Pathway-Specific Defects in T, B, and NK Cells and Age-Dependent Development of High IgE in Mice Heterozygous for a CADINS-Associated Dominant Negative CARD11 Allele

Ag recognition by TCR and BCR complexes triggers multiple branching signaling pathways that are crucial for lymphocyte activation, proliferation, and survival during the adaptive immune response. Impaired signaling through these pathways can lead to immunodeficiency, increasing susceptibility to infections and making them difficult to clear.

CARD11 is a multidomain scaffold protein that serves as a critical hub during Ag receptor signaling, integrating the signals it receives from Ag receptors on the surface of T and B cells and translating them into activation of the NF-κB, JNK, and mTOR pathways (1–10). At rest, CARD11 is kept in a closed, inactive state via an inhibitory domain that functions through four redundant repressive elements (11–13). Ag receptor engagement relieves this autoinhibition through a poorly understood process, allowing CARD11 to undergo a conformational change into an open, active scaffold (14–16). CARD11 then coordinates the recruitment and activation of a variety of cofactors including Bcl10, MALT1, HOIP, and many others (1, 11, 17–21). The assembly of this multiprotein complex and the activation of its components lead to the activation of IKK kinase activity, the phosphorylation and degradation of the inhibitory protein IκBα, and the subsequent nuclear translocation of NF-κB. The CARD11-nucleated complex also dictates the activation of JNK2 and mTORC1 (1).

The signal-inducible scaffold function of CARD11 is essential for normal Ag receptor signaling, and mutations that disrupt this function can impair the adaptive immune response. In the last several years, germline mutations in the CARD11 gene have been found to be responsible for at least three types of primary immunodeficiency in humans: CARD11 deficiency, B cell expansion with NF-κB and T cell anergy (BENTA), and CARD11-associated atopy with dominant interference of NF-κB signaling (CADINS) (1, 22–24). CARD11 deficiency is caused by homozygous loss-of-function (LOF) mutations in CARD11, and the seven patients reported thus far presented with severe Pneumocystis jirovecii pneumonia, hypogammaglobulinemia, and poor in vitro T and B cell activation despite having normal lymphocyte counts (24–29). BENTA is caused by heterozygous gain-of-function (GOF) mutations in CARD11 and is characterized by a massive expansion of B cells as well as recurrent viral infections and weak B and T cell responses to Ag challenges (30–36). CADINS is caused by heterozygous LOF mutations in CARD11 that act in a dominant negative manner to interfere with the function of the coexpressed wild-type (WT) CARD11 allele (37–40). Patients with this condition exhibit an enhanced susceptibility to infections as well as a cohort of allergy-related symptoms known collectively as atopy.

In the last few years, at least 14 different dominant negative CARD11 mutations have been reported in 45 CADINS patients from nearly 20 unrelated families (37–40). The majority (12 out of 14) of these variants are located throughout the CARD and coiled-coil domains of CARD11, whereas the remaining two variants are located in the C-terminal guanylate kinase (GUK) domain. Nearly all of the patients presented with some form of atopy, with the most common symptoms being severe atopic dermatitis, asthma, and food and/or environmental allergies. In addition, laboratory tests revealed elevated serum IgE levels and high eosinophil counts in most patients. Despite having normal lymphocyte numbers, a history of viral skin infections such as molluscum contagiosum was also commonly reported, as well as recurrent respiratory tract infections, bouts of pneumonia, and bronchiectasis. Functional studies in patient lymphocytes have shown defective lymphocyte activation, poor responses to Ag challenges, and attenuated signaling to NF-κB and, to a lesser extent, mTORC1. However, these functional studies have only been performed in less than half of the patients. A fuller understanding of precisely how dominant negative CARD11 alleles lead to immunodeficiency and atopy is needed.

