Activation of GLP-1 Receptor Enhances the Chemosensitivity of Pancreatic Cancer Cells

64 2

Journal of Molecular H Zhao et al. GLP-1R activation inhibits 64:2 103–113 Endocrinology pancreatic cancer RESEARCH Activation of GLP-1 receptor enhances the chemosensitivity of pancreatic cancer cells

He-jun Zhao1, Xia Jiang1, Li-juan Hu2, Lei Yang3, Lian-dong Deng4, Ya-ping Wang4 and Zhi-peng Ren5

1Department of Endocrinology, Tianjin First Central Hospital, Tianjin, China 2Laboratory of Acute Abdomen Disease, Tianjin NanKai Hospital, Tianjin, China 3Clinical Laboratory, Tianjin First Central Hospital, Tianjin, China 4Department of Polymer Science and Technology, School of Chemical Engineering and Technology, Key Laboratory of Systems Bioengineering (Ministry of Education), Tianjin University, Tianjin, China 5Department of Orthopaedics, Tianjin Hospital, Tianjin, China

Correspondence should be addressed to Z Ren: [email protected]

Abstract

This study aimed to determine whether and how the glucagon-like peptide 1 receptor Key Words (GLP-1R) agonist liraglutide affects the chemoresistance and chemosensitivity of ff glucagon-like peptide-1 pancreatic cancer cells to gemcitabine in vitro and in vivo. The GLP-1R and protein receptor kinase A (PKA) levels were compared between the human pancreatic cancer cell line ff pancreatic cancer PANC-1 and the gemcitabine-resistant cell line PANC-GR. The in vitro effects of liraglutide ff chemoresistance on the cell proliferation and apoptosis as well as the nuclear factor-kappa B NF-κB ff gemcitabine expression levels of PANC-GR cells were evaluated. In addition, a mouse xenograft ff apoptosis model of human pancreatic cancer was established by s.c. injection of PANC-1 cells, and the effects of liraglutide on the chemosensitivity were evaluated in vitro and in vivo. In contrast to PANC-1 cells, PANC-GR cells exhibited lower expression levels of GLP-1R and PKA. Incubation with liraglutide dose dependently inhibited the growth, promoted the apoptosis, and increased the expression of GLP-1R and PKA of PANC-GR cells. Similar effects of liraglutide were observed in another human pancreatic cancer cell line MiaPaCa-2/MiaPaCa-2-GR. Either the GLP-1R antagonist Ex-9, the PKA inhibitor H89, or the NF-κB activator lipopolysaccharide (LPS) could abolish the antiproliferative and proapoptotic activities of liraglutide. Additionally, each of these agents could reverse the expression of NF-κB and ABCG2, which was decreased by liraglutide treatment. Furthermore, liraglutide treatment increased the chemosensitivity of pancreatic cancer cells to gemcitabine, as evidenced by in vitro and in vivo experiments. Thus, GLP-1R agonists are safe and beneficial for patients complicated with pancreatic cancer and Journal of Molecular diabetes, especially for gemcitabine-resistant pancreatic cancer. Endocrinology (2020) 64, 103–113

Introduction

Pancreatic cancer is closely related to diabetes mellitus. the prognosis of patients with pancreatic cancer and Diabetes mellitus is not only a risk factor for pancreatic that long-term hyperglycemia is a poor prognostic factor cancer, but it is also one of the clinical manifestations (Lee et al. 2016). Therefore, optimization of the selection of pancreatic cancer (Everhart & Wright 1995, Chari of hypoglycemic drugs in patients complicated with et al. 2008). It is believed that glycemic control affects pancreatic cancer and diabetes mellitus has important

