Butyrate inhibits visceral allodynia and colonic hyperpermeability in rat models of irritable bowel syndrome

Lipopolysaccharide (LPS) or repeated water avoidance stress (WAS) induces visceral allodynia and gut hyperpermeability via corticotropin-releasing factor (CRF) and proinflammatory cytokines, which is a rat irritable bowel syndrome (IBS) model. As butyrate is known to suppress the release of proinflammatory cytokine, we hypothesized that butyrate alleviates these colonic changes in IBS models. The visceral pain was assessed by electrophysiologically measuring the threshold of abdominal muscle contractions in response to colonic distention. Colonic permeability was determined by measuring the absorbance of Evans blue in colonic tissue. Colonic instillation of sodium butyrate (SB; 0.37−2.9 mg/kg) for 3 days inhibited LPS (1 mg/kg)-induced visceral allodynia and colonic hyperpermeability dose-dependently. Additionally, the visceral changes induced by repeated WAS (1 h for 3 days) or CRF (50 µg/kg) were also blocked by SB. These effects of SB in the LPS model were eliminated by compound C, an AMPK inhibitor, or GW9662, a PPAR-γ antagonist, NG-nitro-L-arginine methyl ester, a NO synthesis inhibitor, naloxone or sulpiride. SB attenuated visceral allodynia and colonic hyperpermeability in animal IBS models. These actions may be AMPK and PPAR-γ dependent and also mediated by the NO, opioid and central dopamine D2 pathways. Butyrate may be effective for the treatment of IBS.

However, the effects of butyrate on visceral functions possibly related to the pathophysiology of IBS have been controversial so far. Rectal enema of butyrate decreases pain in response to rectal balloon distention in healthy human volunteers 10 . Moreover, butyrate reduced colonic paracellular permeability and enhanced the barrier in vitro 11,12 . In contrast, several researchers showed that the rectal instillation of butyrate aggravates visceral pain in non-stressed rats [13][14][15][16] , and none of the studies has shown that it exerts beneficial effects on visceral function in an animal IBS model.
In this study, we attempted to determine the effects of the colonic instillation of butyrate on visceral sensation and colonic permeability in rat IBS models, i.e. LPS, repeated WAS or CRF, to explore the possibility of therapeutic application of butyrate in IBS.

Methods
Animals. Adult male Sprague-Dawley rats (Charles River Laboratory, Atsugi, Japan) weighing about 300 g were used. The animals were group-housed (three to four rats per cage) in a regulated environment with illumination (12 h light/dark cycle) and temperature (23 °C−25 °C). Food (Solid rat chow, Oriental Yeast, Tokyo, Japan) and water were given ad libitum. chemicals. Sodium butyrate (SB; Fujifilm Wako Pure Chemical Corporation, Osaka, Japan) was dissolved in phosphate-buffered saline (PBS; 0.14 M NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 and 1.8 mM KH 2 PO 4 ). LPS obtained from Escherichia coli with serotype 055:B5 (Sigma-Aldrich, St. Louis, MO, USA), a rat/human CRF (Peptide Institute, Inc., Asagi, Japan), N G -nitro-L-arginine methyl ester (L-NAME), naloxone hydrochloride and domperidone (Fujifilm Wako Pure Chemical) were dissolved in normal saline. Compound C (dorsomorphin; LC Laboratories, Inc., Woburn, MA, USA), GW9662 (Focus Biomolecules, Plymouth Meeting, PA, USA) and sulpiride (Fujifilm Wako Pure Chemical) were dissolved in dimethyl sulfoxide (Fujifilm Wako Pure Chemical). The doses and routes of administration of the chemicals were determined according to previous publications 5,6,17-20 . Measuring visceral sensation. The conscious rats underwent colonic balloon distention to induce abdominal muscle contractions (visceromotor response, VMR), which were measured by an electromyogram (EMG). This method was previously validated as a quantitative measure of visceral nociception 21 . We evaluated the VMR threshold, defined as the volume (ml) of the distended balloon in the current study, and the experiments were performed as described previously 6 . The method of measurement was described briefly in the following. electrodes implantation and colonic distention balloon placement. Under isoflurane anesthesia, a small skin incision was made for the insertion of EMG electrodes (Teflon-coated stainless steel, 0.05 mm diameter) into the left side external oblique muscle in non-fasted rats. The electrodes were fixed to the muscle and the incised skin by cyanoacrylate instant adhesive. Then, the electrode leads were externalized directly through this closed incision without a subcutaneous (s.c.) tunnel and threaded through a urethane tube. Analgesics or antibiotics were not administered. A distention balloon (6-Fr disposable silicon balloon-urethral catheter, JU-SB0601; Terumo Corporation, Tokyo, Japan) was placed intra-anally with the distal end positioned 2 cm proximal to the anus. colonic distention and abdominal muscle contraction measurement. After the electrodes were fixed and the balloon was inserted, the rats were placed in Bollmann cages. The electrode leads were then connected to an EMG amplifier, and the signals were recorded by a PowerLab system (AD Instruments, Colorado Springs, CO, USA). Colonic distention was performed using the ascending method of limits paradigm with phasic distention by inflating the balloon by water using a syringe. The distention was increased progressively in 0.1 ml increments every 5 s until significant abdominal muscle contractions, i.e. VMR, were detected. The VMR threshold was defined as the distended balloon volume (ml) inducing VMR (Fig. 1A). The threshold was assessed twice (2-min interval), and the mean was calculated for each individual animal. The percentage change threshold was calculated as the threshold value after treatment divided by the basal threshold value and multiplied by 100.
