Candida albicans triggers NADPH oxidase-independent neutrophil extracellular traps through dectin-2

Summary

Candida albicans is one of the top leading causes of healthcare-associated bloodstream infection. Neutrophil extracellular traps (NET) are known to capture and kill pathogens. It is reported that opsonized C. albicans-triggered NETosis is NADPH oxidase-dependent. We discovered a NADPH oxidase-independent NETosis pathway in neutrophil response to unopsonized C. albicans. While CR3 engagement with opsonized C. albicans triggered NET, dectin-2 recognized unopsonized C. albicans and mediated NET formation. Engagement of dectin-2 activated the downstream Syk-Ca²⁺-PKCδ-protein arginine deiminase 4 (PAD4) signaling pathway which modulated nuclear translocation of neutrophil elastase (NE), histone citrullination and NETosis. In a C. albicans peritonitis model we observed Ki67⁺Ly6G⁺ NETotic cells in the peritoneal exudate and mesenteric tissues within 3 h of infection. Treatment with PAD4 inhibitor GSK484 or dectin-2 deficiency reduced % Ki67⁺Ly6G⁺ cells and the intensity of Ki67 in peritoneal neutrophils. Employing DNA digestion enzyme micrococcal nuclease, GSK484 as well as dectin-2-deficient mice, we further showed that dectin-2-mediated PAD4-dependent NET formation in vivo restrained the spread of C. albicans from the peritoneal cavity to kidney. Taken together, this study reveals that unopsonized C. albicans evokes NADPH oxidase-independent NETosis through dectin-2 and its downstream signaling pathway and dectin-2-mediated NET helps restrain fungal dissemination.

Keywords:

Histones – Cell staining – kidneys – Fluorescence microscopy – Fluorescence imaging – Neutrophils – Candida albicans – Intraperitoneal injections

Introduction

Candida albicans is a commensal in the mucosa surface and skin in most humans. Environmental changes in temperature, nutrition, or the presence of serum induce its transformation from yeast to hyphae. C. albicans infection is one of the top leading causes of overall healthcare-associated bloodstream infection in medical centers as well as regional hospitals. Invasive candidiasis affects more than 250,000 people worldwide each year and leads to more than 50,000 deaths. Mortality among patients with invasive candidiasis is as high as 40% even after receiving antifungal therapy [1–3]. Patients with neutropenia or genetic deficiency in NADPH oxidase are susceptible to invasive candidiasis [4, 5], showing that neutrophils and NADPH oxidase activation are indispensable for host defense against C. albicans infection.

NADPH oxidase activation requires the assembly of its regulatory subunits, p40^phox, p47^phox, and p67^phox with its core proteins gp91^phox and p22^phox, resulting in ROS production [6, 7]. In addition to generating ROS, NADPH oxidase activation also induces neutrophil release of nuclear DNA to form a sticky web-like structure named neutrophil extracellular trap (NET) that binds histones, granular proteins and antimicrobial peptides [8]. In vitro studies showed that pathogens trapped by NET are in contact with and killed by concentrated antimicrobial factors [9, 10]. NET is known to capture and kill C. albicans through a NADPH oxidase-dependent mechanism [11]. However, it is also reported that human neutrophils are capable of killing unopsonized C. albicans through a ROS-independent mechanism [12]. Whether C. albicans can induce NET through a NADPH oxidase-independent mechanism is a question to be addressed.

The process of NET formation is called NETosis. Neutrophils undergoing NETosis are characterized by disintegrated nuclear envelop and release of decondensed chromatin into the cytoplasm [8]. Recent study uncovers that cell cycle pathway controls NETosis. NETotic neutrophils have phosphorylated retinoblastoma protein and lamins and express cell cycle marker Ki67 [13]. Chromatin decondensation is the result of protein arginine deiminase 4 (PAD4)-dependent histone citrullination and neutrophil elastase (NE)-mediated histone degradation [14, 15]. NADPH oxidase facilitates both nuclear translocation of NE and PAD4 activation through stimulating myeloperoxidase activation [16]. It has been shown that opsonized C. albicans induces NET through autophagy, ROS, and NE, but not PAD4, apoptosis nor necroptosis [17]. Unopsonized C. albicans is also known to induce NET formation [11]. Since neutrophils use different receptors to recognize serum-opsonized and unopsonized C. albicans [12], it is important to investigate the receptor and the molecular mechanism by which unopsonized C. albicans uses to evoke NET.

Multiple receptors participate in modulating neutrophil anti-C. albicans functions. Fcγ receptor mediates human neutrophil killing of antibody-opsonized fungus through the Syk and PKC signaling pathways, whereas complement receptor 3 (CR3) are involved in killing of unopsonized C. albicans [12]. Mouse neutrophils utilize CR3 for recognition and killing of opsonized C. albicans [18]. Dectin-2 is marginally involved in opsonized C. albicans-induced neutrophil ROS production [19]. Dectin-1 as a phagocytosis receptor for C. albicans yeasts negatively regulates NETosis through interfering with nuclear translocation of granule NE [20]. Although CR3 as a receptor recognizing fibronectin-coated matrix is responsible for C. albicans-induced NETosis [21], which receptor(s) mediates NET formation in response to unopsonized C. albicans alone awaits to be determined.

