Background
Over recent decades the incidence of metabolic disorders, such as obesity and type 2 diabetes mellitus, has increased as a consequence of westernized lifestyle and changes in diet. These conditions are in turn associated with an increased risk of developing cancer [
1‐
3]. Epidemiological studies have demonstrated that obesity and type 2 diabetes are among the top three modifiable risk factors for pancreatic cancer [
2,
4‐
8]. Almost 80% of pancreatic cancer patients present with either new-onset type 2 diabetes or impaired glucose tolerance at the time of diagnosis [
9,
10]. The relationship between type 2 diabetes and pancreatic cancer is complex and it remains unclear whether type 2 diabetes contributes to the development of pancreatic cancer or if precancerous cells cause the diabetes. Individuals with elevated fasting glucose and glycated haemoglobin (H
bA
1c) levels [
11,
12], or with higher c-peptide or insulin levels have a two to four-fold increase in the risk of pancreatic cancer [
1,
7]. Type 2 diabetes patients also demonstrate an increased risk of pancreatic cancer-related death as compared with those without diabetes [
13]. Type 2 diabetes is characterized by hyperglycaemia and peripheral insulin resistance with compensatory hyperinsulinemia. Aside from its metabolic actions, insulin can mediate direct mitogenic effects through the insulin receptor (IR) and insulin-like growth factor I (IGF-I) receptor (IGF-IR). Insulin may also affect the cancer risk indirectly
via increased production and bioavailability of IGF-I [
6,
14]. Additionally, hyperglycaemia can increase the sensitivity to IGF-I [
4], thereby enhancing its mitogenic potential and providing an additional link between type 2 diabetes and cancer.
Insulin-sensitizing and glucose lowering drugs, such as metformin, are used as first-line treatment in the management of type 2 diabetes to improve glycaemic control in patients with insulin resistance. The key metabolic action of metformin involves the inhibition of hepatic glucose secretion, which consequently decreases the hyperinsulinemia. This mechanism is mediated
via activation of the energy-sensing AMP-activated protein kinase (AMPK) in hepatocytes, through the liver kinase B1 (LKB1) signalling pathway [
15]. Although metformin can lower blood glucose, the levels rarely remain within the normal range and as the type 2 diabetes progresses, additional medication such as exogenous insulin is often required to control patients’ hyperglycaemia [
16,
17]. In addition to its anti-diabetic effects, metformin has recently been postulated to have a protective role against cancer. Epidemiological and retrospective studies have demonstrated that diabetic patients taking metformin not only have a lower incidence of pancreatic cancer, but also an improved cancer outcome [
18‐
21]. The indicated anti-neoplastic activity of metformin may relate to reduced plasma insulin concentrations or by direct effects on the tumour cells. Recent studies suggest that metformin-induced AMPK activation at Thr
172 inhibits the central growth control node mammalian target of rapamycin mTOR, thus preventing protein synthesis and cell proliferation [
22]. Metformin has recently been shown to possess anti-tumour effects, both in AMPK-dependent and independent manners [
23‐
25].
Although an increasing number of studies demonstrate the anti-tumour effects of metformin, relatively little is known about the effects and underlying mechanisms of metformin on pancreatic cancer cells. The goal of this study was to examine the direct effects of metformin on human pancreatic cancer cells in the context of normal or elevated glucose levels. Effects on proliferation, apoptosis, AMPK activation and influence on and by the IGF-I pathway were analysed.
Methods
Materials
All chemicals and reagents were purchased from Sigma Aldrich (St. Louis, Mo, USA) unless stated otherwise. Cell culture media, penicillin/streptomycin and fetal bovine serum (FBS) were purchased from Invitrogen (Paisley, UK). IGF-I was purchased from GroPep (Adelaide, Australia). MTT; Cell Proliferation Kit I was derived from Roche (Mannheim, Germany). Anti-cleaved PARP, anti-phospho-AMPKThr172, anti-phospho-AMPKSer485, anti-AMPK, anti-IRS-1, anti-phospho-IGF-IRβ/phospho-IRβ, anti-phospho-AktSer473 and anti-Akt antibodies were purchased from Cell Signaling Technology Inc. (Beverly, MA, USA). Anti-IGF-IRβ was obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA) and anti-GAPDH from Millipore (Temecula, CA, USA).
