Oil of mustard (OM) is a potent neuronal activator that promotes allodynia and hyperalgesia within minutes of application. In this study, OM was used to induce an acute colitis. We also investigated whether intracolonic OM-induced inflammation alters gastrointestinal (GI) function over a longer time frame as a model of postinflammatory irritable bowel syndrome (PI-IBS). Mice given a single administration of 0.5% OM developed a severe colitis that peaked at day 3, was reduced at day 7, and was absent by day 14. At the peak response, there was body weight loss, colon shrinkage, thickening and weight increases, distension of the proximal colon, and diarrhea. Macroscopic inspection of the distal colon revealed a discontinuous pattern of inflammatory damage and occasional transmural ulceration. Histological examination showed loss of epithelium, an inflammatory infiltrate, destruction of mucosal architecture, edema, and loss of circular smooth muscle architecture. OM administration increased transit of a carmine dye bolus from 58% of the total length of the upper GI tract in untreated age-matched controls to as high as 74% when tested at day 28 post-OM. Mice in the latter group demonstrated a significantly more sensitive response to inhibition of upper GI transit by the μ-opioid receptor agonist loperamide compared with normal mice. OM induces a rapid, acute, and transient colitis and, in the longer term, functional changes in motility that are observed when there is no gross inflammation and thereby is a model of functional bowel disorders that mimic aspects of PI-IBS in humans.
- small intestine
- large intestine
- postinflammatory irritable bowel syndrome
- irritable bowel syndrome
- inflammatory bowel disease
irritable bowel syndrome (IBS) as well as inflammatory bowel disease (IBD) have strong neuronal components in their etiology and maintenance (37). Compelling evidence now implicates an association among intestinal inflammation, altered neuronal signaling, and development of symptoms of gastrointestinal (GI) distress and morbidity (12, 27, 31). Functional bowel disorders can result from long-term effects on gut effector systems (i.e., smooth muscle, epithelium) and on their regulation by intrinsic and extrinsic nerves following an inflammatory episode leading to symptoms typical of IBS that can include pain and diarrhea (28, 37, 38). In some patients, functional bowel disorders are the result of a prior history of infection with an intestinal pathogen or allergen that engendered the inflammation (37, 38). Postinfectious and postinflammatory disorders of gut function are now recognized as predisposing incidents in a significant percentage of patients who experience IBS-type symptoms yet have no sign of overt intestinal inflammation and histopathology (28). Whereas experimental models in rodents have led to a significantly enhanced understanding of the effects of acute inflammation on gut function, a paucity of animal models lend themselves effectively to the study of long-term effects that mimic development of the functional gut disorders linked to IBS-like symptoms (6, 37).
Oil of mustard (OM), allyl isothiocyanate, is one of a number of small molecular weight compounds that constitute the pungent components of mustard, horseradish, and wasabi. It is a direct stimulant of small nerve fibers (2), a potent acute inflammatory irritant (14), and has been used experimentally to evoke visceral hyperalgesia following intracolonic administration (15). As a neuronal stimulant, it has been used to successfully study allodynia and visceral hyperalgesia in mice (19, 29) and to elucidate the sensory neural pathways activated in response to its application (3, 20, 26). In the GI tract, intracolonic application of a 0.5% solution of OM rapidly induces visceral hyperalgesia that can be measured as heightened sensitivity to von Frey hairs applied to the abdomen (19) and during graded changes in intraluminal pressure via an indwelling balloon catheter (29). Altered sensitivity to these stimuli occurs within minutes after OM administration. Although a transient influx of neutrophils and eosinophils that peaked at 4 h has been observed (26), the effects of OM as an inflammogen in the GI tract have not been studied beyond 8 h. Our aim was to characterize and investigate the effects of OM as a model of acute colitis in mice. In addition, we hypothesized that this could induce altered gut motility at time points well beyond resolution of and recovery from the initial inflammatory insult and thus provide a potential model for postinflammatory IBS (PI-IBS).
