Stadnyk, Andrew W.*; Carrigan, Svetlana O.†; Otley, Anthony R.*
The inflammatory bowel diseases (IBD), Crohn disease, and ulcerative colitis are relapsing/remitting chronic inflammatory diseases of the gastrointestinal tract. Both are characterized histologically by the infiltration of polymorphonuclear (PMN) and other leukocytes into the intestine, including into the lumen in crypt abscesses. Increases in PMN products such as calprotectin and lactoferrin in patient stool correlate with disease severity and may even be predictive of relapses, suggesting that the migration contributes to the exacerbation of the disease (1–4). From experimental systems, PMN in the lumen are thought to affect epithelial cell functions such as Cl− secretion, the basis of diarrhea (5). We are therefore interested in the mechanisms of transintestinal epithelial migration for the purpose of controlling symptoms and possibly the root cause of IBD. Notwithstanding the progress made during the last decade, much remains unknown about the mechanisms involved in the transintestinal epithelial migration of PMN and the development of crypt abscesses.
PMN transintestinal epithelial migration is achieved by a combination of chemoattractant signals and specific molecular adhesive events. There has been considerable interest in discovering the chemoattractants responsible for the perpetuation of intestinal inflammation and epithelial cells have consequently been identified as a source of PMN chemoattractants, including chemokines (6). The chemokines that potently support PMN migration are found in the family of CXC molecules defined by the presence of a single-variable amino acid separating the first 2 amino terminal cysteine residues. In addition to the conserved cysteines, CXC chemokines active on PMN have a unique “ELR motif” (Glu-Leu-Arg) at the amino terminal. Chemokines from this subclass reported in the inflamed mucosa include CXCL8 (interleukin [IL]-8), CXCL5 (epithelial-derived neutrophil chemoattractant-78 [ENA-78]), CXCL7 (neutrophil-activating peptide-2 or NAP-2), and CXCL1 (growth-related protein-α or Gro-α) (7,8) with IL-8 and ENA-78 confirmed as epithelial-derived based on immunohistochemical evidence from biopsies of patients with IBD (9,10). All of these chemokines can be synthesized by intestinal epithelial cells in cell culture systems. IL-8 has been detected in rectal dialysates recovered from patients with IBD (11), as has the lipid product and chemoattractant leukotriene B4 (LTB4) (12); however, these analyses are confounded by leukocytes in the lumen that secrete chemoattractants. Thus, the evidence that any of these molecules have a direct role in transepithelial migration early in IBD is incomplete.
PMN typically express 2 receptors for the ELR+ CXC chemokines; CXCR1 binds only IL-8 and CXCL6 (GCP-2) with high affinity, whereas CXCR2 binds all of the ELR+ CXC subclass chemokines. Williams et al (13) observed an upregulation of CXCR1 but not CXCR2 in the mucosa of patients with ulcerative colitis, implying that there may be finer modulation of these inflammatory markers than is typically appreciated.
In the course of characterizing adhesion events during PMN transepithelial migration, we observed a peculiar difference in the sensitivity of cells to CXC chemokines during migration across bare filters versus epithelial monolayers. Considering we still do not know the chemoattractant leading to crypt abscesses, we sought to characterize the potency of CXC chemokines using patient PMN in the transintestinal epithelial migration assay, and specifically, the relative contributions of CXCR1 and CXCR2. We show there is indeed a difference in potency between chemokines that use CXCR2 versus IL-8 and that the difference is because of the specific elicitation of adenosine in the assay, which impedes migration specifically across monolayers. The significance of this finding is that differential expression of the chemokines in the inflamed bowel may help rationalize further pharmaceutical controls for the inflammation.
