phenotype among HIV-1-infected men in the Amsterdam Cohort Study. AIDS 14, 1155–61 (2000). 201. Montaner, J. S. G. et al. Expansion of HAART coverage is associated with sustained decreases in HIV/AIDS morbidity, mortality and HIV transmission: the ‘HIV Treatment as Prevention’ experience in a Canadian setting. PLoS One 9, e87872 (2014). 202. Vella, S., Schwartlander, B., Sow, P. S., Eholié, S.-P. & Murphy, R. L. The History of Antiretroviral Therapy and of its Implementation in Resource-Limited Areas of the World. AIDS 26, 1231–41 (2012). 203. Hammer, S. M. et al. A trial comparing nucleoside monotherapy with combination therapy in HIV-infected adults with CD4 cell counts from 200 to 500 per cubic millimeter. AIDS Clinical Trials Group Study 175 Study Team. N. Engl. J. Med. 335, 1081–90 (1996). 204. Hammer, S. M. et al. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of  168 200 per cubic millimeter or less. AIDS Clinical Trials Group 320 Study Team. N. Engl. J. Med. 337, 725–33 (1997). 205. Gulick, R. M. et al. Treatment with indinavir, zidovudine, and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral therapy. N. Engl. J. Med. 337, 734–9 (1997). 206. Attacking AIDS with a ‘Cocktail’ Therapy: Drug Combo Sends Deaths Plummeting. HIV/AIDS News (1999). doi:https://aidsinfo.nih.gov/news/493/attacking-aids-with-a-cocktail-therapy-drug-combo-sends-deaths-plummeting 207. Hirsch, M. et al. A randomized, controlled trial of indinavir, zidovudine, and lamivudine in adults with advanced human immunodeficiency virus type 1 infection and prior antiretroviral therapy. J. Infect. Dis. 180, 659–65 (1999). 208. Staszewski, S. et al. Efavirenz plus zidovudine and lamivudine, efavirenz plus indinavir, and indinavir plus zidovudine and lamivudine in the treatment of HIV-1 infection in adults. Study 006 Team. N. Engl. J. Med. 341, 1865–73 (1999). 209. Gulick, R. M. et al. 3-year suppression of HIV viremia with indinavir, zidovudine, and lamivudine. Ann. Intern. Med. 133, 35–9 (2000). 210. Fischl, M. A. et al. The efficacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex. A double-blind, placebo-controlled trial. N. Engl. J. Med. 317, 185–91 (1987). 211. Yarchoan, R. et al. Administration of 3’-azido-3'-deoxythymidine, an inhibitor of HTLV-III/LAV replication, to patients with AIDS or AIDS-related complex. Lancet 1, 575–80 (1986). 212. Furman, P. A. et al. Phosphorylation of 3’-azido-3'-deoxythymidine and selective interaction of the 5'-triphosphate with human immunodeficiency virus reverse transcriptase. Proc. Natl. Acad. Sci. U. S. A. 83, 8333–8337 (1986). 213. Larder, B. A. & Kemp, S. D. Multiple mutations in HIV-1 reverse transcriptase confer high-level resistance to zidovudine (AZT). Science 246, 1155–8 (1989). 214. Larder, B. A., Kellam, P. & Kemp, S. D. Zidovudine resistance predicted by direct detection of mutations in DNA from HIV-infected lymphocytes. AIDS 5, 137–44 (1991). 215. Flexner, C. HIV-protease inhibitors. N. Engl. J. Med. 338, 1281–92 (1998). 216. Wensing, A. M. J., van Maarseveen, N. M. & Nijhuis, M. Fifteen years of HIV Protease Inhibitors: raising the barrier to resistance. Antiviral Res. 85, 59–74 (2010).  169 217. Venaud, S. et al. Inhibition of HIV by an anti-HIV protease synthetic peptide blocks an early step of viral replication. Res. Virol. 143, 311–319 (1992). 218. Kohl, N. E. et al. Active human immunodeficiency virus protease is required for viral infectivity. Proc. Natl. Acad. Sci. U. S. A. 85, 4686–4690 (1988). 219. Ernest, C. S., Hall, S. D. & Jones, D. R. Mechanism-based inactivation of CYP3A by HIV protease inhibitors. J. Pharmacol. Exp. Ther. 312, 583–91 (2005). 220. Eagling, V. A., Back, D. J. & Barry, M. G. Differential inhibition of cytochrome P450 isoforms by the protease inhibitors, ritonavir, saquinavir and indinavir. Br. J. Clin. Pharmacol. 44, 190–4 (1997). 221. Lillibridge, J. H. et al. Characterization of the selectivity and mechanism of human cytochrome P450 inhibition by the human immunodeficiency virus-protease inhibitor nelfinavir mesylate. Drug Metab. Dispos. 26, 609–16 (1998). 222. Kempf, D. J. et al. Pharmacokinetic enhancement of inhibitors of the human immunodeficiency virus protease by coadministration with ritonavir. Antimicrob. Agents Chemother. 41, 654–60 (1997). 223. Wood, E. et al. Superior virological response to boosted protease inhibitor-based highly active antiretroviral therapy in an observational treatment programme. HIV Med. 8, 80–5 (2007). 224. Lima, V. D. et al. Increased resilience to the development of drug resistance with modern boosted protease inhibitor-based highly active antiretroviral therapy. J. Infect. Dis. 198, 51–8 (2008). 225. De Béthune, M.-P. Non-nucleoside reverse transcriptase inhibitors (NNRTIs), their discovery, development, and use in the treatment of HIV-1 infection: a review of the last 20 years (1989-2009). Antiviral Res. 85, 75–90 (2010). 226. Esnouf, R. et al. Mechanism of inhibition of HIV-1 reverse transcriptase by non-nucleoside inhibitors. Nat. Struct. Biol. 2, 303–8 (1995). 227. Ivetac, A. & McCammon, J. A. Elucidating the inhibition mechanism of HIV-1 non-nucleoside reverse transcriptase inhibitors through multicopy molecular dynamics simulations. J. Mol. Biol. 388, 644–58 (2009). 228. Schauer, G. D., Huber, K. D., Leuba, S. H. & Sluis-Cremer, N. Mechanism of allosteric inhibition of HIV-1 reverse transcriptase revealed by single-molecule and ensemble fluorescence. Nucleic Acids Res. 42, 11687–96 (2014). 229. Steigbigel, R. T. et al. Raltegravir with optimized background therapy for resistant HIV-1 infection. N. Engl. J. Med. 359, 339–54 (2008).  170 230. Steigbigel, R. T. et al. Long-term efficacy and safety of Raltegravir combined with optimized background therapy in treatment-experienced patients with drug-resistant HIV infection: week 96 results of the BENCHMRK 1 and 2 Phase III trials. Clin. Infect. Dis. 50, 605–12 (2010). 231. Hicks, C. & Gulick, R. M. Raltegravir: the first HIV type 1 integrase inhibitor. Clin. Infect. Dis. 48, 931–9 (2009). 232. Evering, T. H. & Markowitz, M. Raltegravir: an integrase inhibitor for HIV-1. Expert Opin. Investig. Drugs 17, 413–22 (2008). 233. Grobler, J. A. et al. Diketo acid inhibitor mechanism and HIV-1 integrase: implications for metal binding in the active site of phosphotransferase enzymes. Proc. Natl. Acad. Sci. U. S. A. 99, 6661–6 (2002). 