NK Cell Activity Differs between Patients with Localized and Diffuse Cutaneous Leishmaniasis Infected with Leishmania mexicana: A Comparative Study of TLRs and Cytokines

Leishmania mexicana causes localized (LCL) or diffuse cutaneous leishmaniasis (DCL). The cause of dissemination in DCL remains unknown, yet NK cells possibly play a role in activating leishmanicidal mechanisms during innate and adaptive immune responses. We had previously shown that Leishmania lipophosphoglycan (LPG) is a ligand for TLR2, activating human NK cells. We have now analyzed NK cells in LCL and DCL patients. NK numbers and effector mechanisms differed drastically between both groups of patients: DCL patients showed reduced NK cell numbers; diminished IFN-γ and TNF-α production; and lower TLR2, TLR1, and TLR6 expression as compared to LCL patients. The altered protein expression found in NK cells of DCL patients correlated with their down-regulation of IFN-γ gene expression in LPG-stimulated and non-stimulated cells as compared to LCL patients. NK cell response was further analyzed according to gender, age, and disease evolution in LCL patients showing that female patients produced higher IFN-γ levels throughout the disease progression, whereas TLR2 expression diminished in both genders with prolonged disease evolution and age. We furthermore show the activation pathway of LPG binding to TLR2 and demonstrated that TLR2 forms immunocomplexes with TLR1 and TLR6. In addition to the reduced NK cell numbers in peripheral blood, DCL patients also showed reduced NK cell numbers in the lesions. They were randomly scattered within the lesions, showing diminished cytokine production, which contrasts with those of LCL lesions, where NK cells produced IFN-γ and TNF-α and were found within organized granulomas. We conclude that in DCL patients the reduced NK-cell numbers and their diminished activity, evidenced by low TLR expression and low cytokine production, are possibly involved in the severity of the disease. Our results provide new information on the contribution of NK cells in Leishmania infections of the human host.


Introduction
Leishmania mexicana causes a wide spectrum of cutaneous diseases, ranging from localized cutaneous leishmaniasis (LCL), characterized by ulcers at sites of parasite inoculation, to diffuse cutaneous leishmaniasis (DCL), where parasites spread throughout the skin forming disfiguring nodules [1]. In Mexico, 400 new patients with cutaneous leishmaniasis are diagnosed each year, where the prevalence of DCL is less than 1% [2]. Although the cause for the uncontrolled parasite spread in DCL patients remains unknown, the early innate immune response against Leishmania possibly plays a pivotal role in determining disease evolution. Leishmania lipophosphoglycan (LPG) is a major surface molecule that activates TLR2 in cells of the innate immunity [3,4]. The carbohydrate composition of LPG characterizes different Leishmania species [5,6]. Murine models of leishmaniasis have linked various TLRs (TLR2, TLR3, TLR4 and TLR9) with enhanced IFN-c and IL-12 production and parasite control [4,[7][8][9][10]. Among the first innate cells capable of early IFN-c and TNFa production are NK cells [11]. They can be divided into 2 subsets: CD56 dim and CD56 bright , yet the roles of these subsets have not been clearly characterized in leishmaniasis [12,13]. We had previously shown that Leishmania LPG activates human NK cells through TLR2 stimulation, leading to IFN-c and TNF-a production [3]. These cytokines synergize in the macrophage to induce iNOS leading to NO production, one of the molecules responsible for intracellular Leishmania destruction [14]. Even though NK cells have been shown to play an important protective role in mouse Leishmania infections [11,15], their response has not been analyzed in patients with LCL and DCL.
In the present study, we comparatively analyzed NK-cell activity as well as their response towards the parasite in LCL and DCL patients. We found that peripheral blood and lesional NK cells of DCL patients were severely reduced in number and produced markedly less IFN-c and TNF-a, as compared to LCL patients. In addition to the reduced cytokine production, NK cells of DCL patients also showed diminished TLR2, TLR1 and TLR6 expression, both in LPG-stimulated and non-stimulated NK cells, which contrasted sharply with the heightened response found in LCL patients. The reduced NK cell cytokine production correlated with a down-regulation of IFN-c gene expression in DCL patients. We further show the activation pathway of TLR2 by Leishmania LPG, and the participation of TLR1 and TLR6 in the binding of LPG.

