The rational design of an HIV-1 vaccine immunogen able to induce potent, cross-reactive, neutralising antibodies remains one of the single greatest challenges in the field of vaccine research today. Roughly a dozen broadly neutralising monoclonal antibodies have been isolated to date, and their epitopes represent important vaccination targets. Interestingly, apart from three that identify over-lapping epitopes in gp41, all of the broadly neutralising monoclonal antibodies target epitopes apparent on different conformations of gp120 (including the epitopes of PG9/PG16). Thus the gp120 monomer remains the most ideal template for immunogen design. Recently, epitopes in the C3-V4 region of gp120 have been shown to be major targets for early strain-specific neutralising antibodies in subtype C infected individuals. Autologous neutralising antibodies identify vulnerable sites on the envelope, and understanding the nature of antigenic “hotspots” on gp120 will help to guide rational vaccine design. This study sought to confirm in four individuals that the C3-V4 epitope was in fact apparent on monomeric gp120, and thereafter to better characterise the nature of viral escape from these antibodies. Using magnetic beads coated with one of 16 different recombinant gp120 proteins it was confirmed that the C3-V4 response was aimed at a monomer-specific epitope in all four cases. In two instances these antibodies were shown to contribute to autologous neutralisation, while in a third the existence of quaternary structure specific antibodies that could not be adsorbed with monomeric gp120 made this link impossible. In the forth instance transfer of the C3-V4 region was shown to expose a normally occluded epitope in the CD4 binding site. This research also provided evidence for other epitopes for autologous neutralising antibodies in C3, overlapping with the CD4 binding site and V5. Lastly, by introducing relevant escape mutations into the parental recombinant gp120s and then comparing the ability of these proteins to adsorb out anti-C3 antibodies, it was shown that while these mutations conferred complete resistance to neutralisation they did not prevent the antibodies from binding to their respective epitopes. The extensive characterisation of C3-related epitopes such as those described in this research should no doubt contribute to the rational design of a gp120 based vaccine immunogen aimed at eliciting broad and potent neutralising antibody responses.

ABSTRACT The magnitude of the AIDS epidemic is well documented. It has been shown that Africa constitutes about 70 % of people infected with HIV worldwide. Efforts to control the AIDS epidemic have focused heavily on studies pertaining to the biology, biochemistry and structural biology of HIV and on the interactions between HIV proteins and new drugs. One of the most challenging problems in AIDS therapy is that HIV develops drug-resistant variants rapidly. Extensive research has been dedicated to designing resistance-evading drugs for HIV-1 protease (predominantly subtype B), which is crucial for the maturation of viral, structural and enzymatic proteins. There are 10 subtypes of HIV-1 within the major group of the virus, with subtype C accounting for about 95 % of infections in South Africa. Since HIV-1 antiretroviral treatment has been developed and tested against the B subtype, which is prevalent in North America, Western Europe and Australia, an important question relates to the effectiveness of these drugs against the C subtype. At this point, however, little is known about inhibitor-resistant mutations in the subtype C. The study, therefore, looked at the two active site mutations (V82A and V82F/I84V) in the South African HIV-1 subtype C protease (C-SA) emerging from the viral population circulating in patients. These mutations are well-characterized within the framework of the subtype B and are known to cause cross-resistance to most of inhibitors currently in clinical use. Protein engineering techniques were used to generate the V82A and the V82F/I84V variants. Comparative studies with the wild-type HIV-1 C-SA protease were performed. The spectral properties of the V82A and the V82F/I84V variants indicated no changes in the secondary structure in the respective variant proteins. Tryptophan and tyrosine fluorescence indicated a major difference in the intensities at the emission maxima for all three proteins. The fluorescence intensity of the V82F/I84V variant, in particular, was significantly enhanced indicating the occurrence of tertiary structural changes at/near the flap region. Both mutations did not impact significantly upon catalytic function. Both variants also had the same Km values comparable to that of the