PMN-activating agents and chemoattractants N-formyl-methionine-leucine-phenylalanine (fMLP) and C5a were purchased from Sigma Chemical (St Louis, MO); the chemokines CXCL8 (IL-8), CXCL5 (8.0 kDa ENA-78), and CXCL1 (7.8 kDa Gro-α) were from Peprotech (Rocky Hill, NJ). LTB4 was purchased from Cayman Chemical (Cedarlane, Hornby, ON). Monoclonal anti-CXCR1 (clone 42705.111, immunoglobulin G2a [IgG2a]), anti-CXCR2, and PE-conjugated anti-CXCR2 (both clone 48311.211, IgG2a) antibodies were from R&D Systems (Minneapolis, MN). Two monoclonal anti-CD18 antibodies were used with identical outcomes; clone 60.3 was from Bristol-Myers Squibb (Seattle, WA), and clone IB4 was provided by Dr Andrew Issekutz, Dalhousie University. Mouse IgG2a isotype control antibody was from Cymbus Biotech (Chandlers Ford, Hampshire, UK), fluorescein isothiocyanate (FITC) -conjugated anti-mouse IgG from Chemicon (Temecula, CA), and PE-conjugated anti-mouse IgG came from Jackson Laboratories (West Grove, PA). Adenosine deaminase (ADA, used at 1 U/mL final concentration) and adenosine (used at 100 μmol/L final concentration) were provided by Dr Jonathan Blay, Department of Pharmacology, Dalhousie University.
Patients were recruited from referrals to the pediatric gastroenterology and nutrition clinic in our tertiary care center. The consent and project information forms were approved by the research ethics board of the IWK Health Centre. Patient ages ranged between 8 and 19 years. A diagnosis of IBD was confirmed histologically in all of the patients used in the study. Data using PMN from patients of either sex and on treatments are pooled with first presentation patients of both sexes.
Donor blood was drawn into acid citrate-dextrose anticoagulant and heparin and processed as described elsewhere (14). Briefly, a portion of this sample was used for platelet-poor plasma (PPP) prepared by centrifugation at 500g at 4°C followed by further centrifugation of the supernatant at 200g at 4°C. PPP was kept on ice until use. The remaining blood was mixed with dextrose, red blood cells were allowed to sediment for 1 hour, and leukocyte-rich plasma (LRP) was recovered. The LRP was centrifuged at 200g at room temperature and the resulting cell pellet resuspended in Ca2+ Mg2+-free Tyrode solution containing 10% PPP. The leukocytes were then labeled using 200 μCi/mL Na251CrO4 (Amersham Corp, Oakville, ON) during a 30-minute incubation at 37°C, with mixing every 5 minutes, then resuspension to a final volume of 4 mL in Tyrode solution with 10% PPP and layered onto a discontinuous Percoll (+10% PPP) density gradient (58%/72%). The gradient was centrifuged at 475g for 25 minutes at room temperature and the PMN recovered from the 58%/72% interface were washed 3 times with Tyrode solution before use in the transmigration assay. The PMN were consistently >95% pure and viable as determined by crystal violet and trypan blue staining.
Flow cytometry was performed on LRP collected as above and on enriched PMN. Residual red blood cell contamination of LRP was removed by lysis. The remaining cells were activated with 10−8 mol/L fMLP or 10−9 mol/L C5a at 37°C for 10 minutes. After treatment, cells were kept on ice for the remainder of the experiment. Cells were stained either with unconjugated anti-CXCR1 antibody followed by a FITC-conjugated secondary antibody (indirect detection), or with PE-conjugated anti-CXCR2 antibody (direct detection). Nonspecific staining was assessed with an isotype-matched unconjugated immunoglobulin followed by a FITC- or PE-conjugated secondary antibody. Fluorescence from CXCR1- and CXCR2-labeled cells was quantified using a FACScan flow cytometer (Becton Dickinson, Franklin Lakes, NJ). PMN were gated using forward and side light scatter properties from the LRP.
The human colonic carcinoma cell line, T84, was maintained in media consisting of a 1:1 mix of Dulbecco Modified Eagle Medium (DMEM) and Ham F12 supplemented with 5% newborn calf serum, 15 mmol/L N-2-hydroxyethylpiperazine-N′-2-ethane sulfonic acid, and 100 U/mL penicillin/streptomycin (Life Technologies, Burlington, ON, Canada). Cells were kept at 37°C in 5% CO2. For migration experiments, T84 monolayers were established on the underside of collagen-coated Transwell filters (0.33 cm2) with a 3-μm pore size (Fisher Scientific, Nepean, ON, Canada), described in detail elsewhere (14). This permits migration across the cells in the basolateral-to-apical direction. Inverted monolayers were used within 1 week and the media was changed the day before use in a migration assay.
