Combined noncanonical NF-κB agonism and targeted BET bromodomain inhibition reverse HIV latency ex vivo

Abstract

Latency reversal strategies for HIV cure using inhibitor of apoptosis protein (IAP) antagonists (IAPi) induce unprecedented levels of latent reservoir expression without immunotoxicity during suppressive antiretroviral therapy (ART). However, full targeting of the reservoir may require combinatorial approaches. A Jurkat latency model screen for IAPi combination partners demonstrated synergistic latency reversal with bromodomain (BD) and extraterminal domain protein inhibitors (BETi). Mechanistic investigations using CRISPR-CAS9 and single-cell RNA-Seq informed comprehensive ex vivo evaluations of IAPi plus pan-BET, bD-selective BET, or selective BET isoform targeting in CD4+ T cells from ART-suppressed donors. IAPi+BETi treatment resulted in striking induction of cell-associated HIV gag RNA, but lesser induction of fully elongated and tat-rev RNA compared with T cell activation-positive controls. IAPi+BETi resulted in HIV protein induction in bulk cultures of CD4+ T cells using an ultrasensitive p24 assay, but did not result in enhanced viral outgrowth frequency using a standard quantitative viral outgrowth assay. This study defines HIV transcriptional elongation and splicing as important barriers to latent HIV protein expression following latency reversal, delineates the roles of BET proteins and their BDs in HIV latency, and provides a rationale for exploration of IAPi+BETi in animal models of HIV latency.

Keywords: AIDS/HIV; Infectious disease; RNA processing; T cells; Transcription.

Figure 1. Evaluation of IAPi-based combination latency reversal regimens in the triple Jurkat model.

Cross titration of IAPi with (A and B) HDACi vorinostat (A); panobinostat (B); (C) PKC agonist ingenol B; (D) TLR-7 agonist GS-9620; (E) disulfiram; (F) GSK3i; (G and H) PRC2-targeted histone methyltransferase inhibitors EED226 (G) and GSK343 (H); and (I) BET inhibitor I-BET151. The y axes represent proviral luciferase reporter induction following 48 hours of drug exposure, normalized to DMSO-treated cells on each plate. Dose-response curves were generated using GraphPad Prism 9 using a log(agonist) versus response variable slope 4 parameter curve fit. Each point represents the average signal from 2 replicate wells from a single targeted screen experiment. Different colored curves represent a fixed concentration, indicated in the figure panels, of LRA combination partner with dose titration of IAPi. Conditions without IAPi treatment were plotted as 5 × 10^–13 M for visualization on the log₁₀ x axis.

Figure 2. Synergistic latency reversal activity for the IAPi and BETi combination in Jurkat N6 model of latency.

(A) Assessment of IAPi+BETi (I-BET151) latency reversal synergy using the Bliss independence model. Error bars represent SEM from pooled replicates across n = 2 independent experiments. (B) UMAP plots depicting overall transcriptome clustering and sample identity (top) and superimposed viral transcript detection (bottom) across treatment conditions. Each dot represents a single cell. Coloring of bottom panel was altered in Adobe Photoshop for ease of visualization of single cells. (C) Gene expression heatmap (red = upregulated; blue = downregulated) indicating the number of significantly differentially expressed genes identified in pairwise comparisons for each treatment condition. Color gradients indicate increasing numbers of statistically significant differentially expressed genes. Differential expression analysis was performed on sctransform normalized values with cutoffs for gene features expressed in at least 10% of cells and a log₂-fold difference of at least 0.25. Statistical significance for DEGs was evaluated with genome-wide Wilcoxon’s ranked sum tests with Bonferroni’s correction.

Figure 3. BET protein KO reveals a primary role for BRD4 in IAPi+BETi combination LRA activity in Jurkat N6 cells.

(A) Representative Western blot following lentiviral transduction of sgRNA targeted to BET family proteins. Data are representative of 7 independent experiments, with each target quantified in at least n = 4 experiments (Supplemental Figure 4). Empty vector (EV) and mock-infected (parent) cells served as negative control conditions (corresponding to DMSO or (-) control labels in B–G). (B) HSA reporter expression on live N6 cells following 48 hours IAPi (100 nM AZD5582) or BETi (1 μM I-BET151) exposure. Error bars represent SEM. (C–G) HSA reporter expression on live single cells following 48 hours of IAPi (100 nM AZD5582) and/or BETi (1 μM I-BET151) drug treatment in the presence or absence of different BET protein or BET protein isoform KOs. (H) Comparison of HSA reporter expression across IAPi+BET protein KO conditions. (I) Comparison of IAPi+BET protein KO and inhibition. For B–I, each dot represents an HSA determination for a sgRNA targeted to the indicated protein(s) or isoform(s). Each protein or isoform was targeted with 2 sgRNAs across n = 6 or n = 7 independent experiments, depending on the target. Dots for drug-treated conditions (i.e., DMSO for negative controls or IAPi±BETi) represent HSA determinations following treatment of parent N6 and empty vector transduced cells. Error bars represent SEM. FDR-adjusted P values for pairwise comparisons using Wilcoxon’s rank-sum tests are indicated as follows: **P < 0.01; ***P < 0.001. (J) Bliss independence analysis of IAPi+BET KO conditions versus IAPi+BETi. Black bars indicate median Bliss index.

