Virchow 2011 or How to ID(H) Human Glioblastoma

Abstract

In 1865, Virchow was the first to describe gliomas pathologically, segregating them into two groups that we now recognize as lowand high-grade gliomas. In 1926, Bailey and Cushing gave the first modern pathologic description of the most malignant brain tumor, the glioblastoma, calling it spongioblastoma multiforme. They differentiated spongioblastoma multiforme from astrocytoma on the basis of their presumption of a different cell of origin, given the different histologic appearance of the two. By 1940, the term glioblastoma multiforme (GBM) was in widespread use, and Scherer was the first to differentiate between primary and secondary GBM, distinguishing those tumors that arose de novo from those that transformed from a preexisting, lower-grade lesion. The term multiforme referred to the extraordinarily diverse histologic features of these neoplasms. Their variegated appearance could range from small, intense blue cells to huge, multinucleated giant cells, and these wide variations could coexist within the same tumor. The extraordinary histologic heterogeneity may prove challenging for the neuropathologist, but these varying histologic subtypes did not affect prognosis or response to known therapies, although Scherer recognized that secondary glioblastoma was associated with longer survival. The advent of modern molecular profiling, however, is changing our understanding of glioblastoma, and the term multiforme may have a new meaning. The first attempts to segregate molecular subtypes of glioblastoma began with the study of primary and secondary glioblastoma, which also have two discrete clinical presentations. Primary GBM represents approximately 95% of all glioblastomas and tends to occur in older patients, whereas secondary GBM arises in younger individuals who have had a previously diagnosed low-grade glioma for many years or even decades before malignant transformation heralds a sudden change in the status of their disease. Primary GBM is characterized by the high frequency of epidermal growth factor receptor (EGFR) gene amplification, whereas p53 mutations characterize secondary GBM. This initial work established that dysfunction of different pathways could lead to histologically identical GBM. The most recent integrated genomic profiling of GBM has revealed four molecular-genetic subtypes that are defined by their most prominent gene expression alterations and are associated with specific mutations and gene copy number alterations. These disease subtypes are referred to as classical (EGFR amplification/ overexpression or mutation, phosphatase and tensin homolog [PTEN] loss or mutation, CDNK2A loss, nuclear export sequence [NES] overexpression, activation of Notch and sonic hedgehog pathways); mesenchymal (NF1 loss, p53 mutation, PTEN loss/mutation, activation of tumor necrosis factor and nuclear factorB pathways); proneural (platelet-derived growth factor receptor-A amplification, isocitrate dehydrogenase 1 [IDH1] mutation, high expression of proneural, eg, Sry-box [SOX], and oligodendroglial, eg, OLIG2, genes); and neural (EGFR amplification/overexpression, expression of neural markers, eg, neurofilament light polypeptide [NEFL]). The subtypes have prognostic significance and defined activity of signaling pathways that characterizes gliomagenesis. Earlier studies identified three subtypes: mesenchymal, proneural, and proliferative. Patients with proneural glioblastomas (as defined by either classification schema) lived longer, particularly compared with those patients with the mesenchymal subtype, in which angiogenesis features prominently. In the article that accompanies this editorial, Lai et al provide additional clarity with respect to GBM subclassification. Mutations in the IDH genes, particularly the IDH1, have been shown by several groups to represent some of the earliest mutations in glioma formation and are particularly characteristic of lowgrade gliomas; they are seen in more than 70% of grade 2 gliomas and in 50% of grade 3 lesions. These data suggest that IDH mutations occur early and may initiate glioma formation. Secondary glioblastomas also have a high propensity to harbor IDH mutations, which confirms their origin from the prior lower-grade tumor. Lai et al explored the presence of IDH mutations in what seem clinically to be primary glioblastomas. They examined 618 primary GBMs for the presence of the most common IDH mutation in glioma, IDH1, and 49 such tumors were identified, suggesting that approximately 8% of apparent primary GBMs seem to arise from clinically silent lower-grade tumors. Moreover, these 49 tumors had clinical and pathologic features that identified them as a discrete subgroup. Histologically, they had less necrosis, more oligodendroglial features, and a significantly greater proportion of tumors with promotor methylation of O-methylguanine DNA methyltransferase (MGMT), which was a facet of a larger pattern of CpG island hypermethylation. Clinically, IDH1 GBMs had more radiographic features that were reminiscent of low-grade glioma, including nonenhancing disease, less edema, larger size, increased prevalence of cystic and diffuse components, and frontal lobe predominance. Most significantly, patients whose GBM contained an IDH1 mutation had a significantly longer survival than those with wild-type IDH1. The authors then examined the transcriptional signatures of both groups of glioblastomas and found that the majority of IDH1 tumors expressed the proneural signature, originally described in 2006, which is substantially similar to the proneural signature that was JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L S