PMN Migration Assay
51Cr-labeled PMN were suspended in DMEM supplemented with 5 mg/mL human serum albumin to a final concentration of 106 cells per milliliter. One hundred microlitres of the cell suspension was added to the upper chamber of the Transwell filters. The filters were then added to a multiwell plate with the lower chambers (the wells) containing DMEM plus 5 mg/mL human serum albumin (600 μL final volume) with or without chemoattractant. Bare filters were soaked in T84 culture media for at least 4 hours before use. When blocking antibodies were used, the PMN were incubated with the antibody at room temperature for 20 minutes before being added to the Transwell. We reported previously that a control antibody was without effect in the assay, so the isotype control was not included in these experiments (14). Migration was allowed to proceed for 70 minutes across bare filters and for 2 hours across monolayers, both at 37°C, 5% CO2. After this time, PMN were collected separately from the top chamber (nonmigrated), lower chamber (migrated), and monolayer/filter (adherent). The upper chamber was washed 3 times with media and the washes were added to the nonmigrated fraction. The monolayer was rinsed once with media to dislodge any loosely adherent cells and these were added to the migrated fraction. The undersides of the bare filters were rinsed with 2 mmol/L ethylenediaminetetraacetic acid to dislodge firmly adherent cells, which were added to the migrated fraction. Any remaining adherent cells on the bare filters or monolayers were collected by submersion in 2 mol/L NaOH overnight. Cells in the bottom chamber were lysed with Triton X 100 (1% final concentration). Radioactivity in each fraction was measured in a Beckman gamma counter. Typically, 90% of the added radioactivity was recovered among all 3 fractions. The number of PMN migrated is reported as the percentage of the radioactivity in the migrated fraction over that of a standard 100-μL aliquot of cells (equal to total cells added).
Parametric methods of comparison were used. When there were only 2 groups in an experiment, the 2-tailed Student t test was used. In experiments with >2 groups, analysis of variance was first performed, and if found significant, then post hoc testing using Tukey was done to compare treatment groups. A P < 0.05 was considered a significant difference and is identified in the figures with an asterisk.
We previously published the migration kinetics and whether β2 integrins (CD18) were critical to migration of healthy adult donor PMN to a panel of chemoattractants (14). We discovered that PMN use CD18-independent mechanisms when migrating to a number of chemoattractants other than fMLP. We thought it was important to use cells from patients with IBD considering there are precedents describing phenotypic and functional differences between patient and healthy donor–peripheral blood PMN (15). We repeated our earlier experiments, including testing whether migration to each chemoattractant required the use of CD18 for migration, using peripheral blood PMN of adolescent patients with IBD. Figure 1A shows that the patient PMN responded to all of the chemoattractants with migration and that the pattern of migration of the patient PMN closely resembles that using healthy adult donors. The pattern included only CD18-dependent migration to fMLP but CD18-independent migration to C5a, IL-8, and LTB4(14). Although Figure 1 shows that migration to IL-8 includes a CD18-independent component, we had previously observed that migration to ENA-78 or Gro-α, which bind 1 receptor in common with IL-8, was entirely CD18 dependent (Fig. 1B). Both of these chemokines have been reported in the inflamed colon and both bind to CXCR2, but only IL-8 binds to CXCR1. Considering there are different activities in PMN that become activated through each receptor (16), we sought to more closely dissect transepithelial migration by patient cells using these receptors.
We first assayed for CXCR1 and CXCR2 expression on patient cells and assessed whether expression was affected by the enrichment technique or by PMN-activating agents. Using flow cytometry, antibodies to both receptors stained essentially all of the PMN among the leukocyte fraction in whole blood (gating using light scatter properties for PMN in Fig. 2A), indicated by the peak fluorescence shifting right of the control antibody. Staining was unchanged by the Percoll enrichment technique (Fig. 2A, enriched PMN). The level of expression was unchanged by treating the enriched PMN for 10 minutes at 37°C with fMLP or C5a (Fig. 2B), stimuli that result in increased surface CD11b/CD18 expression and increased expression of the CD18 activation epitope (17). Because the receptor density did not change either with handling the cells or with common activating agents, we did not measure the numbers of each CXC receptor type on the PMN. We concluded from these results that all of the patient PMN were positive for both receptors and that receptor levels were unlikely to change in the course of an experiment. Furthermore, considering the homogeneity of receptor staining, any chemotactic response is unlikely to be due to subsets of PMN with significant variations in receptor levels. The receptor expression on healthy adult donor PMN was 84 ± 35 and 38 ± 16 fluorescent units for CXCR1 and CXCR2, respectively, and was not statistically significantly different from the adolescent cell expression levels.
