Human pancreatic ductal adenocarcinoma cell lines MIA PaCa-2 (RRID: CVCL_0428) and PANC-1 (RRID: CVCL_0480) were cultured in DMEM (Gibco, 11965) supplemented with 10% FBS (Gibco, 26140) and 1% penicillin–streptomycin antibiotic (Gibco, 15140). The human patient-derived pancreatic ductal adenocarcinoma cell line 339 was cultured in DMEM/F-12 media (Gibco, 11320) supplemented with 10% FBS, insulin (Gibco, 12585), EGF (Gibco, PHG0311), hydrocortisone (Sigma, H0888), bovine pituitary extract (Gibco, 13028), and 1% penicillin–streptomycin antibiotic. Nontumorigenic HPV16-E6E7-immortalized cell line derived from normal pancreatic ductal epithelium (H6c7) (RRID: CVCL_0P38) with near normal genotype and phenotype of pancreatic duct epithelial cells were used for comparison. H6c7 cells were cultured in Serum Free Keratinocyte media (Gibco, 10724) supplemented with EGF (1 ng/mL), bovine pituitary extract (50 mg/mL), and 1% penicillin–streptomycin antibiotic. Human cell lines (MIA PaCa-2, PANC-1, and H6c7) were purchased directly from the ATCC and were passaged for fewer than 6 months after receipt. No additional authentication was performed. The patient-derived cell line (PDX339) was obtained from the Medical College of Wisconsin surgical oncology tissue bank (27). In addition, MIA PaCa-2 cells were transfected with plasmids containing mUNG1 (pMA3790, RRID: Addgen_70110) or HSV-1 UL12.5M185 (pMA4008, RRID: Addgene_70109) gene to generate cells without functional mitochondria (ρ⁰ cells) (28). Mitochondrial expression of DNA repair enzymes mutant Y147A human uracil-N-glycosylase (mUNG1) or Herpes Simplex Virus 1 (HSV-1) protein UL12.5M185 leads to mtDNA depletion and the formation of ρ⁰ cells. ρ⁰ cells were grown in DMEM containing 10% FBS, 50 μg/mL uridine and 1 mM sodium pyruvate. Pyruvate was removed prior to treatment with P-AscH⁻ as pyruvate can scavenge H₂O₂ (29). Frozen cell lines were routinely tested for mycoplasma. Regardless of varying cell type and media components, all cells were treated in fresh 10% DMEM media with ascorbate for 1 h at 37 °C. After the 1 h treatment, media were replaced, i.e., P-AscH⁻ was removed, and cells were allowed to incubate for 48 h. P-AscH⁻ came from a stock solution of 1.00 M (pH 7) made under argon and stored with a tight-fitting stopper at 4 °C. Ascorbate concentration was verified at 265 nm, ε265 = 14,500 M⁻¹ cm⁻¹(30). To enhance rigor and reproducibility, final concentrations of the exposure to cells were calculated in units of moles-per-cell to account for variation in media, cell density, and cellular metabolism (31).

Clonogenic survival assays were performed as previously described (4). Approximately 1–3 × 10⁵ cells were seeded 1 d prior to assay. Cells were treated with P-AscH⁻ (10–30 pmol/cell, 1–2 mM, LD₅₀ for each individual cell line) for 1 h in DMEM/10% FBS followed by SCH772984 (Selleckchem, S7101) or Mdivi-1 (Sigma, M0199) for 24 h. Cells have different susceptibility to P-AscH⁻ (31). Thus, the LD₅₀ for clonogenic survival for each cell line was used as a guide to determine appropriate doses of P-AscH⁻ to use in the various experiments. Cells were trypsinized with TrypLE Express (Gibco, 12604) to form a single-cell suspension and counted using a Countess II automated cell counter (Thermo Fisher) to determine the number of cells plated into each well. Cells were allowed to form colonies for 7 to 14 d before being fixed with 70% ethanol and stained with Coomassie Blue. Colonies containing ≥ 50 cells were scored. Surviving fraction was defined as number of colonies counted/number of cells seeded. Each experimental condition was normalized to their own control to determine a normalized surviving fraction. Experiments were repeated in triplicate.