-19-0186 Journal of Molecular H Zhao et al. GLP-1R activation inhibits 64:2 104 Endocrinology pancreatic cancer clinical significance. Glucagon-like peptide 1 (GLP-1) is ABCG2 expression in pancreatic cancer cells inhibited cell known as an incretin hormone secreted from intestinal migration and invasion, but no effects on cell proliferation endocrine cells, and it augments insulin production in a or apoptosis were observed (Wang et al. 2010). glucose-dependent manner through specific binding to The present study aimed to further elucidate the GLP-1 receptor (GLP-1R) (Meloni et al. 2013). In recent effects of the GLP-1R agonist liraglutide on pancreatic years, GLP-1 and GLP-1R agonists have been widely cancer from the perspective of chemoresistance and used in clinical practice as a new strategy for lowering chemosensitivity, so as to provide a new therapeutic glucose levels. For the first time, our previous work has strategy for pancreatic cancer patients with diabetes shown that a decrease of GLP-1R expression in human mellitus. We investigated whether and how activation pancreatic cancer tissues is associated with the growth and of GLP-1R by liraglutide affects the proliferation and metastasis of cancer as well as a poor prognosis of patients apoptosis of gemcitabine-resistant human pancreatic (Zhao et al. 2014a,b). The GLP-1 agonist liraglutide can cancer cells. Based on previous work by us and others, we activate GLP-1R to inhibit the growth and metastasis of hypothesized that there would be an association between human pancreatic cancer cells, via the inhibition of the GLP-1R activation and chemosensitivity in pancreatic phosphatidylinositol 3-kinase (PI3K)/Akt and nuclear cancer cells, probably by regulation of NF-κB/ABCG2. In factor-kappa B (NF-κB) signaling pathways. Thus, GLP-1R addition, we also explored the role of liraglutide on the agonists may be beneficial for the treatment of pancreatic chemosensitivity of pancreatic cancer cells to gemcitabine. cancer while controlling blood glucose for pancreatic cancer patients with diabetes mellitus. Pancreatic cancer is one of the deadliest common Materials and methods cancers in the world. Approximately 90% of all pancreatic Cell culture malignancies are pancreatic ductal adenocarcinomas, a type of cancer with a mortality rate almost identical to The human pancreatic cancer cell lines PANC-1 and its incidence rate (Siegel et al. 2015). Surgical resection MiaPaCa-2 were purchased from the American Type followed by adjuvant chemotherapy remains the Culture Collection. The gemcitabine-resistant human common treatment choice for early-stage pancreatic pancreatic cancer cell line PANC-GR was established cancer. Gemcitabine, a nucleoside analogue, is used as by subjecting PANC-1 cells to gradually increasing, low a first-line treatment for pancreatic adenocarcinomas of concentrations of gemcitabine. Briefly, the median lethal all stages (Valenzuela et al. 2014). However, a growing dose of gemcitabine for PANC-1 cells was evaluated after number of patients exhibit the development of resistance a week of intervention with different concentrations to gemcitabine within weeks of treatment initiation of gemcitabine and was determined to be 0.08 μg/mL. (Amrutkar & Gladhaug 2017). Although several genetic Initially, the PANC-1 cells were cultured in 0.1 μg/mL and/or epigenetic alterations have been found to contribute gemcitabine for 48 h with a defined drug-free interval. to the chemoresistance of pancreatic adenocarcinomas When the cells adapted to the dose, they were cultured in (Kim & Gallick 2008, Tamburrino et al. 2013), the potential medium containing 0.4 μg/mL gemcitabine. The culture mechanisms involved in gemcitabine resistance still lack was continued with a four-fold increasing concentration clear explanations. Thus, understanding the molecular of gemcitabine until it reached 400 μg/mL. Finally, cells mechanisms of drug resistance in pancreatic cancer as well stably resistant to a high concentration of gemcitabine as seeking high-efficiency, low-toxic new chemotherapy were obtained. Cells were cultured in Dulbecco’s modified sensitizers are critical to develop new effective treatments Eagle’s medium (Life Technologies) supplemented for this deadly disease. with 10% fetal bovine serum at 37°C and 5% CO2 in a ATP-binding cassette subfamily G member 2 (ABCG2), humidified incubator. The MiaPaCa-2-GR cell line was commonly referred to as breast cancer resistance protein established with similar methods as described earlier, with (BCRP), is a 655-amino-acid polypeptide transporter with the median lethal dose of gemcitabine for MiaPaCa-2 cells a wide range of substrates that displays resistance to a determined to be 0.16 μg/mL. variety of drugs, including gemcitabine (Mo & Zhang 2012, Natarajan et al. 2012). Upregulation of BCRP/ABCG2 has Cell proliferation assay been reported to play a role in chemoresistance via the PI3K/Akt and NF-κB signaling pathways in breast cancer Cells were seeded in 96-well plates at a density of 3 × 103/ (Zhang et al. 2011). In another study, modulation of well. After serum starvation for 12 h, the cells were

Cyclic adenosine monophosphate Flow cytometric analysis (cAMP) measurement Apoptosis was determined by annexin V/propidium iodide After counting the cells, the cells were lysed and (PI) staining, as described by the manufacturer (Roche then repeatedly frozen and thawed four times. The Diagnostics). The harvested cells (1.5 × 105) were stained supernatant was collected after centrifugation. The level with 5 μL of Annexin V-FITC and 5 μL of PI solution for of cAMP in the supernatant was measured by an ELISA 15 min in the dark and then analyzed by flow cytometry with a commercially available kit (R&D Systems). The (BD Biosciences, San Jose, CA, USA). On the flow- optical density, which was proportional to the cAMP cytometric figures, the proportion of cells in each quadrant concentration, was determined by using a microplate was calculated as reported previously (Fan et al. 2013).

Table 1 Primer sequences used for RT-PCR.

Primer sequences (5ʹ‒3ʹ) Genes Forward Reverse Annealing temperature (°C) Product length (bp) GLP-1R TCAAGGTCAACGGCTTATTAG TAACGTGTCCCTAGATGAACC 60 480 PKA TCACAGAGTTCATCAGAGAC ACCTGCCTTGTATCCATTG 54 163 NF-κB TCCACCTTCATTCTCAACTT CACCACATCTTCCTGCTTA 54 140 ABCG2 ATACATCAGCGGATACTACAG GAATAAGCCACCATCATAAGG 54 183 GAPDH ACAGTCAGCCGCATCTTCTT CTGGAAGATGGTGATGGGAT 64 87

