Human EOC cell lines OVCAR8 and SHIN3 were kindly provided by Dr Peter Eck (University of Manitoba, Manitoba, Canada) and OVCAR3, OVCAR5, OVCAR10, SKOV3, A2780 and HIO-80 (an immortalized, nontumorigenic human ovarian epithelium cell line) were kindly provided by Dr Thomas Hamilton, or were derived by Dr Andrew K. Godwin, both of the Fox Chase Cancer Center (Philadelphia, USA). SHIN3 cells were maintained in DMEM supplemented with 10% FBS, (both Sigma-Aldrich; Merck KGaA) and 1% penicillin-streptomycin. HIO-80 was cultured in M199/MCDB105 medium (1:1, v/v) containing 4% FBS, insulin (0.3 U/ml) and 2 mM L-glutamine. The remaining cell lines were cultured in PRMI-1640 medium supplemented with 10% FBS and 1% penicillin-streptomycin. All cells were cultured at 37°C with 5% CO₂ and 85–95% humidity. Cell line authentication was carried out by the Clinical Molecular Oncology Laboratory of University of Kansas Medical Center (Kansas, USA) using multiplex short tandem repeat DNA profiling.

L-Ascorbic Acid (Thermo Fisher Scientific, Inc.) was prepared as 1 M stock solutions in sterile water, with sodium hydroxide added drop-wise to adjust the pH to 7.0. Aliquots were stored at −80°C and thawed for single use. Catalase (Sigma-Aldrich; Merck KGaA) was prepared in distilled water at 10,000 units/ml, and was used at a working concentration of 600 units/ml. Olaparib and veliparib were obtained from Selleck Chemicals and were prepared in dimethyl sulfoxide (DMSO) and diluted with cell culture media to working concentrations, for the in vitro experiments. For the in vivo experiments, olaparib was dissolved in PBS containing 10% 2-hydroxy-propyl-betacyclodextrin (Sigma-Aldrich; Merck KGaA). All other reagents and chemicals were obtained from Thermo Fisher Scientific, Inc., unless specifically indicated.

The BRCA1/2 wild-type status was reported previously (30) for all the EOC cell lines used in the present study except SHIN3. The genomic DNA of SHIN3 cells was extracted using a Blood & Cell Culture DNA Mini kit (Qiagen GmbH). The largest and functionally most important exon (exon 11) of both BRCA1 (3,630 bp) and BRCA2 (5,018 bp) was amplified from the genomic DNA template using PCR as previously described (31). The PCR amplicons were submitted to Genewiz, Inc. for DNA sequencing. The primer sequences are provided in Table SI. The thermocycling conditions and Taq enzyme used were as previously described (31). DNA sequences were analyzed using the DNASTAR analysis package (version 8.1; DNASTAR, Inc.). Both nucleic acid and amino acid sequences were aligned using BioEdit (version 7.2) (32).

Cells were seeded at a density of 1×10⁴ cells per well in a 96 well plate, and incubated overnight. Cells were then exposed to a serial dilution of ascorbate (0–3.5 mM), olaparib (0–1,000 µM in SHIN3 cells; 0–800 µM in OVCAR5 cells) and veliparib (0–1,000 µM in SHIN3 cells; 0–800 µM in OVCAR5 cells), or treatment combinations and incubated for 24 or 48 h. In the drug combination groups, either olaparib or veliparib was added 15 min prior to ascorbate treatment. Following treatment, the culture medium was replaced with fresh, drug-free medium, and cells were incubated with MTT for 4 h. Formazan crystals were dissolved using DMSO and the absorbance at 492 nm was measured on a Synergy™ 4 Hybrid microplate reader (BioTek Instruments, Inc.). The half maximal inhibitory concentration (IC₅₀) was determined using a non-linear regression analysis to fit the data to the log₁₀ [inhibitor] compared with a normalized response with a variable slope model.

Different concentrations of ascorbate (ranging from 0–5 mM) were used to avoid drawing conclusions from a single particular concentration. Concentration at IC₅₀ or a concentration range including the IC₅₀ were used. If the treatment time was <48 h, concentrations >IC₅₀ were used, with additional multiple concentrations including at least one close to or lower than the IC₅₀. The in vitro concentration ranges used in the present study are easily achievable in patients by intravenous ascorbate infusion (26).

PAR levels were measured using a HT PARP in vivo Pharmacodynamic assay II (Trevigen, Inc.), and normalized to the protein contents. Protein concentrations of cell lysates were measured using a Pierce bicinchoninic acid assay kit (Thermo Fisher Scientific, Inc.).

