Background: Recent advances in endoscopy have revealed that non-steroidal anti-inflammatory drugs (NSAIDs) often cause ulcers in the human small intestine. However, the mechanism of intestinal ulcer formation is still unclear.
Aims: The role of dietary fibre (DF), intestinal motility and leukotrienes (LTs) in the formation of small intestinal ulcers induced by indomethacin (IND) was investigated in cats.
Methods: Several types of diets containing DF at various percentages were given to animals twice daily during the experiment. IND was administered orally once daily after the morning meal for 3 days, and the area of mucosal lesions in the intestine was measured. Gastrointestinal motility was measured using a telemetry system in conscious cats implanted with force transducers.
Results: In cats fed regular dry food containing 2.8% DF, IND (3 mg/kg, p.o.) significantly increased the motility of the lower half of the small intestine and produced many severe lesions; the total lesion area was 7.7 (SEM 2.0) cm2 (n = 5). The lesions were markedly decreased with the low-DF diet (0.4%) and increased with the high-DF diet (7.2%). The lesion area was 0.1 (SEM 0.1) cm2 (p<0.05) and 18.2 (SEM 4.1) cm2 (p<0.05), respectively. Supplementation with insoluble DF (6% cellulose), but not soluble DF (pectin), in the low-DF diet increased the lesion area significantly. The hypermotility and lesion formation in the small intestine induced by IND were significantly (p<0.05) inhibited by AA-861 (a 5-lipoxygenase inhibitor), pranlukast (a LT receptor antagonist) or atropine.
Conclusions: Insoluble DF, intestinal hypermotility, leukotrienes and cholinergic pathways are implicated in the pathogenesis of small intestinal ulcers induced by NSAIDs.
Statistics from Altmetric.com
It has been reported that non-steroidal anti-inflammatory drugs (NSAIDs) cause ulcers not only in the stomach and duodenum but also in the small intestine in experimental animals and humans.1 2 3 4 Recent progress in endoscopic techniques, such as capsule endoscopy and double balloon (push) endoscopy, has revealed that small intestinal ulcerations induced by NSAIDs in humans are more frequent and more severe than previously thought.5 6 7 8 Ulcerative changes of the small intestine cause bleeding, anaemia, occult blood loss, malabsorption, diarrhoea, mucosal ulceration, occasional stricture due to diaphragmatic disease, and perforation. Furthermore, in contrast with upper gastrointestinal (GI) ulcers, there are few effective agents for prophylaxis and treatment of small intestinal ulcers. NSAIDs are often used in dogs and cats for the treatment of many diseases, and they sometimes produce adverse GI effects, as they do in humans.9 However, very few studies have examined these effects in animals, especially in cats.
Several different factors related to the pathogenesis of NSAID-induced GI ulcers have been proposed,10 11 12 13 such as mucosal defence activities (eg, mucus, mucosal blood flow, mucosal surface barrier, bicarbonate secretion), inflammatory cells (eg, neutrophils), cytotoxic nitric oxides, bile acids, and intestinal flora. Furthermore, the importance of inhibition of both cyclooxygenase (COX)-1 and COX-2 in NSAID-induced GI damage has been pointed out.14 15 Rainsford16 17 proposed that the inhibition of COX by NSAIDs stimulates the production of LTs, which may play a role in the formation of GI lesions. Recently Takeuchi and co-workers18 19 20 found that indomethacin (IND) increased intestinal motility in rats, suggesting the importance of hypermotility in intestinal ulcer pathogenesis. In the present study, we investigated the effects of feeding, dietary fibres (DFs) and intestinal motility on the formation of small intestinal ulcers induced by IND in cats. We also examined the effects of atropine, cimetidine, AA-86121 (a LOX inhibitor; LOX-I), pranlukast22 (cysteinyl LT receptor antagonist; cysLT-RA,), cholestyramine (bile acid-binding resin), neomycin, and erythromycin on the induction of GI lesions by IND.
Materials and methods
Male and female mongrel cats bred for the experiments in the animal house of Tottori University were used after feeding for >6 months.