Ethics Statement
The study was reviewed and approved by the Ethics and Research Committees of the Faculty of Medicine of UNAM (Universidad Nacional Autónoma de México) (FMED/CI/RGG/ 013/01/2008) and guidelines established by the Mexican Health Authorities were strictly followed. All patients and controls were informed and signed a written consent to participate in the study.
Patients and controls 28 patients with LCL and 6 with DCL from La Chontalpa (Tabasco State), an endemic area in southeastern Mexico, were analyzed. Patients were diagnosed by clinical criteria, parasite demonstration in Giemsa-stained smears taken from lesions and intradermal Montenegro hypersensitivity test. LCL patients showed skin ulcers with few parasites, all were positive to the Montenegro test. DCL patients had multiple non-ulcerative nodules covering large areas of the skin that contained heavily parasitized macrophages. All DCL patients were negative for the Montenegro skin test. All patients received anti-Leishmania treatment with Glucantime.
Blood samples were taken from 28 LCL patients (17 males and 11 females), which had a mean age of 28 years and a disease duration ranging from 1.5 to 18 months. The LCL patients were divided according to gender, age (#25 years and #26 years) and evolution time (#3 months and $4 months). The 6 DCL patients (five males and one female) had a mean age of 44 years (24-57 years) with an average disease evolution of 18 years (two patients had 3 years, three had 17 years and one had 35 years). Blood from 21 healthy donors was obtained in a blood bank.

Lipophosphoglycan purification
LPG was purified as previously described [3]. Briefly, parasites were sub-cultured every 4-5 days and grown to a density of 2610 7 /mL. Promastigotes were harvested from stationary-phase cultures and centrifuged at 32006 g for 10 min and washed in PBS. The pellet was extracted with chloroform/methanol/water (4:8:3, v/v) during 30 min at RT. The insoluble material was used for LPG extraction with 9% 1-butanol in water (2650 mL) and the pooled supernatants were vacuum dried. LPG was purified from this fraction by HPLC, using an octyl-sepharose column and a 1-propanol gradient (5-60%) in 0.1 M ammonium acetate. Two octyl-sepharose columns were used to optimize LPG purity. The preparations tested negative for endotoxin with the Limulus sp. amebocyte lysate assay (E-Toxate Kit; Sigma). Additionally, a sample of LPG was analyzed by SDS-PAGE to verify the absence of protein contaminants. 10 mg/mL LPG was used in all experiments.

NK cell purification
NK cells were purified from PBMC of LCL and DCL patients, as well as from healthy donors. Briefly, PBMC were separated by density gradient (Histopaque-1077, Sigma-Aldrich) at 3006 g for 20 min at 20uC. Cells were obtained from the interface, washed twice in cold PBS and placed in RPMI-1640 (Gibco), supplemented with 10% heat-inactivated FBS, 2 mM L-glutamine, 10 nM HEPES, 100 mg/mL penicillin-streptomycin (Gibco), 17 mM NaHCO 3 and seeded in Petri dishes at 37uC, 5% CO 2 during 18 h for adherence of monocytes. Non-adherent cells were removed, washed in PBS and NK cells were purified with an NK cell isolation kit II (Miltenyi Biotec, Bergisch Gladbach, Germany). Briefly, 1610 7 total cells was suspended in 40 mL PBS containing 10 mL of cocktail of biotin-conjugated monoclonal antibodies against CD3, CD4, CD14, CD15, CD19, CD36, CD123 and glycophorin A and incubated for 10 min at 4uC. 30 mL PBS and 20 mL anti-biotin microbeads were added for 15 min at 4uC. The cells were washed with PBS, centrifuged at 3006g for 10 min and passed through a magnetic separation LS column (Miltenyi). NK cells were isolated by negative selection. The purity of the enriched NK cells was assessed by flow cytometry using anti-CD56-PE, anti-CD3-FITC (Coulter Immunotech) antibodies, achieving 97% purity. NK cells were washed and plated in 24-well culture-plates.

