The P2X7 receptor regulates cell survival, migration and invasion of pancreatic ductal adenocarcinoma cells

Five PDAC cell lines were used: AsPC-1, BxPC-3, Capan-1, MiaPaCa-2 and Panc-1. They are genotypically and phenotypically heterogeneous and they are representative of different stages of pancreatic cancer. For example, Panc-1 is derived from epithelioid pancreatic carcinoma, MiaPaCa-2 is a poorly differentiated cell line [34], Capan-1 is a well differentiated cell line derived from liver metastasis [35], and AsPC-1 is a poorly differentiated cell line derived from nude mouse xenografts initiated with cells from the ascites of a patient with pancreatic cancer [36]. All cell lines have mutations in TP53 and KRAS genes, except for BxPC-3 which has wild type KRAS. HPDE, human “normal” pancreatic duct epithelial cell line, transformed using HPV16-E6E7 [37] was used as control.

The primer set for P2X7 detection (284 bp) (Table 1) was designed such that the expression of the isoforms A and B (and also G, H, J, D) would be detected in a real time PCR. The forward primer would recognize the junction between exon 3 and 4, and the reverse primer the junction between the exon 6 and 7. The P2X7R expression among the different cell lines was normalized with respect to HPDE, after normalization of each cell line to average expression of QARS, GUSZ and β-actin housekeeping genes. Figure 1a shows that compared to HPDE cells, there was a significant down-regulation of P2X7R transcripts in all the PDAC cell lines, except for Capan-1 cells. In addition to P2X7R, pancreatic duct cells also express a number of other P2X and P2Y receptors and additional data for the key receptors transcripts are given in Additional file 1: Figure S1 and primers are in Additional file 2: Table S1.

Protein expression of the full length P2X7R A isoform and the C-terminus truncated B isoform was determined using Western blot and immunolocalization (Fig. 1b-c). Figure 1b shows two bands, often seen by other researchers, and the lower band may correspond to the isoform H. The band at 70 kDa, corresponding to the isoform A, appears more abundant in all PDAC cell lines compared to control HPDE cells, but significant increase is detected only for Capan-1 and Panc-1. Figure 1c shows that there was a slightly reduced, but not significant, expression of the 42 kDa band that possibly corresponds to the isoform B in AsPC-1, BxPC-3 and Panc-1; and there is significantly lower expression in MiaPaCa-2 compared to HPDE cells.

Localization of P2X7R was detected by immunofluorescence and confocal microscopy and the results are shown in Fig. 2. The immunofluorescence was observed in HPDE and Panc-1 cells and the immunofluorescence signal resulting from the polyclonal antibody recognizing the extracellular loop, hence the A and B and potentially other isoforms, was localized to the plasma membrane, and weakly also in the cytoplasmic compartment (Fig. 2).

In the following series of experiments we investigated the effect of a sustained activation or inhibition of P2X7 receptor on cell proliferation and/or death using BrdU assay. First, cells were incubated with different concentrations of ATP and the results for the six cell lines are shown in Fig. 3 (black symbols). Generally, with increasing ATP concentration from 100 μM to 5 mM all cell lines showed a decrease in BrdU incorporation, possibly indicating cell death (see below). Minor deviations in responses were observed in Panc-1 cells, which were relatively insensitive to 100 μM ATP, and in Capan-1 cells, which showed an increase in BrdU incorporation with 5 mM ATP, but a reduction at 1 mM. This effect in Capan-1 cells could be due to a selection of resistant cells after long incubation with a high dose of ATP, as shown in murine macrophage RAW 264.7 cells [38]. Lower concentrations of ATP (0.5–10 μM) had small and inconsistent effects on BrdU incorporation, showing a tendency of about 10–20 % increase in BrdU incorporation (Additional file 3: Figure S2).

In order to verify that the above described effects of ATP were due to P2X7R, a more specific agonist, BzATP, was used and the results are shown in Fig. 4. The application of BzATP at 10 μM had no significant effects on all cell lines. With an increase of the concentration of BzATP to 100 μM, a significant reduction of BrdU incorporation was detected in HPDE, AsPC-1 (around 65 %) and BxPC-3 (around 30 %) cells. At high concentration of BzATP (1 mM), all the cell lines showed a significant decrease in BrdU incorporation (between 65 % and 85 %).

