Microtubule-Targeting Combined with HDAC Inhibition Is a Novel Therapeutic Strategy for Diffuse Intrinsic Pontine Gliomas

Diffuse intrinsic pontine glioma (DIPG) is an aggressive brainstem tumor and the most common form of high-grade glioma in children, with a median survival of less than 1 year (1, 2). DIPG is uniformly fatal and remains the leading cause of childhood brain tumor-related death (3). There are no established systemic therapies, and focal irradiation remains the standard treatment (4, 5). However, the application of autopsy and biopsy sampling has aided in understanding the molecular landscape of DIPG tumors over the past decade (6, 7). Approximately 80% of DIPG tumors harbor a mutation in the H3 histone (H3K27M), leading to loss of H3K27 trimethylation (H3K27me3) and aberrant gene expression (8). Moreover, the development of patient-derived cultures and orthotopic models allows the preclinical testing of a variety of epigenetic, targeted agents, and immunotherapies (1, 9, 10).

One of the defining characteristics of DIPG tumors is the infiltrative and invasive growth patterns that contribute to limited therapeutic response (11). Tumor migration, cell division, and structural integrity depend upon microtubules composed of dynamic tubular polymers of α- and β-tubulin (12, 13). Blocking the dynamics of microtubule assembly and disassembly results in disruption of cellular division, mitotic arrest, and induction of apoptosis, thus making this a potential strategy for cancer therapy (14). Trilexium (TRX-E-009–1) is a third-generation benzopyran compound that can cross the blood–brain barrier (BBB) and reach brain tissue. Benzopyran is a naturally occurring compound that has been shown to display different biological effects such as antimicrobial, anti-HIV, anti-inflammatory, anti-coagulant, and antitumor (15–17). As a novel anticancer therapy, TRX-E-009–1 has been shown to initiate various mechanisms of cell death via several pathways, including through inhibition of microtubule polymerization (18). TRX-E-009–1 has been shown to significantly reduce tumor volume in preclinical models of melanoma (19).

Recent therapeutic strategies targeting epigenetic alterations by using epigenetic modifiers have shown promise in preclinical models of pediatric brain cancer (20). These treatment strategies often aim at restoring the normal function of histones by reverting the repressed trimethylation or acetylation phenotype (21). A comprehensive study using a chemical screen in patient-derived DIPG neurospheres along with integrated computational modelling, has highlighted histone deacetylases (HDAC) as promising targets (9). HDACs have been shown to interact and deacetylate microtubules in vivo (22). This led us to a combinatorial approach to enhance efficacy.

Our study demonstrates the antitumor potential of TRX-E-009–1 in combination with HDAC inhibitor SAHA. The combination of both agents inhibited cell proliferation and clonogenic potential and induced G₂–M arrest and apoptosis. Furthermore, this combination also markedly enhanced survival in an orthotopic model of DIPG when combined with irradiation. These preclinical results support the translation of microtubule-targeting compounds into the clinic for DIPG therapy.

Patient-derived DIPG cultures were kindly provided by our collaborators at Stanford University and St John of Hope Hospital, respectively. The cells were cultured in DMEM/F12 and neurobasal media (1:1) supplemented with heparin, B27, human EGF, human FGF, PDGF-AA, PDGF-BB, HEPES, nonessential amino acids, antibiotic/antimycotic, pyruvate and glutamax. Normal human astrocytes and MRC5 cells were purchased from Lonza, and cultured according to the manufacturer's instructions. All cells were maintained at 37°C in 5% CO₂ and humidified atmosphere. Cell lines were validated at the Garvan Molecular Genetics using short tandem repeat DNA profiling and were routinely confirmed as being mycoplasma negative using MycoAlert Mycoplasma Detection Kit from Lonza. TRX-E-009–1 (solutol and liposomal formulations) were provided by Novogen. Solutol formulation was used to assess anti-clonogenic effects in combination with irradiation while all remaining experiments were performed with the liposomal formulation.

