The XPB Subunit of the TFIIH Complex Plays a Critical Role in HIV-1 Transcription, and XPB Inhibition by Spironolactone Prevents HIV-1 Reactivation from Latency

Treatment of latently infected cell lines with SP inhibits transcription of HIV-1.

SP has been previously shown to block acute HIV infection of CD4⁺ T cells (48). Here, to explore whether SP could be used in block-and-lock approaches to an HIV cure, we investigated its ability to inhibit HIV-1 transcription in cell line models of latency. We studied SP activity in two established cell line models of latency, OM-10.1 and ACH-2 cells, in the presence of ART. OM-10.1 cells are a promyelocytic line derived from HL-60 cells that survived an acute HIV-1 infection. Each cell contains a single integrated, full-length provirus (54). SP treatment of these cells resulted in a dose-dependent reduction of HIV-1 capsid p24 in the supernatant, with a half-maximal inhibitory concentration (IC₅₀) of 1.77 μM (Fig. 1A). This inhibition correlated with a dose-dependent degradation of the cellular protein XPB (Fig. 1B). Eplerenone (EPL), a more selective mineralocorticoid receptor (MR) antagonist analog of SP, does not lead to XPB degradation (46, 47, 51, 55). As previously reported (48), no inhibition of HIV-1 was observed with increasing concentrations of EPL (Fig. 1C), suggesting that the mechanism of HIV-1 inhibition is independent of MR antagonism. No cytotoxicity was observed with up to 50 μM SP or EPL following a 72-h incubation (Fig. 1D).

Next, we evaluated the long-term effects of SP treatment. OM-10.1 cells were treated with 10 μM SP and 10 μM EPL or dimethyl sulfoxide (DMSO) for 190 days (Fig. 1E), and capsid p24 released into the supernatant was quantified by p24 enzyme-linked immunosorbent assay (ELISA). In OM-10.1 cells, SP rapidly inhibits viral production by approximately 2.5 logs, with capsid p24 levels stabilizing around the limit of detection (3 pg/ml). As expected, a reduction in the HIV-1 mRNA expression was observed in parallel (Fig. 1F, red line). The mRNA levels of the XPB gene, ERCC3, were either unaffected or slightly elevated by SP treatment (Fig. 1F, brown line). This confirms previous reports that degradation of XPB by SP is mediated by the 26S proteasome and not by reduced XPB mRNA expression (46). Of note, the half-life of SP in long-term culture of OM-10.1 cells with ART was determined by liquid chromatography-tandem mass spectrometry (LC-MS/MS) to be 2.3 (±0.2) days, and cells were split and treated every 3 days. All experiments were performed in the presence of ART unless otherwise stated. Of note, without ART the half-life is 2.9 (±1.7) days. Altogether, these results suggest that, in OM-10.1 cells, SP treatment combined with ART over time inhibits HIV-1 transcription to very low, frequently undetectable levels without obvious cytotoxic effects in the cells.

We next assessed the activity of SP in the ACH-2 cell line, a T cell clone containing a single integrated provirus (56, 57), which harbors a C37T mutation TAR such that the Tat-TAR feedback loop is defective, resulting in low transcription rates (58, 59). SP treatment of these cells resulted in a similar dose-dependent reduction of HIV-1 capsid p24 in the supernatant, with an IC₅₀ of 1.95 μM (Fig. 2A), which correlated with a dose-dependent degradation of XPB (Fig. 2B). Again, EPL was inactive against HIV (Fig. 2C). Cytotoxicity was not observed for up to 20 μM SP for a 72-h period or up to 50 μM with EPL (Fig. 2D). The long-term treatment of ACH-2 cells inhibited viral production by 1 to 1.5 logs, with capsid production stabilizing around 20 to 100 pg/ml (Fig. 2E). As expected, a reduction in cell-associated HIV RNA was also observed (Fig. 2E), while levels of XPB mRNA were either unaffected or slightly elevated (Fig. 2F). No difference in cell viability or cell number was observed over time (Fig. 2G). Differences in the magnitude of SP activity on HIV-1 p24 production between OM-10.1 and ACH-2 cells may be related to the level of XPB degradation in these cell lines, 95% in OM-10.1 cells compared to 81% in ACH-2 cells (Fig. 1B and 2B). Given that SP inhibits the transcription from proviruses with an incompetent Tat-TAR feedback loop, as is the case for ACH-2 cells, the mechanism of action of SP is likely independent of the Tat-TAR interaction.

We also investigated the potential of SP to inhibit SIVmac239 in HUT-78 cells. Potent suppression of simian immunodeficiency virus (SIV) replication with SP but not EPL or DMSO was observed over a period of 12 days (Fig. 2G). This was accompanied by degradation of XPB by SP treatment as measured by Western blotting on day 12 (Fig. 2H, inset panel). A rapid decline in cellular viability, due to cytotoxic effects of virus replication, was observed in DMSO- and EPL-treated cells where virus replication was allowed to proceed, while the viability of cells treated with 10 μM SP was not adversely affected (Fig. 2I). These results suggest that SP may be used in latency studies in nonhuman primates infected with SIV.

In sum, these results suggest that SP inhibits HIV-1 transcription in established cell line models of latency.

RNA silencing of the XPB gene inhibits HIV-1 transcription.

