Two TGCT cell lines obtained from the European Collection of Authenticated Cell Cultures (ECACC) were used. NTERA-2 clone D1 (ECACC Cat# 01071221, RRID:CVCL_3407) is a cell line derived from a human testicular embryonal carcinoma and was first described in 1984 [32]. 577MF (ECACC Cat# 06011802, RRID:CVCL_2290) is a cell line derived from a human testicular teratocarcinoma and was first described in 1980 [33]. Both were cultured following ECACC recommended conditions and incubated in a humidified atmosphere with 5% CO${}_2$ at 37 °C. CDDP-resistance model was developed by growing the sensitive cell line NTERA-2 in increasing sub-lethal concentrations of CDDP in the growth medium. The starting dose was approximately the IC${}_{25}$ (inhibitory concentration, 25%) of the cell line for 72 hours. The medium was then replaced to let the cells recover for a further 72 hours. This development phase was conducted for approximately 8 months, after which the IC${}_{50}$ concentrations were re-assessed. The resistant cell line obtained (NTERA-2R) was cultivated in the same conditions of the parental cell line. Cell lines were negative for mycoplasm contamination (MycoAlert Mycoplasma Detection Kit, Lonza; tested monthly) and were authenticated by short tandem-repeat analysis at the Barretos Cancer Hospital facilities as reported [34].

Cisplatin (CDDP) (PHR1624; Sigma-Aldrich) was prepared at 5 mM in 0.9% NaCl. PCNA-I1 (SML0730; Sigma-Aldrich), ML323 (SML1177; Sigma-Aldrich), T2AA (SML0794; Sigma-Aldrich) and MG-132 (M7449; Sigma-Aldrich) were prepared at 10 mM in DMSO (Sigma-Aldrich). All drugs were aliquoted and stored at –20 °C until use.

Exponentially growing cells were seeded at a density of 5 $\times$ 10${}^3$ cells per well in 96-well plates. After leaving cells to adhere overnight, the drugs were added at zero (vehicle only) and seven increasing concentrations (0.0625, 0.125, 0.5, 1.0, 5.0, 15.0, 30.0, 50.0 $\mu$M). Cells were then incubated for a further 72 hours before measuring viability using CellTiter 96® Aqueous Non-Radioactive Cell Proliferation Assay (Promega). The optical density at 570 nm (OD570) was measured with a Varioskan Flash plate reader (Thermo Fisher Scientific) and expressed as a percentage of the value obtained from control cells. The assays were performed in both technical and biological triplicate.

IC${}_{50}$ concentrations were obtained in GraphPad Prism version 5.0 (GraphPad Software, San Diego, CA, USA), through the equation: log(inhibitor) vs. normalized response, using a Variable slope model.

The NTERA-2R IC${}_{50}$ was also verified after four months straightly in CDDP-free media culture to confirm the resistance phenotype stability.

Evaluation of combination treatment was performed using CDDP IC${}_{50}$ and a range of MG-132 concentrations from 0 to 5 $\mu$M.

Cells were seeded in triplicate in a 12-well plate at a density of 5 $\times$ 10${}^2$ cells/well. After 14 days, cells were fixed and stained with 1 mL of staining solution (0.5% crystal violet; 20% methanol) for 20 minutes. The plate was washed four times with distilled water, inverted on filter paper to remove the remaining water, and then air-dried for 24 hours at room temperature. The colonies were unstained using 1 mL of 100% methanol for 20 minutes, and the plates were read with a Varioskan Flash plate reader (Thermo Fisher Scientific) at 570 nm (OD570). The means and standard deviation of three independent experiments were analyzed.

To evaluate cell migration capacity, the monolayer wound-healing assay was performed. Briefly, a confluent monolayer of cells was seeded in a 6-well plate, and a “scratch” with a p1000 pipet tip was made through the cell layer. After washing several times with PBS, a medium containing 10% FBS was added to each well. Two fields of each wound were photographed in regular periods, from 0 to 64 hours. The wound areas were measured using TScratch software [35]. The experiments were performed in biological and technical triplicate.

