Brain Tumor . Glioma Development . Tumor Invasion . Cancer Genetics . Biomarker

Brain Tumor Biology

Glioma development: from biomarker identification to molecular mechanisms

Gliomas are among the deadliest cancers, with median survivals varying between months for the malignant grade IV glioblastoma (GBM), to decades for low-grade glioma (LGG). Gliomas progress by brain tissue invasion. The aim of our Laboratory is to understand mechanisms underlying glioma invasion. This involves the identification of biomarkers, genetic regulators, signaling networks and molecular effectors of invasion that can ultimately be targeted to control glioma progression.Through an active exchange between clinics and our laboratory, we are collecting freshly resected glioma biopsies for genotyping and ex vivo cell culture. In parallel,we are entering personal, clinical, imaging, histopathological and molecular annotations to construct a comprehensive glioma patient database. This information is useful for classifying gliomas into molecular subsets and allowing identification of biomarkers that may reveal novel glioma pathways.

IDH mutations in low-grade gliomas
IDH neomorphic mutation (IDHmut), found in 80% LGG, catalyzes α-ketoglutarateconversion into 2-hydroxyglutarate, accumulation of which maintains CpG methylation.This results in MGMT epigenetic silencing and TP53 CpG-to-CpA mutational transition. We stratified 210 LGG according to these molecular criteria in a retrospective study. Although IDHmut status is associated with a lower risk of death (HRdeath=0.35, P=0.0023), IDHmut subsets consistently showed higher risks of malignant transformation (MT) than of death. This supports the finding that molecular events relevant to IDH mutation impacts early glioma development prior to MT (see Leu et al., 2016).

The interplay of 3q26 clustered genes SOX2, PIK3CA, MFN1 and OPA1 in GBM cell invasion
Chromosome band 3q26, frequently altered in GBM, contains the genes for transcription factor SOX2, growth factor/AKT signaling activator PIK3CA, and MFN1and OPA1, two effectors of mitochondria fusion, a process linked to inhibition ofcell motility (Fig. 1A). We aimed at determining the roles of these four genes in GBM cell invasion.
Compared to parental LN319 GBM cells, individual 3q26 gene knock-downs consistently shortened mitochondria, and enhanced cell invasion (Fig. 1BC). These phenotype similarities suggested that these 3q26 genes act on a common invasion pathway. Pharmacological inactivation of AKT, downstream of PIK3CA, impairsSOX2 nuclear localization and aggravates SOX2 turnover (Fig. 2A). Chromatin immuno-precipitation and luciferase reporter gene assays show that SOX2 trans-activatesPIK3CA, MFN1 and OPA1 (Fig. 2BC). This indicates a positive regulation loop where AKT signaling activates SOX2 function, which in turn activates PIK3CA,MFN1 and OPA1 transcription. Copy number variations at 3q26 analyzed in 100glioma biopsies show frequent SOX2 gain (29%) and OPA1 loss (32%) (Fig. 3A).SOX2 amplification is consistent with enhanced transcriptional activation of oncogenic targets such as PIK3CA. In contrast, OPA1 needs to be lost to counteract the impact of SOX2 on mitochondria fusion and invasion suppression. Thus,among the genes trans-activated by SOX2, are oncogenic and tumor suppressor genes. While the oncogenic ones (PIK3CA) are kept ‘on’, OPA1 may need tobe turned ‘off’ by deletion. Thus, we provide evidence that a regional interplay between3q26 genes promotes glioma invasion (Fig. 3B). Copy number variations of3q26 genes suggest optimization of these oncogenic activities and are currently being tested for their impact on tumor invasion in glioma patients (see connection to clinical practice).

Combined expression of nestin and SPARC identifies isolated astrocytoma cells in brain tissue
Identification of individual invasive glioma cells in brain tissue requires markers that specifically recognize tumor cells. We tested the presence of proteins involved in glioma development (proliferation, survival, invasion, differentiation, transcription and metabolism) on a tissue micro-array that contains brain sections at various distances from the tumor core and various glioma histology and grade. Glial progenitor marker nestin together with secreted acidic and rich in cysteine (SPARC)represent a specific combination for recognizing glioma cells in a non-neoplastic environment (see Aljammal et al., 2015).

Fig. 1

Fig. 2

Fig. 3