Besides their role in fighting viral infection and tumor resistance, recent studies have shown that NK cells also participate in the immune response against other infectious diseases. The aim of this study was to characterize the possible role of NK cells in the immune response against Paracoccidioides brasiliensis. Purified NK cells from paracoccidioidomycosis patients and healthy individuals were incubated with P. brasiliensis yeast cells or P. brasiliensis-infected monocytes, with or without the addition of recombinant IL-15. We found that NK cells from paracoccidioidomycosis patients exhibit a lower cytotoxic response compared with healthy individuals. NK cells are able directly to recognize and kill P. brasiliensis yeast cells, and this activity seems to be granule-dependent but perforin-independent, whereas the cytotoxicity against P. brasiliensis-infected monocytes is perforin-dependent. These results indicate that NK cells participate actively in the immune response against the P. brasiliensis infection either by directly destroying yeast cells or by recognizing and killing infected cells. Granulysin is the possible mediator of the cytotoxic effect, as the reduced cytotoxic activity against the yeast cells detected in patients with paracoccidioidomycosis is accompanied by a significantly lower frequency of CD56+granulysin+ cells compared with that in healthy controls. Furthermore, we show that NK cells released granulysin in cultures after being stimulated by P. brasiliensis, and this molecule is able to kill the yeast cells in a dose-dependent manner. Another important finding is that stimulated NK cells are able to produce proinflammatory cytokines (IFN-γ and TNF-α) supporting their immunomodulatory role in the infection.
Infections caused by fungi are currently among the most life-threatening diseases. Paracoccidioidomycosis (PCM), a disease caused by the dimorphic fungus Paracoccidioides brasiliensis, is the major cause of systemic mycosis in Latin America (1). The disease presents a wide spectrum of clinical and immunological manifestations, varying from benign and localized forms to severe and disseminated forms. The pattern of the immune response to P. brasiliensis is thought to be a determinant of disease progression and clinical outcome. Effective defense against P. brasiliensis depends mainly upon Th1 response, where TNF-α and IFN-γ play a particularly prominent role (2).
There is increasing evidence to show that the innate immune response is critical to the initial control and the subsequent development of adequate acquired immunity against a number of microorganisms, including fungi (3, 4). NK cells can contribute to the innate resistance to infections through a variety of mechanisms, including the secretion of proinflammatory and immunoregulatory cytokines and chemokines, lysis of infected host cells, and the direct killing of yeast cells (3, 4).
Two subsets of NK cells have been identified in humans according to their CD56 expression, which apparently differs in phenotype and function. Indeed, CD56bright NK cells represent a small population in periphery (∼10%) and are characterized by their proliferative capability and the production of cytokines such as IFN-γ and TNF-α, without showing cytotoxic activity. In contrast, NK cells with low CD56 expression (CD56dim) are abundant in the periphery, exhibit cytotoxic activity, and possess a minor capacity to produce cytokines (5).
In contrast to CTLs, NK cells do not require Ag-specific recognition to kill target cells and are capable of limiting infection prior to the induction of adaptive immune responses. The activation of NK cells is regulated by a balance between stimulatory and inhibitory signals delivered by the interaction of specific receptors and ligands. The killer cell Ig-like receptor family includes the most important inhibitory receptors that can recognize and bind to MHC class I molecules, hence the reduction in these MHC molecules on infected or transformed cells trigger the activation of NK cells. At the same time, there are several receptors described as stimulatory in NK cells, among them NKG2D, which is one of the most studied. This molecule recognizes molecules whose structure is similar to that of MHC molecules, such as MICA/B and UL-16 binding protein molecules (6).
After the activation, granule exocytosis is the major killing mechanism used by NK cells. Cytotoxic granules are composed of several molecules, such as granzymes (mainly A and B), perforin, and granulysin. Studies on Cryptococcus neoformans infection showed that NK cells exhibit fungicidal activity without previous activation and that this activity is dependent on the presence of perforin but not granulysin (7).
Granulysin, the most recently described granule component, interacts with lipids in cell membranes promoting the lysis of bacteria, fungi, and protozoan cells (8). Deng et al. (9) showed that besides its cytotoxic activity, granulysin also exhibits inflammatory and chemotactic properties, contributing to the local immune response. However, granulysin, like granzymes, depends upon perforin to act inside target cells (10).
Tran et al. (11) demonstrated that the presence of IL-15 augments the cytotoxic activity of NK cells against bacteria and fungi. IL-15 is primarily produced by monocytes and dendritic cells. This cytokine play a major role in the activation and the survival of cytotoxic cells, both T CD8+ cells and NK cells (12). In terms of PCM, there are few studies about the participation of NK cells in the immune response. Peraçoli et al. (13) described a diminished cytotoxic response of NK cells in patients with the disease. In mice, it was demonstrated that NK cells exhibit cytotoxic activity against cells infected with P. brasiliensis and were able to inhibit fungal growth in vitro (14).
