Background

Glioblastoma (GBM) is very aggressive and has a median of 12- to 15-month survival with less than 5% of 5 year survival [1]. Patients had a highly different therapeutic response and rates of survival which could be due to tumor heterogeneity [2]. Clinical, pathological examination and imaging techniques are the standard techniques for GBM diagnosis. Invasive tissue biopsy procedure has many risks to those patients, as affecting neurological functions, hemorrhage, etc. [3], with some tumors may be inaccessible due to their location or close to risk organ [4]. Also, imaging techniques cannot discriminate pseudo- and true progression after treatment to prevent unnecessary operations and further useless treatment [5]. Therefore, the need appears for more simple techniques to assess biomarkers from non-surgical samples. Isolation of circulating tumor DNA (ctDNA) from blood has a number of benefits such as decreasing invasive damages and obstacles of getting sufficient tumor tissues. Also, blood sampling is attainable and easy to reiterate when needed which provides a persuasive and achievable method to estimate the characteristics of cerebral tumors [6]. However, few presence of ctDNA in blood is a significant challenge for tumor biomarker testing availability and its translation clinically [7]. Incorporation of histological and genetic evaluation is recommended in GBM, with many genes involved such as isocitrate dehydrogenase 1 and 2 mutations, 1p/19q chromosomal codeletion and point mutations in tumor protein 53, etc. [8]. Using next generation sequencing (NGS) in estimation of genetic alternations in liquid biopsy has been grown rapidly across several solid tumors [9] but still limited in GBM especially in Egypt. Many studies revealed mutations in blood of GBM patients such as Piccioni et al. [10] who has used NGS in ctDNA analysis of advanced glioblastoma patients and has observed mutations in TP53, PDGFRA and NF1 genes, etc. Another study had 33 GBM patients showed mutations in TP53, EGFR and MET genes, etc. [11].

Study objectives

Our study provides insight for the genetic alterations in liquid biopsy of the newly diagnosed GBM patients using targeted next generation sequencing technology together with identification of the microsatellite instability (MSI) status in those patients.

Methods

Participants and sample preparation

This pilot prospective study included 7 newly diagnosed brain tumor patients and was performed from Dec. 2019 to Jun. 2020. DNA was isolated from blood samples by QIAamp® DNA Mini kit—Catalogue Number ID: 51304 as stated by the manufacturer guidance. Concentration, quality and amplifiability of the isolated DNA samples have been tested before further processing [12].

Sequencing and data analysis

Preparation of the libraries was done by Illumina AmpliSeq™ Cancer Hotspot Panel v2, Catalogue Number: 20019161 detecting 50 genetic mutations. Libraries were examined by 2100 Bioanalyzer instrument using DNA 1000 kit—Catalogue Code: 5067-1504 with the anticipated PCR yield 186–277 bp. Patients’ libraries were combined to reach a final sequencing library which ran using MiSeqDx system with read length of 2 × 150 bp and approximately 17 h to finalize the run [9]. Checking each run quality was done by determining specifications depending on PhiX libraries that provide a cluster density of 865–965 k/mm2 clusters passing filter for v2 technology, as well as, run’s quality score is evaluated. The percentage of bases more than the Q30 is averaged over the whole run with a quality score for v2 technology more than 80% bases higher than the Q30 on 2 × 150 bp. Sequence reads was aligned to the Genome Reference Consortium Human Build 37 (GRCh37).

Detection of microsatellite instability (MSI) in studied patients

Mononucleotide markers were recommended in the detection of MSI. Thus, we identified 3 mononucleotide markers: BAT25, BAT26 and NR27, according to manufacturer protocol and data were analyzed using Agilent 2100 Bioanalyzer system [12].

Results

Seven patients were included with 6/7 (85.7%) patients were males. The median patient age was 50 years (range 23–58). Right-sided tumor site was common among our patients 5/7 (71.4%). By MRI brain scan, the median size of the tumor was 5 cm (4–6 cm). The clinicopathological features of our patients are described in Table 1. Variant allele frequency (VAF) of each variant (Table 2) and primary analysis revealed 28 mutations (Table 3). Four variants out of 28 were not found in Catalogue of Somatic Mutations in Cancer (COSMIC) database with 5 non-coding variants were noted in the intron of a transcript and only 1 variant was a SNP in COSMIC database. Across 15 genes, there were (4, 12 and 2) missense, coding silent and intronic mutations, respectively (Fig. 1). Searching in ClinVar database, 13/18 was benign mutations, 1 variant has conflicting interpretations of pathogenicity, 3 mutations were not recorded and only 1 variant was drug responsive one. Regarding missense variants, Tumor protein TP53 (TP53 p.Pro72Arg) was detected in 6 patients (85.7%) which was a drug response mutation resulted from c.215C > G. Mutation in Kinase Insert Domain Receptor (KDR) gene was found in 4 patients (57.1%); 3 patients were glioblastoma multiforme grade IV and one patient was astrocytoma grade II. This mutation was p.(Gln472His) resulting from c.1416A > T and was not recorded in ClinVar. Two patients had mutation in KIT Proto-Oncogene, Receptor Tyrosine Kinase (KIT) gene. It was p.(Met541Leu) variant which results due to c.1621A > C and was known as a benign/likely benign mutation in ClinVar. Only one patient revealed mutation in Janus Kinase 3 (JAK3) gene which was p.(Val718Leu) variant resulting from c.2152G > C and recorded in ClinVar as conflicting interpretations of pathogenicity, likely benign or uncertain significance variant. Benign coding silent variants were noticed in the following genes: Isocitrate dehydrogenase (NADP(+)) 1 (IDH1), APC Regulator of WNT Signaling Pathway (APC), Epidermal Growth Factor Receptor (EGFR), Cyclin Dependent Kinase Inhibitor 2A (CDKN2A), HRas Proto-Oncogene, GTPase (HRAS). These mutations were p.Gly105=, p.Thr1493=, p.Gln787=, p.Arg58=, and p.His27=, respectively, which result from c.315C > T, c.4479G > A, c.2361G > A, c.174A > C and c.81T > C, respectively. MET Proto-Oncogene, Receptor Tyrosine Kinase (MET) gene mutation showed 2 benign variants p.(Ile377=) and p.(Ser178=) due to c.1131C > T and c.534C > T, respectively. Two benign variants were detected in platelet-derived growth factor receptor A (PDGFRA) gene, p.Val824= and p.Pro567= which result from c.2472C > T, and c.1701A > G, respectively. Ret Proto-Oncogene (RET) gene revealed 2 benign variants, p.Ser904 = due to c.2712C > G and p.Leu769 = due to c.2307G > T. Fibroblast growth factor receptor 3 (FGFR3) gene revealed p.(Thr653=) due to c.1959G > A which is not recorded in ClinVar. Two intronic mutations were noticed in Fms-Related Receptor Tyrosine Kinase 3 (FLT3) gene in 5 patients due to c.1310-3T > C and SWI/SNF-Related, Matrix-Associated, Actin-Dependent Regulator of Chromatin, Subfamily B, Member 1 (SMARCB1) gene in 2 cases due to c.1119-41G > A. As regards MSI status, 4/7 (57.1%) cases had MSI-Low and 3/7 (42.9%) cases had MSI-high (Fig. 1).