Cells were lysed in ice-cold radioimmunoprecipitation buffer (Thermo Fisher Scientific, Inc.), supplemented with cOmplete™ Mini Protease Inhibitor Cocktail Tablets (Sigma-Aldrich, Merck KGaA) and Halt™ Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific, Inc.). Protein concentration was determined using the Bradford Protein Assay Kit (Bio-Rad, Inc.). A total of 60 µg protein/lane was resolved on the 4–20% Mini-PROTEAN TGX™ Precast gels (Bio-Rad, Inc.) and transferred onto polyvinylidene difluoride (PVDF) membranes (Bio-Rad, Inc.). The membranes were blocked using 5% skim milk in TBST (20 mM Tris_HCl, pH 7.4, 150 mM NaCl, 0.1% Tween 20) for 1 h at 4°C, followed by incubation at 4°C overnight with specific antibodies against H₂AX (1:500; Cell Signaling Technology, Inc.; cat. no. 7631); p-H₂AX^Ser139 (1:1,000; Cell Signaling Technology, Inc.; cat no. 9718); ATM (1:1,000; Cell Signaling Technology, Inc.; cat. no. 2873); p-ATM^Ser1981 (1:500; Cell Signaling Technology, Inc.; cat. no. 13050); BRCA1 (1:1,000; Cell Signaling Technology, Inc.; cat. no. 14823); BRCA2 (1:1,000; R&D Systems, Inc.; cat. no. MAB2476); RAD51 (1:1,000; Cell Signaling Technology, Inc.; cat. no. 8875); Ku70 (1:1,000; Cell Signaling Technology, Inc.; cat. no. 4588); Ku80 (1:1,000; Cell Signaling Technology, Inc.; cat. no. 2,180); p-DNA-PKcs^Thr2609 (1:350; Thermo Fisher Scientific, Inc.; cat. no. PA5-12913); DNA-PKcs (1:4,000; Santa Cruz; cat. no. sc-9051); β-actin (1:5,000; Thermo Fisher Scientific, Inc.; cat. no. MA5-15739); or vinculin (1:1,000; Cell Signaling Technology, Inc.; cat. no. 13901). Target proteins were visualized using horseradish peroxidase (HRP)-conjugated goat-anti-rabbit IgG (1:5,000; Cell Signaling Technology, Inc.; cat. no. 7,074) or HRP-conjugated horse-anti-mouse IgG (1:5,000; Cell Signaling Technology, Inc.; cat. no. 7076) for 1 h at room temperature with Pierce™ ECL Plus Western blotting substrate (Thermo Fisher Scientific, Inc.). Each western blot analysis was performed. b-actin or vinculin were used as the loading controls. In vivo xenograft mouse model. All procedures were performed in accordance with a protocol (ACUP #2018-2443) approved by the Institutional Animal Care and Use Committee of the University of Kansas Medical Center (Kansas, USA). The intraperitoneal (i.p.) tumor xenografts were established via i.p. injection of 2×10⁶ SHIN3 cells suspended in 200 µl PBS in fifty 4–6-week-old female athymic NCr-nu/nu mice (20–25 g body weight; National Cancer Institute). A total of 2 weeks after cell injection, mice were randomly grouped as follows: i) Control group; i.p. injection of saline solution osmotically equivalent to ascorbate twice daily and olaparib's solvent (PBS containing 10% 2-hydroxy-propyl-betacyclodextrin) once daily in volumes equivalent to the olaparib treated group; ii) ascorbate group, i.p. injection of ascorbate at 4 g/kg twice daily; iii) olaparib group, i.p. injection of olaparib at 50 mg/kg once daily; and iv) combination of ascorbate and olaparib, which were prepared and administered in the same manner as individual drug treatments. After 25 days of treatment, all mice were euthanized by CO₂ inhalation in a closed chamber (20% volume/min) followed by bilateral thoracotomy as approved in the protocol, and gross necropsy was performed with tumor weights and ascites volumes measured, and the number of tumor cells in the ascetic fluids counted. The total tumor burden of each mouse was indicated as total tumor weight at the end of experiment. The liver, kidney and spleen from each group were subjected to histopathological analysis using hematoxylin and eosin (H&E) staining, on 4-µm tissue sections with an automated procedure, as previously reported (33).

