Abstract
The objective of this study was to quantitatively characterize the effects FK506 on the pathophysiology observed in a model of chronic granulomatous colitis in rats and compare these effects to those obtained with cyclosporin A (CyA). Chronic granulomatous colitis was induced in female Lewis rats via intramural (subserosal) injections of peptidoglycan/polysaccharide (PG/PS) into the distal colon. Rats then received daily injections (i.m.) of either vehicle for CyA (0.5 ml/kg cremophor), CyA in vehicle (25 mg/kg), saline (0.5 ml/kg) or FK506 (1 mg/kg in saline), beginning 7 days after PG/PS injection and continuing for an additional 2 weeks. On day 21, we found that the intramural injection of PG/PS produced a chronic colitis that was associated with hepatic and splenic granulomatous inflammation. Daily treatment with CyA or FK506 beginning 7 days after the induction of colitis resulted in significant inhibition in colonic mucosal permeability, colonic myeloperoxidase activity and plasma nitrate/nitrite levels when compared with their vehicle or untreated controls. In some instances, we noticed a significant vehicle-dependent anti-inflammatory activity. The incidence of peritoneal adhesions as well as the presence of hepatic and splenic granulomas induced by PG/PS were also significantly reduced in both the CyA- and FK506-treated groups. Taken together, these data suggest that immunosuppressive therapy is effective at attenuating both the colitis as well as the extraintestinal inflammation induced by PG/PS. We conclude that FK506 may be useful in the treatment of certain types of inflammatory bowel disease.
Patients with active IBD (ulcerative colitis, Crohn’s disease) are usually treated with sulfasalazine, 5-aminosalicylic acid and/or corticosteroids. Patients who become resistant to these standard drug therapies and/or who become steroid dependent may be given certain immunosuppressive agents, such as 6-mercaptopurine or azathioprine. Unfortunately, antimetabolite therapy possesses significant drawbacks including a slow onset of action, myelosuppression, increased risk of malignant transformation and pancreatitis (Sandborn and Tremaine, 1992). Recent clinical studies have suggested that cyclosporine may represent a viable alternative to these types of therapies and may be effective in controlling the chronic gut inflammation observed in corticosteroid-resistant patients with IBD (Sandborn and Tremaine, 1992; Lichtiger et al., 1994; Sartor, 1994). CyA selectively inhibits immune responses mediated by T lymphocytes. Specifically, it is thought that CyA blocks the production of interleukin-2 and interleukin-2 receptors by T cells (Faulds et al., 1993). In addition to this effect, CyA has been shown to inhibit interferon-γ production by T helper cells as well as by inhibiting the production of the B-cell activity factor (Faulds et al., 1993). Furthermore, studies by Kubeset al. (1991) have demonstrated that CyA interferes with neutrophil adhesion to the venular endothelium, which suggests that it may be effective at inhibiting granulocyte infiltration and the subsequent increases in vascular permeability. FK506 is a newly described immunosuppressant drug that shares many of the same pharmacological properties as CyA and has been demonstrated to be approximately 50 to 100 times more potent than CyA in its immunosuppressive effects (Kino et al., 1987; Morris, 1995). Although CyA has been shown to be effective in the treatment of certain types of severe IBD, there have been no clinical or experimental studies which have demonstrated the usefulness of FK506 in the treatment of IBD. Therefore the objective of this study was to quantitatively assess the effects of FK506 on the pathophysiology observed in a model of chronic granulomatous colitis in rats and compare these effects with those observed with a dose of CyA found to be protective in other animal models of chronic inflammation.
Materials and Methods
Induction of colitis.
Female specific pathogen-free Lewis rats (150–175 g) were housed in wire-mesh bottom cages and given water and standard laboratory rat chow ad libitum. A total of 54 rats were randomly assigned to six groups. All rats in each group were anesthetized via inhalation of isoflurane (Aerrane, Anaquest Inc., Liberty Corner, NJ), and the descending colon of each rat was exposed by laparotomy by aseptic technique. Each animal received intramural (subserosal) injections of either sterile saline or PG/PS into the distal colon as described previously (Yamada et al., 1993; Grisham et al., 1994). Animals in group 1 (n = 12) received 9 to 10 intramural (subserosal) injections of sterile saline (50 μl/injection). Each animal in the other five groups (groups 2–6) received 9 to 10 intramural injections of PG/PS (12.5 μg rhamnose/g b.wt.). Rats were allowed to recover from anesthesia and given free access to food and water.
