Differential in vivo expression of mycobacterial antigens in Mycobacterium tuberculosisinfected lungs and lymph node tissues

Infection with Mycobacterium tuberculosis represents a major disease burden globally. There are about 8,7 million new cases of tuberculosis (TB) worldwide and at least 1.4 million deaths per year [1]. Extra pulmonary TB accounts for approximately one-fifth of TB cases among immune-competent individuals and up to one-half in HIV-infected individuals [1‐4]. Though any organ in the body can be involved, lymph nodes are the most common form of extra-pulmonary TB in both adults and children [5, 6]. Immune responses in TB differ between various disease sites and in various forms of disease [7, 8]. TB lymphadenitis is typically a self-contained, while pulmonary TB is usually a rapidly progressive and fatal disease without adequate chemotherapy.

In humans, granuloma formation in response to M. tuberculosis infection is essential for control of mycobacterial infections [9]. Paradoxically, granulomatous inflammation is also associated with the typical immunopathology, tissue damage and symptoms seen in TB. It is believed that mycobacterial antigens are continuously released into the infected tissue and could be responsible for the formation and persistence of lesions [10‐13]. Antigens secreted into the extracellular environment by M. tuberculosis are thought to be immunodominant and to be involved in inducing protective immunity, and in this sense also considered to be the most important antigens during infection [14, 15]. Little is known about the identity of the M. tuberculosis-antigens that are expressed in vivo in the lesions. TB is a chronic and multi-organ disease, and the expression profile of antigens in the infected tissues may be different during various phases of infection and in various organs, making it important to study the differential in vivo-expressed antigens in various organs with different disease manifestation.

Previous work has identified and purified major secretory and somatic M. tuberculosis antigens to near homogeneity from the culture filtrates and bacterial sonicates from M. tuberculosis cultured in wholly synthetic liquid media [16‐18]. Functional rabbit polyclonal antibodies have been generated against these antigens [17, 19, 20]. This collection of antibodies was used to study the in situ M. tuberculosis antigens expression in the infected host tissues. We have previously performed several studies on MPT64 for the development of immunohistochemical method using patient biopsies and aspirate samples to diagnose extrapulmonary TB [21‐24]. In this study, we have focused on exploring the expression pattern of several additional major secreted (MPT32, MPT44, MPT46, MPT51, MPT53, MPT59, MPT63, and MPT64) and somatic mycobacterial antigens (Mce1A, Hsp65, and MPT57) and studied the differential expression in the host tissues during advanced rapidly progressing and fatal pulmonary disease with mainly pneumonic infiltrates, and compared this to the presence of the same antigens in TB lymphadenitis cases, which is mainly chronic and self-limiting disease with organised granulomas.

In this study, we analysed the expression of mycobacterial antigens in infected tissues.

One interesting observation was the differential staining pattern of secreted antigens in the lungs and lymph node lesions. While these antigens were found in high amounts in the pulmonary lesions, their level of staining was very low, or absent, in the lymph node granulomas. This contrasted the findings for somatic antigens, which were found with the same ratio in both types of lesions. The pulmonary lesions exhibited pneumonic infiltrates without proper necrotic granuloma formation, while the lymph node lesions had organised necrotic granulomas. Granuloma formation is known to be a correlate of an effective immune response resulting in control of bacillary multiplication and containment of infection. Relative absence of secreted mycobacterial antigens in the well-formed granulomas indicate that immune response is effective to bring bacillary multiplication under control, while the presence of somatic antigens imply their involvement in the persistence of infection.

We could not detect bacilli by acid-fast staining in the lymph node tissues, which might indicate that the number of intact bacilli is below a detectable level. However, it has been shown that mycobacteria tend to lose of their acid-fastness as they adapt to the stressful intracellular milieu within a granuloma [26‐28]. This can provide an alternative explanation to why no bacilli could be detected in the lymph nodes using this technique. The mycobacterial response to stressful conditions is multifactorial. Other observed change is the down-regulation of a number of secreted antigens that are considered to be highly expressed during normal culture conditions, including Ag85A-C, MPT63 and MPT64 [29‐31]. As the bacteria adapt to persistence and long-term survival within cells of the granuloma, many of the antigens required during, and immediately after, cell invasion, are no longer important for bacterial survival. This could explain why most of the investigated secreted antigens are nearly absent in the lymph nodes, while the somatic antigens were equally detected in both lung and lymph node tissues. Serodiagnostic studies have evaluated specificities of antibodies from sera of TB patients against different mycobacterial antigens including secreted-, membrane-, cell wall associated-, and cytosolic antigens. Although secreted antigens have been most commonly studied, it is the surface exposed components that appear to be most frequently recognised by patient sera [32, 33]. A study that described a lack of host immune response towards the secreted mycobacterial antigen MPT53 in humans and guinea pigs [34], lends support to the theory of a reduced presence of secreted antigens in vivo during stable disease. The question remains why the major secreted antigens in some cases were only detected in the necrotic centres and not in the granuloma cells. One possible explanation is that the discrete spots of staining in the necrotic centres represent cells that harbour multiplying bacilli that produce secreted antigens, but that these cells are gradually removed by cell lysis and necrosis.

There was a notable exception to the almost complete absence of staining of secreted antigens in the lymphadenitis tissues; the antigen MPT64 was identified in all granulomas. The explanation for this observation could be that MPT64 is not cleared away from the tissues and therefore accumulates in the granulomas over time. Structural analyses of this protein, suggest that it harbours a β-grasp motif that can form strong protein-protein interactions [35, 36]. We have previously studied the staining pattern of MPT64 in tuberculous pleural biopsies in HIV co-infected TB patients, where only about half of the biopsies had organised granulomatous inflammation. MPT64 antigen was more frequently detected in the granulomas than in the non-granulomatous inflammation (93% vs. 64%), while mycobacterial DNA was detected more frequently in the non-granulomatous inflammation as compared to granulomas [21]. This antigen therefore seem to preferentially accumulate within granuloma structures, even though the cells of the granuloma are efficiently restricting bacterial growth. The persistence of MPT64 in the tissues suggests a role in bacterial pathogenesis.

In pulmonary lesions without proper granuloma formation, the ratio of cells that stained positive for the somatic antigen Mce1A was lower, and the staining intensity was weaker, compared to the staining pattern of secreted antigens. Furthermore, the staining pattern of Mce1A and secreted antigens did not always overlap; certain regions were not stained by anti-Mce1A, despite being highly stained by antibodies against secreted antigens. In lymph nodes with organised granulomas, Mce1A was detected in all cases, located mainly intracellularly in the granuloma cells rather than necrotic centres as seen with some secreted antigens. The products of the mce1 operon have been shown to be important for granuloma formation [37‐39]. In this study a differential staining pattern of Mce1A in lungs with excessive pathology, but without the typical organised granulomas as compared to lymph node tissue, lend support to the theory of an involvement of this antigen in the formation of organised and protective granulomas.

This study shows that major secreted mycobacterial antigens were found in high amounts in the pulmonary lesions that exhibited pneumonic infiltrates without proper necrotic granuloma formation, while their level of staining was very low, or absent, in the lymph node TB lesions with organised necrotic granulomas and very low bacillary load. One exception was MPT64, which was detected in both pulmonary and lymph node lesions. This indicates a preferential accumulation of this antigen within granuloma structures even if the cells of the granuloma can efficiently restrict bacterial growth, and suggests a role in the persistence of chronic infection. All three somatic antigens were detectable in both pulmonary and lymphadenitis lesions, and their expression pattern was different from that of the secreted antigens. The findings in this study have implications for the development of new diagnostic tools.