Protein samples were isolated and prepared in phosphosafe buffer (EDM Millipore, 71296) containing protease inhibitor cocktail (Sigma Aldrich). Protein content was measured using a DC Protein assay. Total protein (10–40 mg) was loaded on a 4%−20% SDS-PAGE gradient gel. Protein was transferred to Immun-Blot PVDF membrane (Bio-Rad). Membranes were blocked in 5% BSA in PBST. Primary antibodies included: P-ERK (1:1000; #4370, Cell Signaling), ERK (1:1000; #4695, Cell Signaling), Drp1 (1:1000; #5391S, Cell Signaling), P-Drp1 (1:1000; #4494S, Cell Signaling), tubulin (1:2,000; #E7; University of Iowa Developmental Studies Hybridoma Bank, RRID: AB_528499), and GAPDH (1:5,000; #14C10, Cell Signaling). Appropriate horseradish peroxidase–linked secondary antibodies were used at a concentration of 1:20,000. Blots were visualized with SuperSignal West Pico substrate (Thermo Scientific, 34579) on x-ray film. Quantification of the in vivo sample westerns was performed by comparing fraction of p-ERK/ERK. Normalized average integrated densities were determined with ImageJ software (RRID: SCR_003070).

Cells were treated with P-AscH⁻ for 1 h, then incubated for 48 h. Cells were then labeled with 100 nM MitoTracker^© Green FM for 20 min in DMEM phenol-red free medium. Confocal fluorescence images were taken with Zeiss LSM710. Imaris Image Analysis Software (Bitplane) was used to create images of the mitochondrial network and to calculate mitochondrial volume and number of compartments.

Tumor samples were fixed with 4% paraformaldehyde in 0.1 M PBS at 4 ° C overnight. Samples were embedded in OCT and 8 μm sections were cut in a Thermo HM525 Cryostat (Thermo Scientific). OCT sections were warmed up to room temperature and washed with PBS before blocking with 5% normal goat serum (NGS) for 30 min. Sections were incubated with mitochondrial cytochrome c oxidase 1 mouse antibody (1:50, Abcam, RRID: AB_2084810) and p-Drp1 (Ser616) rabbit antibody (1:50, #4494S Cell Signaling) in 1% NGS overnight at 4 ° C. An Alexa Flour 488 conjugated goat anti-mouse (1:200, RRID: AB_10892893) or goat anti-rabbit (Alexa Flour 568 conjugated) was used as secondary antibody. DAPI was used to stain the cell nuclei. Tumor tissue sections were examined with a Zeiss LSM710 confocal microscope (Central Microscopy Research Facility, University of Iowa).

MIA PaCa-2 cells grown on 18 mm glass cover slips in a 12-well plate treated with P-AscH⁻ 1 mM (15 pmol/cell) for 1 h. After treatment, cells were cultured in fresh growth media for 48 h, then were fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.4 overnight at 4 °C. After several 0.1 M cacodylate buffer rinses, cells were stained in 0.1% tannic acid followed by post-fixation in 1% OsO₄ in 0.1 M cacodylate buffer plus 0.8% potassium ferricyanide, and then post fixed with 1% OsO₄ for 1 h. Following serial alcohol dehydration (50%, 75%, 95%, and 100%), the samples were embedded in Epon 12 (Ted Pella). Ultramicrotomy was performed, and ultrathin sections (70 nm) were post stained with uranyl acetate and lead citrate. Samples were visualized with a JEOL 1230 transmission electron microscope (TEM) at the Central Microscopy Research Facility of The University of Iowa.