wild-type enzyme. The catalytic efficiencies and the kinetic constants were lowered 3.6-fold for the V82A mutation and 6-fold for the V82F/I84V mutation relative to the wild-type C-SA protease. Inhibition studies were performed using four inhibitors in clinical use (saquinavir, ritonavir, indinavir and nelfinavir). For the V82A variant, IC50 and Ki values for saquinavir and nelfinavir iv were not affected, whilst those for ritonavir and indinavir were 5- and 9-fold higher than the wild-type C-SA protease, respectively. Against the V82F/I84V variant, however, the inhibition constants were drastically weaker and characterized by IC50 and Ki ratios ranging from 50 to 450. Isothermal titration calorimetry (ITC) was also used to determine the binding energetics of saquinavir, ritonavir, indinavir and nelfinavir to the wild-type C-SA, V82A and V82F/I84V HIV-1 protease. The V82A mutation lowered the Gibbs energy of binding for the respective four clinical inhibitors by 0.4 kcal/mol, 1.3 kcal/mol, 1.5 kcal/mol and 0.6 kcal/mol, respectively, relative to the wild-type C-SA HIV-1 protease. The affinity of V82A HIV-1 protease for saquinavir, ritonavir, indinavir and nelfinavir (Kd = 1.85 nM, 2.00 nM, 12.70 nM and 0.66 nM, respectively, at 25 °C) was in the range of 2- to 13-fold of magnitude weaker than that of the wild-type C-SA protein. The clinical inhibitors exhibited the highest binding affinity to both the wild-type and the V82A enzymes, but were extremely sensitive to the V82F/I84V mutation. The V82F/I84V mutant reduced the binding of saquinavir, ritonavir, indinavir and nelfinavir 117-, 1095-, 474- and 367- fold, respectively. A drop in Kd values obtained for the V82F/I84V in association with saquinavir, ritonavir, indinavir and nelfinavir was consistent with a decrease of between 2.8 - 4.2 kcal/mol in ΔG, which is equivalent to at least 2 to 3 orders of magnitude in binding affinity. Taken together, thermodynamic data indicated that the V82A and V82F/I84V active site mutations in the C-SA subtype lower the affinity of the first-generation inhibitors by making the binding entropy less positive (unfavorable) and making the enthalpy change slightly less favorable.

PhD, Virology, Faculty of Health Sciences, University of the Witwatersrand / HIV-associated cardiomyopathy is a multifactorial disease with a broad spectrum of aetiologies that arise due to chronic immunosupression during HIV infection. The intricate relationship between HIV infection and cardiac co-morbidity was investigated with the aid of HIV-neutralizing aptamers. These synthetic nucleic acid ligands with antibody-like properties are molecular tools with multifunctional applications ranging from drug discovery to diagnostics and therapeutics. The advent of the HIV/AIDS pandemic has naturally married the field of HIV therapy and diagnostics with that of aptamer technology. By employing a HIV-1 neutralizing aptamer, named UCLA1, raised against the viral surface envelope glycoprotein 120, I dissected some of the pathways leading to cardiomyocyte apoptosis in a cell culture system. In chapter one I investigated the potential cytotoxic effects of UCLA1 by comparing it against a panel of 17 antiretrovirals (ARVs) in clinical use with the goal of establishing a safety portfolio geared towards its use as a therapeutic agent. Using cultured human cardiomyocytes and primary peripheral blood mononuclear cells (PBMCs), I selected some of the major biological markers of ARV-induced cytotoxicity and found no measurable deleterious effect, especially when compared to other ARVs used in the same study. In chapter two, the permissiveness of cardiomyocytes to HIV infection as well as the relationship between virus-host interaction and caspase-mediated apoptosis were investigated. Non-productive, receptor and tropism-independent infection was observed, which was arrested after the reverse transcription stage. However, interaction between the virus gp120 and the host’s CXCR-4 chemokine receptor preferentially activated caspase-9 triggering robust mitochondria-mediated apoptosis. A shift from mitochondrial-initiated, caspase-9 mediated to Fas-ligand initiated, caspase-8 mediated was observed when CM were co-cultured with HIV-infected MDM. UCLA1 protected against caspase-9 mediated vii apoptosis but not caspase-8 mediated. Finally in chapter three I provided answers for the shift in caspase activation by showing that supraphysiological levels of IL-1β and IL-6 during HIV infection of MDM augment the effects of tumor necrosis factor (TNF). These observation provide new insight into the complex pathophysiology of HIVCM and highlight the potential of UCLA1 as a novel therapeutic agent to fight HIV and some of its associated diseases.

