Spleen Tyrosine Kinase Is a Critical Regulator of Neutrophil Responses to Candida Species.

Neutrophils are recognized to represent significant immune cell mediators for the clearance and elimination of the human-pathogenic fungal pathogen Candida. The sensing of fungi by innate cells is performed, in part, through lectin receptor recognition of cell wall components and downstream cellular activation by signaling components, including spleen tyrosine kinase (Syk). While the essential role of Syk in macrophages and dendritic cells is clear, there remains uncertainty with respect to its contribution in neutrophils. In this study, we demonstrated that Syk is critical for multiple cellular functions in neutrophils responding to major human-pathogenic Candida species. These data not only demonstrate the vital nature of Syk with respect to the control of fungi by neutrophils but also warn of the potential infectious complications arising from the recent clinical development of novel Syk inhibitors for hematologic and autoimmune disorders.

phils were previously shown to fail to undergo respiratory burst, degranulation, or spreading to integrin signaling and to be incapable of generating ROS intermediates in response to Fc␥R engagement (37,38). In the absence of Syk, PMNs show impaired exocytosis of secondary and tertiary granules, reduced cytokine release, impaired NET formation, and very poor activation of the NADPH oxidase in response to bacteria, including staphylococci and Escherichia coli (39). In studies of neutrophil fungal interactions with Aspergillus, Syk-deficient neutrophils were previously shown to be defective in phagocytosis and killing as well as in ROS production (40,41). Syk has also been shown to be indispensable for ␤-glucan-induced NET formation in neutrophils (42). However, the specific role of Syk in neutrophil-Candida interactions has not been well characterized.
Despite evidence suggesting a crucial role for Syk in PMNs, elucidation of a precise role for Syk in neutrophils has been challenging for a variety of reasons. Deletion of Syk is embryonic lethal. Early studies on Syk-deficient neutrophils have relied on implantation of chimeric Syk-deleted liver cells into wild-type recipients resulting in syk elimination within all hematopoietic cells (37,40). Other studies use murine models with cell lineage-specific deletion of Syk resulting in syk elimination in mature neutrophils or monocytes (39,42). The short life span and terminally differentiated state of PMNs present major obstacles to defining the molecular mechanisms responsible for their responses to Candida (43). To circumvent the limitations posed by primary neutrophils, we used a granulocyte-monocyte progenitor (GMP) cell line that was conditionally immortalized by expression of an estrogen receptor-homeobox B8 (ER-HoxB8) gene fusion as described previously (44,45). The media for these cells contains stem cell factor (SCF) and estradiol, which permits nuclear translocation of the ER-HoxB8 fusion protein, resulting in conditional maturation arrest at the GMP stage. Removal of estradiol from the cell media allowed synchronous differentiation of the GMP into mature neutrophils (46). This method of gene manipulation also permits a complete deletion of the Syk gene in a neutrophil-specific manner.
In this study, we sought to determine the role of Syk in PMN responses against three major human-pathogenic fungal pathogens, C. albicans, C. glabrata, and C. auris. We show that Syk is essential in neutrophil killing of all three Candida species. We further demonstrate that genetic elimination or chemical inhibition of Syk in neutrophils results in diminished tumor necrosis factor alpha (TNF-␣) and ROS production. Our results indicate that neutrophils lacking Syk have decreased mechanical responses, such as phagocytosis, NETosis, and swarming, to all three Candida species. Together, these data suggest that the loss of Syk results in global defects in neutrophil function against Candida species.

RESULTS
Primary neutrophils require Syk to eliminate Candida species. Bone marrowderived primary mouse neutrophils were evaluated for their ability to eliminate C. albicans, C. glabrata, and C. auris in vitro. Primary neutrophils were coincubated with yeast at the indicated multiplicities of infection (MOIs) and evaluated for fungicidal activity using PrestoBlue viability indicator dye as previously described (47,48). Primary neutrophils effectively controlled C. albicans (Fig. 1A) and C. glabrata (Fig. 1B), as well as C. auris (Fig. 1C).
To evaluate the role of Syk, the chemical inhibitor PRT062607 (PRT) (2 M) was included in the neutrophil and Candida coculture. PRT had minimal effect on the viability of yeast or neutrophils alone (viability of Ͼ95% by acridine orange [AO] and propidium iodide [PI] staining; data not shown). Interestingly, PRT suppressed the ability of primary neutrophils to kill all three Candida species across multiple MOIs ( Fig. 1), demonstrating the importance of Syk for neutrophil fungicidal activity.