All statistical analyses were performed using GraphPad Prism version 5.0 (GraphPad Software, Inc.). Multiple comparisons between groups were performed using a one-way ANOVA with a post-hoc Turkey's test with a family-wise error rate of 0.05. Adjusted P<0.05 was considered to indicate a statistically significant difference.

Exon 11 is the largest and most functionally important exon. To the best of our knowledge, the current study was the first to sequence the BRCA1 and 2 genes (both of exon 11) of SHIN3 cells. As indicated in Fig. S1, a single variant of A→G at codon 349 of BRCA1 was detected, which is not known to have a functional outcome or be associated with breast or ovarian cancer. In addition, two variants were detected at codon 2660 A→G and codon 4560 G→C, in exon 11 of BRCA2, which do not result in amino acid changes. The present results indicate that no functional mutations were detected in exon 11 of BRCA1 and 2 in SHIN3 cells. A panel of BRCA1/2 wild-type human ovarian cancer cell lines (A2780, OVCAR10, OVCAR3, OVCAR5, SKOV3, OVCAR8 and SHIN3) (34), and an immortalized, non-tumorigenic human ovarian epithelium cell line (HIO-80) were then screened for sensitivity to ascorbate. As presented in Fig. 1A and Table SII, the IC₅₀ values of ascorbate in the ovarian cancer cells ranged from 0.15–2.80 mM, which is readily achievable by i.v. ascorbate infusion (23,26). In contrast, HIO-80 cells were resistant to ascorbate treatment in the tested concentration range (0.0–3.5 mM), with an IC₅₀ value >3.5 mM (Fig. 1A). When catalase, a H₂O₂ scavenger, was added to the culture medium, SHIN3 cells were protected from the cytotoxic effects of ascorbate (P<0.001; Fig. 1B), suggesting that the ascorbate cytotoxicity was mediated through H₂O₂, consistent with previously published studies (21,23).

The cell lines exhibiting the highest levels of resistance to ascorbate (SHIN3) and a moderately resistant cell line OVCAR5 were selected as representative cell lines, and their sensitivity to the PARPis olaparib and veliparib. In agreement with their BRCA wild-type status, both SHIN3 and OVCAR5 cells exhibited resistance to the PARPi treatments, as the concentrations of PARP required to exert inhibitory effects on cell viability were particularly high (35) (Fig. 1C and D; Table SII).

Since ascorbate-induced H₂O₂ damage to DNA in EOC cells (20) and PARPis impair DNA damage repair, it was hypothesized that the combination of pharmacological ascorbate and PARPis may enhance DNA repair deficiency and improve therapeutic efficacy against EOC. In order to verify this hypothesis, the effects of the combination treatment of pharmacological ascorbate and PARPi olaparib or veliparib in SHIN3 and OVCAR5 cells were determined. Treatment with olaparib (20 µM) or veliparib (20 µM) for 24 and 48 h, respectively, minimally affected the cell viability of SHIN3 and OVCAR5 (Fig. 2A and B). The combined treatment of pharmacological ascorbate with either olaparib or veliparib significantly decreased cell viability compared with either single drug treatment and vehicle control, in both SHIN3 and OVCAR5 cells (Fig. 2A and B). PAR levels were significantly increased by pharmacological ascorbate in SHIN3 cells compared with the control, as early as 15 min (P<0.05; Fig. 3), suggesting that PARP was activated as a cellular response to the H₂O₂-induced DNA damage mediated by ascorbate. Treatment with olaparib significantly decreased PAR levels in the presence of ascorbate compared with the control (P<0.05; Fig. 3), suggesting that PARP-mediated DNA repair was inhibited. Taken together, these results suggest that when PARP activity is inhibited, pharmacological ascorbate significantly potentiates cell death, potentially through enhanced DNA damage in BRCA1/2 wild-type EOC cells.

As pharmacological ascorbate induced H₂O₂ and caused DNA damage in cancer cells (20), the effects of pharmacological ascorbate on DDR in BRCA1/2 wild-type EOC cells was determined in the present study, with a focus on the HR and non-homologous end joining (NHEJ) signaling pathways. Treatment with pharmacological ascorbate for 2 and 6 h both notably increased p-H₂AX^Ser139 levels in both SHIN3 and OVCAR5 cells, a marker of DNA DSBs (Fig. 4). Expression of BRCA1 was downregulated in both tested cell lines 2 h post-ascorbate treatment and further decreased 6 h post treatment (Fig. 4). Notably, the expression of BRCA2 and RAD51 was slightly increased in SHIN3 cells 2 h post treatment, followed by decreases 6 h post treatment (Fig. 4), suggesting that pharmacological ascorbate transiently activated HR repair machinery as part of the stress response, and then inhibited the HR repair machinery. Similar patterns of changes in the expression of BRCA2 and RAD51 were observed in OVCAR5 cells (Fig. 4). Together, these data suggest that pharmacological ascorbate inhibited HR DNA DSBs repair pathway by decreasing the expression of BRCA1, BRCA2 and RAD51 in BRCA1/2 wild-type EOC cells.

