Targeting the Warburg effect with a novel glucose transporter inhibitor to overcome gemcitabine resistance in pancreatic cancer cells

Abstract

Gemcitabine resistance remains a significant clinical challenge. Here, we used a novel glucose transporter (Glut) inhibitor, CG-5, as a proof-of-concept compound to investigate the therapeutic utility of targeting the Warburg effect to overcome gemcitabine resistance in pancreatic cancer. The effects of gemcitabine and/or CG-5 on viability, survival, glucose uptake and DNA damage were evaluated in gemcitabine-sensitive and gemcitabine-resistant pancreatic cancer cell lines. Mechanistic studies were conducted to determine the molecular basis of gemcitabine resistance and the mechanism of CG-5-induced sensitization to gemcitabine. The effects of CG-5 on gemcitabine sensitivity were investigated in a xenograft tumor model of gemcitabine-resistant pancreatic cancer. In contrast to gemcitabine-sensitive pancreatic cancer cells, the resistant Panc-1 and Panc-1(GemR) cells responded to gemcitabine by increasing the expression of ribonucleotide reductase M2 catalytic subunit (RRM2) through E2F1-mediated transcriptional activation. Acting as a pan-Glut inhibitor, CG-5 abrogated this gemcitabine-induced upregulation of RRM2 through decreased E2F1 expression, thereby enhancing gemcitabine-induced DNA damage and inhibition of cell survival. This CG-5-induced inhibition of E2F1 expression was mediated by the induction of a previously unreported E2F1-targeted microRNA, miR-520f. The addition of oral CG-5 to gemcitabine therapy caused greater suppression of Panc-1(GemR) xenograft tumor growth in vivo than either drug alone. Glut inhibition may be an effective strategy to enhance gemcitabine activity for the treatment of pancreatic cancer.

Fig. 1.

Evidence that gemcitabine-resistant pancreatic cancer cells exhibit a strong glycolytic phenotype and that high RRM2 expression underlies gemcitabine resistance. (A) Differential sensitivities of various pancreatic cancer cell lines to gemcitabine by MTT (left) and colony formation (right) assays. Points, means; error bars, SD (n = 6). (B) Quantitative PCR analysis of the mRNA expression levels of various glycolytic enzymes associated with cancer cell aggressiveness and/or drug resistance, including hexokinase 2, glucose phosphate isomerase, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 4, aldolase, PKM2 and pyruvate dehydrogenase kinase 3, in AsPC-1, Panc-1 and Panc-1^GemR cells. Data are shown as means ± SD (n = 3). (C) Western blot analysis of the expression levels of key subunits of the OXPHOS enzyme system: NADH dehydrogenase (complex I), succinate dehydrogenase complex (complex II), UQCRC2 (complex III), cytochrome c oxidase II (complex IV) and ATP5A (complex V). Hsp60 was used as a mitochondrial marker. The immunostaining of OXPHOS enzyme subunits was performed by using a commercial total OXPHOS human western blot antibody cocktail. (D) Western blot analysis of the expression levels of RRM2, RRM1, TS, human equilibrative nucleoside transporter 1, deoxycytidine kinase and p-Akt in the absence (−) or presence (+) of 1 µM gemcitabine (24h exposure) in different pancreatic cancer cell lines. Gem^R, Panc-1^GemR cells. (E) RT–PCR and luciferase reporter assays of the effect of gemcitabine (24h exposure) on mRNA expression of RRM2 (left; 1 µM) and RRM2 gene promoter activity (right) in Panc-1 cells. Data are shown as means ±mSD (n = 3). (F) Effect of shRNA-mediated silencing (left) and ectopic overexpression (right) of RRM2 on gemcitabine sensitivity in Panc-1^GemR and AsPC-1 cell lines, respectively. Western blot analyses of RRM2 expression and MTT assays of cell viability are shown. Scr, scrambled shRNA; pCMV, empty vector. Points, means; bars, SD (n = 6).

Fig. 2.