Syk-deficient neutrophils lack the ability to kill Candida. To further investigate the role of Syk in neutrophil killing and verify the observations made with smallmolecule inhibitors, genetic Syk knockout (KO) neutrophil cell lines were generated using CRISPR single guide RNA (sgRNA) targeting of Syk in a GMP neutrophil cell line constitutively expressing Cas9. Western blotting demonstrated upregulation of Syk protein expression as cells matured from progenitors (with estradiol [ϩestradiol]) to terminally differentiated neutrophils (without estradiol [Ϫestradiol]) ( Fig. 2A). Syk knockout 1 (KO1) and Syk KO2, two independent cell lines with loss of Syk, were generated; Western blotting confirmed the absence of Syk expression in either the progenitor stage or mature neutrophil stage ( Fig. 2A). Despite lacking Syk, the knockout cell lines were capable of differentiation to fully mature neutrophils, demonstrating morphologies by Giemsa staining (Fig. 2B) and ATP production (Fig. 2C) that were similar to those seen with bone marrow-derived primary neutrophils. In addition, Syk knockout cells expressed myeloperoxidase activity levels that were similar to the levels seen with the parental Cas9 neutrophils (Fig. 2D) as well as equivalent levels of surface expression of Ly6G and Mac-1 (Fig. 2E). Similarly to the results seen with chemical inhibition using PRT, Syk-deficient PMN demonstrated significantly attenuated fungicidal activity against C. albicans, C. glabrata, and C. auris (Fig. 2F).
Syk regulates neutrophil cytokine and ROS production in response to Candida. We next evaluated the role of Syk in neutrophil cytokine and ROS responses to Candida. Role of Syk in Neutrophil Responses to Candida ® Neutrophils were stimulated with heat-killed Candida, and production of ROS was measured by lucigenin luminescence. Following 2 h of stimulation, wild-type neutrophils produced robust ROS levels following stimulation with C. albicans hyphae and, to a lesser extent, with C. auris and C. glabrata (Fig. 3). Syk-deficient neutrophils, on the other hand, did not produce ROS in response to any Candida species.
To evaluate TNF-␣ production, neutrophils were stimulated with heat-killed C. albicans, C. glabrata, and C. auris, at a ratio of 10 yeast cells to 1 PMN, and supernatant was analyzed for cytokine stimulation by enzyme-linked immunosorbent assay (ELISA). Primary bone marrow-derived neutrophils produced TNF-␣ against all three Candida species, but production decreased significantly in the presence of 2 M Syk inhibitor PRT (Fig. 4A). Syk knockout neutrophils also produced much lower levels of TNF-␣ against Candida species than the parental Cas9 neutrophil control (Fig. 4B). The addition of PRT to Syk KO PMN did not have an additional effect on TNF-␣ production. These data confirm that the production of ROS and secretion of TNF-␣ in neutrophil responses to Candida are dependent on the presence of Syk.
Syk regulates neutrophil mechanical responses against Candida. We then queried the role of Syk in neutrophil mechanical responses against Candida, including phagocytosis, swarming, and NET formation. Neutrophils were coincubated with heatkilled, fluorescently labeled C. albicans, C. glabrata, and C. auris. Phagocytosis was measured by flow cytometry monitoring the uptake of the AF647-labeled yeast with neutrophils ( Fig. 5A). Cytochalasin D, an actin polymerization inhibitor, was included as a control for nonphagocytic associations. Primary bone marrow-derived neutrophils phagocytosed all three Candida species to different extents by 30 min. In the presence of 2 M Syk inhibitor PRT, phagocytosis of Candida species was almost entirely inhibited to the same activity level as that seen with cytochalasin D (Fig. 5B). Similarly, Syk knockout neutrophils showed decreased phagocytosis of Candida species compared to the parental Cas9 neutrophils (Fig. 5C).
To evaluate neutrophil swarming, representing a coordinated multicellular response to a pathogen, C. albicans, C. glabrata, and C. auris were plated in clusters on glass slides and neutrophil behavior was observed using time-lapse light microscopy. While Cas9 cells swarmed around yeast clusters of all three species and effectively controlled yeast proliferation, use of chemical inhibition of Syk with PRT or Syk-deficient cells revealed  evenly distributed cells throughout the yeast cluster, representing poor yeast control ( Fig. 6A). At 16 h post swarming, yeast clusters were better controlled, with smaller surface areas of fungal growth in the presence of parental, wild-type Cas9 neutrophils, while the Syk-deficient neutrophils resulted in poor control and larger areas of fungal growth (Fig. 6B).
