Exclusive Enteral Nutrition Alleviates Th17-Mediated Inflammation via Eliminating Mechanical Stress–Induced Th17-Polarizing Cytokines in Crohn’s-like Colitis

Abstract

Background and Aims

Exclusive enteral nutrition (EEN) with a liquid diet is the only established dietary treatment for Crohn’s’ disease (CD). However, the mechanism of action of EEN in CD is unclear. T helper 17 (Th17) immune response plays a critical role in CD. We hypothesized that EEN alleviates Th17 response by eliminating mechanical stress–induced expression of Th17-polarizing cytokines.

Methods

A rat model of Crohn’s-like colitis was established by intracolonic instillation of TNBS (65 mg/kg in 250 µL of 40% ethanol). Control rats were treated with saline. We characterized immunophenotypes and molecular changes of the colon in control and colitis rats with and without EEN treatment. Th17 differentiation was determined using coculture assays.

Results

TNBS instillation induced transmural inflammation with stenosis in the inflammation site and a marked increase of Th17-polarizing cytokines interleukin (IL)-6 and osteopontin and the Th17 cell population in the mechanically distended preinflammation site (P-site). EEN treatment eliminated mechanical distention and the increase of IL-6, osteopontin, and Th17 response in the P-site. IL-6 and osteopontin expression was found mainly in the muscularis externa. Mechanical stretch of colonic smooth muscle cells in vitro induced a robust increase of IL-6 and osteopontin. When naïve T cells were cultured with conditioned media from the P-site tissue or stretched cells, Th17 differentiation was significantly increased. Inhibition of IL-6, but not deletion of osteopontin, blocked the increase of Th17 differentiation.

Conclusions

Mechanical stress induces Th17-polarizing cytokines in the colon. EEN attenuates Th17 immune response by eliminating mechanical stress–induced IL-6 in Crohn’s-like colitis.

Graphical Abstract

Graphical Abstract.

Key Messages.

• Exclusive enteral nutrition (EEN) with a liquid diet is an effective treatment to reduce inflammation in Crohn’s disease (CD). However, the mechanism of action of EEN in CD is unclear.

• Our study demonstrates that mechanical stress is increased in a rodent model of CD-like colitis especially at the preinflammation site and induces T helper 17–polarizing cytokine interleukin-6 in the colon. EEN with a liquid diet attenuates T helper 17 immune response by eliminating mechanical stress and interleukin-6 expression.

• Our study reveals a novel mechanism of action of EEN and suggests that mechanical stress may play a critical role in CD.

Introduction

Inflammatory bowel diseases (IBD) such as Crohn’s disease (CD) and ulcerative colitis (UC) are associated with uncontrolled immune response, mediated mainly by effector T helper (Th) cells.¹ Nearly 10 million people experience IBD globally, with an increasing prevalence of IBD in the Western countries. A recent study found that there is a rise in IBD cases in the United Kingdom by 33% from 2006 to 2016.² According to the Centers for Disease Control and Prevention, 1.3% of adults and 80 000 children are living with IBD in the United States. Both life expectancy and quality of life are compromised in IBD patients.³ Among the 2 types of IBDs, CD is characterized by transmural inflammation, fibrosis, and stricture formation in the colon and/or small intestine. The etiology of CD is ambiguous. Surgery is required for nearly 70% of CD patients in their lifetime. Conventional treatments (ie, corticosteroids) are often accompanied with significant side effects, as most of them are involved with the modulation of the immune system. There is a dire need to develop steroid sparring treatments that can be easily accepted by patients.

Dietary treatments are promising candidates in the management of CD patients. Although many different dietary treatments have been tested for IBD treatment, exclusive enteral nutrition (EEN) is the only established dietary treatment for CD.⁴ EEN involves oral or nasogastric tube feeding of a complete liquid diet with the exclusion of normal foods for a defined period (usually 4-8 weeks). Over the last 2 decades, EEN has emerged as a highly effective treatment for the induction of remission in CD. It is low-risk, steroid-sparing, and now the first-line therapy in children with reported remission rates up to 80%.⁵ Health organizations and academic societies (ie, European Crohn’s and Colitis Organization, European Society for Paediatric Gastroenterology Hepatology and Nutrition, North America Society for Pediatric Gastroenterology, Hepatology and Nutrition) now direct clinicians to use EEN to induce remission in young patients.^6,7 It is suggested that corticosteroids, or early immunosuppressive therapy, should be reserved for patients in whom EEN is not an option. Interestingly, there is no therapeutic role for EEN in the treatment of UC, a type of IBD with superficial inflammation and rarely stricture formation. However, the mechanism of action of EEN in CD is still not clear.

