Histone Deacetylase Inhibitor Romidepsin Inhibits De Novo HIV-1 Infections

Source and isolation of primary human cells.

Healthy blood donor buffy coat fractions were obtained from the Department of Clinical Immunology Blood Bank at Aarhus University Hospital, Aarhus, Denmark. Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll density separation. CD4⁺ T cells were enriched by negative depletion from PBMCs using EasySep human CD4⁺ T cell enrichment kit (catalog no. 19052; Stem Cell Technologies) or CD4⁺ T cell isolation kit (catalog no. 130-096-533; Miltenyi Biotec) according to the respective manufacturer's protocol. Monocyte-derived macrophages (MDMs) were generated from monocytes purified from PBMCs by plastic adherence, with subsequent culture for 7 days in RPMI growth medium (RPMI 1640, 50,000 IU of penicillin, 50,000 μg of streptomycin, 10% fetal calf serum, 1% glutamine, interleukin-2 [IL-2; 20 U/ml]) supplemented with 10 ng of macrophage colony-stimulating factor/ml and 1 ng of granulocyte-macrophage colony-stimulating factor/ml (both from PeproTech). Resting CD4⁺ T cells were enriched from cryopreserved PBMCs by negative depletion via a two-step protocol as previously described (20). Briefly, the first step was to enrich CD4⁺ T cells from PBMCs using Miltenyi CD4⁺ T cell isolation kit (catalog no. 130-096-533) according to the manufacturer's protocol. The second step was to further enrich for resting CD4⁺ T cells via depletion of cells expressing CD69, CD25, or HLA-DR (Miltenyi CD69 Microbeads kit II, catalog no. 130-092-355; CD25 Microbeads II, catalog no. 130-092-983; HLA-DR Microbeads, catalog no. 130-046-101). All cell incubations were performed at 37°C unless otherwise noted. The purity of enriched cells populations was assessed by flow cytometry using either a LSR Fortessa (BD Biosciences) or a FACSVerse (BD Biosciences) flow cytometer and FlowJo software (Tree Star).

Viability assay for TZM-bl cells, PBMCs, and activated CD4⁺ T cells.

The viability of TZM-bl cells (21–25) and primary human cells was assessed after exposure to romidepsin (Selleckchem) or panobinostat (Selleckchem) (26–28). We tested a range of drug doses that spanned the physiologically relevant doses of each drug (i.e., 40 nM romidepsin [14] and 30 nM panobinostat [15, 29]) and, in select experiments, subphysiological doses of romidepsin were included along with medium controls to facilitate the observation of dose-dependent effects on HIV infection.

TZM-bl cells in cDMEM (Dulbecco modified Eagle medium, 50,000 IU of penicillin, 50,000 μg of streptomycin, 10% fetal calf serum, 1% glutamine) were seeded (5 × 10³ cells/well) in 96-well plates, incubated for 24 h, and then primed (8 or 18 h) with romidepsin or panobinostat at the indicated range of concentrations (2-fold dilutions). After priming, the cells were washed with cDMEM and incubated for 48 h. TZM-bl cells were treated with trypsin, resuspended in 100 μl of cDMEM, and transferred to 96-well enzyme-linked immunosorbent assay (ELISA) plates, and 20 μl of MTS substrate (CellTiter 96 AQ_ueous One Solution cell proliferation assay; Promega) was added to each well. Plates were incubated for 4 h, and formazan production was then analyzed with an ELISA plate reader (BioTek, ELx808). Cell viability was calculated according to the CellTiter protocol.

PBMCs or CD4⁺ T cells were stimulated in RPMI growth medium supplemented with phytohemagglutinin (PHA; 5 μg/ml) for 48 h. PHA-containing medium was replaced with fresh RPMI growth medium, and cells were incubated for 24 h. The PBMCs or CD4⁺ T cells were then seeded (2 × 10⁵ cells/well) in 96-well plates and primed (8 or 18 h) with romidepsin or panobinostat at the indicated concentrations (2-fold dilutions). After priming, the cells were washed, resuspended in RPMI growth medium, and incubated for 48 h. Cells were resuspended in 100 μl of RPMI growth medium and assayed for formazan production as described above.