In the next series of experiments we tested the effects of P2X7R inhibitors: the allosteric inhibitor AZ10606120 (10 μM) that affects cell proliferation; the competitive antagonist A438079 (10 μM) that inhibits pore formation; and the non-competitive antagonist KN-62 (100 nM) that also inhibits CaM Kinase II [24, 27, 39‐41]. Experiments where AZ10606120 was given without exogenous ATP (i.e. basal conditions) resulted in decreased BrdU incorporation in all cell lines, with the largest decrease in Panc-1 and the lowest decrease in HPDE and Capan-1 cells (Fig. 3, colored symbols). This would indicate either that the inhibitor attenuated proliferation or that it was cytotoxic and promoted cell death. We excluded the cytotoxic effect of the inhibitor with a Trypan blue assay (Additional file 4: Figure S3). Therefore, the inhibitory effect of AZ10606120 in basal conditions was most likely related to a decrease in proliferation. Furthermore, the inhibitor in combination with added ATP concentrations (0.1–1.0 mM), resulted in a further reduction of BrdU incorporation in all PDAC cell lines. Accepting that the inhibitor is not cytotoxic, we presume it reduced BrdU incorporation because it inhibited cell proliferation, as also seen in other cells [17, 27, 41]. At about 5 mM ATP effects of the inhibitor were the smallest, indicating that cell death outweighted cell proliferation. Thus Fig. 3 shows two effects on BrdU incorporation in a population of cells—increasing ATP concentration causing cell death (see below) and AZ10606120 causing a decrease in cell proliferation. Interestingly, AZ10606120 was ineffective in HPDE cells treated with 0.1–5 mM exogenous ATP, indicating that in these conditions cell death was the predominant P2X7R response in these non-cancer cells. In addition to AZ10606120, we also studied the long-term effect of two other P2X7R inhibitors, the pore-inhibitor A438079 and KN-62, and the results are shown in Fig. 5. It appears that these inhibitors improved BrdU incorporation if cells were stimulated with 0.1–1 mM ATP. Thus these inhibitors protected cells against cell death until exogenous concentrations of ATP were 5 mM.

The above data (Figs. 3, 4 and 5) indicated that high concentrations of ATP and BzATP could lead to cell death. We next investigated whether this was due to apoptosis or necrosis. To detect induction of apoptotic pathways, caspase 3/7 assay was used and data are shown in Fig. 6a. The administration of 1 mM ATP (20 h) had no significant effect on caspase 3/7 activation in Panc-1 cells. We moreover did not detect any effect when AZ10606120 was included alone or with 1 mM ATP. As a positive control, it is shown that apoptosis inducer AT101 (see methods) caused marked caspase activation.

In addition to caspase 3/7 assay, an Annexin V-488/propidium iodide (PI) assay was performed (Fig. 6b). Panc-1 cells were incubated with 1 mM ATP (20 h) and images in Fig. 6b show that a number of cells had red nuclei (PI), indicating that the cell membrane was permeabilized as in cells undergoing necrosis, and green Annexin V, indicating phosphatidyl serine exposure on cell surface. The transmission image also shows disrupted cells with uneven surface and nuclear swelling, also indicating a necrotic type of death.