234. Chen, X. et al. Modeling, analysis, and validation of a novel HIV integrase structure provide insights into the binding modes of potent integrase inhibitors. J. Mol. Biol. 380, 504–19 (2008). 235. McColl, D. J. & Chen, X. Strand transfer inhibitors of HIV-1 integrase: bringing IN a new era of antiretroviral therapy. Antiviral Res. 85, 101–18 (2010). 236. Matthews, T. et al. Enfuvirtide: the first therapy to inhibit the entry of HIV-1 into host CD4 lymphocytes. Nat. Rev. Drug Discov. 3, 215–25 (2004). 237. Lalezari, J. P. et al. Enfuvirtide, an HIV-1 fusion inhibitor, for drug-resistant HIV infection in North and South America. N. Engl. J. Med. 348, 2175–85 (2003). 238. Kilby, J. M. et al. Potent suppression of HIV-1 replication in humans by T-20, a peptide inhibitor of gp41-mediated virus entry. Nat. Med. 4, 1302–7 (1998). 239. Liu, S. et al. Different from the HIV fusion inhibitor C34, the anti-HIV drug Fuzeon (T-20) inhibits HIV-1 entry by targeting multiple sites in gp41 and gp120. J. Biol. Chem. 280, 11259–73 (2005). 240. Paxton, W. A. et al. Relative resistance to HIV-1 infection of CD4 lymphocytes from persons who remain uninfected despite multiple high-risk sexual exposure. Nat. Med. 2, 412–7 (1996). 241. Samson, M. et al. Resistance to HIV-1 infection in caucasian individuals bearing mutant alleles of the CCR-5 chemokine receptor gene. Nature 382, 722–5 (1996). 242. Eugen-Olsen, J. et al. Heterozygosity for a deletion in the CKR-5 gene leads to prolonged AIDS-free survival and slower CD4 T-cell decline in a cohort of HIV-seropositive individuals. AIDS 11, 305–10 (1997).  171 243. Dean, M. et al. Genetic Restriction of HIV-1 Infection and Progression to AIDS by a Deletion Allele of the CKR5 Structural Gene. Science 273, 1856–62 (1996). 244. Michael, N. L. et al. The role of viral phenotype and CCR-5 gene defects in HIV-1 transmission and disease progression. Nat. Med. 3, 338–40 (1997). 245. Huang, Y. et al. The role of a mutant CCR5 allele in HIV-1 transmission and disease progression. Nat. Med. 2, 1240–3 (1996). 246. Lu, Y. et al. Genotype and allele frequency of a 32-base pair deletion mutation in the CCR5 gene in various ethnic groups: absence of mutation among Asians and Pacific Islanders. Int. J. Infect. Dis. 3, 186–91 (1999). 247. Galvani, A. P. & Novembre, J. The evolutionary history of the CCR5-Delta32 HIV-resistance mutation. Microbes Infect. 7, 302–9 (2005). 248. Lucotte, G. & Mercier, G. Distribution of the CCR5 gene 32-bp deletion in Europe. J. Acquir. Immune Defic. Syndr. Hum. Retrovirol. 19, 174–7 (1998). 249. Dorr, P. et al. Maraviroc (UK-427,857), a potent, orally bioavailable, and selective small-molecule inhibitor of chemokine receptor CCR5 with broad-spectrum anti-human immunodeficiency virus type 1 activity. Antimicrob. Agents Chemother. 49, 4721–32 (2005). 250. Nichols, W. G. et al. Hepatotoxicity observed in clinical trials of aplaviroc (GW873140). Antimicrob. Agents Chemother. 52, 858–65 (2008). 251. Currier, J. et al. Antiviral activity and safety of aplaviroc with lamivudine/zidovudine in HIV-infected, therapy-naive patients: the ASCENT (CCR102881) study. Antivir. Ther. 13, 297–306 (2008). 252. Yeni, P. et al. Antiviral activity and safety of aplaviroc, a CCR5 antagonist, in combination with lopinavir/ritonavir in HIV-infected, therapy-naïve patients: results of the EPIC study (CCR100136). HIV Med. 10, 116–24 (2009). 253. Wilkin, T. J. et al. HIV-1 type 1 chemokine coreceptor use among antiretroviral-experienced patients screened for a clinical trial of a CCR5 inhibitor: AIDS clinical trial group A5211. Clin. Infect. Dis. 24, 591–595 (2007). 254. Gulick, R. M. et al. Phase 2 study of the safety and efficacy of vicriviroc, a CCR5 inhibitor, in HIV-1-Infected, treatment-experienced patients: AIDS clinical trials group 5211. J. Infect. Dis. 196, 304–12 (2007). 255. Caseiro, M. M. et al. Vicriviroc plus optimized background therapy for treatment-experienced subjects with CCR5 HIV-1 infection: final results of two randomized phase III trials. J. Infect. 65, 326–35 (2012).  172 256. Lalezari, J. et al. Safety, efficacy, and pharmacokinetics of TBR-652, a CCR5/CCR2 antagonist, in HIV-1-infected, treatment-experienced, CCR5 antagonist-naive subjects. J. Acquir. Immune Defic. Syndr. 57, 118–25 (2011). 257. Marier, J.-F., Trinh, M., Pheng, L. H., Palleja, S. M. & Martin, D. E. Pharmacokinetics and pharmacodynamics of TBR-652, a novel CCR5 antagonist, in HIV-1-infected, antiretroviral treatment-experienced, CCR5 antagonist-naïve patients. Antimicrob. Agents Chemother. 55, 2768–74 (2011). 258. Tan, Q. et al. Structure of the CCR5 chemokine receptor-HIV entry inhibitor maraviroc complex. Science 341, 1387–90 (2013). 259. Garcia-Perez, J. et al. Allosteric model of maraviroc binding to CC chemokine receptor 5 (CCR5). J. Biol. Chem. 286, 33409–21 (2011). 260. Garcia-Perez, J. et al. New insights into the mechanisms whereby low molecular weight CCR5 ligands inhibit HIV-1 infection. J. Biol. Chem. 286, 4978–90 (2011). 261. Tilton, J. C. et al. A maraviroc-resistant HIV-1 with narrow cross-resistance to other CCR5 antagonists depends on both N-terminal and extracellular loop domains of drug-bound CCR5. J. Virol. 84, 10863–76 (2010). 262. Duma, L., Häussinger, D., Rogowski, M., Lusso, P. & Grzesiek, S. Recognition of RANTES by extracellular parts of the CCR5 receptor. J. Mol. Biol. 365, 1063–75 (2007). 263. Blanpain, C. et al. Multiple charged and aromatic residues in CCR5 amino-terminal domain are involved in high affinity binding of both chemokines and HIV-1 Env protein. J. Biol. Chem. 274, 34719–27 (1999). 264. Lee, B. et al. Epitope mapping of CCR5 reveals multiple conformational states and distinct but overlapping structures involved in chemokine and coreceptor function. J. Biol. Chem. 274, 9617–26 (1999). 265. Rasmussen, S. G. F. et al. Crystal structure of the β2 adrenergic receptor-Gs protein complex. Nature 477, 549–55 (2011). 266. Gulick, R. M. et al. Maraviroc for previously treated patients with R5 HIV-1 infection. N. Engl. J. Med. 359, (2008). 267. Whitcomb, J. M. et al. Development and characterization of a novel single-cycle recombinant-virus assay to determine human immunodeficiency virus type 1 coreceptor tropism. Antimicrob. Agents Chemother. 51, 566–75 (2007). 