Deficiency of Leishmania Phosphoglycans Influences the Magnitude but Does Not Affect the Quality of Secondary (Memory) Anti-Leishmania Immunity

Despite inducing very low IFN-γ response and highly attenuated in vivo, infection of mice with phosphoglycan (PG) deficient Leishmania major (lpg2-) induces protection against virulent L. major challenge. Here, we show that mice infected with lpg2- L. major generate Leishmania-specific memory T cells. However, in vitro and in vivo proliferation, IL-10 and IFN-γ production by lpg2- induced memory cells were impaired in comparison to those induced by wild type (WT) parasites. Interestingly, TNF recall response was comparable to WT infected mice. Despite the impaired proliferation and IFN-γ response, lpg2- infected mice were protected against virulent L. major challenge and their T cells mediated efficient infection-induced immunity. In vivo depletion and neutralization studies with mAbs demonstrated that lpg2- L. major-induced resistance was strongly dependent on IFN-γ, but independent of TNF and CD8+ T cells. Collectively, these data show that the effectiveness of secondary anti-Leishmania immunity depends on the quality (and not the magnitude) of IFN-γ response. These observations provide further support for consideration of lpg2- L. major as a live-attenuated candidate for leishmanization in humans since it protects strongly against virulent challenge, without inducing pathology in infected animals.


Introduction
Leishmaniasis is a serious global health problem that affects millions of people worldwide, especially in developing tropical and subtropical countries. According to WHO estimate, about 2 million new cases occur every year and over 12 million people are presently infected [1]. In India, around 60,000 deaths were reported in 1999 due to visceral leishmaniasis, but the actual number is thought to be even much higher [2]. A most recent report estimated that 20,000-40,000 leishmaniasis deaths occur per year and most of these deaths occur in only six countries [3]. Various forms of pentavalent antimonial components are used for treatment of human leishmaniasis, but treatment failures and drug resistances are common [4,5]. Hence, there is urgent need for new drugs as well as the development of effective human vaccines to prevent the disease. The development of effective vaccine requires an understanding of the factors that regulate secondary protective immunity.
Following recovery from primary natural or experimental infection with L. major, a state of immunity develops that is able to rapidly protect healed animals (both humans and mice) against secondary challenge [6]. Such infection-acquired immunity is very durable and is mediated by IFN-c-producing effector and central memory-like T cells [7]. Infection-acquired immunity is the underlying principle behind leishmanization, which is still practiced in some countries today [8]. However, the significant morbidity associated with the practice has hampered its use as an acceptable vaccination strategy. Moreover, because leishmanization results in a chronic (latent) infection state, concerns have been raised of the possibility to full blown infection in immunocompromised individuals [9].
An increasing number of parasites lines arising through gene replacement methods have been described which show some promise as vaccine candidates in animal studies [10][11][12]. Among these, phosphoglycan (PG) deficient L. major (termed lpg22) is of particular interest because it does not induce pathology even in immunocompromised mice [13] and persists indefinitely at levels comparable to WT parasites. Persistence has been associated with maintenance of infection-acquired immunity [14,15]. Since lpg22 mutant parasites persist without causing any disease even in the susceptible mice and protects against virulent challenge [12], these attributes make lpg22 mutants a promising live-attenuated Leishmania vaccine candidate and have provoked considerable interests in understanding how it persists and interacts with the host immune system. An unanswered important question is whether lpg22 parasites could induce secondary (memory) immune response comparable to those of WT parasites. Here, we show that despite significant differences in quantity, the secondary anti-Leishmania immunity induced by WT and lpg22 parasites are qualitatively similar. These findings further support

Parasites
The origin of wild type (WT) and phosphoglycan deficient (lpg22) L. major has been previously described [13,[16][17][18]. Parasites were cultured at 26uC in M199 medium (Hyclone, Logan, UT) supplemented with 10% heat-inactivated FBS, 2 mM L-glutamine, 100 U/ml penicillin, 100 mg/ml streptomycin (Hyclone). For selective growth of lpg22 line, Hygromycin B (20 mg/ ml) was added to the culture media. While in a previous study we reported the occurrence of lpg22 revertants lacking LPG but conferring pathology [19], this was not observed in the studies described here.