In this study, we investigate the impact of a dominant negative CARD11 allele on the immune system using a mouse model expressing the CARD11 R30W mutation from the endogenous locus. The CARD11 R30W allele was first characterized in four patients from the same family who all presented with immunodeficiency and atopy (38). Subsequently, as part of a larger study, a fifth patient was identified (39). Studies on patient lymphocytes have shown impaired T cell activation and weak responses to various Ags and mitogens as compared with lymphocytes taken from healthy controls. However, it is difficult to determine whether the observed patient phenotypes are affected by confounding factors, such as environmental exposures, past and current treatments for symptoms, and the possible presence of modifier genes. In this report, we use CARD11^R30W/+ mice to examine how a dominant negative CARD11 mutation disrupts immune cell signaling and effector functions in the absence of these confounding factors. We show that heterozygous presence of the CARD11 R30W allele has diverse effects on both the signaling pathways downstream of CARD11 and the activation and function of T cells, B cells, and NK cells, leading to immunodeficiency and an incompletely penetrant age-dependent increase in serum IgE.

The CARD11 R30W allele was generated using CRISPR/Cas9 genome editing (41). A synthetic CRISPR RNA was designed to target the sequence 5′-TGTAACGGCTCAGCATGTGC-3′. The DNA oligonucleotide used as a template for homology-directed repair was 5′-GACTACATG GAGACGCTGAAGGATGAAGAGGAGGCCCTATGGGATAACGTGGAATGCAACtGGCACATGCTGAGCCGTTACATCAACCCCGCCAAGCTCACCCCCTACCTGCGCCAGTGC-3′ (lowercase letter indicates mutation). Introduction of the mutation removed an endogenous BsrFαI restriction site and introduced a novel BsrI restriction site. The CRISPR RNA, transactivating CRISPR RNA, recombinant Cas9, and homology-directed repair template oligo were injected into C57BL/6J embryos generated by the Transgenic Core Laboratory at the Johns Hopkins University School of Medicine. PCR followed by overnight digestion with BsrI or BsrFαI was performed on DNA from the resulting cohort of 41 live pups to determine which mice carried the CARD11 R30W mutation. The region surrounding R30 in the CARD domain of CARD11 was sequenced in these mice to confirm the presence of the CARD11 R30W allele, and two founders were selected based on the sequencing results. Founders were crossed to C57BL/6J mice, and germline transmission of the CARD11 R30W allele was confirmed in this and all subsequent generations via PCR followed by overnight BsrI or BsrFαI digestion. All mice used in experiments were at least 6 wk of age. Mice were matched by age, sex, and litter within experiments whenever possible. All mice were maintained, and all mouse procedures were conducted in accordance with the Johns Hopkins University Institutional Animal Care and Use Committee.

Abs used for Western blotting and coimmunoprecipitations were as follows: mouse mAb against phospho-IκBα S32/S36 (5A5, 9246), rabbit polyclonal Ab against IκBα (9242), rabbit mAb against phospho-IKKα/β S176/S177 (C84E11, 2078S), rabbit polyclonal Ab against IKKβ (2684), rabbit mAb against GAPDH (D16H11, 5174), rabbit mAb against CARD11 (1D12, 4435), rabbit polyclonal Ab against phospho-SAPK/JNK T183/Y185 (9251), rabbit polyclonal Ab against JNK2 (4672), rabbit mAb against c-Jun (60A8, 9165), rabbit mAb against phospho-S6 S235/S236 (D57.2.2E, 4858), rabbit mAb against phospho-S6 S240/S244 (D68F8, 5364), and rabbit mAb against S6 (5G10, 2217) were from Cell Signaling Technology. Mouse mAb against HOIL-1 (H-1; sc-393754), mouse mAb against CYLD (E-10; sc-74435), and mouse mAb against Bcl10 (331.3, sc-5273) were from Santa Cruz Biotechnology.

Mice were immunized by i.p. injection between 8 and 12 wk of age with 25 μg of DNP-Ficoll (Biosearch Technologies) in PBS or 50 μg of DNP-CGG (Biosearch Technologies) precipitated in Imject Alum (Thermo Fisher Scientific). Blood samples were taken prior to immunization as well as 7 and 14 d after immunization. Serum was isolated from whole blood by allowing blood to clot at room temperature for 20 min, then centrifuging at 2000 × g for 10 min at 4°C. Ab responses were measured by DNP-specific ELISA. Diluted serum was incubated on plates coated with 5 μg/ml DNP-BSA (Biosearch Technologies), and DNP-specific Abs were detected using HRP-conjugated secondary Abs against the indicated mouse isotypes.

