Epithelial cells participate in the immune response of the intestinal mucosa. Extracellular nucleotides have been recognized as inflammatory molecules. We investigated the role of extracellular nucleotides and their associated P2Y receptors in the secretion of cytokines by epithelial cells. The effect of intestinal inflammation on P2Y6 receptor expression was determined by PCR in the mouse, rat, and human. Localization of the P2Y6 receptor was determined by immunofluorescence microscopy in the colon of normal and dextran sulfate sodium-treated mice. The effect of P2Y6 activation by UDP on cytokine expression and release by epithelial cells was determined using a combination of Western blots, luciferase assays, RT-PCR, cytokine Ab arrays, and ELISA. Inflammation up-regulates P2Y2 as well as P2Y6 receptor expression in the mucosa of the colon of colitic mice. In vitro, we demonstrated that UDP could be released by Caco-2/15 cells. We have confirmed the increased expression of P2Y6 by challenging intestinal epithelial cell-6 and Caco-2/15 cells with TNF-α and IFN-γ and showing that stimulation of epithelial cells by UDP results in an increased expression and release of CXCL8 by an ERK1/2-dependent mechanism. The increase in CXCL8 expression was associated with a transcriptional activation by the P2Y6 receptor. This study is the first report demonstrating the implication of P2Y receptors in the inflammatory response of intestinal epithelial cells. We show for the first time that P2Y6, as well as P2Y2, expression is increased by the stress associated with intestinal inflammation. These results demonstrate the emergence of extracellular nucleotide signaling in the orchestration of intestinal inflammation.
Intestinal epithelial cells (IECs)3 participate in the innate immune response of the intestinal mucosa by providing a physical barrier between the exterior environment and the host, as well as by actively secreting and responding to a rainbow of cytokines and releasing other immune active molecules such as defensin and IgA (1). Consequently, intestinal epithelial dysfunction appears to be central to diseases associated with an aberrant gut mucosal immune response such as in the inflammatory bowel diseases (IBDs). IBDs are characterized by an imbalance between anti- and proinflammatory cytokines, disruption of the intestinal epithelial monolayer, infiltration of pathogens, and recruitment and activation of leukocytes (2, 3, 4). Large amounts of extracellular nucleotides are also rapidly released into the extracellular environment at the site of inflammation as the result of cell and tissue damage, stress factors such as hypoxia, and decreased pH as well as leukocyte activation (5, 6).
The physiopathological activities associated with extracellular nucleotides are mediated by the specific activation of P2X1–7 ligand-gated cation channels and the G protein-coupled P2Y1, 2, 4, 6, and 11–14 nucleotide receptors (7, 8, 9). Under normal conditions, the P2Y2, P2Y4, and P2Y6 receptors actively participate in the regulation of Cl−, Na+, and K+ secretion in the small and large intestine (10, 11, 12) as well as in the absorption of water-soluble and high-m.w. compounds from the ileum (13). Under inflammatory conditions, P2Y receptor activation can stimulate cell proliferation, participate in immune cell recruitment, proliferation and differentiation, enhance cytokine and proinflammatory molecule expression and secretion, and increase expression of cell adhesion molecules and cell migration, as has been demonstrated in many other systems (14, 15, 16, 17, 18). In one study using a rat model of chemically induced colitis, it was shown that animals fed a diet supplemented with nucleosides and nucleotides had aggravated colonic injuries, as illustrated by epithelial damage; mucosal erosion; edema; coagulative necrosis of the muscularis propria; and increased leukocyte, macrophage, and lymphocyte infiltration of the colonic tissue (19, 20). The P2Y6 receptor has previously been associated with the inflammatory bowel diseases, in which it was found to be highly expressed on T cells infiltrating inflamed colonic tissues, but absent in T cells of the unaffected bowel (21). However, there is no knowledge concerning the possible implication of P2Y receptors in the epithelial response to intestinal inflammation in mucosal diseases. With an increasing amount of evidence pointing toward a key role for IECs in the pathogenesis of IBDs and the importance of extracellular nucleotides and their associated receptors during inflammation, we show in this study the impact of intestinal inflammation on P2Y receptor expression and the involvement of the P2Y6 receptor on the release and expression of CXCL8 (IL-8) by IECs.