In this study, we propose a functional role and mechanism used by NK cells in the immune response against the P. brasiliensis infection. We found that NK cells can actively participate in the immune response against P. brasiliensis either by destroying yeast cells or by recognizing and killing infected cells. Granulysin is the possible mediator of the cytotoxic effect, as it is produced and released by NK cells. Furthermore, supernatants depleted of granulysin presented decreased fungicidal activity. Our results also support the idea that in addition to their cytotoxic properties, NK cells may play an immunomodulatory role in PCM infection through the production of proinflammatory cytokines IFN-γ and TNF-α, which could influence the subsequent acquired immune response against this fungus.
Materials and Methods
Peripheral venous blood was collected from healthy donors and PCM patients, the latter being divided into two groups: patients with active disease selected before or within the first month of treatment, and patients who had received treatment and had negative serological tests for P. brasiliensis and no clinical signs of the disease. Diagnosis was confirmed by detection of the fungus in clinical specimens and serological tests carried out at the Hospital de Clínicas, State University of Campinas (São Paulo, Brazil). In accordance with the rules of the Medical Research Ethics Committee at the Faculty of Medical Sciences, State University of Campinas, all individuals signed a voluntary consent form agreeing to participate in the study.
Biopsy specimens were taken from 10 patients with PCM who received treatment at the Hospital de Clínicas, State University of Campinas. The biopsies were obtained for diagnostic purposes before the treatment. Immunohistochemical analysis was performed with mAbs against CD56, granulysin, granzyme B, perforin, and IL-15 (Santa Cruz Biotechnology) and the amplification system NovoLink (Max Polymer Detection System; Novocastra Laboratories) according to the manufacturers’ instructions. Negative controls were carried out by undertaking the same procedures on the slides, without the addition of the primary Abs.
Preparation of P. brasiliensis yeast cells
P. brasiliensis (strains Pb18 and Pb265) were maintained at 36°C on Sabouraud dextrose slants (Difco, Detroit, MI) for 5–7 d. They were then collected, diluted in PBS, and the cluster of cells was disrupted by rotating the organisms with 0.5-mm-diameter glass beads in a vortex for 5 min. After this procedure, the cell suspension was maintained at 37°C for 15–30 min. The supernatants containing individual cells were then collected. The number and the viability of yeast cells were determined in a Neubauer chamber in the presence of cotton-blue stain. We only used preparations containing >85% of viable yeast cells.
Purification of NK cells and CD14+ cells
PBMCs were isolated using Ficoll-Hypaque (Pharmacia Biotech, Piscataway, NJ) density gradient centrifugation. NK cells and CD14+ cells were isolated using positive isolation kits (MACs; Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s instructions. The analysis by flow cytometry showed that ∼90–95% of CD56+ cells were CD56+CD3− cells (NK cells), with a minor fraction of CD56+CD3+ cells (NKT cells).
All cells were grown in RPMI 1640 (Invitrogen Life Technologies) supplemented with 10% heat-inactivated AB normal human serum, 100 U/ml gentamicin, 2 mM l-glutamine, and 1 mM sodium pyruvate (all from Sigma-Aldrich, St. Louis, MO).
Determination of the cytotoxic activity of NK cells against P. brasiliensis yeast cells and against target cells
A CFU assay was performed as previously described (15), with minor modifications. In short, P. brasiliensis yeast cells were incubated either with or without 2 × 105 NK cells (ratio 1:250). The number of CFUs of P. brasiliensis was determined after 48 h of culture. After the incubation, a fraction of the supernatants of cultures was collected and stored at −80°C for granulysin detection, and another fraction was diluted in distilled water and spread onto BHI agar plates [supplemented with normal horse serum, P. brasiliensis growth factor, and antibiotic solution (streptomycin/penicillin)].
To evaluate the effect of recombinant granulysin on P. brasiliensis,
Determination of cytotoxic activity against target cells was performed by either 51Cr or lactate dehydrogenase (LDH) release assays. Initially, we determined the cytotoxic activity against target cells (K562 lineage cells) of NK cells from active-disease PCM patients, treated patients, and control individuals by chromo release assays as previously described (13).
To determine the cytotoxic activity of NK cells against P. brasiliensis-infected CD14+ cells, we first cocultured CD14+ purified cells with Pb18 or Pb265 yeast cells for 4 h at 37°C. After this, we obtained an infection ratio of up to 80% (determined by microscopy and flow cytometry). The cytotoxic activity against infected cells was determined by means of LDH release using the kit Citotox 96 (Promega, Madison WI) according to the manufacturer’s instructions.
