Hypoxia potentiates gemcitabine-induced stemness in pancreatic cancer cells through AKT/Notch1 signaling

The human pancreatic cancer cell line PANC-1 was obtained from the American Type Culture Collection (Manassas, VA, USA). The Patu8988 cell line was purchased from Nanjing KeyGen Biotech. Co. Ltd. (Nanjing, China). Both cell lines were cultured in Roswell Park Memorial Institute 1640 medium supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 μg/mL streptomycin, in a humidified incubator with 5% CO₂ at 37 °C. After reaching a 60–80% confluence level, the cells were treated with different concentrations of gemcitabine (Selleck, Houston, TX, USA) for 24 h. To examine the role of the Notch1 or AKT signaling pathway in enhancing stemness, the pancreatic cancer cells were pretreated with 10 μM DAPT (γ-secretase inhibitor; Selleck) for 24 h or 20 μM LY294002 (AKT inhibitor; Beyotime Biotechnology, Shanghai, China) for 2 h before gemcitabine treatment. To clarify the effect of hypoxia on pancreatic cancer cell stemness, the cells were treated with 1% O₂ for different time intervals or with various doses of CoCl₂ (Sigma-Aldrich, St. Louis, MO, USA) for 24 h. To test the synergistic effect of hypoxia and gemcitabine, the cells were co-treated with optimal doses of gemcitabine and CoCl₂ (as indicated in the pertinent figure legends) for 24 h.

Western blot analysis was performed as previously described [13]. In brief, total cell lysates were electrophoresed in a sodium dodecyl sulfate–polyacrylamide gel electrophoresis gel and transferred onto polyvinylidene difluoride membranes (Millipore, Burlington, MA, USA). The membranes were blocked with 5% skim milk and incubated overnight with primary antibodies. After washing, the membranes were incubated with secondary antibodies conjugated with horseradish peroxidase, and the proteins were visualized by adding an enhanced chemiluminescence substrate (Thermo Fisher Scientific, Waltham, MA, USA). Antibodies against Bmi1, Notch1, NICD1, AKT, p-AKT (phosphorylated AKT), and GAPDH (glyceraldehyde 3-phosphate dehydrogenase) were purchased from Cell Signaling Technology (Danvers, MA, USA), and those against Sox2 and HIF-1α (hypoxia-inducible factor-1α) were purchased from Abcam (Boston, MA, USA).

Migration and invasion assays were performed in 24-well Transwell chambers (Corning, Fisher Scientific). For the transwell invasion assay, the upper compartment of the chamber was precoated with Matrigel (Sigma-Aldrich). Equal amounts of approximately 10 × 10⁴ cells were seeded into each upper chamber. The upper and lower chambers were filled with culture medium containing 0.1 and 30% FBS, respectively. After about 24 h, the migratory and invasive cells on the lower surface of the membrane were fixed, stained with 0.1% crystal violet, and then counted in five random fields under a light microscope.

The MTT assay was performed as previously described [30]. After different treatments, the pancreatic cancer cells were seeded into 96-well plates and further incubated with various concentrations of gemcitabine (Selleck) for 48 h. Then, 20 μL of MTT solution (5 mg/mL; Sigma-Aldrich) was added to each well. The plates were incubated for 4 h, after which the medium was replaced with 150 μL of dimethyl sulfoxide (Sigma-Aldrich). The optical density was detected at 490 nm. Each concentration of gemcitabine was set up in five replicate wells.

Flow cytometry analysis was performed as previously described [13]. Anti-CD24–FITC antibody was purchased from BD Pharmingen (San Diego, CA, USA).

The sphere-forming ability assay was performed in stem cell medium (SCM) as previously described [13]. Briefly, after different treatments, the pancreatic cancer cells were washed three times and suspended in SCM, which consisted of Dulbecco’s modified Eagle’s medium/F12 medium supplemented with bovine serum albumin (0.4%; Sigma-Aldrich), Insulin-Transferrin-Selenium (ITS; 1×; Sigma-Aldrich), basic fibroblast growth factor (10 ng/mL; PeproTech, Rocky Hill, NJ, USA), and epidermal growth factor (20 ng/mL; PeproTech). Approximately 1 × 10⁴ cells per well were seeded into ultralow-attachment 6-well plates (Corning), and the medium was changed every 3 days. After 15 to 20 days, the formed spheres (diameter ≥ 50 μm) were counted under a light microscope. The efficiency of sphere formation was calculated on the basis of the ratio of number of spheres to total number of cells.