Various ELR+ CXC chemokines were then examined for potency at inducing patient PMN migration across Transwell filters and in the physiological basolateral-to-apical direction across T84 colonic epithelial monolayers growing on Transwell filters. Migration in the absence of any chemoattractant was consistently around 12% and 2% across bare filters and T84 monolayers, respectively. Migration across bare filters versus T84 monolayers in response to a range of chemokine concentrations is shown in Figure 3. IL-8, ENA-78, and Gro-α effectively mediate migration across bare filters, although the total number of migrating PMN using the CXCR2 ligands was less than the number migrating to IL-8. Based on the T84 monolayer impeding the passive diffusion of the chemoattractant and additional titration curves (not shown), we routinely use chemokine concentrations 10-fold greater for optimal migration across T84 monolayers than for migration across bare filters. The maximal PMN migration to IL-8 across inverted T84 monolayers was similar to the maximum achieved across bare filters; however, migration to the CXCR2 ligands across monolayers was considerably lower than the same chemokine used across bare filters.
A possible explanation for the reduced potency of ENA-78 or Gro-α to mediate migration across monolayers is that the T84 modify the chemokine (18,19). We tested this possibility by preincubating T84 growing on the filter bottom with ENA-78 for 2 hours (the period that migrations are conducted), then recovering and using the media as chemoattractant for PMN across bare Transwell filters. The ENA-78–spiked supernatant proved as potent as a fresh preparation of the chemoattractant at inducing migration across bare filters (data not shown), indicating that the epithelial cells do not directly affect the bioactivity of the chemoattractant.
It was unclear what contribution IL-8 binding to CXCR2 provided to the overall migration when it also bound CXCR1, so we next planned to determine the relative contributions of each receptor in the assay. The preincubation of PMN with blocking antibodies to CXCR1 reduced migration across bare filters by roughly 50% (Fig. 4A), whereas the reduction in migration across T84 was closer to 75% (Fig. 4B), resulting in a final magnitude similar to ENA-78 or Gro-α. Thus, cells migrating to IL-8 but without CXCR1 behave similar to cells migrating to a CXCR2 ligand alone. We planned to conduct the reciprocal experiment, using IL-8 and antibodies to CXCR2 (there are no readily available chemoattractants that bind CXCR1 only). Unfortunately, in specificity control experiments, the commercial antibodies available against CXCR2 (used at 25 μg/mL) did not block migration to concentrations of ENA-78 or Gro-α necessary for optimal migration (not shown), so we were unable to measure the migration mediated by CXCR1 alone. Finally, we compared multiple patients’ PMN migration to IL-8 across bare filters and monolayers, and compared the ratio of migration in the presence of anti-CXCR2 versus in the absence of blocking antibody (reported as percentage of IL-8 migration) and detected a statistically significant difference between the groups (Fig. 5). This finding was also reproduced using PMN from healthy adult donors.
The block in migration across epithelial monolayers unique to CXCR2 ligands could be the result of an event following CXCR1 activation that promotes PMN migration, or an activity unique to CXCR2 activation that impedes migration but which is overcome when CXCR1 is also activated (as IL-8 binds both receptors). Indeed, a number of PMN activities have been attributed to activation of the cells through CXCR1, in particular, the generation of superoxide radicals through the respiratory burst (16). Superoxide anion could act to alter the barrier properties of the epithelial cells, possibly enhancing migration, whereas activation exclusively through CXCR2 may not. Thus, we predicted that blocking the effects or production of superoxide anions would result in lower migration to IL-8 across T84 monolayers. We took 2 approaches to test this hypothesis. In 1 experiment, migration to IL-8 was carried out across T84 monolayers in the presence of 10 μg/mL superoxide dismutase, which had no effect on the magnitude of migration (not shown). In a second experiment, we used PMN from an adolescent with chronic granulomatous disease (who did not have IBD), whose cells are deficient in NADPH oxidase and therefore superoxide anion production. The chronic granulomatous disease patient PMN migrated to IL-8 in a pattern similar to cells of patients with IBD, for example, in similar numbers across bare filters as across T84 monolayers but with greater inhibition using anti-CXCR1 antibody during migration across T84 than across bare filters (Fig. 6). Thus, we could not attribute the differential use of CXCR during migration across T84 monolayers to the selective generation of superoxide radicals. We also ruled out the possibility that IL-8 acting through CXCR1 stimulated metalloproteinases that in turn affected migration, as 2 different inhibitors (metalloproteinase-1 inhibitor and 1,10-phenoanthroline, used at subtoxic concentrations) failed to affect PMN migration across T84 monolayers (3 experiments, data not shown).