In some experiments, prior to the cocultures (with yeast cells or CD14+ infected cells), purified NK cells were treated with 25 mM strontium chloride (Sigma-Aldrich) for 18 h, 10 nM concanamycin A (Sigma-Aldrich) for 2 h, or 4 mM EGTA (Sigma-Aldrich) for 12 h, as previously described (16). The viability of cells was not altered by these treatments as assessed by trypan blue exclusion. The effect of IL-15 on NK cell cytotoxic activity and granule and cytokine mRNA expression was assessed by the addition of the recombinant cytokine (rhIL-15, 5 ng/ml) in cell cultures.
Culture conditions for flow cytometry and quantitative RT-PCR analysis of NK cells and CD14+ cells
Purified NK cells were stimulated with P. brasiliensis yeast cells (strains Pb18 and Pb265 at a ratio of 1 yeast cell to 50 NK cells; 1:50) and/or rhIL-15 (5 ng/ml) for 24 h at 37°C in 5% CO2. For flow cytometry analysis, brefeldin A (1 μg/ml) was added to the cultures for the last 6 h of incubation. We also evaluated the effect of the fungus on surface expression of some molecules in CD14+ purified cells by incubating P. brasiliensis yeast cells with CD14+ cells (ratio 1:50) for 4 h at 37°C in 5% CO2.
To evaluate the number and phenotype of NK cells in the peripheral blood of patients and controls, we used blood samples after red cell lysis. The leukocytes were incubated for 20 min at 4°C with Abs (anti-CD8–allophycocyanin, anti-CD3–PE–Cy7, anti-CD56–PerCP–Cy5.5, and anti-CD16–allophycocyanin–Cy7; all from BD Biosciences, San Jose, CA), washed, and fixed with 2% paraformaldehyde.
Purified NK cells and CD14+ were stimulated as described earlier. For the surface labeling, purified cells (2 × 105 cells) were incubated for 20 min at 4°C with the following Abs. NK cells: anti-CD56 PerCP-Cy5.5 or anti-CD56 PE-Cy7, anti-CD16–allophycocyanin–Cy7, anti-NKG2D–allophycocyanin, anti-CD69–allophycocyanin–Cy7, anti-CD25–PE–Cy7, anti-KIR3DL1–FITC, anti-KIR2DL2/L3/S2–PE, anti-CD8–allophycocyanin–Cy7, or allophycocyanin and anti-CD3–PE–Cy7. CD14++ cells were permeabilized with saponin (0.5% in PBS) for 10 min at 4°C and were incubated with anti–IFN-γ (FITC), anti–TNF-α (PE), anti-granzyme A (PerCP–Cy5.5), anti-granzyme B (FITC), anti-perforin (allophycocyanin), and anti-granulysin (PE) for 30 min at 4°C, followed by another wash and fixation step. The cells were acquired with a FACSCanto flow cytometer, and the analysis was performed using FCS Express Software (Denovo Software).
Granulysin immunoprecipitation and Western blotting
Purified NK cells were incubated with rhIL-15 (5 ng/ml) for 24 h at 37°C in 5% CO2. Cells were lysed by multiple freezing/thawing steps, and the supernatants were centrifuged for removal of cellular debris (10,000 × g for 10 min at 4°C). A fraction of the supernatants was maintained untreated and another was incubated for 1 h at 4°C with anti-granulysin Ab (2 μg/ml; Santa Cruz Biotechnology), after which was added 20 μl of protein A/G agarose beads (Santa Cruz Biotechnology), followed by incubation for 2 h at 4°C under constant agitation. The supernatants were centrifuged (1000 × g for 5 min at 4°C), the pellet of beads was discarded, and the supernatants were analyzed by Western blotting to confirm the depletion of granulysin. Briefly, the supernatants were resolved on polyacrylamide gel (13%, SDS-PAGE) under denaturing conditions and were transferred to polyvinylidene fluoride membranes (Bio-Rad). Membranes were blocked with TBS–BSA (1%) and incubated with goat anti-human granulysin, anti-granzyme B, anti-perforin, or anti-actin (all from Santa Cruz Biotechnology) overnight at 4°C. Membranes were incubated with anti-goat HRP-conjugated Abs for 1 h and analyzed by chemiluminescence in photo documentation equipment (ImageQuant 350; GE LifeScience) using image analyzer software (ImageQuant TL). For CFU assays, P. brasiliensis yeast cells (Pb18 strain, 2 × 103 viable cells) were incubated with either untreated or immunoprecipitated supernatants [undiluted (1:1) or diluted at 1:2 or 1:10 with RPMI 1640]. The number of CFUs of P. brasiliensis was determined after 48 h of culture at 37°C, spreading 100 μl of the culture suspensions onto BHI agar plates (as described earlier).
