NS 105

Glutamate increases pancreatic cancer cell invasion and migration via AMPA receptor activation and Kras-MAPK signaling

Alexander Herner1*, Danguole Sauliunaite1*, Christoph W. Michalski1*, Mert Erkan1, Tiago De Oliveira1, Ivane Abiatari1, Bo Kong1, Irene Esposito2, Helmut Friess1 and Jo¨rg Kleeff1

Abstract

Glutamate has been implicated in tumorigenesis through activation of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors (AMPAR). However, the function of a glutamate-to-AMPAR signal in pancreatic ductal adenocarcinoma (PDAC) has remained elusive. We now show that glutamate-mediated AMPA receptor activation increases invasion and migration of pancreatic cancer cells via activation of the classical MAPK pathway. Glutamate levels were increased in pancreatic cancer accompanied by downregulation of GluR subunits 1, 2, and 4. In pancreatic cancer precursor lesions, pancreatic intraepithelial neoplasia (PanIN), GluR1 subunit levels were strikingly and step-wise increased but its expression was rare in PDAC. Pharmacological inhibition or RNAi-mediated suppression of GluR1 or GluR2 did not affect cancer cell growth but significantly decreased invasion. In a K-ras wildtype cell line, AMPA receptor activation enhanced K-ras activity and—further downstream—phosphorylation of p38 and of p44/42. Preemptive blockade of AMPA receptors in a mouse model of pancreatic cancer inhibited tumor cell settling. AMPA receptor activation thus not only activates MAPK signalling but also directly increases activity of K-ras. Glutamate might serve as a molecular switch that decreases the threshold of K-ras-induced oncogenic signalling and increases the chance of malignant transformation of pancreatic cancer precursor lesions.

Key words: pancreatic cancer, AMPA, AMPA receptor, glutamate *A.H., D.S., and C.W.M. contributed equally to this work.

Introduction

Pancreatic ductal adenocarcinoma (PDAC) is one of the most aggressive human malignancies, with exceptionally low survival rates (5-year survival rate less than 5%).1 Although the cell of origin of PDAC is controversially discussed,2 pancreatic intraepithelial neoplasias (PanIN) of different grades (1A, 1B, 2, 3) in which increasing numbers of molecular alterations are found (i.e., mutations in the Kras, Smad4 and p53 genes), are believed to belong to the multistep progression model of pancreatic cancer.3,4 Neurotransmitters and neuropeptides have been shown to play a role in pancreatic carcinogenesis5–11 and in pancreas-associated diseases8,12–16 however, the role of excitatory neurotransmitters and particularly of glutamate in pancreatic cancer is not well defined. Glutamate activates metabotropic receptors (mGluR; G protein-coupled receptors) and the ionotropic (iGluR) receptors N-methyl-D-aspartate (NMDA), alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and kainate17,18 receptors. AMPA receptors are heteromeric and are assembled from the four subunits GluR1-4 in different combinations. The presence of GluR2 subunit determines AMPA receptor impermeability to Ca2þ.17,19,20 Since AMPA receptor activation regulates differentiation, proliferation and migration of embryonic stem cells,21–24 it has been hypothesized that modulation of AMPA receptor-mediated signals might be involved in carcinogenesis, particularly because it has also been shown that AMPA signals via the MAPK pathway.25–27 This hypothesis has subsequently been proven for some tumor entities such as astrocytoma, glioblastoma, breast carcinoma, lung carcinoma, colon adenocarcinoma, and prostate carcinoma.26,28–31 An extensive search of online available databases on gene expression in pancreatic cancer (www.oncomine.com; www.pancreasexpression.org; http://cgap.nci.nih.gov/SAGE) revealed de-regulation of GluR2 (gene: GRIA2) in cancer as compared to normal pancreas (NP). Since we have recently characterized expression and function of cannabinoids as exemplary inhibitory neurotransmitters in pancreatic diseases, we now set out to define the role of the glutamate system in pancreatic carcinogenesis with an emphasis on AMPA receptors.

Materials and Methods

Sym2206 and S()-50-fluorowillardiine were purchased from TOCRIS Cookson (Ellisville); AMPA, from Biomol (Hamburg, Germany); rabbit polyclonal anti-GluR1, mouse monoclonal anti-GluR2 were purchased from Chemicon International (Temecula, CA) rabbit anti-GluR4 was purchased from Upstate, Lake Placid, NY; anti-p38 MAP kinase, anti-phospho p38-MAP kinase (Thr180/Tyr182), anti-p44/42 MAP kinase, anti-phospho p44/42 MAP kinase (Thr202/ Tyr204) antibodies were purchased from Cell Signaling Technology (Danvers, MA); secondary antibodies—anti rabbit— and anti mouse–HRP-labelled polymer (ready to use), normal rabbit IgG (15 g/l), mouse IgG1, mouse IgG2a, DAB Chromogen System were purchased from DAKO, Hamburg, Germany; ECL-anti-rabbit IgG HRP-linked secondary antibody, from GE Healthcare, UK; ECL, from Amersham Life Science (Bucks, UK); cell culture medium—RPMI 1640, Penicillin/ Streptomycin, 0.25%Trypsin-EDTA were purchased from GIBCO, Invitrogen GmbH (Karlsruhe, Germany) fetal bovine serum, from PAN Biotech GmbH (Aidenbach, Germany). Ethanol, Methanol, H2O2, Naþ-Citrate, BSA, EDTA, SDS, TRIS, NaCl, Formalin, Glycine, Tween-20, TritonX-100, Roticlear were purchased from Carl Roth GmbH (Karlsruhe, Germany). DMSO, Hepes and CaCl2 were purchased from Sigma Aldrich (St. Louis, MO) Mayer’s Hematoxyllin was purchased from Merck (Darmstadt, Germany).