Flow cytometry analysis of apoptosis and cell cycle were performed with FITC Annexin V Apoptosis Detection Kit I (BD Biosciences) and BD Cycletest Plus DNA Kit (BD Biosciences), respectively. Cells were seeded in T25 flasks and treated 24 hours later. After 72 hours, the culture supernatant and the cells were collected, washed twice with 1X PBS, and the specific protocols were followed as recommended by the manufacturer. Cell data was collected using BD Accuri Cytometer. Unstained and single-stained controls were used for color compensation, and at least 10,000 events were collected for each sample. The analysis was performed after two independent experiments, using FCS Express 7 software (De Novo Software, Pasadena, CA, USA).

Expression profile of NTERA-2R and NTERA-2P was performed using nCounter Vantage 3D DNA Damage and Repair panel (Nanostring, USA) as previously reported [36], to evaluate features related to CDDP-resistance. This panel includes 180 genes involved in major DNA damage repair pathways, including base excision repair, nucleotide excision repair, mismatch repair, TLS, and other repair processes (available at: http://nanostring.com).

Total RNA from NTERA-2P and NTERA-2R was isolated in biological triplicate, using TRIzol reagent, according to the manufacturer’s protocol. Probe pools, hybridization buffer, TagSet, and 100 ng total RNA (quantified by Qubit 2.0 Fluorometer) were hybridized for 21 hours at 67 °C, followed by purification and RNA/probe complexes immobilization in nCounter PrepStation (Nanostring, USA) and cartridge scanning in Digital Analyzer (Nanostring, USA), according to the manufacturer’s protocol.

The nSolver Analysis Software version 4.0 (NanoString Technologies, Seattle, WA, USA) was used for quality control assessment, and further steps were carried out in the R statistical environment, version 3.6.3. Gene expression levels and sample distribution were evaluated with quantro package, version 1.18.0 [37]. Data normalization and differential expression were performed in the NanoStringNorm package, version 1.2.1.1 [38]. Data were quantile normalized, and log2 transformed. Differentially expressed genes were defined by the thresholds of fold change $\ge$1.5 and p $\le$ 0.05. Heatmaps of differentially expressed genes were built with the ComplexHeatmap package, version 2.0.0 [39].

Venn diagram was created utilizing Venny 2.1 (RRID:SCR_016561, available at: https://bioinfogp.cnb.csic.es/tools/venny/). The STRING database was used to predict interaction networks from gene expression analysis. Clustering was performed using the K-means clustering method (RRID:SCR_005223, available at: https://string-db.org) [40]. Functional annotation analysis was performed using DAVID (RRID:SCR_001881, available at: https://david.ncifcrf.gov/home.jsp) [41, 42, 43].

Total proteins were extracted after 24 hours of treatment using RIPA buffer, including 10% protease and phosphatase inhibitors (Sigma-Aldrich). After 15 min on ice, samples were centrifuged at 13,000 g for 30 min at 4 °C, and the supernatant was collected. Protein concentration was determined using the Bio-Rad Protein assay, based on Bradford method (Bio-Rad) according to the manufacturer’s instructions. Twenty micrograms of protein were denatured at 95 °C for 5 min in 4X Laemmli buffer (Thermo Fisher Scientific), separated on NUPAGE 10 or 15% bis–tris gels at 90 V and transferred to 0.2 $\mu$m polyvinylidene difluoride (PVDF) membranes (GE Biosciences). Ponceau S solution (Sigma-Aldrich; 0.5% in 5% acetic acid) staining confirmed a correct membrane transfer. Membranes were blocked in 5% non-fat milk in TBS with 0.1% Tween (TBS-T, pH = 7.6) for 1 h and then incubated with primary antibodies overnight at 4 °C. Secondary horseradish peroxidase (HRP)-conjugated antibodies were incubated for 1 h at room temperature (RT). Bands were visualized using the ECL detection system (Cell Signaling), and the ImageQuant LAS 4000 (GE Biosciences) was used for imaging. Primary antibody probing was performed with PARP (Cell Signaling Technology Cat# 5625, RRID:AB_10699459/1:1000), and p-$\gamma$H2A.X (Cell Signaling Technology Cat# 2577, RRID:AB_2118010/1:1000). Anti-$\beta$-actin was used as a loading control (Cell Signaling Technology Cat# 8457, RRID:AB_10950489). Quantification was performed using band densitometry analysis from ImageJ software (version 1.6.1), by comparing the band intensity with the loading control $\beta$-actin. Experiments were performed in biological duplicate.

