Only the genomic features (gene mutations) of a patient’s tumour are analysed at the time of diagnosis due to the limited availability of biopsy material. Although this may identify some genetic characteristics driving tumour growth, few of these mutations are targetable, with even less providing a survival benefit to patients. As DMG tumours are known to be highly heterogeneous (driver mutations vary greatly between and within patient tumours), I-DIMENSIONS seeks to understand the multiple layers influencing DMG across a large selection of tumours so that better-informed treatment strategies can be developed.

I-DIMENSIONS aims to provide medical and research scientists with new data that unlocks effective treatment strategies for DMG. By looking at tumours as a sum of their biological systems (rather than via a single-featured lens), we hope that the developed subtyping system will inform the creation of a therapy selection tool used by clinicians to extend the survival of children diagnosed with brain cancer in the future.

 I-DIMENSIONS is an Integrated Dmg/hgg genomIc Methylomic EpigeNetic Spatial transcrIptomic prOteomic subtypiNg System integrating tumour genomics data (whole genome sequencing–WGS), DNA methylation data (EPIC array) and chromatin landscapes (ATAC-seq) to group patients into methylo-epi-genomic subtypes at diagnosis. Evaluating the spatial heterogeneity (scRNAseq) and the abundance and activity of all proteins (phospho- and proteomic profiling) present in each specimen will provide the most comprehensive picture of the elements sustaining tumour growth, revealing targets to be addressed using precision medicines. Our approach will be informed by utilising 210 samples collected from collaborators worldwide (including about 17 patient samples from the Charlie Teo Foundation Brain Tumour Bank) and be developed by a multi-disciplinary team of international experts.

Critically, I-DIMENSIONS will identify the highly significant influences on the regulation of a tumour’s posttranslational architecture i.e. the non-genomic elements relating to the geographical location of the tumour. The role that endogenous and exogenous microenvironmental influences (such as neurological cues, catecholamines, insulin, growth factors, growth hormones and immune related interactions - scRNAseq), as well as treatment-related neuronal effects from radiotherapy, and corticosteroid use will dictate posttranscriptional and posttranslational effects, that cannot be replicated in in vitro laboratory models.

Hypothesis:

The development of a methylo-epi-genomic subtyping system predictive of the proteomic/phosphoproteomic components of each sample will enable the future stratification of patients into subtypes that are indicative of treatment response.

Aim 1: Develop a novel methylo-epi-genomic subtyping system for DMG.

Aim 2: Identify the key proteomic/phosphoproteomic signatures (drug targets) of each specimen used in the DMG methylo-epi-genomic subtyping system.

Aim 3: Identify the spatial heterogeneity of the disease and relate this back to the proteomic/phosphoproteomic signatures.

This project will generate a large source of genomics data on brain cancers, a crucial step in understanding the genomic variability of brain cancer. Further, Prof Powell will leverage expertise from machine-learning, biology, genetics and advanced biotechnology domains in an effort to unravel the deep complexities of brain cancer and find effective treatment strategies for brain cancer patients.

Brain cancer is one of the most lethal cancers, with few effective treatments available for patients. Moreover, a hallmark of brain tumours is its ability to resist treatment due to its complex and variable genetic make-up. By understanding this genetic complexity, Prof Powell’s research team aims to find the ‘bad actors’ responsible for brain cancers resisting treatment and find effective treatments.

Identifying therapeutic targets from cell state models in gliomas

Glioblastoma is one of the deadliest human cancers, due to its strong resistance to treatment following surgical removal. This resistance is driven by an intrinsic intra-tumoral heterogeneity, which is a hallmark of these tumours. To develop new treatments to overcome this cellular heterogeneity, a more detailed understanding of the nature and drivers of this variation is needed. Single-cell RNA sequencing of human tumours provides a powerful means to systematically interrogate the diversity of malignant and normal cell states.

Recent studies, as well as Prof Powell's work, have highlighted transcriptional cell state diversity across tumour types that is often independent of genetic clonal heterogeneity. His team has developed highly accurate machine learning methods that are able to use the transcriptional signature of a single cell and are able to accurately classify it into a specific cell state. The researchers are able to ‘map’ the within-tumour heterogeneity from thousands of cells in a single patient sample. Doing so over a large clinical cohort across different subtypes of adult diffuse gliomas including astrocytomas, oligodendrogliomas and glioblastomas will allow them to identity the variation in the within-tumour heterogeneity between patients and correlate those cell states with clinical features, recurrence, and treatment resistance. The relationship between cell states and state dynamics will be tested against disease history and clinical features for each cancer classification type (following WHO CNS5 framework).

This project builds on Prof Powell's current research which has been supported by the Charlie Teo Foundation. His team will extend their research into the identification of the genomic drivers of cell states and state dynamics that correlate with clinical outcomes, generating new data on targeted glioblastoma patient cohort, and developing the high-throughput cancer organoid systems to be able to perform high-throughput drug screening against specific tumour cell states.

After decades of failure, scientists finally found effective ways of turning the immune system against cancers, with spectacular results – except for brain cancer. Immunotherapies have continued to fail in brain cancer and the team have discovered that the T-cells are stuck in the bone marrow of brain cancer patients. This project may unlock the reason why immunotherapy has failed in brain cancer and lead to immunotherapy treatments finally being effective.

If the team’s theory is correct, this work will finally allow for the development of effective immunotherapies for patients with brain cancer.

