Fig 1.
HIV-1 infection activates caspases and sheddases.
(A) (Left panel) Enzymatic cleavage of a membrane proximal CD62L peptide by cell membrane-associated proteases isolated from day 6 HIV-1BAL infected (BAL) or uninfected (UI) CD8-depleted PBMC in the presence and absence of 20 μM BB-94. The data are from technical duplicates and the results are representative of two independent experiments. (Right panel) ELISA measurement of soluble CD62L concentration from the infected (BAL) and uninfected (UI) culture media. The statistics are analyzed using technical repeats in a two-tailed student t-test for each experiment with p values <0.01 (**), p < 0.0001 (****). (B) Flow cytometry analysis of caspase activation in CD4+ T cells from HIV-1BAL infected or uninfected PBMC on day 6 of post infection. The cells are stained for intracellular viral capsid protein p24 (top) and activated caspases in p24+ vs p24- T cells (bottom). T cells with activated caspases in p24+ and p24- populations as well as in the uninfected sample are highlighted in red, green and black squares, respectively. (C) Bar diagram showing caspase activation in p24+ (red), p24- (green) populations of the infected sample as well as the uninfected control (black). Caspase activation correlated with CD62L and CD4 loss (S1B Fig). Mann-Whitney nonparametric test ***p < 0.001. The results are representative of three independent experiments. (D) Differential gene expressions between infected and uninfected CD4+ T cells from representative transcriptome analyses by RNA-Seq. HIV, QVD and UI denote cells infected with HIV-1BAL in the absence or presence of QVD-OPH, or uninfected sample. Nef +/- samples are cells infected with Nef-sufficient or -deficient strains of HIV-1NL4-3. HIV-1BAL infection elevated the expressions of many genes involved in inflammatory signaling pathways. (E) Differential gene expressions between HIV-1BAL infected and uninfected CD4+ T cells from 3 independent donors (S1 Table, S2 Fig) showed that the viral infection elevated the expressions of several caspases. (F) FLICA staining of individual caspases in primary CD4 T cells from a single donor following infection by HIV-1BAL in the absence or presence of QVD-OPH. Data are presented in duplicates and analyzed by comparing infected DMSO-treated to uninfected controls using student t-test *p < 0.05, **p < 0.005.
Fig 2.
Caspase activation results in phosphatidylserine exposure and loss of CD62L expression in infected T cells.
(A-C) Representative FACS contour plots on day 6 HIV-1BAL infected PBMC stained with anti-p24 antibody, annexin V, fluorescent inhibitor of caspases (FLICA), and anti-CD62L antibody. The populations exhibiting caspase+/annexin V+ (A) and annexin V+/CD62L- (B) are highlighted in colored quadrants. QVD-OPH (blue gates) but not Belnacasan (orange gates) inhibited caspase activation (A), CD62L loss (B) and the viral infection (C). Histograms corresponding to the FACS data presented in panels A and B are shown in S3E Fig for the caspases and CD62L expressions as well as annexin V staining in the infected and uninfected T cells. (D, E) Quantification of the populations in panels B and C. (F) On cell enzymatic cleavage of a fluorogenic CD62L peptide by HIV-1BAL infected and uninfected PBMC in the presence of QVD-OPH, BB-94 or control DMSO. Data are representative of three independent experiments and statistics are shown as Mann-Whitney nonparametric test *p < 0.05, **p < 0.01.
Fig 3.
Targeting pan-caspase activation suppresses HIV-1 infection.
(A-B) FACS analyses for HIV-1BAL infection level and the loss of CD62L expression in infected CD4 T cells in the presence of specific caspase inhibitors. Primary CD8-depleted PBMC were infected with HIV-1BAL for 6 days in the presence of 50uM indicated caspase inhibitors and analyzed for the accumulation of p24+/CD62L- populations (A) as well as the level of infection using intracellular p24 staining (B). The cells are stained and gated by anti-p24 and anti-CD62L antibodies as well as by FLICA specific to indicated caspases. Single caspase inhibitors used are Z-VDVAD-FMK (casp 2), Z-DQMD-FMK (casp 3), Z-IETD-FMK (casp 8), and Z-LEHD-FMK (casp 9). Oligo-caspase inhibitors used are Belnacasan (casp 1 and 4), Ac-DEVD-CHO (casp 3 and 7), and Z-DEVD-FMK (casp 3,6,7 and 10). QVD-OPH was used as the pan-caspase inhibitor. The results are from two independent experiments. All statistics were calculated with respect to the control DMSO-treated experiments using Mann-Whitney nonparametric test *p < 0.05, **p < 0.01. (C) Bar diagram showing suppression of an X4-tropic HIV-1LAI infection of PBMC by pan-caspase inhibitors, ZVAD-FMK and QVD-OPH. The results are representative of two independent experiments analyzed on day 6 of post infection. Mann-Whitney nonparametric test *p < 0.05. (D) Correlation between caspase activation and infection level. Those treated with DMSO (red) or Belnacasan (green) had the higher infections and caspase activations than QVD-OPH treated (blue). The uninfected controls are shown in open circles. The linear regression resulted in an R square of 0.83 (goodness of fit). (E) Inhibition of cell-to-cell transfer mediated viral transmission. HIV-1BAL infected PBMC were incubated with TZM-BL reporter cells at various density ratios in the presence of Belnacasan, QVD-OPH or DMSO control. The infections are measured luciferase activities from the reporter cells. 2-way ANOVA using Tukey’s method performed on triplicates of each compound per culture ratio. ****p < 0.0001. (F) Bar diagram showing the ratio of infections between 4:1, 2:1 (PBMC to TZM-BL), and that of 1:1 transfer infections. The efficiency of viral transfer following removal of free virus is significantly decreased for QVD-OPH but not Belnacasan. 2-way ANOVA using Tukey’s method performed on triplicates of each compound *p < 0.05, ****p < 0.0001.