The animals were given experimental diets containing DF at various percentages (table 1) twice daily between 9–10 am and 5–6 pm, and the amounts of each diet were based on the manufacturer’s instructions. The following six diet types were used. When an animal stopped consuming the diet during the three study days, the experiment was discontinued and the data were not used.
Dry food containing regular amounts of DF (RFD-Dry) (Canet; Petline, Gifu, Japan; DF = 2.8%), dry food containing low DF (LFD-Dry) (Prescription Diet, feline i/d dry food; Hill’s–Colgate Japan, Tokyo, Japan; DF = 1.2%) and dry food containing high DF (HFD-Dry) (Prescription Diet, feline w/d dry food; Hill’s–Colgate Japan; DF = 7.2%) were used.
Canned food containing no DF (NFD-Can) (Longrun Pilchard Gourmet; Nisshin Pet Food, Tokyo, Japan; DF <0.1%), canned food containing low DF (LFD-Can) Prescription Diet, feline i/d can food; Hill’s–Colgate Japan; DF = 0.4%) and canned food containing high DF (HFD-Can) (Prescription Diet, feline w/d canned food; Hill’s–Colgate Japan; DF = 2.7%) were used.
In experiments examining the effect of DF quality on the formation of ulcers, α-cellulose (insoluble DF; Sigma, St. Louis, Missouri, USA) or pectin (soluble DF from citrus; Wako, Osaka, Japan) was added to LFD-Can.
The following drugs were used: AA-861 (Wako), atropine sulfate (Wako), cholestyramine (Questran; BMS, Tokyo, Japan), cimetidine (Sigma), erythromycin (Wako), indomethacin (Sigma), neomycin sulfate (Wako), pentobarbital sodium (Nembutal; Dainippon, Osaka, Japan), pranlukast (Ono, Osaka, Japan), xylazine (Ceractal; Bayer, Tokyo, Japan). Atropine was dissolved in physiological saline for subcutaneous injection, while other drugs for oral administration were suspended in 1% carboxymethylcellulose; 0.5 ml/kg of each drug was administered. In studies on ulcer formation, IND was administered orally just after the morning meal (10 am) once daily for 3 days. Other drugs were given twice daily just prior to feeding (9 am and 5 pm). In the study on the effect of IND in fasted animals, the full day’s diet was given only once (between 5 and 6 pm), and IND was administered once daily at 10 am for 3 days.
Estimation of mucosal damage in the GI tract
Animals were killed by bleeding from the carotid artery under deep anaesthesia with a combination of xylazine (2 mg/kg, i.m.) and sodium pentobarbital (25–50 mg/kg, i.v.) 24 h after the final dosing of IND. Then, the whole GI tract (from oesophagus to distal colon) was removed and cut along the longitudinal axis. The GI tract was spread out on paper and mucosal lesions were observed macroscopically. The location (corpus, antrum, duodenum or small intestine), grade (erosion or ulcer), and area of lesions were measured. The location of the lesions in the small intestine was expressed as the distance (per cent of the whole small intestine) from the pyloric ring. The whole length of the small intestine from the duodenum to the ileum was divided into 10 parts of equal length. The first part was regarded as the duodenum, and segments of the small intestine was numbered from 2 to 10. The grade of lesions was estimated by at least two persons; “ulcer” was defined as a severe lesion (deep with mucosal detachment) and “erosion” was defined as a superficial or mild lesion. The lesion area (cm2) was obtained from the product of the length and width of the lesions, and the total lesion area (TLA) was obtained by summing the area of individual lesions in each part of the GI tract (corpus, antrum, duodenum and small intestine). Although the duodenum is a part of the small intestine, the TLA in the small intestine was calculated as the sum of the lesion areas in parts 2 to 10 of the small intestine.
