The BCL-2 Inhibitor Venetoclax Augments Immune Effector Function Mediated by Fas Ligand, TRAIL, and Perforin/Granzyme B, Resulting in Reduced Plasma Viremia and Decreased HIV Reservoir Size during Acute HIV Infection in a Humanized Mouse Model

Abstract

The BCL-2 prosurvival protein is implicated in HIV persistence and is a potential therapeutic target for HIV eradication efforts. We now know that cells harboring HIV are preferentially enriched for high BCL-2 expression, enabling their survival, and that the BCL-2 inhibitor venetoclax promotes the death of actively replicating HIV-infected cells in vitro and ex vivo. Herein, we assess the effect of venetoclax on immune clearance of infected cells and show that BCL-2 inhibition significantly enhances target cell killing induced by Fas ligand, TRAIL (tumor necrosis factor-related apoptosis-inducing ligand), and perforin/granzyme B and synergistically enhances autologous NK (natural killer) and CD8 cells' killing of target cells. In a humanized mouse model of acute HIV infection, venetoclax monotherapy significantly decreases plasma viremia and normalizes CD4:CD8 ratios, and results in more mice with undetectable provirus levels than control. In this model, treatment was associated with leukopenia, as has been described clinically in patients receiving venetoclax for other indications. These data confirm meaningful anti-HIV effects of venetoclax during HIV infection but suggest that venetoclax use should be combined with ART (antiretroviral therapy) to reduce toxicity. IMPORTANCE This study is the first to examine the applicability of BCL-2 inhibition in the setting of active HIV infection in vivo. Furthermore, this study demonstrates that venetoclax significantly enhances target cell killing induced by Fas ligand, TRAIL, and perforin/granzyme B and synergistically enhances autologous NK and CD8 cells' killing of target cells.

Keywords: BCL-2 family; human immunodeficiency virus; venetoclax.

FIG 1

Venetoclax augments the HIV-infected cell clearance by FasL and TRAIL. CD4 T cells infected with HIV IIIb, in vitro, were treated with or without the BCL-2 inhibitor venetoclax (1 μM) or a vehicle control (DMSO) for 6 h followed by combination of DMSO/venetoclax with CH11-Fas-stimulating antibody (500 ng/mL), or SuperKiller TRAIL (SKT) (100 ng/mL) for 72 h. (A) Schematic of the experimental workflow for CH11 and SKT experiments. (B) Representative flow plots comparing the effect of DMSO and venetoclax on CH11 and SKT function. (C) Mean (SEM) of 5 experiments, demonstrating the proportion of p24-positive, caspase-3-positive cells with the addition of CH11 with or without venetoclax. (D) Mean (SEM) of 5 experiments, demonstrating the proportion of p24-positive, caspase-3-positive cells with the addition of SKT with or without venetoclax. (E) Mean (SEM) of 5 experiments, demonstrating the proportion of live p24-positive cells with the addition of CH11 with or without venetoclax. (F) Mean (SEM) of 5 experiments, demonstrating the proportion of live p24-positive cells with the addition of SKT with or without venetoclax. (G) Supernatant p24 values for the CH11 cultures (n = 4) relative to vehicle control. (H) Supernatant p24 values for the SKT cultures (n = 4). (I and J) Representations of Bliss independence calculations for CH11 and SKT, respectively; values greater than the predicted combination (dotted line) represent synergy. (K to M) Fold change in the percentage of aCaspase-3 positive cells, relative to DMSO, between inactivated, uninfected CD4 cells and p24-positive cells. Significance for the above plots calculated using a matched one-way ANOVA, with Holm-Sidak correction for multiple comparisons. Significance defined as P ≤ 0.050.

FIG 2

Venetoclax augments the HIV-infected cell clearance by perforin/granzyme B. CD4 T cells infected with HIV IIIb, in vitro, were treated with or without the BCL-2 inhibitor venetoclax (1 μM) or a vehicle control (DMSO) for 6 h, followed by a combination of DMSO/venetoclax with perforin/granzyme B for 45 min. (A) Schematic of the experimental workflow for perforin/granzyme B experiments. (B) Representative flow plots comparing the effect of vehicle and venetoclax on perforin/granzyme B function. (C) Mean (SEM) of 10 independent experiments demonstrating p24-positive target cell death following perforin/granzyme treatments, as determined by Live/Dead staining. (D) Representations of Bliss independence calculations for perforin/granzyme B. Values greater than the predicted combination (dotted line) represent synergy. (E) Fold change in the percentage of Live/Dead positive cells, relative to DMSO, between inactivated, uninfected CD4 cells and p24-positive cells. Significance for the above plots calculated using a matched one-way ANOVA, with Holm-Sidak correction for multiple comparisons. Significance defined as P ≤ 0.050.