Transmural inflammation in CD is associated with Th1 and Th7 immune responses. The Th1 and Th17 cells are differentiated after naive CD4⁺ T cells are exposed to polarizing cytokines such as IL-12 and IL-6, respectively.^8-10 The major cytokines produced by differentiated Th1 and Th17 CD4⁺ T cells are interferon γ (IFNγ) and interleukin (IL)-17/IL-22, respectively. While the importance of Th1 immune response in CD has long been known, mounting evidence suggests that Th17 cytokines play a critical role in gut inflammation in CD.^8-10 Studies have found a positive correlation between IL-17 levels secreted by peripheral blood mononuclear cells and disease activity in IBD patients.¹¹ IBD genome-wide association studies revealed several polymorphisms in Th17-related genes.¹² Th17 cytokines such as IL-17, IL-26, and IL-22 are significantly increased in the inflamed tissues of CD patients. Moreover, Th17 cytokines in turn exacerbate inflammation by promoting Th1 and Th17 immune responses.^13,14 However, current research into the mechanisms of Th17 response in CD is focused mainly on the mucosa and submucosa of the inflamed tissues.^8,15 The sources of polarizing cytokines in the differentiation of Th17 cells are not well understood.

Mechanical stress (MS), associated with edema, inflammatory infiltration, fibrosis, and tissue deformation,¹⁶ is an inevitable feature in transmural inflammation such as CD. This is particularly true in stenotic CD, in which stenosis leads to mechanical distention, a circumferential MS, in the proximal segment.^17,18 We hypothesized that MS might induce upregulation of Th17-polarizing cytokines such as IL-6 and osteopontin (OPN),^{19, 20} contributing to Th17 differentiation in CD. We have developed a TNBS-induced Crohn’s-like colitis model in rodents,^21,22 which mimics well the pathophysiology of CD in humans. In this model, we observed stenosis at the site of TNBS-induced inflammation (denoted as the I-site) and marked lumen distention with fecal retention at the site proximal to I-site (denoted as preinflammation site [P-site]). There was a marked increase of IL-6 and OPN expression not only in the I-site, but also in the mechanically distended P-site, which did not depict superficial inflammation. Strikingly, EEN administration with a liquid diet relieved mechanical distention, attenuated expression of IL-6 and OPN, and significantly reduced inflammation in the CD model. We quantitated Th1 and Th17 immune responses in the I- and P-sites in CD rats treated with normal pellet food and EEN (liquid diet) and tested if MS-induced cytokines contributed to Th17 immune response in CD. If so, the benefit of EEN in CD may rely on its effect to release MS and attenuate MS-induced production of Th17-polarizing cytokines such as IL-6 and OPN.

Methods

Materials

Materials used in the study were obtained from the following sources: TNBS, collagenase D, and Percoll were obtained from Sigma-Aldrich. Flow cytometric antibodies used in the study including anti-mouse TBX21-PE/Cy7, anti-mouse-TCRB-APC, anti-mouse-CD4-PerCP/Cy5.5, anti-mouse-CD8-APC/Cy7, anti-rat CD4-PE-cy5, anti-rat CD3-BV421, anti-rat CD45RA-BV650, anti-rat-IL-17-PE, and anti-rat-IFNγ-FITC were purchased from BD Biosciences, BioLegend, or Thermo Fisher Scientific. The Zombie Aqua fixable viability kit, intracellular staining kit, and nuclear staining kit were from BioLegend.

Animal Experiments

Male 8- to 10-week-old Sprague Dawley rats were purchased from Harlan Sprague Dawley and bred in the animal facility at the University of Texas Medical Branch. Wild-type and OPN^−/− mice were purchased from The Jackson Laboratory (B6.129S6(Cg)-Spp1^tm1Blh/J; stock# 004936). All animal experiments were conducted according to the institutional animal care and use committee guidelines of the University of Texas Medical Branch (#0907051D).

Induction of Crohn’s-like Colitis and Treatments With Liquid Diets

To induce Crohn’s-like colitis, Sprague Dawley rats 8 to 10 weeks old were treated with intracolonic instillation of 250 µL of TNBS (65 mg/kg in 40% ethanol) or saline (sham control) through a catheter inserted 7 cm inside of the anus, as previously described.^21,22 Before the TNBS or saline injection, rats were fasted for 24 hours and treated with bowel cleanser (GoLYTELY) overnight. After the injections, control and TNBS-treated rats were fed ad libitum either with regular pellet food (Picolab Rodent Diet 5053) or EEN liquid diet (Ensure Clear; Abbott). Rats were housed in wire-bottomed cages to prevent their eating of feces or bedding. The body weight and stool consistency of rats were monitored for 7 days. All 4 groups of rats (sham, TNBS, sham + EEN, TNBS + EEN) were euthanized on day 7 for macroscopic and histologic examination. Site-specific colonic tissues were harvested for molecular and functional analysis. The disease activity index (DAI) (on a 6.0 scale) was determined by scoring the extent of body weight change, stool consistency, and hematochezia (each, 0-2 scale) on the day that the animals were euthanized, based on the method of Murthy et al.²³

To study the effect of EEN treatment in CD, we chose the Ensure liquid diet, as it is a complete and balanced nutrition and used as an EEN treatment among CD patients. To determine if other types of liquid diets have the similar effect as the Ensure liquid diet, we carried out a separate study using a rodent liquid diet, a water-suspendible OpenStandard Diet #D11112201L (Research Diets Inc). Control and TNBS-treated rats were fed ad libitum with the rodent liquid diet for 7 days. We assessed the effect of the rodent liquid diet on colon inflammation, distention, and gene expression.

