The following is a grant proposal that was rejected on three separate occasions by the Food Safety Program of the USDA National Research Initiative Cooperative Grants Program (NRICGP).


United States Department of Agriculture

GRANT

Food Safety Program of the USDA

National Research Initiative Cooperative Grants Program (NRICGP)

February 1994

TITLE: Prevalence of Mycobacterium paratuberculosis in Retail Dairy Products

3rd Revision

Investigators: R. J. Chiodini and J. Hermon-Taylor

Table of Contents


Summary

Mycobacterium paratuberculosis is the etiologic agent of paratuberculosis (Johne's disease), a chronic granulomatous ileocolitis of animals including primates. There is increasing evidence that M. paratuberculosis may be the cause of at least some cases of Crohn's disease in humans. It is now known that M. paratuberculosis can be isolated in culture from the tissues of a large proportion of these patients. By the polymerase chain reaction (PCR), we have identified an M. paratuberculosis - specific insertion sequence (IS900) in diseased intestinal tissues of 65% of Crohn's disease patients, 4.3% of ulcerative colitis patients, and 12.5% of controls (cancer). Such a distribution (65% in Crohn's disease) suggests a specific association with that disease and a previously unrecognized environmental distribution in the human population (12.5% in controls). Since M. paratuberculosis may be shed in the milk of infected animals, we postulated that milk might be the environmental source. In a preliminary study conducted in London, we detected IS900 in 20 of 295 (6.8%) samples of retail pasteurized milk samples purchased from local grocery stores. Prevalence follows periodicity whereby the organism is undetectable during summer months, but occurs with high prevalence (25%) during the winter months. We have also isolated viable organisms in culture from some of these samples providing evidence that we are detecting viable organisms rather than only DNA. We have also determined that M. paratuberculosis may survive routine pasteurization methods; perhaps because these are environmental like organisms, more robust than obligate pathogens like M. bovis.

We propose to reproduce and expand on these studies to examine commercial retail dairy products in the United States for the presence of M. paratuberculosis by PCR and cultivation. A 400 base pair product of the 5' region of IS900, for detection of M. paratuberculosis , will be used. Sensitive cultivation methods, together with IS900 PCR and RFLP analysis of successful culture isolates, will be performed concurrently to determine the viability of M. paratuberculosis in dairy products and precisely characterize their subtypes (biovariants).

The results of this study will define the environmental distribution of M. paratuberculosis in retail dairy products in the United States and identify human sources of this organism. The presence of a known animal enteric pathogen, with increasing evidence of involvement in human disease, in retail dairy products has serious public health implications and needs to be addressed. Preliminary findings, if corroborated by these proposed studies, suggest that the introduction of new public health procedures and other measures for the prevention of human disease may be appropriate.

It is imperative in the interest of public health, food safety, and the overall welfare of the US dairy cattle industry, to determine if a similar distribution of M. paratuberculosis exists within retail dairy products in the United States.


Project Description

Address to Critique

In this third revision, we have attempted to address previous criticisms and have made major reductions in the size, scope, and duration of the project. Specifically, we have eliminated examination of dairy products other than milk, despite the comments by previous reviewers that were "impressed with the varied sampling projections". In addition, we have shortened the project duration from 3 years to 2 years. Although one of the reviewers suggested the project be reduced to a single year, the periodicity we have previously determined preclude a study limited to a single year. We have also incorporated into the study, as controls, both M. bovis and M. avium. We would expect that IS902 (M. avium) would be found in most samples (M. avium exists in >50% of all soil and water samples) while M. bovis should be negative in all samples. Furthermore, we have added increased attention to side-by-side comparisons between raw and pasteurized milk (mapping) in all determinations. We believe that these changes greatly improve the quality of the proposal and significantly add to our likelihood of success and meaningfulness of the data generated.

We would also like to point out that previous reviewers have been largely favorable, offering little on which to build a better proposal. The initial proposal received a medium priority largely because of our pasteurization methods, which were never intended to be a major part of the proposal, and since we have "no understanding for the methods of sampling" and our "confusion regarding the meaning of statistical values", presumably because we failed to provide statistical analysis of sample size. Most comments arose from the panel summary, without any explanation, and not from the ad hoc reviewers. The last revision was again criticized for the pasteurization, which had been essentially eliminated from the proposal. The reviewers comments were in line with a review of an appendixed manuscript rather than a review of the proposal. The proposal was given medium priority, because, without explanation, the panel does not believe this is a food safety problem and feel there is no epidemiological evidence to support a link between Johne's and Crohn's disease. We wish to address the panel(s):

It must be recognized that the results of our pasteurization study, which are not a part of this proposal, are irrelevant to the preliminary studies, objectives, goals, and likelihood of success of this proposal. If we assume that all the data generated on pasteurization is meaningless and/or in error due to "methodological shortcomings", such would not affect this proposal. The fact that we have, as clearly explained in preliminary studies, been able to detect and isolate M. paratuberculosis from retail pasteurized milk clearly indicates either survival or widespread contamination. Therefore, the attention given to our previous efforts on pasteurization, regardless of any possible "methodological shortcomings" is unjustified and does not detract from the present proposal.

The suggestion from the last panel that there is "no evidence presented to show that Johne's disease of cattle is ... associated with Crohn's disease in humans" is difficult to understand. There is mounting evidence that M. paratuberculosis is associated with human Crohn's disease and this has (had) been thoroughly reviewed within the introduction as well as an appendixed manuscript. There is little dispute that the association is not firmly established, but sufficient data exists to justify concern. And this Association is receiving more and more attention as reproducible data is being generated from around the world. While the panel believes this is not a food safety problem, many others believe it is, including most ad hoc reviewers. The fact that the Center for Disease Control's Emerging Pathogens Branch has recently placed Crohn's disease on their list of investigative diseases would suggest more than "no association". The interest expressed by Agriculture Canada Food Safety Risk Assessment Unit and the Ministry of Agriculture of the U.K., as well as the National and International Meetings the PI's are requested to speak at, do not reflect a situation in which there is no association.

Despite all the data presented in the introduction and appendixed manuscript, including epidemiology, many veterinarians believe that an association cannot exist because all farmers and veterinarians would have Crohn's disease if this were the case. This is a very simplistic approach! All individuals that drank raw milk did not develop M. bovis infection. Mexican farmers, whom I deal with to obtain M. bovis infected tissues, do not all have tuberculosis although their cattle are heavily infected. If milk is the source of human contact with M. paratuberculosis, as we propose, all epidemiological correlation's between the farm, cattle, and Crohn's disease is destroyed. It is our hypothesis that a susceptible newborn comes into contact with M. paratuberculosis during early infancy via commercial dairy products. The infection of an unnatural host causes the formation of low-proliferating and antigen-poor cell wall deficient forms of the organism. The slow growth and low-level persistence acts as a continuous "insult" to the intestinal lamina propria, thereby setting the stage for a chronic low-grade inflammatory reaction. Over the years, the inflammation increases, without an increase in infection, and the disease progresses to cause the clinical and pathological feature we know today as Crohn's disease. We request that the panel reassess their resolve that "no association" exists.

Lastly, we wish to address our need to obtain or otherwise consult with a food (dairy) microbiologist. We recognize this shortcoming and believe that the proposal would benefit by such an association, particularly if we choose to expand from our PCR and culture methods. A dairy microbiologist would be invaluable to us in terms of procedural matters. Although we recognize the advantages of this collaboration and the added strength such would bring to the proposal, we were unsuccessful in recruiting a collaborator. We contacted many dairy microbiologists in the United States, several whom acknowledged having reviewed this proposal for NRICGP, but all declined to collaborate in any official capacity. Although they all felt the project was important and agreed to provide support and advise, such would only be provided "off the books". The reasons for this refusal were universal - all had special interests (at least one of which was funding) with the dairy industry and felt their "interests" could be jeopardized by collaboration. Therefore, we have been unable to recruit a formal collaborator on this revised proposal. Nevertheless, we recognize the potentially important contribution these individuals would provide and we will continue to seek a qualified individual for collaboration. If one cannot to located, we will accept the "off the books" collaborative arrangements that have been offered to us.