Infection Protocol and Parasite Quantification
Stationary phase promastigotes were washed three times, resuspended in PBS at 10 8 /ml and 50 ml containing 5610 6 (for C57BL/6 infections) or 2610 6 (for BALB/c infections) parasites was injected into the right or left hind footpad. Lesion sizes were monitored weekly with Vanier calipers and parasite burden was determined by limiting dilution assay [20].

Generation of Bone Marrow Derived Dendritic Cells (BMDCs) and in vitro Infection
Bone marrow cells were isolated from the femur of C57BL/6 mice, seeded in 100615 mm Petri dishes at 2610 5 /ml and differentiated using recombinant murine GM-CSF (20 ng/ml, Peprotech, Indianapolis, IN). The culture media were changed twice on day 3 and on day 6, and on day 7, the non-adherent cells were collected and assessed for the expression of CD11c, CD40, CD80, CD86 and MHC class II by flow cytometry. The purity of DCs was between 85-92% (CD11c + cells). BMDCs were incubated with WT parasites for 5 hours at a BMDC to parasite ratio of 1:10. Thereafter, free parasites were washed away and infected cells were used in subsequent co-culture experiments.
T cell Purification, 5-(6-) Carboxyfluorescein Diacetate Succinimidyl Ester (CFSE) Labeling, Adoptive Transfer and Co-culture Experiments T cells were purified from the spleens or dLNs of infected or naive mice by positive selection using CD90.2 coated microbeads and autoMACS (Auburn, CA) according to the manufacturer's suggested protocols and labeled with CFSE dye as described previously [21]. Ten to 30 million cells were adoptively transferred into naïve congenic (Thy1.1) mice by tail vein injection. Recipient mice were subsequently infected with L. major the next day, sacrificed at 5, 14 and 21 days post-challenge to assess proliferation, CD44 and CD62L expression and TNF and IFNc production.

In vitro Recall Response, Proliferation and Intracellular Cytokine Staining
At various times after infection, spleen and dLN cells were cultured in complete DMEM medium at 4610 6 cells/ml (1 ml/ well) in 24-well tissue culture plates and stimulated with soluble Leishmania antigen (SLA, 50 mg/ml) as previously described [22]. After 72 hr, the culture supernatants were collected and stored at -20uC until assayed for cytokines by ELISA. For proliferation, CFSE-labeled cells labeled were resuspended at 10 6 /ml, plated onto 96-well round bottom plates and stimulated with SLA or anti-CD3 and anti-CD28 as previously described [23] or co-cultured with L. major-infected BMDCs at T cell:BMDC of 100:1. After 5 days, proliferation was analyzed by flow cytometry. Some cells were used for intracellular cytokine (IL-4, IL-10, TNF and IFN-c) staining as previously described [24]. Samples were acquired on a FACSCanto II flow cytometer (BD Bioscience, Mississauga, ON, Canada) and analyzed with FlowJo software (TreeStar, Ashland, OR).

Cytokine ELISAs
The levels of IL-4, IL-10, TNF and IFN-c in the culture supernatant fluids were determined by sandwich ELISA using antibody pairs and recombinant cytokine standard (BD Biosciences San Jose, CA) according to the manufacturer's suggested protocols.
Treatment with anti-IFN-c, anti-CD8 Monoclonal Antibody (mAb) and TNFR-Ig in vivo Mice infected with WT or lpg22 L. major were injected intraperitoneally with purified anti-IFN-c mAb (XMG1.2, 2 mg/mouse) or 30 mg/kg Embrel, a TNFR2-Ig fusion protein that inhibits functional activity of murine TNF in vivo [25,26]. The next day, mice were challenged with L. major and antibody or fusion protein treatments were continued weekly for additional 2 weeks. In some experiments, mice were treated with anti-CD8 mAb (TIB210, 1 mg/mouse) 1 day before challenge and for additional 2 weeks (at weekly intervals). This treatment leads to complete and sustained depletion all CD8 + cells throughout the treatment period. All mice were sacrificed after 3 weeks to estimate parasite burden.

