Zinc and Manganese Chelation by Neutrophil S100A8/A9 (Calprotectin) Limits Extracellular Aspergillus fumigatus Hyphal Growth and Corneal Infection Free

This article is featured in In This Issue, p.1

Aspergillus fumigatus causes severe pulmonary and disseminated infections in patients with compromised immune systems because of solid organ and bone marrow transplant, HIV/AIDS, and genetic immunodeficiencies (1, 2). However, Aspergillus and Fusarium species are also major causes of corneal infections (keratitis), which occur in immune competent individuals and are a significant cause of blindness worldwide (3). Spores (conidia) enter the cornea through epithelial abrasions and germinate into filamentous hyphae. Neutrophils are the predominant cell type in early-stage corneal ulcers and in murine models of fungal keratitis (4–6), where they play an essential role in regulating hyphal growth. Previously, we reported that neutrophils combat hyphal growth in the cornea through reactive oxygen species (ROS) and iron limitation and that A. fumigatus possesses antioxidant and iron acquisition mechanisms to allow for survival (7, 8).

In addition to iron, zinc and manganese are also required for fungal growth, and an important immune defense strategy is based on sequestering these metals, termed nutritional immunity (9). Calprotectin (CP) is a heterodimer of S100A8 and S100A9, members of the S100 family of calcium binding proteins, which exhibits antimicrobial effects on bacteria, fungi, and protozoa through sequestration of zinc and manganese at two binding sites formed at the dimer interface (10–12). CP comprises ∼40% of total protein in the neutrophil cytoplasm and is also produced and secreted by other myeloid and nonmyeloid cells, including macrophages, epithelial cells, and keratinocytes under inflammatory stimuli (13–15). Neutrophil CP is also reported to mediate intracellular activities, including NADPH oxidase activation and ROS production, and cytoskeletal rearrangement (16, 17).

In the current study, we identify an essential role for neutrophil CP in regulating growth of the hyphal stage of Aspergillus in vitro and in a murine model of fungal keratitis. We also show CP-dependent chelation of both zinc and manganese is required to compete with the A. fumigatus ZafA-regulated zinc transporter system. In marked contrast, using fluorescent Aspergillus reporter (FLARE) conidia that simultaneously report phagocytic uptake and fungal viability (18), we found no role for CP in neutrophil killing of the conidia stage of A. fumigatus either in vitro or in a murine pulmonary challenge model with conidia.

Taken together, these observations identify a stage-specific role for neutrophil CP in combatting A. fumigatus infections. Given the importance of this pathogen as a cause of severe pulmonary, systemic, and corneal disease, these findings may lead to development of more targeted therapies for these infections.

All animals were used in accordance with the guidelines of the Case Western Reserve University Institutional Animal Care and Use Committee and the Memorial Sloan–Kettering Cancer Center Institutional Animal Care and Use Committee. S100A9^−/− mice were provided by P. Fidel (Louisiana State University, Baton Rouge, LA). CD18^−/− mice were originally provided by C. Doerschuk (University of North Carolina, Chapel Hill, NC). Age- and sex-matched C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME). For pulmonary challenge experiments, S100A9 and wild-type (WT) bone marrow chimeric mice were generated by reconstituting lethally irradiated (9.5 Gy) recipients (C57BL/6.SJL mice) with 2–5 × 10⁶ S100A9^(−/−) or C57BL/6 bone marrow cells and resting them for 6–8 wk prior to use in experiments. Enrofloxacin treatment for 14 d in drinking water was given to prevent bacterial infections. Animal studies were compliant with all applicable provisions established by the Animal Welfare Act and the Public Health Services Policy on the Humane Care and Use of Laboratory Animals.

Strains used in this study are identified in Table I. Strains were grown on Sabouraud dextrose agar (dsRed strains) or Vogel’s minimal medium (VMM) + 2% agar ± 0.5 mM ZnSO₄ (for Zn-deficient strains). Strains were grown at 37°C for 3–5 d for sporulation, and conidia were isolated by disruption in PBS and filtration through sterile cotton gauze.

Mice were anesthetized, the corneal epithelium was penetrated with a 30-gauge needle, and 2 μl conidial suspension (25,000 conidia/μl PBS) was injected into the corneal stroma with a 33-gauge Hamilton syringe. Mice were imaged under a stereomicroscope at 24–48 h postinfection (p.i.). Whole eyes were homogenized in sterile PBS in a Mixer Mill MM300 (Retsch). Serial dilutions were plated in Sabouraud dextrose agar (dsRed strains) or VMM + 2% agar ± 0.5 mM ZnSO₄ (Zn-deficient strains) and incubated 24 h at 37°C. CFU was counted after 24 h. dsRed expressing A. fumigatus eyes were imaged at 24–48 h p.i. and opacity (pixel intensity), and integrated intensity of fluorescence was quantified using Metamorph software as described previously (5).

