Reactive oxygen species (ROS) are essential second-messenger molecules, yet when deregulated, they fuel cancer growth and therapeutic resistance. In high-grade gliomas, including glioblastoma, diffuse hemispheric glioma, and diffuse midline glioma (DMG), genetic, epigenetic, and metabolic alterations drive chronic ROS production and redox imbalance. This oxidative stress promotes DNA damage, epigenetic reprogramming, tumor growth, and immune escape. In DMG, global DNA and histone hypomethylation are amplified by oxidative stress, while ROS-dependent Ras/Raf/mitogen-activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) signaling pathways reinforce tumor survival. Paradoxically, the same ROS create an intrinsic vulnerability as excess ROS can overwhelm defenses and trigger cytotoxicity. Targeting ROS is challenging; however, new strategies, including NADPH oxidase inhibition, metabolic modulation, and ROS-inducing therapies, reveal vulnerabilities. Understanding this redox paradox is critical to exposing therapeutic vulnerabilities and improving outcomes for patients with these deadly cancers.

Abstract Background Diffuse midline glioma (DMG), including tumors originating in the brainstem known as diffuse intrinsic pontine glioma (DIPG), are childhood and adolescent brain tumors that remain uniformly fatal despite extensive genomic and epigenomic characterization. Their unique anatomy, genetic heterogeneity, cell states, immunosuppressive microenvironment, and variable drug penetration within the midline structures create a disease that is biologically complex and difficult to treat with current therapeutic strategies. Methods Spatial multi-omics technologies have emerged as powerful tools for addressing these challenges, particularly given the limited amount of tumor tissue available from biopsy. These methods require minimal material while preserving anatomical context, enabling simultaneous profiling of thousands of transcripts, proteins, and metabolites within a single tissue section. Such approaches can profile tumor cell states, immune interactions, and pharmacological gradients within their native tissue architecture. Results Building on discussions from the 2025 ChadTough Defeat DIPG Research Workshop, we outline a practical roadmap for implementing spatial transcriptomics, spatial proteomics, imaging-based mass spectrometry, and related spatial approaches across the DMG disease course, from diagnostic biopsy through autopsy. Key considerations are discussed, including tissue triage, spatial platform selection, quality control standards, analytical workflows, and data governance frameworks needed to support collaborative analysis. Conclusions Spatial multi-omics offers an opportunity to link tumor biology, microenvironmental features, and therapeutic exposure at cellular or near-cellular resolution. Establishing shared standards and coordinated international frameworks will be essential to translate these technologies from discovery research into clinically meaningful applications for patients with DMG and DIPG. Importance of the study For lethal tumors such as DMG, including DIPG, genetic heterogeneity, anatomical location, diverse malignant cell states, and a profoundly immunosuppressive tumor microenvironment means that no single therapy is likely to achieve durable survival. Understanding how these factors interact within the spatial architecture of the tumor is therefore essential for developing more effective therapeutic strategies. Spatial multi-omics technologies enable high-resolution mapping of tumor biology directly within intact tissue, allowing researchers to identify cellular states, immune interactions, and regional differences in drug exposure that are not captured by bulk molecular profiling approaches. These approaches are particularly valuable in DMG, where biopsy tissue is scarce. By outlining practical considerations and a coordinated framework for implementing spatial multi-omics throughout the DMG disease course, this review highlights how these technologies can accelerate biological discovery and support the development of rational combination therapies for patients with these devastating tumors.

Pediatric high-grade glioma (pHGG), including diffuse midline glioma (DMG), are the most aggressive and fatal pediatric cancers. Mutations and amplifications within the phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3K) pathway drive tumor growth, treatment resistance, and poor outcomes. Although PI3K and mTOR have been identified as genetic dependencies in pHGGs, translating this knowledge into effective treatment remains challenging. The blood-brain barrier (BBB) restricts the delivery of most PI3K/mTOR inhibitors and , hence, often show poor CNS penetration. Even when present in the brain, these agents frequently encounter adaptive resistance mechanisms that blunt efficacy. Side effects, including hyperglycemia, rash, and mucositis, further complicate their use and reduce compliance. Encouragingly, novel brain-penetrant PI3K/mTOR inhibitors offer new opportunities for treatment, but combining these agents with other therapies, including chemotherapy, other small molecules, and immunotherapies, requires careful balancing of toxicity and efficacy. Therefore, achieving optimal dosing for each patient remains a significant hurdle. This review examines the promise and pitfalls of targeting the PI3K/mTOR pathway in DMG, including the limitations of available therapies, mechanisms of resistance, and the critical need for improved regimen design. We propose a roadmap to guide future efforts, emphasizing rational combination strategies and better patient stratification to improve survival for children diagnosed with these devastating cancers.

