Proskauer Beck medium until surface cultures were con- fluent, generally 8 ..... of the antibodies to the recombinant proteins rose consider- ably, anda broad ...
IMMUNITY, Feb. 1991, p. 665-670
Vol. 59, No. 2
Human Immune Response to Mycobacterium tuberculosis Antigens DIANE V. HAVLIR, ROBERT S. WALLIS, W. HENRY BOOM, THOMAS M. DANIEL, KEITH CHERVENAK, AND JERROLD J. ELLNER* Department of Medicine, Case Western Reserve University, and University Hospitals, Cleveland, Ohio 44106 Received 25 June 1990/Accepted 21 November 1990
Little is known about the immunodominant or protective antigens of Mycobacterium tuberculosis in humans. Cell-mediated immunity is necessary for protection, and healthy tuberculin-positive individuals are relatively resistant to exogenous reinfection. We compared the targets of the cell-mediated immune response in healthy tuberculin-positive individuals to those of tuberculosis patients and tuberculin-negative persons. By using T-cell Western blotting (immunoblotting) of nitrocellulose-bound M. tuberculosis culture filtrate, peaks of T-cell blastogenic activity were identified in the healthy tuberculin reactors at 30, 37, 44, 57, 64, 71 and 88 kDa. Three of these fractions (30, 64, and 71 kDa) coincided with previously characterized proteins: antigen 6/alpha antigen, HSP60, and HSP70, respectively. The blastogenic responses to purified M. tuberculosis antigen 6/alpha antigen and BCG HSP60 were assessed. When cultured with purified antigen 6/alpha antigen, lymphocytes of healthy tuberculin reactors demonstrated greater [3H]thymidine incorporation than either healthy tuberculinnegative controls or tuberculous patients (8,113 ± 1,939 Acpm versus 645 ± 425 Acpm and 1,019 ± 710 Acpm, respectively; P < 0.01). Healthy reactors also responded to HSP60, although to a lesser degree than antigen 6/alpha antigen (4,276 + 1,095 Acpm; P < 0.05). Partially purified HSP70 bound to nitrocellulose paper elicited a significant lymphocyte blastogenic response in two of six of the tuberculous patients but in none of the eight healthy tuberculin reactors. Lymphocytes of none of five tuberculin-negative controls responded to recombinant antigens at 14 or 19 kDa or to HSP70. Antibody reactivity generally was inversely correlated with blastogenic responses: tuberculous sera had high titer antibody to M. tuberculosis culture filtrate in a range from 35 to 180 kDa. This is the first systematic evaluation of the human response to a panel of native and recombinant antigens in healthy tuberculin reactors and tuberculous patients. Antigens which stimulated prominent lymphocyte blastogenic responses were identified in seven fractions on T-cell Western blot analysis. Two of these may represent previously characterized proteins; the others may contain immunodominant proteins that will require further characterization.
Tuberculosis remains a major worldwide health problem, resulting in an estimated 3 million deaths each year (25). Mycobacterium bovis BCG, the only vaccine currently available, has shown variable and unpredictable efficacy in several large clinical trials (11). One of the important steps in vaccine development is the identification of immunodominant antigens which may be protective. Through the screening of a M. tuberculosis recombinant Agtll library with monoclonal antibodies derived from immunized mice, a panel of recombinant antigens has been identified (10). Although humans also develop antibodies to these antigens (14, 16) and T-cell clones have been developed which react with several of these antigens (9, 15, 17, 18), their relative importance in human tuberculosis has not been determined. Little, therefore, is known about immunodominant antigens of M. tuberculosis in human infection and disease. Cell-mediated immunity is the critical protective immune response in human disease. In animal models of mycobacterial immunity, protection can be conferred through adoptive transfer of sensitized T cells to immunologically naive animals (19, 20). In humans, tuberculin skin test reactivity is epidemiologically linked with resistance to exogenous reinfection with M. tuberculosis (24). Conversely, persons infected with the human immunodeficiency virus show an increased susceptibility to M. tuberculosis disease (21). While tuberculous patients develop high titers of M. tuberculosis-reactive antibodies, there is no evidence suggesting these have any role in protection. *
