The disease bears the name of the investigator that convincingly distinguished this disease from intestinal tuberculosis in 1932. This distinction was not universally accepted and the notion of a mycobacterial etiology has never been fully dismissed.

Nevertheless, it was 46 years after the distinction of Crohn's disease and intestinal tuberculosis before research attempting to reassociate mycobacteria and Crohn's disease was published. Recently, there has been a surge of interest in the possible association of mycobacteria and Crohn's disease due largely to the isolation of genetically identical pathogenic M. paratuberculosis from several patients with Crohn's disease in the United States, the Netherlands, Australia, and France.

These pathogenic organisms have been isolated from only a few patients and direct evidence for their involvement in the disease process is not clear; however, M. paratuberculosis is an obligate intracellular organism and strict pathogen which strongly suggests some etiologic role.

Immunologic evidence of a mycobacterial etiology, as assessed by humoral immune determinations, have been conflicting, but evaluation of the more relevant cellular immunity have not been performed. Data from histochemical searches for mycobacteria in Crohn's disease tissues have been equally conflicting with acid-fast bacilli detected in 0 to 35% of patients. Animal model studies have demonstrated the pathogenic potential of isolates as well as elucidating the complexity of mycobacteria-intestinal interactions.

Treatment of Crohn's disease patients with anti-mycobacterial agent has not been fully assessed, although case reports suggest efficacy. The similarities in the pathology, epidemiology, and chemotherapy of Crohn's disease and the mycobacterioses are discussed. The issue is froth with controversy and the data generated on the association of mycobacteria and Crohn's disease are in their infantile stages such that a general conclusion of the legitimacy of this association cannot be made. While no firm evidence clearly implicates mycobacteria as an etiologic agent of Crohn's disease, the notion is supported by suggestive and circumstantial evidence, and a remarkable similarities to other known mycobacterial diseases.


What is Crohns Disease?

The disease that Crohn, Ginzberg, and Oppenheimer described in 1932 was a chronic low-grade inflammation of the terminal ileum (59). Earlier cases may have been documented (196, 298), but the authors failed to receive recognition for describing a new disorder. These earlier cases were called nonspecific granulomata and were sorted from those that had previously been termed hyperplastic tuberculosis of the intestine. Although Crohn's disease was first described as a segmental disease of the small intestine, in 1960, it was recognized that the same disorder affected the colon and had been confused with ulcerative colitis (161). In recent years the lesions of Crohn's disease have been recognized in the mouth, larynx, esophagus, stomach, skin, muscle, synovial tissue, and bone (17, 141, 142, 152, 171, 182, 206, 297). Thus, Crohn's disease may be considered a newly recognized disease, with a defined clinical and pathologic description dating back only to the 1960's. Although the terms Crohn's disease, Crohn's colitis, Crohn's ileitis, and regional ileitis have been with us longer, there is uncertainty as to the accuracy of these diagnoses prior to 1960. To this date, Crohn's disease and ulcerative colitis continue to be confused clinically and the term inflammatory bowel disease (IBD) was developed to comprise both diseases.

Patients afflicted with this disorder generally suffer with chronic weight loss, abdominal pain, diarrhea or constipation (obstruction), vomiting, and generalized malaise. Between 70 to 80% of Crohn's disease patients require surgical resection of the diseased intestine (83, 102, 120, 243). Difficulties usually are not ended by surgical intervention, and most patients will suffer recurrences and require further surgical procedures (71, 109, 123, 161, 181, 291). Generally, patients live with chronic pain, in and out of hospitals, throughout their lives. Mortality is approximately 6% (83, 239). Perhaps more important than mortality, is the quality of life of Crohn's disease patients. Less than 50% of patients consider their quality of life to be "good"; suboptimal psychosocial function is recorded in 30-54% of patients (83, 230, 291). These patients with inflammatory bowel disease (an estimated 2 million, 200,000 whom are children, in the United States alone), represent a very unhappy population with little prospect for relief. The etiology remains obscure and medical treatment is supportive at best (102, 238, 243). Twenty-five years after the original description, Crohn and Yarnis wrote (61): "From this small beginning we have witnessed the evolution of a Frankenstein monster that, if not threatening to life, frequently results in serious illness, often prolonged and debilitating."


Is Crohns Disease an infectious process?

In their original disease description, Crohn and co-workers (59) attempted unsuccessfully to infect guinea pigs, rabbits, and chickens with diseased intestine and lymph nodes. Van Patter, W. (Ph.D. thesis, University of Minnesota, 1952) inoculated 131 animals, including guinea pigs, rabbits, cats, rats, and chickens with diseased tissues from 43 patients; all remained normal. Mitchell and Rees (189, 190) and others (42, 43, 270) claimed to have transmitted a granuloma-inducing agent to the footpads of mice by inoculation of diseased tissue filtrates. Others have reported that they transmitted ileitis to rabbits by intraserosal inoculation of tissue filtrates (258). These studies and others could not be reproduced and did not withstand reexamination. The results could not be confirmed by others; granulomas were sometimes produced with filtrates from control tissues, and many of the granulomas were found to contain foreign material such as bone, hair, and synthetic fibers (5, 26, 240, 304). In a multicenter study in which various investigators exchanged material, a transmissible histologic abnormality could not be reproduced (304). In recent years, Das et al (69) has described the production of lymphomas or plasma cell hyperplasia in nude mice by the injection of Crohn's disease tissue. Additionally, these authors claim that sera from Crohn's patients react with certain cells or cell-types in these murine lymphomas as demonstrated by the fluorescent antibody technique (13, 299, 310). This quite fascinating, and unexplained, phenomenon is currently under investigation and still awaits judgment from other laboratories. Thus, over the years, definitive evidence of an infectious (or at least transmissible) agent has not been forthcoming.

Nevertheless, despite the lack of any clear-cut evidence of an infectious etiology, investigators continue to seek agents in diseased tissues, and every few years, a new putative agent emerges. This continued pursuit probably is related to the lack of acceptance of other theories, including a role for allergy, autoimmune disease, and dietary factors. Additionally, the pathology of the disease, the multiple familial occurrences, and the multiple remote lesion sites all suggest an infectious process. Viruses (10, 96, 104, 108, 222, 223, 293, 294, 306) and L-form bacteria (19, 139, 261, 294) have been incriminated most commonly, but in recent years, these all have been discarded as candidate agents. To this date, the etiology of Crohn's disease, being infectious, transmissible, or not, has eluded the scientific community. Now, over 55 years since its distinction from tuberculosis, attention has been driven back to the beginning: "Is Crohn's disease a mycobacterial disease after all?" (97).