For cell viability assay, DIPG cells were seeded into 96-well plates (100 μL per well) and incubated for 72 hours to form neurospheres, whereas NHAs and MRC5 cells were incubated for 24 hours prior to adding the drugs. After incubation, cells were treated with various concentrations of TRX-E-009–1 and SAHA for an additional 72 hours. Resazurin reduction was then used to measure cell viability. Synergistic drug interactions were determined using CalcuSyn software (Biosoft) developed by Chou and Talalay where CI values of less than 1 indicated synergism, CI values more than 1 showed antagonism and CI equal to 1 represented additive effect. IC₅₀ values were determined by nonlinear regression analysis by Graphpad Prism 6 software (Applied BioSystems).

Long-term clonogenic assays were used to assess the effects of TRX-E-009–1, SAHA, and irradiation on the clonogenic activity of DIPG cells. First, 24-well plates were coated by media containing 0.5% soft agar. After 24 hours, cell suspensions in media containing 0.33% agar and drugs or DMSO as a vehicle were added to the bottom layer. Depending on the growth rate of DIPG cultures, 1,000 to 2,000 cells were seeded in each well. Cells were maintained at 37°C in a humidified 5% CO₂ atmosphere. After 2 to 3 weeks, the colonies were stained using 5 mg/mL 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and counted using ImageJ software.

SU-DIPGXVII cells were treated with TRX-E-009–1 for 24 hours. Cells were then collected and fixed with 70% ice-cold ethanol and stained with 25 μg/mL propidium iodide (Sigma-Aldrich, Catalog No. P4864) and 2 μg/mL DNase free RNase (Roche Diagnostics) in 0.1% BSA/PBS. Cells were resuspended in cold 0.1% BSA/PBS and were run on FACSCalibur Flow Cytometer (BD Biosciences). Data were analyzed using Flow analysis software (Flow Jo).

Following treatment, HSJD-DIPG007 cells were lysed using RIPA buffer supplemented with protease and phosphatase inhibitors and protein concentration was determined using the Pierce BCA Analysis Kit per manufacturer's instructions. Thirty micrograms of protein was then incubated for 5 minutes at 95°C in Laemmli loading buffer containing 10% DTT (Bio-Rad). Electrophoresis was conducted at 80V for 2 hours and proteins were transferred to nitrocellulose membranes (Bio-Rad) at 85V for 1 hour. Membranes were blocked in 5% skim milk in PBS containing 0.05% Tween 20 (PBST) and then incubated with primary antibodies diluted in 5% BSA in PBST containing overnight at 4°C [anti-gamma H2A.X (1:1,000; Abcam, #ab26350), p21 Waf1/Cip1 (1:1,000; Cell Signaling Technology, #2946), cleaved caspase-3 (1:1,000; Cell Signaling Technology, #9579), cleaved PARP (1:1,000; Cell Signaling Technology, #5625), GAPDH (1:2,000; Cell Signaling Technology, #2118), β-actin (1:1,000; Cell Signaling Technology, #8457), H3K27me3 (1:1,000; Cell Signaling Technology, #9733), H3K27ac (1:1,000; Cell Signaling Technology, #8173), acetyl-α-tubulin (1:1,000; Cell Signaling Technology, #3971), HDAC6 (1:1,000; Cell Signaling Technology, #7558)]. Membranes were then washed 5× using PBST and incubated with secondary antibody in PBST for 1 hour at room temperature. Proteins were visualized by chemiluminescence (Super Signal; Pierce).