Cyclosporine or FK506 therapy.
Cyclosporine (Sandimmune, cyclosporin concentrate for injection, Sandoz Pharmaceuticals, East Hanover, NJ) was diluted with saline to a concentration of 50 mg/ml and protected from light. FK506 was a gift from Fujisawa Pharmaceuticals (Osaka, Japan) and was diluted with saline to a concentration of 2 mg/ml immediately before use. Group 2 (n = 10) received no treatment and was sacrificed at day 21. Cyclosporine, FK-506 or their respective vehicles were injected intramuscularly into the thigh in a volume of 0.5 ml/day, beginning 7 days after PG/PS injection and continuing for an additional 2 weeks. Group 3 (n = 8) received 0.5 ml/kg/day of Cremophor (polyoxyethylated castor oil, Sandoz Pharmaceuticals; the vehicle for cyclosporine). Group 4 (n = 7) received 25 mg/kg/day of cyclosporine. Group 5 (n = 9) received 0.5 ml/kg/day of saline (vehicle for FK506), and group 6 (n = 8) received 1.0 mg/kg/day of FK506 in saline.
Surgery and mucosal permeability measurements.
Rats were fasted for 24 hr before surgery on day 21 after the intramural injection of saline or PG/PS into the distal colon. The animals were weighed and anesthetized via an intraperitoneal injection of 120 mg/kg sodium 5-ethyl-1 (1′-methyl-propyl)-2-thiobarbiturate (Inactin; Byk-Gulden, Konstanz, Germany). Body temperature was maintained at 37°C with a thermistor-controlled water pad (Aquamatic K-Modules K-20; Baxter, Valencia, CA). The animals underwent a tracheotomy, and the right femoral artery was cannulated for arterial pressure recording and blood sampling. The right femoral vein was also cannulated for injection of the isotope marker.
Both renal vessels were ligated to prevent rapid excretion of the radioisotope marker into the urine. The descending colon was isolated and cannulated at both the splenic flexure and the rectum by use of Silastic tubing (Dow Corning, Arlington, TN; internal diameter, 0.025 and 0.25, respectively) for infusion and correction of the modified Tyrode’s solution as described previously (Yamada et al., 1993). The perfused colon was returned to the abdominal cavity, and the abdominal wall was closed to minimize dehydration of the organs during the experiment. The luminal contents of the colon were removed by perfusion of warm (37°C) modified Tyrode’s solution for 30 min. Mucosal permeability was determined using the blood-to-lumen clearance of 51Cr-EDTA as described previously (Yamada et al., 1993; von Ritter et al., 1988).51Cr-EDTA (100 μCi) (Dupont de Nemours & Co., Boston, MA) was injected via the femoral vein catheter. After a 15-min equilibration period, the perfusate was collected every 10 min for 40 min for the appearance of 51Cr-EDTA. Plasma samples were taken at 40 min for use as reference counts and assessment of circulating NO2 − and NO3 − levels. To estimate the side effects and/or toxicity of cyclosporine and FK506, blood pressure and body weights were recorded.
Tissue preparation and biochemical analysis.
After permeability determinations, the animals were sacrificed viaan overdose of pentobarbital sodium (Butler, Columbus, OH), and the perfused colons were excised. The colons were opened longitudinally. The length and weight of the colons were recorded; the tissue was sectioned and examined for histology. Wet-to-dry weight ratios and MPO determinations were also quantified for the colons. Livers and spleens were also excised, weighed and saved for histology. Colonic wet-to-dry weight ratios were calculated by dividing the wet weight of each sample by its dry weight after a 48-hr incubation at 80°C. Colonic dry weights were determined from the wet weights of the whole samples and wet-to-dry weight ratio. Results were expressed as the dry weight divided by the length of sample. MPO activity was determined as described previously (Yamada et al., 1993). MPO activity was expressed as units per centimeter of colon. For histological analysis, colons, livers and spleens from each group were fixed, dehydrated and embedded in JB-4 (Polysciences, Inc., Warrington, PA). Two-micrometer sections were cut on glass knives and stained with hematoxylin and eosin.
Plasma levels of NO3 − and NO2 − were determined spectrophotometrically with nitrate reductase and the Griess reagent (1% sulfanilamide/0.1% naphthylenediamine dihydrochloride/2.5% H3PO4) as described previously (Grisham et al., 1995).
Statistical analyses.
