Attrition of T-Cell Functions and Simultaneous Upregulation of Inhibitory Markers Correspond with the Waning of BCG-Induced Protection against Tuberculosis in Mice

Mycobacterium bovis bacille Calmette-Guérin (BCG) is the most widely used live attenuated vaccine. However, the correlates of protection and waning of its immunity against tuberculosis is poorly understood. In this study, we correlated the longitudinal changes in the magnitude and functional quality of CD4+ and CD8+ T-cell response over a period of two years after mucosal or parenteral BCG vaccination with the strength of protection against Mycobacterium tuberculosis in mice. The BCG vaccination-induced CD4+ and CD8+ T cells exhibited comparable response kinetics but distinct functional attributes in-terms of IFN-γ, IL-2 and TNF-α co-production and CD62L memory marker expression. Despite a near life-long BCG persistence and the induction of enduring CD4+ T-cell responses characterized by IFN-γ and/or TNF-α production with comparable protection, the protective efficacy waned regardless of the route of vaccination. The progressive decline in the multifactorial functional abilities of CD4+ and CD8+ T cells in-terms of type-1 cytokine production, proliferation and cytolytic potential corresponded with the waning of protection against M. tuberculosis infection. In addition, simultaneous increase in the dysfunctional and terminally-differentiated T cells expressing CTLA-4, KLRG-1 and IL-10 during the contraction phase of BCG-induced response coincided with the loss of protection. Our results question the empirical development of BCG-booster vaccines and emphasize the pursuit of strategies that maintain superior T-cell functional capacity. Furthermore, our results underscore the importance of understanding the comprehensive functional dynamics of antigen-specific T-cell responses in addition to cytokine polyfunctionality in BCG-vaccinated hosts while optimizing novel vaccination strategies against tuberculosis.

antigen-specific T-cell responses in addition to cytokine polyfunctionality in BCGvaccinated hosts while optimizing novel vaccination strategies against tuberculosis.

Introduction
Tuberculosis (TB) is the most devastating bacterial disease of all time and is responsible for over 1.3 million deaths annually [1]. The only vaccine available against TB is Mycobacterium bovis Bacille Calmette-Guérin (BCG),6 an attenuated strain of M. bovis. It is the most widely used vaccine in human history, and is recommended in .157 countries [2]. BCG efficiently prevents severe forms of childhood TB [3,4], but has highly variable efficacy against pulmonary TB in adolescents and adults [5,6]. Therefore, extensive research efforts are ongoing to develop new TB vaccines with superior protective efficacy. Since over 4 billion individuals have received BCG to date, it is generally accepted that the development of new vaccination strategies should consider the extensive BCGvaccination coverage [7][8][9]. Accordingly, future vaccination against TB would involve priming with BCG or an improved live mycobacterial vaccine followed by a boost with subunit vaccine [9,10]. A standard approach for screening new vaccine candidates or strategies to improve BCG-induced immunity starts with the evaluation of immunogenicity and protective efficacy in animal models of TB, in particular, the mouse model. However, despite being at the center of new TB vaccination strategies and its extensive use as a reference standard in many preclinical and clinical vaccine trials, our knowledge of the protective immune responses induced by a near century old BCG vaccine is limited in mice and humans.
Although it is known that BCG induces durable immunity in the mouse model [11,12], understanding of the kinetics of immune response in-terms of expansion, peak and contraction of the T-cell and antibody response is incomplete, and the optimal time to boost BCG-immunity is uncertain. These gaps have hampered the development of effective prime-boost vaccination regimens using BCG. As boosting T-cell responses during the expansion phase may lead to T-cell exhaustion and activation-induced cell death [13,14], boosting would be most desirable after the effector phase of BCG-primed responses and when the memory T-cell population is established [15]. We reasoned that the evaluation of growth and persistence of BCG after different routes of administration in mice and the simultaneous investigation of the kinetics of the magnitude and functional quality of T-cell responses generated, together with the longitudinal changes in the protective efficacy against Mycobacterium tuberculosis (Mtb) challenge, may provide an important insight into the correlates of protection and factors associated with the waning of BCG-immunity.
The BCG vaccine is most commonly administered to newborns by the intradermal route. A recent clinical trial in infants has, however, demonstrated equivalent protective efficacy using intradermal and percutaneous vaccination routes [16]. Furthermore, TB is primarily a respiratory disease and it has been hypothesized that vaccination directed at the respiratory mucosa may provide the best opportunity for protection against infection with Mtb. The animal experimental evidence suggests that local immune defense mechanisms might be crucial for the protective immunity against TB, for example, those stimulated by mucosal immunization via the lower respiratory tract [17,18]. Therefore, we investigated the dynamics of BCG immunity traversing the entire life-span of mice following intranasal (i.n., mucosal) and subcutaneous (s.c., parenteral) vaccination.
Our results showed that BCG-induced CD4 + and CD8 + T-cell responses and protection peak between 12-32 weeks but wanes significantly by 78 weeks regardless of i.n. or s.c. vaccination. The magnitude of CD4 + and CD8 + T cells with sound multi-effector capacity in-terms of type-1 cytokine production, proliferation and cytolytic potential correlated best with protection against Mtb. Notably, increase in the dysfunctional and terminally-differentiated T cells coincided with the waning of BCG-induced protection. Our results provide previously undocumented important insight into how BCG persistence and the immune senescence affect T-cell functionality in mice and highlight the need to develop vaccination strategies that prevent T-cell attrition and exhaustion.