The hesitation expressed by those dairy microbiologists, and perhaps by previous panels, illustrates a lack of understanding of the purpose and motives actual behind this proposal. The PI (RJC) has many contacts with the dairy industry (the actual farmer, not the "industry") and has a high regard for the efforts, profession, and overall well-being. We have requested funding for this proposal for 3-years now, not to destroy the dairy industry, but to help them. The Ministry of Agriculture of the U.K. funded our London study based on our very early results (shown as Figure 1 in the appendix) because they felt the study should be done properly and with sufficient funds to insure a reliable outcome. This data, as presented within this proposal, is planned for submission for publication in December, 1994 (we are delaying publication for various reasons). It is the PI's personal opinion that the release of this data could have, if "found" by the news, a profound effect on the dairy industry.

Despite the study being conducted in Europe, many questions may be asked of the US industry, and until those answers are known, the industry may suffer. We feel the best way to protect the industry is to find the answers and solutions before the questions are asked. We have already proven our research objectives and hypotheses - that M. paratuberculosis is distributed to the population via the milk supply (albeit European milk)- this proposal is designed to answer potentially damaging questions which may arise in the future and hurt the dairy farmer. With this understanding said, we request the panel to look at this proposal in a different light, one that is "pro" and supportive of the industry rather than one that is potentially damaging and controversial.

It is our sincere hope that this revised proposal has adequately addressed previous concerns, is stronger and more attractive that previous versions, and that the panel views this proposal in a different light and recommends support with high priority.

Introduction

Mycobacterium paratuberculosis is the etiologic agent of ruminant paratuberculosis, clinically and commonly referred to as Johne's disease. This enteric pathogen, M. paratuberculosis, is shed in the feces and body fluids of infected animals. The disease is characterized by a chronic granulomatous inflammation of the small and large bowel and regional lymph nodes (1). Infected animals suffer with diarrhea, weight loss, debilitation, and death. The disease most commonly affects ruminants, but under select circumstances, may also infect swine, horses, chickens, and primates (1). While in recent years paratuberculosis has been recognized as a problem in free-ranging wildlife and in zoological collections, its major impact is in domestic livestock, primarily the dairy cattle industry. In the United States, the national disease prevalence has been suggested to be 2.6% of the dairy and 0.8% of the beef cattle (2), although regional surveys suggest local prevalences ranging from 11-17% (3-5). All indications suggest that the prevalence of bovine paratuberculosis is increasing worldwide.

Crohn's disease is a chronic granulomatous ileocolitis of humans, of unknown etiology, which generally affects patients during the prime of life (teens to early 20's) (6). Patients 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. Difficulties usually are not ended by surgical intervention, and most patients will suffer recurrences and require further surgical procedures (6). Medical treatment is supportive at best and patients afflicted with this disorder generally live with chronic pain, in and out of hospitals, throughout their lives. The disease bears the name of the investigator that convincingly distinguished this disease from intestinal tuberculosis in 1932 (7). Twenty-five years after the original description, Crohn and Yarnis wrote (8): "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."

The connection between these two similar diseases can best be appreciated from a historical perspective.

In the late 1890's, an intestinal disease was recognized in humans that was similar to intestinal tuberculosis, but acid fast bacilli could not be visualized in tissues or isolated in culture and the characteristic caseous nodules of tuberculosis were absent. By 1913, cases of intestinal tuberculosis which did not fit the classical pattern were well recognized, but nevertheless, were classified among the tuberculous lesions and the disease was known as hyperplastic tuberculosis (9). By the 1920's, these cases of intestinal tuberculosis which did not contain caseous necrosis or acid fast bacilli were reclassified as nonspecific granulomata (9,10) in an attempt to recognize these cases as distinct from the classical or hypertrophic forms of intestinal tuberculosis. Each description, however, discussed the remarkable resemblance of these cases to intestinal tuberculosis - the only difference being the absence of the pathognomonic lesions of tuberculosis, caseous necrosis and acid fast bacilli. In 1932, the landmark article of Crohn, Ginsberg, and Oppenheimer (7) brought recognition of a disease to be known as regional ileitis as an distinct entity and separated these cases from intestinal tuberculosis. Crohn and his colleagues also recognized the "remarkable resemblance" of their regional ileitis and intestinal tuberculosis. For over 50 years, regional ileitis or Crohn's disease as it is now known, has been regarded as a granulomatous ileocolitis of unknown etiology, distinct from that of intestinal tuberculosis and other mycobacterioses.

As early as 1826, a chronic enteritis was reported in cattle that could not be associated with any currently known cause of diarrhea (1). The characteristic gross thickening and corrugation of the intestinal mucosa in these animals was recognized, but it was not until 1895 that the etiologic agent was discovered to be an acid fast bacillus (11). At that time, the disease was called pseudotuberculous enteritis, and was thought to be caused by the avian tubercle bacillus ( Mycobacterium avium ) since the morphologic characters of the organism resembled that species more than Mycobacterium tuberculosis. The name pseudotuberculosis was given to this disease in recognition of its remarkable similarity to intestinal tuberculosis, but without caseous necrosis. However, unlike M. avium or M. tuberculosis, this organism resisted cultivation. It was fifteen years later, in 1910, that Twort (12) succeeded in isolating the causative agent of this disorder. The isolation was largely the result of a laboratory mistake and a perceptive eye for fine detail - Twort noticed small satellite colonies arising around colonies of the common hay bacillus (Mycobacterium phlei) in culture tubes which had inadvertently not been cleaned for several months.

It is now known that this organism requires an exogenous growth factor, mycobactin, extracted from other species of the genus and requires at least 3-4 months for laboratory growth (1). The organism was named, at the time, Mycobacterium enteriditis chronicae pseudtuberculosae bovis Johne (13). Between the period of its discovery in 1895 and the isolation of its etiologic agent in 1910, pseudotuberculous enteritis was reported from around the world (1). Over the years, the disease became known as paratuberculosis, commonly as Johne's disease, and in 1932, the organism was officially given the name M. paratuberculosis, although synonyms such as M. johnei were used for years after. Since the discovery of paratuberculosis as a disease, little progress has been made in its diagnosis, control, or treatment. For almost 100 years now, paratuberculosis has been spreading through the world's ruminant domestic livestock and wildlife - it remains difficult to diagnose or control, and no animal has ever been cured (1).

The first suggested connection between Johne's disease and Crohn's disease occurred in 1913 when Dalziel (14) described several cases of enteritis in human which, although resembling intestinal tuberculosis, he believed represented a new disorder. Recognizing the incredible similarity between his cases in humans and those of pseudotuberculosis recently described in cattle, he proposed "that the diseases may be the same". He also recognized a major distinction between these two diseases and expressed a view, still largely held to this date, that "In my cases the absence of acid fast bacilli would suggest a clear distinction...". He felt, however, that the histologic lesions were so similar that it justified his proposition even in the absence of demonstrable acid fast bacilli. Dalziel's views were never given much merit and the classification of these human cases as hyperplastic tuberculosis as proposed by Ingram (15) prevailed.