MSc (Med), Faculty of Health Sciences, University of the Witwatersrand, 2011 / BACKGROUND: Human Immunodeficiency Virus (HIV) is a global epidemic with growing numbers of people on highly active anti‐retroviral therapy (HAART) programmes. Effectiveness of treatment needs to be monitored to ensure the uncompromised well being of patients. This is currently done using both Viral Load (VL) and CD4 cell counts for HAART initiation and follow‐up. Although VL is the best predictor of disease progression it is often too expensive for monitoring patients in resource‐limited settings. There is thus a need for a cheaper, more accessible alternative to monitor long term patient response to therapy. METHODS: This study evaluated the use of a recently described flow cytometric assay of CD38 expression (previously developed at the Johannesburg Flow Cytometry Reference Laboratory) in a cohort of HIV+ patients failing 1st line therapy, who were subsequently enrolled onto 2nd line HAART. CD38 and CD8 were “piggy ‐backed” onto the PLG/CD4 protocol and mean fluorescence intensity (MFI) of the CD8/38 expression was monitored longitudinally. Patterns of CD38 expression were compared to 1st line treatment observations to establish equivalence in the predictive power of CD38 expression of fluctuation in viral load on 2nd line treatment patients. In addition, the effect of sample age on assay accuracy was tested before implementation of the CD38 assay at a secondary testing site. RESULTS: The patterns observed in the cohort of 2nd line therapy patients mirrored patterns previously seen in 1st line therapy with 55% of patients showing a continuous decline in CD38 MFI that mimicked changes in VL. The remaining 33% of patients had non‐specific increases in CD38 MFI without concurrent increases in VL and one patient showed irregular VL and CD38 MFI (non‐responder). The CD38 assay showed acceptable accuracy and reproducibility up to 48 hours after venesection (%CV<5%). Implementation at the secondary testing site was successful with 98% similarity (%CV<5%) compared to the reference laboratory. CONCLUSION: CD38 monitoring of 2nd line therapy patients showed comparable patterns to observations in 1st line therapy patients. The assay proved stable over time and easy to implement at another PLG/CD4 testing facility. As such, the CD38 assay offers a cost‐effective, reliable real time supplementary test to long‐term VL monitoring of HIV infected patients on the national ART programme.

The integrin α4β7, which mediates the trafficking of T lymphocytes to the gut associated lymphoid tissue (GALT), a site of rapid HIV replication, has been described as an attachment factor for the HIV envelope protein gp120. While differences in binding affinity of early transmitting and chronic gp120s for α4β7 have previously been noted by others, this has not translated to differences in replication. We aimed to investigate what role this binding interaction has in HIV pathogenesis over time and determine the factors that influence α4β7 reactivity. Understanding the subsets on which α4β7 is expressed may indicate the phenotype of the first host cells infected by HIV. The level of expression of α4β7 on Th17 and Treg CD4+ T cells was of interest as both subsets are important in HIV immunity in the GALT and represented in both the GALT and genital mucosa. All-trans retinoic acid-activated CD4+ T cells isolated from whole blood of healthy donors were incubated with or without HP2/1 (anti-α4 monoclonal antibody) or Act-1 (anti-α4β7 monoclonal antibody) prior to adding virus. Sixty infectious envelope clones were prepared using matched envelope genes from 11 individuals in the CAPRISA 002 Acute Infection cohort representing the T/F virus and variants from 1-39 months post infection (p.i.). Replication was monitored by p24 ELISA over 10 days. ATRA-stimulated CD4+ T cells were subjected to intracellular and surface staining for flow cytometric analysis using a FACSAria to distinguish Th17 and Treg CD4+ T cells and to determine the expression of CD45RA, CCR5, p24, α4β7. The dependence on α4β7 for HIV subtype C replication changes over time and varies across individuals. In three individuals, dependence on α4β7 was higher using T/F viruses and decreased sharply during acute infection (1-3 months post-infection). α4β7 dependence showed an increasing trend in chronic infection over time which was slight in the first year. Factors that influence α4β7 reactivity include glycan distance from α4β7-binding motif, glycan density and length of the V1/V2 region of gp120. Several glycans positioned in conserved regions of gp120 including N234, N332 and N334 were present more or less frequently in viruses with high α4β7 reactivity. There was an association between high dependence on α4β7 for replication at transmission and those T/F viruses that have a S/PDI/V α4β7 binding motif as well T/F viruses found in individuals diagnosed with bacterial vaginosis during acute infection. Several cytokines in the cervicovaginal lavage (CVL) of these individuals during early infection (IL-8, IL-7 and IL-1α) positively correlated with α4β7 dependence for replication. Treg CD4+ T cells expressed slightly higher levels of α4β7 as compared to Th17 CD4+ cells. In addition, Th17 cells in this and other studies were shown to rapidly deplete in vitro following HIV infection while Treg CD4+ T cell were shown to expand. Despite this, Treg CD4+ T cells were significantly more permissive to infection as compared to Th17 CD4+ T cells. Collectively, these data suggest that there is a role for α4β7 in HIV pathogenesis and that the interaction is selected for during transmission by a number of bottlenecks, one of which is the presence of bacterial vaginosis and elevated expression of IL-7, IL-8 and IL-1α. Due to high levels of α4β7 expression and the ability to bind gp120, presence in both the genital tract and the GALT, regulation by ATRA similar to α4β7 upregulation, expansion following HIV infection and elevated permissiveness to acute infection, Treg CD4+ T cells may be a robust vector from the genital mucosa to the GALT shortly after transmission. This population of T cells may be more suited for this function than Th17 CD4+ T cells which are more susceptible to depletion either by HIV or bystander effects. As a result of the α4β7 binding motif being in a highly immunogenic region of gp120 and antibodies directed against this region associated with protection in the only effective vaccine trial to date, further understanding of the interaction between the virus and the integrin provides an opportunity for the development of future vaccine and therapeutic strategies.