Stimulated purified NK cells and CD14+ cells obtained as described earlier were also evaluated by quantitative RT-PCR (qRT-PCR). Total RNA was extracted with TRIzol (Invitrogen, Carlsbad, CA), quantified by spectrophotometry (NanoDrop; Thermo Scientific, Waltham, MA), and treated with human DNase I to eliminate genomic DNA contamination. After DNase treatment, 1 μg total RNA was used for cDNA synthesis, and the cDNA was amplified using SYBR Green PCR Master Mix using Real-Time PCR equipment (StepOne System; Applied Biosystems, Foster City, CA). The results were expressed as relative expression, as described by Pfaffl (17), normalized to the β2-microglobulin gene. The following primers were used: β2-microglobulin, forward 5′-TGCTGTCTCCATGTTTGATGTATCT-3′; reverse 5′-TCTCTGCTCCCCACCTCTAAGT-3′. Granzyme A: forward 5′-AAGAGTTTCCCTATCCATGCTATGA-3′; reverse 5′-TTTGCTTTTTCCGTCAGCTGTA-3′. Granzyme B: forward 5′-TGCAACCAATCCTGCTTCTG-3′; reverse 5′-CGATGATCTCCCCTGCATCT-3′. Perforin: forward 5′-AAGCCCTCCGCCATTCTC-3′; reverse 5′-AACAGCCTCTTGGCCTTCTG-3′. Granulysin: forward 5′-CTGAGCCCTCTCACCTTGTC-3′; reverse 5′-GGAGAGTGGATTCTGGATCG-3′. KIR3DL1: forward 5′-CCCACTGCTTGTTTCTGTCACA-3′; reverse 5′-GGTTACCAGATTTGGAGCTTGGT-3′. KIR2 (CD158b): forward 5′-CTTCGGCTCTTTCCGTGACT-3′; reverse 5′-ACAGAAACAAGCAGTGGGTCA-3′. NKG2D: forward 5′-GAAGAGAGATCCTAAAGGCAATTCA-3′; reverse 5′-CCCCCAGCCCATCCA-3′. IFN-γ: forward 5′-CTAATTATTCGGTAACTGACTTGA-3′; reverse 5′-ACAGTTCAGCCATCACTTGGA-3′. TNF-α: forward 5′-TGGCCCAGGCAGTCAGA-3′; reverse 5′-GGTTTGCTACAACATGGGCTACA-3′. IL-15: forward 5′-GTCTTCATTTTGGGCTGTTTCAGT-3′; reverse 5′-CCTCACATTCTTTGCATCCAGATTCT-3′.
ELISA for cytokines and granulysin
The results for the different groups were analyzed using one-way ANOVA, followed by the Bonferroni multiple comparison test. To compare the effect of the different treatments in each group, we used the ANOVA test for repeated measures or paired Student t tests. A p value ≤0.05 was considered statistically significant. All the statistical analysis was carried out with SigmaStat v1.0 (Jandel).
Number and cytotoxic activity of NK cells
PCM is characterized by a wide range of clinical and immunological manifestations, which are reflected in the clinical forms of the disease. To eliminate these intrinsic differences, we only analyzed patients with the disseminated adult form or the juvenile form of PCM. These patients, despite their differences in clinical manifestation, exhibit similar immunological responses, as demonstrated by previous studies (19, 20).
The percentage of CD56+ cells [NK cells (CD56+CD3− cells) or NKT cells (CD56+CD3+ cells)] was similar in peripheral blood of PCM patients and control individuals (Fig. 1A). However, we were able to show that NK cells from patients with active disease exhibit reduced cytotoxic activity against K562 target cells compared with NK cells from the control group (Fig. 1B). Furthermore, after antifungal treatment, cells of PCM patients showed cytotoxic activity similar to that observed in healthy individuals (Fig. 1B).
Histological analysis showed CD56+ cells (NK or NKT cells) in the inflammatory infiltrate surrounding the granuloma (Fig. 1C) in lesions of PCM patients, indicating their possible participation in the response against the fungus. Controls, performed in the absence of a specific Ab, always produced negative results (Fig. 1D).
Direct fungicidal activity of NK cells against P. brasiliensis yeast cells
To evaluate whether NK cells exhibit cytotoxic activity against P. brasiliensis yeast cells, we performed a number of experiments using NK cells from PCM patients with active disease and healthy individuals. Cells from the control group and PCM patients were able to kill Pb18 and Pb265 yeast cells (Fig. 2A). Additionally, we found that cells from PCM patients exhibited reduced cytotoxic activity compared with cells from healthy individuals. Furthermore, the addition of IL-15 in the cultures augmented the direct cytotoxic activity of NK cells to both Pb18 and Pb265 strains in controls and patients (Fig. 2B).