Materials and Methods
DMEM, penicillin-streptomycin, HEPES, and FBS were purchased from Wisent. Glutamine (GlutaMax) was from Invitrogen Life Technologies. ATP and UTP were from Roche Applied Science. ADP, UDP, suramin, pyridoxal-5-phosphate-6-azophenyl-2′4′-disulfonic acid (PPADS), and N,N″-1,4-butanediyl bis(N′-[3-isothiocynatophenyl]) thiourea (MRS2578) were from Sigma-Aldrich. Recombinant human TNF-α and IFN-γ were purchased from BioShop. The MEK1/2 (UO126), p38 MAPK (SB203580), and JNK (SP600125) inhibitors were acquired from EMD Biosciences. Dextran sulfate sodium (DSS; Mr 36,000–50,000) was obtained from MP Biomedicals. The rabbit polyclonal anti-phospho-p44/p42 (Thr202/Tyr204) and anti-p44/42 were purchased from Cell Signaling Technology. Mouse monoclonal anti-actin (clone C4) and rabbit polyclonal anti-P2Y6 and control peptide Ag were obtained from Chemicon International. HRP-conjugated goat anti-mouse IgG and HRP-conjugated goat anti-rabbit IgG were from GE Health Care BioSciences, and ECL reagent was from Millipore. The human cytokine Ab array was purchased from Panomics, and the human IL-8 (CXCL8) ELISA kit was obtained from RayBiotech.
Induction of colitis
Adult CD-1 mice (30–35 g) were purchased from Charles River Laboratories. Colitis was induced by adding 5% (w/v) DSS to the drinking water for 7 days. Colitis was assessed by clinical scoring, as previously described (22, 23). Scores (maximal score = 4) were expressed as the average of five measured parameters, as follows: variation in weight loss, colon length, stool consistency, visibility of blood in stool, and rectal bleeding. Excised colons were gently washed using PBS and used either for paraffin embedding and immunohistological or immunofluorescence studies or for protein extraction and RNA isolation. All procedures were approved by the Université de Sherbrooke Animal Care Committee and performed according to the Canadian Guidelines for Care and Use of Experimental Animals.
Mouse colonic tissues were fixed in 4% paraformaldehyde overnight at 4°C, embedded in paraffin, cut into 5-μm sections, and applied to Superfrost/Plus slides (Fisher Scientific). H&E staining was performed by the Université de Sherbrooke Hospital Center Service of Pathology. Images were captured on a Leica DMLB2 microscope using a Leica DC300 camera.
Deparaffinized and rehydrated slides were subjected to microwave Ag retrieval by boiling for 6 min in a 10 mM citric acid buffer (pH 6.0) and allowed to cool for 10 min at room temperature (RT). Slides were washed in PBS, then blocked with protein-blocking reagent (Beckman Coulter) for 20 min at RT. The rabbit anti-P2Y6 Ab was diluted in PBS containing 0.1% BSA and 0.2% Triton X-100 (PBT) and incubated with the sections overnight at 4°C. Slides were washed in PBS and then incubated with the Alexa Fluor 568 F(ab′)2 goat anti-rabbit IgG (H+L) (Invitrogen Life Technologies) diluted in PBT for 2 h at RT. Slides were washed in PBS and mounted, and images were captured on a Leica DMRXA microscope using a Leica MPS60 camera.
The rat intestinal epithelial crypt cell line IEC-6 and the human colon carcinoma cell lines Caco-2/15 (24, 25) (provided by J. F. Beaulieu, Université de Sherbrooke, Sherbrooke, Québec, Canada) were grown, as previously described (24, 25, 26). Cells were incubated in serum-free medium for 24 h at 37°C before the experiments. For HPLC assays, Caco-2/15 cells were grown in phenol-free DMEM. Specific kinase inhibitors or P2Y receptor antagonists were added 30 min to serum-free medium before nucleotide stimulation. Inflammatory-like stress was induced in Caco-2/15 cells by adding a combination of 10 ng/ml TNF-α and 10 ng/ml IFN-γ for 6 h at 37°C, as previously described (27).
Transient transfections and luciferase assays
Caco-2/15 cells were washed twice with PBS, and 500 μl of OptiMEM (Invitrogen Life Technologies) was added to each well. Cells were cotransfected using LipofectAMINE 2000 (Invitrogen Life Technologies) with 0.1 μg of the −162/+44 minimal human CXCL8 promoter constructs (provided by A. Brasier, University of Texas Medical Branch, Galveston, TX) (28). After 8 h of transfection, the medium was replaced with complete DMEM. In some experiments, cells were treated with or without 10 μM MRS2578 for 30 min before UDP (100 μM) stimulation for 0–6 h. Luciferase activity was measured, as described previously (26). Results were normalized for Renilla luciferase expression and expressed as fold induction in comparison with control (empty vector) values.
