A Novel Bromodomain Inhibitor Reverses HIV-1 Latency through Specific Binding with BRD4 to Promote Tat and P-TEFb Association

Abstract

While combinatory antiretroviral therapy (cART) can effectively reduce HIV-1 viremia, it cannot eliminate HIV-1 infection. In the presence of cART, viral reservoirs remain latent, impeding the cure of HIV-1/AIDS. Recently, latency-reversing agents (LRAs) have been developed with the intent of purging latent HIV-1, providing an intriguing strategy for the eradication of the residual viral reservoirs. Our earlier studies show that the first-generation, methyl-triazolo bromodomain, and extra-terminal domain inhibitor (BETi), JQ1, facilitates the reversal of HIV-1 latency. BETis have emerged as a new class of compounds that are promising for this HIV-1 "shock and kill" eradication approach. However, when used as a single drug, JQ1 only modestly reverses HIV-1 latency, which complicates studying the underlining mechanisms. Meanwhile, it has been widely discussed that the induction of latent proviruses is stochastic (Ho et al., 2013). Thus, new BETis are currently under active development with focus on improving potency, ease of synthesis and structural diversity. Using fluorous-tagged multicomponent reactions, we developed a novel second-generation, 3,5-dimethylisoxazole BETi based on an imidazo[1,2-a] pyrazine scaffold, UMB-32. Furthermore, we screened 37 UMB-32 derivatives and identified that one, UMB-136, reactivates HIV-1 in multiple cell models of HIV-1 latency with better efficiency than either JQ1 or UMB-32. UMB-136 enhances HIV-1 transcription and increases viral production through the release of P-TEFb. Importantly, UMB-136 enhances the latency-reversing effects of PKC agonists (prostratin, bryostatin-1) in CD8-depleted PBMCs containing latent viral reservoirs. Our results illustrate that structurally improved BETis, such as UMB-136, may be useful as promising LRAs for HIV-1 eradication.

Keywords: BETi; BRD4; HIV-1; JQ1; LRA; UMB-136; latency; reactivation.

Figure 1

UMB-136 is a leading UMB-32 based BETi. (B) J-Lat A2 cells were treated with UMB-32, 58, 59, or 136 (5 μM), or JQ1 (1 μM). GFP-positive cells were acquired and analyzed by flow cytometry. The HIV-reactivated cell population was indicated. (A) Chemical structures of 37 UMB-32 analogs. The structures of UMB-32, 58, 59, 136, and JQ1 were highlighted in red frames.

Figure 2

UMB-136 promotes HIV-1replication by releasing P-TEFb through binding to BRD4. (A) The predicted binding of UMB-58, 59, and 136 with the BRD4 bromodomain (BD1) according to the available data for UMB-32. (B) The superimposition of UMB-136 (violet) onto UMB-32 (green) in complex of BRD4 BD1 (PDB ID: 4WIV; protein: residues in contact with ligand shown in yellow, the rest in gray); hydrogen bonding with crystal waters (red spheres) in dashed red lines, pi-stacking in black dashed line. Graphic representations of (A,B) are generated in PyMol Software. (C) Calculation of the binding energy (PBTOT—relative total binding energy using MM-PBSA; GBTOT—relative total binding energy using MM-GBSA) for UMB-32, 58, 59, 136, and JQ1. (D) Pull-down assay for UMB-136. The lysate of HEK293 cells was incubated with biotinylated UMB-136 or biotin in DMSO at 4°C for overnight with orbital rocking. MyOne™ Streptavidin T1 Dynabeads™ (Invitrogen) were added to the mixture for incubation for 2 h. Beads were washed and boiled to elute the precipitated proteins and subjected to immunoblotting using BRD4 or BRD2 antibody. (E) UMB-136 promotes the association of Tat and P-TEFb. Hela cells stably expressing Tat-Flag were treated with UMB-136 (2.5 μM) or DMSO for 24 h, followed by the incubation with Magnetic Flag Conjugated beads (Clontch) at 4°C for overnight. Beads were washed and boiled to elute the precipitated proteins, which were subjected to immunoblotting using Cyclin T1 antibody. (D,E) were repeated twice individually. (F) Jurkat cells were infected with VSV-G pseudo-typed HIV-1 NL4-3 virus, and treated with UMB-32 (2.5 μM), UMB-136 (2.5 μM), JQ1 (1 μM), prostratin (1 μM), or DMSO. The level of viral transcripts at different length was determined by RT-qPCR using the indicated primers. (G) Jurkat cells were infected with full length NL4-3 (X4 tropic), and treated with UMB-136 (2.5 μM), JQ1 (1 μM), or prostratin (1 μM). The newly produced viruses in the supernatant were measured by qPCR. Results were presented as mean ± s.e.m. of two or three biological replicates ^*p < 0.05, t-test.

Figure 3

UMB-136 reverses HIV-1 latency from multiple cell models of HIV-1 latency. (A) Multiple J-Lat cell clones (6.3, 8.4, 9.2, and 10.4, CD4+ T cell background) were treated with UMB-136 (2.5 or 5 μM) or JQ1 (1 μM). GFP-positive cells were acquired and analyzed by flow cytometry. Percentage of GFP-positive cells was determined. One representative of three independent experiments was shown. (B) THP89GFP cells (monocyte/macrophage background) were treated with UMB-136 (5 μM) or JQ1 (1 μM). GFP-positive cells were acquired and analyzed by flow cytometry. The HIV-reactivated cell population was indicated. (C) HIV-1 latency was established in the CD4+ T cells isolated from the peripheral blood mononuclear cells (PBMCs) of 2 donors, following a previously reported primary cell model (Vicente Planelles). Cells latently infected with wild-type HIV-1 NL4-3 viruses were treated with UMB-136 (2.5 μM) or JQ1 (1 μM). The viral mRNAs from newly produced HIV-1 viruses in supernatant were measured by qRT-PCR assay. (D) The same assay in (D) was conducted using the CD4+ T cells isolated from the tonsillar mononuclear cells (TMCs) of 2 donors. Results were presented as mean ± s.e.m. (n = 3); ^*p < 0.05, t-test.

Figure 4

UMB-136 synergizes with other LRAs to reverse HIV-1 latency in J-Lat cell lines. (A) Multiple J-Lat cell clones (6.3, 8.4, 9.2, and 10.4) were treated with UMB-136 (2.5 μM) or JQ1 (1 μM) in combination with an HDACi (SAHA, 0.5 μM) or a PKC agonist (prostratin, 1 μM). GFP-positive cells were acquired and analyzed by flow cytometry. Percentage of GFP-positive cells was measured. One representative of three independent experiments was shown. (B) The synergistic analysis of LRA combinations. The synergy of LRA combinations in Figure 3A was evaluated using the Bliss independence model (Bliss, 1939). Data are represented as the mean ± SD of biological duplicates. ^*p < 0.05, t-test.

Figure 5

Combined treatment with UMB-136 and Bryostatin-1 reverses HIV-1 latency in patient-derived CD8-depleted PBMCs. (A,B) CD8-depleted PBMCs isolated from seven cART-treated, HIV-positive aviremic patients were stimulated with UMB-136 (2.5 μM), or JQ1 (1 μM) in combination with PKC agonists (prostratin, 250 nM; bryostatin-1, 10 nM) to reactivate latent HIV-1 from reservoir cells. The viral transcripts from newly produced HIV-1 viruses in the supernatant were measured by a quantitative nested RT-qPCR assay. C+: antibodies against CD3 and CD28. ^*P < 0.05, wilcoxon signed rank test.