Xenografts were formed by subcutaneously injecting PANC-1 cancer cells into the right flank of 3- to 4-week-old athymic mice (2 × 10⁶ cells per 100 μL per mouse) (HFK Bioscience Co., Beijing, China). Approximately 6 days after subcutaneous implantation, the mice were randomly separated into the control, GEM (gemcitabine), GEM+DAPT, and DAPT groups (n = 5 per group). Gemcitabine (20 mg/kg) and DAPT (10 mg/kg) were intraperitoneally injected every 3 days and every day, respectively. Tumor volume was measured periodically by using the following formula: Volume = 0.5 × length × width². The experimental protocol complied with the “Guide for the Care and Use of Animals in Wuhan University”.

PANC-1 cells were separated into four groups (control, GEM, GEM+DAPT, and GEM+LY294002) and treated as indicated above. After treatment, approximately 4 × 10⁶ cells suspended in 0.2 mL phosphate-buffered saline were injected into the lateral tail vein of 7- to 8-week-old nude mice (HFK Bioscience Co.; n = 5 per group). After about 4 weeks, the mice were euthanized, and the lungs were completely resected and photographed. For hematoxylin and eosin (H&E) staining, the lungs were fixed with 4% paraformaldehyde and cut into 5-μm sections. The specimens were then stained with H&E, and the number of metastases was detected microscopically. All mice were handled in accordance with the protocols approved by the “Guide for the Care and Use of Animals in Wuhan University”.

The data in our study were expressed as mean ± standard deviation. Student’s t-test was used to compare differences between two groups. Values were considered statistically significant at P < 0.05.

In our previous study, we had shown that low-dose gemcitabine treatment can enhance the stemness of pancreatic cancer cell lines SW1990 and BxPC-3 [13]. In the present study, we further analyzed whether gemcitabine has a similar effect on other pancreatic cancer cell lines such as PANC-1 and Patu8988. Our results revealed that low-dose gemcitabine treatment (1–5 μM) for 24 h, which has a minimal killing effect on pancreatic cancer cells (Fig. 1a), induced the expression of stemness-associated molecules Bmi1 and Sox2 as well as the CSC marker CD24 (Fig. 1b-e). In line with these changes, gemcitabine treatment also enhanced the sphere-forming ability of the evaluated cell lines, which exhibited a greater number of cell spheres and larger microsphere size after treatment (Fig. 1f-h). Although Notch1 signaling has been reported to play an important role in maintaining the stemness and self-renewal ability of CSCs [31], studies on the correlation between gemcitabine and Notch1 signaling are still lacking. Our results revealed that low-dose gemcitabine treatment promoted the expression of both Notch1 and NICD1 in a dose-dependent manner (Fig. 1b and c). Together, our results suggest that low-dose gemcitabine treatment activates Notch1 signaling and induces stemness in pancreatic cancer cells.

To further confirm the role of Notch1 in gemcitabine-enhanced stemness, we pretreated pancreatic cancer cells with 10 μM γ-secretase inhibitor DAPT for 24 h before gemcitabine treatment. The Western blot findings showed that pretreatment with DAPT abolished gemcitabine-induced NICD1 expression (Fig. 2a). Further, Notch1 inhibition dramatically impaired the upregulation of Bmi1, Sox2, and CD24 expression (Fig. 2a-c). In addition, we observed a relative decrease in the number and size of spheres after Notch1 inhibition (Fig. 2d-f). It has been established that the properties of CSCs are connected with enhanced migration and invasion [32]. In the present study, we detected such changes after Notch1 inhibition. Our results showed that gemcitabine treatment increased the migratory and invasive abilities of pancreatic cancer cells, whereas suppression of Notch1 significantly abolished these increases (Additional file 1: Figure S1a-d). Moreover, pretreatment with DAPT dramatically reversed the gemcitabine-induced chemoresistance (Additional file 1: Figure S1e). These results show that gemcitabine promotes pancreatic cancer cell stemness and associated migration, invasion, and chemoresistance partly through Notch1 activation.

Because Notch1 activation was revealed to play a role in gemcitabine-induced stemness and associated malignant traits, we next investigated the effect of supplementation with Notch1 inhibition on chemosensitivity in vivo. As shown in Fig. 3a and b, DAPT treatment significantly reduced the tumor growth rate and size relative to the control at 38 days post-treatment. When combined with gemcitabine chemotherapy, DAPT treatment also synergistically strengthened the killing effect of gemcitabine in pancreatic cancer cells. We further examined the changes in Bmi1 and Sox2 expression and CD24⁺ cell population at the end of treatment. As shown in Fig. 3c-e, gemcitabine chemotherapy increased the expression levels of Bmi1 and Sox2 as well as the proportion of CD24⁺ cells, while combination treatment with DAPT abolished these enrichments. CSCs have an inherent potential for metastasis [33, 34]. Our results, too, revealed an enhanced ability of the cells for lung metastasis after gemcitabine treatment, which was attenuated when combined with DAPT treatment (Additional file 2: Figure S2a-c). These results show that Notch1 inhibition synergistically potentiates the killing effect of gemcitabine and suppresses metastasis in vivo.