Patients and tissue sampling

Tissue samples were collected from patients during pancreatic resections for PDAC (n ¼ 60) or CP (n ¼ 10). Normal pancreatic tissue samples were obtained through an organ donor procurement program, whenever there was no suitable recipient for pancreas transplantation (n ¼ 10). Pancreatic tissues were immediately snap frozen at 80C or formalin-fixed and paraffin-embedded. The use of human tissue for the analysis was approved by the local ethical committee (University of Heidelberg, Germany) and written informed consent was obtained from the patients prior to surgery.

(Quantitative) real time polymerase chain reaction [(Q)RT-PCR]

mRNA and cDNA were prepared using reagents and equipment from Qiagen (Hilden, Germany), following the manufacturer’s instructions. Real time PCR was carried out on a Mastercycler (Eppendorf) Quantitative RT-PCR was carried out on a LightCycler480 (Roche Diagnostics). QuantiTect primer assays (Qiagen) for GluR1-GluR4 subunits were used (for sequences see www.qiagen.com).

Glutamate concentration measurement

Gluamate concentration in the eight pancreatic cancer cell supernatants was determined according to the protocol as previously described.32

Immunohistochemistry

Immunohistochemistry was performed on formalin-fixed, paraffin-embedded tissue sections as previously described.8,15,16,33 The PanIN tissue array contained sections from PanIN1, 2, and 3. Semi-quantitative analysis of PanIN

AMPA receptors in pancreatic cancer

staining was performed as previously described.15 The antibodies and dilutions were as follows: rabbit polyclonal antiGluR1 (200 lg/ml; 1:300), mouse monoclonal anti-GluR2 and rabbit monoclonal anti-GluR4 [1.15 mg/ml; 1:5000; 1:200; all from Chemicon International (Temecula CA)].

Semiquantitative evaluation of AMPA receptor levels in human tissue specimens

Immunoreactivity of AMPA receptors in PanIN structures, cancer cells and/or nerves was quantitatively evaluated according to intensity and area as previously described15: the staining intensity of PanIN structures and cancer cells was recorded as ‘‘no staining’’ (0), ‘‘weak staining’’ (1), moderate staining (2)’’ or ‘‘strong staining (3)’’. The area of stained cells was recorded as <33% (1), 33-66% (2), or >66% (3) of all cancer cells. These numbers were then multiplied resulting in a score of 0–9. Regarding intrapancreatic nerves, only the staining intensity (0–3, as described above) was analyzed due to the generally low number of nerves in pancreatic cancer tissue specimens.

RNAi

Synthetic siRNA oligonucleotides for GluRs were purchased from Qiagen (Hilden, Germany), prepared and stored according to the manufacturer’s instructions. For silencing of GluR1/2 subunit two different RNAi molecules were tested and the higher effect RNAi was chosen. Human GluR1 RNAi (sense 5¢-CCAUGAAGGUGGGAGGUAATT-3¢, antisense 5¢-UUACCUCCCACCUUCAUGGTG-3¢) and GluR2 RNAi (sense 5¢-CGUAUGUUAUGAUGAAGAATT-3¢; antisense 5¢-UUCUUCAUCAUAACAUACGGA-3¢) were used. Control siRNA sequence was UUCUCCGAACGUGUCACGU. Cells were grown to 70% confluency under standard growth conditions. For siRNA transfections, HiPerfect transfection reagent (Qiagen, Hilden, Germany) was used according to the manufacturer’s instructions. The final concentration of the control and the specific oligonucleotides was 10 nM. The efficacy of the RNAi was analyzed by RT-PCR 24 hr after transfection.

Cell viability assay

Cell viability assays were performed as described previously.33,34 In 96 well-plates, 5.000 cells/well were seeded, were grown for 24 hr at 37C, 5%CO2 humidified atmosphere and were exposed to Sym2206 at concentrations of 6.25, 12.5, 25, 50, 75, and 100 lM and to S–(-)-5’-fluorowillardiine at concentrations of 1, 10, 50, 75, and 100 lM for 24 hr. To evaluate the effect of GluR1 and GluR2 RNAi on cell growth, cells were reseeded into 96-well plates 24 hr after RNAi and grown for further 24-48-72-96 hr. Twenty-four hours after reseeding (and as indicated in the results part), some cells were treated with 100 lM of AMPA or the appropriate control. After the indicated time points, yellow tetrazolium salt 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium bromide (MTT; 5 mg/ml in PBS; Sigma Aldrich, St. Louis, MO) was added (50 lg/well) and cells were incubated with MTT for 4 hr. The MTT was subsequently solubilized in acidic isopropanol and was quantified spectrophotometrically by measuring the optical density at 570 nm wave length. The GI50 (the concentration required to achieve 50% growth inhibition) for SYM or FW was calculated by using the formula 100 (T – T0)/(C – T0), where T is the optical density after X h of exposure to the drug, T0 the adsorption at time zero, and C the control after X hr. Alternatively, the growth after RNAi was calculated as percentage of time zero. All experiments were performed in triplicates and repeated at least three times.

