Prognostic value of integrative genomic approaches for IDH-mutant gliomas

The transition from histology- to molecular-based classification of tumor entities has been one of the most transformative trends in neuro-oncology over the last decade. This transformation in the way malignancies are defined and categorized is a testament to the contributions of genomic methods to the study of brain tumors. Genomic and other molecular tests have traditionally aimed at reducing subjectivity in classification of disease and are now key elements in the push toward precision medicine. In neuro-oncology, this trend was codified by World Health Organization classification of central nervous system malignancies in 2021 (reviewed in1). The new classification represented a relatively radical reorganization of central nervous system tumors entities recognized based on salient molecular features, including gene expression or methylation profiles, or the presence of specific highly recurrent mutations. Although the practical clinical applications of the 2021 classification are still debated for some tumor types, the new groupings capture crucial biological differences that have been validated in preclinical models and hold the promise to make therapy more precise and effective. The World Health Organization now classifies adult-type diffuse gliomas into 3 categories that largely reflect established histopathological features and genetic information. These 3 categories are “Astrocytoma, IDH-mutant,” “Oligodendroglioma, IDH-mutant and 1p/19q-codeleted” and “Glioblastoma, IDH-wild type.” This classification reflects the importance of mutations in IDH1—and more rarely IDH2—in the assessment of the aggressiveness of the disease and overall disease progression.2,3 In fact, the 2 IDH-mutant (IDHmut) entities are much less aggressive than IDH-wild-type glioblastoma. IDHmut tumors usually are diagnosed in younger patients and initially respond to treatment, leading to good prognoses and some individuals living decades after diagnosis. Although IDHmut gliomas have generally good prognoses compared to glioblastoma, they are nonetheless heterogeneous malignancies spanning grades 2, 3, and 4 and a wide range of patient outcomes. In this issue, Mamatjan et al.4 asked whether it is possible to identify molecular markers of good and poor prognosis for IDHmut gliomas. They interrogated a cohort accrued locally at the University Health Network in Toronto, Canada, using DNA methylation data. By correlating individual methylation probes with information on outcomes, the authors derived a signature that could stratify patients. The DNA methylation signature was validated across 2 independent patient cohorts: One collected by The Cancer Genome Atlas5 and one by the DKFZ in Germany. A striking feature of their signature of poor prognosis was high DNA methylation at HOX genes. To further investigate this signature of poor prognosis in IDHmut gliomas, the authors employed an integrative genomic approach that included information on DNA methylation, transcriptomes, and genetic variation for samples in The Cancer Genome Atlas cohort. Using what they defined as integrated RNA and methylation (iRM) approach, they clustered IDHmut samples into 4 groups based on transcriptional levels and DNA methylation at HOX loci: (1) Low methylation and low transcription, (2) low methylation and high transcription, (3) high methylation and low transcription, and (4) high methylation and high transcription. They found that the first and fourth categories were informative of overall survival. Specifically, samples with low DNA methylation and low transcriptional levels of HOX genes (low iRM group) were associated with better prognosis, whereas samples with high DNA methylation and high transcriptional levels of HOX genes (high iRM group) were associated with poor prognosis. They also found that the high iRM group displayed increased mutational burden and aneuploidy compared to the iRM low group. Digging deeper into the high iRM group, Mamatjan et al. found that 7 HOX genes were sufficient to establish a signature of poor outcome. These HOX genes included HOXA4, HOXA7, HOXA10, HOXA13, HOXD3, HOXD9, and HOXD10. The HOX signature was predictive of overall survival in both 1p/19q co-deleted and non-deleted cases. The findings of this paper cement HOX gene expression as a strong correlate of poor prognosis in brain tumors, including glioblastoma6,7 and IDHmut gliomas.8 However, the malignant roles of HOX genes extend beyond brain tumors, as they are highly expressed in a variety of blood and solid cancers (reviewed in detail in Bhatlekar et al.9). This work therefore contributes to growing evidence that HOX gene signatures could be predictive of outcomes in many diverse cancer types, although the individual HOX genes in these signatures may depend on tumor site. Integrative approaches to find such tumor type-specific HOX signatures could contribute to patient stratification and aid in patient management in clinical settings. More work will be needed to fully appreciate the biological and clinical implications of the association between HOX-high signatures and poor prognosis. Although mechanistic studies of IDHmut gliomas are made difficult by the lack of patient-derived models, future preclinical work will need to look at combinatorial treatment approaches that specifically target the molecular underpinnings of tumors with high HOX signatures. From a biological standpoint, what is causing the DNA hypermethylation at some HOX genes? The link between the downstream epigenetic effects of the IDH mutation and DNA methylation10 at HOX clusters should be looked at in more detail. It is also possible that the high and low iRM states may be continuous and not discrete, representing a gradation in molecular phenotypes among IDHmut gliomas. If the signature is continuous, then it might be important to look for other molecular correlates that could enable more precise identification of patients at risk. Finally, the association between the high HOX iRM signature and aneuploidy and mutation burden found by the authors is interesting. More work is needed to fully elucidate the mechanistic connection between the iRM signature and mutational profiles. This observation also raises the question of whether the high HOX iRM signature is downstream of aneuploidy, or if the opposite is true. The text is the sole product of the author and no third party had input or gave support to its writing.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.