All results are expressed as mean ± S.E.M. Fisher’s protected least significant difference was used for the comparisons between all groups. Results were considered statistically significant at P < .05.
Results
Intramural (subserosal) injection of PG/PS into the distal colon of genetically susceptible female Lewis rats produces a chronic granulomatous colitis with liver and spleen inflammation. Table1 demonstrates that daily treatment of colitic rats with CyA (25 mg/kg/day) beginning 7 days after the induction of colitis results in decreased body weights of these animals when compared with their vehicle-treated colitic group. Daily treatment with FK506 did not significantly alter body weight compared with their saline-injected counterparts. In addition, we found that after 2 weeks of treatment, mean arterial blood pressure was significantly lower in the CyA-treated group than in their vehicle-treated control group (table 1).
The anti-inflammatory activity of CyA and FK506 was suggested by the ability of these immunosuppressive drugs to inhibit PG/PS-induced increases in liver and spleen weights (table 2) as well as histological inspection of the tissue (figs. 1 and2). Colonic weight-to-dry weight ratios and splenic and liver weights were all significantly reduced in the CyA group compared with the untreated or vehicle-treated groups. FK506 treatment had no significant effect on colonic wet-to-dry ratios or colonic dry weights, but did attenuate the PG/PS-induced increases in spleen and liver weights (table 2). Figure 3 demonstrates that daily administration of either CyA or FK506 beginning 1 week after the induction of colitis resulted in a significant inhibition of the PG/PS-induced increase in colonic mucosal permeability when compared with the untreated colitic rats. Interestingly, there was no statistically significant inhibition of mucosal permeability when compared with vehicle-treated rats. Figure 4 illustrates the effects of vehicle as well as CyA or FK506 treatment on PG/PS-induced increases in granulocyte infiltration into the colon as measured by increases in colonic MPO content. Daily injections of either CyA or FK506 significantly reduced colonic MPO activity when compared with untreated colitic rats (fig. 4). We also found that daily injection of cremophor, but not saline, significantly attenuated the PG/PS-induced increases in colonic MPO activity when compared with the untreated group. Again, we did not observe a statistically significant attenuation in colonic MPO with CyA or FK506 when compared with their vehicle-treated counterparts. We did observe substantial anti-inflammatory activity of the immunosuppressive agents by histological inspection of the tissue (fig. 5). As reported previously, PG/PS-induced colonic, liver and spleen inflammation is associated with enhanced NO production as measured by increases in fasted plasma levels of NO2 − and NO3 − (Yamada et al., 1993; Grishamet al., 1994). We found that the administration of CyA or FK506 significantly inhibited the elevation in plasma NO3 − and NO2 −concentrations compared with their vehicle-treated groups (fig.6). Quantitative morphometric analysis revealed that CyA but not FK506 administration attenuated the PG/PS-induced increases in mucosal thickness compared with untreated colitic animals (fig.7), whereas both CyA and FK506 significantly attenuated submucosal thickness when compared with vehicle-treated colitic rats (fig. 8).
Discussion
Data from a recent randomized, double-blind, placebo-controlled clinical trial demonstrated that high-dose intravenous CyA induces rapid clinical improvement in patients with active ulcerative colitis refractory to corticosteroid therapy (Sandborn et al., 1992;Lichtiger et al., 1994; Sartor, 1994). Although this study presents the exciting possibility that this immunosuppressive agent may be used to treat refractory IBD, high-dose, long-term CyA is associated with toxic side effects including hypertension, paresthesia, headache and elevated serum creatinine levels (Sandborn et al., 1992; Lichtiger et al., 1994; Sartor, 1994). There is little doubt that these toxic effects limit the use of this drug. More recent investigations have demonstrated that another closely related immunosuppressive drug, FK506, is very useful in the treatment of certain immunological conditions such as the tissue rejection that may occur in organ transplantation (Kino et al., 1987; Dawsonet al., 1993; Harding et al., 1989). FK506 has been found to be especially effective for preventing rejection of intestinal transplants (Morris, 1995). This immunosuppressive agent does in fact possess nephrotoxic properties similar to CyA at comparable doses including reduced glomerular flow and renal cortical blood flow, enhanced vascular resistance and arteriolar vascular dysfunction (Morris, 1995). FK506 is 50 to 100 times more potent an inhibitor of lymphocyte activation than is CyA. Thus FK506 may represent potential alternative therapy to the treatment of refractory IBD. As a first step toward determining whether FK506 may represent an effective mode of therapy for the treatment of severe IBD, we systematically quantified the anti-inflammatory properties of this drug in our model of chronic granulomatous colitis and compared these data with those obtained with CyA. Although not identical with either ulcerative colitis or Crohn’s disease, this model does in fact mimic some of the histopathological features of Crohn’s colitis. Indeed, it has recently been reported that PG/PS is present in the colonic bowel wall of patients with active Crohn’s disease, which suggests that PG/PS may play a role in possibly initiating and/or perpetuating the inflammatory response (Klasen et al., 1994). Furthermore, the colitis observed in this model is responsive to oral sulfasalazine, which suggests that this model shares a therapeutic response to distal bowel disease (Grisham et al., 1996).