Diffuse midline glioma (DMG) is a universally fatal high-grade glioma. Its immunologically cold tumor immune microenvironment (TIME) presents a major barrier to durable anti-tumor responses. Dordaviprone (ONC201) is a brain-penetrant DRD2 antagonist and ClpP agonist that disrupts mitochondrial complex II (CII) activity; currently under clinical evaluation for DMG. In melanoma, CII inhibition has been shown to enhance MHC presentation and T cell–mediated killing. Here, we investigated the immunomodulatory effects of dordaviprone in DMG mouse models, patient tumor tissues, and blood samples. Immunocompetent DMG-PPK mice (PdgfraD842V, Trp53DN, H3f3aK27M mutations) were treated with dordaviprone (125 mg/kg, twice weekly) for 1 week (acute) or 3 weeks (chronic) and compared to naïve, sham (orthotopic surgery without tumor engraftment), and vehicle-treated controls. Murine tumors were analyzed by single-cell multiome (scRNA-seq and scATAC-seq) and spatial transcriptomics. Murine blood and bone marrow (BM) and human blood samples were profiled by scRNA-seq and flow cytometry. In models, dordaviprone promoted microglial activation and macrophage infiltration within the TIME. Spatial transcriptomics identified increased expression of the MHC-I subunit B2m in tumor cells and tumor-infiltrating Cd8+ T cells following dordaviprone treatment. Flow cytometry of murine and human peripheral blood confirmed CD4+ and CD8+ lymphopenia at baseline. Dordaviprone increased circulating lymphocytes in models (log2 fold change=1.03; patient data pending), reversing DMG-induced CD3+ T cell sequestration in the BM. Immunohistochemistry confirmed elevated levels of B2M (p=0.0281) and tumor-infiltrating lymphocytes (TILs), including CD45+ (p=0.0275) and CD3+ T cells (p=0.0103) in dordaviprone-treated mice. Despite enhanced TIL recruitment, dordaviprone did not improve survival (n=10), likely due to persistent immunosuppressive myeloid activity, i.e. Havcr2 (TIM-3), Cd74, Cd274 (PD-L1). This study demonstrates that DMG tumors contribute to systemic lymphopenia, partially reversed by dordaviprone. While supporting its immunomodulatory effects, including enhanced antigen presentation and TIL recruitment, our findings highlight future combination strategies to leverage dordaviprone-induced immunoreactive TIME.

Alterations in the FMS-like tyrosine kinase 3 (FLT3) gene are the most frequent driver mutations in acute myeloid leukaemia (AML), linked to a high risk of relapse in patients with internal tandem duplications (FLT3-ITD). Tyrosine kinase inhibitors (TKIs) targeting the FLT3 protein are approved for clinical use, yet resistance often emerges. This resistance is mainly seen following the acquisition of additional point mutations in the tyrosine kinase domain (TKD), resulting in a double mutant FLT3-ITD/TKD, which sustains cell signalling and survival despite the presence of FLT3 inhibitors. Here, we developed a FLT3-mutant AML model with adaptive resistance to type II TKIs, sorafenib, and quizartinib by in vitro drug selection. Through global multiomic profiling, we identified upregulation of proteins involved in reactive oxygen species (ROS) production, particularly NADPH-oxidases, driving cellular 'ROS-addiction', with resistant cells relying on ROS for survival, and genome fidelity preserved by ATM-driven DNA repair. Transcriptomic analysis of adult and paediatric AML (pAML) patients identified high ATM expression as a biomarker for shorter median overall survival in both the de novo and relapsed settings. Inhibition of ATM with clinically relevant therapy WSD-0628 effectively killed TKI- and chemotherapy-resistant AML cells in vitro and significantly extended the survival of mice with sorafenib- and quizartinib-resistant FLT3-ITD AML in vivo. We propose a new treatment strategy to improve survival of patients who develop resistance to sorafenib and quizartinib, as well as relapsed and refractory pAML, exploiting resistance mechanisms to precision therapies and cell-intrinsic features of high-risk cases, highlighting a clinically relevant salvage strategy. ### Competing Interest Statement The authors have declared no competing interest.