Wild-type and OPN^−/− mice of 10 weeks old were injected with 50 μL of freshly prepared TNBS solution (150 mg/kg body weight) in 50% ethanol as described previously.²⁴ The 2 groups of mice were euthanized after 7 days, and full-thickness colonic tissues were taken for immunological analysis. All rats and mice were kept in the animal facility at the University of Texas Medical Branch.

Colonic Tissue Preparations and Quantitative Real-Time Polymerase Chain Reaction

Rats were euthanized 7 days after saline or TNBS treatment, and the colon was collected in fresh carbonated Krebs buffer. The colonic tissues of rats were taken from 3 different sites (P-site, I-site, and site distal to inflammation [D-site]) as described.²² The I-site is the 2- to 3-cm-long segment of the colon where TNBS was instilled into. The P-site tissue was taken ~2 cm proximal to the upper end of inflammation and the D-site tissue was taken ~2 cm distal to the lower end of inflammation. The counterpart colonic tissues were also taken from the sham rats. Full-thickness tissue, muscularis externa (ME), and mucosa/submucosa (M/SM) were collected from each site, as described previously.²⁵

Total RNA was isolated from the P-, I-, and D-sites of the colon using RNeasy Mini Kit according to the manufacturer’s protocol (Qiagen).^21,22 The messenger RNA (mRNA) was then reverse transcribed into complementary DNA using Bio-Rad iScript reverse transcriptase, which was utilized for quantitative polymerase chain reaction using a TaqMan master mix to run on a Bio-Rad CFX9000 thermocycler. TaqMan probes for IL-6 and OPN were purchased from Thermo Fisher Scientific.

In Vitro Culture of Colonic Smooth Muscle Cells

The rat colonic circular smooth muscle cells (SMCs) were isolated as previously described.^25,26 Briefly, muscle tissue pieces of 0.5 × 0.5 cm² size were incubated in sterile digestion buffer containing HEPES buffer with 1.5 mg/mL collagenase type II and 1.0 mg/mL soybean trypsin inhibitor (Sigma-Aldrich) for 45 minutes at 31 °C. The digested muscle tissue is stained and incubated in HEPES buffer without the enzymes. The cells were dissociated into the buffer, subjected to centrifugation, and reconstituted in Dulbecco’s Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS), 100 U/mL of penicillin G, 100 μg/mL streptomycin sulfate, and 0.25 μg/mL amphotericin B (Invitrogen). Cells were plated into a 6-well BioFlex culture plate coated with type I collagen (Flexcell), grown to ~80% confluence, and then subjected to DMEM/1% FBS for 24 hours before stretch. To apply MS on the cells in culture, cells were elongated with 18% static stretch using an FX-5000 Flexercell Tension Plus System for 1 hour (Flexcell).²⁵ The nonstretched control cells were treated similarly but with no elongation applied. After 3 hours, the cells were collected for mRNA isolation. In the similar setup, conditioned media were collected from stretched and nonstretched cells after 24 hours for further analysis.

Ex Vivo Culture of ME

Ex vivo studies were performed by isolating colonic smooth muscle strips (10 mg) from the distended colon segment proximal to the TNBS inflammation site as described previously.¹⁸ Control muscle strips (10 mg) were taken from the colon counterpart of the sham rats. The strips were incubated in 1 mL of DMEM (+1% FBS) for 24 hours, and the conditioned media were collected.

In Vitro Culture of Naïve T Cells

Naïve T cells from the rat spleens were isolated using an EasySep rat T cell isolation kit (STEMCELL Technologies) per manufacturer’s recommendations. Briefly, spleens were isolated from 8-week-old naïve rats, homogenized, and strained for single-cell suspension. After red blood cell lysis, the cell suspension was collected, centrifuged, and reconstituted in magnetic-activated cell sorting buffer. T cells obtained from magnetic separation (STEMCELL Technologies) were stimulated with anti-CD3/anti-CD28 antibodies.

Histologic Analysis

Colonic tissue from all the rat groups (sham, TNBS, sham + EEN, TNBS + EEN) was isolated and flushed with phosphate-buffered saline to eliminate fecal content. Collected colons were processed for formalin fixation and paraffin embedding as described previously.²⁷ Hematoxylin and eosin staining was done in the University of Texas Medical Branch Histopathology Core, and scoring of inflammation of the specimens was performed by 2 independent investigators blindly using the standard scoring system as described.²²

Flowcytometric Analysis

Colonic immune cell enrichment was performed as described previously.^27,28 Briefly, colons were resected from sham, TNBS, sham + EEN, and TNBS + EEN groups of rats, washed, and incubated with digestion buffer containing intravenous collagenase 1 g/L and DNase. After digestion, cells were stained using a 70-μm strainer, and cells were collected by centrifugation at 1500 rpm for 7 minutes. A density gradient is prepared in a 15 mL conical tube with 80% Percoll at the bottom and 40% Percoll-containing cells as a top layer. Density centrifugation was performed for 20 minutes at 2400 rpm without breaks. The cellular ring that appeared at the interphase was collected and washed with fluorescence-activated cell sorting buffer (phosphate-buffered saline, 1% bovine serum albumin, and 0.01% NaN₃) multiple times. Intracellular and nuclear staining was performed as per the manufacturer’s manual. In rat colonic and in vitro differentiation samples, cells were stained with surface markers anti-CD45RA-BV650, CD4-PeCy5, and CD3-BV421 for 20 minutes and intracellular markers IL-17PE and IFNγ-FITC antibodies for 30 minutes. For mouse colonic samples, colonic cells were isolated and stained with TCRB-APC, CD4-PerCP/Cy5.5, TBX21-PE/Cy7, and CD8-APC/Cy7 antibodies. Cells were washed with 200 µL fluorescence-activated cell sorting buffer and analyzed by flow cytometry (BD LSR Fortessa). Data were acquired from 100 000 gated events per sample and further analyzed with the FlowJo software version 10.0.

