Our study demonstrate that it is possible to engender a robust mycobacterium-specific immune response in the lungs by i.n. instillation of a single moderate dose of BCG vaccine, including strong type 1 (IFN-γ and IL-2) lymphocyte responses and macrophage activation (as measured by increased NO release). Our results also demonstrate that i.n. BCG vaccination induces specific T-cell responses in the CLN, which drains the NALT and nasal cavity, while s.c. delivery produces responses almost exclusively in the draining ILN. These observations corroborate the tenet that the immune system is compartmentalized and the early immunological responses tend to be strongest in compartments most proximal to the site of vaccine administration.
Interestingly, the frequency of antigen-specific cytokine-secreting cells found in the lymph nodes (CLNs and ILNs), a secondary lymphoid organ where immune responses develop, was low compared to that found in a nonlymphoid tissue such as the lung. This may be explained by the propensity of effector and memory T-cells to migrate out of LNs and preferentially reside in nonlymphoid organs 
. In our study, long lasting T cell immunity was found to be augmented not only in both inductive (NALT and CLN) and effector sites (lung and spleen) but also in an unrelated cavity (peritoneal cavity) and migratory memory T-cell nesting compartment (BM) 
. The cytokine responses were also observed at the level of MLN that drains the gastrointestinal tract (another mucosal surface), collectively suggesting that i.n. BCG is capable of inducing immunity in multiple local and systemic immune compartments.
The ELISPOT results further demonstrated the ability of i.n. BCG vaccination to induce strong and sustained mycobacterium-specific immunity, directed toward both secreted and subcellular fractions of M. tuberculosis
. These observations, in part, were in agreement with ELISPOT results observed by Goonetilleke et. al 
using PPD as an in vitro
stimulus where peak IFN-γ responses were observed at 12 weeks after parenteral and i.n. BCG vaccination and were sustained until 24 weeks. However, our observations differ from those of Chen et. al. 
who demonstrated peak IFN-γ levels at 3 weeks post i.n. BCG vaccination that declined rapidly after 6 weeks as measured by cytokine ELISA using killed BCG as in vitro
It has been previously shown that early BCG multiplication occurs in the draining ILN following s.c. vaccination at the base of tail 
. BCG then disseminates in the spleen and lungs after 4 weeks with significantly more bacilli counts in the spleen than in the lungs 
. On the contrary, large portion of the total BCG delivered could be cultured from the lungs within 24 h of i.n. BCG vaccination 
. This results in more BCG load in the lungs than in the spleen of mice 
. Although we have not studied BCG persistence and dissemination in different local and distant organs in detail after i.n. or s.c administration, we observed similar BCG growth kinetics in the lungs and spleen as reported in above studies over a course of 12 weeks depending on the route of vaccination (data not shown). These differences in the BCG bacillary antigenic load in the draining lymph nodes, lungs and spleen after i.n. and s.c vaccination may account for the differences observed in the magnitude of specific immune responses in these organs.
Despite observed differences in the frequencies and location of specific cytokine secreting T cells both BCG vaccination routes afforded comparable levels of protection against airway M. tuberculosis challenge in our study. It can be inferred that the frequency of T cells in the lungs elicited by s.c. BCG vaccination was sufficient to afford protection equivalent to that of i.n. BCG in the lungs. Increased ratio of specific IFN-γ to IL-4 SFU's in the lungs after i.n. BCG vaccination also did not contribute toward the enhancement of protection compared to s.c. BCG. These results suggest that the protection imparted by a single moderate dose of BCG vaccine is independent of route of vaccination in our model. It also indicates that the readout of correlate of protection against TB may not be as simple as induction of dominant type 1 cytokine response.
Our results are contrary to the findings of earlier studies in which i.n. BCG afforded better protection than s.c. vaccination 
. However, these results in part correlate with the findings of Goonetilleke et. al. 
