A Novel Histone Deacetylase Inhibitor, AR-42, Reactivates HIV-1 from Chronically and Latently Infected CD4⁺ T-cells

Abstract

Human immunodeficiency virus type 1 (HIV-1) latency is a major barrier to a cure of AIDS. Latently infected cells harbor an integrated HIV-1 genome but are not actively producing HIV-1. Histone deacetylase (HDAC) inhibitors, such as vorinostat (SAHA), have been shown to reactivate latent HIV-1. AR-42, a modified HDAC inhibitor, has demonstrated efficacy against malignant melanoma, meningioma, and acute myeloid leukemia and is currently used in clinical trials for non-Hodgkin’s lymphoma and multiple myeloma. In this study, we evaluated the ability of AR-42 to reactivate HIV-1 in the two established CD4⁺ T-cell line models of HIV-1 latency. In HIV-1 chronically infected ACH-2 cells, AR-42-induced histone acetylation was more potent and robust than that of vorinostat. Although AR-42 and vorinostat were equipotent in their ability to reactivate HIV-1, AR-42-induced maximal HIV-1 reactivation was twofold greater than vorinostat in ACH-2 and J-Lat (clone 9.2) cells. These data provide rationale for assessing the efficacy of AR-42-mediated HIV-1 reactivation within primary CD4⁺ T-cells.

During primary infection, human immunodeficiency virus type 1 (HIV-1) infects permissive cells and converts its single-stranded RNA genome into a double-stranded DNA genome that integrates into the host-cell genome.¹ A subset of the cells harboring integrated HIV-1, termed the latent reservoir, does not actively produce HIV-1 progeny and is thus refractory-to-current antiviral therapy.^2,3 The posttranslational modifications of chromatin, such as histone deacetylation, cause chromatin condensation, which restricts RNA polymerase-mediated HIV-1 transcription and results in viral latency (reviewed in Siliciano and Greene).⁴ Previous reports have demonstrated the ability of histone deacetylase (HDAC) inhibitors, including vorinostat (also known as SAHA) and valproic acid, to reactivate latent HIV-1 through the reversal of chromatin condensation, although there have been inconsistent reports on the effectiveness of valproic acid.^5,6 Clinical studies of vorinostat investigating the kick and kill strategy indicate consistent HIV-1 reactivation from cell lines and HIV-infected patients, but at high dosages.^7,8 Additionally, recent studies with panobinostat and romidepsin in patients on suppressive antiretroviral therapy indicate the potential utility of more potent HDAC inhibitors.⁵

The histone deacetylation activity within chromatin indicates HDAC inhibitors as potentially valuable therapeutic agents for HIV-1 reactivation.^9–11 Currently, the most potent HDAC inhibitors belong to the hydroxamic acid family.¹² This class of HDAC inhibitors includes the US Food and Drug Administration-approved vorinostat and a novel compound AR-42.^5,10 AR-42 is a novel anticancer drug candidate that inhibits deacetylation on both histone and nonhistone proteins.^13,14 AR-42, a modified hydroxamic acid, was rationally designed with an aromatic linker and two Zn²⁺-binding motifs that bind a zinc cation in the catalytic domain of class I and II HDACs with an IC₅₀ value of 30 nM.¹⁵

Published data indicated that AR-42 induces histone H3 acetylation in mouse and canine mast cells.¹⁶ To determine if AR-42 induces acetylation in cells harboring a HIV-1 provirus, we treated chronically and latently infected ACH-2 cells¹⁷ (obtained from Dr. Thomas Folks through the NIH AIDS Research and Reference Reagent Program) with a range of AR-42 (1 nM–5 μM). Following the treatment, cell lysates (15 μg) were electrophoresed on a 10% SDS-PAGE gel and transferred to nitrocellulose. Histone H3 acetylation on lysine 9 was assayed by western blot with the AcH3K9 antibody (Santa Cruz Biotechnology, Inc., 1:1500 dilution) and goat–anti-rabbit immunoglobulin/horseradish peroxidase secondary antibody (cell signaling, 1:5000 dilution). Equivalent protein loading was verified by western blot against actin (cell signaling 4967, 1:1500). Histone acetylation was quantified as a ratio to actin loading control by ImageJ densitometry analysis.

At 10 nM, AR-42 treatment increased histone 3 acetylation, while vorinostat induced acetylation at ~100 nM (Fig. 1). Within the concentrations tested, AR-42-induced histone 3 acetylation was more robust than vorinostat-induced acetylation. As expected, phorbol 12-myristate 13-acetate (PMA)-mediated HIV reactivation did not increase histone 3-acetylation.

Figure 1.