AKT is commonly activated in pancreatic cancer and participates in gemcitabine chemoresistance, and inhibition of AKT could enhance the killing effect of gemcitabine [35]. Our results revealed that gemcitabine treatment promoted the expression of p-AKT (serine 473) in PANC-1 and Patu8988 cell lines (Fig. 4a). To determine the role of AKT in gemcitabine-induced stemness, we pretreated the pancreatic cancer cells with 20 μM LY294002 (an AKT inhibitor) for 2 h before gemcitabine treatment. As indicated in Fig. 4a, AKT inhibition significantly suppressed gemcitabine-induced AKT activation. Subsequently, the expression of Bmi1, Sox2, and CD24 was significantly impaired (Fig. 4a and b). Further, LY294002 pretreatment attenuated the gemcitabine-induced sphere-forming ability of the pancreatic cancer cells (Fig. 4c-e). We further examined the role of AKT in Notch1 activation after gemcitabine treatment. Our results demonstrated that LY294002 attenuated gemcitabine-induced NICD1 expression in both cancer cell lines (Fig. 4a). Then, we analyzed the changes in the stemness-related metastatic, migratory, and invasive abilities of cancer cells after AKT inhibition. Our results showed that pretreatment with LY294002 markedly attenuated gemcitabine-enhanced metastasis in vivo (Additional file 2: Figure S2a-c). It also weakened the migratory and invasive abilities of pancreatic cancer cells (Additional file 3: Figure S3a-d). In total, our results suggest that AKT plays a role in promoting gemcitabine-induced Notch1 activation and stemness.

It is well known that hypoxia is a prominent feature of the microenvironment in pancreatic cancer and that it enhances the chemoresistance against gemcitabine [14]. We, therefore, analyzed the effect of hypoxia on stemness. As shown in Fig. 5a, treatment of pancreatic cancer cells under hypoxic (1%) conditions for 6–12 h significantly promoted the expression of HIF-1α, an important marker in hypoxia response. In addition, stemness-associated molecules Bmi1 and Sox2 were upregulated after hypoxia treatment. We also treated pancreatic cancer cells with CoCl₂, a chemical which stabilizes HIF-1α, to mimic a hypoxic microenvironment. CoCl₂treatment for 24 h, too, promoted Bmi1 and Sox2 expression in a dose-dependent manner in both evaluated cells (Fig. 5b). To further verify the synergistic effect of hypoxia and gemcitabine treatment on stemness induction, we co-treated pancreatic cancer cells with gemcitabine and CoCl₂ for 24 h. The Western blot findings showed that combination treatment with CoCl₂ further reinforced the gemcitabine-inductive effect on Bmi1 and Sox2 (Fig. 5c). In accord, CoCl₂ treatment also enhanced the number and size of gemcitabine-induced spheres (Fig. 5d-f). Our results suggest that hypoxia synergistically enhances gemcitabine-induced stemness.

Because Notch1 has been demonstrated to mediate gemcitabine-induced stemness, we next analyzed the changes on the basis of hypoxic status. The Western blot findings revealed that both hypoxia and CoCl₂ treatment increased the expression of NICD1 (Fig. 6a and b). The synergistically inductive effect of these two treatments was more obvious when combined with gemcitabine treatment (Fig. 6c). We also pretreated pancreatic cancer cells with DAPT before co-treatment with gemcitabine and CoCl₂. The results showed that DAPT significantly attenuated the synergistic enhancement of Bmi1 and Sox2 expression (Fig. 6d). In addition, Notch1 inhibition significantly suppressed the synergistic enhancement of sphere number and size (Fig. 6e-g). These results suggested that Notch1 activation plays a role in the synergistic enhancement of stemness induced by combination treatment with gemcitabine and CoCl₂.

Hypoxia has been reported to activate AKT expression [14]. Our results also demonstrated a similar effect in pancreatic cancer cells (Fig. 6a and b). In addition, synergistically augmented p-AKT expression was more evident after gemcitabine and CoCl₂ co-treatment (Fig. 6c), whereas AKT suppression significantly suppressed the collaboratively inductive effect of the co-treatment on Bmi1 and Sox2 expression (Fig. 6d). In line with these changes, the enhanced sphere-forming ability of the cells was also dramatically reduced after AKT inhibition (Fig. 6e-g). Moreover, the induction of NICD1 expression by gemcitabine and CoCl₂ co-treatment was significantly abolished by pretreatment with an AKT inhibitor in both evaluated cells (Fig. 6d), which suggests the role of AKT regulation in Notch1 activation in cells treated with gemcitabine and CoCl₂. Collectively, these results suggest that AKT/Notch1 signaling plays a role in promoting the synergistic induction of stemness by gemcitabine and hypoxia.