Mycobacteria and Crohns Disease: a historical perspective

Medical historians suggest that Crohn's disease may first have been described as early as 1682-1771, or even earlier (143). Reports of diseases suggestive of Crohn's disease have appeared in 1806, 1813, 1828, 1875, 1907, 1908, 1909, and 1913 (143). Whether these cases actually were Crohn's disease, will remain unknown. Mycobacteria were not discovered until 1874 when Armauer Hansen described acid-fast bacilli in leprosy patients (118). The organism causing tuberculosis, which would be confused with Crohn's disease for years to come, was not discovered until 1882 (149), and intestinal tuberculosis was not recognized until several years later. Nevertheless, a disease was described in the early 1900's that was similar to intestinal tuberculosis, but acid-fast organisms could not be demonstrated. Dalziel (64) in 1913 described several patients with chronic intestinal enteritis which, although very similar to intestinal tuberculosis, was believed to be a new disorder. He drew attention to a recently described disease in cattle called pseudotuberculosis (now known as paratuberculosis) "in which the histological characters and naked-eye appearances are as similar as may be to those we have found in man". Dalziel goes on to state: "In my cases the absence of acid-fast bacilli would suggest a clear distinction, but the histological characters are so similar as to justify a proposition that the diseases may be the same". Also in 1913, contrary to the views of Dalziel, Ignard (cited by 197) wrote that "In many cases of hyperplastic tuberculosis of the intestine, no tubercles, giant cells or bacilli are found. The lesion is consisted of a mixture of variable proportions of tuberculous and inflammatory elements. In certain cases, the last only exists. Nevertheless, these inflammatory tumors should be classified among the tuberculous". The view of Ignard predominated, and these unusual intestinal diseases became known as hyperplastic tuberculosis.

By the 1920's, the belief was fading that intestinal tuberculosis occurred without acid-fast bacilli or caseous necrosis, and a disease known as "non-specific granulomata" emerged (197, 298). In these cases, as well as those described by Crohn et al (59) in later years, the authors each discussed "the remarkable resemblance" of these cases to intestinal tuberculosis. The landmark article by Crohn, Ginzberg, and Oppenheimer (59) recognized regional ileitis as a separate and unique disease entity and displaced the long-held belief of a mycobacterial etiology. We now know that hypertrophic intestinal tuberculosis (3) and tuberculosis without caseation or demonstrable acid-fast bacilli (296) do exist, as well as a distinct disease known as Crohn's disease. Nevertheless, over the years, the notion recurs that Crohn's disease might in fact be mycobacterial in origin.


The search for a mycobacterial etiology

Cultural Data
Attempts to isolate mycobacteria from Crohn's disease patients dates back before the time that this disease was recognized as a distinct entity. During these periods, as discussed, cases of intestinal tuberculosis were documented in which the tubercle bacillus could not be visualized or isolated from diseased tissues. Even after the recognition of Crohn's disease, investigators continued to seek mycobacteria in diseased specimens without success. These attempts were made almost exclusively with Lowenstein-Jensen (LJ) media and seeking Mycobacterium tuberculosis almost exclusively. As knowledge was gained about the cultivation of mycobacteria, investigators came to realize that routine methods and media were not appropriate for all mycobacteria and many species failed to grow under these conditions. Because the clinical and pathological similarities between Crohn's disease and intestinal tuberculosis continued to suggest some form of relationship, researchers adapted different bacteriological methods and immunologic tests in attempts to find an elusive Mycobacterium species responsible for Crohn's disease.

Perhaps the first concerted effort to isolate mycobacteria from Crohn's disease patients was presented by Van Patter, W. (Ph.D. thesis, University of Minnesota, 1952). In these studies, Van Patter reported the results of 1,762 cultures from 43 patients with Crohn's disease. By employing 7 different types of media and incubation periods up to 15 months for some cultures, he isolated acid-fast organisms from 3 patients (7%) after 6, 7.5, and 8 months incubation. These organisms could not be subcultured and were never formally identified.

For the next 25 years, not a single report appeared on the attempt to isolate mycobacteria from Crohn's disease patients. Although undoubtedly attempts were made over the years, such data were presented only as laboratory information related to a case report in order to rule out the possibility of intestinal tuberculosis. In 1962, Golde and McGill (101) addressed the issue of atypical mycobacteria (more appropriately called mycobacteria other than tuberculosis or MOTT) in Crohn's disease and suggested searching for these organisms rather than for M. tuberculosis alone. The ability of these organisms to produce chronic intestinal disease and the similarities of Crohn's disease to tuberculosis and Johne's disease (paratuberculosis) were also noted in their report. It was another 16 years before an original article appeared implicating mycobacteria as etiologic agents in Crohn's disease.

In 1978, a revival of the notion that mycobacteria might be related to Crohn's disease occurred with the articles by Burnham et al (33, 34). These authors described the isolation of M. kansasii from the lymph node of a single patient with Crohn's disease and pleomorphic material, suggestive of cell wall deficient (CWD) organisms, from 22 of 27 Crohn's disease patients, 7 of 13 ulcerative colitis patients, and 1 of 11 controls. It was proposed that CWD-forms of M. kansasii played an etiologic role in both Crohn's disease and ulcerative colitis, however, this theory held a short life.

The most damaging evidence to the theory of M. kansasii as an etiologic agent in Crohn's disease was the failure of Burnham et al (33, 34) to identify the CWD forms and their assumption that they were forms of M. kansasii. This organism is recognized as an opportunistic pathogen causing disease predominantly in individuals with underlying chronic disease (105, 215, 244, 284). It is not a primary pathogen in healthy individuals, and is generally nonpathogenic in animals; a few strains may be pathogenic for mice. Although the natural reservoir of M. kansasii is not known (it is not found in soil or dust), it has been isolated from a variety of water sources (122, 241) and healthy animal tissues including lymph nodes (135, 302). Thus, M. kansasii was not a good candidate for consideration as a primary pathogen. The pleomorphic organisms, however, continued to be investigated. Stanford, J. L. (Proc. 2nd Intl. Workshop on Crohn's Disease, Martinus Nijhoff Publ., 1981, pp. 274-277) cultured patient lymph nodes on many bacteriological media, in addition to LJ and Robertson's cooked meat medium, and reported the isolation of irregular acid-fast masses from 42 of 76 patients with Crohn's disease, 14 of 27 with ulcerative colitis, and 3 of 41 control lymph nodes. Although these masses resembled CWD forms, they could not be identified. In an attempt to indirectly support the notion that these masses were CWD mycobacteria, Stanford chemically induced CWD forms of M. kansasii, filtered them through 0.45 and 0.22 mcm filters, and then inoculated the filtrates onto culture media. After a period of time, abnormal acid-fast forms appeared which were visually indistinguishable from those observed in IBD tissues. Attempts to isolate classical mycobacteria from these experimentally induced acid-fast masses were unsuccessful. In an accompanying paper, White (White, S. A., Proc. 2nd Intl. Workshop on Crohn's Disease, Martinus Nijhoff Publ., 1981, pp. 278-282) presented additional data suggesting that this acid-fast material was of mycobacterial or corynebacterial origin. An examination of culture material by serology and thin-layer chromatography revealed that most material containing these coryneform bacteria or acid-fast forms reacted with highest titers to M. kansasii-related mycobacteria and Corynebacterium antisera. Additionally, White found evidence of mycolic acids in some of the acid-fast masses. From their data, Stanford and White concluded that IBD may be associated with or caused by an organism from the Mycobacterium-Corynebacterium axis.

