Clinical samples

Clinical tumor samples were obtained from eight patients who repeatedly underwent surgery for malignant glioma at University Hospital Hradec Kralove. The study was approved by local ethics committee (Reference No. 201,906 S26P) and patients gave their written consent. All samples included in the study differed in tumor stage and grade and represent a differing number of repeated resections. The material sampled (amount and quality) for RNA analysis of hypoxia markers as well as cell culture derivation varied based on the particular specimen.

The cell derivation procedure and further manipulation with cell cultures was described previously [15]. In this study, three glioma cell primocultures were prepared in sufficient quality and quantity required for further analyses.

Clinical data pertinent to each case were retrospectively or prospectively reviewed including surgical reports, radiological images, histological parameters and therapeutic protocol. Tumor size (before surgery) was evaluated by neuroradiologist and radiation therapist following Magnetic Resonance Imaging (MRI). It was performed by 3 Tesla magnet using a cross-sectional 2D method with the product of the largest perpendicular diameters on T1-weighted contrast-enhanced. Other parameters such as tumor location, survival, response to radiotherapy and TMZ treatment, electrocortigraphy (ECOG), among others were evaluated and correlated too.

Immunohistochemistry and tissue microarray (TMA) construction

Paraffin-embedded specimens, cryopreserved samples and native (non-fixed) tumor tissue samples were retrieved from and handled by Fingerland Department of Pathology. In case of native tumor samples, a minor part was cryopreserved and most of tumor tissue was fixed in 4% buffered formalin solution and paraffin embedded. 1 μm thick sections were cut and stained with hematoxylin-eosin to confirm the diagnosis and select appropriate areas for additional analyses. A TMA was constructed from tumor samples, using TMA Master II system (3DHISTECH Ltd., Budapest, Hungary). For each case, 2-mm-thick sections were cut and routine H&E staining and immunohistochemical studies were performed. The list of antibodies and relevant details of immunohistochemical protocols are summarized in Supplementary Table 1. Heat induced epitope retrieval (HIER) was used for antigen retrieval, employing different pH according to the antibody. For EGFR, proteolytic pretreatment step was used. The sections staining was carried out on Benchmark Ultra stainer manufactured (Ventana/Roche, Tucson, AZ, USA) using either Ventana ultraView Universal DAB detection kit or Ventana OptiView DAB IHC detection kit: both methods use avidin-biotin complex method with horseradish peroxidase as an enzyme and DAB (3,3’-diaminobenzidine) as chromogen. Agilent/Dako Autostainer 48 or Dako Omnis (Agilent, Santa Clara, CA, USA), with EnVision Flex detection kit was used for IDH1 R132H, CD57, ZEB2, TWIST, and HIF1a. For HIF1a and HIF1b, whole sections (WS) from tissue core donor blocks were used. The slides were subsequently counterstained with hematoxylin.

Mutation analysis

DNA from glioma cells was extracted using the commercial DNA Sample Preparation Kit (Roche, Basel, Switzerland). Mutation analysis was performed by multi parallel sequencing (NGS) using the hybrid-capture-based target enrichment. A custom KAPA HyperChoice MAX Library (Roche) for enrichment of the coding and 30 bp upstream and downstream overlaps of selected panel of genes (TP53, EGFR, IDH1, IDH2, PTEN, PIK3CA) was used. Paired-end cluster generation and sequencing was performed by NGS system Illumina MiniSeq. Sequencing data analysis were performed by NextGENe software (Softgenetics) and Varsome Clinical Platform. MGMT methylation analysis was performed on DNA in FFPE using DNA Sample Preparation Kit (Roche). Bisulfite conversion of isolated DNA was performed using the EZ DNA Methylation-Gold Kit (Zymo Research). Detection and quantification of the hypermethylation status of the O (6)-methylguanine-DNA methyltransferase (MGMT) promoter was performed by the methylation-specific real-time PCR method using the CE-IVD marked geneMAP MGMT Methylation Analysis Kit (GenmarkSalgik).