Immunoblot assays

Immunoblot assays were performed as previously described.8,14 Briefly, Su86.86 cells were treated with the drugs according to the experimental procedures and were then lysed in an ice cold lysis buffer (5M NaCl, 1M TrisHCl, 10%Triton X-100, 100 mM EDTA, 200 mM sodium orthovanadate, 200 mM sodium fluoride, 200 mM sodium pyrophosphate, glycerol) supplemented with an EDTA-free mini protease inhibitor cocktail tablet (Roche Diagnostics, Mannheim, Germany). The primary antibodies were used as follows: anti-p38 MAP kinase, anti-phospho-p38 MAP kinase (Thr180/Tyr182), anti-phospho-p44/42 MAP kinase (Thr202/ Tyr204), anti-p44/42 MAP kinase diluted 1:1000 inTBS/ 5%BSA. The blots were subjected to densitometric analysis as previously described.34

Wound healing assay

Su86.86 cells were seeded into petri-dishes in RPMI medium containing 10% FBS and 1% penicillin/streptomycin, and were grown at 37C, 5%CO2 humidified conditions until 90– 95% confluency was reached. A scratch was made with a sterile 10 ll pipette tip and washed twice with sterile PBS. The cells were treated with 10 lM of S–(–)-5¢fluorowillardiine or 10 lM of Sym2206 using NaOH (1:5000) or DMSO (1:10.000) as appropriate controls, respectively. The scratch was photographed using a digital camera (Carl Zeiss Axiocam MRm, Germany) at a magnification of 10 at 2 preselected time points (0 and 24 hr) to observe migration of the cells. Migrated cells were counted using computerized imaging analysis.

Invasion assay

To assess the effects of SYM and FW and to evaluate the influence of specific MAPK antagonists on Su86.86 invasion, BD BioCoat Matrigel Invasion Chambers (BD Biosciences, San Jose, CA) were prepared according to the manufacturer’s instructions. In 500 ll RPMI culture medium containing 0.5% FBS and 10 lM of Sym2206 or control (DMSO 1:10.000), 5 104 cells/ml were seeded. For the specific inhibitor invasion assay, cells were pre-treated for 1 hr with specific p44/42 and p38 inhibitors (U0126 and SB203580 at 10 lM, respectively) alone or in combination and were subsequently stimulated with 10 lM of S–(–)-5¢-fluorowillardiine. Twenty-four hours after treatment, invaded cells were fixed, stained and counted under a light microscope. The invasion index was calculated as the ratio of the number of invaded cells to the number of invaded control cells. All assays were repeated 5 times.

ELISA

Equal amounts (4 105/well) of Su86.86 cells were seeded and grown in 6-well plates in 10% FBS-containing growth medium at 37C, 5%CO2, humidified conditions until adherent. Then, the cells were treated with 10 lM of Sym2006 in 0.5%FBS containing growth medium and were incubated for 24 hr at the same growth conditions. Thereafter, supernatants were analyzed by matrix metallo-proteinase-2 (MMP-2) ELISA according to the manufacturer’s instructions (BD Bioscience, Heidelberg, Germany).

p44/42 and p38 inhibition assay

Su86.86 cells were seeded into 6-well plates in 10% FBS containing growth medium at 37C, 5% CO2, humidified conditions and grown until adherent. The growth medium was then replaced by 0.5% FBS containing growth medium and the cells were incubated at the same conditions with FW and/or SYM (10 lM) for 24 hr. When indicated, cells were preincubated (for 1 hr) with SB203580 (SB, p38 inhibitor, 5 lM (when used in combination with U0126) or 10 lM) and/ or U0126 (p44/42 inhibitor 5 lM (when used in combination with SB) or 10 lM). After 24 hr incubation with the drugs, the supernatants were collected and the amount of MMP-2 was evaluated by ELISA according to the manufacturer’s protocol. The cells were lysed and levels of p44/42/phospho-p44/ 42 and p38/phospho-p38 were evaluated by immunoblot analysis.

Ras activation assay

Su86.86 and BxPC3 cells were seeded into 6 cm petri-dishes at 70–80% confluency in complete growth medium and grown until adherent. Then, cells were treated with 100 lM of AMPA or appropriate control for 24 hr and subjected to a Ras activation assay using the Ras activation assay kit (Milipore, Temecula, CA) according to the manufacturer’s instructions.

Pathway array

Su86.86 were treated with 100 lM of AMPA or the appropriate control for 12 hr. mRNA and cDNA were prepared using reagents and equipment from Qiagen (Hilden, Germany) and were subjected to the RT2 Profiler PCR array (SABiosciences, Frederick) according to the manufacturer’s instructions.

In vivo model of pancreatic cancer

All animal procedures were performed according to local ethical guidelines. In 200 ll of sterile PBS, 106 Su86.86 cells were resuspended and were injected into the subcutaneous issue bilaterally at the sites behind the anterior forelimb of 4-week old athymic nude mice. One day after the injection of tumor cells, treatment with the AMPA receptor antagonist SYM (5 mg/kg, n ¼ 5; s.c. injection once per day during 2 weeks) or DMSO (1:100, control group, n ¼ 5; same injection protocol as for SYM) was started. After 2 weeks, the mice were sacrificed and visible tumors were collected.

