Human Dectin-1 deficiency impairs macrophage-mediated defense against phaeohyphomycosis

Subcutaneous phaeohyphomycosis typically affects immunocompetent individuals following traumatic inoculation. Severe or disseminated infection can occur in CARD9 deficiency or after transplantation, but the mechanisms protecting against phaeohyphomycosis remain unclear. We evaluated a patient with progressive, refractory Corynespora cassiicola phaeohyphomycosis and found that he carried biallelic deleterious mutations in CLEC7A encoding the CARD9-coupled, β-glucan–binding receptor, Dectin-1. The patient’s PBMCs failed to produce TNF-α and IL-1β in response to β-glucan and/or C. cassiicola. To confirm the cellular and molecular requirements for immunity against C. cassiicola, we developed a mouse model of this infection. Mouse macrophages required Dectin-1 and CARD9 for IL-1β and TNF-α production, which enhanced fungal killing in an interdependent manner. Deficiency of either Dectin-1 or CARD9 was associated with more severe fungal disease, recapitulating the human observation. Because these data implicated impaired Dectin-1 responses in susceptibility to phaeohyphomycosis, we evaluated 17 additional unrelated patients with severe forms of the infection. We found that 12 out of 17 carried deleterious CLEC7A mutations associated with an altered Dectin-1 extracellular C-terminal domain and impaired Dectin-1–dependent cytokine production. Thus, we show that Dectin-1 and CARD9 promote protective TNF-α– and IL-1β–mediated macrophage defense against C. cassiicola. More broadly, we demonstrate that human Dectin-1 deficiency may contribute to susceptibility to severe phaeohyphomycosis by certain dematiaceous fungi.


Results
Clinical presentation. A 64-year-old African American male living in Louisiana, United States, first presented to a local hospital in 1984 (at age 27) with a nonhealing growth on his nose following an injury he sustained while working in construction (22). Biopsy revealed mixed yeast and melanin-pigmented fungal filaments, leading to the diagnosis of phaeohyphomycosis, although identification at the fungal species level was not possible. The patient had a normal peripheral white blood cell count and differential and unremarkable chemistries. He was treated with oral ketoconazole (200 mg/day) for 9 months with some improvement, but the infection recurred within 6 weeks of stopping therapy. Three additional months of ketoconazole failed to improve the nasal lesion, prompting a dose increase to 400 mg for 4 more months, which led to marked improvement (22). However, the patient experienced several infection relapses over the subsequent 18 years. In 2004 (at age 47), he presented to an outside hospital with severe facial ulcerations. Despite escalation of antifungal therapy to itraconazole, amphotericin B, and caspofungin, and multiple debridements, the infection progressed, and the patient was transferred to the NIH Clinical Center. He had no other medical conditions and no prior history of infections, including mucocutaneous candidiasis, dermatophytosis, or onychomycosis. His family history was noncontributory. Physical examination revealed extensive ulcerative and necrotic facial lesions ( Figure 1A). Laboratory studies revealed leukocytosis with a white blood cell count of 30,000 cells/μL, of which 89% were neutrophils, 2.8% monocytes, 7.7% lymphocytes, and 0.5% eosinophils. Serum immunoglobulin (including IgE) levels were within normal limits. Biopsy of the facial lesions demonstrated granulomatous inflammation with plasma cells, few neutrophils, and rare eosinophils, with fungal elements engulfed, but not destroyed, within macrophages (Figure 1B and Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/JCI159348DS1). Fungi grew as sterile mycelia and were identified molecularly as C. cassiicola (Supplemental Figure 2). The minimal inhibitory concentrations for amphotericin B, itraconazole, posaconazole, caspofungin, and terbinafine were low at 0.125, 0.06, 0.06, 0.125, and 0.06 μg/mL, respectively. The facial lesions were extensively debrided. While receiving posaconazole and caspofungin, a cerebellar lesion was noted, and biopsy demonstrated neutrophilic microabscesses and granulomatous inflammation and fungal elements. 5-Flucytosine, terbinafine, and subcutaneous IFN-γ were added and the patient eventually improved. Multiple reconstructive surgeries and skin grafting procedures were performed over the subsequent year. The patient has not developed any recurrences over the past 16 years while receiving itraconazole secondary prophylaxis ( Figure 1A).

Introduction
Phaeohyphomycosis is an invasive fungal infection caused by dematiaceous fungi, which are characterized by melanin production and filamentous growth. Phaeohyphomycosis typically affects the subcutaneous tissues following traumatic inoculation and is treatable with antifungal therapy and/or surgical resection. Phaeohyphomycosis may also manifest as cerebral or disseminated disease in immunosuppressed patients or, infrequently, in putatively immunocompetent individuals, and its prognosis is usually poor (1,2). Better understanding of the pathogenesis of phaeohyphomycosis may improve the management and outcome of affected patients.
