Grants

Backgound. Radiotherapy is still the mainstay of the treatment for DIPG. If the majority of children experience an improvement of their neurological condition following irradiation, this effect is not observed in all patients and is universally only transient. Determinants of the response to radiotherapy have yet to be defined. We have described two distinct forms of DIPG according to the type of histone H3 mutated with different response to therapy and survival (Castel et al., Acta Neuropathologica 2015 & 2016). In particular, in a cohort of 67 patients treated at our center, clinical response to radiotherapy was shown to be better in children with DIPG harboring a K27M mutation in HISTH3B gene than in children with a K27M mutation in H3F3A gene (85% vs. 55.3%, p=0.0263, Chi Square test). These data need to be confirmed in a larger study and refined thanks to the availability of DIPG models, to be able to predict response to radiotherapy more accurately. Hypothesis. We hypothesize that radioresistance of DIPG is first an intrinsic phenomenon than can be explained by the genetic background of the disease and second an acquired phenomenon that we aim to describe as oligoclonal or polyclonal. Design and methods. The project proposal aims to bridge biological studies on relevant preclinical models and appropriate patient’s data. The project will take advantage of a unique cohort of biologically defined DIPG patients with full clinical follow-up and treated at a single center as well as a large set of both in vitro and in vivo preclinical models. We have recently published the methods used to generate them and their relevance with respect to the biology of DIPG (Plessier et al., Oncotarget 2017).  We will treat our panel of cellular and murine models with irradiation in order to:  (i) correlate their radiosensitivity with their molecular profile (type of histone H3 mutated at K27 residue and subclonal driver alterations),  (ii) correlate these findings with patient’s data,  (iii) model and depict in vitro and in vivo the molecular aspects of radioresistance and tumour escape from radiotherapy using barcoding and lineage tracing. DIPG stem cells will be infected with lentiviruses encoding fluorescent proteins (a combination of 3 different markers) or unique barcodes. The fluorescence will be used to retrieve the tumor cells by cell sorting and to trace their evolution and development in the mouse brain after clarification. The unique barcodes will be identified by NGS allowing to quantify the clonality of the radioresistance.  Intrinsic mechanisms of radioresistance will be identified by correlating response to radiotherapy and molecular background both in the preclinical models and in the patients’ cohort. Acquired mechanisms of resistance will be studied by cell tracking after lentiviruses transduction and next-generation sequencing. Clinical significance. This project will determine molecular biomarkers associated with radioresistance of DIPG that will be useful in patient’s management. New therapeutic targets will be defined based on association of genetic alterations with initial radioresistance, or acquired genetic modifications after radiotherapy. 

Diffuse intrinsic pontine glioma (DIPG) is a devastating brain tumor which affects children and for which there is no effective treatment at the moment.  We and others have recently shown that DIPG is characterized by a remarkable degree of intratumoral heterogeneity and consists of distinct genetically and phenotypically heterogeneous subclonal populations. It is well recognized that exosomes mediate the cross-talk among the tumor cells and between the tumor and its microenvironment. We hypothesize that the DIPG subclonal network is sustained through paracrine signaling and that exosomes are involved in this intratumor crosstalk promoting tumorigenesis and tumor progression. To understand the role that exosomes play in DIPG tumorigenesis and determine the mechanistic basis of the exosome-mediated interclonal communication, we aim to: i. determine the specific DIPG-exosome signature within the distinct mutational subgroups and within distinct sub-clonal populations; ii. define the exosome-mediated mechanisms of crosstalk between distinct DIPG subclones and determine the functional consequences of such uptake. To address these aims we will specifically isolate and characterize exosomes derived from different DIPG patient primary-derived cells using proteomic and miRNA analysis. These exosomal “signature” will be compared with exosomal signature of glioblastoma patient’s primary cells. Furthermore we will investigate in details the specific profile of exosomes derived from DIPG subclonal populations and we will analyze the functional consequences (effect on growth and invasion) of exosomal uptake in co-culture experiments of DIPG subclones. Our goal is to better elucidate the mechanisms of cell-cell communication that mediate the DIPG growth and invasion.  We believe that this study will lead to the identification of potential new targets and the development of new diagnostic/prognostic tools for patients affected by DIPG. 

Malignant brain tumors are the leading cause of cancer-related mortality in children (1), and diffuse intrinsic pontine glioma (DIPG) is one of the most devastating, with a median survival of <1 year following treatment with radiation therapy (2). Despite more than 250 clinical trials over the past 30 years (3), not a single chemotherapeutic agent has demonstrated survival benefit. There is thus an urgent need for novel treatments for this disease. Our data and two recent high profile studies have implicated the protein EZH2 (enhancer of zeste homologue 2) as a specific therapeutic target in DIPG (4,5). Although EZH2 inhibitors exist and have shown very promising results in phase I clinical trials in other cancers, all the currently available EZH2 inhibitors do not cross the blood brain barrier and thus cannot be effectively employed in brain tumor patients.  
 
We have generated novel brain penetrant inhibitors of EZH2 in conjunction with the drug development institute at MSKCC (Tri-Institutional Therapeutics Development Institute), and the major goal of this project is to validate the in vivo efficacy of these inhibitors using pre-clinical models of DIPG. 
 
We hypothesize that brain penetrant inhibitors of EZH2 will represent effective therapeutics for DIPG based on our data, and recent studies (4,5) which show genetic knockdown of EZH2 has an anti-tumor effect in DIPG. 
 