Here we sought to study the receptor and signaling pathway that mediate unopsonized C. albicans-induced NET formation. Our study revealed the role of dectin-2 and its downstream Syk-Ca²⁺-PKCδ-PAD4/NE pathway in inducing NETosis in a NADPH oxidase-independent manner. Dectin-2 functions to restrain C. albicans spread from peritoneal cavity to kidney through modulating NET.

Results

Both opsonized and unopsonized C. albicans induce NET formation

It has been reported that opsonized C. albicans induces NETosis [17, 22, 23]. Our results revealed that neutrophils released web-like extracellular DNA fibers in response to opsonized as well as unopsonized C. albicans (Fig 1A). NETosis is characterized by disintegration of the nuclear envelope [8]. While stimulation by opsonized C. albicans resulted in nuclear membrane disintegration (Fig 1B), we also observed nuclear envelope breakdown and cytoplasmic membrane rupture following stimulation by unopsonized C. albicans (Fig 1B). Fluorescence images at high magnification clearly demonstrated that similar to opsonized C. albicans stimulation, unopsonized C. albicans hyphae were entangled with histone H3-containing web-like extracellular DNA structure (Fig 1C), although the percentage of NETotic cells in response to unopsonized C. albicans (6.8%) was lower than that to opsonized organisms (13.9%) (Fig 1D). Since opsonin-containing fresh serum also facilitates the germination of C. albicans, we allowed C. albicans yeasts to germinate before adding them to wells containing neutrophils. PicoGreen dsDNA assay showed that the hyphal form but not yeast-locked C. albicans (strain HLC 54) induced NETosis whether it was opsonized or not (Fig 1E), signifying the importance of hyphal formation in triggering NETosis. Our data indicate that not only opsonized but also unopsonized C. albicans in its hyphal form induces NETosis.

Unopsonized C. albicans-induced NET formation is independent of NCF-1

We further explored the requirement of ROS in unopsonized C. albicans-induced NET formation. Results showed that NCF-1 (NADPH oxidase subunit p47^phox)-deficient neutrophils formed histone H3-containing web-like NET structure as readily as NCF-1-sufficient cells upon stimulation by unopsonized C. albicans (Fig 2A and 2B). Time-lapse live cell imaging showed that neutrophils underwent robust NETosis after encountering opsonized C. albicans (Fig 2C and S1 Video) whereas NET formation induced by unopsonized organism was less so (Fig 2C and S2 Video). Importantly, similar to stimulation of NCF-1-sufficient neutrophils with opsonized or unopsonized C. albicans (Fig 2C, S1 and S2 Videos), stimulation of NCF-1-deficient neutrophils by unopsonized C. albicans resulted in loss of lobular shape in the nucleus and chromatin decondensation (Fig 2C and S3 Video). These data demonstrate that NCF-1-deficient neutrophils underwent NETosis after C. albicans challenge. Consistently, Ncf-1^-/- and Ncf-1^+/+ neutrophils releases comparable levels of dsDNA in response to stimulation by unopsonized C. albicans, whereas the response was greatly reduced in Ncf-1^-/- neutrophils upon challenge with opsonized organisms (Fig 2D). Additionally, treatment with MitoTEMPO (mitochondrial ROS inhibitor) did not affect NET formation in neutrophils stimulated by unopsonized C. albicans (Fig 2E). Our results indicate that unopsonized C. albicans-induced NET formation is independent of NADPH oxidase and mitochondrial ROS.

Neutrophil killing of unopsonized C. albicans requires dectin-2-mediated NET formation

We used receptor-deficient neutrophils to identify the receptors that mediate opsonized and unopsonized C. albicans-induced NETosis. Results showed that while CR3 deficiency (Itgam^-/-) reduced dsDNA release triggered by opsonized C. albicans (Fig 3A), dectin-2 deficiency (Clec4n^-/-) reduced that induced by unopsonized organism (Fig 3B). Neither dectin-1 nor MyD88 was involved in NETosis induced by either opsonized or unopsonized organism (Fig 3C and 3D). Confocal microscopic images showed that there was direct contact between dectin-2 and unopsonized C. albicans, whether it is in yeast or hyphal form (Fig 3E). Dectin-2 deficiency abrogated the formation of histone H3-containing NET structure after C. albicans challenge (Fig 3F). Thus, dectin-2 recognition of unopsonized C. albicans by neutrophils results in NETosis.

To study whether NET can kill C. albicans, we added DNA digestion enzyme, micrococcal nuclease (MNase) to the wells at the time when C. albicans was added. While the abilities of WT and CR3-deficient neutrophils to kill unopsonized C. albicans were comparable (WT: 25.8 ± 8.6%; CR3-deficient: 20.3 ± 5.5%), their killing functions were significantly diminished after MNase treatment (WT: 11.2 ± 4.8%; CR3-deficient: 7.7 ± 10.6%) (Fig 3G). It appears that neutrophil killing of unopsonized C. albicans is mediated by NET and independent of CR3 expression. Compared to WT and CR3-deficient cells, dectin-2-deficient neutrophils had lower ability to kill unopsonized C. albicans (6.5 ± 8.9%) (Fig 3G), yet such function was not affected by MNase treatment (7.0 ± 9.6%). These results together reveal that neutrophil killing of unopsonized C. albicans requires dectin-2-mediated NET formation.

