Simon Mallal

Background: Structural and functional studies show that abacavir (ABC) binds non-covalently to HLA B*57:01 and alters the repertoire of peptide presented, some of which may re-stimulate memory CD8+ T cells to cause ABC hypersensitivity. We questioned whether the binding of ABC to HLA-B*5701 might also impair an HLA-B*5701 restricted CD8+ T cell antiviral response. Methods: 52 putative B95-8 Epstein-Barr Virus (EBV) strain HLA-B*57:01 restricted epitopes were identified using NetMHC 3.2 (binding score <400 nM). 7 healthy adult HLA-B*57:01+ donors were screened with EBV peptide pools followed by confirmatory enzyme-linked immunosorbent assay (ELISpot) with individual EBV peptides and avidity testing to 0.001 g/mL peptide. In addition, 4 HLA-B*57:01+ and 31 negative healthy donors were screened with 20-mer peptides overlapping with 15 amino acids from EBNA3A, 3B, BZLF1, and BMLF1. 4 short-term lines were generated from 4 HLA-B*57:01+ ABC-naīve donors by pulsing peripheral blood mononuclear cells (PBMC) with 20 μg/mL VSFIEFVGW for 1 hour, washing and then 9 days culture in RPMI-1640 media containing 10% FCS and 10% T-Stim media and 100 U/mL IL-2. The resulting cell lines contained 5% to 35% peptide specific CD8+ T cells. Cell lines or PBMC taken from HLA-B*57:01+ donors were tested for peptide specific responses ąABC at a range of concentrations previously shown not to affect cell viability, by intra-cellular cytokine re-stimulation assays (ICS) or ELISpot. Peptide concentrations of 0.01, 0.1, 1, and 10 g/mL were used in the ICS assay and 1 g/mL in the ELISpot assay. Expression of interferon (IFN)-γ in the presence of increasing concentrations of ABC relative to baseline (no ABC) and the respective ratios were analyzed using a linear mixed effects model with nesting for multiple peptide concentrations per individual. Results: 4 HLA-B*57:01-restricted epitopes were confirmed of which 2 were independently detected using overlapping peptides: VSFIEFVGW from EBNA3A and VAAHPEIGAW from BMLF1. Responses to VSFIEFVGW were of high avidity and detectable in all 11 donors confirming it as an immunodominant peptide. IFN-γ responses to VSFIEFVGW relative to baseline were reduced relative to baseline with increasing abacavir concentrations using both ICS (Fig a) and ELISpot assay (Fig b). Conclusions: A new immunodominant HLA-B*57:01-restricted EBV epitope was identified. ABC was shown to impair the CD8+ T cell response to this epitope in a dose-dependent manner, suggesting that drugs that bind non-covalently to HLA molecules may modulate immune responses.

HIV/HCV co-infection is associated with diminished HCV specific T-cell responses and with higher HCV RNA levels. During successful HAART, HCV specific T-cell responses increase and HCV RNA levels decrease. However, HCV can evade the host immune response through accumulation of escape variants. Because little is known about the development of escape mutations in co-infected individuals with onset of HAART, we investigated the influence of increased immune pressure during HAART on HCV sequence evolution, in particular within T-cell epitopes. Furthermore, we examined HCV evolution at known drug-resistance sites to the new anti-HCV drugs.

HCV genotype-1 bulk sequences covering the immunogenic HLA-class-I epitopes (HLA-B*0801, HLAA* 0101 in NS3 and HLA-B*2705 in NS5B) were analyzed in HLA-carriers and in HLA-non-carriers before and during HAART (n=69). To study sequence evolution within these epitopes and the occurrence of drug resistance mutations in quasispecies present at >=1%, we performed ultra-deep sequencing using the FLX-454 Roche technology on five subjects before and on HAART.

Bulk sequencing analyses indicated a significant accumulation of immune escape variants in HLA-carriers (16% of sites) compared to non-carriers (only 6%) before the initiation of HAART (p=0.009). These mutations were maintained during HAART (HLA-carriers: 8%; HLA-non-carriers: 6%). The emergence of additional escape mutations with the onset of HAART was rare (7% of sites) and reversions of escape mutations were not seen. However, by utilizing the deep sequencing technique, minor quasispecies could be detected in 88% of 225 analyzed amino acid sites. Within the HLA-B*2705 epitope only, mixtures were found in 24 of 105 sites at the nucleotide level that were not found in bulk sequencing. Furthermore, investigating 35 sites associated with drug resistance in the protease and polymerase genes, six drug resistance mutations could be detected with FLX as minority species but not with Sanger sequencing.

