Inhibition of ALKBH5 demethylase of m6A pathway potentiates HIV-1 reactivation from latency

Abstract

Background: Current latency-reversing agents (LRAs) employed in the "shock-and-kill" strategy primarily focus on relieving epigenetic and transcriptional blocks to reactivate the latent HIV-1. However, their clinical efficacy is limited, partly due to their inability to fully reverse latency and the lack of LRAs specifically targeting post-transcriptional mechanisms. N6-methyladenosine (m⁶A) modification in HIV-1 RNA is emerging as an important post-transcriptional regulator of HIV-1 gene expression, yet its role in latency and reactivation remains largely unrecognized. Here, we explored the potential of small chemical compounds targeting the m⁶A pathway, specifically investigating the inhibition of ALKBH5 and its effect on latent HIV-1 reactivation mediated by the LRA romidepsin.

Methods: We used four in vitro cellular models of latency, primary model of CD4⁺ T cells HIV-1 infection and ex vivo cultures of CD8⁺-depleted PMBCs from ART-treated HIV⁺ patients. We measured latent viral reactivation by evaluating the expression of reporter protein GFP by flow cytometry, viral production by CA-p24 ELISA, and viral transcripts by RT-qPCR. CRISPR/Cas9 method was used to deplete ALKBH5. MeRIP and immuno-RNA FISH were used to address the m⁶A methylation levels on HIV-1 RNA upon ALKBH5 inhibition.

Results: We showed that ALKBH5 inhibitor 3 (ALKi-3) potentiated romidepsin-mediated viral reactivation in in vitro models of latency, primary model of CD4⁺ T cells infected with HIV-1 as well as in ex vivo cultures of CD8⁺-depleted PBMCs from ART-treated HIV⁺ patients. CRISPR/Cas9-mediated depletion of ALKBH5 mimicked the effects of ALKi-3. ALKi-3 increased levels of m⁶A-methylated HIV-1 RNA as shown by meRIP and immuno-RNA FISH.

Conclusion: Our study provides a proof-of-concept for the modulation of the m⁶A pathway in enhancing HIV-1 reactivation. This approach represents a promising adjunct to existing reactivation protocols and provides a concept of "dual-kick", aiming to target transcriptional and post-transcriptional steps in HIV-1 reactivation from latency.

Keywords: ALKBH5; Dual-kick; Epitranscriptomics; HIV-1; Latency; Post-transcriptional mechanisms; Shock-and-kill; m6A.

Fig. 1

ALKBH5 inhibitor 3 (ALKi-3) potentiates viral reactivation mediated by a sub-optimal dose of romidepsin. (A). Chemical structures of the METTL3-14-WTAP activators, FTO inhibitors, and ALKBH5 inhibitors. (B-C) J-Lat A2 cells were either DMSO-treated, treated with 100 µM of indicated compounds alone or in combination with romidepsin [5 nM]. At 24 h post-treatment, viral reactivation was assessed by flow cytometry to quantify the percentages of GFP⁺ cells. Means and standard error of means from three biological replicates in duplicates are indicated. Statistical analysis was performed using a paired Student’s t-test, p ≤ 0.0002 (***)

Fig. 2

Enhanced reactivation potential of romidepsin by ALKi-3 in lymphoid J-Lat 9.2 and promonocytic U1 in vitro latency cellular models. (A-B, F-G) J-Lat 9.2 and U1 cells were either DMSO-treated, treated with increasing doses of ALKi-3 [25-50-100-200 µM] alone or in combination with sub-optimal dose of romidepsin (17.5 nM, and 5 nM for J-Lat 9.2 and U1 cells, respectively). At 24 h post-treatment (A, F) viral reactivation was assessed by measuring the concentration of genomic viral RNA copies/ml in culture supernatant using RT-qPCR. (B, G) Cells metabolic activity was measured by using XTT assay. Results obtained with the mock-treated cells were arbitrary set at a value of 100%. (C-E, H-I) Cells were treated with either DMSO (control), ALKi-3 alone (100 µM), or in a combination with suboptimal dose of romidepsin (17.5 nM, and 5 nM for J-Lat 9.2 and U1 cells, respectively). Viral reactivation was assessed 24 h post-treatment. (C) J-Lat 9.2 cells were subjected to flow cytometry analysis to quantify the percentages of GFP⁺ cells. (D, H) Viral production was estimated by measuring CA-p24 antigen in culture supernatant. (E, I) Total RNA was extracted that was subsequently subjected to quantification by RT-qPCR for TAR- and RRE-containing HIV-1 RNAs. Values were normalized using gapdh primers and were presented as relative fold changes to the values measured in romidepsin + DMSO-treated cells which were arbitrarily set at a value of 1. (A-I) Means and standard errors of the means from three biological repetitions in duplicates are represented. Statistical analysis was performed using a paired Student’s t-test with p-values indicating the significance level: p ≤ 0.05 (*), p ≤ 0.002 (**), p ≤ 0.0002 (***), and p ≤ 0.0001 (****)

