Activation of HIV transcription with short-course vorinostat in HIV-infected patients on suppressive antiretroviral therapy

Abstract

Human immunodeficiency virus (HIV) persistence in latently infected resting memory CD4+ T-cells is the major barrier to HIV cure. Cellular histone deacetylases (HDACs) are important in maintaining HIV latency and histone deacetylase inhibitors (HDACi) may reverse latency by activating HIV transcription from latently infected CD4+ T-cells. We performed a single arm, open label, proof-of-concept study in which vorinostat, a pan-HDACi, was administered 400 mg orally once daily for 14 days to 20 HIV-infected individuals on suppressive antiretroviral therapy (ART). The primary endpoint was change in cell associated unspliced (CA-US) HIV RNA in total CD4+ T-cells from blood at day 14. The study is registered at ClinicalTrials.gov (NCT01365065). Vorinostat was safe and well tolerated and there were no dose modifications or study drug discontinuations. CA-US HIV RNA in blood increased significantly in 18/20 patients (90%) with a median fold change from baseline to peak value of 7.4 (IQR 3.4, 9.1). CA-US RNA was significantly elevated 8 hours post drug and remained elevated 70 days after last dose. Significant early changes in expression of genes associated with chromatin remodeling and activation of HIV transcription correlated with the magnitude of increased CA-US HIV RNA. There were no statistically significant changes in plasma HIV RNA, concentration of HIV DNA, integrated DNA, inducible virus in CD4+ T-cells or markers of T-cell activation. Vorinostat induced a significant and sustained increase in HIV transcription from latency in the majority of HIV-infected patients. However, additional interventions will be needed to efficiently induce virus production and ultimately eliminate latently infected cells.

Trial registration: ClinicalTrials.gov NCT01365065.

Figure 1. Induction of changes in acetylation of histone 3, histone 4 and lysine by vorinostat.

Changes in histone (H) acetylation were quantified using flow cytometry which is shown (for representative participant) as (A) fold change in mean fluorescence intensity (MFI) of antibody to acetylated (Ac) H3, Ac lysine (K) and Ac H4 in lymphocytes by size prior to, during and following vorinostat and (B) histograms of the change in MFI with antibody to Ac H3 following vorinostat. (C) PBMC were analysed by western blot using an antibody to Ac H3 and actin (as a control for total protein). A positive control (pos) of splenocytes from a mouse with acute myeloid leukemia treated with the HDACi panobinostat is also shown. (D) Fold change in acetylated (A) Histone 3 (red), (B) Lysine (orange) and (C) Histone 4 (blue) in total lymphocytes is shown for each study participant (solid circle) and the median (IQR) fold change above baseline is shown. *p<0.01, **p<0.001. Grey shaded box represents the time on vorinostat.

Figure 2. Individual changes in CA-US HIV RNA in blood and tissue.

A) Fold change in CA-US HIV RNA following vorinostat in CD4+ T-cells from blood (left panel) and rectal tissue (right panel) compared to baseline. The maximum fold change in CA-US HIV RNA on study (solid column) and change at day 84 (open column) is shown for CD4+ T-cells from blood; and change at day 14 for rectal tissue is shown for each participant (upper panel) and the median (IQR) change for all participants (lower panel). The grey dashed line indicates 1-fold change. B) Time to reach maximum fold increase in CA-US HIV RNA for each participant. Grey shaded box represents the time on vorinostat. (C) Correlation between baseline CA-US HIV RNA and peak CA-US HIV RNA (left panel) and day 84 CA-US HIV RNA (right panel).

Figure 3. Effects of vorinostat on CA-US HIV RNA, HIV DNA and HIV RNA in blood and tissue.

Changes in (A) CA-US HIV RNA (red), (B) plasma HIV RNA (green) and (C) HIV DNA (blue) is shown for each study participant (solid circle) and the median (IQR) at each time point in CD4+ T-cells from blood (left panel) and rectal tissue (right panel). Open circles represent data when at least one of the replicates were below the lower limit of detection (LLOD). The mean fold change in each parameter is also shown using a generalised estimating equation analysis (middle panel; boxes represent the median, 50^th and 75^th percentiles and whiskers represent the range). Grey shaded box represents the time on vorinostat. *p<0.01, **p<0.001.

Figure 4. Changes in virological and immunological parameters from one patient with viral rebound on study.

Changes in (A) CA-US HIV RNA (red), (B) plasma HIV RNA (green) and (C) CA-HIV DNA (blue) are shown as the mean± SD for replicates of CA-US HIV RNA and HIV DNA. Programmed death-1 (PD1) expression on CD4+ and CD8+ CD45RA- T-cells is shown. Grey shaded box represents the time on vorinostat. (D) Dot plot analysis of flow cytometry for co-expression of PD-1 and CD45RA on CD4+ (top panel) and CD8+ (lower panel) T-cells at baseline, after 7 days of vorinostat and at day 84 of follow up.

Figure 5. Vorinostat induced changes in the adaptive immune response.