In HCV/HIV co-infected individuals, the increase in HCV specific T-cell pressure during HAART appears to have only a modest effect on HCV escape. However, more sensitive sequencing techniques do identify additional escape variants at lower levels in these individuals suggesting changes in the host’s immune pressure on HCV following the commencement of HAART. It remains to be determined if these low frequency viral escape species affect overall HCV-specific immune responses or disease outcome. Furthermore, deep sequencing identifies biologically relevant low-level drug resistant mutations in HCV treatment naïve subjects that are not detected by conventional population-based sequencing.

The immune responses which drive the earliest selection of viral adaptations in heterosexual transmission provide insights into the responses most important for acute natural control of founder viruses, which should be harnessed by preventative vaccines. Similarly, those viral adaptations which revert early after transmission will reflect sites of strong constraint against escape from natural and vaccine-induced responses. Here we characterize in detail the virological and immunological events in an individual who presented with HIV (subtype A) infection only days after epidemiologically proven sexual transmission, and pre-seroconversion (Fiebig stage II). Quantitative changes to intra-patient viral sequences were evaluated with deep pyrosequencing of full-length HIV-1 genomes in both donor and recipient. Longitudinal Sanger and FLX 454 sequencing of HIV-1 was performed on 8 plasma samples from day 13 to 467 post-transmission. HIV-specific CD8 T cell responses were evaluated in 9 PBMC samples (day 35-656) using HLA-class 1 matched peptides in the IFNg ELISpot assay. The donor and recipient shared an HLA-C allele (HLA-C*04) but were discordant at all other HLA class I loci. Evolution of escape in an immunodominant HLA-B*1503-restricted Nef WL9 epitope was detected on day 264 and was present in 41% of viral quasi-species at day 264. T cell responses were detected prior to sequence change and decreased thereafter. Sequence change and CD8 T cell responses were also detected in relation to other published HLA-restricted epitopes. However, early reversion and subsequent re-selection was observed at position 133 of Nef which is not within a known epitope but corresponded to a HLA-C*0401 associated polymorphism in population-based studies. Overall, IFNg responses were detected to 3 of 28 HIV-1 peptides tested on day 35 but broadened to 20 responses while infection became established and reached peak viral load on day 383. T cell responses declined after commencement of antiviral therapy. Adaptive viral changes and reversions in early HIV-1 infection occur at novel sites not captured in immunological studies of chronically infected individuals. This data suggests the presence of novel immune hierarchies and escape mutations in early infection, which may be applied to HIV vaccine design against founder viruses.

Objectives: To determine whether medication adherence and HIV RNA viral load differed between patients grouped by treatment with PIs or NNRTIs on once daily (OD) vs twice daily (BD) antiretroviral therapy (ART). Methods: Clinical information from the WA HIV Cohort is collated electronically. We previously validated a “medication score” calculated as percentage of prescribed tablets taken over the past 4 weeks. Scores are collated with clinical markers in the database and hard copy output allows visualisation of longitudinal treatment and adherence data. Data was collected on 415 patients on OD or BD regimens (2900 visits: mean visits/person=7; PI: 298/1380(22%) OD; NNRTI: 328/1520(22%) OD), from June 2003 to Dec 2006. Analyses of medication adherence (score of 100%) and viral suppression (<50 cps/ml) were undertaken by longitudinal logistic modeling to accommodate unbalanced follow-up. Results: Complete adherence was reported at 75% of visits. In patients on PI-based therapies there was no significant difference in adherence in OD compared with BD regimens (OR=1.07, p=0.7). In contrast, for patients on NNRTI therapy adherence was significantly better amongst those on OD regimens (OR=2.96, p<0.0001). Viral suppression was observed at 80% of visits where patients had been on HAART for at least 6 mths. The strongest predictor of VL>50 was CD4 count <200 (p<0.0001). Amongst patients on a NNRTI-based regimen neither calendar time nor number of daily doses were associated with viral suppression (p>0.7). Amongst patients on PIs, suppression was reduced for OD dosing, particularly in earlier years (pre-2006: OR=1.85, p=0.04; 2006: OR=1.47, p=0.3). Conclusions: Once daily dosing was correlated with improved adherence in patients on NNRTI regimens but not in those on PIs. Patient satisfaction with OD ART dosing may have resulted in greater adherence in the NNRTI group, however, in the PI group patient and prescribers’ preference may be for the extra ‘forgiveness’ that a twice daily regimen can provide.