Thus we report here that intracolonic application of a 0.5% solution of OM produced a severe, transient colitis characterized by body weight loss, colon shrinkage, edema and thickening, severe diarrhea, and transmural inflammatory lesions that occasionly penetrated through the seromuscular layers. Colitis was maximum at day 3 and gradually diminished in intensity with less severe symptoms by day 7 followed by complete resolution and recovery by day 14. When examined for long-term changes in GI function, OM-treated mice had significantly accelerated upper GI transit rates measured on days 21 and 28 post-OM application compared with age-matched untreated controls. Mice 28 days post-OM demonstrated increased sensitivity of upper GI transit to inhibition by the μ-opioid receptor agonist loperamide compared with control mice. Our results suggest that OM-induced colitis and OM-induced accelerated upper GI transit in mice can be used as a new experimental model of functional bowel disorders associated with PI-IBS.
MATERIALS AND METHODS
Male CD-1 mice (Charles River Laboratories, Kingston, NC), 9–10 wk old, were used throughout these studies. All treatments were carried out in accordance with the Federal Animal Welfare Act and with methods approved by the Institutional Animal Care and Use Committee of Johnson and Johnson Pharmaceutical Research and Development.
Freshly opened OM (95 or 98% pure allyl isothiocyanate; Sigma-Aldrich St. Louis, MO) was used in each experiment. Carmine dye (cochineal powder; CI 75470) was purchased from Sigma-Aldrich. For immunochemical studies, rabbit polyclonal antibody to myeloperoxidase was obtained from DAKO Cytomation (Carpinteria, CA). A rat monoclonal antibody to mouse F4/80 (Clone A3–1) was obtained from Abcam (Boston, MA), and rat monoclonal antibody to mouse CD45 (leukocyte common antigen; Clone 30-F11) was obtained from Pharmingen (San Diego, CA). Biotinylated goat anti-rabbit and biotinylated goat anti-rat secondary antibodies, and avidin-horseradish-peroxidase biotin were obtained from Vector Laboratories (Burlingame, CA). 3,3′-Diaminobenzidine was purchased from Biomeda (Foster City, CA).
Induction of colitis.
Mice were briefly anesthetized with ketamine/xylasine (Sigma) and held in a vertical position so that 50 μl of a solution of 0.5% OM in 30% ethanol could be administered intracolonically. The OM administration occurred to a depth of 4 cm via a syringe equipped with a ball-tipped 22-G needle.
Body weights were measured daily.
At the termination of experiments, colons were resected, examined for signs of inflammation, weighed after removing fecal contents (which were examined for signs of diarrhea), and the length from the aboral end of the cecum to the anus was determined. These data and observations were assigned a score as indicated in Table 1. The sum of the individual macroscopic indexes was combined into a macroscopic score for each colon, where 0 is normal and 15 is maximally affected.
Segments from the distal colon taken from the first to fourth centimeter from the anus of each animal were removed, rinsed in saline, fixed in 10% neutral buffered formalin, embedded longitudinally in paraffin, sectioned, and stained with hematoxylin and eosin (H&E). The tissues were examined for epithelial damage using light microscopy by an investigator who was blinded to the sample groups and were then scored for epithelial damage, cellular infiltration, and damage or alteration of smooth muscle architecture by a second investigator according to the scale detailed in Table 2. The sum of these individual histology scores comprised a microscopic score for each colon, where 0 is normal and 9 is maximally affected.
Immunohistochemical (IHC) methods were applied to characterize the inflammatory cells in the paraffin-embedded tissues according to previously described methods (9). Briefly, 5-μm tissue sections were mounted on microscope slides and then routinely dewaxed and rehydrated. After a 5-min exposure in a microwave in target buffer (DAKO) and subsequent treatment in 3% H2O2 for 5 min, the slides were routinely processed for IHC. All incubations were performed at room temperature for 30 min. After a 10-min blocking step with normal goat serum, the tissues were incubated with the primary antibodies. Rabbit anti-myeloperoxidase was used to identify neutrophils, rat anti-mouse F4/80 to detect macrophages, and rat anti-mouse CD45 to detect lymphocytes, predominately T cells, in the infiltrated cell populations. After brief washes, the slides were incubated first with a biotinylated goat anti-rabbit or biotinylated goat anti-rat secondary antibody and then with avidin-horseradish-peroxidase. After additional washes between incubations, the slides were developed with 3,3′-diaminobenzadine (DAB; Biomedia, CA) for 2 × 5 min. The slides were then routinely stained using hematoxylin, dehydrated, and placed on a coverslip.
