CDCP1 on dendritic cells contributes to the development of a model of Kawasaki disease

Abstract

The etiology and pathology of Kawasaki disease (KD) remain elusive. Cub domain-containing protein 1 (CDCP1), a cell surface protein that confers poor prognosis of patients with certain solid tumors, was recently identified as one of the most significantly upregulated genes in SARS-CoV-2-infected children who developed systemic vasculitis, a hallmark of KD. However, a potential role of CDCP1 in KD has not previously been explored. In this study, we found that CDCP1 knockout (KO) mice exhibited attenuated coronary and aortic vasculitis and decreased serum CAWS-specific IgM/IgG2a and IL-6 concentrations compared to WT mice in an established model of KD induced by Candida albicans water-soluble fraction (CAWS) administration. CDCP1 expression was not detectable in cardiomyocytes, cardiofiblobasts or coronary endothelium, but a constitutive expression of CDCP1 was observed on dendritic cells (DCs) and upregulated by CAWS stimulation. CAWS-induced IL-6 production was significantly reduced in CDCP1-KO DCs, in association with impaired Syk–MAPK signaling pathway activation. These novel findings suggest that CDCP1 might regulate KD development by modulating IL-6 production from DCs via the Syk-MAPK signaling pathway.

Introduction

Kawasaki disease (KD) is the leading cause of acquired heart disease in infants and young children in developed countries(1). This acute febrile disorder is characterized by systemic vasculitis that often leads to potentially fatal coronary artery pathologies such as dilation and aneurysm(2). Despite decades of research, the etiology of KD remains unknown, although opportunistic bacterial and viral infections and subsequent inflammatory responses have been implicated in the pathogenesis(3–5). Consistent with those observations, KD-like symptoms have been reported in children infected with severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), suggesting that viral infection can trigger this disorder(6, 7). In most KD patients, timely and large doses of intravenous immunoglobulin (IVIG) reduce the incidence of coronary complications. However, approximately 10–15% of KD patients do not respond to IVIG and are at risk of potentially fatal cardiac complications(8–10). Therefore, it is critically important to understand the mechanisms underlying KD and develop new therapeutic approaches for affected children, especially those who are refractory to IVIG treatment. Such mechanistic studies could also provide crucial insights into the treatment of cardiac complications of SARS-CoV-2 infection in children.

CUB domain-containing protein 1 (CDCP1) is a cell-surface glycoprotein that has been studied almost exclusively in tumor biology(11). Significant upregulation of CDCP1 is observed in many solid tumor cells and is associated with a poor prognosis in certain cancer patients(12–14). Activation of CDCP1 on tumor cells leads to activation of the PI3K/Akt pathway, which promotes tumor cell survival and metastasis through intrinsic mechanisms(15). CDCP1 also drives triple-negative breast cancer metastasis by reducing the abundance of lipid droplets and the stimulation of fatty acid oxidation(16). In mice, CDCP1 deficiency may enhance tumor growth by liberating integrin signaling and growth factor receptor cross-talk in detached tumor cells(17).

Although CDCP1 has been established as a key regulator of tumor cell survival and metastasis and is present on many epithelial cells, some hematopoietic stem cells, and mesenchymal stem cells also express CDCP1 under normal conditions(18, 19), few studies have addressed the potential role of this protein in non-tumor cells. We recently identified CDCP1 as a ligand of the T cell marker CD6(20), suggesting that CDCP1 could regulate immune responses and affect immune-mediated diseases either directly or indirectly. Moreover, CDCP1 was identified as one of the most significantly upregulated genes in a recent study of SARS-CoV-2-infected children with KD-like symptoms, suggesting a previously unknown role of CDCP1 in the pathogenesis of KD(21). Despite these insights, it remains unknown whether CDCP1 contributes to the development of KD and, if so, by which underlying mechanism.

In this project, we explored the potential role of CDCP1 in the pathogenesis of a mouse model of KD induced by treatment with Candida albicans water-soluble fraction (CAWS). Using both wild-type (WT) and CDCP1 knockout (KO) mice, we compared the severity of coronary and aortic vasculitis and studied the related immune mechanisms. We first examined CDCP1 expression in the heart and dendritic cells (DCs); we then studied the regulation of CDCP1 expression by CAWS, the effect of CDCP1 on the production of IL-6, a key inflammatory cytokine, and the potential underlying signaling pathways. Our results establish CDCP1 as a novel regulator of cardiac vasculitis development in the mouse model of KD.

Materials and Methods

Mice

WT mice (C57BL/6J ) and CDCP1 KO mice(17) (backcrossed 7 generations into the C57BL/6J background) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA) and maintained under pathogen-free conditions in the animal facilities of the Lerner Research Institute, Cleveland Clinic. Sex- and age-matched mice (sex ratio 1:1, 8–10 weeks) were used for all experiments. All animal care and experimental procedures were approved by the Institutional Animal Care and Use Committee of the Cleveland Clinic and performed following the U.S. Department of Health and Human Services Guide for the Care and Use of Laboratory Animals.