The efforts of Stanford et al were reviewed and up dated at a recent symposium (266). Since 1974, these investigators have examined over 200 surgical specimens and have isolated pleomorphic, variable acid-fast, organisms from 42 of 76 (55%) Crohn's disease patients, 17 of 27 (52%) ulcerative colitis patients, and 3 of 41 (7%) controls. The organisms remain unidentified although efforts are in progress to classify them more precisely. Although antisera prepared against M. kansasii binds strongly to these pleomorphic organisms, mycobacterial genetic probes failed to produce restriction fragment length polymorphism patterns similar to any Mycobacterium species examined to date (McFadden, J. J., J. Thompson, E. Green, S. J. Hampson, J. Stanford, J. Haagsma, R. Chiodini, and J. Hermon-Taylor. Gastroenterol. 94:A294, 1988).

In 1984 there was again a surge of activity on the role of mycobacteria and Crohn's disease, and yet a different Mycobacterium species. From this period on, there have been more reports on mycobacteria and Crohn's disease than in the last 50 years, and as expected, the data are equally conflicting and controversial. In 1984, Chiodini et al published a series of papers describing the isolation of 2 strains of an M. paratuberculosis-like organism from 11 patients with Crohn's disease but not from 3 with ulcerative colitis or 3 with other bowel diseases (49). Detailed techniques and characteristics of the isolates (48) as well as antimicrobial susceptibility profiles (50) were reported in addition to animal susceptibility studies. The isolates were pathogenic for mice by the intravenous or intraperitoneal route, but not for chickens, guinea pigs, rats, or rabbits. Oral inoculation of one of the strains into a newborn goat produced a granulomatous ileocolitis without observable acid-fast bacilli after 5 months. Immunologic studies on this animal failed to show seroconversion, except for an early IgM response which rapidly subsided. Although the authors presented data suggesting skin test reactivity in this animal to M. paratuberculosis PPD but not M. tuberculosis PPD, the level of observed reactivity would not be considered positive in a clinical setting.

The authors concluded that their isolates were strains of M. paratuberculosis or a biovariant of that species and suggested that this organism plays an etiologic role in at least some cases of Crohn's disease. An editorial that accompanied some of these papers (97) suggested that these studies "provide the most intriguing evidence yet generated regarding a possible cause of this important illness" and that "scientists have come closer than ever to fulfilling Koch's postulates and developing a test system for Crohn's disease". On the other hand, this editorial also recognized some of the pitfalls of these studies, the need for further research, and that skepticism would and should exist.

Shortly thereafter, Chiodini et al reported that primary isolation of their putative agent occurred in a CWD form or spheroplast (55). On primary culture these organisms appeared as non-acid fast coccobacillary forms that had the ultrastructural appearance of spheroplasts, and after several months incubation transformed into characteristic M. paratuberculosis-like organisms. Employing genetic techniques, i.e., restriction polymorphism of the 5s ribosomal RNA genes, spheroplasts were found to be identical to the parent bacillary M. paratuberculosis-like forms (55), and were identified as strains of M. paratuberculosis. Additionally, these workers isolated spheroplasts, four of which transformed into M. paratuberculosis, from 16 of 26 patients with Crohn's disease (61%), but not from 13 patients with ulcerative colitis or from 13 with other bowel disorders. Although these spheroplasts remain unidentified, 7 of 10 tested seroagglutinated with specific M. paratuberculosis antisera suggesting that these unidentified forms were also M. paratuberculosis. Some of these CWD forms required up to 1-1/2 years incubation for primary emergence of colonies. The authors indicated that the presence of CWD forms could account for i) the inability to demonstrate acid-fast bacilli in patients' tissues; ii) the failure to demonstrate a strong and consistent immunologic response because CWD mycobacteria are generally of low immunogenicity; and iii) the previous failure to isolate these organisms because of the caustic nature of most other techniques for processing mycobacteriology specimens. Based on available information about mycobacterial spheroplasts, which suggests that only bacillary forms are pathogenic, they postulated that a very slow rate of reversion with subsequent local hypersensitivity-type immunologic responses could account for the chronicity of Crohn's disease.

In a series of studies the organisms were definitively identified as strains of M. paratuberculosis. By both restriction polymorphism of ribosomal 5s genes (R. J. Chiodini and T. J. Yang, Abstr. Annu. Meet. Am. Soc. Microbiol. 1986, U-20, p.122; R. J. Chiodini, Proc. 21st Joint U.S.-Japan Leprosy Tuberc. Conf. 1986, pp. 8-12; R. J. Chiodini, Proc. 22nd Joint U.S.- Japan Tuberc. Conf. 1987, pp.47-51) and studies of random gene sequences (175, 176), restriction patterns were found to be identical between the Crohn's disease isolates and M. paratuberculosis. These papers not only served to confirm the identification of these isolates, but also described the first genetic technique capable of separating the pathogenic M. paratuberculosis from its close relatives of the environmental M. avium-M. intracellulare (MAI) complex. Previously applied techniques, i.e., DNA:DNA hybridization, failed to separate this closely related group of organisms (174, 307). Although taxonomically it has been proposed to classify these related organisms into a M. avium-M. intracellulare-M. paratuberculosis complex (R. J. Chiodini, Proc. 22nd Joint U.S.-Japan Tuberc. Conf., 1987, pp. 47-51), sufficient genetic divergence is present to maintain M. paratuberculosis as a distinct species.

At a research conference on paratuberculosis in Melbourne in 1986, Coloe et al (Coloe, P. J., C. R. Wilks, D. Lightfoot, and F. A. Tosolini. Aust. Microbiol. 7:188, 1986) reported on the isolation of M. paratuberculosis from one of 30 patients with Crohn's disease. The organism was isolated from colonic material after 16 weeks incubation; cultures of the draining lymph nodes were negative. This isolate was identified by biochemical criteria and cellular fatty acid profiles. This work represented the first confirmation of the isolation of M. paratuberculosis from a Crohn's disease patient. At present, Coloe et al have cultured biopsies from approximately 50 Crohn's disease and 50 control (ulcerative colitis, normal tissue) patients. Although some acid-fast growth has been observed on some cultures from Crohn's disease patients, these slow growing isolates have not yet been characterized (P. J. Coloe, personal communication, 1988).

Whitehead examined the serologic activity of sera from the Crohn's disease patient from whom this Australian isolate was obtained, as well as a few additional patients (Whitehead, J. Proc. 2nd Intl. Coloq. Paratuberc., 1988, in press). Employing an immunoblot technique with patient sera on M. paratuberculosis antigens separated by polyacrylamide gel electrophoresis, this investigator found that identical antigen bands were recognized by the sera from the Crohn's disease patients examined and cattle naturally infected with M. paratuberculosis.