Given these associations, we chose to identify immunodominant and potentially protective targets of cell-mediated immunity in healthy tuberculin reactors. We sought to define dominant antigens by systematically comparing responses to a panel of well-defined antigens and to fractions of mycobacterial culture filtrate by using the technique of T-cell Western blot (immunoblot) analysis. Seven peaks of T-cell reactivity were identified in healthy tuberculin reactors. Two of these fractions may represent previously described antigens (antigen 6/alpha antigen and HSP60), while the others may contain novel proteins which are immunodominant in human infection. MATERIALS AND METHODS
Subjects. Nineteen subjects were studied. Five were healthy tuberculin-negative individuals, eight were healthytuberculin positive subjects, and six were patients with tuberculosis monitored at the MetroHealth Medical CenterCuyahoga County Tuberculosis Clinic. All of the tuberculous patients had pulmonary tuberculosis with positive sputum cultures and were receiving standard chemotherapy: one for 6 weeks, two for 2 months, and two for 5 months. Two had been skin tested at the time of diagnosis and were both tuberculin reactive. None of the patients with tuberculosis had evidence of infection with the human immunodeficiency virus. Children and the elderly were excluded from the study. The ages and racial distribution of the healthy tuberculin reactors and patients were similar but were not specifically matched. Antigens. Mycobacterial strain H37Rv was grown on
Corresponding author. 665
HAVLIR ET AL.
Proskauer Beck medium until surface cultures were confluent, generally 8 weeks. The mycobacteria were sedimented, and the culture filtrate was dialyzed against water by using a Spectrapor 2 membrane (Spectrum Medical, Los Angeles, Calif.), filtered through a 0.45-pLm-pore-size filter, and lyophilized. Purified recombinant BCG HSP60 was a gift of J. D. A. van Embden (9). Antigen 6/alpha antigen was purified from H37Ra M. tuberculosis culture filtrate (20a). The filtrate was precipitated with 50% ammonium sulfate, and then DEAEcellulose exchange chromatography was performed as described by Daniel and Ferguson (7). The antigen was eluted with 0.3 M sodium phosphate, lyophilized, and stored at 4°C. PPD was a gift from Lederle Laboratories (Pearl River, N.Y.). Streptolysin 0 was purchased from Difco Laboratories (Detroit, Mich.). Recombinant Xgtll clones SK50 (HSP70), SK44 (14-kDa protein), and SK4 (19-kDa protein) were provided by T. Shinnick (Centers for Disease Control, Atlanta, Ga.). Bacterial lysates containing these phage were prepared in Escherichia coli 1089. Single colonies were inoculated into LB broth and cultured at 30°C. Once the optical density at 600 nm reached 0.5, the temperature was raised to 42°C for 20 min; the cells were then incubated for 1 h at 37°C. The cells were sedimented at 7,000 x g for 5 min and resuspended in 5% of the original culture volume. The cells were frozen in liquid nitrogen and lysed by thawing. The preparations were sonicated 10 s three times, filtered through a 0.22-,um-poresize filter, and stored frozen at -70°C. A native Xgtll E. coli lysate was prepared in an identical manner. T-cell Western blotting. Antigens were prepared for T-cell Western blotting as described by Abou-Zeid (1). In brief, 75 ,ul of H37Rv culture filtrate (15 mg/ml) or recombinant lysate was mixed with 75 ,ul of sample buffer (25% 0.5 Tris hydrochloride, pH 6.8, 4% sodium dodecyl sulfate, 20% glycerol, 10% 2-mercaptoethanol). The sample was heated for 2 min at 100°C and applied to a 9% polyacrylamide-bis gel, washed with phosphate-buffered saline, and stained with Aurodye (Amersham Corp., Arlington Heights, Ill.) overnight. The nitrocellulose paper was then cut into 29 2-mm strips, dried overnight, and dissolved in 1 ml of dimethyl sulfoxide. The nitrocellulose was precipitated with sodium carbonate buffer, pH 9.6. For the recombinant preparations, an 8-mm vertical strip of the nitrocellulose paper was removed prior to Aurodye staining and subjected to Western blotting with murine monoclonal antibodies IT-1 (14 kDa), IT-10 (19 kDa), and IT-11 (HSP70), which were gifts of T. Shinnick. A 2-mm horizontal strip of the Aurodye-stained protein corresponding to the location of the recombinant protein on Western blotting was then removed and processed in the same manner as described for the culture filtrate. Cell culture and assay of blastogenesis. Mononuclear cells were obtained from heparinized blood by density sedimentation over Ficoll Hypaque (Pharmacia, Uppsala, Sweden). Cells were washed three times with medium (RPMI 1640; Whittaker Bioproducts, Walkersville, Md.). Cells were suspended at a density of 2 x 106/ml in complete medium (RPMI with 2 mM L-glutamine, 100 ,ug of gentamicin per ml, 15 mM HEPES [N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid], and 10% pooled human sera), and 2 x 105 cells were placed in each well of a round-bottom microtiter tissue culture plate (Falcon Products, Becton Dickinson, Oxnard, Calif.). Blastogenic response was measured to purified protein derivative (PPD) (10 ,ug/ml), SLO 1:100, antigen 6/alpha antigen (1 p.g/ml), and HSP65 (5 ,ug/ml), with triplicate wells