To visualize formation of NETs, neutrophils were coincubated with C. albicans, C. glabrata, and C. auris at an MOI of 10. NET formation was measured by Sytox fluorescence, demonstrating the release of extracellular DNA from wild-type neutrophils following Candida stimulation but minimal NET formation in the Syk-deficient cells (Fig. 7A). The citrullinated-histone-specific antibody (Ab) Cit-H3 was used to label NETs in conjunction with Sytox staining and was visualized by fluorescence microscopy (Fig. 7B). Electron microscopy images confirmed NET formation in wild-type Cas9 neutrophils but not Syk-deficient PMN in response to Candida (Fig. 7C).

DISCUSSION
While Syk is an important kinase in innate immune cells, its exact role in neutrophil fungicidal activity has not been completely determined. Here, we demonstrated that multiple critical neutrophil molecular mechanisms responsible for Candida control and elimination, including ROS and cytokine production, phagocytosis, NET production, and swarming, are Syk dependent. Chemical inhibition or genetic Syk deficiency results in widespread neutrophil functional defects in response to C. albicans, C. glabrata, and C. auris. Of note, the loss of fungicidal activity was most pronounced in response to C. glabrata, suggesting that Syk may have differential roles depending on the fungal species, most likely related to types of pattern recognition receptors activated by each Candida species. Interestingly, our study suggests that mouse neutrophils kill the South American clade of C. auris very effectively, despite other reports demonstrating significant C. auris resistance against human neutrophil killing (49). These observations likely reflect differences between human and mouse neutrophils, as well as differential responses to C. auris clades, although further investigation will be required to elucidate these potential mechanisms.  We explored how loss of Syk results in dysfunctional neutrophil responses to Candida. Neutrophils are known to produce abundant levels of ROS, a necessary step for neutrophils to form NETs. In addition, PMNs generate cytokines, including TNF-␣, when they encounter Candida species also serving as a neutrophil activator (50). Our results demonstrate a reduction of both ROS and TNF-␣ when Syk is chemically inhibited or genetically deleted from neutrophils. We posit that Syk-dependent receptors are critical for activation of the recognition and response pathways responsible for the core neutrophil functions against Candida species.
Swarming is a cooperative and exponential migratory behavior by which neutrophils can protect healthy tissue by rapidly controlling sites of infection (51). Neutrophil swarming occurs in models of sterile tissue damage as well as in response to fungal pathogens, including Cryptococcus neoformans and C. albicans, and is dependent on the presence of leukotriene B4 (LTB4) (51)(52)(53). We found that neutrophils swarm around clusters of Candida species restricting their growth, and that this ability is compromised in Syk-deficient cells. Thus, in addition to altering lectin receptor signaling, it is likely that Syk plays a role in chemotaxis and locomotion in response to chemoattractants, such as LTB4, a central requirement of swarming (52), though this will require further investigation.
We ensured that the genetic elimination of Syk from the ER-HoxB8 conditionally immortalized neutrophils did not result in a loss of viability that would confound our findings. The Syk-deficient cells demonstrated normal differentiation (as assayed by morphology and cell surface marker expression) as well as preservation of energy production (ATP) and myeloperoxidase (MPO), a central feature of mature neutrophil granules (54). These data suggest that neutrophil dysfunction related to the loss of Syk is not due to maturational cues or differentiation of GMP progenitors into neutrophils.
A limitation of our conditionally immortalized GMP cell line is its restriction to mouse neutrophils. While functions of human and mouse neutrophils are highly conserved, some differences do exist. For example, there are differences in the content of the neutrophil granules, where human PMN have defensins but mouse neutrophils do not (55,56). They differ in their levels of expression of myeloperoxidase, where murine neutrophils have just 10% to 20% the level of myeloperoxidase activity found in human neutrophils (56). Human neutrophils also have a different propensity for NETosis, where murine neutrophils are less efficient at NET formation and, when made, are more compact than human NETs (57,58). However, chemical inhibition of Syk in human neutrophils also results in reduced inflammatory cytokine production (59), phagocytosis (60), and ROS production and NET release (61), suggesting that Syk plays similar roles in the two species and that many of our findings in the primary murine neutrophils and cell lines will likely translate to humans.
