HIV-1 eradication, which entails the full elimination of the viral reservoir is a tall order and has not been achieved so far, thus prompting additional efforts geared towards a functional cure. A functional cure is defined as the permanent suppression of HIV-1 transcription, with undetectable virus replication, normal CD4
+ T cell counts, no disease progression and no viral transmission in the absence of treatment, despite the presence of integrated proviruses. With this approach, integrated HIV DNA is not fully eradicated but viral transcription is absent or low enough that any occurring viral production can be cleared by the immune system [
108,
109]. Affordability, long-lasting effects, and applicability in resource-poor settings, are features envisaged in functional cure approaches when compared to an ART regimen. As mentioned before, HIV-1 transcription is a complex process that involves multiple factors; thus, several targets can be explored for therapeutic purposes.
5.1. Tat
Tat is a small 14 kDa protein that potently activates HIV-1 gene expression. Tat protein is considered a desirable target for drug development because it is expressed early on during infection [
46,
55], it is highly conserved among HIV-1 isolates [
112], it has no cellular homolog, and, finally, its direct inhibition blocks the feedback loop that drives exponential viral production [
113]. Although several compounds have been previously tested against Tat or its interactions (TAR-Tat-P-TEFb), up until recently, these compounds showed low specificity or poor pharmacokinetics, and none has reached the clinic [
114].
Our group first described didehydro-Cortistatin A (dCA) in 2012, an equipotent analog of the natural product Cortistatin A (CA), as a specific and potent Tat inhibitor. dCA binds to the TAR-binding domain of Tat, the basic domain, blocking transcriptional elongation of the HIV-1 promoter, and consequently leading to the epigenetic silencing of the HIV-1 promoter [
115]. In acutely and chronically infected cells, as well as primary CD4
+ T cells, we demonstrated transcriptional inhibition of HIV-1 to undetectable or very low levels of viral RNA, using subnanomolar concentrations of dCA, without cell-associated toxicity [
115]. Of note, dCA inhibits both HIV-1, HIV-2, and SIV Tat-mediated transcription from the viral promoter [
116], given the conservation of these protein’s basic region. Importantly, dCA does not interact with the cellular protein HEXIM that has a similar but structurally different basic region [
117]. Using primary CD4
+ T cells isolated from ART-suppressed individuals, we showed that long-term dCA treatment strongly suppresses viral transcription, eventually driving a state of persistent latency that is refractory to reactivation with LRAs. Importantly, discontinuation of dCA and ART treatment maintains viral suppression, suggesting long-lasting effects [
118]. In agreement, our in vivo studies using the humanized bone marrow-liver-thymus (BLT) mouse model for HIV-1 latency, revealed that dCA treatment decreases viral RNA in tissues and delays viral rebound upon ART interruption [
111]. Long-term dCA treatment does not alter the classic nucleosome position at the LTR promoter. Instead, a tighter nucleosome/DNA association is observed, which correlates with increased deacetylated histone 3 (H3) occupancy at Nuc-1. This repressive state of chromatin structure at Nuc-1 prevents RNAPII recruitment and elongation, even with stimulation with LRAs [
119]. Low levels of PBAF complex is observed at the HIV promoter with dCA treatment, which is crucial for Nuc-1 remodeling and Tat-activated transcription, while the BAF complex is increased. Our results suggest that dCA accelerates the establishment of latency and, importantly, that dCA activity is directly correlated with Tat-TAR competent proviruses with no apparent off-target effects [
119]. Importantly, our group reported that when intraperitoneally injected in mice, dCA crosses the blood–brain barrier (BBB) and is detected at high levels in the brain, where microglial cells are proposed to serve as an HIV reservoir. Moreover, this study demonstrated that dCA inhibits different HIV-1 clades and decrease extracellular Tat uptake by glial cell lines, which may reduce HIV-1-related neuropathogenesis [
120]. Tat has been shown to potentiate cocaine-mediated reward mechanisms through disfunction of the dopaminergic system [
121]. Using Tat transgenic mice, we also demonstrated the ability of dCA to reduce Tat potentiation of cocaine-mediated reward mechanisms [
120].
