HIV has no functional cure largely due to the persistent reservoir of latently infected CD4+ T cells containing integrated viral DNA. An understanding of this reservoir is central to targeted elimination strategies, but has been limited by a lack of defining markers and use of models or activation strategies that do not reflect infected CD4+ T cells in vivo. Defining the unmanipulated phenotype of the reservoir requires direct ex vivo approaches to identify targets for cure.
We performed single-cell Assay for Transposase Accessible Chromatin with Select Antigen Profiling by sequencing (ASAPseq) to identify HIV+ cells using accessible proviral DNA and their coordinate cell surface markers. We tested this strategy with three settings of increasing difficulty: (1) in vitro infection, peripheral blood CD4+ T-cells; (2) chronic infection, lymph node memory CD4+ T-cells; and (3) ART treated, peripheral blood memory CD4+ T-cells. Cells were enriched by bead separation, labeled with oligo-tagged antibodies, and prepared for transposition before single-cell partitioning, library construction and sequencing. Reads were processed using cellranger-atac, AMULET, kallisto, and hiv-haystack for analysis in R (Seurat and ArchR). Viral alignments were made against consensus or autologous sequences.
With the in vitro infection model, we identified 1320 HIV+ cells out of 7069 cells (18.7%) versus 5.1% p24+ cells via flow cytometry. Most HIV+ CD4+ T-cells had an activated/late differentiated phenotype with upregulated markers including CCR5, SLAM, and CD2. There was concordant and significant enrichment of Fos and Jun motifs in open chromatin of HIV+ cells consistent with preferential infection of proliferative and activated cells. In the chronic LNMC, 23 of 3559 cells were HIV+ (0.65%) and were enriched in CD71 and ICOS protein expression, with most being Tfh cells. In on-ART PBMC, we found 36 of 40542 cells were HIV+ (0.09%) largely in late differentiated states, with upregulation of markers including CD2, CCR5, and PD-1.
These findings show that ASAPseq can be used as a direct genomic strategy to identify HIV+ cells ex vivo and define key phenotypes of infected cells. Importantly, ASAPseq does not require stimulation and is done at single-cell resolution, though with limited information on provirus competency. The combined epigenetic and surface antigen profiling of infected cells is a valuable resource for discovery and testing of HIV reservoir elimination strategies.