The primary objectives of this project are: 1) Validate the in vivo efficacy of novel brain penetrant EZH2 inhibitors we have generated, using preclinical models of DIPG  2) Characterize critical epigenetic effects of histone mutations (found in DIPG) on self-renewal and differentiation, to gain understanding of how these mutations promote tumors 
    
Impact and Innovation: Aim 1 is directly translational, and our aim is to have a small molecule to enter IND (investigational new drug) studies in preparation for phase 1 clinical trials in patients with DIPG. EZH2 has been strongly implicated as a target in other brain tumors e.g. CNS atypical teratoid rhabdoid tumors, and various types of brain metastases; thus successful development of a brain penetrant EZH2 inhibitor will also be beneficial for this group of patients.  Aim 2 is directly mechanistic; and we anticipate that the proposed experiments will reveal insights into how H3K27M mutations promotes the formation of tumors.  
 
Feasibility: All the techniques outlined in this proposal, are already well established in the Allis Laboratory and will allow prompt initiation and completion of the project. It is anticipated that further optimization of the novel brain penetrant EZH2 inhibitors will be required, and further chemical biology support for continued optimization of the already generated brain penetrant inhibitors will be provided by the Tri-Institutional Therapeutics Discovery Institute (Tri-I-TDI). The ultimate goal will be to have a small molecule to enter IND (investigational new drug) studies within a period of 12-18 months in preparation for phase 1 clinical clinical trials in DIPG patients. 
 
Expertise: This project is a collaboration between Dr. Richard Phillips is a brain tumor specialist and physicianscientist at MSKCC and Dr. David Allis a world-leader in the study of epigenetics at Rockefeller. The Allis Lab has been a leader in dissecting the mechanisms by which the H3K27M mutation promote DIPG (6,7). Dr. Richard Phillips MD PhD has spear-headed a collaboration with chemical biology experts at the MSKCC drug development core (Tri-I-TDI) which has led to the development of two novel brain penetrant inhibitors of EZH2 which will be further characterized in this study.   The combination of these expertise will facilitate the successful completion of this project. 

Single-cell genomic profiling is revolutionizing our understanding of tumor biology, as it enables for the first time a genome-wide interrogation of tumor programs at cellular resolution, offering unprecedented insights into tumor cells and their micro-environment at a depth that was unthinkable even a few years ago. I expect that defining the genetic and cellular programs of DIPG with these techniques directly deployed in patient samples will reveal the key programs that underlie DIPG biology. By additionally leveraging the CRISPR/cas9 system to perform targeted and specific genetic knock-outs, we will functionally test candidate regulators and expect to identify novel tumor vulnerabilities and dependencies.  
 
Disease Impact   Applying single-cell genomic technologies to precious patient-derived tumor samples at biopsy as well as autopsy will shed unprecedented light on unique tumor vulnerabilities and dependencies that are not identifiable by bulk genetic studies alone. The proposed research will provide an unparalleled view of the cellular architecture and transcriptional networks underlying DIPG biology, and moreover, reveal novel tumor vulnerabilities that could rapidly enter pre-clinical and clinical trials.  
 
Innovation   The proposed study is the first ever to systematically study DIPG at cellular resolution, and is expected to reveal the interplay between genetically and developmentally driven programs in DIPG. By using the combination of cutting-edge single cell genomic technologies, genome editing tools and computational analysis available at our institutions and at the Broad Institute, these novel and complementary approaches will shed light on the molecular pathways driving DIPGs.  
 
Feasibility  Single-cell RNA sequencing as well as genome editing technologies have been established in the laboratory of Dr. Suva.  A close collaboration with computational scientists at the Broad Institute is also already in place.  The novel findings contributed by this study will provide a rational basis for renewed attempts at improving clinical care of DIPG patients. 
 
Expertise  Dr. Suva and Dr. Filbin have unique and complementary expertise that makes them the ideal investigators to complete the proposed research.  Dr. Filbin is a pediatric neuro-oncologist at Dana-Farber Cancer Institute and research fellow in Dr. Suva’s laboratory at Massachusetts General Hospital (MGH) and the Broad Institute.  Her previous work includes the discovery of a novel combinatorial targeted treatment for glioblastoma, which led directly to clinical trials in adults and pediatric patients with gliomas.  Dr. Suva is a faculty scientist and neuropathologist at MGH and the Broad Institute and has led ground-breaking research deploying single-cell genomics in adult gliomas. 

