A CELL-MEDIATED IMMUNE MODEL OF INFLAMMATORY BOWEL DISEASE IN THE RABBIT

Transcription

1 / $02.00/0 GASTROENTEROLOGY 75:29-33, 1978 Copyright 1978 by the American Gastroenterological Association Vol. 75, No.1 Printed in U.8A. A CELL-MEDIATED IMMUNE MODEL OF INFLAMMATORY BOWEL DISEASE IN THE RABBIT BRUCE S. RABIN, M.D., PH.D., AND SUSAN J. ROGERS, M.D. Division of Clinical Immunopathology, Department of Pathology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania A model of ulcerative colitis related to cellular immunity has previously been produced by sensitization of the colon of guinea pigs with dinitrochlorobenzene (DNCB). We have confirmed and extended these studies in rabbits. Rabbits were skin-sensitized to DNCB and challenged 10 days later with an intrarectal instillation of DNCB. The histology of the colon was examined at various time intervals. Three days after challenge the mucosa was ulcerated; there were crypt abscesses and epithelial basophilia with increased mitotic figures. The increased inflammatory infiltrate in the lamina propria contained polys as well as mononuclear cells. Submucous edema was common in this phase. This was compatible with active ulcerative colitis. These lesions changed to those consistent with the healing phase of ulcerative colitis at 2 weeks, and at 5 weeks there was still mild crypt disorganization and loss along with a paucity of goblet cells and submucous fibrosis. Rabbits that received repeated instillations of DNCB converted from a negative delayed hypersensitivity skin response to rabbit-colon extract to a positive response. Thus, a nonspecific cellular immune reaction in the colon produces histological changes compatible with ulcerative colitis and leads to the production of lymphocytes which are sensitized to colonic antigen. A number of experimental systems have been proposed to investigate the pathogenesis of inflammatory bowel disease (IBD).! One model has indicated that a cellular immune response, occurring within the wall of the colon, is capable of duplicating some of the histological changes associated with IBD.2-4 The antigen to which the immune response is directed need not be normally present in the colon. Determination of the role of the immune system in the pathogenesis of IBD is important in order to formulate a logical approach to therapy in these patients. If, in some patients, the immune system is the primary pathogenic agent, and it is directed against nonintestinal antigens, elimination of those antigens would be important. However, if the immune system has lost its regulatory capability and is manifesting an immune response against autologous antigens, regulation of the immune system would be important. 5 We have found that a chronic antigenic stimulus eliciting a cell-mediated immune response in the colon can produce changes compatible with IBD and can also lead to the production of colon-sensitized lymphocytes. Such lymphocytes may also account for some of the extraintestinal manifestation of IBD. However, as will be discussed, the actual role of lymphocytes which are Received November 1,1977. Accepted January 3,1978. Address requests for reprints to: Bruce S. Rabin, M.D., Ph.D., Department of Pathology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania This study was supported by Grant AM from the National Institutes of Health, National Institute of Arthritis, Metabolism and Digestive Diseases, United States Public Health Services. sensitized to colon antigens in the pathogenesis of IBD remains speculative. Materials and Methods Dinitrochlorobenzene (DNCB) sensitization. New Zealand white rabbits weighing approximately 3 kg were used. The fur on a small area of the flank was shaved. DNCB (Eastraan Organic Chemicals, Rochester, N. Y.) was dissolved in acetone at a concentration of 2000 mg per 0.1 ml. DNCB is a chemical hapten which binds to tissue proteins. When this occurs, T lymphocytes will produce a delayed hypersensitivity response against the hapten. To sensitize the rabbits to DNCB, 0.1 ml of the DNCB solution was placed on the rabbits' skin in a l-cm circle. The acetone was evaporated with hot air directed from a hair dryer. Mter an interval of 10 days, animals sensitized in this manner show a cutaneous delayed hypersensitivity response when 0.1 ml of acetone containing 100 mg of DNCB is placed on the skin. Both erythema and induration are present at the reaction site. Ten days after the animals had been sensitized to DNCB, 100 mg of DNCB in 0.5 ml of acetone were introduced into the lumen of the colon through a 12-cm long plastic tube attached to a I-ml syringe. The rabbits were anesthetized with pentobarbital for this procedure. Therabbits were killed 48 hrs, 2 and 5 weeks later. In some rabbits the DNCB was instilled every 10 days for a total of five times. These rabbits were killed 2 and 5 weeks after the last DNCB instillation. When the animals were killed, the tubing used to instill the DNCB into the colon was put into place. The colonic tissue at the end of the tubing was taken for histology as was colonic tissue approximately 20 cm proximal to this. Sections of duodenum, ileum, liver, and kidney were also taken for histology. The tissue was fixed in buffered formalin and stained with hematoxylin and eosin. 29