To confirm that HIV-1 inhibition by SP was truly dependent on XPB degradation, and not the result of an off-target effect, we knocked down the XPB gene, ERCC3, by short hairpin RNA (shRNA) and assessed viral replication. OM-10.1 cells were transduced with a lentiviral vector carrying either an ERCC3-specific shRNA or an shRNA targeting the green fluorescent protein (GFP) gene (shGFP) as the control (since these cells do not contain GFP). In parallel, cells transduced with shGFP were treated with an IC₅₀ of 1.8 μM and a 90% inhibitory concentration (IC₉₀) of 6.3 μM of SP, as well as 10 μM SP, to compare XPB levels and HIV replication. After 48 h of culture, cells were placed back on ART and selected with puromycin to enrich for cells containing the shRNA. After 10 days of puromycin selection, HIV-1 capsid production in the supernatant was measured by p24 ELISA, and cells were collected for RNA and Western blotting analysis.

Knockdown efficiencies of 60% and 80% at the protein and mRNA levels, respectively (Fig. 3A and B), were achieved with the shRNA targeting ERCC3 while well tolerated by cells (Fig. 3C). Notably, SP treatment resulted in a slight upregulation of ERCC3 mRNA (Fig. 3B) as previously observed (Fig. 1F), which we suspect is a cellular compensatory mechanism for the loss of XPB protein. When compared to cells transduced with the shGFP control, the ERCC3 knockdown resulted in suppression of HIV gag mRNA and p24 production to levels between the IC₅₀ and IC₉₀ concentrations of SP (Fig. 3D and E).

Importantly, when we compared expression of additional TFIIH components, we observed substantial downregulation of XPD, CDK7, and cyclin H with SP treatment (Fig. 3F). With ERCC3 knockdown, despite XPB downregulation similar to the IC₅₀ and IC₉₀ concentrations of SP, XPD was not downregulated (Fig. 3F). However, although not to the same degree as SP, loss of CDK7 and cyclin H was still detected (Fig. 3F).

Altogether, our results strongly support that the inhibition of HIV-1 transcription observed with SP treatment is the result of XPB degradation.

Treatment of latently infected cells with SP reduces HIV-1 reactivation from latency.

We next investigated whether SP treatment could block the reactivation of latent HIV-1 proviruses when cells were exposed to LRAs. The specific LRAs and doses chosen to reactivate each cell model were selected based on the ability to potently stimulate HIV-1 replication with minimal toxicity. Prostratin, phorbol myristate acetate (PMA), and tumor necrosis factor alpha (TNF-α) activate transcription by induction of NF-κB, while the histone deacetylase (HDAC) inhibitor suberoylanilide hydroxamic acid (SAHA) promotes chromatin acetylation, namely, at Nuc-1 on the HIV-1 LTR (60–63).

OM-10.1 and ACH-2 cells were treated for 90 and 15 days, respectively, with 10 μM SP or DMSO, and the levels of HIV-1 capsid production were monitored over time by p24 ELISA (Fig. 4A and B). Cell number and viability remained stable over time (as in Fig. 1 and 2). Once HIV levels had stabilized at maximal inhibition, the cells were treated with a panel of LRAs in the presence of ART and either 10 μM SP or DMSO. After an 8-h incubation period, cells were collected and viral transcription was assessed by reverse transcriptase quantitative PCR (RT-qPCR) to HIV nef mRNA. Robust reactivation was observed in DMSO+ART-treated OM-10.1 and ACH-2 cells (Fig. 4C and D, blue bars) with all LRAs tested. While some reactivation of HIV in SP-treated OM-10.1 and ACH-2 cells was observed in the presence of LRAs, these were non-statistically significant (Fig. 4C and D, red bars). Similar results were observed in the Jurkat latency models (J-Lat A1 and A2) (data not shown).

Next, we assessed the activity of SP in primary cells from ART-adherent individuals living with HIV. Resting CD4⁺ T cells were isolated from peripheral blood mononuclear cells (PBMCs) from seven individuals with undetectable HIV viral loads, adherent to ART for at least 1 year with no viral load “blips” in the last 6 months. These cells were treated for 24 h with either 10 μM SP or vehicle only in the presence of the entry inhibitor T20, to prevent novel infections. In the final 6 h of treatment, cells were challenged with PMA/ionomycin for T cell activation. SP treatment reduced HIV-1 reactivation an average of 1.4 log (27.6-fold) measured by RT-qPCR per million cells (Fig. 4E). The expression of a housekeeping gene POLR2A was also measured and showed modest yet significantly increased levels (an average of 0.4 log; 2.5-fold) (Fig. 4F) (64). These results demonstrate that SP inhibits reactivation from latency in primary CD4⁺ T cells explanted from virally suppressed individuals.

In sum, these results show that treatment of latently infected cell lines, and ex vivo stimulated primary CD4⁺ T cells, with 10 μM SP dramatically reduces reactivation of HIV-1 from latency when exposed to potent LRAs.

SP treatment alone blocks viral reactivation; however, long-term SP treatment does not prevent viral rebound upon treatment interruption.