BCG persists and induces a durable cell-mediated immune response in mice
To understand the growth and persistence of BCG after mucosal or parenteral vaccination, BALB/c mice were vaccinated with 10 6 BCG Copenhagen colony forming units (CFU) by i.n. and s.c. route, and BCG titers were determined in the lung and spleen at 7 different time points. BCG vaccination by the i.n. route resulted in a high bacillary load in the lung. About 0.2-0.5% of the total BCG administered following i.n. vaccination was detected in the lung at 48 h, while BCG could not be detected in the lung at 48 h after s.c. vaccination and only a low numbers of bacilli (45¡14 CFU) were detected in the lung at week 3 (Fig. 1A). In the spleen, peak BCG load was found at week 3 after s.c. vaccination while it took 6 weeks for BCG to be detected following i.n. vaccination. The BCG CFUs declined by week 12 to 32 but persisted at very low or undetectable levels until week 52. However, by week 78, BCG burden resurged in the two organs of both vaccinated groups. These results demonstrate the chronic persistence of BCG in mice.
In parallel, we determined the magnitude and kinetics of BCG-induced cellmediated immune response using Mtb whole-cell lysate (WCL) for in vitro stimulations of lung and spleen cells. The frequencies of WCL-specific IFN-c, IL-17 and IL-4 spot-forming units (SFU) were measured using a cultured ELISPOT assay. Although i.n. vaccination-induced total cytokine SFU peaked earlier in the lungs (i.e., at week 12) than those produced by s.c. vaccination (which peaked at week 32), the magnitude of WCL-specific total cytokine response in two organs was statistically comparable between the routes of vaccination when the total SFU at 7 different time points were compared ( Figure 1B). The total cytokine response by either route was dominated by higher frequencies and proportions of IFN-c SFU, although at the peak of response the frequencies of IL-17 SFU were significantly greater compared to early (6 week) and late (78 week) time points in the lung. These results suggest that though BCG vaccination induced type-1 immune response declines with age, the nature of response continues to be predominantly type-1 after i.n. or s.c vaccination When the antibody response was investigated in the sera of two vaccinated groups, we observed similar kinetics for WCL-specific IgG-antibody response by ELISA, but it was significant only in the sera of s.c. BCG-vaccinated mice (Fig.  S1A). The WCL-specific IgG response was characterized by greater proportions of IgG2a and IgG2b subclass antibodies (Fig. S1B). Overall, these results demonstrate that i.n. and s.c. BCG vaccination induces a strong cell-mediated response following early bacillary load and is maintained for more than 8 months. The peak of immune response in the lung and spleen coincides with decrease in the BCG burden and persistence of bacilli at very low levels.
BCG vaccination-induced CD4 + and CD8 + T cells exhibit distinct cytokine profile To understand the temporal changes in the magnitude and quality of BCGinduced CD4 + and CD8 + T-cell responses, we evaluated the frequencies of WCLand short term culture filtrate (STCF)-specific IFN-c, IL-2 and TNF-a-producing CD4 + and CD8 + T cells in the lung, spleen, draining and distant lymph nodes (LNs) by polychromatic flow cytometry. The WCL and STCF represent a complete repertoire of Mtb antigens as opposed to individual purified antigens or cocktails of few immunodominant antigens used for in vitro stimulations. We found comparable magnitudes of WCL-specific cytokine-producing CD4 + T cells between i.n. and s.c. BCG-vaccinated groups ( Fig. 2A), when the frequencies of IFN-c, IL-2, or TNF-a-expressing CD4 + T cells were analyzed individually in the lung and spleen. The magnitudes of WCL-specific cytokine-producing CD8 + T cells were also comparable between the two vaccinated groups. The T-cell responses peaked at week 32 in the lung and spleen and gradually waned thereafter, with similar expansion and contraction trend in the two BCGvaccinated groups.
When the polyfunctionality of WCL-specific T cells was analyzed in-terms of co-expression of IFN-c, TNF-a and IL-2 by individual cells [19,20], we found that the CD8 + T-cell response in all organs investigated was dominated almost exclusively by IFN-c single-producers ( Fig. 2B and 2C). In contrast, CD4 + T cells were polyfunctional at the peak, and the response was dominated by higher magnitudes of IFN-c and TNF-a single-and double-positive cells. These results demonstrate that BCG vaccination generates CD4 + and CD8 + T-cell responses of distinct quality. Contrary to the belief that BCG vaccination induces a weak CD8 + T-cell response [21], these data demonstrate the significant presence of IFN-cproducing CD8 + T cells at peak time points. Although the percentages of WCLspecific total cytokine-producing cells among CD4 + and CD8 + T cells were comparable, the frequencies of CD4 + T cells in the T lymphocytes (CD3 + ) of the organs investigated were approximately three-fold higher than the frequencies of CD8 + T cells, thereby confirming the dominance of CD4 + T-cell response after BCG vaccination during the entire response kinetics in mice.
The BCG-induced early T-cell responses tend to be strongest in the compartments most proximal to the site of vaccine administration [22], and it is believed that the distribution of specific T cells may become uniform in the local and distant compartments as the vaccination-induced immune response matures. Therefore, we investigated the specific T-cell responses simultaneously in the draining and distant LNs after i.n. and s.c. BCG vaccination at 5 different time points. Vaccination by the i.n. route induced higher WCL-specific cytokineproducing CD4 + and CD8 + T cells in the draining superficial cervical lymph nodes (CLNs), while s.c. vaccination induced more T-cell responses in the inguinal lymph nodes (ILNs) which drains the flank (the site of s.c. vaccination) (Fig. 2C). Responses in the draining CLNs of i.n.-vaccinated and ILNs of s.c.-vaccinated group also peaked at week 32, but, significantly higher magnitudes (p,0.001) of WCL-specific T cells were observed only in the corresponding draining LNs as compared to the distant LNs at all 5 time points investigated (Fig. 2C). These results demonstrate a near lifelong compartmentalization of T-cell responses in the draining LNs depending on the route of BCG vaccination.
Furthermore, the mean frequencies of WCL-specific T cells waned after week 32 in the two vaccinated groups with significantly higher polyfunctional CD4 + T cells at week 32 (p,0.05) compared to week 78 in the three organs investigated, demonstrating decline in the magnitude and quality of BCG vaccination-induced T-cell response. By week 104, the magnitude of WCL-specific cytokine-expressing CD4 + and CD8 + T cells had waned to levels similar to those observed at week 3 ( Fig. 2C), and we observed similar T-cell cytokine response dynamics using Mtb STCF ( Fig. S2A and S2B). However, in contrast, we did not find any significant difference in the magnitude of WCL-specific cytokine-expressing T cells in young (8-week-old) and old (104-week-old) mice, 6 weeks after BCG vaccination (Fig.  S2C), indicating that the decline in the immune response with time after BCG vaccination was not merely due to age-related senescence. Together, these data demonstrate the gradual waning of BCG-induced CD4 + and CD8 + T-cell responses after 32 weeks of vaccination, regardless of the route of administration.

