Quisinostat is a brain-penetrant radiosensitizer in glioblastoma

Abstract

Histone deacetylase (HDAC) inhibitors have garnered considerable interest for the treatment of adult and pediatric malignant brain tumors. However, owing to their broad-spectrum nature and inability to effectively penetrate the blood-brain barrier, HDAC inhibitors have failed to provide substantial clinical benefit to patients with glioblastoma (GBM) to date. Moreover, global inhibition of HDACs results in widespread toxicity, highlighting the need for selective isoform targeting. Although no isoform-specific HDAC inhibitors are currently available, the second-generation hydroxamic acid-based HDAC inhibitor quisinostat possesses subnanomolar specificity for class I HDAC isoforms, particularly HDAC1 and HDAC2. It has been shown that HDAC1 is the essential HDAC in GBM. This study analyzed the neuropharmacokinetic, pharmacodynamic, and radiation-sensitizing properties of quisinostat in preclinical models of GBM. It was found that quisinostat is a well-tolerated and brain-penetrant molecule that extended survival when administered in combination with radiation in vivo. The pharmacokinetic-pharmacodynamic-efficacy relationship was established by correlating free drug concentrations and evidence of target modulation in the brain with survival benefit. Together, these data provide a strong rationale for clinical development of quisinostat as a radiosensitizer for the treatment of GBM.

Keywords: Brain cancer; Drug therapy; Oncology; Radiation therapy; Therapeutics.

Figure 1. QST exhibits low nanomolar efficacy against human GSC cultures.

(A) Dose-response curves with QST (10–1000 nM). Cell viability was measured across 7 patient-derived GSCs and 1 serum-grown long-term glioma line (U87) 3–5 days after treatment with QST. (B) Table illustrating the IC₅₀ of QST for each cell line tested. (C) Immunofluorescent staining of GSC lines BT145 (left) and GB126 (right) 72 hours after treatment with QST at the IC₅₀ concentrations. Control and drug-treated cells were stained for Ki67 and cleaved caspase-3 (CC3) to assess cell proliferation and cell death, respectively. Original magnification, ×20. Scale bar: 20 μm. (D) Quantification of Ki67-positive and CC3-positive cells (n = 3 mice per cell line). The dots or squares indicate values, the bar indicates the mean value, and the error bars indicate SEM. *P < 0.05, **P < 0.01 by unpaired, 2-tailed Student’s t test. For each cell line, the data are compiled from at least 3 independent experiments.

Figure 2. Short-term treatment with QST induces apoptosis and cell cycle arrest in human GSC cultures.

(A) Flow cytometry analysis of apoptosis through annexin V staining in BT145 (top) and GB126 (bottom). Representative flow cytometry dot plots of cells stained for annexin and ViaDye Red counterstain in DMSO- and QST-treated cells. The dots or squares indicate values, the bar indicates the mean value, and the error bars indicate SEM. (B) Mean proportion of cells in each phase of the cell cycle in BT145 and GB126 cells 24 hours after treatment with DMSO or QST, assessed by propidium iodide staining through flow cytometry (n = 3 mice per cell line). BT145-DMSO: G₁, 61%; S, 19%; G₂, 18%. BT145-QST: G₁, 36%; S, 17%; G₂, 46%. GB126-DMSO: G₁, 57%; S, 33%; G₂, 10%. GB126-QST: G₁, 42%; S, 55%; G₂, 3%. (C) Representative immunoblots showing dose-dependent increase in histone H3 acetylation and p21 levels in GSC lines after 24-hour treatment with QST. Fold change (FC) values are indicated above H3K9/14ac bands to indicate changes in acetylated histone H3 relative to DMSO-treated cells. *P < 0.05, **P < 0.01 by unpaired, 2-tailed Student’s t test. For each cell line, the data are compiled from at least 3 independent experiments.

Figure 3. QST sensitizes GSCs to IR.

Immunofluorescent staining of BT145 (A) and GB126 (B) showing an increase in γ-H2AX foci 72 hours after treatment in QST-treated cells but not in DMSO-treated cells. The mean number of γ-H2AX foci quantified in each cell per treatment condition is shown to the right for BT145 and GB126. Representative immunoblots demonstrating that QST treatment in BT145 (C) and GB126 (D) results in accumulation of γ-H2AX over time (left) and but not after drug washout (right). For each cell line, the data are compiled from at least 3 independent experiments. Original magnification, ×63. Scale bars: 20 μm. ****P < 0.0001 by unpaired, 2-tailed Student’s t test.