In some experiments, recombinant CP or PBS (10μl) was injected into the subconjunctival space at the time of infection.

Intratracheal infections were performed as described earlier (18). Briefly, a blunt 20-gauge needle was used to deliver conidia into the trachea of anesthetized mice. Bronchoalveolar lavage (BAL) and lung tissues were processed for flow cytometry as previously described previously (19). Briefly, single-cell suspension of lungs and BAL cells were stained with the following Abs: anti-Ly6G (clone 1A8), anti-CD11b (clone M1/70), anti-CD45.1 (clone A20), anti-CD45.2 (clone 104), and anti-Ly6B.2 (clone 7/4). Neutrophils were identified as CD45⁺CD11b⁺Ly6C^loLy6G⁺Ly6B.2⁺ cells. The data were collected on a BD LSR II flow cytometer and analyzed on FlowJo, version 9.7.6 (Tree Star, Ashland, OR).

A. fumigatus strain 293 (AF293) was genetically modified to express dsRed, and FLARE conidia were labeled as described previously (18). In all experiments, percentage of uptake = (dsRed⁺AF633⁺ + dsRed⁻AF633⁺/total) and percentage of viability = (dsRed⁺AF633⁺/dsRed⁺AF633⁺ + dsRed⁻AF633⁺). For in vitro experiments, FLARE conidia were added at multiplicity of infection 2 to 1 × 10⁶ peritoneal neutrophils in RPMI 1640 medium + 5% FBS for 8 h at 37°C. Cells were washed and resuspended in FACS buffer. Data were collected on a BD Accuri C6 flow cytometer and analyzed on Accuri C6 and FlowJo software.

Mice were injected i.p. with 3% thioglycollate at 18 and 3 h prior to euthanasia. The peritoneal cavity was lavaged with sterile PBS, and neutrophils were purified by negative selection using an EasySep mouse neutrophil enrichment kit (StemCell Technologies, Vancouver, BC, Canada). Neutrophil purity of >90% was verified by Wright–Geimsa staining. Neutrophils were resuspended in RPMI 1640 medium. To obtain neutrophil lysates, purified neutrophils at 2 × 10⁶ cells/ml in RPMI 1640 were freeze-thawed at −80°C for 3 cycles. Lysates were spun at 10,000 × g for 10 min, and supernatants were collected.

Whole blood was collected from healthy donors between age 18 and 65 y in accordance with the Declaration of Helsinki guidelines and the Institutional Review Board of the University of California (Irvine, CA). RBCs were separated in 3% dextran (Sigma-Aldrich, St. Louis, MO) PBS, and neutrophils were purified from remaining cells by overlay on a Ficoll (GE Healthcare) density gradient and centrifugation at 500 × g for 25 min. Remaining RBCs were lysed, and neutrophils were resuspended in RPMI 1640 medium. Purity (>90%) was assessed by flow cytometry using anti-human CD16 and CD66b Abs (eBioscience, San Diego, CA). Live/Dead fixable violet (Thermo Fisher) was used to gate live cells, and positive populations were identified using fluorescence minus one controls.

Conidia (3000 conidia/well) were grown in VMM ± ZnSO₄ (Zn-deficient strains) for 6 h, at which time they were at the hyphal stage. Hyphae were washed twice in PBS. Live neutrophils, lysates, or recombinant CP in RPMI 1640 medium were incubated with hyphae for 16 h at 37°C. For some experiments ZnSO₄ (83265; Sigma-Aldrich) or MnSO₄ (M7899; Sigma-Aldrich) was added to media. Hyphae were stained with 50 μl Caclofluor white stain (undiluted)/well (18909; Sigma-Aldrich) and washed three times in ddH₂O. Fluorescence at 360/440 nM was assessed using a Synergy HT plate reader (Bio-Tek, Winooski, VT). Percentage of fungal mass was calculated as a percentage of total hyphae when grown under experimental conditions compared with growth in RPMI 1640 medium alone (experimental/control × 100).

