In Vitro Evaluation of a Soluble Leishmania Promastigote Surface Antigen as a Potential Vaccine Candidate against Human Leishmaniasis

PSA (Promastigote Surface Antigen) belongs to a family of membrane-bound and secreted proteins present in several Leishmania (L.) species. PSA is recognized by human Th1 cells and provides a high degree of protection in vaccinated mice. We evaluated humoral and cellular immune responses induced by a L. amazonensis PSA protein (LaPSA-38S) produced in a L. tarentolae expression system. This was done in individuals cured of cutaneous leishmaniasis due to L. major (CCLm) or L. braziliensis (CCLb) or visceral leishmaniasis due to L. donovani (CVLd) and in healthy individuals. Healthy individuals were subdivided into immune (HHR-Lm and HHR-Li: Healthy High Responders living in an endemic area for L. major or L. infantum infection) or non immune/naive individuals (HLR: Healthy Low Responders), depending on whether they produce high or low levels of IFN-γ in response to Leishmania soluble antigen. Low levels of total IgG antibodies to LaPSA-38S were detected in sera from the studied groups. Interestingly, LaPSA-38S induced specific and significant levels of IFN-γ, granzyme B and IL-10 in CCLm, HHR-Lm and HHR-Li groups, with HHR-Li group producing TNF-α in more. No significant cytokine response was observed in individuals immune to L. braziliensis or L. donovani infection. Phenotypic analysis showed a significant increase in CD4+ T cells producing IFN-γ after LaPSA-38S stimulation, in CCLm. A high positive correlation was observed between the percentage of IFN-γ-producing CD4+ T cells and the released IFN-γ. We showed that the LaPSA-38S protein was able to induce a mixed Th1 and Th2/Treg cytokine response in individuals with immunity to L. major or L. infantum infection indicating that it may be exploited as a vaccine candidate. We also showed, to our knowledge for the first time, the capacity of Leishmania PSA protein to induce granzyme B production in humans with immunity to L. major and L. infantum infection.


Introduction
Leishmania (L.) are intracellular protozoa that cause a wide spectrum of human diseases, ranging from self-healing cutaneous to lethal visceral leishmaniasis. Cutaneous leishmaniasis (CL) is widely distributed in the Americas, North Africa, the Middle East and Central Asia, causing considerable morbidity [1]. In Tunisia, CL is zoonotic and is mainly due to L. major [2]. In Peru, L. braziliensis is responsible for nearly all the zoonotic CL cases [3].
Anthroponotic Visceral Leishmaniasis (VL) due to L. donovani is mainly distributed in the Indian subcontinent with over 300 000 annual cases [4], whereas zoonotic VL caused by L. infantum is present mainly in Mediterranean basin, Middle east, central Asia and Brazil with up to 50 000 annual cases [5][6][7]. Treatment of leishmaniasis mainly relies on chemotherapy and the control of the disease is challenged by serious side effects of existing drugs and the widespread emergence of drug resistant strains. Several lines of evidence are in favor of the feasibility of a vaccine in humans.