Diffuse midline glioma (DMG) is a universally fatal CNS tumor, with a median overall survival of <12-months. Recent studies, including our own, have identified overexpression of the mitochondrial protease ClpP in DMG patient samples. ClpP plays a pivotal role in regulating energy production. Brain-penetrant ClpP agonists, including ONC201, ONC206, and TR107, induce non-specific degradation of the mitochondrial electron transport chain. However, there is considerable variability in the response of DMG cell lines to these therapies, with only about half reaching IC50. Notably, TP53-mutant cell lines exhibit reduced sensitivity. Further analysis using quantitative proteomics of TP53-mutant DMG cell lines exhibited hallmark activation of ClpP, increased mitochondrial stress and dysfunction. Importantly, ClpP agonism increased the activity of the transcription factor NRF2, driving expression of NQO1 (p=0.0017). NQO1 is a cytoprotective enzyme, promoting redox balance and cell survival. This highlights a potential adaptive response to ClpP agonist therapy in DMG cells. Analysis of DMG patient samples through publicly available RNA sequencing data identified ubiquitously high NRF2 expression compared to other pediatric CNS tumors. This led us to hypothesize that a clinically relevant, selective NQO1 inhibitor could elicit a strong response to ClpP agonism. Therefore, we examined ARQ761, a CNS-active therapy that disrupts NQO1 leading to the formation of an unstable hydroquinone, causing oxidative DNA damage and apoptosis. When ARQ761 was used concurrently with a ClpP agonist, the combination showed additive effects, average Bliss Score of 9.91 (n=4). Interestingly, administering ARQ761 24 h after ClpP agonist exposure combined synergistically, average Bliss Score of 15.87 (n=4). This synergy was further potentiated using ARQ761 48 h after ClpP agonism, when NQO1 expression peaked, resulting in an average Bliss Score of 20.30 (n=4). This study has unveiled an intrinsic antioxidant defense mechanism in DMG, which facilitates evasion from treatments inducing oxidative stress. This defense mechanism appears particularly robust in cell lines harboring TP53-mutations, which are also known to exhibit resistance to radiotherapy.

Diffuse midline glioma (DMG), including tumors diagnosed in the brainstem (diffuse intrinsic pontine glioma - DIPG), is the primary cause of brain tumor-related death in pediatric patients. DIPG is characterized by a median survival of <12 months from diagnosis, harboring the worst 5-year survival rate of any cancer. Corticosteroids and radiation are the mainstay of therapy; however, they only provide transient relief from the devastating neurological symptoms. Numerous therapies have been investigated for DIPG, but the majority have been unsuccessful in demonstrating a survival benefit beyond radiation alone. Although many barriers hinder brain drug delivery in DIPG, one of the most significant challenges is the blood-brain barrier (BBB). Therapeutic compounds must possess specific properties to enable efficient passage across the BBB. In brain cancer, the BBB is referred to as the blood-brain tumor barrier (BBTB), where tumors disrupt the structure and function of the BBB, which may provide opportunities for drug delivery. However, the biological characteristics of the brainstem's BBB/BBTB, both under normal physiological conditions and in response to DIPG, are poorly understood, which further complicates treatment. Better characterization of the changes that occur in the BBB/BBTB of DIPG patients is essential, as this informs future treatment strategies. Many novel drug delivery technologies have been investigated to bypass or disrupt the BBB/BBTB, including convection enhanced delivery, focused ultrasound, nanoparticle-mediated delivery, and intranasal delivery, all of which are yet to be clinically established for the treatment of DIPG. Herein, we review what is known about the BBB/BBTB and discuss the current status, limitations, and advances of conventional and novel treatments to improving brain drug delivery in DIPG.