In sum, our studies using dCA as a latency promoting agent (LPA) highlight the potential of Tat inhibitors for use in block-and-lock approaches. Adding Tat inhibitors to an ART regiment may further reduce cell-to-cell transmission, viral reactivation, and spontaneous blips, as well as drive the virus into a state of deep latency. The expectation is that long lasting transcriptional suppression may lead to prolonged or permanent epigenetic silencing of the provirus, allowing safe interruption of therapy. These studies are currently underway.
Other inhibitors of Tat-mediated transcription have been explored, namely, Triptolide (TPL), a diterpenoid epoxide isolated from
Tripterygium wilfordii Hook F (TwHF), a natural product used for the treatment of rheumatoid arthritis [
122]. TPL has shown anti-HIV-1 activity by blocking Tat function at nanomolar concentrations [
123]. TPL enhances the proteasomal degradation of Tat and, consequently, the in vitro suppression of viral transcription [
123]. However, the clinical potential of TPL has been debated. Wang et al. discovered that TPL induces proteasome-dependent degradation of RNAPII, inhibiting global gene transcription [
124]. Another study demonstrated that TPL covalently binds to xeroderma pigmentosum group B (XPB), a subunit of the transcription factor TFIIH, and inhibits its ATPase activity, which blocks RNAPII-mediated transcription initiation [
125]. Thus, due to its global inhibition of transcription, probably by interfering with important cellular functions, its clinical application may be limited by safety concerns.
Recently, a screen of an FDA-approved compound library identified Levosimendan as a potential LPA [
126]. Levosimendan normally is used for the treatment of acutely decompensated heart failure [
127]. Hayashi et al. discovered that Levosimendan blocks HIV-1 Tat-LTR mediated transcription. Using a PI3K inhibitor, 3-MA, they were able to overcome the inhibitory effect of levosimendan in a dose-dependent manner, suggesting that this compound is possibly involved in the Akt/PI3K pathway to inhibit HIV-1 transcription [
126]. However, the specific mechanism by which this compound mediates the inhibition of HIV-1 transcription and reactivation is still under investigation and needs to be further elucidated. Additionally, they showed that it also suppresses HIV-1 reactivation from latency, using several HIV-1 latency cell lines, primary CD4
+ T cell models of HIV-1 latency, and primary CD4
+ T cells isolated from HIV-1-infected individuals on ART. On the same screen, Hayashi et al. also found Spironolactone (SP) as another anti-HIV-1 agent [
126]. SP can promote the degradation of the XPB subunit of TFIIH [
128] and it was shown to inhibit acute HIV-1 infection of cell lines and primary CD4
+ T cells by blocking HIV-1 transcription [
129]. Recently, our group demonstrated that long-term SP treatment rapidly reduces ongoing transcription in latently infected cell line models in what appears to be a Tat-TAR independent mechanism and was associated with a reduction in RNAPII recruitment to the HIV-1 genome (Mori, L. et al., in press). SP treatment potently reduced HIV-1 reactivation with exposure to a range of LRAs and, importantly, blocked HIV-1 reactivation in ex vivo stimulated primary CD4
+ T cells. Unfortunately, SP inhibition of HIV-1 transcription was not long lived, with viral rebound occurring rapidly upon treatment interruption and XPB replenishment. It will be important to assess the effects of SP alongside other longer-lasting LPA, such as dCA, for example. Since both SP and Levosimendan are FDA-approved compounds, it may accelerate their investigation in humans.