Transfection of siRNA was carried out using a Lipofectamine 2000 Reagent (Life Technologies) according to manufacturer’s direction. Briefly, cells (1.5 × 10⁵ /well) were seeded in 6-well tissue culture plates, Lipofectamine and siRNA (100 nmol) were diluted in Opti-MEM reduced serum medium before mixing; the transfection complex was incubated at room temperature for 5 min before adding to the cells. Clonogenic assays and western blotting were performed 48 h after transfection. Drp1 siRNA was purchased from Santa Cruz Biotechnology, siERK (p44/42) and control siRNA were from Cell Signaling.

The animal protocols were reviewed and approved by the Animal Care and Use Committee of The University of Iowa. Thirty-day-old female athymic nude mice (Foxn1^nu) were obtained from Envigo (RRID: RGD_5508395). Animals were acclimated in the unit for 1 week before any manipulations were performed. For experiments growing xenografts with MIA PaCa-2 cells (2 × 10⁶), tumors were delivered subcutaneously into each flank region of nude mice with a 1-mL tuberculin syringe equipped with a 25-gauge needle. Tumor size was measured every two to three days by means of a vernier caliper, and tumor volume was estimated according to the following formula: tumor volume = Π/6 x L x W², where L is the greatest dimension of the tumor, and W is the dimension of the tumor in the perpendicular direction Tumors were allowed to grow to approximately 5 mm in diameter before experimental treatment began. For tumor isolation experiments for Drp1 colocalization, mice were treated P-AscH⁻ (4 g/kg, i.p., b.i.d.) or an equivalent volume of saline (1 mol/L) for 5 d. Mice were euthanized by CO₂ asphyxiation, and tumors were harvested and processed for experimental analyses. For survival experiments, mice were treated daily for 21 d with P-AscH⁻ (4 g/kg, i.p.) while mitochondrial division inhibitor-1 (Mdivi-1) was administered i.p. 50 mg/kg. Mice body weights were measured twice per week during the treatment period.

Data are presented as the mean ± SEM. For statistical analyses of two groups, unpaired two tailed Student t tests were utilized. To study statistical differences between multiple comparisons, significance was determined using one-way ANOVA analysis with Tukey multiple comparisons test. For survival curve analysis, Log-rank (Mantel-Cox) test was used to compare median survival. All analyses were performed in GraphPad Prism (GraphPad Software, Inc.; RRID: SCR_002798).

Phosphorylated ERK1/2 (p-ERK) is the active form of ERK known to activate Drp1, a GTPase protein which alters mitochondrial morphology by inducing fission. Western blot analysis showed that treatment for 1 h with P-AscH⁻ (1–2 mM) induced an increase in p-ERK 48 h after treatment in the 339 (Figure 1A) and the MIA PaCa-2 (Figure 1B) PDAC cell lines. This increase in p-ERK was not observed in the nontumorigenic H6c7 pancreatic ductal epithelial cell line (Figure 1C). A time course experiment using the MIA PaCa-2 cell line confirmed that the increases in p-ERK (Figure 1D) and p-Drp1 (Figure 1E) do not occur until 48 h after treatment with P-AscH⁻. Previous studies have shown that catalase reverses the cytotoxic effects of P-AscH⁻, consistent with a H₂O₂-mediated mechanism of selective toxicity to PDAC cells (4). Treatment of MIA PaCa-2 cells with catalase (100 μg/mL) prior to treatment with P-AscH⁻ reversed the increase in p-Drp1 and p-ERK (Figure 1F). These data suggest that the P-AscH⁻-induced increases in p-ERK and p-Drp1 are mediated by H₂O₂. To investigate if the P-AscH⁻-induced increases in p-ERK occur in vivo, established tumor PANC-1 xenografts were treated with P-AscH⁻ (4g/kg i.p., b.i.d.) or saline (1.0 M IP daily) for 5 d. Tumor samples harvested from P-AscH⁻-treated mice demonstrated increases in p-ERK protein (Figure 1G) and increases in the p-ERK/ERK ratio compared to tumor samples from mice treated with saline (Figure 1H).