Dectin-2 mediates NET formation through Syk-Ca²⁺-PKCδ signaling pathway in response to unopsonized C. albicans

We used pharmacological inhibitors to inhibit activation of signaling molecules and found that inhibition of Syk, Ca²⁺ influx, and PKCs significantly diminished NET formation (Fig 4A, 4B and 4C). While different isoforms of PKC family have their unique roles in modulating NETosis [24], our results showed that inhibition of PKCδ dose-dependently, but not PKCα and PKCβ, reduced the level of NETosis (Fig 4D). These results together indicate that Syk, Ca²⁺ influx, and PKCδ are involved in unopsonized C. albicans-induced NET. Furthermore, cells treated with Syk inhibitor had lower levels of Ca²⁺ influx, less Ca²⁺-positive cells (S1 Fig and Fig 4E) and lowered the level of phosphorylated PKCδ (Fig 4F). Cells treated with Ca²⁺ chelator had lower levels of phosphorylated PKCδ but not that of phosphorylated Syk (Fig 4G). In line with the observation that dectin-2 deficiency reduced the levels of phosphorylated Syk and PKCδ after stimulation (Fig 4H), our results clearly demonstrate that unopsonized C. albicans induces NETosis through dectin-2 and its downstream Syk-Ca²⁺-PKCδ pathway.

NE nuclear translocation is involved in NCF-1-independent NETosis through Syk-Ca²⁺-PKCδ

Neutrophils treated with neutrophil elastase inhibitor sivelestat had reduced NET formation upon stimulation by unopsonized C. albicans (Fig 5A). Immunofluorescence images showed that NE was distributed in the cytoplasm and was separated from the nuclear region before stimulation (Fig 5B). Responding to unopsonized C. albicans challenge, NE aggregated into larger puncta and started translocating to the nucleus, especially to the decondensed area by 1 h of stimulation (Fig 5B). At 2 h after stimulation, the granules containing NE began to disintegrate, and NE was localized in the nucleus of cells that was ready for NETosis (2 h, Fig 5B). By 3 h after stimulation when NETotic structure began to form, NE was released to the extracellular space and bound to extracellular DNA fibers (3 h, Fig 5B). While nuclear translocation of NE occurred after stimulation with unopsonized C. albicans (Ctrl. in Fig 5C), neutrophils treated with pharmacological inhibitor to Syk, Ca²⁺ influx, PKCδ or NE exhibited condensed chromatin structure and their NE remained in the perimeter of the nucleus (Fig 5C), suggesting that Syk, Ca²⁺ influx, PKCδ and NE activity regulate chromatin decondensation as well as NE nuclear translocation. Together these data indicate that dectin-2 downstream signaling pathway mediates NE translocation. Importantly, NCF-1 deficiency did not affect NE nuclear translocation upon stimulation with unopsonized C. albicans (2 h, S2 Fig) although the number of NE-aggregated puncta was reduced (1 h, S2 Fig). NE was released along with DNA fibers in Ncf-1^-/- cells by 3 h after stimulation (S2 Fig). It appears that NCF-1 is not involved in NE nuclear translocation but may participate in granule aggregation after stimulation by unopsonized C. albicans.

Unopsonized C. albicans-induced NET formation is dependent on PAD4 enzymatic activity

Treatment with inhibitor to PAD1-4 or PAD4 reduced unopsonized C. albicans-induced NET formation (Fig 6A). PAD4 inhibitor also impeded NE nuclear translocation and chromatin decondensation (Fig 6B) as well as the formation of histone H3-containing web-like NET structure (Fig 6C). Thus, PAD4 activity is required for unopsonized C. albicans-induced nucleus decondensation and NE nuclear translocation. Interestingly, inhibition of PKCδ reduced the level of citrullinated H3 (Fig 6D). Since histone H3 citrullination is catalyzed by PAD4 [14], these results show that recognition of unopsonized C. albicans by dectin-2 leads PAD4-dependent NET formation.

NET formation in peritoneal cavity after C. albicans infection

A C. albicans peritonitis model was established to study the role of NET formation in vivo. Since there are few neutrophils in the peritoneal cavity of normal mice, we gave mice two intraperitoneal injections of casein 18 h apart to enrich the neutrophil population. Mice were given C. albicans yeasts intraperitoneally 4 h after the second injection of casein when peritoneal neutrophils constituted about 78.8% of the whole peritoneal cell population (Fig 7A). In vivo imaging system (IVIS) spectrum images showed that C. albicans infection induced release of extracellular DNA into peritoneal cavity as early as 1.5 h after infection and remained at relatively the same level until 3 h later (Fig 7B). Web-like DNA structures that were positive for Ki67 (a novel marker for mature neutrophils that undergo NETosis [13]), histone H3, and Ly6G cells were observed in peritoneal exudates from mice given C. albicans (Fig 7C). Cells on the mesenteric tissues collected from infected mice also stained positive for Ki67 and Ly6G (Fig 7D). In the peritonitis candidiasis model, we observed NET formation in the peritoneal cavity and NETotic cells on the mesenteric tissues.