Whole eyes were fixed in 10% phosphate buffered formalin, paraffin embedded and sectioned. PASH and GMS staining were performed by the Case Western Reserve University Visual Science Research Center histology core. For immunohistochemistry, sections were treated with proteinase K (2015-011; DakoCytomation Carpinteria, CA) and blocked in 1.5% serum. Mouse S100A9 polyclonal Ab (2 μg/ml, AF2065; R&D Systems, Minneapolis, MN), mouse S100A8 polyclonal Ab (2 μg/ml, AF3059; R&D Systems), or anti-mouse neutrophil Ab NIMP-R14 (20 μg/ml) were used, followed by staining with Alexa Fluor 488 chicken anti-goat IgG (1:2000, A2467; Life Technologies, Grand Island, NY) or Alexa Fluor 488 goat anti-rat IgG (1:250, A11006; Life Technologies). Slides were imaged at an original magnification of ×200–400. Neutrophil quantification from histological sections was obtained by measuring percentage of NIMP-R14 positive area/cornea using Metamorph software.

Corneas were dissected at 24–48 h p.i. and treated with 1× collagenase for 1–2 h. Cell suspensions were washed in 1× PBS + 5% FBS, incubated with anti-mouse CD16/CD32 (clone 93, 16-0161-86; eBioscience) (to block FcRs) and stained with anti-mouse neutrophil (Ly6G) Ab NIMP-R14-PE (ab125259; Abcam, Cambridge, MA) or isotype for 1 h and live/dead fixable far red stain (Thermo Fisher). For intracellular staining, cells were fixed and permeabilized using an intracellular staining kit (eBioscience). Cells were incubated with mouse S100A9 polyclonal Ab (AF2065; R&D Systems), mouse S100A8 polyclonal Ab (AF3059; R&D Systems), or isotype for 45 min, washed, and incubated with Alexa Fluor 488 chicken anti-goat IgG (1:2000, A2467; Life Technologies) for 30 min, washed, resuspended in FACS buffer, and analyzed in a BD Accuri C6 (BD Biosciences, San Jose, CA). Analysis was performed using Accuri C6 software. Cells were gated on forward light scatter/side scatter (of light), followed by live cells.

Corneas were dissected at 6–48 h p.i. and homogenized in 150 μl 1× PBS using a Mixer Mill MM300 (Retsch), and lysates were analyzed for mS100A8 (R&D DY3059), mS100A9(R&D DY2065), mCXCL1/KC(R&D DY453), mCXCL2/MIP-2(R&D DY452), or mMPO(R&D DY3667), according to the manufacturer instructions. For neutrophil lysates, purified C57BL/6 neutrophils at a concentration of 1 × 10⁶/ml were lysed in 0.5% Triton X-100 in PBS, and 10-fold serial dilutions were analyzed for mS100A8 and mS100A9.

Recombinant human WT CP and the CP∆Zn/Mn and CP∆Mn mutants were expressed, purified, and tested for activity as described previously (12).

A Mann–Whitney U test was used for unpaired comparison of two groups for all in vivo and biological replicates. In vitro experiments are shown as one representative experiment with a minimum of three technical replicates per experiment and groups compared using the Student t test. Survival data were analyzed by log-rank test. All statistical analyses were performed with GraphPad Prism software (version 6.0c; GraphPad, La Jolla, CA). A p value < 0.05 was considered significant (*p ≤ 0.05; **p ≤ 0.01; *** p ≤ 0.001).

To evaluate the role of CP in regulating A. fumigatus hyphal growth, corneas of C57BL/6 and CP-deficient S100A9^−/− mice were infected with A. fumigatus dsRed conidia, and hyphal growth was monitored at 24 and 48 h p.i., which we reported is the peak time of corneal inflammation and neutrophil infiltration (5). Fungal burden was measured by dsRed imaging and CFU.

S100A9^−/− mice had elevated dsRed fluorescence at 24 and 48 h and significantly elevated CFU at 48 h compared with C57BL/6 (Fig. 1A–C). Because CP is reported to mediate neutrophil chemotaxis (20), we examined whether S100A9^−/− mice have impaired neutrophil recruitment during infection. Neutrophils from infected corneas were quantified by flow cytometry of corneal cell suspensions, immunohistochemistry (IHC) in corneal sections and myeloperoxidase ELISA on corneal lysates. There was no significant difference in neutrophil numbers between C57BL/6 and S100A9^−/− (Fig. 1D, Supplemental Fig. 1C, 1D), indicating that the increased hyphal burden in S100A9^−/− mice is not due to impaired neutrophil migration. We also measured corneal opacity as an indicator of inflammation and found no significant difference between S100A9^−/− and C57BL/6 mice (Supplemental Fig. 1A, 1B). Furthermore, infected corneas of C57BL/6 and S100A9^−/− mice had comparable levels of the neutrophil chemokines CXCL1 and CXCL2 (Supplemental Fig. 1E, 1F), which are essential for neutrophil chemotaxis in pulmonary and cornea infection models (5, 21).