Statistical analysis

Statistical analyses were performed using the GraphPad Prism 4 Software (GraphPad Software). Experimental results are expressed as mean þ/- SEM unless indicated otherwise. The level of significance was set at p < 0.05. Results Analysis of glutamate levels in fresh frozen human NP, chronic pancreatitis (CP) and PDAC tissues demonstrated a striking increase of glutamate in CP and PDAC samples (Fig. 1a, p < 0.0001). Because GluR2 receptor subunit was found to be de-regulated in PDAC, we hypothesized that the increase in glutamate levels induces a pro-invasive and antiapoptotic signal via activation of AMPA receptors. To this end, we determined AMPA receptor expression in pancreatic tissues: while there was a trend toward increased GluR (1–4) expression levels in CP, GluR1, 2, and 4 were down-regulated in bulk pancreatic cancer tissues (Fig. 1b, p ¼ 0.0084 for GluR1, p ¼ 0.0423 for GluR2, and p ¼ 0.0025 for GluR4). While in NP, acinar cells (which form more than 80% of the total cellular mass of the organ) expressed GluRs at a low to medium level, pancreatic cancer cells (which form less than 30% of the total tumor mass) were strongly immunopositive for GluR2. These findings might explain the discrepancies between RT-PCR/cDNA microarray (down-regulation of GluR2) and immunohistochemistry results. To determine the ‘‘source’’ of the decrease in GluR subunit transcripts, we performed extensive immunohistochemical analyses on pancreatic tissue sections using antibodies against GluR1, GluR2, and GluR4 subunits. In NP, the GluR1 subunit was found to be expressed mostly in islets and some faint staining was rarely seen at the apical membrane of duct cells (Fig. 1c and inset, NP), whereas with an increasing PanIN grade, increasing staining intensities were found (Fig. 1c and insets, PanIN1, 2 and 3); we thus performed GluR subunit stainings on PanIN tissue microarrays and evaluated these semi-quantitatively, confirming the gradual increase in immunoreactivity (Fig. 1c, graph). In pancreatic cancer cells, GluR1 subunit expression was again rarely found with a presumed nuclear retention if at all visible (Fig. 1c, PDAC). GluR2 subunit staining was seen at the basolateral membrane of acinar cells of NP tissues (Fig. 1d, NP). Both in low grade and high grade PanINs, as well as in pancreatic cancer, strong GluR2 subunit expression was found on the membranes of the precursor and the malignant cells (Fig. 1d, PanIN1, PanIN2, and PDAC). While a few PanIN1 localized GluR2 to the basolateral membrane of the columnar-shaped AMPA receptors in pancreatic cancer cells, its expression was mostly seen on all membrane compartments (Fig. 1d PanIN1 and inset). Although in some PanINs, only a minority of the cells were GluR2-positive, nearly all of the cancer cells revealed strong GluR2 immunoreactivity. Thus, no semi-quantitative evaluation was performed. The GluR4 subunit was found to be expressed in islets in the NP, CP, and PDAC tissue samples. Few CP tissue samples showed GluR4 subunit expression in stromal components (i.e., fibroblasts), accompanied by GluR4 subunit expression in infiltrating immune cells. In PDAC tissue samples, a polymorphic GluR4 expression pattern is seen, with a trend towards increased expression in samples with acinarto-ductal metaplasia as well as in less differentiated cancer samples (Fig. 1e). Silencing of AMPA receptors as well as their pharmacological inhibition decreases invasion and migration of pancreatic cancer cells To determine the function of glutamate-AMPA receptor signalling in advanced pancreatic cancer—as reflected by high genetic instability, epigenetic de-regulation and (re-)activation of developmentally important and active pathways—we chose eight well established and characterized pancreatic cancer cell lines (i.e., known mutation status of K-ras and p53, morphology and culture conditions). These cells were subjected to enzymatic measurement of glutamate levels as well as to conventional RT-PCR using oligonucleotides for the amplification of mRNAs encoding GluR subunits 1–4. Glutamate was released by all cell lines with the highest amounts found in the supernatants of Su86.86 cells (Fig. 2a); while all GluR subunits (1–4) were expressed by the (K-ras mutated) cell line Su86.86, the other tested cell lines expressed fewer or no transcripts of these receptors (Fig. 2b). Thus, for the subsequent experiments, Su86.86 cells were used (unless otherwise indicated). Interestingly, the amounts of secreted glutamate in cancer cell supernatants were much lower compared to the freshly prepared pancreatic cancer tissue. This might probably be explained by the tissue celularity content. Other cells than tumor cells (fibroblast, nerves) may also contribute to the overall glutamate levels. Furthermore, we hypothesized from these experiments that there