2 30 RABIN AND ROGERS Vol. 75, No.1 Skin tests. A solution oflyophilized rabbit colon extract was prepared by dissolving 100 mg of the extract in 1 ml of phosphate-buffered saline. An intracutaneous injection of 0.1 ml of this material was given into the shaved flank of the rabbit. In addition, normal rabbit serum and phosphatebuffered saline (ph 7.4) was also injected intracutaneously. A group of 12 rabbits was obtained and skin-tested with the three test materials. Induration and erythema were measured at 24 and 48 hr. These animals were then allowed to sit without any further treatment for 5 weeks. Mter this time they were re-skin tested with the same material. Other rabbits, which received five intrarectal instillations of DNCB, were skin-tested 48 hr before the animals were killed. These animals were only skin-tested once. Active immunization of rabbits. Rabbits were immunized with 10 mg of guinea pig, rabbit, or germ-free colon extract contained in 1 ml of complete Freund's adjuvant. Injections were given into the skin of the intrascapular area every 3 weeks for 4 months as has been previously described. 6 Direct immunofluorescence. Direct immunofluorescence of the colon of rabbits given repeated DNCB instillations was performed as previously described. 6 Fluorescent antisera containing antibody to rabbit IgG and to rabbit light chains were used (Cappel Laboratories, Downingtown, Pa.). Results Single intrarectal instillation of DNCB. Thirty-two rabbits which had been skin-sensitized to DNCB received one intrarectal instillation of DNCB. Seven rabbits were killed 3 days after the intrarectal instillation, 20 rabbits at 14 days, and 5 rabbits at 35 days. Active ulcerative colitis was found in each of the animals killed 3 days and 2 weeks after the intrarectal instillation of the DNCB. Active ulcerative colitis was considered to be the presence of crypt abscesses, superficial mucosal ulcerations, depletion of mucus in the cells lining the crypts, edema of both the mucosa and submucosa, and infiltration of the laminia propria with plasma cells, lymphocytes, and polymorphonuclear leukocytes including eosinophils (figs. 1 and 2). At 5 weeks changes compatible with regeneration were present in all of the rabbits killed at this time period. Regenerative changes were characterized by the absence of crypt abscesses and ulceration. However, crypt disarray, crypt loss, mucosal edema, and a mild increase in perivascular infiltrates were present. The basilar laminia propria was found to contain plasma cells, lymphocytes, and mast cells with some eosinophils present. There also appeared to be an increase in the fibroblastic component of the laminia propria and the submucosal tissue. Goblet cell mucous depletion was still present. The ileum, duodenum, proximal colon, and liver were histologically normal in all of the rabbits regardless of the interval between DNCB instillation and killing. Five animals which received DNCB into the colon without prior skin sensitization showed no histological abnormalities at 3 days. Three animals which were skin-sensitized to DNCB but did not receive DNCB in the colon were histologically normal. Three rabbits which had been skin-sensitized to DNCB and were then given only acetone into the colon had no histological abnormalities. All slides were coded and interpreted without knowledge as to the experimental group. As an indication of the uniformity of results, the animals could be assigned to their proper groups on the basis of the tissue histology. Repeated instillation of DNCB. Ten rabbits received five intrarectal instillations of DNCB at 10-day intervals. Six of these rabbits were killed 2 weeks after the last intrarectal instillation and 4 rabbits at 5 weeks. The histological changes in each of the rabbits were identical to those in rabbits receiving only one intrarectal instillation of DNCB and killed at 2 and 5 weeks. With direct immunofluorescence, 7 of these rabbits were found to have IgG bound to the colon epithelial cell membranes. The ileum and duodenum of these rabbits were found to have mild to moderate mucosal and submucosal edema with prominent lacteals. Inflammatory changes were not noted. The colon taken proximal to the area where DNCB was instilled showed edema of both the mucosa and submucosa, but no increase in inflammatory cells was present. In 8 of 10 rabbits receiving repeated instillations of DNCB, the liver showed an inflammatory ifiltrate in the portal tracts which was composed primarily of lymphocytes and plasma cells. Aggregates of mononuclear cells were found in the parenchyma of the midzone and antral regions. Changes in hepatocytes which include cell enlargement with homogenization of the cytoplasm and vacuolization were present. All of the rabbits which received repeated instillations of DNCB showed a positive delayed hypersensitivity skin test to a saline extract of rabbit colon. Animals which did not receive DNCB were skin test-negative. The histology of the delayed hypersensitivity skin reaction showed a small vessel vasculitis wit a cellular and endothelial response in the deep dermis and subcutaneous tissue (fig. 3). The major features were foci of mononuclear cells with only a few polymorphonuclear leukocytes present and lack of fibrosis or necrosis. Active immunization plus DNCB sensitization. Forty-five rabbits were immunized with saline extracts of guinea pig, rabbit, or germ-free rat colon. The associated histological changes have been previously described. 