Molecular Basis of Cisplatin Resistance in Testicular Germ Cell Tumors

The emergence of cisplatin (CDDP) resistance is the main cause of treatment failure and death in patients with testicular germ cell tumors (TGCT), but its biologic background is poorly understood. To study the molecular basis of CDDP resistance in TGCT we prepared and sequenced CDDP-exposed TGCT cell lines as well as 31 primary patients’ samples. Long-term exposure to CDDP increased the CDDP resistance 10 times in the NCCIT cell line, while no major resistance was achieved in Tera-2. Development of CDDP resistance was accompanied by changes in the cell cycle (increase in G1 and decrease in S-fraction), increased number of acquired mutations, of which 3 were present within ATRX gene, as well as changes in gene expression pattern. Copy number variation analysis showed, apart from obligatory gain of 12p, several other large-scale gains (chr 1, 17, 20, 21) and losses (chr X), with additional more CNVs found in CDDP-resistant cells (e.g., further losses on chr 1, 4, 18, and gain on chr 8). In the patients’ samples, those who developed CDDP resistance and died of TGCT (2/31) showed high numbers of acquired aberrations, both SNPs and CNVs, and harbored mutations in genes potentially relevant to TGCT development (e.g., TRERF1, TFAP2C in one patient, MAP2K1 and NSD1 in another one). Among all primary tumor samples, the most commonly mutated gene was NSD1, affected in 9/31 patients. This gene encoding histone methyl transferase was also downregulated and identified among the 50 most differentially expressed genes in CDDP-resistant NCCIT cell line. Interestingly, 2/31 TGCT patients harbored mutations in the ATRX gene encoding a chromatin modifier that has been shown to have a critical function in sexual differentiation. Our research newly highlights its probable involvement also in testicular tumors. Both findings support the emerging role of altered epigenetic gene regulation in TGCT and CDDP resistance development.

No seminoma cell line was available from the official cell-line collections. For the experiments we thus used T-Cam cell line donated to our department but due to the later uncertainties regarding its origin and characteristics, the results obtained on this cell line were finally excluded from the analyses and statistics.
All cell lines were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin (all from ThermoFisher Scientific, Waltham, MA, USA) at 37 °C in a humidified cell culture incubator with 5% carbon dioxide. Prior to passaging experiments, the cells were washed with PBS and trypsinized with 0.25% Trypsin in phosphate buffered saline (PBS).

Treatment with Cisplatin
NCCIT and Tera-2 cells were treated with increasing concentrations of cisplatin (CDDP; Ebewe Pharma, Unterach am Attersee, Austria) starting with the concentration of 0.1 μM. When the lethality of the cells reached 80%, the cells were left to recover over four passages without CDDP. The final concentration of CDDP used for establishment of resistant cell lines was 2 μM for NCCIT and only 0.1 μM for Tera-2 cell line. The same cell lines were also long-term cultivated without CDDP to detect aberrations that occur spontaneously due to prolonged cultivation and distinguish them from those induced by CDDP treatment.
Resistance to cisplatin was assessed from IC50 as determined by MTS proliferation assay performed repeatedly every 3 months. For every cell line, the analysis was performed in parallel with CDDP-treated and -untreated cells. The final data were collected from 3 independent experiments in case of Tera-2 and from 4 independent experiments in case of NCCIT (each performed at least in 3 replicates). The cells were incubated with serial concentrations of CDDP for 72 hours and their proliferation/metabolic activity was measured by Cell Titer 96® Aqueous One Solution Cell Proliferation Assay (Promega, Madison, WI, USA).
The proliferation and cell cycle assays were performed with the established cell lines when cultured in media without the presence of CDDP; the apoptotic assays were performed both without CDDP and with CDDP (10 µ M) added to the culture medium. The analyses were performed in triplicates.

Cell Proliferation Assay
To determine the proliferation rate of TGCT cell lines, we assessed the total number of cells arising from the equal initial number of seeded cells over the period of two weeks. Measurements were carried out repeatedly in triplicates.

EdU Staining
The cell lines were cultured for 1.5 hours with 10 μM EdU (Click-iT® Plus EdU Imaging Kits, ThermoFisher Scientific). After centrifugation, the cells were treated with 10% formaldehyde for 10 minutes at room temperature and 0.1% Triton for another 15 minutes at 37 °C. The cells were then washed with 4% FBS in PBS, spun down and stained with Alexa fluor 647 at room temperature for 30 minutes. Proliferation of the cells was measured using LSR II instrument (BD Biosciences, Franklin Lakes, NJ, USA).

Pyronin/Hoechst Staining
After resuspension in phosphate-citrate buffer solution (pH 4.8), the cells were incubated at room temperature for 20 minutes in the dark, spun down, washed repeatedly with PBS, spun down, washed once with 1 ml staining solution (PBS with 1.5 μg/mL pyronin Y and 2 μg/mL Hoechst 33342), spun down and incubated with 200 μL of the same staining solution for 10 minutes on ice. Cell cycle analysis was performed using LSR II instrument (BD Biosciences). All centrifugation steps were performed in the same way (1000 rpm, 5 min).