To determine whether neutrophils are the source of CP during infection, neutrophils from infected corneas were examined by flow cytometry, and adjacent corneal sections were immunostained with NIMP-R14 or with S100A8 and S100A9 Abs. S100A8 and S100A9 staining increased from 24 to 48 h and coincided with NIMP-R14 staining (Fig. 1E). Neutrophils, but not S100A8 or S100A9 were detected in corneal sections or single-cell suspension from infected S100A9^−/− mice (Supplemental Fig. 1G, 1H), which is consistent with previous reports that S100A8 is not expressed in the absence of S100A9 (22). Although S100A8 and S100A9 are also produced by epithelial cells (15), we did not detect S100A8 or S100A9 in the corneal epithelium. Total S100A8 and S100A9 proteins in infected corneas were quantified by ELISA and found to be elevated at 6, 24, and 48 h p.i. compared with corneas from naive C57BL/6 mice (Fig. 1F). As an additional indicator that infiltrating myeloid cells rather than resident epithelial cells are the source of S100A8 and S100A9 in infected corneas, we infected corneas of CD18^−/− mice, which lack the β2 subunit of CD11/CD18 integrin required for neutrophil trans-endothelial migration to the cornea, and found significantly lower levels of both proteins (Fig. 1F, Supplemental Fig. 1G) (8).

As an additional approach to investigate if neutrophils are the primary source of S100A8 and S100A9, corneal cell suspensions were stained for intracellular S100A8 and S100A9 and cell surface NIMP-R14 and examined by flow cytometry. We found that >90% S100A8⁺ and S100A9⁺ cells were also NIMP-R14⁺ (Fig. 1G, Supplemental Fig. 1H).

Finally, to ascertain if exogenous CP impairs hyphal growth during infection, S100A9^−/− mouse corneas were infected with A. fumigatus-dsRed, and 1 μg total recombinant CP was administered into the subconjunctival space simultaneously where it can diffuse into the corneal stroma. Fig. 1H and 1I shows that mice administered CP had significantly less hyphal mass than those given vehicle only, as measured by dsRed fluorescence, thereby demonstrating that CP has a direct protective effect during A. fumigatus corneal infection.

Collectively, these data indicate that neutrophil-derived CP has an essential role in regulating A. fumigatus infection in the cornea.

To examine whether neutrophil CP directly limits A. fumigatus hyphal growth, peritoneal neutrophils from C57BL/6 and S100A9^−/− mice were incubated with A. fumigatus hyphae. Fungal mass was measured following staining with Calcofluor white, which binds cell wall chitin and can be quantified by fluorimetry (8). We found that 2 × 10⁵ live neutrophils was the minimum number that significantly inhibited fungal growth in vitro (Supplemental Fig. 1I), although lysates from 1 × 10⁵ neutrophils also significantly inhibited fungal growth (Supplemental Fig. 1J). The difference is likely a consequence of partial CP release from live neutrophils compared with total CP in cell lysates. We also quantified S100A8 and S100A9 in neutrophil lysates by ELISA, and found that murine neutrophils contained ∼0.076 pg/cell (Supplemental Fig. 1K).

Hyphae grown in medium alone exhibited normal branching morphology; however, in the presence of C57BL/6 neutrophils, hyphal morphology was distinct, with short hyphal filaments and increased branching (Fig. 2A). In contrast, hyphae incubated with S100A9^−/− neutrophils resemble hyphae grown in RPMI 1640 medium alone (Fig. 2A), indicating a direct effect of CP on growing hyphae. Calcofluor quantification revealed fungal mass similar to media alone when incubated with S100A9^−/− compared with C57BL/6 neutrophils (Fig. 2B, Supplemental Fig. 1I). S100A8/A9 is reported to mediate production of ROS through NADPH oxidase, which is the dominant source of ROS and requires live cells (8, 17). Therefore, to determine whether there is a contributing role for ROS in CP-mediated hyphal killing, neutrophil lysates were incubated with hyphae. We found that as with live neutrophils, fungal mass was not reduced in relation to growth in media alone (100%) when hyphae were incubated with S100A9^−/− lysates, whereas growth was completely blocked when hyphae were incubated with C57BL/6 neutrophil lysates (Fig. 2C), indicating that neutrophil CP activity occurs independently of ROS.

