Natural killer cells are pivotal for in vivo protection following systemic infection by Sporothrix schenckii

Summary

Natural killer (NK) cells are one of the first cell types to enter inflammation sites and have been historically known as key effector cells against tumours and viruses; now, accumulating evidence shows that NK cells are also capable of direct in vitro activity and play a protective role against clinically important fungi in vivo. However, our understanding of NK cell development, maturation and activation in the setting of fungal infections is preliminary at best. Sporotrichosis is an emerging worldwide‐distributed subcutaneous mycosis endemic in many countries, affecting humans and other animals and caused by various related thermodimorphic Sporothrix species, whose prototypical member is Sporothrix schenckii. We show that following systemic infection of BALB/c mice with S. schenckii sensu stricto, NK cells displayed a more mature phenotype as early as 5 days post‐infection as judged by CD11b/CD27 expression. At 10 days post‐infection, NK cells had increased expression of CD62 ligand (CD62L) and killer cell lectin‐like receptor subfamily G member 1 (KLRG1), but not of CD25 or CD69. Depletion of NK cells with anti‐asialo GM1 drastically impaired fungal clearance, leading to a more than eightfold increase in splenic fungal load accompanied by heightened systemic inflammation, as shown by augmented production of the pro‐inflammatory cytokines tumour necrosis factor‐α, interferon‐γ and interleukin‐6, but not interleukin‐17A, in the spleen and serum. Our study is, to the best of our knowledge, the first to demonstrate that a fungal infection can drive NK cell maturation in vivo and that such cells are pivotal for in vivo protection against S. schenckii.

Abbreviations

APC

allophycocyanin

BM

bone marrow

cNK

conventional NK

DC

dendritic cell

dpi

days post‐inoculation

GM‐CSF

granulocyte–macrophage colony‐stimulating factor

IFN‐γ

interferon‐γ

IL‐17Rα

interleukin‐17 receptor α

KLRG1

killer cell lectin‐like receptor subfamily G member 1

MFI

median fluorescence intensity

NCR

natural cytotoxicity receptor

NK

natural killer

NLRP3

nucleotide‐binding oligomerization domain‐like receptors

PE

phycoerythrin

PerCP‐Cy5

peridinin chlorophyll protein‐Cychrome 5

PRR

pattern recognition receptor

TLR

Toll‐like receptor

TNF‐α

tumour necrosis factor‐α

Introduction

Natural killer (NK) cells are one of the first cell types to enter inflammation sites and are capable of rapidly eliminating allogeneic or tumour cells without the need for previous sensitization and in the absence of recognition by recombination activating gene receptors, like the T‐cell receptor and antibodies. These cells originate in the bone marrow (BM) and migrate through the bloodstream to the spleen, liver, lungs and many other organs, keeping a dynamic distribution owing to their ability to re‐circulate between them.1, 2 NK cell development from the earliest NK cell‐committed precursor to functionally competent mature NK cells is a stepwise process that involves the sequential acquisition of a series of surface molecules.3 As proposed by Gotthardt et al.,4 NK precursors lose expression of CD127 [interleukin‐17 receptor α (IL‐7Rα)] and then sequentially acquire expression of CD122 (IL‐2/15Rβ), NK1.1, CD49b, NKp46, CD11b, and then, finally, of CD43 and killer cell lectin‐like receptor subfamily G member 1 (KLRG1). Additionally, murine NK cells can be further dissected into four distinct maturation stages defined by the differential surface expression of CD11b (α _M integrin) and CD27 [a tumour necrosis factor (TNF) receptor family member], from the least to the most mature, as follow: CD11b⁻ CD27⁻ → CD11b⁻ CD27⁺ → CD11b⁺ CD27⁺ → CD11b⁺ CD27⁻.5 As they mature, NK cells lose proliferative potential and become more cytotoxic.3, 5 Activation by cytokines, microbial products or products expressed on abnormal or infected cells can lead to loss of CD27 or to increased expression of CD11b, Ly6C, CD43, KLRG1, CD25 (IL‐2Rα), CD69 (very early antigen), and stem cell antigen 1, as well as to NK cell proliferation and increased cytotoxicity and cytokine production.6, 7, 8, 9 However, our current understanding of NK cell biology in disease conditions comes almost entirely from the setting of tumours and viral infections, with data on NK cell modulation by fungal infections being preliminary at best.