Recording of GI motility
The animals were anaesthetised with xylazine (2 mg/kg, i.m.) and pentobarbital (25 mg/kg, i.v.), and the abdominal cavity was opened under aseptic conditions. Four strain-gauge force transducers (size 8×5 mm, F-08IS-P; Star Medical, Tokyo, Japan) were sutured onto the serosal surface of the gastric antrum, duodenum, jejunum and ileum (fig 1) to enable measurement of the contractile force of the circular muscle at these locations. The lead wires of the force transducers were exteriorised from the abdominal cavity and then brought through a subcutaneous tunnel via a skin incision between the scapulae. After surgery, a jacket-type protector was placed on each cat in order to protect the lead wires. The lead wires were connected to a transmitter (IMT-40TA, Star Medical) placed in the pocket of the protective jacket. Changes in GI muscle contractile activity were then recorded continuously on a personal computer (Dell Inspiron 2200), linked to a receiver (IMT-40RA; Star Medical) using a telemetry system (Model GTS-400; Star Medical).
After the animals recovered from surgery (at least 7 days) the study was started. GI contractions were measured continuously for 3 days during each study, and drugs were administered on the second day. The area of contraction for each site was measured hourly using a computer software program (Analize ΙΙ.dll; Star Medical). The total area of contraction was expressed as the percentage of the mean total area 3 h prior to the morning meal. This was defined as the motor index (%).
All data are expressed as mean with the SEM. Differences between groups were analysed by the Student t test for two-group comparisons, or analysis of variance (Dunnett’s multiple range test) if more than two variables were considered, with the significance level set at 5% (p<0.05).
Effect of feeding on the formation of GI lesions induced by IND
In the fasted state, IND (3 mg/kg, p.o.) caused mild lesions (erosions) in the antrum and duodenum, but almost no lesions in the small intestine. IND given after feeding of RFD-Dry caused lesions in the duodenum and lower half of the small intestine. As seen in fig 2, lesions were often observed encircling in the duodenum and longitudinally along the mesenteric margin in the small intestine. The location of individual lesions in the small intestine observed in each animal given IND in the fasted period (four cats) or after feeding (five cats) is shown in fig 3A. The mean TLAs of the gastric corpus, gastric antrum, duodenum and small intestine in the fasted or fed groups are shown in fig 3B. In the fasted group, the TLAs in the antrum, duodenum and small intestine were 0.2 (SEM 0.0) cm2, 1.1 (SEM 0.6) cm2 and 0.3 (SEM 0.1) cm2 (n = 4), respectively. In the fed group, the TLAs in the duodenum and small intestine were 0.8 (SEM 0.2) cm2 and 7.7 (SEM 2.0) cm2 (n = 5), respectively.
Effects of various diets on lesion formation in the duodenum and small intestine
Several different types of dry and canned foods (three each) containing DF at various percentages were examined for their effects on ulcer formation in the duodenum and small intestine. The results, according to the concentration of DF, are shown in fig 4. As mentioned above, in the group given RFD-Dry (DF 2.8%) IND produced many ulcerative lesions in both the duodenum and small intestine. Diets containing low amounts (<1.2%) of DF (NFD-Can, LFD-Can, LFD-Dry) decreased the TLA of the small intestine significantly (p<0.05), whereas a high-DF (7.2%) diet (HFD-Dry) significantly (p<0.05) increased the TLA. In the duodenum, NFD-Can, LFD-Dry and HFD-Can decreased the TLA significantly (p<0.05), but in contrast with results in the small intestine, HFD-Dry did not increase the TLA.
Effects of DFs on lesion formation in the duodenum and small intestine
To examine the effects of DFs on lesion formation in greater detail, modified diets consisting of LFD-Can supplemented with cellulose (3% or 6%) or pectin (3%) were used. In the group given LFD-Can, IND caused mild lesions in the duodenum but almost no lesions in the small intestine (fig 5A). Addition of 6% cellulose to LFD-Can increased lesions in the small intestine but decreased ulcers in the duodenum (fig 5A). The TLAs of the small intestine in the groups fed 3% and 6% cellulose were 4.5 (SEM 2.3) cm2 and 12.1 (SEM 3.6) cm2 (n = 4), respectively, and the TLA in the 6% cellulose group was significantly (p<0.05) larger than that in the LFD-Can group (fig 5B). The TLAs of the duodenum in the groups given 3% and 6% cellulose were 0.3 (SEM 0.2) cm2 and 0.3 (SEM 0.2) cm2 (n = 4), respectively, and the TLAs were smaller than that in the LFD-Can group.