The effect of the TRX-E-009–1 on cell-free tubulin polymerization was conducted using a Tubulin Polymerization Assay Kit (Cytoskeleton, #BK006P) as per manufacturers’ instructions. Tubulin polymerization assay on HSJD-DIPG007 cells was conducted as described previously (23). Briefly, HSJD-DIPG007 cells were seeded at a density of 1,000,000 cells/well in 6-well plates. Cells were incubated at 37°C for 3 days and then were either left untreated or treated with TRX-E-009–1. Cells were collected after 24 hours following treatment. Briefly, cells were washed with PBS twice and centrifuged at 3,000 rpm for 5 minutes at 4°C. Cells were then lysed in protease inhibitor and Hypotonic buffer 1 (20 mmol/L Tris-HCl, pH 6.8, 1 mmol/L MgCl₂, 2 mmol/L EGTA, 0.5% Nonidet P-40, 2 mmol/L PMSF, 200 U/mL Aprotinin, 100 μg/mL soybean trypsin inhibitor, 5 mmol/L e-amino caproic acid, and 1 mmol/L benzamidine) and incubated at 37°C in the dark. Samples were then vortexed and centrifuged at 16,000 × g for 10 minutes. Supernatant (depolymerized tubulin) was then separated from the cell pellet (polymerized tubulin) and transferred into new tubes. The cell pellets were resuspended in hypotonic buffer 2 containing 10 mmol/L Tris, pH 7.5, 1.5 mmol/L MgCl₂, 10 mmol/L KCl, 0.5% Nonidet P-40, and the protease inhibitor. Western blot analysis was conducted as described above using α-Tubulin (1:1,000, Cell Signaling Technology; 11H10 #2125).

All animal experiments were performed in accordance with the Animal Care and Ethics Committee of the University of New South Wales and the Australian Code of Practice for Care and Use of Animals for Scientific Purposes. All mice were maintained in a temperature-controlled environment with a 12-hour light/dark cycle. DIPG orthografts were generated as described previously (24). Briefly, female NOD/SCID and female balb c/nude mice (7–8 weeks, Animal Resources Centre) were stereotactically injected with 200,000 HSJD-DIPG007 or SU-DIPGVI in 2 μL of Matrigel in the 4th ventricle/pons. All animals were monitored daily for weight loss and neurological symptoms due to drug treatments and/or tumor engraftment. If ataxia, circling, head tilting, and/or weight loss of >20% occurred, the mice were humanely euthanized.

For a single-agent study, Balb/c nude mice received 50 mg/kg TRX-E-009–1 intravenously thrice a week for 4 weeks. Treatment started 4 weeks post-intracranial HSJD-DIPG007 cells inoculation and upon drug treatment.

The combination study treatment commenced at 5 weeks after intracranial injection of the cells. At week 5, specific cohorts were treated with irradiation at 2Gy per day for 4 days. TRX-E-009–1 and SAHA treatment as single agents or combination started on week 6 where TRX-E-009–1 was given three times a week (M/W/F) intravenously at 40 mg/kg, and SAHA was administered intraperitoneally at the dosage of 35 mg/kg 5 days per week (M-F). The treatment continued for four weeks, and 24 hours after the final dose, two animals were euthanized, and brain samples were subjected to IHC. Control animals received the vehicles of both drugs. Plasma and brain tissue samples were collected from the animals two hours after the last administration of TRX-E-009–1. Pharmacokinetic (PK) analysis of the collected samples was outsourced to Jubilant Biosys.

Excised brain tissues were fixed in formalin, embedded in paraffin, and 5 μmol/L sections were cut and mounted on Superfrost slides. The slides were dried in a 60°C oven for 1 hour and deparaffinized in xylene for 20 minutes. Slides were then re-hydrated with 100%, 90%, 70% ethanol, and MilliQ water. Antigen retrieval was performed by immersing the slides in 100 mmol/L tri-sodium citrate buffer at pH 6 and heating for 20 minutes with a cooling interval of 10 minutes following every 5 minutes of heating. The endogenous peroxidases were then inactivated by immersing the slides in 3% hydrogen peroxide. Nonspecific binding of immunoglobulin was blocked by 2.5% BSA, 2.5% FCS, and 0.2% Triton-X for 1 hour at room temperature. Slides were then incubated with primary antibodies [SSRP1 (1:1,000; Origene, #TA308461), SUPT16H (1:1,000; Thermo Fisher Scientific, #MA5–27214), Ki67 (1:1,000; Abcam, #ab209897), H3K27me3 (1:500; Abcam, #ab6002), H3K27ac (1:500; Abcam, #ab177178), anti-gamma H2A.X (1:1,000; Abcam, #ab11174)] at 4°C overnight followed by incubation with secondary antibodies for 1 hour at room temperature. Finally, the slides were washed in MilliQ water and dehydrated sequentially with 70%, 90%, and 100% ethanol.