Induction of KD by CAWS

The CAWS complex was prepared as previously described(22). To induce KD, WT and CDCP1 KO mice were intraperitoneally injected with 4 mg of CAWS(23). Mice were euthanized on day 28 post-injection. Whole hearts were isolated, perfused and fixed in a 4% formaldehyde solution followed by embedding in paraffin for subsequent sectioning and histopathological analysis. Coronary artery and aortic root lesions and myocardial inflammation were identified in serial sections (5 μm) stained with hematoxylin and eosin (H&E) or Masson’s trichrome (MT). The histopathological tissue examinations and inflammation severity scoring were performed by a pediatric rheumatologist experienced in the CAWS model of KD, who was blinded to the genotypes or experimental groups. Inflammation severity was assessed using the following scoring system: 0, no inflammation; 1, rare inflammatory cells; 2, scattered inflammatory cells; 3, diffuse infiltrate of inflammatory cells; and 4, dense clusters of inflammatory cells(24). The inflammation incidence rate was determined by the presence of any coronary, aortic or myocardial inflammation (i.e., inflammation score ≥1).

Preparations of mouse bone marrow-derived dendritic cells (BMDCs) and human monocyte-derived DCs

BMDCs were prepared from the femurs and tibias of age and sex-matched WT and CDCP1 KO mice according to previously described protocols, with minor modifications(25). In brief, 1×10⁶ cells/mL were suspended in RPMI-1640 complete medium containing 10 ng/mL murine GM-CSF (PeproTech, Rocky Hill, NJ, USA). The cells were cultured in 10 cm culture dishes, and the medium was replaced on days 3 and 5. On day 7, loosely adherent cells were harvested, washed in Dulbecco’s phosphate-buffered saline (DPBS) and resuspended in RPMI-1640 complete medium. The harvested cells were subjected to flow cytometry to analyze the surface expression of CD11b and CD11c. This method yielded cell populations comprising 80–85% CD11b⁺CD11c⁺ DCs.

Human DCs were propagated from peripheral blood mononuclear cells using a commercially available Human Monocyte-derived Dendritic Cell Differentiation Kit (R&D System) following manufacturer-provided protocols.

Isolation and primary culture of adult mouse cardiac fibroblasts

Mouse cardiac fibroblasts (MCFs) were prepared from the hearts of age and sex-matched WT and CDCP1 KO mice according to previously described protocols, with minor modifications(26). In brief, four hearts from freshly euthanized adult mice were minced and digested in Hanks’ balanced salt solution supplemented with 0.1% trypsin (Millipore Sigma, Billerica, MA, USA) and 100 U/mL collagenase II (Worthington Biochemical, Lakewood, NJ, USA) at 37°C under constant stirring. Supernatants were collected and neutralized with DMEM/F12 complete medium every 5 min before centrifugation to collect the cells. The harvested cells were then incubated at 37°C with 5% CO₂ for 3 h. Adherent cells were washed with warm DPBS, replenished with fresh medium and cultured until 90% confluence was reached (typically after 1 week). Cells were then subjected to flow cytometry analysis to detect the surface expression of CD90.2, an established mouse fibroblast marker.

Flow cytometric analysis of inflammatory cell profiles

Splenic samples were prepared by homogenizing the tissues between two frosted slides in 10 mL of ice-cold DPBS. After erythrocyte lysis, the splenocyte suspensions were filtered and counted. Next, 1×10⁶ cells were centrifuged and resuspended in an Fc receptor blocking solution (Biolegend, San Diego, CA, USA) for 10 min on ice before staining. For peripheral blood sample analyses, 30 μL of blood from the tail vein were mixed with 1 mL of DPBS. After erythrocyte lysis, the leukocytes were resuspended in an Fc receptor blocking solution (Biolegend) for 10 min on ice.

The cells were resuspended in flow cytometry buffer (PBS with 1% BSA). Cell surface staining was performed in the dark for 30 min on ice, after which the cells were washed twice with flow buffer before the analysis. Fluorochrome-conjugated antibodies specific for the following surface markers were used (1:100 dilution; all purchased from Biolegend unless indicated otherwise): APC-CD3 (clone 17A2), PE- CD19 (clone 6D5), PerCP-Cy5-CD4 (clone GK1.5), PE-CD4 (clone GK1.5), FITC-CD8a (clone 53–6.7), PE-CD8a (clone 53–6.7), PE-NK-1.1 (clone PK136), PerCP-Cy5.5-CD11b (clone M1/70), FITC-CD11c (clone Bu15), Alexa Fluor 647-Gr-1 (clone RB6–8C5), APC-Ly-6G (clone 1A8), Alexa Fluor 700-Ly-6C (clone HK1.4), PE-CD90.2 (clone 20-H12), and FITC-Dentin-2 (clone KVa7–6E7; 1:50; Miltenyi Biotec, Somerville, MA, USA). All data were collected by a BD Fortessa Flow Cytometer (BD Biosciences, Franklin Lakes, NJ, USA) and analyzed using FlowJo software (TreeStar, Eugene, OR, USA). Fluorescence minus one (FMO) controls were used for the gating analyses to distinguish positively from negatively stained cell populations. Compensation was performed using a total antibody compensation bead kit (Invitrogen, Carlsbad, CA, USA). The gating strategy followed previously described protocols(27, 28).