Diffuse midline gliomas (DMGs) are devastating pediatric brain tumors recognized as the leading cause of cancer-related death in children. DMGs are high-grade gliomas (HGGs) diagnosed along the brain's midline. Euchromatin is the hallmark feature of DMG, caused by global hypomethylation of H3K27 either through point mutations in histone H3 genes (H3K27M), or by overexpression of the enhancer of zeste homolog inhibitory protein. In a clinical trial for adults with progressive HGGs, a 22-year-old patient with a thalamic DMG, H3 K27-altered, showed a remarkable clinical and radiological response to dordaviprone (ONC201). This response in an H3 K27-altered HGG patient, coupled with the lack of response of patients harboring wildtype-H3 tumors, has increased the clinical interest in dordaviprone for the treatment of DMG. Additional reports of clinical benefit have emerged, but research defining mechanisms of action (MOA) fall behind dordaviprone's clinical use, with biomarkers of response unresolved. Here, we summarize dordaviprone's safety, interrogate its preclinical MOA identifying the mitochondrial protease "ClpP" as a biomarker of response, and discuss other ClpP agonists, expanding the arsenal of potential weapons in the fight against DMG. Finally, we discuss combination strategies including ClpP agonists, and their immunomodulatory effects suggestive of a role for the tumor microenvironment in DMG patient response.

Abstract BACKGROUND Diffuse midline glioma (DMG) is a devastating pediatric high-grade glioma (pHGG) located along the midline structures of the brain and spine. Palliative radiotherapy is the only approved treatment outside of clinical trials, with patients given 9–11-months survival post-diagnosis. Over the last 10-years, it has become common for patients to be prescribed precision therapies targeting genetic alterations, however largely, DMG harbor genetic alterations that are untargetable. METHODS To improve DMG patient outcomes, we are performing multi-omic analysis on 210 DMG, normal controls and other pHGG samples to functionally characterize the genomic landscape of these tumors and understand the protein-controlled functional consequences of recurring somatic alterations. RESULTS Genomic sequencing has been performed on 40/210 pHGG samples, identifying 290 unique mutations and several recurrent chromosomal alterations. Additionally, proteomic/phosphoproteomic profiling identified both unique and similar features of post-translational landscapes across genotypes. Examining H3.3K27M DMG cell-lines, SU-DIPG-XIII and UON-VIBE5, genomic sequencing identified 11 and 6 mutations respectively, however, only UON-VIBE5 possessed a genomically-predicted target, a PDGFRA-D842V mutation, potentially targetable by Avapritinib. Conversely, investigation of proteomes/phosphoproteomes revealed several potential therapeutic vulnerabilities. SU-DIPG-XIII possessed high enrichment of HDAC/BRD and CDK pathways and UON-VIBE5, ATM-centric pathways. In vitro proliferation assays of Avapritinib treated UON-VIBE5 failed to elicit a response at clinically relevant doses, while all proteomically-predicted therapies (HDAC – Fimepinostat, BRD – JQ1, CDK – Ribociclib and ATM – WSD-0628) successfully influenced tumor cell growth whilst leaving microglia controls cells unaffected. Avapritinib and WSD-0628 (CNS-active ATM inhibitor) drug predictions were further validated in the RA055 DMG model, which possesses similar genomic and proteomic features to UON-VIBE5. Avapritinib alone failed to influence tumor growth, while WSD-0628 notably reduced cell proliferation. The combination of these drugs, however, synergistically inhibited RA055 proliferation. CONCLUSIONS These findings thus provide preliminary evidence supporting the need to take a multi-omic approach in treating DMGs.