Graham and co-workers in 1987 reported their results of culture and DNA hybridization (106). They described the isolation of mycobacteria from 47.6% of 105 specimens including those from Crohn's disease, ulcerative colitis, and control patients. Mycobacteria were isolated from 9 of 59 patients with Crohn's disease, 9 of 19 with ulcerative colitis, and 18 of 27 non-IBD controls. Most isolates were of the MAI complex and M. fortuitum complex, with a single M. kansasii isolate and one they claimed to be similar to M. paratuberculosis isolated from an ulcerative colitis patient (genetic studies conducted after publication suggested that this isolate was not M. paratuberculosis). They did not find any specific association between mycobacteria and Crohn's disease and brought to light the widespread occurrence of mycobacteria in diseased tissues. Yoshimura et al (307) failed to find any association between mycobacteria and Crohn's disease by the use of DNA:DNA hybridization methods. Application of this technique to 31 biopsy specimens revealed mycobacteria-related sequences in 10 of 19 patients with Crohn's disease, 2 of 6 with ulcerative colitis, and 1 of 6 controls. Again their results did not support the notion that mycobacteria are etiologically related to Crohn's disease. The authors recognized that their findings did not rule out mycobacteria as etiologic factors in Crohn's disease, but the results may have been limited by low sensitivity of the methods employed (42% positive rate is equal to the sensitivity they achieved by cultural methods). In their article, Yoshimura et al (307) presented additional data supporting the identification of the Crohn's disease-isolated mycobacteria previously reported (48) as M. paratuberculosis.

The culture results of Graham and co-workers (106) provide some useful and important information and illustrate the ubiquitousness of some Mycobacterium spp. These workers applied tissue processing techniques of lower stringency than that recommended for the isolation of M. paratuberculosis, and these methods probably account for their results. Although the authors suggest that their method using 0.1% hexadecylpyridinium choride (HPC) is that recommended by the National Animal Disease Center (NADC), this laboratory actually recommends 0.1% benzalkonium chloride (279) or more recently, 0.75% HPC (187). At concentrations less than 0.75% HPC, contaminants, which include environmental mycobacteria, commonly overgrow cultures from clinical specimens. Since organisms of the MAI and M. fortuitum complex are widespread in the environment, processing techniques of low stringency would result in the isolation of these species from a variety of sources. The results of Graham et al (106) are comparable to those obtained from environmental sources. MAI complex can be isolated from 26-63% of soil samples, 50% of tap water samples, 13% of dust samples, and 35% of air samples (27, 95, 122, 241, 289). M. fortuitum can be found in 39% of soil samples, 63% of dust samples and 25% of air samples (137, 241, 284).

An interesting feature of the culture results of Graham et al (106) is the specimen type from which mycobacteria were isolated. Except for a single strain of M. fortuitum complex, mycobacteria could not be isolated from resected tissues; but 35 strains of mycobacteria (predominantly MAI and M. fortuitum) were isolated from biopsy specimens of aphthous ulcers. These ulcers provide a suitable micro-environment for the propagation of such environmental organisms. On the other hand, Graham et al (106) isolated as yet unidentified spheroplasts primarily from resected tissues of Crohn's disease patients rather than aphthous ulcers. Retrospectively, these culture data appear to support, rather than refute, a CWD mycobacterial etiology. Thus, as in all other diseases, the area from which material is obtained for culture is of great importance, as are the techniques applied to tissue processing.

At about the same time, Haga (115) briefly reported that he was unable to isolate mycobacteria from 17 fecal specimens, 5 bowel resection, or 9 biopsy specimens from Crohn's disease patients; although, he did report the isolation of acid-fast coccoid bodies from a Crohn's disease patient which could not be identified or subcultured. S. R. Pattyn, F. Portaels, and Y. Van Maercke presented their culture results at the meeting of the International Working Group on Mycobacterial Taxonomy (IWGMT) held in Bithoven, The Netherlands in September of 1987 (S. R. Pattyn, personal communication) and later published their data in the form of a letter (L. J. Colemont, S. R. Pattyn, P. P. Mitchielsen, J. H. Pen, P. A. Pelckmans, Y. M. Van Maercke, and F. Portaels. Lancet 1:294-295, 1988). These workers examined tissues from 32 patients with Crohn's disease and demonstrated acid fast bacilli in 11 (34%) by acid-fast staining. Cultivation attempts yielded 2 strains of M. chelonei which were said to be mycobactin-dependent; mycobacteria could not be isolated from the remaining 9 cases in which acid-fast bacilli were observed. The authors acknowledged that their processing technique, i.e., 0.15% benzalkonium chloride and 0.5% NaOH, may have been too deleterious for recovery of other acid-fast bacilli; M. paratuberculosis is known not to survive exposure to NaOH decontamination.

In 1986 and 1987 a few additional reports on mycobacteria and Crohn's disease appeared, but these were not research papers. Tytgat and Mulder (278) presented a review on the etiology of Crohn's disease and were the first to report, in other than abstract form, the isolation of M. paratuberculosis from a patient in the Netherlands. Data published in this review represented the first independent duplication of previous efforts and confirmation that M. paratuberculosis may be isolated from some cases of Crohn's disease. While all the active theories on the etiology of Crohn's disease were addressed, the authors considered that "a microbial aetiology, particularly mycobacterial, seems the most promising".

Further data from some of the previously mentioned investigators were recently presented at an international research symposium sponsored by the National Foundation for Ileitis and Colitis, Inc. late in 1987.

Haagsma et al presented data on cultivation of mycobacteria from patients with Crohn's disease (113) and the presence of M. paratuberculosis antibodies in Crohn's disease patients (114). They cultured 66 surgical specimens and isolated M. paratuberculosis from 1, M. fortuitum from 1, and acid-fast material from 2. Colonies of M. paratuberculosis emerged after 11 and 16 months incubation on Herrold's egg yolk and Ogawa media, respectively. These studies, however, had two major flaws: the processing protocol changed sometime during the study and control tissues were not cultured. Their isolates of M. paratuberculosis were found to be genetically identical to those isolated by Chiodini in the United States (56). These investigators have recently isolated an additional strain of M. paratuberculosis from a Crohn's disease patient, bringing the number of Crohn's disease-associated M. paratuberculosis isolates to 2 out of 88 specimens examined (Haagsma, J., personal communication, 1988).

Yoshimura et al (308) presented data on characterization of some of their previous isolates (106). Their isolates were predominantly of the MAI complex, probably serovar 19, but some remain unidentified by biochemical or genetic methods. Many of their isolates could not be fully characterized due to their slow growth, but of those examined, none were M. paratuberculosis.