Single-round, VSVg-pseudotyped HIV infection assay.

Vesicular stomatitis virus G protein (VSVg)-pseudotyped HIV virions were produced by transfecting HEK293T cells with a packaging system consisting of four plasmids: pCCL-PGK-eGFP, pMD.2G, pRSV-REV, and pMDlg/p-RRE. Viral supernatants were harvested after 48 and 72 h and concentrated through a 20% sucrose cushion (25,000 × g for 2 h). Viral pellets were resuspended in phosphate-buffered saline and pooled, and aliquots were stored at −80°C.

PBMCs or enriched CD4⁺ T cells were stimulated in RPMI growth medium supplemented with PHA (5 μg/ml) for 48 h. The PHA-containing medium was replaced with fresh RPMI growth medium, and the PBMCs or enriched CD4⁺ T cells were incubated for 24 h. MDMs were grown as described above. Cells were seeded (2 × 10⁵ cells/well) in 48-well plates with RPMI growth medium and pulsed for 2 h with romidepsin (10 or 40 nM) or panobinostat (30 nM). VSVg-pseudotyped HIV virions were then inoculated into the cell cultures, along with Polybrene (6 μg/ml), followed by incubation for an additional 6 h. The cells were then washed to remove unbound virions and resuspended in fresh RPMI growth medium and incubated for 48 h. The cells were then harvested by aspiration for the PBMCs and T cells or after EDTA treatment for the MDMs. Cells were incubated on ice with a Live/Dead fixable violet dead cell stain kit 405 (Life Technologies) according to the manufacturer's instructions. Data were collected using a LSR Fortessa flow cytometer. Viable cells exhibiting enhanced green fluorescent protein (eGFP) expression were characterized as infected.

HIV spreading assay.

Enriched CD4⁺ T cells were stimulated in RPMI growth medium supplemented with PHA (5 μg/ml) for 48 h. PHA-containing medium was replaced with fresh RPMI growth medium, and cells were incubated for 24 h. CD4⁺ T cells were pulsed for 2 h with romidepsin (10 or 40 nM) or panobinostat (30 nM). The cells were then infected with HIV-1_HXB2 (30). After an additional 4 h of incubation, the HDACi and virus inoculum were removed, and the cells were resuspended in fresh RPMI growth medium. At the indicated time points, culture supernatants were harvested and evaluated for p24 antigen by ELISA to determine the replication kinetics of HIV as previously described (9).

Romidepsin and IFN-stimulated genes (ISGs) in PBMCs.

PHA-activated PBMCs were seeded (10⁵ cells/well) into a conical 96-well plate and primed for 2 or 8 h with romidepsin (10 or 40 nM) or beta interferon (IFN-β; 1,000 IU/ml). After priming, RNA was collected from the cells using a High Pure RNA isolation kit (catalog no. 11828665001; Roche) according to the manufacturer's instructions. The RNA expression levels of 39 different genes were subsequently analyzed by using a Fluidigm 48-48 Multiplex Gene Array Biomark system, including TaqMan probes, according to the manufacturer's instructions.

Resting CD4⁺ T cell proliferation and viability assay.

Resting CD4⁺ T cells were incubated with eFluor-670 (catalog no. 65-0840-85; eBioscience) according to the manufacturer's instructions. Dyed resting CD4⁺ T cells (10⁶ cells per well) were placed into a 24-well plate in cRPMI+IL-2 medium (RPMI 1640 with l-glutamine, 50,000 IU of penicillin, 50,000 μg of streptomycin, 10% fetal calf serum, IL-2 [10,000 U/ml], and conditioned medium from a mixed-lymphocyte reaction culture as described previously [31]) plus PHA (0.5 μg/ml) and/or romidepsin (4 or 40 nM). After an 18-h incubation, the cultures were washed twice (2-h interval) with cRPMI+IL-2 medium and then maintained in 2 ml of cRPMI+IL-2 medium for 4 additional days. Cells were incubated on ice with a Live/Dead Fixable Green Dead Cell Stain Kit 488 (L-23101; Life Technologies) according to the manufacturer's instructions. The data were collected using a FACSVerse flow cytometer. Viable cells exhibiting a dilution of eFluor-670 were characterized as having proliferated during the culture period. The fold proliferation was determined relative to the respective human donor “medium-only” culture. The percentage of viable cells represents the proportion of resting CD4⁺ T cells, which excluded the fixable green dead cell stain.

