GASTROENTEROLOGY
1890;81:815-625
Misoprostol Provides a Colonic Mucosal Protective Effect During Acetic Acid-Induced Colitis in Rats RICHARD N. FEDORAK, LONNIE R. EMPEY, COLIN MACARTHUR, and LAURENCE D. JEWELL Department of Medicine, Division of Gastroenterology, University of Alberta, Edmonton, Alberta, Canada
This study determined if intracolonically applied prostaglandin E, analogue (misoprostol) had a mucosal protective effect in rats with 4% acetic acidinduced colitis. The effects of misoprostol were compared with those of 5-aminosalicylic acid and betamethasone. A single application of 4% acetic acid induced an experimental colitis which was maximal at 2 days and showed spontaneous macroscopic and histologic healing by 12 days. Misoprostol (100 pg/kg), but not 5-aminosalicylic acid or betamethasone, administered 30 min before induction of colitis, provided macroscopic and histologic coionic mucosal protection but not protection of in vivo fluid absorption. The mucosal protective effect of misoprostol was time, dose, and diluent volume dependent. In the presence of misoprostol-induced colonic morphologic but not functional absorptive mucosal protection, in vitro unidirectional sodium and chloride flux measurements showed protection of theophylline-stimulated chloride secretion but not sodium absorption. Protection of in vivo colonic fluid absorption, in addition to morphologic protection, could be achieved when misoprostol was administered between 2 and 16 min before induction of colitis or when the highest dose (1000 pg/kg) of misoprostol was examined. We conclude that intracolonic misoprostol administration provides unique mucosal protective effects in experimental colitis.
T
he products of arachidonic acid metabolism, eicosanoids, are important mediators of the intestinal inflammatory response as well as direct stimulators of intestinal fluid and electrolyte secretions (l-3). Therefore, it is suggested that eicosanoids may mediate the gastrointestinal inflammatory response in inflammatory bowel disease and may also contribute to the pathogenesis of diarrhoea (1.3-5). Indeed, in-
and Department
of Pathology,
flamed colonic mucosa in patients with active ulcerative and Crohn’s colitis has been shown to have enhanced levels of both leukotrienes (LT), products of 5-lipoxygenase metabolism, and prostaglandins (PC), products of cycle-oxygenase metabolism (5-17). Studies have reported that luminal levels of PG are inversely related to fluid and electrolyte absorption and that PG and LT levels are also positively correlated with disease activity, as determined by clinical, endoscopic, and histological gradings (12). Agents effective in the treatment of inflammatory bowel disease, such as corticosteroids, 5-aminosalicylic acid (5-ASA), and sulphasalazine, alter arachidonic acid metabolism pathways to varying degrees (12,18-20). However, the individual roles of cycle-oxygenase and 5-lipoxygenase products in gastrointestinal inflammation have not been fully elucidated. It has been suggested that the elevation of endogenous prostaglandin levels seen in vivo during inflammatory disease states may be beneficial, assisting in protecting the mucosa from insult (21-23). Certainly, exogenous PG in the stomach have been shown to have mucosal protective effects, preventing the gastric necrosis produced by such agents as ethanol, hydrochloric acid, sodium hydroxide, hypertonic sodium chloride, taurocholate, and thermal injury (24-27). Protection against experimentally induced colitis has been observed with intracolonic administrations of PGE, and l6,16 dimethyl PGE, (28,29,30). PGI, has inhibited indomethicin-induced small intestinal lesions in a dosedependent manner (16). Furthermore, drugs that selectively inhibit endogenous prostanoid production, such
Abbreviations used in this paper: I-AM, 5-amino~~alicylic acid; BETA, hetamethasone; LT, leukotriene; MISO, misoprostol; PG, prostaglandin. 0 1888 by the American Gastrwnter&gical Association OOM-5085/80/$3.88
616
GASTROENTEROLOGY Vol. 98, No. 3
FEDORAK ET AL.
as indomethacin
and flurbiprofen, are of no therapeutic benefit to patients with inflammatory bowel disease (31). Such agents, associated with reduced levels of PG in the urine and rectum, may actually aggravate colonic mucosal inflammation (32-35). In this study, we investigated if intraluminally applied PGE, analogue (misoprostol] would provide a colonic epithelial mucosal protective effect to the gastrointestinal mucosa of rats undergoing 4% acetic acid-induced colitis. We also compared the effects of misoprostol administration with those obtained with the administration of 5-ASA and betamethasone (BETA]. Materials and Methods Materials Misoprostol (MISO; Cytotec) [(+)methyl (lla, 13 E)-11,16-dihydroxy-l6-methyl-9-oxoprost-l3-en-l-oate] is a synthetic PGE, methyl ester analogue. The misoprostol used in this study, lot CD206-36-12, reference 02420, was a gift from G. D. Searle and Company, Skokie, Ill. It was prepared from the stock solution supplied by Searle to a concentration of 1 mg/ml in absolute ethanol and stored at -70%. Its stability at this temperature is greater than 1 yr (36). A dilution of 100 pg/ml for experimentation was prepared into aliquots on a weekly basis in an ethanol buffer vehicle containing 20% ethanol, 104 mM Na,HPO,, and 16 mM NaH,PO,. The pH of this solution was 7.4, stored at -70°C. A 5-ASA rectal suspension (Salofalk] (67 mg/ml, pH 4.1, stored at 4’C] was purchased from Interfalk Canada Inc. Mont Saint-Hilaire, Quebec. A betamethasone disodium phosphate rectal suspension (Betnesol) (50 bg/ml, pH 7.5, stored at 4X) was purchased from Glaxo Laboratories, Toronto, Ontario, Canada. Acetic acid was purchased from Fisher Scientific, Nepean, Ontario, Canada, and prepared to a 4% solution in water (pH 2.4). All remaining chemicals were reagent grade and purchased from Sigma Chemical Co., St. Louis, MO. lntraluminal Administration of Misoprostol, 5-Aminosalicylic Acid, and Betamethasone Misoprostol (a dosage of 100 pg/kg except where otherwise specified), was prepared for intraluminal administration immediately before experimentation by mixing the appropriate volume of the 100 pg/mI ethanol-buffer stock solution with 154 mM sodium chloride to a total volume of 2 mL (pH 7.4). For a 250-g rat, the concentration of misoprostol administered was thus 12.5 pg/ml in a 2.5% ethanol-buffer vehicle. 5-ASA, 67 mg/mL, and BETA, 50 pg/ml, were prepared for intraluminal administration by directly removing 2 ml from the pharmaceutically supplied enema solutions. Ethanol-buffer vehicle, 2.5%, (pH 7.4) was used for control animals. Nonfasting male Sprague-Dawley rats (250-275 g, Biotron, University of Alberta, Edmonton, Alberta, Canada) were anesthetized with pentobarbitol(55 mg/kg) and atropine (0.5
mg/kg), administered i.p. Through a sterile midline incision, the colon was isolated, and the junction of the cecum and ascending colon was occluded via a reversible ligature, without compromise of neural or vascular integrity. The colon was cleansed of its luminal contents with a warmed 154-mM sodium chloride solution, and the residual fluid was manually expressed out through the rectum. Two milliliters of the chosen agent (MISO, 5-ASA, BETA, or ethanol-buffer vehicle] was then injected into the colonic lumen just distal to the occluding ligature via a 25-gauge needle passed obliquely through the outer muscle layer. The colon was only mildly distended and all mucosal regions were equally coated. A rectal plug [Extra, Wrigley Co., Toronto, Ontario, Canada] was inserted to prevent fluid loss and to ensure an even mucosal contact of the intraluminal agent. The abdominal incision was then closed and the anesthetized rat placed prone in an incubator prewarmed to 37’C for 30 min (unless otherwise specified) before instillation of acetic acid.