6 These rabbits were then skin-sensitized to DNCB and a single intrarectal instillation of DNCB was given 10 days after the skin sensitization occurred. The animals were killed 3 days, 2 weeks, and 5 weeks after the intrarectal instillation ofdncb. The histology of the colon at the site of DNCB instillation was the same in animals receiving DNCB without prior colon immunization and in actively immunized rabbits which received DNCB. The changes in liver histology were also similar in both groups of rabbits. Discussion The development of an experimental immune model for the study of IBD has generally been unsuccessful,i Dogs receiving passive transfer of antisera do develop a transient bloody diarrhea, but no permanent histological changes. Significantly, the dogs begin to produce

3 July 1978 IMMUNE MODEL OF INFLAMMATORY BOWEL DISEASE 31 FIG. 1. Histology of colon 48 hr after the intrarectal instillation of dinitrochlorobenzene into the colon of a dinitrochlorobenzene-sensitized rabbit. A, deep mucosal ulceration. There is a marked accumulation of inflammatory cells, predominantly polymorphonuclear leukocytes but with some lymphocytes and eosinophils present. The crypts are disrupted and destroyed. Edema and an inflammatory infiltrate are present in the submucosa (original magnification, x 40). B, surface erosin with a markedly increased inflammatory infiltrate (original magnification, x 160). C, edema of the submucosa, engorgement of vessels, and an inflammatory infiltrate present (original magnification, x 40). D, mucin depletion, crypt splaying, and crypt abscesses. The epithelium shows degeneration and flattening (original magnification, x 160). E, crypt abscess with inflammatory cells concentrated about the crypt and in the lumen of the crypt. There is cuboidilization and destruction of the crypt epithelium (original magnification, x 400). antibody to their own colon. 7 The fact that the dogs with the persistence of newly synthesized antibody to colon do not have long term disease provides an argument against the role of humoral antibody in the disease pathogenesis. Rabbits immunized with rabbit or rat intestine s,9 or bacteria obtained from patients with IBDlo may produce antibody to the intestine, but do not develop pathological changes.

4 32 RABIN AND ROGERS Vol. 75, No.1 FIG. 3. Biopsy of delayed hypersensitivity skin reaction. Rabbit received repeated intrarectal instillations of dinitrochlorobenzene and was skin-tested with a saline extract of rabbit colon. An increased inflammatory infiltrate with nests of inflammatory cells is present (original magnification, x 40). FIG. 2. Histology of colon 14 days after the intrarectal instillation of dinitrochlorobenzene into the colon of a dinitrochlorobenzenesensitized rabbit. A, decrease of mucin-secreting cells, crypt disarray, and a residual crypt abscess. The lining epithelium is disorganized, cuboidal, and flattened (original magnification, x 160). B, widely spaced crypts with abundant inflammatory infiltrate (original magnification, x 160). Cellular immunity has been shown capable ofproducing mucosal edema, mucosal hemorrhage, mucosal ulceration, and heavy mononuclear infiltration in the colon of animals undergoing a cellular immune response to DNCB This would suggest that, if lymphocytes become reactive to an antigen found in the colon, the resulting tissue damage and lymphocytic infiltration could be part of the etiology or pathogenesis of ulcerative colitis. Study of the cellular immune response to DNCB in rabbits has extended our observations on animals which have been actively immunized to colonic antigen. 6 DNCB is a chemical hapten capable of stimulating a cell-mediated immune response. During the acute phase of the cell-mediated immune response in the colon to DNCB, tissue changes occur which are compatible with those found in ulcerative colitis. After the acute cellular response is over, reparative and regenerative changes occur in the colon and, again, these changes are compatible with the reparative and regenerative changes associated with ulcerative colitis. The antigen which stimulates the cellular response must constantly be present for the perpetuation of the acute phase of the reaction. When rabbits were given a series of five intrarectal DNCB instillations, the histology of the colon was not found to differ from that of animals only given one instillation of DNCB. Thus the reaction becomes chronic only if antigen persistence is present. In animals given one DNCB instillation, the proximal colon, ileum, and duodenum were histologically normal. However, edema without inflammation was found in the duodenum, ileum, and proximal colon of those animals receiving repeated DNCB. The reason for the edema without active cellular infiltration may possibly be attributed to the production of antibody binding to intestinal antigens inducing changes in the permeability of the