Two patients in this study, their family members, and six ethnically matched healthy volunteers provided written informed consent for participation in the study, which was approved by the Clinical Research Ethics Committee of the Peking University First Hospital. The patients permitted us to use their images and medical information.

By use of the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany), genomic DNA was extracted and purified from approximately 1 cm² of fungal elements. Cells were disrupted with glass beads (425–600 µm) (Sigma-Aldrich, Zwijndrecht, The Netherlands) and TissueLyser II (Qiagen). The nuclear genes, ITS and BT2, were amplified by PCR. Details of PCR amplification and sequencing primers can be found in the reference paper (6). Amplification was performed with the 2×EasyTaq PCR SuperMix protocol (Trans Gen Biotech, Beijing, China). Add template DNA (50–100 ng) and forward and reverse primers (0.2–0.4 µM each) to a total reaction volume of 25 µL. DNA amplification was performed in a Mastercycler (Eppendorf, Hamburg, Germany), the process of which includes machine-preheating at 94°C for 5 min, 30 cycles of denaturation at 94°C for 30s, annealing at 54°C for 30s, extension at 72°C for 30s, and final extension at 72°C for 10 min. Amplified bands were visualized using the Gel Doc XR+ system (BioRad, Hercules, CA, USA) with Trans2K Plus DNA Marker (Trans Gen Biotech) indicating size and concentration. PCR products were sequenced by Sangon Biotech Co. Ltd. (Shanghai, China). The alignments and phylogenetic reconstructions were performed in accordance with Li et al. (5).

Human peripheral blood mononuclear cells (PBMCs) were collected from whole blood by density-gradient centrifugation using Ficoll-Paque Plus (GE Healthcare, Chicago, IL, USA) as previously described (6). Specifically, fresh venous blood was drawn into 10 mL EDTA tubes and diluted with phosphate-buffered saline (PBS). PBMCs were isolated using Ficoll-Paque density-gradient centrifugation, meanwhile, erythrocytes in the pellet were lysed. PBMCs were flushed twice with PBS at 800 × g for 8 min and reconstituted in RPMI 1640 medium (HyClone, Logan, UT, USA) supplemented with 100 U/mL penicillin and 100 μg/mL streptomycin (HyClone).

P. expanda isolates were cultured on potato dextrose agar (PDA; BD Biosciences, San Jose, CA, USA) for 3–14 days at 28°C to harvest conidia. Heat-killed (HK) conidia for stimulation were prepared for 30 min at 99°C in a water bath.

The PBMCs were incubated in 96-well plates at a final concentration of 2.5 × 10⁶/mL in a total volume of 200 μL RPMI 1640 medium with 10% autologous serum per well. The PBMCs were stimulated with different pattern recognition receptor agonists, including lipopolysaccharide (LPS, a Toll-like receptor 4 agonist, 100 ng/mL), Trehalose-6,6-dibehenate (TDB, a Mincle agonist, 100 μg/mL), β-glucan (a Dectin-1 agonist, 50 μg/mL), and fungal particles (HK or viable resting conidia, 10⁷ particles/mL) at 37°C in a 5% CO₂ atmosphere. Innate ELISA assays were carried out with stimulated culture supernatants or cells after 24 h. In order to detect the adaptive immune response, PBMCs were stimulated with the corresponding HK fungal particles for 6 days, and cells or culture supernatants were gathered for further ELISA assays and FACS.

After human PBMCs were stimulated for 24 h or 6 days, cell culture supernatants were collected and stored at -70°C. According to the instructions provided by the manufacturer, we used commercial ELISA kits (R&D Systems, Minneapolis, MN, USA) to measuring the cytokines of the patients and healthy controls.

PBMCs were incubated with different HK fungal stimuli for six days, then residual living cells were re-stimulated with Cell Stimulation Cocktail (eBioscience, San Diego, CA, USA) for 5 h in an incubator to promote the intracellular accumulation of secreted cytokines. Human PBMCs were surface stained with FITC-conjugated anti-human CD4 antibody, fixed and permeabilized by Cytofix/Cytoperm solution (BD Biosciences), and stained with PE-conjugated anti-human IL-22 and Alexa^®Fluor647-conjugated anti-human IL-17A antibodies. Data were collected on a BD FACS Calibur system and analyzed with FlowJo7.6 software.

