Sacral nerve stimulation enhances early intestinal mucosal repair following mucosal injury in a pig model

Abstract

Key points

• Reducing intestinal epithelial barrier (IEB) dysfunctions is recognized as being of major therapeutic interest for various intestinal disorders.

• Sacral nerve stimulation (SNS) is known to reduce IEB permeability.

• Here, we report in a pig model that SNS enhances morphological and functional recovery of IEB following mucosal injury induced via 2,4,6‐trinitrobenzenesulfonic acid. These effects are associated with an increased expression of tight junction proteins such as ZO‐1 and FAK.

• These results establish that SNS enhances intestinal barrier repair in acute mucosal injury. They further set the scientific basis for future use of SNS as a complementary or alternative therapeutic option for the treatment of gut disorders with IEB dysfunctions such as inflammatory bowel diseases or irritable bowel syndrome.

Abstract

Intestinal epithelial barrier (IEB) dysfunctions, such as increased permeability or altered healing, are central to intestinal disorders. Sacral nerve stimulation (SNS) is known to reduce IEB permeability, but its ability to modulate IEB repair remains unknown. This study aimed to characterize the impact of SNS on mucosal repair following 2,4,6‐trinitrobenzenesulfonic acid (TNBS)‐induced lesions. Six pigs were stimulated by SNS 3 h prior to and 3 h after TNBS enema, while sham animals (n = 8) were not stimulated. The impact of SNS on mucosal changes was evaluated by combining in vivo imaging, histological and functional methods. Biochemical and transcriptomic approaches were used to analyse the IEB and mucosal inflammatory response. We observed that SNS enhanced the recovery from TNBS‐induced increase in transcellular permeability. At 24 h, TNBS‐induced alterations of mucosal morphology were significantly less in SNS compared with sham animals. SNS reduced TNBS‐induced changes in ZO‐1 expression and its epithelial pericellular distribution, and also increased pFAK/FAK expression compared with sham. Interestingly, SNS increased the mucosal density of neutrophils, which was correlated with an increase in trypsin and TGF‐β1 levels compared with sham. Finally, SNS prevented the TNBS‐induced increases in IL‐1β and IL‐4 over time that were observed with sham treatment. In conclusion, our results show that SNS enhances mucosal repair following injury. This study highlights novel mechanisms of action of SNS and identifies SNS as a new therapy for diseases with IEB repair disorders.

Key points

• Reducing intestinal epithelial barrier (IEB) dysfunctions is recognized as being of major therapeutic interest for various intestinal disorders.

• Sacral nerve stimulation (SNS) is known to reduce IEB permeability.

• Here, we report in a pig model that SNS enhances morphological and functional recovery of IEB following mucosal injury induced via 2,4,6‐trinitrobenzenesulfonic acid. These effects are associated with an increased expression of tight junction proteins such as ZO‐1 and FAK.

• These results establish that SNS enhances intestinal barrier repair in acute mucosal injury. They further set the scientific basis for future use of SNS as a complementary or alternative therapeutic option for the treatment of gut disorders with IEB dysfunctions such as inflammatory bowel diseases or irritable bowel syndrome.

Abbreviations

ENS

enteric nervous system

FAK

focal adhesion kinase

GI

gastrointestinal

HPS

hemalun‐phloxine‐saffron

IBD

inflammatory bowel disease

IEB

intestinal epithelial barrier

IEC

intestinal epithelial cell

IFN

interferon

IL

interleukin

pCLE

probe‐based confocal laser endomicroscopy

SNS

sacral nerve stimulation

TNBS

2,4,6‐trinitrobenzenesulfonic acid

UCEIS

ulcerative colitis endoscopic index of severity

Introduction

The intestinal epithelial barrier (IEB) achieves two major and distinct tasks: it enables the absorption of nutrients, and at the same time it controls the passage of pathogens and/or toxins (Groschwitz & Hogan, 2009). The IEB is formed by a monolayer of proliferating and differentiating intestinal epithelial cells (IECs). IECs constitute an intrinsic barrier forming a cell monolayer maintained by complex protein networks that mechanically links adjacent cells and regulates the passage of molecules between (paracellular permeability) and across (transcellular permeability) IECs. This function is reinforced by the ability of IECs to secrete mucus, electrolytes and antimicrobial peptides (Johansson et al. 2008; Pastorelli et al. 2013). Tight junctions are major actors in this network, controlling paracellular permeability and are, interestingly, also involved in repair processes and IEC polarity (Neunlist et al. 2013).

Dysfunctions of the IEB, characterized by an increase in paracellular and/or transcellular permeability and defects in IEB repair, have been associated with various intestinal disorders, including inflammatory bowel diseases, coeliac disease, irritable bowel syndrome and diarrhoeal infection (Groschwitz & Hogan, 2009; Catalioto et al. 2011). Reducing barrier dysfunction, or enhancing repair processes, is increasingly recognized as being of major therapeutic interest (Neurath & Travis, 2012; Bischoff et al. 2014; Lopetuso et al. 2015). For instance, reducing paracellular permeability with zonulin peptide inhibitor was shown to reduce the development of intestinal inflammation in interleukin (IL)‐10^−/− mice (Arrieta et al. 2009). Pharmacological blockade of tight junctions has also been shown to reduce pain in a model of stress‐induced irritable bowel syndrome (Annaházi et al. 2013). Favouring IEB repair is also of interest in inflammatory bowel diseases (IBDs), as several pharmacological strategies have been developed to enhance IEB repair, such as in vivo delivery of exogenous Annexin‐A1 mimetic peptide after biopsy‐induced injury (Leoni et al. 2015), or delivery of an active molecule onto an IEC layer in vitro (Andújar et al. 2013; Neurath & Travis, 2012).