Healing is generally associated with the development of a life-long immunity to re-infection. Moreover, a large majority of immunocompetent exposed individuals develop asymptomatic or subclinical infections rather than a severe form of leishmaniasis (symptomatic) and acquires a robust and durable immunity to reinfection. Resistance to infection is based on a Th1 dominant response with interferon (IFN)-c production resulting in macrophage activation and parasite killing [8][9][10]. In healed individuals, it has been shown that recovery and resistance to re-infection also correlate with the development of antigen-specific Th1 cell responses and IFN-c production [11][12][13][14]. Based on these data, one vaccine strategy developed against Leishmania infection has been focusing on the use of Th1 responses elicited in immune individuals by defined parasite antigens as indicators of protection. A number of leishmanial proteins have been characterized and evidences from studies in animal models indicate that variable levels of protection can be achieved using defined protein as vaccines [15][16][17]. However, significant differences between mice and humans immune system, in both its innate and adaptive arms exist [18]. Very few candidates have progressed beyond the experimental stage and designing an effective vaccine against leishmaniasis is still a matter of research. One of the candidate vaccine antigens is the PSA which belongs to a unique Leishmania family of membrane-bound and secreted proteins. PSA main signature is a specific Leucin Rich Repeats known to be involved in protein-protein interactions and in pathogen recognition [19]. It is an abundant glycolipid-anchored protein on the surface of the promastigote form of most Leishmania species [19][20][21][22][23][24][25][26]. A watersoluble form released in promastigote culture supernatants of L. major [24] and of L. infantum [26] has also been described. PSA is involved in parasite attachment and invasion of macrophages [27]. It has been reported to induce Th1-mediated protection against murine leishmaniasis when used as a vaccine [28,29]. PSA is also specifically recognized by Th1 cells of humans with a history of self-healing CL [30] and by human sera of VL patients [31]. This protein is present in both promastigote and amastigote stages [32,33] and the presence of PSA family was demonstrated in several Leishmania species [19,21,34,35]. PSA proteins are strongly over-expressed in metacyclic promastigotes suggesting an association with the virulence status of the parasites [32]. The known role of PSA proteins in resistance to complement lysis further supports this hypothesis [36]. The interest for the PSA is also supported by our previous [37,38] and recent studies (unpublished data) showing that sera of dogs vaccinated with L. infantum excreted/secreted antigens (ESA) in combination with MDP (muramyl dipeptide), recognized immunodominant antigens belonging to PSA protein family, from L. infantum ESA (LiPSA-54S; GenBank under Accession number: FJ974055) and from L. amazonensis ESA (LaPSA-38S; GenBank accession number: FJ974054). In the present study, the recombinant native form of L. amazonensis PSA protein (LaPSA-38S), produced in a L. tarentolae expression system was evaluated for its ability to induce cellular and humoral immune responses. We took advantage of a multidisciplinary consortium to assess, using the same tools and samples from individuals with acquired immunity to L. major, L. braziliensis, L. infantum or L. donovani, for PSA immunogenicity and its potential use as a vaccine candidate against the different forms of leishmaniasis.

Ethics statement
The recruitment and sampling collection of different groups of volunteers, was done in accordance with Good Clinical Practice (GCP), based on the recommendations and approval of the local ethical Committee of each Institute involved in this study (Institut  de Recherche pour le Développement (IRD) for France (the  recruitment of blood donors and the sample collection were  performed at the EFs human blood bank of Bordeaux), Instituto  de Salud Carlos III (ISCIII) for Spain (Comité de ética de la investigacion y de bienestar animal), National Institute of Pathology (NIP) for India (Office of ethical committee VMMC & SAFDARJANG hospital New Delhi and Indian council of medical research), Instituto de Medicina Tropical ''Alexander von Humboldt'' (IMTAvH) for Peru (comité de investigacion y ética de la unidad de capacitacion) and Institut Pasteur de Tunis (IPT) for Tunisia (Comité d'éthique de l'Institut Pasteur de Tunis). A written informed consent was obtained from all subjects involved in this study.

Study population and samples
Well-characterized endemic foci, blood banks and medical structures (hospitals, medical departments and public health centers) were identified in each country (India, Peru, Spain, France, and Tunisia). Human groups (cured individuals and healthy individuals with no history of leishmaniasis but with a probable asymptomatic infection and therefore considered as immune) were recruited from endemic areas for VL and CL, based on the following defined inclusion criteria (i) living in endemic foci to L. major, L. infantum, L. braziliensis or L. donovani (ii) well documented medical records for cured groups (iii) presence of typical scars for cured CL group and (iv) high IFN-c response to Soluble Leishmania Antigens (SLA) (.300 pg/ml) for healthy individuals with a probable asymptomatic infection. Patients with active CL or VL were identified in health structures in charge of patient management and had specific clinical symptoms. Healthy individuals recruited in low or non-endemic areas, or in blood banks with no history of leishmaniasis and no or low IFN-c response to SLA (,100 pg/ml) were considered as non immune/ naives. Exclusion criteria were immunosuppressive diseases other than leishmaniasis, long term treatment and pregnancy. The different human groups used in this study are detailed in table 1. Heparinized blood was collected from a total of 104 donors and 82 controls (Table 1).