BCG-induced CD4 + and CD8 + T cells express cytolytic potential that wanes with time
To further characterize the temporal changes in the quality of BCG-induced T cells, we assessed the degranulation potential of CD4 + and CD8 + T cells, measured as the surface mobilization of CD107a and CD107b [19,23,24]. The WCL-specific CD4 + and CD8 + T cells in the lung and spleen of BCG-vaccinated mice expressed CD107 (Fig. 3A), and the route of administration did not affect their degranulation potential. Furthermore, both CD4 + and CD8 + T cells purified from the two organs of vaccinated groups induced cytolysis in target cells based on the up-regulation of apoptosis marker annexin-V on the surface of WCL-pulsed bone-marrow-derived dendritic cells ( BCG-induced CD4 + and CD8 + T cells exhibit distinct memory phenotype in-terms of CD62L expression and their proliferative potential decreases during waning Next, we assessed the association between contracting T-cell responses and other functional abilities such as proliferative potential and up-regulation of memory markers by T cells. Using CFSE-dilution assay, we found the highest WCL-specific proliferative ability of CD4 + and CD8 + T cells in vitro at the expansion phase (week 12) after i.n. or s.c. vaccination ( Fig. 4A and 4B). Thereafter, the proliferation decreased at week 32 and further waned significantly by week 52.
Simultaneously, we investigated the memory phenotype by assessing the expression of CD44 and CD62L or CCR7 [25], and found that .90% of lung and spleen WCL-specific IFN-c + CD4 + T cells expressed CD44 + CD62L 2 activated effector or effector memory (T EM ) phenotype at week 12 and 52 ( Fig. 4C and 4D). Higher percentages (22.45-41.60%) of WCL-specific IFN-c + CD8 + T cells expressed CD44 + CD62 + central memory (T CM ) phenotype (Fig. 4D), but, both CD4 + and CD8 + T cells did not produce significant IL-2 or express another classical T CM marker, CCR-7, during the contraction phase ( Fig. S4A-C). These results suggest generation of predominantly effector/T EM responses but weak T CM responses following BCG vaccination.

T-cell dysfunction markers are up-regulated during waning of BCG responses
Chronic infections often lead to impaired T-cell responses characterized by a progressive loss of T-cell functions, terminal differentiation and eventual T-cell exhaustion [26][27][28][29]. Since BCG chronically persists in mice, we sought to determine whether up-regulation of T-cell inhibitory receptors and markers of Tcell dysfunction corresponded with the contraction of BCG-induced T-cell response and simultaneous decrease in T-cell functionality. We investigated the surface expression of cytotoxic T-lymphocyte antigen 4 (CTLA-4), programmed cell death protein 1 (PD-1), and killer cell lectin-like receptor subfamily G member 1 (KLRG-1), and the intracellular production of IL-10 by T cells. The KLRG-1 and CTLA-4 expression on CD4 + and CD8 + T cells of vaccinated mice increased significantly by week 78 compared to week 12, and compared to agematched naïve mice (p,0.05) (Fig. 5A-5D), suggesting that these changes are not simply age related [30]. Following WCL stimulation, .97% of CTLA-4 + or KLRG-1 + T-cell subsets did not produce IFN-c, IL-2 or TNF-a (data not shown). However, we did not observe any significant increase in the percentages of PD-1expressing CD4 + or CD8 + T cells during the contraction phase in BCG-vaccinated mice (Fig. 5C). Apart from the inhibitory-receptor signaling, cytokines like IL-10 also exert suppressive effect and mycobacterial reactivation [31,32], and we found increased WCL-specific IL-10-producing CD4 + and CD8 + T cells at week 52 as compared to week 12 in BCG-vaccinated mice ( Fig. 5E and 5F). Collectively, these results suggest that the waning of BCG-induced T-cell responses corresponds with the increased frequency of inhibitory markers-expressing CD4 + and CD8 + T cells.
Of importance, in vitro rIL-2 treatment of the lung and spleen cells significantly increased WCL-specific IFN-c and IL-17 SFU and the total cytokine (IFN-c, IL-2 or TNF-a) response at week 78 ( Fig. 5G and 5H), indicating that the dysfunctional T-cell responses could be partly rescued by IL-2 therapy.