Mouse cytokine/chemokine array assay

Serum samples were aliquoted and stored at −80°C prior to a multiplex bead-based analysis by Eve Technologies (Calgary, AB, Canada). The multiplex array assessed 44 cytokine/chemokines, including CX3CL1, CXCL10, CXCL9, CXCL5, CXCL2, CXCL1, CCL22, CCL20, CCL19, CCL17, CCL12, CCL11, CCL5, CCL4, CCL3, CCL2, TIMP1, EPO, VEGF, LIF, TNF-α, IFN-γ, IFN-β, GM-CSF, M-CSF, G-CSF, IL-20, IL-17, IL-16, IL-15, IL-13, IL-12p70, IL-12p40, IL-11, IL-10, IL-9, IL-7, IL-6, IL-5, IL-4, IL-3, IL-2, IL-1β, and IL-1α.

Mouse cytokine enzyme-linked immunosorbent assays (ELISA)

Serum samples, cell culture supernatants, and tissue lysates were aliquoted and stored at −80°C until use. Each sample was centrifuged at 13,000 rpm and 4°C for 10 min prior to the analysis. IL-6, IL-1α, TNF-α, IFN-γ, IL-1β and IL-10 were measured using mouse-specific sandwich ELISA kits according to the manufacturer’s instructions (Biolegend).

Measurement of CAWS-specific immunoglobins

To measure the induction of CAWS-specific Igs, an ELISA plate was coated with 5 μg/mL CAWS and incubated overnight at 4°C. After washing with PBST, the plate was blocked for 1 h at room temperature with 1% bovine serum albumin (BSA) in DPBS. Subsequently, diluted serum (1:1000) was added to each well, and the plate was incubated for 2 h at room temperature and then washed with PBST. To measure total IgG and total IgG+IgM, horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (Biolegend) and goat anti-mouse IgG+IgM (Millipore Sigma) were added to the respective wells and incubated for 1 h at room temperature. After washing, 50 μL of TMB substrate (BD Biosciences) was added to each well for the colorimetric reaction, and the reaction development was stopped by adding 50 μL of 1 M phosphoric acid. Finally, the optical density at 450 nm was measured in each well. For IgA, IgM, and IgG subclass measurements, Mouse Monoclonal Antibody Isotyping Reagents (Sigma-Aldrich, Milwaukee, WI, USA) were used according to the manufacturer’s protocol.

Reverse transcription-polymerase chain reaction (RT-PCR)

Total RNA was isolated from BMDCs using RNeasy Mini Kits (Qiagen, Germantown, MD, USA) according to the manufacturer’s protocol. Reverse transcription was performed using the SuperScript First-Strand Synthesis System for RT-PCR (Thermo Fisher Scientific, Portsmouth, NH, USA) according to the manufacturer’s instructions. The following primers were used to amplify the mouse CDCP1 transcript: mCDCP1 forward 5′-CTCGCAGACACAGGCCACA-3′ and reverse 5′-CAGCTCTCGCCTGGTCCAA-3′. RT-PCR was performed using Go Taq Master Mixes (Promega, San Luis Obispo, CA, USA). The products were examined on 3% agarose gels.

Western blotting

Cells and tissues were homogenized and lysed using the RIPA Lysis Buffer System (Santa Cruz Biotechnology, Dallas, TX, USA). The protein concentrations of lysates were quantified using the DC Protein Assay Kit (Bio-Rad, Hercules, CA, USA) according to the manufacturer’s protocol. Fifteen micrograms of total protein lysate were loaded on 4–20% PAGE Gels (GenScript, NY, USA), separated, transferred to PVDF membranes, and detected by Western blotting as previously described. Briefly, the membranes were incubated overnight at 4°C with antibodies specific for the following proteins: CDCP1 (dilution 1:1000, 4115, Cell Signaling or 1:1000, PA5–17245, Invitrogen), GAPDH (1:7000, MAB374, Millipore Sigma), β-actin (1:5000, 4970, Cell Signaling), Dectin-2 (1.5 μg/mL, MAB1525, R&D Systems, Minneapolis, MN, USA), P-Src (Tyr416; 1:1000, 6943, Cell Signaling), Src (1:1000, 2109, Cell Signaling), P-Syk (Tyr525/526; 1:1000, 2710, Cell Signaling), Syk (1:1000, 2712, Cell Signaling), NLRP3 (1:1000, 15101; Cell Signaling), P-p65 (Ser536; 1:1000, 3033, Cell Signaling), p65 (1:1000, 8242, Cell Signaling), P-Erk1/2 (Thr202/Tyr204; 1:1000, 9101, Cell Signaling), Erk1/2 (1:1000, 4695, Cell Signaling), P-p38 (Thr180/Tyr182; 1:1000, 4511, Cell Signaling), p38 (1:1000, 9212, Cell Signaling), P-JNK (Thr183/Tyr185; 1:1000, 4668, Cell Signaling), JNK (1:1000, 9252, Cell Signaling). HRP-conjugated anti-rabbit IgG (1:5000, 4030–05, Southern Biotech, Birmingham, AL, USA), anti-rat IgG (1:3000, 6180–05, Southern Biotech), and anti-mouse IgG (1:4000, 405306, Biolegend) were used as secondary antibodies. Chemiluminescent images of membranes were acquired using darkroom development techniques.