Spleens were harvested into RPMI 1640 media supplemented with 10% FBS and 2% penicillin/streptomycin/l-glutamine and dissociated using the frosted side of glass microscope slides. Single-cell suspensions were obtained by passing dissociated organs over a 70-μm cell strainer. RBCs were lysed with ACK lysis buffer (Quality Biological), and the final cell pellet was resuspended in 1× PBS for counting. CD4⁺ T cells, B cells, or NK cells were purified from whole splenocytes by negative selection using the mouse CD4⁺ T cell Isolation Kit, Pan B Cell Isolation Kit II, or NK Cell Isolation Kit (Miltenyi Biotec) according to the manufacturer’s instructions. T cells and B cells were stimulated directly after isolation, whereas NK cells were expanded in supplemented RPMI containing 50 ng/ml recombinant mouse IL-15 (210-15; PeproTech) for 6 d prior to stimulation. A total of 2 × 10⁶ T, B, or NK cells were resuspended in supplemented RPMI and allowed to rest for 30 min at 37°C, then stimulated for the indicated times at 37°C with 1 μg/ml hamster anti-mouse CD3 (145-2C11, 553058; BD Pharmigen), 2 μg/ml hamster anti-mouse CD28 (37.51, 553295; BD Pharmigen), and 1 μg/ml mouse anti-hamster IgG1 (HIG-632, 550637; BD Pharmigen), 25 nM PMA (Sigma-Aldrich), and 62.5 nM ionomycin (Sigma-Aldrich), or 0.1 μg/ml anti-mouse IgM (16-5092-85; Invitrogen). Cells were incubated for 10 min on ice, spun at 5000 rpm for 5 min, and lysed in 30 μl of immunoprecipitation lysis buffer (11) containing protease inhibitor mixture (P8340; Sigma-Aldrich) alone or with Halt phosphatase inhibitor mixture (Thermo Fisher Scientific). Samples were heat-treated at 95°C in SDS loading buffer, resolved via SDS-PAGE, and transferred to polyvinylidene difluoride membranes. Western blotting was performed with the indicated Abs and protein bands were quantified using ImageJ software. The results shown are representative of at least two replicates of each experiment.

Spleens were harvested and processed as described above. T cells or B cells were purified from whole splenocytes by negative selection using the mouse CD4⁺ T cell Isolation Kit or Pan B Cell Isolation Kit II (Miltenyi Biotec) according to the manufacturer’s instructions. Purified T or B cells were resuspended in supplemented RPMI and allowed to rest for 30 min at 37°C, then stimulated with 100 nM PMA (Sigma-Aldrich) and 1 μM ionomycin (Sigma-Aldrich) for 30 min at 37°C. Cells were incubated for 10 min on ice, spun at 1400 rpm for 5 min, and lysed in 1 ml of immunoprecipitation lysis buffer with protease inhibitor mixture (P8340; Sigma-Aldrich). Lysates were incubated with 2 μg of mouse anti-Bcl10 Ab for 4 h at 4°C with rotation, then incubated with 20 μl of bed volume of native protein A-conjugated Sepharose (GE Healthcare) overnight at 4°C with rotation. Immunoprecipitates were washed four times for 10 min at 4°C with rotation. Immunoprecipitates and cell lysate input samples were heat-treated at 95°C in SDS loading buffer, resolved by SDS-PAGE, and transferred to polyvinylidene difluoride membranes. Western blotting was performed with the indicated Abs.

