Evaluation of in situ methods used to detect Mycobacterium avium subsp. paratuberculosis in samples from patients with Crohn's disease.

In common with other diagnostic tests, detection of mycobacteria in tissue by microscopic examination is susceptible to spectrum bias. Since Crohn's disease is defined by the absence of detectable pathogenic organisms, the use of in situ techniques to search for Mycobacterium avium subsp. paratuberculosis in Crohn's disease samples requires validation of methods in a paucibacillary setting. To generate paucibacillary infection, C57BL/6 mice were artificially infected with Mycobacterium avium subsp. paratuberculosis strain K10 and M. tuberculosis H37Rv, yielding tissues harboring fewer than one bacillus per oil immersion field. Serial sections of organs were then studied by cell wall-based staining techniques (Ziehl-Neelsen and auramine rhodamine) and nucleic acid-based staining techniques (in situ hybridization [ISH] and indirect in situ PCR [IS PCR]). Microscopic examination and measurement of morphometric parameters of bacilli revealed that for all methods, Mycobacterium avium subsp. paratuberculosis bacilli were observed to be shorter, smaller, and less rod shaped than M. tuberculosis bacilli. Ziehl-Neelsen, auramine rhodamine stains, ISH targeting rRNA, and IS-PCR targeting the IS900 element afforded comparable sensitivities, but for all methods, visualization of individual bacterial forms required magnification x1,000. Auramine rhodamine staining and IS-PCR generated positive signals in negative controls, indicating the nonspecificity of these assays. Together, our results indicate that detection of Mycobacterium avium subsp. paratuberculosis bacilli in tissue requires oil immersion microscopy, that rRNA-ISH provides sensitivity and specificity comparable to those of Ziehl-Neelsen staining, and that the microscopic detection limit for Mycobacterium avium subsp. paratuberculosis in tissue is governed more by bacterial burden than by staining method.