Production and purification of LaPSA-38S recombinant protein
Genomic DNA of L. amazonensis (MHOM/BR/76/LTB-012/ clone 1), cultivated in RPMI 1640 medium (Gibco-BRL, UK) and supplemented with 10% heat-inactivated FBS (Gibco-BRL, UK), 10 mM HEPES, 100 U/ml penicillin, and 100 mg/ml streptomycin (Gibco-BRL, UK) at 26uC, was extracted by phenol: chloroform: isoamyl alcohol extraction (1:1:1) and ethanol precipitation. The LaPSA-38S gene (GenBank accession number: FJ974054) was amplified with forward primer, F-PSA-38S (59-CCATGGCGCAGTGCGTGCGTCGG-39) and reverse primer, R-PSA-38S (59-GCGGCCGCGTGATGGTGATGGTGAT-GATCGTGGTTCGCCAG-39), containing NcoI and NotI restriction sites in each 59 end (underlined) by PCR reaction. The purified PCR product was cloned in pCR2.1-TOPO TA vector using TOPO TA cloning Kit (Invitrogen) following the manufacturer's procedures. The transformed cells E. coli TOP10 (Invitrogen) were screened for the presence of recombinant plasmid with the LaPSA-38S insert by gene-specific PCR and analysis with NcoIand NotIrestriction enzymes and isolated positive clones were sequenced. The insert was removed by NcoIand NotI digestion and subcloned into the NcoI and NotI insertion site of Leishmania expression vector pF4X1.4sat1 allowing selection with the antibiotic Nourseothricin (Jena Bioscience, Germany) to create the recombinant pF4X1.4-LaPSA-38S plasmid. The resulting construct, encoding a full sequence of LaPSA-38S secreted protein fused to a C-terminus His 6 tag. L. tarentolae promastigotes were successfully maintained in continuous culture by successive passages of 5610 5 flagellates/ml every week into 10 ml of completely chemically defined medium (CDM/LP) free of serum, macromolecules, proteins and cell contaminants as previously described [39,40]. For stable integration of the expression cassette into the 18S ribosomal RNA (ssu) locus, 10 mg of pF4X1.4sat1 plasmid containing LaPSA-38S gene was digested by SwaI restriction enzyme. Transfections of L. tarentolae promastigotes were performed by electroporation in 2 mm cuvettes using a Gene Pulser II (Biorad) and a single pulse (5-6 msec) with the settings 450 V and 450 mF. After pulse, cells were transferred into a fresh CDM/LP medium. 100 mg/ml of Nourseothricin was added 24 h after electroporation to select stable transformants. After a week of culture only nourseothricin-resistant cells survived. Transgenic cells were selected as single colonies on the supplemented CDM/ LP-agar medium containing 100 mg/ml Nourseothricin. To confirm the integration of the LaPSA-38S containing cassette into the Leishmania genome, PCR was performed on the genomic DNA of wild type and transgenic cells with ssu forward primer F3001 and aprt reverse primer A1715 (Jena bioscience, Germany). Mass culture and protein purification were manufactured by Virbac in GMP conditions. Culture amplifications were performed on CDM/LP medium. The culture supernatant containing L. amazonensis excreted secreted LaPSA-38S, released by the parasite during its growth, was recovered at the late stationary phase of growth. The only proteins contained in the medium were of parasitic origin and were found with their native conformation because they were naturally excreted-secreted by Leishmania parasites. Excreted secreted PSA was purified from concentrated culture supernatant by Ni-NTA affinity chromatography.