Induction
of Acetic-Acid
Colitis
Acetic-acid colitis was induced as previously described (37). Briefly, the abdominal incision was reopened and the colon isolated. Residual intracolonic fluid was expressed out the rectum and 2 ml of 4% acetic acid (unless otherwise specified) was injected into the lumen of the colon through a &gauge needle passed obliquely through the colonic wall just distal to the occluding ligature. Immediately thereafter, 10 ml of air was injected to clear most of the acetic acid from the colon. The occluding ligature was removed and the midline incision closed. The animals were allowed to recover from anesthesia and were housed in a light-cycled room with free access to standard rat chow pellets (5001, Purina Mills Inc., St. Louis, MO.) and water. Two days later (unless otherwise specified) the rats were killed with a pentobarbitol overdose (240 mg/kg, i.v.). and their colons were removed for experimentation.
Macroscopic
Studies
The colon was rapidly excised, opened along its mesenteric border, and gently rinsed of its luminal contents with an iced solution of 154 mM sodium chloride. The colon was then placed flat, mucosal surface upwards, on a glass plate chilled to 4°C. A transparent acetate was placed 5 mm above the mucosal surface and the area of ulceration and total surface area were traced by a single observer (L.R.E.J. The respective area in cm’ was then calculated using a Zeiss computerized videoscope (Videoplan, Carl Zeiss Co., Canada, Toronto, Ontario].
Histologic Studies Representative sections, taken from the colon at 3,6. 9, and 12 cm distal to the cecal-ascending junction, were fixed in 4% buffered formaldehyde. For light microscopy, &pm, formaldehyde-fixed, paraffin-embedded sections were stained with H&E. Histological grading was determined in a blinded fashion by two observers (L.D.J. and C.M.) and represented the
March 1999
COLONIC MUCOSAL PROTECTION IN COLITIS
numerical sum of two criteria. Mucosal ulceration was graded as follows: 0, no ulceration: 1, focal ulceration: 2, multifocal ulceration; and 3, diffuse ulceration. Depth of injury was graded as follows: 0, no injury; 1, mucosal involvement only; 2. mucosal and submucosal involvement; and 3, transmural involvement. Therefore, there was a minimum histologic grade of 0 and a maximum histologic grade of 6.
Unidirectional
In Viva CoJonic Fluid Absorption
Studies
Nonfasting rats were anesthetized with pentobarbito1 and atropine and kept warm with a thermostatic heat lamp. The intestinal tract was exposed through a midline abdominal incision. An occluding ligature was placed at the cecal-ascending colon junction, and the luminal contents were flushed out the rectum with a warm 154-mM sodium chloride solution instilled via a cannula inserted through an incision just distal to the proximal occluding ligature: residual saline was emptied by gentle manual expression. An intestinal loop of -15 cm in length, beginning 2 cm below the cecal-colonic junction and extending distally to the peritoneal reflection, was created with ligatures. In isolating the loop, care was taken not to compromise mesenteric, vascular, or neural continuity. A 27-gauge needle was inserted obliquely through the outer muscle layer along the antimesenteric border, and 2 ml of 154 mM sodium chloride, prewarmed to 37°C. was instilled into the empty loop. No fluid leakage was detected, and the loop was only mildly distended. The viscera were returned to the abdominal cavity and the incision was closed. The rats were allowed to recover from the anesthesia. Sixty minutes after abdominal closure, the animals were killed by pentobarbitol overdose, and the colonic loop was removed. The length of each loop was measured and each was weighed, both full and emptied, to determine the residual intraluminal volume. The loop was then dried in a 50°C oven for 24 h. Results were expressed as the difference between initial and residual luminal loop volume per gram dry weight or per centimeter of bowel.
In Vitro Electrical
Measurements
In vitro Ussing chamber studies were performed on 6 l-cm segments of colon from each rat, 3 immediately proximal and 3 immediately distal to the middle of the transverse colon. These tissues were removed immediately after the animal was killed, cut open along the mesenteric border, and incubated with a solution of the following ionic composition: Na. 143 mM; K, 5 mM; Mg, 1.1 mM; Ca, 1.25 mM; Cl, 124 mM; HCO,, 25 mM; HPO,, 1.65 mM; H,PO,, 0.3 mM; and fructose, 20 mM. The pH was 7.4 when gassed with 5% CO, in 95% 0, at 37°C. Colonic segments were placed serosal side up on a Plexiglas dissecting board and stripped of their serosa and underlying longitudinal muscle layer using flat-edged dissecting forceps. Mucosal strips were then mounted in Ussing chambers, and transmural electric potential difference, resistance, and intestinal short-circuit current were determined as previously described (38).
617
Flux Measurements
Transepithelial unidirectional fluxes of Na and Cl from mucosa to serosa and from serosa to mucosa were measured under short-circuited conditions. For flux measurement, 1 &i of %l and 1 pCi of =Na were added to either mucosal or serosal reservoirs after the tissues were mounted. Tissue pairs that differed in electrical resistance by greater than 25% were discarded. All flux experiments were conducted in bicarbonate Ringer’s solution as listed above. Effects on colonic ion transport of theophylline (10e3 M) were determined in sequential periods on the same tissue. Steady-state unidirectional basal fluxes were measured over 15 min, beginning 30 min after addition of tracers. Steadystate unidirectional theophylline fluxes were measured over 20 min. beginning 20 min after addition of theophylline. Fluxes were calculated as previously described (38).
Statistical
Analyses
Statistical analyses of data were performed by analysis of variance. This was done by multivariate profile analysis of repeated measures using a microcomputer statistical software package (EPISTATZ, Round Rock, Tex.]. When overall analysis showed significant interaction, Students’ t-test was used to examine the significance of differences.
Results Effect of Acetic
Acid on CoJonic Mucosa
Macroscopic changes. Intracolonic application of acetic acid produced an immediate blanching and mild muscular contraction of the colonic mucosa which resolved spontaneously over 20-30 min. At the time of death, the serosal surface of the colon was found to be hyperemic; however, in no instance was perforation identified. Figure 1 illustrates the macroscopic colonic mucosal lesions found at varying time intervals following 4% acetic acid-induced colitis. Mucosal lesions consisted of circumferential ulcerations covered with a thick white exudate and were located primarily in the region of the transverse and descending colon. The ascending colon and rectum were less affected. The reason for this unequal distribution of injury is unclear, although it may represent a pooling of residual acetic acid in a dependent segment of colon during anesthetic recovery. Macroscopic ulceration was maximal at 2 days and was completely resolved 12 days following the application of 4% acetic acid. Histologic changes. As shown in Figure 2, histological injury grade was maximal within a period of 1 day following the 4% acetic acid application and had returned to near control levels [grade = 0) by 12 days. Figure 3 illustrates the specific light microscopic features of 4% acetic acid-induced colitis. One day
618
FEDORAK
ET AL.