The long-term treatment of HIV-1-infected cells with the Tat inhibitor dCA results in the accumulation of epigenetic marks at the HIV-1 promoter that render reactivation from latency very difficult to occur upon treatment interruption or stimulation with LRAs (41, 43, 44). Thus, we investigated whether SP treatment over time would lead to similar effects. First, we tested whether 10 μM SP alone was sufficient to maintain HIV transcription suppression in the absence of ART. As such, we washed off ART on day 69 from the OM-10.1 cell culture (three independent experiments) and measured viral p24 production in the supernatant (Fig. 5A, black arrows). A nonsignificant increase in HIV-1 p24 production was observed in DMSO-treated cells in the absence of ART (Fig. 5A, light blue line), while no such increase was observed in SP-treated cells (Fig. 5A, orange line), suggesting that SP treatment alone is able to block HIV replication in the absence of ART. Next, we assessed whether pretreatment with SP would delay HIV rebound upon all treatment interruption; as such, on day 78, DMSO and SP were washed off and cells maintained on ART alone (Fig. 5B, black arrow). As expected, HIV p24 levels remained stable in cells pretreated with DMSO; however, within 12 days, a rapid increase in virus replication back to DMSO levels was observed in cells pretreated with SP (Fig. 5B, orange line). Concomitantly, within 3 days of treatment interruption, we observed a rapid rebound in XPB protein levels as well as other TFIIH components (XPD, CDK7, and cyclin H) (Fig. 5C). In fact, XPB levels rebounded above the level of the DMSO control and remained elevated for at least 9 days post treatment interruption (Fig. 5C, orange lanes versus light blue lanes).

While pretreatment with SP did not result in long-lasting epigenetic control of viral transcription upon treatment interruption, SP treatment alone seems sufficient to maintain HIV transcriptional inhibition in the absence of ART. These important results suggest that SP blocks HIV transcription as a single therapy and holds great potential to be used in combination with other latency-promoting agents (LPAs), with the goal of establishing an epigenetic lock on the HIV genome.

SP treatment reduces RNAPII recruitment to the HIV-1 genome.

To confirm that transcriptional inhibition is the mechanism of action of SP against HIV, we investigated RNAPII recruitment to the HIV-1 genome. As such, we performed chromatin immunoprecipitation (ChIP) to RNAPII followed by qPCR using primers recognizing the HIV-1 genome. We sought to identify differences in occupancy of total as well as phosphorylated Ser5 or Ser2 RNAPII (pSer5, pSer2 RNAPII).

Three independent long-term treatments (∼70 days) of OM-10.1 cells with DMSO+ART or SP+ART (Fig. 4A) were treated or not with 1 μM prostratin for 8 h, resulting in a mean 4.3-fold increase in p24 production and a mean 147-fold increase in HIV mRNA (data not shown) in DMSO+ART-treated cells, with no detectable changes in p24 production in SP-treated cells observed (but a mean 5.3-fold increase in HIV mRNA) (data not shown). These cells were cross-linked and prepared for RNAPII ChIP. Without prostratin, with or without SP, a small peak of RNAPII was observed at the HIV promoter (position of ∼450 nt) with very little occupancy throughout the genome (Fig. 6A). Modest reductions in occupancy of total pSer5 and pSer2 RNAPII in unstimulated SP+ART-treated cells were observed compared to those of the DMSO+ART control (Fig. 6B and C, red versus dark blue lines). Stimulation with prostratin led to a significant increase of total RNAPII occupancy in DMSO+ART-treated cells, with a peak at the promoter and substantial occupancy throughout the genome (Fig. 6A, light blue line). As expected, a strong peak of pSer5 RNAPII was observed at the HIV promoter in DMSO+ART-treated cells stimulated with prostratin, with a slight reduction in the gene body (Fig. 6B, light blue line). A peak of pSer2 RNAPII was observed at the HIV promoter of DMSO+ART-treated cells stimulated with prostratin, drastically decreasing between nucleotide (nt) positions 1000 and 3500 but increasing again toward the 3′ end between positions 6000 and 9000 (Fig. 6C, light blue line). This is consistent with the role of pSer2 in relief of promoter-proximal pausing (mediated by P-TEFb [65, 66]) and recruitment of splicing and RNA processing machinery (mediated by CDK13/12 [66–69]). A significant reduction of RNAPII occupancy (total, pSer5 and pSer2) was observed on the HIV genome when prostratin reactivation was performed in the presence of SP (Fig. 6A to C, orange lines). This is apparent at the HIV promoter but most noticeable along the gene body. Almost no pSer5 RNAPII is observed along the HIV genome beyond the promoter in SP-treated cells, while total and pSer2 RNAPII is still detected at both the promoter and toward the 3′ end of the genome. Importantly, no significant differences were observed in the occupancy of RNAPII (total, pSer5, pSer2) on the promoter of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or in the gene body of RPL13A (Fig. 6D to F).

In sum, these results indicate that long-term treatment of cells with 10 μM SP significantly reduces RNAPII occupancy on the HIV genome and potently suppresses RNAPII recruitment upon prostratin stimulation. These effects are not observed on the promoter or open reading frame (ORF) of housekeeping genes.

Long-term effects of SP treatment on the cellular transcription apparatus and cell cycle progression.

Previous reports have suggested that short-term treatment of cells with 10 μM SP selectively degrades the XPB subunit of TFIIH, leaving other components mostly unaffected (46, 48). Additionally, in short-term studies using transcriptome sequencing (RNA-seq) and uridine analog 5-ethynyluridine (EU) incorporation assays, Alekseev et al. did not observe global transcriptional defects in SP-treated cells (51). We thus investigated the long-term effects of SP treatment on components of the TFIIH complex as well as on a few important CDKs and respective cyclins. No significant differences were observed in the mRNA expression of any of the TFIIH components in OM-10.1 cells treated with SP for over 100 days (Fig. 7A, cells from Fig. 1E). However, at the protein level, although XPB was by far the most downregulated protein (over 95%), a clear downregulation of other components of TFIIH (namely, XPD, p62, CDK7, and cyclin H) was observed at all time points compared to that for DMSO and EPL (Fig. 7B). In addition, in ACH-2 cells treated for 15 days with 10 μM SP, a similar but more modest reduction in XPD, cyclin H, and CDK7 subunits of TFIIH was also observed (Fig. 7C). This downregulation of TFIIH components was detected as early as 72 h after treatment of OM-10.1 cells with 10 μM SP and did not appear to magnify over time (Fig. 7D).