To explore the differential transcript profile between the two patients, we conducted RNA-seq using the Illumina platform. We used the biopsy specimens from skin lesions of the patients to extract RNA using Total RNA Kit I (R6834-01, OMEGA, USA). Raw counts were normalized to balance the sequencing depth using fermented palm kernel meal (FPKM). Then, the parameters (q ≤0.01 and |log₂ Fold change| ≥1) were used to identify differentially expressed genes (DEGs, Patient2 versus Patient1) using the Limma package (version 3.42.2) in R. To reveal the pathway activities of DEGs, we performed pathway enrichment analyses for up-regulated DEGs (q ≤0.01 and log₂ Fold change ≥1) and down-regulated DEGs (q ≤0.01 and log₂ Fold change ≤-1) using the ClusterProfiler package (version 3.14.3). Antifungal immunity-related genes were obtained from previous studies. Default statistical methods in respective packages or software were used to conduct the test the significance.

The biopsy specimens from skin lesions of the two patients were fixed in 10% buffered formalin and paraffin embedded for immunohistochemical analysis. According to the instructions of the IHC antibody, we performed immunohistochemical staining. Stained slides were photographed by microscope.

Card9-KO mice (C57BL/6 background) were generously provided by Xin Lin (Tsinghua University School of Medicine, Beijing, China, and MD Anderson Cancer Center, Houston, TX). In this study, 6−8-week-old Card9-KO and C57BL/6 WT mice (Vital River Laboratories, Beijing, China) were maintained in specific pathogen-free facilities at the Institute of Clinical Pharmacology of Peking University. The WT and Card9-KO mice were injected at two hind footpads subcutaneously with 100 μL viable P. expanda (1×10⁹ particles/mL). Skin biopsy specimens were obtained, and slides were stained with periodic acid-Schiff for histopathological analysis.

Main components of in vitro culture medium were MgSO₄·7H₂O, KH₂PO₄, NH₄NO₃, Biotin, Thiamine, Glycerol, and Nicomycin; pH:5.5−6.0. To this medium, 100 μL viable P. expanda spores (1×10⁷/mL) were added and placed in an incubator for 45−50 days, setting the temperature to 35−36 °C.

Data were plotted using GraphPad Prism 7.0 software (La Jolla, CA, USA) and analyzed with unpaired t-tests in SPSS 22.0 software (Chicago, IL, USA). P values of <.05 were considered statistically significant.

Patient1(P1) was a 55-year-old woman from northeast China, born to non-consanguineous parents. She presented to our clinic with an erythematous plaque on her face, that had been gradually enlarging for the past 30 years. At the age of 23, she noted several red papules on her forehead; the area gradually expanded with occasionally itching and without pain. She was diagnosed with “fungal granuloma” at a local hospital and treated with “itraconazole 400 mg/day” for approximately 1.5 years. However, the skin lesions enlarged and spread to her forehead, cheek, and eyelid. Dermatological examination revealed a protuberant dark red mass on her face with erosion, crust, and pus, which covered the forehead and cheek with a clear boundary ( Figure 1A ). A skin biopsy from the lesion showed pseudoepitheliomatous hyperplasia of the epidermis and intense dermal inflammatory infiltrations with the presence of multiple sclerotic cells within the dermis. ( Figure 1B ) A smear of the lesion scrapings showed brown sclerotic cells with cross septa ( Figure 1B ). Therefore, a diagnosis of chromoblastomycosis was made.

Cultures of the biopsied tissue specimens showed filamentous fungus, whose colony was dense, brown to black after 21 days on potato dextrose agar at 28 °C. Further microscopic examination of the cultured organisms revealed vase-shaped sporogenous cells with collar-like structures and flower-like arrangements of small conidia, which implied it belongs to Phialophora spp. ( Figure 1C ). According to our previous phylogenetic analysis of Phialophora verrucosa complex, there are seven species in this complex, including P. verrucosa, P. americana, P. expanda, P. chinensis, P. tarda, P. ellipsoidea and P. macrospora (5). To further identify the isolate, we sequenced the ITS and BT2 rRNA genes using PCR. Then, we constructed the phylogenetic tree based on ITS and BT2 and found this species belonged to the P. expanda clade, which has never been reported to cause disease in the literature. Amphotericin B plus itraconazole was used for therapy. There were some improvements after treatment, which to some extent control the progress of the disease, but she was not completely cured and was still under treatment. Based on previous reports and the recalcitrant character of this patient, we extracted her peripheral blood DNA to sequence for CARD9‐coding exons, and found a homozygous frameshift mutation in exon 6 (c.819‐820insG, p. D274fsX60).