268. Fätkenheuer, G. et al. Subgroup analyses of maraviroc in previously treated R5 HIV-1 infection. N. Engl. J. Med. 359, 1442–55 (2008).  173 269. Hardy, W. D. et al. Two-Year Safety and Virologic Efficacy of Maraviroc in Treatment-Experienced Patients With CCR5-Tropic HIV-1 Infection: 96-Week Combined Analysis of MOTIVATE 1 and 2. J. Acquir. Immune Defic. Syndr. 55, 558–64 (2010). 270. Saag, M. et al. A double-blind, placebo-controlled trial of maraviroc in treatment-experienced patients infected with non-R5 HIV-1. J. Infect. Dis. 199, 1638–47 (2009). 271. Cooper, D. A. et al. Maraviroc versus efavirenz, both in combination with zidovudine-lamivudine, for the treatment of antiretroviral-naive subjects with CCR5-tropic HIV-1 infection. J. Infect. Dis. 201, 803-13 (2010). 272. Reeves, J. D. et al. Impact of mutations in the coreceptor binding site on human immunodeficiency virus type 1 fusion, infection, and entry inhibitor sensitivity. J. Virol. 78, 5476–85 (2004). 273. Wilkin, T. J. et al. Reanalysis of coreceptor tropism in HIV-1-infected adults using a phenotypic assay with enhanced sensitivity. Clin. Infect. Dis. 52, 925–28 (2011). 274. Gulick, R. M. et al. Five-year safety evaluation of maraviroc in HIV-1-infected treatment-experienced patients. J. Acquir. Immune Defic. Syndr. 65, 78–81 (2014). 275. Cooper, D. A. et al. Efficacy and safety of maraviroc vs. efavirenz in treatment-naive patients with HIV-1: 5-year findings. AIDS 28, 717–25 (2014). 276. Swenson, L. C. et al. Genotypic Analysis of the V3 Region of HIV from Virologic Nonresponders to Maraviroc-Containing Regimens Reveals Distinct Patterns of Failure. Antimicrob. Agents Chemother. 57, 6122–30 (2013). 277. Tsibris, A. M. N. et al. Quantitative deep sequencing reveals dynamic HIV-1 escape and large population shifts during CCR5 antagonist therapy in vivo. PLoS One 4, e5683 (2009). 278. Moncunill, G., Armand-Ugón, M., Pauls, E., Clotet, B. & Esté, J. A. HIV-1 escape to CCR5 coreceptor antagonism through selection of CXCR4-using variants in vitro. AIDS 22, 23–31 (2008). 279. Rachlis, A. et al. Canadian consensus guidelines for the optimal use of maraviroc in the treatment of HIV-infected adults. Can. J. Infect. Dis. Med. Microbiol. 21, 159–172 (2010). 280. Abel, S., Russell, D., Taylor-Worth, R. J., Ridgway, C. E. & Muirhead, G. J. Effects of CYP3A4 inhibitors on the pharmacokinetics of maraviroc in healthy volunteers. Br. J. Clin. Pharmacol. 65 Suppl 1, 27–37 (2008).  174 281. Abel, S., Back, D. J. & Vourvahis, M. Maraviroc: pharmacokinetics and drug interactions. Antivir. Ther. 14, 607–18 (2009). 282. Hyland, R., Dickins, M., Collins, C., Jones, H. & Jones, B. Maraviroc: in vitro assessment of drug-drug interaction potential. Br. J. Clin. Pharmacol. 66, 498–507 (2008). 283. Vandekerckhove, L. P. R. et al. European guidelines on the clinical management of HIV-1 tropism testing. Lancet Infect. Dis. 11, 394–407 (2011). 284. Aberg, J. a et al. Primary Care Guidelines for the Management of Persons Infected With HIV: 2013 Update by the HIV Medicine Association of the Infectious Diseases Society of America. Clin. Infect. Dis. 58, e1–e34 (2014). 285. Williams, I. et al. British HIV Association guidelines for the treatment of HIV-1-positive adults with antiretroviral therapy 2012 (Updated November 2013). HIV Med. 15 Suppl 1, 1–85 (2014). 286. Reeves, J. D., Coakley, E., Petropoulos, C. J. & Whitcomb, J. M. An Enhanced-Sensitivity Trofile HIV Coreceptor Tropism Assay for Selecting Patients for Therapy with Entry Inhibitors Targeting CCR5: A Review of Analytical and Clinical Studies. J. Viral Entry 3, 94–102 (2009). 287. Fouchier, R. A., Brouwer, M., Broersen, S. M. & Schuitemaker, H. Simple determination of human immunodeficiency virus type 1 syncytium-inducing V3 genotype by PCR. J. Clin. Microbiol. 33, 906–11 (1995). 288. Jensen, M. A. et al. Improved Coreceptor Usage Prediction and Genotypic Monitoring of R5-to-X4 Transition by Motif Analysis of Human Immunodeficiency Virus Type 1 env V3 Loop Sequences. J. Virol. 77, 13376–88 (2003). 289. Thielen, A. & Lengauer, T. Geno2pheno[454]: a Web server for the prediction of HIV-1 coreceptor usage from next-generation sequencing data. Intervirology 55, 113–7 (2012). 290. Harrigan, P. R. et al. Screening for HIV tropism using population-based V3 genotypic analysis: a retrospective virological outcome analysis using stored plasma screening samples from MOTIVATE-1. in 5th IAS Conference on HIV Pathogenesis, Treatment and Prevention WELBA101 (2009). 291. Swenson, L. C. et al. Deep sequencing to infer HIV-1 co-receptor usage: application to three clinical trials of maraviroc in treatment-experienced patients. J. Infect. Dis. 203, 237–45 (2011).  175 292. Swenson, L. C. et al. Improved detection of CXCR4-using HIV by V3 genotyping: application of population-based and ‘deep’ sequencing to plasma RNA and proviral DNA. J. Acquir. Immune Defic. Syndr. 54, 506–10 (2010). 293. Kagan, R. M. et al. A genotypic test for HIV-1 tropism combining Sanger sequencing with ultradeep sequencing predicts virologic response in treatment-experienced patients. PLoS One 7, e46334 (2012). 294. Swenson, L. C. et al. Use of cellular HIV DNA to predict virologic response to maraviroc: performance of population-based and deep sequencing. Clin. Infect. Dis. 56, 1659–66 (2013). 295. Verhofstede, C. et al. Concordance between HIV-1 genotypic coreceptor tropism predictions based on plasma RNA and proviral DNA. HIV Med. 12, 544–52 (2011). 296. Soulié, C. et al. Factors associated with proviral DNA HIV-1 tropism in antiretroviral therapy-treated patients with fully suppressed plasma HIV viral load: implications for the clinical use of CCR5 antagonists. J. Antimicrob. Chemother. 65, 749–51 (2010). 297. Seclén, E., Del Mar González, M., De Mendoza, C., Soriano, V. & Poveda, E. Dynamics of HIV tropism under suppressive antiretroviral therapy: implications for tropism testing in subjects with undetectable viraemia. J. Antimicrob. Chemother. 65, 1493–6 (2010). 298. Pou, C. et al. HIV-1 tropism testing in subjects achieving undetectable HIV-1 RNA: diagnostic accuracy, viral evolution and compartmentalization. PLoS One 8, e67085 (2013). 299. Lee, G. Q. et al. Limited Evolution of Inferred HIV-1 Tropism while Viremia is Undetectable During Standard HAART Therapy. PLoS One 9, e99000 (2014). 