Animals and tumor implantation experiments The T-test and one-way ANOVA were used for testing statistical differences among groups, respectively. A P All animal experiments were approved by the Peking value <0.05 was considered statistically significant. University Animal Care and Use Committee. Five-week- old male athymic nude mice (Vital River Animal Center, Beijing, China) were used in this study. After adaption, Results the mice were subcutaneously injected with PANC-1 cells Expression levels of GLP-1R and PKA in PANC-GR cells (1 × 106 in 100 μL of PBS) in the right flank. Starting at 3 and PANC-1 cells days after cell injection, the mice were intraperitoneally injected with 100 μL of PBS (vehicle), 0.2 mg/kg liraglutide, It is generally believed that GLP-1R activation is involved 80 mg/kg gemcitabine, or liraglutide + gemcitabine in adenylyl cyclase-mediated cAMP generation, which, in 100 µL of PBS, twice daily for 4 weeks (n = 5 in each in turn, leads to cAMP-dependent activation of PKA in group). The weights of the mice were recorded. When pancreatic islet β-cells, and finally augments insulin release the experiment was terminated, the mice were killed, in a glucose-dependent manner (Kolic & MacDonald the tumors were excised, and the weights and volumes 2015). First, we investigated whether the expression of of the tumors of each mouse were recorded. The tumor GLP-1R and PKA in PANC-GR cells was different from that volume was calculated using the following equation: in PANC-1 cells. In contrast to PANC-1 cells, PANC-GR V = (a × b2) × 0.5236, where ‘a’ is the larger dimension and cells exhibited lower expression levels of GLP-1R, at both ‘b’ is the perpendicular diameter (Wolf et al. 2006). The the mRNA and protein levels (both P < 0.05; Fig. 1A and tumor sections were paraffin-embedded and preserved for B). Also, the PANC-GR cells had a lower level of cAMP immunohistochemical analysis. than the PANC-1 cells (P < 0.05; Fig. 1C). Similarly, the expression levels of PKA, at both the mRNA and protein levels, were significantly lower in PANC-GR cells than in Immunohistochemical analysis PANC-1 cells (both P < 0.05; Fig. 1D and E). The tumor tissue was cut into a 5 × 5 × 1 mm3 volume, We also tested the expression of GLP-1R and PKA in which was embedded in paraffin and cut into a thickness MiaPaCa-2 versus MiaPaCa-2-GR cells. Similarly, MiaPaCa- of 4–5 µm. The slides were blocked with nonimmune goat 2-GR cells exhibited a significantly lower expression of serum for 1 h and then incubated with primary antibody GLP-1R as well as PKA as compared with MiaPaCa-2 cells against rabbit anti-human Ki-67 (dilution, 1:200; (P < 0.05; Supplementary Fig. 1A and B, see section on Bioworld Technology, Dublin, OH, USA) or Bax (dilution, supplementary materials given at the end of this article). 1:200; Cell Signaling Technology) at 4°C overnight. After washing, the slides were incubated with biotinylated Liraglutide decreases the growth and enhances the goat anti-rabbit IgG antibody as the secondary antibody apoptosis of PANC-GR cells (dilution, 1:100; Zhongshan Goldbridge Biotechnology, China). Positive immunostaining, which was visualized as We evaluated the effects of liraglutide on GLP-1R and PKA yellow or brown granules in the cytoplasm and/or nuclei, expression as well as cAMP generation in PANC-1 cells. was evaluated under a light microscope. Incubation with liraglutide (0 nM, 10 nM, 100 nM, or 1000 nM) for 48 h resulted in enhanced expression of GLP-1R and PKA and elevated levels of cAMP in PANC-GR cells in Statistical analysis a dose-dependent manner (all P < 0.05; Fig. 2A, B and C). A All statistical analyses were performed using SPSS11.5 similar effect of liraglutide was observed in MiaPaCa-2-GR software (SPSS Inc.). Data are presented as the mean ± s.d. cells as well (Supplementary Fig. 1C and D).

Apoptosis was determined with annexin V/PI staining using flow cytometry. Compared with the control exposed to 0 nM liraglutide, incubation with liraglutide for 48 h increased the apoptosis of cultured PANC-GR cells (P < 0.05; Fig. 2F), and this effect was shown to occur in a dose-dependent manner. Consistently, the expression levels of apoptosis-associated proteins, including Bax and caspase-3, were significantly increased in cultured PANC-GR cells that were pretreated with liraglutide for 48 h, especially at a high concentration (P < 0.05; Fig. 2G). We also found that pretreatment with liraglutide led to increased expression levels of Bax and caspase-3 in MiaPaCa-2-GR cells (Supplementary Fig. 1F and G).

Liraglutide inhibits the growth and enhances the apoptosis of PANC-GR cells via GLP-1R/PKA-mediated downregulation of NF-κB and ABCG2 We evaluated the difference in the expression levels of the transcriptional regulation factor NF-κB and the multidrug resistance gene product ABCG2 in PANC-GR cells versus PANC-1 cells. PANC-GR cells had significantly higher expression of NF-κB and ABCG2 as compared with PANC-1 cells, at both the mRNA and protein levels (Supplementary Fig. 2). Next, we investigated the effects of liraglutide on NF-κB and ABCG2 expression. Incubation with liraglutide led to decreased expression of NF-κB and ABCG2, at both Figure 1 the mRNA and protein levels, in PANC-GR cells (P < 0.05; Expression levels of GLP-1R and PKA in PANC-GR cells and PANC-1 cells. Fig. 3A and B). In addition, these effects were shown to PANC-GR cells exhibited lower expression levels of GLP-1R, cAMP, and occur in a dose-dependent manner when the liraglutide PKA. (A and B) The expression levels of GLP-1R. Top, the expression of GLP-1R as detected by RT-PCR (A) and Western blot (B); Bottom, concentration was 0 nM, 10 nM, 100 nM, or 1000 nM. GAPDH-corrected band intensities. (C) The cAMP expression level Interestingly, treatment with the GLP-1R antagonist Ex-9 (pmol/1 × 105 cells). (D and E) The expression levels of PKA. Top, the could block the effect of liraglutide on PKA, NF- B, and expression of PKA as detected by RT-PCR (D) and Western blot (E); κ Bottom, GAPDH-corrected band intensities. *P < 0.05 (treatment vs ABCG2 expression, as detected by RT-PCR and Western control). blot (treatment with liraglutide + Ex-9 vs liraglutide alone, P < 0.05; Fig. 3C and D). However, treatment with Ex9 The effects of liraglutide on cell proliferation and alone did not affect the expression of PKA, NF-κB, or apoptosis of PANC-GR cells were also determined. The ABCG2 (treatment with Ex-9 alone vs control, P > 0.05). PANC-GR cells were cultured with liraglutide at 0 nM, Moreover, cell culture with liraglutide could significantly 10 nM, 100 nM, or 1000 nM for 48 h. Liraglutide decrease the cell growth, although Ex-9 could reverse significantly inhibited the growth of PANC-GR cells in a the inhibitory effect of liraglutide on the growth of concentration-dependent manner (all liraglutide groups PANC-GR cells (Fig. 3E). By annexin V/PI staining using vs control, P < 0.05; Fig. 2D). Also, liraglutide inhibited the flow cytometry, we found that liraglutide significantly colony formation of PANC-GR cells in a dose-dependent promoted the apoptosis of PANC-GR cells (P < 0.05 vs manner, with 1000 nM liraglutide inhibiting colony control), whereas Ex-9 could reduce liraglutide-induced formation by an average of 70% as compared with the apoptosis (treatment with liraglutide + Ex-9 vs liraglutide control exposed to 0 nM liraglutide (Fig. 2E). Similarly, alone, P < 0.05; Fig. 3F). liraglutide inhibited the growth of MiaPaCa-2-GR cells in a Similarly, treatment with either the PKA inhibitor concentration-dependent manner (Supplementary Fig. 1E). H89 or the NF-κB activator LPS could effectively reverse