The NHEJ pathway was also investigated. The data revealed that pharmacological ascorbate treatment minimally affected the expression of Ku70 and Ku80 in SHIN3 cells within the first 2 h. After 6 h of treatment, a dose-dependent decrease in the expression of Ku70 and Ku80 was observed following ascorbate treatment (Fig. 4). Ascorbate treatment also decreased the expression of Ku70 and Ku80 in a dose-dependent manner in OVCAR5 cells at either 2 or 6 h of treatment (Fig. 4). Ascorbate treatment increased the levels of p-DNA-PKcs^Thr2609, a product of activated NHEJ, in SHIN3 and OVCAR5 cells after of 2 h treatment as part of the stress responses. After 6 h, p-DNA-PKcs^Thr2609 levels were decreased in OVCAR5 cells, but not in SHIN3 cells (Fig. 4). However, the expression of total DNA-PKcs was decreased by ascorbate in a dose-dependent manner in SHIN3 cells after both 2 and 6 h treatment. In OVCAR5 cells DNA-PKcs was first upregulated (2 h) and then downregulated (6 h) following ascorbate treatment (Fig. 4). These data suggest that the regulation of pharmacological ascorbate on the NHEJ DNA repair proteins (Ku70, Ku80 and DNA-PKcs) was cell-line dependent with an overall tendency of inhibition.

As presented in Fig. 5A, treatment with ascorbate alone or olaparib alone increased the expression of p-ATM^Ser1981, an early marker of oxidative DNA damage, and the expression of p-H₂AX^Ser139, a marker of DNA DSBs, in a dose-dependent manner in SHIN3 cells treated for 2 h. The combination treatment of ascorbate and olaparib further enhanced the expression of p-ATM^Ser1981 and p-H₂AX^Ser139 compared with both drugs alone. Treatment with catalase decreased the levels of p-ATM^Ser1981 and p-H₂AX^Ser139 when it was added prior to treatment with olaparib and ascorbate, suggesting that catalase protects cells from oxidative DNA damage. Similar observations were also observed in OVCAR5 cells (Fig. S2).

In addition, the time-dependent expression levels of p-ATM^Ser1981 and p-H₂AX^Ser139 were detected following incubation with 3 mM ascorbate in SHIN3 cells (Fig. 5B). Consistently, the combination treatment of ascorbate and olaparib further increased the expression of p-ATM^Ser1981 and p-H₂AX^Ser139 compared with either single agent treatment alone at the same time points (Fig. 5B).

In order to assess the therapeutic efficacy of combination pharmacological ascorbate and olaparib in vivo, a validated xenograft mouse model of ovarian cancer was used (20). This model mimics the later stages of EOC with clinical observations of abdominal tumor dissemination and ascites formation in patients. After 25 days of treatment, ascorbate (4 g/kg) alone significantly decreased tumor weight by 32.5% compared with the vehicle control (P<0.05; Fig. 6A). Olaparib (50 mg/kg) alone had minimal effects (P>0.05) on the tumor weight of these BRCA wild-type xenografts (Fig. 6A). The combination treatment of ascorbate and olaparib significantly decreased tumor weight by 48.6% compared with the vehicle control (P<0.05; Fig. 6A). Treatment with ascorbate or olaparib alone did not significantly decrease the ascites volumes or number of tumor cells in the ascites at the end of treatment, the combination treatment resulted in a 73.8% decrease in ascites volume (P<0.05; Fig. 6B) and a 71.7% decrease in the number of tumor cells in the ascetic fluids (P<0.05; Fig. 6C) relative to the vehicle control. The decrease was significant compared with either treatment alone. All treatments were well tolerated and did not result in weight loss (Fig. 6D). H&E staining demonstrated no pathological changes in the livers, kidneys or spleens of the animals in all treatment groups, suggesting that ascorbate, olaparib and the combination treatment were of low toxicity at the tested concentrations (Fig. 6E).