GASTROENTEROLOGY
Vol. 98, No. 3
Figure 1. Macroscopic colonic mucosal lesions at varying time intervals after intracolonic application of 4% acetic acid. The caecum (CM) was removed and colons opened along the mesenteric border. (A) 6 h; (B) 1 day; (C) z days; (D) 4 days; (E) 6 days; and (F) 12 days after acetic acid induction, respectively. Macroscopic ulcerations were maximal at 2 days, markedly improved by 4 days, and completely resolved by 12 days after induction of colitis.
after the application of 4% acetic acid, the acute inflammatory response extended into the submucosa, which was edematous and contained extravascular neutrophils. Small blood vessels were plugged with
6r
in T L
11 0
0
1 2
DAYS
4
6
FOLLOWING
8
ACETIC
10
12
ACID
Figure 2. Colonic histological grade at varying time intervals following intracolonic application of 4% acetic acid. Histological grade was determined by summation of mucosal ulceration and depth of injury score. Maximal injury = 6 Unite; minimal injury (control) = 0 Units. Line drawn by linear regression analysis. Light-microscopic injury following 4% acetic-acid colitis paralleled macroscopic changes although histologic recovery lagged behind macroscopic improvement. Points represent mean * S.E.M. of I)determinations for n = 4 rats.
polymorphonuclear cells. In some animals the colon was entirely necrotic, whereas in others the muscularis propria appeared viable. By the second day, regions of mucosal regeneration were evident, with restoration of mucosal integrity apparent in occasional areas. However, there were prominent areas of necrosis with evidence of hemorrhage and transmural inflammation. Healing was noted to be progressing at 4 and 6 days postapplication. Although return of mucosal integrity was widespread, focal areas of residual deep ulceration persisted. By day 12, the acute inflammation had almost completely resolved, and mucosal regeneration was complete. However, a mild architectural disorder of focal lamina propria fibrosis and bifid and irregular glands remained. Rats instilled intraluminally with a control solution of saline showed normal microscopic architecture at all time intervals. However, compared with gross macroscopic changes, mucosal histological injury following induction of acetic-acid colitis persisted at 4 and 6 days, despite a significantly improved macroscopic appearance. This is similar to findings in human ulcerative colitis, where histological recovery lags behind macroscopic improvement (39). Fluid absorption in vivo. As shown in Figure 4, intracolonic acetic acid produced a concentrationdependent decrease in net colonic fluid absorption.
COLONIC MLJCOSAL PROTECTION
March 1990
IN COLITIS
619
Figure 3. MIcrographs iHuetratlng the coume of untrasted 4% acetic add-induced colitis over a 12-dayperiod. A. Control colon before application of acetic acid.
the
B. Extensive mucosal necrosis with submucosal edema 1 day after application of acetic acid. C. Commencing mucosal regeneration is evident at 2 days. D. Restitution of an essentially normal mucosa is complete by 12 days (H&E; original magnification x40).
Maximal transport injury was present at a 4% concentration of acetic acid and remained constant thereafter: this was the case whether the results were expressed as pi/g dry weight (data not shown] or as J/cm of bowel [Figure 4). Therefore, subsequent experiments were performed using an intracolonic acetic-acid concentration of 470, and macroscopic, histologic, and functional observations made when injury is found to be maximal, i.e., 2 days after the application of acetic acid.
-100
1
j
1 ACETIC
1
3
5
ACID CONCENTRATION
t
f
7
9
-
(X)
Figure 4. Effect of varying lntracolonic acetic acid concentrations on in vlvo net colonic fluid absorption per centimeter. Acetic acid was appIied, and colonic fluid absorption was determined 2 days later. Maximal alteration of net fluid and electrolyte transport was seen at a 4% concentration of acetic acid. Values are mean + SEM for n 256 rats.
Effects of Misoprostol, 5-Aminosalicylic Acid, and Betamethasone on Acetic Acid-Induced Colitis After determining the effectiveness of acetic acid in inducing colitis, we next examined the mucosal protective effect of intraluminal MISO, 5-ASA, and BETA. Each agent was administered 30 min before application of 4% acetic acid. Rats were then killed 2 days after 4% acetic acid application, and the experiments were performed. Macroscopic changes. As shown in Figure 5, pretreatment with intraluminal misoprostol completely prevented the formation of macroscopic coIonic lesions, whereas intraluminal pretreatment with 5-ASA and BETA did not. Figure 6A shows the effect of each pretreatment agent on the surface area of colonic ulceration. Four percent acetic acid applicart 7.0% of tion alone resulted in ulceration of 43.1% the total colonic surface area. Pretreatment with misoprostol reduced acetic acid-induced ulceration to 1.9% * 1.7% (p < O.OOl),whereas pretreatment with 5-ASA or BETA did not (33.5% k 9.5% and 33.7% t 7.3570, respectively). Administration of the ethanolbuffer vehicle alone, which was used to maintain the MIS0 in solution, had no effect on the degree of ulceration. Histologic changes. As shown in Figure 6B, MISO-pretreated animals has a significant improvement in histologic grade compared with 5-ASA- and
620
FEDORAK
ET AL.
GASTROENTEROLOGY
Vol. 98, No. 3
Figure 5. Effect of MISO, 5-ASA, and BETA pretreatment on macroscopic colonic mucosal appearance induced by 4% acetic acid. Agents were administered intraluminally 20 min before 4% acetic acid induction, and colons were excised 2 days later. The caeca (CM) have been removed and the colons opened along the mesenteric border. (A) Vehicle-; (B] BETA- (50 &ml); (C) 5-ASA-(W mg/ml); and (D) MISO(100 rm/lU, -12.5 rg/ml) pretreated colons. Misoprostol pretreatment completely protected the colonic mucosa from 4% acetic acid-induced macro&opic injury.
BETA-pretreated rats. From micrographs of colonic mucosa, acetic acid-induced colitis in which the bowel was pretreated with vehicle only showed severe mucosal and submucosal inflammation with commencing epithelial regeneration. Colon pretreated with BETA or 5-ASA did not differ significantly in the extent of damage. By contrast, pretreatment with MIS0 afforded nearly complete histological protection from 4% acetic acid-induced mucosal injury. Fluid absorption in vivo. As shown in Figure 7, sham-operated, age-matched, control rats had a net colonic fluid absorption. Two days after 4% aceticacid application, rats showed net colonic fluid secretion. Pretreatment with MISO, 5-ASA, or BETA did not alter this net secretory state, whether results were expressed as pi/g dry weight (data not shown) or as hi/cm of bowel (Figure 7). These results of altered colonic fluid absorption 2 days after induction of colitis with MIS0 pretreatment contrast with the near-normal macroscopic and microscopic appearance seen with misoprostol pretreatment. To examine this further, we determined in vivo colonic fluid absorption at 12 days following 4% acetic acid application. Under all experimental conditions, i.e., whether rats were pretreated with MISO, 5-ASA, BETA, or ethanol-buffer vehicle, macroscopic and light-microscopic histologic changes had spontaneously returned to near normal by 12 days (Figures 1 and 2), whereas colonic fluid transport alterations persisted (Figure 7). However, rats pretreated with misoprostol or 5-ASA, showed a significantly greater degree of return to normal colonic absorption than vehicle- or BETA-pretreated rats [Figure 7).