Upper GI transit.
Mice received 250 μl of a 6% solution of carmine dye in 0.5% methyl cellulose (wt/vol) by oral gavage. After 20 min, the mice were rapidly euthanized. The large and small intestines were resected, starting with the distal colon first and working toward the pylorus until the entire small intestine, cecum, and colon were removed intact. The resected bowel was arranged lengthwise, and the length of the entire small intestine was recorded. The distance that the carmine dye front traveled was also measured and then converted into a percentage of length of the entire small intestine.
Loperamide (Sigma-Aldrich) was tested for effects on upper GI transit and was administered intragastrically by oral gavage 1 h before carmine dye administration.
Experimental groups were analyzed for significance of differences between the means of treatment groups and control groups by ANOVA with Bonferroni's test using GraphPad Prism (version 3.03, GraphPad software, San Diego, CA).
OM induces body weight loss and severe colitis.
After intracolonic instillation of 0.5% OM, body weights were measured daily. The mice rapidly lost weight during the first three days post-OM, with maximum mean body weight losses as high as 11% by day 3. The mice then gradually regained their weight until by day 7 their weights were indistinguishable from those of untreated mice (Fig. 1). Mice treated with ethanol vehicle did not show acute weight losses.
Because the most severe body weight loss occurred at day 3, we examined the tissues of mice at that time for evidence of colitis. Distal colons from OM-treated mice were shortened and thickened compared with colons from untreated mice and from mice given only 30% ethanol vehicle, and exhibited bloody lesions (Fig. 2). The distal colon and portions of the proximal colon were more distended than those of normal mice. Proximal colons from many other mice studied were highly distended and contained thick mucus or watery diarrhea.
Data shown in Table 3 reveal that at 3 days post-OM, the distal colons were shorter (by 2.4 ± 0.5 cm; P < 0.05) and had a greater mass (by 0.126 ± 0.070 g; P < 0.05) compared with their untreated controls. The change in length represented a 22% decrease, and the change in colon weight represented a 36% gain compared with untreated controls. The mice exhibited diarrhea scores of 2.2 ± 0.5 (P < 0.05 vs. untreated). Colon damage scores for OM-treated mice were 3.0 ± 0.6 (P < 0.05 vs. untreated), with a number of mice exhibiting bloody penetrating lesions in the distal colon at sites that did not correspond to the site where OM was applied. Values obtained for ethanol-treated mice were unchanged compared with untreated mice.
The proximal colon tissues exhibited a generalized mild erythema and thickening, but no overt macroscopic lesions were observed.
Histological evaluation of tissues from OM-treated mice.
Tissues collected at day 3 for H&E staining revealed discontinuous lesions with loss of epithelial architecture, penetration through the submucosa, and loss of smooth muscle architecture (Fig. 3). There was little or no histological damage observed in tissues treated only with the 30% ethanol vehicle.
In tissues from OM-treated mice, a cellular infiltrate was observed that was most intense in regions of epithelial and smooth muscle damage but was also noticeable in the submucosa and muscularis mucosa of areas that did not exhibit overt pathology in the mucosa or muscularis externa. Immunohistochemical staining of the cellular infiltrates in the lesioned sections from OM-treated mice (Fig. 4) show the epithelial, submucosal, and smooth muscle regions invaded by neutrophils (myeloperoxidase positive) and lymphocytes (leukocyte common antigen; CD45 positive). Macrophages (F4/80 positive) were concentrated in the smooth muscle below these lesioned areas. In other areas of the distal colon near a lesioned area, but where the epithelium was intact, the smooth muscle was hyperplastic. Lymphocytes could be observed in the epithelium but were especially dense in the submucosa and mucosa muscularis as well as in the smooth muscle layer. Neutrophils were essentially found in high density in the submucosa, whereas F4/80+ macrophages were evident in the epithelial layer and were highly prominent in the submucosa and smooth muscle layer. No recruited cells and no immunohistochemical staining for lymphocytes, macrophages, or neutrophils was observed in the smooth muscle or submcosa in tissues taken from mice treated with the 30% ethanol vehicle. Epithelium in these tissues exhibited some LCA+ lymphocytes and F4/80+ stained macrophages, but no MPO+ neutrophils. These data demonstrate that the inflammatory response is engendered by OM as a fully integrated one composed of apparently large numbers of lymphocytes, macrophages, and neutrophils.