Figure 4. QST radiosensitizes GSCs in vitro.

Dose-response curves combining QST and IR treatment in BT145 (A) and GB126 (B). Matrices illustrate the zero interaction potency synergy scores when combining QST with increasing doses of IR in BT145 (C) and GB126 (D). Representative immunoblots show protein levels of γ-H2AX in BT145 (E) and GB126 (F) 1, 6, and 24 hours after treatment with either QST alone (25 nM), IR alone (4 Gy), or both QST and IR (25 nM and 4 Gy, respectively). For each cell line, the data are compiled from at least 3 independent experiments.

Figure 5. QST is effective in slowing tumor growth in a flank model of human GBM.

(A) Schematic illustrating the experimental design. Athymic nude mice were treated with a single dose of QST (10 mg/kg) through IP injection, SC injection, or OG. Blood samples were collected at 0.5, 1, 2, 4, 6, 8, and 24 hours after dosing and analyzed by LC-MS/MS. (B) Total plasma concentration–time curve for QST administered through various routes. Values for AUC_last were calculated for each route to illustrate plasma QST exposure (bottom). Error bars indicate SEM. (C) Schematic illustrating the treatment regimen for mice with flank tumors. When the tumors reached a mean volume of 100 mm³, mice were randomized into 4 groups: vehicle, 10 mg/kg QST, IR alone (6 Gy), or combination treatment (6 Gy IR and 10 mg/kg QST) (n = 10 mice in each cohort). IR was given in fractionated doses (2 Gy MWF) only during the first week of treatment, with or without QST. Following completion of IR, mice in the monotherapy and combination cohorts continued to receive QST alone on MWF until the tumors reached the indicated volume threshold. (D) Weekly mean volume measurements of U87 flank tumors from mice treated with vehicle, QST, IR, or a combination of QST and IR (QST+IR) (n = 10 mice in each cohort). Error bars indicate SEM. (E) Mean weights of mice from each cohort throughout the study duration. Error bars indicate SEM. (F) Total levels of QST in plasma and flank tumors of mice treated with QST and QST+IR (n = 3 or 4 mice per cohort). Error bars indicate SEM. (G) Immunoblotting of protein lysate–derived homogenized flank tumors from each cohort (n = 3 mice per group). Membranes were probed for H3K9/14ac, H3K27ac, γ-H2AX, and β-actin. Differences were assessed using ordinary 1-way ANOVA with Dunnett’s multiple-comparison test.

Figure 6. QST is a brain-penetrant HDACi.

(A) Stability of QST (10 nM and 500 nM) in mouse plasma at 37°C. (B) Stability of QST in mouse nonperfused and perfused brain homogenate (1:7 weight/volume in PBS) at 37°C. (C) Stability of QST in human plasma at 37°C. (D) Stability of QST in human brain homogenate at 37°C. In A–D, values are the mean of triplicate measurements, and error bars indicate SEM. (E) Schematic illustrating the design of the treatment study in non–tumor-bearing athymic nude mice. Mice received treatment with vehicle or QST on MWF for 2 consecutive weeks and were euthanized 2 hours after the last dose of drug for PK and PD analyses. (F) Mean weights of mice in the vehicle- and QST-treated cohorts throughout the study duration (n = 10 mice per cohort). (G) Total and unbound levels of QST in normal brain tissue in QST-treated mice (n = 6 mice per cohort). (H) Immunoblotting of protein lysates derived from homogenized brains from each cohort (n = 3 mice for vehicle cohort, n = 6 mice for QST cohort). Membranes were probed for H3K9/14ac and GAPDH. Quantification of expression of H3K9/14ac (normalized to GAPDH) in QST- and vehicle-treated homogenized brain samples is shown to the right. **P < 0.01 by unpaired, 2-tailed Student’s t test. In G and H, circles and squares indicate values, bars indicate mean values, and error bars indicate SEM.

Figure 7. PK and PD analysis of QST in an orthotopic PDX model of GBM.

(A) Schematic illustrating the experimental design of the treatment study in an orthotopic PDX model of GBM (GB126) to perform PK-PD analyses after short-term treatment (1 week) with QST and/or IR. Total (B) and unbound (C) levels of QST in tumor tissue and brain tissue contralateral to the tumor in mice treated with QST and QST+IR (n = 4 or 5 mice per cohort). Values are the mean of triplicate measurements. Circles and squares indicate values, bars indicate mean value, and error bars indicate SEM. (D) Immunoblotting of protein lysates derived from homogenized brain tumors from each cohort to assess changes in histone H3 acetylation and DNA damage (n = 3 mice per cohort). Membranes were probed for H3K9/14ac, H3K27ac, γ-H2AX, and β-actin. (E) Assessment of cell death as indicated by cleaved poly (ADP-ribose) polymerase (PARP) in homogenized brain tumors from each cohort (n = 3 mice per cohort).

