Development of accurate methods for predicting progression of tuberculosis (TB) from the latent state is recognized as vitally important in controlling TB, because a majority of cases develop from latent infections. Past TB that has never been treated has a higher risk of progressing than does latent Mycobacterium tuberculosis infection in patients who have previously received treatment. Antibody responses against 23 kinds of M. tuberculosis proteins in individuals with past TB who had not been medicated were evaluated. These individuals had significantly higher concentrations of antibodies against Antigen 85A and mycobacterial DNA-binding protein 1 (MDP1) than did those with active TB and uninfected controls. In addition, immunohistochemistry revealed colocalization of tubercle bacilli, antigen 85 and MDP1 inside tuberculous granuloma lesions in an asymptomatic subject, showing that M. tuberculosis in lesions expresses both antigen 85 and MDP1. Our study suggests the potential usefulness of measuring antibody responses to antigen 85A and MDP1 for assessing the risk of TB progression.
alpha-crystallin like protein (also called HspX)
antigen 85 complex proteins (mycolyltransferases)
antigen 85B (also called alpha antigen)
area under the curve
Mycobacterium bovis bacillus Calmette-Guérin
10 kDa culture filtrate protein
dormancy survival regulator
sensor histidine kinase of DosR
6 kDa early secretory antigenic target of Mycobacterium tuberculosis
M. tuberculosis uninfected healthy controls
heat-stress-induced ribosome binding protein A (also called Acr2)
interferon gamma release assay
latent Mycobacterium tuberculosis infection
mycobacterial DNA-binding protein 1 (also called Mt-HLP, HupB, LBP)
- M. tuberculosis
purified protein derivative
receiver operating characteristic
tuberculin skin test
World Health Organization
The WHO reports that M. tuberculosis latently infects 30% of the world's population and that nearly 8.8 million new cases of TB and 1.1 million deaths occur worldwide .
The American Thoracic Society classifies M. tuberculosis-infected individuals as symptomatic and asymptomatic . They divide the asymptomatic group into those with past TB and those with LTBI. Past TB includes a history of active TB or abnormal stable radiographic findings of TB. LTBI denotes a positive TST or IGRA and no clinical and no radiographic evidence of active disease. Patients with both past TB and LTBI have no bacteriological evidence of the disease. Individuals with past TB are more than likely to harbor persisting M. tuberculosis because, once infection is established, human host immunity alone rarely eradicates this organism. Individuals with past TB not currently receiving medication have a higher risk of TB progression than do those with LTBI [3-5]. Identification of people at risk of TB progression is important in that it leads to provision of appropriate preventative medication. However, recent diagnostic methods, such as TST and IGRAs do not permit assessment of the risk of TB progression [4, 5].
As the WHO warns, currently available commercial kits for TB serodiagnosis and identification of symptomatic M. tuberculosis infection are of questionable value because of their broad and erratic results for samples from subjects with paucibacillary forms of TB. In contrast antibody responses, which correlate with bacterial burden, have the potential to track TB progression from asymptomatic infections . During stable asymptomatic infection or after vaccination with Mycobacterium bovis BCG, tests for antibody production are largely negative [6-8]. By contrast, antibody titers to M. tuberculosis antigens increase prior to TB progression [9-12].
Identification of specific antibodies in subjects with asymptomatic M. tuberculosis infection is an important step in identifying those at risk of TB progression. It is conceivable that there are heterogeneous populations of tubercle bacilli during dormancy and actively multiplying phases within granulomatous lesions at the pre-recurrence stage. In this study, we assessed antibody responses to 23 kinds of major M. tuberculosis proteins, including those expressed by growing and dormant bacilli [13-19] obtained from individuals with past TB who were not currently receiving TB chemotherapy. The aim of this study was to explore biomarker-targeted antibody responses in those who are at risk of TB progression.
MATERIALS AND METHODS
The following groups of individuals were enrolled (Table 1).
- A HC group of 17 students (ages 20–24 years, male/female ratio 9/8) from Osaka City University Medical School (Osaka, Japan). These participants were negative for TB according to chest X-ray films and immune-based assessment (TST and IGRAs), and had no known risk factors for M. tuberculosis infection such as HIV infection, close contact with infected individuals or suggestive chest X-ray findings.