Athymic nude mouse model

Female Foxn1-nu athymic immunodeficient mice weighing 27–30 g were purchased from Velaz, Czech Republic. They were given a standard sterilized diet and water ad libitum. Cells suspensions intended for implantations were obtained from cultivated cells. Prepared 100 µl of cell suspension (1mil. cells/application) was injected subcutaneously into each 6 weeks old Foxn-1nu female mice on the right and left side of the back. The administration of treatment (TMZ) began two weeks after implantation (daily from day 15th to day 28th) as based on the approved project plan (ethical committee, project number: MSMT- 18,525/2021-3 attached as supplementary Fig. 4). TMZ was dosed orally (maximum volume 100 µl) at the therapeutic range of 0.9 mg/kg. Each experimental group consisted of 2–3 mice. The size of the tumors and health condition of mice were periodically checked. On the day of last application of TMZ, twenty minutes after treatment, mice were anesthetized with isoflurane, sacrificed and tumors were weighed and preserved until further analysis (in formalin at room temperature for IHC analysis, in Trisol at -80 °C for RT-PCR analysis and in lysis buffer at -20 °C for western blot analysis). Heart, liver, brain and plasma were also collected and stored at -80 °C before MS analysis.

Cell line

Human malignant glioma cell line U87MG was purchased from ATCC (LGC Standards, Poland). Fresh cells from frozen batch were used for every set of experiments (lasted 3–9 weeks). Cultures were grown in EMEM supplemented with 10% FBS and 0.5% penicillin/streptomycin and cells were maintained in incubators with a humidified atmosphere containing 5% CO2 at 37 °C. The absence of mycoplasma contamination was periodically checked.

Crispr/Cas STAT3 knockout cell model

Glioma cells U87MG grown to 50–70% confluence were transfected with transfection mixture (gRNA vectors in Opti-MEM I, the donor DNA and Turbofectin 8.0 - the ratios of 3:1 for Turbofectin: DNA) as based on manufacturer’s protocol (STAT3 Human Gene Knockout Kit (CRISPR), CAT#: KN204922, Origene). Cultures were split every 3 days (2–4 times in total) to dilute out cells containing non-integrated donor DNA. Then, puromycin selection was performed; cells of split P5 were grown directly in the puromycin containing complete media (the range of puromycin concentrations was 1 to 10 µg/ml) and medium was changed every 2–3 days. The puromycin resistant cells (approx. P10 split cells) were analyzed for genome editing, WB was used to measure gene knockdown and genomic PCR was carried out to verify the integration of the functional cassette. With primer pair of 5 F and 3R, both alleles of donor inserted and non-edited/indel were amplified. Thus established cell line was labeled as U87MG STAT3 KO.

Proliferation

The inhibitory effect of chemotherapeutic TMZ at various concentrations on viability of U87MG and its STAT3 KO variant as well as viability of primocultures was evaluated by WST-1 assay for 48 h. Principle of this colorimetric test is based on the cleavage of the tetrazolium salt to colored formazan by mitochondrial dehydrogenases in viable cells. At the end of tested interval, cells were rinsed with PBS and WST-1 solution (diluted according to manufacturer’s recommendations) was added to each well for further 2 h. Quantification of mitochondrial enzymes activity was carried out at 450 nm with 650 nm of reference wavelength by Tecan Infinite M200 spectrophotometer (Tecan, Switzerland).

RNA extraction, cDNA synthesis, primer design, quantitative real-time RT-PCR

Total RNA was isolated from glioma primocultures and both variants of U87MG– STAT3 expressed and STAT3 KO variant - using Direct-zol RNA MiniPrep kit according to the manufacturer’s instructions (ZymoResearch, Irvine, CA, USA). Tumors from cryopreserved tissues and excised from mice (both variants of U87MG tumors) were homogenized using Tissue lyzer (2 cycles − 25 vibration/s; 4 °C; 1 min; Qiagen, USA) in TriReagent. RNA concentration and its purity were measured using NanoDrop 2000 (Thermo Fisher Scientific). All samples had absorption ratio A260/A280 greater than 1.8. RNA integrity number (RIN) was determined using Agilent 2100 Bioanalyzer and cell line samples with RIN higher than 9.0, resp. tissue samples with RIN higher than 8.0 were used for further analysis. First strand cDNA synthesis and qPCR analysis were performed in LightCycler® 96 Instrument (Roche Life Science) as described in [16]. Primers were designed manually, and their sequences are attached as a supplementary file (Supplementary Tables 2 - primer sequences). Calculations were based on delta-delta Cq method [17] and expressed as fold change of the treated groups relative to the control. Beta-2-microglobulin (B2M) or TATA box binding protein (TBP) were used as reference genes for mRNA analysis.

