Lay Summary

The survival time for patients with pediatric high-grade glioma (pHGG) remains devastatingly short, in large part because pHGG tumors become resistant to radiation, the only available therapy with confirmed efficacy. This radioresistance is triggered by regions of hypoxia (low oxygen) that develop within the tumor as it grows. We found that dopamine receptor (DRD2) antagonists dramatically reduce mitochondrial metabolism in adult radio-resistant glioma cells, thereby increasing oxygen levels and restoring radio-sensitivity. Combining radiation with DRD2 antagonist treatment substantially improved survival in mice with radioresistant tumors created from adult patient cells. Importantly, DRD2 antagonists reach the brain but do not damage healthy brain cells. In fact, several DRD2 antagonists are FDA-approved for use as antipsychotic agents in children. Moreover, the recently discovered small molecule ONC201, a DRD2 antagonist drug candidate that also reduces glioma growth by inhibiting mitochondrial metabolism, is being tested for pediatric H3K27M glioma in a clinical trial. However, pHGG tumors vary greatly in phenotype, which often determines the efficacy of treatments. We will examine a panel of pHGG cells obtained from patients to identify phenotypes that are susceptible to DRD2 antagonists, with and without radiation, in cell culture and in mice with tumors created from these patient cells.

The need to discover therapeutics for pHGGs is enormous. Repurposing of FDA-approved drugs is one of the most promising and readily available options for these tumors. DRD2 antagonists appear to be ideal candidates for radiosensitization of pHGG, considering that these agents 1) are already FDA approved and very well tolerated by young children and 2) efficiently cross the blood-brain barrier, which is key for effective treatment of brain tumors. Interestingly, our preliminary studies indicate that DRD2 antagonists not only sensitize glioma cells to radiotherapy but even protect normal astrocytes from consequential damage caused by this treatment. Identifying tumors with mitochondrial-dependent cancer metabolism that predicts the greatest benefit from this targeted therapy will be important for patient stratification in future clinical trials. Overall, we expect to rapidly translate our findings to clinical trials of DRD2 antagonists and radiation therapy for pHGG and thus benefit children with pHGG.

Executive Summary

Central nervous system (CNS) tumors constitute the second most common cancer in children and adolescents, and approximately half of these tumors are gliomas. Although pediatric high-grade gliomas (pHGGs) are rare (8-12% of all primary CNS tumors in children), overall survival (OS) for children with pHGG remains devastatingly short at 9-15 months, despite treatment (1-3). Recent reports indicate that pHGG tumors are a range of distinct entities that might require differential treatment based on specific genetic and epigenetic features (4-8). For example, histone H3 K27M mutations are frequent in diffuse midline glioma (DMG), a pHGG that arises in the brainstem, while histone H3 pG34R/V mutations occur mainly in non-brainstem (NBS) pHGGs . Such markers are being investigated as therapeutic targets; however, little is known about the global metabolomic alterations that characterize pHGG progression. Metabolites define the phenotype of all cells; accordingly, metabolite levels provide the most telling readout of cell function and dysfunction (9,10). The complex relationship of the metabolome to phenotype is poorly understood in pHGGs but must be addressed to exploit them therapeutically.

Our laboratory has made several pivotal discoveries regarding the metabolic and bioenergetics pathways involved in chemo- and radio-resistance in adult glioma. Our analysis of GBM, the most aggressive glioma in adults revealed that 20-25% of GBM tumors are intrinsically resistant to chemo- and radiotherapy (11). Moreover, cells within these tumors are highly dependent on oxidative phosphorylation (OXPHOS) for survival, and this phenotype is regulated by an isoform switch in one subunit of cytochrome c oxidase (CcO), a crucial mitochondrial electron transport chain (12) complex. Specifically, expression of CcO subunit 4 isoform 1 (COX4-1) increases CcO activity and causes GBM cells to become more aggressive and resistant to treatment (13-16). We are currently conducting a prospective multi-center biomarker trial of CcO in GBM, under the NeuroNext infrastructure (NCT02997423). Significant to this proposal, we have identified FDA-approved dopamine receptor D2 (DRD2) antagonists that block CcO activity and are effective against OXPHOS glioma cells in vitro and in vivo (17) and improve their sensitivity to therapy. Importantly, in vitro treatment with DRD2 antagonists enhances radiation-induced toxicity in OXPHOS tumor cells but protects normal human astrocytes from radiation damage.

The recently discovered small molecule DRD₂ antagonist ONC201 is effective against high-grade gliomas in preclinical models. In addition, radiographic regressions have been reported in adults with recurrent H3 K27M-mutant glioma treated with single agent ONC201. Interestingly, ONC201 efficacy is mainly mediated by inhibition of mitochondrial respiration. Moreover, ONC201 is only effective in OXPHOS-dependent cancer cells; cells dependent on glycolysis, were resistant to ONC201 (18).