To investigate the possible mechanism used by NK cells directly to kill P. brasiliensis yeast cells, NK cells were treated with specific inhibitors of granules [strontium chloride, concanamycin A (CMA), and EGTA] prior to the addition of P. brasiliensis yeast cells. The results showed that the presence of cytotoxic granules is imperative to direct cytotoxic activity, as the treatment with SrCl abolished this activity. On the contrary, the treatment with either CMA or EGTA did not interfere with the cytotoxic activity, indicating that perforin does not participate in this function (Fig. 2C).
Cytotoxic activity of NK cells against CD14+ cells infected with P. brasiliensis yeast cells
To evaluate if NK cells were able to recognize and kill infected monocytes, we infected CD14+ cells with P. brasiliensis (Pb18 or Pb265) and cocultured these cells with different concentrations of autologous purified NK cells. Fig. 3A shows that NK cells from the control individuals were able to kill CD14+ cells infected with both strains of P. brasiliensis yeast cells in a number-dependent manner. In contrast, NK cells from PCM patients only killed CD14+ cells infected with the low-virulence strain of P. brasiliensis (Pb265), also in a number-dependent manner. Furthermore, the addition of IL-15 to the cultures resulted in an increased capability of NK cells to kill infected cells (Fig. 3B).
To evaluate the mechanism used by NK cells to kill the infected monocytes, we carried out experiments using NK cells from control individuals and monocytes infected with P. brasiliensis (Pb18). The treatment of NK cells with the inhibitors of cytotoxic components showed that, like direct cytotoxicity, the killing of infected cells was dependent on the presence of granules. However, in this case, the presence of perforin was also important, as the treatment with CMA and EGTA reduced their ability to kill target cells (Fig. 3C).
Expression of cytotoxic granule components (mRNA and protein) by NK cells
As shown earlier, both direct cytotoxicity and cytotoxicity against target cells were dependent on the presence of the granules, and the addition of IL-15 to the cultures augmented these activities. To evaluate whether the stimulus of NK cells with P. brasiliensis yeast cells and/or IL-15 modifies the expression of these molecules, we evaluated the expression of mRNA by NK cells of individuals from the control group. According to Fig. 4A–D, NK cells stimulated with yeast cells from both strains (Pb18 and Pb265) show an increased expression of the mRNA for all the cytotoxic molecules analyzed. Moreover, the addition of IL-15 to the cultures resulted in a huge increment of the expression of all the mRNAs (Fig. 4A–D). The combination of IL-15 and yeast cells in the cultures did not result in an additional effect other than that observed in cultures with IL-15 alone (data not shown).
To confirm if mRNA expression was in accordance with protein expression, we analyzed the expression of granzymes (A and B), perforin, and granulysin by flow cytometry in CD56+ cells of controls and PCM patients stimulated with P. brasiliensis yeast cells and/or IL-15. We observed that these cytotoxic molecules were predominantly expressed by CD56dim cells, notably granzyme B and perforin, whereas granzyme A and granulysin were detected in both CD56dim and CD56bright populations (data not shown). However, after the stimulus with IL-15, both NK cell subpopulations expressed these molecules (except perforin) (data not shown). Fig. 4E–H shows the frequency of total CD56+ cells (CD56dim plus CD56bright) for the indicated cytotoxic molecules. The IL-15 increased the frequency of positive cells for all the molecules except granzyme A in NK cells from PCM patients and individuals from the control group; the same occurred after being stimulated by P. brasiliensis yeast cells. In addition, a higher frequency of CD56+granzyme B+ cells (with all stimuli) and CD56+perforin+ cells (IL-15 stimulus) can be seen in PCM patients in comparison with the control group (Fig. 4F, 4G). In contrast, there was a much higher frequency of CD56+granulysin+ cells in the control group than in PCM patients (Fig. 4H) in all conditions, including nonstimulated cells. The combination of IL-15 and yeast cells did not result in a synergistic effect (data not shown). Furthermore, the immunohistochemical analysis showed granulysin+, granzyme B+, and perforin+ cells in the inflammatory infiltrate surrounding P. brasiliensis yeast cells, with some of them in close contact with the fungi (Fig. 4M–O).
Additionally, when the P. brasiliensis yeast cells were used as a stimulus, the measurement of granulysin in the NK cell culture supernatants showed a higher production of this molecule in cells from healthy individuals than in cells from PCM patients (Fig. 4I), and the addition of IL-15 in the cultures augmented the production and secretion of granulysin. To analyze the effect of granulysin on P. brasiliensis yeast cells, we treated the fungal cells with different concentrations of the recombinant protein (at a concentration range similar to that released in culture supernatants) for 48 h and observed a dose-dependent fungicidal effect on both strains (Fig. 4J). To confirm the role of granulysin in the direct cytotoxic activity of NK cells on P. brasiliensis yeast cells, we performed a CFU assay in which the supernatants of IL-15–stimulated NK cells were depleted of granulysin by an immunoprecipitation procedure, which removed granulysin but maintained the presence of perforin and granzyme B (Fig. 4K). In Fig. 4L it is possible to observe that control supernatants killed P. brasiliensis yeast cells in a dose-dependent manner, whereas granulysin-depleted supernatants were unable to kill the yeast cells.