PhD - Pathology / The entry stage of the HIV-1 viral life cycle has become a prime target for preventing HIV-1 infection. This has led to the development of a new class of anti-retroviral agents termed entry inhibitors, which have proven effective in vitro and in the clinic. These new agents target three different stages of entry, namely CD4 binding, coreceptor interaction with either CCR5 or CXCR4 and the fusion process. Here we studied isolates from patients with HIV-1 subtype C infection and the effectiveness of different coreceptor and fusion inhibitors in vitro. Further we examined resistance profiles to the first licensed entry inhibitor, T-20. In Chapter 2 we examined coreceptor usage of HIV-1 subtype C isolates and their sensitivity to CCR5 and CXCR4 inhibitors. Twenty-nine viral isolates with different coreceptor usage profiles were isolated from patients with advanced AIDS. The CCR5-specific agents, PRO140 an anti-CCR5 monoclonal antibody and RANTES, the natural ligand for CCR5 inhibited all 24 R5 isolates, while the two X4 and the three R5X4 viruses were sensitive to the CXCR4-specific inhibitor, AMD3100. The five X4 or R5X4 viruses were all able to replicate in peripheral blood mononuclear cells (PBMC) that did not express CCR5 confirming their ability to use CXCR4 on primary cells. When tested using coreceptor-transfected cell lines, one R5 virus was also able to use CXCR6, and another R5X4 virus could use CCR3, Bob/GPR15 and CXCR6. The R5X4 and X4 viruses contained more diverse V3 loop sequences with a higher overall positive charge, compared to the R5 viruses. Hence, HIV-1 subtype C viruses are able to use CCR5, CXCR4 or both for entry, and they are sensitive to specific inhibitors of entry via these coreceptors. In Chapter 3 we analyzed isolates from 10 acutely infected individuals, who were followed longitudinally for up to three years. Two of these individuals (Du151 and Du179) underwent a coreceptor switch and were studied more intensively. The other eight individuals retained CCR5 usage throughout the duration of the study. The initial 4 isolates from Du151 were able to utilize CCR5 but the later isolates were able to use both CCR5 and CXCR4 (R5X4). Du179 used both CCR5 and CXCR4 (R5X4) initially, but the later isolate was found to be monotropic and used CXCR4 (X4) exclusively. Viral isolates were tested for their sensitivity to small molecule inhibitors of CCR5, CXCR4 and the fusion process in a PBMC assay. All of the R5 isolates were sensitive to RANTES and PRO140 and insensitive to the two CXCR4 coreceptor inhibitors AMD3100 and T-140. There was a tendency for later isolates to become slightly less sensitive to the CCR5 inhibitors and more sensitive to the CXCR4 entry inhibitors. None of the R5X4 Du179 isolates were effectively inhibited by PRO140 and RANTES, but the X4 isolate of Du179 became sensitive to CXCR4 entry inhibitors. Both Du151 and Du179 underwent amino acid changes in their V3 sequences that included an increased charge associated with CXCR4 usage. This indicates that coreceptor switching can occur in subtype C infections and is associated with changes in the V3 loop. However, both Du151 and Du179 were subsequently found to be dually infected with another subtype C strain, which may account for some of the phenotypic and genotypic changes seen in these individuals including the appearance of CXCR4-virus variants. In Chapter 4 we explored two HIV-1 isolates (CM4 and CM9) able to use alternate HIV-1 coreceptors for entry (i.e. coreceptors other than CCR5 or CXCR4) on transfected cell lines. These isolates were tested for their sensitivity to inhibitors of HIV-1 entry on primary cells. Both isolates were from patients with cryptococcal meningitis, a severe AIDS defining condition. CM4 was able to use CCR5 and Bob/GPR15 efficiently in transfected cells. This isolate grew in D32/D32 CCR5 PBMC in the presence of AMD3100, indicating that it used a receptor other than CCR5 or CXCR4 on primary cells. It was insensitive to the CCR5 entry inhibitors RANTES and PRO140, but was partially inhibited by vMIP-1, a chemokine that binds CCR3, CCR8, Bob/GPR15 and CXCR6. The coreceptor used by this isolate on primary cells is thus currently unknown. CM9 used CCR5, CXCR4, Bob/GPR15, CXCR6 and CCR3 on transfected cells and was able to replicate in the presence of AMD3100 in D32/D32 CCR5 PBMC. It was insensitive to vMIP-1, eotaxin and I309 used individually, but was inhibited