What can metabolites tell us about gliomas?

The uncontrolled proliferation of cancer cells requires corresponding metabolic adjustments. A deregulated bioenergetic metabolism is thus considered a core feature of malignant neoplasms.^1,2 This viewpoint has been extensively validated² in different types of tumors over the past decade, and the understanding of its molecular foundation has advanced considerably, in part due to the power of metabolomics. Along with identifying oncometabolites, such as (R)-2-hydroxyglutarate (R-2-HG),³ metabolomic studies have revealed tumor-specific signatures, provided novel mechanistic insights into cancer phenotypes, and enhanced our understanding of drug action.⁴

As reported in this issue of Neuro-Oncology, Björkblom et al⁵ used multi-platform metabolomics to expand our knowledge related to glioma metabolism. The authors interrogated the characteristics of a large series of adult glial tumors (n = 224) using gas chromatography- and liquid chromatography mass spectrometry (GC-MS and LC-MS). Quantification of 1132 metabolic features, with 240 unique metabolites identified, was integrated with genetic, clinical, and pathological analyses, thereby offering a comprehensive view of the metabolic phenotypes for six WHO classified adult glioma subtypes.

A key finding of the study is that isocitrate dehydrogenase (IDH)-wildtype tumors (n = 172) and IDH-mutated tumors (n = 52) had a significantly different metabolic profile, even if 2-HG, present in high levels in IDH-mutated tumors, was excluded from the analysis. While the result might not seem surprising, it has several interesting implications.

Tumor metabolism has a cell-autonomous core, but it is also highly influenced by extrinsic factors and can, in turn, influence the microenvironment. Autonomous determinants include genetic aberrations and oncogene addiction, whereas environment-driven effects can be traced back to nutrient and oxygen availability and interactions with non-cancerous cells as well as with the extracellular matrix.⁶ The fact that the WHO classified adult glioma subtypes have different metabolic profiles suggests robustness and specificity of the cell-autonomous metabolic component for each subtype and confirms the potential of metabolic phenotypes as major glioma descriptors. The metabolic separation also indicates the possibility that, after careful evaluation of sensitivity and specificity in prospective settings, metabolomics data may reveal yet elusive diagnostic biomarkers for one or several subtypes of malignant gliomas.

Furthermore, the metabolism of cancer cells has been shown to influence several non-malignant cells in the microenvironment. For instance, endothelial and immune cells can be affected by nutrient depletion or by tumor cell-derived metabolites. Such influence opens the possibility that glioma subtypes with different metabolic profiles could elicit different immune responses. Indeed, observations from experimental models of IDH-mutated gliomas suggest that R-2-HG may at least partially contribute to decreased intratumoral T-cell infiltration and activity.^7,8 Of note, the work by Björkblom et al also uncovered IDH status-independent metabolic differences between astrocytomas, oligodendrogliomas, and glioblastomas, for instance, the level of acylcarnitines. It is possible that such differences might also induce yet unrecognized changes in the tumor microenvironment.

Another important finding of the study is that the metabolic profile of IDH-wildtype glioblastomas in the <45-year-old age group was significantly different from that in the >70-year-old age group. Again, this could reflect a difference in intrinsic factors such as overall mutational burden, or, given the similarities between IDH-wildtype glioblastomas in the <45-year-old group and lower-grade astrocytomas (IDH-wildtype), it could also reflect a difference in cell-of-origin. However, extrinsic factors could also play an important role. Future studies integrating plasma/serum metabolomics, genomics, metagenomics, and dietary details might help dissect the influence of systemic metabolism on age-specific metabolic phenotypes.

Like other -omics approaches, exploratory metabolomics have inherent limitations. They do not inform causal relationships and need to be validated in mechanistic and prospective clinical studies. For genomics data, such validation has successfully introduced a layer of molecular information to the WHO classification of gliomas. It will be exciting to see what metabolomics can offer in this regard.