https://orcid.org/0000-0002-7277-2646
Abstract
Candida albicans and Candida glabrata are the 2 most prevalent Candida species causing bloodstream infections. Patterns of innate immune activation triggered by the 2 fungi differ considerably.
To analyze human natural killer (NK) cell activation by both species, we performed ex vivo whole-blood infection assays and confrontation assays with primary human NK cells.
C. albicans was a stronger activator for isolated human NK cells than C. glabrata. In contrast, activation of blood NK cells, characterized by an upregulated surface exposure of early activation antigen CD69 and death receptor ligand TRAIL, as well as interferon-γ (IFN-γ) secretion, was more pronounced during C. glabrata infection. NK cell activation in blood is mediated by humoral mediators released by other immune cells and does not depend on direct activation by fungal cells. Cross-talk between Candida-confronted monocyte-derived dendritic cells (moDC) and NK cells resulted in the same NK activation phenotype as NK cells in human blood. Blocking experiments and cytokine substitution identified interleukin-12 as a critical mediator in regulation of primary NK cells by moDC.
Activation of human NK cells in response to Candida in human blood mainly occurs indirectly by mediators released from monocytic cells.
Natural killer (NK) cells belong to the group 1 innate lymphoid cells and are characterized by the presence of the neural cell adhesion molecule CD56 and the absence of CD3 [1–3]. They recognize and kill tumor or virus-infected cells via missing-self signaling, induced by a downregulation of major histocompatibility complex class I molecules [4]. NK cells have different effector mechanisms to eliminate target cells: (1) exocytosis of cytotoxic granules containing perforin and granzymes, (2) death receptor-induced target-cell apoptosis, and (3) antibody-dependent cellular cytotoxicity mediated by Fcγ-receptor III (CD16) [5]. NK cells also rely on signals from other cells to exhibit their cytotoxicity and produce cytokines. Several studies have demonstrated the functional link between NK cells and DC and showed the reciprocal activation of these cells by soluble factors or direct interaction [6]. In addition to their cytotoxic activity, NK cells participate in early responses to infection through production of cytokines, especially interferon-γ (IFN-γ) [7]. The importance of NK cells in fungal infections has been shown in vitro and in vivo. NK cells exhibit antifungal activity against Candida albicans, Aspergillus fumigatus, Cryptococcus neoformans, and other species [8–11]. Several receptor molecules for fungal pathogens have been identified on NK cells, including NKp30-recognizing β-1,3-glucan, NKp46, and CD56 [12–15]. Depending on the host immune status, NK cells play an essential role on the outcome of systemic C. albicans infection and invasive aspergillosis in murine infection models [16, 17]. Although C. albicans and C. glabrata account for the majority of cases of candidiasis worldwide [18, 19], the interplay of the 2 species with the human immune system is rather different. Distinct targeting of immune-cell populations results in a strong neutrophil-driven response during C. albicans infection [20]. Neutrophils rapidly phagocytose C. albicans upon confrontation and are the only immune cells able to inhibit intracellular filamentation [21]. In contrast, C. glabrata induces a low-grade inflammatory response, recruiting and activating monocytes rather than neutrophils in human blood and during murine infection [20]. Focusing on the interaction between NK and fungal cells, recognition of C. albicans is mediated by the natural cytotoxicity receptor NKp30 [11], while NKp46 and its mouse ortholog NCR1 are involved in binding and damage of C. glabrata [13]. Direct contact between C. albicans and primary NK cells results in release of perforin and granzyme B from cytotoxic granules and cytokine secretion, thereby directing fungal damage and enhancing neutrophil functionality in cocultivation experiments, respectively [22]. In addition, cross-talk between NK cells and other immune cells during fungal infection has been demonstrated [23, 24].