The gastrointestinal tract (GI) mucosa, and in particular the IEB, is highly innervated, with a dense neuronal network of both intrinsic and extrinsic origin (Furness, 2012). From a functional point of view, activation of the enteric nervous system (ENS) has been shown to enhance IEB functions and repair (Neunlist et al. 2013). In this context, neurostimulation of the GI tract emerges as of major interest to target IEB functions. For instance, vagal stimulation reduced barrier dysfunctions induced in an animal model of skin burn, probably by activating key components of the ENS, i.e. enteric glial cells (Costantini et al. 2010). We have consistently shown that sacral nerve stimulation (SNS) reduces paracellular permeability in the rectum in a porcine model as early as 3 h after beginning SNS (Meurette et al. 2012). Furthermore, SNS was shown to reduce IEB dysfunctions induced ex vivo by mucosal challenge with PAR‐2 agonists (Provost et al. 2015).

Nevertheless, if SNS regulates permeability, its implication in the repair processes following mucosal lesions remains to be demonstrated. In this context, the objective of this study was to characterize the impact of SNS on IEB integrity in a model of 2,4,6‐trinitrobenzenesulfonic acid (TNBS)‐induced lesions of the rectal mucosa in the pig.

Methods

Ethical approval

All experiments were conducted in accordance with the French Veterinary Regulations and Ethics Committee standards and approved by the local animal care committee (CEEA‐PdL 2015.23) which follows French (R214‐87 to R214‐137) and European (directive 2010/63/UE) guidelines. All the investigators understand the ethical principles under which the journal operates and that their work complies with the animal ethics checklist of the journal.

Animals

A total of 17 large white pigs (Sus scrofa, strain Large White × Landrace), each weighing between 30 and 40 kg, were used in this study. All animals were reared and transported under conditions specified in the French Animal guidelines 2013. Pigs were housed in groups of two in standard laboratory conditions. Animals had ad libitum access to food and water. Experiments were performed in a dedicated surgical facility of the research laboratory Ecole de chirurgie/LGA (Laboratoire des Grands Animaux, Inserm U1064; Agreement: D44010). All pigs were 6 months old and had achieved sexual and brain maturity (Lind et al. 2007). Animals underwent surgery under general anaesthesia in a right lateral position (pre‐anaesthasia with 15 mg kg^–1 Zolazepam/Tiletamine intramuscularly and anaesthesia induced with 5% isoflurane and 60% nitrous oxide, thereafter maintained with 2% isoflurane infusion). SNS was started after at least a mean 30 min of steady state anaesthesia to minimize the influence of drugs of the induction phase on the central or peripheral nervous systems. To reduce pain of the animal after anaesthesia (period T_6–24h), each animal received analgesia (paracetamol 25 mg kg⁻¹, Nalbuphine 0.2 mg kg⁻¹; i.v.). At the end of the protocol, animals were killed in accordance with Annex IV of the European Directive 2010/63/EU, i.e. under general anaesthesia with the animal receiving an intravenous injection of pentobarbital (3.644 g). The sequence of experiments is summarized in Fig. 1.

Figure 1.

Sequence of experiments

Bilateral SNS

The SNS procedure was performed as previously described (Meurette et al. 2012). Briefly, after cleaning with an iodine‐based solution and draping, percutaneous bilateral placement of electrodes was performed using the conventional peripheral nerve evaluation test kit for humans (Medtronic number 041828‐004; Medtronic Inc., Minneapolis, MN, USA) and using surface landmarks at S3. Stimulation was performed in eight animals (SNS group) for 6 h, 3 h before and 3 h after TNBS enema, while the other animals (sham group) had electrodes placed without stimulation. Stimulation parameters were 14 Hz for 210 μs and voltage was adjusted between 0.5 and 1.5 V for each animal depending on the threshold for SNS‐induced anal contraction.

TNBS enemas

TNBS was used to induce mucosal lesions. The TNBS concentration and contact time were selected from previous rodent studies, namely 15 mg ml⁻¹, 50% ethanol, 15 min (Negaard et al. 2010; Brenna et al. 2013), while the volume (100 ml) was estimated in comparison with human rectal anatomy (Gandarillas & Bas, 2009).To limit the passage of stools and to delimit the contact area of TNBS, a latex plug (consisting of a glove filled with sterile compress) was inserted 15 cm from the anal margin and an anal strapping (anal encirclement with temporary suture) was used to prevent anal leakage of TNBS. A urinary Foley catheter was positioned in the anus before closing, to allow instillation and removal of TNBS.

Endoscopic and endomicroscopic analyses

The rectal mucosa was analysed by white light rectoscopy (Fuginon) and by confocal endomicroscopy (pCLE; Cellvizio 488, Mauna Kea Technologies, Paris, France), before, and 1, 3 and 24 h after TNBS enema. The animals received 25 ml sodium fluorescein solution at 1% Faure (Laboratoire Théa, France) intravenously before pCLE image acquisition. An optical fibre probe was inserted through the operative channel of the endoscope and was gently applied onto the mucosa to record the fluorescence signal, as previously described in pig and in humans (Musquer et al. 2013; Coron et al. 2015). Endoscopy images were analysed in a blinded manner by a gastroenterologist using the ulcerative colitis endoscopy index of severity (UCEIS; Dignass et al. 2012). Using specific software (IC viewer), pCLE images were exported and assembled for further quantitative analyses. Based on criteria used in previous studies (Kiesslich et al. 2007; Li et al. 2010; Musquer et al. 2013; Coron et al. 2015), we developed a method of quantifying parameters that included: (1) crypt – shape and regularity, ratio between long and short axis; (2) mucosal structure – space between crypts and crypt density; and (3) fluorescein intensity in crypt wells and in the lamina propria.