Because less is known about PMN activities unique to stimulation through CXCR2, we next considered mediators that may be constitutively made by PMN but which could differentially affect migration. PMN are a potent source of adenosine, which is generally regarded as an anti-inflammatory agent (20). Speculating that endogenous adenosine was either suppressing migration to ENA-78 or enhancing migration to IL-8, we added ADA to the transepithelial migration assay. Figure 7A shows that the presence of ADA had no effect on the migration of PMN to ENA-78 or IL-8 across bare filters. Including ADA in the transepithelial assay had no significant effect on migration of PMN across T84 monolayers to IL-8; however, a significantly greater number of PMN migrated to ENA-78 across T84 monolayers in the presence of ADA compared with migration to ENA-78 in the absence of ADA (Fig. 7B). This enhancement remained significantly different from the numbers of PMN migrating to IL-8. This enhancement indicated that adenosine in the assay was impeding migration of PMN to ENA-78. Adenosine can have inflammatory and anti-inflammatory activities on PMN depending on the concentration, with low doses acting on 1 receptor type to be inflammatory, whereas higher concentrations act on a second receptor type to effect anti-inflammatory activities (20). We thought to pulse PMN with adenosine before adding the cells to the assay and still run the assay in the presence of ADA. In this combination, adenosine would affect the PMN in the absence of the chemoattractant, whereas the ADA should continue to reduce adenosine concentrations during the migration period. The adenosine pulse had no statistically significant consequence on PMN migration to IL-8 across monolayers (Fig. 7C), yet the combination of treatments resulted in PMN migration in numbers resembling migration to ENA-78 across bare filters (and IL-8 in Fig. 7C). We conclude from this pattern of results that adenosine acts through >1 mechanism to influence PMN migration specifically across the T84 monolayers to CXCR2-specific ligands.
The presence and presumed destructive potential of PMN in the bowel mucosa have long been appreciated as part of IBD pathology, yet effective means of specifically inhibiting PMN infiltration have not been forthcoming. Infiltration across the epithelium into abscesses may be equally important, because this seems to harm the epithelium. That ELR+ CXC chemokines may mediate this migration is supported by evidence showing increased concentrations in the inflamed mucosa and from lumen sampling. This includes 1 study providing clues for differential expression of the ELR+ CXC chemokines during IBD. Both IL-8 and GCP-2 bind CXCR1 and CXCR2, yet IL-8 was produced by peripheral blood mononuclear cells stimulated ex vivo (although production was reduced in cells of patients with Crohn disease), whereas GCP-2 was not produced by the blood cells but was detected mainly localized to endothelial cells and leukocytes in ulcerated tissue (21). In the same study, gut endothelial samples were negative but tissue leukocytes positive for ENA-78 (21). Despite such reports of ELR+ CXC chemokines, whether there is a gradient of effects with different chemokines acting through CXCR1 versus CXCR2 is not known. Indeed, a challenge to using human material is that the first wave of leukocytes to reach the mucosa and lumen likely also secrete chemoattractants (eg, the study by Gijsbers et al (21) identified leukocytes positive for chemokines). When these cells become crypt abscesses, they may become the source of the chemoattractant that continues to influence further infiltration, confounding the goal of identifying the chemoattractants responsible for crypt abscesses from lumen samples.
