Clinical Samples
Data from adult patients who underwent surgery for glioma in our institution between January 2020 and December 2022 were analyzed. Consensus about diagnosis, treatment, and related information was obtained under written informed consent approved by our Institution’s Principal Institutional Review Board. This study adheres to PROBE 2023 guidelines for reporting observational studies. All methods were carried out following relevant guidelines and regulations. Patients were enrolled according to the following criteria: age > 18 years; preoperative magnetic resonance imaging (MRI) is suggestive for glioma; no previous surgery; no previous radiotherapy; at least 18 months of follow-up; patients that undergo a standard STUPP protocol [3] starting from the 30th day after surgery.
Patients were excluded if they had a histologic diagnosis other than glioma, had not fully performed therapeutic or diagnostic follow-up, and had radiological material not available on PACS.
All the patients included underwent a preoperative brain MRI scan included a high field 3 Tesla volumetric study with the following sequences: T2w, FLAIR, isotropic volumetric T1-weighted magnetization-prepared rapid acquisition gradient echo (MP-RAGE) before and after intravenous administration of paramagnetic contrast agent [20, 21].
All participating patients underwent surgery to completely diagnose and remove the tumor mass. The procedures were carried out using an infrared-based Neuronavigator (Brainlab, Kick® Purely Navigation) in a standard neurosurgical theater, equipped with a standard operative microscope (Leica, model OH4). Our institution’s surgical protocol was followed [22], where the extent of resection (EOR) was deemed complete when the white matter was free of disease in all aspects of the surgical cavity. The excision was discontinued by the surgical operator when, despite direct visualization or navigation remnants, neuromonitoring or intraoperative neuropsychological testing indicated a potential for postoperative motor complications.
In the postoperative day, patients underwent a CT-scan to assess major early complications and a volumetric Brain MRI scan to evaluate the EOR. For gross total resection (GTR), “tumor progression” was defined as the initial MRI scan showing the presence of pathologically enhancing tissue, characterized by an MRI pattern inconsistent with cerebral radiation injury, which is considered a “pseudo-progression.” In the case of incomplete resections (< 95% volume reduction), a volumetric increase of the residual disease detected at the first postoperative MRI scan was considered as disease progression, and this was used to calculate the PFS. Our institution had a dedicated neuro-imaging follow-up program that included a standard early (maximum 24 hours after surgery) postoperative volumetric brain MRI, a volumetric brain MRI scan repeated at one month from surgery (25–35 days) for the first step follow-up control to provide information for the radiation treatment planning, and a volumetric brain MRI scan performed every three months.
Clinical and Pathological Analysis
For all the included patients we recorded age, gender, IDH R132H, Ki67, p53 and EGFR expression status. The expression of IDH1 R132H, p53, EGFR and Ki67 in formalin-fixed paraffin embedded (FFPE) tumor tissues was analysed by standard immunohistochemistry technique carried out in the Department of Neuropathology of our University Hospital [23]. The following antibody were used: anti-p53 (DO-7, 1:40; Cell Marque, Hot Springs, AZ, USA), which detect both wild type and mutant p53 protein [24]; anti-IDH1 R132H (DIA-H09, 1:50; Dianova, Hamburg, Germany); anti-EGFR (clone H11; 1:200; Dako, Glostrup, Denmark); anti Ki-67 (MIB-1, 1:50; Dako, Glostrup, Denmark) [23]. Following counterstaining with hematoxylin, slides were dehydrated, mounted and observed under light microscope. At least 200 tumor cells from different fields (from 5 to 10) were reviewed. Scoring was performed by semi-quantitative scoring, independent to diagnosis, with not-expressed /negative (no staining observed), expressed (up to 50% cells are stained) and highly expressed (˃50% of cells stained). %Ki67 was measured with the “hot spot method” where the field with highest apparent Ki67 index was selected and up to 500 cells scored [25].
Histological diagnoses were performed according to the updated version of the 2021 WHO guidelines [26]. PFS and OS was recorded in months; it was measured from date of diagnosis to date of death or date of last contact if alive. Clinical information was obtained by the digital database of our Institution, whereas OS data, were obtained by telephone-interview.