Addition of pectin to LFD-Can did not increase IND-induced lesion formation in either the duodenum or small intestine but, rather, both were significantly (p<0.05) decreased (fig 5B). The TLAs of the duodenum and small intestine were 0.1 (SEM 0.1) cm2 and 0.0 (SEM 0.0) cm2 (n = 4), respectively.
Effects of drugs on lesion formation in the duodenum and small intestine
The effects of drugs on lesion formation were studied in animals given RFD-Dry (table 2). AA-861 (30 mg/kg, p.o.), pranlukast (100 mg/kg, p.o.) and atropine (0.3 mg/kg, s.c.) all inhibited lesion formation both in the duodenum and small intestine. The inhibitory effects on the small intestine were significant (p<0.05). Cimetidine (20 mg/kg, p.o.) inhibited lesion formation in the duodenum significantly (p<0.05) by 75%, and increased intestinal lesions by 43%. Intestinal lesion formation was also significantly (p<0.05) inhibited by pre-treatment with cholestyramine (300 mg/kg, p.o.) and neomycin (20 mg/kg, p.o.), and mildly by erythromycin (20 mg/kg, p.o.).
Effects of IND and other drugs on GI motility
Effect of IND on GI motility
In the fasting period, gastric antrum showed strong contractions continuously or intermittently (fig 6A,B). Both duodenum and jejunum showed relatively regular continuous contractions, but ileum showed a group of strong contractions which occurred one to three times per hour. When the animals started to eat RFD-Dry, the strong contractions seen in the antrum and ileum immediately disappeared and relatively regular contractions were observed (fig 6A). When IND (3 mg/kg, p.o.) was administered during the fasted period, it did not affect GI motility. However, IND given just after feeding of RFD-Dry caused an increase in motility, both in the amplitude and frequency of contraction, of the ileum (fig 6B,C). The effect began 1 h after dosing with IND and continued for more than 6 h. However, IND did not affect upper GI motility (fig 6C). When the animals were given LFD-Can instead of RFD-Dry, IND increased ileal motility, as was seen with RFD-Dry.
Effects of drugs on ileal hypermotility induced by IND
The effects of various drugs on ileal hypermotility induced by IND were examined. Typical responses to each drug are shown in fig 7A, and the results of quantitative analysis of each drug’s effects are shown in fig 7B.
In the control group given RFD-Dry, motility was slightly increased for 4 h after feeding. The mean motor index during the first 2–5 h was 109.7 (SEM 4.4) (n = 4). The administration of IND (3 mg/kg, p.o.) markedly increased motility; the motor index was 170.0 (SEM 15.8) (n = 4, p<0.05 vs control). Pretreatment with AA-861 (30 mg/kg, p.o.), pranlukast (100 mg/kg, p.o.) or atropine (0.3 mg/kg, s.c.) significantly (p<0.05) inhibited the motor stimulating effect of IND. The mean motor indices were 124.5 (SEM 5.2) (n = 4), 127.7 (SEM 15.2) (n = 4) and 95.9 (SEM 11.2) (n = 4), respectively.