Measurement of colonic permeability. Colonic permeability measurement was performed as described previously 6 . Briefly, the rats were anaesthetised by the administration of the mixture of medetomidine hydrochloride (Orion Pharma Ltd., Dhaka, Bangladesh; 0.15 mg/kg), midazolam (Sandoz, Tokyo, Japan; 2 mg/kg) and butorphanol tartrate (Meiji Seika Pharma, Tokyo, Japan; 2.5 mg/kg) intraperitoneally (i.p.) and underwent laparotomy. The colon was ligated at the junction with the cecum, and an open-tipped catheter (3-Fr, Atom, Tokyo, Japan) was inserted into the proximal colon through the hole made by a puncture using needle. The colon was gently flushed with PBS using the catheter in order to wash out all stools, and later, another ligation was added on the colon at approximately 4 cm from the proximal one. Then, 1 ml of 1.5% Evans blue in PBS was instilled into the colon segment through the catheter. After 15 min, the rats were euthanized by terminal exsanguination under deep isoflurane anesthesia. The colons were excised, washed with PBS and 6 mM N-acetyl-cysteine, and placed in 2 ml N,N-dimethylformamide for 12 h. Permeability was calculated by measuring the Evans blue concentration in the supernatant using a spectrophotometer at 610 nm.
Butyrate enema. The non-anesthetized rats placed in Bollmann cages underwent intra-anal insertion of a catheter (JU-SB0601; Terumo) with the distal end positioned 7 cm proximal to the anus. As the catheter was customized, the solution can diffuse from the distal end of the catheter. The rats received 0.5 ml SB solution (2, 6 and 16 mmol/l) once daily, i.e. at doses of 0.37, 1.1 and 2.9 mg/kg/day, for 3 consecutive days through the catheter. The controls were treated with the vehicle (PBS). The basal VMR threshold was measured 30 min after surgery for implanting EMG electrodes and placing the balloon, i.e. 24 h after the last enema. Then, LPS (1 mg/kg) or the vehicle was administered. Later, surgery and balloon placement were performed again, and the threshold was measured 3 h after injection followed by the measurement of colonic permeability. (C) Protocol for determining the effects of SB on repeated WASinduced visceral changes. Thirty minutes after SB or the vehicle enema, the rats were subjected to either WAS or sham stress for 1 h. These treatments, i.e. enema and stress, were performed for 3 consecutive days. The basal threshold was measured before the initial treatments. The second threshold measurement was performed 24 h after the last stress session followed by the measurement of colonic permeability. (D) Effects of SB on CRF (50 µg/kg)-induced visceral changes were also determined. The measurements of the second threshold and colonic permeability were examined 4 h after injection of CRF or the vehicle. placed on a plastic platform (height, 8 cm; length, 6 cm; width, 6 cm) positioned in the middle of a plastic cage, which was filled with water up to 7 cm of the platform height. Control animals were individually placed in the same plastic cage, which was not filled with water (sham stress). experimental procedures. Six groups of five to eleven rats were used. After 24 h from the last colonic instillation of SB with different concentrations or PBS, the basal VMR threshold was assessed. Next, the electrodes and distention balloon were removed, and either the vehicle or LPS (1 mg/kg) was s.c. injected (Fig. 1B). The rats were returned to their home cages, and after 3 h, the second measurement of threshold was implemented followed by the measurement of colonic permeability 5 .
