The results reported in this paper show that an IFN-γ+ T cell immune response to PPE44 can be detected by ELISpot in all healthy individuals naturally PPD+ and, to a lower extent, in subjects vaccinated with BCG; CD4+ T lymphocytes account for IFN-γ secretion in PPE44-responder subjects, as shown by ICC analysis. By the same approaches, our study has highlighted the presence of a strong CD4+ T-cell epitope in the NH2-terminus of the PPE44 molecule localized at the aa position 1-20. Conversely, no significant IFN-γ+ CD4+ T cell response to PPE44 or its immunodominant peptide p1L could be detected in most patients (7 out of 8) with newly diagnosed active TB.
The PPE44 immunodominat T-cell epitope detected in the present study has been previously reported as the antigenic target of an IL-2-induced IFN-γ+ response in mice in which immunization with PPE44-subunit vaccines conferred protective immunity in an experimental model of TB . The data reported in this paper suggest that IFN-γ+ T-cell responses to PPE44 may be associated to immune protection also in human M. tuberculosis infection: indeed, IFN-γ+ T-cells specific for the immunodominant PPE44 peptide p1L were detectable in all individuals whose immune system is likely to have determined the containment of infection and prevented progression to active TB disease (PPD+ healthy subjects), as well as in a proportion of BCG-vaccinated subjects. On the other hand, most patients with active TB, i.e., those individuals whose immune system failed to contain TB infection, did not respond to PPE44 or p1L. In this respect, however, it has to be considered that TB patients enrolled in our study were under TB chemotherapy, which might have decreased the M. tuberculosis-specific IFN-γ responses [12, 13]; another explanation might be that PPE44-specific T cells are sequestered at the site of mycobacterial replication, usually the lung. Alternatively, it is tempting to speculate that the poor T-cell immune responsiveness of TB patients to p1L might be related to a dynamic antigen display due to differential expression of PPE44 during active infection; indeed, in a previous paper, we reported great variations in the expression of the gene coding for PPE44 among M. tuberculosis isolates and that only about one third of patients with active TB produced antibodies to PPE44 . A last attractive hypothesis could be that a T cell response to p1L/PPE44 helpes individuals to contain TB infection, while those who do not mount such a response are more prone to develop active disease.
One of the promising features of p1L is that it was recognized by all 5 PPD+ healthy individuals tested, as shown by ELISpot, suggesting that p1L is most probably able to bind a number of human HLA-DR alleles. It also proved to be immunodominant in two different species, being a T-cell epitope also in the C57BL/6 strain of mice . "Promiscuous" helper peptides are peptides that can bind a wide range of MHC class II alleles. Within their sequence, they typically have a motif, called P1-P6, where position 1 can be an aromatic or a hydrophobic aa whereas position 6 can be a small or hydrophobic aa . Indeed, such motif can be found in 3 positions in p1L, namely 1-6, 3-8 and 10-6. Promiscuous peptides have been searched for and described both in mycobacterial antigens  and in other antigens, such as the malarial circumsporozoite protein . They allow to overcome the problem of the high degree of polymorphism of the HLA-DR molecules expressed in the human population and for such a reason they are ideal candidates for subunit vaccine design and as diagnostic tools. To this aim, future studies will attempt to establish the HLA class II restriction elements binding p1L.
Two other PPE proteins of M. tuberculosis have proven capable of inducing protection against M. tuberculosis in experimental models, namely i) the PPE14 (Rv0915c/Mtb41), that has shown promising vaccine potential in human clinical trials , and ii) the PPE18 (Mtb39A/Rv1196), that is a component of the subunit vaccine Mtb72F. The latter has recently been investigated in clinical trials showing good tolerability and immunogenicity in humans [19, 20]. Dillon et al.  have reported proliferative response towards aa 1-20 of PPE18 in PBMC from PPD-positive human subjects, that is exactly the PPE region were our studies have mapped the CD4+ T-cell epitope. Indeed, the immunodominant p1L domain shows 60 to 85% aa homology with the corresponding sequences of 30 PPE proteins of M. tuberculosis and, in particular, p1L shares 14 identical aa with the NH2-terminal 20-aa sequence of the protective antigens PPE18 and PPE14. These observations raise the possibility that cross-reactivity might have contributed to the strong immunogenicity of the conserved and homologous NH2-terminal regions of the PPE proteins. These considerations make PPE proteins, especially their immunodominant NH2-terminal domains, promising antigen candidates for TB subunit vaccine development.
Latent TB infection is conventionally screened for by the more-than-100-year-old tuberculin skin test, that measures in vivo reactivity to tuberculin or PPD, a mixture of mycobacterial antigens, some of which common to non-tuberculous mycobacteria and to the vaccine strain M. bovis BCG. Recently, assays based on release of IFN-γ by PBMC exposed in vitro to M. tuberculosis-specific antigens, such as ESAT-6 and CFP-10, have emerged as attractive specific alternatives to tuberculin skin test to distinguish between M. tuberculosis infection and BCG vaccination/reactivity to non-tuberculous mycobacteria [22, 23]. However, the sensitivity of both tuberculin skin test and IFN-γ-release assays is suboptimal, and none of these tests distinguish between latent infection and active disease . In this context, PPE44 might turn out as a useful reagent for the immunological diagnosis of latent TB and p1L could prove even more useful than the whole recombinant protein becauseT cell reactivity, especially in thawed PBMC, has often been reported to be higher towards synthetic peptides than to recombinant proteins . Our data indicate that a PPE44- or p1L-specific IFN-γ+ T cell-response occurs in naturally PPD+ individuals, who are likely to harbour latent TB infection, and in a proportion of BCG vaccinees tested, but it is not detectable in most of our patients with active TB. These results, although very preliminary, would make p1L a good candidate, in association with the other TB-specific antigens available, to distinguish between latent infection and active disease.