In the present study, we examined the effects of feeding, DF (soluble and insoluble), and intestinal motility in the formation of small intestinal lesions induced by IND in cats. The administration of IND in fasted cats caused lesions in the gastric antrum and duodenum but almost no lesions in the small intestine. However, IND administered after feeding of regular dry food (RFD-Dry, 2.8% DF) caused many severe lesions in the lower half of the small intestine in addition to the duodenum. The formation of intestinal lesions by IND was decreased tremendously when the animals were given either dry or canned diets containing low amounts of DF, ie, ranging from 0.1% to 1.2%. In contrast, IND caused lesions in animals given HFD-Can (DF, 2.7%), and very severe lesions in animals fed diet containing high amount of DF (HFD-Dry, DF, 7.2%). In addition, IND caused lesions in the small intestine in animals fed LFD-Can (DF, 0.4%) supplemented with 3% or 6% cellulose (insoluble DF), but not with pectin (soluble DF). We previously observed that small intestinal lesions induced by IND in dogs and rats depended on the amount of DF consumed.23 24 In all species (cats, dogs and rats) tested to date, the formation of small intestinal lesions seems to depend on feeding, especially the amount of DF in the diets, indicating that DFs may play an important role in the intestinal lesions by NSAIDs in humans. In recent studies on NSAID-induced intestinal lesions in humans, the incidence of lesions varied from 8.4% to 70%.2 3 4 5 6 7 8 12 The difference in incidence may be due, at least in part, to differences in the diets (especially with respect to DF), in addition to differences in the target diseases, NSAIDs used, and duration of NSAID use.
Takeuchi and co-workers18 19 20 found that IND caused hypermotility of the small intestine in rats, suggesting the pathogenic importance of this effect in lesion formation, ie, hypermotility may break down the mucus layer and accelerate the invasion of intestinal bacteria and irritants into the mucosa. Anthony et al25 suggested that contractions of the intestine induced by IND in rats cause ischaemia–reperfusion injury in the mucosa. In the present study in conscious cats, IND given after feeding of RFD-Dry caused a marked increase in motility of the lower intestine, where the lesions were often observed. IND also increased motility in animals given a low DF diet (LFD-Can), but it did not produce lesions in animals given LFD-Can. Kunikata et al26 reported that IND decreased the mucus content of the small intestine in rats. Under conditions where the surface mucus is decreased, it is possible that insoluble DF may cause physical damage to the mucosal surface when the intestine is strongly contracted. These results suggest that both hypermotility and insoluble DF play an important role in the formation of lesions in the lower small intestine. Various irritants, such as intestinal bacteria, bile acids, NSAIDs and foods, may then aggravate the lesions.
Recent studies using video capsule endoscopy showed that intestinal lesions induced by NSAIDs in humans are often seen in the middle and lower parts of the small intestine.27 It is uncertain as to why IND given after feeding causes lesions in the lower intestine but not in the rest of the upper intestine. The data obtained in this study suggests several possibilities. The first possibility relates to insoluble DFs, ie, soluble components of the diet are absorbed in the upper and middle parts of the small intestine, and as a result, the amount of insoluble (undigested) solid components harmful to the mucosa is increased in the lower intestine. The second possibility is that IND has a selective motor-stimulating effect on the lower intestine; ie, IND given after feeding stimulates motility of the ileum but not the upper GI tract (fig 5C). The third possible mechanism may involve differences in the distribution of intestinal bacteria, ie, the distribution of bacteria in the upper intestine will be less than that in the lower intestine due to the presence of many bactericidal factors, such as gastric acid and bile acids. Furthermore, there is also the possibility that bacteria are refluxed from the caecum and colon into the lower intestine. Therefore, the lower intestine may be more vulnerable to bacterial invasion which can aggravate the lesions. This was partially supported by the present study, in which intestinal lesions were decreased by pre-treatment with neomycin and erythromycin. In addition, it is possible that high concentrations of bile acids28 and NSAIDs29 30 are achieved locally in the lower intestinal mucosa due to entero-hepatic circulation; this possibility is partially supported by the present data showing that cholestyramine, a resin that binds bile acids, significantly decreased the lesions. Intestinal bacteria and bile acids are not harmful to the mucosa under normal physiological conditions, but they may act to aggravate mild mucosal lesions.