Preparation of total soluble Leishmania promastigote antigens (TSLA)
All antigen extracts were prepared from promastigote stationary phase parasite cultures of L. donovani, L. braziliensis, L. major and L. infantum. Each consortium partner used local TSLA (L. major TSLA in Tunisia, L. infantum TSLA in France and Spain, L. braziliensis TSLA in Peru and L. donovani TSLA in India) and TSLA from L. donovani (TSLA Ldd8) (MHOM/IN/80/Ldd8Cl 2 L. donovani strain (Ldd8)), as a standard preparation. TSLA were obtained from washed parasites in 1x phosphate-buffered saline (PBS), centrifuged at 10006g/10 min at 4uC and supernatants were removed. The pellets were resuspended in lysis buffer (50 mM Tris/5 mM EDTA/HCl, pH7. 1 mL/1610 9 parasites), subjected to three rapid freeze/thaw cycles followed and to three pulses of 20 seconds/40W with sonicator. Samples were centrifuged at 50006g for 20 min at 4uC, and supernatants were collected, aliquoted and stored at 280uC until use. Protein quantification was performed using Bradford method.

Cell culture and stimulation
Peripheral blood mononuclear cells (PBMC) were isolated from blood by density centrifugation through Ficoll-Hypaque (GE Healthcare Bio-Sciences AB, Uppsala, Sweden).

Cytometric Bead Array assay (CBA)
IFN-c, granzyme B, tumor necrosis factor-alpha (TNF-a) and Interleukin-10 (IL-10), were detected and quantified from culture supernatants (50 ml) of PBMC exposed for 120 h to the following stimuli: PHA-M (10 mg/ml), TSLA (10 mg/ml) or LaPSA-38S (10 mg/ml), using the BD CBA Human Soluble Protein Flex Set system and according to the instructions of the manufacturer (BD Biosciences). It should be noted that granzyme B was not analyzed in individuals immune to L. braziliensis infection. In order to quantitate samples, the BD CBA Human Soluble Protein Flex Standard was performed for each cytokine and in each experiment. Data were acquired by flow cytometry (FACSCalibur or FACSCanto) using 2-color detection. Flow Cytometric Analysis Program Array (FCAP Array; BD Biosciences) software was required for analyzing samples [41].

Cell phenotyping
Freshly isolated PBMC were stimulated with PMA (50 ng/ml)/ ionomicyn (10 26 M) or PHA at 10 mg/ml for 6 h (positive control) or with TSLA at 10 mg/ml or with LaPSA-38S at 10 mg/ml for 120 h and, or kept with medium alone. Cells were treated with Golgistop (BD Biosciences) for the last 6 hours of culture, then washed and incubated with antibodies: FITC CD3, PerCPcy5.5 CD4, APC-H7 CD8 or PerCPcy5.5 CD8, and PE-Cy7 CD69 or PE CD69 (BD Biosciences), for 20 minutes at 4uC. For intracellular IFN-c detection, cells were fixed and permeabilized using BD Cytoperm/cytofix plus kit (BD Biosciences) according to manufacturer's instructions and labeled with PE-anti-IFN-c mAb (intracellular formulation) (BD Biosciences). Analysis was performed with FACSCalibur using CellQuest Pro software or FACS canto II flow cytometer using DIVA software. FITC mouse IgG1, PerCPCy5.5 mouse IgG1, mouse IgG1PE, PE mouse IgG2a (intracellular formulation) (Biosciences) were used as isotype controls for acquisition by FACSCalibur. For acquisition by FACSCanto, BD CompBeads Set Anti-Mouse Ig, k (Anti-Mouse Ig, k/Negative Control (FBS) Compensation Particles Set) (BD Biosciences) were used.