We are interested in whether chemokines that act on only CXCR1 or CXCR2 are active and possibly distinguish subsets of (IBD) pathology. To this end, we discovered that peripheral blood PMN from healthy adult donors and pediatric patients with IBD show different potencies to direct chemotaxis across epithelial monolayers depending on whether CXCR2 (ligand ENA-78) or both receptors (ligand IL-8) are used compared with migration across bare (acellular) filters. This difference in potency across monolayers could be the result of a direct impedance of chemotaxis to CXCR2 ligands or an activity unique to activating cells through CXCR1. Because there is considerably more known about PMN activities because of stimulation through CXCR1, we first investigated the hypothesis that stimulation of PMN rendered the epithelium “migration permissive,” in contrast to cells stimulated to migrate through CXCR2. For example, PMN activated through CXCR1 induces production of superoxide radicals (16). Superoxide radical production was ruled out in the mechanism behind PMN migration across T84 monolayers to fMLP (22). We likewise ruled out an influence of superoxide radical production on the pattern of migration to ENA-78 across monolayers, also by using cells from a chronic granulomatous patient as well as superoxide dismutase. We additionally ruled out a role for metalloproteinases, which may be made by either PMN or the T84 cells, possibly influencing migration to ENA-78. We ultimately show that the difference in potency of chemotaxis across monolayers versus bare filters can be explained by adenosine influencing PMN chemotaxis through CXCR2.
Adenosine has recently garnered interest in a variety of inflammatory diseases because of its broad anti-inflammatory properties. That adenosine is anti-inflammatory is well demonstrated in 1 study in which an adenosine kinase inhibitor was used to treat experimentally induced colitis. The authors used GP515 (heightens endogenous adenosine levels) during a 10-day exposure of mice to dextran sodium sulfate, which causes acute colitis. The mice receiving the kinase inhibitor experienced significantly less pathology at the end of the study (23). This success was followed by similar studies using a drug regimen that specifically stimulated 1 adenosine receptor type in 2 additional models of rodent colitis (24). This is despite the seemingly contradictory data showing that chronic ileitis results in upregulation of other adenosine receptors that are not necessarily anti-inflammatory (25) and the fact that high adenosine can have detrimental consequences on the circulatory system.
Although these whole-animal experiments are encouraging, the study of anti-inflammatory versus possible inflammatory mechanisms of adenosine and PMN is complicated by PMN possibly possessing all 4 receptor types and acting as a potent source of adenosine. PMN have been confirmed to possess A1 and A2 (A2A and A2B) adenosine receptors, whereas the presence of A3 receptors is disputed (20). At relatively low concentrations of adenosine, A1 receptors impart signals resulting in inflammatory activities by PMN, whereas higher adenosine concentrations impart signals through A2A receptors that manifest as anti-inflammatory activities, for example, by blocking cytokine secretion (20,26). Relevant to PMN, adenosine reportedly inhibits fMLP upregulation of β2 integrin molecules necessary for engaging cells during transmigration (27,28). This inhibition is likely linked to adenosine inhibition of phospholipase D activity (29), which, we have shown, is critical for CD18-dependent PMN transmigration across T84 monolayers (17). Because PMN migration to CXCR2 ligands is similarly CD18-dependent (Fig. 1), the ADA in our assay would catabolize adenosine liberated by PMN, permitting greater upregulation of β2 integrin receptor numbers and affinity in the presence of ENA-78, and favoring greater migration. This mechanism can explain the significant enhancement of migration across T84 monolayers in the presence of ADA.
Another possible explanation for the adenosine exclusively influencing ENA-78–mediated migration may be our using identical concentrations of ENA-78 and IL-8, with IL-8 perhaps being more potent and therefore masking any effect of adenosine. This could be examined by using a range of lower (albeit suboptimal) concentrations of IL-8 combined with the adenosine. The additive effect of CD18-independent migration of PMN to IL-8 (Fig. 1A) may also mask an effect of adenosine acting on the CD18-dependent component. Nevertheless, we discovered the conditions that result in PMN migration to ENA-78, ultimately achieving levels similar to migration because of IL-8.
Migration of PMN at levels similar to those using bare filters (and not statistically different from IL-8 used with monolayers) was achieved by combining the use of ADA in the assay with a pulse exposure of the PMN to adenosine before adding the cells to the coculture (and therefore in the absence of chemoattractant). This effect of the adenosine pulse is more difficult to reconcile mechanistically. It is possible that the adenosine pulse in the absence of chemoattractant activated CD18-independent adhesive mechanisms that further heightened the magnitude of PMN migrating to ENA-78 beyond the CD18-dependent levels alone.