per variable. The cultures were incubated for 5 days at 37°C in 5% CO2. [3H]thymidine (1 ,Ci; specific activity, 6.7 Ci/mmol; ICN, Costa Mesa, Calif.) was added to each culture well for the final 24 h. Cells were harvested with a PHD harvester (Cambridge Technology, Cambridge, Mass.), and [3H]thymidine was measured in a scintillation counter. Results were expressed as mean counts per minute of the triplicates. A response was considered positive if the following two criteria were met: the stimulation index (defined as counts per minute in antigen-stimulated cultures/counts per minute in unstimulated cultures) was greater than 2 and the A&cpm (counts per minute in stimulated cultures - counts per minute in control wells) was greater than or equal to 3,000. Western blotting. Western blotting of M. tuberculosis H37Rv culture filtrate and SK4, SK44, and SK70 recombinant antigen lysates with human sera was performed on samples from all subjects by previously published methods (26). Serum was diluted 1:100 to 100,000 with RPMI 1640 with 1% bovine serum albumin (Sigma Chemical Co., St. Louis, Mo.) and allowed to react overnight at 4°C. For the Western blots of recombinant antigen lysates, the human sera was reabsorbed for 1 h at 4°C with native Agtll E. coli lysate prior to blotting. After washing, blots were incubated for 4 h with alkaline phosphatase-conjugated anti-human polyvalent immunoglobulins (Sigma). Alkaline phosphatase activity was detected by using Nitro Blue Tetrazolium and 5-bromo-4-chloro-3-indoyl-phosphate in 100 mM Tris, pH 9.5-10 mM NaCl-5 mM MgCl2. Statistics. One-way analysis of variance (ANOVA) was used to compare three or more samples; Student's two-tailed t test was used to compare 2 groups. RESULTS T-cell blastogenic response to native and recombinant M. tuberculosis antigens. Each of the seven healthy tuberculin reactors tested showed significant blastogenic responses to multiple fractions of M. tuberculosis culture filtrate by using T-cell Western blot analysis. A representative blot of two tuberculin reactors and one control tuberculin nonreactor is shown in Fig. 1. None of the tuberculin-negative controls responded significantly to any of the fractions of M. tuberculosis according to the criteria for positive responses outlined above. Although there was substantial heterogeneity among healthy tuberculin reactors, some of the fractions stimulated lymphocyte blastogenesis in a majority of the subjects. While some of these peaks of reactivity correlated with Coomassie blue-stained bands in culture filtrate (shown to the left of the graph), others did not. Lymphocytes of tuberculous patients also showed blastogenic reactivity to fractions of M. tuberculosis culture filtrate, although these were lower in magnitude than the healthy tuberculin reactors. Representative T-cell blot responses.for a tuberculous patient and a healthy control are shown in Fig. 2. Results of the T-cell Western blot analysis are summarized in Fig. 3. Seven fractions of M. tuberculosis filtrate (88, 71, 64, 57, 44, 37, and 30 kDa) contained antigens which stimulated lymphocyte blastogenic responses in a majority of healthy tuberculin reactors. None of these fractions were recognized by a majority of the tuberculous patients; the responses that were present tended nevertheless to be directed against these same fractions. The exception was the 30-kDa fraction, to which none of the tuberculosis patients responded. Three of the seven immunodominant fractions corre-
VOL. 59, 1991
HUMAN IMMUNE RESPONSE TO M. TUBERCULOSIS
I r'. C: 11
FIG. 1. T-cell Western blot responses to M. tuberculosis culture and -) filtrate for two healthy PPD-positive subjects (right, -). The horizontal and one healthy PPD-negative control (left, axis shows T-cell blastogenic activity to each of the fractions. A Coomassie-stained gel of M. tuberculosis filtrate is shown at the left. ---
sponded to previously characterized M. tuberculosis proteins. The 71- and 64-kDa antigen fractions corresponded to the recombinant antigens HSP70 and HSP60, while the 30-kDa fraction corresponded to antigen 6/alpha antigen.
