The replication-competent HIV reservoir is a genetically restricted, younger subset of the overall pool of HIV proviruses persisting during therapy, which is highly genetically stable over time

Abstract

Within-host HIV populations continually diversify during untreated infection, and this diversity persists within infected cell reservoirs during antiretroviral therapy (ART). Achieving a better understanding of on-ART proviral evolutionary dynamics, and a better appreciation of how the overall persisting pool of (largely genetically defective) proviruses differs from the much smaller replication-competent HIV reservoir, is critical to HIV cure efforts. We reconstructed within-host HIV evolutionary histories in blood from seven participants of the Women's Interagency HIV Study who experienced HIV seroconversion, and used these data to characterize the diversity, lineage origins, and ages of proviral env-gp120 sequences sampled longitudinally up to 12 years on ART. We also studied HIV sequences emerging from the reservoir in two participants. We observed that proviral clonality generally increased over time on ART, with clones frequently persisting long term. While on-ART proviral integration dates generally spanned the duration of untreated infection, HIV emerging in plasma was exclusively younger (i.e., dated to the years immediately pre-ART). The genetic and age distributions of distinct proviral sequences remained stable during ART in all but one participant, in whom there was evidence that younger proviruses had been preferentially eliminated after 12 years on ART. Analysis of the gag region in three participants corroborated our env-gp120-based observations, indicating that our observations are not influenced by the HIV region studied. Our results underscore the remarkable genetic stability of the distinct proviral sequences that persist in blood during ART. Our results also suggest that the replication-competent HIV reservoir is a genetically restricted, younger subset of this overall proviral pool.IMPORTANCECharacterizing the genetically diverse HIV sequences that persist in the reservoir despite antiretroviral therapy (ART) is critical to cure efforts. Our observations confirm that proviruses persisting in blood on ART, which are largely genetically defective, broadly reflect the extent of within-host HIV evolution pre-ART. Moreover, on-ART clonal expansion is not appreciably accompanied by the loss of distinct proviral lineages. In fact, on-ART proviral genetic composition remained stable in all but one participant, in whom, after 12 years on ART, proviruses dating to around near ART initiation had been preferentially eliminated. We also identified recombinant proviruses between parental sequence fragments of different ages. Though rare, such sequences suggest that reservoir cells can be superinfected with HIV from another infection era. Overall, our finding that the replication-competent reservoir in blood is a genetically restricted, younger subset of all persisting proviruses suggests that HIV cure strategies will need to eliminate a reservoir that differs in key respects from the overall proviral pool.

Keywords: HIV; genetic stability; molecular dating; persistence; phylogenetics; rebound.

Fig 1

Participant sampling timeline. Time zero denotes ART initiation. The right arrow denotes enrolment into the cohort. The asterisk denotes the clinically estimated date of infection, defined as the midpoint between the last negative and first positive HIV tests. Gray shading denotes ART. Here and throughout all figures, closed circles denote pre-ART plasma HIV RNA sampling. Open circles denote post-ART plasma HIV RNA sampling. Diamonds denote proviral sampling on ART. Participants 1, 3, and 7 correspond to participants 1, 3, and 6 in reference (32).

Fig 2

Proviral clonal distribution and dynamics during ART. (A) Total number of env-gp120 proviral sequences collected per participant (shown inside the donut), and the percentages that were observed only once (white) versus those observed more than once (i.e., clones; black). (B) Proviral clonality by time point on ART. Gray slices denote clones observed only at that time point (each clone in a distinct shade of gray); colored slices link clones isolated across time points. (C) Percent proviral clonality over time on ART, with regression line.