All statistical analyses were performed using GraphPad Prism 6 software. For in vitro and IHC experiments, one-way or two-way ANOVAs (Tukey test and Dunnett multiple comparison test) were used. A Log-rank (Mantel–Cox) test was employed for survival analyses of orthotopic xenografts. Images were analyzed using ImageJ Software (ImageJ). P values <0.05 were considered statistically significant.

The data are available in the Article, Supplementary Information, or available from the authors upon request.

To determine whether targeting tubulin polymerization is an effective strategy against DIPG, TRX-E-009–1 was tested in cytotoxicity assays against six DIPG patient-derived cultures. TRXE-009–1 potently decreased DIPG cell viability, with IC₅₀ value ranging from 0.1 to 0.8 μmol/L, whereas minimal effects were observed on human fibroblast cells (MRC5) and normal human astrocytes (NHA) indicating wide therapeutic index (Fig. 1A). Likewise, in a soft agar colony formation assay on three DIPG cultures, TRX-E-009–1 potently reduced clonogenicity at submicromolar concentrations (Fig. 1B).

Given that radiotherapy is currently the only effective treatment for DIPG, we next sought to determine if the effect of TRX-E-009–1 could be further enhanced when combined with irradiation. When HSJD-DIPG007 cells were treated with either 0.5 or 1 μmol/L TRX-E-009–1 alone, or in combination with 2, 4, and 6 Gy irradiation, a significant and synergistic decrease in colony formation was observed for all combinations. TRX-E-009–1 combined with irradiation at 2 and 4 Gy synergistically reduced colony numbers indicated by CI values <1, whereas combination with 1 μmol/L TRX-E-009–1 and 6 Gy abrogated colony formation altogether (Fig. 1C). Together, these results show that TRX-E-009–1 is highly potent against DIPG neurosphere cultures and that this effect is enhanced through combination with irradiation.

Previous studies have shown that TRX-E-009–1 induces mitotic arrest in various cell lines (19). Upon treating HSJD-DIPG007 cells with TRX-E-009-1, we observed a marked increase in p21 protein—a marker associated with linking DNA damage to cell-cycle arrest, suggesting that the cell-cycle arrest in DIPG cells was mediated by p21 over-expression (Fig. 2A; Supplementary Fig. S1).

To determine whether the cytotoxic effect of TRX-E-009–1 is due to the induction of apoptosis in DIPG cells, we examined the protein levels of two key regulators of apoptosis. Treatment of HSJD-DIPG007 cells with 1 and 5 μmol/L TRX-E-009–1 caused a significant increase in cleaved caspase-3 and cleaved PARP (Fig. 2A). Together, these results show that TRX-E-009–1 reduces DIPG cell proliferation and triggers apoptosis.

TRX-E-009–1 has been shown to exert its cytotoxic activity through disruption of the microtubule network (19). Therefore, we evaluated the effect of TRX-E-009–1 treatment on the relative levels of polymerized (insoluble) and depolymerized (soluble) α-tubulin in HSJD-DIPG007 cells (Fig. 2B). In untreated cells, the ratio of polymerized to depolymerized α-tubulin was roughly 0.7, whereas with TRX-E-009–1 treatment (1, 5, 10 μmol/L), this ratio decreased to 0.1, 0.3, and 0.1 respectively.

These results were further confirmed in a cell-free assay measuring tubulin polymerization, where TRX-E-009–1 treatment visibly reduced the polymerization of tubulin compared with control. Reference compounds, paclitaxel (known to stimulate tubulin polymerization) and colchicine (known to bind to tubulin thus preventing polymerization) were also tested (Supplementary Fig. S2).