Cell cultures and lentiviral infection
Primary GB neurosphere cultures were obtained after mechanical dissociation from high-grade gliomas freshly resected from patients. In brief, the tissue was first washed in HBSS plus penicillin–streptomycin (1%) to remove excess debris and blood, and the tumor has been cut and mechanically minced before digested with deoxyribonuclease. The digested tissue was titrated and passed through a cell strainer. Finally, cells were pelleted by centrifugation (300 X g for 10 min) and cultured in Neurobasal medium supplemented with B27 without vitamin A (2%), penicillin–streptomycin (1%), L-glutamine (1%), N-Acetyl-L-Cysteine (60 ng/ml) human EGF (20ng/ml) and human FGF (20ng/ml).
All the GB cell lines were validated by short tandem repeat (STR) DNA profiling performed by Eurofins Genomic Europe (Ebersberg, Germany) in November 2023 and preserved in liquid nitrogen to preserve authenticity. Mycoplasma contamination in cell cultures was routinely detected by using PCR detection kit (Applied Biological Materials, Richmond, BC, Canada).
For MEX3A genetic depletion, lentiviral particles were generated in HEK293 cells transfected with packaging and envelope plasmids (pCMV-dR8.74 and VSV-G/pMDG2), pGFP-pLKO.1 plasmids (shCTR TR30021; shMEX3A TL308061B (#1), TL308061C (#2), Origene, Rockville, MD, USA). GB tumor spheroids were dissociated in cell dissociation solution (C5789, Merck, Darmstadt, Germany) and GB6 cells were infected with purified lentiviral particles resuspended in complete medium for 72 h.
For animal study, GB6 cells were infected with both lentiviral particles expressing shCTR or shMEX3A #2 and the luciferase reporter (pLenti CMV Puro Luc w168-1, Addgene, Watertown, Massachusetts, USA), generated as described above.
mRNA expression analysis
Total RNA was isolated from gliomas tissues and peritumoral brain normal tissues, using TRIzol reagent (Invitrogen/Life Technologies, Carlsbad, CA, USA), and reverse-transcribed with a SensiFAST cDNA Synthesis Kit (Bioline Reagents Limited, London, UK). Quantitative real-time PCR (qRT-PCR) analysis of mRNA expression for the indicated genes was performed by using the ViiATM 7 Real-Time PCR System (Life Technologies). A reaction mixture containing cDNA template, SensiFAST™ Probe Lo-ROX mix (Bioline Reagents Limited) and Taqman Gene Expression Assays (Thermo Fisher Scientific Waltham, MA, US) was amplified using standard qPCR thermal cycler parameters and mRNA quantification was performed by using SDS version 2.3 software. Each sample was amplified in triplicate, and the average of the three threshold cycles was used to calculate the number of transcripts. Data were normalized to the endogenous controls (GAPDH and HPRT), which yielded similar results. And expressed as the fold change respect to the control sample value.
To compare the mRNA expression of MEX3A between the patients samples we considered the value of 2ˆ(-ΔΔCT). For the in vitro and in vivo experiments in primary gliomas cell lines mRNA expression of the indicated genes was expressed as the fold change respect to the control sample value.
The following TaqMan Gene Expression Assays (Thermo Fisher Scientific) were used: MEX3A, Hs00863536_m1; Nanog, Hs02387400_g1; POU5F1, Hs04260367_gH; Sox2, Hs04234036_s1; Gapdh, Hs02786624_g1; Hprt, Hs02800695_m1.
Immunohistochemistry
FFPE brain tissues from the orthotopic GB6 implanted mouse were cut into 5µm sections for MEX3A, Ki67 and sex determining region Y-box2 (SOX2) immunohistochemical staining. FFPE slides were deparaffinized and subjected to heat-induced antigen retrieval at low or high pH buffer and blocked for 30 min with 5% PBS/BSA. Then, the slides were incubated overnight at 4°C with monoclonal antibodies against MEX3A (Rabbit anti-MEX3A ab79046, 1:100; Abcam, Cambridge, UK), Ki67 (Rabbit anti-Ki67 SP6, MA5-14520, 1:100; Thermo Fisher Scientific, Waltham, MA, US) and SOX2 (Rabbit-anti SOX2 ab97959, :100; Abcam, Cambridge, UK). The day after, the slides were incubated for 20 min with secondary antibodies coupled with peroxidase (Dako, Glostrup, Denmark). Bound peroxidase was detected with diaminobenzidine (DAB) solution and EnVision FLEX Substrate buffer containing peroxide (Dako, Glostrup, Denmark). After counterstaining with hematoxylin, sections were dehydrated in a graded water-ethanol series, mounted and observed under light microscope. Cell quantification was performed on collected sections using the imaging software NIS-Elements BR 4.00.05 (Nikon Instruments Europe B.V., Italy).
