Mechanism of Metformin-dependent Inhibition of Mammalian Target of Rapamycin (mTOR) and Ras Activity in Pancreatic Cancer

Cell Lines, Antibodies, and Reagents

Human Panc1 pancreatic cancer cells were purchased from American Type Culture Collection (Manassas, VA). Panc28 and L3.6pL pancreatic cancer cells that express Ras mutation were provided by The University of Texas M.D. Anderson Cancer Center. All three cell lines were maintained in DMEM/F-12 (Sigma) supplemented with 0.22% sodium bicarbonate, 0.022% bovine serum albumin, 5% fetal bovine serum, and 10 ml/liter 100× antibiotic, antimycotic solution (Sigma) at 37 °C in the presence of 5% CO₂. Sp1 antibody was purchased from Millipore (Temecula, CA); Sp3 and Sp4 antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). FAS, Ras, p-mTOR, mTOR, p-4EBP, 4EBP, S6 ribosomal protein, and phospho-S6 ribosomal protein were purchased from Cell Signaling Technology (Danvers, MA). Metformin was purchased from Calbiochem (EMD Millipore, Billerica, MA). Chemiluminescence reagents (Immobilon Western) for Western blot imaging were purchased from Millipore. Active Ras detection assay kit was purchased from Cell Signaling Technology. Nuclear and cytoplasmic extraction kit was purchased from Thermo Scientific.

Cell Proliferation Assay

Panc28 and L3.6pL pancreatic cancer cells (3 × 10⁴/well) were seeded in 12-well plates with 2.5% charcoal-stripped FBS, allowed to attach for 24 h, and treated with different concentrations of NVP-BEZ235, a dual PI3K/mTOR inhibitor. Cells were then trypsinized and counted after 24 and 48 h using a Coulter Z1 cell counter. Each experiment was determined in triplicate, and results are expressed as mean ± S.E. for each set of experiments.

Nuclear and Cytoplasmic Extraction and Western Blot

Panc28 and L3.6pL cells (3 × 10⁵/well) were seeded in DMEM/Ham's F-12 medium in 6-well plates. After 24 h, cells were treated with different concentrations of metformin. Nuclear and cytoplasmic contents were extracted using nuclear and cytoplasmic extraction kit (Thermo Scientific) according to the manufacturer's protocol. Cells were lysed using high salt lysis buffer containing 50 mmol/liter HEPES, 0.5 mol/liter sodium chloride, 1.5 mmol/liter magnesium chloride, 1 mmol/liter EGTA, 10% (v/v) glycerol, 1% Triton X-100, and protease inhibitor mixture, 1:1000 (Sigma). Lysates were collected and vortexed every 15 min for 1 h, centrifuged at 20,000 × g for 10 min at 4 °C, and quantified with Bradford reagent. Western blot analysis was carried out by separating the proteins by SDS-PAGE at 120 V for 4 h. Proteins were then transferred onto PVDF membranes (Bio-Rad) by wet electroblotting, and membranes were blocked with 5% milk in TBST buffer containing 1.576 g/liter Tris, 8.776 g/liter sodium chloride, and 0.5 ml/liter Tween 20. The PVDF membranes were then probed with primary antibodies followed by incubation with horseradish peroxidase-conjugated secondary antibodies. Immobilon Western chemiluminescence substrates (Millipore) were used to develop the membrane, and images were captured on a Kodak 4000 MM Pro image station.

siRNA Interference Assay

Panc28, Panc1, and L3.6pL pancreatic cancer cells were seeded (1 × 10⁵/well) in 6-well plates in DMEM/Ham's F-12 medium supplemented with 2.5% charcoal-stripped FBS without antibiotic and left to attach for 24 h. Knockdown of Sp1, Sp3, and Sp4 along with iLamin as control was carried out using Lipofectamine 2000 reagent according to the manufacturer's instructions and as described previously (35). Small inhibitory RNAs were prepared by (Sigma-Aldrich).