In this study, we showed different NK-cell stimulation patterns depending on the milieu and Candida species. Whereas activation of blood NK cells was stronger with C. glabrata, contact-dependent stimulation of expanded primary NK cells was higher in the presence of C. albicans and triggered degranulation of secretory granules. The immunophenotype of blood NK cells with upregulation of death receptor ligand TRAIL and CD69 was also observed for NK cells in the presence of DC-derived cytokines released upon fungal contact, proving an indirect NK-cell activation by soluble factors during blood infection. Our data demonstrate that differences in interleukin-12 (IL-12) secretion by DC and in blood induced by the 2 Candida species are responsible for different levels of NK-cell activation.
METHODS
Ethics Statement
This study was conducted in accordance with the Declaration of Helsinki. All protocols were approved by the ethics committee, University Hospital Jena (permit number, 273-12/09).
Fungal Strains and Culture
C. albicans strain SC5314 [25] and C. glabrata strain ATCC2001 [26] were grown overnight in yeast peptone d-glucose (YPD) medium at 30°C and 37°C, respectively. Both fungal species were reseeded in fresh YPD medium and grown until they reach mid-log phase. When required, they were cultured overnight in M199 medium, pH 4 at 37°C, reseeded, and grown in M199 medium, pH 8 at 37°C, which induced filamentous growth in C. albicans. Candida cells were killed by incubation in 0.1% thimerosal (Sigma-Aldrich).
Human Whole-Blood Infection Model
Human peripheral blood was collected from healthy donors after written informed consent. The whole-blood infection assay was performed as described previously [27]. Briefly, live or inactivated fungal cells were added to Hirudin-anticoagulated blood (1 × 106 fungi/mL) and incubated at 37°C for indicated time points. In some experiments, aliquots of whole blood were preincubated with a blocking antibody against IL-12 (clone 24910; R&D Systems) or the corresponding isotype control (mouse IgG1, clone MOPC-21; BioLegend).
Isolation and Expansion of Human NK Cells
Human peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats of healthy donors by standard Ficoll gradient centrifugation (Biochrom AG). Untouched NK cells were isolated from PBMCs by negative selection using a NK-cell isolation kit (MiltenyiBiotec). NK-cell purity was checked by flow cytometry (>95% CD56+/CD3−/CD14− cells). Primary human NK cells were either used directly after isolation or further expanded by cytokine (all ImmunoTools) treatment. To generate cytokine-primed NK cells, freshly isolated NK cells were seeded at a concentration of 2 × 106/mL in stem cell growth medium (SCGM; Cell Genix) containing 10% human serum (Sigma-Aldrich) and 100 U/mL IL-2. After 3 days of incubation (37°C and 5% CO2), half of the medium was exchanged with SCGM plus 10% human serum supplemented with 100 U/mL IL-2, 50 ng/mL IL-15, 1000 U/mL IFN-α, and 2000 U/mL IFN-β, and incubated for another 3 days. NK cells were harvested and resuspended in fresh SCGM plus 10% human serum.
Monocyte Isolation and Generation of Monocyte-Derived Dendritic Cells
Monocytes were isolated from PBMCs by positive selection using CD14 microbeads (MiltenyiBiotec). Freshly isolated monocytes were used to generate monocyte-derived dendritic cells (moDC) by incubation in Roswell Park Memorial Institute medium (RPMI; Biochrom AG) containing 10% heat-inactivated fetal bovine serum (hiFBS; Biochrom AG) in the presence of 800 U/mL granulocyte-macrophage colony stimulating factor, sargramostim (Leukine) and 1000 U/mL IL-4 (MiltenyiBiotec) for 7 days. Differentiation of monocytes into moDC was checked by flow cytometry, with results of >85% CD1a+/CD14− cells.
NK-Cell Confrontation Assay
Confrontation of cytokine-primed NK cells with fungi was performed in SCGM plus 10% human serum at a multiplicity of infection (MOI) of 0.5 for 2 hours or 4 hours at 37°C and 5% CO2. Due to prior cytokine priming of NK cells, longer coincubations were not performed. To investigate the direct effect of IL-12 on the regulation of NK-cell surface marker expression, 1 × 106/mL freshly isolated NK cells were incubated with 2 μg/mL IL-12 (ImmunoTools) for 40 hours at 37°C and 5% CO2.