Figure 2 The effects of liraglutide on the cell growth and apoptosis of PANC-GR cells. PANC-GR cells were treated with Lira at 0 nM (control), 10 nM, 100 nM, or 1000 nM for specified durations. (A) The expression of GLP-1R as detected by western blot (top), and band intensities relative to GAPDH (bottom). (B) The cAMP expression level (pmol/1 × 105 cells). (C) The expression of PKA as detected by Western blot (top), and band intensities relative to GAPDH (bottom). (D) Cell proliferation was evaluated at 24 h and 48 h by using the Cell Counting Kit-8 assay. #P < 0.05 (10 nM vs control); $P < 0.05 (100 nM vs control); &P < 0.05 (1000 nM vs control). (E) The number of colonies formed was compared among groups exposed to different concentrations of Lira at 14 weeks. (F) Flow cytometric quantification of apoptosis with annexin V and PI staining (top), and the percentage of apoptotic cells (bottom). (G) The expression of Bax and caspase-3 as detected by Western blot (top), and their band intensities relative to GAPDH (bottom). *P < 0.05 (treatment vs control). Lira, liraglutide. A full colour version of this figure is available athttps:// doi.org/10.1530/JME-19-0186. the decreased expression of NF-κB and ABCG2 caused by gemcitabine. PANC-1 cells were treated with liraglutide, liraglutide treatment (treatment with liraglutide + H89 or gemcitabine, or their combination for 48 h. We observed a LPS vs liraglutide alone, P < 0.05; Fig. 4A and B), almost further decrease of cell growth when the cells were exposed to a level comparable to the control (P > 0.05 vs control). to gemcitabine + liraglutide as compared with gemcitabine Furthermore, treatment with either H89 or LPS could alone (P < 0.05, Fig. 5A). Additionally, liraglutide promoted partially restore the in vitro growth of PANC-GR cells, the proapoptotic effect of gemcitabine, as evidenced which was inhibited by liraglutide (Fig. 4C). In addition, by an increased proportion of apoptotic cells as well as the proapoptotic effect of liraglutide was reduced by either enhanced expression of Bax and caspase-3 in PANC-1 cells H89 or LPS treatment (treatment with liraglutide + H89 or (liraglutide + gemcitabine vs gemcitabine alone, P < 0.05; LPS vs liraglutide alone (P < 0.05; Fig. 4D). Fig. 5B and C). We further verified the synergistic effect of liraglutide in nude mice. Liraglutide treatment enhanced the Liraglutide enhanced the chemosensitivity of chemosensitivity of pancreatic cancer cells to gemcitabine, PANC-1 cells to gemcitabine as gemcitabine + liraglutide further reduced the tumor We investigated whether liraglutide had synergistic volume and tumor weight (Fig. 5D, E, F and G), as antiproliferative and proapoptotic activities with compared with gemcitabine alone. Also, the combination

Figure 3 Liraglutide inhibits the growth and enhances the apoptosis of PANC-GR cells via GLP-1R-mediated downregulation of NF-κB and ABCG2. (A and B) PANC-GR cells were incubated with Lira at 0 nM, 10 nM, 100 nM, or 1000 nM for 48 h. The expression of NF-κB and ABCG2 as detected by RT-PCR (A) and Western blot (B), and their band intensities relative to GAPDH (right panel). (C, D, E and F) PANC-GR cells were treated with Lira (100 nM), the GLP-1R antagonist Ex-9 (500 nM), or Lira (100 nM) + Ex-9 (500 nM). The expression of PKA, NF-κB, and ABCG2 as detected by RT-PCR (C) and Western blot (D), and their band intensities relative to GAPDH (right panel). (E) Cell proliferation was evaluated at 24 h and 48 h by using the CCK-8 assay. ※P < 0.05 (Lira vs control); &P < 0.05 (Lira + Ex-9 vs control); #P < 0.05 (Lira vs Lira + Ex-9). (F) Flow cytometric quantification of apoptosis with annexin V and PI staining (left), and the percentage of apoptotic cells (right). *P < 0.05 (vs control). Lira, liraglutide; CCK-8, Cell Counting Kit. A full colour version of this figure is available at https://doi.org/10.1530/JME-19-0186. of gemcitabine and liraglutide led to decreased Ki-67 a GLP-1R agonist, inhibited the growth and increased but increased Bax immunoreactivity in tumor tissues by the apoptosis of murine CT26 colon cancer cells via immunohistochemical analysis (Fig. 5H). upregulating intracellular cAMP levels while inhibiting ERK1/2 in a PKA-dependent manner (Koehler et al. 2011). In another study, GLP-1 was found to be a potent inducer Discussion of cAMP, thus inhibiting breast cancer cell proliferation Patients with pancreatic cancer have a high prevalence of (Ligumsky et al. 2012). However, Koehler et al. found that diabetes mellitus (68%), and glycemic control affects the GLP-1R activation by exendin-4 did not affect the growth prognosis of patients with pancreatic cancer (Aggarwal and apoptosis of GLP-1R-expressing pancreatic cancer et al. 2013). Therefore, it is of great clinical significance to cells, although it could promote the production of cAMP optimize the selection of hypoglycemic drugs in patients and inhibit the phosphorylation of ERK1/2 (Koehler & with pancreatic cancer and diabetes mellitus. As a new Drucker 2006). We previously found that a lower GLP-1R hypoglycemic agent, GLP-1 has its own advantages, expression in human pancreatic cancer tissues was including a low risk of hypoglycemia and weight gain associated with the growth, metastasis, and poor prognosis as well as cardiovascular and cerebrovascular protection of patients (Zhao et al. 2014a). Also, activation of GLP-1R (Vilsboll et al. 2007). In addition, studies have shown that by liraglutide inhibits growth and promotes apoptosis of GLP-1 or a GLP-1R agonist can inhibit a variety of tumor human pancreatic cancer cells in vitro, accompanied by cells (Koehler et al. 2011, Ligumsky et al. 2012, Zhao et al. cAMP production and resultant inhibition of the Akt and 2014b). For example, activation of GLP-1R by exendin-4, ERK1/2 signaling pathway (Zhao et al. 2014b).