Vorinostat and AR-42 increase histone acetylation. Cellular lysates (15 μg) from ACH-2 cells were loaded per lane and probed with antibodies against acetylated histone H3 and actin. PMA treatments (0.1% DMSO and 100 ng/mL) were negative controls. AR-42 and vorinostat concentrations range from 1 nM to 5000 nM. Densitometry quantification of the actin-loading control and histone 3 acetylation (ImageJ) is displayed as the ratio of histone acetylation intensity to actin-loading control intensity.

An outcome of histone acetylation in latently and chronically infected CD4⁺ T-cells is the reactivation of HIV-1. Expanding on AR-42’s ability to acetylate histone 3 (Fig. 1), we determined AR-42-induced HIV-1 reactivation within two well-established CD4⁺ T-cell models of HIV-1 latency.^17,18 ACH-2 cells were maintained in Roswell Park Memorial Institute medium with 10% fetal bovine serum and penicillin–streptomycin at 37°C under 5% CO₂. ACH-2 cells were treated with the indicated concentrations of vorinostat or AR-42 for 48 hours, in triplicate, at a final dimethyl sulfoxide (DMSO) concentration of 0.1%. A total of 100 ng/mL PMA (Sigma-Aldrich), also in 0.1% DMSO, was used as a positive control. After incubation, 10 μL of culture supernatant was removed, frozen at −80°C, thawed at room temperature, and then assayed for reverse transcriptase (RT) activity assays as described in Ball et al.¹⁹ HIV-1 reactivation was quantified using density (counts/mm²) counts computed by the Typhoon Scanner (GE Healthcare Life Sciences) and the Quantity One software (Bio-Rad Life Science Research). In the ACH-2 cell model, AR-42 reactivated HIV-1 in a dose-dependent manner, while vorinostat achieved a plateau at 500 nM (Fig. 2A). Although both AR-42 and vorinostat have similar potency (460 ± 0.05 nM and 408 ± 0.04 nM, respectively), at higher concentrations, AR-42 is twofold more efficacious than vorinostat in ACH-2 cells.

Figure 2.

AR-42 more effectively induces HIV-1 reactivation and expression from latently infected CD4⁺ T-cells than vorinostat. (A) RT activity of treatment over% PMA activation after 48 hours (average ± SD, n = 3). Calculated EC₅₀ values for both AR-42 and vorinostat are depicted. (B) HIV-1 latently infected J-Lat cells (clone 9.2) were treated with AR-42 or vorinostat at the indicated concentrations for 24 hours, and GFP-positive cells were scored by flow cytometry. The maximum% of GFP-positive cells was determined with the positive control TNF-α (10 ng/ml), which was set to 100%, and the percentage of activation induced by each drug relative to TNF-α is presented.

The second T-cell model, Jurkat CD4⁺ T-cell-derived J-Lat cells (full length clone 9.2),¹⁸ was obtained from Dr. Eric Verdin through the NIH AIDS Research and Reference Reagent Program. J-Lat cells (clone 9.2) were cultured for 24 hours in the presence of 0.1% DMSO with or without AR-42 or vorinostat. Treatment with tumor necrosis factor alpha (TNF-α) (10 ng/mL) served as a positive control.¹⁸ Following the treatment, the cells were washed, fixed in 4% paraformaldehyde, and quantified by flow cytometry using Guava EasyCyte Mini (EMD Millipore). HIV-1 reactivation [green fluorescent protein (GFP) expression] was determined using the FlowJo software (Tree Star) with the gate equivalent to 0.1% DMSO-treated control cells. Additionally, the PRISM software was used to determine the half maximal effective concentration (EC₅₀) for AR-42 and vorinostat. Flow cytometry analysis determined that in the J-Lat (clone 9.2) cell model, AR-42 is 2.4-fold more potent at HIV-1 reaction than vorinostat (EC₅₀ values of 3200 ± 100 nM and 7800 ± 100 nM, respectively; Fig. 2B). Together, the ACH-2 and J-Lat (clone 9.2) data demonstrate that AR-42 can be more potent and efficacious than vorinostat in these HIV-1 reactivation cell line models.

To determine the effect of treatments on cell viability, AR-42-treated cells were assayed using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)/3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assay. The effects of AR-42 and vorinostat were tested for 48 hours and 24 hours, respectively, in ACH-2 and J-Lat (clone 9.2) cells. In ACH-2 cells, both vorinostat and AR-42 caused approximately similar reduction in MTT/MTS activity at 5 μM; although at lower treatment concentrations, vorinostat did not lower MTT/MTS activity >0.1% DMSO after 48 hours (Fig. 3A). In the J-Lat cells (clone 9.2), after 24 hours of treatment, the half cytotoxicity concentration (CC₅₀) of AR-42 was 300 ± 100 nM, while that of vorinostat was 1300 ± 100 nM (Fig. 3B).