Activated ERK is known to activate Drp1 via phosphorylation of Ser 616. Once activated, Drp1 is recruited to the mitochondria and found in the mitochondrial outer membrane (32). ERK phosphorylation of Drp1 promotes mitochondrial fission (23). Since P-AscH⁻ increased p-Drp1 levels 48 h after treatment in PDAC cells, we sought to test whether P-AscH⁻ would induce mitochondrial fission 48 h after treatment. Mito Tracker staining was used to measure the number of compartments and total volume of mitochondria 48 h after PDAC cells were treated with P-AscH⁻ (10 pmole/cell, 1 mM) P-AscH⁻ decreased total volume and increased the number of compartments consistent with mitochondrial fission (Figure 2A, B). These studies were repeated in the PDX339 PDAC cell line demonstrating a 1.4-fold increase in the number of compartments with P-AscH⁻ (data not shown). To support that these changes in mitochondrial morphology were consistent with fission, electron microscopy (Figure 2C) demonstrated decreased length of mitochondria following P-AscH⁻ compared to controls. Quantification demonstrated significant decreases in mitochondrial length in P-AscH⁻ treated cells (Figure 2D), also consistent with mitochondrial fission. Finally, to determine if these increases in mitochondrial fragmentation correlated with the activation of Drp1, we measured colocalization of p-Drp1 to the mitochondria in tumor xenografts from mice treated with P-AscH⁻ or saline. Tumor xenografts were stained for p-Drp1 and mitochondrial cytochrome c oxidase (Figure 2E). In tumor xenografts treated with P-AscH⁻, p-Drp1 colocalization with cytochrome c oxidase was increased compared to saline treated animals (Figure 2F), signifying increased Drp1 colocalization to the mitochondria. These data suggest that P-AscH⁻-induced increases in mitochondrial fission that correlates with activation of Drp1.

To determine the role of P-AscH⁻ in the activation of ERK and Drp1, we generated cell lines using siRNA to inhibit both ERK and Drp1. A significant decrease in p-ERK protein in MIA PaCa-2 cells generated using the siRNA transfection was observed (Figure 3A). Because there is a regulatory role for ERK-signaling communicating to mitochondrial fission-fusion process and Drp1 activation (16,17), the siERK knockout also demonstrated decreases in the activation of Drp1 (Figure 3A).

Parental and siRNA cell lines were then treated with P-AscH⁻ (7 pmole/cell) for 1 h to determine clonogenic survival. Genetic inhibition of ERK significantly increased the effects on P-AscH⁻-induced reduction in clonogenic survival (Figure 3B) compared to siNEG and controls. We then generated cells with inhibition of Drp1 using siRNA technology; western blotting verified decreased levels of activated Drp1 (Figure 3C). Clonogenic survival demonstrated that cells with genetic inhibition of Drp1 were significantly more sensitive to P-AscH⁻ compared to their siNEG and controls (Figure 3D). These data suggest that the activation of ERK/Drp1 may be a survival response when exposed to P-AscH⁻, and inhibition of this pathway may render cancer cells more sensitive to P-AscH⁻.

No change in p-ERK or p-Drp1 following P-AscH⁻ in ρ⁰ cells lacking oxidative phosphorylation--ρ⁰ cells derived from the MIA PaCa-2 cell line were utilized to evaluate the effects that P-AscH⁻ might have on p-ERK and p-Drp1 expression in the absence of mitochondria (28). Two different ρ⁰ clones, clone 3 and clone 4, were treated with 10 pmole/cell (1 mM) P-AscH⁻ for 1 h and then allowed to proliferate for 48 h prior to protein isolation. Unlike MIA PaCa-2 and PANC-1 cells that demonstrate dose- and time-dependent increases in p-ERK and p-Drp1 following treatment with P-AscH⁻, the ρ⁰ cells did not demonstrate any change in p-ERK or p-Drp1 protein expression compared to control (Figure 4). These results suggest that P-AscH⁻ acts at the mitochondrial level to induce cell stress and promotes mitochondrial fission in cancer cells. Previous studies have demonstrated that these ρ⁰ cell lines are resistant to P-AscH⁻, further suggesting that the p-ERK/p-DRP1 pathway is dependent upon functional mitochondria and thus may be partially responsible for cancer cell resistance to P-AscH⁻ treatment (28).

