Multi‐omic integration of DNA methylation and gene expression data reveals molecular vulnerabilities in glioblastoma

Abstract

Glioblastoma (GBM) is one of the most aggressive types of cancer and exhibits profound genetic and epigenetic heterogeneity, making the development of an effective treatment a major challenge. The recent incorporation of molecular features into the diagnosis of patients with GBM has led to an improved categorization into various tumour subtypes with different prognoses and disease management. In this work, we have exploited the benefits of genome‐wide multi‐omic approaches to identify potential molecular vulnerabilities existing in patients with GBM. Integration of gene expression and DNA methylation data from both bulk GBM and patient‐derived GBM stem cell lines has revealed the presence of major sources of GBM variability, pinpointing subtype‐specific tumour vulnerabilities amenable to pharmacological interventions. In this sense, inhibition of the AP‐1, SMAD3 and RUNX1/RUNX2 pathways, in combination or not with the chemotherapeutic agent temozolomide, led to the subtype‐specific impairment of tumour growth, particularly in the context of the aggressive, mesenchymal‐like subtype. These results emphasize the involvement of these molecular pathways in the development of GBM and have potential implications for the development of personalized therapeutic approaches.

MOTL: enhancing multi-omics matrix factorization with transfer learning

“…Glioblastoma is a rare, heterogeneous, and aggressive cancer type. Multi-omics datasets offer an important opportunity to better characterize glioblastoma subtypes, identify biomarkers, and propose novel therapeutic options (Santamarina-Ojeda et al, 2023). However, large collections of glioblastoma tissue samples are difficult to obtain due to the relative scarcity of the disease and the challenges involved in acquiring samples via invasive biopsies.…”

“…In Santamarina-Ojeda et al (2023), the authors conducted a multi-omics profiling (mRNA expression, DNA methylation) for four normal brain samples and nine patient-derived glioblastoma stem cell (pd-GBSC) cultures. The nine cancer samples had been previously classified into three subtypes thanks to transcriptome-based signatures: classical (CL), proneural (PN), and mesenchymal (MS).…”