Immunoprecipitation
We further determined whether the recognition of LPG by TLR2 in NK cells also led to binding of TLR1 and/or TLR6 and additionally analyzed the activation pathway of TLR2 in NK cells of control subjects by immunoprecipitation of TLR2-MyD88, MyD88-IRAK-1, MyD88-TRAF-6, IRAK-1-TRAF-6 and TRAF-6-IKK-a. After purification, NK cells were suspended in RPMI-1640 with 10% heat-inactivated FBS and incubated for 2 h at 37uC with 5% CO 2 . Thereafter, 10610 6 NK cells were incubated with LPG (10 mg/mL) for 1 h at 37uC, 5% CO 2 and the same number of NK cells were incubated in RPMI alone. Cells were washed twice with cold PBS and lysed in 250 ml modified radioimmunoprecipitation (RIPA) buffer (Tris-base, pH 7.4 10 mM, NaCl 150 mM, EDTA 1 mM, NaF 10 mM, NP-40 1%, PMSF 1 mM, aprotinin 10 mg/mL, leupeptin 1 mg/mL, Na 3 VO 4 pH = 10 10 mM, DTT 1 mM) and incubated for 30 min on ice. Cell lysates were centrifuged at 10 0006g for 10 min at 4uC and the supernatants were collected. Protein concentration was determined using DC Protein Assay Reagents Package (Bio-Rad Laboratories, Hercules, CA, USA).

Total and nuclear protein extraction
To obtain the total protein extract, 2610 6 NK cells were incubated with LPG 10 mg/mL for 15, 30, 45 and 60 min or for 15 min with PMA. Cells were washed twice with PBS and lysed in 50 ml of RIPA modified buffer for 30 min. Cellular extracts were centrifuged at 10 0006g for 10 min at 4uC and the supernatants were collected. For nuclear proteins extracts 3610 6 NK cells were incubated with LPG 10 mg/mL for 1 h. Cells were washed twice with PBS and lysed by incubating them for 10 min in a detergentfree hypotonic buffer (10 mM Tris, pH = 7.6. 10 mM NaCl, 1.5 mM MgCl 2 , 0.5 mM EDTA, 1 mM DTT, 1.5 mg/mL leupeptin, 0.7 mM PMSF). Extracts were centrifuged at 4uC for 10 min at 9566g. The supernatants were discharged and intact nuclei were incubated in extraction buffer (20 mM Tris, pH = 8.0, 450 mM KCl, 0.05 mM EDTA, 1 mM DTT, 1.5 mg/mL leupeptin, 5 mM spermidine, 25% glycerol) for 45 min under constant agitation at 4uC. DNA pellets were eliminated by centrifugation for 15 min at 13 5006g at 4uC. Total and nuclear protein extracts were quantified with DC Protein Assay Reagents Package.

Immunohistochemistry (IHC) for detection of NK cells, TLRs and cytokines in NK cells from LCL and DCL patients
Skin punch biopsies (4-6 mm) were taken from the lesions of DCL and LCL patients. The tissues were embedded in paraffin, cut into 5-mm thick slices. Some of the slides were stained with H&E to evaluate the inflammatory characteristics. For Immunohistochemistry analysis the slides were hydrated and antigenically reactivated in a citrate buffer (0.01 M citric acid, 0.01 M sodium citrate) for 10 min at 95uC. Endogenous peroxidase was blocked with methanol/H 2 O 2 3% for 10 min and nonspecific antigenic sites were blocked with 3% bovine serum albumin dissolved in Tris-HCl pH = 7.6 with 0.1% Triton X-100 for 60 min at RT. Thereafter, samples were stained with mouse anti-CD57 (Zymed 08-0167) overnight at 4uC, washed, incubated with secondary antibody biotin anti-mouse (Zymed 62-6540) for 30 min at RT and with streptavidin AP (AB complex/AP, DAKO K0376,) for 30 min at RT. Tissues were washed and color development was assessed after incubation with AP Red substrate kit (Zymed 00-2203) or Stay Green/AP kit (Abcam antibodies: ab-156428) at RT.
TLRs and cytokines were detected by double staining. For this, the samples were washed and endogenous peroxidase and nonspecific antigens were blocked as described above. Thereafter, the samples were incubated 1:50 with goat anti-TLR1, mouse anti-TLR2, rabbit anti-TLR6, mouse anti-IFN-c or mouse anti-TNF-a 1:100 (sc-8687, sc-21759, sc-30001, ab-11866 and sc-1350, respectively) for 30 min at RT, washed and secondary antibodies were incubated for 30 min at RT. The secondary antibodies included: mouse and rabbit specific HRP/AEC detection IHC kit (Abcam ab94705) for TLR6, biotin-labelled rabbit anti-goat (Zymed 81-1640) antibodies for TLR1 and TNF-a and biotinlabelled goat anti-mouse (Zymed 62-6540) was used for detecting TLR2 and IFN-c. Thereafter, tissues were washed and incubated with horseradish peroxidase (Zymed 43-4311) for 30 min at RT. For TLR and IFN-c detection samples were washed and color development was assessed after incubation with DAB Black kit (Biocare Medical BRI40 H, L). For TNF-a, a DAB Substrate kit was used (Roche Cat. 11718096001). The slides were counterstained with Mayers haematoxylin (Biogenex, CA, USA). Digital images of tissue sections were captured using a light microscope and an AxioCam MRc5 camera (Zeiss, Germany). In order to obtain the number of single and double positive cells in these lesions, cells were counted in 8 pictures of each tissue were taken with a final area corresponding to 1 mm 2 of 3 LCL and 3 DCL patients. Controls for primary and secondary antibodies were negative.