(A) SEB-specific (upper panel) and gag-specific (lower panel) CD8+ T-cells were quantified by intracellular cytokine staining and changes in (B) frequency of regulatory T-cells using flow cytometry. Significant changes between time points were determined using a Wilcoxon sign-rank test. P values<0.05 were considered significant. All statistically significant differences are shown.

Figure 6. Vorinostat induced a transcriptional burst and chromatin perturbations that are recurrent with subsequent dosing.

(A) ANOVA (F-test) heatmap of top 50 differentially expressed genes (DEGs) from matched donor supervised analysis (n = 9) comparing gene expression two hours (2 h), eight hours (8 h) and one day following the initial dose of vorinostat. Gene expression was adjusted for baseline expression and represented as a gene-wise standardized expression (Z-score), with p-values<0.05. DEGs are annotated with colored arrowheads: red, transcription coactivators and adaptors; blue, DNA damage and ER stress response; purple, apoptosis resistance; black, chromatin remodeling factors; orange, mSIN3a histone-deacetylase complex subunits. (B) Checkerboard map of DEG two hours after the initial dose of vorinostat compared to baseline showing the top 10 enriched pathways on the x-axis and leading edge analysis (gene members contributing most to enrichment) plotted along the y-axis. Scale represents log₂ fold change where red corresponds to up- and blue down-regulated genes at two hours versus baseline. Genes associated with viral transcriptional activity are annotated with colored arrowheads: red, BAF component SMARCB1 (SNF5) and CDK9; black, splicesome and nuclear export proteins; orange, mSIN3A HDAC subunits.

Figure 7. Changes in host genes were associated with an increase in CA-US HIV RNA after vorinostat.

(A) Heatmap showing the fold change in gene expression using linear regression analysis between CA-US HIV RNA and DEG at two hours following the initial dose of vorinostat (n = 9). CA-US HIV RNA is plotted as a continuous variable (ranging from low to high – light green to dark green) and correlated with distinct gene expression profiles. The copy number of CA-US HIV RNA per million cells two hours following vorinostat for each participant is listed next to the patient identification code at the bottom of each column. The top 50 regression features (of a total of ∼2000 at nominal p-value<0.05) are shown. (B) Pathway analysis was performed on the regression features and a checkerboard map showing the top common enriched pathways on the y-axis and leading edge analysis (gene members contributing most to enrichment) plotted along the x-axis. Up- and down-regulated genes at two hours versus baseline are plotted as log₂ fold change (FC); red squares correspond to upregulated gene expression and blue downregulated gene expression. Genes associated with MAPK signal transduction pathways are annotated with black arrows and cell cycle regulators annotated with green arrows. Genes associated with the ER stress response and apoptosis are annotated with red arrows.

Figure 8. Similar gene expression and pathway activity after the 1st and 7th daily dose of vorinostat.

(A) ANOVA (F-test) heatmap of top 50 DEGs from matched donor supervised analysis (n = 5) comparing gene expression at 2 hours following the first dose (day 0+2 hours), day 7 (7d; prior to the seventh dose), and two hours following the seventh dose (day 7+2 hours). Gene Expression was adjusted for baseline expression and represented as a gene-wise standardized expression (Z-score), with p-value<0.05. The heatmap shows hierarchical clustering of day 0+2 hour and day 7+2 hour timepoints for each subject indicative of highly congruent gene expression profiles. Checkerboard maps of DEG (B) two hours following the initial dose of vorinostat compared to baseline and (C) two hours following the seventh dose of vorinostat compared to prior to the seventh dose. The top 10 common enriched pathways are shown on the y-axis and leading edge analysis (gene members contributing most to enrichment) is plotted along the x-axis. Scale represents log₂ fold change. Genes up- or down-regulated similarly between the two time-points are indicated by black arrowheads whereas genes highlighted in red boxes are inversely or differentially regulated between time points.

Figure 9. Changes in gene expression over the duration of study with most early changes occurring in T-cells.

(A) ANOVA (F-test) heatmap of top 50 differentially expressed genes (DEGs) from matched donor supervised analysis (n = 9) comparing gene expression one, 14 and 84 days following the initial dose of vorinostat. Gene Expression was adjusted for baseline expression and represented as a gene-wise standardized expression (Z-score), with p-values<0.05. (B) Checkerboard map of differentially expressed genes 84 days after the initial dose of vorinostat (70 days post cessation of drug) compared to baseline showing the top 10 enriched pathways using gene subset enrichment analysis (GSEA) on the y-axis and leading edge analysis (gene members contributing most to enrichment) plotted along the x-axis. Scale represents log₂ fold change where red corresponds to up- and blue down-regulated genes respectively. (C) Pathway heatmap illustrates enrichment of gene expression in PBMC subsets at different timepoints compared to baseline. Red and blue represent up and down regulated expression of gene subsets respectively. (D) Checkerboard map of DEG at each timepoint compared to baseline. The cell subsets (modules) are plotted on the y-axis and gene members contributing to enrichment plotted on the x-axis. Scale represents log₂ fold change. Red and blue boxes represent up and down gene regulation respectively. mDC = myeloid cells; pDC = plasmacytoid dendritic cells; NK = natural killer cells.