Our data illustrate an essential role for Syk in neutrophil fungicidal activity against Candida species. Primary neutrophils treated with PRT and Syk KO conditionally immortalized PMN lines demonstrated concordant results. Syk appears to influence a variety of pathways, including ROS generation, cytokine production, phagocytosis, swarming, and NET formation. Our studies were limited to Candida species necessitating further work to understand the role of Syk when neutrophils engage with other fungi such as dimorphic species and molds causing primary human mycoses. These data raise awareness with respect to future patients who may be treated with Syk inhibition for autoimmune diseases and their increased susceptibility to invasive fungal disease.

MATERIALS AND METHODS
Reagents. PRT062607 (PRT), a specific Syk inhibitor, was purchased from Selleckchem (Houston, TX). . Bright-field images taken after 9.5 h of incubation are shown for each condition. Fungi grew well on the arrays when incubated alone (first row), but Cas9 cells (second row) were able to restrict fungal growth. Syk KO cells (third row) and PRT-treated cells (fourth row) were less capable of containing fungal growth after swarming. (B) The area covered by fungal growth is quantified as "Area (pixel 2 )" on the y axis. Data represent comparisons of Cas9 (wild-type), Syk KO, and PRT-treated Cas9 conditions. n ϭ Ն27 swarms across three independent experiments. *, P Յ 0.05; **, P Յ 0.01; ****, P Յ 0.0001 (ANOVA with Tukey's posttest for C. albicans and C. auris, Kruskal-Wallis for C. glabrata). Scale bars, 100 m.
Primary neutrophils. Primary neutrophils were isolated as previously described (46). Male and female mice used for bone marrow harvests were at least 8 weeks old. Briefly, fresh bone marrow was harvested from mouse femurs and tibias, treated with 0.2% followed by 1.6% hypotonic sodium chloride solution for 20 s each to remove red blood cells, and purified by density gradient centrifugation.
Neutrophil cell lines. To circumvent the limitations posed by primary neutrophils, we used a granulocyte-monocyte progenitor (GMP) cell line that was conditionally immortalized by expression of an estrogen receptor-homeobox B8 (ER-HoxB8) gene fusion as described previously (44). The cell media for these cells contain stem cell factor and estradiol, which permits nuclear translocation of the ER-HoxB8 fusion protein, resulting in a conditional maturation arrest at the GMP stage. The removal of estradiol from the cell medium allows synchronous differentiation of the GMP into mature neutrophils (46). Cas9 GMP cell lines were immortalized from the Cas9 transgenic mouse (63), with expression at high (Ͼ98%) efficiency (data not shown). All GMP cell lines were cultured in complete RPMI media containing 2% stem cell factor-conditioned media and 0.5 M ␤-estradiol and grown in a humidified incubator at 37°C in the presence of 5% CO 2 .
Yeast. Wild-type C. albicans (SC5314) was a gift from Eleftherios Mylonakis (Brown Medical School, Providence, RI). Wild-type C. glabrata (ATCC 2001) was purchased from the American Type Culture Collection (ATCC, Manassas, VA). A sequence-confirmed clinical isolate of C. auris belonging to the South American clade was obtained from the MGH microbiology lab (Boston, MA) (64). Yeast cultures were grown overnight in liquid YPD at 30°C with shaking, washed three times in phosphate-buffered saline (PBS) after collection, counted with a Luna automated cell counter (Logos Biosystems, Annandale, VA), and resuspended in PBS at the desired inoculum. CRISPR-Cas9 system and sgRNA design. To construct the lentiviral sgRNA Cas9 vector, single guide RNAs (sgRNAs) were cloned into lentiCRISPR vector in the BsmBI site (46). Lentivirus production and purification were performed as previously described (65). sgRNAs were designed using the CRISPR tool (66) (http://crispr.mit.edu) to minimize potential off-target effects. lentiCRISPR with sgRNAs targeting Syk were cloned using the following sequences: for Syk sgRNA1, ACAGAGACATACCTGCCAGA; for Syk sgRNA2, GTCTTGGGCTGTACTCCCGG; and for Syk sgRNA3, ACACCACTACACCATCGAGA.