Development of Beta-arrestin 2 small molecule inhibitors for brain cancer therapy

Cancers of the intracranial (IC) compartment carry unique therapeutic challenges and offer grim outcomes, regardless of tissue type. These cancers include primary malignancies, such as universally lethal glioblastoma (GBM), as well as far more common brain metastases. Our group unveiled the ability of IC tumours to cause a dramatic plunge in the number of circulating T-cells, the main cells that help drive the body’s defences against cancer. Immunotherapy is a novel modality of cancer therapy that has gained momentum in recent years, which works by harnessing and ramping up the body’s own immune system, especially T-cells (the immune system’s effector arm) to fight cancer. However, without T-cells, there is nothing for immunotherapy to act upon. Our group tracked the missing T-cells in patients with IC cancers and discovered them ‘trapped’ in the bone marrow. We showed that this phenomenon is accompanied by loss of sphingosine 1-phosphate receptor 1 (S1P1) from the T-cell surface. S1P1 is a molecule which would normally act as an ‘exit visa’ allowing T-cells to exit from the bone marrow. Without surface S1P1, T-cells instead are locked in, unable to circulate out of the marrow. To solve this problem, our group demonstrated that mice lacking Beta-arrestin 2 (βARR2), an adaptor protein responsible for S1P1 internalisation, do not experience sequestration, and their T-cells are instead able to travel to the brain, and, with the help of immunotherapy, fight their cancers. Our studies in mouse models provided the proof of concept that blocking βARR2 and preventing loss of S1P1 could promote T-cells travelling to the brain to produce anti-tumour responses. However, we’ve now found that inhibiting βARR2 does far more to combat cancer, and we’re finding unprecedented responses against many types of cancer when we knock out βARR2. We think we’ve uncovered an entirely novel cancer therapeutic, and we’ve partnered with a Nobel Laureate who is an expert in this molecule to help develop a pharmacological strategy to block βARR2. We hope that our continued work will allow for the development of more effective immunotherapies for patients with brain cancer.

This research will build upon the work of our Cancer Genomics – The Next Level project as the data collected on the energy sources needed by brain cancer cells will be integrated and matched with the genetic data, creating the ultimate roadmap of what and how cancer cells grow, survive and ultimately how they can be destroyed.

It will enable us to understand how GBM tumours are metabolically programmed, providing the critical data for future studies in developing new therapies targeting these metabolic vulnerabilities.

Developing a metabolic roadmap to discover novel therapeutic avenues to starve GBM

Glioblastoma (GBM) is the most aggressive primary brain tumour with an overall survival prognosis of less than 15 months - a devastating diagnosis for both the patient and their family. Over many years, researchers have tried to map the genetics of the disease, in order to determine new ways to treat GBM based on genetic changes such as mutations. Despite this, there remains no effective treatments for GBM, with surgery, chemotherapy and radiation therapy the current options, suggesting we need to try an alternative avenue.

While genetic changes in the DNA have a profound effect on cancer cells, these effects are actually carried out by proteins which the DNA codes for. In turn these proteins, including nutrient transporters and enzymes, combine together to mediate the uptake and metabolism of nutrients in order to facilitate the rapid growth of the cancer cells. This is critical, as a cancer cell needs to essentially double all its DNA, RNA, proteins and fats in order to make two new cells during cell division. This cell division occurs rapidly in GBM, increasing the size of the tumour and ultimately tumour relapse post-treatment.

In this pilot study, we will undertake metabolic profiling of GBM patient samples taken at surgery, to understand how they use and distribute their essential nutrients. This will develop a metabolic roadmap which can be combined with genetic information being developed through the Cancer Genomics – The Next Level project. Ultimately, success in this project will enable us to understand how GBM tumours are metabolically programmed, providing the critical data for future studies developing new therapies targeting these metabolic vulnerabilities. These future studies would include using our patient-derived laboratory models to test currently available metabolic drugs, thereby determining whether they can successfully starve the GBM cells.

Experimental models have been extremely useful for learning about cancer and how we might treat it. For brain cancer, however, these models have failed to provide sufficient insight for a breakthrough. This project will use data to help understand what drives brain cancer from models of naturally occurring brain cancer that have never before been explored.

This work could be the beginning of a significant new frontier in brain cancer treatment for children. Striving to treat and cure dogs with cancer, learning what works best and importantly why it works, can inform our own therapy regimens. This provides an important opportunity to improve a child’s brain cancer outcome.

Targeting regions of converging synteny and loss of heterozygosity in paediatric and canine glioma

Brain cancers such as glioma occur in dogs at rates comparable to humans, with short-snouted breeds such as boxers being more susceptible than others. In helping to treat dogs diagnosed with brain cancer, the research team compared the molecular and cellular characteristics of glioma in dogs to their human counterparts and found extensive similarity in particular to aggressive glioma observed in children.

In this project, the team will leverage the similarities between glioma in dogs and in children to further sharpen the lens as to what is causing these cancers. They will mine large datasets on dog and children’s gliomas to precisely define the hundreds of molecular abnormalities found in both disease types. The team will perform a large screen and functionally eliminate each molecular abnormality one by one and evaluate the effects of this depletion on cancer-associated features such as cell growth in human and canine cell models of glioma. This advanced screen is enabled by the versatile and cutting-edge imaging platform and computational expertise at the Jackson Laboratory.

The team expects to see convergence of the most impactful molecular abnormalities on brain cancer’s evolutionary mechanisms, which will implicate that these mechanisms are candidates for the development of new treatments.