Because flavonoids have epigenetic activities including histone modification and changing DNA methylation patterns in various cancers (25–27), we evaluated the effect of TRX-E-009–1 on H3K27me3 and H3K27ac proteins. TRX-E-009–1 increased the levels of both the acetylated and methylated forms of H3K27 in HSJD-DIPG007 cells (Fig. 2C). Given the potent cytotoxic activity of TRX-E-009–1, we subsequently evaluated its antitumor efficacy against an orthotopic patient-derived xenograft model of HSJD-DIPG007. Mice were treated at 4 weeks postimplantation with 50 mg/kg TRX-E-009–1 intravenously three times a week (M,W,F) for 4 weeks (Fig. 3B). TRX-E-009–1 treatment significantly extended survival from 72 to 78 days, P = 0.0013 (Fig. 2D). Analysis of brain samples from Balb/C nude mice xenografted with the same HSJD-DIPG007 cells, 2 hours after the last treatment of TRX-E-009–1, revealed good penetration of TRX-E-009–1 drug in the brainstem region at a dose of 1.07 μmol/L (Fig. 2E). The concentration of drug in the plasma was approximately 10-fold higher, with a brain-to-plasma ratio of 0.1 (Supplementary Fig. S3).

Because TRX-E-009–1 affects H3K27 acetylation and trimethylation, we hypothesized that HDAC inhibitors would further potentiate the activity of TRX-E-009–1. We chose the class I HDAC inhibitor SAHA because it has activity against the HDAC6 isoform, which promotes hyperacetylation of α-tubulin (28, 29). Treatment of HSDJ-DIPG007 cells with SAHA (10 μmol/L) reduced HDAC6 expression by 50% and significantly increased hyperacetylation of α-tubulin (****, P < 0.0001; Supplementary Fig. S3). Consistent with our hypothesis, the combination of TRX-E-009–1 and SAHA synergistically inhibited cell proliferation and reduced clonogenicity against DIPG cultures, as indicated by combination indexes (CI; Fig. 3; Supplementary Fig. S4).

We next evaluated the effect of combining TRX-E-009–1 and SAHA on apoptosis and epigenetic markers. Combination treatment significantly increased the cleavage of both PARP and caspase-3 in HSJD-DIPG007, indicating induction of apoptosis (Fig. 4A). In addition, the combination also markedly increased H3K27 acetylation and H3K27 trimethylation, compared with single-agent treatment (Fig. 4B).

We next sought to determine if the addition of irradiation could further enhance the effect of the combination of TRX-E-009–1 and SAHA. Although colony formation of HSJD-DIPG007 cells in vitro was not affected by treatment with 0.005 μmol/L TRX-E-009–1 or 0.01 μmol/L SAHA alone, combining irradiation with TRX-E-009–1 or SAHA reduced colony numbers compared with control, with a CI value of 0.267. Furthermore, the triple combination of 2 Gy irradiation, TRX-E-009–1 and SAHA further inhibited colony formation, compared with TRX-E-009–1 and SAHA (*, P < 0.05; CI = 0.328; Fig. 5A).

The triple combination was then assessed in vivo using an orthotopic mouse model of patient-derived HSJD-DIPG007 cells. Compared with the survival of vehicle-treated animals (61 days), irradiation alone had no significant effect (63 days). Although TRX-E-009–1 (67 days) and SAHA (64 days) as single agents prolonged survival (P = 0.0007 and 0.0212, respectively), the antitumor efficacy of both TRX-E-009–1 (68 days) and SAHA (69 days) was increased when combined with irradiation (P = 0.0518 and 0.0106, respectively). Consistent with the in vitro data, the triple combination of TRX-E-009–1, SAHA and irradiation further prolonged survival (91 days; Fig. 5B). Furthermore, histologic analysis revealed a significant reduction in the proliferation of tumor cells in the treatment groups and an increase in H3K27ac, H3K27me3, and γ-H2AX in the triple combination group compared with the other cohorts (Fig. 6; Supplementary Table S1).