PrestoBlue killing assay. Neutrophils were plated at 1 ϫ 10 5 cells per well in complete RPMI medium in a 96-well tissue culture plate. PRT (2 M) was added to pretreat appropriate wells for 20 min at 37°C with 5% CO 2 . After Syk inhibitor pretreatment, various multiplicities of infection (MOIs) of Candida species were coincubated with the neutrophils (C. albicans for 2 h, C. auris for 4 h, and C. glabrata for 24 h). A standard curve was generated with serial dilution of yeast cells. Following incubation, all PMNs were lysed with a 1% NP-40 solution containing 10 mM Tris HCl, 150 mM sodium chloride, and 5 mM magnesium chloride, titrated to pH 7.5. After neutrophil lysis, optimized yeast growth medium was added to supplement Candida growth (complete RPMI medium for C. albicans, YPD medium for C. glabrata, and RPMI-MOPS for C. auris). Finally, 10% PrestoBlue cell viability reagent (Thermo Fisher Scientific) was added to each well and fluorescence was read every hour for 18 h at 30°C using a SpectraMax i3x reader (Molecular Devices, Sunnyvale, CA) (67). Standard yeast growth curves were generated to accurately determine the number of viable yeasts. Fluorescence was plotted versus time, and the time to the midcurve (inflection point) was determined using GraphPad Prism 7 software (La Jolla, CA). Percent killing was determined as follows: 1 Ϫ (viability of yeast plus neutrophils/viability of yeast only).
Reactive oxygen species (ROS) production. Reactive oxygen species production was measured as previously described (25). Briefly, neutrophils were plated at 5 ϫ 10 5 cells per well in complete RPMI medium in 96-well white-walled plates (Grenier Bio-One, Monroe, NC). Cells were placed on ice, and heat-killed Candida species were added at an MOI of 10. A lucigenin solution was added to each well for a final concentration of 15 M lucigenin. Luminescence was measured at 37°C every 5 min for 4 h in a SpectraMax i3x reader (Molecular Devices, San Jose, CA), and the results were expressed as arbitrary luminescence units.
TNF-␣ enzyme-linked immunosorbent assay (ELISA). Neutrophils were plated in 96-well tissue culture dishes at 1 ϫ 10 5 cells per well in complete RPMI medium and were cocultured with heat-killed Candida at an MOI of 10. Where appropriate, wells were treated with Syk inhibitor PRT at a concentration of 2 M. After 24 h of stimulation at 37°C with 5% CO 2 supplementation, the concentration of TNF-␣ in the supernatant was measured with ELISA per the instructions of the manufacturers (Duoset; R&D Systems, Minneapolis, MN, and BD Pharmingen, San Jose, CA).
Phagocytosis. To assess phagocytosis, neutrophils were coincubated with AF647-stained heat-killed Candida yeast cells at a ratio of 3 yeast cells per PMN for 5 h at 37°C in a 1.5-ml tube. Samples were fixed in 10% formalin. Neutrophils were labeled with CD11b-fluorescein isothiocyanate (CD11b-FITC), and samples were divided into three wells of a 96-well U-bottom plate. Where appropriate, cells were pretreated with the Syk inhibitor PRT at a concentration of 2 M or with cytoskeletal inhibitor cytochalasin D (Sigma) at a concentration of 30 M. Flow acquisition was done with a high-throughput plate adaptor attached to a FACSCalibur flow cytometer (Becton, Dickinson, San Jose, CA, USA), using CellQuest software (Becton, Dickinson). Percent phagocytosis was measured by selecting for doubly positive CD11b ϩ and AF647 ϩ neutrophils. Flow data were analyzed using FlowJo 10 software (FlowJo, Ashland, OR).
Cellular ATP level. Live cells (2 ϫ 10 4 per well) were plated in PBS on a 96-well white-walled plate. CellTiter-Glo (Promega, Madison, WI) was added at a 1:1 volume ratio, and the cells were incubated for 5 min in the dark and assessed for total luminescence using a SpectraMax i3x reader.
Myeloperoxidase (MPO) activity. Purified live neutrophils (5 ϫ 10 5 per well) were plated in PBS on a 96-well tissue culture plate and serially diluted. Triton X-100 (Sigma, St. Louis, MO) was added at 0.1% for cell lysis. A TMB (3,3=,5,5=-tetramethylbenzidine) substrate reagent set (BD Biosciences, San Jose, CA) was then added at a 1:1 volume ratio to the lysed cells. Endpoint absorbance was read at 605 nm on the SpectraMax i3x reader, and the results were expressed as arbitrary units corresponding to MPO activity levels.
Flow cytometric cell surface marker analysis. Cells (1 ϫ 10 5 ) were labeled with PE-labeled or APC-labeled MAb for 30 min at 4°C in phosphate-buffered saline (PBS)-1% fetal bovine serum (FBS)-1 mM EDTA, washed, and resuspended in the same buffer. Flow cytometry data were acquired with the CellQuest program on a FACSCalibur. Live cells were gated for analysis of forward and side scatter signals. Surface marker expression was measured as geometric mean fluorescence and analyzed using FlowJo 10 software.