G. Gitnick, et al (98) described their efforts to isolate mycobacteria from resected Crohn's disease tissues and inoculation of animals with their organisms. These authors cultured tissues from 27 patients with Crohn's disease, 29 with ulcerative colitis, and 26 with other bowel diseases. Three strains of mycobacteria were isolated, of which two remain uncharacterized. One isolate from a patient with Crohn's disease was identified as M. chelonei, another was said to be similar to M. paratuberculosis, and the third isolate from a cancer patient remains uncharacterized. The two isolates from Crohn's disease required 3 and 12 months incubation, respectively. Acid-fast spherules were isolated from a few Crohn's disease patients as well as controls. The M. chelonei isolate was inoculated orally into newborn goats which subsequently developed a transient diarrhea. Three animals died 5-10 days post-inoculation. Intestinal lesions were limited to mild inflammation and colonic infiltration with polymorphonuclear cells. Animals receiving the uncharacterized M. paratuberculosis-like organism remained clinically and pathologically normal. Of interest is the apparent acute diarrheal disease produced by M. chelonei since this organism is generally associated with immunocompromised hosts or with traumatic wounds (31, 110, 111, 205, 263). An acute intestinal disorder produced by M. chelonei could have significant meaning to both the veterinary and medical professions. However, the authors did not adequately rule out other neonatal diseases of goats as a possible cause of the observed diarrhea and acute bowel inflammation. The latest data from this laboratory indicates that they have isolated mycobacteria from 3 out of 27 patients with Crohn's disease, 1 out of 31 ulcerative colitis patients, and 1 out of 27 controls. Two of the 3 isolates from Crohn's disease are M. paratuberculosis (one genetically confirmed); the other is the strain of M. chelonei reported above. The strains from controls and an ulcerative colitis patient are slow-growers, as of yet unidentified, but do not appear to be M. paratuberculosis (Gitnik, G., personal communication, 1988).

Lastly, at the American Gastroenterological Association meeting in May 1988, the abstracts related to mycobacteria and Crohn's disease were only genetic studies on some of the isolates. McFadden et al (McFadden, J. J., J. Thompson, E. Green, S. J. Hampson, J. Sranford, J. Haagsma, R. Chiodini, and J. Hermon-Taylor. Gastroenterol. 94:A294, 1988) presented data showing that the M. paratuberculosis organisms isolated by Chiodini et al (48-50, 55), a M. paratuberculosis organism isolated independently in the Netherlands (113), a M. paratuberculosis strain isolated from primates (172), and wild-type M. paratuberculosis associated with disease in ruminants were all identical. Additionally, they found that spheroplasts isolated from Crohn's disease and other patients by Stanford (266) were not M. paratuberculosis but a heterologous group of organisms. Hampson et al (Hampson, S. J. J. McFadden, J. Thompson, E. Green, M. Moss, F. Portaels, and J. Hermon-Taylor. Gastroenterol. 94:A170, 1988) used DNA probes to examine mycobacteria isolated from AIDS patients, patients with atypical mycobacteriosis, and healthy individuals. They used a specific M. paratuberculosis DNA probe and were unable to identify this species in any of their material. They concluded that the absence of M. paratuberculosis in their study population confirms that this organism has only been isolated from Johne's disease and Crohn's disease, two pathologically similar disease processes.

Lastly, the Bovine Pathology Laboratory of the Lyon Veterinary School in France isolated a strain of M. paratuberculosis from a 45-year old woman with Crohn's disease. This isolate was identified by numerical taxonomic methods at the Laboratoire Central de Recherches Veterinaires. Drs. Descos and Perard of the Lyon-Suds Hospital and Lyon Veterinary School, respectively, have initiated a study to attempt isolation from fecal and biopsy specimens from approximately 50 patients with Crohn's disease.


Discussion of Cultural Data
It is now clear that a host of different mycobacteria can be isolated from Crohn's disease patients, as well as control populations, and that diseased tissue may be a suitable micro-environment for colonization of some of these species (Table 1). Most of these organisms are environmental opportunists (Table 2), although a few investigators have isolated the pathogenic M. paratuberculosis (Table 3). Unfortunately, in all studies reported to date, different methods have been used (Table 4) and as would be expected, many different results have been obtained. This is true not only for the cultural studies, but for immunological studies as well. Thus, no consistent data are available that support the role of mycobacteria in Crohn's disease.

This controversy is perhaps heightened by the evidence that M. paratuberculosis may be an etiologic agent in Crohn's disease because the organism itself is controversial. M. paratuberculosis has never been subjected to numerical taxonomic methods and has been ignored by the IWGMT. This organism is the slowest growing of the culturable mycobacteria and has a variety of sensitive growth requirements for cultivation (51-54). It generally takes years to become fully proficient at working with this species. Often even the most experienced mycobacteriologists have difficulty growing it because conventional methods are not appropriate. In addition some laboratory and other strains of M. paratuberculosis that are being used for study are actually MAI. The lack of any previous suggestion that M. paratuberculosis had public health significance has also added to its disregard by medical mycobacteriologists. These workers are now studying M. paratuberculosis but lack the background to cope successfully with the peculiarities of this species. The experience and expertise is in the hands of veterinary mycobacteriologists who generally do not have access to human tissue. It is interesting to note that all investigators who have been successful in isolating M. paratuberculosis from Crohn's disease patients were trained originally in veterinary mycobacteriology and had years of experience dealing with this peculiar species.

The microbiologic data generated on mycobacteria and Crohn's disease is rather similar to that observed in leprosy. Although leprosy is caused by M. leprae, a host of other mycobacteria have been associated with the lesions. Such mycobacteria, which are termed leprosy-associated mycobacteria (LAM) or armadillo-derived mycobacteria (ADM), can be isolated from leprosy skin lesions of humans and armadillos (72, 74, 200, 224-226). Isolation rates in armadillos average about 50% for naturally infected, experimentally infected, and non-infected animals. Organisms isolated include MAI, M. scrofulaceum, M. gordonae, M. terrae, and several groups of unclassified, difficult to grow mycobacteria (72, 226). Draper (74) suggested that infection with M. leprae favored the multiplication of environmental and other culturable mycobacteria within the lesions. These organisms are considered by most to be "insignificant" or "contaminants" (74, 226).

In addition to LAM or ADM, leprosy lesions are associated with large numbers of coryneform bacteria, termed leprosy-derived corynebacteria (LDC) (58, 136, 232, 234). Although it was once thought that these organisms were non-acid fast forms of M. leprae, it is now known that they are not related to mycobacteria and probably play a role similar to the LAM (74, 232). Unlike the LAM, LDC are primarily associated with the lesions and are rarely isolated from non-infected tissues. Some investigators believe that the LDC and LAM have a symbiotic relationship with M. leprae, while others believe these organisms represent opportunistic superinfection of the leprosy lesion. Regardless of which view is correct, it is clear that the LAM and LDC have no significance to the disease, their isolation in culture should be disregarded and they should be considered contaminants. It is also relevant to point out that even with the large numbers of M. leprae, LAM, and LDC present in leprous tissues, many cultivation attempts are negative.