has been associated with impaired Dectin-1-dependent cellular responses (19,20,24,25). The other allele had the missense mutation c.668T>G (p.I223S), which is predicted to be deleterious (PolyPhen-2 score 1.0 [http://genetics.bwh.harvard.edu/pph2/. Grocott's methenamine silver-stained (GMS-stained, left) and hematoxylin and eosin-stained (H&E-stained, right) section of soft tissue biopsy demonstrating C. cassiicola being engulfed but not destroyed in macrophages (black arrowheads) within granulomas (×100 magnification shown in Supplemental Figure 1). Scale bars: 20 μm. (C) Chromatograms from CLEC7A sequencing on healthy control and our patient, over the site of mutation in each allele. (D) Representative FACS histograms showing Dectin-1 surface expression in our patient and 2 healthy control patients. Histograms were gated on CD14 + monocytes isolated from peripheral blood. (E) Representative protein immunoblot images of Dectin-1 expression in PBMCs from our patient and a healthy volunteer (HV). β-Actin was used as loading control. (F) TNF-α production by PBMCs after 48 hours of stimulation with either purified particulate β-glucan or α-mannan. Each data point represents an individual well; at least 2 separate blood draws were analyzed in the Dectin-1-deficient patient (each tested in 2-3 technical replicates) and compared to 2 different healthy controls (each tested in 2-3 technical replicates). A single blood draw from each of the CARD9-deficient patients was analyzed in 3-5 technical replicates. Data in panel F were analyzed by 2-way ANOVA with Bonferroni's correction. ***P < 0.005, ****P < 0.0001. NS, not significant. (G) Photograph of a previously reported CARD9-deficient patient (CARD9.02) (21) at presentation at the NIH (age 12). (H) Volume rendering of computed tomography data of patient CARD9.02 at age 12 emphasizing bone, which reveals erosions of the frontal bone (black arrows) and loss of maxillofacial structures (white arrow), including the hard palate, resulting in a common oronasal cavity. (I) Parasagittal T1-weighted magnetic resonance imaging of patient CARD9.02 at age 12 obtained following i.v. gadolinium-based contrast agent administration, which reveals epidural abscess with adjacent cerebritis (black arrow) and tissue loss (white arrow) resulting in a common cavity encompassing the nasopharynx, oropharynx, nasal cavity, oral cavity, and portions of the paranasal sinuses. (J) GMS-(upper) and H&Estained (lower) section of soft tissue biopsy demonstrating granulomatous inflammation with C. cassiicola engulfed within macrophages (accompanying images from brain biopsy shown in Supplemental Figure 4). Both images are from consecutive cuts of the same biopsy sample. Scale bars: 20 μm. tal material). Therefore, p.Y238* does not decrease Dectin-1 protein production but it alters its extracellular β-glucan-binding C-terminal domain. To understand the functional consequences of these mutations, we stimulated PBMCs from the patient and healthy volunteers with purified β-glucan (a Dectin-1 ligand; ref. 26) or α-mannan (a Dectin-2 ligand; ref. 27) and measured TNF-α production, since phaeohyphomycosis is a reported complication of TNF-α-targeted biologics (28). The patient's cells had blunted TNF-α production in response to β-glucan but not in response to α-mannan, which does not bind Dectin-1 ( Figure 1F). As controls, PBMCs from 2 CARD9-deficient patients, one with invasive candidiasis (CARD9.1; ref. 16) and another with C. cassiicola phaeohyphomycosis (CARD9.2; ref. 21), did not respond to either β-glucan or α-mannan ( Figure 1F).
We examined the consequences of these CLEC7A mutations for Dectin-1 expression. The patient's monocytes had significantly reduced Dectin-1 surface expression by flow cytometry with an antibody that recognizes the extracellular, β-glucan-binding, C-terminal domain where the Dectin-1 mutations reside ( Figure  1D and Supplemental Figure 3). By contrast, Dectin-1 expression in PBMCs was normal by immunoblot analysis with an antibody that recognizes the extracellular stalk region of Dectin-1, which is unaffected by the patient's CLEC7A mutations ( Figure 1E   soluble Dectin-1 molecules (29). Indeed, C. cassiicola had Dectin-1 binding sites at filamentous branching points and leading ends, where conidia are released and bud scars form ( Figure 2A). We next examined the dependence on Dectin-1 for mounting proinflammatory cytokine responses to this fungus by stimulating PBMCs from the Dectin-1-deficient patient with C. cassiicola. PBMCs from CARD9-deficient patients and heathy volunteers served as controls. PBMCs from the Dectin-1-deficient patient produced significantly less TNF-α and IL-1β in response to C. cassiicola ( Figure 2B), similar to CARD9-deficient PBMCs ( Figure  2C). By contrast, responses to the Toll-like receptor (TLR) agonists LPS and Pam3CSK4 were normal, confirming selective impairment of antifungal responses (Supplemental Figure 6). Collectively, these data indicate that Dectin-1 binds and mediates proinflammatory cytokine responses to C. cassiicola.
Dectin-1 and CARD9 deficiencies predispose to subcutaneous C. cassiicola phaeohyphomycosis in mice. Human CARD9 deficiency underlies progressive subcutaneous C. cassiicola phaeohyphomycosis (4,21). We modeled subcutaneous C. cassiicola phaeohyphomycosis in mice to determine whether Dectin-1 is important for defense against this fungus in a controlled experimental setting. Following subcutaneous fungal inoculation, wild-type (WT) mice exhibited swollen footpads during the first week, which resolved over the next 2-3 weeks ( Figure 3A). Patient CARD9.2 initially developed subcutaneous C. cassiicola phaeohyphomycosis at age 4 in Colombia (21) and was admitted to the NIH Clinical Center at age 12. She had extensive, uncontrolled disease involving subcutaneous tissues of her face, adjacent bones, and brain ( Figure 1, G-I). As with our index patient, biopsy of infected lesions demonstrated ineffective granulomatous inflammation, with fungal elements engulfed within macrophages ( Figure 1J and Supplemental Figure 5). Over a 13-month hospital admission, the patient developed progressive fatal infection despite treatment with amphotericin B, posaconazole, micafungin, and terbinafine, repeated surgical debridements, and allogeneic hematopoietic stem cell transplantation. Thus, our index patient is compound heterozygous for CLEC7A mutations, which cause functional Dectin-1 deficiency associated with production of a protein with an altered extracellular C-terminal domain. Mechanistically, our Dectin-1-deficient patient is clinically and histologically similar and similarly defective in TNF-α production to a CARD9-deficient patient who developed fatal C. cassiicola phaeohyphomycosis.