Diffuse intrinsic pontine gliomas (DIPG) are infiltrative, highly aggressive pediatric brainstem tumors with limited therapeutic options. Despite international efforts to improve outcome, DIPG show poor response to conventional radiation and chemotherapeutic strategies. Only within the last decade have studies really begun to decipher the molecular mechanisms behind DIPG tumorigenesis, with the goal of identifying novel therapeutic targets for this lethal disease. Next generation sequencing studies of DIPGs have identified canonical mutations, most frequently K27M substitutions in either H3F3A (H3.3) or HIST1H3B (H3.1) as well as TP53, ACVR1, PIK3CA and PDGFRA, among others. These results highlight the importance of epigenetic dysregulation in the pathogenesis of DIPG. However, the fact that H3K27M tumors typically harbor additional genetic aberrations and that modeling efforts using histone mutations alone have failed to induce transfomation, strongly supports the notion that H3K27M per se is likely insufficient to drive malignant transformation.  Additional mechanisms likely exist and are essential for DIPG commencement and/or progression beyond epigenetic and genetic modifications in nDNA.       Unlike normal cells, cancer cells have the capability to make an energy adaption through metabolic reprogramming from oxidative phosphorylation (OXPHOS) to aerobic glycolysis and other metabolic pathways. This occurs regardless of oxygen abundance (the Warburg effect) in order to facilitate their proliferation particularly under unfavorable microenvironments. Abnormal OXPHOS and aerobic metabolism as a result of mitochondrial dysfunction have long been hypothesized to contribute in diverse ways to the multistep process of tumor progression. In recent years, a large number of somatic mutations in the mitochondrial genome (mtDNA) and aberrant mtDNA amount have been increasingly detected in a broad spectrum of primary human cancers. Due to decreased expression of mtDNA-encoded polypeptides and impaired function of respiratory enzyme complexes, quantitative change in mtDNA may decrease mitochondrial respiratory activity and lead to persistent defects in the OXPHOS system. Decreased mtDNA copies and defective mitochondrial function have been strongly linked to neoplastic transformation, tumor progression, metastasis, chemo/radioresistance, and poor prognosis in several types of solid tumors. Our recent work funded by the DIPG Collaborative demonstrated that somatic mtDNA mutations and reduced mtDNA content are highly frequent events in DIPG.  Moreover, partial depletion of mtDNA to DIPG-like levels in immortalized NHA (iNHA) cells significantly increased tumorigenicity in vivo. These findings led us to hypothesize that mitochondrial dysfunction and incompetent oxidative metabolism owing to lower mtDNA copies by themselves, or in a coordinate fashion with nDNA alterations, may be involved in DIPG tumorigenesis and DIPG cells may be characterized by “Warburg rearrangement for metabolism”. Work proposed in this study aims to revert the Warburg metabolism and rebuild normal OXPHOS activity in a panel of patient-derived primary DIPG cell lines by targeting reduced mitochondrial number through pharmacological stimulation of PGC-1α-mediated mitochondrial biogenesis (Aim 1). The potential in vivo efficacy of three mitochondrial biogenesis drugs (AICAR, resveratrol and metformin) that have excellent profiles and are being used or trialled to treat mitochondrial disorders in other human diseases (e.g. Alzheimer's disease, Type 2 diabetes) will be evaluated in both transgenic murine and human DIPG xenograft models. Through the use of high-throughput synergy drug screening, we will produce preclinical data identifying the best clinically approved compounds to be used in combination with mitochondrial targeting in the chemoadjuvant setting (Aim 2). Our innovative approach of manipulating mtDNA quantity to correct compromised mitochondrial function thereby reversing the malignant phenotype of DIPG, either alone or in combination with conventional therapies, will yield important preclinical data for designing novel therapeutic strategies. Our group is a leader in the field of DIPG and being part of the largest pediatric neuro-oncology team in Canada and active contributors in the DIPG collaborative, we are ideally situated to translate discoveries made through this project into phase I/II clinical trials for this devastating pediatric cancer.   

DIPG is characteristically infiltrative (i.e. diffuse and intrinsic), and this infiltrative/invasive behavior is destructive both in the brainstem and in other areas of the central nervous system to which DIPG spreads during the course of the disease. In work previously funded by the Cure Starts Now and the DIPG Collaborative, we discovered that neuronal activity promotes DIPG cell invasion through activity-regulated secreted factors that includes an endogenous antagonist to the Nogo receptor (NgR). NgR signaling is a key mechanism that restricts the plasticity and regeneration of normal brain cells, and blocking the Nogo pathway results in increased motility of cells and cellular processes such as axonal outgrowth. In the normal brain, this signaling pathway plays a role during early brain development and may continue to play a role in ongoing brain plasticity. DIPG cells express NgR and in preliminary studies we have found that blocking NgR signaling with recombinant antagonist or by deleting the NgR gene via CRISPR gene editing from patient-derived DIPG cells dramatically increases DIPG invasion in vitro. In the proposed studies, we will expand these observations to a larger number of patient-derived DIPG cell cultures to determine how universal this mechanism may be, and will assess the importance of NgR signaling to DIPG invasion in vivo using genetic models. If NgR signaling proves to be an important mechanism controlling DIPG invasion, then stimulating the NgR receptor may be an innovative therapeutic strategy to control infiltration of DIPG cells throughout the brainstem and prevent spread more diffusely to the cerebrum and spinal cord.

The regulation of the NgR signaling by neuronal activity represents one of the many ways that experience shapes brain development and plasticity. The effects on DIPG invasion represents yet another way that DIPG cells hijack normal mechanisms of brain development to promote disease progression. Ultimately, by understanding the ways DIPG subverts mechanisms of childhood brain development and adaptability, we hope to develop effective and tumor-specific strategies to disrupt the ability of the tumor cells to use these crucial signals in the tumor microenvironment.

Whole genome sequencing on DIPG biopsy and autopsy samples have identified specific mutations in Histone H3, ACVRI, TP53 and ATRX causing heterochromatin silencing, genetic instability and alterations in DNA damage response pathways. Using a unique panel of patient derived DIPG cultures, we have further examined the molecular profile of DIPG to understand key oncogenic drivers within these tumours and identify novel therapeutic targets. Excitingly, we have found overexpression of multiple regulators which control the cell cycle. Of these, Polo-like Kinase 1 (PLK1) represents the most promising cell cycle factor for targeting in DIPG. PLK1 not only activates the key intermediaries which promote tumour cell growth and division, but also controls cellular response to DNA damage, inducing cell cycle arrest to promote DNA repair and avoid cell death while regulating the mechanisms used to escape arrest and continue tumour growth. Importantly, we hypothesise the rapid repair of DNA and recovery from cell cycle arrest controlled by PLK1 may account for the resistance to radiotherapy commonly seen in children with DIPG. We propose that inhibiting PLKI activity may impede cell cycle progression, block DNA repair and recovery following cell cycle arrest, radiosensitise the tumour and increase the levels of single stranded DNA above the threshold required to induce DNA lethality resulting in tumour cell death. We also hypothesise that the activity of PLK1 inhibitors may be synergistically enhanced through the development of rational combination therapies.

