2.1 Cell lines and cell culture

Human OS cell lines, U2OS and 143B, were obtained from Cell Bank, RIKEN BioResource Research Center (Tsukuba, Ibaraki, Japan). A nonmalignant human lung fibroblast line, IMR-90 (ATCC CCL-186), was obtained from the American Type Culture Collection (ATCC). U2OS cells were maintained in McCoy's 5A medium (Thermo Fisher Scientific), 143B cells in Dulbecco's Modified Eagle Medium (Thermo Fisher Scientific), and IMR-90 cells in Eagle's Minimum Essential Medium (Thermo Fisher Scientific), cultured at 37°C in a humidified atmosphere containing 5% CO₂. All cell culture media were supplemented with 10% heat-inactivated fetal bovine serum (FBS; JRH, Nichirei Biosciences, Tokyo, Japan), penicillin (100 units/ml; Thermo Fisher Scientific), and streptomycin (100 µg/mL; Thermo Fisher Scientific). Cisplatin was obtained from FUJIFILM Wako Pure Chemical Corporation. N-acetyl-cysteine (NAC) and ascorbate were purchased from Sigma-Aldrich (Merck Millipore). Two cisplatin-resistant U2OS cell lines were established as described previously for non-small cell lung cancer ³⁰ with a modification of cisplatin dose. Briefly, 1 × 10⁶ U2OS cells were cultured on three 10-cm dishes in the presence or absence of cisplatin. The cisplatin concentration was increased from 1 to 30 μmol/L over 6 months. The medium was changed every other day with occasional passage to maintain appropriate cellular confluency. Cisplatin resistance was confirmed by a cell proliferation assay.

2.2 Cell viability assay

Cellular viability was assessed using Cell Count Reagent SF (WST-8) (Nacalai tesque). U2OS, 143B, and IMR-90 cells were seeded on 96-well plates (1500 cells/well), 12 hours before starting treatment with cisplatin (1 nmol/L, 10 nmol/L, 100 nmol/L, 1 μmol/L, 10 μmol/L, and 30 μmol/L), ascorbate (1 nmol/L, 10 nmol/L, 100 nmol/L, 1 μmol/L, and 10 μmol/L), or cisplatin plus ascorbate. Ninety-six hours after treatment, 10 µL of WST-8 (2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium) was added to each well and incubated for 2 hours at 37°C, after which the absorbance at 450 nm was measured immediately on a microplate reader (iMark; Bio-Rad). The background readings were subtracted from each original reading. The cell viability assay was performed in triplicate and repeated at least three times. The IC₅₀ was calculated from the curves constructed by plotting cellular viability vs drug concentration. A drug combination effect between cisplatin and ascorbate was evaluated by Combination Index.³¹

2.3 Measurement of intracellular ROS

Intracellular ROS levels were measured by dihydroethidium (DHE) staining. U2OS cells were treated with cisplatin, ascorbate, or cisplatin plus ascorbate, or left untreated for 96 hours. The cells were trypsinized, collected in McCoy's 5A medium supplemented with 10% FBS, centrifuged at 1200 rpm at 4°C for 5 minutes, and then resuspended and incubated in phosphate-buffered saline (PBS) supplemented with 10 μmol/L DHE and 1% FBS at 37°C for 30 minutes in dark. After washing with PBS/1% FBS, cells were resuspended in 200-µL PBS/1% FBS and analyzed by flow cytometry (Canto II; BD Biosciences) and FlowJo (Becton Dickinson and Company). Intracellular ROS levels were determined by measuring the mean fluorescence intensity (MFI) of DHE-positive cells. The MFI of the treated cells was expressed relative to MFI of the untreated cells. ROS measurements were performed in triplicate and the entire experiment was repeated at least three times.