NCF-1-independent NETosis restrains C. albicans spread from peritoneal cavity to kidney

To monitor C. albicans spread, we infected mice with dTomato-expressing C. albicans when neutrophils were enriched in the peritoneal cavity. IVIS images showed that the intensity of fluorescence in the peritoneal cavity remained at relatively the same level at 1 and 2 h after infection and decreased thereafter (Fig 8A). Coinciding with decrease in intensity of dTomato, fungal burden in the peritoneal cavity also decreased by 3 h after infection and in the meantime, it was increased in the kidney (Fig 8B). We then treated mice with NET digestion enzyme micrococcal nuclease (MNase) or its heat-inactivated form (h.i. MNase) intraperitoneally and discovered that treatment with MNase compared to h.i. MNase decreased fungal burden in the peritoneal cavity [from (3.4 ± 2.3) × 10⁵ to (1.6 ± 0.7) × 10⁵ CFU] and increased that in the kidney [from (2.1 ± 1.0) × 10⁴ to (3.6 ± 0.7) × 10⁴ CFU/kidney) (Fig 8C). These results demonstrate that NET functions to restrain C. albicans in the peritoneal cavity and keep it from spread to the kidney. In the meanwhile, Ncf-1^-/- mice were infected by C. albicans intraperitoneally to assess whether NCF-1 participates in inducing NETosis in vivo. Data in S3A and S3B Fig showed that NCF-1 deficiency did not affect NET formation in peritoneal exudate nor did it affect the Ki67⁺Ly6G⁺ population in mesenteric tissues, indicating that an NCF-1-independent NETosis response to C. albicans occurs in vivo. Interestingly, however, Ncf-1^-/- mice had significantly higher fungal burden in the peritoneal cavity but a comparable level in the kidney compared to Ncf-1^+/+ mice at 3 h after infection (S3C Fig). Since ROS is important to phagocytic cell clearance of C. albicans [18], these results indicate that unopsonized C. albicans-induced NCF-1-independent NETosis that restricts fungal spread does occur in vivo, but fungal clearance involves more than just NETosis.

PAD4 is important to prevent fungal spread

Our data in Fig 6 showed that PAD4 is involved in NE nuclear translocation and inducing NET formation in vitro. To explore the role of PAD4 in C. albicans infection, we treated mice with PAD4 inhibitor GSK484. Mice were given GSK484 before intraperitoneal injection of C. albicans. Results showed that inhibition of PAD4 reduced the formation of web-like structure and Ki67 expression in peritoneal Ly6G⁺ cells and Ki67⁺Ly6G⁺ cell population in mesenteric tissues (Fig 9A and 9B). Flow cytometric analysis also revealed that GSK484 treatment reduced the percentage and the level of Ki67 in peritoneal infiltrating neutrophils (Fig 9C). In addition, GSK484 treatment decreased fungal CFU in the peritoneal cavity and increased that in the kidney by 3 h after infection (Fig 9D). Results of our in vitro (Fig 6) and in vivo studies together demonstrated that PAD4 regulates NETosis response to C. albicans.

C. albicans-induced NET formation is dectin-2-dependent

We then determined whether dectin-2 is involved in NETotic response to C. albicans in mice. Peritoneal neutrophil-enriched WT and dectin-2-deficient mice were intraperitoneally infected with C. albicans. While dectin-2 deficiency did not affect neutrophil recruitment to the peritoneal cavity (S4A Fig), peritoneal infiltrating neutrophils from infected dectin-2-deficient mice had less Ki67⁺Histone H3⁺ web-like structures than that from sufficient mice (Fig 10A). Compared to sufficient mice, dectin-2-deficient mice had significantly less Ki67⁺Ly6G⁺ population in mesenteric tissues (Fig 10B) and reduced Ki67 expression in neutrophils (Fig 10C) after C. albicans infection. Furthermore, dectin-2-deficient mice had significantly lower fungal burdens in the peritoneal cavity [(3 ± 1.5) × 10⁵ CFU] but greater burdens in kidneys [(5.9 ± 2.0) × 10⁴ CFU/kidney] than WT mice [(5.5 × ± 2.0) × 10⁵ CFU] in peritoneal cavity; (2.4 ± 1.5) × 10⁴ CFU/kidney] (Fig 10D). Digesting extracellular DNA by MNase did not affect the fungal burdens in the peritoneal cavity and kidney in dectin-2-deficient mice (Fig 10E), supporting the notion that the function of NET in restraining fungal spreading is through dectin-2. These results demonstrate the importance of dectin-2-mediated NETosis in keeping C. albicans from spreading to the kidney.