Data are presented as means $\pm$ the standard deviation of the means. Statistical analysis was performed using the GraphPad Prism 5.0 software. Unpaired Student’s t-test and non-parametric (Mann–Whitney) two-tailed tests were used among all assays, as necessary. A p-value $\le$ 0.05 was considered statistically significant.

We first performed a cell viability assay to determine the half-maximal inhibitory concentration (IC${}_{50}$) value of CDDP in two TGCT cell lines. Fig. 1 presents the two cell lines evaluated and their respective cell viability after treatment with increasing CDDP concentrations, ranging from 0.0625 $\mu$M to 50 $\mu$M. IC${}_{50}$ for NTERA-2 was 0.524 $\mu$M and for 577MF was 2.911 $\mu$M.

Fig. 1.

CDDP viability effect on NTERA-2 and 577MF TGCT cell lines. (A) Representative images of the cell lines (Objective: 20x). (B) CDDP at increasing concentrations was used to treat cells for 72 hours, and cell viability was evaluated. IC${}_{50}$ values are indicated on the graph.

To better understand CDDP-resistance mechanisms, we chose the most sensitive cell line in our previous analysis (NTERA-2) and established a CDDP-resistance model (NTERA-2R). After eight months of CDDP treatment with incremental doses, IC${}_{50}$ values were re-evaluated. A significant increase (p 0.01) was observed in the concentration of CDDP required to inhibit 50% of NTERA-2R cells (3.812 $\mu$M), compared to NTERA-2P (0.524 $\mu$M) after 72 hours treatment (Fig. 2A). Subsequently, cells were maintained in culture for four months in CDDP-free media, and the IC${}_{50}$ was re-assessed. The IC${}_{50}$ fold-increase in NTERA-2R (6.3 fold-increase) was similar to the IC${}_{50}$ previously found (7.2 fold-increase), confirming the CDDP-resistance phenotype stability, even without CDDP continuous exposure. We then performed a phenotypical characterization of NTERA-2R and found that it exhibited a higher aggressive phenotype when compared to NTERA-2P, demonstrating a significant increase in cell proliferation capacity (Fig. 2B), increased clonogenic survival (Fig. 2C,D), and higher migration rates (Fig. 2E,F).

Fig. 2.

Development and phenotypic characterization of TGCT CDDP-resistance model. (A) Cell viability of NTERA-2P and NTERA-2R after 72 hours of CDDP treatment. IC${}_{50}$ values are indicated on the graph. (B) Comparison of cell proliferation capacity, (C,D) cell survival, and (E,F) migration ability in both cell lines. Images are representative of the assays. * p 0.01; ** p 0.001; *** p 0.0001.

Levels of CDDP-induced apoptosis were assessed after 0.6 $\mu$M CDDP treatment for 72 hours. As expected, a significant increase in the apoptotic response to CDDP was seen in NTERA-2P (Annexin +/7AAD - and Annexin +/7AAD +), while almost no change was observed in NTERA-2R (Fig. 3A). Furthermore, cell cycle changes in response to 0.6 $\mu$M CDDP treatment for 72 hours were more prominent in NTERA-2P, presenting a markedly decrease of cells in G1 and an increase in S-phase and G2 (Fig. 3B). This data suggests the essential role of cell cycle control mechanisms in TGCT CDDP-resistance. Next, we treated both cells with their respective CDDP IC${}_{50}$ for 24 hours and performed an immunoblotting assay to verify $\gamma$-H2AX expression levels. We observed no difference in protein levels (Fig. 3C), indicating that the NTERA-2R resistance phenotype was not due to pre-target mechanisms, such as decreased CDDP intracellular accumulation caused by alterations in transporters or increased drug efflux.

Fig. 3.