Detection of antibodies specific for LaPSA-38S
An enzyme-linked immunosorbent assay (ELISA) test was performed to detect specific antibodies to TSLA Ldd8 and LaPSA-38S in individual serum samples from the different studied groups ( Table 2). In order to improve antibodies detection, ELISA protocol was optimized by combining antigens concentrations and anti-serum dilutions. Briefly, 96 wells plates (NuncMaxisorp Immuno Plates, PolyLabo, Strasbourg, France) were coated with 100 ml/well of 2 or 5 mg/ml LaPSA-38S and TSLA overnight at 4uC. Plates were washed three times with PBS, 0.1% Tween 20 (PBS-T), pH 7.4 and blocked with 200 ml/well of PBS containing 0.1% Tween 20 and 3% BSA for 1 h at 37uC. After washing with PBS-T, diluted sera (1/100, 1/150 and 1/200 in PBS-T) were added (100 ml/well) and incubated for 1h30 at 37uC. Plates were then washed with PBS-T and 100 ml/well of diluted HRPconjugated anti-human IgG in washing buffer were added for 30 min at 37uC (1/5000). Plates were developed with the 100 ml/ well of Sigma Fast o-phenylene diamine dihydrochloride (OPD) for Phenotypic analysis of LaPSA-38S and SLA-specific CD4+ and CD8+ was performed for PBMC among CD3+ population, in groups in which we observed a significant and specific LaPSA-38S-induced IFN-c response (CCLm: n = 7, HHR-Lm: n = 6, HHR-LiS: n = 7). Naive individuals (HLR-S: n = 14) were also analyzed. Data are means percentage 6 SD.
doi:10.1371/journal.pone.0092708.t002 15 min. The reaction was stopped with 50 ml/well of 2 N HCl. The absorbance was measured at 492 nm. All sera were tested in duplicate/triplicate, and the mean value was recorded.

Statistical analysis
Data from all partners were collected into a unique database using the open source software Epidata 3.1 (Epidata Association, Denmark http://www.epidata.dk/) and checked for consistency. Individuals with missing data or exclusion criteria were deleted. Data analysis was then performed with Stata statistical software (StataCorp. 2009. Stata Statistical Software: Release 11. College Station, TX: StataCorp LP.). We used non-parametric statistical tests because of the low number of individuals in some groups and of the heterogeneity of the standard deviation between groups. Results are expressed as mean6standard deviation (SD) and a pvalue of ,0.05 was considered significant in all cases. In the statistical analysis of cytokine level and phenotyping, we used Wilcoxon signed-rank test to compare median levels of cytokines or percentage of cells producing IFN-c after different stimulations in paired data. Kruskal-Wallis rank test was used for inter-groups analysis on normalized data after deducting the non-stimulated value. Correlations were performed using pairwise correlations of variance test. For humoral response analysis, we used Kruskal-Wallis rank test to compare the median production of total IgG between different stimulations. We estimated a threshold of positivity for each country by calculating a cut-off using the following formula: Individuals with a level of IgG after stimulation higher than t pos were considered as positive responders. For the statistical analysis of sub-classes, results were compared only for positive individuals

Production of LaPSA-38S in L. tarentolae
In this study, we generated a recombinant L. tarentolae strain expressing exogenously the L. amazonensis LaPSA-38S gene. The resulting construct encoded a full sequence of LaPSA-38S secreted protein fused to a C-terminus His 6 tag. LaPSA-38S protein purity was assessed by SDS-PAGE using 12% polyacrylamide Ready Gels (Bio-Rad) under reducing and non reducing conditions and after silver staining. Immobilized Ni 2+ affinity chromatography of culture supernatant containing excreted/secreted extracts from recombinant L. tarentolae, yielded a preparation featuring a purified recombinant LaPSA-38S protein migrating as a 45 kDa species in SDS-PAGE (Figure 1). The extra band observed in figure 1 was identified as a dimer form of LaPSA-38S as suggested by a proteomic study of purified LaPSA-38S using a combination of 1D gel electrophoresis and LC-MS/MS (coupling between liquid chromatography and tandem mass spectrometry) (Bras-Gonçalves et al., manuscript in preparation).  TSLA Ldd8 stimulation (Figure 2a and b). It should be noted that no significant differences were observed in IFN-c and granzyme B responses between local TSLA and TSLA Ldd8 (Figure 2a and b). Significant TNF-a levels (p#0.02) were observed in CCLb  (Figure 3d). IL-10 levels were significantly higher in CCLm, HHR-Lm and HHR-LiT compared to corresponding naive groups (HLR-T) (p# 0.009). No IL-10 was observed in response to TSLA. No difference was observed in the cytokine levels between immune groups, except for LaPSA-38S-induced granzyme B levels which were significantly higher in HHR-LiS compared to CCLm and HHR-Lm (p = 0.02 and 0.04, respectively).