Taken together with the in vivo studies, these data indicate that neutrophils have a nonredundant role in controlling A. fumigatus infection through CP-mediated inhibition of hyphal growth.

To examine if there is a direct effect of CP on A. fumigatus hyphal growth in vitro and to assess the relative contribution of Zn and Mn chelation in CP activity, WT A. fumigatus hyphae were incubated with recombinant CP in the presence of 1 or 5 μM MnSO₄ and with increasing concentrations of ZnSO₄.

Hyphae incubated with 25 μg/ml recombinant CP exhibited the short hyphal filaments and increased branching as shown above in the presence of C57BL/6 neutrophils (Figs. 3A, 2A). To determine if hyphal growth can be rescued by Zn or Mn, we measured A. fumigatus hyphal growth following incubation with CP in the presence of ZnSO₄ and MnSO₄. We found that hyphal mass in the presence of 50 μg/ml (∼1.4 μM) CP was significantly lower than hyphae grown in RPMI 1640 medium alone (Fig. 3B). As 1 mol CP can bind either 2 mol Zn or 1 mol Zn plus 1 mol Mn (12), rescue of A. fumigatus hyphal growth to 100% was found to require 4 μM ZnSO₄ when added alone, whereas 2 μM ZnSO₄ was sufficient to rescue hyphal growth in the presence of 1 or 5 μM MnSO₄. Conversely, an excess (5 μM) of Mn alone only partially rescued growth (Fig. 3B), which is likely due to Mn occupying only one of the two CP binding sites.

As a complementary approach to determine the relative contribution of Zn and Mn, we performed experiments with recombinant CP constructs with mutated binding sites for either Mn alone or for Zn and Mn (12): CP∆Mn, lacking high-affinity binding of Mn but with completely normal binding of Zn, and CP∆Zn/Mn, lacking high-affinity binding of both Zn and Mn. Fig. 3C shows that whereas hyphal mass in the presence of WT CP was ∼10% of growth in medium alone, hyphal mass in the presence of CP∆Mn was significantly higher, though still inhibitory. In contrast, hyphal growth was unimpaired when incubated with the CP∆Zn/Mn mutant, indicating complete loss of antifungal activity (Fig. 3D).

To determine if neutrophil anti-fungal activity depends on metal chelation, 2 × 10⁵ purified human neutrophils were incubated with WT A. fumigatus hyphae in vitro ± ZnSO₄ or MnSO₄ and fungal growth was measured. Neutrophils inhibited fungal growth >95%, which was reversed in the presence of ZnSO₄ (5 μM), but not MnSO₄ (5 μM) (Fig. 3E, 3F).

A. fumigatus expresses the zinc sensitive transcription factor ZafA under Zn limiting conditions, which upregulates the expression of the zinc transporters ZrfA, ZrfB, and ZrfC and the putative zinc binding protein Aspf2 (23–25). To determine whether ZafA mediated zinc uptake is necessary to compete with CP, parent strain (WT) and ∆zafA mutant A. fumigatus hyphae were incubated with a 5 × 10⁴ C57BL/6 or S100A9^−/− neutrophils or with recombinant CP, and hyphal growth was measured following Calcofluor staining.

This lower number of C57BL/6 neutrophils did not inhibit growth of the WT A. fumigatus hyphae; however, growth of ∆zafA hyphae was significantly less (Fig. 4A), indicating increased sensitivity of this mutant. In contrast, growth of ∆zafA mutants was not inhibited when incubated with the same number of S100A9^−/− neutrophils (Fig. 4A). Consistent with these data, growth of the ∆zafA mutant was significantly less than WT when incubated with recombinant CP (Fig. 4B), indicating that ZafA is required for optimal hyphal growth in the presence of CP.

To examine if ZafA is required for growth in the cornea, C57BL/6 mice were infected with WT or ∆zafA mutant conidia, and CFU/eye were assayed after 48 h. Corneas infected with ∆zafA had less hyphae in the corneal stroma than those infected with WT, as detected by GMS staining (Fig. 4C), and significantly lower CFU than WT (Fig. 4D). There was no significant difference in the number of neutrophils in the corneas of mice infected with WT versus ∆zafA A. fumigatus as quantified by NIMP-R14 positive cells (Fig. 4E, 4F). Growth of ∆zrfC, ∆zrfAB, and ∆aspf2 A. fumigatus in the cornea was not significantly different from the WT (Supplemental Fig. 2D–F). Furthermore, although there were fewer CFU in the ∆zrfABC mutant, this strain also did not grow as well as the WT or ∆zafA in vitro (Supplemental Fig. 2A–C).