Fig 4.
Caspase inhibition reduced HIV viral release.
(A) Infections of replication-incompetent JRFL (CCR5-tropic) and SF33 (CXCR4-tropic) luciferase-expressing pseudo-HIV to PBMC in the presence of caspases inhibitors, Belnacasan (orange), QVD-OPH (blue), HIV-1 entry inhibitor, T20 (black), or control DMSO (red). The levels of infections are measured by luciferase activities in relative light units. Statistical analyses are performed using Mann-Whitney nonparametric test *p < 0.05. (B) Total HIV RNA per 106 cells was determined by RT-PCR of day 6 HIV-1BAL infected primary CD4 T cells in the presence of Belnacasan, QVD-OPH or control DMSO. Significant reduction of total HIV RNA was observed in QVD-OPH treated infections. (C) Viral transcription levels were measured by the ratio between copies of HIV-1 RNA and genomic DNA (gDNA) in day 6 HIV-1BAL infected primary CD4 T cells. HIV gDNA copies were measured with LTR primer by real time PCR using ACH-2 cells that containing a single integrated copy of HIV as standard. The results are from two independent experiments. Mann-Whitney nonparametric test **p < 0.01. (D) Pie charts showing frequency (size of slice) of gene expressions and their ratio (color of the heatmap) between HIV-1BAL infected primary CD4+ T lymphocytes in the presence and absence of QVD-OPH for ADAM and caspase genes from representative duplicates of RNA sequencing data (Fig 1D). (E) RT-PCR analyses of the caspase expressions using caspases 1, 3, and 8 specific primers from day 3 HIV-1BAL infected PBMC in the presence of QVD-OPH or DMSO. The gene expression level is calculated relative to that of control beta actin. (F-H) Representative viral release assays from two independent experiments. Total infectious units produced from day 6 HIV-1BAL infected PBMC in the presence of Belnacasan (orange), BB-94 (green), QVD-OPH (blue), or DMSO (red) in the form of both free virus released into media (F) and cell surface-associated virus recovered through trypsinization (G) were determined using TZM-BL reporter cells. The infectious ratio between cell-associated and free supernatant viruses measures relative viral association on cells (H). While BB-94 and QVD-OPH reduced both the number of free released and cell-associated viruses, QVD-OPH treated infections showed proportionally more cell-surface associated viruses. The statistical analyses were performed using one-way ANOVA **p < 0.01, ***p < 0.001, ****p < 0.0001.
Fig 5.
Caspase inhibition prevents HIV-1 release and reduces viral fitness.