Waning of BCG-induced protection against Mtb coincides with decreased T-cell functionality
To understand the kinetics and durability of BCG-induced protection, we challenged i.n. and s.c. BCG-vaccinated mice with Mtb Erdman at different time points after vaccination, and Mtb load was determined in the lung and spleen 6 weeks later. Our data demonstrate that the route of vaccination did not Dynamics of BCG-Immunity in Mice significantly alter the protective efficacy against Mtb at all eight infection time points investigated compared to age-matched naïve controls (Fig. 6A). The protection was found between 6-52 weeks with maximum efficacy at week 32 (log 10 CFU reduction 1.77-2.1 in lung and 2.34-2.56 in spleen), but eventually waned by week 78. Previously elderly (2-year-old) mice have been shown to effectively control early Mtb infection [33], and in our model, we consistently found low CFU burden in the lung or spleen of 104-week-old elderly naïve mice compared to simultaneously infected 12-week-old young control mice 6 weeks Dynamics of BCG-Immunity in Mice after challenge (data not shown), confirming that the low CFU levels in week-104 groups (Fig. 6A) were not due to failed infections. The frequency of ESAT-6+CFP-10-specific cytokine-producing T cells in the lung and spleen evaluated as a surrogate of Mtb load after challenge [34] correlated inversely with the degree of protection ( Fig. 6B and S5A), with significantly more ESAT-6+CFP-10-specific CD4 + T cells in week-3 and -78 as compared to week-32 groups, 6 weeks postchallenge. Furthermore, Mtb challenge during the expansion phase of BCGinduced response resulted in a better recall of WCL-specific T-cell responses in the two organs ( Fig. 6C and S5B). Conversely, challenge during the contraction phase demonstrated decreased frequency of WCL-specific cytokine-producing CD4 + and CD8 + T cells compared to the levels observed pre-challenge. Together, these results suggest that the degree of BCG-induced protection correlates with the magnitude and intact functional capability of specific CD4 + and CD8 + T cells (Fig. 6D).

