Researching the Role of Mycobacterium avium paratuberculosis in the Etiology of Crohn's Disease

G&H Can you describe the current theory of Crohn's disease etiology in terms of endogenous and exogenous factors?

MB It has been convincingly demonstrated that there is a genetic predisposition to Crohn's disease (CD), which constitutes an endogenous susceptibility. In addition, it has been shown, albeit indirectly, that exogenous factors also play a role due to the incomplete penetrance of the genetic predisposition. From past work on the NOD2 genetic variant, it has been shown that only 1–5% of people with the predisposing mutations develop CD. The other 95–99% remain disease-free. This implies that there could be exogenous factors that determine which of these people with susceptible alleles develop clinical disease.

G&H Is there a possibility that the presence of a combination of susceptibility genes is necessary for disease onset and that they have not yet been identified?

MB As human genetic studies recruit more and more patients, including in meta-analyses, additional CD susceptibility genes are continually uncovered. However, to my knowledge, no interaction among CD susceptibility genes has been demonstrated to show that if a patient has variation A and variation B, their risk of CD development is 100%.

G&H Can you describe the current theories of exogenous factors that may trigger CD onset in susceptible individuals?

MB Theorized contributing exogenous factors include a variety of nonmicrobial factors, such as birth control pills and cigarette smoking, all of which may or may not modify disease behavior in a variety of ways. As a microbiologist, my research interest is in microbial exogenous factors, which can be broken down into two groups. The first consists of what is termed normal fecal flora, or intestinal microbiota found in most healthy humans. The second consists of infection with abnormal pathogenic microbes.

Research, primarily in animals, has shown that some animal models of colitis do not occur without the presence of normal intestinal flora. Currently, I am not aware of substantive data in humans that demonstrate the same phenomenon.

In any case, I would also argue that normal flora alone is an unlikely candidate for the trigger that progresses patients from susceptibility to active disease. Using the same statistical reasoning as discussed above regarding NOD2, 95–99% of people with these alleles do not develop disease. Therefore, normal flora, alone, do not seem to have the capacity to trigger illness.

With regard to microbial pathogens, there are several infections that are likely cleared by the majority of hosts and typically do not cause disease. In patients with CD, it is theorized that the genetic variations allow these same organisms to affect the immune response and manifest as active disease. The two candidate pathogens currently receiving the most research attention in this regard are Mycobacterium avium paratuberculosis (MAP) and adherent invasive Escherichia coli strains.

G&H What are the technical and practical challenges in studying MAP as an exogenous cause of CD onset?

MB Although the technical challenges in screening for MAP are not insurmountable, they are still real. If a sputum sample is sent to the laboratory to screen for the agent of Mycobacterium tuberculosis, that lab benefits from a century of experience looking for this mycobacterium in the sputum. There is a standardized, Centers for Disease Control–approved protocol for processing that sample to eliminate other organisms and allow the mycobacterium to grow. There are then standardized ways to take that processed sample, perform microscopy, submit it for culture in culture media, or submit it for nucleic acid amplification, using commercial kits. In the case of MAP, there is no accepted standard for processing the sample for culture media, polymerase chain reaction (PCR), or for microscopy. If a physician has a patient biopsy that they want to test for MAP, and they call the clinical lab in their hospital, there is no standard protocol available.

Further, development of the protocol is not as simple as transferring and adapting existing protocols for tuberculosis. There are conceptual challenges that make consistent protocol development more difficult. When sputum samples are processed for tuberculosis, there may be anywhere from 104 to 106 oral salivary bacteria present, which may interfere with the test. However, the bacterial contamination encountered with intestinal samples, which are needed to screen for MAP, is much greater.

MAP is an intracellular pathogen. It is not thought to be freely circulating in the fecal stream and is not expected to be captured in stool samples, as would Clostridium difficile or Shigella. Also, to say simply that MAP is intra-cellular does not address the question of which cells carry it. It may be found in macrophages, dendritic cells, or another cell type. We currently do not know if we should biopsy the mucosa, the submucosa, the mesenteric nodes, or some combination of sites to screen definitively.

When looking for mycobacterial tuberculosis in the sputum, the premise is that the bacteria have grown in the host, caused rupture in the cells, and are in an extracellular log phase of growth that can be transferred to an extracellular growth media in the lab. If we ask the lab to check for intracellular mycobacteria like MAP, inside the cells of a tissue biopsy, it may not be replicating. Therefore, effective screening for MAP requires a method to disrupt the tissue, liberate the bacteria, and allow them to grow into a culture in the lab.