Statistical Analysis
Results are shown as the mean 6 SEM. A two-tailed Student's t-test was used to compare means of lesion sizes, parasites burden, and cytokine production from different groups of mice. Significance was considered if p#0.05 (*, = p#0.05; **, = p#0.01 and ***, = p#0.001).

lpg2 2 L. major Parasites Induce Memory T cell Population in the Susceptible and Resistant Mice
As previously reported [13], lpg2 2 L. major persists in the footpad of BALB/c mice without causing any lesion for up to 16 weeks post-infection ( Figure 1A and data not shown). Consistent with our previous observation [12], spleens and draining lymph nodes (dLNs) from 13 wk-infected mice contain low but detectable number of IFN-c-producing cells ( Figure 1B, left panel), following short-term (3 days) in vitro restimulation with SLA. Furthermore, CD4 + T cells from lpg2-infected mice (but not those from naïve mice) proliferated strongly in response to SLA stimulation in vitro, indicating that parasites-specific memory T cells are maintained even in the absence of cutaneous lesions ( Figure 1B, right panel).
BALB/c mice do not naturally heal WT L. major infections, which makes it difficult to compare memory responses following lpg22 and WT L. major infections in this mouse strain. Therefore, we utilized C57BL/6 mice to investigate and compare the quality of memory response following infection with WT and lpg22 L. major. As shown in Figure 2A and consistent with our previous report [27], lpg22infected C57BL/6 mice did not develop any lesion while mice infected with WT parasites developed lesions that healed by 12 weeks post-infection. Sixteen (16) weeks after infection, both WT and lpg22 infected C57BL/6 mice contain comparable numbers of parasites (,1000) in their footpads ( Figure 2B). However, the dLNs (data not shown) and spleens of mice infected with lpg22 parasites contained significantly less IFN-c-( Figure 2C), IL-4 and IL-10 ( Figure S1A and S1B)producing cells and produced less IFN-c ( Figure 2D) and IL-10 ( Figure 2E) than those from WT L. major-infected mice following in vitro restimulation with SLA.
Because WT and lpg22 L. major differentially activate infected dendritic cells [28], we tested whether the impaired IFN-c response in lpg22 L. major-infected mice was related to suboptimal antigen presentation following SLA restimulation. Therefore, we restimulated cells from lpg22 infected mice in vitro with L. majorinfected bone marrow-derived dendritic cells (BMDCs) and assessed proliferation and IFN-c production. Similar to SLA stimulation, proliferation and IFN-c production by T cells from lpg2infected mice were significantly lower than those from WTinfected mice ( Figure 2F and 2G). Interestingly, the frequency of TNF-producing T cells from WT and lpg22 infected mice following in vitro restimulation with infected DCs was comparable ( Figure S2), suggesting that the difference in IFN-c recall responses was not related to global impairment or reduction in numbers of Leishmania-specific memory cells in lpg2infected mice. Collectively, these results suggest that lpg22 parasites are less effective in inducing IFN-c-producing memory T cell responses than WT parasites.

Impaired Proliferation and IFN-c Production by T cells from lpg22 L. major-infected Mice are Maintained in vivo
To further determine whether the impaired secondary (proliferation and IFN-c production) responses observed in T cells from lpg22 L. major infected mice are strictly related to T cells and in a more physiological environment, we adoptively transferred equal numbers of CFSE-labeled highly enriched T cells from healed WT and lpg22 L. major-infected or naive Thy1.2 (CD90.2 + ) mice into naïve congenic Thy1.1 (CD90.1 + ) recipients and challenged them with WT L. major. Challenged mice were sacrificed on days 5, 14 and 21 post-challenge and donor (CD90.2 + ) T cells from the dLNs and spleens were assessed for cell proliferation and IFN-c production by flow cytometry (Figure 3A). At all time points tested, cells from lpg2-infected mice were less proliferative ( Figure 3B) and produced significantly (p,0.05) less IFN-c ( Figure 3C) in response to virulent L. major challenge than those from WT-infected mice, suggesting that the defect in memory responses observed in lpg2-infected mice is strictly related to T cells.

