PURPOSE:This study was undertaken to assess the electron microscopic and microbiologic findings in tissue biopsy samples from patients with pouchitis and to compare them with findings in patients with normal pouches, conventional ileostomies, and normal ileum.METHODS:Tissue samples were obtained from 78 patients: 23 patients with normal pouches endoscopically and histologically (Group 1), 12 patients with endoscopic and histologic evidence of inflammation (pouchitis) (Group 2), 14 patients who had either endoscopic or histologic evidence of inflammation but not both (Group 3), 20 patients with conventional ileostomies (Group 4), and 9 patients without ileostomies from whom biopsy samples of normal ileum were obtained (Group 5).RESULTS:The mean total aerobic facultative counts in the biopsy samples from the pouchitis patients were significantly higher when compared with biopsy samples from Groups 4 and 5 (P < 0.05). There were no significant differences in the mean anaerobic counts among the five groups. Positive cultures were obtained in 90 percent of patients with pouches compared with 69 percent of patients with conventional ileostomies or normal ileum (P < 0.05). Intramural bacteria were observed on electron microscopy in biopsy specimens of 47 percent patients with pouches compared with 14 percent of patients with conventional ileostomies or normal ileum (P < 0.05). However, the proportion of patients with positive cultures or intramural bacteria was not increased in the pouchitis group compared with the normal pouch group.CONCLUSION:These data suggest that intramural aerobic facultative bacterial counts are elevated in patients with pouchitis and may play a role in the pathogenesis of pouchitis.

Twenty-two percent of 117 biopsies of human intestinal tissues had ultrastructural images of classical regulated secretion from eosinophils in vivo i.e. eosinophil granule extrusion (EGE). Replicate intestinal biopsies that were positive for bacteria had EGE more often than not (p < 0.05); 77% of the isolates were Staphylococci. Some of the intestinal biopsies also had damaged nerves; all that had EGE and damaged enteric nerves also had positive bacterial cultures. The EGE that we observed could not account for all enteric nerve damage, suggesting multifactorial mechanisms for nerve damage in gut tissues. Among the possibilities are release of neurotoxic eosinophil granule proteins by an alternate secretory route, i.e., piecemeal degranulation, direct toxicity of tissue invasive bacteria and/or damaged nerves of unknown etiology such as those that are regularly present in uninvolved tissues of patients with Crohn's disease.

Strains of Bacteroides vulgatus from a variety of sources were tested for their abilities to enhance the inflammatory response in an experimental model for ulcerative colitis. Although there were considerable differences noted in inflammatory responses when guinea pigs were immunized with the various strains, there did not appear to be any correlation between the source of the isolates and the severity of the carrageenan-induced lesions. Strains from patients with ulcerative colitis were no more active in the model system than were strains from patients with antibiotic-associated colitis or strains from a healthy human source. The antibody titer to the strain used for immunization did not correlate with the severity of the cecal ulcerations, as determined by histopathologic evaluation.

The ability of Escherichia coli to colonize the large bowels of animals is related to many factors inherent to the intestinal environment and the bacterium. The use of germfree mice eliminates the competition between E. coli and the other microflora and allows most E. coli strains to colonize. We found that E. coli K-12 strains differing in chromosomal antibiotic resistance could monoassociate in germfree mice in large numbers. However, when two or more strains were in competition with each other, we detected quantitative differences in the abilities of the strains to colonize. The order of colonizing ability was as follows: nalidixic acid resistance greater than streptomycin resistance greater than rifampin resistance. We also found that a nalidixic acid-resistant strain bearing plasmid pBR322 colonized less efficiently and at lower levels when in competition with the nalidixic acid-resistant strain. Studies of the membrane proteins of the various strains indicated that changes in membrane proteins occurred concomitantly with altered resistance to antimicrobial agents. These results suggest that chromosomally linked alterations in antimicrobial sensitivity may also reflect changes in membrane proteins and a decreased ability to colonize mammalian intestines in otherwise isogenic bacterial strains.

Stool specimens from patients with ulcerative colitis (UC) and other inflammatory bowel diseases were analyzed for long-chain fatty acid (LCFA) content. Samples of cecal contents from guinea pigs with carrageenan induced colitis and gnotobiotic guinea pigs associated with a microflora known to cause cecal ulcerations (Bacteroides vulgatus) were also evaluated by identical techniques. Samples were extracted by the Folch procedure. The organic phase was analyzed by gas liquid chromatography (GLC) and a chemotactic assay using human polymorphonuclear leukocytes. Normal guinea pig cecal contents and stool specimens from patients without UC contained only trace amounts of LCFA and no chemotactic activity (CA). Cecal contents from carrageenan fed guinea pigs showed small amounts of LCFA and detectable CA. UC patient stool and cecal contents from gnotobiotic guinea pigs associated with B. vulgatus revealed large LCFA concentrations with peaks detected in the C18 and C16 regions by GLC. These samples were almost as active at concentrations of less than .6 µg/ml in the Chemotaxis assay as the C5a control. No KDO or heptose was detected in extracted samples and less than 1% of the material was carbohydrate or protein. A more detailed study of the C18 region revealed that C18:2 and C18:3 were present in the stool of patients with UC and in gnotobiotic guinea pigs with B. vulgatus. These unsaturated fatty acids were chemotactic when oxidized. These data indicate that LCFA may serve as active chemotactic factors and as distinct markers of UC in humans.