completely when vMIP-1 or I309, the ligand for CCR8, were combined with AMD3100. These results strongly suggest that this isolate can use CCR8 on primary cells. Collectively these data suggest that some HIV-1 isolates can use alternate coreceptors on primary cells, which may have implications for strategies that aim to block viral entry using coreceptor inhibitors. In Chapter 5 we examined the effectiveness of T-20 to inhibit HIV-1 subtype C isolates. T-20 blocks the fusion stage of the viral entry cycle and it is the first entry inhibitor to be licensed for clinical use. T-20 consists of 36 amino acids and was designed based on the HR-2 region of HIV-1 subtype B. A total of 23 HIV-1 subtype C isolates were tested for their ability to replicate in the presence and absence of T-20. This included five isolates with multiple genotypic drug resistance mutations to reverse transcriptase and protease inhibitors. Among the 23 subtype C isolates there were 10-16 amino acid changes in the 36 amino acid region corresponding to T-20. However, all isolates were effectively inhibited by T-20 at 1 mg/ml, including the 5 isolates resistant to other anti-retroviral drugs. The gp41 region was sequenced and the HR-1 and HR-2 amino acids analyzed. All isolates had the amino acids GIV at positions 36-38 in gp41, which are associated with sensitivity to T-20. One X4 had a GVV motif but this did not affect its sensitivity. Thus, T-20 inhibited subtype C viruses despite significant genetic differences in the HR-2 regions of subtypes B and C. These data suggest that T-20 would be highly effective in patients with HIV-1 subtype C infection including those failing existing anti-retroviral drug regimens. In Chapter 6 we examined the in vitro resistance patterns of HIV-1 subtype C to T-20. Resistance to T-20 is a consequence of persistent exposure to the antiretroviral peptide. To establish if patterns of resitance to T-20 were similar to resistance mutations occurring in subtype B viruses, 11 subtype C and 4 subtype B viruses were passaged in the presence of increasing concentrations of T-20. The subtype C isolates showed varying levels of replication at 1 mg/ml T-20 by day 18, but by day 29 all replicated efficiciently at 10 mg/ml T-20. All isolates showed evidence of genotypic changes in gp41 HR-1 following exposure to T-20 that included G36S/E/D, I37T, V38M/A/L/E, N42D, N43K/S, L45R/M and A50T/V. Five viruses had compensatory changes in the HR-2 region, which corresponds to the T-20 sequence, and two isolates had changes in the V3 region. Mutational patterns among the 4 subtype B viruses were similar to those for subtype C and those previously published in the literature. These data indicate that in vitro resistance to T-20 develops rapidly among HIV-1 subtype C isolates. In general, mutational patterns for subtype C were similar to those described for subtype B, suggesting that the mechanism of action for T-20 is similar for HIV-1 subtype B and C isolates. Observations from these studies indicate that HIV-1 subtype C predominantly use the CCR5 coreceptor to enter cells. CXCR4 usage is rare compared to other subtypes, although such isolates are found in patients with advanced AIDS. The two cases of coreceptor switching reported here were dually infected. Subtype C isolates were sensitive to coreceptor and fusion inhibitors except for two isolates able to utilize alternate coreceptors. However, alternate coreceptor usage is very rare and unlikely to impact on the utility of coreceptor inhibitors. Given the propensity for CCR5 usage this may imply that CCR5 coreceptor inhibitors may be more effective in countries where HIV-1 subtype C predominate. Entry inhibitors may be useful for prevention and treatment strategies and have the potential to provide sterilizing immunity. These agents could be used as microbicides and as an adjunct to existing antiretroviral therapies for use in HIV-1 subtype C infected individuals. However resistance to entry inhibitors can emerge and should be used in combination with other antiretrovirals to minimize this outcome. While entry inhibitors provide a new line of defence against HIV-1, their cost may prevent their use in developing countries in the immediate future. Nevertheless, this is the first comprehensive study of the sensitivity of HIV-1 subtype C isolates to entry inhibitors providing a data-driven rationale for their use in individuals infected with HIV-1 subtype C.