might be an autocrine loop of glutamate-toAMPAR signalling in Su86.86 cells. Thus, we suppressed GluR1 or GluR2 receptor subunit expression in Su86.86 cells using specific RNAi. In contrast to what would be expected from the known function of Glu receptors, neither silencing of the GluR1 nor the GluR2 receptor subunit had any effect on cancer cell growth (Fig. 2c, black vs. green dots/lines); in addition, incubation of the cells with the AMPA receptor ligand AMPA did not affect proliferation (Fig. 2c, red vs. blue dots/lines). However, both GluR1 and GluR2 RNAi reduced the invasiveness of Su86.86 cells in a Boyden chamber assay by 50% (Fig. 2d; white vs. black bar); accordingly, stimulation with AMPA doubled the number of invaded cells (Fig. 2d; p ¼ 0.0143 and p ¼ 0.0272, respectively; white bars). Furthermore, knockdown of either the GluR1 or GluR2 receptor subunit expression was sufficient to inhibit the proinvasive capacity of AMPA (Fig. 2d, black bars), which underlines the specificity of AMPA toward its receptor [receptor subunit(s)]. To corroborate these findings and to further confirm the specificity of the glutamate-to-AMPAR signal, we used the small molecule AMPA receptor agonist S(–)-50-fluorowillardiine (FW) and the AMPA receptor antagonist Sym2206 (SYM). Comparable to the experiments using AMPA and/or RNAi, neither FW nor SYM had any effect on proliferation of Su86.86 cells (both in MTT assays (Fig. 2e and 2f, left panels) and as determined by Annexin/PI stainings for apoptotic cells (data not shown); however, FW slightly increased migration (scratch assay analysis: (Fig. 2e middle panel) and doubled the number of invaded cells in the Boyden chamber assay (Fig. 2e, right panel). Accordingly, AMPA receptor antagonism reduced cancer cell migration and invasion (Fig. 2f middle and right panels, p ¼ 0.0049). Because invasion through the ‘‘basement membrane-like’’ matrigel-coated chamber is—among others—mediated by increased levels of the MMP-2, we determined its concentration in cell culture supernatants of FW- and SYM-treated Su86.86 cells. In line with the results from the invasion assays, FW increased whereas SYM reduced MMP-2 protein levels (Fig. 2e and 2f right panels, p ¼ 0.0161 and p ¼ 0.0366, respectively). AMPA receptor activation increases K-ras activity and signals into the MAPK pathway To determine the effector pathways activated by the AMPA receptors, Su86.86 cells were stimulated with AMPA for 12 hr. The isolated mRNA was subjected to a ‘‘signal transduction’’ PCR array revealing downregulation of ‘‘stress’’ pathway genes (Fig. 3a, green bars) but upregulation of TGFbeta, Jak-Stat and Ca2þ/PKC pathway genes (Fig. 3a, red bars). These results lead us to hypothesize that—apart from the presumed MAPK pathway induction mediated by GluR activation (see below)—there might also be a direct signal into ras-GTP induction. This assumption is supported by recent reports on AMPA signal transduction in neurons, which have demonstrated that AMPA activation leads to activation of small GTPases of the ras family and also to Ca2þ-mediated Ras-GRF/Ras-GRP activation which again induce rasGTP35,36; thus, K-ras activity was determined in a not K-rasmutated cell line (BxPC3; expressing GluR2). Consistent with our hypothesis, K-ras was strikingly activated following stimulation with AMPA (Fig. 3b) in K-ras nonmutated cells. At the same time, phosphorylation of p44/42 was induced following stimulation with FW (Fig. 3c and densitometry), which was reversible by pre-incubation with SYM (Fig. 3c; p ¼ 0.0319). These results demonstrated a direct effect of a glutamate-to-AMPAR receptor signal into the K-ras-MAPK pathway in non-Kras-mutated pancreatic cancer cells. Inhibition of p44/42 and p38 pathways abrogates GluR-mediated pancreatic cancer cell invasion To substantiate the finding that the MAPK pathways are crucial for AMPA receptor-mediated cancer cell invasion, we performed invasion assays using FW and specific inhibitors: U0126 (p44/42 inhibitor) and SB203580 (SB, p38 inhibitor). Comparable to the findings above, incubation of Su86.86 cells with FW enhanced p38 and p44/42 phosphorylation (Fig. 3d, left panel; completely abrogated by preincubation with the respective inhibitors or a combination hereof) and significantly increased the number of invaded cells (Fig. 3d, right panel) as well as the amount of secreted MMP-2 (Fig. 3e). Following pre-incubation with either U0126 or SB, these increases were completely blocked; additionally, preincubation of the cells with a combination of U0126 and SB suppressed the number of invaded cells and the amount of secreted MMP-2 far below the base-line (Fig. 3e right panels). AMPA receptor antagonism inhibits pancreatic tumor growth in vivo Having thus found in vitro evidence that a glutamate-AMPA signal decreases pancreatic cancer cell