Stable integrant-specific differences in bimodal HIV-1 expression patterns revealed by high-throughput analysis

Abstract

HIV-1 gene expression is regulated by host and viral factors that interact with viral motifs and is influenced by proviral integration sites. Here, expression variation among integrants was followed for hundreds of individual proviral clones within polyclonal populations throughout successive rounds of virus and cultured cell replication. Initial findings in immortalized cells were validated using CD4+ cells from donor blood. Tracking clonal behavior by proviral \"zip codes\" indicated that mutational inactivation during reverse transcription was rare, while clonal expansion and proviral expression states varied widely. By sorting for provirus expression using a GFP reporter in the nef open reading frame, distinct clone-specific variation in on/off proportions were observed that spanned three orders of magnitude. Tracking GFP phenotypes over time revealed that as cells divided, their progeny alternated between HIV transcriptional activity and non-activity. Despite these phenotypic oscillations, the overall GFP+ population within each clone was remarkably stable, with clones maintaining clone-specific equilibrium mixtures of GFP+ and GFP-cells. Integration sites were analyzed for correlations between genomic features and the epigenetic phenomena described here. Integrants inserted in genes sense orientation were more frequently found to be GFP negative than those in the antisense orientation, and clones with high GFP+ proportions were more distal to repressive H3K9me3 peaks than low GFP+ clones. Clones with low frequencies of GFP positivity appeared to expand more rapidly than clones for which most cells were GFP+, even though the tested proviruses were Vpr-. Thus, much of the increase in the GFP-population in these polyclonal pools over time reflected differential clonal expansion. Together, these results underscore the temporal and quantitative variability in HIV-1 gene expression among proviral clones that are conferred in the absence of metabolic or cell-type dependent variability, and shed light on cell-intrinsic layers of regulation that affect HIV-1 population dynamics.\n\nSummaryVery few HIV-1 infected cells persist in patients for more than a couple days, but those that do pose life-long health risks. Strategies designed to eliminate these cells have been based on assumptions about what viral properties allow infected cell survival. However, such approaches for HIV-1 eradication have not yet shown therapeutic promise, possibly because much of the research underlying assumptions about virus persistence has been focused on a limited number of infected cell types, the averaged behavior of cells in diverse populations, or snapshot views. Here, we developed a high-throughput approach to study hundreds of distinct HIV-1 infected cells and their progeny over time in an unbiased way. This revealed that each virus established its own pattern of gene expression that, upon infected cell division, was stably transmitted to all progeny cells. Expression patterns consisted of alternating waves of activity and inactivity, with the extent of activity differing among infected cell families over a 1000-fold range. The dynamics and variability among infected cells and within complex populations that the work here revealed has not previously been evident, and may help establish more accurate correlates of persistent HIV-1 infection.