Effect of Varying Doses of Misoprostol To further examine the mucosal protective effect of misoprostol, we next determined the effect of varying doses of intraluminal misoprostol on colonic ulceration and in vivo absorption. Misoprostol was administered 30 min before induction of colitis with 4% acetic acid: experiments were then conducted 2 days after acetic-acid application. Macroscopic changes. As shown in Figure 8A, intraluminal misoprostol preadministration showed a dose-dependent mucosal protective effect against coionic macroscopic ulceration. However, only the highest test dose used (1000 pg/kg) completely prevented ulceration. There was a single colonic ulcer in 1 of 6 rats in the lOO-pg/kg dose group, whereas at the 1 pg/kg dose all colons were ulcerated and the ulcer size fluctuated widely (range, X2%-39.5% total surface area). The macroscopic mucosal protective ED,, of misoprostol was approximately 1 pg/kg. Fluid absorption in vivo. In contrast to the macroscopic mucosal protection seen with low doses of misoprostol, net in vivo fluid absorption remained secretory, and only at the highest test dose (1000 pg/kg) was net colonic fluid secretion partially reversed to net fluid absorption (Figure 8B). As well, misoprostol pretreatment at 1000 pg/kg did not alter basal colonic fluid absorptive rates in control rats (data not shown). Onset and Duration of Action of Misoprostol Treatment We next determined the onset and duration of action by pretreatment with misoprostol(lO0 pg/kg) at
March 1990
COLONIC MUCOSAL PROTECTION
CONTROL
AA
MIS0
5ASA
BETA
IN COLITIS
621
was administered 0.5 or 1 min before 4% acetic acid, net colonic fluid transport remained secretory (Figure 9B) despite the normal macroscopic appearance described. On the other hand, misoprostol administration between 2 and 16 min before 4% acetic-acid application conferred a mucosal protective effect on in vivo colonic fluid transport, reversing net fluid secretion to net fluid absorption. Colonic net fluid transport once again became secretory, despite a normal macroscopic appearance, when misoprostol pretreatment occurred more than 32 min before application of the acetic acid.
Volume of Dilutent
CONTROL
AA
MIS0
5ASA
BETA
Figure 6. Effect of MISO, L-ASA, and BETA pretreatment on 4% acetic acid-induced colitis. (A) Colonic ulceration expressed as percent of total colonic surface area; (B) histologic grade, determined by summation of mucosel ulceration and depth of damage score: maximal injury = 6 U, minimal injury (control) = 0 U. Agents were administered intraluminally 30 min before 4% acetic acid, and rats were killed 2 days later. AA, pretreated with vehicle only; MISO, pretreated with MIS0 (100 &kg, -12.5 rg/mlk 5-ASA, pretreated with S-MA (67 n&ml); BETA, pretreated with BETA (50 fig/ml]. Results are compared with sham-operated, age-matched controls. Misoprostol pretreatment, but not 5-ASA or BETA, reduced 4% acetic acid-induced colonic ulceration and histologic injury to near control values. Values represent means * SEM of n = 6 rats. (*)p t0.001 compared with all other groups.
Misoprostol offered greater protection against colonic ulceration when diluted in 1 or 2 ml of dilutent than when diluted in 0.5 or 0 ml of dilutent [Figure 10). Indeed, misoprostol in 0.5 ml dilutent was only slightly more useful than no misoprostol at all. It was observed that the 0.5 ml volume remained localized at the site of instillation and was almost completely absorbed within the 30-min pretreatment time interval. The l- and 2-ml volumes were found to cover all regions of colonic mucosa uniformly.
0 ENTROL m MIS0
r
!Xl &I3
5ASA BETA
50
0
-50 1
varying time intervals before the application of 4% acetic acid. Experiments were then conducted 2 days after acetic-acid application. Macroscopic changes. As shown in Figure 9A, when misoprostol was administered between 0.5 and 32 min before 4% acetic-acid application, complete mucosal protection was observed, with the area of colonic ulceration being reduced to near zero. Colonic mucosal protection diminished when misoprostol was administered more than 32 min before acetic acid. These results are similar to those of the onset and duration of action of PG in ethanol-induced gastric lesions (24). Fluid absorption in vivo. When misoprostol
*
1
-100 0
‘A 12
2 DAYS
FOLLOWING
ACETIC
ACID
Figure 7. Effect of MISO, 5-ASA, and BBTA pretreatment on in viva net colonic fluid absorption, determined at varying time intervals following 4% acetic acid-induced colitis. Vehicle only (AA); MISO, 100 &kg, -12.5&ml; S-AM, 67mg/ml; or BETA, 50 fig/ml was preadministered intraluminrrliy 30 min before 4% acetic acid, and rats were killed either 2 days or 12 days later. Net colontc fluid absorption is secretory in all groupn at 2 days. By 12 days following acetic-acid colitis, MISO- and 5-ASA-pretreated groups showed signikant improvement in net fluid abnorption. Values are mean * SEM for n z 6 rata and compared with sham-operated, age-matched controls. (‘lp to.01 compared with all groups at 2 days.
622
FEDORAK
GASTROENTEROLOGY
ET AL.
lated net colonic chloride secretion prove net sodium absorption.
Vol. 98, No. 3
but did not im-
Discussion
0
1
10 MISOPROSTOL
100
The present study was undertaken to investigate the mucosal protective effects of misoprostol on acetic acid-induced colitis and to compare the effects of MIS0 with those of 5-ASA and BETA. In active human inflammatory bowel disease, levels products of 5-lipoxygenase metabolism, LT, and products of cyclooxygenase metabolism, PG, increase (5-7). The pattern of arachidonate metabolism and mucosal concentrations of LTB, and PGE, in acetic acidinduced colitis in rats is similar to that in human inflammatory bowel disease (37). The similarities in the pattern of arachidonic acid metabolism in acetic acid-induced colitis and in human inflammatory bowel disease reflect the general principle that, although gastrointestinal inflammatory responses can be initiated by a wide variety of stimuli, they are modulated
1000
(&kg)
Figure 8. Effect of varying doses of misoprostol pretreatment on 4% acetic acid-induced colitis. Misoprostol was administered intraluminally in 2 ml of vehicle 30 min before 4% acetic acid, and experiments were performed 2 days later. (A) Colonic ulceration expressed as percent of total colonic surface area; (B] in viva net colonic fluid absorption per centimeter. MIS0 pretreatment, 20 min before acetic acid induction, showed a dose-dependent mucosa-protective effect on macroscopic colonic mucosal ulceration, whereas net fluid and electrolyte absorption was only partially protected at the highest test dose. Values are mean f SEM for n 2 4 rats. (*)p co.01 compared with all other doses.