Protein Expression Analysis

IL-6 and OPN levels were determined by enzyme-linked immunosorbent assays²⁷ in the conditioned media collected from the ex vivo culture of muscle strips and media collected from the stretched or nonstretched SMC culture according to the manufacturer’s protocol. IL-6 and OPN enzyme-linked immunosorbent assay kits were purchased from R&D Systems.

In Vitro Th17 Differentiation Study

T cells obtained from magnetic separation (STEMCELL Technologies) were stimulated with anti- CD3 (1 ng/mL)/anti-CD28 (3 µg/mL) antibodies. Replicates of a minimum of 3 wells were plated with T cells in different conditions; control with T cell culture media, test wells plated with 1:2 ratio of T cell culture media, and conditioned media obtained from stretched cells (STRCM) and nonstretched cells (NSCM). Separately, NSCM and STRCM were treated with molecular intervention IL-6–neutralizing antibodies (R&D Systems) at a concentration of 0.5 µg/mL. In the small interfering RNA (siRNA) studies, colonic SMCs were transfected with scrambled negative control and siRNAs (siIL-6 and siOPN: titrated concentration 75 nM) according to the manufacturer’s recommendations using RNAiMAX (Invitrogen).²⁹ After transfection, the cells were stretched, and 24 hours later, the conditioned media were utilized for culturing T cells to determine the potential role of the specific molecules. Recombinant rat IL-6 was used as a positive control (20 ng/mL). In the ex vivo setup, T cells were cultured with conditioned media obtained from muscle strips of sham and TNBS P-site. After 72-hour culture of T cells in T cell culture medium, NSCM and STRCM with and without molecular interventions T cells were activated using BioLegend T cell activation cocktail with brefeldin A for 5 hours. Then, the cells were harvested and analyzed for immunophenotyping using antibodies; CD32-purified, Live/Dead Aqua, CD4-PEcy5, CD3-BV421, IFNγ-FITC, and IL-17-PE as described previously.

Statistical Analysis

All values are expressed as mean ± SEM. Statistical analyses were performed using GraphPad Prism (version 9.1.2). Analysis of variance with nonrepeated measures (by Student-Newman-Keuls test) for comparisons of multiple groups and Student’s t test for comparisons of 2 groups was performed to determine the significance. P values ≤.05 were considered statistically significant.

Ethical Considerations

All animal experiments were conducted according to the institutional animal care and use committee guidelines of the University of Texas Medical Branch (Protocol #0907051D).

Results

EEN Reduces Inflammation in Crohn’s-like Colitis in Rats

In rats fed with regular pellet food, intracolonic instillation of TNBS led to localized inflammation and stenosis at the injection site of the colon (I-site) and lumen distention in the segment proximal to the I-site (P-site) as previously described.²² To evaluate the effect of EEN, some sham and TNBS-administered rats were fed exclusively with the Ensure liquid diet for 7 days. We chose the Ensure liquid diet as an EEN treatment in the study, as it is a complete and balanced nutrition and used as an EEN treatment among CD patients. When fed with regular food, rats treated with TNBS displayed significant loss of body weight compared with control rats by day 3 to day 7. In comparison, EEN treatment to TNBS administered rats significantly improved body weight changes (Figure 1A). EEN also reduced DAI in TNBS colitis rats (Figure 1B). Histopathological analysis showed significant transmural inflammation at the I-site with immune cell infiltration across full thickness of the colon wall in the TNBS colitis rats, whereas EEN treatment reduced the distention at P-site and reduced inflammatory infiltration at I-site (Figure 1C, D). The inflammation score was 2.7 ± 0.2 in TNBS colitis rats. However, EEN treatment significantly reduced intestinal inflammation in I-site, with an inflammation score of 1.2 ± 0.1 (Figure 1C). These results suggest that EEN treatment significantly reduces gut inflammation and disease activity in the rat model of CD.

Figure 1.

Beneficial effect of exclusive enteral nutrition (EEN) in the rodent model of Crohn’s-like colitis. (A) Body weight change for 7 days, (B) disease activity index (DAI) (day 7 data), (C) Inflammation index in sham colon (S) and different sites (distended colon site proximal to inflammation [P], inflammation site [I], and nondistended site distal to inflammation [D]) of TNBS colitis (day 7) after regular diet (solid bar) and EEN treatment (streaked bar). (D) Representative microscopic views of sham and different sites of TNBS colitis after regular diet (a-d) and EEN treatment (e-h). n = 5 or 6 in each group. *P < .05 vs sham group control. #P < .05 vs TNBS/regular diet.