Cell culture
The human pancreatic adenocarcinoma cell lines AsPC-1, BxPC-3, PANC-1 and MIAPaCa-2 were purchased from ATCC-LGC Standards (Manassas, VA, USA). The cells were maintained in RPMI1640 or DMEM supplemented with 10% FBS and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin) in a humified 5% CO2 atmosphere at 37°C. All experiments were performed in glucose-free RPMI1640 or DMEM supplemented with 5 mM (normal) or 25 mM (high) D-glucose, 2 mM L-glutamine and antibiotics as above (serum-free media; SFM), unless stated otherwise.
MTT proliferation assay
Cells were plated (10 × 103 cells/well) in 96-well plates in growth media with 5 mM glucose for 24 h before switching to SFM with 5 mM or 25 mM glucose for another 24 h. Cells were subsequently dosed with increasing concentrations of metformin (0–20 mM) in SFM with 5 mM or 25 mM glucose in sextuplicates (n = 6 wells). SFM with either 5 mM or 25 mM was used as control. Following incubation for 24–72 h, cell proliferation was assessed by MTT according to the manufacturer’s instructions. The samples were measured on a Labsystems Multiskan Plus plate reader (test wavelength 595 nm, reference wavelength 660 nm) using the DeltaSoft JV software (BioMetallics Inc., Princeton, NJ, USA).
Western immunoblotting
Cells were cultured (6 × 10
5 cells/well) in 6-well plates for 24 h. After an additional 24 h in normal glucose SFM, the cells were dosed with metformin (0–20 mM) in SFM or 1% FBS SFM with 5 or 25 mM glucose for 24 h. Cells were then spiked with IGF-I (100 ng/ml) as indicated for the final 15 min of incubation. Cells were lysed as previously described [
26]. Protein concentrations were determined using BCA protein assay reagent kit (ThermoFisherScientific, Waltham, MA, USA). Lysates were dissolved in Laemmli buffer, boiled for 5 minutes and separated (60–65 μg protein per lane) by SDS-PAGE (8% or 12%) and transferred to 0.2 μm Hybond-C extra nitrocellulose membrane (Amersham Biosciences, Buckinghamshire, UK). The membranes were blocked with 5% (w/v) milk in Tris-buffered saline Tween-20 (TBST) and probed overnight (4°C) with the indicated antibodies, all used at dilutions of 1:1000. Immunoblotted proteins were detected using HRP-conjugated secondary antibodies and visualized by SuperSignal West Extended Duration Substrate (ThermoFisherScientific) using BioRad Chemidoc XRS + system and Image lab software.
Statistical analysis
Proliferation data are expressed as means ± SE of six replicate wells. Densitometry analyses of Western blot data were performed using Image J software (NIH, USA) and are expressed as means ± SE of three individual experiments, unless stated otherwise. Statistical analyses were performed by one- or two-way ANOVA with Bonferroni post hoc test using GraphPad prism software. A P-value of <0.05 was considered statistically significant.
Discussion
Type 2 diabetes or impaired glucose tolerance frequently occurs in pancreatic cancer patients. Compared to other treatments, diabetic patients on metformin have a reduced risk of approximately 40% of developing several types of cancer, including pancreatic cancer [
19‐
21]. However, the molecular relationships underlying the metabolic and suggested anti-cancer actions of metformin remain poorly understood. Additionally, the importance of optimal glucose control for the anti-tumour effects of metformin has not been fully established. In this study, we describe direct anti-proliferative actions by metformin using
in vitro models of pancreatic cancer. In addition, we demonstrate that elevated glucose levels impair AMPK activation and reduce the efficacy of metformin. Importantly, we show a novel role for metformin on human pancreatic cancer cells that may contribute to its indicated anti-cancer actions among type 2 diabetic patients.
Metformin is believed to act mainly through activation of the energy-conserving LKB1-AMPK pathway. Physiological activation of the AMPK metabolic checkpoint in response to nutrient depletion and energy stress suppresses energy-consuming cellular processes such as protein synthesis and cell division. We found that metformin during normal glucose conditions significantly reduced proliferation and promoted apoptosis through PARP cleavage in pancreatic cancer cells with functional LKB1 (MIAPaCa-2, BxPC-3 and PANC-1), while being incapable of suppressing growth under the same conditions in AsPC-1 pancreatic cancer cells. AsPC-1 cells have previously been reported to carry an epigenetic inactivation of LKB1 [
27]. Our findings are consistent with prior observations, showing pro-apoptotic actions on breast cancer cells [
28,
29] and that a functional LKB1 was required for the
in vitro anti-proliferative effect of metformin [
25,
30].