One of the hallmarks of P2X7R is the ability to form pores permeable to larger molecules up to 900 Da, as shown in the above experiments. Therefore, we investigated effect of high ATP concentration (5 mM) on membrane permeabilization using YoPro-1 dye in a physiological buffer and results are shown in Fig. 7. The cell lines tested showed an increase in YoPro-1 uptake after stimulation with 5 mM ATP. The maximum YoPro-1 uptake with 5 mM ATP was detected after 20 min in AsPC-1 and Panc-1 cells, and after 50–60 min in BxPC-3 cells. We tried to inhibit P2X7R and P2X7R-mediated pore formation using two inhibitors: the allosteric inhibitor of P2X7R AZ10606120 (10 μM); and the competitive antagonist A438079 (10 μM), the latter known as the pore inhibitor. All cell lines tested showed a small but significant reduction in pore formation when A438079 was used: for YoPro-1 uptake at 60 min, this inhibition was around 6 % in BxPC-3, 13 % in AsPC-1 and 4.5 % in Panc-1 cells. A438079 protected Panc-1 cells from pore-formation in the short-term experiment (Fig. 7) and from cell death in long-term experiment where BrdU incorporation was monitored (Fig. 5). The allosteric inhibitor, AZ10606120, had no significant effect on YoPro-1 uptake in Panc-1 and AsPC-1 cells and showed a small inhibition of 6 % only in BxPC-3 cells. In an additional series of experiments we also tested 1 mM ATP on YoPro-1 uptake, but effects on pore formation were even smaller than with 5 mM ATP and these are shown in Additional file 5: Figure S4. Interestingly, although 1–5 mM ATP did not cause very pronounced pore-formation in PDAC cells in given experimental conditions and A438079 had relatively small protective effects, the inhibitor could nevertheless protected PDAC cells from cell death during long-term incubation as shown in Fig. 5.

In order to determine effect of P2X7R on cell migration, a scratch assay was used and cell proliferation was inhibited by aphidicolin. Here we focus on Panc-1 cells as a representative cell line that showed most significant effects of AZ10606120 on cell proliferation (Fig. 3) and the fastest migration among the PDAC cells (Additional file 6: Figure S5). In the first set of experiments we used a mechanical 96-pin wound maker. As shown in Fig. 8a-b, the inhibitor AZ10606120 (10 μM) did not show any effect on Panc-1 cells migration under these experimental conditions. We reasoned that the scratch procedure caused cells damage and ATP release that would obscure a possible effect of P2X7R on migration, because there might be other P2 receptors that contribute to cell migration as shown in studies on other cells [42, 43]. Therefore, we used another more gentle method, where the wound was created by removal of a silicone insert and the results are shown in Fig. 8c-d-e and Additional files 7 and 8. With this new protocol it was possible to detect that AZ10606120 caused a significant, about 30 % reduction in Panc-1 migration compared to the control.

In order to evaluate whether the effect of P2X7R on PDAC cells migration was associated with invasion, i.e. the ability to degrade the ECM, a classical trans-well invasion assay was performed with Panc-1 cells (Fig. 9). Cells were incubated for 24 h with AZ10606120 (10 μM), BzATP (10 μM) and ATP (100 μM). Results in Fig. 9 show that the inhibitor caused about 20 % reduction in Panc-1 cell invasion compared to the control. BzATP and ATP stimulation increased cell migration by about 27 % and 55 %, respectively.

Human pancreatic ductal adenocarcinoma cell lines were purchased from ATTC (Manassas, VA, USA). AsPC-1 (CRL-1682) and BxPC-3 (CRL-1687) were grown in RPMI-1640 medium, Capan-1 (HTB-79) in Iscove’s Modified DMEM (IMDM), Panc-1 (CRL-1469) in Dulbecco’s Modified Eagles Medium (DMEM) and MiaPaCa-2 (CRL-1420) in DMEM/Ham’s F12. Cell culture media contained stable glutamine, 10 % (20 % for Capan-1) fetal bovine serum (FBS) (PAA Laboratories; A15-151 Gold) plus 2.5 % of horse serum for MiaPaCa-2 (Biochrom), 100 U/ml penicillin and 100 μg/ml streptomycin. The human pancreatic duct epithelial cell line HPDE6-E6E7 (H6c7) [60, 61], which we refer to as HPDE, was obtained from Dr. Ming-Sound Tsao. HPDE cells were grown in KBM Basal Medium (Lonza, CC-3101), which is part of the KGM-Bullet Kit (Lonza, CC-3111). Cells were grown at 37 °C in a humidified atmosphere with 5 % CO₂. All standard chemicals were purchased from Sigma-Aldrich unless otherwise stated.