Figure 4 Liraglutide inhibits the growth and enhances the apoptosis of PANC-GR cells via PKA/NF-κB- mediated downregulation of ABCG2. PANC-GR cells were treated with Lira (100 nM), Lira (100 nM) + the PKA inhibitor H89 (10 μM), or Lira (100 nM) + the NF-κB activator LPS (100 ng/mL). The expression of NF-κB and ABCG2 as detected by RT-PCR (A) and Western blot (B), and their band intensities relative to GAPDH (right). (C) Cell proliferation was evaluated by using the CCK-8 assay. ※P < 0.05 (Lira vs control); &P < 0.05 (Lira + H89 vs Lira); #P < 0.05 (Lira + LPS vs Lira). (D) Flow cytometric quantification of apoptosis with annexin V and PI staining (left), and the percentage of apoptotic cells (right). *P < 0.05 (vs control); §P < 0.05 (vs Lira). Lira, liraglutide; CCK-8, Cell Counting Kit. A full colour version of this figure is available athttps://doi.org/10.1530/ JME-19-0186.

In this study, for the first time, we investigated the malignant behaviors of pancreatic tissues. In the present expression and function of GLP-1R in gemcitabine- study, we found a further decrease of GLP-1R expression resistant human pancreatic cancer cells. Compared in gemcitabine-resistant human pancreatic cancer cells, to PANC-1 cells, we found a lower GLP-1R expression probably indicating the special role of GLP-1R in the onset in PANC-GR cells. Activation of GLP-1R by liraglutide and progression of pancreatic cancer. In this study, we inhibited the growth and promoted the apoptosis of further confirmed the antiproliferative and proapoptotic PANC-GR cells in a dose-dependent manner, accompanied effects of liraglutide on gemcitabine-resistant human by decreased expression of NF-κB and ABCG2; these pancreatic cancer cells, likely related to upregulation of effects could be abolished by either the GLP-1R antagonist the GLP-1R/cAMP/PKA signaling pathway, which needs Ex-9, the PKA inhibitor H89, or the NF-κB activator LPS. further investigation. Moreover, liraglutide increased the chemosensitivity of It is reported that GLP-1 exerts anti-inflammatory pancreatic cancer cells to gemcitabine, irrespective of in effects in chronic inflammatory diseases including type vitro or in vivo evidence. Together with our previous work 2 diabetes by inhibiting the NF-κB signaling pathway (Zhao et al. 2014a,b), we provided data supporting that (Lee & Jun 2016). In one study, liraglutide decreased activation of GLP-1R has a direct antitumor effect against NF-κB activation, which could be reversed by the PKA pancreatic cancer as well as synergistic antiproliferative inhibitor H-89, indicating the involvement of the and proapoptotic activity with gemcitabine. cAMP-PKA pathway in inhibition of NF-κB activation; GLP-1R is expressed in various tissues and organs of liraglutide treatment subsequently led to the attenuation the human body. In addition, it has been reported that of ovalbumin-induced airway inflammation (Zhu GLP-1R is expressed in a variety of endocrine tumors, but et al. 2015). In fact, NF-κB is not only an important not in carcinomas or lymphomas (Korner et al. 2007). Our transcription factor in inflammation, immune response, previous work showed that GLP-1R is expressed in normal cell proliferation, and apoptosis (Baker et al. 2011), but it pancreatic tissues; in contrast, pancreatic cancer tissues also has been found to be a key regulator of tumor growth exhibit low or absent expression of GLP-1R. Furthermore, and chemotherapy resistance (Pramanik et al. 2018). negative GLP-1R expression was more commonly observed Activation of NF-κB is frequently observed in pancreatic in advanced pancreatic cancer with a larger diameter and/ ductal adenocarcinoma, and an altered NF-κB signaling or lymphatic metastasis (Zhao et al. 2014a). This evidence pathway contributes to therapeutic resistance (Prabhu suggests a possible relationship between GLP-1R and the et al. 2014). Silencing NF-κB may induce apoptosis and

