Human pancreatic cancer cells under nutrient deprivation are vulnerable to redox system inhibition

Abstract

Large regions in tumor tissues, particularly pancreatic cancer, are hypoxic and nutrient-deprived because of unregulated cell growth and insufficient vascular supply. Certain cancer cells, such as those inside a tumor, can tolerate these severe conditions and survive for prolonged periods. We hypothesized that small molecular agents, which can preferentially reduce cancer cell survival under nutrient-deprived conditions, could function as anticancer drugs. In this study, we constructed a high-throughput screening system to identify such small molecules and screened chemical libraries and microbial culture extracts. We were able to determine that some small molecular compounds, such as penicillic acid, papyracillic acid, and auranofin, exhibit preferential cytotoxicity to human pancreatic cancer cells under nutrient-deprived compared with nutrient-sufficient conditions. Further analysis revealed that these compounds target to redox systems such as GSH and thioredoxin and induce accumulation of reactive oxygen species in nutrient-deprived cancer cells, potentially contributing to apoptosis under nutrient-deprived conditions. Nutrient-deficient cancer cells are often deficient in GSH; thus, they are susceptible to redox system inhibitors. Targeting redox systems might be an attractive therapeutic strategy under nutrient-deprived conditions of the tumor microenvironment.

Keywords: auranofin; cancer therapy; chemical biology; drug discovery; drug screening; glutathione; metabolism; oxidation reduction (redox); oxidative stress; papyracillic acid; penicillic acid; redox regulation; thioredoxin.

Figure 1

PCA and PPA inhibit cell growth of human pancreatic cancer under nutrient-deprived conditions.A, schematic representation of the screening procedure for microbial metabolites that preferentially inhibit PANC-1 cell growth under nutrient-deprived conditions. B and C, isolation procedure and chemical structure of PCA (B) and PPA (C). D and E, preferential cytotoxicity of PCA (D) and PPA (E) in PANC-1 cells. PANC-1 cells were incubated with PCA (D) or PPA (E) for 24 h in NDM or DMEM. F, preferential cytotoxicity of PCA and PPA in human pancreatic cancer cells. G and H, colony formation of PANC-1 cells. PANC-1 cells were incubated for 10 days in DMEM after treatment with 6 μm PCA (G) or 4 μm PPA (H) for 24 h in NDM or DMEM. I, detection of apoptotic cells. PANC-1 cells were incubated with 30 μm PCA or 44 μm PPA for 24 h, and apoptotic cells stained with annexin V and PI were detected by flow cytometry. J and K, caspase 3/7 activity in PANC-1 cells. PANC-1 cells were incubated with 59 μm PCA (J) and 44 μm PPA (K) for 6 h. The data are presented as means ± S.D. of three independent experiments. p values were determined by two-tailed Student's t test. *, p < 0.05.

Figure 2

PCA and PPA bind to GSH and decrease cellular GSH.A, effects of PCA on intracellular concentration of metabolites involved in the primary metabolic pathway. Intracellular metabolites of PANC-1 cells treated with 59 μm PCA were measured by CE–TOF MS. B–D, intracellular GSH levels. PANC-1 cells were incubated with 59 μm PCA and 88 μm PPA, and intracellular GSH levels and the GSH/GSSG ratio were measured by a GSH/GSSG-Glo assay (B). PANC-1 cells were incubated with various concentrations of PCA and PPA for 2 h (C). Human pancreatic cancer cells were treated with 59 μm PCA or 88 μm PPA for 2 h (D). The data are presented as means ± S.D. of three independent experiments. p values were determined by two-tailed Student's t test. *, p < 0.05; **, p < 0.01; ***, p < 0.001. E, PCA directly binds to GSH. The PCA–GSH conjugate was detected by LC–MS. F, PPA directly binds to GSH. G, formation of the PCA–GSH conjugate in PANC-1 cells. Intracellular metabolites of PANC-1 cells treated with 59 μm PCA were measured by LC–MS. The data are presented as means ± S.D. of three independent experiments. p values were determined by two-tailed Student's t test. **, p < 0.01. H, schematic model to explain the PCA- and PPA-induced decreases in GSH.