Cytokine Ab array
Cytokines released by Caco-2/15 cells were determined using the Panomics cytokine human Ab array kit. Caco-2/15 cells were seeded onto 0.4-μm pore-size membrane inserts (Ultident Scientific) in six-well plates and grown until 4 days postconfluence, giving rise to polarized enterocyte-like cells (24, 26). Cells were serum starved for 24 h and then stimulated for 6 h with 100 μM UTP or UDP with or without MRS2578, SB203580, or UO126, as described above and in the figure legends. Cell medium was collected, cleared of debris, and used for the cytokine Ab array, as recommended by the manufacturer. Proteins were detected by ECL, and signals were captured on autoradiographic films.
CXCL8 release was quantified using the RayBiotech ELISA kit for human IL-8. Caco-2/15 cells were treated as described for the cytokine Ab array, and the ELISA were performed as recommended by the manufacturer.
RT-PCR analysis and real-time PCR quantification
Total RNA was isolated from IECs and mice colonic tissues with the RNeasy mini kit (Qiagen). cDNA was then synthesized from purified RNA by reverse transcription using the SuperScript II system (Invitrogen Life Technologies). Five percent of the synthesized cDNA was used as a template for PCR using the TaqDNA polymerase system (Qiagen). Oligonucleotide primers were designed to selectively amplify specific subtypes of P2Y receptor cDNA, as previously described (29). The sequence-specific primers for CXCL8 and GAPDH were as follows: 5′-GGA ACC ATT CTC ACT GTG TGT-3′ and 5′-CCT ACA ACA GAC CCA CAC AAT-3′ for CXCL8; 5′-CGG AGT CAA CGG ATT TGG TCG TAT-3′ and 5′-AGC CTT CTC CAT GGT GGT GAA GAC-3′ for human GAPDH; and 5′-GGT GAA GGT CGG TGT CAA CGG ATT-3′ and 5′-GAT GCC AAA GTT GTC ATG GAT GAC C-3′ for rat GAPDH. The resulting PCR products were resolved on a 1.5% agarose gel containing 10 μg/ml ethidium bromide and photographed under UV illumination. Quantification for P2Y2 and P2Y6 receptors and CXCL8 expression in human tissues was determined by quantitative PCR using the TissueScan Real-Time Crohn’s and colitis disease panel from Origene Technologies. Real-time PCR was performed according to the manufacturer’s guidelines using the following oligonucleotides: 5′-CGG TGG ACT TAG CTC TGA GG-3′ and 5′-GCC TCC AGA TGG GTC TAT GA-3′ for the human P2Y2; 5′-CCT GCC CAC ACA GCC ATC TT-3′ and 5′-GGC TGA GGT CAT AGC AGA CAG TG-3′ for the human P2Y6; and 5′-TCT GCA GCT CTG TGT GTG AAG G-3′ and 5′-AAT TTC TGT GTT GGC GCA GT-3′ for CXCL8. β-actin (primers provided by Origene) amplification was used as a control to normalize input amounts.
After treatment and/or stimulation, cells or mouse colonic tissue samples were washed with ice-cold PBS and lysed in Triton buffer (40 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.2 mM sodium orthovanadate, 40 mM β-glycerophosphate, 0.1 mM PMSF, and protease inhibitor mixture from Sigma-Aldrich). Protein concentration was determined using the Bio-Rad Protein Assay reagent from Bio-Rad. Samples were sonicated for 2 s and heated for 5 min at 96–100°C, subjected to 10% SDS-PAGE, and transferred to polyvinylidene fluoride membranes for protein immunoblotting, as previously described (26, 29, 30, 31). Immunoblotting for phospho-ERK1/2 was performed using a 1/1,000 dilution of rabbit polyclonal anti-phospho-p44/p42 MAPK (Thr202/Tyr204), and P2Y6 immunoblotting was performed with a 1/1,000 dilution of rabbit anti-P2Y6 receptor. Specific protein bands were detected using a 1:10,000 dilution of HRP-conjugated anti-rabbit IgG and visualized on autoradiographic film using the Millipore chemiluminescence system. For normalization of the signal, membranes were stripped of Ab by 30-min incubation at 37°C in Restore Western blot stripping buffer (Fisher Scientific), washed in TBST (20 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 0.1% Tween 20) and reprobed with a 1/1,000 dilution of rabbit anti-ERK1 (K-13) MAPK or a 1/10,000 dilution of mouse monoclonal anti-actin and a 1/10,000 dilution of HRP-conjugated anti-rabbit IgG or HRP-conjugated anti-mouse IgG as the secondary Ab, respectively.