50 0” c
_
40
A
30
da $3 u_l-
G
20
sa 10
i
0
Effect of Misoprostol Measurements
on Transmural
Flux
As described, MIS0 pretreatment 30 min before 4% acetic acid induction conferred a macroscopic but not functional colonic absorptive mucosal protection. To determine whether the altered in vivo net absorption was a consequence of enhanced colonic secretion or of reduced absorption, basal and theophylline-stimulated intestinal short-circuit current and unidirectional sodium and chloride transmural flux measurements were examined on MISO-pretreated and nonpretreated colons. Misoprostol was administered intraluminally 30 min before 4% acetic acid induction, and the rats were killed 2 days later for measurements. As shown in Table 1, in control tissues, theophylline (lop3 M) decreased sodium absorption and stimulated net chloride secretion, a result similar to that previously described (40). Animals treated with 4% acetic acid showed a diminished net basal sodium absorption and did not show theophylline-stimulated chloride secretion. Misoprostol pretreatment of acetic acid-induced colitis maintained theophylline-stimu-
150 5 i=
100
k
50
z
1890;81:815-625
Misoprostol Provides a Colonic Mucosal Protective Effect During Acetic Acid-Induced Colitis in Rats RICHARD N. FEDORAK, LONNIE R. EMPEY, COLIN MACARTHUR, and LAURENCE D. JEWELL Department of Medicine, Division of Gastroenterology, University of Alberta, Edmonton, Alberta, Canada
This study determined if intracolonically applied prostaglandin E, analogue (misoprostol) had a mucosal protective effect in rats with 4% acetic acidinduced colitis. The effects of misoprostol were compared with those of 5-aminosalicylic acid and betamethasone. A single application of 4% acetic acid induced an experimental colitis which was maximal at 2 days and showed spontaneous macroscopic and histologic healing by 12 days. Misoprostol (100 pg/kg), but not 5-aminosalicylic acid or betamethasone, administered 30 min before induction of colitis, provided macroscopic and histologic coionic mucosal protection but not protection of in vivo fluid absorption. The mucosal protective effect of misoprostol was time, dose, and diluent volume dependent. In the presence of misoprostol-induced colonic morphologic but not functional absorptive mucosal protection, in vitro unidirectional sodium and chloride flux measurements showed protection of theophylline-stimulated chloride secretion but not sodium absorption. Protection of in vivo colonic fluid absorption, in addition to morphologic protection, could be achieved when misoprostol was administered between 2 and 16 min before induction of colitis or when the highest dose (1000 pg/kg) of misoprostol was examined. We conclude that intracolonic misoprostol administration provides unique mucosal protective effects in experimental colitis.
T
he products of arachidonic acid metabolism, eicosanoids, are important mediators of the intestinal inflammatory response as well as direct stimulators of intestinal fluid and electrolyte secretions (l-3). Therefore, it is suggested that eicosanoids may mediate the gastrointestinal inflammatory response in inflammatory bowel disease and may also contribute to the pathogenesis of diarrhoea (1.3-5). Indeed, in-
and Department
of Pathology,
flamed colonic mucosa in patients with active ulcerative and Crohn’s colitis has been shown to have enhanced levels of both leukotrienes (LT), products of 5-lipoxygenase metabolism, and prostaglandins (PC), products of cycle-oxygenase metabolism (5-17). Studies have reported that luminal levels of PG are inversely related to fluid and electrolyte absorption and that PG and LT levels are also positively correlated with disease activity, as determined by clinical, endoscopic, and histological gradings (12). Agents effective in the treatment of inflammatory bowel disease, such as corticosteroids, 5-aminosalicylic acid (5-ASA), and sulphasalazine, alter arachidonic acid metabolism pathways to varying degrees (12,18-20). However, the individual roles of cycle-oxygenase and 5-lipoxygenase products in gastrointestinal inflammation have not been fully elucidated. It has been suggested that the elevation of endogenous prostaglandin levels seen in vivo during inflammatory disease states may be beneficial, assisting in protecting the mucosa from insult (21-23). Certainly, exogenous PG in the stomach have been shown to have mucosal protective effects, preventing the gastric necrosis produced by such agents as ethanol, hydrochloric acid, sodium hydroxide, hypertonic sodium chloride, taurocholate, and thermal injury (24-27). Protection against experimentally induced colitis has been observed with intracolonic administrations of PGE, and l6,16 dimethyl PGE, (28,29,30). PGI, has inhibited indomethicin-induced small intestinal lesions in a dosedependent manner (16). Furthermore, drugs that selectively inhibit endogenous prostanoid production, such
Abbreviations used in this paper: I-AM, 5-amino~~alicylic acid; BETA, hetamethasone; LT, leukotriene; MISO, misoprostol; PG, prostaglandin. 0 1888 by the American Gastrwnter&gical Association OOM-5085/80/$3.88
616
GASTROENTEROLOGY Vol. 98, No. 3
FEDORAK ET AL.
as indomethacin
and flurbiprofen, are of no therapeutic benefit to patients with inflammatory bowel disease (31). Such agents, associated with reduced levels of PG in the urine and rectum, may actually aggravate colonic mucosal inflammation (32-35). In this study, we investigated if intraluminally applied PGE, analogue (misoprostol] would provide a colonic epithelial mucosal protective effect to the gastrointestinal mucosa of rats undergoing 4% acetic acid-induced colitis. We also compared the effects of misoprostol administration with those obtained with the administration of 5-ASA and betamethasone (BETA]. Materials and Methods Materials Misoprostol (MISO; Cytotec) [(+)methyl (lla, 13 E)-11,16-dihydroxy-l6-methyl-9-oxoprost-l3-en-l-oate] is a synthetic PGE, methyl ester analogue. The misoprostol used in this study, lot CD206-36-12, reference 02420, was a gift from G. D. Searle and Company, Skokie, Ill. It was prepared from the stock solution supplied by Searle to a concentration of 1 mg/ml in absolute ethanol and stored at -70%. Its stability at this temperature is greater than 1 yr (36). A dilution of 100 pg/ml for experimentation was prepared into aliquots on a weekly basis in an ethanol buffer vehicle containing 20% ethanol, 104 mM Na,HPO,, and 16 mM NaH,PO,. The pH of this solution was 7.4, stored at -70°C. A 5-ASA rectal suspension (Salofalk] (67 mg/ml, pH 4.1, stored at 4’C] was purchased from Interfalk Canada Inc. Mont Saint-Hilaire, Quebec. A betamethasone disodium phosphate rectal suspension (Betnesol) (50 bg/ml, pH 7.5, stored at 4X) was purchased from Glaxo Laboratories, Toronto, Ontario, Canada. Acetic acid was purchased from Fisher Scientific, Nepean, Ontario, Canada, and prepared to a 4% solution in water (pH 2.4). All remaining chemicals were reagent grade and purchased from Sigma Chemical Co., St. Louis, MO. lntraluminal Administration of Misoprostol, 5-Aminosalicylic Acid, and Betamethasone Misoprostol (a dosage of 100 pg/kg except where otherwise specified), was prepared for intraluminal administration immediately before experimentation by mixing the appropriate volume of the 100 pg/mI ethanol-buffer stock solution with 154 mM sodium chloride to a total volume of 2 mL (pH 7.4). For a 250-g rat, the concentration of misoprostol administered was thus 12.5 pg/ml in a 2.5% ethanol-buffer vehicle. 5-ASA, 67 mg/mL, and BETA, 50 pg/ml, were prepared for intraluminal administration by directly removing 2 ml from the pharmaceutically supplied enema solutions. Ethanol-buffer vehicle, 2.5%, (pH 7.4) was used for control animals. Nonfasting male Sprague-Dawley rats (250-275 g, Biotron, University of Alberta, Edmonton, Alberta, Canada) were anesthetized with pentobarbitol(55 mg/kg) and atropine (0.5
mg/kg), administered i.p. Through a sterile midline incision, the colon was isolated, and the junction of the cecum and ascending colon was occluded via a reversible ligature, without compromise of neural or vascular integrity. The colon was cleansed of its luminal contents with a warmed 154-mM sodium chloride solution, and the residual fluid was manually expressed out through the rectum. Two milliliters of the chosen agent (MISO, 5-ASA, BETA, or ethanol-buffer vehicle] was then injected into the colonic lumen just distal to the occluding ligature via a 25-gauge needle passed obliquely through the outer muscle layer. The colon was only mildly distended and all mucosal regions were equally coated. A rectal plug [Extra, Wrigley Co., Toronto, Ontario, Canada] was inserted to prevent fluid loss and to ensure an even mucosal contact of the intraluminal agent. The abdominal incision was then closed and the anesthetized rat placed prone in an incubator prewarmed to 37’C for 30 min (unless otherwise specified) before instillation of acetic acid.