Transwell Assay
Either 1 × 106 monocytes or moDC were seeded into the upper compartment of a Transwell plate (Corning) and mock-infected or confronted with 5 × 105 inactivated C. albicans and C. glabrata, respectively (moDC:Candida ratio of 2:1; MOI = 0.5) in RPMI plus 10% hiFBS. The upper compartment was separated by a microporous membrane (pore size, 0.4 μm) from the lower compartment that contained 1 × 106 freshly isolated NK cells in RPMI plus 10% hiFBS. When indicated, 10 ng/mL mouse IgG1 or anti–IL-12 antibody were added to the upper compartment.
Flow Cytometry
Analyses of isolated immune cells (NK cells and moDC) and NK cells in whole blood were performed using differential fluorescence-activated cell sorting (FACS) staining and subsequent measurement with a BD FACSCanto II. NK cells were specifically identified by a CD3− and CD56+ staining. Changes in surface marker expression were investigated using antibodies listed in Supplementary Table 1. Stained blood samples were treated with BD FACS lysing solution followed by washing and harvesting cells in BD CellWASH solution. Analysis of flow cytometry data was performed using FlowJo 7.6.4 software.
Quantification of Secreted Proteins
Protein quantification within supernatants of isolated cells or in plasma samples obtained from whole-blood experiments was performed using Luminex technology (ProcartaPlex and High Sensitivity ProcartaPlex Immunoassay; ThermoFisher Scientific).
Statistical Analyses
All statistical analyses were performed using GraphPad Prism software. P values were determined using 1-way ANOVA followed by multiple comparison tests.
RESULTS
NK-Cell Stimulation Is Dependent on the Milieu and Candida Species
Previous experiments using cytokine-activated primary human NK cells have shown that C. albicans predominantly induces downregulation of CD16 and increase in CD107a surface exposure, related to active release of cytotoxic granule content [22]. In agreement, we observed a 1.7-fold lower downregulation for CD16 and 1.5-fold higher CD107a expression with C. albicans compared to C. glabrata after 2 hours of confrontations (Supplementary Figure 1). Differences were even more marked 4 hours postinfection: 2.5-fold lower downregulation for CD16 and 2.7-fold higher CD107a expression with C. albicans, indicating a weaker NK-cell activation potential for C. glabrata (Figure 1A). No regulation of death-inducing ligands TRAIL and FasL was detected. However, experiments with isolated cells do not necessarily reflect in vivo conditions due to missing interactions with other components of the host response [28]. Thus, to further analyze NK-cell activation we used a human whole-blood model of infection [27]. In contrast to stimulation of isolated NK cells, surface levels of CD16 and CD107a did not change after confrontation of whole blood with either C. albicans or C. glabrata (Figure 1B). However, both species induced a markedly increased surface exposure of death-inducing ligand TRAIL and early activation antigen CD69 (Figure 1B), while these markers were only slightly increased on isolated NK cells (Figure 1A). No changes in surface expression of FasL during NK-cell activation could be detected for any tested condition (data not shown). Furthermore, in sharp contrast to data obtained for primary NK cells, confrontation with C. glabrata induced a significantly stronger activation of blood NK cells than with C. albicans. Whole-blood infection with inactivated fungal cells revealed the same differences in NK-cell activation as induced by live C. albicans and C. glabrata (Supplementary Figure 2). Furthermore, both C. albicans wild type and nonfilamentous cph1Δ/efg1Δ mutant showed comparable NK-cell stimulation phenotype, indicating that filamentation is not responsible for a C. albicans-specific NK-cell activation pattern (Supplementary Figure 3).