300. Rozera, G. et al. Massively parallel pyrosequencing highlights minority variants in the HIV-1 env quasispecies deriving from lymphomonocyte sub-populations. Retrovirology 6, 15–27 (2009). 301. Rozera, G. et al. Quasispecies tropism and compartmentalization in gut and peripheral blood during early and chronic phases of HIV-1 infection: possible correlation with immune activation markers. Clin. Microbiol. Infect. 20, O157–66 (2014). 302. Perelson, A. S., Neumann, A. U., Markowitz, M., Leonard, J. M. & Ho, D. D. HIV-1 Dynamics in Vivo: Virion Clearance Rate, Infected Cell Life-Span, and Viral Generation Time. Science 271, 1582–6 (1996). 303. Coffin, J. M. HIV population dynamics in vivo: implications for genetic variation, pathogenesis, and therapy. Science 267, 483–9 (1995).  176 304. Maldarelli, F. et al. HIV populations are large and accumulate high genetic diversity in a nonlinear fashion. J. Virol. 87, 10313–23 (2013). 305. Mansky, L. M. & Temin, H. M. Lower in vivo mutation rate of human immunodeficiency virus type 1 than that predicted from the fidelity of purified reverse transcriptase. J. Virol. 69, 5087–94 (1995). 306. Wood, N. et al. HIV evolution in early infection: selection pressures, patterns of insertion and deletion, and the impact of APOBEC. PLoS Pathog. 5, e1000414 (2009). 307. Fraser, C. et al. Virulence and pathogenesis of HIV-1 infection: an evolutionary perspective. Science 343, 1328-35 (2014). 308. Richman, D. D. HIV chemotherapy. Nature 410, 995–1001 (2001). 309. Brumme, Z. L. et al. HLA-associated immune escape pathways in HIV-1 subtype B Gag, Pol and Nef proteins. PLoS One 4, e6687 (2009). 310. Ji, J. & Loeb, L. A. Fidelity of HIV-1 reverse transcriptase copying RNA in vitro. Biochemistry 31, 954–8 (1992). 311. Boyer, J. C., Bebenek, K. & Kunkel, T. A. Unequal human immunodeficiency virus type 1 reverse transcriptase error rates with RNA and DNA templates. Proc. Natl. Acad. Sci. U. S. A. 89, 6919–23 (1992). 312. Jetzt, A. E. et al. High Rate of Recombination throughout the Human Immunodeficiency Virus Type 1 Genome. J. Virol. 74, 1234–40 (2000). 313. Zhuang, J. et al. Human Immunodeficiency Virus Type 1 Recombination: Rate, Fidelity, and Putative Hot Spots. J. Virol. 76, 11273–82 (2002). 314. Neher, R. A. & Leitner, T. Recombination rate and selection strength in HIV intra-patient evolution. PLoS Comput. Biol. 6, e1000660 (2010). 315. Jung, A. et al. Recombination: Multiply infected spleen cells in HIV patients. Nature 418, 144 (2002). 316. Chen, J. et al. High efficiency of HIV-1 genomic RNA packaging and heterozygote formation revealed by single virion analysis. Proc. Natl. Acad. Sci. U. S. A. 106, 13535–40 (2009). 317. Hu, W. S. & Temin, H. M. Genetic consequences of packaging two RNA genomes in one retroviral particle: pseudodiploidy and high rate of genetic recombination. Proc. Natl. Acad. Sci. U. S. A. 87, 1556–60 (1990).  177 318. Hu, W. & Temin, H. Retroviral recombination and reverse transcription. Science 250, 1227–33 (1990). 319. Stuhlmann, H. & Berg, P. Homologous recombination of copackaged retrovirus RNAs during reverse transcription. J. Virol. 66, 2378–88 (1992). 320. Malim, M. H. & Bieniasz, P. D. HIV Restriction Factors and Mechanisms of Evasion. Cold Spring Harb. Perspect. Med. 2, a006940 (2012). 321. Mangeat, B. et al. Broad antiretroviral defence by human APOBEC3G through lethal editing of nascent reverse transcripts. Nature 424, 99–103 (2003). 322. Bishop, K. N., Holmes, R. K., Sheehy, A. M. & Malim, M. H. APOBEC-mediated editing of viral RNA. Science 305, 645 (2004). 323. Yu, Q. et al. Single-strand specificity of APOBEC3G accounts for minus-strand deamination of the HIV genome. Nat. Struct. Mol. Biol. 11, 435–42 (2004). 324. Saiki, R. K. et al. Enzymatic amplification of beta-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science 230, 1350–4 (1985). 325. Mullis, K. B. et al. United States Patent No. 4683195. (1987). 326. Atkinson, M. R., Deutscher, M. P., Kornberg, A., Russell, A. F. & Moffatt, J. G. Enzymatic synthesis of deoxyribonucleic acid. XXXIV. Termination of chain growth by a 2’,3'-dideoxyribonucleotide. Biochemistry 8, 4897–904 (1969). 327. Sanger, F. & Coulson, A. R. A rapid method for determining sequences in DNA by primed synthesis with DNA polymerase. J. Mol. Biol. 94, 441–8 (1975). 328. Sanger, F., Nicklen, S. & Coulson, A. R. DNA sequencing with chain-terminating inhibitors. 1977. Proc. Natl. Acad. Sci. U. S. A. 74, 5463–5467 (1977). 329. Sanger, F., Coulson, A. R., Barrell, B. G., Smith, A. J. & Roe, B. A. Cloning in single-stranded bacteriophage as an aid to rapid DNA sequencing. J. Mol. Biol. 143, 161–78 (1980). 330. Prober, J. M. et al. A system for rapid DNA sequencing with fluorescent chain-terminating dideoxynucleotides. Science 238, 336–41 (1987). 331. Jorgenson, J. W. & Lukacs, K. D. Zone electrophoresis in open-tubular glass capillaries. Anal. Chem. 53, 1298–1302 (1981). 332. Cohen, A. S. et al. Rapid separation and purification of oligonucleotides by high-performance capillary gel electrophoresis. Proc. Natl. Acad. Sci. U. S. A. 85, 9660–63 (1988).  178 333. Ansorge, W., Sproat, B., Stegemann, J., Schwager, C. & Zenke, M. Automated DNA sequencing: ultrasensitive detection of fluorescent bands during electrophoresis. Nucleic Acids Res. 15, 4593–602 (1987). 334. Kambara, H., Nishikawa, T., Katayama, Y. & Yamaguchi, T. Optimization of Parameters in a DNA Sequenator Using Fluorescence Detection. Bio/Technology 6, 816–21 (1988). 335. Smith, L. M. et al. Fluorescence detection in automated DNA sequence analysis. Nature 321, 674–9 (1986). 336. Slatko, B. E. et al. ‘First generation’ automated DNA sequencing technology. Curr. Protoc. Mol. Biol. Chapter 7, Unit7.2 (2011). 337. Drossman, H., Luckey, J. A., Kostichka, A. J., D’Cunha, J. & Smith, L. M. High-speed separations of DNA sequencing reactions by capillary electrophoresis. Anal. Chem. 62, 900–3 (1990). 338. Luckey, J. A. et al. High speed DNA sequencing by capillary electrophoresis. Nucleic Acids Res. 18, 4417–21 (1990). 339. Swerdlow, H. & Gesteland, R. Capillary gel electrophoresis for rapid, high resolution DNA sequencing. Nucleic Acids Res. 18, 1415–9 (1990). 340. Huang, X. C., Quesada, M. A. & Mathies, R. A. DNA sequencing using capillary array electrophoresis. Anal. Chem. 64, 2149–54 (1992). 