We also investigated whether long-term SP treatment affected the expression and function of other critical CDKs and their cyclin partners, such as P-TEFb (CDK9 and cyclin T1) and cell cycle proteins CDK2, cyclin E1, and cyclin A. No significant differences were observed between SP-, vehicle-, or EPL-treated OM-10.1 cells (Fig. 7B). Recently, CDK7 was shown to be the primary kinase responsible for CDK9 Ser175 phosphorylation thought to bias CDK9’s binding from BRD4 to Tat (30). Additionally, it was proposed that CDK7 phosphorylates CDK9 at Thr186, which is required for CDK9 activity, though CDK9 also autophosphorylates at this residue (30, 70). Given SP reduction of CDK7 protein expression (Fig. 7B and D), we compared the levels of pThr186 and pSer175 CDK9 by Western blotting in SP- and vehicle-treated OM-10.1 cells treated for 24 days (Fig. 7E). While potent depletion of XPB and approximately 60% reduction in CDK7 was observed, the levels of total, pThr186, and pSer175 CDK9 remained unchanged. Furthermore, by propidium iodide (PI) staining, no differences in cell cycle progression between treatment groups were observed (Fig. 7F).

In sum, although long-term treatment of cells with 10 μM SP results in a reduction in protein levels of additional components of the TFIIH complex besides XPB, we did not observe differences in the expression of other important kinases, cyclins, housekeeping genes, phosphorylation of CDK9, or in cell cycle progression.

Treatment of latently infected cells with SP reduces HIV-1 transcription without global dysregulation of cellular mRNA expression.

Due to the critical nature of TFIIH for RNAPII-mediated transcription and the significant loss of some of the components of TFIIH in addition to XPB (Fig. 3F, 5C, and 7), we investigated global cellular mRNA synthesis defects that may occur with SP treatment. As such, we treated OM-10.1 cells for 15 days with either SP+ART or DMSO+ART (once maximal HIV inhibition had been achieved in the SP-treated cells), followed by RNA extraction and mRNA-seq analysis. For this study, we included cells stimulated or not with the protein kinase C (PKC) agonist prostratin to assess effects on activation-dependent genes. HIV encoding genes were among the most significantly downregulated genes in cells treated with SP compared to those of the DMSO control, particularly upon stimulation with prostratin (Fig. 8A, left and right, red points). Some of the most significantly affected cellular genes are highlighted and labeled in blue [pAdj < 1e−10 and abs(log2-fold change) > 1 for SP versus DMSO comparisons (Fig. 8A, left and right; see also Tables S1 and S2 in the supplemental material) and pAdj < 1e−30 with an abs(log2- fold change) > 2.5 for stim versus no stim comparisons (Fig. 8A, middle two panels; see also Tables S3 and S4 in the supplemental material)]. In unstimulated cells, treatment with SP resulted in the upregulation of 49 genes and downregulation of 25 compared to those of the DMSO control [Fig. 8A left panel; pAdj ≤ 0.05, log2 FC ≥ abs(1)]. Prostratin stimulation led to the differential expression of 758 genes in DMSO-treated cells and 823 genes in SP-treated cells (546 of which were differentially expressed between conditions) (Fig. 8A, middle panels). SP treatment, while specifically affecting the expression of some genes, does not appear to affect all RNAPII-mediated transcription, which is in support of previous studies suggesting XPB depletion by SP does not affect global cellular transcription (51). To determine more rigorously what patterns of gene signaling pathways were affected by SP treatment, we performed gene set enrichment analysis (GSEA) (71). A bubble-lattice plot was constructed to highlight gene sets affected by SP treatment, where the color of the bubble represents the normalized enrichment score (NES) (indicating the number and differential intensity of the assessed genes in each indicated pathway), and statistical significance is indicated by the size of the bubble (Fig. 8B). HIV-encoded genes were significantly downregulated compared to DMSO control, particularly under conditions of prostratin stimulation, indicating the strong specific effect of SP treatment on HIV transcription (Fig. 8B). Of note, with this type of bubble representation of the data, it is important to keep in mind the size of the gene set; for instance, the MYC targets have modest downregulation with SP treatment, but since there are 200 genes in this set, the NES score is of greater magnitude. Conversely, downregulation of HIV genes is greater, but because there are a lot fewer genes in this set, the NES is of lower magnitude. Genes related to DNA packaging (histone- and chromatin-related genes) were significantly upregulated in both unstimulated and prostratin stimulated SP-treated cells (Fig. 8B). Conversely, MYC targets, E2F targets, RNAPII transcription termination, and type I interferon (IFN-I) receptor binding genes were downregulated with SP treatment, which increased in significance upon prostratin stimulation (Fig. 8B). Furthermore, cellular (endogenous) genes involved in the late phases of the HIV life cycle (reactome gene sets [72]) were modestly downregulated upon stimulation, which include genes involved in transcription, RNA processing, and export from the nucleus (Fig. 8B). IFN-γ response genes were slightly upregulated in unstimulated SP-treated cells, gaining significance upon stimulation (Fig. 8B). Other gene sets related to transcription of the HIV genome, RNAPII elongation, IFN-α response, housekeeping, and apoptosis were not substantially affected (Fig. 8B). Heatmaps were constructed from transcripts per million (TPM) data to facilitate direct comparison of gene expression with different lengths (Fig. 8C; see also Table S5 in the supplemental material). All HIV genes were clearly downregulated upon SP treatment and, although prostratin stimulation increased HIV expression, this was substantially reduced compared to that of DMSO-treated cells (Fig. 8C, left). Because of the modest downregulation in genes involved in the late phase of the HIV life cycle observed in the GSEA (Fig. 8B), we examined the expression of individual genes within this set. Expression of transcription-related genes, such as TFIIH, P-TEFb, the SEC, NELF, DSIF, and RNAPII genes do not appear to be substantially affected by SP treatment (Fig. 8D, left). Furthermore, there were no obvious differences in expression of RNA splicing, export, capping, or cleavage genes between SP- and DMSO-treated cells (Fig. 8C, middle). These results are consistent with previous results, in which SP did not affect the transcriptional machinery and cell cycle progression (Fig. 7) but blocked HIV reactivation (Fig. 4). Nonetheless, SP treatment does curiously seem to upregulate a large subset of DNA packaging complex genes (Fig. 8C, right panel), which is often associated with defects in cell cycle progression (73). However, this upregulation might not be biologically significant since, as mentioned above, no obvious differences in cell division were observed in SP-treated cells (Fig. 7).