Our data shows PLKl is essential for DIPG tumour cell growth and that PLKl inhibition provides one of the most promising potential treatment applications for children with DIPG to date. We have found single agent PLK1 inhibition using clinically available therapies has extremely potent cytotoxic effects against DIPG cultures in vitro, with near complete eradication of DIPG cells, inhibition of colony formation and a significant induction of G2 cell cycle arrest all at exceptionally low nanomolar potency (IC50 <10nM). Using our most aggressive patient-derived in vivo model of DIPG, we have examined the activity of 2 different PLKl inhibitors as single agents and found that both therapies significantly and potently enhanced the survival of all treated animals (P=<0.0005). Importantly, PLKl inhibitors have shown the most potent anti-DIPG effect in vitro and in vivo compared with any other therapy we have tested in our models to date. We have also found that PLKI inhibition targets one of the major oncogenic drivers in DIPG, the Phosphoinositide 3-kinase (Pl3K) pathway. We have preliminary evidence PLK 1 inhibition is synergistically enhanced upon addition of the Pl3K inhibitor BKM120, providing a highly potent combination therapy for DIPG (Combination Index 0.0004). We have seen a similar effect of PLK1 inhibitors in combination with Temsirolimus (an inhibitor of Mechanistic Target of Rapamycin). As a result, we seek to conduct an analysis of PLK1 inhibitors in combination with inhibitors which target key oncogenic pathways in DIPG to identify the two most potent and synergistic combinations, their mechanism of action and their efficacy in our two orthotopic models of DIPG, providing the necessary quantum of data to move these treatments to the clinic.

Our team has the necessary skills and expertise to ensure success of the proposed project and successful clinical translation. PI Franshaw is an early career Research Officer at the Children's Cancer Institute with expertise in establishing and maintaining DIPG cultures, drug discovery, in vivo models of DIPG, advanced imaging and cancer cell biology. Since joining the Institute late 2012, she has helped establish multiple DIPG cultures, 2 in vivo models of DIPG and identified multiple novel treatments against DIPG currently under investigation. Her most recent discovery of the potency of PLKI inhibitors against DIPG is one of the most promising findings in the field to date. Co-PI Ziegler is a paediatric oncologist and director of the clinical trials unit at the Kids Cancer Centre (Sydney Children's Hospital), Group Leader at the Children's Cancer Institute and a Conjoint Associate Professor at the University of New South Wales. He is a national leader in early phase clinical trials in paediatric oncology, testing novel therapeutic agents in children with relapsed and refractory malignancies for whom there are no known effective treatment strategies. Support by A/Prof Ziegler will ensure success of the project and the rapid movement of effective therapies to the clinic to directly benefit children with DIPG.

Validation of TAK228 as an Effective Drug for Treating Children Diagnosed with DIPG Diffuse intrinsic pontine glioma (DIPG) is an untreatable childhood cancer.  DIPGs develop in pons making them surgically unresectable.  We and others have identified mutations in histone and their obligate partner mutations that seem to drive tumorigenicity. While epigenetic drugs targeting histone mutation have shown positive outcome in preclinical setting, these drugs have yet to show effectiveness in the clinic.  Moreover, >70% of DIPGs are driven by AKT gain or PTEN loss.  Our collaborative team at Children’s National and Johns Hopkins School of Medicine have achieved robust preliminary success in targeting AKT and PTEN pathways using TAK228, also known as sapanisertib, MLN0128 and INK128. We also find that TAK228 in combination with radiation leads to increased DIPG apoptosis. We also find preliminary evidence of synergy between panobinostat and TAK228, consistent with findings in other tumor types that mTOR inhibition in combination with panobinostat leads to improved cell killing (20, 21). We hypothesize that combinatorial regimen of TAK228, radiation, and panobinostat can effectively target H3.3 and H3.1K27M mutant cells in vitro and in vivo and that any mechanism of drug resistance are identifiable using cutting edge protein and gene studies. In this study, two Specific Aims will be tested: Specific Aim 1. Assess TAK228 efficacy in H3.3 and H3.1 DIPG models, alone and in combination with radiation and panobinostat.   Specific Aim 2. Determine mechanisms of cellular resistance to TAK228.  Our collaborative team has all necessary resources to carry the outlines experiments. The experiments detailed here will provide the pre-clinical justification of a phase I/II clinical trial of TAK228 in patients with newly diagnosed DIPG. 
 
Innovation. Our collaborative proposal is highly innovative in the following aspects: - Strong preliminary data indicating the efficacy of TAK228 for targeting DIPG tumor 

- Robust data indicating TAK228 administration more than doubles the median survival of mice bearing a well-accepted murine DIPG orthotopic model 

- Ability to target both mTORC1 and mTORC2 pathways using TAK228 

- mTORC1 inhibition by commonly used drugs (e.g. rapamycin and everolimus) often leads to upregulation of mTORC2, which may contribute to the relative lack of effectiveness of mTORC1 inhibitors. TAK228, which is a dual mTORC kinase inhibitor may have broader therapeutic use than rapalogs, due to its ability to inhibit both mTORC1 and mTORC2.

- Ability to inhibit DIPG growth and induce apoptosis at nanomolar concentrations.

- The agent is available for clinical trials in the United States through the CTEP program at NCI. 

- Collaborative proposal between two leading institutions in childhood brain tumor research.

- Access to robust and rare xenograft, allograft, and in vitro models of DIPG avenues.

- Potential to assess and identify combinatorial therapy of TAK228 for direct translational application of this agent. 

- Potential for identification of tumorigenic pathways and thus further allowing for novel therapeutic combinations. 
 