H&E-stained tissues from proximal colons revealed edema and swelling of the muscularis externa, although no prominent cellular infiltrate was observed in this region of the bowel (data not shown).
Time course of OM-induced colitis.
The colitis induced by OM did not extend beyond 7 days. The macroscopic damage scores in animals that were examined for symptoms of colitis were greatest on day 3 (8.8 ± 1.4, n = 5, P < 0.05 vs. untreated), were reduced at day 7 (3.4 ± 0.7; n = 5, P < 0.05 vs. untreated), and restored to normal by day 14 ( Fig. 5). The mean histology score on day 3 was 4.8 ± 1.7 (P < 0.05 vs. untreated), declined by day 7 to 2.2 ± 0.6 (P < 0.05 vs. untreated), and declined further by day 14 to 1.0 ± 0.6 (P = NS vs. untreated). Ethanol essentially had no effect on macroscopic or microscopic scoring. These data therefore demonstrate that intracolonic OM produces an acute, transient, and severe tissue response in the distal colon. The response at day 3 appears to have many of the hallmarks of colitis but resolves within days of its induction.
OM induces a delayed postinflammatory accelerated upper GI transit.
Our original objective was to determine whether intracolonic OM could result in altered GI motility at a point in time that was distant from the original insult. We therefore examined upper GI transit at times when there was no longer any evidence of inflammation, starting with day 14, and extending to days 21 and 28. Upper GI transit was increased at days 21 and 28 post-OM administration compared with untreated controls (Fig. 6). In these studies, upper GI transit increased significantly (P < 0.05) from a mean of 58 ± 1% in untreated controls to a mean of 73 ± 2% at day 28 post-OM. Upper GI transit was slightly less at day 21 (70 ± 2% ) vs. day 28 but was still significantly higher compared with transit in untreated mice (P < 0.05). In mice that had been given only the ethanol vehicle 28 days before, we found that upper GI transit was no different than that seen in untreated mice. These results clearly demonstrate that intracolonic OM reproducibly induces accelerated GI transit in the small intestine, a site far removed from the fourth centimeter of the mid-distal colon where OM application occurred and at a time several weeks following the acute inflammatory response. This accelerated upper GI transit developed gradually and suggests that progressive tissue or receptor remodeling may have occurred as a consequence of OM-induced inflammation.
Inhibition of accelerated upper GI transit by loperamide.
GI inflammation in mice induced by oral adminstration of croton oil has previously been reported to increase GI transit rates and to increase sensitivity to inhibition by opiates (32, 33). We thought that the post-OM increase in GI transit might serve as a PI-IBS model, and to further explore this premise, we determined the activity of loperamide, a peripherally active μ-opioid receptor agonist, in this model. Results show that (Fig. 7) loperamide significantly inhibited transit at the lowest dose tested (0.1 mg/kg) in the day 28 post-OM group (P < 0.001 vs. OM control, P < 0.05 vs. loperamide-treated normal mice), but not in the untreated mice, and exhibited a shift toward greater inhibition at higher doses compared with normal mice. These results demonstrate altered gut motility in OM-treated mice that is associated with increased sensitivity to a μ-opioid receptor agonist.
The data presented in this report signify two separate actions that result from intracolonic OM application that may have important implications for colitis studies and for postinflammatory GI function studies. First, we demonstrated that intracolonic OM induces a severe acute colitis that includes body weight loss, colon shrinkage, and weight gain with diarrhea, mucus hypersecretion, and distension in the distal and proximal colon. Tissue damage in the distal colon was transmural, included penetrating ulcers, and was discontinuous. Histological studies revealed that the damage in the distal colon included loss of epithelium, destruction of crypts and of smooth muscle architecture, and extensive cellular infiltration into the lesions. In areas showing severe epithelial damage, lymphocytes and neutrophils could be observed in the epithelium, submucosa, and smooth muscle, whereas macrophages were concentrated in the smooth muscle below these lesions. All three cell populations predominantly infiltrated the submucosa in other areas where there was submucosal edema and muscle thickening, but where there was no evidence of epithelial damage. The epithelium and smooth muscle in these sections also contained lymphocytes and macrophages but no neutrophils. These data demonstrate that the inflammatory response induced by OM is composed of lymphocytes, macrophages, and neutrophils in high proportions, consistent with a fully integrated immune response and not confined to a single cellular compartment. For an immune response such as this to have occurred, a complex signaling cascade of cytokines and chemokines and upregulation of cell adhesion molecules is required. This further strengthens the concept that OM-induced colitis, even if acute and rapidly resolved, might be an adequate model for human clinical disease. Although the overt and microscopic tissue damage occurred principally in the distal colon, there was edema in the proximal colon, but no significant inflammatory lesions were found there.