In 1984, we proposed that Mycobacterium paratuberculosis might be the cause of at least some cases of Crohn's disease. This hypothesis originated from the isolation of M. paratuberculosis from 3 patients with Crohn's disease but not from controls (16,17). These original studies suggesting an etiologic role of M. paratuberculosis - like organisms in Crohn's disease were expanded, and it was determined that primary isolation of this organism was in the form of a spheroplast or cell wall deficient form (18). These spheroplast-like agents were isolated from 16 of 26 patients with Crohn's disease (61%), but not from 13 patients with ulcerative colitis or from 13 patients with other bowel disorders (18). Of these 16 Crohn's disease - associated spheroplasts, 4 had transformed into classical bacillary forms resembling M. paratuberculosis. One of these cultures required 5 years for the emergence of visible colonies. Using a variety of molecular techniques, including DNA:DNA hybridization, 5s ribosomal DNA analysis (18,2D), and restriction fragment length polymorphism (22), these organisms were definitively identified as M. paratuberculosis . Recently, the remaining 12 unidentified spheroplasts from Crohn's disease patients and several from controls were examined by PCR for the presence of IS900 (a species-specific insertion sequence). An additional 6 cultures from Crohn's disease patients were identified as containing M. paratuberculosis, bringing the total number isolates to 10 out of 26 or 38% of all Crohn's disease cultures(23). Since our original isolations, other laboratories have also isolated this organism from the diseased tissues of patients with Crohn's disease including investigators in California, Texas, France, Australia, England, the Netherlands, and the Czech Republic (6).

Although the status of investigations on the role of mycobacteria and Crohn's disease has recently been reviewed in detail (6), there have been many achievements in the last few years. These advances have largely been made possible by the discovery of IS900, a species specific insertion sequence of M. paratuberculosis (24), and the application of the polymerase chain reaction (PCR) and other molecular approaches. Readers are referred elsewhere for detail discussions of previous findings (6;appendix).

Many investigators (6;appendix) have been successful in isolating unidentified acid-fast organisms from tissues of both Crohn's disease patients and controls. The slow growing mature of these organisms, however, precluded their identification by conventional methods. Because their identity was unknown, even if mycobacterial in origin, the widespread distribution of mycobacteria in the environment preclude any suggestions related to etiology or pathogenic potential. The discover of IS900 (24) as a species-specific marker of M. paratuberculosis allowed for the first time the ability to specifically identify this species.

Combined with the amplification of PCR, a powerful tool emerged. Examination of unidentified cultures from Crohn's disease and controls revealed IS900 in 38% (10 of 26) of Crohn's disease cultures in one study (23) and in 33% (5 of 15) in another (25,26), thereby suggesting that about 1/3 of cultures from Crohn's disease patients have detectable levels of M. paratuberculosis (27). Based on the hybridization signals, it was estimated that these cultures contained between 3 and 30 individual mycobacterial cells (genomes) per culture. Such results suggest that these strains either fail to grow in vitro or grow at an extremely slow rate. Using a more cross-reactive probe against the mycobacterial 65 Kd heat shock protein, mycobacterial sequences of unknown species were detected in Crohn's disease as well as control cultures (23), again illustrating the widespread nature of mycobacteria in general and the need to define your objectives and organisms at the species level.

Thus, of the 3 laboratories actively engaged in the isolation of M. paratuberculosis from the diseased intestines of patients with Crohn's disease, isolation rates are currently 20%, 33%, and 38% from Crohn's disease and only 0.8% (1 out of 121) controls (27).

Another application of molecular biology has been the direct detection of mycobacterial DNA in tissue specimens. Earlier studies using DNA:DNA hybridization techniques to search for mycobacteria in Crohn's disease patients suggested the presence of mycobacterial DNA in 53% (10 of 19) of Crohn's disease patients, 33% (2 of 6) of ulcerative colitis patients and 17% (1 of 6) of normal controls (20). These results were compatible to studies on the environmental distribution of mycobacteria. Without the ability to differentiate species, these findings had little etiologic meaning.

With IS900, the ability to specifically search for this species was possible but efforts to detect M. paratuberculosis in tissue specimens from patients with Crohn's disease by Southern blotting were unsuccessful (28). The advent of PCR technology, however, provided the sensitivity needed to detect an organism that was below the detection level of acid-fast light microscopy, immunohistochemistry, or standard DNA techniques.

Application of PCR and IS900 to surgical resected specimens revealed the presence of M. paratuberculosis in 65% (26 of 40) of Crohn's disease patients, 4.3% (1 of 23) of ulcerative colitis patients, and 12.5% (5 of 40) of non-IBD controls (Figure 1; 29). Based on the known sensitivity of the PCR assay and assuming 100% efficiency in DNA extraction, we estimate that the bacillary load in many of these specimens may be as low as 300-500 bacilli per gram tissue. Positives in Crohn's disease were from small bowel and colon, whereas positives from the few controls were colon only.

FIGURE 1. Autoradiographs showing detection of M. paratuberculosis DNA in 4 of 6 Crohn's disease tissue DNA extracts. Samples submitted to 33 cycles PCR amplification in triplicate with one buffer-only PCR control (o). 10fg of M. paratuberculosis DNA was amplified as a positive control (Mp). PCR products were run on 2% agarose gel, southern blotted onto a nylon membrane, and hybridized at high stringency with 32 P labeled pcr400 product. Exposed for 12 hours on Kodak X-omat film. CD26, CD27, and CD29 show triplicate positive signals. CD30 shows a single positive signal suggesting a lower abundance of M. paratuberculosis genomes. The additional higher molecular weight band in CD26 is seen with some M. paratuberculosis isolates.

These results have several important implications. The finding of M. paratuberculosis in control colon indicates a previously unrecognized environmental distribution of this organism, suggesting that environmental contact with M. paratuberculosis may be more common than previously suspected. The low presence of M. paratuberculosis in ulcerative colitis patients suggest that opportunistic colonization of the disrupted or abnormal colonic mucosa by M. paratuberculosis does not occur. Therefore, the high prevalence in Crohn's disease patients would suggest an etiologic role for this known chronic enteric pathogen in humans as in many other animal species.

Our findings have been confirmed by Dr. Patrick Berche of the Hopital Necker-Enfants Malades in Paris, France. Examining pediatric cases of Crohn's disease, M. paratuberculosis (IS900) was found in 75% of pediatric patients but in less than 20% of controls (30).

These various studies, and others, conclusively establish that M. paratuberculosis occurs within the diseased intestinal tissues of a high proportion of Crohn's disease patients.

Preliminary studies

We have attempted to address criticisms that the high prevalence of M. paratuberculosis (IS900) found within Crohn's disease patients may be related to the high sensitivity of the methods used and that similar prevalences would likely be found with other pathogens. To address this issue, we examined DNA extracts from our original study (29) with IS902 (M. avium)-PCR and found IS902 in approximately 12% of all patient groups - a finding consistent with an environmental opportunistic pathogen. Likewise, examination of these samples with a probe directed against the mycobacterial 32 kDa antigen revealed no distribution or prevalence differences between Crohn's disease and controls. Similar findings were found using mycobacterial 65 kDa hsp on culture material from Crohn's disease patients and controls (23), whereby all patient groups had similar prevalences and distribution. Thus, we believe that the high frequency of IS900 in Crohn's disease is not related to assay sensitivity, as other organisms are found in equal frequency in all patient groups, and that these additional findings further support a specific role for M. paratuberculosis in Crohn's disease.

