IDH-Mutant Gliomas: DNA Hypomethylation Linked to Persistence of Stem-Like States
Background: Tracking cell-state changes as tumors progress
A study published in Nature Genetics reports that reduced DNA methylation in isocitrate dehydrogenase (IDH)-mutant gliomas is associated with an increase in stem-like states and less cellular differentiation. The analysis included 36 tumor samples from 19 patients. The researchers investigated how epigenetic changes accompanying tumor progression relate to the maintenance of and transitions between cell states.
IDH-mutant gliomas may grow slowly at first, but undergo genetic changes and changes in cellular composition as they progress. In some progressing tumors, the glioma CpG island methylator phenotype (G-CIMP) shifts from a high to a low state, a change known to be associated with shorter survival. However, the relationship between reduced methylation and tumor-cell differentiation had not been adequately explained. Key limitations of earlier single-cell studies included small sample sizes, a lack of matched samples collected from the same patients at different time points, and limited characterization of genetic features.
Key findings: Fewer transitions toward differentiation in hypomethylated tumors
The researchers analyzed frozen tumor samples collected at three institutions. Of the 36 samples, 32 were longitudinally matched samples from 15 patients; the remaining four were unmatched initial or recurrent tumor samples. They combined an assay that reads RNA and DNA methylation in the same nucleus with bulk whole-exome analysis. In a separate large-scale single-nucleus RNA analysis, 158,143 nuclei from 32 samples passed quality control, and these data were used to assess cellular composition.
Recurrent tumors had a higher proportion of stem-like states and a lower proportion of differentiated states. G-CIMP scores, which indicate methylation levels, also tended to decline at recurrence. Within the same tumor, however, methylation levels were generally similar across distinct malignant cell states. This makes it difficult to explain tumor-wide hypomethylation solely by an increase in stem-like cells that were already less methylated.
The researchers reconstructed cell lineage trees from DNA methylation patterns and applied PATH, an analytical framework for inferring the maintenance of and transitions between cell states within lineages. In all five patients with matched samples whose tumors shifted from G-CIMP-high to G-CIMP-low, the probability that stem-like states in recurrent tumors would transition to differentiated states was estimated to be lower. This was an inference combining lineage trees and cell-state information, rather than a real-time observation of cellular changes. State heritability in this analysis refers to the degree to which a state is maintained within a tumor-cell lineage.
The study also identified candidate links between hypomethylation and stem-like states: hypomethylation at regions targeted by polycomb repressive complex 2 (PRC2), increased expression of stem-cell-related genes, and enrichment of key transcription-factor binding motifs in hypomethylated regions. None of these observations individually establishes a causal relationship involving a specific pathway.
Significance and outlook: Potential prognostic marker and remaining validation
A reanalysis of The Cancer Genome Atlas (TCGA) data found that, in IDH-mutant oligodendrogliomas as well, the subgroup with lower methylation scores was associated with worse overall survival. This provides a basis for extending investigation of the prognostic relevance of hypomethylation from astrocytomas to oligodendrogliomas.
The findings support a model in which hypomethylated tumors show less differentiation and maintain stem-like states. Limitations remain, including the limited number of samples and cells available for joint analysis, as well as the many genomic locations not measured in single-cell methylation data. The data used were also not suited to identifying point mutations in individual cells. The causal relationship between hypomethylation and cell-state changes, and whether methylation scores can predict treatment response, require validation in larger cohorts and follow-up experiments.
The study design and results are available in the Nature Genetics paper and the full text at Europe PMC.
Nature Genetics, Published online: 22 June 2026; doi:10.1038/s41588-026-02642-7The authors use single-cell multi-omics to interrogate longitudinal samples of isocitrate dehydrogenase mutant gliomas. Their analyses couple progression and hypomethylation with changes in differentiation topologies and increased emergence of stem-like states in aggressive disease.
This study provides a rationale for adding DNA methylation information to pathological grade and genetic alterations in research evaluating prognosis in IDH-mutant gliomas. For example, methylation scores could be measured in independent cohorts of patients with astrocytoma and oligodendroglioma to test whether they add information on survival prediction to existing classifications. Predictive performance and reproducibility would need to be assessed separately before use as a clinical test.
Drug-development studies could examine the relationship between methylation status before treatment with mutant IDH inhibitors and differentiation and response after treatment. The authors also identify the potential to stratify patients according to treatment response as a topic for future work. This study alone does not establish a basis for predicting drug resistance or selecting treatment, nor does it validate performance for early detection of recurrence or for use as a companion diagnostic.