Latently-infected CD4+ T cells are considered the main barrier to a cure for HIV-1. While these cells do not produce HIV constitutively, they can be induced to produce infectious virus upon activation. During viral suppression under therapy, a proportion of infected cells remain transcriptionally-active and these are predictive of time to viral rebound after antiretroviral therapy (ART) cessation. Considerable technical challenges are posed by the low frequency of transcriptionally-active HIV-1 reservoir cells and the fact that many of those cells reside in lymphoid tissues such as the gut. Finally, there are currently no known biomarkers that reliably distinguish latently-infected cells from uninfected cell populations in vivo.
To address these limitations, we developed “HIV-Seq”, a new single-cell (sc)RNAseq approach that enables simultaneous characterization of the transcriptome and surface proteome of unstimulated HIV-infected cells from blood and gut tissue from people living with HIV (PWH). Using custom-designed HIV-specific capture sequences and DNA-barcoded antibodies directed to key cell surface proteins (CITE-seq) introduced into a single cell RNAseq workflow (10X Genomics), we describe an in-depth combined scRNAseq/CITE-seq analysis of HIV reservoir cells from blood in the context of both viremia and ART suppression.
This new approach was applied to longitudinal samples collected at Week 0 (prior to commencing ART) of acute infection, and Week 24 or 45 after ART suppression [n=5]. CD4+ T cells were enriched using bead-based isolation and stained with DNA-tagged antibodies. HIV capture sequences were incorporated during library preparation. We additionally characterised total immune cells (CD45+) and T cells (CD3+) from the blood and gut of one ART-suppressed individual. scRNAseq and CITE-seq analyses were performed and sequences were aligned to a constructed subtype B consensus reference sequence.
Based on viremic sample data, HIV-seq enables 32-72% increased capture of HIV RNA+ cells, relative to no capture. In total, we identified 1232 HIV RNA+ cells from viremic timepoints and 26 HIV RNA+ cells from the ART-suppressed timepoints representing the transcriptionally-active reservoir.
Our HIV-seq method enables efficient identification and characterization of HIV-infected cells including in the context of ART suppression, allowing for in-depth transcriptomic and surface phenotypic analysis of transcriptionally-active reservoir cells.