25 20 15 fraction FIG. 3. T-cell Western blot reactivity for all tuberculosis patients (n = 5) and PPD-positive subjects (n = 7) to M. tuberculosis culture filtrate. Vertical bars represent the percentage of subjects with positive responses (stimulation index > 2 and Acpm > 3,000) for each fraction tested.
25 30 0
10 cpm FIG. 2. T-cell Western blot responses to M. tuberculosis culture ) and one healthy PPD filtrate for one tuberculous patient ( negative control (--- -). Prominent antigens were identified at 109, 64, and 41 kDa by T cells of the tuberculous patient. Note, however, that the scale of the horizontal axis is reduced compared with that in Fig. 1.
Western blot analysis using monoclonal antibodies to the recombinant heat shock proteins and antigen 6/alpha antigen confirmed the presence of these proteins in the expected fractions of culture filtrate. We next tested the activity of these previously characterized antigens in the same three groups of patients (Fig. 4). Purified antigen 6/alpha antigen elicited significant blastogenic activity in the T cells of seven of eight healthy tuberculin reactors (mean [3H]thymidine, 8,113 + 1,939 Acpm; P < 0.01 by ANOVA). None of the tuberculin-negative controls or the tuberculosis patients responded to this antigen. Three of the eight healthy tuberculin reactors responded to HSP60, as did one of five tuberculous patients. All of the healthy tuberculin reactors and tuberculous patients responded to PPD, although responses of the tuberculous patients were significantly lower (P < 0.01). We also tested blastogenic responses to the 71-, 14-, and 19-kDa recombinant antigens in a semipurified form on nitrocellulose paper. The 14- and 19-kDa antigens specifically were included because T-cell Western blots, performed using 9% acrylamide gels, may have been inadequate to detect responses to low-molecular-weight proteins. The 14and 19-kDa antigens failed to elicit a blastogenic response in the T cells of any of the subjects tested. There were positive responses to the 71-kDa antigen in 2 tuberculous subjects. Antibody responses to M. tuberculosis culture filtrate and recombinant antigens. Conventional Western blotting was
HAVLIR ET AL.
HTR _ TB
0 L.. ._
FIG. 4. Blastogenic response to soluble M. tuberculosis antigens. Lymphocytes were cultured with antigen (PPD [10 ,ug/ml], BCG HSP60 [5 ,ug/ml], or antigen 6/alpha antigen [1 p.g/ml]) for 5 days; [3H]thymidine was assessed during the final day. Groups marked * differed significantly by ANOVA with P < 0.01; those marked + differed with P < 0.05.
performed in parallel with the T-cell blots. Healthy tuberculin reactors showed heterogeneity in their antibody responses to culture filtrate. In general, M. tuberculosis antibody was of low titer (1/100 dilution). The antigens identified were not consistent among donors, nor did they correspond to the peaks identified in T-cell Western blot analysis. A representative Western blot is shown in Fig. 5. Both control and healthy-tuberculin-reactor sera identified bands at approximately 80 kDa and in the 20- to 25-kDa range. In contrast, tuberculous patients had high titers of M. tuberculosis-reactive antibody (up to 1/10,000 dilution), in a range of molecular sizes from 35 to 180 kDa. Three of the recombinant proteins (14, 19, and 71 kDa) were identified by antibody but not T cells. Even though only two of study subjects mounted T-cell responses to any of these antigens, all 15 (including tuberculin-negative controls) had antibodies reactive with recombinant 14-kDa ctrl 180
zs .:i:,, z.::'.