In addition to the PMN secreting adenosine and possessing multiple adenosine receptors, the T84 cell line and native colonic epithelial cells have adenosine receptors, and the prevailing effect of adenosine on epithelial cells is inflammatory. In this setting, adenosine acts through A2B receptors on the epithelial cells to stimulate Cl− secretion (30). In fact, others using the Transwell paradigm showed that this stimulation of Cl− secretion was induced by transmigrated PMN acting back on the epithelium. Transmigrated PMN secreted 5′-adenosine monophosphate, which was converted to adenosine by CD73 on the apical membrane of the T84 cells, and thus the PMN indirectly are the source of adenosine that drives Cl− secretion underlying diarrhea (5). The adenosine additionally induced epithelial cell secretion of IL-6, which in turn activated the PMN in the model crypt lumen (31). Our data build on this “reciprocal secretion” (31) by showing that endogenous adenosine in the Transwell paradigm has anti-inflammatory potential by affecting the capacity of PMN to transmigrate in the presence of CXCR2 ligands and likely other chemoattractants that elicit CD18-dependent adhesion mechanisms. Collectively, these mechanisms support the concept that preventing PMN transepithelial cell migration will reduce the inflammation seen in colitis and that reducing PMN transepithelial migration may be a contributing mechanism behind the anti-inflammatory property of adenosine in the gut.
The authors thank Wendy Hughes, Hana James, and Barbara Christensen for their help with patient recruitment and technical assistance during the experiments, and Dr A. Issekutz, Department of Pediatrics, for identifying and recruiting the patient with CGD.
1. Tibble JA, Bjarnason I. Fecal calprotectin as an index of intestinal inflammation. Drugs Today (Barc) 2001; 37:85–96.
2. Kane SV, Sandborn WJ, Rufo PA, et al. Fecal lactoferrin is a sensitive and specific marker in identifying intestinal inflammation. Am J Gastroenterol 2003; 98:1309–1314.
3. Gouni-Berthold I, Baumeister B, Wegel E, et al. Neutrophil-elastase in chronic inflammatory bowel disease: a marker of disease activity? Hepatogastroenterology 1999; 46:2315–2320.
4. Fagerberg UL, Loof L, Myrdal U, et al. Colorectal inflammation is well predicted by fecal calprotectin in children with gastrointestinal symptoms. J Pediatr Gastroenterol Nutr 2005; 40:450–455.
5. Madara JL, Patapoff TW, Gillece-Castro B, et al. 5′-adenosine monophosphate is the neutrophil-derived paracrine factor that elicits chloride secretion from T84 intestinal epithelial cell monolayers. J Clin Invest 1993; 91:2320–2325.
6. Stadnyk AW. Intestinal epithelial cells as a source of inflammatory cytokines and chemokines. Can J Gastroenterol 2002; 16:241–246.
7. Brandt E, Müller-Alouf H, Desreumaux P, et al. Circulating growth-regulated oncogene α contributes to neutrophil priming and interleukin-8-directed mucosal recruitment into chronic lesions of patients with Crohn's disease. Eur Cytokine Netw 1998; 9:647–653.
8. Z’Graggen K, Walz A, Mazzucchelli L, et al. The C-X-C chemokine ENA-78 is preferentially expressed in intestinal epithelium in inflammatory bowel disease. Gastroenterology 1997; 113:808–816.
9. Imada A, Ina K, Shimada M, et al. Coordinate upregulation of interleukin-8 and growth-related gene product-α is present in the colonic mucosa of inflammatory bowel disease. Scand J Gastroenterol 2001; 36:854–864.
10. Keates S, Keates AC, Mizoguchi E, et al. Enterocytes are the primary source of the chemokine ENA-78 in normal colon and ulcerative colitis. Am J Physiol 1997; 273:G75–G82.
11. Keshavarzian A, Fusunyan RD, Jacyno R, et al. Increased interleukin-8 (IL-8) in rectal dialysate from patients with ulcerative colitis: evidence for a biological role for IL-8 in inflammation of the colon. Am J Gastroenterol 1999; 94:704–712.
12. Lauisten K, Laursen LS, Bukhave K, et al. Effects of topical 5-aminosalicylic acid and prednisolone on prostaglandin E2 and leukotriene B4 levels determined by equilibrium in vivo dialysis of rectum in relapsing ulcerative colitis. Gastroenterology 1986; 91:837–844.