Abstract Diffuse midline glioma (DMG), including diffuse intrinsic pontine glioma (DIPG), are uniformly fatal brain cancers affecting children, adolescents, and young adults. These tumors are characterized by ‘oncohistone’ mutations in histone H3 genes (H3F3A, HIST1H3B, HIST1H3C), resulting in the substitution of lysine 27 to methionine (H3K27M), resulting in dominant negative hypomethylation at lysine 27 (H3K27) and a global loss of gene silencing. CRISPR-Cas9 loss-of-function gene deletion screens identified mutation-independent dependencies on PIK3CA and MTOR genes for tumor growth and proliferation, highlighting a targetable molecular dependency across DMG patient-derived models (n=38). However, systemic PI3K/mTOR inhibition increased blood glucose and insulin levels, promoting hyperinsulinemia/hyperglycemia and reduced efficacy in vivo. To exploit genetic dependencies while maintaining compliance and therapeutic benefit, we optimized combinations of clinically relevant PI3K inhibitors (paxalisib, GCT007, everolimus) with the antiglycemic drug metformin to restore glucose homeostasis and decrease DMG insulin receptor activity in vivo, extending the survival of DIPG xenograft models. Phosphoproteomic profiling of DIPG models treated with PI3K/mTOR inhibitors promoted calcium-activated PKC signaling. The brain-penetrant PKC inhibitor enzastaurin in combination with paxalisib, synergistically extended the survival of DIPG xenograft models, including those mimicking disease progression, with further benefits potentiated using metformin in a multimodality approach. Combined PI3K-PKC inhibition in vivo promoted differentiation of OPC-like DMG to more OC-like cells, enhancing PDGFRA-JAK/STAT proinflammatory signaling, showing features of demyelination and MHC-II+ antigen expression, including PD1/PDL1. In parallel, combined PI3K/mTOR and PDGFRA inhibition using paxalisib and avapritinib synergistically extended the survival of patient-derived xenograft models, showing the preclinical relevance of simultaneously targeting these vulnerabilities. Together, we reveal that therapeutic inhibition of PI3K/mTOR and compensatory PKC signaling, while mitigating side effects with metformin, altered the chromatin architecture, leading to cellular differentiation and opening the door for clinically relevant checkpoint inhibitors. This combination strategy addresses DMG genetic dependencies, cellular and systemic responses to precision therapies, while harnessing the immune system for potential clinical translation.

Abstract BACKGROUND Less than 10% of patients with diffuse midline glioma (DMG) survive 2-years post-diagnosis, with no approved treatments other than palliative radiotherapy. Immune-checkpoint inhibition has failed in the clinic, likely due to a lack of tumor infiltrating lymphocytes (TILs) and limited/no expression of immune checkpoint proteins or proinflammatory cytokines in the tumor immune microenvironment (TIME). Similarly, glioblastoma patients suffer from a ‘cold’ TIME, linked to the sequestration of T cells in the bone marrow (BM) through β-arrestin-induced internalization of S1PR1. β-arrestin is activated by Dopamine receptors (DRD2); therefore, we hypothesized that systemic DRD2 antagonism using the antagonist ONC201, might reverse lymphopenia, increasing TIL populations. METHODS Flow cytometry analysis was performed on bloods from DMG patients at diagnosis. Circulating immune cells and BM from the immunocompetent PPK model +/-ONC201 and tumor-naïve mice were analyzed by a hematology analyzer and flow cytometry. Immunocompromised SU-DIPG-VI and HSJD-DIPG-007 and immunocompetent XFM murine models were treated +/-ONC201 and tumors analyzed by immunoblotting, immunohistochemistry, and flow cytometry. scRNAseq and scATACseq analysis was performed on tumors from immunocompetent PPK models +/-ONC201. RESULTS DMG patients displayed low levels of total lymphocytes, including CD4+, CD8+ and NK subsets (n=9). Similarly, DMG-bearing mice displayed lower levels of lymphocytes (CD45+, CD4+, CD8+ subsets, n=3) compared to sham-engrafted mice. ONC201 treated mice showed lymphocyte levels comparative to sham-engrafted mice, reversing tumor-induced lymphopenia. A significant increase of T cells expressing S1PR1+ was seen in the BM following ONC201. Tumor immune cell infiltration following ONC201 was confirmed in both immunocompromised and immunocompetent models. Multiomics analysis identified ONC201 to promote pro-inflammatory polarization of microglia/macrophages (MHC II and CD86) and increase expression of immune checkpoints (VISTA and PD-L1). CONCLUSIONS DMG patients suffer from low peripheral lymphocytes, likely contributing to cold TIME. ONC201 reversed tumor-induced lymphopenia, promoting TIL recruitment within the tumor, warming the TIME.

