Abstract
The development of therapies to eliminate the latent HIV-1 reservoir is hampered by our incomplete understanding of the biomolecular mechanism governing HIV-1 latency. To further complicate matters, recent single-cell RNA sequencing (scRNA-seq) studies reported extensive heterogeneity between latently HIV-1-infected primary T cells, implying that latent HIV-1 infection can persist in greatly differing host cell environments. We show here that transcriptomic heterogeneity is also found between latently infected T cell lines, which allowed us to study the underlying mechanisms of intercell heterogeneity at high signal resolution. Latently infected T cells exhibited a dedifferentiated phenotype, characterized by the loss of T cell-specific markers and gene regulation profiles reminiscent of hematopoietic stem cells (HSC). These changes had functional consequences. As reported for stem cells, latently HIV-1-infected T cells efficiently forced lentiviral superinfections into a latent state and favored glycolysis. As a result, metabolic reprogramming or cell redifferentiation destabilized latent infection. Guided by these findings, data mining of single-cell RNA-seq data of latently HIV-1-infected primary T cells from patients revealed the presence of similar dedifferentiation motifs. More than 20% of the highly detectable genes that were differentially regulated in latently infected cells were associated with hematopoietic lineage development (e.g., HUWE1, IRF4, PRDM1, BATF3, TOX, ID2, IKZF3, and CDK6) or were hematopoietic markers (SRGN; hematopoietic proteoglycan core protein). The data add to evidence that the biomolecular phenotype of latently HIV-1-infected cells differs from that of normal T cells and strategies to address their differential phenotype need to be considered in the design of therapeutic cure interventions. IMPORTANCE HIV-1 persists in a latent reservoir in memory CD4 T cells for the lifetime of a patient. Understanding the biomolecular mechanisms used by the host cells to suppress viral expression will provide essential insights required to develop curative therapeutic interventions. Unfortunately, our current understanding of these control mechanisms is still limited. By studying gene expression profiles, we demonstrated that latently HIV-1-infected T cells have a dedifferentiated T cell phenotype. Software-based data integration allowed the identification of drug targets that would redifferentiate viral host cells and, by extension, destabilize latent HIV-1 infection events. The importance of the presented data lies within the clear demonstration that HIV-1 latency is a host cell phenomenon. As such, therapeutic strategies must first restore proper host cell functionality to accomplish efficient HIV-1 reactivation.