Quantitative Differences in Memory T cells from WT and lpg22 L. major-infected Mice
To directly determine whether there are quantitative differences in numbers of memory T cells in mice infected with WT and lpg22 parasites, we assessed the expression of CD62L and CD44 on T (CD3 + ) cells from WT-and lpg2-infected mice directly ex vivo by flow cytometry. CD44 is a marker of previous T cell activation and hence is expressed by all memory T cells [29] whereas CD62L is a lymph node homing receptor for lymphocytes, which is downregulated upon lymphocyte activation. These markers discriminate between central memory-like T cells (CD44 hi CD62L hi , Tcm) and effector memory-like T cells (CD44 hi CD62L lo , Tem) [24]. Our direct ex vivo results show that the percentages of CD4 + memory-like T cells (Tcm and/or Tem populations) in the draining lymph nodes of lpg22infected mice were much lower than those from WT-infected mice ( Figure 4A).
Next, we used the highly sensitive adoptive transfer studies to determine whether there were differences in CD62L expression on proliferating (Leishmania-experienced) donor cells from WT and lpg22 L. major-infected mice in vivo. At day 14 post-challenge, donor CD3 + T cells from both groups proliferated and downregulated their CD62L expression ( Figure 4B), although these events were more pronounced in cells from WT-infected mice. Thus, despite lower proliferative response, Leishmania-reactive cells from lpg2-infected mice could downregulate their CD62L expression, suggesting that they could potentially home to the site of infection to mediate effector functions.

Infection with lpg2 2 L. major Protects Against Virulent Challenge Despite Poor DTH Response
To test whether cells from lpg22infected mice could confer protection to naïve mice following adoptive transfer, we challenged WT and lpg2infected C57BL/6 mice with virulent WT L. major parasites and after 72 hr, measured delayed-type hypersensitivity (DTH) response. We found that WT L. majorinfected mice exhibited strong DTH response whereas lpg2infected mice did not exhibit any significant DTH response following challenge ( Figure 5A). Interestingly, despite the impaired proliferation, IFN-c and DTH responses, lpg2-infected mice displayed comparable protection to WT-infected mice ( Figure 5B). Furthermore, adoptive transfer of highly purified T cells from both WT and lpg2-infected mice conferred comparable protection to naïve mice against virulent L. major challenge ( Figure 5C). Taken together, these results indicate that despite quantitative differences in recall responses, cells from lpg22infected mice are qualitatively as efficient as those from WT-L. major-infected mice in mediating secondary protective immunity.
Protection in lpg22infected Mice is Dependent on IFN-c but is Independent of CD8 + T cells We previously showed that infection with lpg22 L. major induced a strong primary CD8 + T cell proliferation and IFN-c production [28]. Since CD8 + T cells are important in both draining lymph node cells and splenocytes from lpg22 infected BALB/c mice were labeled with CFSE, restimulated in vitro with SLA for 5 days, stained intracellularly for IFN-c and assessed for proliferation by flow cytometry, the numbers showed on dot plots represent percentages in total T cell population (B). Data presented is a representative of 3 independent experiments with similar results. doi:10.1371/journal.pone.0066058.g001 Phosphoglycans Regulate Immunity to L. major PLOS ONE | www.plosone.org primary [23,30] and secondary [31] anti-Leishmania immunity, we speculated that secondary resistance in lpg2-infected mice might be dependent on CD8 + T cells. As shown in Figure 6A and 6B and similar to WT infection, depletion of CD8 + T cells had no effect on lpg2-induced protection, suggesting that CD8 + T cells are dispensable for both WT and lpg22 L. major-induced immunity. Because lpg2-infected mice were protected against secondary L. major challenge despite showing significantly impaired IFN-c recall response in vitro and in vivo, we investigated whether protection following lpg2-infection is independent of IFN-c. The data in Figure 6C and 6D show that IFN-c is stringently required for lpg2-induced protection as neutralization Protection in lpg2 L. major-infected Mice is not due to Intact TNF Production In addition to IFN-c TNF has been shown to play a critical role in resistance to L. major [32]. Unlike the impaired IFN-c  expression, the percentage of TNF-expressing CD4 + T cells from lpg2infected mice was comparable to those from WT-infected mice ( Figure 7A) and these cells produced similar amounts of TNF in cultures ( Figure 7B). Interestingly, the majority (.55%) of cytokine-producing cells in WT-infected mice co-expressed IFN-c and TNF, suggesting that polyfunctional cells predominate in mice infected with WT but not in those infected with lpg2 2 parasites. However, treatment of lpg2-infected mice with Embrel (soluble TNFR-Ig to block binding of TNF to its cellular receptors) prior to and during secondary L. major challenge did not affect lpg2induced resistance ( Figure 7C), suggesting that lpg2-induced resistance is not mediated by TNF.