Diffuse midline glioma (DMG), including tumors diagnosed in the brainstem (diffuse intrinsic pontine glioma; DIPG), are uniformly fatal brain tumors that lack effective treatment. Analysis of CRISPR/Cas9 loss-of-function gene deletion screens identified PIK3CA and MTOR as targetable molecular dependencies across patient derived models of DIPG, highlighting the therapeutic potential of the blood-brain barrier-penetrant PI3K/Akt/mTOR inhibitor, paxalisib. At the human-equivalent maximum tolerated dose, mice treated with paxalisib experienced systemic glucose feedback and increased insulin levels commensurate with patients using PI3K inhibitors. To exploit genetic dependence and overcome resistance while maintaining compliance and therapeutic benefit, we combined paxalisib with the antihyperglycemic drug metformin. Metformin restored glucose homeostasis and decreased phosphorylation of the insulin receptor in vivo, a common mechanism of PI3K-inhibitor resistance, extending survival of orthotopic models. DIPG models treated with paxalisib increased calcium-activated PKC signaling. The brain penetrant PKC inhibitor enzastaurin, in combination with paxalisib, synergistically extended the survival of multiple orthotopic patient-derived and immunocompetent syngeneic allograft models; benefits potentiated in combination with metformin and standard-of-care radiotherapy. Therapeutic adaptation was assessed using spatial transcriptomics and ATAC-Seq, identifying changes in myelination and tumor immune microenvironment crosstalk. Collectively, this study has identified what we believe to be a clinically relevant DIPG therapeutic combinational strategy.

Diffuse midline gliomas (DMGs) diagnosed in the pons (DIPG) are universally fatal central nervous system tumors and are the leading cause of cancer-related death in children. Palliative radiotherapy is the only recognized treatment, with median overall survival just 9-11 months. The brain-penetrant, small molecule, ONC201, shows early-stage clinical trial efficacy, extending survival by ~9-11 months compared to historic controls. However, studies to determine the mechanisms behind the temporary clinical response to ONC201 are needed. Here, we have used a systems-biological approach to investigate whether genomic features influenced ONC201 response. DMGs harboring PIK3CA mutations were more sensitive to ONC201, whereas those harboring TP53 mutations were less sensitive. Quantitative proteogenomics identified that ONC201 elicits potent agonism of the mitochondrial protease, ClpP, driving proteolysis of electron transport chain and tricarboxylic acid proteins, leading to mitochondrial dysfunction. However, metabolic adaptation to ONC201 is promoted by the spare redox-signaling capacity of cells harboring WT-PIK3CA that was counteracted using the brain-penetrant PI3K/Akt/mTOR inhibitor, paxalisib. ONC201 and paxalisib combinations extended survival of orthotopic DIPG xenograft mouse models (SU-DIPG-VI, p=0.0027; SF8626, p=0.0002; HSJD-DIPG-007, p=< 0.0001). The combination in the first three recorded patients; two at progression following re-irradiation, and one at diagnosis following the completion of radiation, resulted in dramatic reductions in tumor area, dramatically extending overall survival for all three patients (25 months, 30 and 31 months continuing). The DIPG patient receiving the combination since diagnosis, remains in progression free survival (MR axial diagnosis scan = 1554 mm2, current tumour area = 306 mm2, ~80% reduction). The patient continuing to receive the combination at progression and following reirradiation also experienced a marked decrease in tumor size (MR axial diagnosis scan = 1248 mm2, current tumour area = 315 mm2, ~75% reduction), 10 months following radiological detection of progression. These data inform the phase II clinical trial (NCT05009992).