341. Huang, X. C., Quesada, M. A. & Mathies, R. A. Capillary array electrophoresis using laser-excited confocal fluorescence detection. Anal. Chem. 64, 967–72 (1992). 342. Kircher, M. & Kelso, J. High-throughput DNA sequencing--concepts and limitations. Bioessays 32, 524–36 (2010). 343. Margulies, M. et al. Genome sequencing in microfabricated high-density picolitre reactors. Nature 437, 376–80 (2005). 344. Wheeler, D. A. et al. The complete genome of an individual by massively parallel DNA sequencing. Nature 452, 872–6 (2008). 345. Rodrigue, S. et al. Whole genome amplification and de novo assembly of single bacterial cells. PLoS One 4, e6864 (2009). 346. Tawfik, D. S. & Griffiths, A. D. Man-made cell-like compartments for molecular evolution. Nat. Biotechnol. 16, 652–6 (1998). 347. Dressman, D., Yan, H., Traverso, G., Kinzler, K. W. & Vogelstein, B. Transforming single DNA molecules into fluorescent magnetic particles for detection and  179 enumeration of genetic variations. Proc. Natl. Acad. Sci. U. S. A. 100, 8817–22 (2003). 348. Williams, R. et al. Amplification of complex gene libraries by emulsion PCR. Nat. Methods 3, 545–50 (2006). 349. Ronaghi, M., Karamohamed, S., Pettersson, B., Uhlén, M. & Nyrén, P. Real-time DNA sequencing using detection of pyrophosphate release. Anal. Biochem. 242, 84–9 (1996). 350. Ronaghi, M. DNA SEQUENCING:A Sequencing Method Based on Real-Time Pyrophosphate. Science 281, 363–5 (1998). 351. Mardis, E. R. The impact of next-generation sequencing technology on genetics. Trends Genet. 24, 133–41 (2008). 352. Rothberg, J. M. et al. An integrated semiconductor device enabling non-optical genome sequencing. Nature 475, 348–52 (2011). 353. Shendure, J. & Ji, H. Next-generation DNA sequencing. Nat. Biotechnol. 26, 1135–45 (2008). 354. Quail, M. A., Swerdlow, H. & Turner, D. J. Improved protocols for the illumina genome analyzer sequencing system. Curr. Protoc. Hum. Genet. Chapter 18, Unit 18.2 (2009). 355. Chen, F. et al. The history and advances of reversible terminators used in new generations of sequencing technology. Genomics. Proteomics Bioinformatics 11, 34–40 (2013). 356. Illumina Inc. Illumina Technology. (2015). Accessed September 24, 2015, at: <http://www.illumina.com/technology.html> 357. Bentley, D. R. et al. Accurate whole human genome sequencing using reversible terminator chemistry. Nature 456, 53–9 (2008). 358. Glenn, T. C. Field guide to next-generation DNA sequencers. Mol. Ecol. Resour. 11, 759–69 (2011). 359. Loman, N. J. et al. Performance comparison of benchtop high-throughput sequencing platforms. Nat. Biotechnol. 30, 434–9 (2012). 360. Eid, J. et al. Real-time DNA sequencing from single polymerase molecules. Science 323, 133–8 (2009). 361. Korlach, J. et al. Real-time DNA sequencing from single polymerase molecules. Methods Enzymol. 472, 431–55 (2010).  180 362. Quail, M. A. et al. A tale of three next generation sequencing platforms: comparison of Ion Torrent, Pacific Biosciences and Illumina MiSeq sequencers. BMC Genomics 13, 341 (2012). 363. Archer, J. et al. Use of four next-generation sequencing platforms to determine HIV-1 coreceptor tropism. PLoS One 7, e49602 (2012). 364. Ewing, B., Hillier, L., Wendl, M. C. & Green, P. Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res. 8, 175–85 (1998). 365. Harrigan, P. R. et al. Performance of ReCall basecalling software for high-throughput HIV drug resistance basecalling using ‘“in-house"’methods.”’ in The XIV international AIDS Conference (2002). 366. Woods, C. K. et al. Automating HIV drug resistance genotyping with RECall, a freely accessible sequence analysis tool. J. Clin. Microbiol. 50, 1936–42 (2012). 367. Luscombe, N. M., Greenbaum, D. & Gerstein, M. What is bioinformatics? A proposed definition and overview of the field. Methods Inf. Med. 40, 346–58 (2001). 368. Altman, R. B. & Dugan, J. M. Structural Bioinformatics. (John Wiley & Sons, 2009). at <https://books.google.com/books?hl=en&lr=&id=4H_ai7ivRIcC&pgis=1> 369. Goodman, N. Biological data becomes computer literate: new advances in bioinformatics. Curr. Opin. Biotechnol. 13, 68–71 (2002). 370. Pop, M. & Salzberg, S. L. Bioinformatics challenges of new sequencing technology. Trends Genet. 24, 142–9 (2008). 371. Searls, D. B. Bioinformatics tools for whole genomes. Annu. Rev. Genomics Hum. Genet. 1, 251–79 (2000). 372. Hwang, S. S., Boyle, T. J., Lyerly, H. K. & Cullen, B. R. Identification of the envelope V3 loop as the primary determinant of cell tropism in HIV-1. Science 253, 71–4 (1991). 373. Cann, A. J. et al. The region of the envelope gene of human immunodeficiency virus type 1 responsible for determination of cell tropism. J. Virol. 66, 305–9 (1992). 374. Fouchier, R. A. et al. Phenotype-associated sequence variation in the third variable domain of the human immunodeficiency virus type 1 gp120 molecule. J. Virol. 66, 3183–7 (1992). 375. De Jong, J. J., De Ronde, A., Keulen, W., Tersmette, M. & Goudsmit, J. Minimal requirements for the human immunodeficiency virus type 1 V3 domain to  181 support the syncytium-inducing phenotype: analysis by single amino acid substitution. J. Virol. 66, 6777–80 (1992). 376. Jensen, M. A. & van ’t Wout, A. B. Predicting HIV-1 coreceptor usage with sequence analysis. AIDS Rev. 5, 104–12 (2003). 377. Garrido, C. et al. Evaluation of eight different bioinformatics tools to predict viral tropism in different human immunodeficiency virus type 1 subtypes. J. Clin. Microbiol. 46, 887–91 (2008). 378. Low, A. J. et al. Current V3 genotyping algorithms are inadequate for predicting X4 co-receptor usage in clinical isolates. AIDS 21, F17–24 (2007). 379. De Mendoza, C. et al. Performance of a population-based HIV-1 tropism phenotypic assay and correlation with V3 genotypic prediction tools in recent HIV-1 seroconverters. J. Acquir. Immune Defic. Syndr. 48, 241–4 (2008). 380. Gribskov, M., McLachlan, A. D. & Eisenberg, D. Profile analysis: detection of distantly related proteins. Proc. Natl. Acad. Sci. U. S. A. 84, 4355–8 (1987). 381. Beerenwinkel, N. et al. Geno2pheno: interpreting genotypic HIV drug resistance tests. IEEE Intell. Syst. 16, 35–41 (2001). 382. Beerenwinkel, N. et al. Diversity and complexity of HIV-1 drug resistance: a bioinformatics approach to predicting phenotype from genotype. Proc. Natl. Acad. Sci. U. S. A. 99, 8271–6 (2002). 383. Beerenwinkel, N. et al. Geno2pheno: Estimating phenotypic drug resistance from HIV-1 genotypes. Nucleic Acids Res. 31, 3850–5 (2003). 