Our analysis of RNA-sequencing libraries from >250 pHGG tissues/cell types revealed that the COX4i1 transcript is the most commonly upregulated gene in NBS pHGGs but not DMG. Overall, gene signatures reflecting high levels of OXPHOS metabolism and mitochondrial transcriptional activity are strongly expressed in NBS pHGGs but weakly expressed in DMGs.

Metabolic radiosensitization is a therapeutic concept that targets the metabolic demand for O₂ by downregulating mitochondrial oxidative metabolism. DRD2 antagonists appear to be ideal candidates for clinical metabolic radiosensitization of pHGG considering that: 1) many DRD2 antagonists are already FDA approved; 2) they show no considerable toxicity, and 3) they do effectively cross the blood-brain barrier. However, the key to the use of such a radiosensitizer in children will be identifying those tumors with OXPHOS-dependent cancer metabolism that would be predicted to benefit the most from this targeted therapy.

We hypothesize that inhibition of mitochondrial metabolism by DRD2 antagonists increases tumor oxygenation and radiation response in OXPHOS pediatric high-grade gliomas (pHGG).

In this application, we propose to validate high CcO activity/COX4-1 expression as a biomarker of response to DRD2 antagonist therapy in pHGGs. Using a panel of previously characterized patient-derived tumor xenograft cells representing the major pHGG subtypes, including non-brainstem (NBS) and diffuse midline gliomas (DMG), we will:

Aim 1: Determine the metabolic profile, mitochondrial complexes activity level, DRD₂ and COX4 isoforms expressed in pHGG tumors and determine the susceptibility of these tumors to DRD₂ antagonists. (Timeline: months 0 to 6)

The objective of this aim is to define the metabolic phenotype(s) of pHGG and to determine how phenotype affects therapy sensitivity, outcome and susceptibility to DRD₂ antagonists as single agents or in combination with radiation. We hypothesize that OXPHOS metabolism (indicated by high CcO activity/COX4-1 expression as a companion biomarker) renders NBS pHGG cells sensitive to DRD₂ antagonists ± radiation therapy. 

We will determine mitochondrial complexes activity and O₂ consumption by spectrophotometric analysis, high-resolution respirometry and using the XF24 analyzer. Protein expression will be determined by Western blot analysis. For metabolic profiling, we will perform LC/MS analysis of cellular metabolites and stable-isotope tracing (in collaboration with Dr. Eric Taylor, Director of the Metabolomics Core facility). For cell viability testing, we will use two DRD₂ antagonists already approved for pediatric use, and ONC201, a drug candidate currently in a clinical trial for use in children. The IC₅₀ for each drug and cell line will be determine. To assess cell proliferation and cell death, we will perform flow cytometry analysis of 5-ethynyl-2′-deoxyuridine (EdU) incorporation and Annexin V/propidium iodide staining. To test the effect of the drugs as radiosensitizers, drug-treated cells will be exposed to radiation and results will be evaluated by colony-forming assays. To investigate the O₂ dependence, cell viability test  will be studied in aerobic (20% O₂) and hypoxic (1%-5% O₂) states.

Expected Outcomes. Results from Aim 1 should: 1) allow us to determine the expression of the potential biomarker in pHGG subtypes;  2) provide a rigorous understanding of pHGG tumor metabolism. We expect to identify at least two metabolic subgroups, low-OXPHOS (probably DMG) and high-OXPHOS (probably NBS) pHGGs. Low-OXPHOS will be defined by glycolytic metabolism, and high-OXPHOS will be defined by reliance on OXPHOS, supported by glutamine and fatty acid oxidation. It will also provide the first insight into the connection between CcO activity/COX4 isoform expression, metabolic flexibility, and tumorigenicity between subtypes. Identifying metabolic biomarkers and pathway alterations in pHGGs may provide novel insights into the metabolic vulnerabilities that can be targeted therapeutically; 3) provide the first insight into the connection between CcO activity, COX4 isoform and DRD₂ expression, with the susceptibility to DRD₂ antagonists and radiation therapy in pHGG. We expect that some of the proposed drugs will show synergistic activity with radiation in killing high CcO activity/COX4-1⁺ NBS pHGG cells. The results of testing ONC201 as a radiosensitizer will be of particular interest, as this drug has been tested only in pediatric patients with previously irradiated tumors (NCT03416530). Aim 1 will also determine if DRD₂ antagonists protect NHA from radiation damage.