- An active TB group of 15 individuals (aged 35–71 years, male/female ratio 13/2) whose diagnoses of active TB were based on microbiologic examination of sputum specimens yielding either positive cultures for M. tuberculosis or positive DNA amplification tests specific for M. tuberculosis (TRC Test; TRCRapid-160, Tosoh, Tokyo, Japan). All members of this group had positive IGRAs.
- A past TB group of 15 patients with definitive histories of pulmonary TB more than 5 years previously. Their sputum samples were negative on culture and nucleic acid amplification M. tuberculosis tests. Their chest X-ray films showed sclerotic lesions and stable cavities. The cavitary lesions indicated radiographic diagnoses of TB because they had no surrounding infiltrating shadows. In this group, 33% of individuals were IGRA positive.
|HC||Active TB||Past TB|
|Number of participants||17||15||15|
|Age, mean (years) ± SD||21.82 ± 1.13||255.93 ± 11.25||69.40 ± 12.73|
|Age range (years)||20–24||35–71||42–91|
|IGRA positive (%)||0||100||33.33|
Subjects were excluded from this study when disease due to nontuberculous mycobacteria was confirmed by repeated cultures and the findings satisfied the American Thoracic Society guidelines . The serum specimens were assayed without knowledge of the patients' clinical characteristics. This study was approved by the Research and Ethical Committees of the National Toneyama Hospital and Osaka City University Graduate School of Medicine and informed consent was obtained from all subjects.
Materials and reagents
pET-21b and Bugbuster HT were obtained from Novagen (Darmstadt, Germany), Escherichia coli BL21 (DE3) cells from Toyobo (Osaka, Japan), Lowenstein–Jensen Luria-Bertani medium and carbenicillin from Sigma (St. Louis, MO, USA), isopropyl-1-thio-beta-d-galactopyranoside and Ni-NTA agarose from Qiagen (Gaithersburg, MD, USA), skimmed milk from Morinaga (Tokyo, Japan), horseradish peroxidase-conjugated anti-human IgG, IgA, or IgM antibodies and Envision kits from Dako (Carpinteria, CA, USA), sureBlue reserve TMB microwell peroxidase substrate from KPL (Gaithersburg, MD, USA) and monoclonal Acr antibody from HyTest (Turku, Finland).
Recombinant protein preparation
A pET-21b-based expression vector containing the full coding sequence for HBHA (Rv0475) was generated by a PCR-based approach and maintained in E. coli BL21 (DE3) cells. The DNA sequence encoding HBHA was amplified from M. tuberculosis H37Rv DNA using appropriate primers and cloned into the NdeI and HindIII sites of pET-21b. The vectors expressing Acr, HrpA (Rv0251c), ESAT-6 (Rv3875), CFP10 (Rv3874), Ag85A (Rv3804c) and Ag85B (Rv1886c) were produced similarly by a PCR-based approach using a bacterial chromosome. Each PCR product containing coding regions was designed to allow expression of C-terminal, 6 × histidine-tagged variants of the recombinant proteins following ligation into pET-21b. After construction, expression vectors were confirmed by DNA sequencing. Recombinant M. tuberculosis proteins were purified by utilizing a Ni-NTA column (1 mL bed volume, GE Healthcare Bio-Science, Piscataway, NJ, USA) according to the manufacturer's instruction.
Enzyme-linked immunosorbent assay
Concentrations of IgG, IgA and IgM antibodies were determined by ELISA using recombinant proteins. Ninety-well microplates (Sumilon Type H, LMS, Tokyo, Japan) were coated with each recombinant antigen in bicarbonate buffer, pH 9.6 (0.5 µg/well) overnight at 4°C. The plates were blocked with PBS containing 0.05% Tween 20 and 5% skimmed milk for 12 hr at 4°C and washed four times with PBS containing 0.05% Tween 20. The plates were then washed and human serum samples diluted 1:100 in PBS containing 0.05% Tween 20 and 0.5% skimmed milk were incubated for 12 hr at 4°C. After washing the wells, HRP-conjugated anti-human IgG, IgA, or IgM antibodies were added at a 1:5000 dilution. After 1 hr incubation at 37°C, the plates were washed four times before 100 µL of SureBlue reserve-TMB was added to each well. The reactions were stopped after 3 min by adding 50 µL of 0.1 M HCl and the plates read at 450 nm using a Multiskan (Thermo Fisher Scientific K.K., Yokohama, Japan).
