Isotretinoin promotes elimination of translation-competent HIV latent reservoirs in CD4T cells

Abstract

Development of novel therapeutic strategies that reactivate latent HIV and sensitize reactivated cells to apoptosis is crucial towards elimination of the latent viral reservoir. Among the clinically relevant latency reversing agents (LRA) under investigation, the γc-cytokine IL-15 and the superagonist N-803 have been shown to reactivate latent HIV ex vivo and in vivo. However, their clinical benefit can be hindered by IL-15 promoting survival of infected cells. We previously identified a small molecule, HODHBt, that sensitizes latently infected cells to death upon reactivation with γc-cytokines through a STAT-dependent pathway. In here, we aimed to identify and evaluate FDA-approved compounds that could also sensitize HIV-infected cells to apoptosis. Using the Connectivity Map (CMap), we identified the retinol derivative 13-cis-retinoic acid (Isotretinoin) causes similar transcriptional changes as HODHBt. Isotretinoin enhances IL-15-mediated latency reversal without inducing proliferation of memory CD4 T cells. Ex vivo analysis of PBMCs from ACTG A5325, where Isotretinoin was administered to ART-suppressed people with HIV, showed that Isotretinoin treatment enhances IL-15-mediated latency reversal. Furthermore, we showed that a combination of IL-15 with Isotretinoin promotes the reduction of translation-competent reservoirs ex vivo. Mechanistically, combination of IL-15 and Isotretinoin increases caspase-3 activation specifically in HIV-infected cells but not uninfected cells. Our results suggest that Isotretinoin can be a novel approach to target and eliminate translation-competent HIV reservoirs.

Fig 1. Identification and confirmation of Isotretinoin as a latency reversal agent.

A: Workflow to identify 5 FDA-approved compounds with transcriptional profiles similar to lead LRA using CMap analysis. Figure created with BioRender. B: CD69 MFI of matched cultured T_CM (n = 9) were measured by flow cytometry after treatment with 30 IU/mL IL-2 or IL-2 plus 100μM HODHBt or 10μM of indicated compounds for 72 hours. Significant p values over cells treated with IL-2 alone. C: Reactivation of latent HIV in T_CM (n = 7) measured by flow cytometry after treatment with 100μM HODHBt or 10μM Isotretinoin alone, plus 100 ng/mL IL-15, or αCD3/CD28. D: Calculation of synergy for LRA combinations using the Bliss independence model (n = 7). Data is presented as the difference between the observed and expected fractional response. E: Viability of uninfected or latently infected matched conditions (n = 7). F: Memory CD4 T cells were stained with CellTrace yellow and treated with 100μM HODHBt or 10μM Isotretinoin alone or in combination with 100 ng/mL IL-15, or αCD3/CD28 for 7 days. Proliferation was measured by flow cytometry (n = 5). Wilcoxon matched-pairs signed rank test was used to calculate p values (*p < 0.05; **p < 0.01).

Fig 2. Effects of Isotretinoin on latent viral reactivation.

A: CD4, CD8, and NK cell activation analysis after stimulating with indicated conditions for 96 hours (n = 10 per arm). B: p24 levels were measured in the supernatants of PBMCs (blue = pre-Isotretinoin treatment, red = post-16 weeks of isotretinoin treatment) left unstimulated or treated with either 100 ng/mL IL-15 or αCD3/CD28 for 96 hours. Open circles indicate participants where the raw p24 values were non-quantifiable as described in the methods. Dashed line denotes the assay average limit of detection (aLOD) and median values are shown. C: % Increase in released p24 over the unstimulated condition for both IL-15 and αCD3/CD28 reactivation conditions. Wilcoxon matched-pairs signed rank test was used to calculate p values (*p < 0.05; **p < 0.01).

Fig 3. Effects of Isotretinoin on the translation-competent viral reservoir ex vivo in cells from PWH.

P24 levels were measured in supernatants (A) and cell lysates (B) after culture for 96 hours with the labeled conditions (n = 8) for αCD3/CD28 responders. For the supernatants and cell lysates, dashed line denotes the HIV-negative baseline (aHNB) and average limit of detection (aLOD). Horizontal lines display median values for each treatment condition. C: Cytotoxicity of treatment conditions for each participant was measured compared to unstimulated controls. P24 levels were measured in supernatants (D) and cell lysates (E) after culture with the labeled conditions (n = 4) for αCD3/CD28 non-responders. F: Cytotoxicity of treatment conditions for each participant was measured compared to unstimulated controls. Wilcoxon matched-pairs signed rank test was used to calculate p values (*p < 0.05; **p < 0.01).

Fig 4. Isotretinoin increases levels of pro-apoptotic proteins FAS and Noxa and increases caspase-3 activation in HIV-infected cells.

A: HIV-1 reactivation in latently infected T_CM (n = 3) previously reactivated at day 17 with 100ng/mL IL-15 alone, IL-15 with 100μM HODHBt, IL-15 with 10μM Isotretinoin. Second reactivation from days 24–26 with αCD3/CD28. Paired t-test was used to calculate p values (*p < 0.05; **p < 0.01). B: CD95 MFI of matched cultured uninfected T_CM (n = 9) was measured by flow cytometry after treatment with 30 IU/mL IL-2 or IL-2 plus 100μM HODHBt or 10μM of indicated compounds for 72 hours. Significant p values over cells treated with IL-2 alone. C: NOXA levels were measured in cultured uninfected T_CM treated with 100μM HODHBt or 10μM Isotretinoin alone, plus 100 ng/mL IL-15 (n = 2–3). Paired t-test was used to calculate p values (*p < 0.05; **p < 0.01). D: Caspase-3/7 activation measured in uninfected (CD4+p24-) and infected (CD4-p24+) CD4 T cells by flow cytometry after treatment for 48 hours with 100μM HODHBt or 10μM Isotretinoin alone, plus 100 ng/mL IL-15. Histograms represents uninfected (CD4+p24-) and infected (CD4-p24+) populations in one representative donor. Graphs indicate fold induction of PE-Caspase 3/7 MFI values for treatments over DMSO controls for both uninfected (CD4+p24-) and infected (CD4-p24+) populations (n = 8). Wilcoxon matched-pairs signed rank test was used to calculate p values (*p < 0.05; **p < 0.01).