Gene expression of TLR2, IFN-c and TNF-a by Real-time PCR
Gene expression of TLR2, IFN-c and TNF-a were analyzed by real-time PCR in 5 healthy controls, 6 LCL and 3 DCL patients. Total RNA from non-stimulated and LPG-stimulated NK cells (18 h) was retro-transcribed using High-Capacity cDNA Archive kit (Applied Biosystems), according to manufacturers instructions. Quantitative Taqman PCR analysis was performed with the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems) containing 16 Taqman Universal Master Mix (Applied Biosystems) and 16 probes and primers sets Hs00610101_m1 (TLR2), Hs00174128_m1 (TNF) and Hs00989291_m1 (IFN). The thermal profile was as follows: 95uC for 10 min and 40 cycles at 95uC for 15 s and 60uC for 1 min. All amplification reactions were done in duplicate and the relative quantification of TLR2, TNF-a and IFN-c gene expression were calculated using the comparative Ct method (2 2DDCT ) [16]. Levels of mRNA expression were assessed after normalization, using GAPDH as internal control.

Statistical analysis
Statistical differences between groups were obtained using Mann-Whitney U-test or Student's T-test. The Kruskal-Wallis test was used for the comparison of more than two groups of data. Correlation analyses were performed by Spearmans test. Data are presented as mean 6SEM, p,0.05 was considered statistically significant. These statistical analyses were done using the Prism 5 software (GraphPad Software, San Diego, CA, USA).

LPG bindsTLR1 and TLR6 and activates the TLR signaling pathway
Both TLR1 and TLR6 proteins are expressed in non-stimulated and LPG-stimulated NK cells from healthy controls ( Fig. 1A and  1B). In order to analyze whether TLR2 binds to TLR1 and TLR6 in NK cells, we immunoprecipitated with anti-TLR2 and Western blotted with a-TLR1 or a-TLR6. A recognition band was found both experimental conditions: in non-stimulated and LPGstimulated NK cells. It is noteworthy, that in both cases, a more intense recognition was observed in non-stimulated NK cells ( Fig. 1C and 1D, lane a) as compared to LPG-stimulated NK cells ( Fig. 1C and 1D, lane b). We speculate that the reduced binding of TLR1 or TLR6 to TLR2 after incubation with LPG is possibly due to partial obstruction of the docking site after LPG binds to TLR2.
Through immunoprecipitations we analyzed the binding of proteins involved in the TLR signaling pathway in non-stimulated and LPG-stimulated NK cells. The immunoprecipitations included: TLR2-MyD88, MyD88-IRAK-1, MyD88-TRAF-6, IRAK-1-TRAF-6 and TRAF-6-IKK-c. We observed protein binding in all immunoprecipitations, with an increase when the NK cells were stimulated with LPG (Fig. 1E, second, third and fourth blots). Only the binding of TRAF-6-IKK-c decreased (Fig. 1E, fifth blot). This decrease in binding was expected, since the kinases need to be degraded to induce NF-kB translocation to the nucleus. To determine whether LPG induces NF-kB nuclear translocation, we used nuclear extracts of NK cells to analyze p50 and p65 isoforms and their nuclear translocation. We found nuclear translocation of both isoforms in LPG-stimulated NK cells (Fig. 1F, lane b).
Finally, we analyzed the kinetics of phosphorylation of the kinases IKK and IkB in non-stimulated NK cells and in cells stimulated with PMA 10 mg/mL or LPG 10 mg/mL during 15, 30, 45 and 60 min. We observed an increase in phosphorylation of these kinases at 45 and 60 min after stimulation with LPG (Fig. 1G). With these data we were able to confirm that LPG activates NK cells through the TLR2 pathway.