FIG 3

Venetoclax increases HIV-infected cell clearance by NK cells. CD4 T cells infected with HIV IIIb, in vitro, were treated with or without the BCL-2 inhibitor venetoclax (1 μM) or a vehicle control (DMSO) for 6 h followed by a combination of DMSO/venetoclax with autologous NK cells (1:1 E:T ratio) in duplicate, for 24 h. (A) Schematic of the experimental workflow for NK cell experiments. (B) Representative flow plots comparing the effect of vehicle and venetoclax on NK cell function. (C) Mean (SEM) of 8 experiments, demonstrating the proportion of p24-positive, target cells staining with Live/Dead stain. (D) Mean (SEM) of 8 experiments, demonstrating the proportion of live p24-positive cells in the NK cell cocultures. Significance for the above plots calculated using a matched one-way ANOVA, with Holm-Sidak correction for multiple comparisons. (E) Representations of Bliss independence calculations. Values greater than the predicted combination (dotted line) represent synergy. (F) Supernatant p24 values for the NK cocultures (n = 7) relative to vehicle control, significance calculated using a Friedman’s test, with Dunn’s test for multiple comparisons. Significance defined as P ≤ 0.050.

FIG 4

Venetoclax increases HIV infected cell clearance by cytotoxic T cells (CTLs). CD4 T cells infected with HIV IIIb, in vitro, were treated with or without the BCL-2 inhibitor venetoclax (1 μM) or a vehicle control (DMSO) for 6 h. The cells were subsequently cocultured with autologous, peptide-expanded CD8 (1:2 effector: target [E:T] ratio) for 48 h. (A) Schematic of the experimental workflow for CTL cell experiments. (B) Representative flow plots comparing the effect of vehicle and venetoclax on CTL function. (C) Mean (SEM) of 5 experiments, demonstrating the proportion of p24-positive target cells staining with Live/Dead stain. (D) Representations of Bliss independence calculations. Values greater than the predicted combination (dotted line) represent synergy. (E) Supernatant p24 values for the CTL cocultures (n = 5) relative to vehicle control. Significance for the above plots calculated using a matched one-way ANOVA, with Holm-Sidak correction for multiple comparisons. Significance defined as P ≤ 0.050.

FIG 5

Venetoclax treatment reduces HIV load in vivo. Two groups of mice were treated with either venetoclax or vehicle control, with a total of 25 mice in each group, starting at day −1. The mice were infected with 200 ng of HIV-1 (NL4-3 HIV1/Clade B; X4 tropism virus) by intraperitoneal injection on day 0 and followed for 10 weeks. Blood obtained periodically was assayed for human T-cell counts and percentages by flow cytometry and HIV RNA by PCR, and splenocytes obtained after sacrifice were examined for intact proviral DNA by IPDA. (A) Schematic representing experimental workflow. (B) Mean (SEM) of plasma HIV RNA, expressed as copies per milliliter, over time. Significant reductions in plasma viral RNA were observed from week 6 onwards. Values reported as undetermined or less than the limit of detection were assigned the value 2,000 (the LOD for the assay, denoted by the dotted line). At 6 weeks, 20/23 control mice had HIV RNA above the LOD, compared to 12/18 venetoclax-treated mice; at 8 weeks, it was 20/22 (vehicle) compared to 11/18 (venetoclax); and at 10 weeks 20/22 (vehicle) compared to 10/16 (venetoclax). (C) Mean (SEM) of the CD4 count per mL over time. Significance calculated using multiple t tests with false-discovery rate (FDR) of <0.05. Significance defined as P ≤ 0.050. (D) Linear regression analysis comparing the mean number of viral copies per CD4 cell. (E) Linear regression analysis comparing plasma viremia and CD4 count in the vehicle-treated group. (F) Linear regression analysis comparing plasma viremia and CD4 count in the venetoclax-treated group demonstrating significantly greater rate of decline in viral load compared to CD4 count. (G) Seven of nineteen mice in the venetoclax group had undetectable intact proviral HIV DNA, compared to 1/22 in the vehicle-treated group. Significance calculated with Fisher’s exact test. (H to K) Comparison of intact, 3′-defective, 5′-defective, and total proviral DNA in the mice with detectable viral load as determined by IPDA.

FIG 6

Cellular dynamics following venetoclax treatment in vivo in HIV infection. (A) Mean (SEM) of total CD45 cell count expressed as copies per milliliter, over time. (B) Mean (SEM) of the CD8 count per mL over time. (C and D) CD4:CD8 ratios in the vehicle- and venetoclax-treated groups, respectively. Significance calculated using a one-way t test compared to a hypothetical value of 1. (E) Percentage of live CD4 cells over time. (F) CD4 count at day 0 (n = 10) compared to vehicle (n = 10), calculated using an unpaired two-tailed t test.