Search

Summary: Background: The malignant childhood brain tumour, medulloblastoma, is classified clinically into molecular groups which guide therapy. DNA-methylation profiling is the current classification ‘gold-standard’, typically delivered 3–4 weeks post-surgery. Pre-surgery non-invasive diagnostics thus offer significant potential to improve early diagnosis and clinical management. Here, we determine tumour metabolite profiles of the four medulloblastoma groups, assess their diagnostic utility using tumour tissue and potential for non-invasive diagnosis using in vivo magnetic resonance spectroscopy (MRS). Methods: Metabolite profiles were acquired by high-resolution magic-angle spinning NMR spectroscopy (MAS) from 86 medulloblastomas (from 59 male and 27 female patients), previously classified by DNA-methylation array (WNT (n = 9), SHH (n = 22), Group3 (n = 21), Group4 (n = 34)); RNA-seq data was available for sixty. Unsupervised class-discovery was performed and a support vector machine (SVM) constructed to assess diagnostic performance. The SVM classifier was adapted to use only metabolites (n = 10) routinely quantified from in vivo MRS data, and re-tested. Glutamate was assessed as a predictor of overall survival. Findings: Group-specific metabolite profiles were identified; tumours clustered with good concordance to their reference molecular group (93%). GABA was only detected in WNT, taurine was low in SHH and lipids were high in Group3. The tissue-based metabolite SVM classifier had a cross-validated accuracy of 89% (100% for WNT) and, adapted to use metabolites routinely quantified in vivo, gave a combined classification accuracy of 90% for SHH, Group3 and Group4. Glutamate predicted survival after incorporating known risk-factors (HR = 3.39, 95% CI 1.4–8.1, p = 0.025). Interpretation: Tissue metabolite profiles characterise medulloblastoma molecular groups. Their combination with machine learning can aid rapid diagnosis from tissue and potentially in vivo. Specific metabolites provide important information; GABA identifying WNT and glutamate conferring poor prognosis. Funding: Children with Cancer UK, Cancer Research UK, Children’s Cancer North and a Newcastle University PhD studentship.

Abstract The most common malignant brain tumour in children, medulloblastoma (MB), is subdivided into four clinically relevant molecular subgroups, although targeted therapy options informed by understanding of different cellular features are lacking. Here, by comparing the most aggressive subgroup (Group 3) with the intermediate (SHH) subgroup, we identify crucial differences in tumour heterogeneity, including unique metabolism-driven subpopulations in Group 3 and matrix-producing subpopulations in SHH. To analyse tumour heterogeneity, we profiled individual tumour nodules at the cellular level in 3D MB hydrogel models, which recapitulate subgroup specific phenotypes, by single cell RNA sequencing (scRNAseq) and 3D OrbiTrap Secondary Ion Mass Spectrometry (3D OrbiSIMS) imaging. In addition to identifying known metabolites characteristic of MB, we observed intra- and internodular heterogeneity and identified subgroup-specific tumour subpopulations. We showed that extracellular matrix factors and adhesion pathways defined unique SHH subpopulations, and made up a distinct shell-like structure of sulphur-containing species, comprising a combination of small leucine-rich proteoglycans (SLRPs) including the collagen organiser lumican. In contrast, the Group 3 tumour model was characterized by multiple subpopulations with greatly enhanced oxidative phosphorylation and tricarboxylic acid (TCA) cycle activity. Extensive TCA cycle metabolite measurements revealed very high levels of succinate and fumarate with malate levels almost undetectable particularly in Group 3 tumour models. In patients, high fumarate levels (NMR spectroscopy) alongside activated stress response pathways and high Nuclear Factor Erythroid 2-Related Factor 2 (NRF2; gene expression analyses) were associated with poorer survival. Based on these findings we predicted and confirmed that NRF2 inhibition increased sensitivity to vincristine in a long-term 3D drug treatment assay of Group 3 MB. Thus, by combining scRNAseq and 3D OrbiSIMS in a relevant model system we were able to define MB subgroup heterogeneity at the single cell level and elucidate new druggable biomarkers for aggressive Group 3 and low-risk SHH MB.