HPLC measurement of UDP released by Caco-2/15 cells
After the induction of a stress-like insult to Caco-2/15 cells, cell medium was collected, cleared of cellular debris, and centrifuged on a CentriconPlus-20 column (Millipore; 5-kDa cutoffs) to clear the medium from high m.w. proteins. Detection of UTP and UDP in the cell culture medium was determined by reverse-phase HPLC (Varian Vista series) using an Agilent Zorbax 300SB-C18 column (300 Å) equilibrated with a mixture (98:2, v/v) of solution 1 (50 mM ammonium phosphate (pH 5.0), containing 5 mM tetrabutilammonium sulfate) and solution 2 (methanol containing 5 mM tetrabutilammonium sulfate). Cell medium samples (1 ml) or standard samples (100 nM in 1 ml of either UTP or UDP) were injected, and the ratio of solution 2 kept at 2% (v/v) for 15 min, and then linearly increased to 20% (v/v) over 40 min. The phase was than reversed for 5 min with 50% (v/v) of solution 2, and then linearly decreased to 2% (v/v) over 10 min and equilibrated for 10 min with 2% (v/v) of solution 2. Nucleotides were detected by absorbance at 260 nm. Retention time for UDP released by Caco-2/15 cells was compared with those of standard samples, and quantification of the concentration of UTP or UDP released by Caco-2/15 cells was evaluated by calculating the area under the curve (AUC) and compared with those obtained for the standard UTP and UDP samples.
Measurement of ecto-nucleotidase activities
Data were expressed and analyzed as indicated in the figure legends.
Intestinal inflammatory stress increases P2Y receptor expression
As shown in Fig. 1,A, the chemical induction of colitis in CD-1 mice using 5% DSS in the drinking water increased the expression of P2Y2 and P2Y6 receptor mRNA by >3-fold after 7 days of treatment. Enhanced P2Y1 receptor mRNA expression could also be measured, but was not statistically significant, whereas no difference in the P2Y4 receptor transcript could be noticed between normal and DSS-treated mice. Although several oligonucleotide primer sets were used, we could not detect transcripts for P2Y12, 13 and P2Y14 receptors (data not shown), whereas the P2Y11 receptor gene has not yet been cloned in the mouse or rat. This increase in P2Y2 and P2Y6 mRNA receptor expression correlated with an increase in the clinical score with r values of 0.928 and 0.987 (p < 0.05) for P2Y2 and P2Y6, respectively. Based on these results, we further analyzed the expression of the P2Y2 and P2Y6 receptor transcripts as well as CXCL8 in noninflamed, Crohn’s, and ulcerative colitis human tissues by quantitative PCR (Fig. 1,B). As in the mouse, P2Y2 receptor mRNA expression was increased 2-fold in Crohn’s and colitis tissues as compared with noninflamed samples and by 3- to 4-fold for the P2Y6 receptor (Fig. 1,B). The relative mRNA expression for CXCL8 was 21.9 ± 7 (p < 0.01 as compared with noninflamed samples) for Crohn’s samples, 13.3 ± 5.3 (p < 0.05) for ulcerative colitis samples, and 1.8 ± 0.9 for noninflamed samples (data not shown). Characterization for the contribution of the colonic epithelium to the increase in P2Y6 receptor expression was assayed in vivo by indirect immunofluorescence and in vitro by inducing an inflammatory stress-like insult to IEC lines. The colon of normal mice was characterized by well-defined colonic crypts, an uninterrupted monolayer of epithelial cells, and a normal submucosa layer (Fig. 2,A). In contrast, DSS treatment resulted in colonic crypt shortening, disruption of epithelial barrier integrity, and a thicker submucosa layer, as indicated by arrowheads on mildly affected regions (Fig. 2,B) and more visibly on severe regions of inflammation, as shown by the strong infiltration of immune cells and crypt ulceration (Fig. 2,C). Indirect immunofluorescence studies were performed to determine the expression pattern and relative expression of the P2Y6 receptor along the colonic crypt of normal (Fig. 2, E and G) and DSS-treated mice (Fig. 2, F and H). Negative controls using the preabsorbed Ab with the control Ag peptide showed single cell labeling at the level of the submucosal layer (Fig. 2,D). The P2Y6 receptor is located in the bottom third of the colonic crypt for both normal (Fig. 2,E) and DSS-treated mice (Fig. 2,F). At the cellular level, using a higher magnification, we were able to locate the expression of the P2Y6 receptor in the basolateral compartment of normal colonic tissues (Fig. 2,G) and in both the apical and basolateral compartments of the epithelial cells in the DSS-treated tissues (Fig. 2,H). Finally, using the same time of exposure, our results suggest an increase in P2Y6 expression at the bottom of the colonic crypt in DSS colitic mice (Fig. 2,H) as compared with normal mice (Fig. 2,G). These data were further supported by in vitro assays, in which an inflammatory stress-like insult was induced in IEC-6 (Fig. 2,I) and 4-day postconfluent Caco-2/15 cells (Fig. 2,J) by challenging the cells with a combination of TNF-α (10 ng/ml) and IFN-γ (10 ng/ml) (27). Immunoblotting for P2Y6 receptor expression using a commercially available Ab against the P2Y6 receptor revealed three bands at ∼36, 40, and 55 kDa for IEC-6, and 34, 40, and 55 kDa for Caco-2/15 cells (Fig. 2, I and J). All three bands were lost or strongly diminished after preincubation of the Ab with the blocking peptide (data not shown). Induction of inflammatory-like stress resulted in an increase of >30% in P2Y6 receptor expression after 6 h of challenge (Fig. 2, I and J), as determined by densitometry analysis (data not shown).
