Epac‐2 ameliorates spontaneous colitis in Il‐10 −/− mice by protecting the intestinal barrier and suppressing NF‐κB/MAPK signalling

Abstract Intestinal barrier dysfunction and intestinal inflammation interact in the progression of Crohn's disease (CD). A recent study indicated that Epac‐2 protected the intestinal barrier and had anti‐inflammatory effects. The present study examined the function of Epac‐2 in CD‐like colitis. Interleukin‐10 gene knockout (Il‐10 −/−) mice exhibit significant spontaneous enteritis and were used as the CD model. These mice were treated with Epac‐2 agonists (Me‐cAMP) or Epac‐2 antagonists (HJC‐0350) or were fed normally (control), and colitis and intestinal barrier structure and function were compared. A Caco‐2 and RAW 264.7 cell co‐culture system were used to analyse the effects of Epac‐2 on the cross‐talk between intestinal epithelial cells and inflammatory cells. Epac‐2 activation significantly ameliorated colitis in mice, which was indicated by reductions in the colitis inflammation score, the expression of inflammatory factors and intestinal permeability. Epac‐2 activation also decreased Caco‐2 cell permeability in an LPS‐induced cell co‐culture system. Epac‐2 activation significantly suppressed nuclear factor (NF)‐κB/mitogen‐activated protein kinase (MAPK) signalling in vivo and in vitro. Epac‐2 may be a therapeutic target for CD based on its anti‐inflammatory functions and protective effects on the intestinal barrier.


| INTRODUC TI ON
Crohn's disease (CD) is a chronic, recurrent and non-specific transmural type of inflammation commonly found in the digestive tract. 1 CD is characterized by abdominal pain, diarrhoea, anaemia and associated malnutrition, which detrimentally impact the patient's quality of life. 2 The incidence of CD has increased steadily with improvements in the standard of living. 3 The predisposing factors and specific pathological mechanisms of CD are not fully understood.
However, imbalances in immune homeostasis, intestinal barrier dysfunction and bacteria are major contributors to the pathogenesis of CD. 4 Improvements in intestinal barrier function positively impact the treatment and recovery of CD. 5 When the intestinal barrier is disrupted, bacteria enter the human body through the intestinal mucosa, which leads to local or systemic inflammation. 6 The structure of the epithelial barrier is essential for resisting microbial entry and maintaining homeostasis, and a major goal in the treatment of CD is to promote intestinal epithelial repair. 7 The mucosal barrier includes the epithelial barrier, which hinders the spread of macromolecular substances. 8 Epithelial cells in the intestinal barrier connect cell boundaries via tight junctions (TJs), and the functions of the intestinal barrier are primarily performed by TJs, the mucus layer and antibiotic peptides. 9 Many bacterial pathogens destroy TJs by inhibiting the expression of zonula occludens-1 (ZO-1) and occludin, and the destruction of TJ structure and function in epithelial cells leads to the loss of barrier integrity. 10 Inflammation induces the pathophysiological mechanism of intestinal injury. 11 The activation of cytokines (interleukin (IL)-17A, tumour necrosis factor-alpha (TNFα) and IL-1β) may inhibit the expression of TJ protein and lead to the destruction of intestinal barrier function. 12 The exchange protein Epac-2 is activated by cAMP, and it is the link between Ras and Rap-1. Epac-2 is the major Epac isoform expressed in the gut. 13,14 Epac-2 activation upregulates the expression of Rap-1. 15 Recent studies showed that Rap-1 robustly regulated the functions of adherens junctions (AJs) and TJs in vitro and in vivo and improved the function of the epithelial barrier by promoting the expression of ZO-1 and occludin. 16 Epac-2 activation reverses the attenuation of TJ proteins and inhibits macrophage production of proinflammatory cytokines via Rap-1 activation. 17 Inhibition of Rap-1 expression reduces the activity of macrophages, 18 and the role of Rap-1 in barrier restoration was further described. Anna and colleagues showed that Epac-2 activation-induced overexpression of Rap-1 inhibited inflammation by attenuating p38 mitogenactivated protein kinase (MAPK) and nuclear factor kappa B (NF-κB) p65 signalling. 19,20 These data suggest that Epac-2 plays key roles in regulating macrophage production of proinflammatory cytokines and maintaining the function of the intestinal barrier in CD patients.
The Il-10 −/− model is the closest animal model to the CD phenotype. 21 We investigated the clinical results and effective mechanisms of Epac-2 in Il-10 −/− mice treated with Me-cAMP (an Epac-2 agonist) or HJC-0350 (an Epac-2 antagonist). Colonic expression of Epac-2 was abnormal in CD patients, and Epac-2 activated the Rap-1 pathway and inhibited the phosphorylation of NF-κB/MAPK in macrophages, which increased the levels of ZO-1 and occludin in intestinal epithelial cells. These results suggest that the targeting of Epac-2 will improve the current treatment of CD.