Wright-Giemsa staining. Wright-Giemsa staining was performed as previously described (44). Cytocentrifuge preparations of cells (Thermo Shandon, Pittsburgh, PA) were stained 4 min in Wright stain, followed by 12 min in 20% Giemsa stain. Staining was visualized with light microscopy.
Neutrophil swarming microscopy. Utilizing a microarray printing platform (Picospotter; PolyPico, Galway, Ireland) and arrays of spots prepared with a solution of poly-L-lysine (Sigma-Aldrich, Saint Louis, MO), ZETAG and FITC (added to visualize the spots by microscopy) were printed onto ultraclean glass slides (Fisher Scientific, Hampton, NH). Slides were screened for accuracy and then dried for at least 2 h before use. ProPlate 16-well plates (Grace Bio-labs, Bend, OR) were attached to the glass slides with printed arrays. A 50-l volume of the desired target, consisting of live C. albicans, C. glabrata, or C. auris inocula suspended in PBS, was added to each well, and each suspension was incubated with rocking for 5 to 10 min. Following incubation, the wells were thoroughly washed out with water to remove unbound yeast from the glass surface. Wells were screened to ensure appropriate patterning of targets onto the spots with minimal nonspecific binding before use. Swarming was observed using a Nikon Ti-E microscope equipped with an electron-multiplying charge-coupled-device (EM-CCD) camera (Hamamatsu; C9100-13). The microscopy chamber, housed in a Plexiglas environmental chamber, was humidified and maintained at 37°C (68). The swarming targets to be observed during the experiment were selected and saved using the multipoint function in Metamorph (Molecular Devices, Downingtown, PA) prior to loading. Neutrophils (5 ϫ 10 5 ) were then added to each well. All selected points were optimized for perfect focus before launching of the experiment. For the GMP cells treated with 2 M PRT, the neutrophils were preincubated for 30 min before use. Swarming area analysis was done by manually outlining the swarms or areas of fungal growth in ImageJ software. Images were analyzed using Adobe Creative Cloud Photoshop and Illustrator (Adobe, San Jose, CA).
Neutrophil extracellular trap (NET) formation quantification. Neutrophils were plated at 5 ϫ 10 5 cells per well in complete RPMI medum in a 96-well tissue culture plate. Where appropriate, wells were treated with PRT at a concentration of 2 M. Candida species at an MOI of 10 were coincubated with the neutrophils for 1 h at 37°C. After the incubation, Sytox red (Invitrogen, Carlsbad, CA) was added at a final concentration of 160 nM to detect extracellular DNA. Sytox red fluorescence was read every hour for 18 h at 30°C using a SpectraMax i3x reader, and the results were expressed as arbitrary fluorescence units. Staining for NET histones was carried out as previously described (69). Briefly, neutrophils were incubated with C. albicans hyphae or C. glabrata and C. auris yeast cells for 3 h. Samples were then stained with anti-histone H3 citrulline R2ϩR8ϩR17 (Abcam, Cambridge, MA; 0.014 mg/ml) followed by donkey anti-rabbit IgG Cy3 (Jackson Immunoresearch, West Grove, PA; 0.0075 mg/ml) secondary antibody. Samples were then imaged at ϫ40 on Nikon Ti-E microscope.
Scanning electron microscopy. A coverslip model was used for scanning electron microscopy experiments, as previously described (70,71). Candida yeasts (1 ϫ 10 6 ) were added to poly-L-lysinetreated plastic 13-mm-diameter coverslips (Thermanox; Thermo Fisher Scientific, Waltham, MA) and incubated for 1 h at 30°C. Neutrophils (5 ϫ 10 5 ) were added to wells, and cocultures were incubated for 4 h at 37°C. After washing was performed, coverslips were fixed overnight in 4% formaldehyde-1% glutaraldehyde, followed by washing and treatment with 1% osmium tetroxide. Samples were then washed, dehydrated through a series of ethanol washes followed by critical-point drying, and coated with a 14-nm-thick platinum layer. Microscopy was completed on a LEO 1530 scanning electron microscope at 3 kV.
Statistics. Statistical calculations were performed using GraphPad Prism 7 software (La Jolla, CA). Data were analyzed by two-tailed unpaired t test or by one-way analysis of variance (ANOVA), where appropriate, and were considered significantly different for P values of Յ0.05.