A similar phenomenon occurring in Crohn's disease would clearly account for the numerous environmental mycobacteria isolated from Crohn's disease and control tissues, and also account for the coryneform bacteria isolated from Crohn's disease patients (S. White and J. Stanford (Proc. 2nd Intl. Workshop on Crohn's disease, Martinus Nijhoff Publ., 1981). Because > 106 viable bacteria are required to yield a single colony when some mycobacteria are subcultured in vitro (presumably more bacteria are needed for primary culture) (224), low numbers of a pathogenic strain or species overgrown by environmental mycobacteria and coryneform bacteria would be difficult to isolate particularly if the organism has peculiar in vitro growth requirements. In addition, organisms present in numbers < 106/gram tissue are not detectable by acid-fast staining and light microscopy. If Crohn's disease patients are infected with low numbers of M. paratuberculosis or some other Mycobacterium species and super-infected with organisms similar to the LAM and LDC, such a situation could account for the data generated.


Immunological Data
The use of immunologic responsiveness to specific antigens is a well recognized method of determining the etiology of infectious disease. Generally these determinations are based on the demonstration of rising antibody titers, but in some diseases, particularly chronic conditions, such is often not demonstrable. Diagnostic assays of chronic disease are therefore more generally based on cell-mediated immunity (CMI) or delayed-type hypersensitivity (DTH) rather than humoral responses. Despite the appropriate evaluation of cellular immunity in chronic conditions, such as Crohn's disease, most studies to date have examined humoral immunity and these have been quite limited. Except for a few scattered reports, immunologic studies related to mycobacteria and Crohn's disease have been conducted either in direct response to bacteriologic data (presented under Cultural Evidence), or involved the use of mycobacterial antigens in accessing general immunologic functions.

Morganroth and Watson (195) examined delayed cutaneous reactions and precipitating antibodies in Crohn's disease patients to antigens of atypical mycobacteria of the Runyon groups I, II, and III, as well as standard purified protein derivative (PPD). No increased incidence of sensitivity to these antigens was detected in 22 Crohn's disease patients as compared to controls. Unfortunately, the authors did not describe the species of mycobacteria examined or the nature of the antigens used. Thayer et al (272) examined skin test reactivity of Crohn's disease patients to tuberculin PPD, in addition to several other non-mycobacterial antigens, to evaluate anergy in Crohn's disease. These authors also failed to find an increased skin reaction to PPD in Crohn's disease patients and found no evidence of anergy as assessed by skin test reactivity. Bird and Britton (23) also failed to find increased responses to M. tuberculosis in Crohn's disease patients by the lymphocyte blastogenesis assay. Matthews et al (168) examined sera from 24 Crohn's disease patients in an agglutination assay with antigens from M. paratuberculosis and M. avium, in addition to antigens from non-mycobacterial microbes. As antigen, these authors used phenol-killed whole cells of the mycobacteria which displayed wide cross-reactivity. The majority of sera from Crohn's disease patients (79 to 96%) agglutinated M. paratuberculosis and M. avium cells, but such reactivity was also observed in an equal number of controls. Thus, there was no clear difference observed between Crohn's disease patients and controls.

In conjuction with their isolation of M. kansasii in culture, Burnham et al (33, 34) determined that in skin tests with antigens prepared from M. kansasii a high proportion of Crohn's disease patients showed an increased response as as compared to controls. No differences in reactivity between controls and Crohn's disease patients were noted with antigens prepared from 16 other Mycobacterium species. White et al (292) in Burnham's group also found increased reactivity of Crohn's disease patients to M. kansasii antigens. By the use of an indirect fluorescent antibody technique, positive responses were found in 9 of 11 sera from patients with Crohn's disease but not in any of 33 control sera. Based on these and their culture data, these authors suggested CWD M. kansasii as an etiologic agent in Crohn's disease. During the same year, however, Whorwell et al (294) reported their inability to demonstrate M. kansasii in tissues by immunofluorescence, and by 1980, members of Burnham's group reported that they were unable to duplicate their original immunologic findings. Although increased responsiveness to skin tests with M. kansasii antigens was stll observed in Crohn's disease patients was, they also found increased responsiveness in their control population (77). Also in 1980, Grange et al (107) reported increased IgA and IgM antibodies to M. tuberculosis in patients with Crohn's disease. They noted that responses of tuberculosis patients were predominantly of the IgG class rather than IgA and IgM as found in Crohn's disease.

In conjunction with the isolation studies reported by Chiodini et al (48,49), these authors (273) presented data suggesting increased serologic reactivity to M. paratuberculosis antigens in Crohn's disease by the ELISA technique. Patients with Crohn's disease had a statistically significant increase in antibody titer to a protoplasmic antigen of M. paratuberculosis as compared to controls. Examining cross-reactivity between antigens, these authors found 52.5% and 39% cross reactivity of their antigen with M. kansasii and M. tuberculosis, respectively. As a result of this cross-reactivity, a significant proportion of Crohn's disease patients' sera also reacted to M. kansasii antigens. These results have not been duplicated in any other laboratory.

Cho et al (57), examined seroreactivity of Crohn's disease and control patients to common mycobacterial antigens and a species-specific peptidoglycolipid of M. paratuberculosis. Increased reactivity of Crohn's disease patients in comparison to controls was not observed with either antigen. They concluded that, as in paratuberculosis, seroreactivity is not a reliable method of examining the relationship between Crohn's disease and mycobacteria. Haga et al (116) also failed to duplicate the results of Thayer et al (273) and reported that the antibody titers to M. paratuberculosis of 32 Crohn's disease patients, 37 ulcerative colitis patients, or 48 non-IBD controls did not differ in any immunoglobin class.

Cho et al (57) recognized that these inconsistencies in seroreactivity did not necessarily contradict the on-going theories, particularly when organisms related to the MAI complex were involved. The MAI group is so widespread that all individuals, healthy and diseased, are likely to be exposed to these organisms and their antigens. Seroreactivity of the general population to MAI antigens is expected, not only because of their ubiquitous nature, but also because antigens in the Order Actinomycetales are highly conserved. Common antigens, particularly those of major cellular components, exist between all Family members of the Order including Streptomycetacae, Nocardiaceae, Actinomycetaceae, Mycobacteriaceae, in addition to Corynebacteriaceae. Thus, when unpurified antigens, such as sonicated whole cells, are used a wide range of reactivity among normal and diseased populations would likely be found. Some studies have shown that 40 to 60% of the general public reacts to MAI antigens, probably related to their constant exposure to these agents in the environment (211, 302). Most of the "common" mycobacterial antigens, such as lipoarabinomannan, are cell wall components which would be lacking in a CWD form. Unless a specific antigen can be located that will not cross-react with MAI and other related organisms, a difference among populations is not likely to be noted. Although Cho's use of a species-specific antigen from M. paratuberculosis (57), was the proper approach, animals naturally infected with M. paratuberculosis do not respond to this antigen and this antigen has not been found in wild-type strains isolated from clinical cases, suggesting that this antigen does not exist or is not expressed in wild-type strains (37). Furthermore, this species-specific antigen was later found not to be species-specific but identical to that of MAI serovar 2 (38) and some recent data suggest that these laboratory strains in which this specific antigen was detected are not M. paratuberculosis but rather MAI (R. J. Chiodini, Proc. 22nd Joint U.S.-Japan Tuberc. Meet., 1987, pp. 47-51).