Milieu and Candida species influence NK-cell stimulation. A, Primary expanded NK cells (MOI = 0.5, 4 hours) or (B) human whole blood (1 × 106 fungal cells/mL, 8 hours) were either mock infected or confronted with C. albicans and C. glabrata and analyzed for surface expression of NK cell activation markers. In general, C. albicans induced higher activation of primary expanded NK cells compared to C. glabrata, whereas NK-cell activation in whole blood was stronger during confrontation with C. glabrata than with C. albicans. Changes in the surface expression levels of NK cell activation markers CD16, CD107a, CD69, and TRAIL are shown after confrontation with C. albicans (black line) and C. glabrata (grey line). Filled bars indicate basal expression of NK cells from mock-infected samples. Data from 1 of at least 9 independent experiments using cells from different donors with virtually identical results are shown. Quantitative analysis was performed for each surface marker using 1-way ANOVA followed by multiple comparison tests. Data are means ± SEM normalized to basal levels of mock-infected NK cells (set to 100%, filled bars). C, IFN-γ secretion was measured in supernatants collected following 4 hours infection of primary expanded NK cells from 9 independent experiments (left) and plasma collected after 8 hours of whole-blood infection from 8 independent donors (right) with either HBSS (mock infected; squares), C. albicans (triangles), or C. glabrata (circles). C. albicans induced significant higher release of IFN-γ by primary NK cells, in contrast to the higher secretion of IFN-γ by blood NK cells in response to C. glabrata. Quantitative analysis was performed using paired 1-way ANOVA followed by multiple comparison tests. Lines connect datapoints from the same donors. * P < .05, ** P < .01, *** P < .001. Abbreviations: C.a., C. albicans; C.g., C. glabrata; exp., expanded; NK, natural killer; MOI, multiplicity of infection; IFN, interferon; HBSS, Hanks balanced salt solution.
Altogether, our results identified differences in the stimulation of NK cells, dependent on the milieu and Candida species, between ex vivo whole-blood infection and expanded primary cells. This was further supported by analysis of IFN-γ release after fungal infection. The proinflammatory cytokine was mainly released by NK cells and induction was stronger by C. albicans during primary NK-cell confrontation compared to the higher levels detected after blood infection with C. glabrata (Figure 1C).
Different Levels of Cytokines Are Induced by C. albicans and C. glabrata During Whole-Blood Infection
We previously demonstrated that activation of primary human NK cells by C. albicans depends on direct contact, whereas the fungus exclusively interacts with neutrophils and monocytes during whole-blood infection, suggesting an indirect induction of NK-cell effector mechanisms by soluble blood components [22, 27]. To further investigate which stimuli could be involved in the activation of blood NK cells, the release of cytokines into plasma generated from whole-blood infection experiments in response to C. albicans and C. glabrata was quantified. Interestingly, release of monocyte cytokines, such as IL-1β, IL-6, tumor necrosis factor-α (TNF-α), and IL-12, was significantly higher during 4 hours of incubation with C. glabrata (Figure 2A), which is in line with our previously published data showing a greater association of C. glabrata with monocytes in the whole-blood assay [20]. Whereas differences in the IL-6 and TNF-α secretion profile disappeared after 8 hours, differences in plasma levels of IL-1β induced by the 2 Candida species were still clearly present (blood plus C. albicans, 436 ± 98 pg/mL; blood plus C. glabrata, 1413 ± 268 pg/mL) and IL-12 (blood plus C. albicans, 92 ± 35 pg/mL; blood plus C. glabrata, 596 ± 161 pg/mL) (Figure 2B). Plasma collected from whole blood infected with inactivated fungal cells presented a lower secretion of cytokines compared to live fungal cells, but the cytokine profile was comparable to that induced by live C. albicans and C. glabrata 8 hours postinfection (Figure 2C). Even after infecting whole blood with the nonfilamentous C. albicans cph1Δ/efg1Δ mutant, the same cytokine secretion pattern was maintained, especially the higher IL-12 secretion with C. glabrata infection compare to the wild-type and nonfilamentous C. albicans (Figure 2D). Other studies have demonstrated a role for IL-12 in stimulation of IFN-γ production by NK cells, thus making IL-12 an obvious candidate for a putative blood-specific stimulus regulating NK-cell function [29]. In addition, we checked the secretion of other cytokines particularly involved in NK-cell priming and activation and found no (IL-15, IL-23, IL-27) or only slightly increased (IL-2, IL-18) plasma levels compared to mock-infected blood upon fungal confrontation, with no differences between the 2 Candida species (Supplementary Figure 4). Another central factor involved in the immune response against Candida infections is complement activation [30]. Among the complement products, C5a induces a range of proinflammatory effects [31–33]. C5a secretion was measured in plasma collected from whole blood infected either by C. albicans or C. glabrata 4 hours postinfection. C5a was produced during both fungal infections but its secretion was higher with C. albicans (180 ± 18 hours, n = 8) than C. glabrata (69 ± 6 hours, n = 8) (Supplementary Figure 5).