Next, four groups of five to six rats were used to evaluate the effects of SB on WAS model (Fig. 1C). First, the basal threshold was measured. Then SB or the vehicle enema followed by WAS or sham stress for 1 h was implemented for 3 consecutive days. The measurement of the second threshold followed by colonic permeability was performed 24 h after undergoing the last stress session 7 .
The effects of SB on CRF model were also tested using four groups of five to six rats. The second threshold was measured 4 h after injection of CRF (50 µg/kg, i.p.) or the vehicle followed by the measurement of colonic permeability (Fig. 1D) 6 .
Next, to evaluate the mechanisms of actions of SB, the effects of compound C (2 mg/kg s.c.), GW9662 (3 mg/ kg s.c.), L-NAME (10 mg/kg i.p.), naloxone (1 mg/kg s.c.), sulpiride (200 mg/kg s.c.) or domperidone (10 mg/kg s.c.) were tested. The groups in these experiments consisted of five to six rats. These drugs were given together with SB or the vehicle enema.
Statistical analysis. Statistical analyses were performed using SYSTAT 13 software (Systat Software, Chicago, IL, USA). Data were presented as means ± SEM. Multiple comparisons were performed by one-or two-way analysis of variance (ANOVA) followed by Tukey's honestly significant difference (HSD) test. Comparisons between the two groups were performed using Student's t-test.

ethical statement. For all studies, approval was obtained from the Research and Development and Animal
Care Committees at Asahikawa Medical University (#17149, approved on August 2, 2017). ethical approval. All applicable international, national, and/or institutional guidelines for the care and use of animals were followed. Compound C reversed the effects of SB on the LPS model. As SCFAs were reported to activate AMP-activated protein kinase (AMPK) signalling 23 , the role of AMPK on the effects of SB was determined. First, we evaluated the effects of compound C, an AMPK inhibitor, on the basal threshold, and the LPS-induced visceral allodynia and colonic hyperpermeability. Three s.c. injections of compound C altered neither the basal threshold (0.60 ± 0.020 ml for compound C, n = 10 vs. 0.60 ± 0.015 ml for vehicle, n = 10, p > 0.05) nor the visceral response L-NAME or naloxone reversed the effects of SB on the LPS model. Butyrate was shown to increase nitric oxide (NO) production from macrophages 25 , and NO is known to modulate pain response 26,27 . Therefore, we determined the role of NO. L-NAME, an NO synthesis inhibitor, did not alter either the basal threshold (0.60 ± 0.017 ml for L-NAME, n =

Discussion
The effects of butyrate on visceral sensation are controversial. Bourdu et al. 13 reported that SB enema induced visceral hypersensitivity in rats for the first time and advocated that this response by SB could be used for an experimental animal model of IBS. After that, this phenomenon has been well reconfirmed by several other researchers, and it is now recognized as one of the common rat IBS models [14][15][16] . In contrast, a few human studies showed that butyrate improved abdominal pain in patients with IBS 30,31 . Additionally, repeated WAS decreases butyrate-producing microbiota, and visceral hypersensitivity induced by stress is alleviated by the supplementation of butyrate-producing bacteria in rats 32 , which may be indirect evidence suggesting the ameliorative effect of butyrate in visceral pain.
In the study of Bourdu et al. 13 , the rats underwent 1 ml SB enema twice daily for 3 consecutive days, and the tested concentrations of SB solution ranged from 8 to 1000 mmol/l. Under this protocol, SB induced visceral allodynia in a dose-responsive manner. Meanwhile, butyrate concentrations in the cecal fluid and colonic contents of rats, pigs and monkeys were reported to be 3 to 7 mmol/l when diets contained little or no fermentable dietary fiber, and as high as 40 mmol/l when the diets provided ample fermentable fiber 33 . Given the evidence above, the concentrations of SB solution in that study seemed to be extremely high. Thus, we used the physiologic concentration of SB solution for enema, i.e. 2, 6 and 16 mmol/l. The doses of SB using these solutions were 0.37, 1.1 and www.nature.com/scientificreports www.nature.com/scientificreports/ 2.9 mg/kg/day, whereas the doses in that study ranged from 8.4 to 1048 mg/kg/day, which means that the highest tested dose in the current study equaled only about one third of the minimum tested dose by Bourdu et al.
The current study showed for the first time that SB abolished visceral allodynia in rat IBS models, which was completely different from the findings by the previous studies above. Our results strongly suggest that a high dose of SB may induce visceral hypersensitivity, and physiologic concentrations of SB may improve visceral pain. Intraluminal administration of physiologic doses of butyrate into the distal colon for 7 days decreases visceral pain and discomfort in response to colonic distention in healthy humans 10,34 , which may further support our results and the notion above.