Jiwa et al (130) described IgG serum antibodies to mycobacterial PPD's in Crohn's disease patients. These investigators examined seroreactivity to PPD's prepared from M. tuberculosis, M. kansasii, M. phlei, M. paratuberculosis, and M. smegmatis and found that Crohn's disease patients have elevated antibody titers to all species examined. Serologic studies conducted with a crude antigen and 3 antigenic fractions of M. paratuberculosis also showed a slight, but insignificant, increased antibody titer in Crohn's disease patients as compared to controls. Such widespread reactivity to PPD, probably based on a ubiquitous cross-reactive antigen, is highly indicative of sensitization by environmental organisms gaining immune access through a defective mucosal barrier.

Kobayashi et al (147) sought antibodies to mycobacteria in Crohn's disease patients by the ELISA method using lipoarabinomannan (LAM) and a protoplasmic antigen preparation of M. paratuberculosis as antigen. These authors failed to find any significant elevation in IgA, IgG, or IgM antibody levels in Crohn's disease as compared to controls. Based on their findings these authors concluded that, since all chronic infections have an associated serologic response to the etiologic agent, their failure to find a response in Crohn's disease greatly diminishes the likelihood of mycobacteria as etiologic agents. This statement is not entirely correct. For example, patients with the tuberculoid form of leprosy, by definition fail to mount a humoral immune response (46, 119, 158, 202). Additionally, although these authors attempted to address and correct errors made in other studies, their results with control antigens did not agree with those previously reported. Crohn's disease patients have a generalized increased antibody response to enteric organisms (11, 25, 86, 154), but in the study by Kobayashi et al (147), antibody titers to lipid A were not increased. The use of LAM as a broad mycobacterial antigen may also be inappropriate because most normal individuals have demonstrable LAM titers, probably related to exposure to environmental mycobacteria. Although patients with mycobacterial diseases such as leprosy and tuberculosis generally have LAM titers higher than control groups, it is unclear if other mycobacterioses produce similar responses. Additionally, this study was the first to attempt duplication of the serologic results of Thayer et al (273) using a similar protoplasmic antigen. Some questions have been raised regarding the nature of their preparation, however, since Kobayashi et al reported that this antigen had at least 20 SDS-PAGE bands while Thayer et al reported that their antigen contains only 5 SDS-PAGE bands (W. R. Thayer, J. A. Coutu, R. J. Chiodini, and H. J. Van Kruiningen. Gastroenterol. 88:1613, 1986).

D. C. Markesich et al (167) have studies the interaction of peripheral blood monocytes with M. paratuberculosis to determine if monocytes from Crohn's disease patients react differently to mycobacteria as compared to controls. These investigators found that macrophages from Crohn's disease patients inhibited growth more efficiently than controls and that the survival of M. paratuberculosis in Crohn's disease monocytes was significantly less than in controls. The authors noted that their study was conducted with a limited number of patients and the possibility or effect of increased activated macrophages in Crohn's disease patients (180, 207) was not assessed.

Das et al (Das, P. K., J. L. G. Blaauwgeers, A. W. Slob, J. Spies, A. Chand, A. Kolk, and H. J. Houthoff. Gastroenterol. 94:A88, 1988) examined the possible relationship of mycobacteria and Crohn's disease by using immunoblot analysis and a lymphoproliferative assay. They found that sera from patients with Crohn's disease reacted with various mycobacterial and gut-associated antigens, and that many sero-reactive epitopes were shared between mycobacteria and human gut tissue. Thus they concluded that the pathogenesis of Crohn's disease could involve either cross-reactive epitopes or idiotypes, without the persistent presence of viable mycobacteria. Their lymphoproliferative assay showed that the lymphocytes from 5 out of 6 patients with Crohn's disease reacted specifically to M. paratuberculosis antigens, whereas those from controls, ulcerative colitis, and bowel cancer patients did not. Other than the brief, inconclusive, report by Thayer et al (Thayer, W.R., J. A. Coutu, R. J. Chiodini, and H. J. Van Kruiningen. Gastroenterol. 90:1662, 1986), these preliminary studies represent the only investigation of CMI responsiveness to mycobacterial antigens in Crohn's disease.

Ajitsu et al (Ajitsu, S., S. Mirabella, and H. Kawanishi. Gastroenterol. 94:A4, 1988) examined the immunologic responsiveness of murine intestinal tissues by studying the response of gut-associated lymphoid tissues to orally administered M. paratuberculosis antigens in young and old mice. They found that cells from old mice responded to M. paratuberculosis antigens to a much greater extent than those from young mice (stimulation index > 10 vs. < 3). Additionally, cells from old mice responded by producing IgG, IgM, and IgA, while cells from young mice produced only low levels of IgA. These authors concluded that the oral tolerance to M. paratuberculosis antigens in old mice is impaired, that the gut mucosal immunity in these older mice is hyper-reactive, and that these age-associated features are due in part to impaired antigen-specific T suppresser cells with over-reactive antigen specific B and T helper cells. Although not suggested by the authors, their observation that young animals respond poorly to oral M. paratuberculosis challenge, as compared to aged mice, could explain why M. paratuberculosis appears only to successfully infect only young animals. The hyper-reactive mucosal immunity in the aged could explain the age-dependent resistance to M. paratuberculosis infection (51).