caspase-3 (Pan et al. 2008, Wang et al. 2012, 2016). In this study, we found that GLP-1R activation with liraglutide decreased the expression of NF-κB, which could be abolished by the GLP-1R antagonist Ex-9, the PKA inhibitor H89, or the NF-κB activator LPS, consequently suppressing the growth and enhancing the apoptosis of gemcitabine- resistant human pancreatic cancer cells. Consistent with our results, Iwaya et al. reported that the GLP-1R agonist exendin-4 could attenuate the proliferation of breast cancer cells through activation of GLP-1R and subsequent inhibition of NF-κB activation (Iwaya et al. 2017). Taken together, GLP-1R agonists may have an antitumor effect in gemcitabine-resistant human pancreatic cancer cells, at least in part by inhibiting activation of the NF-κB signaling pathway. Pancreatic cancer is a malignant tumor with a high mortality and a poor prognosis. Congenital or acquired drug resistance is the main reason for chemotherapy failure. Chemotherapeutic resistance is attributed to a variety of factors, among which drug efflux pumps mediated by ABC transporters is an important cause (Choi & Yu 2014). The ABCG2 protein is a member of the G subgroup of human ABC transporter proteins, and its gene is located on 4q22 of the human chromosome (Natarajan et al. 2012). ABCG2 has been found to confer the side population phenotype and therefore is considered as a marker of cancer stem cells (Ding et al. 2010). Studies have shown that the expression of ABCG2 is normally low or even absent in pancreatic tissues, but it is high in human pancreatic cancer cells (Wang et al. 2010, Skrypek et al. 2015). Further research has demonstrated that ABCG2 is a downstream gene of NF- B, which can Figure 5 κ Liraglutide enhances the chemosensitivity of pancreatic cancer cells to directly bind to the promoter region of the ABCG2 gene, gemcitabine. (A, B and C) PANC-1 cells were incubated or cultured with thereby transcriptionally activating the expression of GEM (0.2 μM), Lira (100 nM), or GEM (0.2 μM) + Lira (100 nM). (A) Cell ABCG2 and subsequently contributing to the occurrence proliferation was evaluated at 24 h and 48 h by using the Cell Counting Kit-8 assay. ※P < 0.05 (GEM vs control); &P < 0.05 (Lira vs control); §P < 0.05 and development of tumors (Shojaie & Ghaffari 2016, Sui (GEM + Lira vs control); #P < 0.05 (Lira + GEM vs GEM). (B) Flow cytometric et al. 2016). However, the underlying mechanism has not quantification of apoptosis with annexin V and PI staining (left), and the been clearly elucidated. Gemcitabine has been reported to percentage of apoptotic cells (right). (C) The expression levels of Bax and caspase-3 as detected by western blot (left), and their band intensities upregulate the expression of ABCG2/BCRP in pancreatic relative to β-actin (right). (D, E, F, G and H) Mouse xenograft models were cancer cells, which initiates the onset of chemoresistance established by subcutaneously injecting PANC-1 cells into the right flank (Sun et al. 2016). In the present study, we found that of each nude mouse. (D) Representative images of tumors from mice treated with PBS, GEM, Lira, or GEM + Lira. The weight gains of mice (E), PANC-GR cells had significantly higher expression of tumor volume (F), and tumor weights (G) were evaluated. The expression ABCG2 as compared with PANC-1 cells. Interestingly, levels of Ki-67 and Bax in tumor tissue were determined by GLP-1R activation with liraglutide led to decreased immunohistochemical analysis. Magnification, 100×. P* < 0.05 (vs control). Lira, liraglutide; GEM, gemcitabine. A full colour version of this figure is ABCG2 expression in PANC-GR cells and enhanced available at https://doi.org/10.1530/JME-19-0186. chemosensitivity of PANC-1 cells to gemcitabine, both in vitro and in vivo. These results suggest that GLP-1R increase the effectiveness of gemcitabine in pancreatic activation not only suppresses tumor progression, but cancer, accompanied by inhibition of Bcl-2 and the cell it may also attenuate the chemoresistance of pancreatic cycle regulatory protein cyclin D1 as well as activation of cancer cells to gemcitabine. Thus, our research provides a