invasiveness and migration, we subcutaneously injected athymic nude mice with PDAC cells (Su86.86) and performed ‘‘tumor prevention’’ experiments. Two weeks after tumor cell injection and treatment with SYM and DMSO as a control, subcutaneous tumors were found in four mice in the control group (n ¼ 5) (size 1 2 mm–2 2 mm), whereas in the SYM treated group (n ¼ 5), no tumors were visible (Fig. 3f). This is of importance because it demonstrates that abrogation of glutamate-AMPA signalling seems to inhibit the settling of the injected tumor cells in the subcutaneous mouse tissue, which might be due to inhibition of an autocrine glutamate signal. Discussion Taken together, our results unravel the glutamate-AMPA axis as an important signal transducer towards a more aggressive and invasive pancreatic cancer phenotype. In vivo, both autocrine and paracrine mechanisms may contribute to the perpetuation of a glutamate signal, which could also be responsible for the immune response and cytokine release observed in pancreatic cancer. This assumption would be in line with the findings from our pathway array where ‘‘intrinsically,’’ AMPA-induced upregulation of, that is, IL-1a, PECAM and myc also suggested an AMPA-induced stress response. Importantly, the functional and morphological findings of this study suggest that in contrast to the NP, the increased tissue glutamate directly activated AMPA receptors on PanIN and/or cancer cells to switch on invasive and migratory programs probably via activation of the K-ras/MAPK cascade (Fig. 4). This is of particular interest since it has been shown by a number of reports that there is a link between AMPA and K-ras signalling,36–38 however, in this case, the link does not seem to be calcium-dependent because RNAi of both a calcium-permeable (GluR1) and calcium-impermeable (GluR2) AMPA subunit was associated with significantly decreased invasiveness. Furthermore, the direct increase of Kras activity by glutamate-mediated AMPA receptor activation may contribute (at least in not K-ras-mutated cancer cells or in precursor lesions) to a lowering of the mitogenic threshold necessary for malignant transformation (Fig. 4). First of all, this is particularly important in the ‘‘early’’ steps of carcinogenesis in which the level of K-ras activity seems to determine whether or not a cell may undergo malignant transformation.39 Secondly and because it has recently been demonstrated that inflammation might be instrumental in pancreatic carcinogenesis and progression40–42 or, depending on the mouse model used, was even a prerequisite for K-rasdependent oncogenic transformation,43 the (over-)activation of pathways (i.e., inflammation response signals) which directly increase K-ras activity might contribute to the loss of epithelial homeostasis and integrity. AMPA receptor activation may constitute such a molecular switch through increases in glutamate levels since these were found ‘‘already’’ in CP tissues, where, in contrast to PDAC, AMPA receptor expression was as high as and partially even higher than in the NP. Thus, AMPA receptor activation might be important in the transformation of PanIN to cancer and in the growth of K-ras wild type tumors while in later stages, K-ras independent (but mutated) cancers, high K-ras activity levels may probably not be further increased (as also demonstrated in Panc1 cells, Fig. 3b). This assumption is supported by a recent report demonstrating that one class of pancreatic cancer cells required K-ras to maintain viability while others were K-ras-independent.44 The more ‘‘mesenchymal’’ the cancer cells were, the less they depended on K-ras; at the same time, K-ras-independency correlated with resistance to apoptosis following K-ras inhibition. This observation supports the conclusion that in ‘‘later stage’’ (i.e., more mesenchymal) cancers, a pro-proliferative capacity is exchanged for a more robust invasive/migratory phenotype. Such a supposition would be in line with our finding (which is in contrast with reports from other tumors) that AMPA receptor activation does not increase proliferation of (probably ‘‘late’’ stage, because in vitro cultured/passaged) pancreatic cancer cells but significantly enhanced their invasion and migration. However, our findings are derived from in vitro experiments and may thus only be an imperfect picture of the in vivo situation. This is particularly important when taking into consideration the completely different effects of activation of calcium-permeable versus –impermeable AMPA receptors (i.e., those containing a GluR2 subunit).45 An adequate dissection of the contribution of the GluR subunits to the observed effects on pancreatic cancer cells and to putative effects in carcinogenesis would only be possible using conditional gene targeting in mice (i.e., using the well-established genetically engineered mouse model of pancreatic cancer which uses expression of an oncogenic Kras from its endogenous locus specifically in the pancreas; plus/ minus deletion of the respective GluR subunits). Such experiments will have to be a major part of further analyes of the glutamate/AMPA system in pancreatic diseases. Conclusion In conclusion, the glutamate-to-AMPAR signal shown in this report might serve as a molecular switch that decreases the threshold of Kras-induced oncogenic (in this case, invasive/migratory; Figs. 2 and 4) signalling and thus increases the chance of malignant transformation of pancreatic cancer precursor lesions. References 1. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T, Thun MJ. Cancer statistics. CA Cancer J Clin 2008;58:71–96. 2. Esposito I, Seiler C, Bergmann F, Kleeff J, Friess H, Schirmacher P. Hypothetical progression model of pancreatic cancer with origin in the centroacinar-acinar compartment. Pancreas 2007;35:212–7. 3. Hruban RH, Goggins M, Parsons J, Kern SE. Progression model for pancreatic cancer. Clin Cancer Res 2000;6:2969–72. 4. Maitra A, Adsay NV, Argani P, IacobuzioDonahue C, De Marzo A, Cameron JL, Yeo CJ, Hruban RH. Multicomponent analysis of the pancreatic adenocarcinoma progression model using a pancreatic intraepithelial neoplasia tissue microarray. Mod Pathol 2003;16:902–12. 5. Friess H, Ding J, Kleeff J, Fenkell L, Rosinski JA, Guweidhi A, Reidhaar-Olson JF, Korc M, Hammer J, Buchler MW. Microarray-based identification of differentially expressed growth- and metastasis-associated genes in pancreatic cancer. Cell Mol Life Sci 2003;60:1180–99. 6. Friess H, Zhu Z, Liard V, Shi X, Shrikhande SV, Wang L, Lieb K, Korc M, Palma C, Zimmermann A, Reubi JC, Buchler MW. Neurokinin-1 receptor expression and its potential effects on tumor growth in human pancreatic cancer. Lab Invest 2003;83:731–42. 7. Hartel M, di Mola FF, Selvaggi F, Mascetta G, Wente MN, Felix K, Giese NA, Hinz U, Di Sebastiano P, Buchler MW, Friess H. Vanilloids in pancreatic cancer: potential for chemotherapy and pain management. Gut 2006;55:519–28. 8. Michalski CW, Maier M, Erkan M, Sauliunaite D, Bergmann F, Pacher P, Batkai S, Giese NA, Giese T, Friess H, Kleeff J. Cannabinoids reduce markers of inflammation and fibrosis in pancreatic stellate cells. PLoS ONE 2008;3:e1701. 9. Mijatovic T, Gailly P, Mathieu V, De Neve N, Yeaton P, Kiss R, Decaestecker C. Neurotensin is a versatile modulator of in vitro human pancreatic ductal adenocarcinoma cell (PDAC) migration. Cell Oncol 2007;29:315–26. 10. Schuller HM, Al-Wadei HA, Majidi M. GABA B receptor is a novel drug target for pancreatic cancer. Cancer 2008;112:767–78. 11. Takehara A, Hosokawa M, Eguchi H, Ohigashi H, Ishikawa O, Nakamura Y, Nakagawa H. Gamma-aminobutyric acid (GABA) stimulates pancreatic cancer growth through overexpressing GABAA receptor pi subunit. Cancer Res 2007;67: 9704–12. 12. Carracedo A, Gironella M, Lorente M, Garcia S, Guzman M, Velasco G, Iovanna JL. Cannabinoids induce apoptosis of pancreatic tumor cells via endoplasmic reticulum stress-related genes. Cancer Res 2006;66:6748–55. 13. Fogli S, Nieri P, Chicca A, Adinolfi B, Mariotti V, Iacopetti P, Breschi MC, Pellegrini S. Cannabinoid derivatives induce cell death in pancreatic MIA PaCa2 cells via a receptor-independent mechanism. FEBS Lett 2006;580:1733–9. 14. Michalski CW, Laukert T, Sauliunaite D, Pacher P, Bergmann F, Agarwal N, Su Y, Giese T, Giese NA, Batkai S, Friess H, Kuner R. Cannabinoids ameliorate pain and reduce disease pathology in ceruleininduced acute pancreatitis. Gastroenterology 2007;132:1968–78. 15. Michalski CW, Oti FE, Erkan M, Sauliunaite D, Bergmann F, Pacher P, Batkai S, Muller MW, Giese NA, Friess H, Kleeff J. Cannabinoids in pancreatic cancer: correlation with survival and pain. Int J Cancer 2008;122:742–50. 16. Michalski CW, Shi X, Reiser C, Fachinger P, Zimmermann A, Buchler MW, Di Sebastiano P, Friess H. Neurokinin-2 receptor levels correlate with intensity, frequency, and duration of pain in chronic pancreatitis. Ann Surg 2007;246:786–93. 17. Hollmann M, Heinemann S. Cloned glutamate receptors. Annu Rev Neurosci 1994;17:31–108. 18. Kim CH, Lee J, Lee JY, Roche KW. Metabotropic glutamate receptors: Phosphorylation and receptor signaling. J Neurosci Res 2007. 19. Rao VR, Finkbeiner S. NMDA and AMPA receptors: old channels, new tricks. Trends Neurosci 2007;30:284–91. 20. Seeburg PH. The TiPS/TINS lecture: the molecular biology of mammalian glutamate receptor channels. Trends Pharmacol Sci 1993;14:297–303. 21. Behar TN, Scott CA, Greene CL, Wen X, Smith SV, Maric D, Liu QY, Colton CA, Barker JL. Glutamate acting at NMDA receptors stimulates embryonic cortical neuronal migration. J Neurosci 1999;19: 4449–61. 