Dectin-1 binds and mediates proinflammatory cytokine responses to C. cassiicola. We further investigated the immunological relevance of Dectin-1 deficiency in our index patient. We asked whether Dectin-1-sensing pathogen-associated molecular patterns were present on the C. cassiicola surface by staining with Card9 -/mice exhibited footpad swelling that was significantly greater than that of WT mice ( Figure 3, A and B). We were unable to quantify live fungal organisms from the infected footpads by culture due to the mycelial growth of the fungus impeding CFU-based quantification. We therefore first used histological analysis, which revealed large granulomas with neutrophils surrounded by lymphocytes and histiocytes and significantly increased C. cassiicola invasion within granulomas of Card9 -/mice. The increased presence of fungal organisms was confirmed by fungal qPCR (Figure 3, B-D). Notably, Clec7a -/mice also exhibited significantly greater footpad swelling, C. cassiicola invasion within granulomas, and footpad fungal burden by , and Card9 -/-(n = 7) mice infected in the hind footpad with 5 × 10 6 CFU of C. albicans SC5314. Data in panels E-G were pooled from 2 independent experiments and analyzed by 2-way ANOVA with Bonferroni's correction. Fungal burden data in panel G were analyzed by Mann-Whitney U test. *P < 0.05; **P < 0.01; ***P < 0.005.
qPCR compared with WT mice, albeit to a lesser extent than Card9 -/mice ( Figure 3, A-D). Therefore, Dectin-1 is necessary for defense against experimental C. cassiicola infection.
To define which immune cells mediate protective immunity during subcutaneous C. cassiicola phaeohyphomycosis, we infected monocyte/macrophage-specific CARD9-deficient Card9 fl/fl Cx3cr1 CreER mice (15) and found that CARD9 deficiency in monocytes/macrophages enhanced infection susceptibility relative to WT controls ( Figure 3E). By contrast, Rag1 -/mice that lack lym-phocytes were not susceptible ( Figure 3F). Thus, monocyte/macrophage-dependent responses protect, while adaptive immunity is dispensable in this model.
Next, we examined the fungal specificity of the Dectin-1 and CARD9 dependence for defense during subcutaneous infection. Our patient did not manifest mucocutaneous candidiasis and Cle-c7a -/mice control oropharyngeal candidiasis normally (30). During subcutaneous candidiasis, Card9 -/mice were susceptible, showing increased footpad swelling and fungal burden compared with WT and Card9 -/mice (n = 6), normalized to the WT controls (n = 6). Each data point represents an individual mouse; data were pooled from 2 independent experiments and analyzed by 1-way ANOVA with Dunnett's correction. (E) Results of an in vitro C. cassiicola killing assay with bone marrow-derived macrophages, prestimulated for 24 hours with TNF-α, IL-1β, or both. Killing was determined by measuring β-D-glucan levels in the culture supernatant. Bar graph shows the mean ± SEM for 2 independent experiments; overlaid dot plot shows technical replicates from one of these experiments. (F) Schematic representation of the proposed model of anti-C. cassiicola immunity in the footpad of WT, Dectin-1-deficient, and CARD9-deficient mice. Data in panels A and B were analyzed by 2-way ANOVA with Bonferroni's correction. Data in panel E were analyzed by unpaired, 2-tailed t test. *P < 0.05; **P < 0.01; ***P < 0.005; ****P < 0.0001.
By contrast, the levels of IL-1β and TNF-α were significantly decreased within infected footpad homogenates of Clec7a -/and Card9 -/mice compared with WT mice ( Figure 4A). To determine whether decreased IL-1β and TNF-α production during infection may contribute to the impaired protection seen in Clec7a -/and Card9 -/mice, we infected IL-1 receptor type 1-(IL-1R1-), IL-1β-, and TNF-α-deficient mice and found them to exhibit increased footpad swelling after C. cassiicola infection ( Figure 4B). Collectively, these data indicate that IL-1β and TNF-α contribute to protection during experimental C. cassiicola phaeohyphomycosis, and their production depends on Dectin-1/CARD9 signaling in mice and humans.
IL-1β and TNF-α enhance macrophage killing of C. cassiicola. Since macrophages, IL-1β, and TNF-α contribute to anti-C. cassiicola defense, we examined how these cytokines may promote protective macrophage-dependent immunity. Macrophage accumulation did not differ in C. cassiicola-infected subcutaneous tissue between WT, Clec7a -/-, and Card9 -/mice ( Figure 4C). Similarly, Dectin-1 and CARD9 were dispensable for neutrophil accumulation in C. cassiicola-infected subcutaneous tissues ( Figure 4C). Thus, CARD9 promotes neutrophil recruitment in a fungus-and tissue-specific manner, as it mediates neutrophil recruitment in C. albicans-infected brain and Phialophora-infected subcutaneous tissues (15,18). Notably, although macrophage accumulation was unaffected, macrophage production of TNF-α and IL-1β was significantly decreased in the absence of Dectin-1 or CARD9 in C. cassiicola-infected subcutaneous tissue ( Figure 4D). We hypothesized that TNF-α and IL-1β directly enhance the fungal killing capacity of macrophages, since these cytokines are known to enhance macrophage antimicrobial killing pathways (35), and mice. By contrast, Clec7a -/animals controlled subcutaneous candidiasis similarly to WT ( Figure 3G). Therefore, mouse Dectin-1 deficiency confers susceptibility to C. cassiicola but not Candida albicans subcutaneous infection, reflecting the fungus-specific susceptibility of our Dectin-1-deficient patient to C. cassiicola infection.