2.4 Live-cell metabolic assay

Cellular metabolism and mitochondrial function were measured as the extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) using a Seahorse Bioscience XFe96 Extracellular Flux Analyzer (Agilent Technologies). U2OS cells were seeded in the wells (1 × 10⁴ cells/well) of a standard XFe96 microplate (Agilent Technologies) and incubated at 37°C. Twelve hours later, the cells were treated with cisplatin, ascorbate, or cisplatin plus ascorbate, or left untreated for 96 hours and analyzed using the flux analyzer. Analysis medium supplemented with 1-mmol/L pyruvic acid, 10-mmol/L glucose, and 2-mmol/L glutamine (pH 7.4) was freshly prepared according to the manufacturer's protocol. For the measurement of mitochondrial and glycolytic functions, the assay conditions, including the final concentrations of FCCP (1.0 µmol/L), oligomycin (1.0 µmol/L), rotenone plus antimycin A (ROT/AA; 0.5 µmol/L), and 2-deoxy-D-glucose (2-DG; 50 mmol/L), were chosen based on our preliminary experiments. We routinely confirmed no significant changes in the number of cells in each well. The flux analyzer measurements were performed in triplicate and repeated at least three times.

2.5 Gamma-H2AX immunofluorescence, imaging, and image analysis

U2OS cells were seeded at a density of 1 × 10⁴ cells/well on a sterile round cover glass (Matsunami) set into a well of a 12-well plate. Cells were treated with cisplatin, ascorbate, or cisplatin plus ascorbate, or left untreated for 96 hours and then fixed with 4% paraformaldehyde in phosphate buffer solution (FUJIFILM Wako) for 5 minutes at room temperature. Fixed cells were permeabilized with Tris-buffered saline (TBS)/0.1% Triton X-100 at room temperature for 5 minutes. After 1 hour of blocking with TBS containing 2% bovine serum albumin at 37°C, the cells were incubated with mouse anti-γH2AX (pS139) antibody (1/10 dilution; BD Biosciences) at 4°C overnight. After washing with TBS/0.1% Triton X-100 three times, the cells were incubated with Alexa 488-conjugated anti-mouse antibody (1/200 dilution, Life Technologies) at room temperature for 1 hour. After washing again with TBS/Triton X-100 three times, cover glasses were mounted on glass slides using an antifade mounting medium (Thermo Fisher Scientific) containing DAPI to counterstain DNA. The slides were imaged under a fluorescence microscope (BZ-X710; Keyence, Osaka, Japan). For quantitative analysis, the number, size, and intensity of γH2AX foci in the nuclei were measured by image analysis software (Keyence). At least 10 views were randomly chosen from each slide for the analysis. γH2AX staining experiments were performed in triplicate and repeated three times.

2.6 Sphere formation assay

For the sphere formation assay, U2OS cells were resuspended in 0.33% agar gel and plated onto the top of a 0.66% agar gel layer in a 6-well plate. McCoy's 5A medium containing cisplatin, ascorbate, or cisplatin plus ascorbate was added to cover the agar gel layers. Ninety-six hours later, the medium containing the regents was removed and changed to normal McCoy's 5A medium. The medium was changed every other day and the number of spheres quantified 21 days posttreatment with cisplatin, ascorbate, or cisplatin plus ascorbate using inverted phase contrast microscopy (BZ-X710, Keyence). At least 15 fields in each well were randomly picked and the size and number of spheres were measured. The sphere formation assays were performed in triplicate and repeated six times.

2.7 Real-time RT-PCR

Total RNA was extracted from U2OS cells using the Monarch Total RNA Miniprep kit (New England Biolabs) and reverse-transcribed using the iScript^TM cDNA Synthesis Kit (Bio-Rad). cDNAs were subjected to quantitative real-time PCR using Luna Universal qPCR Master Mix (New England Biolabs). qPCR and data collection were performed using a StepOnePlus Real-time PCR system (Thermo Fisher Scientific). The following primer pairs were used: human HK2, 5ʹ-TGGAGATGGAGAATCAGA-3ʹ/5ʹ-CCAGGAAACTCTCGTCTA-3ʹ; G6PD (glucose-6-phosphate dehydrogenase), 5ʹ-CCGGATCGACCACTACCTGGGCAAG-3ʹ /5ʹ-GTTCCCCACGTACTGGCCCAGGACCA-3ʹ; LDHA (lactate dehydrogenase A), 5ʹ-GGTTGAGAGTGCTTATGA-3ʹ/5ʹ-AACACTAAGGAAGACATCA-3ʹ; MCT4 (monocarboxylate transporter 4), 5ʹ-ATTGGCCTGGTGCTGCTGATG-3ʹ/5ʹ- CGAGTCTGCAGGAGGCTTGTG-3ʹ; and HRPT1 (hypoxanthine phosphoribosyltransferase 1), 5ʹ-TTTGCTTTCCTTGGTCA-3ʹ /5ʹ-GCTTGCGACCTTGACCATCT-3ʹ.