Interestingly, strong positive correlations were found between IFN-c and granzyme B production in CCLm group (r = 0.82; p = 0.0000) and between granzyme B and IL-10 production in HHR-LiS group (r = 0.7; p = 0.03), in response to LaPSA-38S. IFN-c and granzyme B productions were also highly correlated in response to TSLA Ldd8 in CCLm and HHR-Li groups (r = 0.67; p = 0.0002 and r = 0.76; p = 0.01; respectively).

Humoral response to LaPSA-38S
Anti-LaPSA-38S total IgG responses were analyzed by using ELISA tests in sera from (i) patients with active disease (aVLd, aVLiT, aVLiS, aCLb), (ii) cured individuals (CVLd, CCLm, CCLb), (iii) healthy individuals with high IFN-c response to TSLA Ldd8 as immune with no clinical symptoms (HHR-Lm, HHR-LiT, HHR-LiF, HHR-LiS) and healthy individuals with low IFN-c response to TSLA Ldd8 as naïve controls ( Table 3). The cut-off value of reactivity for LaPSA-38S antigen was calculated as the mean plus 3 SD of the OD values observed in naive controls from each country. These cut-off values allowed us to distinguish between positive and negative sera and consequently to estimate performance parameters of each ELISA test (Table 3). Globally, the results indicated that LaPSA-38S antigen does not constitute a target for the humoral response in the different studied groups. The percentage of positivity was very low ranging from 0% to 12.5%. However, an exception was observed with the aVLiS group where LaPSA-38S was recognized by 50% of tested sera. In contrast and as expected, highest positivity was observed using TSLA antigen. Best results were obtained with sera from individuals with active VL diseases which 100% react with TSLA Ldd8.

Discussion
Few vaccine candidates have progressed beyond the experimental stage in animal models of leishmaniasis (mice, hamsters). This is consistent with the difficulty in developing a suitable animal model that reproduces the features of the human disease. Direct extrapolation of data from animal models to human disease is often controversial and the protective immune responses controlling experimental infection may not reflect those required to prevent leishmaniasis in endemic areas. The purpose of this study was to evaluate the ability of a PSA antigen to elicit an ex vivo cellular immune response in PBMC of non exposed and previously exposed human subjects with and without evidence of protective immunity. The rationale is that antigens that elicit immune responses correlating with protection or cure (e.g. significant IFN-c production in previously exposed and cured individuals) may be good candidates for prophylactic and/or therapeutic vaccines.
We have previously demonstrated that naturally ESA purified from culture supernatants of L. infantum promastigotes formulated with an adjuvant were able to induce a long lasting Th1-mediated protection against experimental and natural canine VL [19,32,37,38]. Interestingly, a native soluble L. infantum PSA has been identified as the active constituent of ESA that reproduce the observed protective immunity when used as a vaccine in dogs (unpublished data). PSA is present in both promastigote and amastigote stages and is over-expressed in stationary phase parasites [19,32,33]. It is highly conserved in its N-terminal and C-terminal parts, among Leishmania species, which is a requisite for ensuring cross-species protection (unpublished data). PSA isolated from L. amazonensis promastigotes as well as recombinant vaccinia viruses expressing this protein have been demonstrated to elicit a protective immune response against infection with L. amazonensis in BALB/c mice [28,42]. Moreover, PSA-2 polypeptides purified from L. major and plasmid DNA encoding PSA-2, but not recombinant PSA-2 purified from E. coli, provided significant protection against L. major infection in BALB/c mice [29,43]. Our aim was to identify a human vaccine candidate targeting Leishmania species that cause CL and VL. We used a recombinant native L. amazonensis PSA (LaPSA-38S), as a potential candidate vaccine. Despite the fact that the amino sequence alignment of LaPSA-38S with PSA proteins from several Leishmania species, shows that these proteins are not 100% identical, the N-terminal and C-terminal regions present highly conserved segments. However, the number of Leucine Rich Repeat motifs located in the central region was variable [19]. Nevertheless, these proteins belong to the same unique PSA family according to various well known criteria which constitute the PSA family unique signature [19]. Thus, all PSA might be considered as highly similar regarding their protein primary structure since they correspond to the same architecture depicted by a precise domain organization which is well conserved among PSA from different Leishmania species [19].
