A glutamatergic biomarker panel enables differentiating Grade 4 gliomas/astrocytomas from brain metastases

Abstract

Background: The differentiation of high-grade glioma and brain tumors of an extracranial origin is eminent for the decision on subsequent treatment regimens. While in high-grade glioma, a surgical resection of the tumor mass is a fundamental part of current standard regimens, in brain metastasis, the burden of the primary tumor must be considered. However, without a cancer history, the differentiation remains challenging in the imaging. Hence, biopsies are common that may help to identify the tumor origin. An additional tool to support the differentiation may be of great help. For this purpose, we aimed to identify a biomarker panel based on the expression analysis of a small sample of tissue to support the pathological analysis of surgery resection specimens. Given that an aberrant glutamate signaling was identified to drive glioblastoma progression, we focused on glutamate receptors and key players of glutamate homeostasis.

Methods: Based on surgically resected samples from 55 brain tumors, the expression of ionotropic and metabotropic glutamate receptors and key players of glutamate homeostasis were analyzed by RT-PCR. Subsequently, a receiver operating characteristic (ROC) analysis was performed to identify genes whose expression levels may be associated with either glioblastoma or brain metastasis.

Results: Out of a total of 29 glutamatergic genes analyzed, nine genes presented a significantly different expression level between high-grade gliomas and brain metastases. Of those, seven were identified as potential biomarker candidates including genes encoding for AMPA receptors GRIA1, GRIA2, kainate receptors GRIK1 and GRIK4, metabotropic receptor GRM3, transaminase BCAT1 and the glutamine synthetase (encoded by GLUL). Overall, the biomarker panel achieved an accuracy of 88% (95% CI: 87.1, 90.8) in predicting the tumor entity. Gene expression data, however, could not discriminate between patients with seizures from those without.

Conclusion: We have identified a panel of seven genes whose expression may serve as a biomarker panel to discriminate glioblastomas and brain metastases at the molecular level. After further validation, our biomarker signatures could be of great use in the decision making on subsequent treatment regimens after diagnosis.

Keywords: biomarker; brain metastasis; epilepsy; glioblastoma; glutamate; glutamate receptors.

Figure 1

Expression of ionotropic glutamate receptors in Grade 4 gliomas/astrocytomas and brain metastases. RNA was isolated from grade 4 glioma/astrocytoma (H, n=35) and brain metastasis samples (M, n=20), and reverse-transcribed in cDNA as described in the material and methods section. Subsequently, the mRNA expression of (A) NMDA receptors, (B) AMPA receptors, (C) KA receptors and house-keeping control TATA-box binding protein (TBP) was assessed by real-time PCR. The box-and-whisker plots represent relative amounts ( ${10}^3\times 2^{-\Delta Ct}$ ) of target mRNA. Median is shown as a black-coloured line and the mean is red; *p<0.05 (Mann-Whitney U test).

Figure 2

Expression of metabotropic glutamate receptors and key players of glutamate homeostasis in Grade 4 gliomas/astrocytomas and brain metastases. RNA was isolated from grade 4 glioma/astrocytoma (H, n=35) and brain metastasis samples (M, n=20) and reverse-transcribed in cDNA. Subsequently, the mRNA expression of (A) metabotropic receptors (GRM1-8) and (B) key players of glutamate shuttling (SLC1A2 and SLC7A11 coding for EAAT2 and xCT antiporter respectively) and metabolism (IDH1, BCAT1, GLUL), and house-keeping control TBP was quantified by real-time PCR. The box-and-whisker plots represent relative amounts ( ${10}^3\times 2^{-\Delta Ct}$ ) of target mRNA. Median is shown as a black-coloured line and the mean is red; *p<0.05 (Mann-Whitney U test).

Figure 3

Glutamate receptors expression in human brain tumour slices. AMPA receptor subunits GluA1 and GluA2 (shown in green) and metabotropic receptor mGluR3 (red) were determined in the tumour area. Additionally, nuclei were counterstained with DAPI (blue). (A) The fluorescence levels of glutamate receptors and DAPI were used to quantify receptor expression in glioblastoma (n=5) and metastasis (n=5) tissues as described in detail in the material & methods section. Data are presented as mean ± SEM; *p<0.05 (Student’s t-test). (B) Negative controls of Alexa Fluor 488 and Cy5-conjugated secondary antibodies. (C) Representative images are based on microscopic photographs that were taken at 100× magnification. Bars represent 200 μm.

Figure 4

Binary classification of Grade 4 glioma/astrocytoma and brain metastasis by receiver operating characteristic analyses. ROC analyses were performed on a total of 55 tissue samples obtained from surgery (n=35 grade 4 glioma/astrocytoma, n=20 brain metastases). Only genes with an AUC>0.8 are presented in (A₁) encoding for glutamate receptors and (A₂) glutamate homeostasis. (B) Both, GRIA3 and SLC7A11 failed to reach an AUC>0.8 and were excluded as biomarker candidates (see supplementary 6 and 7 for the remaining genes with AUC<0.8).

Figure 5

Schematic presentation of the pathophysiological function of the biomarker panel in glioma cells. Biomarker candidates are highlighted in red colour and their functions are illustrated. Briefly, in the cytoplasm, glutamate (Glu) is synthesised from α-ketoglutarate (α-KG) and branched-chain essential amino acids by BCAT1. In glioblastoma, glutamate is primarily released via cystine/glutamate antiporter solute carrier family 7 member 11 (xCT). Cystine is an essential precursor for glutathione synthesis, to counteract oxidative stress in fast-growing tumours. In addition, glutamate is catalysed to glutamine (Gln) by glutamine synthetase (GS). With respect to glutamate receptors, the metabotropic receptor mGluR3 is coupled to downstream signalling pathways like the PI3K/AKT pathway and contribute to migration and survival of the tumour cells. In part, AMPA receptors (GluA1/2) may also contribute to an activation of downstream signalling pathways due to calcium influx. Little is known on the function of kainate receptors (GluK1/4). Leu, Leucine; Ile, isoleucine; Val, Valine.