This study has demonstrated that targeting microtubule polymerization with a BBB permeable compound TRX-E-009–1 is a potential novel therapeutic strategy for DIPG. Our results show for the first time that this anti-microtubule agent inhibits cell proliferation and colony formation in DIPG. Anti-microtubule agents can result in either tubulin stabilization or destabilization (12). We found potent tubulin depolymerization in DIPG by TRX-E-009–1 alongside induction of apoptosis. These data clearly identify tubulin as a valid molecular target in DIPG and suggest that the use of TRX-E-009–1 represents a potentially valuable therapeutic approach in this disease.

Previous work with TRX-E-009–1 and other similar microtubule polymerization inhibiting compounds have shown links to additional mechanisms of action, including a significant increase in mitochondrial reactive oxygen species (ROS) production (30, 31). Targeting mitochondria has been a focus of interest in DIPG to promote cytotoxic ROS in tumor cells selectively (2). Furthermore, studies have shown that these compounds can also affect cell-cycle regulation and apoptosis, as well as modulate the activity of other signaling pathways (32, 33). These findings suggest that these compounds may have potential as multitargeted therapies, with the ability to simultaneously target multiple pathways involved in cancer progression. Further research is needed to fully understand the mechanisms of action of these compounds and to determine their potential as effective treatments for cancer.

The current standard therapy for DIPG is irradiation. Adding irradiation to the TRX/SAHA combination greatly extended survival in this aggressive tumor model, from 72 to 91 days.

The in vivo efficacy exhibited in this study shows, for the first time, that the combined use of TRX-E-009–1 and SAHA significantly enhances survival in an aggressive orthotopic model of DIPG. We have demonstrated enhanced efficacy of epigenetic modifying agents with irradiation, however, given that DIPG is an extremely aggressive brain tumor, additional combinatorial studies are warranted to enhance further effectiveness that ultimately may lead to greater benefit for the patients.

In conclusion, our findings provide the first evidence that combining a tubulin polymerization inhibitor with a HDAC inhibitor represents a novel therapeutic approach for DIPG. The cytotoxic effects are mediated through G₂–M arrest, depolarized tubulin and subsequent apoptosis, and the dual targeting approach was found to combine effectively with irradiation in vitro and in vivo. TRX-E-009–1 is not currently in clinical development, however these results provide the proof of principle for this class of compounds—that is, targeting microtubule polymerization with a BBB permeable agent represents a new potential treatment strategy. Both anti-microtubule agents (NCT01729260) and HDAC inhibitors have been tested in clinical trials for DIPG (34). Although limited efficacy has been seen to date with single agents, our data suggest that the combination of these therapies may represent a novel and promising combinatorial therapeutic strategy for DIPG.

D.S. Ziegler reports personal fees from Bayer, Astra Zeneca, Novartis, Day One, FivePhusion, Amgen, Alexion, and Norgine; grants and personal fees from Accendatech outside the submitted work. No disclosures were reported by the other authors.

A. Ehteda: Formal analysis, investigation, methodology, writing–original draft. A. Khan: Formal analysis, investigation, writing–original draft, writing–review and editing. G. Rajakumar: Investigation, methodology. A.S. Vanniasinghe: Investigation. A. Gopalakrishnan: Investigation. J. Liu: Investigation. M. Tsoli: Resources, supervision, writing–original draft, writing–review and editing. D.S. Ziegler: Resources, supervision, writing–original draft, writing–review and editing.

We would like to thank A/Prof Monje and Dr Angel Montero Carcaboso for generously supplying the SU-DIPG and HSJD-DIPG cells respectively, and David Brown from Novogen for providing TRX-E-009–1. We also thank Han Shen, Laura Franshaw, Swapna Joshi, Cecilia Chang, Eleanor Ager, along with the Scientific Services Group and Flow Cytometry at UNSW for their technical support and advice. This research was supported by the Kids Cancer Project, a Cancer Institute NSW Translational Program Grant (2019/TPG2037), a NHMRC Investigator grant (GNT2017898), and a NHMRC project grant (2014/GNT1085411).

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.