Cell proliferation assay
For IncuCyte® experiments, infected GB cells were seeded in 96-well plates (20 x103 cells/well for each cell lines; 12 wells for each condition) in complete medium and treated with IncuCyte® NucLight Rapid Red Reagent (#4717, Essen BioScience, 1:1000). Plates were transferred into the IncuCyte® S3 Live Cell Analysis Systems and incubated at physiological conditions (37°C, 5% CO2), over 96 h. Images were collected every 6 h, and proliferation was evaluated as the ratio number of infected cells on NucLight® positive cells. The experiments were performed in triplicate, and data were analyzed by using the IncuCyte® software package (Essen BioScience, Ann Arbor, MI, USA).
Neurospheres formation assay
Dissociated primary GB cells for each experimental groups (shCTR, shMEX3A#1 and shMEX3A#2) were plated in 96-well plates in decreasing numbers (50, 25, 10, 5 cells; 12 well for each condition, 12 replicates per cell/density number) in neurospheres culture media. The plates were incubated in a 37°C, 5% CO2, humidified incubator. At ten days, any well that contains neurospheres was scored.
Extreme limiting dilution analysis was conducted using the software available at http://bioinf.wehi.Edu.au/software/elda/.
Orthotopic xenograft study
Female NOD/SCID gamma (NSC) mice (6 weeks old) were used. All described procedures involving experimental animals were performed in agreement with standard guidelines under a protocol approved by local ethic authorities (Ministry of Health) and conducted in accordance with Italian Governing Law (D.lgs 26/2014).
For establishing intracranial GB, 3 x 105 cells from GB6 cell line transduced with lentiviral particles expressing shMEX3A#2 or shCTR, and the luciferase reporter, were stereotaxically implanted into the striatum of the mice device (coordinates: 2 mm anterior, 2 mm lateral, 3 mm depth from the dura).
Tumors were analysed by luminescence imaging (IVIS Lumina III, PerkinElmer, USA) each week. Before imaging, mouse underwent an intraperitoneal injection of D-luciferin (10 µl g− 1, XenoLight RediJect D-Luciferin, PerkinElmer, USA). Mice were sacrificed upon signs of tumor formation (weight loss, hunching, rough coat, level of consciousness and activity) and brains were fixed in 4% formaldehyde and paraffin embedded for immunohistochemical analysis.
Statistical methods
Statistical analyses were performed with GraphPad Prism software version 9.5.1 (GraphPad, San Diego, CA, USA). The median value of MEX3A expression was used as cutoffs to divide the patients in the two group of “high” or “low” MEX3A expression level. MEX3A expression in different groups was compared using two-tailed Student’s t test or one-way ANOVA. The association between MEX3A expression and clinicopathological and molecular characteristics of the patients was assessed by χ2 test. The sensitivity and specificity of MEX3A in glioma diagnosis was evaluated using receiver operating characteristic (ROC) curve. Survival curves were carried out using Kaplan-Meyer method, assessing differences in OS and PFS between the indicated groups by log-rank test. Univariate and multivariate survival analysis were performed using the Cox proportional hazard regression model. Only the features with prognostic significance in univariate analysis were included in the subsequent multivariate analysis. The threshold of statistical significance was considered p < 0.05.
Potential source of Bias and Study size
A potential source of bias is expected from the exiguity of the sample, which, nevertheless, in regards to the endpoints selected, presents an excellent post-hoc statistical estimated power (1- β = 0.9402 for α 0.05 and effect size “f” = 0.34), thus providing extremely reliable conclusions.