In sum, SP treatment results in robust suppression of HIV transcripts both in unstimulated or prostratin stimulated cells, consistent with the observations in Fig. 1 and 4. While the expression of a few specific gene sets was significantly impacted by SP treatment, we hypothesize that their fold change may not be biologically significant since no major defects were observed in cell growth and viability. Nevertheless, the effects of SP treatment at concentrations sufficient to potently degrade XPB will need to be assessed in vivo.

Cell lines and cell culture.

OM-10.1 and ACH-2 cells were obtained through the NIH AIDS Reagent Program (from S. Butera, no. 1319, and T. Folks, no. 349, respectively). Cells were maintained in RPMI 1640 (Gibco) supplemented with 2.0 mM l-glutamine/100 U/ml penicillin/100 μg/ml streptomycin (Gibco) and 10% heat-inactivated fetal bovine serum (FBS) (Atlas Biologicals; catalog no. FS-0500-AD) (complete RPMI) and cultured at 37°C and 5% CO₂. Cells were passaged every 3 days, and medium was replaced with the indicated concentrations of SP, EPL, or DMSO and an ART cocktail (200 nM lamivudine [AIDS reagent], 200 nM raltegravir [Selleck Chemicals; MK-0518], 100 nM efavirenz [AIDS reagent]). SP was purchased from Sigma-Aldrich (no. 1619006) and dissolved in DMSO (Fisher Scientific; no. BP231-100). EPL was purchased from Sigma-Aldrich (no. E6657-10MG) and dissolved in DMSO. Cell concentration and viability was measured using the TC20 automated cell counter (Bio-Rad; no. 1450102) and trypan blue staining.

Infection of Hut78 cells with SIVmac239.

Hut78 cells were infected as previously described (90). Briefly, cells were infected overnight with SIVmac239. Cells were then washed with phosphate-buffered saline, split into 3 flasks, and treated with DMSO, EPL, or SP (10 μM) for 12 days. SIV replication was assessed every 3 days postinfection by measuring the viral capsid in the supernatant with p27 ELISA (Advances Bioscience Laboratories; no. 5436).

Determination of SP half-life in cell culture.

Stability of SP was evaluated in cell culture using an AB Sciex 5500 mass spectrometer run in multiple reaction monitoring mode. Cells were treated with 10 μM SP (with or without ART), and the concentration of SP was evaluated before and after 72 h of incubation at multiple time points. Cells were resuspended and treated with 2× vol/vol acetonitrile to lyse the cells, precipitate cellular protein, and inactivate the virus. Samples were filtered through a 0.2-μm filter plate prior to injection onto the LC-MS/MS. LC conditions used a Thermo BetaSil C₁₈ 5-μm, 50- by 2.1-mm column with a linear gradient of water + 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B) (10% B from 0 to 0.5 min, 90% B at 2 to 3 min, 10% B at 4 to 6 min). Detection of SP used the mass transition of 417.3→341. Half-life was calculated by plotting the natural log of the percent remaining compound verses time where t_1/2 = ln(2)/slope of the linear regression.

p24 ELISA.

The amount of HIV p24 capsid protein produced in the supernatant of cultured cells was quantified using the antigen capture assay kit from Advanced BioScience Laboratories, Inc. (no. 5447) and performed according to the manufacturer’s protocol.

MTT cell proliferation assay.

The viability and proliferation of cells treated with increasing concentrations of SP and EPL were measured using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay according to the manufacturer’s protocol (ATCC; no. 30-1010K). Cells were cultured in 96-well plates (in triplicate per drug concentration) for 72 h prior to addition of the MTT reagent.

Western blotting for OM-10.1, ACH-2, and HUT-78 cells.

