Effector memory differentiation increases detection of replication-competent HIV-l in resting CD4+ T cells from virally suppressed individuals

Abstract

Studies have demonstrated that intensive ART alone is not capable of eradicating HIV-1, as the virus rebounds within a few weeks upon treatment interruption. Viral rebound may be induced from several cellular subsets; however, the majority of proviral DNA has been found in antigen experienced resting CD4+ T cells. To achieve a cure for HIV-1, eradication strategies depend upon both understanding mechanisms that drive HIV-1 persistence as well as sensitive assays to measure the frequency of infected cells after therapeutic interventions. Assays such as the quantitative viral outgrowth assay (QVOA) measure HIV-1 persistence during ART by ex vivo activation of resting CD4+ T cells to induce latency reversal; however, recent studies have shown that only a fraction of replication-competent viruses are inducible by primary mitogen stimulation. Previous studies have shown a correlation between the acquisition of effector memory phenotype and HIV-1 latency reversal in quiescent CD4+ T cell subsets that harbor the reservoir. Here, we apply our mechanistic understanding that differentiation into effector memory CD4+ T cells more effectively promotes HIV-1 latency reversal to significantly improve proviral measurements in the QVOA, termed differentiation QVOA (dQVOA), which reveals a significantly higher frequency of the inducible HIV-1 replication-competent reservoir in resting CD4+ T cells.

Fig 1. Ex vivo differentiation of resting enriched CD4+ T cells prior to QVOA increases detectable reactivation of latent HIV-1.

a, Schematic of assay outline for QVOA and differentiation QVOA (dQVOA). The relative distribution of central memory (T_CM), effector memory (T_EM), and transitional memory (T_TM) cells during each phase of the assays are represented as blue, orange, and green characters. b, Flow cytometric analysis showing the memory T cell subset gating strategy on samples from RAVEN participant 2274-R either after resting enrichment (top panel, cells that enter QVOA) or after 7 days of differentiation (bottom panel, cells that enter dQVOA). The numbers in each gate represent the percentage of events arising from its parental population. c, Column plot showing the proportions of naïve, T_CM, T_TM, T_EM and T_EMRA cells as a percentage of live CD3+Lineage-CD8- lymphocytes before and after 7 days of differentiation. NIH 2–20 and NIH 2–37 were assessed for memory subset distribution using a different flow cytometry antibody panel compared to the RAVEN participant samples and therefore were not included in this analysis. d, XY plots showing HIV-1-Gag expression level in each well from QVOA (top panel) and dQVOA (bottom panels) over time. The rCD4+ T cell source was RAVEN participant 2147-R. Stimulation was performed on day 0 for QVOA and dQVOA and when noted ART (EFV/RAL) was included in all media in the assay. Limiting dilutions performed for each assay type are shown. e, Bar graph showing the proportions of QVOA wells (black) found to be HIV-1-Gag+ compared to dQVOA wells (grey). f, Column graph showing the latent HIV-1 reservoir measured as infectious units per million rCD4+ T cells (IUPM) through QVOA and dQVOA. Limit of detection for each assay is represented by red bars. Wilcoxon matched-pairs signed rank tests were used and * denotes p < 0.05, ** denotes p < 0.01, and *** denotes p < 0.001.

Fig 2. Ex vivo differentiation drives HIV-1 reactivation in a reproducible manner.

a, Column plot showing the measured IUPM for QVOA (n = 7) and dQVOA (n = 3) performed on RAVEN participant 2147-R. Error bars are 95% confidence intervals generated by the program IUPMStats (http://silicianolab.johnshopkins.edu). %CV represents the coefficient of variation for each assay type. (p = 0.0004; unpaired t test with Welch’s correction). Black brackets indicate assays initiated on the same day and performed in parallel. %CV over each bracket represents the %CV for each batched assay set-up. b, Column plots showing the frequency of p24 positive wells after complete QVOA (left panel, white bars), differentiation alone as part of dQVOA (right panel, black bars), or after differentiation, activation, and outgrowth in complete dQVOA (right panel, white bars). For both panels, gray bars indicate the frequency of wells that remain p24 negative after completion of each assay. Frequency of HIV-1-Gag+ wells was determined using the top assay dilution for each assay.

Fig 3. Reservoir measurements derived from dQVOA directly correlate with QVOA, pre-ART highest viral load reported, and effector memory T cells.

a, The cross-correlation matrix (Pearson r value) of clinical measurements from linear regression. Analysis includes all available data from participants across the NIH and RAVEN cohort. *** denotes p < 0.001. b, XY Graphical representation of correlations found to be significant in panel a. c, Correlation with dQVOA IUPM from linear regression (Pearson r value) for each individual measurement. Analysis includes available data from participants in the RAVEN cohort. * denotes p < 0.05; ** denotes p < 0.01; *** denotes p < 0.001.

Fig 4. Differentiation drives reactivation of clonally-expanded, replication-competent HIV-1.

P6-PR-RT single-genome sequencing[78] was performed on HIV-1-Gag+ culture supernatants from dQVOA on NIH participant 2–20 (n = 2), NIH participant 2–37 (n = 1), and RAVEN participant 2147-R (n = 1). Circles represent single RNA genomes in the supernatants. Sequences obtained from different wells are shown in different colors. Identical sequences across different wells may result from latency reversal of infected cell clones. Sequences within wells that are different by a single nucleotide or two from a large rake of identical sequences are likely the result of RT error upon viral replication in the QVOA wells rather than from induction of different proviral variants.