Induction
of Acetic-Acid
Colitis
Acetic-acid colitis was induced as previously described (37). Briefly, the abdominal incision was reopened and the colon isolated. Residual intracolonic fluid was expressed out the rectum and 2 ml of 4% acetic acid (unless otherwise specified) was injected into the lumen of the colon through a &gauge needle passed obliquely through the colonic wall just distal to the occluding ligature. Immediately thereafter, 10 ml of air was injected to clear most of the acetic acid from the colon. The occluding ligature was removed and the midline incision closed. The animals were allowed to recover from anesthesia and were housed in a light-cycled room with free access to standard rat chow pellets (5001, Purina Mills Inc., St. Louis, MO.) and water. Two days later (unless otherwise specified) the rats were killed with a pentobarbitol overdose (240 mg/kg, i.v.). and their colons were removed for experimentation.
Macroscopic
Studies
The colon was rapidly excised, opened along its mesenteric border, and gently rinsed of its luminal contents with an iced solution of 154 mM sodium chloride. The colon was then placed flat, mucosal surface upwards, on a glass plate chilled to 4°C. A transparent acetate was placed 5 mm above the mucosal surface and the area of ulceration and total surface area were traced by a single observer (L.R.E.J. The respective area in cm’ was then calculated using a Zeiss computerized videoscope (Videoplan, Carl Zeiss Co., Canada, Toronto, Ontario].
Histologic Studies Representative sections, taken from the colon at 3,6. 9, and 12 cm distal to the cecal-ascending junction, were fixed in 4% buffered formaldehyde. For light microscopy, &pm, formaldehyde-fixed, paraffin-embedded sections were stained with H&E. Histological grading was determined in a blinded fashion by two observers (L.D.J. and C.M.) and represented the
March 1999
COLONIC MUCOSAL PROTECTION IN COLITIS
numerical sum of two criteria. Mucosal ulceration was graded as follows: 0, no ulceration: 1, focal ulceration: 2, multifocal ulceration; and 3, diffuse ulceration. Depth of injury was graded as follows: 0, no injury; 1, mucosal involvement only; 2. mucosal and submucosal involvement; and 3, transmural involvement. Therefore, there was a minimum histologic grade of 0 and a maximum histologic grade of 6.
Unidirectional
In Viva CoJonic Fluid Absorption
Studies
Nonfasting rats were anesthetized with pentobarbito1 and atropine and kept warm with a thermostatic heat lamp. The intestinal tract was exposed through a midline abdominal incision. An occluding ligature was placed at the cecal-ascending colon junction, and the luminal contents were flushed out the rectum with a warm 154-mM sodium chloride solution instilled via a cannula inserted through an incision just distal to the proximal occluding ligature: residual saline was emptied by gentle manual expression. An intestinal loop of -15 cm in length, beginning 2 cm below the cecal-colonic junction and extending distally to the peritoneal reflection, was created with ligatures. In isolating the loop, care was taken not to compromise mesenteric, vascular, or neural continuity. A 27-gauge needle was inserted obliquely through the outer muscle layer along the antimesenteric border, and 2 ml of 154 mM sodium chloride, prewarmed to 37°C. was instilled into the empty loop. No fluid leakage was detected, and the loop was only mildly distended. The viscera were returned to the abdominal cavity and the incision was closed. The rats were allowed to recover from the anesthesia. Sixty minutes after abdominal closure, the animals were killed by pentobarbitol overdose, and the colonic loop was removed. The length of each loop was measured and each was weighed, both full and emptied, to determine the residual intraluminal volume. The loop was then dried in a 50°C oven for 24 h. Results were expressed as the difference between initial and residual luminal loop volume per gram dry weight or per centimeter of bowel.
In Vitro Electrical
Measurements
In vitro Ussing chamber studies were performed on 6 l-cm segments of colon from each rat, 3 immediately proximal and 3 immediately distal to the middle of the transverse colon. These tissues were removed immediately after the animal was killed, cut open along the mesenteric border, and incubated with a solution of the following ionic composition: Na. 143 mM; K, 5 mM; Mg, 1.1 mM; Ca, 1.25 mM; Cl, 124 mM; HCO,, 25 mM; HPO,, 1.65 mM; H,PO,, 0.3 mM; and fructose, 20 mM. The pH was 7.4 when gassed with 5% CO, in 95% 0, at 37°C. Colonic segments were placed serosal side up on a Plexiglas dissecting board and stripped of their serosa and underlying longitudinal muscle layer using flat-edged dissecting forceps. Mucosal strips were then mounted in Ussing chambers, and transmural electric potential difference, resistance, and intestinal short-circuit current were determined as previously described (38).
617
Flux Measurements
Transepithelial unidirectional fluxes of Na and Cl from mucosa to serosa and from serosa to mucosa were measured under short-circuited conditions. For flux measurement, 1 &i of %l and 1 pCi of =Na were added to either mucosal or serosal reservoirs after the tissues were mounted. Tissue pairs that differed in electrical resistance by greater than 25% were discarded. All flux experiments were conducted in bicarbonate Ringer’s solution as listed above. Effects on colonic ion transport of theophylline (10e3 M) were determined in sequential periods on the same tissue. Steady-state unidirectional basal fluxes were measured over 15 min, beginning 30 min after addition of tracers. Steadystate unidirectional theophylline fluxes were measured over 20 min. beginning 20 min after addition of theophylline. Fluxes were calculated as previously described (38).
Statistical
Analyses
Statistical analyses of data were performed by analysis of variance. This was done by multivariate profile analysis of repeated measures using a microcomputer statistical software package (EPISTATZ, Round Rock, Tex.]. When overall analysis showed significant interaction, Students’ t-test was used to examine the significance of differences.
Results Effect of Acetic
Acid on CoJonic Mucosa
Macroscopic changes. Intracolonic application of acetic acid produced an immediate blanching and mild muscular contraction of the colonic mucosa which resolved spontaneously over 20-30 min. At the time of death, the serosal surface of the colon was found to be hyperemic; however, in no instance was perforation identified. Figure 1 illustrates the macroscopic colonic mucosal lesions found at varying time intervals following 4% acetic acid-induced colitis. Mucosal lesions consisted of circumferential ulcerations covered with a thick white exudate and were located primarily in the region of the transverse and descending colon. The ascending colon and rectum were less affected. The reason for this unequal distribution of injury is unclear, although it may represent a pooling of residual acetic acid in a dependent segment of colon during anesthetic recovery. Macroscopic ulceration was maximal at 2 days and was completely resolved 12 days following the application of 4% acetic acid. Histologic changes. As shown in Figure 2, histological injury grade was maximal within a period of 1 day following the 4% acetic acid application and had returned to near control levels [grade = 0) by 12 days. Figure 3 illustrates the specific light microscopic features of 4% acetic acid-induced colitis. One day
618
FEDORAK
ET AL.