Most studies show that both human and murine NK cells exhibit in vitro activity against diverse fungi, either directly through perforin and granzyme secretion or indirectly through secretion of constitutively produced interferon‐γ (IFN‐γ), whereas studies using NK cell‐depleted mice have revealed a pivotal role for these cells in the host response against Cryptococcus neoformans, Aspergillus fumigatus, Candida albicans and Histoplasma capsulatum.10, 11 Sporotrichosis is an often neglected emerging opportunistic subcutaneous mycosis endemic in many countries but most commonly in those from tropical and subtropical climate regions,12 which is caused by various thermodimorphic fungi of the Sporothrix genus, including Sporothrix brasiliensis, Sporothrix globosa, Sporothrix luriei and Sporothrix schenckii sensu stricto.13 The disease follows the traumatic inoculation of the fungus through soil‐inoculating injuries, inhalation of conidia, or through zoonotic transmission, especially from cats, affecting immunocompromised individuals in an opportunistic fashion.14

Despite mounting evidence, NK cells are not regarded in the general literature as having a major role in the host immune response against fungal pathogens in general and, specifically for Sporothrix, no previous assessment of their role exists, either in vitro or in vivo. More importantly, we could not find any assessment of how NK cell maturation is modulated in the setting of fungal infections in vivo. We therefore aimed to determine the role played by NK cells in the host immune response against S. schenckii by assessing key NK cell maturation/activation markers, as well as the ability of the host to clear the infection following NK cell depletion with anti‐asialo GM1.

Materials and methods

Animals

Male 5‐ to 6‐week‐old BALB/c mice were obtained from the Multidisciplinary Centre for Biological Research (CEMIB), University of Campinas, São Paulo, Brazil. The animals were housed in individually ventilated cages in an ambient, controlled temperature and 12 : 12 hr light/dark cycles. Clean water and food were offered ad libitum. All animal procedures were performed according to the guidelines of the Brazilian College of Animal Experimentation (COBEA) and were approved by the research ethics committee of Araraquara's School of Pharmaceutical Sciences from UNESP University (approval no. 17/2015).

Microorganism and culture conditions

Sporothrix schenckii sensu stricto ATCC 16345, originally obtained from a human case of diffuse lung infection (Baltimore, MD) and kindly provided by the Oswaldo Cruz Foundation (Rio de Janeiro, Brazil), was used for all experiments. For infection of mice, a piece of the fungal mycelium grown on Mycosel agar tubes was transferred to an Erlenmeyer flask containing 100 ml of brain–heart infusion broth (Difco Laboratories, Detroit, MI.) and then cultured for 6 days at 37° with constant shaking at 150 r.p.m. Then, an aliquot containing 10⁷ yeast cells was transferred to a fresh medium and cultured for a further 5 days under the same conditions to achieve maximum mycelium‐to‐yeast conversion in a logarithmically growing culture.

Animal infection and NK cell depletion

Animals were inoculated intraperitoneally with 10⁶ S. schenckii yeast cells in sterile phosphate‐buffered saline (PBS), pH 7·4 (hereafter, PBS) or an equal volume of PBS alone and then killed at 5, 10 or 15 days post‐inoculation (dpi), or at 10 dpi only for selected experiments. Alternatively, S. schenckii‐infected mice were depleted of NK cells through the intraperitoneal inoculation of 40 μl of anti‐asialo GM1 (Wako Chemicals, Neuss, Germany) at −1, 3 and 7 dpi and then killed at 10 dpi; control groups consisted of S. schenckii‐inoculated or PBS‐inoculated mice that received an equal volume of PBS at the same schedule. Assessment of the fungal burden was performed by plating an aliquot of the spleen macerate on Mycosel agar and counting the recovered colony‐forming units after 3 days of incubation at ambient temperature.