Expression of activation markers by NK cells
The results described above point to a differential activation pattern in NK cells from healthy individuals and PCM patients after the challenge with yeast cells and IL-15. To confirm these data, we analyzed the expression of activation markers by flow cytometry in both CD56bright and in CD56dim populations. The IL-15 addition resulted in an increase in the expression of CD69 and CD25 by NK cells (CD56bright and CD56dim) in both control and patient cells (Fig. 5). However, the increase in NK cells from healthy individuals was more prominent than that from PCM patients. In contrast, stimulus with P. brasiliensis yeast cells only induced the activation of NK cells from individuals of the control group, mainly of the CD56dim population (Fig. 5A, 5C).
Expression of inhibitory and stimulatory receptors by NK cells
As shown, NK cells are able to recognize and kill P. brasiliensis-infected monocytes. The recognition of target cells by NK cells is made by several stimulatory and inhibitory receptors expressed by NK cells and their ligands expressed by target cells. To investigate if the stimuli with P. brasiliensis and/or IL-15 altered the expression of these receptors on NK cells and their ligands on monocytes, we first analyzed the expression of the mRNA for NKG2D and inhibitory receptors KIR2 and KIR3DL1 in NK cells purified from control individuals. Fig. 6A shows that P. brasiliensis yeast cells and IL-15 augmented the expression of NKG2D mRNA in NK cells. We observed a great variability regarding the expression of mRNA for inhibitory receptors, although the stimulus with P. brasiliensis yeast cells as well as with IL-15 appears to increase their expression (Fig. 6B, 6C). To confirm the data obtained from the mRNA, we analyzed the expression of these receptors in NK cells from PCM patients and healthy individuals. NKG2D was detected in both CD56dim and CD56bright cells from PCM patients and control individuals (data not shown). Nevertheless, the expression of NKG2D in NK cells only increased after the stimulus with IL-15 (Fig. 6D), with similar results in both groups analyzed.
In relation to the inhibitory receptors, we found that their expression was restricted to the CD56dim population and that the stimuli used did not alter their initial levels (Fig. 6E, 6F). However, the expression of KIR2DL2/L3/S2 was higher in cells from PCM patients than in cells from healthy individuals. It is also interesting to note that NK cells from all PCM patients analyzed express the receptor KIR3DL1, whereas only a part of the control individuals presented this receptor (data not shown).
We next examined the expression of some ligands for the stimulatory and inhibitory receptors in monocytes infected with P. brasiliensis yeast cells. The results demonstrated that after the stimulus with P. brasiliensis yeast cells, there was a decrease in the number of MHC class I molecules on the surface of monocytes (Fig. 6G), which was associated with more monocytes expressing MICA/B molecules (Fig. 6H).
Production of proinflammatory cytokines by NK cells
Besides their role as cytotoxic cells, NK cells are also an important source of IFN-γ and TNF-α, which enable them to participate in the modulation of the subsequent immunological response. P. brasiliensis-stimulated or IL-15–stimulated NK cells express significant amounts of IFN-γ and TNF-α mRNAs (Fig. 7A, 7B), notably after the cytokine stimulus. Flow cytometric analysis showed that NK cells only augmented their IFN-γ and TNF-α production after being stimulated by IL-15 (Fig. 7C–E). Furthermore, we found that both the CD56bright cells and CD56dim cells are capable of producing IFN-γ (Fig. 7C and 7D, respectively), although CD56bright cells were the major source, and only CD56bright produced TNF-α. We also found a lower frequency of CD56+IFN-γ+ cells in PCM in patients compared with control patients after the addition of IL-15 (Fig. 7C, 7D), as well as after being stimulated by P. brasiliensis yeast cells (Pb18 and Pb265). The analysis of the culture supernatants showed that NK cells from healthy individuals produce and release large amounts of IFN-γ and TNF-α after treatment with IL-15 (Fig. 7F, 7G). The P. brasiliensis yeast cells, as a stimulus, also induced the production of IFN-γ but not TNF-α by NK cells.