2.8 RNA sequence analysis

RNA was extracted from cisplatin-resistant and parental U2OS cells using an RNeasy Micro Plus kit (Qiagen) according to the manufacturer's protocol. RNA integrity was assessed on a Bioanalyzer (Agilent), and high-quality samples were used for library preparation. RNAs were purified using the NEBNext Poly(A) mRNA Magnetic Isolation Module (New England Biolabs) and oligo-dT beads, and the libraries were prepared using the SMARTer RNA-seq kit (Takara Bio USA Inc). After normalization and pooling, the libraries were sequenced on a Nextseq 500 (Illumina) using 50 bp paired-end reads to a depth of >30 million reads per sample. The quality of the RNA-seq results was assessed using FastQC (ver. 0.11.7). The raw reads were trimmed and quality-filtered using Trim Galore! (ver. 0.4.4), Trimmomatic (ver. 0.36),³² and cutadapt (ver. 1.16) software. Clean reads were aligned using STAR (ver. 2.6.1a),³³ the count matrix generated using the featureCounts tool (ver. 1.6.1),³⁴ and differential expression analysis performed using DESeq (ver. 1.30.0) ³⁵ with standard settings. Genes with |log₂FC| ≥ 2 and a P-value < .05 were subjected to Gene Ontology analysis using clusterProfiler (ver. 3.6.0).³⁶ We compared the gene expression levels from parental and cisplatin-resistant U2OS cells and picked the genes with significant expression in gene set enrichment analysis (GSEA).³⁷ The RNA sequence data have been deposited in the DDBJ database (accession number: DRA009407).

2.9 Statistical analysis

Differences between groups were analyzed by an unpaired two-tailed Student's t test. Multiple groups were analyzed by one-way analysis of variance. Results are presented as the mean ± standard deviation. P < .05 was considered significant.

3.1 Ascorbate enhances the cytotoxicity of cisplatin in human OS cells

To assess the effect of ascorbate on cisplatin-induced cytotoxicity, we measured cellular viability after 96 hours of continuous cisplatin, ascorbate, or cisplatin plus ascorbate treatment. Cisplatin treatment decreased the viability of U2OS cells in a dose-dependent manner with an IC₅₀ of 15.5 μmol/L (Figure 1A). In contrast, ascorbate treatment alone did not significantly affect the viability of U2OS cells at doses between 0.001 and 10 μmol/L. At 100 μmol/L, ascorbate treatment markedly reduced cellular viability (Figure 1B). We next tested the chemosensitizing effect of ascorbate (1-30 μmol/L) on cisplatin. Although ascorbate treatment alone did not affect cellular viability at these doses, it enhanced the cytotoxic effect of cisplatin (Figure 1C). The IC₅₀ values for cisplatin upon combined treatment with cisplatin and ascorbate were 6.62 μmol/L with 1-μmol/L ascorbate, 1.90 μmol/L with 10-μmol/L ascorbate, and 0.06 μmol/L with 30-μmol/L ascorbate. The Combination Index was 0.47 with 1-μmol/L ascorbate and 0.56 with 10-μmol/L ascorbate, showing the synergistic effect of the combined treatment. In 143B cells, cisplatin treatment slightly decreased cell viability, with an IC₅₀ value of 532 μmol/L. The chemosensitizing effect of ascorbate on cisplatin was also observed in 143B cells. The IC₅₀ values for combined treatment with ascorbate were 90.5 μmol/L with 1-μmol/L ascorbate, 88.0 μmol/L with 10-μmol/L ascorbate, and 80.7 μmol/L with 30-μmol/L ascorbate. In contrast, ascorbate treatment did not affect the sensitivity of nonmalignant human lung fibroblast, IMR-90 cells to cisplatin (Figure 1G). These data indicate that ascorbate treatment synergistically enhanced the cytotoxic effect of cisplatin in a dose-dependent manner in human OS cells.