5.3. mTOR Complex (mTORC)
mTOR, a serine/threonine kinase that forms two complexes, mTOR1 and mTOR2, is involved in a variety of cellular processes, such as the regulation of glucose metabolism, cell growth, energy balance and viability [
133,
134]. Rapamycin is a bacterial macrolide used for the treatment of renal transplantation rejection and is an allosteric inhibitor of mTOR kinase, that selectively inhibits mTORC1. Heredia et al. demonstrated that rapamycin can inhibit HIV-1 replication in vitro by down-regulation of CCR5 expression, the HIV-1 major co-receptor [
135]. Later, the same team demonstrated that INK128, an inhibitor of both mTORC1 and mTORC2, successfully suppressed HIV-1 replication in vivo [
136]. INK128 inhibited both basal and induced (by PMA) HIV-1 transcription [
136]. Inhibition of HIV-1 transcription is consistent with inhibition of mTORC2, which is essential for phosphorylation of PKC isoforms [
137], required for NF-κB induction of HIV-1 transcription [
138]. Later, in 2017, Besnard et al. supported the idea that the inhibition of mTORC1 and mTORC2 can suppress HIV-1 reactivation. Through a human genome-wide shRNA screen, the authors uncover the mTOR pathway as a modulator of HIV-1 latency. Results showed that mTORC1 and mTORC2 inhibition strongly suppress reactivation of HIV-1 in a CD4
+ T cell line model and in HIV-1-infected patient cells. MLST8 knockdown, a subunit shared by mTORC1 and mTORC2, suppressed HIV-1 reactivation under PMA stimulation, but not upon BET and HDAC inhibitor treatment. This supports the role that PKC-dependent NF-κB activation seems to be an important target of mTOR and its relation to HIV-1 latency [
139]. In another study, Ji Shan et al. adapted a genome-wide CRISPR screening approach in a T-cell-based latency model and discovered that TSC1 and DEPDC5, two mTORC1 natural inhibitory genes, are potentially involved in HIV-1 latency. Their results suggested that TSC1 and DEPDC5 suppress the AKT-mTORC1 pathway activity and hamper the initiation of HIV-1 transcriptional translation to maintain latency [
140]. In 2019, two clinical trials, NCT02990312 and NCT02440789, were initiated to evaluate the impact of rapamycin on HIV-1 persistence and immune activation/inflammation.
Targeting cellular proteins could be an attractive approach to overcome HIV-1 drug resistance. However, the mTOR pathway controls key cellular processes and inhibitors may come with some off-target activity. The feedback from clinical trials will determine their future clinical use.
5.5. Heat Shock Protein 90 (HSP90)
HSP90 is a heat-shock chaperone that localizes at the HIV-1 promoter and regulates its expression [
142]. Joshi et al. demonstrated that the cytosolic HSP90 isoform is an essential HIV-1 host factor, and its inhibition by RNA interference blocks HIV-1 replication in primary human T cells [
143]. HSP90 was shown to be involved in HIV-1 reactivation through modulation of the NF-κB signaling pathway and stimulation of Tat-mediated HIV-1 transcription [
144,
145].
An in vivo study using Hsp90 inhibitors (AUY922 or 17-AAG) reported that viral rebound was blocked up to 11 weeks after treatment interruption in mice pretreated with a reverse transcriptase inhibitor (EFdA) plus AUY922 or 17-AAG [
146]. Infectious viruses were successfully recovered by heat shock or cell activation from PBMCs and spleen, suggesting that these cells were latently infected [
146].
GV1001, an MHC class II-restricted peptide vaccine designed to induce T-cell immunity to telomerase, can reduce the levels of HSP90 inside the cell and on the cell surface [
147]. It was previously shown that GV1001 suppressed the replication of the Hepatitis C virus (HCV) through the interaction with HSP90 [
148]. Kim et al., likewise, reported that GV1001 can significantly suppress HIV-1 replication, as well as viral reactivation from latently infected cells. Of note, GV1001 suppressed NF-κB activation and, consequently, downregulated Tat-dependent transcriptional activity and HIV-1 RNA production [
144]. These results suggested that HIV-1 suppression by GV1001 is HSP90-dependent, since treatment with an anti-HSP antibody resulted in the loss of compound activity.
It seems clear that HSP90 is implicated in NF-κB activity [
145]. Investigating the underlying molecular mechanism of NF-κB suppression is needed. Together, these data suggest that HSP90 inhibitors can also be studied as potential LPAs to achieve a functional HIV-1 cure.