To determine if other types of liquid foods have the similar effect as the Ensure liquid diet, we studied the effect of a rodent liquid diet, #D11112201L (Research Diets Inc), in the model of Crohn’s-like colitis. Although the nutrient contents among the regular pellet food, Ensure, and rodent liquid diets are not identical, the calories from the daily consumption of pellet food, Ensure, and rodent liquid diets are similar (~ 80 kcal). We found that the rodent liquid diet had a very similar effect as the Ensure liquid diet on colon inflammation, mechanical distention, and gene expression. These results are reported in the Supplementary Appendix (Supplementary Figure 1). We then used Ensure liquid diet as the EEN treatment through the study.

EEN Eliminates MS in CD

MS, which is associated with tissue edema, inflammatory cell infiltration, and extracellular matrix deposition, is an inevitable pathological change in gut inflammation. MS becomes more significant in CD than in superficial gut inflammation. Transmural inflammation and stenosis are hallmark changes in CD, which lead to lumen distention, a circumferential mechanical stretch to the tissue.^17,18 These changes are clearly depicted in the rat model of CD (Figure 2A, B). As shown in Figure 2, TNBS colitis rats revealed a marked colonic lumen distention at the site proximal to the inflammation site (Figure 2A, left panel). However, EEN treatment with liquid diet released the lumen distention completely in P-site and reduced inflammation at I-site (Figure 2A, right panel). Compared with sham control, the colon circumference was significantly increased in P-site in TNBS colitis rats when fed with the regular diet. However, when rats were fed with the EEN liquid diet, the colon circumference returned to normal levels in the TNBS-treated rats (Figure 2B). These results depict that EEN treatment eliminates MS in the inflamed colon. In the sham rats, EEN treatment significantly reduced fecal mass and pellet size, and increased water content of the fecal pellets (Figure 2C).

Figure 2.

Exclusive enteral nutrition (EEN) relieves lumen distention in Crohn’s-like colitis rats (7 days). (A) Outlook view of colons (top) and macroscopic view of distal colons (bottom). (B) EEN prevents colon circumference increase in TNBS colitis rats. (C) EEN treatment in sham control rats significantly decreased fecal production. Daily fecal production in 1 rat taking regular diet (a) and EEN diet (b) is shown. Average fecal weight (c), number (d), and % water content (e) are shown (n = 4 or 5). *P < .05 vs control rats fed with regular diet. The yellow boxes indicate different sites of colon tissue for the study. D, nondistended site distal to inflammation; I, localized inflammation site; P, distended colon site proximal to inflammation site; S, sham control.

EEN Reduces Th17 Immune Response in CD

We then investigated the changes of the inflammatory cellular milieu in the colon in sham and colitis rats fed with regular food and an EEN diet. For that, we performed immunophenotyping analysis of mononuclear cells in the different sites of the colon. Flow cytometric analysis of immune cells revealed that the frequency of the Th1 CD4⁺CD3⁺IFNγ⁺ population was significantly increased at the I-site of TNBS colitis rats compared with sham control rats (16.9 ± 3.2 vs 4.3 ± 1.5) (Figure 3A-C). EEN treatment reduced Th1 cell population to 10.9 ± 4.1 at the I-site. On the other hand, the Th17 CD4⁺CD3⁺IL-17⁺ population was found significantly increased at the P-site in the colitis rats as compared with sham control rats (23.4 ± 6.2 vs 2.4 ± 0.3). EEN treatment significantly reduced the Th17 cell population at the P-site (9.8 ± 4.3) as compared with TNBS colitis rats when fed with a regular diet (Figure 3C). These results suggest a strong link between MS and Th17-mediated inflammation in CD.

Figure 3.

Site-specific changes of T helper 1 (Th1) and Th17 cell populations in Crohn’s disease–like colitis. Colonic mononuclear cells were isolated from control and TNBS-treated rats with and without exclusive enteral nutrition (EEN) treatment, and flow cytometric analysis was performed. (A) Representative dot plots of flow cytometric analysis of CD3⁺ T lymphocytes gated on live CD4⁺ cell populations. Representative of pooled experiments with 7 rats. (B) Representative dot plots. (C) Summary analysis of the frequency of Th1, CD4 + 3 + interferon γ (IFNG) + population and Th17, CD4 + 3 + interleukin (IL)-17 + population from rat colonic localized inflammation sites (I) and distended colon sites proximal to inflammation site (P) after 7 days of TNBS treatment with and without EEN administration. *As compared with sham control. #As compared with TNBS control. FSC, forward scatter; FSC-A, forward scatter-area; FSC-H, forward scatter-height; S, sham control; SSC, side scatter.