UDP is released by epithelial cells during inflammation
As for the cytokines and chemokines, inflammation has been reported to stimulate the release of extracellular nucleotides (5, 6, 32, 33). To determine whether IECs could release nucleotides, UDP in particular, under an inflammatory insult, we induced an inflammatory-like stress to the Caco-2/15, as previously described, collected the cell medium, and ran reverse-phase HPLC to detect the presence of UDP. Cells challenged with the cytokines showed an increase in UDP released (110–125 nM and a retention time of 11 min) as compared with nonchallenged cell medium in which the level of UDP was below the detection capacity of our system (<10 nM) (Fig. 3). Furthermore, because UDP could also be generated throughout the hydrolysis of UTP by E-NTPDases (34), we measured that Caco-2/15 has a UTPase activity of 1.02 ± 0.11 μmol/min/mg protein, in which the UTPase activity is the average ± SE of four experiments done in triplicate.
UTP and UDP activate CXCL8 expression and release by IECs
Intestinal inflammation is characterized by an increase in proinflammatory cytokine and chemokine release. It has been well accepted that IECs participate actively in that process (2, 4, 35, 36, 37). In that regard, we have investigated whether the activation of the P2Y2 and P2Y6 receptors by UTP and UDP could regulate cytokine and chemokine secretion by IECs. Caco-2/15 cells were grown to 15 days postconfluence on Transwell membranes and exposed to 100 μM UTP or UDP for various times, and cell medium was collected and used immediately for cytokine Ab array assays. UTP and UDP stimulate CXCL8 release after 1 h of treatment (data not shown), and peak secretion was visible after 6 h of stimulation as compared with unstimulated cells (Fig. 4,B). Besides the increased secretion of CXCL8, UTP and UDP transiently increased the release of the anti-inflammatory cytokine IL-4 and proinflammatory cytokine IL-12 (Fig. 4,B). Finally, addition of MRS2578 before UDP stimulation totally abolished the release of CXCL8. It is noteworthy that unstimulated Caco-2/15 cells showed a basal release of IL-12. CXCL8 release could be correlated with an increase in CXCL8 mRNA expression, as shown in Fig. 4, C and D. Indeed, as for CXCL8 release by Caco-2/15 cells, UTP and UDP significantly (p < 0.01) increased CXCL8 mRNA expression 3-fold. To determine which of the P2Y2 or P2Y6 receptors was involved in the expression and release of CXCL8 by IECs, we took advantage of P2Y receptor pharmacology using different antagonists. Suramin (P2Y2 and P2Y6 antagonists) and PPADS (P2Y1 and P2Y6 antagonists and a weak P2Y4 antagonist) had no significant inhibitory effect on UTP-stimulated cells (Fig. 4,C). However, addition of MRS2578 (a specific P2Y6 antagonist) significantly inhibited the stimulatory effect of UTP, but more significantly UDP by >12-fold (Fig. 4,D). To a lesser extent, suramin and PPADS could also antagonize the stimulatory effect of UDP (Fig. 4 D).