Immunofluorescence staining

Each sacrificed mouse was perfused with 10 mL of ice-cold 0.01 M PBS, followed by 20 mL of ice-cold 4% paraformaldehyde (PFA). Hearts and colons were removed, further fixed with 4% PFA for 24 h, and washed in a 30% sucrose until the tissues sank in the solution. Fixed hearts and colons were embedded in OCT medium (Thermo Fisher Scientific, Waltham, MA, USA) and stored at −80°C until processed. Sagittal sections (14 μm thickness) were prepared and processed for morphological studies.

For immunofluorescence staining, sections were first rehydrated and then incubated for 1 h in PBS containing 0.3% Triton X-100, 10% normal donkey serum, 1% BSA, and M.O.M blocking solution (Vector Laboratories, Burlingame, CA, USA), followed by incubation with a mouse anti-CDCP1 primary antibody (Sigma, clone 9A2) diluted in pre-absorption solution (1:500) overnight at 4°C. After three washes in PBS, the sections were incubated with a FITC-conjugated anti-mouse IgG (secondary antibody; dilution, 1:500, Jackson ImmunoResearch, West Grove, PA, USA) for 2 h. The sections were then mounted with DAPI Fluoromount-G (Abcam, Cambridge, UK). Images were obtained using a BZ-X700 microscope (Keyence, Osaka, Japan).

BMDCs isolated from WT and CDCP1 KO mice were stimulated in the presence or absence of 5 μg/mL CAWS at 37°C for 24 h. Cells were fixed with ice-cold 100% methanol for 15 minutes at −20°C, then rinsed three times in 1X PBS for 5 minutes each. After blocking for 1h in 1× PBS containing 10% normal donkey serum, 0.3% Triton™ X-100, 1% BSA, and M.O.M., the cells were incubated overnight with a mouse anti-CDCP1 (Sigma, clone 9A2) primary antibody or mouse IgG isotype control diluted in pre-absorption solution overnight at 4°C. The cells were then washed with PBS and incubated with FITC-conjugated anti-mouse IgG (secondary antibody) for 2 h. Finally, the cells were mounted on slides and counterstained with DAPI for nuclear visualization.

Statistical analyses

The data were analyzed using GraphPad Prism 7 (GraphPad Software, La Jolla, CA, USA) and are presented as group means ± SEM. The Mann–Whitney U test or Student’s unpaired t-test were used as appropriate. P values <0.05 were considered significant.

Results

CDCP1 KO mice develop attenuated coronary and aortic vasculitis after CAWS induction

To explore the potential role of CDCP1 in the molecular pathogenesis of KD, KD-like vasculitis was induced in age- and sex-matched WT and CDCP1-KO mice by injecting CAWS intraperitoneally. Hearts were harvested on day 28; histopathological examinations of the aortic root, including coronary arteries were performed (29). Although the incidence of coronary and aortic vasculitis was not significantly different between the two groups of mice, CDCP1 KO mice had significantly lower overall coronary and aortic vasculitis histopathological scores than did WT mice after KD induction (Fig. 1A–B). The absence of severe KD in CDCP1 KO mice suggested that CDCP1 is a novel regulator of this model of KD, and led to subsequent studies to explore the cellular expression and role of CDCP1 in CAWS-associated coronary and aortic vasculitis.

Figure 1:

CDCP1 deficiency reduces the severity of vasculitis in a mouse model of Candida albicans water-soluble fraction (CAWS)-induced Kawasaki disease. Wild-type C57BL/6J (WT) and CDCP1 knockout (KO) mice were injected intraperitoneally with 4 mg of CAWS and euthanized 28 days later. (A–B) Representative hematoxylin and eosin (H&E) and Masson’s trichrome (MT)-stained horizontal sections of the aortic root areas of WT and KO mice at day 28. The coronary vasculitis incidences were included in the parentheses (A) and the inflammation scores of individual mice were indicated in the graph (B). Data were pooled from three experiments. Scale bars = 200 μm. *P < 0.05. (C) Representative spleen samples from WT and KO mice at day 0 (naive) and day 28. (D) The graph depicts the relative weights of the spleens of individual mice. Data are pooled from two experiments.

CDCP1 KO mice develop splenomegaly with a decreased splenic B cell compartment after KD induction

Splenomegaly and inflammatory cell infiltration at the aortic root are hallmarks of CAWS-induced KD(30, 31). Interestingly, despite attenuated inflammatory cell infiltration at the aortic root, CDCP1 KO mice developed splenomegaly comparable to that in WT mice (Fig. 1C), and no significant intergroup difference in the spleen/body weight ratio was observed at day 28 after KD induction (Fig. 1D). Flow cytometric analysis of splenic and peripheral leukocyte profiles demonstrated reduced percentages of CD3⁻CD19⁺B cells in the CDCP KO mouse spleens but not in peripheral blood (Suppl. Fig. 1–2). Other leukocyte populations, including total CD3⁺ T cells, or CD4⁺ T cells and CD8⁺ T cell subsets, CD11b⁺CD11c⁺ DCs, CD11b⁺Gr-1^HiSSC^Int (spleen) or CD11b⁺Ly6G⁺(peripheral blood) neutrophils, CD11b⁺Gr-1^Lo-negSSC^Lo (spleen) or CD11b⁺Ly6c^Hi-negLy6G⁻SSC^Lo (peripheral blood) monocytes and macrophages, and CD11b⁺Gr-1^Lo-negSSC^Hi (spleen) or CD11b⁺Ly6C^Lo-negLy6G⁻SSC^Hi (peripheral blood) eosinophils, were comparable between at day 28 after KD induction regardless of CDCP1 protein status (Suppl. Fig. 1–2).