3.2 Synergistic ROS induction and DNA damage upon combined treatment with cisplatin and ascorbate

To gain insight into the potential mechanisms underlying the chemosensitizing effect of ascorbate on cisplatin treatment, we measured ROS production by DHE-based flow cytometry. U2OS cells were continuously exposed to cisplatin or cisplatin plus ascorbate at the indicated doses for 96 hours and intracellular ROS levels were measured. Cisplatin treatment increased intracellular ROS levels in a dose-dependent manner (Figure 2A). In addition, ROS levels significantly increased in the cells treated with cisplatin plus ascorbate compared to cisplatin treatment alone. To evaluate the kinetics of intracellular ROS production in response to treatment with cisplatin and ascorbate, we measured ROS levels after 24-, 48-, and 96-hour exposure. Although ascorbate treatment alone did not increase intracellular ROS levels, the combined treatment results in an increase after 24 hours exposure, with further increase over time (Figure 2B). Hence, cisplatin and ascorbate together enhance intracellular ROS production in U2OS cells.

To evaluate the biological effect of the cisplatin-induced increase in ROS, we examined double-strand breaks in U2OS cells treated with cisplatin or cisplatin plus ascorbate by immunostaining for phosphorylated histone H2AX (γH2AX), which accumulates at the sites of double-strand DNA damage. After 96 hours treatment of U2OS cells with cisplatin, ascorbate, or cisplatin plus ascorbate, we measured the number of γH2AX staining dots per cell. The number of γH2AX foci was comparable between untreated control cells and ascorbate-treated cells, but significantly higher in cisplatin-treated cells (Figure 2C,D). Combined treatment with cisplatin and ascorbate further increased the number of γH2AX foci than cisplatin treatment alone (Figure 2C,D). Less than 10% of cisplatin has been reported to bind DNA,³⁸ but it is possible that the increase in γH2AX foci is due to this direct effect, enhanced by ascorbate. Thus, to confirm that ROS production significantly contributes to the observed DNA damage, we examined the effects of a ROS scavenger, NAC, on γH2AX foci formation. The addition of NAC markedly suppressed γH2AX foci formation (Figure 2D). We can conclude, therefore, that cisplatin-induced ROS production leads to an increase in DNA damage, and that this effect is exacerbated upon addition of ascorbate.

3.3 Enhanced mitochondrial respiration altered the metabolism of U2OS cells

An altered cellular redox state due to increased ROS can shift the balance of metabolic processes in the cell. In particular, ROS can mediate pathophysiological effects via modulation of the glycolytic and mitochondrial respiratory pathways. Having established that cisplatin and cisplatin with ascorbate increase ROS production in OS cells, we thus examined the effect of these chemicals on glycolysis and mitochondrial function. Basal glycolysis was significantly increased by the combined treatment compared to no treatment or cisplatin treatment alone (Figure 3A). Compensatory glycolysis (assayed in the presence of mitochondrial inhibitors) was unchanged among U2OS cells treated with cisplatin, ascorbate, or cisplatin plus ascorbate, with a slight but insignificant increase observed in the cisplatin plus ascorbate-treated cells (Figure 3A). In contrast to changes observed for basal glycolysis, mitochondrial function was decreased upon combined treatment with cisplatin and ascorbate compared to cisplatin treatment alone and no treatment (Figure 3B). U2OS cells treated with both chemicals exhibited lower basal respiration, ATP production, and levels of maximal respiration. The data indicate that cisplatin and ascorbate, together, increase ROS to levels that induce mitochondrial damage. A reduction in oxidative respiration is countered by elevated glycolytic flux in the treated U2OS cells, presumably as a circumventive measure.

3.4 Cisplatin and ascorbate treatment influences gene expression in the glycolytic and pentose phosphate pathways

Since combined treatment with cisplatin and ascorbate increases basal glycolysis, we sought to confirm and explain this increase by examining the expression of several genes involved in glucose metabolism. Firstly, we examined the levels of mRNA for HK2, the first enzyme in the glycolytic pathway, which phosphorylates glucose to produce glucose-6-phosphate. Cisplatin treatment increases HK2 mRNA levels (Figure 4). Subsequently, we examined MCT4 (required for extracellular excretion of lactic acid produced by glycolysis), LDHA (converts pyruvate to lactic acid in the glycolytic pathway), and G6PD. G6PD is the first enzyme in the PPP, a pathway which provides NADPH for the synthesis of fatty acids, steroids, and ribose (for nucleotide and nucleic acid formation) and for maintaining reduced glutathione (an antioxidant). MCT4 and G6PD were upregulated in U2OS cells treated by cisplatin, and the expression levels were further increased by the combined treatment with cisplatin plus ascorbate. LDHA was upregulated in U2OS cells treated with cisplatin but not significantly enhanced by the combined treatment (Figure 4).