The production of cytokines by NK cells is an important mechanism in the cross talk between these cells and other innate immune cells, such as monocytes, macrophages, and dendritic cells, which in turn produce cytokines like IL-15 as a positive feedback mechanism. As shown in all of the experiments, the presence of IL-15 in the cultures augmented the cytotoxic capacity of NK cells, as well as their production of proinflammatory cytokines. Therefore, we analyzed the ability of monocytes to produce IL-15 after being stimulated by P. brasiliensis yeast cells. As shown in Fig. 7H, monocytes infected with both strains of P. brasiliensis augmented their expression of IL-15 mRNA, and the stimulus with the Pb265 strain seems to be more effective in inducing this expression. The flow cytometric analysis of membrane-associated IL-15 in monocytes infected with P. brasiliensis yeast cells from both strains showed an increased frequency of CD14+IL-15+ cells, in comparison with uninfected cells, notably when infected with the Pb265 strain (Fig. 7I, 7J–L). Additionally, the immunohistochemical analysis of lesions from PCM patients showed IL-15+ giant cells and macrophages surrounding the granulomatous reactions (Fig. 7M).
The major aim of this study was to evaluate the possible role of NK cells in the initial immunological response against P. brasiliensis. Among the few studies about this issue, the study carried out by Peraçoli et al. (13) showed more NK cells in the peripheral blood of PCM patients, although exhibiting a diminished cytotoxic activity. Our data partially confirm these data, given that we observed that NK cells from patients with active disease did exhibit a decreased cytotoxic response against target cells. In relation to the number of CD56+ cells, we found no differences between PCM patients and healthy individuals. Additionally, we observed that after the treatment, the cytotoxic capability of NK cells from patients was similar to that observed in the control group. These data indicate that during the development of the disease, it is possible that the acquired immunological response interfere with the activation of NK cells. PCM patients with active disease are characterized by the production of large amounts of suppressive cytokines such as IL-10 and TGF-β (20–22) and by an elevated number of regulatory T cells in the circulation (23). These factors can interfere with both the acquired immunological response against the fungus and the activation of NK cells (24, 25) and can be responsible for some of the characteristics observed in NK cells from PCM patients, such as the diminished production of granulysin and the increased expression of inhibitory receptors.
This is one of the first studies, to our knowledge, to demonstrate the direct cytotoxic activity (in vitro) of NK cells on P. brasiliensis yeast cells. This activity has been studied for others pathogens, such as Cryptococcus neoformans, Toxoplasma gondii, and Candida albicans (26, 27). We also showed that NK cells are able to recognize and kill monocytes infected with P. brasiliensis yeast cells. In both activities, NK cells from PCM patients showed diminished responses compared with cells from healthy individuals. Currently, we are conducting experiments to determine the nature of the interference in the activation of NK cells in PCM patients.
Our data demonstrated that either the direct cytotoxicity or the cytotoxicity against target cells is dependent on the release of granules. However, whereas the direct cytotoxicity seems to be independent of perforin, the cytotoxicity against infected monocytes is dependent, at least partially, on the presence of this molecule. A previous study showed that perforin participates in the infection caused by Histoplasma capsulatum, as perforin-knockout mice show diminished survival rates associated with the elevation of fungal burden (10). The mechanism by which perforin acts on target cells is to some extent controversial. Currently, it is hypothesized that perforin causes an osmotic stress by forming pores in plasma membranes. Because of the repair mechanism elicited, granulysin, granzymes, and perforin pores are cointernalized into endocytic vesicles; in the cytosol, perforin disrupts endosomal vesicles, allowing access for granzymes and granulysin into cytosol (28).
The cytotoxic effect of NK cells on P. brasiliensis yeast cells was dependent on granules, but independent of perforin. The major components of cytotoxic granules, along with perforin, are granzymes (A and B mainly) and granulysin. Our results showed that besides a reduced cytotoxic activity against the yeast cells, the frequency of CD56+granulysin+ cells was much lower in PCM patients than in healthy controls. These results are in agreement with a previous study carried out by our group that showed that the serum levels of granulysin are significantly lower in PCM patients compared with healthy individuals (18). It was noteworthy that NK cells from PCM patients stimulated by IL-15 recovered their cytotoxic activity against P. brasiliensis yeast cells and infected monocytes and that recovery was accompanied by the increase in granulysin expression. In the current study, we demonstrate that NK cells release granulysin in cultures after the stimulus with P. brasiliensis yeast cells and that this molecule is able to kill P. brasiliensis yeast cells in a dose-dependent manner. Furthermore, granulysin-depleted supernatants did not exhibit fungicidal activity. For the aforementioned reasons, we believe that granulysin is the main mediator of cytotoxicity against P. brasiliensis yeast cells.
Granulysin acts on cells through the interaction with membrane lipids causing its disruption (8). This molecule plays an important role in infections caused by mycobacteria, both in tuberculosis and leprosy (29, 30). Besides its role in killing pathogens, granulysin acts as a proinflammatory molecule, attracting other cells to the infection site (9) or activating dendritic cells through TLR4 stimulation (31).