Feasibility. The proposed project is a highly feasible project as shown by our robust preliminary data and published manuscript (Cancer Letters, 2017). Our collaborative team of basic researchers and clinicians have the expertise and the resources to fully execute the proposed project.                                                     
 
Expertise. Dr. Nazarian (PI) at CNHS has extensive expertise studying the molecular profiling (mRNA, methylation, proteomics) of DIPGs. Dr. Raabe (Co-PI) a pediatric oncologist at Johns Hopkins University School has extensive experience in cell culture, molecular biology, immunochemical analysis and xenografting. Dr. Raabe’s laboratory had developed a primary DIPG line (JHH DIPG1) extensively used for preclinical testing across many laboratories.

We and others co-discovered the presence of activating mutations in ACVR1 in 25% of human DIPGs in 2014. Subsequently, my laboratory has developed a murine DIPG model incorporating R206H ACVR1 and observed that R206H ACVR1 significantly accelerates brainstem gliomagenesis. In addition, short-term treatment with a bone morphogenetic protein pathway inhibitor (LDN-212854), significantly reduced proliferation in this model. Furthermore, we have preliminary data that the MAPK pathway is upregulated in human DIPGs with ACVR1 mutations and that the MAPK pathway remains activated in our murine DIPG model despite BMP pathway inhibition.

Therefore, we hypothesize that inhibiting both the BMP pathway and the MAPK pathway will significantly prolong survival of ACVR1 mutant DIPG-bearing mice. Here are proposing to test two treatment strategies: 1) LDN-212854 (a BMP pathway inhibitor) in combination with a MEK inhibitor (PD-0325901), and 2) E6201, a small molecule inhibitor that inhibits both the BMP pathway and the MAPK pathway. Of note, both PD-0325901 and E6201 are already in clinical trials in adult cancers.

Disease Impact

DIPG is an incurable brain tumor that arises in children. The standard of care today in 2017 is the same as it was in the 1960s. This is due, in part, to the frequent “recycling” of adult brain tumor treatment approaches, which is becoming less common due to the discovery of significant genetic differences between adult and pediatric gliomas, including DIPG. In this proposal, we seek to change this. Here we are proposing to evaluate the in vivo efficacy of a bone morphogenetic protein (BMP) pathway inhibitor in combination with a MEK (MAPK/ERK Kinase) inhibitor in our murine DIPG model that harbors the R206H ACVR1 and H3.1K27M mutations. We will evaluate two treatment regimens; one treatment regimen will be a combination of two drugs (LDN-212854 and PD-0325901) and another will be one drug that inhibits both the BMP and MAPK pathways (E6201). Of note, we will evaluate both treatment regimens in combination with radiation, the standard of care for children with DIPG. Results from this proposal will guide future preclinical work as well as guide the development of a clinical trial for children with DIPG through the PBTC.

Innovation

We have developed a murine DIPG model that incorporates mutant ACVR1 (R206H) and H3.1K27M (two genetic alterations that often co-occur in human DIPG specimens). Using this model, we have observed that R206H ACVR1 accelerates brainstem gliomagenesis and that H3.1K27M accelerates brainstem gliomagenesis only in the presence of ACVR1 R206H. We propose to evaluate inhibitors of the BMP and MAPK pathways in this genetic mouse model to help prioritize therapies for clinical trials in children with DIPG. Furthermore, we will also investigate mechanisms of resistance by comparing the RNAseq of treated and untreated tumors.

Feasibility

We have developed this murine DIPG model and have already started making these tumors at Northwestern. With regards to LDN-212854, we have an ongoing collaboration with Paul Yu (a researcher who studies FOP at Harvard) who provides drug to our lab for preclinical testing. We will purchase PD-0325901 from Selleck Chemicals and will receive E6201 from Strategia Therapeutics (See letter of Support).

Expertise

My laboratory has developed these murine models of DIPG and we are experienced in preclinical testing of targeted therapies using this model.

Diffuse intrinsic pontine glioma (DIPG) is a currently incurable childhood brain tumor that, despite many past clinical trials, has never been shown to respond to systemic chemotherapy. Radiation therapy (RT) is effective in extending life but is not curative; median overall survival is 11 months and long-term survival is extremely rare.1 Preliminary evidence from compassionate use of convection-enhanced delivery (CED) of carboplatin shows this approach is likely to be efficacious at local tumor control and prolonging survival, but distant CNS disease develops (Gill S, DIPG Symposium, 2017). Therefore, a combined local and systemic therapy approach is likely to be required for long-term cure. Major questions related to both chemotherapy and RT must be answered before such a treatment approach can be planned. For chemotherapy, no studies have measured intratumoral drug penetration to determine whether the failure of systemic chemotherapy is due to failure to reach the tumor. Trials continue to test systemic chemotherapy alone, taking resources away from the study of CED and combined approaches. For RT, the current standard of care is focal treatment only, although DIPG clearly spreads throughout the neuraxis,2 which is likely to become a bigger impediment to survival as local tumor control improves. We have established multiple murine orthotopic patient-derived xenograft (PDX) models of DIPG that show local and disseminated tumor, and we have measured intratumoral drug penetration of the chemotherapy agents gemcitabine and selinexor in these models. We have also opened a phase 0 trial of gemcitabine given prior to biopsy in newly-diagnosed DIPG, and we are developing a similar trial with selinexor. We hypothesize that a multimodality approach that includes chemotherapy via CED, along with craniospinal RT (CSI) with a focal boost, will produce long-term cures in DIPG. At this stage, the questions related to this hypothesis that we can answer through clinical and preclinical trials are 1) Can systemic chemotherapy penetrate DIPG adequately to produce a therapeutic effect; 2) Do PDX models adequately reflect drug penetration in humans; 3) Does CED improve drug penetration over systemic delivery; and 4) Does the addition of CSI improve the control of disseminated tumor over focal RT alone. Our objective in this proposal is to advance DIPG research toward an effective, multimodal treatment approach. To pursue this objective, we will accomplish these specific aims and have the following anticipated results: 
 Specific Aim 1: Compare intratumoral chemotherapy delivery between systemic delivery and CED, and between clinical trial subjects and PDX models. We anticipate that systemic chemotherapy delivery will be inadequate for therapeutic efficacy in both clinical trial subjects and PDX models, and will be improved by CED. We will use the two phase 0 trials of gemcitabine and selinexor discussed above and will compare the results to parallel trials run in murine orthotopic PDX models of DIPG, with both drugs given systemically and by CED. 
 Specific Aim 2: Measure the impact of the addition of CSI to focal RT on survival and disseminated disease in PDX models of DIPG. We anticipate that adding CSI to focal RT will decrease the amount of disseminated disease in PDX models of DIPG, although it will not increase survival. We will use a patient-derived PDX model of DIPG to compare disseminated disease, as measured by high-resolution brain MRI and histological review of brain sections from necropsy, as well as survival in mice receiving CSI with a focal boost versus focal RT alone