Cytokine secretion during whole-blood infection with Candida albicans and C. glabrata. Secretion of cytokines was measured in plasma samples after 4 hours (A) and 8 hours (B) confrontation of human whole blood with either HBSS (mock infected), C. albicans, or C. glabrata and revealed significantly different levels between the Candida species. Plasma was also collected from whole blood infected with thimerosal-inactivated fungal cells (C) and nonfilamentous C. albicans cph1Δ/efg1Δ (D), and cytokines secretion was measured. Bars show means ± SEM of at least 4 independent experiments with blood from different donors. P values were determined using 1-way ANOVA followed by multiple comparison tests. * P < .05, ** P < .01, *** P < .001. Abbreviations: C.a., C. albicans; C.g., C. glabrata; ctrl, control; HBSS, Hanks balanced salt solution; IL, interleukin; TNF-α, tumor necrosis factor-α.
Dendritic Cell-Derived Soluble Factor(s) Induce Differential NK-Cell Activation
Monocytes and dendritic cells can release large amounts of cytokines upon activation. Unfortunately, isolated monocytes showed a limited lifespan upon confrontation with inactivated Candida cells (Supplementary Figure 6) and were not suitable for cross-talk experiments with NK cells using the Transwell system. Therefore, we further investigated the induction of NK-cell effector mechanisms by cytokines released from stimulated moDC. Interestingly, there was an evident upregulation of CD69 and TRAIL on primary NK cells as shown for the blood NK cells (Figure 3A). In each case, C. glabrata confrontation increased surface levels more than C. albicans (eg, CD69 increase with C. glabrata 901 ± 107% and with C. albicans 598 ± 76%). In contrast, CD16 and CD107a expression showed only a slight response, indicating no role in the NK-cell effector mechanisms induced by moDC-derived cytokines. In addition, we analyzed other markers of NK-cell activation and found CD38, an adhesion molecule that triggers NK-cell cytotoxicity, to have significantly higher induction on the NK-cell surface in response to C. glabrata than C. albicans (Figure 3A). Other surface markers, such as CD56, HLA-DR, and NKp30, were upregulated in response to cytokines released by moDC upon fungal infection, but no differences in the expression between C. albicans and C. glabrata infection were detected. The lower NK-cell stimulation induced by cytokines released from moDC in response to C. albicans also resulted in lower IFN-γ levels present in the NK-cell compartment. Despite the high donor-dependent variance, IFN-γ secretion by primary human NK cells was stronger during NK cell-moDC-C. glabrata coincubation in the majority of cases (Figure 3B).
Surface activation marker levels on primary NK cells changed in response to DC-derived soluble factor(s) induced by Candida. A, NK cells were harvested from the lower compartment of the Transwell system after 40 hours of incubation without moDC or coincubation with moDC located in the upper compartment that were either mock infected or infected with C. albicans or C. glabrata. Changes in the surface expression levels of NK cell activation markers CD69, TRAIL, CD16, CD107a, CD38, HLA-DR, CD56, and NKp30 are shown after normalization to basal levels of NK cells in the absence of moDC (set to 100%). Bars show means ± SEM of at least 3 independent experiments with cells isolated from different donors and P values were determined using 1-way ANOVA followed by multiple comparison tests. B, Supernatants were collected from the lower compartment of the Transwell system that contained primary NK cells and analyzed for IFN-γ secretion in response to moDC-derived soluble factors. Before-after graph shows higher IFN-γ release after 40 hours of confrontation with C. glabrata (circles) compared to C. albicans (triangles) in 12 independent experiments with cells isolated from different donors. Lines connect datapoints from the same donors. P values correspond to the means ± SEM of all experiments obtained with paired 1-way ANOVA followed by multiple comparison tests. * P < .05, *** P < .001. Abbreviations: C.a., C. albicans; C.g., C. glabrata; IFN-γ, interferon-γ; moDC, monocyte-derived dendritic cells; NK, natural killer.