Abstract Recurring activating mutations, amplifications and allelic loss of the negative repressors of the phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3K) signaling genes are overarching contributors to the poor survival of patients with H3K27-altered diffuse midline gliomas (DMG) diagnosed in the pons (diffuse intrinsic pontine glioma – DIPG), highlighting the clinical potential of therapies that target this axis. Employing a loss-of-function CRISPR-Cas9 gene deletion screen, we identified PIK3CA and MTOR as survival dependences in DMG/DIPG patient derived cell line models (n=36), highlighting the clinical potential of targeting PI3K/Akt/mTOR. The brain-penetrant, PI3K/mTOR inhibitor paxalisib is currently under clinical investigation in DIPG (NCT03696355 and NCT05009992), but dosing limiting toxicities including hyperglycaemia, mucositis and skin irritation limit patient compliance. To exploit this therapeutic vulnerability, we optimized the preclinical use of paxalisib, identifying both precision medicine combination strategies and systemic interventions that reduced side effects and elicited profound tumor control in vivo. Mimicking the patient experience, paxalisib elevated blood glucose levels, promoting the phosphorylation of the insulin receptor (p-INSR) in patient derived DIPG xenograft models. Combining paxalisib with the anti-glycemic drug metformin, synergistically extended the survival of two DIPG xenograft models, decreasing p-Akt, p-mTOR and p-INSR in vivo. Phosphoproteomic profiling of DIPG cells identified paxalisib induced intracellular calcium-regulated PKC signaling. The brain penetrant PKC inhibitor enzastaurin extended the survival of DIPG xenograft models alone, and synergistically combined with paxalisib, enhanced using metformin in three DIPG xenograft models, as well as the immunocompetent H3f3aK27M, PdgfraD842Vmut, Trp53-/- murine brainstem glioma model. Mechanisms of adaptation/plasticity were assessed by ATAC-Seq and quantitative proteomic profiling xenograft tumors refractory to treatment. Here we address the intrinsic neoplastic sequela of DIPG, by combined targeting of PI3K/Akt/mTOR using paxalisib, compensatory PKC signaling using enzastaurin, coupled with strategies to manage treatment-related side-effects and reduced efficacy using metformin; providing the preclinical rationale for the addition of metformin to NCT05009992.

Abstract Diffuse midline gliomas (DMGs) diagnosed in the pons (DIPG) are universally fatal central nervous system tumors and are the leading cause of cancer-related death in children. Palliative radiotherapy is the only recognized treatment, with median overall survival just 9-11 months. The brain-penetrant, small molecule, ONC201, shows early-stage clinical trial efficacy, extending survival by ~9-11 months compared to historic controls. However, studies to determine the mechanisms behind the temporary clinical response to ONC201 are needed. Here, we have used a systems-biological approach to investigate whether genomic features influenced ONC201 response. DMGs harboring PIK3CA mutations were more sensitive to ONC201, whereas those harboring TP53 mutations were less sensitive. Quantitative proteogenomics identified that ONC201 elicits potent agonism of the mitochondrial protease, ClpP, driving proteolysis of electron transport chain and tricarboxylic acid proteins, leading to mitochondrial dysfunction. However, metabolic adaptation to ONC201 is promoted by the spare redox-signaling capacity of cells harboring WT-PIK3CA that was counteracted using the brain-penetrant PI3K/Akt/mTOR inhibitor, paxalisib. ONC201 and paxalisib combinations extended survival of orthotopic DIPG xenograft mouse models (SU-DIPG-VI, p=0.0027; SF8626, p=0.0002; HSJD-DIPG-007, p=< 0.0001). The combination in the first three recorded patients; two at progression following re-irradiation, and one at diagnosis following the completion of radiation, resulted in dramatic reductions in tumor area, dramatically extending overall survival for all three patients (25 months, 30 and 31 months continuing). The DIPG patient receiving the combination since diagnosis, remains in progression free survival (MR axial diagnosis scan = 1554 mm2, current tumour area = 306 mm2, ~80% reduction). The patient continuing to receive the combination at progression and following reirradiation also experienced a marked decrease in tumor size (MR axial diagnosis scan = 1248 mm2, current tumour area = 315 mm2, ~75% reduction), 10 months following radiological detection of progression. These data inform the phase II clinical trial (NCT05009992).

All Supplementary Figures and their captions.