Dectin-1-and CARD9-dependent IL-1β and TNF-α contribute to protection during subcutaneous C. cassiicola phaeohyphomycosis. We next asked which cytokines protect during C. cassiicola phaeohyphomycosis and their dependence on Dectin-1/CARD9 signaling for production in vivo. We focused on IL-1α, IL-1β, TNF-α, and IFN-γ, whose production was abrogated in our index patient's PBMCs upon C. cassiicola stimulation, and on IL-17A, which was produced normally by the patient's PBMCs but promotes mucocutaneous antifungal immunity in other settings ( Figure 2B and Supplemental Figure 7) (31,32). The levels of IL-1α, IFN-γ, and IL-17A were not decreased within infected footpad homogenates of Clec7a -/and Card9 -/mice relative to WT mice ( Figure 4A). Although IFN-γ levels were similar, we examined its role further since recombinant IFN-γ had been administered as part of our index patient's successful treatment. We found normal IFN-γ production by CD4 + T cells and intact expression of IFN-γ receptor 1 (IFN-γR1) by myeloid cells in the footpads of Clec7a -/and Card9 -/mice relative to WT mice (Supplemental Figure 8, A and B). Moreover, we found similar fungal burdens in WT and Ifng -/mouse footpads (Supplemental Figure 8C), collectively indicating that IFN-γ is dispensable for control of C. cassiicola infection in this model. These data are consistent with the lack of reported phaeohyphomycosis in patients with inherited IFN-γR deficiencies, neutralizing autoantibodies against IFN-γ, or receiving the IFN-γtargeting monoclonal antibody emapalumab (23, 33, 34). (A) Relative frequency of deleterious CLEC7A mutations with high CADD scores (>20) at the population level (calculated using data from 1000 Genomes) and in the index patient and 17 additional unrelated patients with severe forms of phaeohyphomycosis who were enrolled consecutively over an 8-year period at the NIH (see Table 1 for details). Data were analyzed by Fisher's exact test. ****P < 0.0001. (B) Mean fluorescence intensity (MFI) values for Dectin-1 surface expression in CD14 + monocytes isolated from 8 healthy donors (black dots, see Supplemental Table 1 for details), 5 phaeohyphomycosis patients without CLEC7A mutations (gray dots), and 6 phaeohyphomycosis patients carrying CLEC7A mutations (green, red, and blue dots as indicated). FACS staining was performed using an antibody that targets the C-terminus where the CLEC7A mutations reside. No PBMCs were available for testing in the other patients with severe phaeohyphomycosis, including the one carrying the c.547C>T (p.Leu183Phe) variant (Table 1). FACS histograms show representative Dectin-1 staining for healthy control (black line), a patient homozygous for p.Y238* (red line), and patients heterozygous for p.Y238* (green line) or p.I223S (blue line) relative to isotype staining control (filled gray histogram). (C) TNF-α production by PBMCs stimulated with particulate β-glucan for 48 hours, from the same patients as shown in panel B. One blood draw was tested per healthy donor or patient in 3-12 technical replicates and the mean value per individual is depicted. Data were analyzed by 1-way ANOVA with Dunnett's correction. *P < 0.05, **P < 0.01. NS, not significant. J Clin Invest. 2022;132(22):e159348 https://doi.org/10.1172/JCI159348 fungal killing. Moreover, in an additional 12 out of 17 unrelated patients with severe forms of phaeohyphomycosis, we identified heterozygous or biallelic deleterious CLEC7A mutations associated with an altered Dectin-1 extracellular β-glucan-binding domain, impairing TNF-α production in response to fungal stimulation. Collectively, these data underscore the importance of Dectin-1 in antifungal immunity to phaeohyphomycetes.
C. cassiicola causes multiple diseases in plants (39). It also rarely causes human phaeohyphomycosis, affecting diabetic (40), CARD9-deficient (4,21,41), and apparently immunocompetent patients (42)(43)(44). Human Dectin-1 deficiency caused by p.Y238* homozygosity was previously associated with vulvovaginal candidiasis and onychomycosis (19). This susceptibility was linked to impaired Dectin-1-mediated recognition of Candida and decreased production of IL-17A and TNF-α (19). The p.Y238* mutation has also been associated with increased risk of mucosal Candida colonization and invasive aspergillosis in the setting of allogeneic hematopoietic stem cell transplantation, linked to impaired production of proinflammatory cytokines by PBMCs of heterozygous and homozygous patients (20,38). We show here that whereas p.Y238* Dectin-1 protein is produced at adequate levels, it has an altered extracellular, β-glucan-binding, C-terminal domain, which could affect the innate recognition of C. cassiicola. We found that the C-terminal domain of Dectin-1 bound C. cassiicola at branching buds and hyphal tips, representing one of the first insights to our knowledge into innate recognition of this specific fungus. Whether other CARD9-coupled CLRs, TLRs, or other pattern recognition receptors can engage ligands within the C. cassiicola cell wall and initiate protective immunity, either alone or collaboratively, remains to be determined. Based on our data, we propose that human Dectin-1 deficiency contributes to severe C. cassiicola phaeohyphomycosis, and that deleterious CLEC7A (and CARD9) mutations should be sought in affected patients.