Figure 3.

AR-42 reduces the viability of latently infected CD4⁺ T-cells. (A) ACH-2 latently infected cells (48 hours). (B) J-Lat (clone 9.2) latently infected cells (24 hours). MTT or MTS cell viability assays were tested using vorinostat (SAHA) as a positive control. DMSO (0.1%) was used as a vehicle control. (C) Early apoptosis and necrosis (annexin V and propidium iodide staining) were tested in ACH-2 latently infected cells (black dotted) treated with 0.1% DMSO ± AR-42 for 48 hours.

In addition to MTT/MTS cell viability analysis, early apoptosis and necrosis studies were performed on AR-42-treated ACH-2 cells using annexin V and propidium iodide staining. Flow cytometry parameters for annexin V and propidium iodide were set based on heat-killed cells (incubated at 50°C for one hour) and performed using Beckman Coulter Cytomics FC500. Similar to the MTT/MTS results, AR-42 reduced the cell viability of ACH-2 cells at the CC₅₀ of 217 ± 1 nM (Fig. 3C). These data suggest that AR-42 is more toxic than vorinostat in these two HIV-infected cell lines.

This study was designed to assess the ability of a novel HDAC inhibitor (AR-42) to reactivate HIV-1. We observed the following: AR-42 more potently induces histone 3 acetylation than vorinostat, AR-42 is more efficacious and equipotent than vorinostat in its ability to induce HIV-1 gene expression, and AR-42 is more toxic than vorinostat in two CD4⁺ T-cell line models of HIV-1 latency.

In the cellular models of schwannoma and meningioma, AR-42 inhibited cellular growth (IC₅₀ values between 250 nM and 1 μM, depending on the cell line).²⁰ In several models of non-Hodgkin’s lymphoma, AR-42 significantly enhanced the anti-tumor activity of HB22.7, an anti-CD22 monoclonal biologic.²¹ AR-42 is currently in two clinical trials: one for the treatment of non-Hodgkin’s lymphoma (NCT01798901) and the other for multiple myeloma (NCT01129193, www.clinicaltrials.gov). In the multiple myeloma phase I trial, a 40-mg dose of AR-42 achieved a maximum concentration (C_max) of 1 μM, a concentration that is sufficient to reactivate HIV in the ACH-2 model.^22,23 In the MT-2 and C8166 cellular models of cancers associated with the deltaretrovirus human T-lymphotropic virus type 1 (HTLV-1), AR-42 induces both histone acetylation and apoptosis; this study did not assess the ability of AR-42 to reactivate HTLV-1 gene expression.¹¹ Furthermore, in a mouse model of HTVL-1-associated adult T-cell leukemia/lymphoma, AR-42 significantly increased animal survival compared to vehicle-treated control animals.¹¹ Thus, AR-42 has promising activity against the cancers of various etiologies.

AR-42 treatment decreased MTT activity and cell viability at the treatment concentrations of 250 nM–1000 nM, although the cellular damage would not be attributed solely to drug treatment, because AR-42-induced HIV-1 release can also result in cell death. Previous studies have indicated that activated latently infected cells are presumed to die due to viral pathogenic effects, apoptosis, or pyroptosis.^4,24 A strength of this study is that rather than assessing the supernatant-associated HIV RNA concentration following the reactivation, we assessed either intracellular GFP production (J-Lat cells clone 9.2) or RT activity deposited into the supernatant (ACH-2); both of these assays would not be confounded by HIV RNA or DNA, which could be liberated by cell death.

HIV-1 latently infected cell line models, as used in this study, have proven to be useful in investigating the induced reactivation of HIV from latently infected cells.²⁵ Recognizing that individual HIV-1 latently infected cell models have limitations, we tested the ability of AR-42 to reactivate the HIV-1 gene expression in both the J-Lat cells (clone 9.2) and the ACH-2 models. Although there are slight differences between the results from the two cell lines, compared to vorinostat, AR-42 had at least one favorable pharmacological attribute in each model [ie, better efficacy in ACH-2 and better potency in J-Lat cells (clone 9.2)].

In summary, AR-42 potently induces histone acetylation in the ACH-2 cells and HIV-1 gene expression in the two models of latently infected CD4⁺ T-cells. These results (ie, favorable efficacy and toxicity profiles), combined with the ongoing AR-42 clinical studies, suggest that AR-42 should be tested in the primary cell models of HIV-1 latency.²⁶