We have also been successful in obtaining expression and purification of recombinant p43 (the antigen produced by IS900). We used synthetic peptides of p43 to examine serologic reactivity of Crohn's disease patients and controls to M. paratuberculosis (IS900). The positive strand of p43 was mapped and 80 overlapping biotynolated 15 residue synthetic peptides were prepared spanning the entire p43 sequence. Each peptide was bound 7.5 u g/ml to duplicate streptavidin-coated wells in microtiter plates with streptavidin-only and conjugate-only controls. Human sera, diluted 1/40 with PBS containing protease inhibitors were added to the wells and incubated for 1 hr. After washing, plates were reacted with rabbit anti-human Ig followed by ABTS substrate. Plates were read on an automated plate reader and results recorded by a computer program which expressed results as the duplicate mean optical density for immunorecognition of each peptide by Crohn's disease sera against the mean OD + 2 SD for each peptide using normal control sera. By this method we were able to demonstrate the binding of Crohn's disease sera to specific sequence determinants within p43. Sera from 8 of 12 (75%) of normal subjects showed weak immunorecognition to one or more of 4 p43 domains: peptides 22-24, 34-37, 40-43, and 64-67. On the other hand, sera from all 9 (100%) Crohn's disease patients so far examined demonstrated significant immunorecognition (>mean + 2SD) to one or more of 4 different domains: peptides 28-31, 54-58, 71-73, and 74-80. All Crohn's disease patients, but no controls, recognized and significantly reacted (>3 SD above mean) with the major carboxyterminal epitope of p43 (peptides 74-80). IS900 has proved uniquely specific for M. paratuberculosis and these findings, although limited and preliminary, support the concept that most of the human population has been exposed to M. paratuberculosis , but some members of the population react to specific epitopes on p43, perhaps the carboxyterminal epitope, that results in an abnormal hyper-reactivity within the intestinal mucosa. These are the first studies to suggest a switch in antigen recognition by Crohn's disease patients which may account for "susceptibility" in some members of the population.

Of particular relevance to the proposed studies, based on the finding of M. paratuberculosis by IS900 PCR in 65-78% of Crohn's disease patients and in 12-20% of controls, we sought to examine potential sources of this organism. Several pieces of data provided an indication of what this source(s) might be:

  1. Restriction fragment length polymorphisms using IS900 have established that there are at least 2 distinguishable variants of M. paratuberculosis , a bovine strain and an ovine-caprine strain (31,32). Applying these techniques to the human isolates of M. paratuberculosis identifies their source as bovine.
  2. In the United States, the national disease prevalence of bovine paratuberculosis has been estimated to be 2.6% of the dairy and 0.8% of the beef cattle (1,2), although regional surveys suggest local prevalences in dairy cattle ranging from 11-17% (1,3-5). As such, M. paratuberculosis is prevalent among the US dairy population.
  3. Patients with Crohn's disease and controls used in our studies did not appear to have any direct contact with domestic cattle. Therefore, exposure to a bovine pathogen must be indirect, possibly through bovine products.
  4. Cattle infected with M. paratuberculosis shed organisms in feces often at concentrations approaching 10/ 8 colony forming units (cfu) per gram (1). In addition to feces, other body fluids may also shed M. paratuberculosis , including the mammary gland. Some studies have suggested that as many as 35% of clinically infected cattle have detectable quantities of M. paratuberculosis in their milk using only conventional culture (33). As many as 11.6% of asymptomatic carriers may likewise shed detectable quantities of M. paratuberculosis in milk (34).
  5. Studies on survivability of M. paratuberculosis during both the Standard Holder (bulk) and the Higher Heat Shorter Time (HHST) methods indicated that M. paratuberculosis, particularly those strains isolated from humans, may survive both methods of pasteurization (35, Appendix).
  6. In addition to the resistance of M. paratuberculosis to complete killing by standard pasteurization methods, pasteurized milk may also become contaminated with M. paratuberculosis or other pathogens by contamination of pasteurized milk with raw milk (36) or with excrement from the environment. The presence of coliforms in pasteurized bottled milk suggests that post-pasteurization contamination occurs frequently (37).
  7. Since M. paratuberculosis is not generally found in the environment, but appears at least in humans to have an environmental-like distribution, a possible public source could be the milk supply. The commercial distribution of milk., i.e., regional, could coincide with the childhood space clustering observed in Crohn's disease (38).
  8. The incidence of Crohn's disease is much higher in non-breast fed children suggesting that either breast milk provides some form of protection or non-breast fed children are exposed to a causative factor with increased frequency (39,40).

With these facts in mind, we initiated a study in London to examine retail milk samples from throughout Southern England for the presence of M. paratuberculosis. As reported in the a previous revision, applying PCR methods described, IS900 was detected in 12 of 40 (30%) samples obtained from retail milk (Figure 2). In addition to the demonstration of M. paratuberculosis by PCR, several of the PCR positive samples yielded viable organisms on cultivation on Herrold's egg yolk medium.

FIGURE 2. Autoradiographs showing detection of M. paratuberculosis DNA in a sample of bottled pasteurized whole milk. Phases (cream, whey, pellet) were created by centrifugation at 41,000 x g for 60 minutes at 1oC. DNA was extracted and samples submitted to 33 cycles PCR amplification in triplicate with one buffer-only PCR control and triplicate phase separation process controls. 10fg of M. paratuberculosis DNA was amplified as a positive control (M. paratuberculosis). PCR products were run on 2% agarose gel, southern blotted onto a nylon membrane, and hybridized at high stringency with 32 P labeled pcr400 product. Exposed for 12 hours on Kodak X-omat film. Positive signals were obtained in triplicate samples of the cream and cell pellet. This milk sample yielded viable M. paratuberculosis in culture.

Since our original application, we have expanded those studies, and to date have processed 295 bovine retail milk samples. Overall prevalence, based on total samples, is now estimated to be 6.8% of retail milk. Several interesting observations, however, were noted. Our initial success was followed by a long (5-month) period of negative testing (Figure 3). During this negative period, we exhaustively evaluated our methodology, and although previously tested samples (frozen) were still positive, we continued to fail in detecting IS900 in any of the 85 samples tested during that period. As noted in Figure 3, we are currently in a period of high detection with 25% of milk samples being positive. We have not been able to determine the reason(s) for this periodicity, although as noted in Figure 3, it appears to follow a predictable pattern with high detection during Aug-Sep and Jan-Feb. We are anxious to see if this high prevalence period is again followed by a 5-month lag.

FIGURE 3. Detection of M. paratuberculosis in retail bovine milk by IS900 PCR during an 18-month period. Detection appears to have a distinct periodicity during the winter months with prolonged intervals of negative testing. Total units of milk tested: 295; Number test positive: 20; Percent positive: 6.8% 2 3

The detection of organisms in the pellet and lipid (cream) phases suggest a bacillary form (solid and hydrophobic) rather than hydrophilic free DNA which would reside in the water (whey) phase. Early spiking experiments confirmed this: DNA resided in the whey while whole bacilli were found in the cream and pellet. We have further confirmed this suggestion (i.e., bacillary form) by the use of immuno-capture on magnetic beads and the ability to grow viable M. paratuberculosis organisms from some of the milk samples.

These preliminary findings strongly suggest that pasteurized milk, and possibly other dairy products, may be contaminated with viable M. paratuberculosis and provide a source of human contact with this organism. In the interest of public health and food safety, as well as the dairy industry, these studies need to be expanded and environmental sources of M. paratuberculosis (a well recognized enteric pathogen with increasing evidence of involvement in a human disease) in the United States be determined.

To determine potential sources of human contact with M. paratuberculosis, we propose:

  1. To define the prevalence of the specific intestinal pathogen, Mycobacterium paratuberculosis, in retail dairy products using a highly specific polymerase chain (PCR).
  2. To determine the viability of M. paratuberculosis in PCR positive samples by in vitro cultivation
  3. To characterize the subtypes Df M. paratuberculosis isolates by restriction fragment length polymorphism (RFLP).

Rationale and Significance

Mycobacterium paratuberculosis is the etiologic agent of paratuberculosis (Johne's disease), a chronic granulomatous ileocolitis of animals including primates. There is increasing evidence that M. paratuberculosis may be the cause of at least some cases of Crohn's disease in humans. It is now known that M. paratuberculosis can be isolated in culture from the tissues of a large proportion of these patients. By the polymerase chain reaction (PCR), we have identified an M. paratuberculosis - specific insertion sequence (IS900) in diseased intestinal tissues of 65% of Crohn's disease patients , 4.3% of ulcerative colitis patients, and 12.5% of controls (cancer). Most recently, IS900 has been detected in 75% of pediatric cases of Crohn's disease. Such a distribution (65-75% in Crohn's disease) suggests a specific association with that disease and a previously unrecognized environmental distribution in the human population (12.5% in controls). Since M paratuberculosis may be shed in the milk of infected animals, we postulated that milk might be the environmental source. In a study conducted in London, we detected IS900 in 20 of 256 (6.s%) samples of retail whole pasteurized milk. Prevalence appeared to have a periodic pattern with some months having detection prevalences of up to 25%. Viable organisms were also isolated from some of these samples. We have also determined that M. paratuberculosis may survive routine pasteurization methods. A reason for this may be because these are environmental like organisms, more robust than obligate pathogens like M. bovis.