Evidence that gemcitabine activates RRM2 expression through E2F1 upregulation in pancreatic cancer cells. (A) RT–PCR (left) and western blot (right) analyses of the expression levels of RRM2 vis-à-vis E2F1, Sp1 and NF-YA in Panc-1 cells in the absence (−) or presence (+) of 1 µM gemcitabine (24h exposure) versus those in Panc-1^GemR (Gem^R) cells. (B) Left, western blot analysis of the expression of E2F1 and RRM2 in normal pancreatic epithelial cells (NPC) versus different pancreatic cancer cell lines, including AsPC-1, BxPC-3 and Panc-1. Right, quantitative PCR analysis of the relative mRNA abundance of E2F1 and RRM2 in NPC versus the aforementioned cancer cell lines after normalization to β-actin. Data are shown as means ±mSD (n = 3). (C) Western blot analysis of the expression of E2F1, RRM2 and Sp1 in paired pancreatic tumor tissues and adjacent normal tissues from pancreatic cancer patients. (D) Western blot and luciferase reporter assays of the effect of shRNA-mediated knockdown of E2F1 on the expression of RRM2, CDK2 and cyclin A (left) and RRM2 gene promoter activity (right) in gemcitabine-treated Panc-1 cells (1 µM, 24h). Data are shown as means ±mSD (n = 3). (E) Evidence that TS is an E2F1-targeted gene. Western blot analysis of the effects of ectopic expression and shRNA-mediated knockdown of E2F1 on the expression of E2F1 and TS in AsPC-1 and Panc-1 cells, respectively.

Fig. 3.

Evidence that the glucose inhibitor CG-5 suppresses RRM2 expression through E2F1 downregulation. (A) Left, structure of CG-5 (upper) and evidence that CG-5 is a pan-Glut inhibitor exhibiting comparable potencies for inhibition of glucose uptake in Panc-1 cells overexpressing various Glut isoforms [lower; data are shown as means ±mSD (n = 3)]. Right, differential sensitivities of various pancreatic cancer cell lines and NPC to CG-5 by MTT assays. Points, means; error bars, SD (n = 6). (B) Western blot analysis of the expression of p-Akt, p-mTOR, p-AMPK, Skp2, cyclin D1, Sp1, RRM2 and E2F1 in CG-5-treated Panc-1 cells. (C) RT–PCR and luciferase reporter assays of the effect of CG-5 (24h exposure) on mRNA expression of RRM2 and E2F1 (upper) and RRM2 gene promoter activity (lower), respectively, in Panc-1 cells. Data are shown as means ±mSD (n = 3). (D) Western blot and luciferase reporter assays of the effect of ectopic overexpression of E2F1 on the CG-5-induced reductions in RRM2 expression (left) and RRM2 gene promoter activity (right), respectively, in Panc-1 cells. For the promoter-luciferase assays, cells were treated with 5 µM CG-5 for 24h. Data are shown as means ±mSD (n = 3). (E) Western blot analysis of the effect of ectopic overexpression of Sp1 on the CG-5-induced reduction in RRM2 expression in Panc-1 cells. (F) Western blot analyses of the effects of 2-DG and glucose (Glu) starvation on the abundance of E2F1 and RRM2 (upper) and the protective effect of high exogenous glucose levels (10 versus 2%) on CG-5-mediated downregulation of E2F1 and RRM2 expression (lower) in Panc-1 cells. Cells were treated for 24h unless otherwise noted. (G) Evidence that E2F1-mediated RRM2 upregulation underlies the acquisition of gemcitabine resistance in BxPC-3 and AsPC-1 cells. Left, MTT assays of the concentration-dependent effects of gemcitabine on the viability of BxPC-3^GemR (B^G) and AsPC-1^GemR (A^G) cells, vis-à-vis the parental BxPC-3 (B) and AsPC-1 (A) cells, after 24h of treatment. Points, means; bars, SD (n = 6). Right, western blot analysis of the expression levels of E2F1 and RRM2 in BxPC-3^GemR and AsPC-1^GemR cells relative to the respective parental cell lines. (H) Western blot analysis of the dose-dependent suppressive effect of CG-5 on the expression of E2F1 and RRM2 in AsPC-1^GemR cells after 24h of exposure.

Fig. 4.