https://jme.bioscientifica.com © 2020 Society for Endocrinology https://doi.org/10.1530/JME-19-0186 Published by Bioscientifica Ltd. Printed in Great Britain Downloaded from Bioscientifica.com at 09/27/2021 03:47:55PM via free access Journal of Molecular H Zhao et al. GLP-1R activation inhibits 64:2 112 Endocrinology pancreatic cancer scientific basis for elucidating the role and mechanism of Baker RG, Hayden MS & Ghosh S 2011 NF-kappaB, inflammation, and the GLP-1R signaling pathway in NF- B/ABCG2-mediated metabolic disease. Cell Metabolism 13 11–22. (https://doi. κ org/10.1016/j.cmet.2010.12.008) chemotherapeutic resistance as well as searching for Chari ST, Leibson CL, Rabe KG, Timmons LJ, Ransom J, De Andrade M molecular targets to reverse the chemotherapy resistance & Petersen GM 2008 Pancreatic cancer-associated diabetes mellitus: of pancreatic cancer. prevalence and temporal association with diagnosis of cancer. Gastroenterology 134 95–101. (https://doi.org/10.1053/j. In conclusion, these data provide the first evidence gastro.2007.10.040) that liraglutide alone displays significant antitumor Choi YH & YU AM 2014 ABC transporters in multidrug resistance and effects against gemcitabine-resistant human pancreatic pharmacokinetics, and strategies for drug development. Current Pharmaceutical Design 20 793–807. (https://doi.org/10.2174/1381612 cancer cells, and it further potentiates the therapeutic 82005140214165212) effectiveness of gemcitabine via modulation of the Ding XW, Wu JH & Jiang CP 2010 Abcg2: a potential marker of stem NF- B signaling pathway and downstream ABCG2. cells and novel target in stem cell and cancer therapy. Life Sciences κ 86 631–637. (https://doi.org/10.1016/j.lfs.2010.02.012) Thus, GLP-1R agonists seem to be safe and beneficial for Everhart J & Wright D 1995 Diabetes mellitus as a risk factor for patients complicated with pancreatic cancer and diabetes, pancreatic cancer. A meta-analysis. JAMA 273 1605–1609. especially for gemcitabine-resistant pancreatic cancer. Fan J, Xie Y, Xue J, Zhang Y & Yang Q 2013 Cellular apoptosis of hemocytes from Dendrolimus tabulaeformis Tsai et Liu larvae induced However, more research is still needed. with the secondary metabolites of Beauveria brongniartii (Sacc.) Petch. PLoS ONE 8 e71600. (https://doi.org/10.1371/journal.pone.0071600) Iwaya C, Nomiyama T, Komatsu S, Kawanami T, Tsutsumi Y, Hamaguchi Y, Horikawa T, Yoshinaga Y, Yamashita S, Tanaka T, et al. Supplementary materials 2017 Exendin-4, a glucagonlike peptide-1 receptor agonist, attenuates This is linked to the online version of the paper at https://doi.org/10.1530/ breast cancer growth by inhibiting NF-kappaB activation. Endocrinology JME-19-0186. 158 4218–4232. (https://doi.org/10.1210/en.2017-00461) Kim MP & Gallick GE 2008 Gemcitabine resistance in pancreatic cancer: picking the key players. Clinical Cancer Research 14 1284–1285. (https://doi.org/10.1158/1078-0432.CCR-07-2247) Declaration of interest Koehler JA & Drucker DJ 2006 Activation of glucagon-like peptide-1 The authors declare that there is no conflict of interest that could be receptor signaling does not modify the growth or apoptosis of perceived as prejudicing the impartiality of the research reported. human pancreatic cancer cells. Diabetes 55 1369–1379. (https://doi. org/10.2337/db05-1145) Koehler JA, Kain T & Drucker DJ 2011 Glucagon-like peptide-1 receptor activation inhibits growth and augments apoptosis in murine CT26 Funding colon cancer cells. Endocrinology 152 3362–3372. (https://doi. This work was supported by grants from the National Natural Science org/10.1210/en.2011-1201) Foundation of China (81602139, 81602403), the Natural Science Foundation Kolic J & Macdonald PE 2015 cAMP-independent effects of GLP-1 on of Tianjin City (16JCQNJC10100), and the Tianjin Hospital Science and beta cells. Journal of Clinical Investigation 125 4327–4330. (https:// Technology Foundation of Tianjin City (TJYY1508). doi.org/10.1172/JCI85004) Korner M, Stockli M, Waser B & Reubi JC 2007 GLP-1 receptor expression in human tumors and human normal tissues: potential for in vivo targeting. Journal of Nuclear Medicine 48 736–743. (https:// Data availability doi.org/10.2967/jnumed.106.038679) The datasets generated and analyzed during the present study are Lee YS & Jun HS 2016 Anti-inflammatory effects of GLP-1-based available from the corresponding author on reasonable request. therapies beyond glucose control. Mediators of Inflammation 2016 3094642. (https://doi.org/10.1155/2016/3094642) Lee W, Yoon YS, Han HS, Cho JY, Choi Y, Jang JY & Choi H 2016 Prognostic relevance of preoperative diabetes mellitus and the degree Author contribution statement of hyperglycemia on the outcomes of resected pancreatic ductal H Z and Z R contributed to the conception and design of the research. adenocarcinoma. Journal of Surgical Oncology 113 203–208. (https:// L H, X J, L Y, L D, and Y W performed the experiments. H Z and Z R analyzed doi.org/10.1002/jso.24133) the data. L H and L Y interpreted the experimental results. H Z prepared Ligumsky H, Wolf I, Israeli S, Haimsohn M, Ferber S, Karasik A, the figures. H Z and X J drafted the manuscript. L H edited and revised the Kaufman B & Rubinek T 2012 The peptide-hormone glucagon-like manuscript. All authors approved the final version of the manuscript. Z R is peptide-1 activates cAMP and inhibits growth of breast cancer cells. the principal investigator of this work. Breast Cancer Research and Treatment 132 449–461. (https://doi. org/10.1007/s10549-011-1585-0) Meloni AR, Deyoung MB, Lowe C & Parkes DG 2013 GLP-1 receptor References activated insulin secretion from pancreatic beta-cells: mechanism and glucose dependence. Diabetes, Obesity and Metabolism 15 15–27. Aggarwal G, Kamada P & Chari ST 2013 Prevalence of diabetes mellitus (https://doi.org/10.1111/j.1463-1326.2012.01663.x) in pancreatic cancer compared to common cancers. Pancreas 42 Mo W & Zhang JT 2012 Human ABCG2: structure, function, and its role 198–201. (https://doi.org/10.1097/MPA.0b013e3182592c96) in multidrug resistance. International Journal of Biochemistry and Amrutkar M & Gladhaug IP 2017 Pancreatic cancer chemoresistance to Molecular Biology 3 1–27. gemcitabine. Cancers 9 E157. (https://doi.org/10.3390/ Natarajan K, Xie Y, Baer MR & Ross DD 2012 Role of breast cancer cancers9110157) resistance protein (BCRP/ABCG2) in cancer drug resistance.