EEN Attenuates Mechanotranscription of Proinflammatory Mediators

MS applied to cells induces molecular changes such as transcriptional regulation of mechanosensitive genes, a process referred to as mechanotranscription,^17,18 which further activates downstream signaling pathways leading to functional changes. To determine if mechanotranscription of certain proinflammatory mediators may be involved in gut inflammation and immune phenotype in CD, we determined gene expression of OPN and IL-6 in different sites of the colon in the rat model of CD. Both OPN and IL-6 have been reported to be involved in the development of the Th17 immune response.^20,21 We found that the expression of IL-6 and OPN mRNAs was upregulated dramatically in the I-site and the mechanically distended proximal site (P-site) but not in the D-site (Figure 4A-a, B-a). This result indicates that the upregulation of IL-6 and OPN in P-site is due to MS. This was further confirmed by studies with EEN treatment with a liquid diet. When rats were treated with EEN, which released MS as shown above (Figure 2), gene expression of IL-6 and OPN in colitis rats especially in the P-site was attenuated to nearly the level as in control rats (Figure 4A-b, B-b).

Figure 4.

Effect of exclusive enteral nutrition (EEN) on mechanosensitive gene expression. (A) Interleukin (IL)-6 and (B) osteopontin (OPN) messenger RNA (mRNA) expression was markedly induced in localized inflammation sites (I) and distended colon sites proximal to inflammation site (P), but not the nondistended site distal to inflammation (D) in TNBS rats (7 day) when fed with a regular diet (a). However, EEN treatment blocked upregulation of IL-6 and OPN (b). (a) n = 5; (b) n = 4. *P < .05 vs sham control (S). CD, Crohn’s disease.

MS Induces Upregulation of IL-6 and OPN in the Colonic SMCs In Vivo and In Vitro

To determine the cellular sources of increased IL-6 and OPN in the P-site, we compared their expression in M/SM and ME. It was shown that the increased expression of IL-6 and OPN mRNAs was not detected in M/SM, but rather only in the smooth muscle layer (Figure 5A). Moreover, we isolated colonic smooth muscle strips (10 mg per sample) from the P-site of TNBS colitis rats and the counterpart of sham control rats. Protein expression analysis revealed that IL-6 and OPN levels increased significantly in the conditioned media obtained from the muscle strips of the P-site compared with that of the control tissue (Figure 5B). These results suggest that MS induces mechanotranscription of IL-6 and OPN, mainly in the SMCs.

Figure 5.

Mechanical stress–induced expression of interleukin (IL)-6 and osteopontin (OPN) in colonic smooth muscle cells in vivo and in vitro. (A) Expression of IL-6 and OPN messenger RNAs (mRNAs) was markedly increased in muscularis externa at the proximal segment of TNBS rats (7 day) when rats were fed with regular diet. (a) n = 5; (b) n = 5. *P < .05 vs sham control (S). (B) Protein expression of IL-6 and OPN increased significantly in conditioned media obtained from the muscle trip culture of TNBS distended colon site proximal to inflammation site (P) as compared with the muscle strip culture of sham colons. Mechanical stress–induced expression of IL-6 and OPN (C) mRNA and (D) protein in the culture of rat colonic smooth muscle cells. Cells were encountered with mechanical stretch (STR) and without stretch, nonstretched (NS). n = 6. *P < .05 vs S. M/SM, mucosa/submucosa; ME, muscularis externa.

To further determine if MS is a factor independent of inflammation in the induction of IL-6 and OPN in the gut, we applied MS directly on the primary culture of rat colonic SMCs.³⁰ Colonic SMCs were isolated from naïve rats. The primary culture of colonic SMCs was subjected to the static mechanical stretch of 18% elongation for 1 hour using the Flexercell stretching system. Gene expression analysis was performed in nonstretched (NS) and stretched (STR) SMCs. We found that MS significantly increased mRNA expression of IL-6 (2.2-fold) and OPN (3.4-fold) compared with NS control samples (Figure 5C). Protein expression was also determined in conditioned media obtained from NS and STR samples 24 hours after 1 hour of mechanical stretch. Enzyme-linked immunosorbent assay analysis depicted a significant increase of IL-6 (239.4 ± 16.3 pg/mL vs 119.3 ± 5.9 pg/mL; P < .05) and OPN (971.3 ± 366.0 pg/mL vs 260.8 ± 28.0 pg/mL; P < .05) in stretched samples compared with NS control samples (Figure 5D). These results combined with the in vivo and ex vivo findings suggest that MS is an independent factor in the induction of certain Th17-polarizing cytokines (eg, IL-6 and OPN), especially in gut SMCs.

MS-Induced IL-6 Plays a Crucial Role in Th17-Mediated Inflammation

We next determined if MS-induced IL-6 and/or OPN in the primary culture of SMC are involved in Th17 differentiation. Naïve T cells were isolated from the spleen of naïve rats and cultured in the presence of T cell culture media (control) and NSCM and STRCM muscle cells. Our results revealed a significant increase in the CD4⁺CD3⁺IL-17⁺ population with STRCM (4.5 ± 0.8) as compared with NSCM (1.7 ± 0.4) (Figure 6A), suggesting that MS is an independent factor promoting Th17 differentiation. To determine if IL-6 or OPN is involved in MS-induced Th17 differentiation, we conducted molecular inhibition studies using siRNA against IL-6 and OPN (siIL-6 and siOPN, respectively) as well as a neutralizing antibody against rat IL-6. Recombinant IL-6 was used as a positive control. We found that neutralization or silencing of IL-6 in SMCs showed a significant decrease in MS-induced Th17 differentiation (Figure 6B, C) as compared with SMCs treated with respective control samples. However, silencing OPN in SMCs did not exert any effect on MS-induced Th17 differentiation (Figure 6D).