The rarity of C. cassiicola phaeohyphomycosis is notable relative to the frequency of human Dectin-1 deficiency, as approximately 5 and approximately 0.1 out of 1,000 individuals are homozygous for p.Y238* and p.I223S, respectively. Therefore, these deleterious CLEC7A variants in themselves do not represent a primary immunodeficiency per se. By contrast, these population CLEC7A variants are relatively benign unless the carrier is traumatically exposed to C. cassiicola or certain other phaeohyphomycetes to instigate disease, after which impaired Dectin-1-dependent immune responses appear to contribute to suboptimal infection control. The contribution of common population variants to infection susceptibility has been previously shown. For example, common heterozygous CFTR mutations are associated with increased risk of respiratory infections (45,46). Furthermore, the common homozygous CCR5Δ32 mutation (~1% of Whites) increases the risk of severe West Nile virus infection upon exposure to the virus by mosquitos (47)(48)(49). Moreover, the common homozygous TYK2 P1104A mutation (~1 out of 600 of Europeans) increases the risk of tuberculosis upon exposure to Mycobacterium tuberculosis (50). Indeed, most cases of human C. cassiicola infections are reported to develop in the setting of physical injuries, as with our index patient (42). In addition, the low virulence of C. cassiicola associated with its marked growth restriction at temperatures greater than 30°C (39) and its preferential habitat on crop plants and debris (51) further limits the avail-we had observed poor containment of fungal growth within macrophages in the soft tissue biopsy of our index patient ( Figure 1B and Supplemental Figure 1). Indeed, stimulation of macrophages with both IL-1β and TNF-α, but not either cytokine alone, significantly increased C. cassiicola killing compared with unstimulated cells ( Figure 4E). Of note, macrophage killing of C. cassiicola did not depend on ROS, since blocking the production of ROS did not impair macrophage killing of the fungus in vitro (Supplemental Figure 9). Together, our data support a model in which Dectin-1mediated C. cassiicola recognition by CARD9-expressing macrophages promotes TNF-α and IL-1β production, which enhances macrophage nonoxidative fungal killing during experimental C. cassiicola phaeohyphomycosis ( Figure 4F).
Deleterious CLEC7A mutations associated with impaired Dectin-1 responses are frequent in patients with severe phaeohyphomycosis. To examine the potential broader implications of impaired Dectin-1 responses in human phaeohyphomycosis, we evaluated 17 other unrelated, putatively immunocompetent individuals with severe forms of phaeohyphomycosis who were enrolled consecutively over an 8-year period at the NIH Clinical Center. We found CLEC7A mutations with high CADD scores (>20) in 12 of them (70.9%) ( Table 1). Combined with our index patient with C. cassiicola infection, 13 out of 18 (72.2%) patients with severe phaeohyphomycosis carried deleterious CLEC7A mutations: 5 with biallelic and 8 with heterozygous CLEC7A mutations. This represents a significantly greater frequency than that seen at the population level, evidenced by comparing to healthy individuals recorded in 1000 Genomes (https://www.internationalgenome. org/. Accessed January, 2021) (263 out of 2,504, 10.5%; P < 0.0001, Fisher's exact test) ( Figure 5A). Moreover, we found significantly reduced binding of the C-terminus-targeted anti-Dectin-1 antibody on the surface of monocytes and decreased TNF-α production upon β-glucan stimulation by PBMCs of patients with phaeohyphomycosis carrying CLEC7A mutations, consistent with impaired Dectin-1-dependent responses in these patients ( Figure  5, B and C). Impaired TNF-α production was observed across Dectin-1 p.Y238* homozygous (red symbols) and heterozygous (green symbols) and Dectin-1 p.I223S heterozygous (blue symbols) patient PBMCs ( Figure 5C), similar to the previously reported decreased proinflammatory cytokine production in heterozygous and homozygous Dectin-1 p.Y238* PBMCs (19,20,(36)(37)(38), suggesting that p.Y238* and p.I223S may act in a dominant-negative manner. Collectively, these data indicate that impaired Dectin-1 responses may have contributed to severe phaeohyphomycosis in more than 50% of patients in our cohort with this infection.

Discussion
We report a patient with Dectin-1 deficiency with severe C. cassiicola phaeohyphomycosis, an infection previously reported in inherited CARD9 deficiency. Therefore, Dectin-1 deficiency may phenocopy CARD9 deficiency with respect to susceptibility to this dematiaceous fungus in the setting of traumatic fungal inoculation. We show that Dectin-1 binds to C. cassiicola and promotes proinflammatory cytokine production by human PBMCs. In C. cassiicola-infected mice, Dectin-1 and CARD9 deficiencies impaired fungal control and decreased production of TNF-α and IL-1β, which contribute to protective immunity via macrophage gillosis in an immunosuppressed patient (55). Additional studies will be required to examine the mechanisms by which p.Y238* and p.I223S may act in a dominant or co-dominant manner and to model and study phaeohyphomycosis caused by Cladophialophora, Bipolaris, Curvularia, and other melanin-bearing agents of phaeohyphomycosis in WT, Dectin-1-deficient, CARD9-deficient, and heterozygous and homozygous p.Y238* and p.I223S mice. Moreover, the identification of deleterious CLEC7A mutations in approximately 70% of patients with severe phaeohyphomycosis in our cohort requires validation in future studies with more patients manifesting various fungus-and tissue-specific forms of severe phaeohyphomycosis. Furthermore, the role of prolonged antifungal therapy and/or long-term secondary antifungal prophylaxis in patients with severe phaeohyphomycosis who carry deleterious CLEC7A mutations requires further study.
In summary, we show that Dectin-1 deficiency may contribute to susceptibility to severe phaeohyphomycosis by certain dematiaceous fungi upon traumatic inoculation. Our findings provide mechanistic insights into the pathogenesis of phaeohyphomycosis and may help improve the diagnosis and management of immunocompetent patients who develop this infection.

Methods
Whole-exome sequencing. Sequencing libraries were generated using the TruSeq DNA Sample Prep Kit (Illumina), following the NimbleGen SeqCap EZ Library SR User's Guide, v4.0 (Roche Nimblegen, Inc.). Briefly, 1 μg of each gDNA was sheared using the Covaris instrument and Covaris microTUBEs (Covaris, Inc.). The following shearing conditions were used: duty cycle, 10%; intensity, 5.0; bursts per second, 200; duration, 120 seconds; mode, frequency sweeping; temperature, 5.5°C to 6.0°C. Each sample was prepared with a specific indexing adapter to facilitate multiplex pooling in the exon enrichment procedure. Exome enrichment was performed using the SeqCap EZ Human v3.0 Exome Enrichment kit on 200 ng of each TruSeq library. Libraries were combined to create a 6-plex reaction for the enrichment, yielding a total DNA library mass of 1,200 ng. One microgram of the 6-plex pooled libraries was hybridized to capture target oligonucleotides for 48 hours, as specified in the manufacturer's protocol.