Biochemical Pharmacology 83 1084–1103. (https://doi.org/10.1016/j. acting human glucagon-like peptide-1 analog, given as monotherapy bcp.2012.01.002) significantly improves glycemic control and lowers body weight Pan X, Arumugam T, Yamamoto T, Levin PA, Ramachandran V, Ji B, without risk of hypoglycemia in patients with type 2 diabetes. Lopez-Berestein G, Vivas-Mejia PE, Sood AK, Mcconkey DJ, et al. Diabetes Care 30 1608–1610. (https://doi.org/10.2337/dc06-2593) 2008 Nuclear factor-kappaB p65/relA silencing induces apoptosis and Wang Y, Zhou Y, Zhou H, Jia G, Liu J, Han B, Cheng Z, Jiang H, Pan S & increases gemcitabine effectiveness in a subset of pancreatic cancer Sun B 2012 Pristimerin causes G1 arrest, induces apoptosis, and cells. Clinical Cancer Research 14 8143–8151. (https://doi. enhances the chemosensitivity to gemcitabine in pancreatic cancer org/10.1158/1078-0432.CCR-08-1539) cells. PLoS ONE 7 e43826. (https://doi.org/10.1371/journal. Prabhu L, Mundade R, Korc M, Loehrer PJ & Lu T 2014 Critical role of pone.0043826) NF-κB in pancreatic cancer. Oncotarget 5 10969–10975. (https://doi. Wang F, Xue X, Wei J, An Y, Yao J, Cai H, Wu J, Dai C, Qian Z, Xu Z, et al. org/10.18632/oncotarget.2624) 2010 hsa-miR-520h downregulates ABCG2 in pancreatic cancer cells to Pramanik KC, Makena MR, Bhowmick K & Pandey MK 2018 inhibit migration, invasion, and side populations. British Journal of Advancement of NF-kappaB signaling pathway: a novel target in Cancer 103 567–574. (https://doi.org/10.1038/sj.bjc.6605724) pancreatic cancer. International Journal of Molecular Sciences 19 E3890. Wang Y, Wu X, Zhou Y, Jiang H, Pan S & Sun B 2016 Piperlongumine (https://doi.org/10.3390/ijms19123890) suppresses growth and sensitizes pancreatic tumors to gemcitabine Shojaie N & Ghaffari SM 2016 Simultaneous analysis of Wnt and in a xenograft mouse model by modulating the NF-kappa B NF-kappaB signaling pathways in doxorubicin sensitive and pathway. Cancer Prevention Research 9 234–244. (https://doi. methotrexate resistant PLC/PRF/5 cells. Cell Journal 17 730–739. org/10.1158/1940-6207.CAPR-15-0306) (https://doi.org/10.22074/cellj.2016.3845) Wolf I, O’Kelly J, Rubinek T, Tong M, Nguyen A, Lin BT, Tai HH, Siegel RL, Miller KD & Jemal A 2015 Cancer statistics, 2015. CA: A Cancer Karlan BY & Koeffler HP 2006 15-Hydroxyprostaglandin Journal for Clinicians 65 5–29. (https://doi.org/10.3322/caac.21254) dehydrogenase is a tumor suppressor of human breast cancer. Cancer Skrypek N, Vasseur R, Vincent A, Duchene B, Van Seuningen I & Research 66 7818–7823. (https://doi.org/10.1158/0008-5472.CAN-05- Jonckheere N 2015 The oncogenic receptor ErbB2 modulates 4368) gemcitabine and irinotecan/SN-38 chemoresistance of human Zhang W, Ding W, Chen Y, Feng M, Ouyang Y, Yu Y & He Z 2011 pancreatic cancer cells via hCNT1 transporter and multidrug- Up-regulation of breast cancer resistance protein plays a role in resistance associated protein MRP-2. Oncotarget 6 10853–10867. HER2-mediated chemoresistance through PI3K/Akt and nuclear (https://doi.org/10.18632/oncotarget.3414) factor-kappa B signaling pathways in MCF7 breast cancer cells. Acta Sui H, Zhou LH, Zhang YL, Huang JP, Liu X, Ji Q, Fu XL, Wen HT, Biochimica et Biophysica Sinica 43 647–653. (https://doi.org/10.1093/ Chen ZS, Deng WL, et al. 2016 Evodiamine suppresses ABCG2 abbs/gmr050) mediated drug resistance by inhibiting p50/p65 NF-kappaB pathway Zhao H, Wang L, Wei R, Xiu D, Tao M, Ke J, Liu Y, Yang J & Hong T in colorectal cancer. Journal of Cellular Biochemistry 117 1471–1481. 2014a Activation of glucagon-like peptide-1 receptor inhibits (https://doi.org/10.1002/jcb.25451) tumourigenicity and metastasis of human pancreatic cancer cells via Sun Y, Gu M, Zhu L, Liu J, Xiong Y, Wei Y & LI F 2016 Gemcitabine PI3K/Akt pathway. Diabetes, Obesity and Metabolism 16 850–860. upregulates ABCG2/BCRP and modulates the intracellular (https://doi.org/10.1111/dom.12291) pharmacokinetic profiles of bioluminescence in pancreatic cancer Zhao H, Wei R, Wang L, Tian Q, Tao M, Ke J, Liu Y, Hou W, Zhang L, cells. Anti-Cancer Drugs 27 183–191. (https://doi.org/10.1097/ Yang J, et al. 2014b Activation of glucagon-like peptide-1 receptor CAD.0000000000000315) inhibits growth and promotes apoptosis of human pancreatic cancer Tamburrino A, Piro G, Carbone C, Tortora G & Melisi D 2013 cells in a cAMP-dependent manner. American Journal of Physiology: Mechanisms of resistance to chemotherapeutic and anti-angiogenic Endocrinology and Metabolism 306 E1431–E1441. (https://doi. drugs as novel targets for pancreatic cancer therapy. Frontiers in org/10.1152/ajpendo.00017.2014) Pharmacology 4 56. (https://doi.org/10.3389/fphar.2013.00056) Zhu T, Wu XL, Zhang W & Xiao M 2015 Glucagon like peptide-1 (GLP- Valenzuela MM, Neidigh JW & Wall NR 2014 Antimetabolite treatment 1) modulates OVA-induced airway inflammation and mucus for pancreatic cancer. Chemotherapy 3 137. (https://doi. secretion involving a protein kinase A (PKA)-dependent nuclear org/10.4172/2167-7700.1000137) factor-kappaB (NF-kappaB) signaling pathway in mice. International Vilsboll T, Zdravkovic M, Le-Thi T, Krarup T, Schmitz O, Courreges JP, Journal of Molecular Sciences 16 20195–20211. (https://doi. Verhoeven R, Buganova I & Madsbad S 2007 Liraglutide, a long- org/10.3390/ijms160920195)

Received in final form 23 October 2019 Accepted 17 December 2019 Accepted Manuscript published online 17 December 2019