Figure 6.

Role of colonic smooth muscle cell (SMC)–derived interleukin (IL)-6 and osteopontin (OPN) in mechanical stress–enhanced T helper 17 (Th17) differentiation in vitro and in Th17 immune response in Crohn’s-like colitis in rats. (A) Summary of the frequency of Th17, CD4⁺CD3⁺IL-17⁺ population from the T cells cultured in the presence of conditioned media obtained from stretched (STRCM) and nonstretched (NSCM) colonic SMCs. (B) Effect of neutralizing antibody against IL-6 on mechanical stress–enhanced Th17 differentiation. As a positive control, recombinant rat IL-6 (20 ng/mL; Sigma-Aldrich) was used separately in the study. Frequency of the CD4⁺CD3⁺IL-17⁺ population from the T cells cultured in the presence of conditioned media from gut SMCs treated with (C) small interfering IL-6 and (D) small interfering OPN . (A-D) n = 6. *P < .05 vs NSCM in A and B. *P < .05 vs scramble small interfering RNAs (siRNAs) in C and D. (E) T cells were isolated from control rat spleens and cultured in the presence of conditioned media harvested from 24-hours culture of muscle strips excised from sham rats and P-site of TNBS treated rats (TNBS P site). Shown are summary of the frequency of Th1 (CD4⁺CD3⁺IFNγ⁺) and Th17 (CD4⁺CD3⁺IL-17⁺) population in the samples. n = 8. *P < .05 vs sham. (F) Effect of neutralization of IL-6 on Th17 differentiation in T cells treated with conditioned media obtained from samples of sham and TNBS distended colon site proximal to inflammation site (P). Recombinant IL-6 was used as a positive control. n = 8. *P < .05 vs control, #P < .05 vs TNBS P site. Scr, scrambled.

To determine if IL-6 is involved in the increased Th17 response in vivo in the CD-like colitis model, we isolated smooth muscle strips from the P-site of colitis rats and the counterpart of sham control. Conditioned media were collected after 24-hour incubation of the tissue strips and applied to naïve T cells for another 72 hours. T cell differentiation to Th17 and Th1 populations was determined by flow cytometry using intracellular markers IL-17 and IFNγ, respectively. Conditioned media from the P-site did not significantly increase the number of IFNγ-expressing T cells (Figure 6E). However, media from the P-site significantly increased IL-17–expressing cells (CD4⁺CD3⁺IL-17⁺ population) as compared with sham control media (27.8 ± 2.3 vs 16.6 ± 4.2; P < .05). Importantly, treatment with anti-IL-6–neutralizing antibody attenuated the increase of IL-17–expressing cells by the conditioned media of the P-site (Figure 6F).

To further confirm if OPN plays a role in Th17 differentiation, TNBS colitis was induced in wild-type and OPN KO mice. After 7 days, mouse colons were collected for immunophenotyping studies. We found that there was no change in Th17 population (TCRb⁺CD4⁺RORg⁺ cells) between sham wild-type and OPN KO mice. There was no significant difference in Th17 immune response in the inflamed colons of OPN KO mice as compared with that of the wild-type mice (Supplementary Figure 2). These results indicate that it is IL-6, but not OPN, that mediates MS-induced Th17 differentiation in the colon.

Discussion

Although the benefits of EEN in reducing inflammation and inducing remission in active CD have long been confirmed, the mechanisms underlying the benefits are still poorly understood. As IBD may be associated with microbiota dysbiosis in the gut, many people had suspected that the benefits of EEN in CD might be due to its effect to correct dysbiosis. However, although all forms or brands of EEN might be effective in inducing remission of inflammation, they led to vastly different microbiome changes in the gut.^31-34 More importantly, EEN treatment reduces bacterial diversity and richness in the gut.^33,35 These changes are unlikely to account for the benefits of EEN in CD, as reduced bacterial diversity and richness would predispose the gut to more inflammatory changes. Other investigators observed that EEN treatment might improve barrier function in CD³⁶ and reduce the production of proinflammatory cytokines such as IL-6 and IL-8.^31,37 However, the mechanisms for the improved barrier function and reduced cytokine production are unclear.

Our study suggests a novel mechanism underlying the beneficial effect of EEN on CD. We have found that EEN eliminates MS and reduces the mechanotranscription of proinflammatory mediators such as IL-6 and OPN in a CD-like colitis model. We propose that these effects may represent a key mechanism of action of EEN in CD. CD is characterized by transmural inflammation and fibrosis- or inflammation-associated stenosis.³¹ These changes further lead to mechanical distention with massive retention of intraluminal contents proximal to inflammation or stenosis.³⁸ Our rodent model of CD-like colitis mimics well these features in CD, with stenosis in the inflammation site and distention prior to stenosis. MS is increased in the tissues of the inflammation site, and particularly in the segment prior to inflammation (P-site).^18,22 We found that EEN treatment with a liquid diet dramatically reduces fecal mass and significantly increases the water content in the stools. These led to little, if any, fecal accumulation in the colon in control rats and significantly reduced fecal retention in the colitis rats. Consequently, mechanical distention in the P-site is released and stenosis is improved, leading to released MS in the tissue. Further quantitative measurements show that released MS in colitis is associated with attenuated mechanotranscription of proinflammatory mediators and reduced inflammation in the CD model.