NK cells in peripheral blood and lesions of LCL and DCL patients
Quantitation of NK cells in blood and tissues of patients with both clinical forms revealed that LCL patients had more NK cells in blood as well as in lesions, as compared to DCL patients. NK cells in PBMC were analyzed by flow cytometry ( Fig. 2A and 2B), showing that the percentage of NK cells in peripheral blood of LCL patients ranged from 1.4-27.8% (mean: 8.4361.22), whereas in DCL patients they ranged from 0.5-3.1% (Fig. 2C). No correlation was found between the number of NK cells in peripheral blood and disease duration, age or gender.
Immunohistochemical stains of NK cells in skin biopsies showed enhanced numbers of NK cells in lesions of LCL patients, as compared to DCL patients (Fig. 2D images 1 and 3). H&E stains showed granuloma formations in LCL tissues (Fig. 2D, image 2), whereas in DCL patients, a randomly scattered distribution of inflammatory cells was found (Fig. 2D, image 4). NK-cell counts in tissues showed a mean number 210 cells/mm 2 for LCL patients, whereas DCL patients showed 67 cell/mm 2 (Fig. 2E). Taken together, LCL patients had significantly more NK cells as compared to DCL patients, both in peripheral blood as well as in infected lesions.

TLR2 expression on NK cells of peripheral blood
NK cells were isolated by negative selection achieving a purity of 97%, as shown by flow cytometry (Fig. 3A, middle image). TLR2 expression was analyzed in non-stimulated as well as LPGsimulated NK cells of 28 LCL patients, six DCL patients and 21 healthy controls. All NK cells expressed TLR2, albeit with different intensity. NK cells of LCL patients expressed significantly higher levels of TLR2, as compared to DCL cells or to healthy controls ( Fig. 3B and 3C), yet no differences were found in TLR2 expression when comparing non-stimulated with LPG-stimulated NK cells within each patient group (Fig. 3C).
To ascertain whether the increased TLR2 expression found in LCL patients was related to gender, disease duration or age, we subdivided the 28 patients according to these parameters. Thus, we analyzed the TLR2 expression in 11 females and 17 males. We found that NK cells of male LCL patients expressed significantly higher levels of TLR2, as compared to females, both in nonstimulated as in LPG-stimulated cells (Fig. 4A). When analyzing TLR2 expression in NK cells according to disease progression, we found that both in males and females the expression of TLR2 diminishes significantly after four months of disease duration (Fig. 4B). The analysis of TLR2 expression according to age revealed that males $26 years always express higher levels of TLR2, as compared to female LCL patients of the same age group, both in non-stimulated as well as in LPG-stimulated cells (Fig. 4C).