Increased CXLC8 expression by P2Y6 is mediated at the transcriptional level
Having established that the P2Y6 receptor mediates the increase in CXCL8 mRNA expression and release by IECs, we further investigated the regulating mechanisms underlying this increase. Caco-2/15 cells were transfected with the minimal human CXCL8 promoter coupled to the luciferase gene (CXCL8-Luc) for 48 h, as described in Materials and Methods. As previously described, the −162/+44 minimal CXCL8 promoter construct has potential binding sites for AP-1 (−126 to −120 bp), NF-IL-6 (−94 to −81 bp), and NF-κB (−80 to −70 bp) as well as a TATA box located at position −20 to −13 bp (28). Cells were then stimulated with 100 μM UDP or PBS (controls) for 30 min, 1 h, 3 h, and 6 h before luciferase assays (Fig. 5,A). Addition of 100 μM UDP significantly increased luciferase activity by >2- to 3-fold after 3 h (p < 0.05) and 6 h (p < 0.01) of stimulation, respectively (Fig. 5,A). Addition of MRS2578 before UDP stimulation inhibited the induction of luciferase activity back to the control level (CXCL8-Luc construct alone) (Fig. 5 B). This observation strongly supports that CXCL8 expression by UDP is mediated at the transcriptional level in IECs.
CXCL8 expression by P2Y6 is partially dependent on ERK1/2 phosphorylation
To understand the mechanism by which UDP induces the expression and release of CXCL8 by IECs, we investigated the impact of specific inhibitors of the ERK1/2 MAPK (UO126), p38 MAPK (SB203580), JNK1/2 (SP600125), and NF-κB (SN-50) (data not shown) signaling pathways on CXCL8 mRNA expression and secretion in Caco-2/15 (Fig. 6,A). Inhibition of the MEK1/2-ERK1/2 pathway by UO126 added 30 min before UDP stimulation resulted in a significant decrease of CXCL8 mRNA expression by 60% (Fig. 6,A). However, no significant inhibition of CXCL8 transcript expression was observed in SB203580 and SP600125- as well as SN-50 (data not shown)-treated cells. Similarly, pretreatment of Caco-2/15 cells with UO126 before UDP stimulation abolished the release of CXCL8 as compared with control cells stimulated with 100 μM UDP (Fig. 6 B). As for CXCL8 mRNA expression, addition of SB203580 did not show any significant inhibition of CXCL8 released.
ELISA were also performed to quantify the impact of our different treatments on CXCL8 released by Caco-2/15 cells. As for the cytokine arrays, both UTP and UDP significantly increased the release of CXCL8 as compared with nonstimulated cells (Fig. 6,C). Treatments with MRS2578 and UO126 significantly decreased the UDP-induced CXCL8 release as compared with UDP-stimulated cells (Fig. 6,C). In line with the cytokine arrays, the inhibition of the p38 MAPK signaling pathway with SB203580 did not significantly antagonize the release of CXCL8 by Caco-2/15 cells stimulated by UDP (Fig. 6 C).
P2Y6 receptor-mediated ERK1/2 activation
To determine whether the P2Y6 nucleotide receptor can mediate ERK1/2, we stimulated IEC-6 and Caco-2/15 cells with UDP in a time-dependent manner (Fig. 7). ERK1/2 phosphorylation was significantly increased after 5 min of stimulation in both cell lines, as shown by Western blot and densitometric analysis. Pretreatment of cells with the P2Y6 antagonist MRS2578 inhibited ERK1/2 phosphorylation (data not shown), indicating that P2Y6 can activate ERK1/2. No significant increase in p38 or JNK1/2 phosphorylation could be observed following UDP stimulation (data not shown).
Findings in this study demonstrate for the first time that UDP and its associated P2Y6 receptor are not only involved in the regulation of ion transport by IECs under normal conditions (12, 38), but also actively participate in the pathogenesis of intestinal inflammation by stimulating the expression and release of CXCL8. The involvement of the P2Y6 receptor in the inflammatory bowel diseases has previously been suggested, its expression being increased in human T cells infiltrating the bowel of affected patients (12, 21). In accordance with a recent study (12), the P2Y6 receptor was located on the colonic mucosa in the bottom third of mice colonic crypts. The relative expression of the P2Y receptor was enhanced on the mucosa of DSS-treated mice. Up-regulation of P2Y6 receptor expression was confirmed in vitro by challenging IEC-6 and Caco-2/15 cells with a combination of TNF-α and IFN-γ to induce an inflammatory stress-like insult. As previously reported, Western blotting revealed specific signals at ∼34/36 and 40/42 kDa (39). We also detected a specific signal at 55 kDa. The difference in m.w. could be attributed to the different nature of the cells as well as species from which they were isolated. The exact cellular localization of the P2Y6 receptor is still being debated. Indeed, in previous studies, the endogenous P2Y6 receptor was localized at the basolateral side of colonic epithelial cells of the rat (12), whereas the recombinant human P2Y6 receptor bearing a hemagglutinin tag was localized at the apical side of Caco-2 cells in culture (12, 40). In this study, we show that the P2Y6 receptor can be localized in both the apical and basolateral compartments of colonic epithelial cells, suggesting that P2Y6 cellular localization may be dependent on the physiopathological condition, the species studied, tissue processing, and the biochemical and histological tools available.