Figure 8. QST is a potent radiosensitizer in an orthotopic PDX model of GBM.

(A) Schematic illustrating the experimental design of the treatment study in an orthotopic PDX model of GBM (GB126) to assess survival benefit across all treatment cohorts. Mice received IR only in the first week of treatment. Mice continued to receive QST treatment alone until they displayed neurological symptoms (the survival PK and PD endpoint). QST and IR were administered MWF. (B) Representative heatmap images of bioluminescence intensity across all treatment cohorts 22 days after treatment initiation. (C) Mean photon flux (p/s) measured through bioluminescence imaging across all cohorts throughout the entire duration of the treatment. Error bars indicate SEM. (D) Kaplan-Meier survival analysis of mice treated with vehicle, QST (10 mg/kg), IR (6 Gy), or QST+IR (6 Gy IR and 10 mg/kg QST) mice. Total (E) and unbound (F) levels of QST in tumor tissue and brain tissue contralateral to the tumor in mice treated with QST and QST+IR (n = 4 or 5 mice per cohort). Circles and squares indicate values, bars indicate mean values, and error bars indicate SEM. **P < 0.01, ***P < 0.001, ****P < 0.0001 by unpaired, 2-tailed Student’s t test (C) or Kaplan-Meier method with the Mantel-Cox log-rank test (D).

Figure 9. End-stage PD assessment of QST in an orthotopic PDX model of GBM.

(A) Immunoblotting of protein lysates derived from homogenized brain tumors from each cohort (n = 3 mice per cohort). Membranes were probed for H3K9/14ac, H3K27ac, and β-actin. (B) Normalized levels of H3K9/14ac and H3K27ac protein in all cohorts. Circles and squares indicate values, bars indicate mean values, and error bars indicate SEM. *P < 0.05, ***P < 0.001, ****P < 0.0001 by unpaired, 2-tailed Student’s t test. NS, not significant.

Figure 10. Combined treatment with QST and IR induces cell cycle arrest and a neuron-like cell fate in vivo.

Venn diagrams show overlap in genes upregulated (A) or downregulated (B) in response to QST monotherapy, IR, or combination treatment (QST+IR). Gene numbers in each section are shown in parentheses. (C) Volcano plots showing the –log₁₀(P value) and log₂(fold change) for transcripts detected by RNA-seq analysis of endpoint orthotopic GB126 xenograft tumors treated with QST (left), IR (middle), or QST+IR (right). Significantly up- and downregulated genes (false discovery rate < 0.05, 2-fold) are marked in red and blue, respectively. GO analysis of genes upregulated (D) or downregulated (E) in GB126 tumors due to QST, IR, or QST+IR treatment. (F) Reverse transcription quantitative real-time PCR analysis of the expression of neuronal genes in GB126 tumors treated with either QST monotherapy or combination therapy. Circles, squares, and triangles indicate values; bars indicate mean values, and error bars indicate SEM. *P < 0.05, **P < 0.01 by unpaired, 2-tailed Student’s t test. NS, not significant.

Figure 11. QST is a brain-penetrant HDACi that sensitizes GBM cells to radiation treatment.

Summary of the major findings in this study. Through a careful PK-PD–guided approach, we determined that QST can cross the BBB and exert its intended PD effect (increase in histone acetylation) in normal brain tissue as well as GBM cells. A second-generation HDACi, QST has high subnanomolar isoform selectivity for HDAC1 and HDAC2, which are class I HDAC isoforms that are primarily responsible for mediating histone deacetylation. Free (non–protein bound) QST inhibit the function of HDAC1 and HDAC2, resulting in widespread histone hyperacetylation in GBM cells. QST treatment alone in GBM cells also results in increased levels of DNA damage and oxidative stress. When QST is combined with IR treatment, genes involved in DNA damage repair pathways and cell division are downregulated, whereas genes that regulate neuronal development and function are significantly upregulated. These findings suggest that combination therapy of QST and IR provides therapeutic benefit through decreased cell proliferation, dampening of pathways involved in the DNA damage repair response, and a shift toward a neuron-like cell fate. ROS, reactive oxygen species.