Total splenocytes

Spleens were aseptically removed and passed through a 100‐μm cell strainer into a Petri dish containing 2 ml of RPMI‐1640 medium supplemented with 20 μm of 2β‐mercaptoethanol, 100 U/ml of penicillin and streptomycin, 2 mm of l‐glutamine and 5% fetal calf serum (hereafter, RPMI) with the aid of a syringe plunger. For red cell lysis, 6 ml of a 0·17 m ammonium chloride solution was added to the resulting suspension and then incubated on ice for 5 min. Splenocytes were then separated from the supernatant by centrifugation at 300 g for 5 min at 4°, washed once with 3 ml of RPMI and then resuspended in 1 ml of the same medium. Cell concentration was determined by microscopy using the Trypan blue exclusion test and then adjusted as required for each experiment.

Cytokines

Cytokines were measured using BD™ cytometric bead array (BD Biosciences, San Jose, CA) according to the manufacturer's instructions in the serum – obtained from blood collected by cardiac puncture – and spleen supernatant collected after maceration but before the red cell lysis described above.

Flow cytometry

The following monoclonal antibodies were used: anti‐CD16/CD32 purified (clone 93), anti‐CD3 fluorescein isothiocyanate (clone 17A2), anti‐CD4 allophycocyanin (APC) (clone RM4‐5), anti‐CD49b APC (clone DX5), anti‐CD8 peridinin chlorophyll protein‐Cychrome 5.5 (PerCP‐Cy5.5) (clone 53‐6.7), anti‐CD27 phycoerythrin (PE) (clone LG.7F9), anti‐CD127 PE (clone A7R34) and anti‐CD25 PE (clone PC61‐5) from eBiosciences (San Diego, CA); anti‐CD8 APC (clone 53‐6.7), anti‐CD11b PerCP‐Cy5.5 (clone M1/70), anti‐CD62L PerCP‐Cy5.5 (clone MEL‐14), anti‐NKp46 PerCP‐Cy5.5 (clone 29A1.4), anti‐CD69 PE (clone H1.2F3) and anti‐CD19 PE (clone 1D3) from BD Biosciences; and anti‐KLRG1 PerCP‐Cy5.5 (clone 2F1/KLRG1) and anti‐CD122 PE (clone TM‐β1) from BioLegend (San Diego, CA). NK cells were gated either as CD3⁻ CD49b⁺ SSC^low or, for the depletion experiments, as CD3⁻ CD122⁺ CD49b⁺ cells. For all staining procedures, only freshly isolated, unstimulated splenocytes were used. The events were acquired using a BD Accuri C6 flow cytometer (BD Biosciences) and analysed with the flow cytometer's proprietary software. At least 100 000 events were effectively included in each analysis. All histograms were scaled for comparison purposes.

Statistical analysis

Statistical analysis was performed in graphpad prism ver. 6.01 by applying Student's t‐test or one‐ or two‐way analysis of variance followed by Tukey's or Sidak's multiple comparisons test, respectively, as indicated. Differences were considered statistically significant when P ≤ 0·05. The data are expressed as the means ± SD. Each experiment was performed with four to ten (but mostly five) mice; the exact number used in each experiment can be found in the respective Figure legend.

Results

Natural killer cells expand in the spleen and become more mature following infection by S. schenckii

Following systemic infection of BALB/c mice with S. schenckii, we assessed the kinetics of NK cell expansion and change in maturation status in the spleen, a major compromised organ following systemic infection by this pathogen. As shown in Fig. 1(a,b), although the frequency of NK cells did not increase following infection – and even slightly diminished at 10 dpi, their absolute numbers were significantly higher in infected mice as early as 5 dpi; thereafter, their numbers continuously increased through 15 dpi, indicating that NK cells expand in the spleen at a rate equivalent to the increase in cellularity of this organ (data not shown), closely resembling NK cell expansion from birth to adulthood under steady state in mice.15 Likewise, NK cells were found to be already more mature at 5 dpi, although, similar to the steady state, the less mature CD11b⁻ CD27⁺ NK cells comprised the predominant subset (Fig. 1c). With the advance of the infection, the fully mature CD11b⁺ CD27⁻ phenotype became the predominant NK cell subset at 10 and 15 dpi (Fig. 1d,e). Our results therefore indicate that systemic infection by S. schenckii drives NK cell maturation and expansion in the spleen.