Our data showed that NK cells are able to recognize and be directly activated by P. brasiliensis yeast cells. It is interesting to note that low-virulence (Pb265) and high-virulence (Pb18) strains activate NK cells in a different manner. The major difference between these strains is the composition of their cell wall; the low-virulence strain has more β-glucans in the composition of its cell wall compared with the high-virulence strain (32). Several receptors expressed by cells of the immune system are able to recognize β-glucans, such as TLR2, dectin-1, and the complement receptor 3 (CR3) (33).
Despite a number of controversies, recent studies have demonstrated that NK cells express TLR2, but not dectin-1, and are activated by its ligands (34, 35). It was shown that NK cells can recognize mycobacteria via TLR-2 leading to increased cytotoxic activity and production of IFN-γ (35). NK cells also express CR3, and some studies have shown that the activation of NK cells by CR3 augments their cytotoxic capacity (36, 37). We are currently carrying out experiments to verify the mechanism used by NK cells to recognize the P. brasiliensis yeast cells.
Our results showed that NK cells from PCM patients exhibited higher amounts of granzyme B and perforin after stimulation than cells from healthy individuals. These data, associated with the diminished cytotoxic response, could indicate that the activation of NK cells in this group is somehow impaired. As shown, the expression of the activation markers CD69 and CD25 were higher in cells from healthy individuals than in PCM patients, notably after being stimulated by IL-15.
The presence of IL-15 is essential to the proper activation and maintenance of cytotoxic cells, both NK cells and CD8+ T lymphocytes. In this study, we confirmed the importance of IL-15 in the activation of NK cells, as in all experiments the addition of this cytokine augmented the activity of NK cells. In Candida albicans infection, IL-15 acts as a potent stimulus to the proinflammatory and antifungicidal activity of human neutrophils and monocytes (38). Neutrophils treated with IL-15 also show improved activity against P. brasiliensis yeast cells associated with the elevated production of reactive oxygen intermediaries (15).
IL-15 also induces the expression of the stimulatory NK cell receptor NKG2D (39, 40). In the current study, we confirmed these data, as both IL-15 and P. brasiliensis yeast cell stimuli lead to an increased expression of this receptor by NK cells. We also showed that NK cells from PCM patients exhibited a higher expression of the inhibitory receptors KIR2DL2/L3/S2 and KIR3DL1 compared with cells from the control group. Furthermore, we showed that monocytes infected with P. brasiliensis yeast cells have an augmented expression of MICA/MICB molecules, which are the ligands to the stimulatory receptor NKG2D, in addition to a reduction in the expression of MHC class I molecules (ligands to the inhibitory NK receptors). These data indicate a possible mechanism by which NK cells recognize and are activated to kill the infected monocytes, but further studies will be necessary to determine the exact mechanism used by NK cells to kill P. brasiliensis-infected target cells, including other receptors not included in this study.
Another interesting result was the capacity of NK cells to produce proinflammatory cytokines (IFN-γ and TNF-α) in response to stimuli. The production of IFN-γ by NK cells is induced by the direct engagement of their receptors or as a response to the presence of inflammatory cytokines such as IL-12, IL-15, and IL-18 (41, 42). We found that monocytes infected with P. brasiliensis yeast cells augmented their expression of IL-15 mRNA, as well as their expression of the membrane-associated form of this cytokine. We can therefore make the assumption that the interaction of NK cells and APCs is important in the response against P. brasiliensis infection, as described for other infectious diseases (43–45).
In conclusion, our results demonstrated that NK cells are able to participate actively in the immune response against the P. brasiliensis infection, either by directly killing the yeast cells or by recognizing and killing the infected cells. Granulysin is the possible mediator of the cytotoxic effect, as the reduced cytotoxic activity against the yeast cells detected in patients with PCM is accompanied by a significantly lower frequency of CD56+granulysin+ cells compared with healthy controls. Furthermore, we show that NK cells released granulysin in cultures after being stimulated by P. brasiliensis yeast cells, and this molecule is able to kill P. brasiliensis yeast cells in a dose-dependent manner. Another important finding is that NK cells are able to produce IFN-γ and TNF-α, which could influence the subsequent acquired immunological response by stimulating other cells such as dendritic cells, macrophages, and lymphocytes.
The authors have no financial conflicts of interest.
This work was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo, Conselho Nacional de Desenvolvimento Científico e Tecnológico, and Fundação Salvador Arena.
Abbreviations used in this article:
- concanamycin A
- complement receptor 3
- lactate dehydrogenase
- quantitative RT-PCR.
- Received September 7, 2011.
- Accepted May 21, 2012.
- Copyright © 2012 by The American Association of Immunologists, Inc.