We propose to reproduce and expand on these studies conducted in London to examine commercial retail dairy products in the United States for the presence of M. paratuberculosis by PCR and cultivation. A 400 base pair product of the 5' region of IS900 probed with a 228bp internal sequence will be used for detection of M. paratuberculosis. Sensitive cultivation methods, together with IS900 PCR and RFLP analysis of successful culture isolates, will be performed concurrently to determine the viability of M. paratuberculosis in dairy products and precisely characterize their subtypes (biovariants). Insertion elements from M. bovis (IS1081) and M. avium (IS901) will be used as controls.

The results of this study will define the environmental distribution of M. paratuberculosis in retail dairy products in the United States and identify human sources of this organism. The presence of a known animal enteric pathogen, with increasing evidence of involvement in human disease, in retail dairy products has serious public health implications and needs to be addressed. Preliminary findings, if corroborated by these proposed studies, suggest that the introduction of new public health procedures and other measures for the prevention of human disease may be appropriate.

Experimental Plan

1. Number and Samples. All products to be examined will be purchased from retail grocery stores. Although our original proposal planned to evaluate a variety of milk products, we have elected to limit this revision to the examination of only dairy milk. Milk testing will be continued at a rate of 30 per month throughout the calendar year so that any seasonal effect will be identified (as with clinical paratuberculosis in cattle, we have observed a seasonal prevalence of M. paratuberculosis in milk).

In addition, as suggested by a previous reviewer, raw milk will also be tested as available throughout the study period. An effort will be made to obtain raw milk directly from local processing plants, with a corresponding sample of bottled milk, for side-by-side testing. These will be obtained at a rate of 10 side-by-side samples per month. Attempts will be made, through collaborations, to increase the number of agreeable plants and side-by-side samples.

Statistical Considerations. Statistical analysis of sample size is difficult in view of our preliminary studies which suggest periodicity to the prevalence of M. paratuberculosis in dairy products. However, applying binomial distribution analysis to total sample size, at a 95% confidence level, suggests the ability to detect 0.3% total prevalence in our proposed dairy products. Of milk samples, there exists a 95% confidence level of detecting a 0.8% prevalence in regional samples. However, since there appears to be monthly periodicity, on an individual monthly basis, at 30 samples per month, a prevalence of 9.5% is needed to obtain a 95% probability of detecting at least 1 positive sample. While higher prevalences are needed when viewed on a monthly (periodic) basis, it is important to recognize that sample size cannot be increased within the scope of this study and that the preliminary data (see Figure 3) was derived with a confidence level of only 63% suggesting true prevalence is higher. It is emphasized that total sample size provides us a 95% confidence level of detecting a 0.3% prevalence.

2. Processing of Samples. All samples will be processed under aseptic conditions in a biological safety cabinet separate from other procedures in this proposal.

The outer surfaces of retail milk containers will be thoroughly washed, and then rinsed with phenolic-disinfectant followed by 10% hypochorite solution. A 30 ml sample of milk will be obtained from the middle of the container by syringe aspiration. A 13.5 ml aliquot will then be transferred to a centrifuge tube and centrifuged at 41,000 x g for 60 minutes. This results in 3 phases: a sediment (pellet), cream (top), and whey (middle). These phases are easily separated after stabilization by cooling for 30 minutes at 1oC. A portion of each phase will be processed by PCR and culture as described below.

3. Pentane Extraction. Mycobacteria, due to their rich lipid content, are attracted to organic solvents and extraction of material with pentane has been successfully used to concentrate M. paratuberculosis from large volumes of organic matter (1). Samples to be extracted with pentane will be homogenized (if necessary) in a grinder and then 1/4 volume of pentane added to the mixture. The pentane mixture will be vortexed and phases allowed the form (pentane & solid/aqueous) by incubation at room temperature for 30 minutes. The pentane layer is then removed and the pentane evaporated by incubation at 37oC under a stream of nitrogen. The resulting pellet will be resuspended in PBS-Tween (0.05%) and processed for culture and DNA extraction.

4. DNA extraction. To a 500 µl sample of liquid, 50 µl of chilled lysis buffer (1% TritonX100, 10mM EDTA, 3mM dithiothreitol) will be added and mixed to disrupt solid material. The mixture will then be centrifuged at 1500 X g to pellet large debris and the supernatant removed. The supernatant will be subjected to enzymatic DNA extraction with subtilisin (10mg/ml) for 12 hours at 37oC, lysozyme (50mg/ml) for 6 hours at 50oC, and pronase (3mg/ml) for 48 hours at 37oC. DNA will then be purified by phenol-chloroform extractions, ethanol precipitation, and recovered by centrifugation at 12,000 X g for 15 minutes. The DNA will be suspended in 100 µl TE (10mM Tris/HCl, 0.1mM EDTA) for PCR. A control buffer-only tube will be included in every homogenization/DNA extraction.

Centrifugation and pentane pellets will be suspended in 100 µl PBS-Tween 80 in a 0.5 ml screwcap tube and boiled for 10 minutes. After boiling, the sample will be added to another tube containing glass beads (0.1mm) and the tube shaken vigorously for 3 minutes using a mini-bead beater. The base of the centrifuge tube will be pierced with a fine needle and the lysate recovered by centrifugation into a 1.5 ml tube. This lysate will be used directly in PCR reactions. All samples will be run in triplicate with appropriate controls as described below.

5. PCR Reaction. Oligonucleotide primers will be selected to amplify a 400 base pair fragment (pcr400) from the 5' region of IS900 (nucleotides 22-421). This reaction is uniquely specific for M. paratuberculosis . The primer sequences will be 5'-GTTCGGGGCCGTCGCTTAGG-3' (primer 90) and 5'GAGGTCGATCGCCCACGTGA-3' (primer 91). PCR reaction mixtures will consist of PCR buffer (67mM Tris-HCL, 16.6mM NH4SO4, 3.3mM MgCl2, 1.7mg/ml bovine serum albumin in TE [pH 8.8), 10mM B -mercaptoethanol, 200µm deoxynucleotide triphosphates, 6ng/µl primers, 5 units Taq DNA polymerase, and TE buffer (10mM Tris/HCl, 0.1mM EDTA, pH 8.8). Five µl of the final DNA extract will be added to each PCR reaction pre-mix for a total volume of 50 µl. PCR reactions will be performed for 33 cycles consisting of denaturation at 93oC for 1 minute, annealing at 58oC for 1 minute, and extension at 72oC for 3 minutes. Each sample will be run in triplicate with a buffer-only PCR negative control. Each experiment will include one positive control containing 10fg of M. paratuberculosis DNA. Similar methodology will be applied to detect the insertion elements IS902 (M. avium) and IS1081 ( M. bovis ) [41,42]. These additional reactions and analyses are performed as additional controls and to monitor for specificity of the reactions.

