Using Gene Expression Changes for Monitoring Glioblastoma WHO CNS Grade 4 Recurrence: Validation of NGS Candidate Genes via qRT-PCR

Glioblastoma WHO CNS Grade 4 (GBM) is associated with poor prognosis and high recurrence rates despite the current therapeutic interventions. Conventional imaging bears restrictions in detecting tumor recurrence and discriminating it from tissue necrosis and post-operative scar tissues. In a previous study, we conducted a whole-transcriptome screening using next-generation sequencing (NGS) on whole blood samples (n = 33) from seven patients and tumor biopsies to identify tumor-specific gene expression patterns. This revealed downregulation post-surgery, with a return to pre-surgery baseline levels at the time of tumor recurrence. In the screening phase, we were able to identify 374 genes across all patients, and in this study, we aimed to validate the most predictive genes identified in the screening phase using the same samples with qRT-PCR (customized TaqMan array cards). A subset of candidate genes was selected for each patient to reduce the number of screening genes to be validated, resulting in a total of 94 genes for validation using qRT-PCR. All genes were measured in all samples, thus allowing a search for genes methodologically confirming the NGS candidate genes as well as examining qRT-PCR-based gene expression patterns, which were not captured by NGS in certain patients. Filter criteria included a post-surgery fold change (FC) ≤ 0.6 across all or most time points and already cited GBM-related genes. From whole blood, total RNA was isolated, converted into cDNA, and 6,270 gene expression measurements were performed employing custom TaqMan array cards (qRT-PCR). We were able to successfully validate NGS candidate genes for three patients out of seven. Differential gene expression (DGE) of both methods was comparable for these patients (e.g., r2 = 0.76, P < 0.001 for patient #1). Between 38-62% of NGS candidate genes per patient could be successfully validated, but only in patients where the screening comprised ≥20 NGS candidate genes. QRT-PCR also detected tumor recurrence-associated expression patterns in genes that were not captured by NGS, but 2-9-fold more candidates were observed for NGS-selected genes. Furthermore, in a patient receiving additional chemotherapy (CTx) during the tumor recurrence phase, a reduction in DGE of 13 genes was observed. In another patient, DGE indicated recurrence 13 days before radiological examinations confirmed it. The evidence, based on screening and validated, suggests a potential association, but a coincidental link cannot be ruled out. Our findings may point to the possible utility of gene expression measurements in peripheral whole blood as a simplified liquid biopsy method for GBM tumor recurrence. However, further prospective cohort studies with improved control of sampling during clinical follow-up and radiological examinations are needed to strengthen the presumed causality and the clinical value of early gene expression changes as an indicator of tumor recurrence.