Active Ras Detection Assay

Pancreatic cancer cells (Panc28 and L3.6pL, 3 × 10⁵/well) were seeded in Dulbecco's modified Eagle's medium/Ham's F-12 medium in 6-well plates. After 24 h, cells were treated with or without metformin (15 mm) for 36 h. Cells were harvested under nondenaturing conditions and rinsed with ice-cold PBS. Cells were lysed using lysis buffer. Affinity precipitation of active Ras was performed using active Ras detection assay kit according to the manufacturer's protocol. Cell lysates (500 μg) were treated with GTPγS or GDP to activate or inactivate Ras, which acts as a positive or a negative control, respectively. The lysates were then incubated with GST-Raf1-RBD fusion protein in glutathione resin (where RBD is Ras-binding domain). The eluted samples were electrophoresed and immunoblotted using Ras mouse monoclonal antibody.

Orthotopic Nude Mice Study and Immunohistochemical Staining

Male athymic nude mice (NCI-nu) (at least four per treatment group) were injected with suspensions of L3.6pL cells consisting of single cells with >90% viability directly into the pancreas as described previously (31, 32). Mice were treated (orally) with vehicle (control) or 250 mg/kg of metformin daily and sacrificed 4–5 weeks after injection, and body weights were recorded. Primary tumors in the pancreas were excised, measured, and weighed. For immunohistochemistry, tumor tissue specimens were fixed in 10% formaldehyde embedded in paraffin and sectioned into 3–5-mm-thick slices. In this study, we used slides generated from an orthotopic tumor model (31). Slides were deparaffinized with xylene, dehydrated with alcohol, and rinsed with distilled water. After antigen retrieval, sections were incubated with the following primary antibodies: anti-p-mTOR antibody (1:500, Cell Signaling, 2971), anti-IGF-1R antibody (1:100, Cell Signaling, 3027), overnight at 4 °C. Slides were then incubated with horseradish peroxidase-conjugated secondary antibody (30 min) and large volume HRP polymer (45 min) for sections incubated with p-mTOR and IGF-1R primary antibody, respectively. The substrate diaminobenzidine was added followed by hematoxylin counterstaining. After dehydration, the slides were soaked in xylene for 3–5 min. Immunostaining of IGF1 receptor was carried out using Lab Vision Autostainer (Thermo Scientific). Sections from normal pancreas were used as positive control. The negative control sections were incubated with antibody dilution buffer without primary antibody (31).

Metformin and Sp Down-regulation Inhibit mTOR Signaling

Fig. 1A shows that NVP-BE235, an mTOR and PI3 kinase inhibitor, also decreases proliferation of Panc28 and L3.6pL cells. Similar results were observed after treatment with 15 mm metformin, and L3.6pL cells were more sensitive than Panc28 cells to the growth inhibitory effects of both compounds. Treatment of Panc28 and L3.6pL cells with 50 nm NVP-BE235 decreased activation (phosphorylation) of both mTOR and AKT, and 5–20 mm metformin also inhibited phosphorylation of mTOR and AKT (Fig. 1B), demonstrating that like NVP-BE235, metformin blocks mTOR signaling in pancreatic cancer cells. Similar results were observed in Panc1 cells; however, in this cell line, metformin also decreased total mTOR and AKT protein levels. Immunostaining of pancreatic tumors and normal pancreatic tissue from an orthotopic mouse model using L3.6pL cells (31) also showed that the enhanced staining of phospho-mTOR in tumors was decreased in pancreatic tumors from mice treated with metformin (Fig. 1C). Moreover, Western blot analysis showed that phospho-AKT and phospho-mTOR decreased in tumors from mice treated with metformin as compared with vehicle-treated mice (Fig. 1D). The lack of a significant decrease in phospho-mTOR was due to a single animal outlier.

