Background

The latent HIV reservoir represents the major barrier to a cure. One curative strategy is targeting diseased cells for elimination based on biomarkers that uniquely define these cells. Single-cell RNA sequencing (scRNA-seq) has enabled the identification of gene expression profiles associated with disease at the single-cell level. Because HIV provirus in many cells during latency is not entirely silent, it became possible to determine gene expression patterns in a subset of cells latently infected with HIV.

Objective

The primary objective of this study was the identification of the gene expression profiles of single latently infected CD4+ T cells using scRNA-seq. Different conditions of latency establishment were considered. The identified profiles were then explored to prioritize the identified genes for future experimental validation.

Methods

To facilitate gene prioritization, three approaches were used. First, we characterized and compared the gene expression profiles of HIV latency established in different environments: in cells that encountered an activation stimulus and then returned to quiescence, and in resting cells that were infected directly via cell-to-cell viral transmission from autologous activated, productively infected cells. Second, we characterized and compared the gene expression profiles of HIV latency established with viruses of different tropisms, using an isogenic pair of CXCR4- and CCR5-tropic viruses. Lastly, we used proviral expression patterns in cells from people with HIV to more accurately define the latently infected cells in vitro.

Results

Our analyses demonstrated that a subset of genes is expressed differentially between latently infected and uninfected cells consistently under most conditions tested, including cells from people with HIV. Our second important observation was the presence of latency signatures, associated with variable conditions when latency was established, including cellular exposure and responsiveness to a T cell receptor stimulus and the tropism of the infecting virus.

Conclusion

Common signatures, specifically genes that encode proteins localized to the cell surface, should be prioritized for further testing at the protein level as biomarkers for the ability to enrich or target latently infected cells. Cell- and tropism-dependent biomarkers may need to be considered in developing targeting strategies to ensure that all the different reservoir subsets are eliminated.