Histopathology of granulomatous lesions
Paraffin-embedded lung sections were stained with monoclonal anti-Acr (1:500), polyclonal anti-MDP1 (1:500) and polyclonal anti-Ag85 (1:1000) according to the manufacturer's instructions and examined with the Dako EnVision system.
Optical density differences between study groups were determined using the Mann–Whitney's U-test. ROC curve analysis and the AUC with 95% s CI for each antigen were calculated with the GraphPad Prism version 5.0 (GraphPad Software, San Diego, CA, USA). In all analyses, P < 0.05 was considered statistically significant.
Concentrations of serum antibodies against 6 kDa early secretory antigenic target of Mycobacterium tuberculosis and 10 kDa culture filtrate protein
We examined the serum concentrations of antibodies against ESAT-6 and CFP10, which are produced from M. tuberculosis during the growth phase, by ELISA. We found that the concentrations of IgG-class antibodies to ESAT-6 and CFP10 were significantly higher in the active TB group than in the HC group (P < 0.05, Fig. 1). In addition, the concentration of IgG against CFP10 was significantly higher in the active TB group than in the past TB group (P < 0.05).
Concentrations of antibodies against 16 proteins encoded by the dormancy survival regulator (Rv3133c) region of Mycobacterium tuberculosis
In order to clarify antibody responses to latency-associated antigens of M. tuberculosis, we next examined the levels of serum antibodies against 16 proteins encoded by the DosR regulon, namely Rv0574, Rv0079, Rv1998, Rv2005, Rv2029, Rv2030, Rv2031c (Acr), Rv2032, Rv2623, Rv2624, Rv2628, Rv2629, Rv3127, Rv3129, Rv3132 (DosS), and Rv3134. With the exception of Acr and DosS, levels of IgG antibodies to DosR regulon-encoded proteins were low or absent in the three groups (Fig. S1, Fig. 2a, b). The concentrations of IgG antibody to Acr were significantly higher in the active TB and past TB groups than in the HC group. However, there were no differences between the past and active TB groups in concentrations of IgG to either Acr or DosS. In contrast to IgG antibodies, IgM and IgA antibody responses to the proteins were below measurable limits in all the tested individuals (data not shown).
Concentrations of antibodies to non-dormancy survival regulator regulon proteins
We next examined concentrations of IgG to immunogenic M. tuberculosis proteins other than ESAT-6, CFP10, and DosR-regulon proteins. Although the concentrations of IgG to HBHA and HrpA were higher in the past TB group than in the active TB group, we found no difference between the past TB and HC groups (data not shown). In contrast, the concentrations of IgG antibody against Ag85 and MDP1 were higher in the past TB group than in the HC and active TB groups (P < 0.01, Fig. 2c–e).
Sensitivity of immunoglobulin G response to tuberculosis status
Based on these data, we performed ROC analyses on data from the past TB and HC groups (Fig. 3 and Table 2). Acr, Ag85A, and MDP1 produced ROC curves acceptable for diagnostics with AUCs of 0.992 (95% CI 0.972–1.013, P < 0.0001), 0.965 (95% CI 0.910–1.019, P < 0.0001), and 0.973 (95% CI 0.925–1.024, P < 0.0001), respectively. Moreover, the AUCs for Ag85A and MDP1 were 0.809 (95% CI 0.642–0.976, P = 0.0040) and 0.858 (95% CI 0.720–0.995, P = 0.0008), respectively, in a ROC analysis between the past TB and active TB groups.