Up-regulation of the P2Y6 receptor in response to inflammatory insult also represents a new finding. The involvement of the P2Y2 receptor in the inflammatory response, as well as the up-regulation of its expression following inflammatory stress, have been reported in multiple systems (see Abbracchio and Burnstock (7) and Dubyak and el-Moatassim (9) for complete reviews). However, we show for the first time that P2Y2 mRNA expression is up-regulated during intestinal inflammation in our animal model as well as in human colonic samples of Crohn’s disease and ulcerative colitis (Fig. 1). More interestingly, P2Y6 receptor mRNA expression was also up-regulated in colon tissues isolated from DSS-treated mice and in human intestinal samples of IBDs (Fig. 1). To our knowledge, this is the first study reporting the P2Y2 and P2Y6 receptors to be concomitantly up-regulated in the context of inflammatory insult, such as intestinal inflammation. Because both genes are located within <4 cM on the human chromosome 11q13.5 (41) and on the mouse chromosome 7 E3, it would not be surprising if both genes shared common regulatory sequences, mechanisms, and molecules. Whether this increase is specific for the inflammatory bowel diseases is not known, but would be unlikely.
Although the levels of extracellular nucleotides in inflamed colonic mucosa have neither been reported nor measured, stressed and dying cells, activated leukocytes, and the acidic and hypoxic microenvironment provided by inflammation favor the release of nucleotides such as UTP and UDP (5, 6, 32). In intestinal inflammation, epithelial cells could be one possible source of extracellular nucleotides, including UDP. Indeed, we determined that the UDP concentration in the extracellular environment of Caco-2/15 cells, under inflammatory stress, was in the range of 110–125 nM/ml medium, whereas in nonchallenged cells the level of UDP was below our detection capacity (<10 nM) (Fig. 3). The measured levels of UDP in the medium samples were below the threshold values that promote P2Y6 receptor activation (42). However, recent studies suggest that measurements of nucleotide levels in bulk medium samples underestimate the local accumulation of nucleotides in the vicinity of P2 cell surface receptors (33, 43). For example, using a fusion system to anchor the ATP-specific enzyme luciferase to the outer cell surface of human platelets, it was observed that accumulation of ATP in the bulk medium from thrombin-stimulated platelets underestimates by at least one order of magnitude the ATP concentration in the environment near the cell surface (44). This result suggests that extracellular nucleotides, including UDP, can be found in sufficient amounts to stimulate P2Y receptors in an autocrine manner in the vicinity of the nucleotide released. Unfortunately, there is no such system to quantify the level of UTP or UDP at the cell surface level. Hence, in IBDs, extracellular nucleotides could be secreted not only by IECs, but also by other cell types, such as leukocytes, platelets, and smooth muscle cells, as well as from dying cells in which the cytoplasmic content, rich in nucleotides, is released into the extracellular environment. Furthermore, the absence of signal for the presence of UTP released by challenged Caco-2/15 cells in our samples could be due to its hydrolysis to UDP by E-NTPDases (34). In fact, it has been reported that the dephosphorylation of UTP by plasma membrane E-NTPDases significantly increases the accumulation of UDP, thereby favoring P2Y6 receptor activation (34, 45).
In the present study, we have shown that both UTP and UDP stimulate the expression and release of CXCL8 by the colonic epithelial cell line Caco-2/15, suggesting that both P2Y2 and P2Y6 receptors might be involved. However, the lack of a significant stimulating effect of ATP (data not shown), the poor antagonist activity of suramin and PPADS on ADP (data not shown) and UTP stimulation of CXCL8 mRNA expression, and the inhibition of UDP-induced CXCL8 expression by MRS2578 strongly suggest that P2Y6 may be the nucleotide receptor involved in the expression and secretion of CXCL8 by IECs. The stimulating effect of UTP on CXCL8 expression and secretion may be due to its hydrolysis into UDP by the E-NTPDase expressed by epithelial cells of the digestive tract (45, 46). The UTP effect on CXCL8 secretion could also come from minute amounts of UDP in the UTP preparation, as we observed by HPLC (data not shown). This contaminating UDP, in combination with the NTPDase activity of IECs that hydrolyze UTP to UDP, may be sufficient to stimulate the P2Y6 receptor.