OM is known to be proinflammatory, but that activity extends mainly to its ability to induce plasma extravasation, which is principally due to the release of substance P and to neurokinin 1 receptor (NK1R) activation (1, 14, 15, 24). OM is known to signal via NK1R by studies showing diminished plasma extravasation after NK1R antagonist treatment. However, NK1R activation is an early event because its antagonism was effective only in the 15 min immediately following OM application (14). Plasma extravasation still occurred in NK1R knockout mice and in denervated mice, suggesting a nonneuronal component to the mechanism activated by OM (20, 40). Plasma extravasation in this model was also not dependent on mast cells, because OM-induced changes did not occur in mast cell-deficient mice, and the mast cell-derived products serotonin and histamine do not appear to play a role in OM skin inflammation (14). OM-induced plasma extravasation was not inhibited by antagonism of CGRP (24) or by l-NAME inhibition of NO synthase (14).
There are earlier reports suggestive of direct tissue damage by OM (10, 44). In one study (26), within hours of intracolonic administration, OM was shown to induce a cellular infiltrate that was primarily composed of neutrophils and eosinophils. However, that infiltrate diminished between 4 and 8 h, and there were no further examinations for effects on tissues at later time points as were performed in the present study. In other investigations, topical OM application to mouse ears was reported to cause tissue damage but only by OM concentrations >5% (14), whereas the present studies were conducted with 0.5% OM. Therefore, it is not clear whether the damage observed in the present study was exclusively neuronally initiated or was via direct effects on tissue or was due to both.
The time course of the colitis we observed was unique compared with that seen in other models of colitis. OM-induced colitis peaked at 3 days and then resolved between days 7 and 14 postapplication. In comparison, trinitrobenzene sulfonic acid (TNBS)-induced colitis continues to develop past day 3 and increases in severity through days 7 to 14, after which there is resolution of damage. Lesions in TNBS colitis are continuous, although the severity is similar to that seen in OM colitis. Dextran sulfate sodium (DSS)-induced colitis does not begin to manifest symptoms until 4 or 5 days have elapsed and is dependent on daily DSS intake. Overt tissue damage in DSS colitis is discontinuous but not as severe as that seen in OM colitis.
The second important result originating from our studies with intracolonic OM is the demonstration that there were functional bowel changes that were spatially and temporally removed from the induction of colitis and when inflammation could no longer be detected. Thus upper GI transit was significantly accelerated compared with transit seen in untreated mice. Oral administration of croton oil or castor oil, which are proinflammatory substances, induce an immediate accelerated upper GI transit (8, 33), and two daily doses of croton oil will induce accelerated upper GI transit up to 4 days later (35). It is therefore noteworthy that the accelerated small bowel transit induced by a single intracolonic OM administration was observable at days 21 and 28. Because it was our intention to explore postinflammatory changes, we did not pursue functional bowel changes at earlier time points when there was active inflammation. This is an animal model in which postinflammatory functional bowel changes manifested as accelerated upper GI transit develop weeks after the inflammatory episode was initiated. In addition, accelerated upper GI transit was dose responsively reduced by the μ-opioid agonist loperamide, with greater sensitivity to the compound than seen when administered to normal mice. This is reminiscent of, and consistent with, results reported for opiate-mediated correction of upper GI transit dysfunction following oral croton oil administration and had been considered to be a reflection of increased opiate receptor expression (32, 35). This finding is also consistent with other experimental models of inflammation-induced gut dysfunction that have been characterized by persistent changes in neuronal and smooth muscle function several weeks after recovery from an initial inflammatory insult to the gut epithelium (6, 7, 41, 43), but these functional changes followed a more chronic intestinal infection. Long-term changes included impaired neurotransmitter release from enteric nerve endings (7, 41) and enhanced neuromuscular contractile activity (43).