|HC (n = 17)||Active TB (n = 15)||Past TB (n = 15)||AUC||95% CI||P-value|
|ESAT-6||0.137 ± 0.094||0.297 ± 0.325||0.210 ± 0.065||0.782||0.62–0.95||0.0066|
|CFP10||0.077 ± 0.096||0.662 ± 0.836||0.151 ± 0.305||0.702||0.52–0.88||0.0519|
|Acr||0.147 ± 0.058||0.989 ± 1.419||0.333 ± 0.145||0.992||0.97–1.01||< 0.0001|
|DosS||0.090 ± 0.052||0.136 ± 0.072||0.144 ± 0.058||0.773||0.61–0.94||0.0087|
|Ag85A||0.984 ± 0.570||1.899 ± 0.928||2.918 ± 0.839||0.965||0.91–1.02||< 0.0001|
|Ag85B||0.836 ± 0.437||1.110 ± 0.630||1.611 ± 0.497||0.886||0.77–1.00||0.0002|
|HBHA||0.890 ± 0.207||0.485 ± 0.388||0.923 ± 0.347||0.526||0.31–0.74||0.8061|
|HrpA||0.260 ± 0.304||0.172 ± 0.122||0.341 ± 0.205||0.651||0.46–0.85||0.1460|
|MDP1||0.112 ± 0.145||0.269 ± 0.141||0.509 ± 0.238||0.973||0.92–1.02||< 0.0001|
Table 3 summarizes the sensitivities to antigens in each group. These data show that IgG responses to CFP10 and Acr occurred more often in the active TB than the past TB or HC groups. The sensitivities for CFP10 and Acr in the active TB group were 53.3% and 60.0%, respectively. By contrast, the sensitivities of IgG responses to Ag85A (86.7%) and MDP1 (60.0%) were greater in the past TB than in the HC groups.
|Antigens||HC (n = 17)||Active TB (n = 15)||Past TB (n = 15)|
|ESAT-6||0.0% (1)||20.0% (3)||6.7% (1)|
|CFP10||6.7% (1)||53.3% (8)||33.3% (5)|
|Acr||0.0% (1)||60.0% (9)||60.0% (9)|
|DosS||0.0% (1)||6.7% (1)||13.3% (2)|
|Ag85A||6.7% (1)||33.3% (5)||86.7% (13)|
|Ag85B||0.0% (1)||13.3% (2)||33.3% (5)|
|HBHA||6.7% (1)||6.7% (1)||20.0% (3)|
|HrpA||0.0% (1)||0.0% (1)||0.0% (1)|
|MDP1||0.0% (1)||13.3% (2)||60.0% (9)|
Colocalization of antigen 85 complex proteins and mycobacterial DNA-binding protein 1 with Mycobacterium tuberculosis in human pulmonary granulomatous lesions
In order to confirm expression of Ag85 and MDP1 in persistent M. tuberculosis, we histopathologically examined a biopsy of a solitary pulmonary coin lesion that was visible in chest X-ray films of a patient who was negative for sputum bacteriology (both smear and culture). Microbiologic examination of this biopsied specimen showed M. tuberculosis infection in a tuberculoma. We stained the biopsy section with HE (Fig. 4a), Ziehl–Neelsen (Fig. 4b, c), and antibodies to control (Fig. 4d), Ag85 (Fig. 4e) and MDP1 (Fig. 4f). The HE stain showed typical granuloma formation with central caseation (Fig. 4a). Ziehl–Neelsen staining revealed acid fast bacilli located in the center of the granuloma (Fig. 4b, c). The lesion was positive for antibodies against Ag85 and MDP1 in the same areas as were positive for Ziehl–Neelsen. Thus, for the first time we demonstrated colocalization of Ag85, MDP1 and persistent M. tuberculosis bacilli.
Asymptomatic M. tuberculosis infection represents a large pathogen pool that can cause reactivation. Therefore, identifying those at risk of TB progression is an important challenge for successful control of TB. Although the accuracy of serodiagnosis of active TB is controversial, several lines of evidence suggest it has the potential to track disease progression. For example, several studies have reported higher antibody concentrations in sputum smear-positive than in smear-negative subjects [6, 21] and increased antibody production when disease progresses in both macaques  and HIV-infected human cohorts [9-12]. Since antibody titers and targets differ from each other [22, 23], identification of several antigens that are targeted by antibodies in people who are at risk of TB progression is the first critical step. Because people with radiographic evidence of past or spontaneously cured TB have a higher risk of disease progression than LTBI alone, we selected subjects with past TB for this study and examined their concentrations of antibodies against M. tuberculosis proteins.
