Metformin Disrupts Crosstalk between G Protein–Coupled Receptor and Insulin Receptor Signaling Systems and Inhibits Pancreatic Cancer Growth

Ductal adenocarcinoma of the pancreas is a devastating disease, with an overall 5-year survival rate of only 3% to 5%. The incidence of this disease in the United States has increased recently to >37,000 new cases each year, and is now the fourth leading cause of cancer mortality in both men and women. As the current therapies offer very limited survival benefits, novel therapeutic strategies are urgently required to treat this aggressive disease.

G protein–coupled receptors (GPCR) and their cognate agonists are increasingly implicated as autocrine/paracrine growth factors for multiple solid tumors, including small cell lung cancer, colon, prostate, breast, and pancreas (1–3). Pancreatic cancer cell lines express multiple functional GPCRs using a Ca²⁺ mobilization assay as indicators of productive ligand-receptor interactions (4). A variety of GPCR agonists, including neurotensin, angiotensin II, and bradykinin, stimulated DNA synthesis in pancreatic cancer cell lines, including PANC-1 and MIAPaCa-2 (4–7). Furthermore, a broad-spectrum GPCR antagonist inhibited the growth of pancreatic cancer cells either in vitro or xenografted into nu/nu mice (8). Other studies showed increased expression of angiotensin II and neurotensin GPCRs in pancreatic cancer tissues (9–12). More recently, we identified a novel crosstalk between insulin/insulin-like growth factor-I receptors and GPCR signaling systems in pancreatic cancer cells, leading to the enhancement of GPCR-induced early signaling (13), including Ins(1,4,5)P₃ generation and increase in the intracellular Ca²⁺ concentration ([Ca²⁺]_i). Insulin-induced potentiation of GPCR signaling was mediated through the phosphatidylinositol-3-kinase/Akt/mTOR signaling module (13), a key pathway in insulin/insulin-like growth factor action (14). These findings assume an added importance in view of the large number of epidemiologic studies linking long-standing type 2 diabetes, obesity, and metabolic syndrome, characterized by peripheral insulin resistance and compensatory overproduction of insulin, with increased risk for developing pancreatic cancer (see ref. 15 for review). Given the complexity of the pancreatic microcirculation (16), and the close topographical relationship between the islets and small ducts in these diseases (17), the locally overproduced insulin is thought to act directly on ductal pancreatic cancer cells.

Metformin (1,1-dimethylbiguanide hydrochloride), one of the most widely prescribed drugs for the treatment of type 2 diabetes, stimulates AMP kinase (AMPK) in intact cells (18). AMPK activity is switched on by phosphorylation on Thr¹⁷² in the activation loop of the catalytic subunit by LKB-1/STK11, the product of the Peutz-Jegher syndrome tumor-suppressor gene (19). Major downstream targets of AMPK include TSC2 and raptor (20, 21). The AMPK-mediated phosphorylation of these targets leads to the inhibition of mTOR complex 1 activity. Interestingly, a recent epidemiologic report linked the administration of metformin with a 62% reduced risk of pancreatic cancer in patients with type 2 diabetes mellitus (22). In addition, metformin prevented carcinogen-induced pancreatic cancer induction in hamsters maintained on high-fat diets (23) and inhibited the growth of breast and colon cancer cells (24, 25). Despite its potential clinical implications, there is no understanding of the precise mechanism(s) by which metformin inhibits the proliferation of cancer cells, and it is not known whether metformin has any direct effect on pancreatic cancer growth.

In this study, we show that metformin disrupts the crosstalk between insulin receptor and GPCR signaling systems in pancreatic cancer cells. Specifically, metformin prevented insulin-induced augmentation of Ca²⁺ signaling, DNA synthesis, and anchorage-independent proliferation in response to stimulation with GPCR agonists in PANC-1 and MIAPaCa-2 pancreatic cancer cells. Metformin inhibited these signaling systems through AMPK. Furthermore, our results show that metformin administration inhibits the growth of PANC-1 and MIAPaCa-2 tumor xenografts in vivo.