384. Sing, T. et al. Geno2pheno[coreceptor]: a Tool for Predicting Coreceptor Usage from Genotype and for Monitoring Coreceptor-associated Sequence Alterations. in 3rd European HIV Drug Resistance Workshop (European AIDS Clinical Society, 2005). Accessed September 24, 2015, at: <http://pubman.mpdl.mpg.de/pubman/faces/viewItemOverviewPage.jsp?itemId=escidoc:1328242> 385. Sing, T., Beerenwinkel, N. & Lengauer, T. Learning Mixtures of Localized Rules by Maximizing the Area Under the ROC Curve. in Proceedings of the ROCAI-2004 International Workshop on ROC Analysis in Artificial Intelligence (eds. Hernandez-Orallo, J., Lachiche, N. & Flach, P.) 89–96 (2004). Accessed September 24, 2015, at: <http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.2.2786> 386. Vandekerckhove, L., Verhofstede, C. & Vogelaers, D. Maraviroc: integration of a new antiretroviral drug class into clinical practice. J. Antimicrob. Chemother. 61, 1187–90 (2008).  182 387. Low, A. J., McGovern, R. A. & Harrigan, P. R. Trofile HIV co-receptor usage assay. Expert Opin. Med. Diagn. 3, 181–91 (2009). 388. Lehmann, C. et al. Stable coreceptor usage of HIV in patients with ongoing treatment failure on HAART. J. Clin. Virol. 37, 300–4 (2006). 389. Poveda, E. et al. Design and validation of new genotypic tools for easy and reliable estimation of HIV tropism before using CCR5 antagonists. J. Antimicrob. Chemother. 63, 1006–10 (2009). 390. Dong, W. et al. Performance of RE_Call basecalling software for high-throughput HIV drug resistance basecalling using in-house methods. in 11th Annual Canadian Conference on HIV/AIDS Reserach 138P (2002). 391. Brooks, J. I. et al. Evaluation of an automated sequence analysis tool to standardize HIV genotyping results. in 18th Annual Canadian Conference on HIV/AIDS Research O061 (2009). 392. Valdez, H. et al. Weighted OBT susceptibility score (wOBTSS) is a stronger predictor of virologic response at 48 weeks than baseline tropism result in MOTIVATE 1 and 2. in 48th Annual ICAAC/IDSA 46th Annual Meeting H–1221 (2008). 393. McGovern, R. A. et al. Optimization of clinically relevant cutpoints for the determination of HIV co-receptor usage to predict maraviroc responses in treatment experienced (TE) patients using population V3 genotyping. in 12th European AIDS Conference PE3.4/8 (2009). 394. Swenson, L. C. et al. "Deep” sequencing to identify treatment-experienced patients who respond to maraviroc (MVC). in 12th European AIDS Conference PE3.3/2 (2009). 395. Su, Z. et al. Response to vicriviroc in treatment-experienced subjects, as determined by an enhanced-sensitivity coreceptor tropism assay: reanalysis of AIDS clinical trials group A5211. J. Infect. Dis. 200, 1724–8 (2009). 396. Lengauer, T. Genotypic basis of viral tropism. in 4th International Workshop on Targeting HIV Entry. (2008). 397. MacArthur, R. D. & Novak, R. M. Reviews of anti-infective agents: maraviroc: the first of a new class of antiretroviral agents. Clin. Infect. Dis. 47, 236–41 (2008). 398. Connor, R. I., Sheridan, K. E., Ceradini, D., Choe, S. & Landau, N. R. Change in coreceptor use correlates with disease progression in HIV-1--infected individuals. J. Exp. Med. 185, 621–8 (1997).  183 399. Xiao, L., Rudolph, D. L., Owen, S. M., Spira, T. J. & Lal, R. B. Adaptation to promiscuous usage of CC and CXC-chemokine coreceptors in vivo correlates with HIV-1 disease progression. AIDS 12, F137–43 (1998). 400. Soriano, V. et al. Optimal use of maraviroc in clinical practice. AIDS 22, 2231–40 (2008). 401. Saag, M., Heera, J. & Goodrich, J. Reanalysis of the MERIT study with the Enhanced Trofile Assay (MERIT-ES). in 48th ICAAC/IDSA, 46th Annual Meeting H–1232a (2008). 402. Lengauer, T., Sander, O., Sierra, S., Thielen, A. & Kaiser, R. Bioinformatics prediction of HIV coreceptor usage. Nat. Biotechnol. 25, 1407–10 (2007). 403. McGovern, R. A. et al. Population-based V3 genotypic tropism assay: a retrospective analysis using screening samples from the A4001029 and MOTIVATE studies. AIDS 24, 2517–25 (2010). 404. Portsmouth, S. et al. Virologic outcome by V3 loop genotypic population sequencing and 454 ‘deep sequencing’in clade B and non-B virus in MERIT at 48 and 96 weeks. in 8 th International AIDS Conference TUPE0134 (2010). 405. Skrabal, K. et al. Determining human immunodeficiency virus coreceptor use in a clinical setting: degree of correlation between two phenotypic assays and a bioinformatic model. J. Clin. Microbiol. 45, 279–84 (2007). 406. Swenson, L. et al. Large-scale application of deep sequencing using 454 technology to HIV tropism screening. in 17th Conference on Retroviruses and Opportunistic Infections M–268 (2010). 407. Zhang, L., Huang, Y., He, T., Cao, Y. & Ho, D. D. HIV-1 subtype and second-receptor use. Nature 383, 768 (1996). 408. MacArthur, R. D. & Novak, R. M. Reviews of anti-infective agents: maraviroc: the first of a new class of antiretroviral agents. Clin. Infect. Dis. 47, 236–41 (2008). 409. Swenson, L. C. et al. Deep V3 sequencing for HIV type 1 tropism in treatment-naive patients: a reanalysis of the MERIT trial of maraviroc. Clin. Infect. Dis. 53, 732–42 (2011). 410. Wang, C., Mitsuya, Y., Gharizadeh, B., Ronaghi, M. & Shafer, R. W. Characterization of mutation spectra with ultra-deep pyrosequencing: application to HIV-1 drug resistance. Genome Res. 17, 1195–201 (2007). 411. Nieuwkerk, P. T. Limited Patient Adherence to Highly Active Antiretroviral Therapy for HIV-1 Infection in an Observational Cohort Study. Arch. Intern. Med. 161, 1962 (2001).  184 412. Knapp, D. J. H. F. et al. ‘Deep’ sequencing accuracy and reproducibility using Roche/454 technology for inferring co-receptor usage in HIV-1. PLoS One 9, e99508 (2014). 413. Crow, J. F. & Kimura, M. An introduction to population genetics theory. (Harper & Row, 1970). 414. Westby, M. et al. Emergence of CXCR4-using human immunodeficiency virus type 1 (HIV-1) variants in a minority of HIV-1-infected patients following treatment with the CCR5 antagonist maraviroc is from a pretreatment CXCR4-using virus reservoir. J. Virol. 80, 4909–20 (2006). 415. McGovern, R. A. et al. Next generation deep sequencing to evaluate viral tropism in HIV-1 patients exposed to maraviroc add-on therapy for 8 days. in 6th IAS Conference on HIV Pathogenesis TUPDA0102 (2011). 