Exome-enriched libraries were quantified using the KAPA Library Quantification Kit (KAPA Biosystems), clustered on the cBot Cluster Station, and sequenced as 2 × 100 bp reads on the HiSeq 2500 instrument, according to the manufacturer's protocol (Illumina).
High-quality, trimmed paired-end sequence reads were mapped to the human genome reference consortium GRGh37 (hg19) using Bowtie 2 and the default parameters with mixed mode disabled (57). Multiply mapped reads and PCR duplicates were removed using SAMtools (sequence alignment/map) (57). Additional filtering of the prealigned BAM files to remove reads with low-quality base, mapping, and alignment scores was performed prior to variant detection using Strand NGS 2.1 software (Strand Genomics, Inc.) and the following parameters: mapping quality threshold 20 or greater, base quality 17 or greater, and alignment score 85 or greater. Aligned reads were also base quality score recalibrated and locally realigned around indels. Reads were filtered against the SeqCap EZ Exome v3.0 target region (64 Mb exonic sequences) and reads greater than 100 bp outside of the targeted region were excluded from further analyses. The average coverage depth was 75× from approximately 1.9 Gb. Identification of able opportunities for this fungus to cause human infection, and thus may additionally explain the rarity of this infection despite the relatively high frequency of deleterious CLEC7A variants in the general population. Of note, these deleterious CLEC7A mutations are enriched in the San population of South Africa, a region where subcutaneous mycoses are common (19,52). Whether genetic variation in the CLR/CARD9 pathway may explain, at least partly, the increased prevalence of subcutaneous mycoses in certain subtropical regions warrants investigation. Our findings, together with the accompanying report by Hsu and colleagues (53), imply that Dectin-1 may be critical to the optimal development of antifungal defense following traumatic inoculation by low-virulence, dematiaceous fungi (Corynespora), or inhalational exposure to highly virulent fungi in certain geographic areas (Coccidioides).
To study the immunopathogenesis of C. cassiicola phaeohyphomycosis, we modeled it in mice. We used footpad swelling as a surrogate measure of infection control, which we found correlated to tissue fungal burden, although differential swelling in different gene-deficient mice may also be due to inappropriate inflammation and inability to form or resolve granulomas. Using this model, we found that Dectin-1 and CARD9 were each required for anti-C. cassiicola defense, which depended on macrophages, and TNF-α and IL-1β were critical for protection and macrophage fungal killing. The mechanism by which macrophages mediate C. cassiicola killing remains to be determined, although our data indicated that ROS was either not involved or could be compensated by additional nonoxidative mechanisms. These data are consistent with the normal oxidative burst of our index patient and the lack of reported phaeohyphomycosis in patients with chronic granulomatous disease who lack phagocyte oxidative burst (34,54). Moreover, future studies will be required to determine whether human monocyte-derived macrophages also rely on TNF-α and IL-1β in an interdependent manner for anti-C. cassiicola defense. By contrast, neutrophil accumulation was unaffected by Dectin-1 and CARD9 deficiency in this model, although we cannot rule out functional deficits in these cells or that reduced IL-1β and TNF-α may have had an effect on fungal uptake and killing by neutrophils; these questions will require further study using neutrophil-specific Dectin-1-and CARD9-deficient mice. However, we found that monocyte/macrophage-specific CARD9-deficient mice had enhanced susceptibility to C. cassiicola infection, indicating that macrophage defects are the primary driver of the disease in this model. In addition to intact neutrophil recruitment, we also found that IL-17A production within C. cassiicola-infected footpads were Dectin-1 and CARD9 independent. This contrasts with Phialophora and Exophiala phaeohyphomycoses, where neutrophil recruitment and IL-17A production were shown to be CARD9 dependent (3,18). Thus, CLR/CARD9-dependent protection against phaeohyphomycosis may utilize differential fungal species-specific mechanisms, which merit further investigation. Future studies should decipher potential additional Dectin-1-and CARD9-dependent anti-C. cassiicola protective cellular and molecular mechanisms and define which Dectin-1-independent CARD9-dependent CLRs may also participate in defense against C. cassiicola and other dematiaceous fungi in mice and humans. Of note, a recent report indicated that impaired Dectin-2-dependent immune responses may have contributed to the development of fatal invasive asper-counting with trypan blue exclusion. Washed PBMCs were used for stimulation assays or flow cytometry.
PBMC stimulation assay. PBMCs (5 × 10 5 ) were incubated in a round-bottom 96-well plate (Corning) at 37°C in a 5% CO 2 incubator in complete RPMI, supplemented with either 200 ng/mL LPS (Sigma-Aldrich), 1 μg/mL Pam3CSK4 (Sigma-Aldrich), α-mannan (see below), β-glucan (see below), or heat-killed (incubation at 70°C for 1 hour) C. cassiicola, which was added at a 1:12 dilution to the cells. After 48 hours, PBMCs were pelleted and the supernatant was collected and stored at -80°C until analysis. Cytokine levels were analyzed via conventional ELISA ( Figure 5B; TNF-α DuoSet, R&D Systems, DY210) or Luminex ( Figure 1F). Luminex analysis was done via a multiplex bead array assay with antibodies and cytokine standards (R&D Systems, Peprotech). Individual Luminex bead sets (Luminex) were coupled to cytokine-specific capture antibodies according to the manufacturer's protocols and biotinylated polyclonal antibodies were used at twice the recommended concentrations for a classical ELISA according to the manufacturer's instructions. The assay was run with 1,200 beads per set of cytokines in a volume of 50 μL. The plates were read on a Luminex MAGPIX platform where more than 50 beads were collected per bead set. The median fluorescence intensity of the beads was then measured for each individual bead, which was analyzed with the Millipex software using a 5P regression algorithm.
