NEW YORK (GenomeWeb) – The DNA methylation patterns found in glioblastoma (GBM) tumors may offer survival hints for individuals with the brain cancer, new research suggests.
As they reported in Nature Medicine yesterday, researchers from the Medical University of Vienna and other centers did reduced representation bisulfite sequencing (RRBS)-based methylation profiling on 112 formalin-fixed, paraffin-embedded (FFPE) tumor samples from well-characterized GBM cases in a national patient registry, including a subset of cases with matched primary and progression tumors.
"[O]ur study establishes a rich and openly available resource characterizing the DNA methylation dynamics of glioblastoma progression in a highly annotated clinical cohort with matched [magnetic resonance imaging] and detailed histopathological data," co-senior authors Christoph Bock and Adelheid Woehrer, at the Medical University of Vienna, and their colleagues wrote.
In samples collected over time and/or from different tumor sections in each individual, for example, the team saw variable and shifting methylation profiles that appeared to reflect progression events as well as methylation links to the tumor microenvironment. It also detected methylation differences that seemed to coincide with patient outcomes, including lower-than-usual methylation at the promoters of MGMT and Wnt signaling genes in individuals with shorter survival times.
"Given that robust protocols are available for measuring DNA methylation in routine clinical diagnostics, epigenetic biomarkers are likely to contribute to improved diagnosis, prognosis, and personalized therapy in glioblastoma and other cancers," the authors wrote.
The team began with extensively profiled, IDH mutation-free primary tumors from individuals in the Austrian Brain Tumor Registry who were sampled two to four times, including a subset of patients with tested after progression.
Using RRBS, the researchers profiled an average of more than 1.8 million CpG sites in 283 FFPE tumor samples collected at different time points or tumor sections from the same patient, setting the data alongside MRA, pathology, clinical, and other data collected for the same individuals at the same time point.
To that, they added data for another 105 individuals with IDH-wildtype GBM tumors and samples from more than a dozen individuals with IDH-mutated secondary GBM, oligodendroglioma, or astrocytoma.
In the individuals with available progression tumor samples, the team detected progression-related DNA methylation gains at the promoters of genes implicated in processes such as apoptosis and neural development, along with waning methylation in the MGMT promoter and in Wnt signaling gene promoters, particularly in individuals with poorer survival patterns.
In addition, the researchers showed that it was possible to bioinformatically obtain copy number insights and other genetic or transcriptomic clues related to tumor subtype based on RRBS data, with a machine learning approach built from the DNA methylation, histological clues, clinical annotation, MRI patterns, and other data in the original set of well-characterized GBM tumors.
"The RRBS data allowed us to infer a broad range of tumor properties, including known biomarkers such as G-CIMP status, MGMT promoter methylation, and chromosome 1p19q co-deletion," the authors noted. "We also established the utility of RRBS data for predicting glioblastoma transitional subtypes, thus extending this candidate biomarker to routinely collected FFPE samples."