OM is a neuronal stimulant (2, 10, 16, 17, 45, 46). As such, it has been used to great effect to study neuronal signaling in transduction of pain in peripheral tissues including viscera (4, 13, 19, 29, 40, 42). Intracolonic OM is reported to induce hyperalgesic responses to intracolonic balloon distension and also hyperalgesic responses to abdominally applied von Frey hairs (19, 29). The latter is an example of referred pain, and signifies neuronal signaling from visceral sites distant from the somatic structures that are stimulated but which share the same spinal nerve communication to the central nervous system. The visceral hyperalgesia in those reports occurred within 30 min of OM administration, as did the referred anterior abdominal wall hyperalgesia. The referred pain occurred at a site that was distant from the OM application site. Our observation of accelerated upper GI transit is similar, in that the OM application to the mid-distal colon resulted in alterations in gut function in the small intestine. There is no direct evidence yet to demonstrate that the effects by OM that we observed are due to neuronal activation, however.
The fact that we observed accelerated small-bowel transit consistently only after a 3- to 4-wk delay implies that gradual receptor remodeling, or tissue remodeling with receptor redistribution, might have occurred within the enteric nervous system and/or within visceral afferent pathways. There is precedent for neuroplastic alterations for NMDA receptor expression on nociceptive neurons as a consequence of OM stimulation, but these were short-term studies (4, 30). Various neuronal receptors, among them cannabinoid, NMDA, NK1, P2X, GABA, and μ-opioid, and soluble mediators that include nitric oxide and substance P, have been identified as being upregulated as a consequence of OM treatment (3, 4, 13–15, 23–26, 42, 46). Altered neuronal function in IBD includes morphological remodeling of intrinsic nerves expressing receptors for nerve growth factor essential to regrowth of nerves and repair of tissue damage but also overexpression of nerve growth factor receptor on fibroblasts, smooth muscle, and epithelial cells in response to inflammatory cytokines (34, 36, 39). Nerve growth factor regulates the expression of sensory neuropeptides (i.e., tachykinins and CGRP) (39) and expression of vanilloid VR1 receptor and purinergic P2X3 receptor, which have been implicated in postinflammatory hyperalgesia (34, 36). The signaling pursuant to OM application leading to colitis is not clearly understood yet, and the receptors or cell types that might be altered to change intestinal motility in this context have not yet been identified. In terms of motility changes, it may therefore be instructive to note studies showing that soon after intracolonic TNBS administration to mice, there were alterations in enteric Type II AH neuron activity and excitability in the ileum and increases in amounts of serotonin secreted by enterochromaffin cells in that region of the bowel (21, 22). Data demonstrating altered expression for serotonin receptors and serotonin reuptake transporter have recently been obtained in tissues from patients with IBD and IBS (5, 11). Those data represent a postinflammatory outcome and thus support the idea that intracolonic OM may similarly induce remodeling in the upper GI tract. OM has recently been demonstrated to signal through a newly described member of the transient receptor potential family, ANKTM1 (18), and other studies reveal OM-induced signaling involves cannabinoid, opioid, GABA, and purinergic receptors (3, 13, 26, 42). Whether alterations in local concentrations of enteric neurotransmitters, receptors, or other factors are responsible for the post-OM-accelerated GI transit remains to be determined.
Taken together, the data in the present study suggest that intracolonic administration of OM to mice can be used to study short-term development of colitis, possibly through a neurogenic origin, and also could be applied to studying development of dysfunctional bowel syndromes at a postinflammatory point in time when there is no gross inflammation. Whereas it is attractive to hypothesize that OM-induced colits and the subsequent IBS-like accelerated upper GI transit originate via neurogenic stimulation, there is as yet insufficent evidence to directly support this concept. Nonetheless, this model may have potential as an animal model for PI-IBS or other functional bowel disorders.
The authors thank Craig Schneider and Nate Wallace, Enterology Research Team [Johnson & Johnson Pharmaceutical Research and Development (JJPRD), Spring House] for expert technical assistance in these studies and Brenda Hertzog and Danielle Lawrence (Target Validation Team, JJPRD, Spring House) for efforts in histology sample preparation and evaluation.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society