GASTROENTEROLOGY
Vol. 98, No. 3
Figure 1. Macroscopic colonic mucosal lesions at varying time intervals after intracolonic application of 4% acetic acid. The caecum (CM) was removed and colons opened along the mesenteric border. (A) 6 h; (B) 1 day; (C) z days; (D) 4 days; (E) 6 days; and (F) 12 days after acetic acid induction, respectively. Macroscopic ulcerations were maximal at 2 days, markedly improved by 4 days, and completely resolved by 12 days after induction of colitis.
after the application of 4% acetic acid, the acute inflammatory response extended into the submucosa, which was edematous and contained extravascular neutrophils. Small blood vessels were plugged with
6r
in T L
11 0
0
1 2
DAYS
4
6
FOLLOWING
8
ACETIC
10
12
ACID
Figure 2. Colonic histological grade at varying time intervals following intracolonic application of 4% acetic acid. Histological grade was determined by summation of mucosal ulceration and depth of injury score. Maximal injury = 6 Unite; minimal injury (control) = 0 Units. Line drawn by linear regression analysis. Light-microscopic injury following 4% acetic-acid colitis paralleled macroscopic changes although histologic recovery lagged behind macroscopic improvement. Points represent mean * S.E.M. of I)determinations for n = 4 rats.
polymorphonuclear cells. In some animals the colon was entirely necrotic, whereas in others the muscularis propria appeared viable. By the second day, regions of mucosal regeneration were evident, with restoration of mucosal integrity apparent in occasional areas. However, there were prominent areas of necrosis with evidence of hemorrhage and transmural inflammation. Healing was noted to be progressing at 4 and 6 days postapplication. Although return of mucosal integrity was widespread, focal areas of residual deep ulceration persisted. By day 12, the acute inflammation had almost completely resolved, and mucosal regeneration was complete. However, a mild architectural disorder of focal lamina propria fibrosis and bifid and irregular glands remained. Rats instilled intraluminally with a control solution of saline showed normal microscopic architecture at all time intervals. However, compared with gross macroscopic changes, mucosal histological injury following induction of acetic-acid colitis persisted at 4 and 6 days, despite a significantly improved macroscopic appearance. This is similar to findings in human ulcerative colitis, where histological recovery lags behind macroscopic improvement (39). Fluid absorption in vivo. As shown in Figure 4, intracolonic acetic acid produced a concentrationdependent decrease in net colonic fluid absorption.
COLONIC MLJCOSAL PROTECTION
March 1990
IN COLITIS
619
Figure 3. MIcrographs iHuetratlng the coume of untrasted 4% acetic add-induced colitis over a 12-dayperiod. A. Control colon before application of acetic acid.
the
B. Extensive mucosal necrosis with submucosal edema 1 day after application of acetic acid. C. Commencing mucosal regeneration is evident at 2 days. D. Restitution of an essentially normal mucosa is complete by 12 days (H&E; original magnification x40).
Maximal transport injury was present at a 4% concentration of acetic acid and remained constant thereafter: this was the case whether the results were expressed as pi/g dry weight (data not shown] or as J/cm of bowel [Figure 4). Therefore, subsequent experiments were performed using an intracolonic acetic-acid concentration of 470, and macroscopic, histologic, and functional observations made when injury is found to be maximal, i.e., 2 days after the application of acetic acid.
-100
1
j
1 ACETIC
1
3
5
ACID CONCENTRATION
t
f
7
9
-
(X)
Figure 4. Effect of varying lntracolonic acetic acid concentrations on in vlvo net colonic fluid absorption per centimeter. Acetic acid was appIied, and colonic fluid absorption was determined 2 days later. Maximal alteration of net fluid and electrolyte transport was seen at a 4% concentration of acetic acid. Values are mean + SEM for n 256 rats.
Effects of Misoprostol, 5-Aminosalicylic Acid, and Betamethasone on Acetic Acid-Induced Colitis After determining the effectiveness of acetic acid in inducing colitis, we next examined the mucosal protective effect of intraluminal MISO, 5-ASA, and BETA. Each agent was administered 30 min before application of 4% acetic acid. Rats were then killed 2 days after 4% acetic acid application, and the experiments were performed. Macroscopic changes. As shown in Figure 5, pretreatment with intraluminal misoprostol completely prevented the formation of macroscopic coIonic lesions, whereas intraluminal pretreatment with 5-ASA and BETA did not. Figure 6A shows the effect of each pretreatment agent on the surface area of colonic ulceration. Four percent acetic acid applicart 7.0% of tion alone resulted in ulceration of 43.1% the total colonic surface area. Pretreatment with misoprostol reduced acetic acid-induced ulceration to 1.9% * 1.7% (p < O.OOl),whereas pretreatment with 5-ASA or BETA did not (33.5% k 9.5% and 33.7% t 7.3570, respectively). Administration of the ethanolbuffer vehicle alone, which was used to maintain the MIS0 in solution, had no effect on the degree of ulceration. Histologic changes. As shown in Figure 6B, MISO-pretreated animals has a significant improvement in histologic grade compared with 5-ASA- and
620
FEDORAK
ET AL.
GASTROENTEROLOGY
Vol. 98, No. 3
Figure 5. Effect of MISO, 5-ASA, and BETA pretreatment on macroscopic colonic mucosal appearance induced by 4% acetic acid. Agents were administered intraluminally 20 min before 4% acetic acid induction, and colons were excised 2 days later. The caeca (CM) have been removed and the colons opened along the mesenteric border. (A) Vehicle-; (B] BETA- (50 &ml); (C) 5-ASA-(W mg/ml); and (D) MISO(100 rm/lU, -12.5 rg/ml) pretreated colons. Misoprostol pretreatment completely protected the colonic mucosa from 4% acetic acid-induced macro&opic injury.
BETA-pretreated rats. From micrographs of colonic mucosa, acetic acid-induced colitis in which the bowel was pretreated with vehicle only showed severe mucosal and submucosal inflammation with commencing epithelial regeneration. Colon pretreated with BETA or 5-ASA did not differ significantly in the extent of damage. By contrast, pretreatment with MIS0 afforded nearly complete histological protection from 4% acetic acid-induced mucosal injury. Fluid absorption in vivo. As shown in Figure 7, sham-operated, age-matched, control rats had a net colonic fluid absorption. Two days after 4% aceticacid application, rats showed net colonic fluid secretion. Pretreatment with MISO, 5-ASA, or BETA did not alter this net secretory state, whether results were expressed as pi/g dry weight (data not shown) or as hi/cm of bowel (Figure 7). These results of altered colonic fluid absorption 2 days after induction of colitis with MIS0 pretreatment contrast with the near-normal macroscopic and microscopic appearance seen with misoprostol pretreatment. To examine this further, we determined in vivo colonic fluid absorption at 12 days following 4% acetic acid application. Under all experimental conditions, i.e., whether rats were pretreated with MISO, 5-ASA, BETA, or ethanol-buffer vehicle, macroscopic and light-microscopic histologic changes had spontaneously returned to near normal by 12 days (Figures 1 and 2), whereas colonic fluid transport alterations persisted (Figure 7). However, rats pretreated with misoprostol or 5-ASA, showed a significantly greater degree of return to normal colonic absorption than vehicle- or BETA-pretreated rats [Figure 7).