We have recently developed a hybridization capture technique for IS900 (and other) PCR reactions. The overall sensitivity of the PCR is greatly reduced due to the vast excess of non-target DNA and inhibitory substances in the sample. In attempts to circumvent these limitations, pPN14 plasmids containing IS900 were digested with SmaI/SacI to remove the optimal M. paratuberculosis -specific PCR 413 target site from the 5'region of IS900 and then religated. Primers p8 (5'-GCGCTCGAGTAGCCGCGTTC-3') and p21 (5'-GCGCTCGATAGCCGCGTTC-3') were used to amplify 5'-biotynolated 513bp capture probes from the 3' end of IS900. In the test, 50 µl of buffer containing excess of 5'-b capture probe is added to 450 ml of sample DNA extract, boiled for 5 min, incubated overnight at 65oC, cooled, and 25 µl of a streptavidin-coated M280 Dynadeads added. After 2 hrs, the beads are captured magnetically and the supernatant discarded. IS900-PCR is then performed directly on the beads. In initial tests in which feces from 10 bovines suspected of being infected and shedding small numbers of M. paratuberculosis were examined, all samples were negative by standard IS900-PCR. The samples were then "captured" as described and 6 of the 10 samples were positive. Although we describe the standard PCR methodology in this proposal, we will likely use this capture method if further analysis and comparisons suggest superior sensitivity.

6. Southern Blot Hybridization. PCR products will be run on 2% agarose gels and blotted onto nylon membranes by the methods of Southern (43) as previously described (21). Membranes will be hybridized overnight with a 32 P-labeled 228 bp probe internal to the 400 bp amplification product and then washed at high stringency. The membranes will then be exposed to X-ray film for 4-60 hours (depending on intensity as measured with a geiger counter) at -80oC. A simplified dot blot system validated in London may also be adapted thereby permitting a larger number of samples to be routinely tested.

7. Controls and Contamination Precautions. Stringent precautions will be taken throughout all steps to avoid false positives due to laboratory contamination. These will include the use of widely separated laboratories for each individual procedure, UV irradiation of buffers and surfaces, and meticulous cleaning of equipment and surfaces with 10% hypochlorite, 1M HCl, and 100% ethanol. PCR procedures will be performed in a separate dedicated facility. Capped tubes will be opened only once for 5 µl sample addition. Each assay run in triplicate will contain negative and positive controls which will undergo all procedures described above. Negative controls will include buffer samples that will be subjected to the complete cycle of sample preparation and also PCR reagents without template processed in a similar fashion. Any experiment in which a negative control is positive will be discarded.

In addition to negative controls, we will incorporate a positive control in each sample to monitor for PCR inhibition. Modified templates for IS900 have been synthesized by site-directed mutagenesis and will be used as an added control (23). These templates were created by first performing a standard PCR reaction on IS900 to create a "wild-type" PCR product. This product was then used as a template in two PCR reactions using IS900-1 primers + IS900 MT-2 primer, and IS900-2 primers + IS900 MT-1 primer. The product of these reactions were mixed and used in a PCR reaction with IS900 PCR primers 1 and 2. The resulting product was electrophoresed through 1.5% agarose gel, excised and purified, the ends filled with Klenow fragment, and then cloned into the SmaI site of PUC18. The modified template contains a 2 base substitution in the HaeIII site at 385bp. DNA extracted from this clone will be used as a modified template to spike PCR reactions at a concentration no greater than 0.1 fg (approximately 25 targets).

In order to analyze these samples, prior to electrophoresis, PCR products are digested with HaeIII. Analysis of these IS900 products produces bands at 217bp modified product and at 172bp for the IS900 "wild-type". Negative samples would contain only the 217bp band. Inhibition of the assay would be indicated by the absence of any bands. At least one PCR with modified template will be performed with each sample.

Thus, all samples will be subjected to 4 PCR reactions, one of which will contain modified template. In the event inhibition is suggested in the modified PCR, the original samples will be further purified with guanidium isothiocyanate and DNA purified on silica particles as previously described (21,44), before repeating the PCR.

8. Analysis of Data. Samples containing 400bp PCR fragments of IS900, no bands in negative controls, and 217bp and 172bp products in the modified sample will be considered positive for M. paratuberculosis. Samples containing only the spiked 172bp modified product will be considered negative for M. paratuberculosis . Samples containing no bands will be considered to contain PCR inhibitor(s) and either be re-run or discarded if the problem cannot be corrected.

We do expect some conflict in the methods of testing, i.e., IS900 PCR vs culture. Since PCR, by the methods employed, is more sensitive than cultivation and is not dependent upon cellular condition or viability, we do expect some IS900 PCR positive samples that are culture negative. In these instances, samples will be considered positive for M. paratuberculosis , but of unknown viability. We do not foresee this as a problem or hindrance to the project since, as previously stated, culture is used only to complement IS900 PCR and illustrate that we are detecting organisms rather than just DNA, as we have already been successful in demonstrating.

9. Cultivation methods. Although our primary objective is to determine viability and prevalence of M. paratuberculosis , cultivation methods will be expanded to include other Mycobacterium species. Cultivation of M. paratuberculosis will be performed by methods previously described (1,16-18,21,45-47) and will include both direct cultivation of the sample and of the pentane extracts. Organisms will be decontaminated with 0.25% and 0.75% hexadecylpyridinium chloride (0.75% for M. paratuberculosis and 0.25% for other mycobacteria) and cultured on Herrold's egg yolk medium supplemented with 2mg/L mycobactin, Ogawa medium, Lowenstein Jensen, ATS, and 7H10. Cultures on 7H10 will be washed in PBS prior to inoculation. M. paratuberculosis will be identified by the presence of IS900, M. avium by the presence of IS902, M. bovis by the presence of IS1081, and other species by standard methods (48).

In addition to standard cultivation methods described above, milk samples will be digested with lipase to allow the milk to be passed through a 0.45mcm filter. Approximately 100 ml of milk is proposed to be filtered, the membrane filter removed, washed with PBS containing 0.5% Tween 80, and the wash cultivated on media as described above. Employing all these cultivation methods, ie., standard culture, pentane extraction, and membrane filtration, will insure the recovery of M. paratuberculosis if it is present in viable form.

10. RFLP analysis. All M. paratuberculosis isolates will be subjected to RFLP analysis. Mycobacterial DNA will be obtained by fractionation of M. paratuberculosis in a Hughes press at -80oC and 15-25,000 psi, chloroform:isoamyl alcohol (24:1) extraction, and precipitation with ethanol as previously described (21). Restriction enzymes will be obtained from Bethesda Research Laboratories (BRL), Gaithersburg, MD and include: Ava I, BamH I, Bcl I , Bgl II, BseE II , EcoR I, EcoR II, Hinc II, Hind III, Pst I, Pvu II , Sal I, Sst I, and Xho I. Restriction enzymes underlined have previously been shown to produce polymorphisms within the IS900 sequence (31,32). All restrictions, electrophoresis, Southern blots, hybridizations, and autoradiography will be performed by routine methods as previously described (21). The IS900 400bp sequence, radiolabelled with 32 P will be used as a probe.

11. Evaluation of alternate pasteurization methods. It is important to the overall objectives of this proposal to not only identify a potential health problem in dairy products, but to also investigate potential solutions. Preliminary studies have determined that standard methods of pasteurization may not be effective in killing M. paratuberculosis, particularly those strains of isolated from humans (35, Appendix). However, these studies also suggest that effective killing should be achievable with minor changes in temperature and/or time since these strains are not totally resistant to current methods. Details regarding methodology were deleted at the advise of the Program Director since it caused the major criticisms and was never meant to represent a major part of this proposal. Therefore, as an adjunct, but not part of this proposal, we will attempt to determine the minimum pasteurization requirements to effectively kill M. paratuberculosis .

Although we have not been successful in recruiting a dairy (food) microbiologist for reasons discussed in the address to critique, we recognize the importance of such collaboration to insure proper procedural matters. Therefore, we will not proceed with any experiments beyond the PCR and cultural methods, particularly a re-evaluation of pasteurization methods, without first obtaining collaborative arrangements with a food microbiologist familiar with the procedural matters of these determinations.