Cells and culture conditions. The human pancreatic cancer cell lines PANC-1, MIAPaCa-2, BxPC-3, and AsPC-1 were obtained from the American Type Culture Collection. PANC-1 and MIAPaCa-2 cells were grown in DMEM with 2 mmol/L of glutamine, 1 mmol/L of sodium pyruvate, 100 units/mL of penicillin, and 100 μg/mL of streptomycin, and 10% fetal bovine serum (FBS) at 37°C in a humidified atmosphere containing 10% CO₂. BxPC-3 cells and AsPC-1 cells were grown in RPMI 1640 containing 100 units/mL of penicillin, 100 μg/mL of streptomycin, and 10% FBS at 37°C in a humidified atmosphere containing 5% CO₂.

Measurement of [Ca²⁺]_i. Ca²⁺ signaling was determined in cells grown on glass coverslips for 4 to 5 d, as previously described (13).

[³H]Thymidine incorporation into DNA. PANC-1, MIAPaCa-2, BxPC-3, and AsPC-1 cells (5 × 10⁴) were plated and grown in 3.5 cm tissue culture plates in DMEM with 2 mmol/L of glutamine, 1 mmol/L of sodium pyruvate (PANC-1 and MIAPaCa-2 cells) or RPMI 1640 (BxPC-3 and AsPC-1 cells), and 10% FBS. After 5 d, the cultures were washed twice with PBS and incubated in serum-free medium. After 24 h, the cultures were transferred to fresh medium containing the specified concentration of agonist and/or inhibitor for 17 h and then pulse-labeled for 6 h with [³H]thymidine (0.25 μCi/mL). The acid-insoluble radioactivity was determined as previously described (7).

Western blot analysis. Confluent cultures of PANC-1 cells grown on 6 cm dishes were washed twice with PBS and incubated with serum-free medium for 3 h. Metformin was added at 5 mmol/L and the cultures were incubated at 37°C for the specified times (0 min–24 h). To detect the activation of AMPKα, the cultures were washed in cold PBS and directly lysed in 2× SDS-PAGE sample buffer [200 mmol/L Tris-HCl (pH 6.8), 2 mmol/L EDTA, 0.1 mol/L Na₃VO₄, 6% SDS, 10% glycerol, and 4% 2-mercaptoethanol]. The lysates were subjected to SDS-PAGE on 10% gels and separated proteins were transferred to Immobilon-P membranes (Millipore). Western blots were then performed on membranes incubated overnight with phosphorylated AMPKα (Thr¹⁷²) or phosphorylated ERK1/2 (Thr²⁰² and Tyr²⁰⁴) monoclonal antibodies at a dilution of 1:1,000 in PBS containing 5% nonfat dried milk and 0.1% Tween 20. The immunoreactive bands were detected with enhanced chemiluminescence reagents (Amersham). The same membranes were subsequently stripped and probed in a similar fashion with anti-AMPKα polyclonal antibody or anti-ERK at a dilution of 1:2,000 in PBS containing 5% nonfat dried milk and 0.1% Tween 20.

Anchorage-dependent cell proliferation. PANC-1 and MIAPaCa-2 cells (2 × 10⁴) were plated on tissue culture 12-well plates in DMEM containing 1% FBS. After 24 h of incubation at 37°C, 5 nmol/L of neurotensin, in the absence or presence of 10 ng/mL of insulin, was added with or without 1 mmol/L of metformin. The cultures were then incubated for 7 to 9 d, and the total cell count was determined from a minimum of four wells per condition using a Coulter counter, after cell clumps were disaggregated by passing the cell suspension 10 times through a 19-gauge, and subsequently, a 21-gauge needle.

Anchorage-independent cell proliferation. Cells were plated on tissue culture 12-well plates coated with polyhydroxyethylmethacrylate [poly-(HEMA)] as previously described (7). Under these conditions, cells are denied attachment to the substratum. Cultures of PANC-1 or BxPC-3 cells, 3 to 5 d after passage, were washed and suspended in DMEM (PANC-1) or RPMI 1640 (BxPC-3 cells). Cells were then disaggregated by two passes through a 19-gauge needle into an essentially single-cell suspension as judged by microscopy. Cell numbers were determined using a Coulter counter, and 2 × 10⁴ cells were seeded in either DMEM containing 1% FBS on the poly-(HEMA)–coated dishes (PANC-1 cells) or in RPMI containing 1% FBS (BxPC-3 cells). After 24 h of incubation at 37°C, 5 nmol/L of neurotensin (PANC-1 cells) or 5 nmol/L of bradykinin (BxPC-3 cells), in the absence or in the presence of 10 ng/mL of insulin, was added with or without 1 mmol/L of metformin. The cultures were then incubated in a humidified atmosphere containing 10% CO₂ at 37°C for 7 to 9 d, and the total cell count was determined from a minimum of four wells per condition using a Coulter counter, after cell clumps were disaggregated by passing the cell suspension 10 times through a 19-gauge, and subsequently, a 21-gauge needle.