Effect of Varying Doses of Misoprostol To further examine the mucosal protective effect of misoprostol, we next determined the effect of varying doses of intraluminal misoprostol on colonic ulceration and in vivo absorption. Misoprostol was administered 30 min before induction of colitis with 4% acetic acid: experiments were then conducted 2 days after acetic-acid application. Macroscopic changes. As shown in Figure 8A, intraluminal misoprostol preadministration showed a dose-dependent mucosal protective effect against coionic macroscopic ulceration. However, only the highest test dose used (1000 pg/kg) completely prevented ulceration. There was a single colonic ulcer in 1 of 6 rats in the lOO-pg/kg dose group, whereas at the 1 pg/kg dose all colons were ulcerated and the ulcer size fluctuated widely (range, X2%-39.5% total surface area). The macroscopic mucosal protective ED,, of misoprostol was approximately 1 pg/kg. Fluid absorption in vivo. In contrast to the macroscopic mucosal protection seen with low doses of misoprostol, net in vivo fluid absorption remained secretory, and only at the highest test dose (1000 pg/kg) was net colonic fluid secretion partially reversed to net fluid absorption (Figure 8B). As well, misoprostol pretreatment at 1000 pg/kg did not alter basal colonic fluid absorptive rates in control rats (data not shown). Onset and Duration of Action of Misoprostol Treatment We next determined the onset and duration of action by pretreatment with misoprostol(lO0 pg/kg) at
March 1990
COLONIC MUCOSAL PROTECTION
CONTROL
AA
MIS0
5ASA
BETA
IN COLITIS
621
was administered 0.5 or 1 min before 4% acetic acid, net colonic fluid transport remained secretory (Figure 9B) despite the normal macroscopic appearance described. On the other hand, misoprostol administration between 2 and 16 min before 4% acetic-acid application conferred a mucosal protective effect on in vivo colonic fluid transport, reversing net fluid secretion to net fluid absorption. Colonic net fluid transport once again became secretory, despite a normal macroscopic appearance, when misoprostol pretreatment occurred more than 32 min before application of the acetic acid.
Volume of Dilutent
CONTROL
AA
MIS0
5ASA
BETA
Figure 6. Effect of MISO, L-ASA, and BETA pretreatment on 4% acetic acid-induced colitis. (A) Colonic ulceration expressed as percent of total colonic surface area; (B) histologic grade, determined by summation of mucosel ulceration and depth of damage score: maximal injury = 6 U, minimal injury (control) = 0 U. Agents were administered intraluminally 30 min before 4% acetic acid, and rats were killed 2 days later. AA, pretreated with vehicle only; MISO, pretreated with MIS0 (100 &kg, -12.5 rg/mlk 5-ASA, pretreated with S-MA (67 n&ml); BETA, pretreated with BETA (50 fig/ml]. Results are compared with sham-operated, age-matched controls. Misoprostol pretreatment, but not 5-ASA or BETA, reduced 4% acetic acid-induced colonic ulceration and histologic injury to near control values. Values represent means * SEM of n = 6 rats. (*)p t0.001 compared with all other groups.
Misoprostol offered greater protection against colonic ulceration when diluted in 1 or 2 ml of dilutent than when diluted in 0.5 or 0 ml of dilutent [Figure 10). Indeed, misoprostol in 0.5 ml dilutent was only slightly more useful than no misoprostol at all. It was observed that the 0.5 ml volume remained localized at the site of instillation and was almost completely absorbed within the 30-min pretreatment time interval. The l- and 2-ml volumes were found to cover all regions of colonic mucosa uniformly.
0 ENTROL m MIS0
r
!Xl &I3
5ASA BETA
50
0
-50 1
varying time intervals before the application of 4% acetic acid. Experiments were then conducted 2 days after acetic-acid application. Macroscopic changes. As shown in Figure 9A, when misoprostol was administered between 0.5 and 32 min before 4% acetic-acid application, complete mucosal protection was observed, with the area of colonic ulceration being reduced to near zero. Colonic mucosal protection diminished when misoprostol was administered more than 32 min before acetic acid. These results are similar to those of the onset and duration of action of PG in ethanol-induced gastric lesions (24). Fluid absorption in vivo. When misoprostol
*
1
-100 0
‘A 12
2 DAYS
FOLLOWING
ACETIC
ACID
Figure 7. Effect of MISO, 5-ASA, and BBTA pretreatment on in viva net colonic fluid absorption, determined at varying time intervals following 4% acetic acid-induced colitis. Vehicle only (AA); MISO, 100 &kg, -12.5&ml; S-AM, 67mg/ml; or BETA, 50 fig/ml was preadministered intraluminrrliy 30 min before 4% acetic acid, and rats were killed either 2 days or 12 days later. Net colontc fluid absorption is secretory in all groupn at 2 days. By 12 days following acetic-acid colitis, MISO- and 5-ASA-pretreated groups showed signikant improvement in net fluid abnorption. Values are mean * SEM for n z 6 rata and compared with sham-operated, age-matched controls. (‘lp to.01 compared with all groups at 2 days.
622
FEDORAK
GASTROENTEROLOGY
ET AL.
lated net colonic chloride secretion prove net sodium absorption.
Vol. 98, No. 3
but did not im-
Discussion
0
1
10 MISOPROSTOL
100
The present study was undertaken to investigate the mucosal protective effects of misoprostol on acetic acid-induced colitis and to compare the effects of MIS0 with those of 5-ASA and BETA. In active human inflammatory bowel disease, levels products of 5-lipoxygenase metabolism, LT, and products of cyclooxygenase metabolism, PG, increase (5-7). The pattern of arachidonate metabolism and mucosal concentrations of LTB, and PGE, in acetic acidinduced colitis in rats is similar to that in human inflammatory bowel disease (37). The similarities in the pattern of arachidonic acid metabolism in acetic acid-induced colitis and in human inflammatory bowel disease reflect the general principle that, although gastrointestinal inflammatory responses can be initiated by a wide variety of stimuli, they are modulated
1000
(&kg)
Figure 8. Effect of varying doses of misoprostol pretreatment on 4% acetic acid-induced colitis. Misoprostol was administered intraluminally in 2 ml of vehicle 30 min before 4% acetic acid, and experiments were performed 2 days later. (A) Colonic ulceration expressed as percent of total colonic surface area; (B] in viva net colonic fluid absorption per centimeter. MIS0 pretreatment, 20 min before acetic acid induction, showed a dose-dependent mucosa-protective effect on macroscopic colonic mucosal ulceration, whereas net fluid and electrolyte absorption was only partially protected at the highest test dose. Values are mean f SEM for n 2 4 rats. (*)p co.01 compared with all other doses.
50 0” c
_
40
A
30
da $3 u_l-
G
20
sa 10
i
0
Effect of Misoprostol Measurements
on Transmural
Flux
As described, MIS0 pretreatment 30 min before 4% acetic acid induction conferred a macroscopic but not functional colonic absorptive mucosal protection. To determine whether the altered in vivo net absorption was a consequence of enhanced colonic secretion or of reduced absorption, basal and theophylline-stimulated intestinal short-circuit current and unidirectional sodium and chloride transmural flux measurements were examined on MISO-pretreated and nonpretreated colons. Misoprostol was administered intraluminally 30 min before 4% acetic acid induction, and the rats were killed 2 days later for measurements. As shown in Table 1, in control tissues, theophylline (lop3 M) decreased sodium absorption and stimulated net chloride secretion, a result similar to that previously described (40). Animals treated with 4% acetic acid showed a diminished net basal sodium absorption and did not show theophylline-stimulated chloride secretion. Misoprostol pretreatment of acetic acid-induced colitis maintained theophylline-stimu-
150 5 i=
100
k
50
z