β-Glucan and α-mannan preparations. Carbohydrates were purified as described previously (see main text for references), and then prepared for PBMC stimulation. In brief, 5 mg of particulate β-glucan was added to 1 mL water, sonicated for 30 seconds, pelleted by centrifugation at 10,000g for 10 minutes and resuspended in complete RPMI for use in stimulation experiments at the indicated concentrations. For α-mannan, 5 mg was dissolved in PBS, mixed at room temperature for 2 hours with shaking, centrifuged at 10,000g for 10 minutes, and resuspended in complete RPMI at indicated concentrations for stimulation experiments.
Dectin-1 expression by flow cytometry. PBMCs were isolated as above, resuspended in PBS, and Fc receptors blocked with anti-CD16/32 (BD Biosciences) on ice for 10 minutes prior to staining with anti-human CD14-FITC (Biolegend, clone HCD14) and anti-human Dectin-1-PE (Biolegend, clone 15E2), or the appropriate isotype control (Biolegend, IgG2a clone MOPC-173), on ice for 30 minutes. Stained samples were washed in PBS and acquired using a BD LSR, equipped with BD FACSDiva software. FlowJo (BD) was used for the final analysis.
Dectin-1 immunoblot analysis. Whole-cell lysates were suspended in 1× RIPA buffer containing protease and phosphatase inhibitors (Thermo Fisher Scientific). The lysates were centrifuged at approximately 14,000g for 15 minutes at 4°C. Supernatants were collected and protein concentrations were determined using the Bradford Protein Assay (Bio-Rad), according to the manufacturer's protocol. The proteins (20 μg) were resolved in 12% SDS-PAGE and electrotransferred onto 0.2 μm PVDF. Membranes were blocked in 5% BSA and incubated with primary antibody against Dectin-1 (Invitrogen, PA5-34382) and β-actin (Cell Signaling Technology, clone D6A8), followed by secondary anti-rabbit IgG, HRP-linked antibodies (Cell Signaling Technology). Chemiluminescence detection was performed with Clarity Western ECL Blotting Substrate (Bio-Rad), using the Chemi-Doc MP Imaging System (Bio-Rad). Quantification was obtained by densitometry image analysis using Image Lab 5.2 software (Bio-Rad).
SNPs and indels was performed using Strand NGS software. Strand NGS utilizes a modified Bayesian variant calling method adapted from the MAQ SNP calling algorithm which compares the nucleotides present on aligned reads against the reference at each position in the genome. The dbGaP study accession numbers are phs001899.v2.p1 and phs001561.
CLEC7A sequencing. Genomic DNA was isolated from peripheral blood leukocytes using a Gentra Puregene DNA isolation kit (Qiagen) and amplified using Platinum Taq PCR SuperMix High Fidelity (Thermo Fisher Scientific) following the manufacturers' protocols with primers and cycling conditions listed in Supplemental Table 2. PCR products were purified with Exo-SAP IT (Thermo Fisher Scientific) and sequenced using BigDye Terminators v3.1 (Applied Biosystems) per the manufacturers' instructions using primers listed in Supplemental Table 3. Reactions were purified over Performa DTR plates (Edge Biosystems) and resulting products run on an ABI 3730XL capillary sequencer. Chromatograms were analyzed using Sequencher (Gene-Codes) and compared to the NCBI reference sequence, NM_197947.
C. cassiicola culture. The C. cassiicola strain used in all experiments was isolated from our Dectin-1-deficient patient. C. cassiicola was stored in 50% glycerol at -80°C, and routinely grown on potato dextrose agar (PDA; Sigma-Aldrich) plates at 20°C for 7 days, harvested, and replated onto fresh PDA plates and grown for an additional 7 days before use. To harvest C. cassiicola, 10 mL sterile tissue-culture grade water (Gibco) was added to each plate, and a soft paint brush (sterilized in 70% ethanol and rinsed in sterile water) used to brush the colonies and lift fungal cells. The fungi and water were pipetted off the plate, and a further wash/brush repeated with 5 mL of water. The resulting suspension was centrifuged (1,800g for 4 minutes) to pellet C. cassiicola, which was resuspended in sterile PBS. Each suspension was adjusted to an OD 600 of 2.0 for use in experiments. Microscopy analysis and analysis of Dectin-1 ligand exposure was performed as outlined below.
Dectin-1 ligand staining. C. cassiicola was prepared as above and 50 μL used for staining. Pelleted fungal cells were resuspended in 1 mL 2% paraformaldehyde and incubated at room temperature for 10 minutes. Fungi were centrifuged again and supernatants discarded, before washing in 1 mL PBS. Cell pellets were resuspended in 400 μL 10% mouse serum/PBS and incubated for 15 minutes at room temperature to block, and then washed in 1 mL PBS. Cells were resuspended in 100 μL PBS and 1 μL Fc-Dectin-1 (29) added to the cells, incubating at 4°C for 1 hour with continual slow rotation. Fungi were washed twice in PBS, and then stained with 2 μg/mL anti-human IgG-PE secondary in 100 μL PBS for 30 minutes at 4°C. Cells were washed twice in PBS, resuspended in 20 μL PBS, added to microscope slides, and analyzed with a Leica fluorescence microscope equipped with ZEN software. C. albicans germ tubes (induced in 20% FBS/water for 1 hour at 37°C) were used as a positive control, and secondary-only-stained samples used as a negative control.