It is noteworthy that the increased expression of mechanosensitive proinflammatory mediators IL-6 and OPN in the mechanically distended P-site was mainly from the ME but not from mucosa and submucosa. This is consistent with previous findings that gut SMCs are exceptionally sensitive to mechanical distention, compared with other types of cells in the bowel wall.¹⁷ In fact, we found that mechanical stretch in vitro induced robust expression of IL-6 and OPN in the primary culture of colonic SMCs. These data highlight the pathogenic significance of gut smooth muscle in the mechanically distended prestenotic site in CD. CD may be a broader and deeper problem than what we have thought. It is a broader problem, as the seemingly normal but distended bowel segment prior to inflammation or stenosis is the main site to produce mechanosensitive molecules such as IL-6 and OPN, and a root cause of inflammatory response. It is a deeper problem, as we found that the deep ME is the main tissue source of IL-6, OPN, and other mechanosensitive molecules in the CD model. The mechanosensitive changes in the distended proximal site may very likely contribute to the recurrence of inflammation, which occur most frequently in the prestenotic region in CD.^39,40 Unfortunately, the impact of MS in CD has not been well recognized, as the deep ME layer in the seemingly normal P-site is not accessible by endoscopy or conventional biopsy.

Mounting evidence suggests that both Th1 and Th17 immune responses are involved in the development of chronic inflammation in CD.^8,15 Current research into the mechanisms of Th1 and Th17 responses in CD has focused mainly on the mucosa and submucosa of the inflamed site. However, when we sought to compare the immune responses in different sites of the CD-like colitis model, we found a Th1-dominant immune response in the I-site and a Th17 immune response at the distended P-site. Unlike the I-site, the P-site shows marked distention but no clear mucosa inflammation. Thus, the site-specific immune responses in the I- and P-sites of our CD model indicate that MS may be responsible for Th17 differentiation. This is confirmed by the results of EEN treatment, as EEN treatment with liquid diet releases mechanical distention in the P-site. We found that EEN treatment significantly reduced Th17 immune response at the P site. These in vivo studies suggest that MS plays a critical role in the development and maintenance of T cell immune responses, particularly Th17 response in the gut.

To distinguish the effect of MS from that of inflammation in the model of CD-like colitis, we took a unique approach to compare the site-specific changes in the I- and P-sites and focus mainly on the mechanically distended P-site. However, this is an inflammation model, and inflammation may exert systemic responses. To confirm the role of MS in Th17 differentiation, we turned to the in vitro approach, in which cells were exposed directly to MS. We found that conditioned media from mechanically stretched colonic SMCs in primary culture significantly increased CD4 cells’ differentiation into Th17 cells, compared with the media from nonstretched control cells. Together with the findings in the in vivo studies, our work demonstrates convincingly that MS is an independent factor promoting Th17 cell differentiation in the gut. As we found that expression of Th17-polarizing cytokines IL-6 and OPN is potently induced by MS, we used the in vitro setting to test if IL-6 and/or OPN mediate MS-induced Th17 differentiation. It is found that the knockdown of IL-6, but not of OPN, in the SMCs attenuated MS-induced Th17 differentiation, suggesting that MS induces Th17 differentiation via mechanotranscription of certain proinflammatory mediators such as IL-6.

To confirm the pathogenic role of MS-induced IL-6 in the development of the Th17 immune response in the CD model, we focused on the tissues from the P-site of colitis rats in which IL-6 and Th17 responses are found significantly increased compared with sham. Conditioned media of the P-site tissue increased Th17 differentiation, which was significantly attenuated by the anti-IL-6 antibody. An additional study showed that OPN knockout does not attenuate the Th17 immune response in colitis. These data confirm that MS-induced IL-6, but not OPN, plays a critical role in the development of Th17 immune response in the CD model.

Taken together, our in vivo, in vitro, and ex vivo studies demonstrate that MS plays a crucial role in the development of Th17 immune response by mechanotranscription of IL-6 in CD-like colitis. These data suggest that MS is an independent factor in the development of the Th17 immune response in the gut. EEN treatment with a liquid diet releases MS, eliminates MS-induced expression of IL-6, and reduces the Th17 immune response in the CD model. These novel findings suggest that the benefits of EEN in reducing inflammation in CD may rely on its effects in releasing MS and eliminating MS-induced proinflammatory mediators in the gut.

Funding

This study was supported in part by the U.S. Department of Defense (W81XWH2010681 to X-ZS) and the National Institutes of Health (R01 DK124611 to X-ZS).

Author Contributions

All the authors have read and approved the final version of the paper and contributed to the work. XZS obtained the funds, designed the study, performed the research, and drafted the paper. RG, KZ, YML, and JCJ performed the research, analyzed data. RG also contributed to draft manuscript. YC and SC contributed to the establishment of the animal model, data explanation, inflammation scoring, and manuscript revision.

Conflicts of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Data Availability

Data are available upon request.