Discussion

Opsonized C. albicans, through interaction with CR3, activates downstream Syk-dependent NADPH oxidase activation [18] and NADPH oxidase is required for opsonized C. albicans-induced NET formation [11]. Unlike NET formation induced by lipopolysacchride, phorbol 12-myristate 13-acetate (PMA) and Shigella flexneri, PAD4 has been reported not to be involved in opsonized C. albicans-induced NET formation in human and mouse [17, 23, 25]. Opsonized C. albicans-induced NET is inhibited by PKC inhibitor, but not Ca²⁺ chelator [17]. Thus, it appears that CR3 recognition of opsonized C. albicans sends signals to activate NADPH oxidase-dependent NET formation through Syk-PKC-ROS cascade, but PAD4 and Ca²⁺ do not take part in NET formation. Our study employing confocal microscopy, fluorescence microscopy, transmission electron microscopy, live cell imaging and PicoGreen dsDNA assay shows that unopsonized C. albicans triggers neutrophils to undergo a NADPH oxidase-independent NETosis. Distinct from opsonized organisms, unopsonized C. albicans-induced NET is through ligation of dectin-2 that drives Syk-Ca²⁺-PKCδ-NE/PAD4 signaling pathway. Our study together with those of others reveal that opsonized and unopsonized C. albicans utilize different receptors and different signaling pathways to trigger NETosis.

Employing PAD4^-/- and WT neutrophils Guiducci et al. showed in their recent publication [23] that in response to opsonized C. albicans, PAD4 deficiency reduced histone H3 citrullination (confocal microscopy of neutrophils in vitro). However, they found that PAD4 was not required for opsonized C. albicans-induced NETosis (quantified by SytoxGreen assay, immunofluorescence staining for confocal microscopic imaging and electron microscopic imaging). We showed that GSK484 treatment of neutrophils challenged with unopsonized C. albicans reduced NE translocation (immunofluorescence staining for confocal microscopic imaging), Sytox Organge⁺histone H3⁺ web-like structure formation (immunofluorescence staining for fluorescence microscopic imaging) and NETosis (quantified by Quant-iT PicoGreen dsDNA assay). Therefore, it appears that unlike challenge with opsonized C. albicans, PAD4 activation results in NETosis when neutrophils are challenged with unopsonized organisms in vitro. Furthermore, it is shown that PAD4 deficiency increased fungal CFU in kidney on day 3 and 7 after intravenous infection, decreased that in the tongue on day 1 after sublingual infection but no other time points [23]. We treated mice with PAD4 inhibitor GSK484 before intraperitoneally infected them with C. albicans. Inhibition of PAD4 reduced Ki67 expression and web-like structure formation in Ly6G⁺ cells in peritoneal exudate and Ki67⁺Ly6G⁺ cell population in mesenteric tissues (intracellular Ki67 straining followed by flow cytometric analysis and immunofluorescence staining for fluorescence microscopic imaging). In addition, GSK484 treatment decreased fungal CFU in the peritoneal cavity and increased that in the kidney at 3 h after infection. These data clearly demonstrated that PAD4 is important to C. albicans-induced NETosis in vivo. Regarding the role of PAD4 in host defense, our work with peritonitis infection together with that reported by Guiducci et al. with sublingual infection show that PAD4 functions to restrain fungal spread from the inoculation site to distal site during early phase of C. albicans infection [23]. We speculate that since NETosis-mediated restrain of fungal spread [16, 26] does not affect eventual fungal clearance, PAD4 is not required for control of fungal infection.

C-type lectin receptor engagement elicits proinflammatory cytokine response to stimulation by fungal ligand through Syk-mediated PKCδ activation [27]. PKCδ activity modulates CARD9/Malt1/Bcl10 signalosome formation to facilitate downstream NFκB translocation and subsequent cytokine production [27]. Deletion of Prkcd, but not Prkca nor Prkcb genes, abolishes TNF, IL-6 and IL-1β production by dendritic cells upon zymosan, curdlan or C. albicans stimulation [27], indicating the unique role of PKCδ in fungal challenge. We use pharmacological inhibitors for different PKC isoforms and uncover the importance of dectin-2 downstream PKCδ in histone H3 citrullination and NETosis in response to unopsonized C. albicans. Conventional PKCs are known to mediate NADPH oxidase-dependent ROS-mediated NETosis [16]. Our results reveal a PKCδ (novel PKC isoform)-mediated signaling pathway that is involved in unopsonized C. albicans-induced NADPH oxidase-independent NETosis. Our finding also suggests that CARD9/Malt1/Bcl10 signalosome which is downstream of PKCδ may function to mediate NE translocation and PAD4 activation, histone H3 citrullination and trigger NETosis.

C-type lectin dectin-1 has been reported to interfere with C. albicans-induced NET formation in human neutrophils through promoting phagocytosis [20]. NE is normally associated with granule membrane [28]. Upon phagocytosis, it is delivered to phagosome and sequestered within C. albicans-containing phagosome [20]. After which, its access to decondensed chromatin is blocked [20]. Metzler et al. observed in human neutrophils that NE is dissociated from granule membrane via ROS production to gain access to decondensed chromatin [28]. C. albicans yeast is known to be taken up and cleared by human neutrophils and the neutrophils remain intact [29]. Mouse neutrophils, however, allow engulfed C. albicans to germinate, resulting in membrane rupture and eventual cell death [29]. Our data show that dectin-1 and NADPH oxidase are not involved in NETotic response to unopsonized C. albicans. We also found that inhibiting NE reduces unopsonized C. albicans-induced NET formation, and that NE translocates to the nucleus is independent of NCF-1. We further demonstrated that Syk-Ca²⁺-PKCδ-PAD4 pathway modulates NE nuclear translocation and its access to decondensed chromatin without the involvement of ROS. Thus, it appears that chromatin decondensation and NE translocation as a result of dectin-2 downstream Syk-Ca²⁺-PKCδ-PAD4 signaling pathway are important to NADPH oxidase-independent NETosis.