C. cassiicola killing assay. Bone marrow cells from the femurs/ tibias of C57BL/6 females were flushed out with sterile 2 mM EDTA/ PBS, washed, and resuspended in RPMI (with GlutaMAX; Invitrogen) with 20% FBS (Sigma-Aldrich), 1% penicillin/streptomycin (Invitrogen), and 40 ng/mL M-CSF (Biolegend). Bone marrow cells were incubated in 75 cm 2 tissue culture flasks (Corning) for 5 days, replacing the media on day 3. On day 5, the media were replaced with ice-cold 2 mM EDTA/PBS and flasks incubated on ice for 10 minutes. Adherent cells (macrophages) were gently lifted using a cell scraper (Gibco) and counted using trypan blue exclusion. Macrophages were then seeded into flat-bottom 96-well plates (4 × 10 5 per well) in RPMI supplemented with 10% FBS and 1% penicillin/streptomycin. Some wells were additionally supplemented with 10 ng/mL recombinant murine TNF-α (Biolegend) and/or 10 ng/mL recombinant murine IL-1β (Biolegend), or 5 mM N-acetyl cysteine (Sigma-Aldrich). Macrophages were incubated overnight at 37°C and 5% CO 2 . The following day, C. cassiicola (prepared as above to OD 600 = 2.0) was added to the macrophages (50 μL fungal suspension added per well to a final volume of 200 μL) and the incubation at 37°C continued for a further 3 hours. Culture supernatants were then collected and frozen at -20°C. Concentrations of β-D-glucan in the supernatants were measured using the Glucatell (1→3)-β-D-glucan detection reagent kit (Cape Cod Associates), as per the manufacturer's instructions.
Footpad histology and cytokine determination. The infected left hind foot of WT, Card9 -/-, and Clec7a -/mice was removed and fixed in 10% formalin for 24 hours before embedding in paraffin wax. Tissue sections were stained with periodic acid-Schiff (PAS). To estimate C. cassiicola burden in the footpad, pictures were taken of the entire footpad section (at least 2 per mouse) and the total area covered by fungal cells measured using ImageJ software (NIH). For cytokine analysis, infected footpads were removed with a scalpel and added to 0.5 mL PBS supplemented with 0.05% Tween 20 and a protease inhibitor cocktail (Roche). Footpads were homogenized and cell debris removed by centrifugation at 1,100g for 5 minutes at 4°C. The resulting supernatant was snap-frozen on dry ice and stored at -80°C until analysis. Samples were analyzed by Luminex array as described above, and cytokine concentrations were determined per milligram of footpad tissue.
Determination of fungal burden in infected mouse footpads with qPCR. On day 10 after infection, infected footpads of WT, Card9 -/-, Clec7a -/-, and Ifng -/mice were harvested and homogenized in 1 mL PBS using Omni tip homogenizers (Omni International). To isolate and quantify DNA from pure C. cassiicola, the fungus was grown as described above and the fungal suspension was adjusted to an OD 600 of 2.0 in 1 mL PBS. The footpad homogenates and pure fungal suspensions were centrifuged at 5,000g, supernatants were discarded, and the pellets were utilized to isolate DNA as described previously (58). For qPCR, 20 ng of total DNA, PerfecTa Fast mix (Quanta Biosciences), fungus-specific Fungi-Quant probe/primers targeting the fungal 18s rRNA gene, and specific cycling program were utilized as described previously (59). Fungal DNA amount per footpad was interpolated from a series of serially diluted DNA from pure C. cassiicola. HEK293 transfection. WT Dectin-1 expression plasmid (Origene, clone SC307610, NM_197947) was used for site-directed mutagenesis to generate the patient variant, c.714T>G; p.Y238*. HEK293 cells (ATCC) were plated at 5 × 10 5 cells/well in a 6-well plate and allowed to rest overnight. The following day, cells were transfected using 2 μg WT or p.Y238* plasmid and Fugene HD reagent (Promega) at a 3:1 ratio. Twenty-four hours later, cells were harvested for immunoblot analysis.
Mice. Eight-to 12-week-old female mice were maintained in individually ventilated cages under specific pathogen-free conditions at the NIH. The following strains (and their respective WT controls/ littermates) were obtained from the NIAID Taconic contract; Rag1 -/-, Il1r1 -/-, Tnfa -/-, and Ifng -/-. All other strains and their respective controls/littermates were bred in-house at the NIH (Card9 -/-, Clec7a -/-, Il1b -/-, Card9 fl/fl Cx3cr1 Cre-ER+/-). In experiments using Cre-expressing mouse lines, mice were pretreated with 3 mg tamoxifen (dissolved in Miglyol 812, The Warner Graham Company) daily by intraperitoneal injection for 5 days prior to infection. Animals were infected and analyses performed as outlined below. Animals were euthanized by cervical dislocation following administration of ketamine/xylazine cocktail at defined time points after infection (see figure legends).
Footpad infections. C. cassiicola was prepared as above and 50 μL injected into the left hind footpad of each mouse using a 21-G needle. Footpad swelling was measured once per week using calipers, measuring the height of the foot just below the toes at the widest part of the foot. The right uninfected footpad was measured and swelling expressed as a percentage relative to the uninfected footpad for each individual mouse. For C. albicans footpad infections, C. albicans SC5314 was grown in YPD broth overnight at 30°C with shaking at 200 rpm, washed in PBS, and counted with a hemocytometer. Yeast cells (5 × 10 6 ) were injected into the left hind footpad in a 50 μL volume and footpad swelling measured every 2 days as described above. C. albicans footpad burdens were measured by removing the infected footpad with a scalpel, homogenizing in 0.5 mL sterile PBS, and plating on YPD agar plates. Viable colonies were counted after incubation at 37°C for 24 hours and expressed as CFU per gram of footpad tissue.