, who demonstrated that a single i.n. dose of BCG did not differ in the degree of protection afforded and that two sequential i.n. BCG doses are required to afford superior protection compared to parenteral BCG vaccine administered using similar double dosing strategy. In different studies, intravenous, oral, rectal and aerosol BCG administration using single dose imparted similar protection against M. tuberculosis 
. It has been shown that BCG vaccination using wide range of dose induces varying magnitude and quality of immune responses but imparts similar levels of protection 
. Thus, the route of application and consequently the dissemination and tissue localization of BCG following a single vaccination may have minor impact on protective immunity against TB 
However, similar conclusion may not be drawn for non-multiplying subunit or killed vaccines and detailed studies evaluating effect of route, dose and delivery system on the protection imparted by individual subunit vaccines are required. Mice vaccinated via i.n. route using culture filtrate proteins-based subunit vaccine were found to be better protected as compared to s.c. route 
, and the dose of a subunit vaccine formulation has been shown to critically influence its immunogenicity and protective efficacy 
. Furthermore, one cannot totally discount the benefit of mucosal immunity induced by a single i.n. BCG as the compartmentalization of immune responses generated in the lungs and respiratory tract has been shown to have positive implications for the development of booster subunit vaccines and effective prime-boost vaccination strategies 
Central to the development of a subunit vaccine would be identification of M. tuberculosis
proteins that can effectively prime mycobacterium-specific immunity in the respiratory tract. Increased immune responses in the lungs and CLN after i.n. vaccination enabled us to identify M. tuberculosis
proteins recognized by the immune cells of these organs. Among the nine proteins evaluated in this study, Apa and GroEL were found to be highly antigenic in BCG-vaccinated mice as measured by IFN-γ response in different immune compartments. Other polypeptides which induced prominent responses following i.n. BCG vaccination were Ag 85 complex proteins and Pst-S1. The magnitude and hierarchy of these antigen-specific responses varied at different time points, which may be due to stage specific differential expression of antigens in vivo
or variation in subcellular localization of bacilli resulting in the temporal differences in the antigen profile recognized. Of note, Apa, a 45/47 kDa cell surface or secreted glycoprotein, GroEL, a 65 kDa cytosolic heat shock protein and Ag85 complex, 30–32 kDa cell surface or secreted polypeptides, have all been previously shown to be highly antigenic in humans 
with different stages of M. tuberculosis
The Apa complex is composed of mannosylated proteins 
with up to nine identified glycoforms 
. APA is also known as M. tuberculosis
protein (MPT)-32 (calculated molecular mass 32 kDa), ModD (putative involvement in molybdate uptake) or fibronectin attachment protein (FAP), and can mediate bacterial attachment to host cells as a potential adhesin 
. Both Apa and Ag85 complex proteins have been found to be released into phagosomes and other subcellular compartments of M. bovis
BCG infected macrophages 
. The recombinant Apa homolog has also been found to activate dendritic cells and induce Th1 polarization 
. It is noteworthy that Apa is predominantly recognized following vaccination with live but not dead BCG organisms 
. Moreover, monoclonal antibody against Apa has been previously shown to abrogate the attachment and internalization of BCG by human bladder tumor cells and stable binding of BCG to bladder mucosa via FAP was necessary for the expression of BCG-induced antitumor activity 
. Overall, this suggests that APA-specific responses might contribute to protective responses induced by live multiplying BCG.
In parallel with the responses following i.n. BCG vaccination, Apa was also highly immunogenic after intranasal multicomponent subunit vaccination. The immune responses induced by Apa were characterized by strong in vitro type1 and type 2 cytokine response. The Apa also induced strong antibody response characterized by elevated specific serum IgG and nasal lavage IgA as determined using ELISA and cytotoxic T-cell response (mean cytotoxicity 30%) as evaluated by neutral red uptake assay (data not shown).
It has been previously shown that deglycosylation of Apa affects its capacity to stimulate in vitro
proliferation of T-cells from guinea pigs following s.c. BCG vaccination 
. However, i.n. recombinant (unglycosylated) Apa subunit vaccinated mice could significantly inhibit M. tuberculosis
growth in the lungs and spleen similar to that of Ag85A-based subunit vaccine (this study) and a s.c. prime-boost vaccination strategy using recombinant MVA expressing Apa has been previously shown to protect guinea pigs against virulent M. tuberculosis
. These observations along with the strong antigenicity and immunogenicity obtained from the recombinant (unglycosylated) Apa used in this study warrants further evaluation in comparison with its native (glycosylated) counterpart.
The significant level of protection afforded in animal models by Ag85A, another fibronectin binding protein, after i.n. vaccination in previous studies 
and strong T cell mediated immunogenicity elicited by Apa vaccination in the respiratory and other immune compartments leading to comparable protection against M. tuberculosis
challenge in the current study bodes well for use of these molecules as components of future mucosal TB vaccines. Overall, findings of this study strongly support further evaluation of mucosally targeted Apa-based vaccine to prevent tuberculosis.