ABSTRACT Introduction Circulating levels, and the ex vivo production, of the chemokines CCL3, CCL4, CCL5, CXCL8 and CXCL12 (known to play an important role in the pathogenesis of either human immunodeficiency virus type 1 (HIV-1) or tuberculosis (TB)) were examined in the context of both single infections with HIV-1 or Mycobacterium tuberculosis (Mtb) and coinfection with both organisms. We hypothesised that CCL3L1 gene copy number (known to affect CCL3 production, associated with susceptibility to and disease progression of HIV-1) would be associated with mother-to-child transmission (MTCT) of HIV-1, and that the IL8-251T→A single nucleotide polymorphism (SNP) (associated with enhanced CXCL8 production and susceptibility to TB in African Americans) would be highly represented in the South African Black population. Methods Samples used included (i) plasma, DNA samples and cell culture supernatants from control, HIV-1, TB and HIV-1/TB groups, (ii) DNA samples from mothers and their infants (grouped as HIV-1 exposed-uninfected, infected in utero, or infected intrapartum), and (iii) DNA samples from a populationbased study cohort. Chemokines were quantified by enzyme-linked immunosorbent assay (ELISA), CCL3L1 gene copy numbers were determined by real-time polymerase chain reaction (PCR), and a real-time PCR method was developed for identification of the IL8-251T→A SNP. DNA sequencing was used for confirmation. Results We found reduced ex vivo chemokine production in response to phytohaemagglutinin (PHA) together with increased plasma levels of chemokines in HIV-1 and TB patients. In contrast to that seen in Caucasians (median CCL3L1 copy number of 2), in Black individuals (median CCL3L1 copy number of 5) circulating levels of CCL3 did not correlate with CCL3L1 gene copy number; in addition, a high proportion of Black individuals were found to have CCL3L1 copy numbers below their population-specific median. Using MTCT as a model for studying HIV-1 transmission, infants who became infected with HIV-1 had significantly reduced CCL3L1 gene copy numbers. IL8-251A allele frequencies were found to be 0.41 for Caucasian groups, and 0.85 for Black groups; due to study limitations, the possible association of IL8-251T→A with TB susceptibility could not be addressed. Discussion The increased plasma levels of chemokines seen in HIV-1 and TB, likely due to chronic immune activation in vivo, may result in T cell anergy which in turn might be the cause of reduced PHA-stimulated ex vivo chemokine production. Our results suggest that Black South Africans may be at particularly high risk for acquiring HIV-1 (at least with respect to CCL3L1 gene copy number), and further imply the presence of other genetic polymorphisms which may influence plasma CCL3 levels. In addition, the high IL8-251A allele frequency (if indeed associated with TB in South African populations) in Black individuals suggests a greater risk for infection with Mtb. It will be important, in larger studies, to gain a more in-depth understanding of the relationships between host genotype and chemokine production phenotype, and to relate these measures to infection outcomes. Conclusions Together, these results highlight the importance of gaining an understanding of the effects of host genotype on the development of innate and acquired immunity to HIV-1 and TB, which will be key in the design of efficient therapies and prevention strategies.