antigen when tested at a 1/100 dilution by Western blotting (Table 1). One of three tuberculin-negative controls, two of seven tuberculin reactors, and one of five tuberculous patients demonstrated antibody reactivity with the 19-kDa antigen. A larger number of subjects responded to recombinant HSP70 (three controls, four tuberculin reactors, and five tuberculous patients). Thus, there was marked discordance between T- and B-cell responses. DISCUSSION Several antigens of M. tuberculosis have been identified and cloned by using murine monoclonal antibodies (2, 22). Although these are capable of eliciting blastogenic responses in human T-cell clones (4, 9, 18), their relative importance in human M. tuberculosis infection can only be determined in comparison to a full repertoire of mycobacterial antigens. Collins and Young have highlighted the immunogenic potential of M. tuberculosis culture filtrate (6). It and PPD differ in many ways: the virulence of the mycobacterial strains (H37Rv versus H37Ra), the medium (Proskauer Beck versus TABLE 1. T-cell blastogenesis and seroreactivity to M. tuberculosis native and recombinant antigens
58.5 48.5 36.5 26.6
PPD 7lkDa 64 kDa
30 kDa 19 kDa 14 kDa 4 5 1 3 2 FIG. 5. Western blot of M. tuberculosis culture filtrate with human sera. Molecular weight markers are in far left column. Lane 1, Serum from a healthy tuberculin-negative control, diluted 1/100; lane 2, serum from a healthy tuberculin reactor, diluted 1/100; lanes 3 to 5, serum from a representative tuberculous patient (samples diluted 1/100, 1/1,000, and 1/10,000, respectively).
0 0 0 0 0 0
ND ++ ND ND + ++++
ND ++++ ND ND 0 ++++
ND ++++ ND ND + ++++
0 + +++ 0 0
0 0 0 0
a Blastogenic responses to native (PPD and antigen 6/alpha antigen [30-
kDal) and recombinant (14-, 19-, 64-, and 71-kDa) antigens was assessed as [3Hlthymidine incorporation by blood mononuclear cells from healthy tuberculin reactors (HTR), tuberculous patients (TB), and tuberculin-negative controls (CTRL). Seroreactivity was assessed by Western blot. Responses are indicated according to the following scale: +, 25 to 50% positive; + +, 51 to 75% positive; +++, 76 to 99% positive; ++++, 100% positive; ND, not done.
VOL. 59, 1991
Middlebrook 7H-9), and the use of detergents and continuous agitation (both in the preparation of PPD). Most importantly, culture filtrate is neither autoclaved nor treated by ammonium sulfate precipitation. Proteins of culture filtrate are thus less likely to undergo hydrolysis and more likely to retain the tertiary structure that may be important in antibody binding. We therefore chose to analyze the responses to fractions of M. tuberculosis culture filtrate using the technique of T-cell Western blotting. We found that a majority of healthy tuberculin reactors responded to seven fractions, with molecular sizes of 30, 37, 44, 57, 64, 71, and 88 kDa. A potential disadvantage of this technique is that the fractions are not adjusted for protein content, and one might hypothesize that these peaks merely correlated with major protein bands. This did not appear to be the case, as some abundant proteins failed to elicit significant T-cell responses, and some of the peaks of T-cell reactivity were not associated with prominent areas of Coomassie blue staining. Our results are in partial agreement with those of Barnes et al., who performed T-cell Western blot analysis of M. tuberculosis sonicate by using T-cell lines reactive to a M. tuberculosis cell wall preparation (3). He found prominent responses to proteins at 10, 19, 23, 28, 30, 40 to 50, and 65 kDa. Three of these fractions (30, 44, and 64 kDa) correspond to prominent antigens in our investigation, suggesting that some of these proteins may be secreted as well as cell wall associated. Barnes used a gradient gel; our study was performed with a uniform 9% polyacrylamide gel. This may have precluded our identification of the smaller proteins, particularly that at 10 kDa. Only 5 to 10% of individuals infected with M. tuberculosis ever develop clinical illness; the nature of their apparently unique susceptibility is not well understood. One might hypothesize that the immune systems of such susceptible individuals are unable to recognize important mycobacterial antigens. If this were true, one might find "holes" in the repertoire of responses to antigens in tuberculous patients when compared with healthy tuberculin reactors. In this study, one fraction, at 30 kDa, failed to elicit blastogenic responses in any of the tuberculous patients, despite the fact that it contained an antigen immunodominant for healthy tuberculin reactors. Similarly, purified antigen 6/alpha antigen, which is represented in the 30-kDa fraction, failed to elicit a blastogenic response in any of the tuberculous patients, despite