Cells were cultured to confluence in a Petri dish and then RNA was isolated with RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. DNase I, Qiagen was used. The cDNA was synthesized using the RevertAid First Strand cDNA synthesis kit (Fermentas). For RT-PCR cDNA, Jumpstart Taq DNA, primers and dNTP mix were used; the amplification parameters were as follows: one cycle at 95 °C for 1 min, 45 cycles at 95 °C for 20 s, 57 °C for 30 s, 72 °C for 1.15 min, one final cycle at 72 °C for 5 min. The transcripts were run electrophoretically in 1.2 % agarose gel. The real time PCR reactions were run using Roche FastStart Universal SYBR Green Master (ROX) with parameters as follows: one cycle at 50 °C for 2 min and one cycle at 95 °C for 10 min followed by 40 amplification cycles at 95 °C for 15 s, 57 °C for 1 min, 72 °C for 1 min and at the end a dissociation step at 95 °C for 15 s, 60 °C for 15 s, 95 °C for 15 s. Data were evaluated with SDS 2.4 software (Applied Biosystems). Primers were designed using Primer-BLAST (NCBI) and synthesized by TAG Copenhagen A/S, DK. Three housekeeping genes that were found quite stable in normal and cancerous pancreatic tissues [62] were used for normalization: β-actin, β-glucuronidase (GUSB) and glutaminyl-tRNA synthetase (QARS). The relative amount of the target mRNA was calculated from a standard curve constructed from the results of a serial dilution run on each plate, using Pfaffl method [63].

Cells were cultured to confluence in a Petri dish and lysed by adding 5x diluted lysis buffer (50 mM Tris-base, 0.25 M NaCl, 10 mM EDTA, 0.5 % NP40, 1 % TritonX-100, 4 mM NaF, pH 7.5) with 1x protease inhibitor. The final lysate was centrifuged at 15,000 g at 4 °C for 15 min. Protein concentration was estimated with a Coomassie protein assay. Protein samples were reduced by heating at 98 °C for 10 min with 50 mM DTT. 50 μg of proteins were loaded and run in 10 % polyacrylamide precast gels (Invitrogen) and blotted to PVDF membranes (Invitrogen). Membranes were blocked with 5 % skim milk solution in TBS-Tween 0.1 % buffer for 1 h at room temperature and incubated overnight at 4 °C with primary antibodies against P2X7R extracellular loop (1:200 rabbit polyclonal, Alomone, APR-008) and β-Actin (1:1000 mouse monoclonal C4, Santa Cruz, Sc-47778). Blots were then incubated with appropriate HRP-conjugated antibodies and developed with EZ-ECL (Biological Industries) and visualized on Fusion FX (Vilber Lourmat) and band intensity was calculated using Bio1D software.

Cells were grown on glass coverslip, and then fixed in 4 % paraformaldehyde for 15 min at room temperature. Cells were treated with 0.1 M TRIS-glycine (pH 7.4) for 15 min; permeabilized 0.3 % Tween-20 and blocked with 5 % BSA for 30 min. Cells were incubated with antibody against P2X7R extracellular loop (1:100 rabbit polyclonal, Alomone, APR-008) overnight at 4 °C. Then preparations were incubated with a secondary antibody conjugated to Alexa 568. DAPI (Molecular Probes) was used for nuclear staining. Fluorescence was examined with 40x 1.3 NA objective in Leica SP 5X MP confocal laser scanning microscope. Images and overlays were analyzed in Leica LAS AF software.

Cells were plated in white, clear bottom 96-well plates. ATP and 2′-3′-O-(4-benzoylbenzoyl)-ATP (BzATP) were dissolved in media and the pH was adjusted to 7.4 with NaOH. The indicated concentrations of ATP or BzATP were added with and without P2X7R inhibitors (10 μM AZ 10606120, 10 μM A438079 and 100 nM KN-62, Tocris) in 1 % serum media (0 % of serum for the experiments with low ATP concentrations—from 0.5 μM to 10 μM). After 60 h of incubation at 37 °C in a humidified atmosphere with 5 % CO₂, cells were treated according to the manufacturer’s instruction of the Roche Cell Proliferation ELISA, BrdU chemiluminescence kit. Luminescence was read in FLUOstar OPTIMA (BMG Labtech).