Morris et al. have described a novel selective ERK1/2 inhibitor, SCH772984, that demonstrated tumor regression in xenograft models of BRAF, NRAS, and KRAS mutant tumors (33). This inhibitor also effectively inhibited MAPK-signaling in tumor cells that had developed resistance to inhibitors of BRAF or MEK (33). Pancreatic cancer and H6c7 cells were treated with P-AscH⁻ (1 mM for 1 h), SCH772984 (1 μM for 24 h), or a combination of P-AscH⁻ and SCH772984. SCH772984 prevented the P-AscH⁻-induced increased phosphorylation of ERK (Figure 5A), but also prevented the increased phosphorylation of downstream Drp1 (Figure 5B). Cancer cells treated with P-AscH⁻ alone or SCH772984 alone had decreased clonogenic survival compared to controls in both cancer cell lines MIA PaCa-2 (Figure 5C) and PDX339 (Figure 5D); clonogenic survival was further reduced when the two compounds were used in a combined treatment. The non-tumorigenic pancreatic ductal epithelial cell line H6c7 cells did not demonstrate a decrease in clonogenic survival with treatment by either P-AscH⁻ alone, SCH772984 alone, or with the P-AscH⁻ and SCH772984 combination treatment (Figure 5E). These data combined with the results of the genetic manipulation by siERK indicate that inhibition of the ERK pathway sensitizes cancer cells to P-AscH⁻.

The mitochondrial division inhibitor-1 (Mdivi-1) has been shown to inhibit mitochondrial fission, increase mitochondrial fusion, and thereby increase mitochondrial length (34). Mdivi (25 and 50 μM) reversed the P-AscH⁻-induced phosphorylation of Drp1 in MIA PaCa-2 cells (Figure 6A). PDAC and H6c7 cells were then treated with P-AscH⁻ (1 mM for 1 h), Mdivi-1 (50 μM for 24 h), or P-AscH⁻ + Mdivi-1 to determine clonogenic survival. Like the findings with the genetic inhibition of Drp1, combining P-AscH⁻ with Mdivi-1 resulted in significant decreases in clonogenic survival compared to either treatment alone (Figure 6B, C). As seen in the previous experiments, H6c7 cells were not sensitive to any of the treatments (Figure 6D).

Mdivi-1 has previously been shown to inhibit the growth of pancreatic cancer in vivo (35). To test whether the combination of P-AscH⁻ and Mdivi-1 would increase survival, MIA PaCA-2 xenografts were generated. When tumors were 5 mm in diameter, mice were randomized to receive daily treatment of either saline, P-AscH⁻ (4 g/kg i.p.), Mdivi-1 (50 mg/kg i.p.), or P-AscH⁻ + Mdivi-1 for 21 d. Compared to controls, mice in the combination treatment group had the longest median survival (54 vs. 33 days, *p < 0.05) (Figure 7A). The treatments were well tolerated with no signs of toxicities and no significant differences in body weight between groups (Figure 7B). Also, Mdivi-1 + P-AscH⁻ decreased phosphorylated Drp1 in tumor xenografts as measured by immunofluorescence (Figure 7C) and quantified by immunofluorescence (Figure 7D) (*p < 0.05 vs. control, means ± SEM). These survival data are consistent with our in vitro models of genetic and pharmacological Drp1 inhibition, suggesting that pharmacological inhibition of Drp1 in vivo sensitizes cancer cells to P-AscH⁻.