Therapy for Latent HIV-1 Infection: The Role of Histone Deacetylase Inhibitors

Abstract

Persistence of HIV-1 in latently infected CD4⁺ T-cells prevents eradication in HIV-infected treated patients. Latency is characterized by a reversible silencing of transcription of integrated HIV-1. Several molecular mechanisms have been described which contribute to latency, including the establishment and maintenance of repressive chromatin on the HIV-1 promoter. Histone deacetylation is a landmark modification associated with transcriptional repression of the HIV-1 promoter and inhibition of histone deacetylase enzymes (HDACs) reactivates latent HIV-1. Here, we review the different HDAC inhibitors that have been studied in HIV-1 latency and their therapeutic potential in reactivating latent HIV-1.

Introduction

Current antiretroviral therapy potently suppresses HIV-1 replication, however proviral infection still persists in a small pool of latently infected cells. This reservoir, primarily composed of resting CD4⁺ T-cells, must be targeted in any attempt to eradicate HIV-1 infection. Mechanisms leading to HIV-1 latency in CD4⁺ T-cells are not completely understood, but the chromatin environment is critical for the establishment and maintenance of latency [1–4]. While there are many different factors that contribute to HIV-1 latency and a repressive chromatin environment [5–10], one that is of great current interest due to the potential for the therapeutic intervention is the recruitment of histone deacetylases (HDACs) to the 5′ long terminal repeat (LTR) of HIV-1 [11–16]. There are four classes of HDACs – class I, II, III and IV – but primarily class I HDACs, which includes HDACs 1, 2 and 3, are recruited to the HIV-1 LTR. Class I HDACs are recruited to the LTR by host transcription factors, including yin yang 1 and the late SV40 factor, nuclear factor κB p50 homodimers, Sp1, c-myc, activator protein 4, and C-promoter binding factor-1 [11,13–18]. Many studies have emerged over the past decade analysing the usefulness of HDAC inhibitors (HDACis) in reactivating latent virus in cultured cell models of latency, ex vivo in resting CD4⁺ T-cells from patients, and most recently in vivo in patients with antiretroviral (ART)-suppressed viraemia [1,19–24]. Here, we will focus on recent studies of HDACis as therapeutic agents to disrupt latent HIV-1 infection.

Global HDAC inhibition

In an early attempt to use HDAC inhibition as a tool to disrupt latent infection valproic acid (VPA), or Depokote ER®, was added to the therapy of a small cohort of patients in the context of ART intensified by enfuvirtide. Due to wide use of VPA for other medical indications in HIV-infected patients on ART, the lack of significant drug–drug interactions of this weak HDAC inhibitor was well established. Following the demonstration that VPA induced the expression of latent infection ex vivo in resting CD4⁺ T-cells from aviraemic patients [20,22], studies investigating the effect of VPA on reducing the latent HIV-1 reservoir in vivo were attempted. An initial study found depletion of latent infection in all patients, which was significant in three out of four patients, observed over 3 months of study duration [21]. However, two follow-up observational studies in patients on long-term VPA and standard ART found no consistent effect, in one study using PCR-based measures of viral DNA [25], and in another measuring replication-competent virus with a detection limit of 0.5 infected cells per million [26]. Another follow-up study also did not observe a durable or cumulative depletion of latent infection in patients prospectively administered VPA in the context of standard ART or ART intensified by T-20 or raltegravir using sensitive viral outgrowth assays [22,27]. Finally, a recent randomized, multicentre study of the addition of VPA to HAART in virally suppressed patients reported no effect on the frequency of replication-competent virus within CD4⁺ T-cells [28].

One explanation for these disappointing results is that VPA may not function as an effective HDAC inhibitor in vivo at the concentrations achieved in patients under standard regimen. However, ex vivo modelling studies had suggested that the doses employed would have induced proviral expression [20]. In addition, VPA may not have induced sufficiently robust proviral expression to result in the death or clearance of reactivated cells, and/or host antiviral immune responses in these patients might have been insufficient to allow clearance of infection following induction.

Class-specific HDAC inhibition

Following this setback, more selective and class-specific HDAC inhibitors were tested for their ability to induce both LTR expression in cell models of latency and HIV-1 production in resting CD4⁺ T-cells from aviraemic patients on ART [23,24,29–33]. A study using class-specific HDACis showed that class I, but not class II, inhibitors induce chromatin changes at the HIV-1 LTR in model cells, and that class I selective inhibitors effectively stimulate LTR expression [24]. It further demonstrated that selective inhibition of HDACs 1, 2 and 3 resulted in a more efficient activation of the HIV-1 LTR and that inhibition of HDACs 1 and 2 alone failed to activate the LTR and induce viral outgrowth in resting CD4⁺ T-cells from patients. The importance of HDAC 1, 2 and 3 inhibition in activating latent HIV-1 was further supported by Huber et al. [34]. It is likely that HDACs 1, 2 and 3 function as part of several multiprotein complexes that are independently recruited to the HIV promoter and mediate its transcriptional repression.