Ex vivo assessment of rectal paracellular and transcellular permeability

Before (each animal was its own control), and 1, 3 and 24 h after TNBS enemas, serial surgical biopsies were sampled and placed in ice‐cold Krebs solution before transfer to the laboratory for immediate processing and analysis. For each specimen, the mucosa was separated from the submucosa (in the plane of the submucosal blood vessels). Specimens of rectal mucosa were then mounted in dedicated Ussing chambers (Transcellab, TBC, Paris, France) exposing a mean surface of 0.0314 cm². Specimens were bathed on each side in 1.5 ml Ham's Nutrient Mixture (HAM/F12, Invitrogen, Paris, France).The medium was continuously oxygenated and maintained at 37°C with a gas flow of 95% O₂ and 5% CO₂. After an equilibration period of 30 min, 275 μl apical medium was replaced with 150 μl fluorescein–5.6 sulfonic acid (1 mg ml⁻¹, 400 D, Invitrogen) and 75 μl horseradish peroxidase (10 mg ml⁻¹, 44 kDa; Sigma Aldrich, Saint‐Quentin‐Fallavier, France). The fluorescence levels of 150 μl aliquots taken at the basolateral side were measured at 30 min intervals over a period of 180 min using a fluorimeter (Varioskan, Thermo SA, France). Paracellular permeability was determined by calculating the slope of the fluorescence intensity over time. Transcellular permeability was assessed by measuring the fluorescence resulting from the enzymatic activity of HRP, measured on a fluorimeter from 20 μl samples recovered at 1 h intervals (during a total of 3 h) from the basolateral side.

Histological evaluation

Following microdissection, mucosal specimens were fixed in paraformaldehyde (4% in PBS) for 8 h. After tissue washing in PBS, the specimens were dehydrated and embedded in paraffin. Sequential sections of mucosa (5 μm) were obtained and stained with hemalun‐phloxine‐saffron (HPS). Each section was observed under an Olympus IX 50 microscope (Olympus Inc., Centre Valley, PA, USA). Pictures were acquired using a DP71 digital camera (Olympus) connected to a computer through a frame grabber card (Cell B software, Olympus).

Immunohistochemistry

Following microdissection, mucosal specimens were fixed in paraformaldehyde (4% in PBS). Immunohistochemical methods for ZO‐1 staining were adapted from Clairembault et al. (2015).

Quantification of mRNA by real‐time quantitative PCR

Total RNA from mucosa was extracted from cells using Nucleospin RNA II (Macherey‐Nagel, Hoerdt, France) columns according to the manufacturer's instructions. Reverse transcription was performed on purified RNA. The RNA level was determined using Nanodrop software. Samples were processed to obtain 1 μg RNA in a final volume of 15 μl. Total RNA, (n6) random hexamers (1 μl; 265 ng μl⁻¹; GE Healthcare, Orsay, France) with dNTPs (1 μl; 10 mm; Gibco, Cergy‐Pontoise, France) were used to synthesize single‐stranded cDNA using Superscript II Reverse Transcriptase (0.5 μl, 200 U μl⁻¹; Invitrogen) according to the manufacturer's instructions, in a total volume of 25 μl. Incubation was performed at 42°C for 60 min. Amplification conditions for ZO‐1, Claudin‐1, Occludin and S6 templates were optimized for the RotorGene 2000 device (Ozyme, Saint‐Quentin en Yvelines, France). PCRs were performed with 2 μl cDNA, 0.1 μl solution of SYBR Green I diluted 1:100 (Sigma), 1 μl (10 μm) each primer, 1 μl (10 mm) dNTPs and 0.4 μl Titanium Taq DNA Polymerase kit (Ozyme), according to the manufacturer's instructions. Cycling conditions were as follows: 5 min at 95°C, with amplification for 35 cycles with denaturation for 5 s at 95°C, annealing 15 s at 60°C and extension 20 s at 72°C. Primers were chosen on separate exons to amplify cDNA. A standard curve was generated with serial dilutions of control cDNA by plotting the relative amounts of these dilutions against the corresponding Ct (cycle threshold) values. The level of expression of each gene was calculated from these standard curves using Rotor Gene software (Ozyme). Samples were tested in duplicate and the mean values were used for quantification using the 2^−ΔΔCt method, as previously described.

Western blot

Proteins were extracted from sample lysates containing RA1, according to the manufacturer's instructions (Macherey‐Nagel). The proteins were precipitated, washed and resuspended in Laemmli buffer (SDS, β‐mercaptoethanol and bromophenol blue). The samples were subjected to electrophoresis on SDS‐polyacrylamide gels and analysed by immunostaining. Antibodies used were anti‐ZO‐1 (Life Technologies, Carlsbad, CA, USA), anti‐pY397‐FAK (Ozyme), anti‐FAK (Santa Cruz Biotechnologies, Santa Cruz, CA, USA) and anti‐β‐actin (Sigma Aldrich).

Multiplex cytokine assay

Blood samples were taken after each biopsy, at the biopsy site. Serum was obtained after centrifugation (15 min, 3000 g, without brake) and stored at −80°C prior to utilization. Cytokine concentration was measured in 10 μl of serum using the Swine cytokine 7‐plex Mag panel (Life Technologies; LSC0001M) or TGF‐β1 Single Plex Magnetic bead kit (Merck‐Millipore, Billerica, MA, USA; TGFBMAG‐64K‐01), accordingly to the manufacturers’ instructions. The analysis was performed with a Bio‐plex 200 (Bio‐Rad, Hercules, CA, USA) at the Therassay facility (Nantes, France).

Mucosal elastase and trypsin activity assays of biopsy supernatants

Neutrophil elastase activity was measured using the substrate methyl‐O‐succinyl‐Ala‐Ala‐Pro‐Val‐AMC (elastase substrate V; EMD Chemicals, Billerica, MA, USA). Then, 50 μl elastase (200 mU ml⁻¹) in Hanks’ balanced salt solution containing 10 mm Hepes, pH 7.0, 1.5 mm MgCl₂, 1.5 mm CaCl₂ and 0.2% Nonidet P‐40 was incubated with 50 μl elastase‐specific substrate, methyl‐O‐succinyl‐Ala‐Ala‐Pro‐Val‐AMC (200 μm) (elastase substrate V; EMD Chemicals). The fluorescence from the free AMC, as an index of proteolytic activity, was measured every 10 min for 60 min using a Victor X4 fluorescence plate reader (PerkinElmer Life Sciences, Waltham, MA, USA). Trypsin activity was measured using the substrate, t‐butoxycarbonyl‐QAR‐AMC, with similar conditions as for neutrophil elastase.