Figure 1.

Natural killer (NK) cells expand in the spleen and become more mature following infection by Sporothrix schenckii. BALB/c mice were inoculated intraperitoneally with 10⁶ S. schenckii yeast cells or phosphate‐buffered saline (PBS) and then killed at the indicated time‐points for assessment of NK cell frequency and maturation status by flow cytometry. (a, b) Frequency and absolute number of NK (CD3⁻ CD49b⁺ SSC^low) cells in the spleen, respectively. (c–e) Frequency of splenic NK cell subsets in each maturation stage as defined by the expression of CD11b and CD27. (f, g) Representative plots from 10 days post‐inoculation. Statistical significance was determined by two‐way analysis of variance using Sidak's multiple comparisons test and a 95% confidence interval. *P < 0·05, **P < 0·01, ***P < 0·001 and ****P < 0·0001 for comparisons with the control group in each time‐point. The results are presented as the mean ± SD of five mice.

CD62L and KLRG1 are substantially up‐regulated on NK cells from S. schenckii‐infected mice

Next, to address whether NK cells would become activated following systemic infection by S. schenckii, we assessed a number of activation markers whose expression is expected to increase as a result of NK cell activation. NK cells displayed a substantially up‐regulated surface expression of CD62L (l‐selectin) and KLRG1, but not of CD25 or CD69, as judged by their median fluorescence intensity at 10 dpi (Fig. 2b). On the other hand, the frequency of CD62L‐, CD25‐, KLRG1‐, and CD69‐expressing NK cells was similar between control and infected mice (Fig. 2a). We also characterized NK cells as conventional (cNK) or unconventional (thymic) based on the expression of CD127 to determine the source of the NK cells accumulating in the spleen following the S. schenckii infection. As seen in Fig. 2(d), the frequency of thymus‐originated CD127⁺ NK cells was reduced almost fourfold in infected mice, suggesting that the accumulation of NK cells in the spleen occurred mostly through in situ proliferation or the infiltration of BM‐originated cells. Furthermore, the lack of CD25 coupled with CD69 at only steady‐state levels and up‐regulated KLRG1 expression suggest a late, more mature phenotype for the NK cells found in infected mice at 10 dpi.

Figure 2.

CD62 ligand (CD62L) and killer cell lectin‐like receptor subfamily G member 1 (KLRG1) are substantially up‐regulated in natural killer (NK) cells from Sporothrix schenckii‐infected mice. BALB/c mice were inoculated intraperitoneally with 10⁶ S. schenckii yeast cells or phosphate‐buffered saline (PBS) and then killed at 10 days post‐inoculation for assessment of NK cell activation by flow cytometry. (a, d) Frequency of each indicated marker (stained as follows: CD127, CD62L/CD25 and CD69/KLRG1) on splenic NK (CD3⁻ CD49b⁺ SSC^low) cells. (b) Median fluorescence intensity (MFI) of each indicated marker on splenic NK cells. (c) Representative plots (grey‐filled and open histograms indicate control and infected mice, respectively). Statistical significance was determined by two‐way analysis of variance using Sidak's multiple comparisons test (a, b) or Student's t‐test (d) and a 95% confidence interval. ****P < 0·0001 for comparisons with the control group or as indicated. The results are presented as the mean ± SD of four or five mice.

NK cells play a pivotal role in the in vivo elimination of S. schenckii

The importance of NK cells for clearing S. schenckii in vivo was determined by depleting mice of these cells through the intraperitoneal inoculation of anti‐asialo GM1 at −1, 3 and 7 dpi and then killing them at 10 dpi to determine the splenic fungal load. Depletion efficiency reached 78·9% ± 3·1% at 10 dpi, enough to reduce NK cell frequency to < 0·5% of the total splenic cells and their absolute numbers to about half those found at steady state (Fig. 3a,b). Furthermore, more than twice as many of the few remaining NK cells in the spleens of NK cell‐depleted infected mice were immature CD3⁻ CD122⁺ CD49b⁺ NKp46⁻ cells compared with non‐depleted infected mice, although NKp46 expression on the NKp46⁺ NK cells was not significantly affected (Fig. 3d,e). As a result, the fungal load in the spleens of NK cell‐depleted infected mice increased more than eightfold (Fig. 3c), which was accompanied by an augmentation of the size and cellularity of this organ (the author's observations and data not shown, respectively). Hence, NK cells play a pivotal role in the in vivo elimination of S. schenckii.