416. McGovern, R. A. et al. Population-based sequencing of the V3-loop can predict the virological response to maraviroc in treatment-naive patients of the MERIT trial. J. Acquir. Immune Defic. Syndr. 61, 279–86 (2012). 417. Jabara, C. B., Jones, C. D., Roach, J., Anderson, J. A. & Swanstrom, R. Accurate sampling and deep sequencing of the HIV-1 protease gene using a Primer ID. Proc. Natl. Acad. Sci. U. S. A. 108, 20166–71 (2011). 418. Raymond, S. et al. Development and performance of a new recombinant virus phenotypic entry assay to determine HIV-1 coreceptor usage. J. Clin. Virol. 47, 126–30 (2010). 419. Gonzalez-Serna, A. et al. TROCAI (tropism coreceptor assay information): a new phenotypic tropism test and its correlation with Trofile enhanced sensitivity and genotypic approaches. J. Clin. Microbiol. 48, 4453–4458 (2010). 420. Genebat, M. et al. Correlation between the Trofile test and virological response to a short-term maraviroc exposure in HIV-infected patients. J. Antimicrob. Chemother. 64, 845–9 (2009). 421. Ruiz-Mateos, E. et al. Virological Response after Short-Term CCR5 Antagonist Exposure in HIV-Infected Patients: Frequency of Subjects with Virological Response and Associated Factors. Antimicrob. Agents Chemother. 55, 4664–9 (2011). 422. Fätkenheuer, G. et al. Efficacy of short-term monotherapy with maraviroc, a new CCR5 antagonist, in patients infected with HIV-1. Nat. Med. 11, 1170–2 (2005). 423. Schürmann, D. et al. Antiviral activity, pharmacokinetics and safety of vicriviroc, an oral CCR5 antagonist, during 14-day monotherapy in HIV-infected adults. AIDS 21, 1293–9 (2007).  185 424. McGovern, R. a et al. Maraviroc treatment in non-R5-HIV-1-infected patients results in the selection of extreme CXCR4-using variants with limited effect on the total viral setpoint. J. Antimicrob. Chemother. 1–8 (2013).  425. Gonzalez-Serna, A. et al. Correlation of the Virological Response to Short-Term Maraviroc Monotherapy with Standard and Deep-Sequencing-Based Genotypic Tropism Prediction Methods. Antimicrobial Agents and Chemotherapy 56, 1202–7 (2012). 426. Katoh, K., Misawa, K., Kuma, K. & Miyata, T. MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acids Res. 30, 3059–66 (2002). 427. Stamatakis, A. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688–90 (2006). 428. Shao, W. et al. Analysis of 454 sequencing error rate, error sources, and artifact recombination for detection of Low-frequency drug resistance mutations in HIV-1 DNA. Retrovirology 10, 18 (2013). 429. Brodin, J. et al. PCR-induced transitions are the major source of error in cleaned ultra-deep pyrosequencing data. PLoS One 8, e70388 (2013). 430. McGovern, R. A., Harrigan, P. R. & Swenson, L. C. Genotypic inference of HIV-1 tropism using population-based sequencing of V3. J. Vis. Exp. 42, 2531 (2010). 431. Veselinovic, M. et al. HIV pre-exposure prophylaxis: mucosal tissue drug distribution of RT inhibitor Tenofovir and entry inhibitor Maraviroc in a humanized mouse model. Virology 464-465, 253–63 (2014). 432. Neff, C. P., Kurisu, T., Ndolo, T., Fox, K. & Akkina, R. A topical microbicide gel formulation of CCR5 antagonist maraviroc prevents HIV-1 vaginal transmission in humanized RAG-hu mice. PLoS One 6, e20209 (2011). 433. Blakney, A. K., Ball, C., Krogstad, E. A. & Woodrow, K. A. Electrospun fibers for vaginal anti-HIV drug delivery. Antiviral Res. 100 Suppl, S9–16 (2013). 434. Hütter, G. et al. Long-term control of HIV by CCR5 Delta32/Delta32 stem-cell transplantation. N. Engl. J. Med. 360, 692–8 (2009). 435. Symons, J. et al. Dependence on the CCR5 Coreceptor for Viral Replication Explains the Lack of Rebound of CXCR4-Predicted HIV Variants in the Berlin Patient. Clin. Infect. Dis. 59, 596–600 (2014). 436. Allers, K. et al. Evidence for the cure of HIV infection by CCR5Δ32/Δ32 stem cell transplantation. Blood 117, 2791–9 (2011).  186 437. Yukl, S. A. et al. Challenges in detecting HIV persistence during potentially curative interventions: a study of the Berlin patient. PLoS Pathog. 9, e1003347 (2013). 438. Hütter, G. More on Shift of HIV Tropism in Stem-Cell Transplantation with CCR5 Delta32/Delta32 Mutation. N. Engl. J. Med. 371, 2437–8 (2014). 439. Perez, E. E. et al. Establishment of HIV-1 resistance in CD4+ T cells by genome editing using zinc-finger nucleases. Nat. Biotechnol. 26, 808–16 (2008). 440. Badia, R., Riveira-Muñoz, E., Clotet, B., Esté, J. A. & Ballana, E. Gene editing using a zinc-finger nuclease mimicking the CCR5Δ32 mutation induces resistance to CCR5-using HIV-1. J. Antimicrob. Chemother. 69, 1755–9 (2014). 441. Didigu, C. A. et al. Simultaneous zinc-finger nuclease editing of the HIV coreceptors ccr5 and cxcr4 protects CD4+ T cells from HIV-1 infection. Blood 123, 61–9 (2014).   187 Appendix 1 Chapter Three Supplementary Material  Supplmental Figure 3.1a.  The percentage of patients with undetectable plasma HIV following the start of maraviroc compared to efavirenz at the FPR cutoff points suggested for population-based C) & D) and next generation genotyping A) & B). Patients were retrospectively screened for tropism by population-based V3 genotyping.  Adjusting g2p FPR cutoff points patients were re-stratified and virological response to therapy compare by tropism.  The yellow line represesnts patient response to efavirenz; the blue line represents patient response to maraviroc.  188  Supplemental Figure 3.1b.  The percentage of patients with undetectable plasma HIV following the start of maraviroc compared to efavirenz at the more conservative FPR cutoff points suggested primarily by European E) & F) and German G) & H) guidelines. The yellow line represents patient response to efavirenz; the blue line represents patient response to maraviroc.  189  Supplemental Figure 3.2a.  Change in plasma viral load following the start of maraviroc twice-daily (b.i.d.) compared to efavirenz at the FPR cutoff points suggested for population-based C) & D) and next generation genotyping A) & B). The yellow line represents patient response to efavirenz; the blue line represents patient response to maraviroc.  190  Supplemental Figure 3.2b.  Change in plasma viral load following the start of maraviroc twice-daily (b.i.d.) compared to efavirenz at the more conservative FPR cutoff points suggested by European E) & F) and German G) & H) guidelines. The yellow line represents patient response to efavirenz; the blue line represents patient response to maraviroc.  