TLR1, TLR2 and TLR6 expression on purified NK cells of LCL and DCL patients
Having shown that LPG is a ligand for TLR1, TLR2 and TLR6 and that TLR2 can bind with either TLR1 or TLR6, we analyzed whether specific heterodimers could be associated with the clinical forms of the disease. We analyzed the expression of all three TLRs in non-stimulated and LPG-stimulated NK cells in three groups of individuals: nine LCL patients (two females and seven males, who had a mean age 2565 years, and a mean disease duration of five months); four DCL patients (one female and three males, who had a mean age of 52 years, and a disease duration 19 years), and four healthy controls without a history of leishmaniasis.
Healthy controls and LCL patients expressed significantly higher levels of all three TLRs, as compared to DCL patients. LPG stimulation showed no significant increase in TLR expression between LCL and DCL patients (Fig. 5A).
Only LCL patients were analyzed for TLR1, TLR2 and TLR6 expression, according to age, due to the reduced number of DCL patients. The nine LCL patients were divided into in two groups: #25 years (n = 5) and $26 years (n = 4). A significant reduction of all three TLR receptors was observed in patients $26 years, as compared to patients #25 years, since the later expressed only half the values of TLR1, TLR2 and TLR6, both in non-stimulated as well as in LPG-stimulated NK cells (Fig. 5B).
In an attempt to analyze whether this non-responsiveness of all three receptors was specific for LPG, we analyzed the expression of these TLRs in NK cells of both groups of patients after stimulation with additional TLR2 agonists: PGN (peptidoglycan) and Pam 3 Cys-Ser. Significant differences in the expression of all three receptors on NK cells were observed between the two patient groups: whereas NK cells of LCL patients showed an enhanced expression of all three TLR receptors, all of which tended to increase with TLR2 agonists, DCL patients showed significantly lower levels of TLR1, TLR2 and TLR6 expression, which remained unchanged despite stimulation with various TLR2 ligands (S1). These data show that the unresponsiveness of NK

TLR1, TLR2 and TLR6 expression on NK cells of tissue lesions of LCL and DCL patients
To assess the phenotype and distribution of NK cells expressing these TLRs in the lesions of LCL and DCL patients, we performed double immunostaining: CD57 and TLR1 (or TLR2 or TLR6) (Fig. 6A). The mean number of NK cells expressing TLR1, TLR2 or TLR6 in LCL patients was 1156l5, 151631 and 125617 cells/mm 2 , whereas in DCL patients the mean number was 7068, 6568 and 58620 cells/mm 2 , respectively (Fig. 6B). Thus, LCL patients showed significantly higher numbers of NK cells (CD57 + / TLR + ) expressing TLR2, TLR1 or TLR6 on NK cells, as compared to DCL patients.

TLR1, TLR2 and TLR6 expression in NK subsets CD56 dim and CD56 bright
Since NK cells are subdivided phenotypically according to their function into CD56 dim (cytotoxic) and CD56 bright (cytokine producing) cells, we were interested in analyzing TLR1, TLR2 and TLR6 expression in both NK subsets of LCL and DCL patients ( Fig. 7A and 7B). We therefore analyzed the expression of these receptors in 9 LCL and 4 DCL patients, before and after stimulation with LPG. Only NK CD56 bright cells expressed high levels of TLR1, TLR2 and TLR6, which were significantly higher in LCL patients (4 to 8-fold), as compared to DCL patients. There were also significant differences in TLRs expression between both NK subsets: CD56 dim expressed significantly lower levels of TLRs as compared to CD56 bright (Fig. 7C, 7D and 7E). However, no significant differences were found between non-stimulated and LPG-stimulated NK cells in both subsets of NK cells of the different groups.

