Early and accurate diagnosis of leprosy and MDT are considered essential to disrupt M. leprae transmission and leprosy incidence. Due to the complex and varying immune responses that characterize leprosy spectrum, it is likely that immune diagnosis will only be achieved by antigens that specifically induce cellular and humoral responses [8, 19]. Identifying antigens that are targets of the proinflammatory T cell response could also reveal those that are associated with limiting M. leprae burden and therefore useful for vaccination. In this study a panel of recombinant M. leprae proteins was evaluated for the ability to induce CMI and antibody responses to identify antigens that are specifically reactive in leprosy patients. Although many antigens were identified as cellular targets by their ability to induce IFNγ secretion in WBA of TT/BT leprosy patients and HHC, a minority of the screened proteins induced serum IgG responses in BL/LL patients, such that few proteins induced specific responses across the leprosy spectrum.
In silico analysis predicted the presence of T cell and B cell epitopes within all of the M. leprae proteins tested in this study. The most promiscuous T cell epitope predicted was located in ML0405 protein, which was estimated to contain T cell epitopes recognized by 50 of 51 of HLA-DR alleles along with 24 potential B cell epitopes. Consistent with this prediction, the ML0405 protein was recognized by IFNγ release in WBA using TT/BT leprosy patient blood and by serum IgG from LL/BL leprosy patients. However, despite the prediction that all of the evaluated proteins would possess multiple HLA-DR epitopes, more than half (17 of 33; 52%) of the proteins were not recognized by leprosy patients or controls. Our results are in agreement with previous studies that found some of these proteins were not immunogenic [4, 9, 19–25]. Although we cannot exclude the possibility that the proteins investigated were not recognized because individuals with the appropriate HLA were not recruited, we consider this improbable given the large degree of promiscuity that was predicted. Several other possibilities could explain the lack of reactivity. Although we consider it unlikely given the recognition of positive responses to many proteins, the simplest explanation would be that recombinant expression in E. coli leads to significantly different folding and processing of proteins than occurs during native expression in M. leprae. A more likely possibility is that these proteins, although present in the genome of M. leprae, may not be translated . Another explanation could be that limited antigen-presentation or T cell suppression may occur during leprosy in order to prevent nerve damage via T cell-mediated killing of M. leprae -infected Schwann cells. We are aware that the softwares employed in this study identify and predict HLA binding regions from antigen sequences without predicting interactions with the TCR. Therefore the use of other in silico prediction softwares that also include the probability of the antigen being processed, presented in the context of a certain HLA allele and recognized by TCR could lead to different conlusions.
In leprosy endemic regions the high rates of concomitant exposure to M. tuberculosis and to other non-pathogenic environmental mycobacteria could confound the determination of M. leprae -specific immune responses. Several of the M. leprae antigens tested in this study have homologues in other mycobacteria with greater than 50% identity (Table 1). The routine M. bovis BCG vaccination of newborns in leprosy-endemic countries such as Brazil could stimulate responses that could cross-react with M. leprae. M. bovis BCG showed high homology with immunogenic M. leprae proteins (from 62 to 88%) This high homology was not, however, associated with neither cross -reactivity nor with a lack of specificity. For example, the ML2331 protein possesses over 80% identity with M. tuberculosis and M. bovis BCG proteins but serological and cellular responses to this antigen were highly specific to leprosy patients (or at-risk contacts). Antigen expression levels and bioavailability may determine the immune dominant antigens of each mycobacterial infection, such that differing responses to similar proteins may distinguish each disease. As a further demonstration of the limited ability of in silico predictions to indicate specificity, despite having no identified homologue in M. tuberculosis, the ML2346 protein induced strong IFNγ responses in all of the study groups, including TB patients. Based on these findings we suggest that, at present, in silico identification of M. leprae proteins with high identity to other mycobacterial proteins should not be used as a definitive criterion to exclude them from further diagnostic or vaccine evaluations.
Several studies have reported antigen-specific cellular and antibody responses during leprosy, but few studies have consolidated data to determine if particular antigens are differentially recognized across the disease spectrum [3, 4, 9, 19–25]. Araoz et al. described M. leprae recombinant proteins recognized by antibodies produced by lepromatous patients that that could also induce CMI among tuberculoid patients . Some of the proteins analyzed in our study (ML1632, ML1685 and ML1556) induced strong and specific production of IFNγ in TT/BT and HHC groups, demonstrating their potential application for identification of those at risk of developing tuberculoid leprosy. These proteins were not, however, recognized by antibodies from LL/BL patients. Our parallel screening suggested that the antibody response to M. leprae recombinant proteins was dependent upon their ability to induce cellular responses, and indicates only a limited number of M. leprae antigens contained T cell and B cell epitopes that are immune reactive in the context of disease (ML0405, ML2055 and ML2331). These antigens could be considered priority diagnostic antigens. In order to optimize the screening of new M. leprae antigens for leprosy diagnosis, our data suggest that it may be beneficial to conduct WBA testing before serological assays. If the protein is not able to induce a T cell response it could be de-prioritized from further testing.