CDCP1 KO mice exhibit lower serum levels of CAWS-specific IgM and IgG2a after KD induction

The formation of immuno-complexes has been implicated in the pathogenesis of KD(32). Given the observed reduction in the B cell numbers in CDCP1 KO mouse spleens treated with CAWS, we analyzed serum concentrations of CAWS-specific Igs in the WT and CDCP1 KO mice after KD induction. Lower concentrations of total CAWS-specific IgM and IgG were observed in sera from CDCP1 KO mice relative to WT mice at day 28 (Fig. 2). Sequential ELISA using mAbs specific for IgM and IgG subclasses showed significantly reduced serum concentrations of CAWS-specific IgM and IgG2a in CDCP1 KO mice relative to WT mice. In contrast, the concentrations of CAWS-specific IgA, IgG1, IgG2b, and IgG3 were comparable between groups (Fig. 2).

Figure 2:

Reduced serum levels of Candida albicans water-soluble fraction (CAWS)-specific IgM and IgG2a in CDCP1 knockout (KO) mice compared to WT in Kawasaki disease (KD) mice. Levels of CAWS-specific immunoglobin subtypes and IgG subclasses were measured in sera from naïve mice or KD-induced wild-type (WT) and CDCP1 KO mice by ELISA. Data are pooled from four experiments.*P <0.05.

CDCP1 KO mice develop lower serum concentrations of the pro-inflammatory cytokine IL-6 after KD induction

Inflammatory cytokines are critical to the development of KD and KD models(33–37). We used a cytokine array to analyze 44 cytokines/chemokines in sera from 5 paired WT and CDCP1 KO mice 28 days after CAWS treatment (Fig. 3A). IL-6 and IL-1α were the only 2 cytokines identified in this array whose levels were significantly different between the 2 groups that contain 5 mice each. However, in following validation studies using respective conventional ELISAs to measure serum concentrations of IL-6 and IL-1α in all serum samples, only IL-6 was confirmed to have significantly reduced serum concentrations in CDCP1 KO mice after KD induction (Fig. 3B). We, therefore, focused on IL-6 in our sequential studies.

Figure 3:

Production of the pro-inflammatory cytokine IL-6 alone is reduced in CDCP1 knockout (KO) mice after Kawasaki disease (KD) induction. (A) Concentrations of 44 cytokines/chemokines in sera from 5 wild-type (WT) and 5 CDCP1 KO mice at day 28 after Candida albicans water-soluble fraction (CAWS) injection were measured using cytokine/chemokine arrays. (B) Concentrations of IL-6 in sera from naïve or KD-induced WT and CDCP1 KO mice (day 28) were measured by ELISA. Data are pooled from four experiments.**P <0.01. (C) Local IL-6 concentrations were measured in tissue lysates prepared from areas around the aortic root in each heart of WT and KO mice at 28 days after CAWS injection by ELISA. Data were normalized against the concentrations of the total proteins. Data are pooled from two experiments. *P <0.05.

Because previous studies indicated that local pro-inflammatory cytokines are critical for the development of cardiac vasculitis in models of KD (35, 38), the levels of IL-6 were quantified in heart protein lysates prepared from CAWS-treated WT and CDCP1 KO mice. Again, the local IL-6 levels were significantly lower in the hearts of the CDCP1 KO mice (Fig. 3C).

CDCP1 is not detectable in naïve mouse hearts

To test the hypothesis that CDCP1 expression in the heart is required for the local IL-6 production and the development of cardiac vasculitis, protein lysates from heart and colon tissues (CDCP1-positive control) were probed to detect CDCP1 using western blotting. No CDCP1 was detected in the heart lysates whereas CDCP1 bands were clearly detectable using lysates from WT but not CDCP1 KO colons (Suppl Fig. 3A) This finding was confirmed using immunohistochemistry as well (Suppl Fig. 3B).

The heart is mainly composed of cardiac myocytes, as well as endothelial cells and fibroblasts(39). CDCP1 is expressed on many types of epithelial cells, and is also detectable on several types of fibroblasts and endothelial cells, and is increased after IFN-γ stimulation(20, 40). To rigorously establish the distribution of CDCP1 in native cardiac cells, primary cardiac fibroblasts were isolated from WT and CDCP1 KO mice as published (26). After confirming that the purity of the isolated fibroblasts (~90%) by flow cytometry (Suppl Fig. 3C), cells were cultured with or without IFN-γ or CAWS for 2 days, and cell lysates subjected to western blotting to detect CDCP1. CDCP1 could not be detected in cell lysates prepared from cultured cardiac fibroblasts (from WT mice), regardless of the stimulation status (Suppl Fig. 3D). To determine the expression of CDCP1 in cardiac endothelial cells, aortas were isolated from naïve WT and CDCP1 KO hearts and protein lysates prepared. CDCP1 was again undetectable using western blotting (Suppl Fig. 3A).

