Heterosexual transmission of multiple highly conserved viral variants in HIV-1 subtype C-infected seronegative women

Viral population diversity was assessed in samples collected from five HIV-infected women who were RNA positive and antibody negative. Similar to studies in men, highly conserved viral variants were detected (mean nucleotide diversity of 0.11% for p17p24, 0.32% for C2C3). In two individuals diversity was uncharacteristically lower inenvC2C3 than ingagp17p24, suggesting selection inenv at this very early stage of infection. Although transmitted HIV-1 diversity has implications for the design of interventions such as vaccines and microbicides, this has been understudied, particularly in women. The majority of available data on virological determinants impacting on sexual transmission are derived from investigations on recently infected men showing homogeneous viral population transmission, even when donors harboured genetically complex viral populations [1–5]. The limited number of studies performed on women suggest sex differences, showing that women could be more susceptible to infection with multiple HIV-1 variants compared with men [1,6–9]. Most transmission studies have investigated HIV antibody-positive individuals; however, these may not accurately reflect the original transmitted population as diversification may occur from the time of infection to seroconversion as a result of host selective pressures and high viral turnover [10]. This study characterized HIV-1 env and gag populations during the pre-seroconversion stage in heterosexually infected South African female sex workers. These samples provide unique and valuable insights into the mechanisms of transmission and early replication during primary subtype C HIV-1 infection of women. Samples were obtained from a female sex worker cohort screened monthly for seroconversion [11]. The date of infection was estimated as the midpoint between the last RNA-negative and the first RNA-positive sample, except for Du23 in which the date of infection was estimated as 14 days before the first RNA-positive sample. For our retrospective study, samples from two visits before the seroconversion sample were screened for HIV RNA using the nucleic acid sequence-based amplification (NASBA) viral load assay. RNA was extracted and a reverse transcriptase–nested polymerase chain reaction (PCR) was performed for amplification of both the env C2C3 (511 base pairs; bp) and gag p17p24 (621 bp) regions. PCR-based limiting dilution assays were performed to ensure that the number of template copies in a single reverse transcriptase–nested PCR reaction were sufficient to ensure that the analysis of 20–40 colonies would be representative of the majority of viral populations (minimum of 30 copies) [12,13]. The amplifiable copy number per sample ranged from more than 32 (n = 1) to over 320 (n = 4) was calculated using QUALITY (Quantitation Using A LImiting diluTion assaY) [13]. Env C2C3 and gag p17p24 diversity was assessed using the heteroduplex tracking assay (HTA) [14]. Two clones representing each HTA migration pattern were sequenced and sequences were analysed using distance-based phylogenetic methods. Intra-individual diversity was calculated by adjusting the sequence ratio to reflect the migration pattern frequency and DNA distances were expressed as a percentage of nucleotide diversity using the Kimura 2-parameter model. The phenotype was predicted using two different methods [15,16]. Five HIV-seronegative women (Du23, 114, 145, 204 and 421) were identified as HIV-1 RNA positive, all of whom seroconverted within a month of HIV-1 RNA identification. Samples were collected from individuals infected for an estimated 7 days (Du114, Du204) to 14 days (Du23, Du45, Du421). Characteristic of acute infection, the plasma viral loads were high, ranging from 1.6 × 104; 3.9 × 104; 9.3 × 104; 5.6 × 105 to 5.1 × 106 copies/ml for Du23, 145, 204, 114 and 421, respectively. Four out of five individuals produced HTA heteroduplexes exhibiting very little to no shifts (Fig. 1a,c) in both the p17p24 as well as the C2C3 region. In one individual, Du204, four different p17p24 intermediary shift heteroduplex migration patterns were observed (Fig. 1b). In the C2C3 region, Du145 exhibited two different heteroduplex migration patterns (Fig. 1d), the result of a 6 bp insertion (intermediary shift) and 99 bp deletion (large shifts) in 5 and 15% of the viral population, respectively. Unlike C2C3, p17p24 sequence analysis showed no insertions or deletions.Fig. 1. Heteroduplex tracking assay illustrating intra-individual diversity of the p17p24 region (a and b) and C2C3 region (c and d) demonstrating infection with highly homogenous viral populations (a and c) and infection with divergent viral populations (b and d).: Lane P contains the probe and lane TP the total viral population annealed to the probe. The following 20 lanes are clones hybridized to the intra-individual probe. Empty arrows indicate intermediary shifts in mobility and filled arrows indicate large shifts in mobility. p17p24 heteroduplex separation was enhanced by the addition of 20% urea. (a) Du23 p17p24 heteroduplex tracking assay (HTA) representative of an infection highly homogenous in gag. (b) Du204 p17p24 HTA representative of an infection with four different gag variants. (c) Du114 C2C3 HTA representative of an infection highly homogenous in env. (d) Du145 C2C3 HTA representative of an infection with two different env variants.The mean intra-individual p17p24 DNA distances ranged from 0.06 to 0.2% (overall mean 0.11%), with a maximum nucleotide distance of 1.2%. The mean C2C3 intra-individual diversity ranged from 0.02 to 0.6% (overall mean 0.32%), with a maximum distance of 1.8%. Interestingly, in Du114 and Du204, mean intra-individual diversity was uncharacteristically lower in C2C3 (0.02 and 0.1%, respectively) than in p17p24 (0.08 and 0.2%, respectively). All env and gag sequences grouped within subtype C and no evidence of dual infection was detected (results not shown). The phenotype was predicted to be R5 in all cases. In conclusion, recently infected women in the preseroconversion period of infection harboured highly homogeneous, but not identical, viral populations. The genetic distance between viruses infecting an individual exceeded the expected rate of diversification considering the short time from transmission (as much as 1.8% in C2C3 and 1.2% in p17p24) and indicates selective intra-person transmission of multiple closely related variants. The viral populations in recently infected women in this study were highly homogenous compared with other studies in women. This region of the genome is quite heterogeneous; our study showed a mean inter-individual genetic diversity of 8.6% for C2C3 and 7.1% for p17p24. A direct comparison of intra-individual diversity from this study to a similar investigation in women showed the overall median C2C3 difference of 0.4% for our study, compared with 1.2% in the Kenyan study [6,7]. It would be of interest to extend the current study to include the V1 to V5 region, as this would enable more extensive comparisons with other studies carried out in women and men [17]. Whereas the choice of sample (that is DNA versus RNA) may have influenced the results, it is unlikely to have influenced the conclusion of this study, as the time available for diversification of archival DNA is short. There are multiple host and viral factors that could account for the observed differences in transmitted variants between these studies: the presence of ulcerative sexually transmitted infections, contraception use [18], as well as differences in the predominantly circulating subtypes, A and D in Kenya, and C in South Africa. Two women, sampled within 7 days post-infection, had unexpected lower diversity in env compared with gag, suggesting that very early selection forces were focussed on properties in envelope. This supports cell tropism or viral fitness as playing a major role in the selective transmission of variants resulting in reduced env diversity. Early cytotoxic T-cell lymphocyte immunological selection pressure could have resulted in diversification in gag. It is likely that selective transmission and selective amplification both contribute to this early viral population selection and restriction. An understanding of the mechanisms governing transmission could assist in the planning and design of much needed interventions such as vaccines and microbicides. Acknowledgements The authors are very grateful to the study participants and the field research team. Informed consent was obtained from all participants and human experimentation guidelines of IRB of the Nelson R Mandela School of Medicine of Natal University (137/95), SA and University of Cape Town were followed (003/99; 004/99).