Statistical analyses

Data are reported as means ± SD. The Mann–Whitney test for comparing intergroup data and the Wilcoxon matched‐pairs test for intragroup data were used where appropriate. A conventional P value < 0.05 was considered to be statistically significant. The program Prism 4.0 (Graph Pad Prism 5, La Jolla, CA, USA) was used for statistical analyses except for two‐way repeated measures ANOVA and post hoc pairwise comparisons using a t test with pooled SD analyses for which the program R3.2.3 (R Foundation, Vienna, Austria) was used.

Results

Observation of clinical parameters

From the initial 17 animals included in this study, three were excluded. Two died (one from each group) at wake‐up following the anaesthesia period. The third was excluded for technical failure of the stimulators. For all other animals, the electrode implantation procedure and the SNS protocol occurred normally and all surgical procedures were performed without complications. During the SNS and TNBS‐induced‐lesion procedures, heart rate, oxygen saturation and core body temperature showed no difference between the two groups (Table 1). Similarly, 24 h after TNBS administration no difference was observed between groups or from initial baseline values for any of the parameters measured. Serum C‐reactive protein was below the detection limit irrespective of the experimental conditions.

Table 1.

Clinical parameters observed before (T0) and 24 h after TNBS enema in the sham and SNS groups

		Sham	SNS
Heart rate (bpm)	T0	97.8 ± 5.1	99.6 ± 5.3
	T24h	92.7 ± 5.5	92.4 ± 6.1
$P_{O_2}$ (%)	T0	97.4 ± 0.2	95.7 ± 0.6
	T24h	96.8 ± 0.3	97.5 ± 0.3
T (°C)	T0	35.4 ± 0.4	36.1 ± 0.9
	T24h	37.4 ± 0.4	37.2 ± 0.1

The results are means ± SEM. (bpm, beats per minute; P, partial pressure; T, temperature).

SNS enhances restoration of barrier function

As we have previously described a potential protective effect of SNS on the IEB, we first analysed the impact of SNS on mucosal lesion induced by TNBS. Paracellular permeability was significantly increased 1 h after TNBS enema in both groups (sham and SNS) as compared to control (T0; n = 0.001). This increase was similar between both groups (n = 0.359; two‐way repeated ANOVA, Fig. 2 A). Transcellular permeability was also significantly increased 1 h after TNBS enema in both groups as compared to control (T0; n = 0.001). This increase was similar between both groups (n = 0.765; two‐way repeated ANOVA, Fig. 2 B).

Figure 2. Sacral nerve stimulation restores intestinal barrier functions and mucosal architecture following TNBS enema in pigs .

A and B, TNBS‐induced increase of paracellular and transcellular permeability in sham (filled bars) and SNS (empty bars) groups 1h after enema. C and D, follow‐up of paracellular and transcellular permeability in sham (filled bars) and SNS (empty bars) groups 1h, 3h and 24h after TNBS enema. Data in A–D are means ± SEM (n = 8 in sham group and n = 6 in SNS group, *P < 0.05 using non‐parametric paired Student's t‐test compared to their respective T0 and ^§ P < 0.05 using non‐parametric Student's t‐test comparing both groups). E, representative probe‐based confocal laser endomicroscopy (pCLE) images taken at different time points prior to and following TNBS enemas. Note the similar aspect of pCLE images 1 h after TNBS enema in the sham and SNS groups (increased fluorescence intensity in the crypt; distorted crypts) but increased architectural recovery after 24 h in the SNS group compared to sham animals. F, representative HPS staining images taken at different time points prior to and following TNBS enemas (scale bar: 100 μm). G, representative white‐light rectoscopy images revealing, macroscopically, the lesions induced by TNBS. Note the less damaged aspect of the mucosa in SNS animals.

We next analysed whether SNS enhanced restoration of the barrier functions following TNBS‐induced lesions. Paracellular permeability was significantly decreasing over time in both groups (two‐way repeated ANOVA, n = 0.001, Fig. 2 C) and no significant effect of SNS upon this decrease was shown (two‐way repeated ANOVA, n = 0.081). Interestingly, at T3h paracellular permeability was significantly reduced following SNS as compared to sham (n = 0.011). Transcellular permeability remained increased over time in the sham group while SNS significantly reduced it over time (two‐way repeated ANOVA, n = 0.037, Fig. 2 D). As compared to sham, in the SNS group, transcellular permeability was lower but not significantly at T3h (n = 0.085) and at T24h (n = 0.050) after TNBS‐induced mucosal injury.

SNS enhances the restoration of mucosal morphology

We next aimed to determine whether these functional changes were associated with alterations in morphological parameters of the mucosa. We first used an in vivo approach with probe‐based confocal laser endomicroscopy (pCLE). For each time point, multiple sequences of pCLE‐acquired images with a mean surface area of 23.5 ± 1.7 mm² were analysed.

Prior to TNBS enema in both groups the images of the mucosa were composed of dark, circular crypts with low fluorescence intensity in the lamina propria (Fig. 2 E, T0). Following TNBS enema, the fluorescence intensity in the crypt increased dramatically in the sham group (one‐way ANOVA, n = 0.039). In contrast, in the SNS group the fluorescence intensity in the crypt remained low over time, and significantly lower than in the sham group (Table 2, Crypt brightness, n = 0.031). Concerning morphological parameters, at 24 h crypt size was significantly higher in the sham than in the SNS group. In contrast, other morphological parameters studied (Table 2) remained similar between both groups. HPS staining revealed a more preserved mucosal and, in particular, crypt organization in the SNS group as compared to sham (Fig. 2 F).

Table 2.