22. Ikonomidou C, Bosch F, Miksa M, Bittigau P, Vockler J, Dikranian K, Tenkova TI, Stefovska V, Turski L, Olney JW. Blockade of NMDA receptors and apoptotic neurodegeneration in the developing brain. Science 1999;283:70–4. 23. Joo JY, Kim BW, Lee JS, Park JY, Kim S, Yun YJ, Lee SH, Lee SH, Rhim H, Son H. Activation of NMDA receptors increases proliferation and differentiation of hippocampal neural progenitor cells. J Cell Sci 2007;120:1358–70. 24. Komuro H, Rakic P. Modulation of neuronal migration by NMDA receptors. Science 1993;260:95–7. AMPA receptors in pancreatic cancer 25. Hayashi T, Umemori H, Mishina M, Yamamoto T. The AMPA receptor interacts with and signals through the protein tyrosine kinase Lyn. Nature 1999; 397:72–6. 26. Ishiuchi S, Yoshida Y, Sugawara K, Aihara M, Ohtani T, Watanabe T, Saito N, Tsuzuki K, Okado H, Miwa A, Nakazato Y, Ozawa S. Ca2þ-permeable AMPA receptors regulate growth of human glioblastoma via Akt activation. J Neurosci 2007;27:7987–8001. 27. Stepulak A, Sifringer M, Rzeski W, Endesfelder S, Gratopp A, Pohl EE, Bittigau P, Felderhoff-Mueser U, Kaindl AM, Buhrer C, Hansen HH, StryjeckaZimmer M, et al. NMDA antagonist inhibits the extracellular signal-regulated kinase pathway and suppresses cancer growth. Proc Natl Acad Sci U S A 2005; 102:15605–10. 28. Abdul M, Hoosein N. N-methyl-Daspartate receptor in human prostate cancer. J Membr Biol 2005;205:125–8. 29. Rzeski W, Turski L, Ikonomidou C. Glutamate antagonists limit tumor growth. Proc Natl Acad Sci U S A 2001;98: 6372–7. 30. Takeda M, Haga M, Yamada H, Kinoshita M, Otsuka M, Tsuboi S, Moriyama Y. Ionotropic glutamate receptors expressed in human retinoblastoma Y79 cells. Neurosci Lett 2000;294:97–100. 31. Yoshioka A, Ikegaki N, Williams M, Pleasure D. Expression of N-methyl-Daspartate (NMDA) and non-NMDA glutamate receptor genes in neuroblastoma, medulloblastoma, and other cells lines. J Neurosci Res 1996;46:164–78. 32. Hauser R, Gos T, Krzyzanowski M, Goyke E. The concentration of glutamate in cerebral tissue as a factor for the assessment of the emotional state before death. A preliminary report. Int J Legal Med 1999;112:184–7. 33. Erkan M, Kleeff J, Gorbachevski A, Reiser C, Mitkus T, Esposito I, Giese T, Buchler MW, Giese NA, Friess H. Periostin creates a tumor-supportive microenvironment in the pancreas by sustaining fibrogenic stellate cell activity. Gastroenterology 2007; 132:1447–64. 34. Reiser-Erkan C, Erkan M, Pan Z, Bekasi S, Giese NA, Streit S, Michalski CW, Friess H, Kleeff J. Hypoxia-inducible protooncogene Pim-1 is a prognostic marker in pancreatic ductal adenocarcinoma. Cancer Biol Ther 2008;7:1352–9. Int. J. Cancer: 129, 2349–2359 (2011) VC 2011 UICC 35. Hu H, Qin Y, Bochorishvili G, Zhu Y, van Aelst L, Zhu JJ. Ras signaling mechanisms underlying impaired GluR1-dependent plasticity associated with fragile X syndrome. J Neurosci 2008;28:7847–62. 36. Zhu JJ, Qin Y, Zhao M, Van Aelst L, Malinow R. Ras and Rap control NS 105 AMPA receptor trafficking during synaptic plasticity. Cell 2002;110:443–55.
37. Qin Y, Zhu Y, Baumgart JP, Stornetta RL, Seidenman K, Mack V, van Aelst L, Zhu JJ. State-dependent Ras signaling and AMPA receptor trafficking. Genes Dev 2005;19:2000–15.
38. Tian X, Feig LA. Age-dependent participation of Ras-GRF proteins in coupling calciumpermeable AMPA glutamate receptors to Ras/Erk signaling in cortical neurons. J Biol Chem 2006;281:7578–82.
39. Ji B, Tsou L, Wang H, Gaiser S, Chang DZ, Daniluk J, Bi Y, Grote T, Longnecker DS, Logsdon CD. Ras activity levels control the development of pancreatic diseases. Gastroenterology 2009;137:1072–82,82e1–6.
40. Carriere C, Young AL, Gunn JR, Longnecker DS, Korc M. Acute pancreatitis markedly accelerates pancreatic cancer progression in mice expressing oncogenic Kras. Biochem Biophys Res Commun 2009; 382:561–5.
41. Khasawneh J, Schulz MD, Walch A, Rozman J, Hrabe de Angelis M, Klingenspor M, Buck A, Schwaiger M, Saur D, Schmid RM, Kloppel G, Sipos B, et al. Inflammation and mitochondrial fattyacid beta-oxidation link obesity to early tumor promotion. Proc Natl Acad Sci U S A 2009;106:3354–9.
42. Strobel O, Dor Y, Alsina J, Stirman A, Lauwers G, Trainor A, Castillo CF, Warshaw AL, Thayer SP. In vivo lineage tracing defines the role of acinar-to-ductal transdifferentiation in inflammatory ductal metaplasia. Gastroenterology 2007;133: 1999–2009.
43. Guerra C, Schuhmacher AJ, Canamero M, Grippo PJ, Verdaguer L, Perez-Gallego L, Dubus P, Sandgren EP, Barbacid M. Chronic pancreatitis is essential for induction of pancreatic ductal adenocarcinoma by K-Ras oncogenes in adult mice. Cancer Cell 2007;11: 291–302.
44. Singh A, Greninger P, Rhodes D, Koopman L, Violette S, Bardeesy N, Settleman J. A gene expression signature associated with ‘‘K-Ras addiction’’ reveals regulators of EMT and tumor cell survival. Cancer Cell 2009;15:489–500.
45. Ishiuchi S. New roles of glutamate receptors in glias and gliomas. Brain Nerve 2009;61:753–64.