Knockdown of Sp transcription factors by RNAi did not decrease expression of mTOR protein in pancreatic cancer cells (31); however, there is evidence from RNAi studies that Sp transcription factors regulate activation of kinases such as AKT (39) and, therefore, we further investigated the effects of Sp silencing on inhibition of mTOR and downstream kinases. Fig. 2A shows that in Panc28 cells transfected with oligonucleotides targeting Sp1 (iSp1), Sp3 (iSp3), and Sp4 (iSp4), there was specific knockdown of the individual Sp proteins as described previously (35), and this was accompanied by decreased expression of phospho-mTOR and AKT but not total mTOR and AKT proteins. A similar approach was used for L3.6pL cells (Fig. 2B), and silencing of Sp1, Sp3, and Sp4 also decreased phosphorylated mTOR and AKT, indicating that all three Sp transcription factors regulated activation of both kinases. There was some variability in these responses and this was due, in part, to the efficiency of Sp1, Sp3, and Sp4 knockdown. A comparison of the effects of metformin with Sp silencing showed that both treatments decreased activation (phosphorylation) of mTOR-regulated S6RP and 4EBP gene products in Panc28 (Fig. 2C), L3.6pL (Fig. 2D), and Panc1 (Fig. 2E) cells. Metformin decreased activation of mTOR and mTOR-regulated genes in all three cell lines; however, there were some concentration-dependent differences in these effects. These results suggest that metformin-induced down-regulation of Sp1, Sp3, and Sp4 plays a role in inhibiting activation of the mTOR pathway.

Metformin and Sp Down-regulation Inhibit Lipogenic Genes

mTOR activation is also important for lipogenesis and enhances cleavage of sterol regulatory element-binding protein 1 (SREBP-1) to give the cleaved (active) transcription factor SREBP-1c, which in turn regulates expression of FAS (40–42). Treatment of Panc28 cells with metformin decreases SREBP-1c expression in both nuclear and cytosolic fractions (Fig. 3A). Similar results were observed in L3.6pL cells (Fig. 3B); however, the concentration-dependent decrease in p-SREBP-1c was also cell context-dependent. It has been previously reported that both SREBP-1c and FAS are Sp1-regulated genes in breast and colon cancer cells (43), and silencing of Sp1, Sp3, and Sp4 or all three proteins combined (iSp1/3/4) in Panc28 and L3.6pL decreased expression of phosphorylated SREBP-1c, whereas total SREBP-1c expression was not significantly decreased by the treatments (Fig. 3C). The results suggest that in pancreatic cancer cells, SREBP-1c is not directly regulated by Sp transcription factors, and the observed decrease in phospho-SREBP-1c is due to inactivation of mTOR by Sp silencing (Fig. 2, A and B). FAS protein expression is also decreased by metformin in pancreatic and other cancer cell lines (43), and we observed that silencing of Sp transcription factors in Panc28 and L3.6pL (Fig. 3D) cells also decreased FAS expression. Thus, the decreased expression of FAS by metformin is due to both direct effects from the loss of Sp proteins and also decreased activation of SREBP due to inhibition of mTOR. Knockdown of Sp1, Sp3, or Sp4 alone or in combination also induced poly(ADP-ribose) polymerase-2 (PARP) cleavage, a marker of apoptosis.

Metformin Inhibits Activation of mTOR through Down-regulation of IGF-1R

IGF-1R and other growth factor receptors are upstream activators of mTOR in pancreatic cancer cells (19, 22, 44), and IGF-1R is essential for proliferation of pancreatic cancer cells (44). It has been reported that IGF-1R expression is regulated by Sp1 in some cancer cell lines (37, 39, 45). Fig. 4A shows that metformin decreases IGF-1R expression in Panc28 and L3.6pL cells, and the role of IGF-1R in regulating the mTOR pathway was investigated by silencing of IGF-1R by RNAi in Panc28 and L3.6pL cells. Down-regulation of IGF-1R in these cells resulted in decreased phosphorylation of mTOR and AKT (Fig. 4B), and this was accompanied by decreased phosphorylation of S6RP and SREBP-1c and decreased expression of FAS protein (Fig. 4C). These results confirm the important role of IGF-1R in regulating mTOR and is consistent with inhibition of mTOR by metformin through down-regulation of IGF-1R. Transfection with iSp1, iSp3, iSp4, or iSp1/3/4 oligonucleotides also decreased IGF-1R expression in pancreatic cancer cells, confirming that IGF-1R is an Sp-regulated gene that is down-regulated by metformin (Fig. 4D). The overall contribution of IGF-1R to cell proliferation was confirmed by knockdown of IGF-1R by RNAi, which significantly decreased Panc28 and L3.6pL cell growth (Fig. 4E). Thus, inhibition of mTOR activation by metformin in pancreatic cancer cells is due to metformin-induced down-regulation of Sp1, Sp3, and Sp4 and the Sp-regulated IGF-1R gene.