All of these additional steps require new methods that have been validated and outlined in protocol. Until they have been, different labs in different places will try different methods, none of which is likely to prove reproducible across the board. Labs may perform effective screening on the wrong sample or ineffective screening on the right sample. Therefore, there are both pre-ana-lytic and analytic issues that must be resolved in order to achieve uniform screening.

G&H Is there a possibility that MAP could prove an exogenous factor for CD development when found in combination with another pathogen?

MB It is most certainly possible that MAP is only part of the story and could be one of several microbes playing a role in CD etiology, in one of two potential scenarios. The first is that some humans are predisposed to MAP infection whereas others are predisposed to a different bacterium and both of these pathogens are independent triggers for CD onset in susceptible individuals. They would thus represent different agents on different pathways with the same end result. The other scenario is that MAP infection alters our immune surveillance to other microbes such that the mycobacterium and the other microbe, together, synergistically work in the susceptible host.

G&H Are there currently any findings to support a correlation between levels of MAP concentration and CD disease activity or severity?

MB A refined assay could show how much mycobacterium is present in the tissue and potentially allow us to gauge disease activity and treatment response, a laudable goal. Unfortunately, it is a goal that is still at the inception. In the absence of a reliable assay as described above, it is unlikely that a more sensitive assay is forthcoming. However, in the 1980s, infectious disease specialists were challenged to develop a screening procedure to detect HIV in the blood. By the late 1990s, that assay was available and sophisticated enough to quantify viral load in infected individuals. Once we understand where and how to look for MAP, the engineering steps to quantify it may simply require modifications of the first assay. Unfortunately, we are not yet sure if we should be looking in the blood or in the tissue, directly for the mycobacteria or indirectly for an immune response raised against the mycobacteria. These more fundamental issues remain before we can begin testing for mycobacterial concentration.

G&H If a definitive link were made between MAP infection and CD onset, how might it change our understanding of CD as an autoimmune disorder?

MB For certain autoimmune diseases, like type 1 diabetes, thyroiditis, rheumatoid arthritis, and lupus, there is a measurable antibody response in humans that is directed against self. These patients make antibodies that recognize and react to something in humans, such as rheumatoid factor or antinuclear antibodies. Unlike these diseases, CD has never been formally demonstrated to be an autoimmune disease. There is indirect evidence of auto-immunity, in that CD is treated with immunosuppressive drugs, but it has never formally fulfilled the criteria for autoimmune disorders because, to my knowledge, there is no clearly demonstrable antibody against self that we can say all CD patients produce. Linking CD to MAP infection would further differentiate CD from the other autoimmune disorders.

G&H How might a confirmed link between MAP and CD onset affect approaches to treatment?

MB It has been theorized that if MAP has a role in CD, treating MAP should cure CD. In this discussion, it is important to remember that our understanding of how to cure tuberculosis is well advanced. There have been antibiotics against tuberculosis nearly as long as there have been antibiotics. MAP is a veterinary pathogen. The natural host of MAP does not receive antibiotics because it is more economical to cull the host than to treat it. Therefore, we have limited data on antibiotic treatment for this organism. If we can establish that the organism is related to the disease, we cannot necessarily leap to treating the organism and curing disease. That step will require a series of translational research activities. We will need to know what strains are circulating, what is the antibiotic profile of these strains, and how antibiotics work in in vivo models, before moving forward with clinical trials.

G&H What are the next steps in research of MAP as it relates to CD?

MB I believe that the next fundamental advance in CD will come from the biologic understanding of how CD susceptibility genes mediate our immune response to microbial challenges and whether these microbial challenges are broad-spectrum or instead these genes have a particular role in guiding immune responses to certain microbial agents. We know that CD is a genetic disease, and we know many of the involved genes. Understanding what they normally do and how they are dysfunctional in CD will represent a clear advance.

At the epidemiologic level, we need a standardized assay that can be applied not only in niche labs but on the bench of any clinical lab around the world and can be used in relatively standard studies of cases and controls, finding similar results across the board. I do not think that collecting the patients is a limiting factor in this regard. The CD community has already demonstrated a capacity in all of the genetic studies to collect large numbers of phenotypically characterized patients with CD/colitis/ other controls and to biobank their serum, tissue, and DNA. With a good assay, we can take these well-described populations and say that there is or is not an enrichment of mycobacteria among the group with CD versus controls. Currently, there are two meta-analyses that suggest an association. Despite that, I would say that the tissue PCR that is being used and described in the meta-analy-ses is still not a protocol that will yield consistent results at the average clinical lab. The tests remain difficult and challenging and, on the epidemiologic level, we still need a simple, reproducible assay.