intestine. Indeed, direct immunofluorescence of the colon where DNCB had been applied revealed deposition of IgG on the epithelial cells in some rabbits which received five instillations of DNCB. The histology of the liver showed changes compatible with chronic active hepatitis when five DNCB instillations were given. We have previously shown 14 that actively immunized rabbits developed a chronic active hepatitis which is postulated to be on the basis of shared antigens between the liver and colon. Evidence that repeated exposure to DNCB elicits lymphocytes reactive to colonic antigen was obtained by delayed hypersensitivity skin testing. The rabbits repeatedly receiving DNCB became skin test-positive to colonic antigen but not to normal rabbit serum. Rabbits which did not receive DNCB but were maintained in our rabbit facility for the same length of time as the animals receiving DNCB were skin test-negative to this

5 July 1978 IMMUNE MODEL OF INFLAMMATORY BOWEL DISEASE 33 antigen. This suggests that a cellular immune response occurring within the colon, but not initially involving colonic antigen, may lead to the production of lymphocytes reactive to the colon. To determine whether the presence of colon-sensitized lymphocytes could perpetuate the tissue lesions initiated by DNCB, rabbits actively immunized to colon antigen were sensitized to DNCB and then had DNCB instilled into the colon. The proximal colon of these animals showed histological changes compatible with those which we have already described. 6 At the site of the DNCB instillation the tissue changes reflected those seen when DNCB was instilled into animals which were not actively immunized to colonic antigen. Thus, having sensitized lymphocytes present did not perpetuate the tissue lesions or cause them to be more severe. These studies indicate that a cellular immune response occurring within the wall of the colon of rabbits is capable of producing tissue changes compatible with ulcerative colitis. The lesions are only present as long as the exogenous antigen stimulating the immune response is present and the continued presence of the antigen leads to the production of colon-sensitized lymphocytes. There has been the suggestion that an infectious agent may be present in the intestine of patients with IBD.I5-17 This agent may be capable of eliciting a cellular immune response which would reproduce the same type of immunological activity as did the DNCB in our experimental system. In addition, the cell-mediated graft versus host reaction 2 or a cellular immune response against intestinal parasites 3 produces changes in the intestine compatible with some of the changes of ulcerative colitis and celiac disease. Thus, as long as an exogenous agent is present and a cellular immune response occurs in the wall of the colon, the lesion of IBD may persist. If this hypothesis is correct, then elimination of the offending agent should lead to regression of the tissue lesion. 18 REFERENCES l. MacPherson B, Pfeiffer CJ: Experimental colitis. Digestion 14: , Elson CO, Reilly RW, Rosenberg IH: Small intestinal injury in the graft versus host reaction: an innocent bystander phenomenon. Gastroenterology 72: , Ferguson A, Jarrett EEE: Hypersensitivity reactions in the small intestine. 1. Thymus dependent of experimental 'partial villous atrophy.' Gut 16: , Holden RJ, Ferguson A: Histopathology of cell mediated immune reaction in mouse colon-allograft rejection. Gut 17: , Allison AC, Denman AM, Barnes RD: Cooperation and controlling functions of thymus derived lymphocytes in relation to autoimmunity. Lancet 2: , Rabin BS, Rogers SJ: Nonpathogenicity of anti-intestinal antibody in the rabbit. Am J Pathol 83: , LeVeen HH, Falk G, Schatman B: Experimental ulcerative colitis produced by anti co Ion sera. Ann Surg 154: , Holborow EJ, Asherson GL, Wigley RD: Autoantibody production in rabbits. Immunology 6: , Watt J, Marcus R: Experimental ulcerative colitis of the colon in animals. Gut 14: , Asherson GL, Holborow EJ: Autoantibody production in rabbits. Immunology 10: , Bicks RO, Rosenberg EW: A chronic delayed hypersensitivity reaction in the guinea pig colon. Gastroenterology 46: , Bicks RO, Brown G, Hickey HD, et al: Further observations on a delayed hypersensitivity reaction in the guinea pig colon. Gastroenterology 48: , Bicks RO, Azar MM, Rosenberg EW, et al: Delayed hypersensitivity reactions in the intestinal tract. 1. Studies of 2,4-dinitrochlorobenzene-caused guinea pig and swine colon lesions. Gastroenterology 53: , Rabin BS, Rogers SJ: Pathologic changes in the liver and kidney produced by immunization with intestinal antigens. Am J Pathol 84: , Cave DR, Mitchell DN, Brooke BN: Evidence of an agent transmissible from ulcerative colitis tissue. Lancet 1: , Cave DR, Mitchell DN, Kane SP, et al: Further animal evidence of a transmissible agent in Crohn's disease. Lancet 2: , Cave DR, Mitchell DN, Brooke BN: Experimental animal studies of the etiology and pathogenesis of Crohn's disease. Gastroenterology 69: , Van Der Waaij D, Cohen BJ, Anver MR: Mitigation of experimental inflammatory bowel disease in guinea pigs by selective elimination of the aerobic gram-negative intestinal microflora. Gastroenterology 67: , 1974