In vivo microscopy parameters analysed by probe‐based confocal laser endomicroscopy (pCLE) in the rectal mucosa 1, 3 and 24 h after TNBS enema in the sham and SNS groups

		Sham	SNS
Ma/ma	T1h	0.93 ± 0.10	1.01 ± 0.07
	T3h	1.05 ± 0.08	1.04 ± 0.11
	T24h	1.07 ± 0.06	1.06 ± 0.03
Sphericity	T1h	1.06 ± 0.02	1.28 ± 0.02
	T3h	1.03 ± 0.04	1.12 ± 0.05
	T24h	0.96 ± 0.05	1.05 ± 0.02
Perimeter	T1h	1.19 ± 0.18	0.87 ± 0.18
	T3h	1.00 ± 0.08	0.86 ± 0.09
	T24h	1.23* ± 0.11	0.86* ± 0.04
Fluorescence intensity in the crypt	T1h	2.51* ± 0.56	1.06* ± 0.17
	T3h	1.75 ± 0.40	1.13 ± 0.09
	T24h	1.26* ± 0.15	0.92* ± 0.06
Crypt density	T1h	0.39 ± 0.02	0.42 ± 0.08
	T3h	0.35 ± 0.06	0.42 ± 0.11
	T24h	1.10 ± 0.18	1.16 ± 0.14
Lamina propria area	T1h	0.94 ± 0.04	0.95 ± 0.05
	T3h	0.97 ± 0.05	1.01 ± 0.03
	T24h	0.87 ± 0.04	0.95 ± 0.05
Fluorescence intensity in the lamina propria	T1h	1.64 ± 0.28	1.01 ± 0.06
	T3h	1.21 ± 0.28	0.95 ± 0.09
	T24h	1.09 ± 0.08	0.97 ± 0.03

Ma/ma: major axis/minor axis ratio of crypt lumen; Sphericity: 4π × area/perimeter²; Crypt density: number of crypts per mm²; Lamina propria area: % of visible lamina propria on image. All results are expressed as fold‐change from T0 and as means ± SEM (*P < 0.05; Mann–Whitney).

Finally, at 24 h we used the UCEIS, which is a validated endoscopic score to characterize macroscopically the impact of SNS on TNBS‐induced mucosal lesions. At 24 h after TNBS enema the two groups presented patchy lesions with alternation of yellow zones and normal or edematous mucosae. Superficial ulcerations and petechiae were also seen in both groups (Fig. 2 G and Fig. S1). The mucosa appeared to be less damaged in the SNS group than in the sham group. However, the UCEIS score was not significantly less severe in the SNS compared to the sham group (4.7 ± 0.7 vs. 2.8 ± 0.7; sham and SNS, respectively, n = 0.087).

SNS enhances protein expression of tight junctions

We next determined, 24 h after TNBS enema, whether changes in mucosal barrier functions were associated with changes in the expression of tight junction proteins of the intestinal epithelial lining.

First, epithelial distribution of ZO‐1 was assessed by immunohistochemical methods in the sham and SNS groups (Fig. 3 A). In the sham group, ZO‐1 pericellular network organization was disrupted and appeared mainly punctiform. In contrast, following SNS, areas of mucosa with continuous pericellular distribution of ZO‐1 could be observed. Western blot analysis showed that ZO‐1 expression (normalized to T0) was significantly higher in the SNS group as compared to the sham group (Fig. 3 B, C; 0.76 ± 0.16 vs. 1.44 ± 0.29 ratio to T0; sham vs. SNS, respectively; n = 0.303). These changes in ZO‐1 protein expression were associated with an increase in ZO‐1 mRNA expression (0.68 ± 0.44 vs. 7.19 ± 3.34 ratio to T0; sham vs. SNS; n = 0.009). In contrast, no significant change in the mRNA expression of other tight junction proteins (claudin‐1 or occludin) was measured (Fig. 3 D).

Figure 3. Sacral nerve stimulation enhances expression of tight junctions and activates FAK in pig mucosa following TNBS enema .

A, representative photomicrographs of the rectal mucosa labelled with antibodies against ZO‐1 24 h after TNBS enemas in sham and SNS animals. Scale bar: 100 μm. High‐magnification image of each area marked by red square; scale bar: 10 μm. Note the presence of pericellular labelling of ZO‐1 in the SNS but not the sham animal. B, representative immunoblot obtained from mucosal sample lysates using antibodies against ZO‐1. β‐Actin was used as a loading control. C, optical densities of ZO‐1 immunoreactive bands, normalized to the optical densities of β‐actin immunoreactive bands in the same samples, and expressed as percentages of their value at T0, prior to TNBS enemas. Data correspond to mean ± SEM of six samples for sham animals (filled bars) and five samples for SNS animals (empty bars). *P < 0.05 (Mann–Whitney). D, transcriptomic analysis of mRNA expression of ZO‐1, occludin and claudin‐1 in mucosal samples 24 h after TNBS enema in sham animals (filled bars; n = 6) and SNS animals (empty bars; n = 5). Data are means ± SEM (*P < 0.05; Mann–Whitney).

SNS modulates the early mucosal inflammatory response and activates focal adhesion kinase

We next aimed to determine whether changes in IEB functions induced by SNS were associated with changes in the mucosal inflammatory response.

We determined the impact of SNS on the expression of key mediators involved in early gut inflammatory processes, measured in serum obtained at the site of mucosal biopsy. Among the cytokines measured, IL‐1β, IL‐4, IL‐10, interferon (IFN)α and IFNγ, but not tumour necrosis factor (TNF)‐α or IL‐8, significantly increased as early as 1 h after TNBS enema in the sham group as compared to T0. In contrast, no changes in these cytokines occurred at any time in the SNS group (Table 3). In addition, IL‐1β and IL‐4 levels were significantly lower in SNS compared to sham animals (Fig. 4 A, B).

Table 3.