Urban et al. employed pulmonary and subcutaneous C. albicans infection models to investigate NET formation [30]. In vivo NET formation in infected tissues was identified as structures that stained positive for cell-impermeable DNA dye (pre-injected before animals were killed), myeloperoxidase and histone. NET formation in subcutaneous tissue was observed on 6 days after subcutaneous infection and in the lungs at 24 h after intranasal infection [30]. NADPH oxidase deficiency greatly reduces NET formation in pulmonary A. fumigatus infection [31]. It appears that NADPH oxidase is involved in host NETosis response to pulmonary fungal infection in vivo. Ki67 is recently established as a NETosis marker [13]. Neutrophils undergoing NETosis in the lungs of C. albicans-infected mice as well as that in the brain of fungus-infected humans are Ki67-positive [13]. We enriched neutrophils in the peritoneal cavity of a mouse by injecting casein twice before C. albicans infection. Abundant peritoneal neutrophils allow a robust NET response to C. albicans infection. IVIS imaging showed extracellular DNA release as early as 1.5 h after infection and histone H3⁺Ki67⁺Ly6G⁺ neutrophils with web-like structures were observed in peritoneal exudates. The % of Ly6G⁺ neutrophils undergoing NETosis was quantified by flow cytometric analysis of intracellular Ki67. This C. albicans peritonitis model where neutrophils could be easily obtained without tissue homogenization and enzyme digestion is useful in quantifying NET and studying their function.

The role of dectin-2 in host defense against C. albicans is well documented. Dendritic cells utilize dectin-2 to recognize C. albicans for IL-6, IL-1β, and IL-23 production [32]. Dectin-2 signaling induces cytokine production through Syk-CARD9-NF-κB pathway [32]. While dectin-2-deficient dendritic cells have greatly reduced ability to produce cytokines in response to fungal α-mannan, dectin-2 deficiency also dampens host cytokine response to C. albicans [32]. Systemic C. albicans infection results in higher mortality in dectin-2-deficient mice than WT mice [32]. Neutrophils are major effector cells that kill C. albicans [5]. Whether dectin-2 plays a role in neutrophil response to C. albicans is an interesting question. It is reported that the role of dectin-2 in neutrophil ROS response to opsonized C. albicans is only marginal [19]. Our results show that dectin-2 recognizes unopsonized C. albicans for NETotic response, and dectin-2 deficient mice have reduced ability to restrain fungal spread from peritoneal cavity to kidney. These results suggest that C. albicans injected to the peritoneal cavity remain unopsonized at least for a short period of time to be recognized by dectin-2 and reveal a new role for dectin-2 in neutrophil anti-C. albicans functions.

In summary, this study showed that recognition of unopsonized C. albicans by dectin-2 triggers NET formation through a NADPH oxidase-independent pathway. Signaling pathway leading to NETosis involves Syk-Ca²⁺-PKCδ-NE/PAD4. Dectin-2-mediated NET as revealed in the C. albicans peritonitis model functions to control fungal spread from peritoneal cavity to kidney. Our work provides a better understanding of the molecular mechanism involved in NADPH oxidase-independent NET formation and sheds light on the role of dectin-2 in neutrophil anti-C. albicans function.

Materials and methods

Fungus and infection

C. albicans strain SC5314 (ATCC MYA-2876), its isogenic mutant strain HLC54 (yeast-locked strain, efg1/efg1 cph1/cph1), GFP-expressing strain OG1 [33], and dTomato-expressing strain CFA2-dTomato [34] were used in this study. All strains were cultured on yeast-peptone-dextrose (YPD) agar (DIFCO) plate at 30°C. Mice were injected intraperitoneally with C. albicans yeasts prepared in HBSS buffer. Unopsonized C. albicans was prepared in phenol red free HBSS buffer for experiments. To opsonize, C. albicans yeasts were added to phenol red free HBSS containing 10% fresh mouse serum and let stand at room temperature for 30 min. To induce hyphal formation, C. albicans yeasts were incubated in RPMI 1640 medium at 37°C for 4 h before use.

Mice

Wild-type (C57BL/6), Itgam^-/-, Ncf-1^-/- (originally purchased from the Jackson Laboratories, Bar Harbor, ME, USA), Clec7a^-/- (from Dr. Gordon Brown, University of Cape Town, Cape Town, South Africa) [35], Clec4n^-/- [19] and MyD88^-/- (from Dr. Tsung-Hsien Chuang, National Health Research Institutes, Taiwan) mice were bred and maintained at the Laboratory Animal Center of National Taiwan University College of Medicine. All mice used in this study were maintained under specific pathogen-free conditions. Mice at 6–12 weeks of age were used in all of the experiments.