Discussion
We found increased BCG load in the lung as compared to spleen following i.n. vaccination while s.c. vaccination resulted in relatively higher levels in the spleen as compared to lung by 3 weeks. Similar results have been reported earlier after i.n. vaccination [35], and it is known that following s.c. vaccination initial BCG multiplication occurs in the draining ILN and results in higher bacterial dissemination to the spleen than the lung by week 4 [36]. It has been suggested that BCG chronically persists in the draining lymph nodes of mice after s.c. vaccination [12], and may serve as a reservoir for dissemination to the lung and spleen. In our study BCG sporadically disseminated to the lung in low numbers during early and late time points after s.c. vaccination. Previous studies of the shorter duration have suggested that BCG never or only rarely disseminates to the lung after s.c. vaccination [37], while others have reported that BCG disseminates to the lung in low numbers during a 20-week study [36]. In view of the recent report of bone marrow stem cells as a niche for dormant Mtb, further studies are required to characterize reservoirs of persistent BCG in mice [38]. Despite a significant decline in the BCG-burden by week 12 and the persistence of bacteria in low numbers thereafter at week 32-52, BCG-induced Mtb-specific T-cell responses and the degree of protection against subsequent Mtb challenge peaked at week 12-32 in the lung and spleen in our study. This occurred regardless of the early differences in the tissue localization of BCG-burden and the pronounced differences in the location of specific T cells in the draining LNs depending on the route of vaccination. We did not find any significant difference in the strength of protection against Mtb challenge between i.n. and s.c. BCG vaccination at all-time points investigated using a moderate vaccine dose. The dose and route of BCG administration and consequently the tissue localization have been previously suggested to have a minor impact on the protective immunity against TB [39,40]. Likewise, in our model, no direct correlation was found between the extents of BCG antigenic load and the magnitudes of BCG-induced T-cell responses or the levels of protection against Mtb. In fact, the peak of T-cell responses succeeded early BCG load and waned despite relatively higher BCG titers at week 78.
The BCG-induced T-cell responses in our study were characterized predominantly by IFN-c and/or TNF-a-producing CD4 + T cells and single IFN-c-positive CD8 + T cells. Although the importance of CD4 + T cells in protection against Mtb has been widely accepted, the role of subunit vaccine-induced CD8 + T-cell response has been recently debated [41,42]. Nonetheless, BCG vaccinationinduced CD8 + T-cell response has been previously shown to play an important part in the protection against Mtb [21,43,44]. Multiple studies have shown that IFN-c and TNF-a cytokines are required for immunity against Mtb infection [45,46]. In contrast, several other studies have shown that the magnitudes of IFNc-producing T cells or in vitro cytokine levels do not correlate with the degree of protection against TB [34,40]. Recently, IFN-c, TNF-a, and IL-2 co-expression by individual polyfunctional CD4 + T cells has been proposed as a marker of BCGinduced protective immunity [20]. In our study, the frequency of Mtb-specific total or polyfunctional Th1 cytokine-producing CD4 + T cells was about two-fold more at the peak (week 32) compared to the expansion phase (week 12), but the levels of protection after Mtb challenge was not significantly different. Furthermore, the protection was higher at week 12 compared to week 78, despite comparable magnitudes of BCG-induced T-cell responses. Hence, the correlation between the magnitude of total cytokine-producing T cells or polyfunctional CD4 + T cells and the degree of protection against Mtb challenge, although positive, appears to be rather weak. Previously, it has been reported that the frequency of BCG-specific CD4 + , CD8 + , and cd T cells and polyfunctional Th1 cytokine-producing CD4 + T cells in the blood does not correlate with the protection against TB in BCG-vaccinated children [47]. The increased BCGburden during the contraction phase and the concurrent decrease in the magnitude and functionality of type-1 cytokine-producing specific T cells is in agreement with previously described negative association between the functional capacity of Mtb-specific T cells and Mtb load in the sputum of TB patients [48,49].
It is known that persistent, low-dose antigenic stimulation throughout life may compromise antigen-specific T cells in two ways. First, they may become functionally exhausted and lose essential functional activity that is necessary for immune protection. Second, repeated T-cell stimulation can lead to a loss of the replicative capacity of some antigen-specific T cell populations, as a result of telomere erosion and/or unrepaired DNA damage, a process known as replicative senescence (reviewed in [30]). The induction of senescence in T cells is also known to limit the long-term persistence of immune responses during aging [50]. Therefore, the attrition of specific T-cell responses discerned in our study can be attributed in part to the onset of immune senescence associated with the advancing age in BCG-vaccinated mice. The terminally-differentiated and senescent T cells are characterized by the up-regulation of KLRG-1 expression [51], while functionally exhausted T cells are defined by the early loss of IL-2 production, decreased proliferation, and the increased expression of inhibitory markers like CTLA-4 and PD-1 with low levels of T CM markers CD62L and CCR-7 [27,52,53]. Although the terms senescence and exhaustion are often used interchangeably when referring to highly differentiated T cells, available data suggest that both processes are regulated independently of each other [30]. It should be noted that the onset of the contraction of specific T-cell responses in our study was chronicled at the week 52 in BCG-vaccinated mice i.e., at the adult mid-life, and significantly increased expression of KLRG-1 and CTLA-4 was found during the contraction phase on the T cells of BCG-vaccinated mice compared to those of age-matched naïve mice. In addition, we found comparable magnitudes of specific T-cell cytokine responses in elderly (104-week-old) and younger (8-week-old) mice, 6 weeks post BCG vaccination. Old mice (24-month-old) have been previously shown to effectively control early Mtb infection [33,54], and the BCG vaccination has been found to be equally protective in young, adult and old guinea pigs against TB [55]. Therefore, decreased functional abilities of T cells during the contraction phase of BCG-response observed in our study may not be only due to the advanced-age-related senescence in mice, and it is probable that the chronic persistence of BCG with low grade continuous antigen stimulation throughout life contributes to the attrition of T-cell functions and precipitates immune senescence.