Figure 3.

Natural killer (NK) cells play a pivotal role in the in vivo elimination of Sporothrix schenckii. BALB/c mice were inoculated intraperitoneally with 10⁶ S. schenckii yeast cells whilst depleted of NK cells through the intraperitoneal inoculation of 40 μl of anti‐asialo GM1 (ASGM1) at −1, 3 and 7 days post‐inoculation (dpi) and then killed at 10 dpi; control groups consisted of S. schenckii‐ or phosphate‐buffered saline (PBS) ‐inoculated mice that received an equal volume of PBS at the same schedule. NK cell depletion efficiency was determined by flow cytometry. (a, b) Frequency and absolute number of NK (CD3⁻ CD122⁺ CD49b⁺) cells in the spleen, respectively. (c) Fungal burden in the spleen of NK cell‐depleted (ASGM1) or non‐depleted (PBS) S. schenckii‐infected mice, as determined by the number of colony‐forming units grown on Mycosel agar. (d, e) Frequency and median fluorescence intensity (MFI) of NKp46 on splenic NK cells. (f, g) Representative plots. Statistical significance was determined by one‐way analysis of variance using Tukey's multiple comparisons test (a, b, d and e) or Student's t‐test (c) and a 95% confidence interval. *P < 0·05, **P < 0·01 and ****P < 0·0001 for comparisons with the control group or as indicated. The results are presented as the mean ± SD of five (a, b, d, and e) or 9–10 (c) mice from two separate experiments.

NK cell depletion results in augmented systemic production of inflammatory cytokines

Using the same experimental approach as above, we sought to address why NK cell depletion was so detrimental to the ability of the host to control the infection by assessing the in vivo production of a panel of cytokines at 10 dpi. As shown in Fig. 4, the production of the pro‐inflammatory cytokines TNF‐α, IFN‐γ and IL‐6 was greatly increased in both the spleen and serum of NK cell‐depleted infected mice compared with the non‐depleted mice. Additionally, IL‐2, IL‐4, IL‐10 and IL‐17A were all below detection levels across the three experimental groups (data not shown), suggesting that inflammation was contained through an IL‐10‐independent mechanism in non‐depleted infected mice, whereas the exacerbated inflammation seen in NK cell‐depleted infected mice was not the result of the development of a pathogenic T helper type 17 response. As cytokine production in the spleen was measured in supernatants obtained by maceration, an approach that requires the organ to be processed in a certain amount of liquid, the difference between the spleen and serum concentrations of each cytokine should not be interpreted at face value. As a whole, our results show that the defective fungal clearance following NK cell depletion was accompanied by heightened systemic inflammation, although it remains to be determined whether such augmented inflammation is cause or consequence, or maybe a combination of both, of the impaired host immune response.

Figure 4.

Natural killer (NK) cell depletion results in augmented systemic production of inflammatory cytokines. BALB/c mice were inoculated intraperitoneally with 10⁶ Sporothrix schenckii yeast cells whilst depleted of NK cells through the intraperitoneal inoculation of 40 μl of anti‐asialo GM1 (ASGM1) at −1, 3 and 7 days post‐inoculation (dpi) and then killed at 10 dpi; control groups consisted of S. schenckii‐ or phosphate‐buffered saline (PBS)‐inoculated mice that received an equal volume of PBS at the same schedule. Cytokine concentration was determined using BD™ cytometric bead array. (a–f) Concentration of tumour necrosis factor‐α (TNF‐α), interferon‐γ (IFN‐γ) and interleukin‐6 (IL‐6) in the spleen or serum, as indicated. Statistical significance was determined by one‐way analysis of variance using Tukey's multiple comparisons test and a 95% confidence interval. *P < 0·05, **P < 0·01, ***P < 0·001 and ****P < 0·0001 for comparisons with the control group or as indicated. The results are presented as the mean ± SD of five mice.