LC-MS analysis

Frozen tissues were homogenized in 4 volumes of cold PBS (w/v) using Fastprep-24 5G sample disruption instrument. Thawed plasma or homogenized tissues in the volume of 100 µl, was mixed with the same volume of methanol and acetonitrile, vortexed for 15 min and centrifuged at 14,000 g for 3 min. Supernatant was then filtered through 0.22 μm PTFE syringe filter into the vial and measured.

Detection of TMZ and its metabolites AIC and MTIC was performed on the Agilent 1290 Infinity II UHPLC system coupled to the Agilent 6470 QqQ mass spectrometer. Chromatographic conditions were maintained at gradient elution of 0.4 ml/min by 0.1% formic acid in water and methanol (0-0.5 95:5, 0.5-3.0 gradient to 5:95, 3.0–4.0 5:95, 4.0–5.0 95:5), thermostated autosampler set to 15 °C and column thermostat equipped with the Zorbax Eclipse plus RRHD C18 2.1 × 50 mm, 1.8 μm (PN 959757-902) column kept to 30 °C. MS source parameters were set to the following: drying gas 200 °C at 2 l/min, sheath gas 400 °C at 12 l/min, nebulizer pressure 25 psi, capillary voltage 2500 V and nozzle voltage 0 V. TMZ transitions of [M + H] + ions m/z were detected with setting of dwell time 50 ms, cell accelerator 4 V, fragmentor 88 V for 195→138 and 55 (collision energy– CE 8 and 28 V). MTIC transitions of [M + H] + ions m/z were detected with setting of dwell time 50 ms, cell accelerator 4 V, fragmentor 88 V for 169→109 and 43 (CE 20 and 40 V). AIC transitions of [M + H] + ions m/z were detected with setting of dwell time 50 ms, cell accelerator 4 V, fragmentor 88 V for 127→110, 82 and 55 (CE 20 V, 20 V and 40 V).

Morphology

The TMA was evaluated for presence of the tumor tissue. Only TMA cases with at least one representative core were included for further analyses. The immunohistochemistry results were first digitalised using Leica Aperio AT2 slide scanner (Leica Biosystems, Buffalo Grove, IL, USA) and then evaluated with Aperio ImageScope software (Leica Biosystems, Buffalo Grove, IL, USA). The assessment of immunohistochemistry was performed by an experienced neuropathologist (JS). Percentage of positive cells and most prevalent staining intensity (1– weak, 2– moderate, 3– strong) was noted. Average percentage of positive cells and modified H score (mHS = percentage * intensity) was used for the analysis as reported previously [18]. Whole sections were analyzed using the same approach.

Fluorescence microscopy

U87MG and U87MG-STAT-KO glioblastoma cells were fixed with 2% paraformaldehyde (20 min, 25 °C), rinsed with PBS permeabilized, and blocked with 1% Triton X and 5% BSA in PBS (30 min, room temperature). The cells were incubated with a primary antibody against STAT3 (D3Z2G® Rabbit mAb, Cell Signaling) at 4 °C overnight. Then, the cells were washed three times with cold PBS (5 min, 25 °C) and were incubated for additional 1 h (room temperature) with Alexa Fluor 488-labelled anti-rabbit antibody. Thereafter, the cells were rinsed three times with PBS and labelled with DAPI (10 µg/mL). The specimens were mounted into the Prolong Gold mounting medium (Invitrogen-Molecular Probes, Inc., Carlsbad, California, CA, USA) and examined using fluorescence microscopy technique (Nikon Eclipse E 400 (Nikon Corporation, Kanagawa, Japan)). The results were analyzed using LUCIA DI Image Analysis System LIM 4.2 (Laboratory Imaging Ltd., Prague, Czech Republic). All the samples were tested in duplicates in three independent experiments.

Statistics

Data in all tests used are expressed as an average ± SD from at least two experiments. The concentration of chemotherapeutic TMZ causing a 50% decrease of cell viability (IC50 value) was determined by Graph Pad Prism 7.0. Statistical analysis of the data from RT-PCR analysis, proliferation WST-1 assay and LC-MS analysis was carried out using TWO-WAY analysis of variance (ANOVA) followed by Sidak’s multiple comparison test significant at level of p˂0.05. Other data analysis was done with GraphPad Prism 7.0.