The absence of histologic and microbiologic evidence of microorganisms in the sarcoidosis granulomas has led to questions about a role for infectious agents in sarcoidosis immunopathogenesis. This report describes molecular and immunologic support for the presence of the virulence factor, Mycobacterium sodA, in systemic sarcoidosis.
The strength of molecular analysis for sodA lies in the ability to identify pathogenic Mycobacterium species . The detection of sodA nucleic acids yielded important information in this study. Although the primers in region 1 and region 4 were capable of detecting nontuberculous mycobacteria, sequence analysis of the amplicons was most consistent with members of MTB complex, thus revealing mycobacteria capable of inducing granuloma formation. The majority of the sequences possessed 100% positional identity with MTB (GenBank No. AF061030); however, sequence analysis of the amplicons also revealed sequences that were genetically distinct (Table 3, Figure 1B). To confirm a polymorphism, we insured that the nucleotide substitution was present on both strands and that an actual peak was visualized on the chromatogram. The detection of the polymorphism, A302G, in two sarcoidosis specimens processed and analyzed at separate time points also suggests that this polymorphism is not due to PCR error.
Phylogenetic analysis of the amplicons derived from the sarcoidosis granulomas revealed three major points: 1) There is a clear association between the sarcoidosis amplicons and the sequences of members of MTB complex, demonstrating that they share a recent common ancestor, and precluding the possibility that the sarcoidosis sequences are derived from distantly related mycobacterial or Nocardia species. 2) The sequences from Sarcoidosis 2 and 15 form a distinct clade, emphasizing that these sequences may belong to a single mycobacterial species (Figure 1B). Whether these mutations reflect a distinct member of MTB complex or M. tuberculosis strain with a mutation in sodA is unclear. A sodA mutation, unique from that found in the two sarcoidosis specimens, has been described in an isoniazid-resistant MTB strain . 3) The substantial evolutionary distances between the sarcoidosis amplicons and the atypical mycobacteria eliminate the possibility that we have sequences of environmental mycobacteria as a contaminant.
Superinfecting mycobacteria preferentially home to tuberculous granulomas . This observation has been used to explain the detection of mycobacterial nucleic acids in sarcoidosis granulomas. We included in this analysis 10 granulomatous control specimens of known etiology. SodA DNA was detected in only two of them. One of the specimens was a lymph node containing Histoplasma capsulatum in an individual with rheumatoid arthritis. Prior reports have noted an association with latent tuberculosis in such patients . The detection of mycobacterial sodA in sarcoidosis specimens in a significantly different proportion compared to granulomatous negative control specimens suggests that the presence of sodA sequences does not simply reflect preferential localization of environmental mycobacteria to established granulomas.
The inclusion of sarcoidosis and control specimens that were frozen or paraffin-embedded was designed to address the possibility that mycobacteria are introduced into sarcoidosis specimens during tissue fixation. There are reports of environmental mycobacteria in hospital water systems, and well as fiberoptic bronchoscopes [20, 21]. We chose snap frozen and paraffin-embedded tissue specimens among the sarcoidosis and PPD- controls specimens. The absence of sodA in the majority of the negative control specimens, as well as the presence of sodA nucleic acids in snap frozen and paraffin-embedded sarcoidosis tissues, suggests that the mycobacterial nucleic acids detected are not present due to contamination during tissue fixation. Identification of immunogenic sodA peptides has not been previously reported in tuberculosis or sarcoidosis subjects. Despite the absence of mycobacteria by histologic staining and culture of sarcoidosis specimens, the immune response to sodA among the sarcoidosis subjects more closely reflected the response among PPD+ subjects than PPD- volunteers. There was not a significant difference between the sarcoidosis and tuberculosis subjects in the percentage of subjects recognizing these peptides or the distribution of the T cell frequencies (Figure 2). Many of the sarcoidosis and PPD+ subjects recognized more than one peptide, suggesting that the immune response observed in these subjects can be elicited by multiple sodA epitopes (Figure 3). Due to limitations in the number of PBMC, we were unable to test all 40 sodA peptides, so it is possible that sodA epitopes other than the four that we tested exist.
The dual assessment for mycobacteria in sarcoidosis pathogenesis provides interesting information. Among the 12 subjects, 10 possessed molecular or immunologic evidence for sodA. This illustrates the importance of using complementary approaches to assess for microorganisms in pathologic specimens of unknown etiology. The two sarcoidosis subjects who were negative by both methods suggest that mycobacteria do not have a role in all sarcoidosis specimens, supporting other reports of the inappropriate diagnosis of sarcoidosis in subjects with disease with close phenotypic and pathologic features, such as chronic beryllium disease . Thus sarcoidosis may be a common pathologic phenotype that can be caused by infectious or non-infectious etiologies.
PCR of archival specimens was more sensitive for sodA than ELISPOT analysis of peripheral blood (Table 5). There was a time lapse between acquisition of the blood and tissue samples in a given patient; this differential timing of biopsies and blood testing is a major limitation of the analysis. The common observance in the dual analysis was a positive PCR analysis for sodA but a lack of recognition by ELISPOT analysis. This result is likely due to the preservation of nucleic acids by tissue fixation, as opposing to an evolving host immune response. Some of the patients had resolved their disease between the diagnostic biopsy and acquisition of PBMC. Concordantly, there exists evidence that immune recognition of mycobacterial antigens cease with clearance of mycobacteria [17, 23, 24]. Future studies using blood and tissue of sarcoidosis subjects obtained simultaneously will be performed. Another limitation is that the PCR involves analysis of a site of active sarcoidosis involvement [ie the granuloma], whereas the immunologic analysis involves systemic responses. Sensitivity for immune recognition might improve by using lymphocytes from sites of active inflammation such as bronchoalveolar lavage fluid or directly from lymphocytes derived from fresh sarcoidosis granulomas. Finally, the sensitivity of the ELISPOT may have improved if we had tested all 40 sodA peptides. It is unlikely that the four peptides are recognized by all HLA types .