Like Mtb infection [25], BCG-induced responses in our study were dominated by CD4 + effector/T EM subset secreting IFN-c and/or TNF-a [56], and the frequency of IL-2 + T cells was much lower compared to that reported previously after protective subunit vaccination of mice [57]. Even though CD44 + CD62L + expression on CD8 + T cells in our study suggested an acquisition of the T CM phenotype [58], these cells did not produce IL-2 during the contraction phase and showed concurrent impaired proliferative capacity contrary to the earlier reported characteristics of classical T CM cells [59,60]. Multiple human and mouse studies have also shown that BCG primarily promotes effector/T EM cells [12,61]. Many factors may influence development of a long-lived and high-quality memory T cells and of these, antigen load and persistence may be the most important ones [62]. Our study provides a direct evidence for the chronic persistence of BCG load and its correlation with the poor central memory T-cell responses in mice. Previous observations of improved T CM responses after clearance of Mtb or BCG in mice and their association with protection provide additional evidence for the correlation between chronic persistence of BCG and poor T CM responses [63][64][65]. Paradoxically, the protective efficacy of BCG in mice also appears to be dependent on the presence of viable bacilli [36] and consequently on the T EM responses. Recently, greater CD4 + T CM responses induced by recombinant BCG DureC::hly vaccine (rBCG) have been shown to be responsible for superior protection compared to canonical BCG against Mtb and the T CM responses are shown to coincide with the faster clearance of rBCG in mice [66]. It has been suggested that the canonical BCG is cleared more rapidly in humans than mice and influence the memory phenotype (especially T CM phenotype) of BCG-specific CD4 + T cells [15]. However, in this study [15], specific CD4 + T cells expressed characteristics of both central memory (CCR7 + expression) and effector memory (predominant cytokine IFN-c and cytotoxic marker expression) following contraction and exhibited a more complex and different phenotype than the classical IL-2 + T CM cells. The growth and persistence of BCG in humans is not known with certainty, and reports of BCG-lymphadenitis or dissemination years after BCG vaccination are used to support a relatively long life-span of this vaccine in humans [67][68][69]. On the other hand, one would argue that disseminated BCG infections in HIV infected patients and children are not frequent complications or common findings [70]; supporting the viewpoint that viable BCG probably does not persist in children for a long period of time after vaccination. We agree that some aspects of the BCG growth observed in mice such as the resurgence of BCG-burden during the contraction phase of immune response may not necessarily occur in immunocompetent humans and rather this may be a phenomenon seen only in susceptible animal models and children with immune gene deficiency [71]. Like reported observations in mice [7], BCG-induced immunity is thought to wane with time after vaccination at birth in humans and is generally considered to last no more than 10-15 year [72]. However, the time frame for the waning of BCG in humans is still a subject of debate, and a long-term protective efficacy of BCG in some populations has been reported [73]. It is generally agreed, nevertheless, that the waning of BCG-induced protection through childhood and young adult life coincides with the gradual increase in TB incidence in highly TB-endemic regions, which reaches peak in the 25-35-year-old age group [8]. Overall, caution is merited while comparing longitudinal changes in BCG-induced immune responses in mice with those observed in humans.
The persistent antigen load in chronic viral infections and active TB in humans is known to be associated with a progressive impairment of T-cell responses [26,53,74], and up-regulation of markers of T-cell dysfunction, terminal differentiation and exhaustion [52,75,76]. In our study, BCG-induced response during the contraction phase was characterized by an increased expression of KLRG-1, CTLA-4 and IL-10 production, reduced ability to recall T-cell responses following Mtb challenge and loss of protection against Mtb. However, we did not observe significant changes in the PD-1 expression on CD4 + or CD8 + T cells during the three phases of BCG-induced response and our results are at odds with the previous study [77], which documented PD-1 expression on T cells after a high-dose intravenous BCG infection of mice. It has been previously shown that the BCG-vaccinated adults harbor early-stage-differentiated, low PD-1-expressing antigen-specific effector memory CD4 + T cells while BCG-vaccinated, latent TB infected (LTBI) adults had late-stage-differentiated PD-1 + T cells [78], and the advanced TB patients harbor PD-1 + severely exhausted T cells with compromised IFN-c production [74,79]. It is tempting to speculate that the resurgence of BCG growth at the late time point (78 weeks) in our study could be due to the lack of immune control and attrition of T-cell functions. However, since BCG CFUs were not enumerated at 104-week time point, this needs further investigation. Overall, our results suggest that the multifactorial functional ability of BCG-induced T cells in-terms of type-1 cytokine production, proliferative and cytotoxic potential and ability to recall immune-mediators post Mtb challenge likely correlate with the superior protection against TB. These results also suggest a need to look beyond the quantitative measures of few cytokine markers of vaccine-induced Th1 cells, with improved focus on the combined functional abilities of different immune subsets.
Our results have several implications for the development of effective vaccination strategies using BCG or other live TB vaccines. Although parenteral BCG vaccination via skin was sufficient to induce a durable protective immunity in the lung, our observation of the life-long route-specific compartmentalization of T-cell responses in the draining LNs has important ramifications for the development of effective prime-boost vaccination strategies. The administration of a booster vaccine by a homologous route, involving the same draining LNs as the primer vaccine, may be an effective strategy to induce a superior booster effect in prime-boost scenarios. The ideal time to boost BCG-primed T-cell responses in the mouse model might be after the expansion and peak phase i.e. at week 52 or thereafter. This viewpoint is further supported by a recent data from our laboratory, where boosting with Apa (Rv1860)-based subunit vaccine during the contraction phase offered a stronger boost and imparted significant protection [80]. The data also reinforces the fact that in most published vaccine studies, mice are challenged far too quickly [9] and the longer vaccination-to-challenge intervals, such as 6-8 months after BCG vaccination, will provide better information about the protective efficacy of preclinical vaccines against Mtb infection in the mouse model. Further, the loss of protection during the chronic phase of BCG growth and the corresponding increase in KLRG-1 + and CTLA-4 + T cells highlight the need to develop strategies which can overcome these inhibitory signals and promote a highly distinct or superior quality T-cell responses. One such approach based on our observations could be a rIL-2 therapy to rescue dysfunctional T cells [81]. Recently, it was demonstrated that BCG-mediated protection is lost as Mtb infection enters the chronic phase in C57BL/6 mice, but a single boost with H1-subunit vaccine could selectively promote CD4 + KLRG-1 2 IL-2 + T CM cells and impart long-term protection [82], which suggests that strategies to prolong BCG-efficacy without T-cell exhaustion could be pursued.
Overall, our study provided a comprehensive portrait of longitudinal changes in the specific T-cell responses following BCG vaccination in the mouse model. The data suggest an association between the functional capacities of BCG-induced T cells and levels of protection against Mtb and provide a rationale for future longitudinal studies optimizing novel prime-boost vaccination strategies and understanding the mechanisms of loss of BCG-induced protection in young adults from the regions of high prevalence of Mtb infections. It is a wellestablished fact that BCG provides the lowest level of protection against TB in the regions with the highest prevalence of latent infection [5], and the caveat is that the certain aspects of the dynamics of BCG-immunity in mice may not inevitably embody human responses. In the light of results presented here, however, it is reasonable to envisage that repeated Mtb and environmental mycobacteria exposure and prevalent latent infections in these human populations may further exacerbate T-cell attrition and exhaustion leading to breakdown in BCG-induced anti-TB immunity. Since BCG vaccination-induced T-cell responses mirror those reported by an actual Mtb infection in mice and humans, our results indicate that a successful novel TB vaccine should re-educate the immune system rather than reproducing the natural immune responses generated by Mtb infection.