Recent studies have further explored the effect of class-I-specific HDAC inhibitors on the disruption of HIV-1 latency. Suberoylanilide hydroxamic acid (SAHA), or vorinostat (Zolinza™), a class-I-selective HDAC inhibitor of higher potency and greater selectivity than VPA, was tested for its ability to induce latent HIV-1 expression. SAHA induces HIV promoter activity in cell models of latency and also induces HIV-1 production from resting CD4⁺ T-cells from patients suppressed on ART [23,35]. Similarly, oxamflatin [(2E)-5-[3-[(phenylsufonyl) aminol phenyl]-pent-2-en-4-ynohydroxamic acid], a newly identified hydroxamate-based agent with clinical applications in certain malignancies, effectively increases expression of HIV-1 in latency model cells [32]. ITF2357, another hydroxamic acid-type HDACi with anti-inflammatory and anti-tumour properties in vitro and in vivo [36–38], induces HIV-1 in latently infected model cells [31].

A number of newly synthesized and novel class-I-selective HDAC inhibitors, metacept-1 and metacept-3 [29], CG05 and CG06 [30], and NCH-51 [33] have been reported. Each of the inhibitors is similar in structure to the hydroxamic acid agents oxamflatin and SAHA, and induce reactivation of HIV-1 in cell models of latency [29,30,33]. Each compound showed increased HIV promoter activity and higher potency to class I HDACs compared to VPA but mixed results, collectively, when compared to SAHA. Studies of the potential for these newer HDAC inhibitors, oxamflatin, or ITF2357, to induce viral outgrowth from the resting CD4⁺ T-cells of aviraemic patients or disrupt the viral reservoir in vivo, have not yet been performed.

So far, SAHA is the only class-I-selective HDAC inhibitor shown to induce viral outgrowth in ex vivo studies in resting CD4⁺ T-cells from suppressed patients on ART [39]. This provided rationale for the study of SAHA in vivo and set it apart as a promising therapeutic agent. Significantly, administration of a single dose of SAHA, or vorinostat, to HIV-infected patients on ART with clinically undetectable levels of viraemia, was sufficient to upregulate HIV RNA within resting CD4⁺ T-cells [39]. This study was carefully designed to minimize the risk to patients and considered the ethics of exposing a patient to an experimental agent with potential risk and no clinical benefit. Patients were screened by a 6 h in vitro exposure to vorinostat and eight of the eleven patients in whom HIV RNA could be measured received a single 200 mg dose of vorinostat to determine tolerability. This was followed by a 400 mg dose of vorinostat two to four weeks later to analyse safety and tolerability and also validate the pharmacokinetics of the drug exposure, biomarker effect and clearance. Four to five weeks later, a second 400 mg dose of vorinostat was administered and resting CD4⁺ T-cells collected during a window of 4–7 h after dosing. A mean 4.8-fold increase of unspliced HIV-1 gag RNA expression was observed and was significantly increased compared to pre-exposure expression over baseline in all eight patients. Importantly, the limited exposure to vorinostat was well tolerated by all the patients and no adverse events attributable to vorinostat were reported. This is the first study to demonstrate proof-of-concept that a molecular mechanism regulating HIV-1 latency can be therapeutically targeted in vivo, validating the concept of the use of class-I-selective HDAC inhibitors to disrupt HIV-1 latency in humans. This provides a stepping-stone into further studies that target factors that establish and maintain HIV-1 latency.

Future directions

One of the most important considerations for future studies using HDAC inhibitors already approved for other therapeutic targets is the testing of clinically achievable drug concentrations and exposures. It is of concern that use of higher concentrations to achieve increased HIV-1 reactivation may lead to off-target effects [34]. For example, at lower, clinically achievable concentrations, SAHA is a class-I-specific inhibitor, but at higher concentrations, SAHA loses its specificity, targets other classes of HDACs, and induces cell differentiation [34,40–42]. Additionally, SAHA can induce toxicity and apoptosis at higher concentrations (micromolar-range) [43–45].

As the field advances, it is important to consider whether a single agent like SAHA or a combination of agents is sufficient to activate HIV-1 in a clinically significant way given the heterogeneity of the HIV-1 reservoir. The vorinostat study described above was the first of its kind to demonstrate proof-of-concept that latent virus can be induced in humans, where the HIV-1 reservoir is most heterogeneous [39]. Recent studies have analysed synergistic, additive, or cooperative effects between class I HDAC inhibitors and other drugs such as prostratin, 5-Aza-2′-deoxycitydine (5-Azac), or tumour necrosis factor-α, in reactivation of latent HIV-1 [32,46,47]. Since several distinct mechanisms contribute to the establishment and maintenance of latency, synergies between drugs targeting different mechanisms may provide a way to minimize the toxicity of individual drugs. While the recruitment of HDACs to the HIV-1 promoter certainly contributes to a repressive chromatin environment, it is possible that other cellular factors must be targeted in order to adequately disrupt latency. Two recently published reviews by Hakre et al. [48] and Barton et al. [49] give a comprehensive summary of other areas of potential therapeutic interest. Additionally, high-throughput screening techniques have recently been used to identify new targets that cooperate with HDAC inhibitors to activate latent HIV-1 [50]. New techniques such as high-throughput and virtual screening are especially valuable considering the vast library of small molecules available. Finally, new systems to analyse chromatin changes in HIV-1 latency such as the viral outgrowth assay described by Ylisastigui et al. [20], latently infected primary cells developed by a number of groups [51–53], and in vivo mouse and human studies are of vital importance at this stage of developing a pipeline of HDAC inhibitors as adjuvant therapy to target the HIV-1 latent reservoir.