Data obtained in the present study demonstrate that both CyA and FK506 have similar but not identical patterns of anti-inflammatory activity in our model of chronic colitis in rats. We found that CyA and, to a lesser extent, FK506 attenuated the grossly visible signs of inflammation in the colon (adhesions, bowel wall thickness, nodules) as well as in the liver and spleen (data not shown). Furthermore, we found that daily administration of either CyA or FK506 beginning 7 days after the induction of colitis inhibited the PG/PS-induced increase in mucosal permeabilities (fig. 3) and colonic MPO activities (fig. 4) when compared with rats with untreated colitis. Histological inspection and quantitative morphometric analyses of the liver, spleen and colon confirmed the anti-inflammatory activity of the immunosuppressive drugs, which demonstrated that CyA and FK506 were effective at attenuating leukocyte infiltration (figs. 1, 2 and 5). Furthermore, CyA was effective at attenuating the PG/PS-induced increases in mucosal and submucosal thickness as well as inhibiting the increase in crypt depth (figs. 7, 8, 9). However, daily administration of CyA did appear to have an adverse effect on the rats, as noted by their loss of body weight and decreased systemic blood pressure (table 1). Major side effects that can develop during cyclosporine therapy in humans include nephrotoxicity, hypertension, paresthesia, headache, nausea and vomiting, hypertrichosis and anaphylaxis. Specific toxicities of FK506 noted in the rat, baboon (Ohara et al., 1990), dog (Ochiaiet al., 1987) and rabbit (Blackham and Griffiths, 1991) are nephrosis, anorexia and weight loss. The weight loss and lower blood pressure in the CyA treatment group of this study may be explained by anorexia caused by the toxicity of immunosuppressants and consequent decrease of circulating blood volume. The low value of colonic wet-to-dry ratios (table 2) in the group administered cyclosporine may be attributed to the dehydration caused by the toxicity of immunosuppressant.
Unexpectedly, we found that daily injection with cremophor or saline did provide a certain degree of anti-inflammatory activity when compared with untreated colitic rats. For example, daily injections of cremophor significantly reduced colonic MPO activity as well as crypt depth when compared with untreated colitic animals (figs. 4 and 9). In other experiments, daily injection with cremophor or saline tended to reduce the PG/PS-induced increases in mucosal permeability and MPO activities, although these differences were not statistically significant. Because of the tendency for the vehicles to attenuate some of the indices of inflammation coupled with the fact that relatively large groups were used for statistical comparisons (i.e., four groups in each case), we were unable to observe statistically significant alterations in PG/PS-induced mucosal permeabilities, MPO activity or mucosal thickness when compared with vehicle-treated colitis. The mechanisms by which cremophor or saline, presumably “inert” vehicles, significantly attenuated some of the PG/PS-induced inflammation are not certain; however, there are at least two possibilities. First, daily injections of these vehicles may induce a substantial stress response that results in the release of corticosteroids. Enhanced circulating levels of corticosteroids induced by stress could conceivably dampen the inflammatory response. A second possibility may be related to the reported bioactivity of cremophor. This lipophilic substance has been shown to inhibit prostacyclin synthesis in isolated tissue thereby promoting vasoconstriction (Brunkwall and Bergqvist, 1993; Besarab et al., 1987). Because leukocyte accumulation and possibly other mediators of inflammation are blood flow dependent, it is conceivable that cremophor-induced vasoconstriction may attenuate leukocyte accumulation by limiting the delivery of cells and growth factors viadecreases in blood flow.