Diffuse intrinsic pontine gliomas (DIPG) are highly aggressive brain tumors found in an area of the brainstem called the pons, which controls many of the body’s most vital functions such as breathing, blood pressure, and heart rate. DIPG is the leading cause of pediatric death by brain tumor. Median survival after diagnosis is only 9 months, and 5-year survival is less than 1%. Unfortunately, complete surgical removal is not an option in the treatment of these tumors due to their extensively infiltrative nature and related severe neurological damages to the most vital functions of the body caused by the surgery. The only effective traditional limited-field radiation produces responses in more than 90% of DIPG patients but only transiently. Despite intensive multimodality treatment, refractory disease and relapses are frequent events in these tumors. Numerous trials to increase the dose of radiation have been performed and have not improved patient survival. Experimental chemo- and biologic agents in addition to radiation are actively investigated. More than 250 clinical trials have been conducted in the past decades and several trials evaluating new agents are either underway or have been recently completed but none of these trials have shown any benefit in the likelihood or the median length of survival. Thus, new agents that can be used for adjuvant DIPG therapy are desperately needed.  Using DIPG cells established from newly diagnosed and recurrent patients, which retain stem-like cells features including the ability to self-renew and differentiate into multiple neural cell lineages, we performed a repurposing drug screen aiming at evaluating the potential of recycling of old known drugs for DIPG therapy. Repurposing drugs that are already approved to treat certain diseases or conditions to see if they are safe and effective for treating other diseases provides quicker translation from bench to bedside. We identified that a subset of quinoline class of anti-malaria agents to have significant DIPG killing properties. Among them, we selected the FDA-approved off-patent compound mefloquine due to its efficacy and potency in killing DIPG cells, with an IC50 less than 0.5 µM in all the tested cultures. Mefloquine remains a useful anti-malarial drug for many patients and is the drug of choice for U.S military deployed overseas and travelers. However, a growing body of evidence suggests that mefloquine causes adverse central nervous system effects, including posttraumatic stress disorder-like syndromes such as nightmares and paranoia, because after being absorbed into the bloodstream, it accumulates in the brain at relatively high concentrations, which compromises its use.    
 
Identifying new indications for existing drugs, even taking advantage of the adverse effects of these compounds, is not uncommon in drug development.   In this proposal, we will take advantage of this unique characteristic of mefloquine and use low dose (2.5 - 7.5mg/kg in mouse) of this compound to achieve enough amounts in the brain (300 nM - 1 µM) and target the infiltrative DIPG reservoir, while avoiding the neurological side effects.  Anti-malaria studies suggested that this group of drugs induce cell apoptosis in host and parasites through increased production of reactive oxygen species (ROS). We will evaluate ROS pathway as a potential mechanism in mefloquine-induced DIPG cell death.  Our preliminary research using mefloquine at low doses restricted DIPG xenograft growth and improved the survival of mice. We will take advantage of the adverse side effect of mefloquine, when used as an anti-malaria drug, and investigate the possibility of achieving desired concentration in DIPG using low dose of this compound. Further, the ROS stress induced by mefloquine will specifically target the invasive DIPG cells in the brain.  We hypothesize that low dose of mefloquine can cross the blood-brain barrier, achieve the desired local concentration, and induce DIPG cell death through increased ROS production, thus halt tumor growth and prolong animal survival.  Investigating the efficiency and safety of mefloquine in animal models is the first step to move this agent to the clinic to treat DIPG patients.  Upon completion of the proposed project, we expect that we will have the necessary evidence to move this drug into a clinical trial. 