Activation of latent HIV in ACH2 cells.

ACH2 cells (12, 32, 33) were incubated for 18 h with or without PHA (0.5 μg/ml) or romidepsin (4 or 40 nM), washed twice (using a 2-h interval) with cRPMI+IL-2 medium, and then incubated in fresh cRPMI+IL-2 medium for 48 h. Viral p24 antigen was measured in culture supernatants by ELISA as an indicator of the ability of either romidepsin or PHA to induce virus production from this latent HIV cell model (9).

Modified viral outgrowth assay.

The viral outgrowth assay was performed essentially as described previously (34). Briefly, resting CD4⁺ T cells were isolated from healthy blood donor buffy coat fractions. On day 0, 10⁶ resting CD4⁺ T cells per well from each donor were plated into a 24-well plate, together with the indicated number of ACH2 cells in cRPMI+IL-2 medium with PHA (0.5 μg/ml) or romidepsin (4 or 40 nM). After 18 h, the cultures were washed twice (using a 2-h interval) with cRPMI+IL-2 medium, and then the cultures were moved into six-well plates and fed with 5 × 10⁵ MOLT-4/CCR5 cells in a final volume of 8 ml of cRPMI+IL-2 medium. On day 7, 6 ml of the culture volume was replaced with fresh cRPMI+IL-2 medium. On day 14, 200 μl of each culture supernatant was transferred to TZM-bl cells in duplicate, and the luciferase activity from these reporter cells was measured on day 16 as an indicator of HIV production within a given culture (15).

Statistics.

Comparisons of infectivity within a single-round infection assay were performed using Student t tests of each drug condition versus dimethyl sulfoxide (DMSO) controls. Differences in replication (using the area under the curve [AUC]) in the HIV spreading assay were evaluated by a one-way repeated-measures analysis of variance (ANOVA). Comparisons between modified viral outgrowth conditions were made using a two-way ANOVA wherein the sources of variation were either the latency reactivation treatment or the number of ACH2 cells per well.

Limited HDACi-induced cell death in primary human cells.

Our research focus was to determine the extent of romidepsin's impact on de novo HIV infection. As a control for the generalizability of our observations with romidepsin to other HDACi, we also evaluated the impact of panobinostat on de novo HIV infection in selected assays. To design the best strategies for evaluating the impact of HDACi on de novo viral infection, we first determined the toxicity of each drug in (i) HIV reporter TZM-bl cells, (ii) primary human PBMCs, and (iii) primary CD4⁺ T cells. The dose range examined included physiologically relevant doses of each drug (i.e., 40 nM romidepsin [14] and 30 nM panobinostat [15, 29]). Furthermore, we designed these analyses to detect potentially delayed toxic effects of drug exposure by measuring viability at both 24 and 48 h after removal of the drug (Fig. 1A, left and middle panels). Romidepsin and panobinostat treatment of TZM-bl cells resulted in only a minor impact on cell viability at 24 h, with the exception of the very high doses of romidepsin. In contrast, cell viability at 48 h was dramatically affected by romidepsin treatment, where cell death exceeded 90% at ≥100 nM (Fig. 1A, middle panel). Cell death was less pronounced for panobinostat, even at high doses (Fig. 1A, left and middle panels). In these experiments, TZM-bl cells were exposed to the HDACi for 8 h. Given that key ex vivo analyses have incorporated drug exposures with these HDACi of 18 h (34); we also evaluated 18 h of drug exposure in TZM-bl cells and observed a notably greater reduction in cell viability (Fig. 1A, right panel). As little as 7 nM romidepsin resulted in >80% of TZM-bl cells dying within 24 h. We observed ∼50% cell death at 30 nM panobinostat and 80% cell death at 125 nM. Next, to explore whether HDACi toxicity also occurred in primary human cells, we determined the viability of both PBMCs and CD4⁺ T cells exposed to romidepsin or panobinostat for a total of 8 h. We found that romidepsin-induced cell death in primary cells was less pronounced versus TZM-bl cells. Specifically, ca. 70 to 80% of PBMCs and CD4⁺ T cells remained viable for 48 h after exposure to the drug concentrations used in our subsequent virus infection assays (Fig. 1B and C).