IFN-c and TNF-a production by NK cells in blood and tissue lesions
Production of IFN-c and TNF-a was analyzed in 28 LCL, 6 DCL patients and 21 healthy donors in LPG-stimulated and nonstimulated purified NK cells. Non-stimulated NK cells from healthy subjects and LCL, DCL patients produced similar basal amounts of IFN-c (with mean values of: 39, 37, 46 pg/mL, respectively). Yet after stimulation with LPG, only NK cells from LCL patients and healthy subjects increased their IFN-c production (mean values: 60 and 77 pg/mL, respectively). In contrast, NK cells from DCL patients reduced their IFN-c production to half their basal value (from 46 to 22 pg/mL), when stimulated with LPG (Fig. 8A). To ascertain whether the increased IFN-c production found in LCL patients was related to gender, disease duration and/or age, we subdivided the 28 patients according to these parameters. Thus, we analyzed the cytokine production in 11 females and 17 males. Female patients produced significantly more IFN-c in non-stimulated (60 pg/mL) and LPGstimulated (103 pg/mL) NK cells, as compared to males, which only showed a slight increase in IFN-c production (21 to 32 pg/ mL) after LPG stimulation (Fig. 8B). In an attempt to associate IFN-c production with disease evolution in female and male LCL patients, we further separated the patients into groups according to their disease duration into #3 months or $4 months. NK cells of female LCL patients with $4 months disease duration showed higher IFN-c production in basal conditions, as well as after LPG stimulation, as compared to those with #3 months or to male LCL patients (Fig. 8C). When analyzing IFN-c production according to age, we found that females, particularly those aged #25 (37% of the women) showed a vigorous response (5-fold increase) when NK cells are stimulated with LPG. Although NK cells of female LCL patients aged $26 showed a higher IFN-c production in nonstimulated NK cells, as compared to younger females, these cells responded only slightly to LPG. This stands in contrast to the minimal response towards this parasite antigen found in male patients of any age, who only showed a minimal increase in cytokine production after LPG stimulation (Fig. 8D). Taken together, our data reveal that NK cells of female patients aged #25 years, with disease duration of $4 months, showed the most vigorous IFN-c production when the cells are stimulated with LPG, whereas NK cells from female patients aged $26 years already come activated and therefore respond only weakly to further stimulus by LPG. This contrasts with the diminished IFN-c production in male NK cells of all age groups, in both nonstimulated and LPG-stimulated conditions, irrespective of disease duration. When comparing IFN-c production in NK cells of lesions in LCL and DCL patients, we found that LCL patients showed higher numbers of double positive cells (CD57 + /IFN-c + ), as compared to DCL patients (Fig. 8E). Our data show that IFN-c production by NK cells in both blood and tissue lesions were markedly reduced in DCL, as compared to LCL patients.
Additionally, we analyzed TNF-a production in NK cells in the same group of subjects and found that LCL patients produced higher levels (mean: 42 pg/mL) as compared to DCL patients (mean: 23 pg/mL), particularly after LPG stimulation, where, instead of enhancing cytokine production as in LCL patients (53 pg/mL), NK cells of DCL patients further reduced their TNFa production (17 pg/mL). The difference of TNF-a production in LPG-stimulated NK cells between LCL and DCL patients was significant (Fig.9A). We also analyzed TNF-a production in LCL patients according to gender, disease duration and age. NK cells of female LCL patients produced significantly more TNF-a than male LCL patients in non-stimulated (56 vs 32 pg/mL) cells. After LPG stimulation, NK cells of both females and males increased their TNF-a production (64 vs 46 pg/mL), albeit the differences were not statistically significant (Fig. 9B). No significant difference in TNF-a production was found when comparing different age groups or disease duration (data not shown). The analysis of TNFa production in lesions of both patient groups showed that LCL patients had more double positive cells (CD57 + /TNF-a + ), as compared to DCL patients (Fig. 9C, black arrows). Interestingly, many large single positive cells (CD57 -/TNF-a + ) are observed in tissues of DCL patients (Fig. 9C, red arrows). Since they are not NK cells, the exact nature of these larger cells withTNF-a + staining found in tissues of DCL remains to be determined.

TLR2, IFN-c and TNF-a gene expression in NK cells in LCL and DCL patients
In order to clarify whether the reduced cytokine production and TLR2 expression in DCL patients was related to reduced gene expression, real-time PCR was done in NK cells of both groups of patients. The gene expression of TLR2, IFN-c and TNF-a was analyzed in non-stimulated and LPG-stimulated NK cells of 6 LCL and 3 DCL patients as well as 5 healthy controls. DCL vs LCL patients were compared using 2 -DDCT method and as a     9) or to healthy controls (Table1, lane 5 and lane 8), albeit these differences were not statistically significant. The same holds true for the expression of TNF-a genes, which also showed a nonsignificant down-regulation in DCL patients, as compared to LCL patients, both in non-stimulated as in LPG-stimulated NK cells (Table 1, lane 6 and lane 9).
Thus, we were able to show that the reduced protein expression of IFN-c in NK cells of DCL patients correlated with the downregulation of its gene expression.