Since increased ROS are known to affect metabolic pathways, we examined whether the enhanced gene expression observed here on treatment with cisplatin (and/or combined cisplatin and ascorbate) is linked directly to ROS production. RNA levels for the genes were, therefore, assessed in the presence of NAC, a ROS scavenger. NAC treatment significantly suppressed the induction of HK2, MCT4, LDHA, and G6PD (Figure 4). These data further support our findings that cisplatin and ascorbate synergistically modify mitochondrial function and glycolytic metabolism in U2OS cells, and that this is achieved, at least in part, by increased ROS production.

3.5 Combined treatment with cisplatin and ascorbate suppresses U2OS sphere formation

A small subset of stem-like cells present in tumors, known as CSCs, are proposed to be responsible for cancer initiation, and tumor recurrence, metastasis, and chemoresistance.³⁹ Here, the sphere formation assay was used to identify a cell population with CSC properties in OS.⁴⁰ The assay is based on the ability of CSCs to form a three-dimensional sphere when grown in a gel matrix. To test whether combined treatment with cisplatin and ascorbate affects the sphere formation capacity of U2OS cells, the cells enclosed in agar gel were treated with cisplatin, ascorbate, or cisplatin plus ascorbate for 96 hours. The medium was changed to complete normal medium, and the culture continued for 21 days. Cisplatin plus ascorbate treatment significantly reduced the number and the average size of spheres, compared to untreated control and cisplatin or ascorbate treatment alone, or to untreated control (Figure 5A-C). These findings suggest that cisplatin-induced cytotoxicity is enhanced in the cells with sphere-forming ability (CSCs) in the presence of ascorbate.

3.6 Ascorbate increases the chemosensitivity of cisplatin-resistant U2OS cells

Acquired chemoresistance of cancer cells is both a huge concern and a limiting factor for chemotherapy programs. Our next objective, therefore, was to establish whether ascorbate treatment restores the cisplatin sensitivity of cisplatin-resistant cells. We first established U2OS cell lines that were less sensitive to cisplatin (CisLS-U2OS) by continuous exposure of parental U2OS cells to increasing doses of cisplatin over 6 months. The IC₅₀ value of 12.4 µmol/L in parental U2OS cells was shifted to 21.4 µmol/L in CisLS-U2OS cells (Figure S1). To characterize the CisLS-U2OS cells, we analyzed the gene expression by RNA sequencing (Figure S2A).⁴¹ A GSEA analysis was performed to identify differentially expressed gene profiles on comparison of the CisLS-U2OS cell line to parental cell lines. GSEA revealed that genes associated with ribosome biogenesis, mitochondrial function (including oxidative phosphorylation and respiratory chain function), and drug metabolism were enriched in CisLS-U2OS cells (Figure 6A and Figure S2B,C). Recent evidence indicates that MYC stimulates ribosomal biogenesis. Hence, using GSEA, we analyzed whether MYC target genes are induced in CisLS-U2OS cells, using two data sets for MYC target genes.^{42, 43} MYC target genes were indeed enriched in CisLS-U2OS cells (Figure 6B).

To further characterize the effects of cisplatin resistance, we assessed glycolysis and mitochondrial function in CisLS-U2OS cells relative to parental cells. Parameters of mitochondrial respiration, such as basal respiration, ATP synthesis, maximal respiration levels, and spare respiration, were markedly increased in CisLS-U2OS cells (Figure 6C, Left), which is consistent with the presence of genes for mitochondrial respiration in the GSEA. The additional observation of enhanced glycolysis in CisLS-U2OS cells suggests that the metabolic reprogramming of both the glycolytic pathway and mitochondrial function is an important process in the development of cisplatin resistance (Figure 6C, Right).

Finally, to test whether ascorbate restores cisplatin sensitivity in CisLS-U2OS cells, we examined the cellular viability of parental and CisLS-U2OS cells treated with cisplatin or cisplatin plus ascorbate. The addition of ascorbate shifted the IC₅₀ value of 21.6 µmol/L (for cisplatin treatment of CisLS-U2OS cells) to 3.61 µmol/L (Figure 6D). Thus, ascorbate overcomes cisplatin resistance.