Romidepsin inhibits VSVg-pseudotyped HIV infection of CD4⁺ T cells.

To understand the impact of HDACi on de novo viral infection, we performed a series of primary human cell experiments to measure de novo infection following HDACi treatment. In the first line of experiments, IL-2/PHA-activated PBMCs were primed with increasing amounts of either romidepsin or panobinostat for 2 h prior to incubation with VSVg-pseudotyped single-round HIV virions expressing eGFP (Fig. 2A). The live cells in each culture were examined using flow cytometry to quantitate the proportion of eGFP positive cells as a measure of infectivity (Fig. 2B). We found that priming cells with panobinostat did not alter the proportion of infected PBMCs (Fig. 2B and C); however, romidepsin significantly reduced the infectivity of VSVg-pseudotyped HIV at a relatively low concentration (10 nM, P ≤ 0.01; 95% confidence interval [CI] = 1.874 to 8.240), as well as at a more physiologically relevant concentration (40 nM, P ≤ 0.001; 95% CI = 2.845 to 9.211) (Fig. 2B and C). Collectively, our experiments with activated PBMCs showed a substantial reduction in infectivity after exposure to HDACi, although this experimental approach did not permit cell lineage-specific data interpretation.

To examine whether the observed inhibition of infectivity by romidepsin specifically affected CD4⁺ T cells and/or macrophages, we sorted CD4⁺ T cell populations from activated PBMCs prior to HDACi priming and generated MDMs. Similar to the results described above, significantly fewer primary CD4⁺ T cells were infected after priming with 10 or 40 nM romidepsin (P ≤ 0.01 [95% CI = 3.017 to 13.32] and P ≤ 0.0001 [95% CI = 5.910 to 16.22], respectively) (Fig. 2C). In contrast, MDMs were generally permissive for infection by the VSVg-pseudotyped single-round HIV virions regardless of HDACi treatment (see Fig. S1 in the supplemental material). Specifically, we observed >20% infection in DMSO-, romidepsin-, and panobinostat-treated MDMs derived from one of the donors whose CD4⁺ T cells exhibited profoundly reduced infection after exposure to romidepsin. Together, these data indicate that de novo infection of activated CD4⁺ T cells by VSVg-pseudotyped HIV was inhibited by romidepsin.

Romidepsin inhibits wild-type HIV infection of CD4⁺ T cells.

Next, we examined whether the mechanism of inhibition was specific to the pH-dependent fusion that characterizes VSVg-pseudotyped virions or whether wild-type HIV envelope fusion was similarly affected by treatment. To do this, HDACi-primed activated CD4⁺ T cells were challenged with replication-competent HIV-1_HXB2—a highly pathogenic, CXCR4-tropic, laboratory-adapted viral isolate (30) (Fig. 3A). CD4⁺ T cell cultures primed with 10 nM romidepsin exhibited significantly reduced infection after 2 days in culture, as measured by Gag^p24 in supernatants (P ≤ 0.05) (Fig. 3B and C). Similar effects were observed for romidepsin at 40 nM (P ≤ 0.05) but not for 30 nM panobinostat. By day 5, reduced HIV replication was only observed in the cultures primed with 40 nM romidepsin (P ≤ 0.001). This pronounced effect was maintained through 7 days of the culture, despite the fact that the drug pulse was relatively short in duration (Fig. 3B and C). The exposure of activated CD4⁺ T cells to physiologically relevant doses of romidepsin significantly decreased viral spreading, and thus the effect of romidepsin on de novo HIV infection is not dependent on the viral route of entry.

Romidepsin selectively activates ISGs in PBMCs.