Background In 2010, 2.2 million adolescents were living with HIV (Human Immunodeficiency Virus) worldwide. This study aimed to describe the socio-demographic and clinical characteristics of the adolescents (10-19 years old) initiating anti-retroviral treatment (ART) and to investigate characteristics that are associated with virological failure in adolescents on ART. Methods This was an analysis of adolescents initiating ART from June 2004-2010 at the Hlabisa Treatment and Care Programme in KwaZulu-Natal, South Africa. Data was collected from two datasets at Africa Centre for Health and Population Studies. Time to outcomes of death and lost to follow up (LTFU) were quantified using Kaplan-Meier estimates. The outcome was virologic response (< 70copies/ml) after at least 6 months on ART and the associations with an unsuppressed viral load were investigated using multivariable logistic regression. Results 543 adolescents, median age 15 years (IQR 12-18), initiated ART; 67.8% (368) were females. Age at treatment initiation showed a bimodal distribution, with a peak at 11 years and another at 17-19 years; 61 females aged 16-19 years initiated ART whilst pregnant. At baseline, median CD4 count was 152 cells/μl (IQR 72-251), 392 (72.2%) had prior TB and 129 (23.8%) a weight-for-age z-score ≤ -2 (i.e. were under-nourished). Numbers of adolescents starting ART increased from 53 in the years 2004-2006 to 196 in 2010. Overall mortality was 36.5 per 1000 person years (95% CI 27.2 - 48.8); LTFU 98.8 per1000 person years (95% CI 82.8-118). Adjusting for age and gender, LTFU was significantly higher in females initiating in late adolescence (15-19 years) (p<0.001) and 24 (39.3%) of those initiating ART whilst pregnant were LTFU. The first viral load after initiation was taken at a median time of 11.25 months (IQR 7.78-16.20). Of the 364 adolescents with a viral load result after at least 6 months of ART, 119 (32.7%) had an unsuppressed viral load (95% CI 27.9- 37.5). Adolescents who initiated in the year 2010 were found to have less odds of an unsuppressed viral load compared to those who initiated between 2004 and 2006 [adjusted Odds Ratio (aOR) 0.29 (95% CI 0.11-0.79)]. Those who had the first viral load test done after > 30 months of ART had higher odds of an unsuppressed viral load compared to those tested after 6-12 months of ART [ aOR 6.88 (95% CI 1.29-36.66)]. Conclusion Despite the yearly increase in adolescents initiating ART, good virological responses can be obtained through increased ART support to both individuals and health care providers. Timely viral load monitoring identifies those in need of increased adherence support on ART and may result in good virological responses. Recommendations Adolescents on ART are a vulnerable group that requires special attention to improve clinical and virological outcomes. Adolescent friendly ART clinics may be useful in providing this service and mitigate the high attrition rates of those on treatment for HIV. Public health awareness campaigns on HIV and its treatment may have a positive impact on virological response to ART and therefore campaigns targeting adolescents must be intensified. Early virological testing after 6 months on ART to monitor treatment responses helps to identify those with sub-optimal response to ART and reduce the progression to virological failure and drug resistance to anti-retroviral drugs.

Student Number : 0114163J - PhD thesis - Faculty of Health Sciences / HIV-1 gains entry into host cells by binding to CD4 and a coreceptor, predominantly CCR5 or CXCR4. Viruses that use CCR5 are termed R5, those able to use CXCR4 are termed X4 while viruses able to use both coreceptors are referred to as R5X4. Accelerated CD4 decline and disease progression within an infected HIV-1 subtype B infected individual is often associated with the emergence of viruses able to use CXCR4. However, CXCR4 coreceptor usage appears to occur less frequently among HIV-1 subtype C viruses, the most predominant strain circulating globally, including South Africa. The aim of this study was to investigate the genetic determinants of CXCR4 usage in HIV-1 subtype C isolates. The V3 region of the envelope glycoprotein is the major determinant of coreceptor usage. In Chapter 2, 32 subtype C isolates with known phenotypes (16 R5, 8 R5X4 and 8 X4 isolates) were assessed using a subtype C specific V3-heteroduplex tracking assay. Results indicated that there were sufficient genetic differences to discriminate between R5 viruses and those able to use CXCR4 (both R5X4 and X4). In general, R5 isolates had a mobility ratio >0.9 whereas CXCR4-using isolates were usually <0.9. Sequence analysis of the V3 region showed that CXCR4-using viruses were often associated with an increased positive amino acid charge, insertions and loss of a glycosylation site, similar to HIV-1 subtype B. In contrast, where subtype B consensus V3 has a GPGR crown motif irrespective of coreceptor usage, all 16 subtype C R5 viruses had a conserved GPGQ sequence at the tip of the loop, while 12 of the 16 (75%) CXCR4-using viruses had substitutions in this motif, most commonly arginine (R). Thus, the rare