Mice xenografts. Early passage PANC-1 or MIAPaCa-2 cells were harvested, and 2 × 10⁶ cells were implanted into the right flanks of male nu/nu mice. The male nu/nu mice were maintained in a specific pathogen–free facility at the University of California at Los Angeles. The UCLA Chancellor's Animal Research Committee approved all the animal experiments.

The animals were randomized into control and treated groups (10 mice per group). Treatment was initiated when the tumors reached a mean diameter of 2 mm, and the first day of treatment in both cases was designated as day 0. For injection, metformin was dissolved in sterile saline and was given once daily i.p. at 250 mg/kg (50 μL/mouse). The control group received vehicle only (50 μL saline). Tumor volume (V) was measured with an external caliper every 4 d and it was calculated as V = 0.52 (length × width²). The treatment was continued until any of the tumors reached ∼1.5 cm in length, when all the animals were sacrificed and the tumors removed. The volume of the excised tumors was calculated as V = 0.52 (length × width × depth).

Chemicals and reagents. DMEM was obtained from Life Technologies. Neurotensin, angiotensin, bradykinin, and insulin were obtained from Sigma Chemical. Metformin was obtained from Sigma-Aldrich. The AMPK inhibitor, compound C, was purchased from Calbiochem. An antibody that detects the phosphorylated state of Thr¹⁷² of AMPKα and an antibody against AMPKα were purchased from Millipore. An antibody that detects the phosphorylated state of Thr²⁰² and Tyr²⁰⁴ of ERK1/2 and an antibody against ERK1/2 were purchased from Cell Signaling Technology. Horseradish peroxidase–conjugated antirabbit IgG and antimouse IgG were from Amersham Biosciences. All other reagents were of the highest grade available.

Statistical analysis. The values obtained are presented as the mean ± SE and analyzed with Student's t test, using SigmaPlot 2000 (SPSS).

Metformin blocks the crosstalk between insulin and GPCR receptor signaling on intracellular Ca²⁺ mobilization. As a first step to investigate whether metformin disrupts the crosstalk between the insulin/insulin-like growth factor and GPCR signaling systems, we determined whether it prevents the stimulatory effect of insulin on GPCR-induced increase in [Ca²⁺]_i in pancreatic cancer cells. As shown in Fig. 1A, the addition of neurotensin (5 nmol/L) to PANC-1 cells induced a rapid increase in [Ca²⁺]_i from a basal level of 142 ± 15 nmol/L (mean ± SE; n = 20) to a peak value of 735 ± 75 nmol/L (n = 20) at 25 to 35 seconds, which subsequently declined toward a plateau phase (Fig. 1A,, left, black trace and right, open column). In contrast, the addition of insulin (10 ng/mL) to PANC-1 cells did not produce any detectable change in [Ca²⁺]_i. However, treatment of PANC-1 cells with 10 ng/mL of insulin for 5 minutes prior to stimulation markedly enhanced neurotensin-induced increase in [Ca²⁺]_i that now reached a peak value of 1,185 ± 134 nmol/L (n = 25; Fig. 1A , gray trace and black column).

Pretreatment with 5 mmol/L of metformin for 1 hour did not have any significant effect on the [Ca²⁺]_i increase induced by neurotensin alone (Fig. 1A,, bottom-middle and striped column), but completely blocked the enhancing effect of insulin on neurotensin-induced Ca²⁺ mobilization (peak value 559 ± 102 nmol/L, top-middle and gray column). Similar results were obtained with either MIAPaCa-2 cells stimulated with neurotensin (Fig. 1B,, left) or BxPC-3 cells stimulated with either bradykinin or angiotensin (Fig. 1B , middle and right). These results indicate that exposure to metformin abrogates the crosstalk between insulin and GPCR signaling systems in a variety of pancreatic cancer cell lines.