LaPSA-38S was produced in a non-pathogenic L. tarentolae recombinant expression system. This expression system allows the conservation of Leishmania-type post-translational modifications including glycosilation, phosphorylation and disulfide bond formation. Moreover, the L. tarentolae expression system, in addition to being non-pathogenic to mammals, is an ideal expression system to produce native proteins of Leishmania [44]. A multi-disciplinary consortium has been set-up including groups from Europe (France and Spain for L. infantum infection) and from different endemic countries where both VL and CL are highly prevalent (India for L. donovani, Peru for L. braziliensis and Tunisia for L. major and L. infantum infections). We demonstrated that LaPSA-38S induced similar significant and specific levels of IFN-c in individuals immune to L. major and L. infantum infection. Healthy high IFN-c responders had equivalent responses to the cured group. A non specific IFN-c response to LaPSA-38S stimulation was observed in L. braziliensis or L. donovani immune individuals. Analysis of the phenotypes of LaPSA-38S-stimulated T cells showed a significant increase in the percentage of IFN-c producing CD4+ T cells detected only in cured individuals. Furthermore, a strong positive correlation was found between the percentage of IFN-c-producing CD4+ T cells and the IFN-c production determined by CBA. Data on the PSA immunogenicity in humans are scarce. One study has shown that PSA from L. amazonensis was able to stimulate T cell proliferation in leishmaniasis patients [45]. To our knowledge, only one study reported that PSA from L. major was able to induce strong proliferative responses and high IFN-c and TNF-a levels in individuals with a past history of self-healing CL [30]. Healing from human leishmaniasis is generally associated with an appropriate Th1 response characterized by IFN-c  IFN-c (Fig. 2a), granzyme B (Fig. 2b), were detected and quantified from culture supernatants of PBMC exposed for 120h to local TSLA (10 mg/ml) and TSLA Ldd8 (10 mg/ml), by Cytokine Beads Array test (CBA) using Flow cytometry. Statistically significant differences between stimulated and non stimulated cultures and between groups (p# 0.03) are showed. doi:10.1371/journal.pone.0092708.g002 Figure 3. LaPSA-38S specific IFN-c, granzyme B, TNF-a and IL-10 responses. IFN-c (Fig. 3a), granzyme B (Fig. 3b), TNF-a (Fig. 3c) and IL-10 ( Fig. 3d) were detected and quantified from culture supernatants of PBMC exposed for 120 h to SLA (10 mg/ml) or LaPSA-38S (10 mg/ml) using Cytokine Beads Array test (CBA). Data were analyzed by Flow cytometry. PHA (10 mg/ml) was used for all cultures as positive control (data not shown). Statistically significant differences between stimulated and non stimulated cultures (p#0.003) and between groups (p#0.01) are showed. doi:10.1371/journal.pone.0092708.g003  The cut-off value of reactivity for LaPSA-38S antigen was calculated as the mean plus 3 SD of the OD values observed in naive controls from each country. These cut-off values allowed us to distinguish between positive and negative sera and consequently to estimate performance parameters of each ELISA test. doi:10.1371/journal.pone.0092708.t003 production. Thus, our finding showing that LaPSA-38S stimulates high levels of IFN-c in individuals with immunity to L. major or L. infantum infection, suggests that this protein may be associated with protective immunity. With regard to IL-10 production, we showed that, in addition to its Th1 response-inducing activity, LaPSA-38S induces significant IL-10 levels in individuals with immunity to L. major or L. infantum infection, and to a lesser extent in naïve groups. IL-10 is an anti-inflammatory cytokine with broad immuneregulatory functions, produced by different cell populations including dendritic cells, macrophages, Th2 and different regulatory CD4+ T subsets [46]. The regulatory effects of IL-10 on immune responses and pathology have been frequently