doing so in T cells of seven of eight of the healthy tuberculin reactors. It is intriguing to speculate that the inability to respond to this antigen may predispose an individual to tuberculosis. The present study does not establish the nature of the relationship between the inability to mount a T-cell response to antigen 6/alpha antigen and the occurrence of active tuberculosis, however. T-cell function (including blastogenesis, interleukin-2 production, and gamma interferon production) is often suppressed by monocytes (or other cells) in patients with tuberculosis (8). The generally low T-cell responses of the tuberculous patients in the present study are consistent with this observation. The particular lack of responses to the 30-kDa fraction might alternatively be due to activation of suppressor mechanisms by antigen 6/alpha antigen or other proteins in this fraction. Nonetheless, the data suggest that antigen 6/alpha antigen may be important in protective immunity and worthy of further investigation. Antigen 6 is a secreted protein of M. tuberculosis. It and alpha antigen have 20 identical N-terminal amino acids and are likely identical. Alpha antigen corresponds to BCG 85B
HUMAN IMMUNE RESPONSE TO M. TUBERCULOSIS
in the Closs crossed immunoelectrophoresis reference system for BCG (16, 27). Huygen et al. assessed lymphoproliferation and gamma interferon response to P32, identified as 85A in the BCG 85 complex (13). They found a mean blastogenic response in 12 tuberculin-positive individuals of 6,105 Acpm versus 209 in controls and 1,915 in tuberculosis patients. Our results support these observations and suggest that the entire BCG 85 complex may be important in the cell-mediated response to infection with M. tuberculosis. In this study, the 64-kDa fraction of M. tuberculosis culture filtrate elicited a response in five of eight healthy tuberculin reactors while purified BCG HSP60 did so in only three. This suggests that other proteins in this fraction may be important targets of T-cell reactivity. It also suggests that while HSP60 is an immunodominant protein in the antibody responses of immunized mice, this may not be true for T-cell responses in infected humans. Among the other recombinant antigens tested, only the 71-kDa antigen (HSP70) elicited a blastogenic response in any of the subjects. Although human T-cell clones reactive to the 14- and 19-kDa proteins have been described, most PPD-driven CD4+ T-cell clones do not respond to any of these antigens (18), suggesting that they are not major targets for the M. tuberculosis T-cell immune response. The antibody and T-cell responses of the subjects of this study differed in both their regulation and the antigens identified. Antibody to many of the recombinant antigens could be detected in the serum of the healthy tuberculin reactors, although T-cell responses could be detected in none (except HSP60). In subjects with tuberculosis, the titer of the antibodies to the recombinant proteins rose considerably, and a broad range of new antigens also were identified in M. tuberculosis filtrate. In contrast, blastogenic responses to these same antigens were significantly reduced. Blastogenic responses to the recombinant proteins could be detected in only two subjects, both tuberculous patients. The 14-, 19-, and 71-kDa proteins seemed mainly to be B-cell antigens. In separate studies of T-cell and antibody responses to mycobacterial antigens in tuberculosis, it has been reported that cell-mediated immunity is suppressed (23) and antibody responses are enhanced (5, 12). This is the first report in which both T-cell and antibody responses to a panel of mycobacterial antigens have been assessed simultaneously. This dissociation between humoral and cell-mediated responses supports the view that antibodies are not protective in the human immune response to M. tuberculosis and that the search for protective antigens should not rely solely on those antigens identified by monoclonal antibodies of immunized animals. In summary, we have systematically evaluated the blastogenic responses of blood mononuclear cells of healthy tuberculin reactors to M. tuberculosis culture filtrate and identified seven dominant peaks of T-cell reactivity. Two of these may represent activity to the previously characterized proteins antigen 6/alpha antigen and HSP60. Blastogenic reactivity to antigen 6/alpha antigen was not detectable in tuberculous patients; whether this represents selective suppression as a consequence of disease or a causal factor in its pathophysiology will require further investigation. Molecular characterization of the reactive fractions which do not correspond to previously described proteins may lead to identification of novel antigens important in the immune response in human tuberculosis.
