Intact HIV-1 proviruses accumulate at distinct chromosomal positions during prolonged antiretroviral therapy

Abstract

Chromosomal integration of genome-intact HIV-1 sequences into the host genome creates a reservoir of virally infected cells that persists throughout life, necessitating indefinite antiretroviral suppression therapy. During effective antiviral treatment, the majority of these proviruses remain transcriptionally silent, but mechanisms responsible for viral latency are insufficiently clear. Here, we used matched integration site and proviral sequencing (MIP-Seq), an experimental approach involving multiple displacement amplification of individual proviral species, followed by near-full-length HIV-1 next-generation sequencing and corresponding chromosomal integration site analysis to selectively map the chromosomal positions of intact and defective proviruses in 3 HIV-1-infected individuals undergoing long-term antiretroviral therapy. Simultaneously, chromatin accessibility and gene expression in autologous CD4+ T cells were analyzed by assays for transposase-accessible chromatin using sequencing (ATAC-Seq) and RNA-Seq. We observed that in comparison to proviruses with defective sequences, intact HIV-1 proviruses were enriched for non-genic chromosomal positions and more frequently showed an opposite orientation relative to host genes. In addition, intact HIV-1 proviruses were preferentially integrated in either relative proximity to or increased distance from active transcriptional start sites and to accessible chromatin regions. These studies strongly suggest selection of intact proviruses with features of deeper viral latency during prolonged antiretroviral therapy, and may be informative for targeting the genome-intact viral reservoir.

Keywords: AIDS/HIV; Infectious disease; T cells.

Figure 1. Simultaneous analysis of near-full-length HIV-1 proviral sequences and corresponding HIV-1 integration sites.

(A–C) Horizontal phylogenetic trees of all intact, near-full-length HIV-1 sequences from 3 study participants (P1–P3). Clonal sequences are listed only once; the number of clones is indicated by circular symbols. Chromosomal integration site coordinates (3′-LTR border) for each sequence are indicated.

Figure 2. Chromosomal positions of intact and defective HIV-1 proviruses.

Circos plots demonstrating chromosomal integration site positioning of intact and defective proviruses from 3 study participants (A, patient 1; B, patient 2; C, patient 3). Color and line coding indicate viral sequence characteristics (intact vs. defective) and orientation of integrated provirus relative to host gene. Targeted genes were identified using Ensembl (v86); gene names are shown according to HUGO classification (https://www.genenames.org). Colored dots indicate the number of clones detected. *Sequences in chromosomal regions associated with multiple genes; ^†mixed orientation among these genes; ^#integration sites in pseudogenes.

Figure 3. Chromosomal annotations of HIV-1 integration sites associated with intact and defective proviral sequences.

(A and B) Pie charts showing proportion of intact and defective HIV-1 sequences located in genic versus non-genic/pseudogenic regions (A), and with the same or opposite orientation relative to host genes (among sequences integrated in genes; B). Integration sites associated with multiple genes and mixed orientations to host genes were not considered for the analysis in B. Significance was tested using 2-tailed χ² tests; nominal P values are reported. (C and D) Pie charts indicating the proportion of intact and defective HIV-1 sequences located in regions with defined repetitive genetic elements (C) (SINE, short interspersed nuclear element; LINE, long interspersed nuclear element; LTR, LTR retrotransposon; DNA, DNA transposon) and in exons or introns (D). (E–G) Ontology analysis of genes harboring defective and intact HIV-1 sequences. Data represent a categorization of genes harboring intact or defective HIV-1 sequences according to defined formal functional entities (E). (F) Top ten canonical pathways predicted by Ingenuity Pathway Analysis for genes containing intact or defective proviruses; x axis shows corresponding –log(P value) for each pathway using right-tailed Fisher’s exact tests, with a threshold of –log(0.05) marked as a dotted line. RhoGDI, Rho GDP dissociation inhibitor. (G) Positioning of intact and defective HIV-1 proviruses in cancer-related genes. Left y axis shows upper limit of the –log(P value) for each indicated category (Ingenuity Pathway Analysis–based right-tailed Fisher’s exact tests); right y axis depicts the number of sites identified in the “Cancer” category in the different gene groups. For A–G, clonal sequences were counted only once.

Figure 4. Distinct chromosomal locations of intact HIV-1 proviruses in study participant 1.

(A and B) Circos plots highlighting ATAC-Seq and RNA-Seq reads in proximity (ATAC-Seq: ±8000 bp; RNA-Seq: ±5000 bp) to integration sites of intact and defective HIV-1 proviruses. (C) Combined individual-value/box-and-whisker plots indicating the chromosomal distance between HIV-1 integration sites and the most proximal TSS listed in Ensembl v86 (databank), or identified through analysis of expressed RNA species located within the boundaries of the host gene, using autologous RNA-Seq data from the indicated cell populations and limiting the analysis to proviruses integrated in expressed genes. (D) Combined individual-value/box-and-whisker plots showing the chromosomal distance between integration sites and the center of the most proximal ATAC-Seq peaks in indicated CD4⁺ T cell populations. (E) Gene expression intensity of host genes harboring intact or defective proviral integration sites, normalized to the chromosomal distance between integration sites and the most proximal TSSs determined using autologous RNA-Seq data as described in C. In C–E, boxes and whiskers represent median, 25% and 75% percentiles, and minimum/maximum levels. Significance was calculated using 2-tailed Mann Whitney U tests; nominal P values are reported. (F) Distance between each integration site and most proximal TSS, plotted against corresponding distance between each integration site and the center of the nearest ATAC-Seq peak. Spearman’s correlation coefficients are shown for each cell population. In A–F, clonal sequences were shown/counted only once. EM, effector memory; CM, central memory.

Figure 5. Chromosomal integration site features in study participants 2 and 3.

(A and B) Circos plots highlighting ATAC-Seq and RNA-Seq reads in proximity (ATAC-Seq: ±20,000 base pairs; RNA-Seq: ±10,000 base pairs) to integration sites of intact and defective HIV-1 proviruses. (C) Combined individual-value/box-and-whisker plots summarizing the distance between integration sites and the most proximal TSS derived from Ensembl v86 (databank) or from autologous RNA-Seq data (as described for Figure 4C, among proviruses integrated in expressed genes). (D) Combined individual-value/box-and-whisker plots showing the distance between integration sites and the center of the most proximal ATAC-Seq peak. (E) Combined individual-value/box-and-whisker plots demonstrating the transcriptional activity of host genes harboring intact or defective proviral sequences, normalized to the distance between integration sites and the most proximal TSSs, as determined for Figure 4C and Figure 5C. In C–E, boxes and whiskers represent median, 25% and 75% percentiles, and minimum/maximum levels. Significance was calculated using 2-tailed Mann Whitney U tests; nominal P values are reported. (F) Distance between each integration site and most proximal TSS, plotted against corresponding distance between each integration site and the center of the nearest ATAC-Seq peak. Spearman’s correlation coefficients are shown for each cell population. For A–F, clonal sequences were shown/counted only once. EM, effector memory; CM, central memory.