Flow-cytometry analysis of CDDP effect on apoptosis and cell cycle distribution. (A) apoptosis and (B) cell cycle distribution in NTERA-2P and NTERA-2R, after 0.6 $\mu$M treatment for 72 hours. Differences are given in percentage. (C) Measurement of $\gamma$-H2AX protein expression in NTERA-2P and NTERA-2R, after CDDP IC${}_{50}$ treatment for 24 hours. The image represents one immunoblotting assay.

To elucidate the molecular changes associated with CDDP-resistance, we performed an analysis of both NTERA-2P and NTERA-2R 72 hours after CDDP (IC${}_{50}$) or vehicle (0.9% NaCl) treatment using the nCounter Vantage 3D DNA Damage and Repair panel from NanoString. Gene expression analysis comparing NTERA-2P and NTERA-2R treated with vehicle, revealed five differentially expressed genes (Supplementary Fig. 1). Among them, three genes were upregulated in NTERA-2R, two of which were related to DNA repair (MGMT and XPC), and one was responsible for a small subunit of DNA polymerase delta (POLD4). One of the genes downregulated in NTERA-2R encodes the catalytic subunit of DNA polymerase epsilon (POLE). Interestingly, high MGMT expression levels significantly correlated with a worst disease-free survival in TGCT cohort from the TCGA database (p = 0.0066), when considering the 15% patients with higher MGMT expression and the 15% patients with lower MGMT expression (Supplementary Fig. 2). We then performed a differential expression analysis of CDDP induced genes in parental (Fig. 4A) and resistant (Fig. 4B) cell lines, and found three genes that were exclusively altered in NTERA-2P and 21 genes exclusively altered in NTERA-2R. In the following analyses, we considered these 24 exclusively altered genes as the most important to explain our in vitro model’s resistant phenotype. In addition, we determined which genes were altered in common between NTERA-2P vs. NTERA-2PT (NTERA-2P treated with CDDP) and NTERA-2R vs. NTERA-2RT (NTERA-2R treated with CDDP) analyzes and nine genes were observed (Fig. 4C), suggesting that the regulation of these genes are independent of cisplatin resistance status.

Fig. 4.

Differential expression analysis of CDDP-induced gene expression. Nanostring DNA Damage and Repair panel revealed genes related to CDDP-resistance in (A) NTERA-2P and (B) NTERA-2R. Only genes with fold change $\ge$1.5 and p $\le$ 0.05 were included. (C) Venn diagram comparing all the genes differentially expressed in the analyses performed in A and B revealed 24 genes exclusively altered in P or R. P, NTERA-2P; R, NTERA-2R; PT, NTERA-2P treated with CDDP; RT, NTERA-2R treated with CDDP.

A list with all genes differentially expressed (fold change $\ge$1.5 and p $\mathrm{\le }$ 0.05) and their respective fold change is presented in Supplementary Tables 1,2.

We used the STRING database to predict the interactions between the 24 genes. We also used DAVID gene ontology (GO) to find the main regulated processes for these genes. These analyses revealed two clusters mainly related to DNA repair mechanisms and a cluster consisting of genes that regulate cell cycle, apoptosis, and cellular signaling (Fig. 5A). DAVID Functional GO-Analysis identified nine pathways significantly associated with the analyzed genes (p $\le$ 0.05), and DNA repair was the most significant (Fig. 5B). Interestingly, among the genes exclusively altered in NTERA-2P after CDDP treatment, none were related to DNA repair, while in NTERA-2R a set of DNA repair genes was differentially expressed. Another interesting finding was that after CDDP treatment, the antiapoptotic gene BCL2 was downregulated in NTERA-2P, while the antiapoptotic gene BCL2L1 was upregulated in NTERA-2R. These results suggest the idea that the main change in NTERA-2R is an increased DNA repair capacity and specific changes in cell cycle control, which may trigger apoptosis evasion and allow cells to proliferate, even in the presence of CDDP adducts.

Fig. 5.