Together, these data indicate that ZafA is essential for virulence during corneal infection, and that ZafA-dependent zinc uptake competes with CP mediated zinc sequestration.

In addition to metal chelation, S100A8 and S100A9 have been shown to promote neutrophil phagocytosis and intracellular killing of bacteria by ROS (26). Therefore, to examine the role of CP in neutrophil phagocytosis and conidial killing in vitro, peritoneal neutrophils from C57BL/6 or S100A9^−/− mice were incubated with FLARE conidia as described previously (18). FLARE conidia express dsRed protein, which acts as an indicator of live conidia. FLARE conidia also incorporate Alexa Fluor 633 extracellularly, which persists even after loss of conidia viability (18). Thus, FLARE conidia can be utilized to monitor conidial uptake and viability in phagocytes as the presence of live (DsRed⁺Alexa Fluor 633⁺) conidia can be distinguished from counterparts that contain killed (DsRed⁻Alexa Fluor 633⁺) conidia (18, 27). Representative images of neutrophils containing live or dead FLARE conidia, in addition to bystander neutrophils, are shown in Supplemental Fig. 3B.

Neutrophils were incubated with FLARE conidia for 8 h, and the total and viable conidia were quantified by flow cytometry. There was no significant difference in either conidial uptake or viability between C57BL/6 and S100A9^−/− neutrophils (Fig. 5A, 5B), indicating that CP is not required for intracellular conidial killing in vitro.

To examine whether CP expression in hematopoietic cells is required for host defense against A. fumigatus pulmonary infection, we generated CD45.2⁺S100A9^−/− → C57BL/6.SJL bone marrow chimeric mice, together with CD45.2⁺S100A9^+/+ → C57BL/6.SJL controls. Both groups of chimeric mice were infected with FLARE conidia by the intratracheal route, and at 12 h p.i., single-cell suspensions from BAL fluid (BALF) and from whole lungs were assessed for conidial uptake and killing. Phagocytosis and viability were indistinguishable between S100A9^−/− neutrophils and control neutrophils (Fig. 5C, 5D). In addition, total S100A9^−/− and S100A9^+/+ neutrophil counts were similar in both groups (Fig. 5E), and mice that lack S100A9 in hematopoietic cells survived as well as control mice (Supplemental Fig. 3A).

Thus, CP expression in radiosensitive hematopoietic cells, including neutrophils, is dispensable for host defense against A. fumigatus conidia. Collectively, these data demonstrate that CP acts in a fungal stage-specific manner to restrict extracellular A. fumigatus hyphal growth and is dispensable for intracellular conidial killing.

Nutritional immunity is an emerging concept that addresses the importance of nutrient regulation by the immune system as a means of controlling pathogen growth and survival (28, 29). Essential nutrients required by pathogens include amino acids, lipids, and transition metals such as iron, zinc, manganese, and copper. For example, iron is an essential cofactor for many cellular processes (30), and we reported that neutrophils regulate A. fumigatus hyphal growth by sequestering iron and that treating infected corneas with iron-binding protein lactoferrin or inhibiting A. fumigatus iron-binding siderophore synthesis with HMG-CoA reductase inhibitors (statins) reduced fungal burden (7). In the current study we extend our understanding of nutritional immunity in A. fumigatus infection to include a role for neutrophil CP.

CP is an abundant cytosolic protein in neutrophils (13), and we found that neutrophils are the primary source of CP during infection of the cornea. Previous work has shown that neutrophils are essential for control of A. fumigatus keratitis in mice (8). We did not observe S100A8/A9 staining in corneal epithelium, despite a report of epithelial S100A8/A9 production in a Pseudomonas keratitis model (31). Another report of A. fumigatus lung infection also showed that leucopenic mice had very low CP expression, indicating that neutrophils are the primary source of CP (32). The high levels of S100A8 and S100A9 protein in infected corneas is consistent with reports that CP can reach levels as high as 1 mg/ml in abscess fluid (10). The ELISA may have underestimated the concentration of CP in the tissue and in neutrophils because of the ability of S100A8 and S100A9 to form higher-order oligomers, which likely still have Zn- and Mn-chelating activity (33). We also demonstrated that CP is essential for control of A. fumigatus growth in the cornea but not in a conidia pulmonary challenge model. Furthermore, we compared the ability of CP-deficient S100A9^−/− neutrophils to inhibit hyphal growth with their ability to kill A. fumigatus conidia using FLARE conidia and found that whereas S100A9^−/− neutrophils could not limit hyphal growth, there was no deficiency in their ability to kill intracellular conidia. This is in contrast to reports that S100A8/A9 promotes phagocytosis and ROS production in response to Escherichia coli and zymosan (16, 26). Our findings describe CP as a specific regulator of extracellular A. fumigatus hyphal growth. Previous work identified neutrophil ROS and iron-chelating enzymes as essential in antifungal responses during keratitis (7, 8). Furthermore, neutrophils contain numerous antimicrobial peptides and enzymes including defensins, cathelicidin, proteases, and chitinases, which can mediate killing and degradation of fungal pathogens (34, 35). Therefore, CP likely inhibits growth early in infection, which complements other antifungal mechanisms.