Taken together, these results show that cytokines released by moDC upon fungal confrontation were able to drive primary human NK-cell activation, which was equal to the differential NK-cell activation levels during whole-blood infection with C. albicans and C. glabrata.
Dendritic Cell-Derived Soluble Factor(s) Are Released in Response to C. albicans and C. glabrata Stimulation
To identify the specific stimulus that mediates NK-cell activation, supernatants collected from the upper compartment of the Transwell system were analyzed to gain further insight into cytokines released by stimulated moDC during Candida confrontation. In agreement with cytokine profiles during whole-blood infection, both species induced secretion of IL-6, TNF-α, and IL-12 that was higher during C. glabrata than C. albicans infection (eg, IL-12, moDC plus C. albicans 2517 ± 47 7pg/mL and moDC plus C. glabrata 6118 ± 972 pg/mL) (Figure 4A). However, IL-12 levels released on moDC activation were markedly higher than induced blood levels (with C. albicans 27-fold higher and with C. glabrata 11-fold higher). The concentration of IL-1β in supernatants of infected moDC was at the limit of detection and comparable to the spontaneous release by mock-infected moDC, indicating no infection-induced secretion like in whole blood. Our data also showed the flow of cytokines into the lower compartment, which maintained the same differences in their supernatant levels between confrontation with C. albicans and C. glabrata (Figure 4B).
Cytokine release by moDC upon Candida albicans and C. glabrata stimulation. Supernatants were collected from both the upper compartment (A), where moDC were either mock treated (light grey filled bars) or confronted with C. albicans (black open bars) and C. glabrata (grey open bars), and the lower compartment (B) of the Transwell system. Bars show means ± SEM of at least 4 independent experiments with cells isolated from different donors. P values were determined using 1-way ANOVA followed by multiple comparison tests. ** P < .01, *** P < .001. Abbreviations: C.a., C. albicans; C.g., C. glabrata; IL, interleukin; moDC, monocyte-derived dendritic cells; NK, natural killer; TNF-α, tumor necrosis factor-α.
MoDC-Derived IL-12 Triggers the IFN-γ Release by NK Cells
IL-12 is a good candidate to be the factor responsible for the different levels of NK-cell activation induced by the 2 fungal species: on the one hand it is differentially produced by moDC after confrontation and on the other hand it has been shown to induce IFN-γ release by NK cells [6]. To determine the role of moDC-derived IL-12, induction of NK effector mechanisms was quantified in the presence of an IL-12 neutralizing antibody. Blocking of IL-12 resulted in inhibition of release of IFN-γ by NK cells in response to either C. albicans (with IgG1, 500 ± 97 pg/mL; with anti–IL-12, 52 ± 14 pg/mL) or C. glabrata (with IgG1, 728 ± 204 pg/mL; with anti–IL-12, 28 ± 3.5 pg/mL) with levels close to background detected during coincubation of NK cells with mock-treated moDC (7.5 ± 3.2 pg/mL) (Figure 5). Furthermore, blocking of IL-12 significantly decreased the strong CD69 upregulation during C. albicans (with IgG1, 580 ± 83%; with anti–IL-12, 256 ± 19%) and C. glabrata (with IgG1, 887 ± 133%; with anti–IL-12, 499 ± 85%) infection, whereas CD69 surface levels were not affected by the isotype-matched control antibody (Figure 6). However, the inhibitory effect of the anti–IL-12 antibody on CD69 regulation was only partial because we could still detect a 5-fold and 2.6-fold higher CD69 expression in response to C. glabrata and C. albicans, respectively. Interestingly, increased CD56, NKp30, and HLA-DR surface expression levels were reduced to almost background levels on nonstimulated NK cells. The effect of IL-12 on induction of CD56, NKp30, HLA-DR, and CD69 on NK cells was confirmed by direct treatment of primary cells with recombinant human IL-12, whereas no changes in the surface expression of TRAIL and CD38 could be detected (Supplementary Figure 7). In line with this, TRAIL and CD38 showed no significant differences between anti–IL-12, IgG1, and nontreated coincubations in response to both Candida species (Figure 6). Checking CD69 and TRAIL expression on NK cells, as well as IFN-γ release during whole-blood infection in the presence of anti–IL-12, the increased CD69 surface levels and secretion of IFN-γ induced by C. glabrata were partially blocked, whereas TRAIL surface exposure was not affected (Figure 7). In contrast, we observed only marginal effects of the blocking antibody during C. albicans infection.