Discussion

We show that NK cells display a more mature phenotype as early as 5 dpi and, as they progress through the maturation pathway through the course of the infection, the fully mature CD11b⁺ CD27⁻ phenotype becomes the predominant subset at 10 and 15 dpi. CD11b⁺ CD27⁻ NK cells have been reported to show increased expression of CD62L, S1P₅ (sphingosine‐1‐phosphate receptor 5), CX₃CR1, CXCR3, CXCR4 and CCR1.16, 17 Accordingly, we found CD62L to be substantially up‐regulated on NK cells from S. schenckii‐infected mice, consistent with the predominance of the CD11b⁺ CD27⁻ subset. Additionally, CX₃CR1 is highly expressed and mainly found on KLRG1⁺ NK cells (considered fully mature CD11b⁺ CD27⁻ cells) identifying an even later maturation stage shown to have impaired IFN‐γ production and cytotoxicity towards YAC‐1 cells under cytokine stimulation.18 Therefore, it is possible that a CX₃CR1⁺ KLRG1⁺ fully mature NK cell subset accounts for a considerable fraction of splenic NK cells following systemic infection by S. schenckii.

Natural killer cell activation by cytokines, microbial products or products expressed on abnormal or infected cells can lead to loss of CD27 or to increased expression of CD11b, Ly6C, CD43, KLRG1, CD25, CD69 and stem cell antigen 1.6, 7, 8, 9 Although both CD25 and CD69 were found only at steady‐state levels in S. schenckii‐infected mice, KLRG1 had its expression up‐regulated more than twofold. KLRG1 is a cadherin‐binding inhibitory receptor of the C‐type lectin‐like family whose expression is associated with the fully mature CD11b⁺ CD27⁻ subset, identifying naive NK cells with reduced effector functions, decreased IL‐15 responsiveness, and proliferative capacity.19 Notably, CD69 has been reported to be only transiently up‐regulated following NK cell activation,20, 21 whereas KLRG1 stays up‐regulated for a prolonged period,22 suggesting the shift from an early, fully activated to a late, more mature phenotype by NK cells in our study.

Natural cytotoxicity receptors (NCRs) are structurally unrelated type I transmembrane proteins belonging to the immunoglobulin superfamily that were grouped together based on their capability to activate NK cells and promote the killing of tumour cells, although most of the known NCR ligands are pathogen‐derived. Of the three known NCRs (NKp46, NKp44 and NKp30), NKp46 (CD335) is the only one conserved between humans and mice, suggesting that NKp46 is the primary NCR involved in pathogen and tumour recognition.23 NKp46 is constitutively expressed on NK cells regardless of their state of activation and is generally believed to be expressed before CD49b (α ₂ integrin) during NK cell development in the BM.24, 25 However, Gotthardt et al. 4 showed that, as opposed to the liver, NKp46 expression is acquired after that of CD49b in the BM and hence the spleen, lymph nodes and blood harbour a considerable fraction of CD49b⁺ NKp46⁻ NK cells in the steady state; this is consistent with our own findings of an average 13·2% ± 2·8% of this subset in control non‐infected mice.

Aside from cytokines, activating stimuli may be delivered to NK cells via Toll‐like receptors (e.g. TLR2 and TLR4), nucleotide‐binding oligomerization domain‐like receptors (e.g. NLRP3), retinoic acid inducible gene 1‐like receptors, or activating receptors such as NCRs, NKG2D, NKG2C/CD94, 2B4 (CD244), NKp80, DNAM1, NTB‐A, CD16 and some Ly49s.6, 26, 27 Several kinds of bacteria/viruses are known to directly activate NK cells through pattern recognition receptors (PRRs), especially TLR2,28 which has been shown to be directly involved in the recognition of Mycobacterium tuberculosis by NK cells in humans and may also be involved in the NK‐mediated response to human cytomegalovirus.29 Furthermore, Li et al.30, 31 showed that NKp30 was responsible for the recognition and killing of the fungal pathogens Cryptococcus and Candida by human NK cells upon binding to β‐1,3‐glucans on the surface of the fungi. The β‐1,3‐glucans are a major component of both S. schenckii and S. brasiliensis cell walls32, 33 and we and others have previously implicated TLR2 and TLR433, 34, 35, 36, 37 and NLRP338 in Sporothrix recognition in vitro and in vivo. All considered, there is a strong case for PRR‐dependent NK cell activation pathways having a role in immune response development following infection by S. schenckii.