Figure 3

PCA and PPA significantly increase ROS levels under nutrient-deprived conditions.A–C, intracellular ROS levels. PANC-1 cells were incubated with 118 μm PCA and 88 μm PPA in DMEM, and intracellular ROS levels were measured by a ROS-Glo H₂O₂ assay (A). PANC-1 cells were incubated with PCA and PPA for 12 h (B). PANC-1 cells were treated with 118 μm PCA or 88 μm PPA for 12 h (C). D, intracellular GSH levels under nutrient-deprived conditions. PANC-1 cells were incubated with 24 μm PCA and 18 μm PPA in NDM or DMEM. E, intracellular ROS levels under nutrient-deprived conditions. PANC-1 cells were incubated with 118 μm PCA and 88 μm PPA for 6 h in NDM or DMEM. F, effects of glucose and amino acids on preferential cytotoxicity of PCA and PPA under nutrient-deprived conditions. PANC-1 cells were incubated with PCA or PPA for 24 h in NDM to which glucose or amino acids were added. G, effects of cystine, glutamine, and glycine on preferential cytotoxicity of PCA and PPA under nutrient-deprived conditions. PANC-1 cells were incubated with 59 μm PCA and 44 μm PPA for 24 h in NDM with cystine, glutamine, and glycine. H, schematic of the mechanism underlying preferential cytotoxicity of PCA and PPA to nutrient-deprived cancer cells. For all graphs, the data are presented as means ± S.D. of three independent experiments. p values were determined by two-tailed Student's t test. *, p < 0.05; **, p < 0.01; ***, p < 0.001; n.s., not significant.

Figure 4

Auranofin preferentially inhibits cell growth of human pancreatic cancer under nutrient-deprived conditions.A, chemical structure of auranofin and TrxR inhibition. B, auranofin inhibits human recombinant TrxR1/2 activity. C, auranofin inhibits TrxR activity in PANC-1 cells. The data are presented as means ± S.D. of three independent experiments. p values were determined by two-tailed Student's t test. *, p < 0.05; **, p < 0.01. D, preferential cytotoxicity of auranofin in PANC-1 cells. PANC-1 cells were incubated with auranofin for 24 h in NDM or DMEM. E, preferential cytotoxicity of auranofin in human pancreatic cancer PANC-1 cells. F, redox state of Trx1 in PANC-1 cells under nutrient-deprived conditions. The redox state was examined by redox Western blotting. G, intracellular ROS levels. PANC-1 cells were treated with 1.5 μm auranofin for 12 h in the absence or presence of 5 mm NAC. H, NAC abrogated the inhibition of cell growth by auranofin. PANC-1 cells were incubated with auranofin for 24 h in the absence or presence of 5 mm NAC. I and J, caspase 3/7 activity in PANC-1 cells. PANC-1 cells were incubated with 1.5 μm auranofin for 12 h (I) or the indicated time (J). The data are presented as means ± S.D. of three independent experiments. p values were determined by two-tailed Student's t test. *, p < 0.05; **, p < 0.01; ***, p < 0.001; n.s., not significant. K, PARP activation. PANC-1 cells were treated with auranofin for 24 h. L, detection of apoptotic cells. PANC-1 cells were incubated with 0.75 μm auranofin for 12 h in the absence or presence of 5 mm NAC. M, antitumor activity of auranofin. Auranofin (12.5 mg/kg, intraperitoneally, n = 5) or cisplatin (12.5 mg/kg, intravenously, once a week, n = 5) was administered to mouse xenograft models of human PSN-1 cancer. p values were determined by two-tailed Student's t test. **, p < 0.01. N, antitumor activity of PCA. PCA (12.5 or 50 mg/mkg, n = 5) was administered intraperitoneally once a week to PSN-1 tumor-bearing mice. p values were determined by two-tailed Student's t test. *, p < 0.05; n.s., not significant. O, schematic of the mechanism underlying preferential cytotoxicity of auranofin to nutrient-deprived cancer cells.