At the same time, we also found that SB improved colonic barrier. Previous studies showed that butyrate in physiologic concentrations can enhance intestinal barrier function, but high-dose butyrate disrupts the barrier using Caco-2 cells in vitro 35 . These findings strongly suggest that an adequate dose of butyrate may improve visceral function.
We have recently demonstrated that LPS-, repeated WAS-or CRF-induced visceral allodynia and colonic hyperpermeability were mediated via peripheral CRF, TLR4 and the proinflammatory cytokine system [5][6][7] . The speculated mechanisms of the visceral changes in these IBS models are considered as follows 6 . Activating peripheral CRF receptors by stress stimulates TLR4 to alter tight junction (TJ) proteins 36 , thereby inducing colonic hyperpermeability. Impaired gut barrier induces bacterial translocation leading to increased LPS to trigger to release proinflammatory cytokines by activating TLR4, which induces visceral allodynia possibly through the activation of visceral afferent neurons 37 . At the same time, the cytokine also increases gut permeability via modifying TJ proteins 38 . Additionally, LPS not only stimulates TLR4 but also activates peripheral CRF receptors 5 . Thus, peripheral CRF and TLR4-cytokine signalling may develop a vicious cycle activating each other to induce these visceral changes. www.nature.com/scientificreports www.nature.com/scientificreports/ In this scenario, visceral allodynia is considered to result from colonic hyperpermeability. In the current and our previous studies showed that visceral allodynia occurred associated with colonic hyperpermeability with no exception 6,17,18,39 . Moreover, Creekmore et al. 40 , demonstrated a positive correlation between the magnitude of visceral pain and paracellular permeability in repeated WAS model, and knockdown of occludin, one of the TJ proteins, induced intestinal hyperpermeability with visceral hypersensitivity. These results further support the notion above. Thus, SB may inhibit peripheral CRF-TLR4-cytokine signaling to improve colonic barrier followed by inhibition of visceral allodynia.
Actually, butyrate was reported to inhibit the expression of proinflammatory cytokines triggered by interferon-γ in RAW 264.7 cells 41 . Moreover, it was also shown that high-fat diet impaired gut barrier with increased serum level of LPS and upregulated the TLR4 gene and proinflammatory cytokines in the liver, which were improved by the intragastric administration of SB in mice 42 .
SCFAs are known to activate AMPK in the liver and muscles 23 . In addition, we have very recently shown that metformin, an AMPK activator, blocked the visceral changes in the same IBS models 18 . Thus, we hypothesized that the effects of butyrate are mediated via AMPK, and it actually happened, i.e. compound C reversed the effects of SB. Several in vitro studies proved that LPS-induced inflammatory response was inhibited by AMPK 43,44 . In this context, the effects of SB may result from the suppression of cytokine signalling via AMPK.
At the same time, it was reported that butyramide, a butyrate-releasing derivative exerts an anti-inflammatory effect via the upregulation of PPAR-γ in dextran sulphate sodium-induced murine colitis 24 . PPAR-γ inhibits the expression of various cytokines in macrophages 45 , and we have also confirmed previously that activating PPAR-γ by pioglitazone abolished the visceral changes in these IBS models 17 . Moreover, butyrate reduces colonic paracellular permeability by PPAR-γ activation in HT-29 cells 11 . In the current study, the effects of SB were reversed by PPAR-γ antagonist, suggesting that butyrate may inhibit cytokine signalling via PPAR-γ to exert the effects.
It is known that there exists a link between AMPK and PPAR-γ. It was shown that LPS increased the expression of TLR4 via the suppression of PPAR-γ in endothelial EA.hy926 cells, and the activation of AMPK prevented the increase of TLR4 protein via the rescue of the decreased PPAR-γ protein 46 , suggesting that PPAR-γ might be a downstream effector of AMPK. In contrast, several studies showed that PPAR-γ mediates the activation of AMPK 47 . In this context, both signalling can modulate each other. Although both signalling modulated the effects of SB in the current study, we did not determine which signalling was upstream. Further studies are needed to explore this issue. www.nature.com/scientificreports www.nature.com/scientificreports/ We also showed that the effects of SB were reversed by L-NAME or sulpiride, but not by domperidone, suggesting that they were mediated via NO and central dopamine D 2 signalling. These findings may support the result that the effects of SB were mediated via AMPK signalling, because the antinociceptive effect by metformin in the LPS model was mediated via NO and central dopamine D 2 pathways, which were shown in our previous study 18 .