Metformin abolished the crosstalk between insulin and GPCR receptor signaling systems on DNA synthesis and cell proliferation in pancreatic cancer cells. We next determined whether crosstalk between GPCR agonists and insulin in pancreatic cancer cells leads to long-term biological responses, including ERK signaling, DNA synthesis, and cell proliferation. Cultures of PANC-1, MIAPaCa-2, or BxPC-3 cells were incubated in medium containing increasing doses of metformin with neurotensin (PANC-1 and MIAPaCa-2) or bradykinin (BxPC-3) in the absence or presence of insulin, as indicated in Fig. 2. After 17 hours of incubation, the cultures were pulse-labeled with [³H]-thymidine. The results in Fig. 2 show that either insulin or GPCR agonists stimulated DNA synthesis in PANC-1 (A), MIAPaCa-2 (C), and BxPC-3 (D) cells. The combination of the GPCR agonists and insulin (black columns) induced a striking enhancement in DNA synthesis in all three cell lines and induced marked prolongation of ERK signaling, as shown by Western blot analysis of lysates of PANC-1 cells treated with neurotensin and insulin (Fig. 2B).

The salient feature of the results shown in Fig. 2 is that metformin abolished the increase in DNA synthesis induced by the combination of neurotensin and insulin in PANC-1 cells (Fig. 2A), and prevented the prolonged ERK signaling induced by neurotensin and insulin in PANC-1 cells (Fig. 2B). Similarly, metformin prevented the stimulation of DNA synthesis induced by neurotensin and insulin in MIAPaCa-2 cells (Fig. 2C), and bradykinin and insulin in BxPC-3 cells (Fig. 2D).

In order to determine whether metformin inhibits the proliferation of pancreatic cancer cells, single cell suspensions of PANC-1 cells were plated on normal culture dishes (anchorage-dependent growth; Fig. 3A) or on culture dishes coated with poly-(HEMA), which prevents the adhesion of the cells to the substratum (anchorage-independent growth; Fig. 3B). The cells were incubated in medium supplemented with neurotensin, insulin, or neurotensin and insulin in the presence or absence of 1 mmol/L of metformin. As illustrated in Fig. 3, the combination of neurotensin and insulin promoted a marked increase in cell number in both anchorage-dependent and anchorage-independent conditions. The addition of metformin markedly inhibited the stimulation of either anchorage-dependent or anchorage-independent growth induced by the combination of neurotensin and insulin in PANC-1. Metformin inhibited the stimulation of anchorage-dependent growth of MIAPaCa-2 cells induced by neurotensin and insulin (Supplementary Fig. S1), and the anchorage-independent growth of BxPC-3 cells induced by bradykinin and insulin (Supplementary Fig. S2). The results show that metformin directly inhibits DNA synthesis and proliferation induced by insulin and GPCR agonists in pancreatic cancer cells.

Metformin inhibits crosstalk between insulin and GPCR signaling systems through AMPK. AMPK activity is switched on by phosphorylation on Thr¹⁷² in the activation loop of the catalytic subunit by LKB-1/STK11. In order to determine whether metformin stimulates AMPK activation in human pancreatic cancer cells, we used Western blot analysis with an antibody that detects the phosphorylated state of the α subunit of AMPK on Thr¹⁷². As shown in Fig. 4A, treatment of PANC-1 cells with metformin induced striking and sustained increase in AMPK phosphorylation on Thr¹⁷², indicating that metformin treatment induces AMPK activation in PANC-1 cells.

Next, we determined whether a specific inhibitor of AMPK (compound C; ref. 26) reverses the inhibitory effects of metformin on insulin-induced augmentation of GPCR-mediated Ca²⁺ signaling. As shown in Fig. 4B, treatment with compound C (5 μmol/L) reversed the inhibitory effect of metformin on neurotensin-induced Ca²⁺ signaling in insulin-pretreated PANC-1 cells. A similar reversal of metformin inhibition was obtained with BxPC-3 cells treated with insulin and stimulated with bradykinin (Supplementary Fig. S3).

