In-depth single-cell analysis of translation-competent HIV-1 reservoirs identifies cellular sources of residual viremia and rebound viruses

BACKGROUND: Clonal expansion of HIV-infected cells is a well-known driver of the long-term persistence of the HIV-1 reservoir in ART-suppressed individuals, though the contribution of cell clones to plasma viremia under ART and upon analytical treatment interruption (ATI) is poorly understood. We developed a single cell assay to simultaneously sequence the proviral genome, matched integration site (IS), and TCR from cells harboring inducible proviruses, called STIP-Seq.
METHODS: Peripheral blood CD4 T cells from 8 ART-treated individuals were stimulated with PMA/ionomycin. Following methanol permeabilization and p24 staining, p24+ cells were single cell sorted. Whole genome amplification allowed for sequencing of the near full-length (NFL) proviral genomes by a 2- or 5-amplicon non-multiplexed PCR, IS analysis by Integration Site Loop Amplification and TCRβ sequencing by a multiplexed PCR. Three out of eight participants underwent an ATI, and phylogenetic analyses were performed to compare plasma-derived env (V1-V3) sequences to STIP-Seq proviral sequences.
RESULTS: A large proportion of p24+ cells stemmed from clonally expanded infected cells (78%, 135/173). NFL sequencing yielded a total of 42 distinct genomes. While only 14% of the genomes were intact and 2% had a large deletion, 83% displayed a small deletion at the 5' end of the genome (<500bp), consistently removing the major splice donor site. In two participants, STIP-Seq identified a clone with a genome-intact provirus matching low-level viremia (LLV) under ART, one of which had a predicted specificity towards M. tuberculosis and an IS in KCNA3, a gene involved in cell proliferation. The other clone (IS in SMG1P2) also matched plasma sequences retrieved during the ATI, suggesting that this clone was already producing LLV during ART before contributing to rebound upon ATI. Finally, we found a match between a plasma sequence under ART and a STIP-Seq sequence with a 5bp deletion covering the MSD, suggesting that MSD-defective proviruses could contribute to LLV.
CONCLUSIONS: This study provides one of the most comprehensive characterizations of the translation-competent HIV reservoir to date. By applying STIP-Seq in the context of an ATI, we identified cellular sources of plasma viremia, showing the relevance of the assay in HIV reservoir studies.