Diffuse intrinsic pontine glioma (DIPG) is a form of brain tumor that arises in the brainstem of children and is nearly 100% fatal with a median survival of 9-12 months after diagnostic. The location of these highly infiltrating tumors in the ventral pons precludes surgical resection and clinical trials with different chemotherapeutic drugs have shown dismal results. Alterations in epigenetic regulation appear to be an important feature in DIPG as whole-genome/exome sequencing revealed recurring heterozygote mutations in H3.3 and H3.1 histones (K27M, G34R and G34V), respectively encoded by H3F3A or HIST1H3B genes. Mutations in H3F3A that result in the K27M substitution occur in 70-80% of DIPG while 11-31% of mutations are found in HIST1H3B. The evidence from mechanistic studies indicates that H3-K27M mutant histones exert their dominant effects through global reduction of methylase activity that may be critical for tumor maintenance as an inhibitor of JMJD3 demethylase has a considerable anti-tumor effect. Presently there are no drugs to target mutant histones specifically. Development of effective therapies for DIPG is hampered by the inability to target the majority of cancer relevant proteins using traditional small molecules. In contrast gene specific silencing using RNA interference, or antisense oligonucleotides (ASO), has no theoretical limitation on genes that can be targeted. Moreover as the nucleotide backbone chemistry determines the PD/PK and the sequence determines target engagement these flexible platforms can be used to rapidly assess the therapeutic potential of new genes or modulation of multiple genes in a network. Using RNA interference with chemically modified ultra-stable siRNAs for cancer therapy is an innovative approach with no apparent target limitations that builds on the growing clinical experience using oligonucleotides to modulate gene expression in neurological diseases. Dr. Anastasia Khvorova in the RNA Therapeutics Institute at UMMS has developed new ultra-stable hydrophobically modified siRNAs (hsiRNA) capable of sustained gene silencing throughout the adult mouse brain after a single injection into the cerebral spinal fluid (CSF). Also our laboratories (Sena-Esteves and Khvorova) found that intracranial delivery of hsiRNA is effective in silencing gene expression in a highly migratory human glioblastoma multiforme (GBM) orthotopic tumor model. Therefore we hypothesize that CSF delivery of hsiRNAs is an effective approach to silence gene expression in orthotopic DIPG tumors and drive transformative therapeutic outcomes with translation potential. Successful completion of this proposal will demonstrate that RNA interference is a potent platform to harness the wealth of information available on gene networks driving DIPG to develop new genetic therapies for this devastating disease that so desperately needs effective treatments. We propose the following aims: Aim 1: Target screening and validation in human DIPG lines. The high frequency of H3-K27M mutations in DIPG suggests that it may have a fundamental role in tumor maintenance/progression despite not being sufficient for tumor initiation. Here we will develop hsiRNAs specific for the H3F3A gene where most H3.3K27M mutations are found through the use of SNPs that vary across the 15 histone genes as well as targeting the unique 3’untranslated region of this gene. Also we will investigate the therapeutic effect of targeting cell cycle proteins with no apparent genetic redundancy such as mitosis cyclins (CCNA2, CCNB1) and others. Initial screening will be conducted in human HEK293 cells using mRNA levels as outcome measure. The biological effect of the most efficient hsiRNAs for each target gene will be assessed in multiple DIPG lines with and without H3.3K27M. 
 Aim 2: Efficacy studies in experimental models of DIPG. The hsiRNAs shown to impact tumor growth/survival in vitro will be tested in orthotopic tumor models generated by stereotaxic injection of firefly luciferase expressing DIPG tumor cells (H3.3 WT or K27M) in the brainstem of athymic nude mice. hsiRNAs will be infused into the CSF via the cisterna magna to maximize brainstem delivery and we will assess impact on tumor growth over time via bioluminescence imaging and survival. 

Brain tumors are the leading cause of mortality and morbidity from cancer in children and young adults. Malignant gliomas, including supra-tentorial high grade gliomas (HGG) and Diffuse Intrinsic Pontine Gliomas (DIPG), remain a major therapeutic challenge in pediatric oncology because they are still incurable. The failure of current treatments is explained by the difficulties to do a complete surgical resection, and by the low penetration of drugs commonly used neuro-oncology. This low penetration is related to the existence of the blood-brain barrier (BBB), that consists partly of endothelial cells intimately linked together by tight junctions and prevents about 98% of small molecules and 100% of large molecules to reach the brain parenchyma.

It is now proven that the application on brain parenchyma of low power pulsed ultrasound (US) in combination with an US microbubble contrast agent (UCA) can safely and transiently open the BBB (Hynynen et al, Radiology 2001). A significant increase in penetration of several antineoplastic drugs, high molecular weight antibodies, DNA plasmid and neural stem cells has been observed after US-induced opening of the BBB. Studies on murine models of gliomas even showed an improved tumor control and increasing animal survival. These studies, initially conducted on small animals, have recently been applied by us in primates with confirmation of previous results (Horodyckid et al, J Neurosurg in press).

The skull represents the main obstacle to the use of US in neuro-oncology, since the bone induces distortions and attenuation of ultrasonic energy. Considering the fact that a surgical procedure is systematically carried out during the treatment of primary brain tumors (resection or biopsy), the research team CarThéra™, directed by Pr Alexandre Carpentier, has recently developed an implantable ultrasonic transducer. The concept was described in a recent publication with opening of the BBB in rabbits (Beccaria et al, J Neurosurg 2013). The team was able to demonstrate a significant increase in intracerebral concentration of Temozolomide and Irinotecan in rabbits following US9induced opening of the BBB (Beccaria et al, J Neurosurg 2015). The first clinical trial is under way in La Pitié Salpétrière Hopistal (Paris, France) under the coordination of Pr A. Carpentier (NCT02253212 ; see also Carpentier et al, Sci Transl Med in press).

A few of molecules have shown a very good anti-tumor efficacy on DIPG primary cell lines but many of them have poor diffusion across the BBB (Grasso et al, Nat Med 2015). One of the most effective drugs, the panHDAC inhibitor Panobinostat has recently shown a strong anti-tumor activity in vitro and in vivo, but this molecule presents a low penetrance through the blood-brain barrier (~10% in mouse), and its therapeutic index is limited by hematotoxicity.

There is therefore a strong rationale to explore panobinostat with unfocused US in a preclinical xenograft model of DIPG deriving from treatment-naive biopsies. After adaptation and validation of US parameters in mice, we will explore the pharmacokinetics of panobinostat in blood and brain (pons, cortex, thalamus) in healthy mice and then in DIPG mouse models after opening of the BBB with US. Depending on the results, we will evaluate the anti9tumor activity of panobinostat combined or not with unfocused US in comparison to panobinostat alone in the preclinical xenograft DIPG models. This project has been validated by the ethical committee CEEA 26 and Research Ministry (approval number 2015071316539600 v3, APAFIS#1141). In case preclinical data and safety data in the brainstem are encouraging in mice, we will test the safety of US-mediated BBB disruption in the brainstem of primates. The goal is to obtain the necessary preclinical data to justify the development of a phase I clinical trial to evaluate the feasibility and safety of the technique in children treated for a HGG and DIPG.

