Study describes new glioma subtypes

Findings of research conducted at the University of São Paulo and published in Cell can help evaluate the prognosis of patients with central nervous system tumors and improve treatment protocols.

By Karina Toledo  |  Agência FAPESP – In an international study conducted in Brazil and published on January 28 in the journal Cell, researchers have identified new glioma subtypes on the basis of epigenetic profiles, i.e., how gene expression is modulated.

According to the authors, the discovery may help evaluate the prognosis of patients with gliomas and may pave the way for personalized treatment. Gliomas are tumors of the central nervous system.

Houtan Noushmehr, one of the principal investigators for the project and a professor at the Genetics Department of the University of São Paulo’s Ribeirão Preto School of Medicine (FMRP-USP), said that epigenetic modifications correspond to a set of chemical processes that shape genome functioning and thus the phenotypic profile by activating or deactivating genes. If the genome is comparable to the hardware of a computer, Noushmehr explained, then epigenetics is the software that makes it work.

Among the most widely known epigenetic mechanisms are DNA methylation, the addition of a methyl group (carbon and hydrogen atoms) to the DNA base cytosine, which potentially prevents the expression of some genes, and histone modification, in which acetyl, methyl or other groups are added to or subtracted from the amino acid residues in histones.

The study conducted at FMRP-USP was designed to evaluate DNA methylation profiles in 1,122 adult glioma samples from The Cancer Genome Atlas (TCGA), a joint effort of the US National Cancer Institute and National Human Genome Research Institute, both of which are linked to the National Institutes of Health (NIH). TCGA collects genomic, epigenomic and clinical data as well as tissue samples from cancer patients in many countries.

“We had various types of data available for analysis, such as complete tumor genome sequences, complete exome sequences [all the protein-coding genes in genomes], RNA sequencing and DNA methylation data, and chromosome copy number alterations. All these data enabled us to map the epigenetic profiles of tumors. We also had the patients’ clinical histories and other details showing how the disease progressed,” Noushmehr said.

The analysis was conducted with the aid of bioinformatics tools as part of the postdoctoral research of Tathiane Malta, a FAPESP grantee, and during the master’s research of Thais Sabedot, both under the aegis of a FAPESP Young Investigators Award coordinated by Noushmehr.

New division

According to the scientists, until recently, adult gliomas were divided according to their aggressiveness into two main groups: slow-growing tumors (grades 1, 2 and 3) and very aggressive, fast-growing tumors (grade 4).

Previous studies published by Noushmehr in 2010, 2013 and 2015 separately evaluated samples from high- and low-grade tumors and identified tumor subgroups by their DNA methylation profiles.

One such subtype, for example, is G-CIMP (Glioma CpG Island Methylator Phenotype), a term coined by Noushmehr to describe a hypermethylation phenotype associated with a better prognosis in high-grade gliomas.

In this most recent study, samples from high- and low-grade tumors were analyzed together.

“Patients with low-grade tumors typically have a mutation in the IDH1 or IDH2 gene. However, we observed that some low-grade tumors didn’t have this mutation, while some high-grade tumors did. We therefore decided to analyze them all together in search of a more accurate classification. Generally speaking, we found that tumors with the IDH1 mutation had a DNA methylation phenotype and a better prognosis,” Malta said.

At the end of the analyses, tumors from patients with the IDH1 mutation and 1p/19q chromosome co-deletion, previously considered a homogeneous group, were divided into two new subgroups with different clinical outcomes: G-CIMP-low (low DNA methylation profile and shorter patient survival) and G-CIMP-high (hypermethylation and longer survival).

In the group without the IDH1 mutation, which should have been the most aggressive cases according to previous criteria, the researchers identified a new subgroup with low aggressiveness that displayed similarities with pilocytic astrocytoma, another type of central nervous system tumor that usually has a good prognosis. Survival for patients with this type of tumor that was similar to astrocytoma was significantly longer than for patients without the mutation.

“In terms of histopathology, i.e., the microscopic anatomical changes in diseased tissue, these are different tumors, but from the molecular standpoint, they’re similar. This finding is very important and will be explored in more detail in future studies,” Malta said.

For Noushmehr, the new findings serve to stratify glioma patients more accurately and will contribute to an enhancement of treatment protocols.

“They serve as a basis for confining more aggressive treatment to patients who genuinely need it,” he said. “With the current approach, some patients don’t respond well to drugs, and this may perhaps be explained by one of these phenotypes we report, such as G-CIMP-low, for example. The findings will also be useful for the development of new drugs because they point to therapeutic targets for patients with the worst prognosis.”

According to Malta, until recently, most cancer research focused on efforts to pinpoint genetic mutations and DNA sequence alterations. Now, however, scientists have also begun to observe how genomic regulation is disrupted in cancer.

“In some cancers, the role of epigenetic regulation, especially DNA methylation, appears to have a strong influence,” Malta said. “There’s evidence that epigenetics may be involved in the genesis of cancer, as well as its development and aggressiveness. But we don’t yet know which is cause and which is effect.”

The paper Molecular Profiling Reveals Biologically Discrete Subsets and Pathways of Progression in Diffuse Glioma (doi: http://dx.doi.org/10.1016/j.cell.2015.12.028), from Michele Ceccarelli et al., can be read by Cell subscribers at www.cell.com/cell/abstract/S0092-8674(15)01692-X.