Ethics Statement
All animal experiments performed were in strict accordance with the guidelines of the U.S. Public Health Service Policy on the Humane Care and Use of Animals and the Guide for the Humane Care and Use of Laboratory Animals. All procedures including euthanasia by exsanguination under isoflurane inhalation anesthesia were approved and supervised by the Institutional Animal Care and Use Committee of CDC (Approval numbers SABMOUC1664, 1847, 1490 and SHIMOUC 1422). We obtained 6-8-week-old female BALB/c mice from the Charles River Laboratories. Experimental mice were housed under specificpathogen-free conditions in the animal biosafety laboratory (ABSL)-II and ABSL-III facilities of the Centers for Disease Control and Prevention (CDC) along with sentinel mice.

BCG Vaccination and Mtb Infection
M. bovis BCG Copenhagen vaccine (Danish 1331 strain, Staten Serum Institute, Copenhagen, Denmark) and Mtb Erdman challenge strain were provided by the Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, MD. For vaccinations of mice, lyophilized BCG was suspended in vaccine diluent (diluted Sauton medium) provided by the supplier. I.n. vaccination was carried out by applying a total of 30 ml of BCG vaccine suspension containing 1610 6 CFU drop-wise to the external nares (15 ml per nostril) using a fine tip micropipette and allowing the mouse to inhale the suspension into the lungs naturally [22]. For s.c. vaccination, 50 ml of a BCG suspension was injected above the gluteus superficialis and biceps femoralis muscles of both hind legs using a 26-gauge needle to deliver a total of 1610 6 CFU. The vaccine diluent-administered or naïve mice served as controls.
BCG-vaccinated and control mice were challenged by the airway i.n. route to deliver 5610 4 CFU of Mtb Erdman [22,80]. At 48 h following challenge, approximately 0.5% (mean 250 CFU) of the total CFU delivered could be cultured from the lungs. In BCG-vaccinated mice bacillary titers were determined by plating serial dilutions of organ homogenates onto Middlebrook 7H10 agar (BD-Biosciences) supplemented with 0?2% (v/v) glycerol and 10% (v/v) Oleic Albumin Dextrose Catalase (OADC) growth supplement. The BCG growth on plates was confirmed by spoligotyping to rule out atypical mycobacterial coinfection in mice. In Mtb-challenged mice, organ CFU were determined on Middlebrook 7H10 agar supplemented with OADC, 10 mg ml 21 ampicillin and 50 mg ml 21 cycloheximide with or without 2 mg ml 21 thiophene 2 carboxylic acid hydrazide (TCH). Mycobacterial colonies were enumerated after 4 weeks of incubation at 37˚C.

ELISPOT assay
Commercially available ELISPOT kits (BD-Biosciences and eBioscience) were used to investigate IFN-c, IL-2, IL-4 or IL-17A secretion by mouse organ cells according to the manufacturer's protocol as described before [22,80]. See Text S1 for details.

Flow cytometry
Mouse organ cells were stained with specific fluorochrome-labeled antibodies to assess the expression of cell surface markers and intracellular cytokines using polychromatic flow cytometry as described earlier [19,80]. See Text S1 for details.

CFSE dilution assay
Cells were stained with 5 mM CFSE (Invitrogen) in a total volume of 1 ml for 5 min in the dark [84]. The reaction was stopped using cold PBS containing 5% FBS. The cells were stimulated with or without Mtb antigens (1 mg ml 21 ) for 72 h in the presence of 5% CO 2 at 37˚C. The cells were washed, stained with desired anti-mouse fluorochrome-labeled antibodies and acquired using FACS Caliber flow cytometer and analyzed using FlowJo software.