12. Data gathering on milk products. As requested by one of the reviewers, we will collect as much data as possible from each dairy product, including pH, specific gravity, and other measurements deemed appropriate. We will also maintain frozen samples of all products should the need arise for further examination or analysis.

The entire product label, especially the UPC symbol if available, will be collected from each product to provide us the ability to track each product if deemed appropriate. With this data, it may be possible to determine the time and place of production as well as possibly the source of the dairy product. i.e., farm or processing plant. For instance, during peak prevalence, it may be important to know if positive samples were from a single or multiple processing plants.

We will also make contact with local and national weather bureaus to keep track of weather and monitor local dairy practices. Based on preliminary data, which suggests high prevalence during the winter months, would suggest that peak prevalence may coincide with winter housing and calving periods - peak periods of increased clinical disease in cattle. By careful monitoring of these parameters, it may be possible to correlate the presence of M. paratuberculosis in milk to seasonal dairy practices.

13. Time course. We anticipate requiring 4-months for the post-doctoral fellow to familiarize him/herself with the methodologies and to optimize processing of sample material. After that period, we expect, based on previous experiences, the ability to process 10 samples per week. This weekly load would include analysis of samples, cultivation, and RFLP, as well as analysis of data. Under such a time frame, we anticipate processing 320 samples the first year and 480 the second year, with the remainder of the time devoted to data analysis and RFLP analyses.


Facilities and equipment

The mycobacteriology laboratory is a limited-access, self-contained, laboratory with all instrumentation required for the proposed studies, either within the unit or available to the proposal, including, but not limited to, freezers, scintillation counter, microtiter readers, standard, high-speed and ultra-centrifuges, Hughes press, electrophoresis apparatuses and power supplies, incubators, MASH II, vacuum apparatuses, chromatography columns, flow cytometry (BD 440), as well as standard laboratory apparatuses. The laboratory has been rated and approved as meeting the requirements of a Biohazard Level III Laboratory.

All bacterial manipulations and tissue culture experiments will be performed in a Class II laminar flow hood in a limited access laboratory. All guidelines set forth in "CDC/NIH Biosafety in Microbiological and Biomedical Laboratories" and "Laboratory Safety: Principles and Practices" will be followed. All procedures involving radioisotopes will be performed in accordance with the "Radiation Safety Policy and Procedure Manual".


Collaborative arrangements

There are no formal consultants or collaborators other than the principal investigators. As discussed in the introduction, we recognize the need to consult and establish a collaboration with a food microbiologist and will make such arrangements prior to any award if this proposal is funded.

One of the previous reviewers expressed concern about the geographic separation of the Principal Investigators and the need of the "proposed research associate to conduct the research between two laboratories, one in Rhode Island and one in England". While we can empathize with this concern, we do not feel it is a deficiency in the proposal, but rather, a strength. We propose, as defined in the budget justification, to send the proposed research associate to London for a period of 1-2 weeks to have "hands-on" experience in PCR analysis of milk. As one can appreciate, it required a great deal of effort to work out the details of these methodologies and "hands-on" is the best way to learn the techniques. Since the laboratory of JHT is actively analyzing milk samples, such would not only provide the best training possible, but save time in laboratory set-up in the United States. After this 1-2 week internship, the proposed Research Associate will no longer need to work between these 2 countries directly. In todays world of telephones, FAX machines, E-mail, and Federal Express, the geographic separation should not be a deficiency, but a strength in terms of bringing together the best investigators to accomplish the objectives of the proposal. The Principal Investigators have previously demonstrated our ability to "break-down" geographic barriers and produce productive collaborative efforts (19,22-24,35, 49-51).


References to Project Description

1. Chiodini, R. J., H. J. Van Kruiningen, and R. S. Merkal. 1984. Ruminant paratuberculosis (Johne's disease): The current status and future prospects. Cornell Vet. 74:218-262.

2. Merkal, R.S., D. L. Whipple, J. M. Sacks, and G. R. Snyder. 1987. Prevalence of Mycobacterium paratuberculosis in ileocecal lymph nodes of cattle culled in the United States. J. Am. Vet. Med. Assoc. 190:676-680.

3. Arnoldi, J.M., S. S. Hurley, and S. Lesar. 1983. Johne's disease in Wisconsin cattle - A survey of cull cows. Proc. First Intl. Colloq. Paratb., Ames, Iowa; pp. 16-21.

4. Chiodini, R.J. and H. J. Van Kruiningen. 1986. The prevalence of paratuberculosis in culled New England Cattle. Cornell Vet. 76:91-104.

5. McPherron T. A., C. L. Coglianese, E. H. Jaster, J. W. Urbance, R. D. Mock, C. R. Oetzel, and D. F. Parrett. 1983. Observations on the economic impact of Johne's disease (paratuberculosis) on cattle herds in Illinois. Proc. Intl. Colloq. Res. Paratb., Ames, IA, June 16-19, pp. 42-51.

6. Chiodini, R. J. 1989. Crohn's disease and the Mycobacterioses: a review and comparison of two disease entities. Clin. Microbiol. Rev. 2:90-117.

7. Crohn, B., L. Ginzburg, and G. Oppenheimer. 1932. Regional ileitis, a pathological and clinical entity. JAMA 99:1323-1329.

8. Crohn, B. B., and H. Yarnis. 1957. Regional enteritis. 2nd ed. Grune & Stratton, New York.

9. Wilensky, A., and E. Moschcowitz. 1927. Nonspecific granulomata of the small intestine. Am. J. Med. Sci. 173:374-380.

10. Moschcowitz, E, and A. Wilensky. 1923. Nonspecific granulomata of the intestine. Am. J. Med. Sci. 166:48-66.

11. Johne, H. A. and L. Frothingham. 1895. Ein eigenthuemlicher Fall von tuberkulose beim Rind. Dtsch. Ztschr. Tiermed. Path. 21:438-454.

12. Twort, F. W. 1911. A method for isolating and growing the lepra bacillus of man and the bacillus of Johne's disease in cattle (preliminary note). Vet. J. 67:118-120.

13. Twort, F. W., and G. L. Y. Ingram. 1912. A method for isolating and cultivating the Mycobacterium enteritidis chronicae pseudotuberculosae bovis , Johne and some experiments on the preparation of a diagnostic vaccine for pseudotuberculous enteritis of bovines. Proc. Roy. Soc. Ser. B84:517-542 [also in Vet. J. 68:353-365].

14. Dalziel, T.K. 1913. Chronic interstitial enteritis. Br. Med. J. 2:1068-1070.

15. Kirsner, J. B. 1982. The "idiopathic" inflammatory bowel diseases. Their cause and pathogenesis. Arch. Dermatol. 118:280-282.

16. Chiodini, R. J., H. J. Van Kruiningen, R. S. Merkal, W. R. Thayer, and J. A. Coutu. 1984. Characteristics of an unclassified Mycobacterium species isolated from patients with Crohn's disease. J. Clin. Microbiol. 20:966-971.

17. Chiodini, R. J., H. J. Van Kruiningen, W. R. Thayer, R. S. Merkal, and J. A. Coutu. 1984. The possible role of mycobacteria in inflammatory bowel disease. I. An unclassified Mycobacterium species isolated from patients with Crohn's disease. Dig. Dis. Sci. 29:1073-1079.

18. Chiodini, R. J., H. J. Van Kruiningen, W. R. Thayer, and J. A. Coutu. 1986. The spheroplastic phase of mycobacteria isolated from patients with Crohn's disease. J. Clin. Microbiol. 24:357-363.

19. McFadden, J.J., Butcher, P.D., Chiodini, R.J., and Herman-Taylor, J. 1986. Determination of genome size and DNA homology between an unclassified Mycobacterium species isolated from patients with Crohn's disease and other mycobacteria. J. Gen. Microbiol. 133:211-214.

