Rapid immunochromatographic assay for diagnosis of tuberculosis: antibodies detected may not be specific.

We read with interest the letter by Grobusch et al. (3) on a rapid immunochromatographic assay for the diagnosis of tuberculosis but wish to present evidence that this test may not be as specific as they portrayed it to be. It was long thought that the 38-kDa antigen of Mycobacterium tuberculosis was confined to organisms of the M. tuberculosis complex (MTBC) (1). As suggested in the letter of Grobusch et al., it has therefore been assumed that detection of antibody to the 38-kDa antigen implies infection (although not necessarily disease) with M. tuberculosis, M. africanum, M. bovis, or M. microti. The lack of antibody development after M. bovis BCG vaccination has been ascribed to the lower expression of the 38-kDa antigen in this variant (9). However, Thangaraj and colleagues reported in 1996 (8) the detection of genes encoding the 38-kDa antigen in a strain of M. intracellulare and the expression of the antigen in this species. They speculated that the possession of a homologue of the M. tuberculosis 38-kDa antigen by M. intracellulare might be associated with virulence, since M. intracellulare has the ability to cause disease in immunocompetent hosts, in distinction to M. avium, which does not possess the antigen. Thangaraj et al. did not detect the 38-kDa antigen or the appropriate gene sequence in M. avium, M. paratuberculosis, M. smegmatis, M. fortuitum, M. chelonei, M. leprae, M. kansasii, M. marinum, or M. vaccae. M. malmoense was not included in these investigations. We have produced a recombinant 38-kDa antigen in Escherichia coli as a histidine-tagged fusion protein (7) and used this to raise a panel of 15 mouse monoclonal antibodies (MAbs) by using standard hybridoma technology (6). Selection was based on reactivity with the recombinant antigen by enzyme-linked immunosorbent assay and/or Western blotting of an M. tuberculosis lysate. To further assess the specificity of the MAbs, they were screened against lysates of M. avium, M. intracellulare, M. kansasii, M. malmoense, M. vaccae, and M. xenopi. Upon immunoblotting, 2 of the 15 MAbs (7 and 29) reacted with a 38-kDa protein in M. intracellulare and 2 MAbs (1 and 7) reacted with a 38-kDa protein in M. malmoense (Fig. 1). All remaining MAbs were negative against all lysates. Twenty-one of 23 M. malmoense lysates examined by immunoblotting consistently reacted with both MAbs. The two isolates of M. malmoense not expressing the 38-kDa antigen are also atypical when examined by other taxonomic methods, including random amplified polymorphic DNA analysis (5). Our results confirm the observations of Thangaraj et al. in relation to M. intracellulare and strongly suggest that M. malmoense also expresses a homologue of the 38-kDa antigen of M. tuberculosis. M. malmoense is a recognized primary pathogen of immunocompetent children and the elderly, and it has been estimated to cause up to 10% of tuberculosis-like pulmonary disease in the United Kingdom and other northwest European countries (2, 4). We have yet to proceed with studies to detect the gene sequence in M. malmoense which encodes for the antigen. If such a sequence is found it will extend the argument of Thangaraj et al. (8) that possession of the antigen may be associated with virulence, demonstrating the presence of a common immunodominant antigen in the three species of mycobacteria (MTBC organisms, M. intracellulare, and M. malmoense) which cause disease in immunocompetent individuals. These findings may have important implications for vaccine studies. However, it is also clear from our results and those of Thangaraj et al. that the detection of antibodies which react with the 38-kDa antigen of M. tuberculosis cannot be taken to indicate infection or disease due to MTBC organisms alone, since such antibodies may also be evoked by infection or disease with M. intracellulare or M. malmoense.