Quantification of cytokines (pg ml⁻¹) in serum obtained at the site of mucosal biopsy before (T0), and 1 and 3 h after TNBS enema in the sham and SNS groups

		Sham	SNS
IL‐1β	T0	650.6 ± 290.5	1208.0 ± 414.2
	T1h	4369.0* ± 1073.0	1197.0 ± 574.4
	T3h	13693.0 ± 6803.0	1270.0 ± 367.4
IL‐4	T0	11.1 ± 2.4	19.8 ± 2.2
	T1h	35.9* ± 2.4	24.1 ± 2.1
	T3h	35.6* ± 5.1	22.7 ± 1.4
IL‐8	T0	448.1 ± 146.0	1427.0 ± 768.2
	T1h	1279.0 ± 301.5	740.9 ± 218.4
	T3h	767.7 ± 356.3	890.8 ± 361.8
IL‐10	T0	23.5 ± 14.8	90.2 ± 15.3
	T1h	32.9* ± 23.7	179.0 ± 93.2
	T3h	19.8# ± 11.9	105.2 ± 15.0
IFNα	T0	2.4 ± 1.7	9.0 ± 1.7
	T1h	3.3* ± 2.4	14.7 ± 5.7
	T3h	1.8* ± 1.1	9.3 ± 1.0
IFNϒ	T0	9.5 ± 1.7	22.7 ± 6.9
	T1h	32.1* ± 2.5	52.3 ± 31.2
	T3h	26.5* ± 2.7	25.5 ± 5.3
TNFα	T0	235.4 ± 13.8	297.9 ± 47.4
	T1h	379.1 ± 147.8	296.2 ± 49.7
	T3h	352.2 ± 107.0	313.3 ± 68.8

The results are means ± SEM (*P < 0.05, Bonferroni's post test after one‐way ANOVA in comparison to T0; ^# P < 0.05, Bonferroni's post test after one‐way ANOVA in comparison to T1h).

Figure 4. Sacral nerve stimulation regulates mucosal production of cytokines in pigs following TNBS enema .

A, time course of mucosal IL‐1β levels (assessed in blood serum taken at the biopsy site) in sham and SNS animals following TNBS enemas. Data correspond to mean ± SEM (*P < 0.05 using Student's t‐test comparing both groups at each time) for sham animals (n = 5; filled bars) and SNS animals (n = 6; empty bars). B, time course of mucosal IL‐4 levels (assessed in blood serum taken at the biopsy site) in sham and SNS animals following TNBS enemas. Data correspond to mean ± SEM (*P < 0.05; Student's t‐test) for sham animals (n = 5; filled bars) and SNS animals (n = 6; empty bars). Note, in the sham group, the concentrations of IL‐1β and IL‐4 were significantly higher than in the SNS group (P < 0.05; two‐way ANOVA). C, time course of mucosal TGF‐β1 levels (assessed in blood serum taken at the biopsy site) in sham and SNS animals following TNBS enemas. Data correspond to mean ± SEM (*P < 0.05; Mann–Whitney) for sham animals (n = 6; filled bars) and SNS animals (n = 6; empty bars). Note the increased level of TGF‐β1 in SNS compared to sham animals (two‐way ANOVA; P < 0.05). D, representative immunoblot obtained from mucosal sample lysates using antibodies against pFAK and FAK. E, optical densities of pFAK and FAK immunoreactive bands were measured, the pFAK/FAK ratios were calculated and then expressed as the ratio of this value at T0, prior to TNBS enema. Data correspond to mean ± SEM of six samples for sham animals (filled bars) and five samples for SNS animals (empty bars). *P < 0.05 (Mann–Whitney).

Interestingly, we also observed a significant increase in the level of TGF‐β1 in the mucosa of SNS compared to sham animals (two‐way ANOVA; n = 0.023). Furthermore, 1 h after TNBS enema the TGF‐β1 concentration was significantly higher in the SNS group compared to sham (Fig. 4 C). As TGF‐β1 is a known activator of the focal adhesion kinase (FAK), we finally determined whether SNS could modulate phosphorylation of FAK (pY397‐FAK), which reflects changes in activity known to be involved in the regulation of IEB maintenance and repair. We observed a significant increase in the pY397‐FAK/FAK ratio in the SNS compared to sham group (0.84±0.13 vs. 3.03±1.46; sham vs. SNS; n = 0.016, Fig. 4 D, E).

SNS induces early mucosal changes in mucosal cellularity

We finally aimed to determine whether changes in IEB functions induced by SNS were associated with changes in mucosal cellularity.

In the sham group, mucosal neutrophil density did not significantly change during the first 3 h after TNBS enema compared to T0 (Fig. 5 A, B). In contrast, in the SNS group, neutrophil density increased significantly as early as 1 h after TNBS enema compared to T0 (Fig. 5 A, B). In addition, neutrophil density in the SNS group was significantly higher than in sham animals over time (two‐way ANOVA, n = 0.032; Fig. 5 B). In contrast, mononuclear cell density did not change over time after TNBS enema in either group (Fig. 5 C).

Figure 5. Sacral nerve stimulation enhances mucosal cellularity and protease activities in pigs following TNBS enema .

A, representative photomicrographs of HPS staining of rectal mucosa 24 h after TNBS enema in sham and SNS animals (10× magnification). A 40× magnification image of each area is marked by black square. Arrows indicate neutrophils. Note the increased density of neutrophils in SNS compared to sham animals. B, evolution of neutrophil density in sham (filled bars) and SNS (empty bars) groups before and after TNBS enema. C, evolution of mononuclear cell density in sham (filled bars) and SNS (empty bars) groups before and after TNBS enema. Data in B and C are means ± SEM (n = 6 in each group, *P < 0.05 using Student's t‐test compared with T0 for each group). D, time course of mucosal levels of trypsin activity (assessed in the supernatants of colonic biopsies) in sham and SNS animals following TNBS enema. Data correspond to mean ± SEM for sham animals (n = 5; filled bars) and SNS animals (n = 4; empty bars). Note the significant increase of trypsin activity in SNS (one‐way ANOVA; P < 0.05) but not sham animals. E, Pearson correlation analyses of neutrophil density and trypsin activity (P < 0.05, r² = 0.69). F, time course of mucosal levels of elastase (assessed in the supernatants of colonic biopsies) in sham and SNS animals following TNBS enema. Data correspond to mean ± SEM for sham animals (n = 5; filled bars) and SNS animals (n = 5; empty bars).