Immunostaining and Western blot analysis of pancreatic tissue from orthotopic mouse model using L3.6pL cells (31) showed intense IGF-1R expression in tumor tissue sections from control animals (Fig. 4F, panels a–c). On the other hand, IGF-1R expression was reduced in the metformin-treated groups (Fig. 4F, panels d and e). However, the immunoreactivity was only minimally changed in tissues from one of the metformin-treated animals (Fig. 4F, panel f), and similar results were observed in Western blots of tumor lysates (data not shown). In normal pancreas, IGF-1R expression was moderate in ductal cells, and acinar cells were devoid of IGF-1R staining.

Metformin and Sp Down-regulation Target EGFR-Ras Signaling

The EGFR is essential for K-Ras signaling and subsequent Ras-dependent pancreatic cancer cell growth (37, 38). Like IGF-1R, knockdown of EGFR by RNAi also decreased proliferation of L3.6pL and Panc28 cells (Fig. 5A), and this is consistent with the role of the EGFR-Ras pathway in pancreatic cancer cell proliferation (37, 38). It has previously been reported that metformin decreases EGFR expression in pancreatic cancer cells (46), and similar results were observed in Panc28 and L3.6pL cells treated with metformin (Fig. 5B). The importance of metformin-mediated down-regulation of Sp transcription factors in decreasing EGFR was confirmed by RNAi where knockdown of Sp1 or all three Sp proteins (iSp1/3/4) significantly decreased EGFR protein levels in Panc28 and L3.6pL cells (Fig. 5C). Down-regulation of EGFR by Sp knockdown was not observed in cells transfected with iSp4, and silencing Sp3 decreased EGFR only in L3.6pL cells, showing that Sp1 is the major regulator of EGFR expression; this has been reported in other cancer cell lines (39, 47).

The effects of metformin and Sp silencing on Ras activity were determined using an active Ras detection assay. Treatment of Panc28 and L3.6pL cells with metformin or transfection of these cells with iSp1/3/4 (combined) decreased active Ras-GTP levels (Fig. 5D), and similar results were observed after silencing EGFR (iEGFR) by RNAi in Panc28 and L3.6pL cells (Fig. 5E). Thus, metformin-induced down-regulation of Sp1, Sp3, and Sp4 and Sp-regulated IGF-1R and EGFR genes resulted in inhibition of both the mTOR and the Ras pathways in pancreatic cancer cells, and metformin also decreased IGF-1R and EGFR expression in pancreatic tumors from an orthotopic mouse model.

Metformin-induced down-regulation of Sp transcription factors in pancreatic cancer cells was dependent on induction of mitogen-activated protein kinase phosphatase 1 (MKP1) and MKP5, and this response was blocked by the phosphatase inhibitor sodium orthovanadate (SOV) (31). Results in Fig. 6A show that SOV also blocks metformin-mediated inhibition of mTOR and AKT phosphorylation and also blocks phosphorylation of 4EBP and S6RP (Fig. 6B) in Panc28 and L3.6pL cells. These results confirm that an important underlying mechanism of action of metformin in pancreatic cancer cells was due to down-regulation of Sp transcription factors and the Sp-regulated IGF-1R and EGFR genes, which results in the inhibition of mTOR and Ras pathways as illustrated in Fig. 6C.