A hallmark feature of this experimental model of chronic colitis as well as human IBD is enhanced production of NO (Yamada et al., 1993; Grisham et al., 1994). In fact, we have demonstrated that NO plays an important role in mediating some of the pathophysiology in our model of granulomatous colitis (Yamada et al., 1993; Grisham et al., 1994). In the present study, we found that both CyA and FK506 significantly inhibited the production of NO in vivo as measured by their ability to attenuate the PG/PS-induced increases in plasma levels of NO3 −/NO2 − when compared with their vehicle-treated controls (fig. 6). Because cremophor treatment significantly reduced MPO activity but did not attenuate the PG/PS-induced increases in plasma NO3 −/NO2 −, we conclude that at least some of the NO formed systemically in this model is being produced by cells other than leukocytes (e.g., epithelial cells, mast cells, endothelial cells). At present, it is difficult to determine which of the three organ systems involved in this model of colitis (i.e., colon, liver or spleen) is the primary generator of NO. However, previous studies from our laboratory have demonstrated that certain NOS inhibitors are effective at attenuating colonic, hepatic and splenic inflammation, which suggests that all three organs produce significant amounts of NO (Grishamet al., 1994).
The mechanisms by which CyA or FK506 may attenuate the multiorgan chronic granulomatous inflammation are not known; however, we suspect that these drugs modulate lymphocyte activation and subsequent cytokine production. It is known that lymphocyte chemotaxis, but not neutrophil or monocyte chemotaxis, are impaired by these types of immunosuppressive agents. Furthermore, recently published studies have shown that these immunosuppressants exert effects on other inflammatory cell types, such as neutrophils (Kubes et al., 1991; Pigattoet al., 1988; Kolb et al., 1990; Wallace et al., 1992; Wenzel-Seifert and Seifert, 1993; Suzuki et al., 1993) and mast cells (Amon, 1992; Hatfield et al., 1992). CyA and an investigational immunosuppressant, L-683,590, have been shown to attenuate the colonic permeability and neutrophil infiltration in a model of acute self-limiting colitis (Higa et al., 1993). The daily dose of FK506 used in this study (1.0 mg/kg) has been shown to inhibit ischemia/reperfusion-induced neutrophil infiltration in the rat liver (Suzuki et al., 1993) and cat small intestine (Kubes et al., 1991). However, in these studies, FK506 was administered before the induction of injury, whereas we administered the drug therapeutically at 7 days after induction of colitis.
Another possible mechanism for the inhibitory effects of these immunosuppressive agents may be their ability to affect NO metabolism (Dawson et al., 1993; Conde et al., 1995; Marumoet al., 1995). It is known, for example, that the immunosuppressive effects of cyclosporine and FK506 are mediated by a selective inhibition of the early-phase T-cell activation genes, including interleukin-2 (Tocci et al., 1989; Bickel et al., 1987). Both cyclosporine and FK506 block thecis-trans-peptidyl-prolyl-isomerase activity of their cytosolic binding proteins (immunophilins), cyclophilin and FK506 binding protein (Furman et al., 1992).cis-trans-peptidyl-prolyl-isomerase may be important for the natural folding of these proteins; however, the common target of the drug-immunophilin complexes could be calcineurin, a protein phosphatase, which plays a critical role in Ca++-dependent T-cell activation (Furman et al., 1992). It has been proposed that calcineurin regulates the phosphorylation and activation of NOS. Thus, CyA and FK506 may prevent the calcineurin-mediated dephosphorylation of NOS and diminish the enzymatic production of NO. Indeed, we have demonstrated that chronic NOS inhibition with two pharmacologically distinct NOS inhibitors provides significant anti-inflammatory activity in our model of chronic granulomatous colitis (Grisham et al., 1994). Taken together, our data suggest that FK506 may be useful in the treatment of chronic gut inflammation at much lower doses than CyA.
Footnotes
-
Send reprint requests to: Dr. Matthew Grisham, Department of Physiology, LSU Medical Center, P.O. Box 33932, 1501 Kings Highway, Shreveport, LA 71130.
-
↵1 Some of the work reported in this study was supported by grant DK 47663 from the National Institutes of Health.
- Abbreviations:
- CyA
- cyclosporin A
- MPO
- myeloperoxidase
- NO
- nitric oxide
- NOS
- nitric oxide synthetase
- PG/PS
- peptidoglycan/polysaccharide
- NO2−
- nitrite
- NO3−
- nitrate
- IBD
- inflammatory bowel disease
- EDTA
- ethylenediaminetetraacetic acid
- Received May 21, 1996.
- Accepted October 21, 1996.
- The American Society for Pharmacology and Experimental Therapeutics