Discussion
The cause of uncontrolled parasite dissemination in DCL patients infected with Leishmania mexicana remains an enigma.
Although much insight has been gained on the importance of a Th1 response for parasite control in mouse models [14], these data cannot be extrapolated to the human disease. The molecules and mechanisms of innate and adaptive immunities, particularly the role of inflammation, need to be further assessed in the physiopathology of human leishmaniasis. One of the cells that possibly play a role in defining disease severity is the NK cell, since this innate cell is able to produce IFN-c and TNF-a, both of which are required to activate the leishmanicidal machinery within macrophages. We had previously shown that Leishmania LPG is a ligand for TLR2 leading to IFN-c and TNF-a production [3]. Our current results show that in addition to TLR2, TLR1 and TLR6 are also present in the binding of LPG.
Most studies of TLRs in leishmaniasis have been done in the murine models, where expression of TLR2, TLR4, TLR7, TLR8 and TLR9 have been analyzed and related to disease outcome, together with other contributing factors such as Leishmania species and genetic background. Yet data on the role of TLRs obtained from experimental murine leishmaniasis remain controversial: on Patients with L. mexicana Have Altered NK Response one hand, enhanced the expression of these TLRs have been related to protection mediated by cytokine production, whereas their absence has been associated with a Th2 response and elevated Leishmania numbers [7,10,[17][18][19][20]. Contrasting results have shown that the absence of TLR2 during the initial stages of the disease lead to reduced parasite burdens in L. amazonensis infected mice, which was associated with more organized granuloma formations [21]. This was supported by data of TLR2 2/2 mice that achieved a better elimination of the parasite due to reduced inflammatory infiltrates [22].
Little is known of TLRs and NK cell activity in patients with different clinical forms of cutaneous leishmaniasis and to what degree these can be related. Previous studies of NK cells in DCL patients have reported a reduction in NK cell numbers that could be restored after treatment and parasite reduction [1, 35], suggesting that the parasite is able to regulate NK cells. Lieke et al. demonstrated that NK cells incubated with L. major or L. aethiopica could lead to death not only of the parasite, but of the NK cell as well [36]. Yet functional modulation of NK cells by this parasite remained to be analyzed. We were therefore interested in analyzing the expression of TLR2, as well as cytokine production by NK cells of patients with LCL and DCL and to evaluate their possible association with disease severity. We were furthermore interested whether specific heterodimers form between TLR2 and TLR1 or TLR6 when binding to LPG and if there is a preferential expression of any of these TLRs in patients with LCL and DCL that could be related to disease severity. We found striking differences in the NK-cell numbers, in the magnitude of TLR expression as well as in IFN-c and TNF-a productions by NK cells of DCL and LCL patients, which correlated with disease severity: DCL patients showed reduced NK cell numbers (possibly due to NK-cell death), down-regulated TLR2, TLR1 and TLR6 expression as well as reduced cytokine production, as compared to LCL patients. In contrast, LCL patients showed enhanced expression of these TLRs, which correlated with augmented IFN-c and TNF-a production by their blood NK cells, in addition to enhanced tissue NK cells with IFN-c and TNF-a staining.
Although tissue lesions of DCL patients showed reduced NK cells with TNF-a staining, they harbored larger NK-negative cells, which stained for TNF-a. Our study did not clarify the nature of these cells, yet due to their size/form, we are tempted to speculate that they could be mast cells, capable of storing preformed TNF-a, that is released upon various stimuli. The elevated amounts of cells containing TNF-a in lesions of DCL patients, possibly account for the intense inflammation and associated tissue damage found in these patients [37]. The tissue damage due to intense inflammation has also been shown in patients with mucocutaneous leishmaniasis [38,39].
A further finding was that the NK cells of LCL patients expressing TLRs were found within granulomas. Highly organized granuloma structures have been related to host resistance during hepatic leishmaniasis of patients with visceral leishmaniasis caused by Leishmania donovani [40]. Within the protective microenvironment created by granuloma formations, NK cells possibly contribute to anti-leishmanial mechanisms by secreting IFN-c and TNF-a, leading to macrophage activation and helping to create a Table 1. Transcript expression by quantitative real time PCR in non-stimulated (NS) and LPG-stimulated NK cells (S). Supporting Information Figure S1 Expression of TLR1, TLR2 and TLR6 in NK cells stimulated with different TLR2 ligands (LPG, PGN and Pam 3 Cys). Cell surface expression is indicated by the geometric mean of fluorescence intensity (MIF). These results are the mean 6SEM. *p#0.05 was considered statistically significant. (TIFF)