reported [47]. Heterogeneity of CD4+ T cell cytokine response has complicated the definition of the immune correlates of vaccinemediated protection [48]. Recently, the balance between proinflammatory IFN-c/TNF-a and regulatory IL-10 cytokines has been shown to be involved in the outcome of human leishmaniasis and in the prediction of vaccine success [17]. Therefore, the capacity of LaPSA-38S to induce IL-10 does not exclude this protein as a potential candidate vaccine. We also detected significant and specific levels of granzyme B in L. major and L. infantum immune individuals. Very few reports have investigated the involvement of granzyme B in human leishmaniasis. While an increase of granzyme B activity have been associated with a good prognosis in human CL caused by L. major or L. mexicana, [49,50], CD8+ granzyme B+ T cells have been reported to mediate tissue injury in CL caused by L. braziliensis [51]. Granzyme B is a serine protease expressed in human T CD8+ and T CD4+ cells as well as in Treg cells [52]. It has been first implicated in apoptosis of target cells [53]. Recently, several studies have highlighted a new inflammatory role of granzyme B through the regulation of cytokine expression and processing and regulatory T cells activity [53][54][55][56]. To our knowledge, this is the first description of a Leishmania antigen-induced granzyme B response in humans immune to Leishmania infection. The increase of granzyme B, in response to a heterologous prime boost vaccination of mice using the LACK p36 antigen suggests its contribution to the disease control [57]. Our data also show strong positive correlations between granzyme B and IFN-c levels in response to LaPSA-38S, in individuals with cured L. major CL, which could be associated with protective immunity. We also showed in our study that LaPSA-38S induced a mixed Th1 and Th2/Treg cytokine profile in individuals with immunity to L. major and L. infantum infection but not in individuals with immunity to L. braziliensis or L. donovani. Data on conserved Leishmania antigens enabling cross-species protection were reported in murine models [58][59][60]. However, absence of cross-species protection was more frequently described [61][62][63]. The absence of cross-protection between Leishmania species could be explained by the biological differences between Leishmania species and by the level of sequence homology of the Leishmania proteins used. Indeed, while the L. major and L. infantum proteins-coding sequences are conserved, the new world L. braziliensis proteins-coding sequences are more divergent [64]. Finally, humoral responses analysis results clearly show that LaPSA-38S antigen did not induce a significant antibody response in individuals that display different clinical manifestations and forms of leishmaniasis. These results are in accordance with other studies describing reactivities of human VL sera to PSA antigen [32]. However, in our case, we cannot exclude that this low recognition may be due to the aminoacids sequence divergence between LaPSA-38S (used in our experiment) and PSA from L. major, L. infantum and L. donovani. In conclusion, we have demonstrated here that, in spite of differences in disease forms induced by L. major and L. infantum, the LaPSA-38S protein was able to induce a similar Th1 and Th2/Treg profile in individuals immune to infection by these two species. To our knowledge, this is the first description of PSA-induced IFN-c, granzyme B, TNF-a and IL-10 response in humans with immunity against L. major and L. infantum infection. The capacity of LaPSA-38S to induce mixed Th1 and Th2/T regulatory responses is an argument for the use of this protein as a potential candidate vaccine. Our approach to investigate the ability for a given antigen, to elicit cellular protective responses in human populations from different endemic areas may help to select vaccine candidates and to prioritize antigens for clinical development of a subunit vaccine against leishmaniasis.