Figure 4. Selective targeting of BRD4 alone or in combination with IAPi in primary CD4⁺ T cells from aviremic donors.

(A) Western blot of BET family proteins showing BRD4-specific degradation with 5 nM of the BRD PROTAC ZXH 3-26. BET degradation was confirmed by Western blot for all (n = 3) donors; a representative blot is shown. (B) Fold change in HIV gag cell–associated RNA (normalized to TBP expression, except for PMA/i due to known TBP upregulation following PMA/i exposure; ref. 78) in total CD4⁺ T cells following 24 hours of exposure to the indicated compounds. Each dot represents the average induction of 4 to 5 replicates of 1 to 2 million CD4⁺ T cells for an individual donor (n = 3). Error bars represent SEM. Bliss index for combination regimens is indicated in the upper left. (C) Total cellular ATP levels normalized to DMSO and (D) percentage of live cells measured by AOPI membrane exclusion dye microscopy-based assay following drug exposure. Note that PMA/i viability with AO/PI staining may be an underestimate of viability due to the formation of large clusters of proliferating viable cells. Each dot represents the average of 4 to 5 replicates (C, CellTiterGlo) or 2 replicates (D, AO/PI staining) for each donor (n = 3). Error bars represent SEM. Donors A-2, B, and C were tested (Supplemental Table 1).

Figure 5. Pan or selective targeting of BET protein BD domains alone or in combination with IAPi.

(A) Dose-response curves for pan-BD and BD-selective BETi in the triple Jurkat model across n = 3 independent experiments. (B–D) Combination activity of IAPi and (B) pan-BETi, (C) iBET-BD1, or (D) iBET-BD2 in the triple Jurkat model. Representative of n = 4 independent experiments. Conditions without IAPi treatment were plotted as (A) 5 × 10^–9 M or (B–D) 1 × 10^–11 M for visualization on the log₁₀ x axis. (E) Resting CD4⁺ T cell HIV gag caRNA (TBP normalized, except for PMA/i due to known TBP upregulation following PMA/i exposure, ref. , I) with parallel measurements of cell-associated p24 protein (F), and (G) culture medium p24 protein induction following 40 hours exposure of IAPi (100 nM AZD5582), pan-BETi (1 μM I-BET151), BD1- or BD2-selective BETi (2 μM), or combinations thereof compared with the positive control PMA/i. Horizontal black lines indicate (E) mean ± SEM or (F and G) geometric mean across all donors. (H) Total cellular ATP levels and (I) cellular viability following 40 hours drug exposure. Note that PMA/i viability with AO/PI staining may be an underestimate of viability due to the formation of large clusters of proliferating viable cells. (J) QVOA following IAPi and pan-BETi exposure relative to the positive control PHA/IL-2. Infectious unit per million resting CD4⁺ T cells (IUPM) for each condition represented as a percentage of the PHA IUPM for each donor (different shapes). Open shapes indicate no positive wells were detected. For QVOA, resting CD4⁺ T cells from donors E-2, G, and D-2 were evaluated (Supplemental Table 1). All error bars represent mean ± SEM. FDR-corrected P values for pairwise comparisons using Wilcoxon’s signed-rank tests are indicated as follows: *P < 0.05; **P < 0.01.

Figure 6. 26.

HIV transcript profiling and p24 protein induction following IAPi and/or BETi exposure in primary CD4⁺ T cells from aviremic donors. (A) Fold change values for different HIV transcripts normalized to μg RNA input in total CD4⁺ T cells following 8 hours of LRA stimulation. Each dot represents the average fold change over DMSO-treated cells for an individual donor. Donors (A-2, D-3, H-2, J, K-2) were tested for IAPi/pan-BETi (I-BET151) and single agents; 4 donors were tested for IAPi, pan-, and BD-selective-BETi (D-3, H-2, J, K-2) (Supplemental Table 1). For 1 donor there was a likely PCR amplification failure for the nef transcript. In 2 donors, there were insufficient cells to evaluate the HDACi suberoylanilide hydroxamic acid (SAHA) in parallel. Filled symbols represent conditions where there were no detectable transcripts above background and data were left censored at 5 copies/μg RNA input, based on the background digital droplet signal observed in no reverse transcriptase and no template control wells run for each donor on each plate for each primer/probe set. Horizontal bars indicate the median fold change across the donors tested for an indicated transcript/drug condition. (B) Bliss synergy indexes for IAPi+BETi drug combinations across different transcripts. Horizontal bars indicate the median index calculated across 4 to 5 donors depending on the transcript/drug condition. (C) Ultrasensitive p24 measurements in culture supernatants following 8 hours drug exposure and washout for 4 different donors (D-3, H-2, J, K-2).