Fig. 3

Depletion of ALKBH5 potentiates viral reactivation in in vitro latency models. The J-Lat 9.2 (A-D) or U1 (E-G) cells were transduced with lentiviral vector targeting ALKBH5 (sgALKBH5) or with control sgRNA (sgNTC). Five days after puromycin selection, ALKBH5-depleted cells were harvested and subjected to romidepsin treatment for additional 24 h. (A, E) Immunoblotting to detect ALKBH5. GAPDH is the protein loading control. (B) J-Lat 9.2 cells were subjected to flow cytometry analysis to quantify the percentage of GFP⁺ cells. (C, F) Viral production was estimated by measuring CA-p24 antigen in culture supernatant. (D, G) Total RNA was extracted and subsequently subjected to quantification by RT-qPCR for TAR- and RRE-containing HIV-1 RNA. Values were normalized using gapdh primers and were presented as relative fold changes to the values measured in romidepsin treated sgNTC-transduced cells which were arbitrarily set at a value of 1. (B, C, F) Means and standard errors of the means from three biological repetitions in duplicates are represented. Statistical analysis was performed using a paired Student’s t-test with p-values indicating the significance level: p ≤ 0.05 (*), p ≤ 0.002 (**), p ≤ 0.0002 (***)

Fig. 4

ALKi-3 treatment potentiates HIV-1 gene expression in primary CD4⁺ T model of HIV-1 infection. (A) Schematic of the experiment. CD4⁺ T cells were isolated and stimulated with 5 µg/ml of PHA-P and 20 units/ml of IL-2 for three days. At day three, CD4⁺ T cells were infected with HIV-1 NL4-3 ∆Env_EGFP and after 24 h of HIV-1 infection cells were treated with 50 µM of ALKi-3 for further 24 h and subjected to HIV-1 expression analyses. (B) Representative flow cytometry dot plot depicting no infection, and cells infected with HIV-1 NL4-3 ∆Env_EGFP either treated with DMSO or ALKi-3. CD4⁺ T cells were subjected to flow cytometry analysis to quantify the percentage of GFP⁺ cells (C) and their MFI (D). (E) Viral production was estimated by measuring CA-p24 antigen in culture supernatant. (C– E) Means and standard errors of the means from three donors are represented. Statistical analysis was performed using two-way ANOVA with p value indicated above the graph (F) Total RNA was extracted and subsequently subjected to quantification by RT-qPCR for TAR-containing HIV-1 RNA. Values were normalized using gapdh primers and were presented as relative fold changes to the values measured in DMSO-treated cells which were arbitrarily set at a value of 1. Means and standard errors of the means from three donors are represented. Statistical analysis was performed using a paired Student’s t-test with p value indicated above the graph

Fig. 5

Impact of ALKi-3 on the levels of m⁶ A methylated HIV-1 RNA. U1 cells were treated either with romidepsin alone or with romidepsin + ALKi-3 and were collected 24 h post-treatment. (A-B) Total RNA was extracted and subsequently subjected to polyA mRNA enrichment followed by meRIP against m⁶A modified RNA. The m⁶A immunoprecipitated RNA was quantified by RT-qPCR for cellular EEF1A1- and RRE-containing HIV-1 RNAs. Values were normalized using input and IgG control and were presented as m⁶A fold enrichment. Means and standard errors of the means from three biological repetitions are represented. Statistical analysis was performed using a paired Student’s t-test with p-values indicating the significance level: p ≤ 0.05 (*), p ≤ 0.002 (**), p ≤ 0.0002 (***), and p ≤ 0.0001 (****). (C-F) Reactivated U1 cells were subjected to RNA-FISH and immunostaining using antibodies against m⁶A modification for subsequent confocal microscopy analysis. (C) Representative image of m⁶A immuno-HIV RNA FISH. HIV-1 ^gagRNA is shown in red, m⁶A in green, and DAPI-stained nucleus in blue. Yellow spots indicate colocalization sites as marked by white arrows. (D-F) The number of m⁶A (D), HIV-1 ^gagRNA (E) and m⁶A-HIV-1 ^gagRNA (F) spots were quantified. Images were acquired with confocal microscopy and spots were quantified in z-stacks from approx. 20 images/biological repetition, n = 3. Results are presented as box and whiskers with 5–95% confidence interval. Median value is shown as a bar, dots are spots outside the whiskers representing outliers, mean value is shown as “+”.Statistical analysis was performed using a paired Student’s t-test with p-values indicating the significance level: p ≥ 0.12 (not significant– ns), p ≤ 0.05 (*), p ≤ 0.002 (**), p ≤ 0.0002 (***), and p ≤ 0.0001 (****)