Patient2 (P2) was a male with previously reported phaeohyphomycosis caused by P. americana, also from the P. verrucosa complex (5, 6). Besides his similar facial rash, his CARD9 mutation site was also identical with that of P1. However, unlike sclerotic bodies observed in the P1 tissue, the morphology of fungus in P2 tissue had short, septate, and dematiaceous hyphae, which supported the diagnosis of phaeohyphomycosis.

Primary peripheral immune cells were isolated from the two patients to determine the functional effects of CARD9 deficiency. We first used PBMCs from the P1 to assess the release of inflammatory cytokines (TNF-α, IL-6, and IL-1β) in response to 24 h stimulations with LPS, TDB, β‐glucan, and fungal particles, including P. expanda conidia (HK or viable yeast). Compared with cells isolated from the healthy control group, the P1 showed comparable defects in IL-6, IL-1β, and TNF-α expression in response to stimulations of LPS, TDB, β‐glucan, and HK or viable P. expanda conidia ( Figure 2A ). We then monitored Th cell responses in the patient and the control PBMCs stimulated with HK fungi for 6 days and found a marked attenuation in IL-22 and IL-17 expression in the patient’s cells. Consistently, the patient’s samples showed a significant reduction in Th22 (IL-22⁺IL-17^-CD4⁺) and Th17 (IL-17⁺CD4⁺) cells ( Figure 2B ). Similar data were obtained in P2 in our previous report, suggesting comparable deficient ex vivo immune responses in these two patients (6).

To explore the mechanisms of heterogeneous fungal morphotypes in the two patients, we performed RNA-seq analysis using the biopsy tissue from the lesions. A total of 4,935 DEGs (P1 versus P2) including 1,972 up-regulated and 2,963 down-regulated DEGs were screened out. Up-regulated DEGs were enriched in several antifungal-related pathways, including cytokine-cytokine receptor interaction, chemokine signaling pathway, and IL-17 signaling pathway ( Figure 3 ), indicating that P. expanda induces higher excessive immune response, thus P1 had a relatively higher inflammatory response than P2. However, down-regulated DEGs only were enriched in the Herpes simplex virus 1 related pathway, Ribosome and Chemical carcinogenesis, and a few irrelevant pathways ( Figure 3 ), suggesting that P. americana caused a relatively weak inflammatory response. Accordingly, pattern recognition related genes (CLEC4E), proinflammatory cytokines related genes (IL-1β, IL-6), chemokine related genes (CXCL1, CXCL2), antimicrobial peptides related genes (S100A8, MMP, DEFB1), and adaptive immunity-related genes (IL-17A, IL-17RA), all showed higher expression in P1 versus P2 ( Figure 4 ). Given the heterogeneous pathway activities between the two patients, we speculate that local immune cell proportions are also different. Indeed, we found that P1 showed higher CD68, CD11c, MPO, CD4 expression by immunohistochemistry, indicating more macrophages, dendritic cells, neutrophils and CD4⁺ T cells infiltrated to the lesion, respectively ( Supplementary Figure 1 ). Different infiltration of the immune cells may in return explain the stronger responses induced by P. expanda than P. americana. Together, these results indicate that these two pathogens induce different immune responses during infections in patients with the same gene mutations.

To study fungal specificity in these two species, we carried out in vivo and in vitro studies to test the morphotypes of them. We first used Card9 KO mice and inoculated fungal spores (P. expanda and P. americana) into the footpad, as previously described (3, 6). The results revealed that both P. expanda and P. americana were able to cause chronic infections in Card9 KO mice, whereas WT mice recovered from the inoculation in 4 weeks ( Figure 5 ). Similar lesions of swelling, abscess, ulceration, and crust were noted in these two isolates infected mice at 6 weeks post-infection. However, unlike plenty of pigmented fungal hyphae in P. americana infected mice, P. expanda infected mice showed lots of sclerotic bodies of fungal elements in tissue ( Figure 5 ). These results indicated that P. americana is prone to causing phaeohyphomycosis, whereas P. expanda causing chromoblastomycosis in Card9 KO mice.

For further verification, we designed induction experiments in vitro to compare the ability to form muriform cells in these two isolates. After 50 days of culture, we successfully induced muriform cells in P. expanda, whereas P. americana showed beaded hyphae. Therefore, we successively confirmed that P. expanda is prone to forming a sclerotic body both in vivo and in vitro, which is consistent with our clinical findings ( Figure 6 ).