IL-12 contributes to IFN-γ release by primary NK cells. MoDC infected with Candida albicans (triangles) or C. glabrata (circles) were either nontreated or incubated in the presence of an isotype control (IgG) or anti-IL-12 blocking antibody (αIL-12). IFN-γ secretion was measured within supernatants obtained from the lower compartment of the Transwell system that contained primary NK cells. Presence of αIL-12 inhibited release of IFN-γ by NK cells in response to both Candida species. Each data point in the before-after graphs represents an independent experiment for a total of 8 experiments. Lines connect datapoints from the same donors. Quantitative analysis was performed using paired 1-way ANOVA followed by multiple comparison tests and P values correspond to the means ± SEM of all experiments. * P < .05, ** P < .01. Abbreviations: C.a., C. albicans; C.g., C. glabrata; IgG, immunoglobulin G; IL, interleukin; IFN-γ, interferon-γ; moDC, monocyte-derived dendritic cells; NK, natural killer.
Effects of IL-12 on NK cell surface marker expression. The surface expression of activation markers on primary NK cells during coincubation with either Candida albicans-infected (triangles) or C. glabrata-infected moDC (circles) in a Transwell system was evaluated in the presence of IL-12 neutralizing antibody (αIL-12) and compared to nontreated or IgG-treated samples. Changes in the surface expression levels of CD69, TRAIL, CD38, HLA-DR, CD56, and NKp30 are shown after normalization to basal levels of NK cells in the absence of moDC (set to 100%, not shown). Blocking of IL-12 resulted in a significant decrease of CD69, CD56, NKp30, and HLA-DR on NK cells during Candida confrontation. Levels of the activation markers were not affected by the isotype-matched control antibody. Quantitative analysis was performed using paired 1-way ANOVA followed by multiple comparison tests. Each data point in the before-after graphs represents an independent experiment of at least 7 experiments. Lines connect datapoints from the same donors. P values correspond to the means ± SEM of all experiments with primary NK cells isolated from different donors. ** P < .01, *** P < .001. Abbreviations: C.a., C. albicans; C.g., C. glabrata; IL, interleukin; moDC, monocyte-derived dendritic cells; NK, natural killer.
Blocking of IL-12 partially decreases NK cell activation in human whole blood. Human whole blood was treated with an anti-IL-12 antibody (αIL-12) prior to inoculation of either Candida albicans (triangles) or C. glabrata (circles) for 8 hours to investigate the impact of IL-12 on activation of blood NK cells. Samples were compared with either nontreated or IgG-containing Candida-infected blood. After whole-blood infection, NK cell surface exposure of CD69 and TRAIL (A) and secretion of IFN-γ within plasma samples (B) were analyzed in at least 7 independent donors. In comparison with non- and IgG-treated infected samples, presence of αIL-12 partially prevented the increase in CD69 expression and plasma levels of IFN-γ in response to C. glabrata. Each data point in the before-after graphs represents an independent experiment for a total of 8 experiments. Lines connect datapoints from the same donors. * P < .05. Abbreviations: C.a., C. albicans; C.g., C. glabrata; IgG, immunoglobulin G; IL, interleukin; IFN-γ, interferon-γ; NK, natural killer.