| Patient specimen preparation
Intestinal specimens were collected from patients with CD (n = 12) who underwent intestinal resection, and uninjured bowel tissue was collected from colon cancer patients (control, n = 16), as described in Appendix S1.

| Drug administration
Il-10 −/− mice with spontaneous colitis were divided into 3 groups (n = 10): Me-cAMP-treated (10 μl, 100 μmol/L, i.p. injection) group, the HJC-0350 (0.5 mg/kg, i.p. injection) group and the Il-10 −/− group as a positive control. 23,24 The mice received treatment every 2 days until the end of 4 weeks. Wild-type C57Bl/6 mice were used as the negative control (WT) group. After 4 weeks, the entire colon of each animal was harvested, and the length of the colon was measured 25 (Appendix S1).

| Colitis symptom assessment
Mice in all groups were evaluated for the extent of colitis using the disease activity index (DAI) score once weekly as previously described. 6 The scoring interval for DAI was 0-5 (Appendix S1).

| Histopathological analysis
Colon tissue was fixed and scored based on the grade of intestinal inflammation. 26 The scoring interval for intestinal inflammation in the mice was 0 to 4 (Appendix S1).

| FITC-dextran permeability assay
After completion of the therapeutic schedule, the mice were fasted under water restriction for 4 h then administered oral FITC-dextran (Sigma, Cat #: F-7250) as previously reported 27 (Appendix S1).

| Bacterial content
Bacteria were isolated from mesenteric lymph nodes (MLNs) and liver tissue using a previously described method 28 (Appendix S1).

2.12
| Cell co-culture assay RAW 264.7 cells were pre-treated with 50 µmol/L Me-cAMP or 0.3 µmol/L HJC-0350 for 1 h then incubated with 1 μg/ml LPS for 24 h. The cell supernatants were collected and centrifuged (12,000 g; 10 min), added to the corresponding Caco-2 cells, and incubated for 24 h (Appendix S1).

| Flow cytometry analysis
The apoptosis rates of Caco-2 cells were determined and analysed using a FACSCalibur flow cytometer (Appendix S1).

| Western blot analysis
Western blotting (WB) was performed as previously described to measure protein levels in tissues and cells. 6 Primary antibodies against occludin, ZO-1, Epac-2, Rap-1, NF-κB p65 (p65), p-NF-κB p65 (p-p65), p-p38 MAPK (p-p38), p38 MAPK (p38) and β-actin were used (Abcam) at dilutions of 1:1,000. Densitometric analysis of protein band intensity was performed with ImageJ (National Institutes of Health, USA). The results are presented as the relative density of each experimental band with respect to the density of the β-actin band normalized to the lowest mean value group.

| Quantitative real-time PCR (qRT-PCR)
qRT-PCR was performed as previously reported to measure the mRNA levels 27 of TNFα, IL-1β, IL-17A and β-actin in tissues. The sequence information for mouse gene-specific primers is provided in the Appendix S1.

| Statistical analysis
The collected experimental data were statistically analysed using SPSS 17.0 (SPSS Inc., Chicago, IL). Independent samples t tests were used to assess data from two groups, and chi-squared tests were used to assess categorical data. Findings with P values less than 0.05 were statistically significant.

