Fig 1.
Viral loads in cervical and plasma samples early in infection.
Viral RNA copies/ml were quantified in cervical (n = 29; red symbols) and plasma (n = 67; blue symbols) samples collected between 0–7 months of infection. Analysis of significance was done using a two-tailed Mann-Whitney test (p<0.001). Each symbol represents an individual subject.
Fig 2.
Higher HIV-1 diversity in the female genital tract than in blood within 7 months of infection.
The HIV-1 C2-V3-C3 env region present in cervical (red symbols and red boxplot) and plasma (blue symbols and blue boxplot) samples were PCR amplified and analysed by NGS. The average p-distance (s/nt) of HIV sequences within C2-V3-C3 region were calculated using MEGA 6 with sequences below 200bp excluded from analysis (A). The number of unique sequences within each sample were identified using online Galaxy software to collapse all sequence reads for each sample (B). Frequency of clones in cervical (C) and plasma samples (D) were determined from the number of unique clones. Average distance (E) and the number of unique sequences (F) in cervical and plasma samples grouped by subtype of infection. Analysis of significance was done using a two-tailed Mann-Whitney test. Each symbol represents an individual subject.
Fig 3.
HIV-1 diversity in cervical samples and blood at very early and early infection.
Average nucleotide p-distance (s/nt) and the number of unique sequences were analyzed following NGS. Average distance and number of unique sequences of cervical (red symbols and boxplot) and plasma (blue symbols and boxplot) samples were grouped into very early (0–3 months) and early infection (3–7 months) (A-D). Analysis of significance was done using a two-tailed Mann-Whitney test. Each symbol represents an individual subject.
Fig 4.
Envelope sequence diversity in the cervical tract and matched plasma samples early in infection.
Paired cervical and plasma samples (n = 21) from individuals infected between 0–7 months were compared for C2-V3-C3 env diversity by calculating genetic distance using MEGA6 software (A). The number of unique sequences was derived by collapsing the number of total sequences (range = 93–2409 sequences) for each individual sample using online Galaxy software (B). Paired cervical and plasma sequences were analyzed for selective pressure by estimating the dN/dS ratio in each sample using SNAP v2.11 and plotted (C). Sequences of dS = 0 or dN = 0, resulting in dN/dS = 0 were included as zeros or no evolution. Statistical significance between matched cervical and plasma samples were determined using a two-tailed Wilcoxon matched-pairs signed rank test. The average of 100 maximum likelihood bootstrapped trees of nucleotide sequences were generated with MEGA6, rooted to the SIVcpz CD.90.ANT sequence, and visualized with FigTree 1.4.2 to highlight sequence heterogeneity. The C2-V3-C3 env sequences from cervical (C) and plasma (P) paired samples aligned to the reference HIV sequences from the Los Alamos Sequence Database (D) and following trees provide bootstrapping values.
Fig 5.
Sequence diversity in plasma influences CD4 T cell decline.
Viral load and CD4 T cell slopes were plotted from up to 7 year longitudinal sampling of study subjects. Each slope was derived from a number of time points between viral set point to initiation of cART (S8 Fig). Spearman Rank correlation were calculated and presented for the average p-distance and viral load change for cervical (A) and plasma samples (B) and for the CD4 T cell loss compared to the HIV diversity in cervical (C) and plasma samples (D). Spearman Rank correlation were also calculated and presented for number of unique sequences versus CD4 T cell decline for cervical (p = 0.6463, r = 0.09446) (E) and plasma samples (0.0762, r = -0.2146) (F).
Fig 6.
Slower CD4 T cell decline is associated with HIV-1 subtype infection.
Declines in CD4 T cell rates per week from longitudinal sampling of study subjects (n = 72) were plotted. Each slope was derived from multiple time points of plasma sampling between viral set point and cART. Scatter dot plots of CD4 T cell declines in plasma stratified by subtype A, C and D infection are depicted on the left Y-axis (A). A marginal model with generalized estimating equation (GEE) approach was used to analyze CD4 T cell decline rates in a larger sample number of Ugandan and Zimbabwean women (n = 286) of the same cohort [24]. Box plots with mean and error bars of GEE calculated CD4 T cell declines grouped by subtype A, C and D are depicted on the right Y-axis (A). Statistical significance was done using a two-tailed Mann-Whitney test and two sample t-test assuming unequal variances. Spearman Rank correlation of the CD4 T cell loss versus average HIV genetic distance in the early plasma samples were separated and plotted by the infecting subtype A (B), subtype C (C) and subtype D (D).
Fig 7.
Models of heterosexual HIV-1 transmission bottlenecks.
Donors infected with HIV-1 have a genetically diverse HIV-1 population in the blood. Within the donor, a genetic bottleneck has been reported to occur between the blood compartment and semen, resulting in less diverse HIV-1 populations. The intact female vaginal mucosa is known to act as an efficient barrier to HIV-1 transmission, resulting in only a single TF variant establishing infection in the recipient (Model 1). Aside from the bottleneck within the donor the most stringent HIV-1 sieve effect is thought to take place in the female genital tract. In this study we show high HIV genetic diversity and distinct HIV-1 clones in the endocervix during early infection suggesting either an infection of the female genital tract with a population of HIV-1 clones from the donor (Model 2) or infection of the female genital tract with a single or limited number of HIV-1 clones which then evolves prior to systemic transmission (Model 3). Model 2 has been expanded in the bottom schematic figure. Transmission processes across the female genital tract: HIV-1 virions from the donor fluid are trapped by mucus or entry is blocked by the physical barrier of the mucosae (A). HIV-1 virions can be captured and internalized by intraepithelial Langerhans cells (B). HIV-1 can infect stromal CD4 T cells leading to productive infection (C). HIV-1 virions can bind to and infect DCs in the stroma and infect CD4 T cells in trans through an infectious synapse (D). Infected CD4 T cells and DCs can disseminate the virus from the mucosa to the blood (E).