We also determined whether the selective inhibitor of AMPK counteracts the inhibitory effects of metformin on DNA synthesis induced by neurotensin and insulin in PANC-1. As shown in Fig. 4C, compound C (5 μmol/L) reversed the inhibitory effect of metformin on DNA synthesis induced by either insulin or by the combination of insulin and neurotensin in PANC-1 cells. The addition of compound C in the absence of metformin did not induce any significant effect [the levels of DNA synthesis were similar to those shown in Fig. 4C  (open columns)]. Treatment with compound C (5 μmol/L) also reversed the inhibitory effect of metformin on DNA synthesis induced by bradykinin and insulin in BxPC-3 cells (Supplementary Fig. S4).

In line with the results presented above, LKB-1/STK11 is expressed in many pancreatic cancer cells, including PANC-1, MIAPaCa-2, and BxPC-3. However, LKB-1 is silenced by promoter hypermethylation in the pancreatic cancer cell line AsPC-1 (27). If the effects of metformin are mediated by AMPK, AsPC-1 cells (in contrast with PANC-1, MIAPaCa-2, and BxPC-3) should be refractory to metformin because these cells do not express LKB-1/STK11, the upstream kinase that phosphorylates and activates AMPK in response to metformin. Treatment of AsPC-1 cells with 1 mmol/L of metformin did not prevent the stimulation of DNA synthesis induced by neurotensin and insulin in these cells, in contrast to results obtained in parallel cultures of PANC-1 cells (Supplementary Fig. S5). Taken together, these results indicate that metformin disrupts the crosstalk between insulin and GPCR signaling systems through AMPK in human pancreatic cancer cells.

Metformin inhibits the growth of PANC-1 or MIAPaCa-2 xenografts in nude mice. Given our results showing the inhibitory effects of metformin on signaling and proliferation in pancreatic cancer cells, we next examined whether this compound could inhibit pancreatic cancer growth using PANC-1 and MIAPaCa-2 tumor xenografts in nude mice. The animals were randomized into control and metformin-treated groups (10 mice per group). Treatment was initiated when the tumors reached a mean diameter of 2 mm. Metformin was given once daily i.p. at 250 mg/kg for the duration of the experiment. As shown in Fig. 5, administration of metformin strikingly decreased the growth of either PANC-1 or MIAPaCa-2 cells xenografted in nude mice. For example, the tumor volumes of PANC-1 xenografts at the end of the experiment (day 41) were 286.11 ± 50.95 mm³ in the control and 98.47 ± 29.20 mm³ in the metformin-treated group (P = 0.0057; Fig. 5A,, inset). MIAPaCa-2 xenografts at the end of the experiment (day 25) were 228.35 ± 40.99 mm³ in the control and 92.67 ± 20.10 mm³ in the metformin-treated group (P = 0.0072; Fig. 5B , inset). These results show, for the first time, that metformin inhibits the growth of human pancreatic cancer cells xenografted into nude mice.

Recently, we identified a novel crosstalk between insulin/insulin-like growth factor-I receptors and GPCR signaling systems in pancreatic cancer cells, leading to the enhancement of GPCR-induced early signaling (13). Insulin-induced potentiation of Gq signaling was prevented by either inhibitors of phosphatidylinositol-3-kinase or by rapamycin, a specific inhibitor of mTOR complex 1. These findings indicated that in addition to its well-established role in the regulation of protein synthesis, the rapamycin-sensitive phosphatidylinositol-3-kinase/TORC1/S6K pathway mediates a novel crosstalk between insulin receptors on GPCR signaling systems (13). The results presented in this study showed that insulin also enhances long-term biological responses induced by GPCR agonists in pancreatic cancer cells, including DNA synthesis, proliferation, and anchorage-independent growth. We conclude that potentiation of Gq signaling by insulin through an mTOR-dependent pathway provides a crosstalk mechanism by which insulin enhances the mitogenic responsiveness of pancreatic cancer cells to Gq-coupled receptor agonists.