Because the corneal stroma is avascular, infiltrating cells are recruited from limbal vessels in the peripheral cornea and migrate through the dense corneal stroma to the site of infection. Thus, neutrophils do not reach the site of infection until hyphae are formed. In contrast, during A. fumigatus pulmonary infection neutrophils are rapidly recruited to the highly vascularized lung, and rapidly phagocytose conidia (19). Thus, in the cornea, antihyphal rather than anti-conidia responses are essential for control of infection.

These results also fit well with the recent finding that neutrophil extracellular traps (NETs) are generated specifically in response to large extracellular pathogens such as fungal hyphae (36). The antifungal activity of NETs has been attributed to CP in Candida albicans dermal abscess and intranasal infection models and in vitro in response to Aspergillus nidulans (37, 38). NETs have been identified in lungs of A. fumigatus–infected mice (39, 40); however, rather than using resting conidia, which are normally inhaled and phagocytosed (and were used in the current study), in those studies, the lungs were instilled with conidia that had been previously incubated for 7 h or with hyphal fragments. Although we are investigating NET formation during fungal corneal infection, those studies and ours indicate that CP, and therefore, CP-laden NETs are important to regulate growth of extracellular hyphae rather than conidia. Interestingly, Urban et al. (37) reported that only ∼30% of CP was contained within NETs, whereas the remaining CP is found in the supernatant or associated with cellular debris, therefore demonstrating NET-independent secretion of CP. In activated monocytes, S100A8 and S100A9 proteins were shown to be released in a nonclassical, energy-dependent manner by associating with microtubules (41). Further studies on CP release by neutrophils and other cell types are warranted.

Several properties have been reported for CP, including direct chemotactic activity and chemoattractant response in neutrophils (20, 42). However, we found no difference in the number of neutrophils recruited to infected corneas of S100A9^−/− mice or in S100A9^−/− neutrophil migration to lungs and no significant difference in corneal opacity between C57BL/6 and S100A9^−/− mice. This observation is consistent with a recent report demonstrating that, although S100A8/A9 can induce neutrophil chemotaxis, its absence does not affect neutrophil recruitment in vivo in a murine model of vaginal candidiasis (15). Furthermore, CP is reported to promote phagosomal ROS production in response to zymosan particles or E. coli (16, 26); however, we found that S100A9^−/− neutrophils had no impaired ability in phagocytosis or cytotoxic activity against conidia. In contrast, lung neutrophils that are defective in NADPH oxidase activity show a profound defect in conidial killing when mice are infected intratracheally with the FLARE conidia (18). In the current study, regulation of hyphal growth by calprotectin does not require NADPH oxidase because our findings with live neutrophils were replicated using neutrophil lysates. Therefore, CP appears to have no role in either neutrophil recruitment or in ROS production.

Instead, we show that CP antifungal activity depends on chelating Zn and Mn. Addition of Zn or Zn and Mn was found to reverse CP-mediated Aspergillus antifungal activity in vitro (32). The current study extends these findings using CP mutants that lack the ability to bind either Zn and Mn or Mn only to show the relative contribution of each of these transition metals. We found that a CP∆Zn/Mn mutant had no antifungal activity, whereas a CP∆Mn mutant had only partial loss of activity. Furthermore, Zn alone rescued fungal growth in the presence of CP or human neutrophils whereas Mn alone did not. These data indicate that Zn chelation is essential for antifungal activity, whereas Mn appears to have a contributing role. This is in contrast to bacteria in which Mn binding has an essential role in the antimicrobial activity of CP (12, 43). CP-mediated Mn chelation has been shown to inhibit superoxide dismutase activity and reduce the virulence of Staphylococcus aureus (44). Zn is required for the function of numerous enzymes and transcription factors in A. fumigatus, whereas Mn function is not well characterized, although it is required for superoxide dismutase function (45).