We next determined the proteolytic activity of the mucosal supernatant following organotypic culture. In the SNS group, we observed a significant increase in trypsin activity in the supernatants of mucosal biopsies taken as early as 1 h after TNBS enema (one‐way ANOVA; n = 0.04; Fig. 5 D). In contrast, no change in trypsin activity was observed in the supernatants of mucosal biopsies taken from sham animals (Fig. 5 D). In addition, trypsin activity levels were significantly higher in the SNS as compared to sham group. Interestingly, trypsin activity was positively correlated with the density of neutrophils (Fig. 5 E; Pearson correlation, n = 0.005, r ² = 0.5459). No change in elastase activity was measured in the supernatants of mucosal biopsies over time in the sham or SNS groups (Fig. 5 F).

Discussion

In the present study we developed a porcine model of TNBS‐induced acute lesions of the rectal mucosa to evaluate the impact of SNS on mucosal repair. We observed that SNS enhanced the recovery from TNBS‐induced increases in transcellular permeability. In vivo confocal endomicroscopy analyses consistently revealed lesser severity of mucosal lesions in the SNS compared to the sham group. Furthermore SNS reduced TNBS‐induced changes in ZO‐1 expression and its epithelial pericellular distribution, and also increased pFAK/FAK expression as compared to sham. Interestingly, SNS increased the mucosal density of neutrophils, which correlated with an increase in trypsin activity and TGF‐β1 levels compared to sham treatment. Finally, SNS prevented the TNBS‐induced increases in IL‐1β and IL‐4 over time that were observed with sham treatment. Together, these results suggest that SNS enhances barrier repair processes following mucosal injury.

A major finding of this study was the ability of SNS to increase significantly transcellular permeability recovery following acute mucosal injury. Concerning paracellular permeability, SNS did not significantly affect its overall recovery after TNBS injury, although paracellular permeability was significantly reduced at 3 h after TNBS induction of mucosal injury. This differential effect on IEB functions could be due to the fact that following injury the restoration of paracellular permeability is mediated by restitution processes (not needing differentiated IEC) while transcellular permeability is controlled by the restoration of IEC functions associated with more mature/differentiated cells. Supporting this hypothesis is the observation that during ischaemia/reperfusion models in human small intestine, massive IEC shedding followed by epithelial lining repair occurs within 4 h after reperfusion (Matthijsen et al. 2009). In addition, in a rat colitis model induced by TNBS, a complete ablation of the epithelium lining was observed within 3 h after enema, but a full re‐epithelialization occurred within 1 day of enema (Salim & Söderholm, 2011). Analysis at earlier time points after injury could have identified stronger SNS‐induced changes in paracellular recovery, although as mentioned earlier, paracellular permeability was significantly reduced in SNS vs. sham group 3 h after TNBS.

Concerning transcellular permeability, we clearly observed that SNS significantly reduced transcellular permeability over time, suggesting an impact of SNS upon differentiation. Indeed, although the epithelial surface lining recovered quickly following acute TNBS‐induced mucosal damage of the small intestine in guinea pig, it was fully differentiated only after 7 days (Pontell et al. 2009). Thus, the ability of SNS to induce differentiation of the epithelium needs to be further explored. Therefore, restoration of paracellular and transcellular permeability induced by SNS could also reflect the ability of SNS to enhance IEC differentiation and maturation following mucosal re‐epithelialization. Restoration of paracellular and transcellular permeability could participate in the recovery of barrier function by limiting electrolyte losses and luminal antigen translocation via IEC to the immune system (Taupin & Podolsky, 2003). Interestingly, GI diseases such as IBD or irritable bowel syndrome are associated with an increased transcytosis of luminal antigens (Schürmann et al. 1999), increased transcellular permeability is correlated with the active phase of IBD (Söderholm et al. 2004) and/or increased paracellular permeability is involved in pain or relapses (Wyatt et al. 1993; Porras et al. 2006; Piche et al. 2009; Salim & Söderholm, 2011; Wilcz‐Villega et al. 2014). Our study thus suggests that SNS could limit the development of ongoing inflammation in this kind of disease.

Twenty‐four hours after TNBS, a reduction in ZO‐1 protein expression and disruption of its pericellular network organization were observed in sham compared with SNS groups. These results could be at the origin of the restorative properties of SNS upon IEB functions. Indeed, in vitro, reversible disruption of IEB by Ca²⁺ deprivation in the culture medium showed that ZO‐1 distribution was involved in the restoration of barrier function and was also accompanied by restoration of enterocyte polarization (Walsh et al. 2001). This increased expression of ZO‐1 24 h after TNBS enema in the SNS group could result from distinct mechanisms: (1) by direct neuromediator‐dependent regulation of ZO‐1 expression, as for instance electrical stimulation of the ENS which has been shown to increase ZO‐1 expression via vasointestinal peptide‐dependent pathways (Neunlist et al. 2003); and (2) as a consequence of SNS‐mediated reduction in cytokines such as IL‐4 which has been shown to reduce ZO‐1 expression in IEC (Colgan et al. 1994).