Discussion of Immunologic Data
Crohn's disease patients do not have any consistent, reproducibly significant, antibody response against mycobacterial antigens. Some studies have demonstrated responses in some patients, while others have not. Patients with mycobacterioses usually have a humoral immune response, therefore the lack of such in Crohn's disease could be strong evidence against the etiologic role of mycobacteria in this disease. While some patients with pulmonary tuberculosis (45, 76, 78, 201) and, occasionally, lepromatous leprosy (46, 47, 119, 202) fail to elicit a humoral immune response, these cases are generally associated with bacterial overload and anergy. Immunologic non-responsiveness could also be caused by advancing age (235, 283), debility (78), or malnutrition (78, 158). Such is not the case in Crohn's disease. However, during certain periods in mycobacterioses, a humoral immune response is not demonstrable. Primary immunity to mycobacteria is cell-mediated. Only as the disease progresses, and there is an increasing bacterial load, does the humoral immune response become activated with the production of antibody. Patients with polar tuberculoid leprosy, in which there is a low bacterial load, fail to elicit a humoral response to M. leprae (119, 276). Most cattle with paratuberculosis have a low antibody response, but it is often not greater than that in noninfected animals (51). Cattle with overt clinical disease, unless anergic, do have a significant antibody response as compared to control cattle; variable immune responsiveness occurs in animals during subclinical disease which is perhaps more relevant to Crohn's disease. Clinical paratuberculosis (severe diarrhea and rapid weight loss) is considered the terminal stage of the disease since animals generally die within a few months after clinical onset (51). During this subclinical period, when animals appear normal but suffer subtle decreased productivity, weight loss, and increased susceptibility to other infections (51, 53, 151), immunologic responses are not readily distinguished from non-infected animals (51). The level of immunity in these animals probably is masked by cross-reactive responsiveness to environmental mycobacteria and related taxa (2, 66, 94, 121, 155, 156, 177-179, 220, 233). Only in recent years has the use of purified antigens been successful in diminishing some of the non-specific cross-reactive responses (1, 305), and shown that each animal species, i. e., cattle, sheep, and goats, responds serologically to different antigenic determinants (248, 249). In humans it may be necessary to examine tissue lymphoid cells rather than those of the peripheral blood (82, 134, 191, 192), to seek antigens which may be masked and therefore not demonstrable (252-255), to purify antigens to reduce non-specific reactions due to environmental mycobacteria (2, 66, 94, 121, 155, 156, 177, 178, 220, 233), and/or define antigenic determinants recognized by the human immune system since they may be different from those of ruminants.

At this time, insufficient information is available on the immune response of early paratuberculosis in cattle to judge whether or not a demonstrable immune response is present and how it is elicited. It is also unclear what type of immune response occurs in animals with tuberculoid-type paratuberculosis, in which acid- fast bacilli are not demonstrable but are culturable (32, 51). A tuberculoid response, i.e., DTH reaction to M. paratuberculosis antigens would be expected in these animals, but more often than not, they fail to mount any humoral, cellular, or DTH reactions (51). Likewise, no information is available on the immunologic responses in primary human intestinal tuberculosis. The only report on CMI responses in intestinal tuberculosis (22) does not define the disease as primary or secondary or whether concurrent pulmonary disease was present. Patient histories suggest that those with ulcerative or ulcerotrophic (secondary) tuberculosis generally respond to intradermal injection of tuberculin PPD, unless pulmonary disease is far advanced (anergy). This response would be expected because most patients with pulmonary tuberculosis are PPD reactive. Patients with primary intestinal tuberculosis caused by M. bovis likewise produce a positive response to PPD, but patients with hypertrophic intestinal tuberculosis are generally non-responsive to PPD if the infection is caused by M. tuberculosis, rather than M. bovis (3). Even in cases where M. tuberculosis has been isolated, the patients do not respond to skin tests. Thus, PPD reactivity is of no diagnostic value in primary hypertrophic intestinal tuberculosis if the causative agent is M. tuberculosis (3). The reason(s) for this lack of reactivity are unknown and have not been investigated. It would be important to determine both humoral and CMI responses, in subclinical tuberculoid paratuberculosis and intestinal tuberculosis, particularly the hyperplastic types. Such information on defined mycobacterial disorders would be invaluable for understanding the lack of consistent immunologic reactivity in Crohn's disease if the etiology is related to a Mycobacterium spp.

Since Crohn's disease is a granulomatous disease, and therefore assumed to be DTH mediated, it would be appropriate to study cell-mediated rather than humoral responses. If an infectious agent is present in Crohn's disease, it is present in low numbers and probably not in sufficient amount to stimulate a humoral immune response. The presence of granulomas in Crohn's disease is evidence of a functional and responsive CMI system, and such responses should be measurable. However, if the agent is similar to M. paratuberculosis, it may be difficult to sort specific responses from cross-reactive responses without using purified antigens or defined antigenic determinants recognized by the human immune system. Although an attempt to demonstrate a CMI or DTH response to mycobacteria in Crohn's disease is the obvious route of investigation, as yet, only limited inconclusive studies have been performed.


Histochemical Data
The ultimate goal in determining the etiologic relationship of an organism to a disease state is to demonstrate the association of that organism with the lesions. Routine acid-fast staining has not been successful in Crohn's disease; the absence of acid-fast bacilli was a criterion for the classification of Crohn's disease as a distinct disease entity. There have been few reports on the search for mycobacterial antigens or acid-fast bacilli in tissues from patients with Crohn's disease, but no good evidence of tissue-lesion associated mycobacteria has been found.

The first published attempt to identify mycobacteria in tissue sections from patients with Crohn's disease was the report of Whorwell et al (294) in 1978. These investigators sought evidence for the presence of M. kansasii, in addition to other pathogenic microorganisms, by immunofluorescence in tissues from patients with Crohn's disease and ulcerative colitis. No evidence of infection was found. Haga (115) also reported that he was unable to demonstrate M. paratuberculosis antigens in 18 formalin-fixed tissues from Crohn's disease patients by immunohistochemistry using anti-M. paratuberculosis antisera.

On the other hand, as discussed under cultural data, Yoshimura et al (307) employing liquid genomic DNA:DNA hybridization was able to detect mycobacteria-related sequences in 53% of Crohn's disease patients, 33% of ulcerative colitis patients, and 17% of controls. None of these sequences, however, were identical to M. paratuberculosis, and some question whether or not the methodology was sensitive and specific enough to determine if the sequences were even mycobacterial in origin.

Van Kruiningen et al (282) used peroxidase-anti-peroxidase immunohistochemistry to demonstrate mycobacteria in Crohn's disease tissues. They examined 50 formalin-fixed, paraffin-embedded, tissues from 15 patients with Crohn's disease and found positive staining in areas of submucosal inflammation in 3 patients. One of the positive specimens was considered possibly to represent phagocytized red blood cells, and another possibly to be due to cross-reactions with non-acid fast organisms around an abscess, but the third could not be accounted for by any artifact. Control, non-Crohn's disease tissues, were not examined in this study. These authors acknowledged that their method, although successful in identifying organisms in positive control specimens, did not demonstrate mycobacteria in tissues from animals with experimentally induced intestinal mycobacterial infections. By conventional acid-fast staining, these tissues were either acid-fast negative or contained very few demonstrable organisms.

Kobayashi et al (148) failed to demonstrate mycobacteria in Crohn's disease tissues by immunohistochemical methods. They examined 67 specimens (from 30 patients with Crohn's disease), either fixed in formalin, periodate-lysine-paraformaldehyde, or not fixed, and reacted with anti-M. paratuberculosis, anti-M. tuberculosis, and monoclonal anti-LAM antisera. Although staining was observed in control tissues; e.g., lymph node from an AIDS patient, pulmonary tissue from two patients with mycobacterial infection, and liver from rats infected with M. kansasii, M. fortuitum, M. paratuberculosis, or spheroplasts of MAI serovar 26; no staining was observed in any of the Crohn's disease specimens. These authors indicated that

on September 21 2002 01:00:32 ·