To check the possible cytotoxicity of the inhibitor AZ10606120 (10 μM), the cell viability, was determined using Trypan blue exclusion. Panc-1 and BxPC-3 cells were grown in 6-well plates in 1 % serum. After 20 h, 40 h and 60 h of incubation at 37 °C in a humidified atmosphere with 5 % CO₂, cells were incubated 10 min with Trypan blue. Bright field images were taken with Leica DMI6000B microscope and 10x objective. In order to determine the mode of two common cell deaths (apoptosis, necrosis) two assays were used. For apoptosis, Panc-1 cells were plated in white 96-well plates with clear bottom. ATP solution adjusted to pH 7.4 was added in 1 % serum media for a final concentration of 1 mM. Apoptosis inducer AT101 [64] was used as a control. After 20 h of incubation, caspase 3/7 activation was determined. Cells were treated according to the manufacturer’s instruction of the Promega Caspase-Glo® 3/7 Assay. Luminescence was read by FLUOstar OPTIMA. For necrosis assay Panc-1 cells were grown in 35 mm glass-bottom dishes, after 20 h of treatment with ATP cells were incubated for 15 min with Annexin V-488 and Propidium Iodide (PI) (Life Technologies). Pictures were taken with 40x 1.3 NA objective in Leica SP 5X MP confocal laser scanning microscope, CLSM. Images and overlays were analyzed in Leica LAS AF software.

Cells were plated in black 96-well plates with clear bottom. When a confluence around 80–90 % was reached, the media was changed to a physiological buffer (140 mM NaCl, 10 mM HEPES, 1 mM MgCl₂, 0.4 mM KH₂PO₄, 1.6 mM K₂HPO₄, 1.5 mM CaCl₂) with 11.1 mM of glucose for AsPc-1 and BxPC-3 and 25 mM for Panc-1. Cells were incubated with AZ10606120 (10 μM) or A438079 (10 μM) for 1 h. Then cells were incubated with Yo-Pro-1 Iodide (2.5 μM, Molecular Probes) for 15 min. ATP was dissolved in a physiological buffer and the indicated concentrations were added. Ionomycin (5 μM) was added at the end of each experiment. Fluorescence was read every 10 min by FLUOstar OPTIMA.

The scratch assay was performed in two different methods. In the first method, cells were grown to confluence in 96-well Essen ImageLock plate, and then incubated with 10 μM AZ10606120 for 1 h and 5 μM aphidicolin (to stop proliferation). The wounds were made simultaneously in all wells with a 96-pin WoundMaker (Essen BioScience), washed with PBS and phase contrast pictures were taken every 2 h for 24 h in the IncuCyte ZOOM Kinetic Imaging System. Pictures were processed with the IncuCyte software using the relative wound density (%RWD) option, which measures the density of the wound with respect to the density of the cell region. The program uses the following formula: %RWD = 100 × w(t) − w(0)/c(t) − w(t); where, w(t) is the density of wounded area at time t and c(t) is the density of cell region at time t. To avoid cells damage, a second method was used. Ibidi biocompatible silicone culture-insert was stuck in the center of a 35 mm Petri dish and cells were plated in the same number in the two chambers. When the confluence was reached, the insert was removed and afterwards media with and without AZ10606120 (10 μM) was added. 10x phase contrast pictures were taken every 1 h for 24 h in the Nikon BioStation IM-Q that allowed a temperature- and gas-controlled environment. Images were analyzed by Nikon NIS-Elements AR 4.13.03 software.

The inserts for 24-well plates (transparent PET membrane, 8.0 μm pore size, Falcon) were coated with Geltrex Matrix (Invitrogen). 30.000 Panc-1 cells with aphidicolin (5 μM) and with/without AZ10606120 (10 μM), BzATP (10 μM) and ATP (100 μM) were placed in the upper chamber in 1 % of serum, and media with 10 % of serum was added in the lower chamber. After 24 h incubation at 37 °C in a humidified atmosphere with 5 % CO₂, cells were fixed in cold methanol and stained with Crystal Violet. Bright field images were taken with 10x objective in Leica DMI6000B microscope. Cells were counted by ImageJ, NIH.

All chemicals/kits were purchased from Sigma-Aldrich unless otherwise stated. Data are shown as mean values ± standard error of mean (SEM) and n represents the number of biological replicates. Statistical analysis on data was performed by using Students t-test and One-way Analysis of Variance (ANOVA) with Holm-Sidak method using SigmaPlot 11.0. Data were analyzed in Origin or Microsoft Excel.