191  Supplemental Figure 3.3a.  Change in plasma viral load following the start of maraviroc once-daily (q.d.) compared to efavirenz at the FPR cutoff points suggested for population-based C) & D) and next generation genotyping A) & B). The yellow line represents patient response to efavirenz; the blue line represents patient response to maraviroc.  192  Supplemental Figure 3.3b.  Change in plasma viral load following the start of maraviroc once-daily (q.d.) compared to efavirenz at the more conservative FPR cutoff points suggested by European E) & F) and German G) & H) guidelines. The yellow line represents patient response to efavirenz; the blue line represents patient response to maraviroc.  193  Supplemental Figure 3.4a.  Time to change in tropism from R5 to non‐R5 virus for patients re‐screened using aV3‐genotypic tropism assay at the FPR cutoff points suggested for population-based C) & D) and next generation genotyping A) & B). Screening plasma samples were rescreened using the V3 genotypic assay, and re‐stratified as being R5 (green line) or non‐R5 (red line) for patients enrolled in the maraviroc twice-daily (b.i.d.) arm or the maraviroc once-daily (q.d.) arm.  194  Supplemental Figure 3.4b.  Time to change in tropism from R5 to non‐R5 virus for patients re‐screened using aV3‐genotypic tropism assay at the more conservative FPR cutoff points suggested by European E) & F) and German G) & H) guidelines. Screening plasma samples were rescreened using the V3 genotypic assay, and re‐stratified as being R5 (green line) or non‐R5 (red line) for patients enrolled in the maraviroc twice-daily (b.i.d.) arm or the maraviroc once-daily (q.d.) arm.  195  Supplemental Figure 3.5. The concordance and discordance between ESTA and V3 genotype, a comparison of genotype testing in singlicate, duplicate and triplicate. The change in plasma viral load (A) and the percentage of patient with suppressed viremia (B) following the start of maraviroc. Green lines represent patients screened R5 by both assays; red lines represent patients screened non‐R5 by both assays. Black lines indicate discordance in tropism call between the two assays.  196 Appendix 2 Chapter Four Supplementary Material Supplemental Table 4.1.  454 GS-FLX chemistry, adaptor linked, tagged primers.  A list of the primer sequences and HXB2 locations used in the generation of V3 sequences on the 454 GS-FLX with standard chemistry.  The adaptor (red), multiplex identifier (MID; blue) and target sequence (black) segments of the primer have been identified.     197 Supplemental Table 4.2.  The three most prominent sequences by genotype for patients with non-R5 HIV-1 before and after the addition of maraviroc to their treatment regimen. Sequences in bold were minority non-R5 variants by 454 “deep” sequencing (FPR cutoff 3.5%) at pre-maraviroc time point(s).  The non-R5 variant of interest from Patient C was identified two days prior to maraviroc at low prevalence.  In some differences are at the nucleotide level.   198   199 Appendix 3 Chapter Five Supplementary Material Supplemental Table 5.1.  454 GS-FLX Titanium chemistry, adaptor linked, tagged primers.  A list of the primer sequences and HXB2 locations used in the generation of V3 sequences on the 454 GS-FLX with Titanium chemistry.  The adaptor (red), multiplex identifier (MID; blue) and target sequence (black) segments of the primer have been identified.   200  Supplemental Figure 5.1.  HIV-1 sequence frequency stratified as a function of g2p FPR during the course of short-term maraviroc exposure in patients screened as having R5 virus. Sequences were placed in five bins according to g2p FPR interpretation of the sequences generated by 454 “deep” sequencing.  The red and yellow lines represent non-R5 variants (non-R5≤3.5), least likely to respond to maraviroc; the green line represents HIV-1 variants with a g2p FPR bordering the R5 or non-R5 tropism call.  The blue and black lines represent R5 variants, the variants most likely to respond to maraviroc.  201  Supplemental Figure 5.2.  Phylogenetic trees illustrating the change in a representative R5 and non-R5 viral population after short-term maraviroc exposure. Phylogenetic trees based on maximum likelihood for a patient screened as having A) R5 virus and B) non-R5 virus by 454 “deep” sequencing at Day 0.  The viral populations from patients 10 and 22 were selected as examples to demonstrate changes in R5 and non-R5 screening populations, respsectively, following eight days of maraviroc exposure.  Blue branches represent R5 variants (R5 FPR>3.5); red branches represent non-R5 variants(non-R5≤3.5); grey branches represent variants existing at the other time point.  The value at each tip indicates the number of times the sequence was identified.  202  Supplemental Figure 5.3a.  Phylogenetic trees illustrating the change in non-R5 HIV-1 viral populations after short-term exposure to maraviroc. Phylogenetic trees based on maximum likelihood for patients experiencing a viral tropism switch, as determined by 454 “deep” sequencing, after eight days of MVC exposure.  Patients 7 and 19 were screened as having non-R5 virus at baseline however after eight days of MVC exposure non-R5 variants were suppressed.  Blue branches represent R5 variants (R5 FPR>3.5); red branches represent non-R5 variants(non-R5≤3.5); grey branches represent variants existing at the other time point.  The value at each tip indicates the number of times the sequence was identified.  203  Supplementary Figure 5.3b.  Phylogenetic trees illustrating the change in non-R5 HIV-1 viral populations after short-term exposure to maraviroc. Phylogenetic trees based on maximum likelihood for patients experiencing a viral tropism switch, as determined by 454 “deep” sequencing, after eight days of MVC exposure.  Patients 13 and 17 were screened as having less than 2% non-R5 virus at baseline however after eight days of MVC exposure the non-R5 population expand to predominate.  Blue branches represent R5 variants (R5 FPR>3.5); red branches represent non-R5 variants (non-R5≤3.5); grey branches represent variants existing at the other time point.  The value at each tip indicates the number of times the sequence was identified.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0166738/manifest

Comment

Related Items