Previous work indicates that metformin functions by activating AMPK at Thr
172 with subsequent downstream inhibition of the growth promoting PI3K/Akt/mTOR pathway [
29,
31,
32]. Similarly, we also found the growth inhibitory properties of metformin to be associated with the activation of AMPK
Thr172 in pancreatic cancer cells. Under hyperglycaemic conditions, the efficacy of metformin was reduced with less anti-proliferative and pro-apoptotic activity observed. Other investigators have reported that lung and colon carcinoma cells were more sensitive to metformin-induced growth inhibition at low glucose concentrations, while no significant effect of metformin on cell death was observed in high glucose conditions [
30]. Similarly, a recent study demonstrated anti-proliferative effects on pancreatic cancer cells by metformin at the low 0.05-1 mM range at normal (5 mM) glucose conditions [
33]. This study is in concordance with our data demonstrating direct anti-tumour effects of metformin and supports our findings of enhanced sensitivity at physiological normal glucose levels. We have now shown that the lower anti-proliferative effect of metformin on pancreatic cancer cells at higher glucose levels correlates to an impaired AMPK
Thr172 activation and a shifted balance from AMPK
Thr172 towards AMPK
Ser485 activation. The role of AMPK
Ser485 in the complex AMPK signaling network is at present not completely clear and conflicting reports exist. A recent study indicated that endogenous protein kinase A (PKA)-induced activation of AMPK
Ser485 in pancreatic beta cells did not affect the phosphorylation status of AMPK
Thr172. However, the activation of Thr
172 and Ser
485 were inversely correlated in response to glucose [
34]. Other studies have proposed that PKB/Akt-induced phosphorylation of AMPK
Ser485 can counteract AMPK
Thr172 activation, thereby reducing the effects of metformin [
31,
32,
35].
Hyperinsulinemia with resulting increased circulating levels of IGF-I have been suggested to play a role in the connection between type 2 diabetes and cancer [
36]. Activation of the IR and IGF-IR result in receptor autophosphorylation and recruitment of insulin receptor substrate (IRS) 1, which in turn activates the PI3K/Akt pathway leading to protein synthesis and cell survival [
14]. The responsiveness to IGF-I can be enhanced by exposure to high glucose concentrations [
4], which may then further promote cancer progression. In pancreatic cancer cells, IGF-I stimulated a pronounced phosphorylation of Akt and also AMPK
Ser485. However, at physiologically normal glucose levels, IGF-I stimulated AMPK
Ser485 phosphorylation did not appear to antagonize pharmacological activation of AMPK
Thr172 by metformin. Instead, we established that metformin under these conditions suppressed IGF-IR/IR phosphorylation causing a downstream inhibition of both basal and IGF-I stimulated Akt phosphorylation. It is well established that IGF-IR
via activation by its ligands transmits mitogenic signals leading to the survival and proliferation of multiple types of cancer. Mechanisms by which metformin inhibits these pathways may thus contribute to the anti-tumour effects previously observed in response to metformin. Studies in other cell types have shown that during normal glucose conditions, AMPK
Thr172 can phosphorylate inhibitory serine residues on IRS-1, which prevents signalling through the PI3K/Akt pathway [
29,
37]. However, studies have also shown that Akt at high glucose conditions can inhibit AMPK by phosphorylation of Ser
485, which prevents activation of Thr
172 and thereby the action of metformin [
29,
31,
35,
37,
38]. In keeping with this, we observed a strong activation of Akt and AMPK
Ser485 following IGF-I stimulation at high glucose, which was sustained after exposure to metformin. At high glucose, IGF-I induced Akt and AMPK
Ser485 phosphorylation appeared to correlate with a further reduction of the already-impaired AMPK
Thr172 phosphorylation by metformin.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AHR designed the study. EK, KS and AHR made significant contributions to the experimental design, acquisition and interpretation of data, manuscript preparation and editing. RA and AHR provided funding for the study and manuscript editing. All authors read and approved the final manuscript.