Cells were collected, washed, and lysed with a 20 mM HEPES (pH 8.0), 100 mM KCl, 0.2 mM EDTA, 5 mM β-mercaptoethanol, 0.1% IGEPAL CA-630, 10% glycerol, and supplemented with complete EDTA-free protease inhibitor cocktail (Roche; no. 4693132001) on ice for 10 min. For pSer175 and pThr186 CDK9 Western blotting, we added phosphatase inhibitor cocktail 3 (Sigma-Aldrich; no. P0044-1ML). Protein concentration was determined using the Bio-Rad protein assay dye (no. 50000006). Equal amounts of total protein extract were loaded, alongside the Precision Plus Protein all blue protein standards (Bio-Rad; no. 1610373) onto a precast 4% to 20% gradient stain-free protein gel (Bio-Rad; no. 4568094). Proteins were then transferred to a nitrocellulose membrane using the Bio-Rad Trans-Blot Turbo system (no. 1704270). Membranes were blocked with 5% milk and probed with the following antibodies: anti-XPB (Sigma-Aldrich; no. X0879, diluted 1:2,000), anti-XPD (Cell Signaling Technology; no. 11963S, diluted 1:1,000), anti-CDK7 (Santa Cruz; sc-7344, diluted 1:500), cyclin H (Cell Signaling Technology; no. 2927S, diluted 1:1,000), p62 (Santa Cruz; sc-25329, diluted 1:500), cyclin A2 (Cell Signaling Technology; no. 4656S, diluted 1:1,000), cyclin E1 (Cell Signaling Technology; no. 4129S, diluted 1:500), CDK2 (Cell Signaling Technology; no. 25465S, diluted 1:1,000), CDK9 (Cell Signaling Technology; no. 2316S, diluted 1:1,000), pThr186-CDK9 (Cell Signaling Technology; no. 2549S), pSer175-CDK9 (kind gift from Jonathan Karn), β-tubulin (Proteintech; no. 66240-1-IG, diluted 1:10,000), and GAPDH (Santa Cruz; sc-32233, diluted 1:1,000) in 3% bovine serum albumin (BSA) (or 5% milk for the β-tubulin antibody). Secondary antibodies were anti-mouse (GE Healthcare; no. NA931-1ML) diluted 1:10,000 or anti-rabbit (Sigma-Aldrich; catalog no. A0545), diluted 1:15,000, both conjugated to horseradish peroxidase. Bands were visualized using the Amersham ECL prime Western blotting detection reagent (GE Healthcare; no. RPN2236) and imaged on the Bio-Rad ChemiDoc XRS+ system.

Quantitative PCR of cell-associated RNA.

Total RNA was extracted from cells using an RNeasy kit (Qiagen; no. 74106). RNA was DNase treated using the TURBO DNA-free kit (Invitrogen; no. AM1907). cDNA was synthesized using random hexamer primers and SuperScript III first-strand synthesis kit (Invitrogen; no. 18080051). Quantitative PCR (qPCR) was performed using SensiFast SYBR No-ROX kit (Bioline; no. BIO-98020). Samples were run in duplicate or triplicate. The mRNA expression was normalized to RPL13A mRNA expression, and the relative abundance was calculated (ΔΔC_T).

qPCR primers.

The following RPL13A primer was purchased from Bio-Rad: PrimePCR SYBR green assay RPL13A human no. 10025637. Other primers used were: XPB-F, GGATGAGTCAGGCACCAAAGT; XPB-R, CTGGAGAGAAGGCTTCCAAGAA; XPD-F, CGACTACATCTACCCCGAGCA; XPD-R, CATGATCAGGGCCAACAGGGA; CCNH-F, TCACCCCAGGATAATAATGCTCA; CCNH-R, CAGTATCTGTTCAAGTGCCTTCT; CDK7-F, GGAGCCCCAATAGAGCTTATACA; CDK7-R, TCCACACCTACACCATACATCC; GTF2H1-F, GACCTTGTTGTGAGTCAAGTGA; GTF2H1-R, CCTGCTTATGATTGGATGTGGAA; GTF2H2-F, CGTATGGGATTTCCTCAGCAC; GTF2H2-R, AGCCTCCTAATGTAAGCCCTG; GTF2H3-F, GAATGGCAGACTTGGAGACTTC; GTF2H3-R, GCAAAGTTTCTGTATGTTGACCC; GTF2H4-F, ACCCCATTTTCCGCCAGAAC; GTF2H4-R, CGGCGTACTTGTCAAGGGAG; GTF2H5-F, AAGACATTGATGACACTCACGTC; GTF2H5-R, GGGAAAAAGCATTTTGGTCCATT; MNAT1-F, GGTTGCCCTCGGTGTAAGAC; MNAT1-R, AGTTGCTCTTTCTGAGTGGAGT; AllHIVmRNA-F, TTGCTCAATGCCACAGCCAT; AllHIVmRNA-R, TTTGACCACTTGCCACCCAT; HIV early gag-F, GCGACTGGTGAGTACGCCAA; HIV early gag-R, CCCCTGGCCTTAACCGAATTT.

ChIP qPCR primers.

123-F, CCCTGATTGGCAGAACTACACAC; 123-R, TCTACCTTATCTGGCTCAACTGGT; 326-F, GACAGCCGCCTAGCATTTCAT; 326-R, CCACGCCTCCCTGGAAAGT; 412-F, CGAGAGCTGCATCCGGAGTACT; 412-R, GAGGCTTAAGCAGTGGGTTCC; 1097-F, GCAGTCCTCTATTGTGTGCATCAA; 1097-R, CCTGTGTCAGCTGCTGCTTG; 1477-F, CCATCAATGAGGAAGCTGCAGAA; 1477-R, GGTGGATTATGTGTCATCCATCCT; 3582-F, CAGAAATACAGAAGCAGGGGCAA; 3582-R, GTGTGGGCACCCTTCATTCTT; 6352-F, GAGCAGAAGACAGTGGCAATGA; 6352-R, CACAGGTACCCCATAATAGACTGTG; 8885-F, CCTGCTGTAAGGGAAAGAATGAGAC; 8885-R, CACCTCTTCCTCCTCTTGTGCT; GAPDH-F, CCTCACGTATTCCCCCAGGTTTA; GAPDH-R, AGCCACACCATCCTAGTTGCCT.

