• National Cancer Institute
  • National Human Genome Research Institute

Posted: February 1, 2011

TCGA Scientists Discover Four Distinct Subtypes of Glioblastoma Distinguished by Gene Expression Patterns and Clinical Characteristics

Catherine Evans

Using its previously published description of genomic changes that drive glioblastoma (GBM) tumor development, The Cancer Genome Atlas (TCGA) researchers have now established the existence of four subtypes of GBM. These subtypes, described in a 2010 Cancer Cell article, are defined by genomic characteristics, survival length, patient age and treatment response. The scientists’ grouping of patients by GBM subtype is an important first step toward developing personalized treatments that target the unique gene changes in each subtype.

The scientists identified distinct groups of patients, separating them according to differing patterns of gene expression. The four groups were named Proneural, Neural, Classical and Mesenchymal. The GBM subtypes were found to differ by the type of genetic abnormalities they carried and by the patients’ clinical characteristics.

GBM Subtypes Link to Patient Outcome

Classical GBM tumors are characterized by abnormally high levels of epidermal growth factor receptor, or EGFR. EGFR is a protein found on the surface of some cells that, when bound by epidermal growth factor, sends signals for the cell to keep growing. The EGFR abnormalities occur at a much lower rate in the three other GBM subtypes. However, TP53, the most frequently mutated gene in GBM, is not mutated in any of the Classical GBM tumors. TP53 is the gene for a protein that normally suppresses tumor growth. Clinically, the Classical group survived the longest of the subgroups in response to aggressive treatment.

Unlike in Classical tumors, TP53 is significantly mutated in Proneural tumors (54 percent). Proneural tumors are also characterized by having the most frequent mutations in the IDH1 gene. IDH1, when mutated, codes for a protein that can contribute to abnormal cell growth. Another gene, PDGFRA, was mutated and expressed in abnormally high amounts only in the Proneural tumors and not in any other subgroups. When PDGFRA is altered, too much of its protein can be produced, leading to uncontrolled tumor growth. Unlike the other groups, whose patients were similar in age on average, the Proneural subgroup was significantly younger. They also tended to survive longer. However, patients in the Proneural group who received aggressive treatment did not survive significantly longer than Proneural patients who did not receive aggressive treatment. Clinicians may be able to use this information in the future to avoid unnecessary treatment regimens for patients in the Proneural subgroups.

The Mesenchymal subgroup contains the most frequent number of mutations in the NF1 tumor suppressor gene (37 percent). Frequent mutations in the PTEN and TP53 tumor suppressor genes also occurred in the group. Patients in the Mesenchymal group had significant increases in survival after aggressive treatment, unlike those in the Proneural, and to an extent, in the Neural subgroups.

While the Neural subgroup had mutations in many of the same genes as the other groups, the group did not stand out from the others as having significantly higher or lower rates of mutations. The Neural group was characterized by the expression of several gene types that are also typical of the brain’s normal, noncancerous nerve cells, or neurons. Patients in the Neural group were the oldest, on average. They also had some improvement in survival after aggressive treatment, but not as much as the Classical and Mesenchymal groups.

These findings will put scientists and clinicians on the path toward more individualized treatment of GBM patients. The results from this study could lead to diagnostic tests that will classify a GBM patient into one of the four subgroups. The added knowledge about what sets each group apart will lead to treatments that are tailored to the unique pattern of genomic changes within each patient’s tumors.


Verhaak, R.G., Hoadley, K.A., Purdom, E., Wang, V., Qi, Y., Wilkerson, M.D., Miller, C.R., Ding, L., Golub, T., Mesirov, J.P., Alexe, G., et al. (2010) Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell. 17(1):98-110. View PubMed abstract.