The biguanide metformin is the most widely prescribed drug for the treatment of type 2 diabetes, worldwide. The primary systemic effect of metformin is the lowering of blood glucose levels through reduced hepatic gluconeogenesis and increased glucose uptake in skeletal muscles and adipose tissue (28). Metformin not only lowers blood glucose but also reduces the hyperinsulinemia associated with insulin resistance. At the cellular level, metformin is known to stimulate the activation of AMPK (18), a conserved regulator of the cellular response to low energy that it is activated when ATP concentrations decrease and 5′-AMP concentrations increase in response to nutrient deprivation, hypoxia, and metformin administration (19). AMP directly binds to the AMPK γ-subunit, causing a conformational change that exposes Thr¹⁷² in the activation loop of the α-subunit, allowing it to be phosphorylated by LKB-1/STK11, the product of the Peutz-Jegher syndrome tumor-suppressor gene (19). Interestingly, the Peutz-Jegher syndrome is characterized by predisposition to various gastrointestinal neoplasms, including pancreatic adenocarcinoma (29) and intraductal papillary-mucinous neoplasms (30).

Because metformin-induced activation of AMPK inhibits mTOR function via TSC2 and raptor phosphorylation (20, 21, 31, 32), we hypothesized that metformin disrupts the crosstalk between insulin receptor and Gq signaling via AMPK. Here, we present several lines of evidence supporting this hypothesis: (a) metformin completely blocked the enhancing effect of insulin on neurotensin-induced Ca²⁺ mobilization in either PANC-1 or MIAPaCa-2 cells stimulated with neurotensin or in BxPC-3 cells stimulated with either bradykinin or angiotensin. (b) Metformin abolished the increase in DNA synthesis induced by neurotensin and insulin in PANC-1 or MIAPaCa-2 cells. Similarly, metformin prevented the stimulation of DNA synthesis promoted by bradykinin and insulin in BxPC-3 cells. (c) Pancreatic cancer cell treatment with metformin produced a marked and sustained increase in AMPK activation, as judged by Thr¹⁷² phosphorylation. (d) Treatment with compound C, a specific inhibitor of AMPK, reversed the inhibitory effect of metformin on neurotensin-induced Ca²⁺ signaling in insulin-pretreated PANC-1 cells. A similar reversal of metformin inhibition was obtained with BxPC-3 cells treated with insulin and stimulated with bradykinin. (e) Compound C also reversed the inhibitory effect of metformin on DNA synthesis induced by the combination of insulin and neurotensin in PANC-1 cells or by the combination of bradykinin and insulin in BxPC-3 cells. (f) Metformin did not disrupt crosstalk between insulin and GPCR signaling systems in the pancreatic cancer AsPC-1 cells which do not express LKB-1/STK11 (27), the upstream kinase that phosphorylates and activates AMPK. These results indicate that metformin disrupts crosstalk between insulin and GPCR signaling systems through AMPK in human pancreatic cancer cells.

Recent epidemiologic reports linked the administration of metformin with a reduced incidence and improved prognosis in patients with cancer (33, 34). A recent study showed a statistically significant association between metformin therapy and reduced risk of pancreatic cancer in patients with type 2 diabetes mellitus (22). In addition to these epidemiologic associations, carcinogen-induced pancreatic cancer in hamsters maintained on high-fat diets was prevented by metformin (23) and metformin inhibited the growth of breast and colon cancer cells (24, 25, 35). However, the precise mechanisms involved remain incompletely understood. In view of the direct inhibitory effects of metformin on signaling and proliferation of pancreatic cancer cells shown in the present study, we examined whether this compound inhibits pancreatic cancer growth using PANC-1 and MIAPaCa-2 tumor xenografts in nude mice. Our results show, for the first time, that metformin markedly inhibits the growth of human pancreatic cancer cells xenografted into nude mice.

In conclusion, our results raise the attractive possibility that treatment with metformin, a widely used agent for counteracting metabolic syndrome and type 2 diabetes, directly inhibits pancreatic cancer cell proliferation. The results provide a basis for novel therapeutic strategies for the treatment of pancreatic cancer, a devastating disease with limited survival options.

No potential conflicts of interest were disclosed.

Grant support: NIH grants R21CA137292, RO1DK56930, RO1DK55003, and P30DK41301 (E. Rozengurt).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.