20. Yoshimura, H. H., D. Y. Graham, M. K. Estes, R. S. Merkal. 1987. Investigation of association of mycobacteria with inflammatory bowel disease by nucleic acid hybridization. J. Clin. Microbiol. 25:45-51

21. Chiodini, R.J. 1990. Characterization of Mycobacterium paratuberculosis and organisms of the Mycobacterium avium complex by restriction fragment length polymorphism of the rRNA gene region. J. Clin. Microbiol. 28:489-494

22. McFadden JJ, Butcher PD, Chiodini RJ, and Hermon-Taylor, J. 1987. Crohn's disease-isolated mycobacteria are identical to Mycobacterium paratuberculosis, as determined by DNA probes that distinguish between mycobacterial species. J. Clin. Microbiol. 25:796-801.

23. Wall, S., Kunze Z. M., Saboor, S., Seechurn P., Chiodini, R., and McFadden, J. J. 1993. Identification of spheroplastÄlike agents isolated from tissues of patients with Crohn's disease Crohn's disease and control tissues by polymerase chain reaction. J. Clin. Microbiol. 31:1241-1245.

24. Green, E. P., M. L. V. Tizard, M. T. Moss, J. Thompson, D. J. Winterbourne, J. J. McFadden, and J. Hermon-Taylor. 1989. Sequence and characteristics of IS900, an insertion element identified in a human Crohn's disease isolate of Mycobacterium paratuberculosis . Nuc. Acid Res. 17:9063-9073
.
25. Moss, M. T., J. D. Sanderson, M. L. V. Tizard, J. Hermon-Taylor, F. El-Zaatari, D. Markesich, and D. Y. Graham. 1992. Identification of M. paratuberculosis by PCR in long term cultures of Crohn's disease tissue. In: Chiodini RJ, Kreeger JM, eds. Proceedings of the Third Internationall Colloquium on Paratuberculosis. Providence: International Association for Paratuberculosis, Inc., Publ, pp. 208-213.

26. Moss MT, Sanderson JD, Tizard MLV, Hermon-Taylor J, el Zaatari FAK, Markesich DC, Graham DY. 1992. Polymerase chain reaction detection of Mycobacterium paratuberculosis and Mycobacterium avium ssp silvaticum in long term cultures from Crohn's disease and control tissues. Gut 33:1209-1213.

27. Chiodini RJ. 1992. Historical overview and current approaches in determining a mycobacterial aetiology of Crohn's disease. In: Mulder CJJ, Tytgat GNJ, eds. Is Crohn's disease a mycobacterial disease? Dordrecht: Kluwer Academic Publishers; pp. 1-15.

28. Butcher, P. D., J. J. McFadden, and J. Hermon-Taylor. 1991. Investigation of mycobacteria in Crohn's disease tissue by Southern blotting and DNA hybridization with cloned mycobacterial genomic DNA probes from a Crohn's disease isolated mycobacteria. Gut 29:1222-1228.

29. Sanderson JD, Moss MT, Tizard ML, Hermon-Taylor J. 1992. Mycobacterium paratuberculosis DNA in Crohn's disease tissue. Gut 33:890-896.

30. Dellisola B, Poyart C, Goulet O, Ricour C, Valcke JC, Berche P. 1992. Is Crohn's disease an infection. Detection of Mycobacterium paratuberculosis in Crohn's disease by means of gene amplification (PCR). Preliminary results. Rev Med Interne 13(7):s401.

31. Collins, D. M., D. M. Gabric, and G. W. de Lisle. 1990. Identification of two groups of Mycobacterium paratuberculosis strains by restriction endonuclease analysis and DNA hybridization. J. Clin. Microbiol. 28:1591-1596.

32. Whipple, D., P. Kapke, and C. Vary. 1990. Identification of restriction fragment length polymorphisms in DNA from Mycobacterium paratuberculosis . J. Clin. Microbiol. 28:2561-2564.

33. Taylor, A.K., C. R. Wilks, and D. S. McQueen. 1981. Isolation of Mycobacterium paratuberculosis from the milk of a cow with Johne's disease. Vet. Rec. 109:532-533.

34. Sweeney RW, Whitlock RH, Rosenberger AE. 1992. Mycobacterium paratuberculosis cultured from milk and supramammary lymph nodes of infected asymptomatic cows. J Clin Microbiol 30:166-171.

35. Chiodini, R. J. and J. Hermon-Taylor. 1993. The effect of pasteurization on the viability of Mycobacterioum paratuberculosis . J. Vet. Lab. Diag. 5:629-631 (Appendix) .

36. Greenwood, M. H., W. L. Hooper, and J. C. Rodhouse. 1990. The source of Yersinia spp. in pasteurized milk: an investigation at a dairy. Epidemiol. Infect. 104:351-360.

37. Greenwood, M. H., P. Gill, E. F. Coetzee, B. M. Ford, W. L. Hooper, S. C. Matthews, S. Patrick. 1988. An appraisal of methods used in the examination of retail samples of cows milk. Epidemiol. Inect. 100:369-378.

38. Ekbom, A., A. Hans-Olov, C. G. Helmick, A. Jonzon, and M. M. Zack. 1990. Perinatal risk factors for inflammatory bowel disesae: a case-control study. Am. J. Edpidemiol. 132:1111-1119.

39. Bergstrand, O., and G. Hellers. 1983. Breast-feeding during infancy in patients who later develop Crohn's disease. Scand. J. Gastroenterol. 18:903-906.

40. Koletzko, S., P. Sherman, M. Corey, A. Griffiths, and C. Smith. 1989. Role of infant feeding practices in development of Crohn's disease in childhood. Br. Med. J. 298:1617-1618.

41. Collins DM, Stephens DM. 1991. Identification of an insertion sequence, IS1081, in Mycobacterium bovis . FEMS Microbiol Lett 67:11-5.

42. Moss MT, Malik ZP, Tizard ML, Green EP, Sanderson JD, Hermon Taylor J. 1992. IS902, an insertion element of the chronic-enteritis-causing Mycobacterium avium subsp. silvaticum . J Gen Microbiol. 138:139-45.

43. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Molec. Biol. 98:503-517.

44. Boom, R., C. J. A. Sol, M. M. M. Salimans, C. L. Jansen, P. M. E. Wertheim-van Dillen, and J. van der Noordaa. 1990. Rapid and simple method for purification of nucleic acids. J. Clin. Microbiol. 28:495-503.

45. Chiodini, R.J. 1986. Biochemical characteristics of various strains of Mycobacterium paratuberculosis . Am. J. Vet. Res. 47:1442-1445.

46. Chiodini, R. J. 1990. Bacteriocidal activity of various antimicrobial agents on human and animal isolates of Mycobacterium paratuberculosis . Antimicrob. Agents Chemother. 34:366-367.

47. Chiodini, R.J. 1991. Antimicrobial synergism of rifabutin in combination with other antimicrobial agents against strains of Mycobacterium paratuberculosis . J. Antimicrob. Chemother. 27:171-176.

48. Vestal, A.L. 1981. Procedures for the isolation and identification of mycobacteria. US Public Health Service Publ No. 81-8230.

49. McFadden, J.J., Butcher, P.D., Thompson, J., Chiodini, R., and Hermon-Taylor, J. 1987. The use of DNA probes identifying restriction-fragment-length polymorphisms to examine the Mycobacterium avium complex. J. Molec. Microbiol. 1:283-291.

50. McFadden, J.J., Butcher, P.D., Herman-Taylor, J., and Chiodini, R.J. 1986. Cloning of DNA extracted from unclassified Mycobacterium isolated from Crohn's disease patient. Gastroenterol. 90:1543.

51. McFadden, J.J., Thompson, J., Green, E., Hampson, S.J., Stanford, J., Haagsma, J., Chiodini, R.J., and Hermon-Taylor, J. 1988. Identification and characterization of further bacteria isolated from Crohn's disease and control tissues. Gastroenterol. 94:A294.