CDCP1 is expressed in DCs and upregulated by CAWS stimulation

The above results demonstrate that constitutive CDCP1 is absent in cardiac cells. However, a recent report showed that CDCP1 is expressed in human monocyte-derived DCs(41). Because murine DCs, but not peritoneal macrophages, bone marrow-derived (BM) macrophages, or other splenocytes can respond to CAWS (34), data mining was used to explore whether CDCP1 is expressed in mouse DCs; CDCP1 RNA transcripts were reported to be abundant in mouse BMDCs in the Gene Expression Omnibus (not shown). To validate the data mining results, BMDCs of WT and CDCP1 KO mice were generated and RNAs purified. CDCP1 transcripts in WT but not CDCP1 KO BMDCs were detected by RT-PCR ( Fig. 4A). To determine if CDCP1 was modulated by CAWS, we stimulated BMDCs with CAWS for use in western blotting and immunofluorescent staining. Consistent with the RT-PCR results, CDCP1 protein was constitutively expressed and was upregulated following exposure to CAWS for 24 hr in WT but not in CDCP1 KO BMDCs (Fig. 4B–C). In addition, infiltrating CD11c+ cells around the aortic root in WT mice after KD induction were also positive for CDCP1 in immunofluorescent staining assays (Fig. 4D). Furthermore, CDCP1 was detectable on monocyte-derived human DCs and its expression upregulated upon CAWS stimulations as analyzed by both flow cytometric (Fig. 4E) and immunofluorescent staining assays (Fig. 4F). These data, taken together, established that CDCP1 is present on DCs and its expression upregulated upon CAWS stimulations.

Figure 4.

Dendritic cells (DCs) constitutively express CDCP1 and upregulate CDCP1 expression following Candida albicans water-soluble fraction (CAWS) stimulation. (A) Total RNA was isolated from bone marrow-derived DCs (BMDCs) prepared from wild-type (WT) and CDCP1 knockout (KO) mice; RT-PCR was used to detect CDCP1 transcripts (519 bp amplicon). NT: water control; +RT: RNA with reverse transcription; -RT, RNA without reverse transcription. (B) Cell lysates were prepared from WT and CDCP1 KO BMDCs stimulated with or without 5 μg/mL CAWS for 24 h. CDCP1 protein levels were determined by Western blotting. Colon lysates from WT and CDCP1 KO mice were included as controls. (C) BMDCs treated as in (B) were stained with an anti-CDCP1 mAb (green) or isotype control, counterstained with DAPI (blue), and examined using fluorescence microscopy. Scale bar: 5 μm. (D) Co-localization of CDCP1 (green) and DC marker CD11c (red) in the infiltrated cells around the aortic root in a WT mouse after KD induction. DAPI was used for nuclear counterstaining (blue). (E) CDCP1 was also detectable on human monocyte-derived DCs and upregulated by CAWS stimulation as measured by flow cytometry, and (F) by immunofluorescent staining. Scale bar: 100 μm.

The absence of CDCP1 impairs the CAWS-induced production of IL-6 in DCs

To determine the effect of CDCP1 on the production of IL-6 by DCs, BMDCs from WT and CDCP1 KO mice were stimulated with a range of concentrations of CAWS (0.1–10 μg/mL) or 100 ng/mL lipopolysaccharide (LPS, a control) for 48 hours. CDCP1 protein expression was quantified using western blotting and IL-6 in culture supernatants was quantified using ELISA. Stimulation with CAWS, but not LPS, upregulated CDCP1 expression (Fig. 5A). LPS stimulated the DCs to produce IL-6 and the absence of CDCP1 increased the LPS-stimulated DC production of IL-6 (Fig 5B). In contrast, CAWS treatment induced the productions of IL-6 from the DCs in a concentration-dependent manner, and the absence of CDCP1 significantly reduced the CAWS-stimulated DC production of IL-6 (Fig. 5C).

Figure 5.

Elevated IL-6 is produced by dendritic cells (DCs) in response to Candida albicans water-soluble fraction (CAWS) stimulation and, was impaired in the absence of CDCP1. Identical numbers of wild-type (WT) and CDCP1 knockout (KO) BMDCs (1×10⁶) were stimulated for 48 h with 100 ng/mL LPS or 0–10 μg/mL CAWS. Lysed cells were subjected to western blotting to detect CDCP1, and supernatants were collected and subjected to ELISA to measure IL-6. (A) western blotting confirmed that CDCP1 was constitutively expressed in WT BMDCs and upregulated by CAWS stimulation, but not by LPS stimulation. Colon lysates from WT mice were included as a control (B) IL-6 in the DC supernatants before and after LPS stimulation as measured by ELISA (C) IL-6 in the DC supernatants before and after CAWS stimulation measured by the same ELISA.