HAVLIR ET AL.
We thank Frits van der Kuyp for his assistance in patient recruitment, Tom Shinnick for providing guidance in use of recombinant antigens, and Pamela Bowman for typing the manuscript. Recombinant BCG HSP60 was made available through J. D. A. van Embden and supported by the UNDP-World Bank-WHO Special Programme for Research and Training in Tropical Diseases. This work was supported in part by Public Health Service grants Al 07024, Al 27243, Al 25076, and Al 24298 from the National Institutes of Health. 1.
REFERENCES Abou-Zeid, C., E. Filley, J. Steele, and G. A. W. Rook. 1987. A simple new method for using antigens separated by polyacrylamide gel electrophoresis to stimulate lymphocytes in vitro after converting bands cut from Western blots into antigen bearing particles. J. Immunol. Methods 98:5-10. Anderson, A. B., A. Worsaee, and S. D. Chaparas. 1988. Isolation and characterization of recombinant Xgtll bacteriophages expressing eight different mycobacterial antigens of potential immunological relevance. Infect. Immun. 56:13441351. Barnes, P. F., V. Mehra, G. R. Hirschfield, S. Fong, C. AbouZeid, G. Rook, S. Wu Hunter, P. J. Brennan, and R. Modlin. 1989. Characterization of T-cell antigens associated with the cell wall protein peptidoglycan complex of Mycobacterium tuberculosis. J. Immunol. 143:2656-2662. Boom, W. H., R. N. Husson, R. A. Young, J. R. David, and W. F. Piessens. 1987. In vivo and in vitro characterization of murine T-cell clones reactive to Mycobacterium tuberculosis. Infect. Immun. 55:2223-2229. Coates, A. R. M., M. J. Nicolai, M. J. Pallen, A. Guy, S. D. Chaparas, and D. A. Mitchison. 1989. The 45 kilodalton molecule of Mycobacterium tuberculosis identified by immunoblotting and monoclonal antibodies as antigenic in patients with tuberculosis. Br. J. Exp. Pathol. 70:215-225. Collins, F. M., J. R. Lamb, and D. B. Young. 1988. Biological activity of protein antigens isolated from Mycobacterium tuberculosis culture filtrate. Infect. Immun. 56:1260-1266. Daniel, T. M., and L. E. Ferguson. 1970. Purification and characterization of two proteins from culture filtrates of Mycobacterium tuberculosis H37Ra strain. Infect. Immun. 1:164168. Ellner, J. J., W. H. Boom, K. L. Edmonds, E. A. Rich, Z. Toosi, and R. S. Wallis. 1990. Regulation of the immune response to Mycobacterium tuberculosis, p. 77-91. In E. M. Ayoub, G. H. Cassell, W. C. Branche, Jr., and T. J. Henry (ed.), Microbial determinants of virulence and host response. American Society for Microbiology, Washington, D.C. Emmrich, F. E., J. Thole, J. van Embden, and S. H. E. Kaufmann. 1986. A recombinant 64 kilodalton protein of Mycobacterium Bovis specifically stimulates human T4 clones reactive to mycobacterial antigens. J. Exp. Med. 163:1024-1029. Engers, H. D., V. Houba, J. Bennedsen, T. M. Buchanan, S. D. Chaparas, G. Kadival, 0. Closs, J. R. David, J. D. A. van Embden, T. Godal, S. A. Mustafa, J. Ivanyi, D. B. Young, S. H. E. Kaufmann, A. G. Khomenko, A. H. J. Kolk, M. Kubin, J. A. Louis, P. Minden, T. M. Shinnick, L. Trnka, and R. A. Young. 1986. Results of a World Health Organization-sponsored workshop to characterize antigens recognized by mycobacteriaspecific monoclonal antibodies. Infect. Immun. 51:718-720. Fine, P. E. 1989. The BCG story: lessons from the past and implications for the future. Rev. Infect. Dis. 11(52):S353-S359. Grange, J. M. 1984. The humoral immune response in tubercu-