Cell cycle stage analysis.

OM-10.1 cells were treated long-term (80 to 100 days) with ART and 10 μM SP, 10 μM EPL, or DMSO, and they were split and treated every 72 h. After 72 h of incubation, 1 × 10⁶ cells was collected per condition (in triplicate) for cell cycle analysis. All centrifugation steps were performed at 500 × g for 5 min. For fixation, cells were washed in 1× Dulbecco's phosphate-buffered saline (DPBS) (Gibco; no. 14190-250) and then fixed in 66% ice-cold ethanol, on ice, for at least 2 h or stored at 4°C until the next day. For staining, fixed cells were equilibrated to room temperature, resuspended, and pelleted. Fixing solution was aspirated and cells were washed with 1× DPBS. The pellet was resuspended in 0.5 ml of FxCycle PI/RNase staining solution (Invitrogen; no. F10797) and stained for 15 to 30 min at room temperature (RT) and protected from light. Cells were acquired on a BD LSR II flow cytometer using the 561-nm laser with the 617/25 band-pass filter.

Transduction of OM-10.1 cells.

The Mission shRNAs targeting ERCC3 and GFP in the pLKO.1-puro plasmid were purchased from Sigma-Aldrich (SHCLNG TRCN0000359264 and SHC204, respectively). First, lentiviral vectors were prepared by transient transfection of 293T cells, with the shRNA vector and packaging plasmids pMD2.g and psPAX2 (gifts from Didier Trono; Addgene plasmid no. 12259 and no. 12260) at a molar ratio of 2:1:1, respectively. Transfections were performed using TransIT-LT1 transfection reagent (Mirus Bio LLC; no. MIR 2305) as per manufacturer’s instructions. Lentiviral virus-like particles (VLPs) were collected and filtered through a 0.45-μm filter, to remove cellular debris, at 48 and 72 h posttransfection. Filtered VLPs were combined, quantified, and stored at −80°C until use. Three days prior to transduction, medium was changed on OM-10.1 cells to remove ART to allow successful transduction with the VLPs. OM-10.1 cells were transduced with equal multiplicities of infection (MOIs) of lentiviral vectors in T25 flasks as follows: pLKO.1-shGFP, pLKO.1-ERCC3, pLKO.1-shGFP + 1.8 μM SP, and pLKO.1-shGFP + 6.3 μM SP in the presence of the indicated concentrations of SP. VLPs were added to cells and a nontransduced control was included. Cells were washed after an overnight incubation, and ART (± the indicated SP concentrations) was added back to the medium. Two days later, cells were selected with 1.5 μg/ml puromycin (Gemini Bio-Products; no. 400-128P). Cells were maintained in puromycin-supplemented medium treated with the indicated concentrations of SP every 3 days. Ten days post-puromycin selection, cells were harvested for XPB protein analysis by Western blotting and RNA expression analysis of ERCC3 and HIV by RT-qPCR. Additionally, supernatant was collected from each transduction for quantification of HIV capsid protein by p24 ELISA. Cells were counted, and viability was determined by trypan blue staining using the TC20 automated cell counter.

Latent HIV-1 reactivation.

OM-10.1 cells (after 45 days of treatment with ART and 10 μM SP or DMSO) and ACH-2 cells (after 15 days of treatment with ART and 10 μM SP or DMSO) were stimulated (or not) with 1 μM prostratin (LC Laboratories; no. P-4462-1MG), 10 ng/ml TNF-α (BioLegend; no. 570104), 1 μM SAHA (LC Laboratories; no. V-8477_1g), or 20 nM PMA (Sigma-Aldrich; no. P1585-1MG). This was done in the presence of ART and either 10 μM SP or DMSO. After 8 h, p24 production was measured in the supernatant by ELISA, and cells were collected for RNA extraction and RT-qPCR analysis.

J-Lat A1 cells were treated with 160 nM PMA or 1 ng/ml TNF-α for 2 h and then treated overnight with 10 μM SP, EPL, or DMSO, after which the mean fluorescence intensity of GFP was measured by flow cytometry. J-Lat 11.1 cells were treated with 8 nM PMA for 2 h following overnight treatment with 10 μM SP, EPL, or DMSO.

Isolation, expansion, and reactivation of primary CD4⁺ T cells.

Resting CD4⁺ T cells were isolated from 7 ART-treated, virally suppressed, HIV-1-infected individuals through magnetic negative depletion as previously described (91) and treated with SP (10 μM) or mock (DMSO) for 24 h. The fusion inhibitor enfuvirtide (T20) was supplemented to the cell culture to prevent new rounds of infection. In the final 6 h of culture, the cells were treated with PMA (50 ng/ml) and ionomycin (1 μM) to induce maximum ex vivo latency reversal as previously described (64). Cell-associated HIV-1 RNA was extracted using TRIzol phase separation and DNase treatment. HIV-1 RNA expression was measured by RT-qPCR as cell-associated polyadenylated HIV-1 levels per million CD4⁺ T cells (92, 93). Cellular gene expression (POLR2A; Thermo Fisher; no. 4331182) was measured as cell-associated polyadenylated POLR2A levels per million CD4⁺ T cells.

RNA polymerase II ChIP.