| Epac-2 expression was decreased in CD patients
Immunohistochemistry revealed that the expression of Epac-2 decreased significantly in colon tissue of the CD group compared with the control group ( Figure 1A,B). Western blotting analyses of Epac-2, occludin and ZO-1 were performed on 8 randomly selected samples from CD and control patients. The data revealed that the expression of Epac-2, occludin and ZO-1 was suppressed in the inflammatory foci of colon tissue from the CD group compared to uninflamed colon tissue from the control group ( Figure 1C,D). This result indicates a connection between the decrease in Epac-2 expression and the pathological process of CD.
F I G U R E 1 Epac-2 expression was decreased in CD patients. (A, B) Immunohistochemistry was performed to analyse the expression of Epac-2 in the colon of the CD group (n = 12) and control group (n = 16). The results are shown as the mean integrated optical density (IOD) ± SD. (C, D) Eight samples were randomly selected from CD patients and control patients for WB. The results revealed the expression of Epac-2, occludin and ZO-1 in the inflammatory foci of colon tissue from the CD group and uninflamed colon tissue from the control group. The relative protein is shown as fold change of protein expression in the CD group relative to the control group. IOD, integrated optical density, * p < 0.05, compared to the control group

| Epac-2 activation did not affect Caco-2 cell monolayers with LPS-induced damage
To was higher than that in the control group (1.52 ± 0.71%) but was not different from that in the Me-cAMP group (15.56 ± 3.88%) or HJC-0350 group (14.22 ± 3.68%, Figure 4A,B  Figure 4C).
Permeability was determined using FD4. The percentage of FD4 transport to the lower layer was not different among the LPS, Me-cAMP and HJC-0350 groups but exceeded the level of the control group by 2.3-, 2.2-and 2.3-fold, respectively ( Figure 4D). These findings demonstrated that Epac-2 activation by Me-cAMP in Caco-2 cells did not ameliorate LPS-induced damage to cell monolayers.

| Epac-2 activation protected intestinal epithelial cells in the Caco-2 and RAW 264.7 cell coculture system
Caco-2 and RAW 264.7 cells were co-cultured in a Transwell plate, and the TEER value of Caco-2 cells in the LPS group (422.34 Ω/cm 2 ) was less than that of cells in the control group (1047.11 Ω/cm 2 ). The

| Epac-2 suppressed the p65 and p38 signalling pathways
Western blotting results indicated that the Epac-2/Rap-1 pathway was activated in the Me-cAMP group compared to the LPS-induced

| DISCUSS ION
To our knowledge, the present study was the first study to suggest The present study demonstrated that the Epac-2 protein exerted significant protective effects against spontaneous enteritis in mice.
The expression of cytokines is upregulated during the impairment of intestinal barrier function. 32 Our surveillance data indicated that Epac-2 decreased the inflammatory score, DAI and cytokine levels.
Thus, we analysed the indicators associated with changes in intestinal barrier function and discovered that Me-cAMP reversed the attenuation of TJ protein expression and weakened intestinal permeability in Il-10 −/− mice. Tight junction proteins are associated with intestinal barrier function, and occludin and ZO-1 partially reflect the degree of inflammation in the intestine. 33,34 Our data suggested that the positive effect of Epac-2 in Il-10 −/− mice was mediated by increasing the expression of TJ proteins in the colon to some extent.

Disturbance of intestinal epithelial barrier function increases intes-
tinal permeability in CD. 35 We observed remarkable reductions in the levels of serum glucan binding and the possibility of bacterial translocation in the Me-cAMP group. These data demonstrate that Epac-2 has an excellent protective effect against CD. Our study still has certain limitations. For example, LPS cannot induce the apoptosis of Caco-2 cells at low concentration (10 ng/ ml), but it can induce the apoptosis at high concentrations (1 μg/ml, Figure S3). Although there was apoptosis in the LPS group, we could not determine whether it was necrosis or not, which might partly affect the reliability of the results. In addition, the results showed that

ACK N OWLED G EM ENTS
This work was supported partly by funding from the Natural Science

CO N FLI C T O F I NTE R E S T
The authors declare no financial conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.