Evidence that CG-5 suppresses E2F1 expression through miR-520f upregulation. (A) Left, the expression of miR-520f in Panc-1 cells treated with CG-5 or gemcitabine for 24h as determined by quantitative RT–PCR. Data are shown as means ±mSD (n = 3). Right, western blot analysis of E2F1 and RRM2 expression in Panc-1 cells expressing miR-520f mimic (upper) or anti-miR-520f in the presence or absence of CG-5 (lower; 5 µM, 24h). (B) Upper, the miR-520f seed region and its target sequences in the 3′UTR of E2F1 of three species. Lower, Sequence of miR-520f and its target region in the 3′UTR of E2F1. The duplex formed by the seven-nucleotide seed region of miR-520f and its complementary target sequence in E2F1 are identified in bold-face type. Unpaired bases are listed above and below the duplex. (C) Diagram of wild-type (wt)- and mutant (mut)-E2F1-3′UTR-containing luciferase reporter constructs (left; mutated target sequence is underscored). Luciferase assays of the effects of a miR-520f mimic on the activation of wt- or mut-E2F1-3′UTR (center) and the effects of anti-miR-520f in the presence or absence of 5 µM CG-5 for 24h on activation of wt-E2F1-3′UTR (right) in Panc-1 cells. Data are shown as means ±mSD (n = 3). (D) Effects of the expression of miR-520f mimic or anti-miR-520f on the proliferation of Panc-1 cells. Points, means; bars, SD (n = 6). (E) Relative ectopic expression levels of miR-520f in a stably transfected Panc-1 clone (clone 2) versus that in control cells (WT) as determined by quantitative RT–PCR (left). Data are shown as means ± mSD (n = 3). Colony formation (center) and western blot analyses (right) of the effects of stable miR-520f expression on cell survival and the expression of E2F1 and RRM2, respectively, in gemcitabine-treated cells. Points, means; bars, SD (n = 6).

Fig. 5.

Evidence that CG-5 increases the sensitivity of pancreatic cancer cells to gemcitabine. (A) Effects of CG-5 on gemcitabine sensitivity in Panc-1 and Panc-1^GemR cells. Colony formation assay results were used for analysis of synergistic interactions between gemcitabine and CG-5 (left). Points, means; bars, SD (n = 6). The effects of CG-5 on gemcitabine-mediated changes in the expression levels of E2F1, RRM2, p-Akt and γ-H2AX were assessed by western blotting (center). Evaluation of the effect of CG-5 on gemcitabine-induced expression of γ-H2AX in Panc-1 cells as determined by immunofluorescent staining (10×) (right). CI, combination index. (B) Western blot (left) and immunofluorescent staining (right; 10×) analyses of γ-H2AX expression in Panc-1 cells treated for 24h with CG-5. (C and D) Western blot analyses of the effects of (C) 2-DG and glucose starvation, and (D) the ribonucleotide reductase inhibitor Triapine on γ-H2AX expression in Panc-1 cells. (E) The protective effects of ectopic RRM2 expression on CG-5-induced γ-H2AX expression in Panc-1 cells. (F) Effects of CG-5 and gemcitabine, each alone and in combination, on the intracellular distribution of RRM2 in Panc-1 cells. Upper, immunofluorescent staining of RRM2 (green); nuclei stained with 4′,6-diamidino-2-phenylindole (blue); red lines indicate planes of cross-sectional analyses of fluorescence intensities, which are shown in two-dimensional histograms (blue, nuclei; green, RRM2). Magnification bar, 10 µm. Lower, western blot analysis of RRM2 expression in cytoplasmic and nuclear fractions of Panc-1 cells treated with CG-5 and gemcitabine, each alone and in combination. α-Tubulin and histone H3 were used as cytoplasmic and nuclear markers, respectively.

Fig. 6.

Antitumor effects of CG-5 and gemcitabine, each alone and in combination, on Panc-1^GemR xenograft tumors in athymic nude mice. (A) Effects of CG-5, gemcitabine and the combination on Panc-1^GemR tumor growth. Arrows indicate days on which gemcitabine or vehicle was administered. Data are expressed as means ± SE. (B) Western blot analysis of the expression of E2F1, RRM2, γ-H2AX and H2AX in Panc-1^GemR tumors from three mice in each treatment group. Left, immunoblot. Right, relative abundance of E2F1 and RRM2 in the tumors from drug-treated mice versus those from vehicle-treated mice after normalization to β-actin. (C) Schematic diagram depicting the mechanism by which CG-5 sensitizes pancreatic cancer cells to gemcitabine.