Discussion
In this study, we investigated the correlates and possible mechanism of lpg2mediated protection in murine cutaneous leishmaniasis. We demonstrated that lpg2-infected mice contain Leishmania-reactive (memory) cells that rapidly proliferate and produce effector cytokines (IFN-c and TNF) in response to Leishmania antigen stimulation in vivo and in vitro. However, lpg2parasites were less effective than WT parasites in inducing and/or maintaining Leishmania-specific memory T cells. Nevertheless, memory T cells generated by lpg2parasites were capable of mediating comparable protection against virulent L. major challenge, suggesting that lpg2induced memory cells are qualitatively and functionally comparable to those induced by WT parasites. Depletion and neutralization studies with mAbs demonstrated that akin to WT parasites, lpg2-L. major-mediated resistance was strongly dependent on IFN-c, but independent of CD8 + T cells.
Because of its critical role in resistance to intracellular pathogens, the production of IFN-c by T cells is widely used as a parameter for assessing vaccine efficiency [33][34][35]. However, our in vitro and in vivo data demonstrated that cells from lpg2-infected mice produced significantly less IFN-c, yet these mice were strongly protected against virulent L. major challenge. These observations suggest that other factors contribute to secondary immunity against virulent L. major challenge in lpg2infected mice. Apart from IFN-c, tumor necrosis factor (TNF) has also been shown to play important role in protective immunity against leishmaniasis [36,37]. Indeed, the frequency of TNF-producing cells from lpg2-infected mice was comparable to those from WTinfected controls ( Figure 7A and 7B). Neutralization of TNF signaling did not affect resistance of lpg2 2 infected mice to virulent L. major challenge, suggesting that TNF does not compensate the defective IFN-c production in lpg2infected mice. Interestingly, TNF neutralization enhances parasite control in lpg2-L. majorinfected mice following virulent L. major challenge. This apparent paradox may be related to reduced inflammation following TNF neutralization and consequent reduction in macrophage recruitment, which would reduce the number of cells available for parasites to infect.
Although a robust IFN-c response is important for primary and secondary resistance to L. major [38,39], recent studies suggest that other factors distinct from the magnitude of IFN-c response might  [40]. In addition, impaired Treg expansion in p110d deficient mice (and not enhanced IFN-c response) contributed to the hyper-resistance of these mice to L. major infection [41]. These observations suggest that in the absence of Treg activation, low levels of IFN-c can efficiently activate macrophages leading to effective intracellular parasite killing in vivo. However, we did not find any difference in the percentage or absolute numbers of Tregs in both spleens and dLNs of WT and lpg2-infected mice (data not shown), suggesting that lower activation of Tregs does not account for the effective primary and/or secondary immunity in lpg2-L. major infected mice in the presence of lower IFN-c response. Interestingly, we found that both primary [12] and secondary ( Figure 2E) infections with lpg2 2 L. major result in suppressed IL-10 response. It is conceivable that akin to L. major infection in p110d deficient mice [41] and infection with L. major clone SD [40], the low IL-10 response could permit low levels of IFN-c to more efficiently activate macrophages leading to effective parasite killing in vivo. Thus, although lpg2-L. major-infected mice also produce significantly less IFN-c, it is conceivable that low levels of IFN-c may be more efficient at activating macrophages for more efficient parasite control when IL-10 levels are also correspondingly low. In line with this, we found that neutralization of IFN-c abolished lpg2-induced immunity, suggesting that this low level of IFN-c is nonetheless required for effective parasite control following secondary virulent challenge.