To gain further insights into a potential mechanism for the HIV inhibition exhibited by romidepsin, we investigated the gene profile in PBMCs challenged with either romidepsin or panobinostat. We found that nearly all tested innate immune sensors, as well as other genes examined, were affected by 2 or 8 h of panobinostat exposure (e.g., the downregulation of MX1, Viperin, APOBEC3G, NF-κB, and IRF7) (Fig. 4). In contrast, whereas the expression of most genes was unchanged after 2 h of exposure to romidepsin, 8 h of 40 nM romidepsin exposure led to major alterations in the gene profile (Fig. 4, right panel). Expression was downregulated for most of the examined genes; however, there were three notable exceptions: IFIT1, ISG15, and STAT1. The expression of each of these genes was increased after 8 h of exposure to 40 nM romidepsin (Fig. 4, right), a finding indicative of differential regulation of innate restriction factors important for viral confinement.

Romidepsin inhibits proliferation of resting CD4⁺ T cells and reduces the sensitivity of viral outgrowth assays.

Given our observations that romidepsin reduced de novo HIV infection and virus spread in culture, we investigated the impact of this drug specifically on resting CD4⁺ T cells. We chose these cells because they are the primary cell type included in the viral outgrowth assays used to quantitate the size of the latent reservoir in HIV patients (20, 31, 34–39). The impact of romidepsin on the proliferation of magnetically enriched, HIV-negative blood donor-derived resting CD4⁺ T cells was profound. We first confirmed a previously reported observation (40) that exposure to romidepsin interferes with T cell proliferation when combined with mitogenic stimulation (see Fig. S2 in the supplemental material). However, it was essential that we also evaluate the impact of romidepsin (18-h drug exposure) in the absence of mitogenic stimulation to more closely approximate the conditions utilized in viral outgrowth assays (34) (Fig. 5A). We found that both the proliferation (Fig. 5B and C) and cell viability (Fig. 5D) of resting CD4⁺ T cells were dramatically reduced by 18 h of exposure to romidepsin.

Latent HIV infection in resting CD4⁺ resting T cells is infrequent (∼1 infectious unit per million resting CD4⁺ T cells tested [IUPM]). Viral outgrowth assays are extended-duration (∼14 days), stimulatory cultures of resting CD4⁺ T cells that are optimized for the quantitation of the latent HIV reservoir (20, 31, 34–39). To test our hypothesis that romidepsin's broad impact on cell viability, cell proliferation, and de novo HIV infection could alter the sensitivity of viral outgrowth assays and to experimentally overcome the limitation posed by the fact that latency events in cART-suppressed HIV patients are rare, we established a modified viral outgrowth assay with 10⁶ resting CD4⁺ T cells per culture from HIV-negative blood donors spiked with a defined number of ACH2 cells. ACH2 cells are a model for HIV latency (12, 32, 33), where cells can be stimulated to produce HIV progeny virions by romidepsin exposure (as with latently infected resting CD4⁺ T cells [14]), but not the mitogenic stimulator PHA (unlike latently infected resting CD4⁺ T cells [Fig. 5E]). Three different treatment conditions were evaluated in this modified viral outgrowth system: medium only, 0.5 μg of PHA, and 40 nM romidepsin (Fig. 5F). When we tested the typical patient condition of 1 IUPM, we observed viral outgrowth in only 17% (1 of 6) of PHA-treated wells tested and 0% (0 of 6) in the medium-only or romidepsin (0 of 8) wells. When a 10-fold-higher IUPM was stimulated, we observed 0% viral outgrowth in the romidepsin wells (0 of 8), whereas 33% of the medium-only wells (2 of 6) and 50% of the PHA-stimulated cultures (3 of 6) exhibited viral outgrowth. When 100 IUPM was evaluated, only 50% of the romidepsin cultures (4 of 8) versus 83% of the medium-only wells (5 of 6) and 100% of the PHA wells (6 of 6) were positive for viral outgrowth. Even at 1,000 IUPM, not all wells treated with romidepsin (5 of 6) were positive for viral outgrowth, as was the case for both medium control (6 of 6) and PHA-treated (6 of 6) wells. We conclude from these data that romidepsin reduced the sensitivity of the viral outgrowth assay and that romidepsin's induction effects are obscured within this assay with the typical patient condition of 1 IUPM.