13. Williams EJ, Haque S, Banks C, et al. Distribution of the interleukin-8 receptors, CXCR1 and CXCR2, in inflamed gut tissue. J Pathol 2000; 192:533–539.
14. Blake KE, Carrigan SO, Issekutz AC, et al. Neutrophils migrate across intestinal epithelium using β2 (CD11b/CD18) independent mechanisms. Clin Exp Immunol 2004; 136:262–268.
15. Anton PA, Targan SR, Shanahan F. Increased neutrophil receptors for and response to the proinflammatory bacterial peptide formyl-methionyl-leucyl-phenylalanine in Crohn's disease. Gastroenterology 1989; 97:20–28.
16. Stillie RM, Farooq SM, Gordon JH, et al. The functional significance behind expressing two IL-8 receptor types on PMN. J Leukoc Biol 2009; 86:529–543.
17. Carrigan SO, Pink DBS, Stadnyk AW. Neutrophil transepithelial migration in response to the chemoattractant fMLP but not C5a is phospholipase D-dependent and related to the use of CD11b/CD18. J Leukoc Biol 2007; 82:1575–1584.
18. Van Den Steen PE, Wuyts A, Husson SJ, et al. Gelatinase B/MMP-9 and neutrophil collagenase/MMP-8 process the chemokines human GCP-2/CXCL6, ENA-78/CXCL5 and mouse GCP-2/LIX and modulate their physiological activities. Eur J Biochem 2003; 270:3739–3749.
19. Nufer O, Corbett M, Walz A. Amino-terminal processing of chemokine ENA-78 regulates biological activity. Biochemistry 1999; 38:636–642.
20. Haskó G, Cronstein BN. Adenosine: an endogenous regulator of innate immunity. Trend Immunol 2004; 25:33–39.
21. Gijsbers K, Van Assche G, Joossens S, et al. CXCR1-binding chemokines in inflammatory bowel diseases: down-regulated IL-8/CXCL8 production by leukocytes in Crohn's disease and selective GCP-2/CXCL6 expression in inflamed intestinal tissue. Eur J Immunol 2004; 34:1992–2000.
22. Nash S, Stafford J, Madara JL. The selective and superoxide-independent disruption of intestinal epithelial tight junctions during leukocyte transmigration. Lab Invest 1988; 59:531–537.
23. Siegmund B, Rieder F, Albrich S, et al. Adenosine kinase inhibitor GP515 improves experimental colitis in mice. J Pharmacol Exp Ther 2001; 296:99–105.
24. Odashima M, Bamias G, Rivera-Nieves J, et al. Activation of A2A adenosine receptor attenuates intestinal inflammation in animal models of inflammatory bowel diseases. Gastroenterology 2005; 129:26–33.
25. Sundaram U, Hassanain H, Suntres Z, et al. Rabbit chronic ileitis leads to up-regulation of adenosine A1/A3 gene products, oxidative stress, and immune modulation. Biochem Pharmacol 2003; 65:1529–1538.
26. McColl SR, St-Onge M, Dussault AA, et al. Immunomodulatory impact of the A2A adenosine receptor on the profile of chemokines produced by neutrophils. FASEB J 2006; 20:187–189.
27. Wollner A, Wollner S, Smith JB. Acting via the A2 receptors, adenosine inhibits the upregulation of Mac-1 (CD11b/CD18) expression on fMLP-stimulated neutrophils. Am J Respir Cell Mol Biol 1993; 9:179–185.
28. Thiel M, Chambers JD, Chouker A, et al. Effect of adenosine on the expression of beta(2)integrins and L-selectin of human ploymorphonuclear leukocytes in vitro. J Leukoc Biol 1996; 59:671–682.
29. Thibault N, Harbour D, Borgeat P, et al. Adenosine receptor occupancy suppresses chemoattractant-induced phospholipase D activity by diminishing membrane recruitment of small GTPases. Blood 2000; 95:519–527.
30. Strohmeier GR, Reppert SM, Lencer WI, et al. The A2b adenosine receptor mediates cAMP responses to adenosine receptor agonists in human intestinal epithelia. J Biol Chem 1995; 270:2387–2394.
31. Sitaraman SV, Merlin D, Wang L, et al. Neutrophil-epithelial crosstalk at the intestinal luminal surface mediated by reciprocal secretion of adenosine and IL-6. J Clin Invest 2001; 107:861–869.