Rainsford16 found that gastric lesions induced by NSAIDs in mice were prevented by various LOX-Is (eg, L651,392) and LT-RAs (eg, FPL 55712). From these results, he proposed that the inhibition of COX by NSAIDs caused diversion of arachidonic acid through the LOX pathway with resultant enhancement of LT production, and the LTs produced may play a role in the formation of gastric lesions. Rainsford17 also reported that lesions of the small intestine induced by IND in mice and rats were significantly prevented by LOX-Is (MK-886 and L-656,224). This was supported by recent findings by Nishio et al31 in rats that intestinal lesions induced by IND were significantly prevented by pre-treatment with cysLT-RA (pranlukast). They also showed that IND markedly decreased the amount of prostaglandin E2 (PGE2), and significantly increased that of cysLTs in the small intestine. From these data, they concluded that cysLTs play an important role in the pathogenesis of small intestinal lesions induced by IND.
Holme et al32 reported that synthesised LTC-1 caused contractions of guinea pig ileal preparations in vitro. Using the same ileal preparation, Yoshikawa et al33 reported that acetylcholine release caused by electrical field stimulation was inhibited by LOX-I (AA-861) and was accelerated by cysLTs, suggesting an activation of cholinergic pathways by LTs. Pawlik et al34 reported that both LTC4 and LTD4 induced strong contractions in the ileum in anesthetised dogs, whereas Garcia et al35 reported that a cysLT-RA (pranlukast) inhibited GI motility in dogs during both the digestive and interdigestive periods. All of these findings suggest that LTs have a stimulatory effect on GI motility, and that some of them do so by activation of cholinergic pathways.
Sanders et al36 found that IND induced spontaneous contractions in isolated dog ileal preparation and also increased the contractile response of the preparations to acetylcholine, suggesting that IND stimulated intestinal motility by removing the inhibitory effects of PGs such as PGI2 and PGE2. Takeuchi et al18 19 found that IND increased motility of the small intestine in rats, and suggested the importance of intestinal hypermotility in the pathogenesis of intestinal lesions induced by IND. In the present study in cats, IND increased motility in the ileum and produced lesions in the lower half of the small intestine, and both of these effects were significantly inhibited by LOX-I (AA-861), cysLT-RA (pranlukast), or an anti-cholinergic agent (atropine). The results are consistent with findings by Rainsford17 and Nishio et al31 in mice and rats, and suggest that LTs play a role in the pathogenesis of intestinal lesion formation by IND, probably by enhancing the effect of IND on intestinal motility through cholinergic pathways.
Whereas cimetidine significantly inhibited the formation of duodenal lesions induced by IND, it increased lesions in the lower small intestine, suggesting that anti-secretory drugs for the treatment of NSAID-induced upper GI ulcers may not be appropriate for the treatment of lower intestinal ulcers. It has been reported that proton pump inhibitors protect the small intestine against the damaging effect of IND in rats via acid-unrelated mechanisms, including anti-inflammatory and antioxidant effects.37 38 Further studies will be necessary to clarify the effects of anti-secretory drugs on intestinal lesion formation. The results of the present study also suggest the importance of the timing of NSAID use with respect to food intake and the risk of NSAID use in patients with inflammatory bowel syndrome who are treated with diets supplemented with insoluble DF.
We conclude that NSAIDs given after feeding cause lesions in the duodenum and lower half of the small intestine, and that indigestible solid components of food (such as insoluble DFs), hypermotility of the intestine, leukotrienes, and cholinergic pathways are implicated in the pathogenesis of small intestinal ulcers induced by NSAIDs.
The authors wish to thank Miss T Hara, and Messrs K Takata, S Matsuzaki, S Matsuura and D Murakawa of our department for their technical collaboration. We are greatly indebted to Dr Y Akiba, CURE/UCLA & BBRI, Los Angeles, California; Dr K Takeuchi, Kyoto Pharmaceutical University, Kyoto; Dr Y Ashida, Takeda Pharmaceutical Co. Ltd, Osaka; and Dr M Fujita, Ono Pharmaceutical Co. Ltd, Osaka, Japan, for their valuable discussions and suggestions.
Competing interests None.
Ethics approval Experimental procedures were approved by the Animal Research Committee, Faculty of Agriculture, Tottori University, Tottori, Japan. In accordance with our institution’s guidelines, the number of animals used in our studies was kept as low as possible; four or five cats were included per group, and the same data from the control group was used in all ulcer experiments.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.