TEXT

Top Abstract Text References

The identification of persons infected with Mycobacterium tuberculosiswho account for one-third of the world's populationhas a high priority in tuberculosis (TB) control programs, second only to the identification and treatment of infectious TB patients (1). To date, the only indicator of latent infection with M. tuberculosis is a positive tuberculin skin test, which measures delayed-type hypersensitivity (DTH) responses to the intradermal injection of the purified protein derivative (PPD) of tuberculin, an ammonium sulfate precipitate of filtrates of heat-inactivated, stationary-phase cultures (24). Unfortunately, the use of PPD imposes serious limitations on skin test accuracy. First, batches of PPD vary in protein composition and potency (13), making it exceedingly difficult to compare results obtained with different individuals, countries, or studies. Second, PPD is highly cross-reactive, for it shares many epitopes with antigens of other mycobacteria. Thus, the specificity of the PPD skin test (i.e., the ability of the test to correctly identify noninfected individuals) is very low for persons who have been vaccinated with Mycobacterium bovis bacillus Calmette-Guérin (BCG) and for persons living in areas that have a high environmental burden of nontuberculous mycobacteria (3, 4).

The premise of our work is that the accurate diagnosis of TB infection by skin testing requires a new tuberculin, one consisting of defined protein antigens that are unique to M. tuberculosis. A new tuberculin should contain many antigens, since cocktails of multiple antigens are needed to elicit strong DTH responses (16) and to cover the broad spectrum of antigen recognition by different individuals (9) typical of TB (23). We have shown that cocktails of M. tuberculosis complex-specific antigens elicit DTH responses that distinguish TB infection from sensitization with nontuberculous mycobacteria (16). The challenge now is to identify antigens unique to M. tuberculosis to develop multiantigen cocktails that allow discrimination between M. tuberculosis infection and vaccination with M. bovis BCG.

We have initiated a search for DTH-active antigens expressed by M. tuberculosis but not by M. bovis BCG. Prior to the present work, only one antigen, ESAT-6 (26), was known to have such properties (another antigen, MPT64 [29], is absent from only some BCG substrains [15, 17], and it evokes a vigorous DTH response to vaccination with some, but not all, substrains of BCG [11]). ESAT-6, which elicits strong, TB-specific DTH responses in guinea pigs (9), in cattle (21), and in humans (22, 27), is encoded by RD1, a DNA region present in the genome of M. tuberculosis and virulent M. bovis but missing from the DNA of all substrains of M. bovis BCG (17). In the present work we show that a second antigen encoded by the RD1 region of M. tuberculosis (gene Rv3874) elicits strong, M. tuberculosis-specific DTH responses in guinea pigs.

The Rv3874 gene product. To identify DTH-eliciting antigens specific to M. tuberculosis, amino acid sequences, which were deduced from the nine putative genes in RD1 (7), were analyzed with a BLAST protein homology program to eliminate proteins having homologs encoded outside the RD1 region, since such proteins are likely to share epitopes with antigens of M. bovis BCG. The products of Rv3874 and Rv3879c shared no homology with proteins encoded elsewhere in the M. tuberculosis genome. Because of previous evidence that the Rv3874 gene is expressed and that the corresponding protein is found in the culture filtrate of M. tuberculosis (culture filtrate protein 10 [5]), we analyzed the Rv3874 product. The Rv3874 open reading frame was amplified from the chromosomal DNA of M. tuberculosis H₃₇Rv by PCR and was cloned into the Escherichia coli expression vector pQE-30 (Qiagen) by following standard protocols (18, 19). Recombinant protein was expressed as a polyhistidine-tagged fusion protein and was purified to near homogeneity by using a three-step chromatography protocol (6). The purified protein had an apparent molecular mass of ~10 kDa as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis, in agreement with a calculated M_r of 10,794 (data not shown).

DTH responses in guinea pigs. The ability of the Rv3874 gene product to elicit DTH was determined by guinea pig skin testing. We tested four groups of female outbred Hartley guinea pigs weighing approximately 300 g each. One group was infected with M. tuberculosis H₃₇Rv by using an aerosol-generating device (Glas-Col, Middlebrook) calibrated to deliver about 100 tubercle bacilli into the lungs of each animal. Two additional, control groups were immunized with a suspension of 10⁶ cells of M. bovis BCG or of M. avium (the most common cause of nontuberculous mycobacterioses in humans [10]) by intradermal injection in the shaved abdomen. A fourth group was mock immunized by intradermal injection of phosphate-buffered saline. This four-group design was utilized in two independent experiments that included two different substrains of M. bovis BCG (BCG Japan and BCG Pasteur).

View larger version (16K): [in this window] [in a new window]   FIG. 1.   The DTH response to MTSA-10 distinguishes TB infection from immunization with BCG or with M. avium. For skin test experiments, M. tuberculosis H37Rv was grown to mid-log phase by shaking at 37°C in liquid glycerol-alanine-salts medium with 0.05% (vol/vol) Tween 80 in a biosafety level 3 laboratory. M. bovis BCG Japan and BCG Pasteur were grown in Sauton's medium with 0.025% (vol/vol) tyloxapol at 37°C. M. avium TMC 724 was grown by shaking at 37°C in Middlebrook 7H9 (Difco) liquid medium with 2% glycerol, 10% oleic acid-albumin-dextrose-catalase supplement, and 0.05% (vol/vol) Tween 80. PPD from M. tuberculosis was prepared at Colorado State University by using a standard protocol (25) and was applied to a Detoxi-gel column (Pierce, Rockford, Ill.) to remove the lipopolysaccharide. Purification of native MPT32 from M. tuberculosis H37Rv was performed by the method of Dobos et al. (8). Purification of recombinant MPT64 from E. coli was described previously (6). Guinea pigs were aerosol infected with M. tuberculosis () or were immunized with BCG () or with M. avium (). Six to eight weeks later, animals were tested intradermally with 1 µg of PPD and 2 µg of purified recombinant antigens. Each point represents 1 animal. Results are expressed as diameter of erythema measured 24 h after antigen injection. (A) In a first experiment, the group sizes were six animals for M. tuberculosis, eight for M. bovis BCG Japan, four for M. avium, and four for saline. Antigens were PPD, MPT64 (BCG Japan produces the M. bovis homolog MPB64; see text), and MTSA-10. (B) In a second experiment, the group sizes were six animals for M. tuberculosis, six for M. bovis BCG Pasteur, six for M. avium, and five for saline. Antigens were PPD, MPT32 (BCG Pasteur does not produce MPB64; see text), and MTSA-10.

Genome analysis of the mtsa-10 gene. The distribution of the mtsa-10 gene in tuberculous and nontuberculous mycobacteria (listed in Table 1) was analyzed by Southern transfer hybridization. The presence of the mtsa-10 gene was visualized as a single hybridization signal of approximately 4 kb (data not shown), indicating that the mtsa-10 gene is present as a single copy in the mycobacterial chromosome. The mtsa-10 gene was found in M. tuberculosis (two laboratory strains and two clinical isolates), in virulent M. bovis, and in Mycobacterium africanum. As expected, the gene was absent from all BCG substrains (Table 2). The mtsa-10 gene was found in only 2 of 13 species of nontuberculous mycobacteria tested, suggesting a limited gene distribution. The mtsa-10 gene was notably absent from M. avium, the most common human pathogen among nontuberculous mycobacteria (10). Among other nontuberculous mycobacteria causing human disease with any significant frequency (10), the mtsa-10 gene was found in Mycobacterium kansasii but not in Mycobacterium fortuitum or Mycobacterium scrofulaceum (Table 2).