Among the candidate cellular pathways activated by SNS that are putatively involved in its ability to enhance mucosal repair following injury is FAK activity, as assessed by the increase in pFAK/FAK induced by SNS as compared to sham treatment. FAK activity regulates intestinal mucosal wound healing by increasing IEC survival, migration and proliferation (Owen et al. 2011). Furthermore, FAK was also shown to increase mucosal wound healing by regulating IEC cell spreading (Van Landeghem et al. 2011). Although FAK deletion in an intestinal epithelial‐conditional FAK knockout mouse model has no significant effect on intestinal development and function under basal conditions, epithelial repair is significantly impaired in knock‐out mice following dextran sodium sulphate colitis. Interestingly, FAK activity was also shown to increase ZO‐1 expression and pericellular localization (Ma et al. 2013), suggesting it could be involved in ZO‐1 changes induced by SNS. Besides IECs, which could have contributed to the SNS‐induced increase in tissue activity of FAK, other cellular sources also increase FAK activity during mucosal repair. In particular, fibroblast migration, which plays a central role in wound healing processes, is under the control of FAK activity (Leeb et al. 2003). Thus, one could speculate that SNS, by activating FAK‐dependent pathways both in IECs and in fibroblasts, could enhance wound healing. Interestingly, IBDs are characterized by a reduced migratory potential of fibroblasts correlated with reduced FAK phosphorylation (Leeb et al. 2003). Enhancing fibroblast migration and IEC repair by anti‐TNF therapies has been suggested to facilitate wound healing in Crohn's disease patients (Di Sabatino et al. 2007). Therefore SNS, by using similar pathways and mechanisms of action, could represent an alternative and/or complementary therapeutic approach to biotherapies in IBD.

Another important finding of this study was the identification of putative factors that could be responsible in part for SNS‐induced repair properties. Indeed, following TNBS enema, SNS was shown to increase levels of TGF‐β1 and trypsin as compared to controls. TGF‐β1 is an immunosuppressive cytokine, which in addition to its immune function is also involved in IEB repair processes. Indeed, TGF‐β1 contributes to the restitution of wounded IECs (Dignass & Podolsky, 1993). In particular, TGF‐β1 enhances the migration of intestinal epithelial cells across the wound margin by upregulating their expression of matrix metalloproteinase (MMP)‐1 and MMP‐10 (Salmela et al. 2004). In addition, in vitro stimulation of IEC with TGF‐β1 has been shown to promote their differentiation through Smad2‐ and Smad3‐dependent pathways (Yamada et al. 2013). Interestingly, from a mechanistic point of view, TGF‐β1, but not fibroblast growth factor, platelet‐derived growth factor or vascular endothelial growth factor, has been shown to increase FAK levels in Caco‐2 and IEC‐6 (Walsh et al. 2008). It is also tempting to speculate that increased levels of trypsin could also participate in the repair process induced by SNS. Indeed, although various studies have shown that trypsin can participate in IEB dysfunction via activation of PAR‐2‐dependent pathways (Cenac et al. 2002, 2004), a recent study showed that trypsin can also reduce paracellular permeability in IECs by activation of protein kinase C‐dependent pathways (Swystun et al. 2009). Furthermore, trypsin could further enhance wound healing by targeting fibroblasts, as, in the skin, trypsin potentiates wound healing in part via its effects in differentiating monocytes to fibroblasts (White et al. 2013).

Besides changes in IEB functions, our study also showed that SNS modified tissue remodelling following TNBS enema. In particular, we observed a significant increase in neutrophil recruitment in the mucosae of SNS animals as compared to sham animals. Neutrophils are central to mucosal healing processes (Fournier & Parkos, 2012). Although their first role is to participate in the initiation phase of inflammation by clearing bacteria or cellular components, in part through the production of reactive oxygen species (Soehnlein & Lindbom, 2010), neutrophils are increasingly recognized as central players in wound healing. Neutrophil recruitment occurs within minutes of injury and peaks at 24–48 h (Fournier & Parkos, 2012). However, whether neutrophils are directly involved in SNS effects, the mechanism of action of the increase in neutrophil recruitment needs to be further studied. Finally, our results show that SNS prevents the IL‐1β and IL‐4 increase observed in the mucosa of sham animals following TNBS enema. Whether these changes reflect a direct immunomodulatory role of SNS, or are the result of enhanced SNS‐induced mucosal healing, remains also to be determined.

In conclusion, our work has shown the ability of SNS to enhance early mucosal repair following mucosal injury. It also identified putative mechanisms of action and factors involved in these effects. This study therefore sets a scientific basis for considering SNS as a novel therapeutic option in the treatment of diseases or symptoms associated with defects in intestinal wound healing.

Additional information

Competing interests

Upon manuscript submission, the authors have no competing interests to disclose.

Author contributions

J.B. performed all experiments and surgery with the support of A.C.C.D.S. E.C performed endoscopies and analysed the results. J.J., P.A. and J.C have contributed to realize specimen dissection, Ussing assay and organotypic culture experiment. N.V. performed elastase and trypsin assays. J.B., G.M. and M.N. designed the research study. J.B. and M.N. analysed the data and wrote the paper. All authors approved the manuscript and this submission.

Funding

The work was supported by Région des Pays de la Loire and Fondation pour la Recherche Médicale.

Translational perspective

Intestinal epithelial barrier (IEB) dysfunctions, such as increased permeability or altered mucosal healing, is increasingly recognized as being central mechanisms of various intestinal disorders. Recent data studies in pig have demonstrated that sacral nerve stimulation (SNS), a validated therapeutic approach for treatment of faecal incontinence, enhances IEB permeability. Nevertheless, the ability of SNS to enhance IEB repair following mucosal lesions remained to be identified. In this current study, we demonstrated that SNS enhanced both morphological and functional IEB repair after acute TNBS‐induced mucosal injury. In addition, SNS also enhanced neutrophil recruitment but reduced intestinal inflammation 24 h after TNBS injury. This works identifies novel mechanisms of action and factors involved in SNS effects upon rectal mucosa. Combining large animal models with the use of SNS stimulators used in clinical settings, our work set the scientific rational for a rapid transfer of SNS as a treatment option of diseases or symptoms associated with defects in intestinal wound healing.

Supporting information

Figure S1. Data set of individual rectoscopy.