A. fumigatus growth in a low zinc environment requires a Zn uptake system comprising the Zn-sensitive transcription factor ZafA, which is required for expression of zinc transporters ZrfA, ZrfB, and ZrfC and the putative Zn binding protein Aspf2 (23–25, 32). ZafA and ZrfC are required for virulence in a pulmonary aspergillosis model in which mice were immunosuppressed (24, 32). This zinc transporter system is also important in C. albicans and Cryptococcus gatti, where it is required for yeast survival in vivo (46, 47). In the current study, we show that ZafA is also essential for virulence during corneal infection and that ZafA-deficient A. fumigatus strain has increased susceptibility to inhibition by neutrophils and CP. In contrast to pulmonary infection, we found no role for ZrfC in vivo during corneal infection. The difference between these and previous findings has yet to be determined; however, the cornea has a higher Zn content than other tissues, including the lung and the blood (48); therefore ZrfA and ZrfB may compensate for ZrfC in this setting as ZrfA and ZrfB expression is elevated in the ∆zrfC mutant (32). A ZrfABC-deficient strain also exhibited less virulence in vivo and significantly impaired growth in vitro. The difference in growth of this mutant compared with ZafA-deficient A. fumigatus may be due to low level constitutive (ZafA-independent) expression of Zrf transporters on the ZafA mutant, which are completely absent in ZrfABC-deficient strain (49). Because of its importance in both pulmonary and corneal infections, ZafA may be a target for antifungal therapies (50).

Whereas the A. fumigatus Zn transport system has been characterized, specific Mn transporters have yet to be identified in A. fumigatus. However, a recent report showed that A. fumigatus siderophores that bind iron can also bind Mn (51).

The zinc transport system is important in A. fumigatus survival in the lungs, although our current findings using the FLARE viability assay showed that there was no significant difference in viable conidia in neutrophils from infected lungs of S100A9^−/− bone marrow chimeric mice compared with C57BL/6 mice, which taken together with in vitro assays indicate that CP is not required for host survival in a pulmonary challenge model in which the immune system is otherwise intact. CP is likely also important in the immunosuppressed condition where hyphae are present, as leucopenic mice have reduced CP expression in the lung and increased fungal burden (32), and patients who develop Aspergillus pulmonary infections are often either neutropenic or have neutrophil dysfunction (52).

Interestingly, there is also a CP-independent mechanism of zinc sequestration that targets intracellular pathogens. Macrophages can limit zinc uptake by fungal pathogen Histoplasma capsulatum by sequestering metals from the phagosome using the transporters Slc30a4 and Slc30a7 and metallothioneins (53). It is unclear whether this mechanism is relevant in A. fumigatus conidial killing in the phagosome; however, because conidia are dormant compared with H. capsulatum yeast, zinc sequestration may not influence conidia survival.

We found that recombinant human CP inhibited Aspergillus growth at concentrations as low as 6.25 μg/ml. In vitro assays with murine neutrophils required >1 × 10⁵ C57BL/6 neutrophils to inhibit growth similarly. Human neutrophils were found to contain an average of 25 pg CP/neutrophil (54), whereas we found that murine neutrophils contained only ∼0.076 pg/cell. Importantly, although human and mouse S100A8 and S100A9 share 59% protein sequence identity, minimum inhibitory concentrations may differ. The high level of sequence identity and the identical metal ligand residues suggest the Zn and Mn affinities are the same; however, dissociation constants have only been measured for recombinant human CP (12). Furthermore, release of CP from live neutrophils is likely <100% and indeed, a lysate from 1 × 10⁵ neutrophils effectively inhibited Aspergillus growth (37).

Treatment of Aspergillus infections is challenging because of the limited range of antifungal agents with variable efficacy and significant associated toxicity. Results from the current study demonstrate that exogenous CP controls fungal growth in S100A9^−/− mice. Therefore, treatment with CP or new inhibitors of fungal Zn and Mn transport could represent a new therapeutic strategy for fungal infection, particularly in neutropenic patients (50).

We thank the Case Western Reserve University Visual Science Research Core, particularly Catherine Doller, Denice Major, and Scott Howell, for excellent technical assistance. We thank Paul Fidel (Louisiana State University) for providing the S100A9^−/− mice. T.M.H. and A.J. thank Karin Bornfeldt (University of Washington) for the S100A9^−/− bones used to make the chimaeras for the lung experiments.

The authors have no financial conflicts of interest.