The development of Leishmania vaccine is a global public health priority because of the enormous morbidity and mortality associated with the disease. Unfortunately, there is currently no clinical approved vaccine for human cutaneous leishmaniasis [42,43]. Interestingly, recovery from natural or experimental infection leads to long-lasting protective immunity against reinfection, an observation that formed the basis for leishmanization still practiced in many countries today [44]. However, the significant morbidity associated with this practice has hampered its acceptance as a vaccination strategy. To overcome these problems, a number of live attenuated mutant parasites have been generated [11,13,[45][46][47]. lpg2parasites have several advantages as a potential live-attenuated vaccine because it does not cause disease in most cases even in the highly immunocompromised (SCID) mice [12]. Although a revertant line of lpg2parasites has been described previously [19], we have yet to detect revertant line in our studies using C57BL/6 mice, suggesting that this may either be a very rare event restricted to the highly susceptible BALB/c mice or due to latent infections or differences in gut microflora resulting from differences in housing environments, a factor that has been shown to affect infection outcome even in animals with same genetic background [48][49][50]. In addition, altered microflora in mice from different vendors (JAX versus The University of Manitoba Central Animal Care Services [originally derived from Charles River, Montreal, Canada]) might also account for the differences from our previous observation regarding the requirement of CpG as adjuvant for generating protective immunity following vaccination with lpg2parasites [27]. Indeed, Dr Beverley's lab has recently completed a sequence comparison of our strain with theirs (where reversion has been observed) and preliminary analysis show that both strains are genetically the same (i.e. still lack the LPG2 gene), further implicating environmental and/or epigenetic factors as the primary cause of these differences. Furthermore, similar to WT mice, lpg2parasites persist indefinitely in vaccinated mice [13], which eliminates the need for repetitive inoculations. Lastly, lpg2parasites protect vaccinated host against virulent challenge without inducing ''nasty'' DTH response. We refer DTH as being ''nasty'' because a huge vaccination-induced swelling (DTH response) on the face following a bite from an infected sandfly would be undesirable.
Overall, our results show that despite poor DTH and IFN-c recall responses, lpg2-L. major parasites induced protective immunity in both BALB/c and C57BL/6 mice is qualitatively comparable to those of WT parasites. We hypothesize that the excellent protection observed in these mice is related to more efficient IFN-c activity in the presence of low IL-10 response. Our findings lend support for the consideration of lpg2parasites as liveattenuated vaccine or leishmanization candidates against cutaneous leishmaniasis, particularly in parts of the world where leishmanization is still practiced with virulent parasites. This would at least reduce the morbidity associated with using virulent organisms for leishmanization since lpg2parasites do not cause any disease. We are currently examining the pathogenesis of lpg2parasites in non-human primates in order to determine whether infection with this avirulent mutant parasite could also confer protection against virulent challenge. Figure 7. Resistance in lpg2 2 L. major-infected mice is not dependent of TNF. CD3 + spleen cells from WT-and lpg2 2 L. major-infected mice were stimulated in vitro with infected BMDCs for 24 hr and assessed for IFN-c and TNF expression by flow cytometer. The culture supernatant fluids were assayed for TNF by ELISA (B). Infected mice were treated with TNFR-Ig (Embrel) 24 hr prior to challenge and further treated once weekly for additional 2 weeks. After 3 weeks mice were sacrificed and parasite burden was determined (C). Data presented are representative of 2 independent experiments (n = 3-5 mice per group) with similar results. doi:10.1371/journal.pone.0066058.g007

Supporting Information
Figure S1 Impaired IL-4 and IL-10 recall response by spleen cells from lpg2-infected mice. C57BL/6 mice were infected with WT and lpg2-L. major and after 16 weeks, mice were sacrificed, the spleen cells were restimulated in vitro with SLA for 72 hr and the frequency of IL-4-(A) and IL-10 (B)-producing cells was determined by flow cytometry. (TIF) Figure S2 lpg2-infected mice are not impaired in their TNF recall response. C57BL/6 mice were infected with WT and lpg2-L. major and after 16 weeks, mice were sacrificed, the spleen cells were restimulated in vitro with infected BMDCs for 72 hr and the frequency of TNF-producing cells was determined by flow cytometry. (TIF)