It has been demonstrated that NO pathway may be both pro-nociceptive and anti-nociceptive 48 . Based on the present study, we would raise a hypothesis that butyrate increases NO production from macrophages 25 , and NO inhibits proinflammatory cytokine genes in various immune cells 49 , thereby improving the IBS model.
Garrido-Gil et al. 50 showed that central dopaminergic depletion increased the level of IL-1β in the colon, suggesting that brain dopamine reduces the vulnerability of gut inflammation. Additionally, butyrate possibly crosses the blood brain barrier 51 , and SB at a dose of 300 mg/kg i.p. protects dopamine neurons to improve the motor deficit in Parkinson's disease model 52 , suggesting the possibility that SB act centrally to modulate dopamine signalling to exert the effects by suppressing cytokine.
However, we did not think that SB directly act on the brain in the current study for the following reasons. Butyrate is the main energy source for colonocytes 23 , and the majority of the luminal butyrate is consumed in the gut resulting in relatively low concentration of butyrate in portal vein. Then it is metabolized in liver and its concentration becomes lower in systemic circulation 53 . Moreover, the brain uptake of intravenous administration of butyrate was reported to be only less than 0.006% in baboons 54 . Therefore, the amount of brain uptake is considered to be negligible in the current experimental settings using physiologic concentration of SB enema. Butyrate is known to activate vagal afferents 55 , and activation of upper gut-innervating vagal afferents induces the release of dopamine from nigral neurons 56 . In this context, it is reasonable to think that SB may act peripherally and indirectly activate brain dopamine D 2 receptor.
As it is known that opioid signalling is involved in altered visceral sensory function by stress 28 , the role of opioid receptor was determined. Opioid receptors are expressed in immune cells and modulate cytokine response 57 . We found that naloxone blocked the effects of SB. There is no direct report indicating that butyrate activates opioid signalling, but Pol et al. 58 demonstrated that NO upregulated the µ-opioid receptor gene transcription in mice gut during intestinal inflammation. Moreover, NO stimulated the neuronal release of endogenous opioids to stimulate opioid receptors in the brain and the spinal cord 26,27 . These results suggest that butyrate may activate opioid receptors via NO to exert the effects. In this context, both peripheral and central opioid signalling may contribute to the effects of butyrate.  A and B), whereas domperidone did not modify the effects (C and D). *p < 0.05 vs. vehicle + vehicle + LPS, # p < 0.05 vs. vehicle + SB + LPS by two-way ANOVA followed by Tukey's HSD test. Each column represents mean ± SEM. The number of rats examined is shown in parentheses (n = 5).
This study has several limitations. We did not show the direct evidence that SB inhibits the production of cytokines. Although the visceral changes observed in these IBS models were mediated via IL-1 or IL-6 5-7 , the colonic mucosal levels of the cytokines were not elevated in the current experimental settings (data not shown). Therefore, we could not test the direct effect of SB on cytokine signalling. Since cytokines may act locally to visceral afferents or TJ inducing visceral changes, increased cytokines in the colonic mucosa were not prerequisite for these changes. In other words, activating local cytokine signaling and the elevated level of cytokines are different issues. Moreover, the sources of the cytokines responsible for the visceral changes were not determined. It is known that various cells other than macrophages or monocytes, such as fibroblasts, endothelial cells, neuronal cells and smooth muscle cells, secrete the cytokines 59 . Additionally, the role of opioid receptor subtypes on the effects of butyrate was not determined either. Before determining the precise mechanisms of actions by SB in a molecular or cellular level, we should clear the issue above.
In spite of these limitations, our results clearly showed that butyrate improved visceral changes in IBS models. Recently, several studies have been demonstrated that perturbations of the intestinal microbiota play a role in the pathophysiology of IBS. Although it is not definitely known that an altered microbiota is a cause or a consequence, it may be involved with the changes in intestinal motility, visceral sensation, mucosal barrier and the expression of pattern recognition receptors 60 . The microbiota generates and releases molecules that can signal to distant organs, which may induce these changes. Together with the evidence above, our results suggest that butyrate is one of the signalling molecules between the microbiota and host in the pathophysiology of IBS.
In conclusion, SB enema blocked visceral allodynia and colonic hyperpermeability in rat IBS models, which may be AMPK and PPAR-γ dependent, and mediated by the NO, central dopamine D 2 and opioid pathways. Butyrate may be useful for the treatment of IBS.

Data availability
All data generated or analysed during this study this study are included in this published article.