We found humoral antibody (IgG) responses to the major pathogenic proteins of M. tuberculosis, ESAT-6 and CFP10 (which are produced in the growth phase) in 20.0% and 53.3%, respectively, of individuals in the active TB group, but in only 6.7% and 33.3%, respectively, of those in the past TB group (Fig. 1 and Table 3). Indeed, only 33.3% of the past TB group was positive for IGRAs using these antigens. This suggests that both humoral and cell-mediated immune responses to ESAT-6 and CFP10 are low in the past TB group, demonstrating the difficulty of diagnosing asymptomatic infection, which has potential to reactivate.
The DosR is considered a key regulator of adaptation of M. tuberculosis to hypoxic conditions [24-26]. T cells of asymptomatic subjects with LTBI DosR reportedly preferentially recognize regulon-encoded proteins [14, 27]. With the exception of Acr and DosS, this study has shown that antibody concentrations are low in patients with past TB group. Possible explanations for the present data are low antigenicity or degrees of expression of DosR regulon-encoded proteins other than Acr and DosS. Another possible explanation is the location of DosR regulon encoded proteins; because antigens encoded by the DosR regulon are not secreted, humoral immune responses to such intracellular antigens cannot be evoked.
In this study, we have identified two candidate antigens, Ag85A and MDP1, for identifying those at risk of TB progression. Ag85A belongs to the Ag85 complex, which consists of three member proteins designated Ag85A, Ag85B, and Ag85C. Ag85 genes are encoded by three genes located at different sites in the M. tuberculosis genome and show cross-reactivity as well as homology at amino acid levels . Amino acid homology between Ag85A and Ag85B is 82%. The Ag85 complex has mycolyl transferase enzymatic activity, which translocates mycolic acids of trehalose 6-monomycolate to free sugars, such as trehalose . Recent studies show that dynamic remodeling of mycolic acid-containing glycolipids in the cell wall occurs in vivo, this process being mediated by the broad substrate specificity of Ag85 [29, 30]. Importantly, immune responses to glycerol monomycolate reportedly occur specifically in people with LTBI or in BCG-vaccinated individuals but not in patients with active TB . Thus, Ag85-dependent exchange of trehalose 6-monomycolate to glycerol monomycolate is likely to occur in LTBI and BCG-vaccinated sites. Collectively, these reports suggest that the detection of immune responses to Ag85 in this study is not surprising and that this protein may be useful for detecting asymptomatic M. tuberculosis infection at risk of TB progression.
Mycobacterial DNA-binding protein 1 has pleiotropic functions and is an essential protein for survival of M. tuberculosis. It influences various biological functions of mycobacteria, from the growth phase to the hypoxic dormant state [32, 33]. The present data suggest the possibility of enhanced expression of this protein in latent rather than in the active disease states (Figs. 2, 3). MDP1 suppresses mycobacterial multiplication by controlling gene expression [33, 34] and also has ferritin-like activity controlling iron homeostasis . Iron is essential for the survival and multiplication of M. tuberculosis inside macrophages, whereas the host utilizes iron to generate bactericidal reactive oxygen intermediates. MDP1 stores iron inside bacteria and prevents iron-dependent generation of oxygen radicals . Such multifunctional activities of this protein are likely involved in downshifting the growth rate and long-term persistence of M. tuberculosis in the latent state, thus accounting for the definite IgG production in the past TB group in the present study.
We thank Ms. Sara Matsumoto and Inori Omura, for assistance with the experiments and their heartfelt encouragement. We would like to thank Mr. Hiroo Yamamoto and Ms. Junko Koizumi (Osaka City University Medical School) for their technical assistance. This work was supported by grants from the Ministry of Education, Culture, Sports, Science, and Technology, Ministry of Health, Labor and Welfare (Research on Emerging and Re-emerging Infectious Diseases, Health Sciences Research Grants), and Japan Health Sciences Foundation.
The authors declare no financial or commercial conflict of interest.