The absence of CDCP1 impairs Syk–MAPK activation in CAWS-stimulated DCs

LPS and CAWS differentially affect DC IL-6 production depending on the presence or absence of CDCP1, and CDCP1 was not necessary for LPS-induction of IL-6 from DCs whereas it attenuated CAWS-dependent DC production of IL-6. These results suggest that LPS and CAWS engage different underlying signaling pathways. TLR4 is the main receptor for LPS(42), whereas Dectin-2 is reported to be the receptor for CAWS on BMDCs(43). Upon ligation, Dectin-2 induces the production of pro-inflammatory cytokines by activating NF-κB and MAPK dependent signaling pathways and NLRP3 inflammasomes(44). To distinguish differences in the underlying mechanisms of CAWS-induced cytokine production from those invoked by LPS stimulation, surface expression of Dectin-2 on WT and CDCP1 KO BMDCs was quantified. Both western blotting and flow cytometry demonstrated that Dectin-2 expression regardless of CDCP1 expression (Suppl. Fig. 4). When WT and CDCP1 KO BMDCs were stimulated with 5 μg/mL CAWS for a range of periods and cell lysates probed with antibodies specific for signal transduction molecules, Syk was significantly phosphorylated (P-Syk) at 15 min in WT BMDCs but not in CDCP1 KO BMDCs. The three MAPK family kinases, P-Erk1/2, P-p38, and P-JNK were significantly activated at 30 min in WT but not in CDCP1 KO BMDCs. NLRP3 activation was not observed. A trend towards higher levels of NF-kb p65 phosphorylation were observed in KO BMDCs relative to WT BMDCs at 30 and 60 min. P-Src phosphorylation was also both impaired and delayed in CDCP1 KO BMDCs relative to the WT counterparts (Fig. 6).

Figure 6.

Syk–MAPK pathway signaling is impaired in the CDCP1-deficient dendritic cells (DCs). Wild-type (WT) and CDCP1 knockout (KO) bone marrow-derived DCs were challenged with 5 μg/mL Candida albicans water-soluble fraction for the indicated time, lysed, and analyzed by western blotting to detect the indicated proteins. The activation of Syk-dependent signaling molecules in the Dectin-2 pathway, include NF-κB (p65), MAPKs (Erk, p38, JNK), the NLRP3 inflammasome, and Src, which is necessary for intrinsic CDCP1 activation. The phosphorylated (activated) and the unphosphorylated total amount were assessed using respective antibodies where appropriate. β-actin was used as the loading control. The uncut gels from the western blots of syk, p38 and p65 are presented in Suppl. Fig. 5.

Discussion

In this study, we found that CDCP1 KO mice developed attenuated cardiac vasculitis after CAWS administration. In addition, the production of both CAWS-specific IgM/IgG2b and IL-6 were reduced in the CDCP1 KO mice after KD induction. Mechanistically, CDCP1 is not expressed constitutively in the heart, but is expressed by DCs and upregulated upon CAWS stimulation. Furthermore, CDCP1 is required for CAWS-induced IL-6 production triggered via the Syk–MAPK signaling pathway. These results demonstrate for the first time that CDCP1 is a novel regulator of the development of CAWS-induced KD, and provide new insight into the etiopathogenesis of KD.

CDCP1 was originally discovered as a prognostic marker in patients with certain solid tumors, and almost all existing studies on this protein have focused on its direct role in tumor cell survival and metastasis(15). Accordingly, the potential roles of CDCP1 in the immune system are obscure. As KD is established as an immune disease, the novel finding that CDCP1 KO mice were protected from cardiac vasculitis in our KD model further supports a novel immune regulatory role for CDCP1. We previously reported the identification of CDCP1 as a new ligand of CD6 and a key regulator of T cell responses, thus providing the first evidence supporting CDCP1 as a novel immune regulator(20). Consistent with those findings, we observed that CDCP1 KO mice were protected in a model of multiple sclerosis, as indicated by attenuated pathogenic T cell responses and reduced T cell infiltration in the spinal cord. Furthermore, we observed significantly upregulated soluble CDCP1 concentrations in the synovial fluid from patients with inflammatory juvenile idiopathic arthritis and rheumatoid arthritis, but not in samples from patients with osteoarthritis, and identified CDCP1 as a T cell chemoattractant(20). These observations suggest that CDCP1 might interact with CD6 on the T cell surface to facilitate T cell activation and migration(20). As noted previously, CDCP1 expression was significantly upregulated in children with SARS-CoV-2 infection who developed acute vasculitis(21). These latest clinical observations, together with our results from our CDCP1 KO mouse model of KD, further indicate a previously unknown role of CDCP1 in immune regulation and strongly indicate that this protein is integrally involved in the pathogenesis of KD.

Animal models are invaluable for understanding the pathogenesis of this disorder and testing new potential treatments(45). Limited availability of specimens from KD patients, combined with ethical considerations such as 1) the need to demonstrate the possibility of direct benefit “first in children” clinical trials and 2) the need to justify direct study in pediatric diseases only when there is no adult equivalent, increase the stringency of the requirement to show premise, prior to administering investigational agents to children. Although no animal model can faithfully replicate every aspect of human KD, CAWS-induced cardiac vasculitis is one of the most commonly utilized KD models(46), together with the LCWE- (47)and NOD1-induced KD models(48). CAWS is composed mainly of β-glucan and α-mannan(22). When injected intraperitoneally, CAWS stimulates the production of inflammatory cytokines, including IL-1, GM-CSF, TNF-α, and IL-6, which induce the cardiac vasculitis that resembles human KD(35–37, 49). Human KD is known to be mediated in part by IL-6 which contributes to fever, followed by anemia, and thrombocytosis. Of the 44 inflammatory cytokines/chemokines studied herein, only IL-6 was significantly downregulated in CDCP1 KO mice challenged to induce KD. The production of both CAWS-specific IgM/IgG2b and IL-6 is reduced in KD mice that lacked CDCP1, though other aspects of immunity were comparable to those in WT mice. Taken together our data suggest that IL-6 is a key cytokine underlying the pathogenesis of KD and is consistent with previous reports showing increased IL-6 production in KD patients and models(37, 50, 51).