Abstract Brain tumors represent the highest cause of mortality in the pediatric oncological population. Diagnosis is commonly performed with magnetic resonance imaging. Survival biomarkers are challenging to identify due to the relatively low numbers of individual tumor types. 69 children with biopsy-confirmed brain tumors were recruited into this study. All participants had perfusion and diffusion weighted imaging performed at diagnosis. Imaging data were processed using conventional methods, and a Bayesian survival analysis performed. Unsupervised and supervised machine learning were performed with the survival features, to determine novel sub-groups related to survival. Sub-group analysis was undertaken to understand differences in imaging features. Survival analysis showed that a combination of diffusion and perfusion imaging were able to determine two novel sub-groups of brain tumors with different survival characteristics (p

Abstract Paediatric brain tumors are becoming well characterized due to large genomic and epigenomic studies. Metabolomics is a powerful analytical approach aiding in the characterization of tumors. This study shows that common cerebellar tumors have metabolite profiles sufficiently different to build accurate, robust diagnostic classifiers, and that the metabolite profiles can be used to assess differences in metabolism between the tumors. Tissue metabolite profiles were obtained from cerebellar ependymoma (n = 18), medulloblastoma (n = 36), pilocytic astrocytoma (n = 24) and atypical teratoid/rhabdoid tumors (n = 5) samples using HR-MAS. Quantified metabolites accurately discriminated the tumors; classification accuracies were 94% for ependymoma and medulloblastoma and 92% for pilocytic astrocytoma. Using current intraoperative examination the diagnostic accuracy was 72% for ependymoma, 90% for medulloblastoma and 89% for pilocytic astrocytoma. Elevated myo-inositol was characteristic of ependymoma whilst high taurine, phosphocholine and glycine distinguished medulloblastoma. Glutamine, hypotaurine and N-acetylaspartate (NAA) were increased in pilocytic astrocytoma. High lipids, phosphocholine and glutathione were important for separating ATRTs from medulloblastomas. This study demonstrates the ability of metabolic profiling by HR-MAS on small biopsy tissue samples to characterize these tumors. Analysis of tissue metabolite profiles has advantages in terms of minimal tissue pre-processing, short data acquisition time giving the potential to be used as part of a rapid diagnostic work-up.

The imaging and subsequent accurate diagnosis of paediatric brain tumours presents a radiological challenge, with magnetic resonance imaging playing a key role in providing tumour specific imaging information. Diffusion weighted and perfusion imaging are commonly used to aid the non-invasive diagnosis of children's brain tumours, but are usually evaluated by expert qualitative review. Quantitative studies are mainly single centre and single modality.The aim of this work was to combine multi-centre diffusion and perfusion imaging, with machine learning, to develop machine learning based classifiers to discriminate between three common paediatric tumour types.The results show that diffusion and perfusion weighted imaging of both the tumour and whole brain provide significant features which differ between tumour types, and that combining these features gives the optimal machine learning classifier with >80% predictive precision. This work represents a step forward to aid in the non-invasive diagnosis of paediatric brain tumours, using advanced clinical imaging. Keywords: Perfusion, Diffusion, Machine learning

IntroductionPhysiologically relevant pre-clinical ex vivo models recapitulating CNS tumor micro-environmental complexity will aid development of biologically-targeted agents. We present comprehensive characterization of tumor aggregates generated using the 3D Rotary Cell Culture System (RCCS).MethodsCNS cancer cell lines were grown in conventional 2D cultures and the RCCS and comparison with a cohort of 53 pediatric high grade gliomas conducted by genome wide gene expression and microRNA arrays, coupled with immunohistochemistry, ex vivo magnetic resonance spectroscopy and drug sensitivity evaluation using the histone deacetylase inhibitor, Vorinostat.ResultsMacroscopic RCCS aggregates recapitulated the heterogeneous morphology of brain tumors with a distinct proliferating rim, necrotic core and oxygen tension gradient. Gene expression and microRNA analyses revealed significant differences with 3D expression intermediate to 2D cultures and primary brain tumors. Metabolic profiling revealed differential profiles, with an increase in tumor specific metabolites in 3D. To evaluate the potential of the RCCS as a drug testing tool, we determined the efficacy of Vorinostat against aggregates of U87 and KNS42 glioblastoma cells. Both lines demonstrated markedly reduced sensitivity when assaying in 3D culture conditions compared to classical 2D drug screen approaches.ConclusionsOur comprehensive characterization demonstrates that 3D RCCS culture of high grade brain tumor cells has profound effects on the genetic, epigenetic and metabolic profiles of cultured cells, with these cells residing as an intermediate phenotype between that of 2D cultures and primary tumors. There is a discrepancy between 2D culture and tumor molecular profiles, and RCCS partially re-capitulates tissue specific features, allowing drug testing in a more relevant ex vivo system.