In the colon, cytokines and other inflammatory molecules are generally produced by cells belonging to the immune system, such as lymphocytes, macrophages, and dendritic cells. Recently, nonclassical immune cells, such as IECs, have been shown to actively participate in the mucosal immune response by not only responding to inflammatory molecules, but also by acting as APCs and by secreting cytokines (4, 36, 37, 47, 48). Several cytokines, such as TGF-α, IL-1, IL-10, IL-15, and IL-18, are constitutively expressed by IECs and may be involved in the basal influx of immune cells into the mucosa and in epithelial cell growth and homeostasis (36). Others, such as IL-1β, IL-6, TNF-α, MCP-1, CCL10, GM-CSF, and CXCL8, are also expressed by normal epithelial cells, but are markedly up-regulated under proinflammatory conditions (4, 36). More recently, UDP activation of the P2Y6 receptor in the myeloid cell lineage resulted in an increase in CXCL8 release (49, 50, 51, 52). In this context, we show that stimulation of P2Y6 on IECs with UDP resulted not only in an increased expression of CXCL8, but also in the stimulation of CXCL8 released by IECs through a ERK-1/2-dependent mechanism (Fig. 6). In HT-29 cells, CXCL8 release has been reported to be regulated by a posttranscriptional mechanism involving the stabilization of the mRNA transcript (53). However, we have shown in this study, using the minimal CXCL8 human promoter coupled to the luciferase gene, that the increase in CXCL8 expression could be regulated at the transcriptional level by stimulating IECs with the P2Y6 agonist UDP in the presence or not of MRS2578, as shown in Fig. 5. Furthermore, the participation of ERK1/2 in IECs in the context of inflammation and CXCL8 expression has previously been reported (53, 54). It is well accepted that cytokine expression is generally regulated by the NF-κB signaling pathway. In the present study, treatment of IECs with the NF-κB inhibitor SN50 (data not shown) before UDP stimulation did not result in any change in CXCL8 mRNA expression, suggesting that P2Y6-mediated CXCL8 expression and release by IECs involve a ERK1/2-dependent, but NF-κB-independent mechanism, as previously reported in nontransformed human colonocytes following flagellin/TLR5 engagement (55).
The contribution of the P2Y6 receptor to the pathogenesis of ulcerative colitis could therefore involve two distinct mechanisms, as follows: first, by actively participating in the onset of the disease, the acute phase, by regulating the immune response of IECs, and secondly, by influencing infiltrating T cell function in the chronic phase of colitis by a yet unidentified mechanism (21). Nonetheless, our results support a key role for extracellular nucleotides, UDP in particular, and the P2Y6 receptor in the innate mucosal response of the intestine, not only by regulating T cell activity in chronic colitis, but also by stimulating the release of CXCL8, a chemokine known for is chemoattractive ability to recruit neutrophils during the acute phase of colitis. Hence, our results support the concept that extracellular nucleotides are alarmins, a group of endogenous danger-signaling molecules rapidly released and displaying potent innate immune-enhancing activities (56, 57).
We thank Benoit Auclair for his technical help with the immunofluorescence studies; Dr. Fernand Gobeil Jr., Dr. Witold Neugebauer, and Elie Simard for their technical help with the HPLC analysis; and Mathieu Darsigny for helpful comments on DSS treatment, as well as Elizabeth Herring and Sophie Tousignant for critical reading of the manuscript.
The authors have no financial conflict of interest.
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.
This work was supported by the Crohn’s and Colitis Foundation of Canada Grant in Aid of Research and an establishment grant from the Fonds de la Recherche en Santé du Québec. F.-P.G. is a scholar from the Fonds de la Recherche en Santé du Québec and a member of the Fonds de la Recherche en Santé du Québec-funded Centre de Recherche Clinique Étienne Lebel.
Abbreviations used in this paper: IEC, intestinal epithelial cell; AUC, area under the curve; DSS, dextran sulfate sodium; E-NTPDase, ecto-nucleoside triphosphate diphosphohydrolase; IBD, inflammatory bowel disease; MRS2578, N,N″-1,4-butanediyl bis(N′-[3-isothiocynatophenyl]) thiourea; PPADS, pyridoxal-5-phosphate-6-azophenyl-2′4′-disulfonic acid; RT, room temperature.