Apoptosis Assay
Apoptosis rate of the cells was in all cell lines determined without any treatment and after 72 hours of treatment with 10 µ M CDDP. The cells were stained with annexin V-Dy647 (RCANXD-T100, Exbio, Vestec, Czech Republic) and propidium iodide (PI) solution (130-093-233, Miltenyi Biotec, Bergisch Gladbach, Germany). The proportion of apoptotic and necrotic cells was measured using LSR II instrument (BD Biosciences).
Nucleic Acid Isolation DNA from the cell lines and patients´ frozen samples was isolated with DNeasy Mini Kit (Qiagen, Hilden, Germany); RNA from the cell lines and patients´ frozen samples was isolated with RNeasy Mini Kit (Qiagen, Hilden, Germany). DNA and RNA from FFPE samples were extracted with RecoverAll Total Nucleic Acid Isolation Kit for FFPE (ThermoFisher Scientific). Concentration, quality, and integrity of nucleic acids were evaluated by spectrophotometry-Nanodrop 2000 and Qubit 2.0 Fluorometer (both ThermoFisher Scientific), capillary electrophoresis-Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and by PCR. Synthesis of cDNA was performed using the iScript kit (Bio-Rad, Hercules, CA, USA) starting from 1 µ g of total RNA. All procedures were carried out according to the manufacturers' instructions.

Whole Exome Sequencing
Exome libraries from cell line DNA samples were prepared using SureSelectXT Human All Exon V5+UTR kit (Agilent Technologies) according to the manufacturer instruction, with starting amount of 50 ng of DNA. The samples were sequenced on NextSeq 500 Instrument (Illumina, San Diego, CA, USA) using a High Output Kit v2 (150 cycles) (Illumina) according to the manufacturer's instructions. Libraries from patients' samples were prepared using Nimblegen SeqCap EZ Human Exome Library v2.0 kit (Roche, Basel, Switzerland) starting with 1 µ g of gDNA and sequencing was performed on HiSeq 2500 (Illumina) in Atlas Biolabs (Berlin, Germany). The paired-end sequencing was employed (n ≥ 1.6 × 10 8 of paired reads) with median Phred quality score above 30. On average, 92% reads mapped to target regions (minimum 72%) with mean coverage 50×.

Whole Transcriptome Sequencing
Whole transcriptome sequencing was performed with RNA isolated from the TGCT cell lines. Transcriptome libraries were prepared using the SureSelectXT RNA Reagent Kit (Agilent Technologies), and carried out on NextSeq500 (Illumina) by using a High Output Kit according to the manufacturer's instruction. Mean coverage was 121.44 (minimum 76.43) with duplication rate below 60%.
Amplicon Sequencing DNA from FFPE samples of TGCT patients was subjected to amplicon sequencing using Ion AmpliSeq Comprehensive Cancer Panel (ThermoFisher Scientific) with all-exon coverage of over 400 genes involved in cancer pathogenesis. The libraries were prepared with Ion AmpliSeq Library Kit 2.0 and 200 bp chemistry (Ion PI Template OT2 200 kit v2 and Ion One Touch 2 System). The samples were sequenced using Ion PI Sequencing 200 Kit v2, Ion PI Chip v2 and the Ion Proton Sequencer platform (all ThermoFisher Scientific, Waltham, MA, USA) performed as described in [27]. All procedures were carried out according to the manufacturer's instructions and detailed protocols are available on manufacturer's websites.

Sequencing Data Analysis
The detected sequences were separated from the technical noise, their quality was checked (FastQC, QualiMap), they were aligned to the reference genome sequence (hg19) (BWA,) [77] and the variants were called (VarScan2 [78], samtools [79]). The alterations detected were further analyzed for their potential clinical effect (Ingenuity), annotated by SnpEff [80] and matched with the publically available databases, genome atlases and other on-line sources (The Cancer Genome Atlas, UCSC, ONCOMINE, CLINVAR). Aberrations were visualized in Integrative Genome Viewer [74], and checked manually for their presence (both strand reads), absence in control samples, frequency, quality and sequence context (e.g., a presence of homopolymers, etc.).
For WES sequencing, only significant variants that passed the filtration criteria were further evaluated (absolute quality reads number ≥5, variant allele ratio in control samples (germ-line or platinum sensitive) ≤0.03, variant allele ratio in studied primary tumor samples ≥0.3, in platinum treated cell lines ≥0.80, synonymous variants excluded). The identified variants were confirmed to be unique for cisplatin-resistant cells by checking their low/no presence in the CDDP-naive long-term in parallel cultured cell lines.
Similarly, significant variants reported in case of amplicon sequencing of primary tumor samples were filtered in if variant allele ratio was ≥0.2 in tumor samples and ≤0.03 in germline controls; variant effect was defined as nonsense, missense, stop/loss, frameshift, and unknown; and variant clinical effect defined as pathogenic, probably pathogenic, drug response, other and unknown.
Tumor mutation burden (TMB) was calculated as a number of called single nucleotide variants (SNVs) filtered in (with absolute quality reads number ≥5, variant frequency ≥0.05, synonymous variants excluded), related to the length of exons covered.
Copy number variants were predicted from WES data using CNV kit [81] with normalization to pooled control samples sequenced on the same instrument using identical library preparation kits. Predicted segmental changes were calculated for all samples and segments with predicted CN different from 2 were selected for further comparison. Selected regions were plotted in heatmap diagrams for all samples.