occurrence of CXCR4-using viruses in subtype C may be due to the highly conserved nature of the GPGQ crown that may limit the potential for the development of X4 viruses. The usefulness of available genotype-based methods for predicting viral phenotypes in subtype C was explored in Chapter 3. Results indicated that commonly used prediction methods could detect R5 viruses, but were not very sensitive at identifying X4 viruses. We therefore developed a subtype C specific predictor based on position specific scoring matrices (PSSM). Similar methodology, as used in developing the subtype B PSSM, was applied on a training set of 280 subtype C sequences of known phenotype (229 NSI/CCR5 and 51 SI/CXCR4). The C-PSSM had a specificity of 94% (C.I. [92%-96%]) and sensitivity of 75% (C.I. [68%-82%]), indicating that the C-PSSM had improved sensitivity in predicting CXCR4 usage. This method also highlighted amino acid positions within V3 that could contribute differentially to phenotype prediction in subtypes B and C. A reliable phenotype prediction method, such as the C-PSSM, could provide a rapid and less expensive approach to identifying CXCR4 variants, and thus increase our knowledge of subtype C coreceptor usage. In Chapter 4 we examined the genetic changes in full-length gp160 envelope genes of 23 sequential isolates from 5 patients followed for two to three years. Three of the patients' isolates used CCR5 at all time points while 2 patients underwent a coreceptor switch with disease progression. The genetic changes observed over time indicated changes in length of variable loops particularly the V1, V4 and V5 and shifting N-glycosylation sites, particularly in the 2 patients that used CXCR4. Changes in the V3 were only noted in the 2 patients’ that used CXCR4 which included substitutions of specific amino acids including those in the crown and increased amino acid charge in the V3 region. Both of these patients were dually infected suggesting that recombination may contribute to the rapid emergence of X4 viruses. The in vitro and in vivo development of CXCR4 usage was analysed in a pediatric patient that experienced a coreceptor switch during disease progression (Chapter 5). Biological and molecular clones were generated and the V1-V5 regions sequenced. Analyses of the V3 region indicated that the evolution to CXCR4 usage happens in a step-wise manner that included increased charge and changes in the crown motif. The intermediate variants with predicted dualtropism were also associated with increased V1-V2 lengths, suggesting that other regions may contribute to coreceptor switching. Furthermore, the development of CXCR4 usage within this patient was due to two mutational pathways, in which one resulted in R5X4 viruses and the other X4 variants. In Chapter 6, the impact and treatment of acute TB on HIV-1 diversity in co-infected patients was investigated, specifically to determine the genetic characteristics of the viral populations present before, during and after TB treatment. Plasma samples from 18 HIV-1 infected patients were analysed using the C2V3 region, six of whom showed a high degree of variation using a V3-HTA and were selected for further analyses. All patients were predicted as R5 with no evidence of coreceptor switching over time. There was no correlation between the degree of genetic diversity and viral load, although both showed fluctuations over time. Phylogenetic and pairwise genetic distance analysis indicated that there was amplification of existing variants in 3 patients while in the other 3 patients there were dramatic shifts in viral populations suggesting selection of viral sub-populations over time. Thus in some co-infected patients, TB can affect HIV-1 genetic heterogeneity although there was no evidence of a shift towards CXCR4 usage despite the presence of an AIDS defining illness. Observations in this study have shown that the V3 region is the major determinant of coreceptor usage within HIV-1 subtype C, similar to HIV-1 subtype B. Characteristics such as increased charge length variability of the V3 region and loss of the glycosylation site within this region are associated with CXCR4 usage. The limited number of X4 viruses in subtype C does suggest some restricting mechanisms for CXCR4 usage. In this study we looked at genetic determinants and found that the rare occurrence of CXCR4-using viruses in subtype C, may be due to the highly conserved nature of the GPGQ crown that may limit the potential for the development of subtype C X4 viruses. Furthermore, the development of CXCR4 usage happened in a step-wise manner, with R5X4 viruses intermediates, in which an increased V1-V2 was observed suggesting that other regions within the envelope protein do contribute to coreceptor usage. Thus, regions such as V1-V2 and V4-V5 did contribute to coreceptor usage, but the V3 region remained the most important determinant of coreceptor usage in HIV-1 subtype C isolates. Collectively these findings have provided important data on the genetic determinants of CXCR4 usage in HIV-1 subtype C and an understanding of how they might evolve within a patient.

Abstract will not load on to DSpace