Natural killer cells can be divided into cNK or thymic based on the expression of CD127. The cNK cells are characterized as CD127⁻ CD69⁻ Ly49⁺ CD11b^high and are the predominant population in the spleen and blood in the steady state, although significant numbers are also found in the liver, kidneys and mucosal tissues. CD127⁺ CD69^high Ly49^low CD11b^low NK cells, in turn, represent a thymus‐originated subset that is less cytotoxic than splenic cNK cells but able to produce more IFN‐γ, TNF‐α and granulocyte–macrophage colony‐stimulating factor (GM‐CSF).25 After the discovery that, as for thymic NK cells, all innate lymphoid cell subsets express CD127 and are dependent on GATA‐3, thymic NK cells may now be classified as type 1 innate lymphoid cells.39 Regardless, our results show that the frequency of CD127⁺ NK cells was reduced almost fourfold in infected mice, suggesting that most NK cells found in the spleen following the S. schenckii infection either proliferate in situ or infiltrate this organ from the BM; this notion is further reinforced by our observation of few CD69‐expressing NK cells in both infected and non‐infected mice.

An antibody‐mediated NK cell depletion animal model using anti‐asialo GM1, whose capability to induce NK cell depletion in vivo was reported by Kasai et al.,40 was used to address whether NK cells were required for clearing the S. schenckii infection. Pilot experiments showed that 24 hr after the intraperitoneal inoculation of 40, 20 and 10 μl of anti‐asialo GM1, splenic NK cell depletion reached, 100% ± 0%, 98·9% ± 1·0% and 80·4% ± 12·1%, respectively (see Supplementary material, Fig. S1a). After 5 days, the inoculation of 40 and 20 μl resulted in the depletion of, respectively, 83·1% ± 7·1% and 69·4% ± 16·4% of splenic NK cells (see Supplementary material, Fig. S1c). Considering that the inoculation of 40 μl resulted in a complete, but not significantly higher, depletion than that obtained with 20 μl, it is improbable that higher doses would significantly enhance this effect. Ultimately, depletion efficiency reached 78·9% ± 3·1% at 10 dpi after three 40 μl inoculations performed at −1, 3 and 7 dpi. Although this was already satisfactory, the impairment of NK cell activity must have been even greater, because more than twice as many of the few remaining NK cells in the spleen of NK cell‐depleted infected mice were immature CD49b⁺ NKp46⁻ cells compared with non‐depleted infected mice; this suggests the rapid infiltration of immature NK cells from the BM, where NKp46 acquisition was shown to occur after that of CD49b.4 Furthermore, as shown by Chiossone et al.,5 during the first days of reconstitution after NK cell depletion, these cells are in a more immature state, so are unable to develop their full effector functions.

Natural killer cells are one of the first responders following infection and play a role in the activation and/or maturation of dendritic cells, macrophages and T cells by producing IFN‐γ and TNF‐α and/or killing immature dendritic cells;41, 42 they are able to kill regulatory T cells, indirectly promoting CD4⁺ and CD8⁺ effector T‐cell responses,41 and to express co‐stimulatory molecules which allow them to directly promote T‐cell proliferation.43 NK cells also have a prominent pro‐inflammatory role during infections owing to their ability to produce IFN‐γ – frequently acting as the main source of this cytokine during infection by several different bacterial pathogens, TNF‐α, GM‐CSF and chemokines such as nacrophage inflammatory protein‐1α.6, 44 Furthermore, Bär et al.45 showed that NK cells are critical for in vivo immunity to systemic Candida albicans infection and that they act by providing GM‐CSF, which promoted the fungicidal activity of neutrophils. Hence, we hypothesize that the severely impaired S. schenckii clearance following NK cell depletion resulted from any combination of the above‐mentioned mechanisms, so leading to exacerbated systemic inflammation as a consequence of the higher fungal burden.