(A-F) Representative electron microscopy images of day 6 HIV-1BAL infected PBMC in the absence (A-C) or presence (D-F) of QVD-OPH. Transmission-EM (A, D) and pseudo-colored Scanning-EM (B, E) images of budding virions on infected cells. Majority virions in the control treated samples (panels A-C) show a uniform distribution of typical morphological sized particles of ~100–150 nm (white arrows) with visible capsids. Smaller sized (<50 nm) aggregating virus-like particles were frequently observed in QVD-OPH treated samples in both TEM (orange arrows) (D) and SEM (E) images. Panels C and F are TEM images of day 6 HIV-1BAL infected PBMC stained with anti-CD62L together with a 10nm gold particle-conjugated secondary antibody. Majority virions were found segregated from gold-labeled CD62L in control DMSO treated cells (C), but colocalized in the presence of QVD-OPH (F). All scale bars = 200 nm with the insert in (F) presented in 4x additional magnification. (G) Bar diagrams with matching representative EM images show size distribution (mature ≥ 100 nm, immature ≥ 100 nm without capsid, and small ≤ 50 nm) of virus-like particles (VLPs) by ratio of total observed particles associated with primary CD4+ T lymphocytes infected with HIV-1BAL in the presence or absence of QVD-OPH quantified from ~5 images per condition over 3 independent experiments. ***p < 0.001, ****p < 0.0001. Statistical analysis performed on pooled data from all donors with no data omitted. (H) Viral fitness assay comparing viral p24 concentration normalized infectious activities among DMSO, QVD-OPH and BB-94 treated infections. The infectious activities of released viral progeny were measured through infection of TZM-BL reporter cells and the concentrations of p24 viral capsid protein in released viruses were quantified by ELISA. Data were presented as triplicates from each treatment and the statistical analysis was done using 2-way ANOVA test *p < 0.05, **p < 0.01, ****p < 0.0001. The results are representative of two independent experiments. (I) Immunoblot of gag p55, p24, and p17 proteins in the infected cell lysate (Lysate) or in released free viruses (Virus) from infections treated with DMSO, QVD-OPH, or Nelfinavir. (J) FACs analysis of the infection levels on samples used for the immunoblots Compared to DMSO controls, QVD-OPH and Nelfinavir significantly reduced p24 levels in both immunoblots and FACS analyses (panel I and J). The results were representative of four independent experiments. The statistical analysis was performed using one-way ANOVA ***p < 0.001. (K) Normalized p24/p55 band intensity ratio to DMSO controls from three independent western blot experiments. Individual p24 and p55 band intensities were determined from cell lysate of the same sample using ImageJ (https://imagej.nih.gov/ij/).
Fig 6.
HIV-1 Nef mediates caspase activation for viral release.
(A) RNAseq analyses for differential gene expression on highlighted gene groups responsible for the cellular signaling pathways leading to caspase activation and apoptosis. RNA sequencing were performed using day 6 Nef+ and Nef- HIV-1NL4-3 infected CD4 T cells from 2 donors. Depicted are genes involved in interferon-induced intrinsic apoptotic pathway, TNF-alpha and death receptor-mediated extrinsic apoptotic pathways, as well as genes involved in inflammasome activation. Nef+ HIV infection resulted in significant upregulation of transcriptions of these pathways compared to Nef- infections. Paired student t-test, *p < 0.05, **p < 0.01, ***p < 0.001. (B) Histograms showing total caspase expressions in Nef+ (red) and Nef- (blue) HIV-1NL4-3 infected PBMC on day 6 post infections.(C) FLICA staining of caspases in p24+ (red) and p24- (black) populations of day 7 infected PBMC using Nef+ and Nef- HIV-1NL4-3 viruses. The results are representative of two independent experiments. 2-way ANOVA ***p < 0.001. (D-E) Percentage of total caspase activation (D) and p24 staining (E) on day 7 of Nef+ or Nef- HIV-1NL4-3 infected PBMC in the presence or absence of QVD-OPH. The statistics are calculated using student t-test **p < 0.01, ***p < 0.001, ****p < 0.0001. (F-G) Level of infection by p24 staining (F) and resulting pan-caspase activation (G) in PBMC’s infected by HIV-1NL4-3 clones expressing different sequences for Nef. One-way ANOVA **p < 0.01, ***p < 0.001, ****p < 0.0001. (H-I) Correlation plots between p24 and caspase activation across Nef variants from HIV (H) or SIV (I). R2 and Pearson correlation denoted ****p < 0.0001.
Fig 7.
Caspase Inhibitor suppresses the ex vivo viral reservoir.
(A) HIV viral release was measured as HIV RNA copies/ml in the presence of various caspases inhibitors or DMSO from CD4 T cells derived from infected individuals but stimulated in vitro with anti-CD3 for 3 and 6 days. PT1 and PT2 are two infected viremic individuals not receiving ART treatment. (B) Effect of QVD-OPH in suppressing HIV release from viral reservoirs. Viral RNA (copies/ml) were quantified from CD4 T cells in day 2 and 4 of anti-CD3 stimulated culture supernatants in the presence of QVD-OPH or DMSO from three aviremic individuals on ART treatment. (C) QVD-OPH inhibited HIV viral release from non-stimulated CD4 T cells derived from three infected individuals. (D) Percentage of HIV released from CD4 T cells with QVD treatment compared to that of DMSO from all infected viremic and aviremic individuals. The p-value is calculated using ordinary two-way ANOVA between QVD and DMSO treated samples.
Fig 8.
Targeting caspase activation suppresses the viral reservoir.
Model of Nef-mediated activation of caspases for promoting PS exchange thus stabilizing sheddase activity to cleave CD62L and permit the release of viral progeny. Inhibition of this activation causes downstream tethering of budding virus by CD62L resulting in a phenotypic change of viral morphology and loss of fitness.