Taken together, these results indicate that differences in IL-12 secretion by moDC in response to C. albicans and C. glabrata are responsible for the differential secretion of IFN-γ by NK cells. However, differences in the exposure of surface activation markers CD69, TRAIL, and CD38 seemed to be mediated by additional factors that may also be important for regulation within blood.
Discussion
This study aimed to compare NK-cell activity against C. albicans and C. glabrata, the 2 common fungal species responsible for invasive candidiasis. Both species are known to differently interact with the innate immune system. In previous studies, we demonstrated that C. albicans induces a strong and rapid immune response in blood, which is predominantly mediated by neutrophils [20, 27, 28]. In contrast, the response against C. glabrata is characterized by slower neutrophil activation with reduced amplitude and a more pronounced activation of monocytes [20, 27].
Our present data reveal that primary NK cells show a significantly stronger response against C. albicans than C. glabrata. This response is characterized by an upregulation of the degranulation pathway and downregulation of Fcγ receptor III, and requires direct contact between NK cells and fungi [22]. However, human NK-cell activation has a low efficiency for C. albicans killing and is followed by destruction of the immune cells due to elongation of engulfed C. albicans [22, 34]. In contrast, isolated mouse NK cells are able to kill C. glabrata upon direct interaction [13]. Neither of these effects could be observed during infection of human whole blood. Blood NK cells were largely activated by soluble factors released by other immune cells. This indirect activation of blood NK cells was considerably more pronounced for C. glabrata compared to C. albicans infection. Furthermore, the activation phenotype differed from that observed in isolated NK cells. Activation of blood NK cells was characterized by an upregulation of CD69 and the death receptor ligand TRAIL, whereas degranulation pathway and Fcγ receptor III were not involved. The blood activation phenotype could be mimicked by coincubation with Candida-activated moDC. Furthermore, the lower degree of activation of blood NK cells during C. albicans infection could be explained by the presence of other and faster immune cells involved in the control of elimination of C. albicans, such as polymorphonuclear cells [27] and by different levels of secreted cytokines, as indicated by the moDC-NK cell-Candida coincubation experiments. The different activation profiles expressed by NK cells dependent on the milieu and Candida species were also reflected in different IFN-γ secretion. IFN-γ was highly secreted by primary NK cells in response to C. albicans infection. For blood NK cells and during NK cell-moDC coculture, levels were higher for C. glabrata. The important role of IFN-γ as a proinflammatory cytokine to control Candida infections was previously demonstrated in vivo [35]. IFN-γ knockout mice are more sensitive to invasive C. albicans infection than wild-type mice, where IFN-γ is able to improve C. albicans phagocytosis and killing by neutrophils and macrophages [36–38]. IFN-γ is also known to enhance the activity of antigen-presenting cells [39], driving among other things the secretion of T helper (Th1)-inducing cytokines.
IL-12 was the most promising candidate for stimulating NK-cell activation and its release was significantly higher during infection of whole blood and moDC with C. glabrata. Furthermore, soluble factors secreted by C. albicans can suppress IL-12 secretion by monocytes [40, 41]. NK cells expose the high-affinity IL-12 receptor (IL-12Rβ1/β2) on their surface. IL-12–IL-12R interaction leads to activation of the Jak-STAT pathway, thereby inducing NK-cell activation and production of IFN-γ [42]. The regulatory function of moDC-derived IL-12 in NK-cell activation was verified using an IL-12 neutralizing antibody that blocked IFN-γ secretion by NK cells during C. albicans and C. glabrata infection, while the expression of surface activation markers was either partial (CD69) or not affected (TRAIL and CD38). In line with this, differences in IL-12 secretion in response to both Candida species are also partially responsible for different levels of NK-cell activation in blood.
Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Financial support. This work was supported by the Deutsche Forschungsgemeinschaft within the Collaborative Research Center CRC124 FungiNet (projects A2 to J. L., B4 to M. T. F. and C3 to O. K.); and by the Jena School for Microbial Communication.
Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