An alternative hypothesis is that the main contribution of NK cells to protection against S. schenckii came instead from their immunoregulatory role, as NK cells are protective in many autoimmune diseases, such as in experimental autoimmune encephalomyelitis where NK cell depletion before disease induction leads to increased severity and mortality.44 Moreover, most direct interactions between NK and T cells result in impaired T‐cell responses owing to NK cell‐secreted IL‐10, competition for IL‐2 after CD25 up‐regulation by NK cells, or the recognition and killing of activated T cells by NK cells.41, 44 However, our results suggest that at least such direct NK–T‐cell inhibiting interactions might not be playing a role, for the following reasons: IL‐10 production was below detection levels across all groups, CD25 expression was not detectable on NK cells from both infected and non‐infected mice, and the frequency of both total CD3⁺ cells and CD8⁺ CD3⁺ cells did not increase in infected mice following NK cell depletion (see Supplementary material, Fig. S2).

Natural killer cells have been implicated in shaping immunity post‐vaccination owing to their ability to regulate adaptive immune responses, such as supporting the development of memory T cells and humoral immunity following immunization in the presence of certain stimuli.46 Furthermore, following infection or vaccination, NK cells themselves can develop into long‐lived memory cells capable of more robust cytotoxic responses and higher IFN‐γ production upon rechallenge – in some cases in an antigen‐specific manner – compared with naive NK cells.47 Also, pro‐inflammatory cytokines (i.e. IL‐12, IL‐15 and IL‐18) alone were found to be capable of supporting splenic NK cell memory properties in the absence of antigen; these cytokine‐induced memory‐like NK cells can be induced by vaccination in response to CD4⁺ T‐cell‐derived IL‐2 and myeloid cell‐derived IL‐12 and type I interferons, and have been implicated in the enhancement of NK cell function ex vivo.47, 48 Most interestingly, NKp46 and NKp30, two key NK cell cytotoxicity receptors, were reported to act as fungicidal PRRs for Cryptococcus and Candida,30, 31, 49 posing the exciting possibility that pathogen‐specific NK cell memory, the existence of which has already been demonstrated for viruses,47 could develop against fungi. Given our own work developing an anti‐Sporothrix vaccine,50, 51 the above underlines the huge clinical value of a better understanding of the NK–Sporothrix relationship.

Lastly, in most non‐immunocompromised individuals, sporotrichosis manifests as a localized cutaneous/subcutaneous disease lacking haematogenous dissemination to visceral organs.13 Also, skin‐tissue‐resident NK cells express lower levels of CD49b than splenic cNK cells, have constitutive CD69 expression, and are also mature CD11b⁺ CD27⁺ cells.25 Therefore, NK cells would be expected to respond differently to such localized infections from how they did in our study, even without considering the many differences in the overall immune response that could indirectly affect NK cell activity. Our decision to use a model of systemic disease was aimed at gaining insight into the immune mechanisms engaged when this disease does become systemic, without having to resort to immunosuppression, which has also the added benefit of showing how NK cells respond in commonly used systemic experimental models. Although apparently not serving as an IFN‐γ source, as we had previously suggested,52 and in a way mostly contrary to our initial expectations, NK cells proved to be an essential component of the immune machinery driving the elimination of S. schenckii.

Author contributions

LSF designed and performed experiments, analysed data and wrote the manuscript. DLP and MCP performed experiments. IZC supervised the project.

Disclosures

The authors declare no commercial or financial conflict of interest.

Supporting information

Figure S1. Pilot depletion experiments.

Figure S2. Frequency and absolute number of selected cell populations in natural killer cell‐depleted mice.

Figure S3. Gating strategy for the flow cytometric analysis of splenic natural killer cells.