Table 1 Clinicopathological features of studied population
Full size table
Table 2 Assessment of variant allele frequency and MSI status
Full size table
Table 3 Activating mutations detected and MSI status in the studied population
Full size table
Fig. 1
figure 1

Different studied activating mutations and MSI status in GBM patients

Full size image

Discussion

Recently, our information about the genetic features of cerebral tumors has raised dramatically by using next generation sequencing platforms [13]. Liquid biopsy has been widely used in solid tumors to identify driver mutations, but is still limited in glioblastoma multiforme (GBM) patients [14]. Our pilot study aimed to assess the activating variants in blood samples of the newly diagnosed GBM using targeted next generation sequencing together with identification of the microsatellite instability (MSI) status. We found the drug responsive variant p.(Pro72Arg) of the tumor suppressor TP53 gene in 6/7 (85.7%) patients. Previous reports showed that TP53 is mutated in 29% of the GBM samples and another study showed TP53 mutation in 38% of gliomas, including 23% of primary glioblastomas and 80% of secondary glioblastomas [15, 16]. Other studies showed that p53 is the commonest mutation noticed in the blood derived ctDNA samples of gliomas [17]. The drug responsive variant p.(Pro72Arg) of the TP53 gene was found in 47.94% ependymoma grade III and also detected in a young medulloblastoma patient [18]. Kinase Insert Domain Receptor (KDR) gene which is a Vascular Endothelial Growth Factor Receptor 2 (VEGFR2) gene has a role in tumor initiation and neovascularization [19]. We found KDR gene mutation in 4 patients (57.1%) which is p.(Gln472His) as a result of c.1416A > T. KDR p.(Gln472His) is a germline variant observed in fifty percent of GBM and forty-seven percent of grade 2–3 astrocytomas [20]. A correlation between KDR p.(Gln472His) and risk of glioma has been observed, as unusual angiogenesis may be implicated in primary tumorigenesis [21] with the more angiogenic activity, the worse the survival rate. Thus, GBM patients with the p.(Gln472His) substitution have poor prognosis and this may be related to increases in micro vessel density [22]. Other reports found better survival in positive KDR p.(Gln472His) head and neck squamous cell carcinomas [23]. KIT mutations have been described in tumor cell proliferation, such as cancer stem cell proliferation, and proliferation of endothelium in gliomas and assisting tumor-related angiogenesis [24]. Here, we observed KIT p.(Met541Leu) variant in 2 (28.6%) patients which result from substitution at c.1621A > C. This variant has been described to enhance the receptor affinity to its ligand, stem cell factor (SCF) [25]. Zaman et al. [20] observed KIT M514L in 43.75% of both patients of GBM and glioma grade 2–3 and this variant may be used as a marker of aggressiveness which result from mechanisms that do not include regulation of angiogenesis. In our study, only one patient revealed JAK3 p.(Val718Leu) variant resulting from c.2152G > C substitution. JAK3 is a gene encodes a protein-tyrosine kinase which functions in cytokine receptor-mediated signal transduction and altered in 1.90% of all tumors [26]. As regards MSI status in our GBM patients, 4/7 (57.1%) had MSI-Low and 3/7 (42.9%) had MSI-High. Viana-Pereira et al. [27] found that 13.5% of high-grade glioma samples presented instability, with (< 1%, 12.5% and 86.8%) are MSI-H, MSI-L stable tumors, respectively. Previous study noticed about 27% MSI in 45 pediatric high-grade gliomas using mononucleotide (BAT25 and BAT26) markers [28], another study did not note MSI in 41 cases using (CAT25, BAT25 and BAT26) [29]. Further studies are needed to explore whether liquid biopsy in brain tumor patients could potentially defeat the natural difficulty developed accompanied by the standard tissue biopsy. Larger sample size and longer follow-up period are recommended to compare genetic mutations and MSI status in liquid based versus tissue-based biopsy by targeted next generation sequencing.

Conclusions

Development of noninvasive or minimally invasive approaches to discover and monitor tumors is a major challenge and still limited in our brain tumor patients. This study identifies the molecular landscape and microsatellite instability status in a sample of Egyptian brain tumor patients, which may have an important role in improving the outcome, survival rate and to develop new personalized treatments.