losis: its nature, biological role and diagnostic usefulness. Adv. Tuberc. Res. 21:1-78. Huygen, K., J. P. van Vooren, M. Turneer, R. Bosmans, P. Dierckx, and J. DeBruyn. 1988. Specific lymphoproliferation, gamma interferon production, and serum immunoglobulin G directed against a purified 32 kDa mycobacterial protein antigen (P32) in patients with active tuberculosis. Scand. J. Immun. 27:187-194. Ivanyi, J., G. H. Bothamley, and P. S. Jackett. 1988. Immunodiagnostic assays for tuberculosis and leprosy. Br. Med. Bull. 4:634-649. Lamb, J. R., J. Ivanyi, A. Rees, R. A. Young, and D. B. Young. 1986. The identification of T-cell epitopes in Mycobacterium tuberculosis using human T-lymphocyte clones. Lepr. Rev. 57(S2):131-137. Matsuo, K., R. Yamasuchi, A. Yamazaki, H. Tasaka, and T. Yamada. 1988. Cloning and expression of the Mycobacterium bovis BCG gene for extracellular antigen. J. Bacteriol. 170: 3847-3854. Mustafa, A. S., F. Oftung, H. K. Gill, and I. Natvig. 1986. Characteristics of human T-cell clones from BCG and killed M. leprae vaccinated subjects and tuberculosis patients. Lepr. Rev.
7(S2):123-130. 18. Oftung, F., A. S. Mustafa, R. Husson, R. A. Young, and T. Godal. 1987. Human T-cell clones recognize two abundant Mycobacterium tuberculosis protein antigens expressed in Escherichia coli. J. Immunol. 138:927-931. 19. Orme, I. M. 1988. Induction of nonspecific acquired resistance and delayed-type hypersensitivity, but not specific acquired resistance in mice, inoculated with killed mycobacterial vaccines. Infect. Immun. 56:3310-3312. 20. Orme, I. M., and F. M. Collins. 1983. Protection against Mycobacterium tuberculosis infection by adoptive immunotherapy. J. Exp. Med. 158:74-83. 20a.Sada, S. D., L. E. Ferguson, and T. M. Daniel. 1990. An enzyme linked immunosorbent assay (ELISA) for the serodiagnosis of tuberculosis using a 30,000 dalton native antigen. J. Infect. Dis. 162:928-931. 21. Selwyn, P. A., D. Hartel, V. A. Lewis, E. Schoenbaum, S. Vermund, R. Klein, A. Walker, and G. H. Frieland. 1989. A prospective study of the risk of tuberculosis among intravenous drug users with human immunodeficiency virus infection. N. Engl. J. Med. 320:545-550. 22. Shinnick, T. M., C. Krat, and S. Schadon. 1987. Isolation and restriction site maps of the genes encoding five Mycobacterium tuberculosis proteins. Infect. Immun. 55:1718-1721. 23. Shiratsuchi, H., and I. Tsuyuguchi. 1984. Analysis of T-cell subsets by monoclonal antibodies in patients with tuberculosis after in vitro stimulation with purified protein derivative of tuberculin. J. Exp. Immunol. 57:271-278. 24. Stead, W. W., J. P. Lofgren, E. Warren, and C. Thomas. 1985. Tuberculosis as an endemic and nosocomial infection among the elderly in nursing homes. N. Engl. J. Med. 312:1483-1487. 25. Styblo, K., and A. Rouillon. 1981. Estimated global incidence of smear positive pulmonary tuberculosis. Unreliability of officially reported figures on tuberculosis. Bull. Int. Union Tuberc. 56:118-125. 26. Wallis, R. S., S. L. M. Aide, D. V. Havlir, M. Amir-Tahmasseb, T. M. Daniel, and J. J. Ellner. 1989. Identification of antigens of Mycobacterium tuberculosis using human monoclonal antibodies. J. Clin. Invest. 84:214-219. 27. Wiker, H. G., M. Harboe, S. Nagai, M. E. Patarayo, C. Ramirez, and N. Cruz. 1986. MPB59, a widely cross reacting protein of Mycobacterium bovis BCG. Int. Arch. Allergy Appl. Immun. 81:307-314.