Cytotoxicity assay
Syngeneic BMDCs were generated using femur bone-marrow cells as described previously [85]. The lung and peritoneal macrophages were obtained after overnight adherence of lung cells or peritoneal lavage of naïve mice. These antigen presenting cells were pulsed with WCL (10 mg ml 21 ) for 4 h at 37˚C in RPMI-1640. Cells were washed (64) and co-cultured with CD4 + or CD8 + T cells purified using Dynabeads-untouched-T-cell purification kits (Life Technologies) from the lung or spleen cells of BCG or naïve mice for 12 h at 37˚C. The target and effector cell ratios of 1:0, 1:1, 1:10 and 1:20 were used. The apoptosis of the target cells after co-culture was assessed using mouse annexin-V staining kit (BD-Biosciences) and polychromatic flow cytometry according to the manufacturer's instructions. In addition to the annexin-V (FITC) and propidium iodide (PI), the cells were stained with antibodies against CD11c (Alexa Fluor-700-clone HL3), CD11b (PE-Cy7-clone M1/70), F4/80 (APC-clone BM8) and CD4 (PerCP-clone RM4-5) or CD8 (PE-clone 53-6.7). The cells were acquired using LSR II flow cytometer and the data was analyzed by Flowjo software.

In vitro inhibitory receptor blockade and IL-2 treatment
We investigated the potential of rIL-2, anti-KLRG-1 and anti-CTLA-4 antibody treatment to reinvigorate T-cell response in in vitro cultures at the 78 week time point after BCG vaccination of mice. We incubated lung or spleen cells with 10 ng mL 21 of rIL-2, with or without WCL stimulation, in ELISPOT and ICS assays. In vitro KLRG-1 or CTLA-4 blockade using corresponding antibodies was also attempted [86,87]. Different concentrations of KLRG-1 antibody (2F1; eBiosceinces) up to 20 mg mL 21 had no detectable effect on the cytokine response in these assays; however, 4F10 CTLA-4 antibody (BD-Biosciences) or purified Fab (20 mg mL 21 , using Pierce Fab Micro Preparation kit, Thermo Fisher) completely inhibited WCL-specific T-cell cytokine response suggesting inhibitory rather than blocking effect of this antibody [88].

Statistical analyses
Differences in immune response between groups or time points were assessed by 1-way ANOVA followed by Tukey's post-test and the protective efficacies between vaccinated and control groups were compared by the nonparametric Kruskal-Wallis test followed by the Dunn's post-test (GraphPad Prism program). Unless indicated, all immune response data are presented after subtracting no antigen control values. The P value ,0.05 was considered significant. Figure S1. Longitudinal changes in the serum antibody response following BCG vaccination. BALB/c mice vaccinated with 1610 6 CFU BCG by i.n. or s.c. route were euthanized at seven different time points by exsanguination and the serum was isolated from the cardiac blood and evaluated for the antibody response by ELISA. The WCL-specific total IgG (A) or three different IgG subclass (B) levels in the serum of vaccinated or naïve mice are plotted as an absorbance at 492 nm. The data are mean ¡ SEM responses of 5-32 individually analyzed mice per time point per group. doi:10.1371/journal.pone.0113951.s001 (TIF) Figure S2. Longitudinal changes in the T-cell responses following BCG vaccination. (A-B) The lung and spleen cells of i.n. and s.c. BCG-vaccinated mice (n54/time point/group) were stimulated with Mtb STCF, and the magnitudes and polyfunctionality of CD3 + CD4 + and CD3 + CD8 + T cells in-terms of IFN-c, IL-2 and TNF-a production were determined using polychromatic flow cytometry.

Supporting Information
The longitudinal changes in the magnitudes of STCF-specific total cytokine producing CD4 + (A) or CD8 + (B) T cells in the lung (top) and spleen (bottom) are plotted as bar graph while changes in the magnitude of STCF-specific single-, double-, triple-cytokine-producers are plotted as an overlay line graph. The data (A, B) are mean ¡ s.e.m. responses of 2 (at week 3, 6, 12, 78 and 104) or 3 (at week 32 and 52) independent experiments. *Significantly less total cytokine response was found compared to week 32 using 1-way ANOVA with Tukey's post-test; * P,0.05 and ** P,0.01. (C) The effect of age and time on the BCG vaccination-induced T-cell response. The magnitudes of total cytokine-producing CD4 + T cells in the lungs and spleen after WCL stimulation were investigated at the age and time indicated in three BCG-vaccinated groups (n54 mice/time point/group). These groups include i) 8-week-old mice vaccinated with BCG and sacrificed 6 weeks later, ii) 8-week-old mice vaccinated with BCG and sacrificed 104 weeks later, and iii) 104-week-old mice vaccinated with BCG and sacrificed 6 weeks later to investigate T-cell response. *Significance in comparison with the group of mice vaccinated at 104 weeks of age and sacrificed 6 weeks later is indicated (1-way ANOVA with Tukey's post-test; * P,0.05 ** and P,0.01). doi:10.1371/journal.pone.0113951.s002 (TIF) Figure S3. BCG vaccination-induced CD4 + and CD8 + T cells express cytolytic potentials. Representative flow cytometry histograms demonstrate annexin-V expression on the cell surface of WCL-pulsed target cells (i.e., CD3 2 F4/80 + macrophages) from the (A) lung or (B) peritoneal lavage after co-culture with the purified CD4 + (left) or CD8 + (right) T cells (at 1:20 ratio) from the pooled lungs (upper panel) or spleens (lower panel) of BCG-vaccinated or age-matched naïve mice (n53/group). (*P,0.05, ** P,0.01, using 1-way ANOVA with Tukey's post-test compared to naïve controls). doi:10.1371/journal.pone.0113951.s003 (TIF)