Fig 3

Participant 1: diversity and inferred integration dates of HIV sequences persisting during ART. (A) Plasma viral load history, with symbols denoting sampling time points. The lower limit of quantification of the viral load assays performed between 2008 and 2010 was 80 or 48 HIV RNA copies/mL depending on the assay; thereafter, it was 20 copies HIV RNA copies/mL. Closed circles denote pre-ART plasma HIV RNA sampling, and open circles denote post-ART plasma HIV RNA sampling. Diamonds denote proviral sampling on ART. Gray shading denotes ART. (B) Example rooted within-host phylogeny, with scale in estimated substitutions per nucleotide site. Asterisks identify nodes supported by posterior probabilities ≥ 70%. The adjacent highlighter plot shows ENV-GP120 amino acid diversity, with colored ticks denoting non-synonymous substitutions with respect to the reference sequence at the top of the phylogeny. (C) Linear model (dashed blue diagonal) relating plasma HIV RNA collection dates (closed colored circles) to their respective root-to-tip distances in the example phylogeny shown in B. This linear model is then used to convert the root-to-tip distances of on-ART sequences of interest (rebound HIV in open circles and proviruses in colored diamonds) to their integration dates. Gray lines trace the phylogenetic relationships between HIV env-gp120 sequences. (D) Integration dates and 95% HPD intervals for distinct on-ART sequences, stratified by collection year, that were derived from averaging results across all QC-passed phylogenies for this participant. The Kruskal-Wallis test was used to compare proviral integration date distributions across all sampling time points. The Mann-Whitney U-test was used to compare plasma and proviral integration date distributions. (E) This panel shows the P-values from the tests for population genetic structure (AMOVA, = analysis of molecular variance; CC, correlation coefficient) that compared the populations of distinct proviral sequences per time point (closed black circle) and the distinct proviral and rebound populations overall (cross symbol). The bars around the CC P-value represent the 95% HPD interval of the P-values derived from all QC-passed phylogenies for this participant.

Fig 4

Participant 1: recombinant env-gp120 proviral and rebound sequences. Colored circles in the sampling timeline (top) denote the year of origin of one or more recombinant sequence fragments detected among provirus or rebound HIV sequences (shown below). This timeline also shows the collection date of the sampled rebound HIV sequences (gray doughnuts) and longitudinal on-ART proviral sequences (black diamonds). The sampled recombinant sequences are grouped by type (proviral or rebound HIV) and year of collection, one sequence per line. The solid and dotted line fragments represent the two parent sequences, colored by year of origin.

Fig 5

Participant 2: diversity and inferred integration dates of HIV sequences persisting during ART. Legend as in Fig. 3, except that E shows P-values from genetic compartmentalization tests applied to serially sampled proviruses only.

Fig 6

Participant 3: diversity and inferred integration dates of HIV sequences persisting during ART. Legend as in Fig. 5.

Fig 7

Participant 4: diversity and inferred integration dates of HIV sequences persisting during ART. Legend as in Fig. 5, except that D shows pairwise comparisons between groups after correction for multiple comparisons, as the overall Kruskal-Wallis test returned P = 0.0004.

Fig 8

Participant 5: diversity and inferred integration dates of HIV sequences persisting during ART. Legend as in Fig. 5.

Fig 9

Participant 6: diversity and inferred integration dates of HIV sequences persisting during ART. Legend as in Fig. 5.

Fig 10

Scaled integration timings of on-ART proviral sequences and rebound HIV. Estimated integration timings of proviruses sampled longitudinally on-ART (diamonds) and rebound viruses (circles), depicted on a scale between the estimated date of infection and ART suppression. For participant 1, the date of ART re-suppression post-rebound was used. Sequences with integration dates after ART suppression are plotted at 1.0. For context, the blue-shaded box marks the 3 years leading up to ART suppression.

Fig 11

Proviral diversity during ART. (A) Grand mean within-host patristic distance separating all pairs of distinct proviral sequences per time point, with a line linking each participant’s values. (B) Same as A but expressed in terms of participants’ mean phylogenetic diversity of distinct sequences. P-values were computed using a Friedman test applied to the first three time points for all participants, as well as to all four time points for participants 1, 4, and 6. (C) Relationship between pre-ART plasma HIV RNA and on-ART proviral diversity, expressed in terms of grand mean patristic distance. (D) same as C but where diversity is expressed in terms of mean phylogenetic diversity.