Tuberculosis (TB) remains one of the main threats to mankind. Despite the efforts of the scientific community toward understanding this disease, ~9 million active TB cases and 2 million deaths occur every year (Dye et al., 1999
; Dye, 2006
). Although immune cell activation is required to limit the growth of Mycobacterium tuberculosis
(Mtb), if uncontrolled it can damage host tissue. The balance between protection and pathological consequences is the crux of TB pathogenesis. It is therefore crucial to understand and discriminate the components of the immune response that are protective from those that are damaging to effectively intervene in TB. Indeed, the hope for improved vaccination relies on the possibility that these responses can be independently manipulated.
The discrimination between protective and pathological components of the immune response during TB is particularly important because M. bovis
bacille Calmette-Guérin (BCG), currently the only available vaccine against TB, has variable efficacy (ranging from 0 to 80%; Colditz et al., 1994
) and its protection may last for only 10 yr (Sterne et al., 1998
; Weir et al., 2008
). As a result of this apparent loss of activity, revaccination with BCG, DNA vaccines, or subunit vaccines is considered as a potential control strategy; however, the safety of this strategy in a highly exposed population is not yet fully defined. The issue of safety arises because of the fact that repeated exposure of Mtb-infected animals to mycobacterial antigens can initiate immune-mediated pathology. Specifically, the exposure of Mtb-infected guinea pigs to either live mycobacteria or mycobacterial antigens results in necrotic inflammation at the site of challenge, a response that has been termed the Koch phenomenon (Koch, 1891
). Although the cellular and molecular determinants of this response are undefined, it appears that the severity of the Koch phenomenon depends on the dose of antigen, as lower doses of antigen are used to induce a delayed-type hypersensitivity response; this is the basis for the current skin test for detecting latently infected individuals (Rich, 1944
). A damaging focal response at the site of initial infection can also be triggered by repeated vaccine (BCG or DNA) challenge of Mtb-infected animals. This leads to the development of severe pathology, including necrosis and increased granulocyte influx in preexisting lesions in the lung (Turner et al., 2000
; Moreira et al., 2002
; Taylor et al., 2003
). The perceived risk of increased pathological consequences as a result of vaccination in previously exposed humans has lead to initial safety screens being performed on novel vaccine candidates (Sander et al., 2009
). However, results from these types of studies should be interpreted cautiously, as the extent of repeated antigen exposure will differ greatly depending on the level of disease in the community. Therefore, there is a concern that adult postexposure vaccination to prevent the reactivation of TB could lead to pathology, particularly in highly exposed populations.
The importance of the cytokine IFN-γ in the protective response to Mtb is well established (North and Young, 2004
); however, the mediator of pathological responses has not been identified. The Koch-like pathologies described in the previous paragraph were consistently associated with neutrophil influx and could therefore be mediated by IL-17 (Miyamoto et al., 2003
; Kolls and Linden, 2004
). Further, although IL-17 is induced during mycobacterial infection, it does not play a significant role during the early period of infection (up to 100 d; Khader and Cooper, 2008
). In addition, IFN-γ is able to regulate the IL-17 response during BCG infection (Cruz et al., 2006
), and in the absence of IFN-γ signaling in the stroma, an increase in neutrophil involvement in the TB granuloma is seen (DesVignes and Ernst, 2009
); these data reflect an important regulatory activity for IFN-γ in the control of pathology in TB. Therefore, we hypothesized that repeated antigen exposure would allow the IL-17 response to overcome the IFN-γ–mediated regulation and thereby mediate immunopathological consequences, and that the Koch phenomenon may result from an unregulated IL-17 response.
To test our hypothesis and investigate the mechanisms underlying the pathogenic response to repeated antigen exposure, we examined the pathological cellular response in an established mouse model of repeated antigen exposure that leads to increased pathology (Turner et al., 2000
; Moreira et al., 2002
; Taylor et al., 2003
). Specifically, Mtb-infected mice were repeatedly vaccinated subcutaneously with BCG to promote an immunopathologic response. We found that enhanced pathology was associated with a dramatic increase in the number of antigen-specific IL-17–producing cells in the lungs of infected and revaccinated animals. Importantly, in the absence of IL-23 or in the presence of anti–IL-17 antibody, the enhanced pathological response was ablated, along with the granulocyte/neutrophil infiltrate. These results support a pathogenic role of IL-17–producing antigen-specific cells during TB. Thus, the development of Th17 responses after vaccination and their impact on chronic infection should be considered.