We now have a wealth of genetic and epigenetic information about DIPG. However, relatively little is known about the molecular consequences downstream from H3K27M mutations and how they can lead to cancer formation. It is known that, as well as DNA hypomethylation, H3K27M mutation leads to a global loss of the repressive chromatin mark H3K27me3. H3.3K27M has also been shown to interact more strongly than wild type (WT) H3.3 with the polycomb repressive 2 (PRC2) complex members EZH2 and SUZ12. However, beyond these observations relatively little is known about how H3K27M mutations might alter the histone code through changes in interactions with writers, erasers and readers. This is a critical step in understanding how the H3K27M mutation might drive DIPG pathogenesis and how to develop appropriate therapies for H3K27M mutant tumors.  To better our understanding of the effects of these H3K27M mutations, we are undertaking a novel approach to the problem by investigating the protein interactions that are altered by H3.3K27M and H3.1K27M. Histone binding partners whose histone affinity may change in the presence of H3K27M are likely to play a significant role in mediating this mutation’s tumorigenicity. As well as extending our molecular understanding of DIPG, focussing on these interactors offers an attractive new avenue in the search for an effective therapy for this fatal disease. The structural and post-translational plasticity of histone tails means directly targeting H3K27M itself with small molecules is unlikely to be effective. However, understanding how this mutation alters the histone interactome offers a unique opportunity to specifically modulate the activity of those proteins immediately downstream from H3K27M. The goal of this project is to exploit novel proteomics technology (Bio ID) to gain new insight into the functional consequences of the H3K27M mutation. We hypothesize that the H3K27M mutation leads to differential interaction with and function of multiple chromatin modifiers and nuclear proteins beyond PRC2. This hypothesis is supported by our preliminary data and will be rigorously tested through the following specific aims: Aim 1: To determine the differential protein interactions imparted by the K27M mutation on both H3.1 and H3.3 Aim 2: To characterize the H3K27M interaction landscape in DIPG model systems Our project is the first to directly address the question of how the H3K27M mutation in both H3.1 and H3.3 impacts on the H3 protein-protein interaction landscape. A deep understanding how the protein landscape of DIPG changes will not only shed further light on the molecular basis of this disease, but it will also open a new approach for developing more effective treatment regimens for this universally fatal disease. The facilities and expertise at The Brain Tumor Research Centre and the Hospital for Sick Children as well as our close collaborations within the DIPG Registry Group will help us rapidly advance our findings through pre-clinical investigations and eventually into clinical trials. 
 

Diffuse intrinsic pontine glioma (DIPG) mainly affects very young children and has the highest mortality of all pediatric solid tumors. The median survival is only 9 months, with most patients dying within 2 years after the initial diagnosis. The clinical management of DIPG is a major challenge in pediatric neuro-oncology. The tumor's location in the brainstem precludes surgical resection, and systemic or local chemotherapy is not effective for DIPG. The standard of care is focal radiation therapy, which only transiently relieves symptoms and delays tumor progression.

 Advances in tumor biology and diagnosis, including genomic analysis of tumors, have resulted in the identification of specific mutations, which provide an opportunity for a targeted approach to DIPG. Unlike many adult brain tumors, DIPG is associated with a specific mutation that is present in greater than 80% of tumor samples, and results in replacement of lysine 27 with methionine in variants of the histone H3 protein. The majority of the mutations occur in the H3F3A gene, which codes for histone H3.3. K27 normally undergoes post-translational modifications that result in a dynamic state of un-, mono-, di-, and tri-methylation, or in acetylation, each of which is associated with distinct epigenetic control, in conjunction with other modifications within the histone tails. The K27 mutation represents a toxic gain of function, resulting in sequestration of the EZH2 subunit of the Polycomb repressive complex 2 (PRC2), the enzyme that catalyzes trimethylation of K27 to silence genes. Thus, K27M is a dominant-negative mutation, which occurs somatically, in a heterozygous state in the tumor cells.

 The goal of this study is to selectively knock down the expression of the mutant allele of H3F3A, by means of antisense technology. Antisense oligonucleotides (ASOs) are precision drugs for targeted therapy, which have already been or are being developed for a variety of indications. One highly promising ASO, nusinersen, is well into phase-3 clinical trials, and is being safely and effectively administered to the CNS by lumbar puncture 2-3 times a year, in infants and children with spinal muscular atrophy. The Pl was an inventor of nusinersen. We propose to develop ASOs to precisely and selectively elicit destruction of the mRNA coding for mutant H3.3 K27M, such that the normal H3.3 encoded by the wild-type allele can carry out its normal functions. We expect that treatment with such an ASO will lead to tumor regression or arrest, either by itself, or in combination with other drugs, including small molecules or biologicals.

 Funds are requested ($100,000 for 1 year) to establish the feasibility of this antisense approach in vitro. ASOs will be systematically designed, screened, and optimized for their ability to selectively reduce expression of H3F3A K27M mutant mRNA and protein in three independent patient-derived DIPG tumor cell lines. Effects on expression of downstream targets, as well as on cell proliferation and survival, will be measured. The lead ASO(s) should be suitable for future IND-enabling, in vivo pre-clinical studies employing orthotopic-xenograft or genetic mouse models of DIPG. The basic approach could also be extended to other dominant-negative mutations in H3F3A or in other genes causally associated with DIPG or with other gliomas.