Three independent treatments of OM-10.1 cells with ART and 10 μM SP or DMSO for ∼70 days were stimulated (or not) with 1 μM prostratin for 8 h. Supernatant and cells were collected for p24 ELISA and RT-qPCR analysis, respectively, and remaining cells were cross-linked with 1% formaldehyde for 10 min and quenched by 0.125 M glycine for 5 min at room temperature. Cells were lysed using an SDS lysis buffer, and lysates were sonicated 16 times for 10-s bursts on ice to generate sheared chromatin of 200 to 400 nucleotides in the presence of complete EDTA-free protease inhibitor cocktail (Roche; no. 4693132001) and phosphatase inhibitor cocktail 3 (Sigma-Aldrich; no. P0044-1ML). The protein concentration of the precleared sonicated lysate was determined using the Bradford protein assay. A 4% fraction of each sonicated lysate was reserved for the input control. Four micrograms (per condition) of each of the antibodies total RNAP II (Millipore; no. 05-623), pSer5 RNAPII CTD (Active Motif; no. 61986), and pSer2 RNAPII CTD (Active Motif; no. 61984) or isotype controls mouse IgG (Invitrogen; no. 31903) or rat IgG (Thermo Fisher Scientific; no. 029602) were conjugated to protein G Dynabeads (Thermo Fisher Scientific; no. 10003D) for 20 min, rotating at RT. A total of 500 μg precleared protein lysate was added to the conjugated beads for immunoprecipitation (IP) overnight with rotation at 4°C. Dynabeads were washed using magnetic separation. Samples were then treated (including input samples) with freshly prepared digestion buffer (10 mM Tris-HCl, 1 mM EDTA, 0.5% SDS, 10 μg/ml RNase A [Thermo Scientific; no. EN0531]) at 37°C for 30 min. Cross-links were reversed for at least 4 h shaking (1,000 rpm) at 65°C with 200 nM NaCl. DNA samples were then treated with proteinase K (Fisher Scientific; no. BP1700100) shaking (1,000 rpm) at 60°C for 1 h. Dynabeads were removed, and DNA was purified using the QIAquick PCR purification kit (Qiagen; no. 28106). qPCR was performed in 384-well format using SensiFast SYBR No-ROX kit (Bioline; no. BIO-98020). The average threshold cycle (C_T) value of the IgG-only background for each primer was subtracted from the C_T value using the HIV-1 primer. The data were then normalized to input values using the following formula: % of input = [2-∂C_T(IP) – 2-∂C_T(IgG)] × 4.

mRNA-seq library preparation.

Total RNA was extracted using the RNeasy kit (Qiagen; no. 74106). RNA was DNase treated using the TURBO DNA-free kit (Invitrogen; no. AM1907). Total RNA was quantified on the Qubit 2.0 fluorometer and run on the Agilent 2100 Bioanalyzer RNA pico chip (Agilent Technologies, Santa Clara, CA) for quality assessment. mRNA was selectively isolated from total RNA (500 ng) using poly-T oligonucleotides attached to magnetic beads according to the manufacturer’s guidelines in the NEBNext poly(A) mRNA magnetic isolation module (catalog no. E7490; NEB, Ipswich, MA). The library preparation from the enriched mRNA was conducted according to NEBNext Ultra II directional RNA kit (catalog no. E7760; NEB, Ipswich, MA). Briefly, the enriched mRNA samples were chemically fragmented at 94°C for 15 min. The fragmented RNA was random hexamer primed and reverse transcribed to generate the first-strand cDNA. The second strand was synthesized after removing the RNA template and incorporating dUTP in place of dTTP. The incorporation of dUTP quenches the second strand during the PCR amplification step later, and therefore the strand information was preserved. The purified double-stranded (ds) cDNA was end repaired and adenylated at their 3′ ends. A corresponding “T” nucleotide on the adaptors was utilized for ligating the adaptor sequences to the ds cDNA. The adaptor ligated DNA was purified and PCR amplified to incorporate unique barcodes to generate the final libraries. The final libraries were validated on the bioanalyzer DNA chips, pooled equally, and loaded onto the NextSeq 500 for sequencing with 2 × 80-bp paired-end chemistry.

Read mapping and differential expression analysis.

Raw FASTQC files were filtered for read quality using trim-galore version 0.6.1 with the following options: –clip_R1 5 –clip_R2 5 –trim-n –paired (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/). Trimmed reads were aligned to the human transcriptome (GRCh38v100) with the HIV-1 genome added as an additional chromosome (REFSEQ release 202). To align transcripts, salmon version 0.14.1 was used with the validateMappings option (94). Aligned transcriptomes were mapped to genes, and differential gene expression was called using tximport version 1.16.1 and DESeq2 version 1.28.1, respectively (95, 96). TPMs were determined using the countsFromAbundance = “lengthScaledTPM” option in tximport. All plots were generated in R. GSEAs were computed from the GSEA app version 4.1.0 using TPMs as counts and publicly available gene sets current as of October 2020 (71). DESeq2 and TPM data can be found in Tables S1 to S5 in the supplemental material.

Statistical analysis.

Statistics were performed using GraphPad Prism, and a P value of <0.05 was considered significant for all comparisons. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. Not significant (NS) unless indicated. Data are presented as the mean of the indicated replicates with error bars indicating the standard deviation (SD). The two-tailed t test was used where applicable, and a paired t test was used for reactivation studies to compare stimulated conditions to unstimulated conditions. Comparisons of multiple conditions were performed using a two-way analysis of variance (ANOVA) repeated-measures test (Dunnett or Holm-Sidak).