Aim 2: Test the therapeutic efficacy of DRD₂ antagonists in combination with radiation in patient-derived pHGG xenografts in orthotopic mouse models. (Timeline: months 4 to 12)

The objectives of this aim are to 1) validate CcO activity/COX4-1 expression as therapeutic targets for the anti-proliferative, radiosensitizing effect of DRD₂ antagonists in pHGG; 2) evaluate the in vivo potential of DRD₂ antagonists as anti-pHGG therapeutic agents alone or in combination with SAARP and 3) to determine the effects of DRD2 antagonists on hypoxia levels in intracranial tumors.

We hypothesize that DRD₂ antagonists will increase tumor oxygenation, reduce intracranial tumor burden and prolong survival of mice bearing OXPHOS NBS pHGGs PDX.  

We will use different luciferase-cell models with different levels of CcO activity conferred by the presence/absence of COX4-1. Whole-brain bioluminescence will be measured for each mouse and mice will be sorted into Control and Treated groups of equal mean bioluminescent signal (12 mice per group will provide 80% power and a 5% significance level to detect the difference in mean OS). Treatment with DRD₂ drugs by i.p. injection will be initiated 2 weeks after tumor cells implantation. Tumor growth and development will be monitored for luciferase expression. Mice will be observed daily and euthanized when they reach a moribund state. 

To determine the efficacy of DRD₂ antagonists in combination with local external beam irradiation in animal models we will use the recently acquired small animal radiation research platform (sarrp), which can focus a beam of radiation with an accuracy of 0.2 mm, recreating confocal beam therapy for humans on the scale of a mouse. Mice bearing intracranial pHGGs will be treated with (a) saline, (b) DRD2 antagonists, (c) 12 Gy (2 Gy/fraction, 3 times per week for 2 weeks); or (d) DRD2 antagonists plus radiation. Tumor growth and anti-tumor efficacy data based on mouse survival/death will be determined. At the end of the study, brains will be harvested and collected in formalin or snap-frozen in liquid nitrogen. Brain tissue will be processed for immunohistochemistry and assessed for apoptosis, proliferation and hypoxia markers. For detection of hypoxic tumor cells, we will use the 2-nitroimidazole(EF5). EF5 is an imaging marker that selectively binds to hypoxic cells and forms adducts that can be detected with a specific antibody conjugated with Alexa Fluor 488. At sacrifice, brains will be collected, frozen in OCT, and stored at -80°C before sectioning. After sectioning, slices will be stained with anti-EF5 and Alexa Fluor 488 conjugated antibodies. A fluorescence microscope at 20× magnification will be used to visualize hypoxic regions, and ImageJ software will be used for calculations.

All animal studies will be performed with the approval of the UIowa-IACUC. Survival curves will be generated by the Kaplan-Meier method and analyzed by log-rank test. Two-way categorical comparisons will be performed using Fisher's exact test. All tests will be two-sided and P<0.05 will be considered statistically significant. This Aim will be performed in collaboration with Dr. John Buatti and Dr. Mariko Sato, pediatric brain tumor specialists involved with the children’s oncology group-COG.

Expected Outcomes. Results from Aim 2 should allow us to 1) determine the therapeutic benefit of DRD₂ antagonists against pHGG in relevant animal models; and 2) provide information on the specificity of DRD₂ antagonists for CcO/COX4-1⁺ cells, as indicated by remission or stable disease (increased OS) in mice bearing biomarker-positive tumors and lack of effect or partial remission in mice bearing biomarker-negative tumors; 3) determine the efficacy of DRD₂ antagonists as radiosensitizing agents; and 4) determine the effects of DRD₂ antagonists on tumor oxygenation in relevant animal models.

Clinical significance: O₂ tension plays a critical role in the response to radiation therapy (RT). Here we propose that hypoxic tumors can be sensitized to radiation therapy by targeting mitochondrial respiration. We identified FDA-approved DRD2 antagonists as mitochondrial CcO inhibitors. A single dose of the drug prior to RT reduces O₂ consumption and significantly enhances the tumor response to radiation.  Surprisingly, we observed a radioprotective effect on well-oxygenated normal human astrocytes.  Because of the absence of any meaningful therapy, repurposing of drugs is one of the most promising and readily available options for pHGGs.

This proposal is significant because: 1)  we offer a novel and exciting biomarker-driven drug-repurposing strategy to improve survival outcomes for children with pHGG; 2) it will delineate the metabolic profiles of pHGGs that may selectively contribute to tumor resistance to therapy. Specifically, we will validate CcO activity/COX4-1 expression as a biomarker for pHGG response to therapy; 3) it will examine the benefits of using DRD₂ antagonists to improve the sensitivity of pHGG tumor cells to radiotherapy while minimizing damaging sequelae and 4)  if successful, results from this study will inform the immediate design of a biomarker-driven clinical trial in children with pHGG tumors under the umbrella of the Pediatric Early Phase-Clinical Trial Network (PEP-CTN).

Description of Research Proposal

Budget

Collaborations and Conflicts of Interest