Bone marrow neutrophils

Bone marrow cells were harvested from the femurs and suspended in dPBS buffer before overlaid on discontinuous percoll gradients (55%, 62%, and 81% in the order from top to bottom) (GE healthcare). After centrifugation at 1,400 × g for 30 min, cells at the interface between 62% and 81% gradients were harvested and washed. Flow cytometric analysis showed that 90–95% of cells were CD11b⁺Ly6G⁺.

NET induction and quantification

Two hundred thousand neutrophils suspended in HBSS were seeded in 96 well-plate before addition of 4 × 10⁵ opsonized or unopsonized C. albicans. The plate was centrifuged at 800 × g for 3 min to spin down cells. Wells were treated with 0.5 U of micrococcal nuclease (MNase, NEB) 3 h later and incubated at 37°C for 10 min to partially digest NET. Supernatants were collected and cell free dsDNA were quantified by Quant-iT PicoGreen dsDNA assay kit (Life technology) according to manufacturer’s instruction.

Treatment with inhibitor

Neutrophils were pre-treated with indicated inhibitors 30 min before addition of C. albicans. Inhibitors SkyI (for Syk), BAPTA-AM (a selective chelator of Ca²⁺), Ro 318220 (for total PKC), Rottlerin (for PKCδ), BB-CI-Amidine (for PAD1-4), GSK484 (for PAD4) were all purchased from Cayman. Ro 6976 (for PKCα+β1), LY 333531 (for PKCβ) were from Millipore.

Flow cytometric analysis of Ca²⁺ influx in neutrophils

One million neutrophils were suspended in 200 μl of phenol red-free HBSS buffer. Loading dye for intracellular Ca²⁺ staining was prepared by adding Calcium Indicator to Signal Enhancer at the ratio of 1:1000 according to the manufacturer’s recommendation (BD Biosciences Calcium Assay Kit, 640176). Two hundred μl of loading dye was added to cells and the mixture was incubated at 37°C for 45 min. After resting in room temperature for 20 min, cells were placed in a 5 ml FACS tube and analyzed by flow cytometry to set the basal level of intracellular Ca²⁺ intensity (30 sec). Unopsonized pre-germinated C. albicans prepared in 10 μl of HBSS was then added to the tube for continuous flow cytometric analysis for additional 300–420 sec. Data were analyzed by the kinetic mode of FlowJo software. All procedures including sample acquisition and data analysis followed that of modified Bio-protocol published by S. Lee [36]. Original FACS contour plot for measurement of Ca²⁺ intensity is shown in S1 Fig.

Immunofluorescence staining

Five hundred thousand neutrophils mixed with C. albicans at a ratio of 1:2 were plated on coverslips and incubated at 37°C for 3 h. Coverslips were fixed in 10% formaldehyde for 15 min and permeabilized with 0.5% Triton X-100. After thorough wash, coverslips were blocked (10% FBS in PBS) and stained with anti-neutrophil elastase (abcam, 1:50) or anti-histone H3 antibody (Cell Signaling, 1:100) at 4°C overnight. Coverslips were stained with cell-permeable DNA dye Hoechst 33258 (1 μg/ml, Invitrogen) or cell-impermeable DNA dye SYTOX Orange (1 μM, Life technology) diluted in blocking buffer and left on ice for 15 min. Coverslips were then mounted by mounting gel and subject to fluorescence microscopic or confocal microscopic analysis.

Live cell imaging

Five hundred thousand neutrophils suspended in HBSS containing SYTOX Orange (0.5 μM) and cell-permeable DNA dye Draq5 (2 μM) were seeded in a chamber (1 μ-Slide 8 well ibiTreat plates, ibidi). After spun down, 2 × 10⁶ of pre-germinated C. albicans OG1 were added. NET release was monitored by inverted confocal microscope LSM 780 AxioObserver Z1 for three-color.

NET fungicidal activity assay

Twenty thousand neutrophils suspended in HBSS were seeded in 96-well plate and allowed to adhere for 30 min before addition of 4 x 10⁵ unopsonized C. albicans yeasts. Wells containing C. albicans yeasts without neutrophils were used as control. Plate was centrifuged at 800 × g for 3 min to spin down yeasts. HBSS containing micrococcal nuclease or heat-inactivated MNase (h.i. MNase) was added to the final concentration of 10 U/ml. Buffer was collected and cold H₂O (pH = 11) was added 3 h later to lyse cells. C. albicans was detached by mini cell scraper and vigorously pipetting. The number of viable fungi was determined by plating the supernatant on yeast-peptone-dextrose agar plate. Colony counts (CFU) were enumerated 2 days later. The ability of neutrophils to kill C. albicans is presented as % killing of C. albicans which was calculated by dividing the difference of CFU counts between the control group (without neutrophils) and neutrophil-added groups by the counts of the control.