Decades of research have revealed the successful clinical blockade of IL-6R or IL-6 as a treatment for autoimmune diseases such as rheumatoid arthritis(52) and systemic juvenile idiopathic arthritis(53). Despite previously published data related to a role for IL-6 in KD and autoimmune diseases, only one small pilot clinical study of 4 patients evaluated the efficacy of IL-6R blockade as a treatment option for KD with inconclusive results (54). The new insights from this study showing the correlation of IL-6 levels with vasculitis severity in our KD mice, and the association of reduced IL-6 production with attenuated coronary vasculitis in CDCP1 KO mice indicate that future detailed evaluations of IL-6-targeted mono- and/or combined therapies for KD are warranted.

Few studies have explored the normal tissue distribution of CDCP1 and its potential role in non-tumor cells, including myeloid cells. Despite that data showing that CDCP1 expression is upregulated in human monocyte-derived DCs upon activation has been published by two different investigative teams(41, 55), no follow-up studies have been reported. We also observed that mouse DCs express CDCP1 and this discovery of CDCP1 expression by DCs further strengthens our novel hypothesis that CDCP1 acts as an immune regulator. Among mouse leukocytes, only DCs respond to CAWS(34). We observed significant upregulation of CDCP1 on DCs in response to treatment with CAWS, but not in response to LPS. Interestingly, LPS and CAWS stimulation differed concerning the production of IL-6 by DCs. These results strongly indicate that distinct stimuli work via different mechanisms of action within the DCs.

CAWS binds to Dectin-2 on DCs to initiate a series of cellular responses beginning with the association of the ITAM-containing adaptor molecule FcRγ, followed by the recruitment of phosphorylated Syk(44). CDCP1 was not detectable in naïve mouse hearts but was present on the infiltrating DCs in the hearts from mice with induced KD. After excluding a possible difference in Dectin-2 expression between the WT and CDCP1 KO DCs, we observed impaired Syk phosphorylation and downstream MAPK activation upon CAWS stimulation in the absence of CDCP1. However, impaired Syk phosphorylation cannot explain the increased NF-kB activation observed previously in CDCP1 KO BMDCs(56). CDCP1 has been implicated as a tumor cell regulator via Src and PKCδ; these kinases then activate FAK tyrosine kinases, which play a major role in integrin signaling(15) and tunes adhesion(57). In general, Src family and Syk kinases tend to cooperate in signaling pathways, such that the Src-family is “upstream” of Syk or activated first in response to pathogen detection(58). Here, we observed delayed and impaired Src phosphorylation in CDCP1 KO DCs. The activation of Syk activation also induces the production of reactive oxygen species(59), which are important for activation of the NLRP3 inflammasome and the enhanced processing of pro-IL-1β into mature IL-1β. We did not observe NLRP3 activation, which explained the lack of IL-1β induction by CAWS(59).

Our results contribute to an expanding body of knowledge about how Kawasaki disease evolves and may also inform future studies in SARS-CoV-2-induced vasculitis in children. CDCP1 KO mice develop attenuated cardiac vasculitis in the context of CAWS-induced KD, and these decreases in severity are associated with a reduction in IL-6 production. Mechanistically, CDCP1 is expressed on DCs and is inducible by CAWS, even though CDCP1 itself is not expressed constitutively in the heart cells. The absence of CDCP1 expression impaired CAWS-induced upregulation of IL-6 production by DCs, which appears to be mediated by reduced Syk phosphorylation and subsequent downstream MAPK activation. Taken together, our results reveal a novel, previously unrecognized role for CDCP1 in the pathogenesis of mouse KD and provide additional evidence supporting the function of CDCP1 as a novel immune regulator. Further, the data suggest that reducing CDCP1-dependent signals may be desirable in the management of KD.

Key points:

1. CDCP1 is detectable on DCs and its expression upregulated by CAWS stimulation

2. Absence of CDCP1 attenuates cardiac vasculitis in a mouse model of Kawasaki disease

Non-standard Abbreviations and Acronyms:

KD

Kawasaki disease

SARS-CoV-2

severe acute respiratory syndrome coronavirus-2

IVIG

intravenous immunoglobulin

CDCP1

CUB domain-containing protein 1

CAWS

Candida albicans water-soluble fraction

WT

wild-type

KO

knockout

DC

dendritic cell

H&E

hematoxylin and eosin

MT

Masson’s trichrome

BMDC

bone marrow-derived dendritic cell

DPBS

Dulbecco’s phosphate-buffered saline

MCF

mouse cardiac fibroblast

FMO

fluorescence minus one

ELISA

enzyme-linked immunosorbent assay

BSA

bovine serum albumin

HRP

horseradish peroxidase

RT-PCR

reverse transcription-polymerase chain reaction

PFA

paraformaldehyde

LPS

lipopolysaccharide