Network and functional predictions of CDDP-induced gene expression. Network and functional predictions indicate the main clusters and pathways of the 24 genes exclusively altered in NTERA-2P and NTERA-2R cells after CDDP treatment. (A) STRING interaction prediction considering the altered genes. Genes primary function was identified, tabulated, and colors are according to STRING clusters. (B) DAVID Functional GO-Analysis identified nine pathways significantly associated with the analyzed genes (p $\le$ 0.05).

Considering the central role of cell cycle regulation and DNA repair mechanisms in our CDDP-resistance model, we hypothesized whether a panel of drugs containing three TLS inhibitors and one proteasome inhibitor, which was previously described as targeting TLS, could be an alternative to overcome GCT resistance. We performed a cell viability assay to determine the IC${}_{50}$ value of PCNA-I1, MG-132, ML-323 and T2AA, and found that all inhibitors induced some degree of cytotoxicity in both NTERA-2P and NTERA-2R (Fig. 6A). These results indicate that these inhibitors may be part of a new strategy to treat TGCTs, and more importantly, they could overcome CDDP-resistance in TGCTs. MG-132 demonstrated the more significant responses among the four inhibitors, showing strong cytotoxic activity in both NTERA-2P and NTERA-2R (IC${}_{50}$: 78.55 nM and 77.52 nM, respectively). Additionally, to confirm this drug’s potential, we verified its effect on the 577MF cell line, which is also a TGCT and presents an intermediate sensitivity to CDDP compared to NTERA-2P and NTERA-2R. The 577MF IC${}_{50}$ found after the cell viability assay was 88.39 nM (Fig. 6B). These data show that MG-132 effectively reduced cell viability at low nanomolar concentrations in all cell lines analyzed.

Fig. 6.

Effect of a panel of drugs in TGCT cell lines viability. (A) IC${}_{50}$ doses of the inhibitors based on MTS data in NTERA-2P and NTERA-2R cells. (B) IC${}_{50}$ dose of MG-132 based on MTS data in 577MF cell line.

In light of these data, we tested the potential of MG-132 and CDDP as a combinatory treatment. We first assessed cell viability treating NTERA-2P and NTERA-2R with a range of MG-132 concentrations combined with CDDP (IC${}_{50}$) for 72 hours. A remarkable enhancement of cytotoxicity was observed in the combinatory treatment for both cell lines. Interestingly, NTERA-2R was more responsive than NTERA-2P to the combinatory treatment (Fig. 7A). We then found that the treatment using the combination of CDDP (IC${}_{50}$) and MG-132 (25 nM – non-toxic concentration), for 72 hours, increased apoptosis when compared to CDDP alone in both cell lines (Fig. 7B). This result agreed with increased cleaved PARP levels found after the treatment with CDDP and MG-132 combination in both cell lines (Fig. 7C). Finally, we performed flow cytometry to compare the cell cycle distribution after treatment with CDDP (IC${}_{50}$) alone and in combination with MG-132 (25 nM) for 72 hours. As presented in Fig. 7D, CDDP alone increased the G2 population in both cell lines. When combining both drugs, we observed an increase in S population and a remarkable increase in the Sub-G0 population compared to CDDP alone in both cell lines. These results indicate that MG-132 treatment in non-toxic doses restores CDDP sensitivity of the resistant cell line.

Fig. 7.

MG-132 and CDDP combination indicate great therapeutic potential in both NTERA-2P and NTERA-2R. (A) Cell viability assay after MG-132 or CDDP (IC${}_{50}$) + MG-132 for 72 hours treatment. IC${}_{50}$ values are indicated on the graph. (B) Apoptosis assay after CDDP (IC${}_{50}$) or CDDP (IC${}_{50}$) + MG-132 (25 nM) for 72 hours treatment. (C) Western blot evaluation of PARP protein levels (total and cleaved) in cells treated for 24 hours with CDDP (IC${}_{50}$), MG-132 (25 nM), or the combination. $\beta$-Actin was used for data normalization. The image is representative of one immunoblotting assay. (D) Changes in cell cycle distribution after 72 hours of treatment with CDDP (IC${}_{50}$) or CDDP (IC${}_{50}$) + MG-132 (25 nM) were statiscally significant for both NTERA-2P and NTERA-2R (p 0.01).