Western blot analysis

Neutrophils stimulated with or without C. albicans hyphae (MOI = 2) were lysed in PhosphoSafe Extraction Reagent (EMD Millipore). Cell lysates were separated by electrophoresis at 10 or 12.5% SDS-polyacrylamide gel and transferred to Immobilon-P membrane (Millipore). Membrane was blocked with 5% nonfat milk (Fluka) for 1 h and incubated in buffer containing rabbit anti-pSyk (Tyr575), -citrullinated histone H3 (R2 + R8 + R17) (Abcam), -pPKCδ (Ser645) (Cell Signaling), and -β-actin, -GAPDH (GeneTex) antibodies at 4 °C overnight. Membrane was then incubated with buffer containing goat anti-rabbit/rat IgG-HRP antibody (GeneTex) for 1 h. Western Chemiluminescent HRP (Millipore) was used as substrate.

Peritoneal C. albicans infection in casein-peritonitis mice

Mice were injected with 1 ml of 9% casein (9 g of casein sodium salt dissolved in 100 ml of hot PBS) intraperitoneally twice with an interval of 18 h. Four hours after the second casein injection, mice were given 1×10⁸ of C. albicans intraperitoneally. For some experiments, HBSS containing 100 U of micrococcal nuclease or heat-inactivated MNase (h.i. MNase) was administered intraperitoneally at the time of infection. At indicated time after infection, peritoneal exudates, mesentery tissues and kidneys were collected and subject to following experiments.

In vivo imaging and quantification of fluorescence

Before imaging, mice were injected with 1 ml of 9% casein intraperitoneally twice with an interval of 18 h. Four hours later, mice were infected by 5 × 10⁸ of C. albicans (SC 5314 or CAF2-dTomato) in 200 μl HBSS intraperitoneally. To monitor NET release, mice were given an additional intraperitoneal injection of 100 μl SYTOX Orange (5 μM) at the time of infection. A Xenogen IVIS Imaging System 200 series (PerkinElemer Inc.) was used to quantify fluorescent C. albicans (CAF2-dTomato) or the release of extracellular DNA over time. Photons in user-specified region of interest (ROI, gated area) during 10 sec exposure were measured by Living Image 3.2 software. Relative intensity of total photons in ROI was calculated based on the intensity at 1 h after infection.

Tissue and peritoneal exudate cell immunofluorescence staining

Mesentery were embedded in O.C.T. and allowed to freeze at -80°C overnight. Five μm thick tissue sections were cut and mounted on gelatin-coated slides. Peritoneal exudates were diluted 1:5 and seeded on coverslips for 1 h. Slides and coverslips were fixed with 4% paraformaldehyde, left on ice for 30 min, permeabilized with 0.5% Triton X-100 and let sit at room temperature for 5 min. After wash, samples were blocked (PBS containing 5% FBS) and stained with anti-histone H3 (Cell Signaling), -Ki67 and -Ly6G (Biolegend) antibodies at 4°C overnight. Nucleus was stained with Hoechst 33258 (1 μg/ml, Invitrogen) for 15 min. Samples were then mounted by mounting gel and subject to confocal microscopic analysis.

Quantification of fungal load

Kidneys were collected and homogenized in a tissue grinder with 1 ml of RPMI 1640 medium. Peritoneal exudate was collected from mice after intraperitoneal injection of 5 ml of HBSS. Homogenates and peritoneal fluids were treated with 0.5 U of micrococcal nuclease and incubated at 37°C for 10 min to digest DNA entangled with fungus. One hundred microliter of kidney homogenates and peritoneal exudates were plated on YPD agar. Colonies were counted after incubation at 30°C for 2–3 days.

Ki67 staining

Peritoneal exudates from naïve and infected mice were collected, treated with 0.5 U/ml MNase and subject to surface staining for neutrophil marker CD11b and Ly6G. After fixation in 4% paraformaldehyde and permeabilization with 1% saponin, cells were stained with anti-Ki67 antibody prepared in staining buffer (0.5% saponin) overnight. Cells were fixed in 1% PFA and subject to flow cytometric analysis.

Statistics

Student t test was used to compare the difference between two groups. Statistical significance was defined as P < 0.05.

Ethics statement

Mouse study was carried out in strict accordance with the recommendations in the Guidebook for the Care and Use of Laboratory Animals, The Third Edition, 2007, published by The Chinese-Taipei Society of Laboratory Animal Sciences. All animal procedures and experimental protocols were approved by AAALAC-accredited facility, the Committee on the Ethics of Animal Experiments of the National Taiwan University College of Medicine (Permit Number: 20140304, 20140533 and 20180013).

Supporting information

S1 Fig [tif]
The contour plot of calcium influx in neutrophils upon unopsonized . stimulation and the gating strategy for % Ca-positive cells.

S2 Fig [green]
Neutrophil elastase (NE) translocation in and neutrophils upon stimulation by unopsonized . .

S3 Fig [a]
NETotic response of and mice to peritoneal . infection.

S4 Fig [a]
NETotic response of and mice to peritoneal . infection.

S1 Video [blue]
Stimulation of neutrophils by opsonized . triggers NET formation.

S2 Video [blue]
Stimulation of neutrophils by unopsonized . triggers NET formation.

S3 Video [blue]
Stimulation of neutrophils by unopsonized . triggers NET formation.