Organs were harvested and processed as described above. For measuring lymphocyte numbers and thymic T cell development, whole splenocytes or thymocytes were washed in FACS buffer (PBS [pH 7.4], 0.5% BSA, 2 mM EDTA, 0.02% sodium azide), then stained in FACS buffer for 30 min on ice with CD19-allophycocyanin-Cy7 (1D3, BD Biosciences), B220-PerCP-Cy5.5 (RA3-6B2, BD Biosciences), CD3-allophycocyanin (145-2C11, BD Biosciences), CD4-BV421 (GK1.5, BD Biosciences), CD8-FITC (53-6.7, BD Biosciences), TER-119–allophycocyanin-Cy7 (TER-119, BD Biosciences), Gr-1–allophycocyanin-Cy7 (RB6-8C5, BD Biosciences), NK1.1-PE (PK136, BD Biosciences), and CD49b-Pacific Blue (DX5, BioLegend). For measuring upregulation of activation markers, T cells or B cells were purified from whole splenocytes by negative selection using the mouse CD4⁺ T Cell Isolation Kit or Pan B Cell Isolation Kit II (Miltenyi Biotec) according to the manufacturer’s instructions. A total of 1 × 10⁵ cells per sample were stimulated in 96-well plates in supplemented RPMI with or without the indicated concentrations of plate-bound hamster anti-mouse CD3 (145-2C11, 553058; BD Pharmigen) and 2 μg/ml hamster anti-mouse CD28 (37.51, 553295; BD Pharmigen) for T cells or anti-mouse IgM (16-5092-85; Invitrogen) for B cells. Cells were washed in FACS buffer and surface stained in FACS buffer for 30 min on ice with CD25-allophycocyanin (PC61, BD Pharmigen) for T cells or CD86-BV421 (GL-1, BioLegend) for B cells. For peritoneal B1 cell staining, peritoneal cavity cells were isolated by peritoneal lavage with 5 ml of 1× PBS + 3% FBS. Cells were washed with FACS buffer, then incubated in Fc block (2.4G2, BD Biosciences) for 10 min at room temperature. After washing again in FACS buffer, cells were stained in FACS buffer for 30 min on ice with CD11b-BV421 (M1/70, BD Biosciences), B220-PerCP-Cy5.5 (RA3-6B2, BD Biosciences), CD5-allophycocyanin (53-7.3, BD Biosciences), and IgM-PE (RMM-1, BioLegend). For staining of marginal zone and follicular B cells, whole splenocytes were washed with FACS buffer, then incubated in Fc block (2.4G2, BD Biosciences) for 10 min at room temperature. After washing again in FACS buffer, cells were stained in FACS buffer with CD19-allophycocyanin-Cy7 (1D3, BD Biosciences), B220-PerCP-Cy5.5 (RA3-6B2, BD Biosciences), IgM-PE-Cy7 (R6-60.2, BD Biosciences), IgD-allophycocyanin (11-26c.2a, BD Biosciences), CD21-FITC (7E9, BioLegend), and CD23-PE (B3B4, BD Biosciences). For regulatory T cell (Treg) staining, 1 × 10⁶ whole splenocytes or thymocytes per sample were plated in 96-well round-bottom plates and stained with Live/Dead Aqua fixable viability dye (Invitrogen) in 1× PBS for 30 min at room temperature, then washed with FACS buffer. Cells were surface stained in FACS buffer for 10 min at room temperature with CD3-FITC (145-2C11, BD Biosciences), CD4-BV-421 (GK1.5, BD Biosciences), CD8-PE-Cy7 (53-6.7, BD Biosciences), and CD25-allophycocyanin (PC61, BD Biosciences), then fixed and permeabilized using the FoxP3/Transcription Factor Staining Buffer Set (Invitrogen). Intracellular staining for FoxP3-PE (FJK-16s; Invitrogen) was performed in 1× permeabilization buffer for 30 min on ice. Cells were washed twice in 1× permeabilization buffer, then resuspended in FACS buffer for data collection. All flow cytometry data were collected on a BD LSR II or BD FACSCelesta flow cytometer and analyzed using FlowJo software (TreeStar).

For cytokine ELISAs, spleens were harvested and processed as described above. CD4⁺ T cells, CD8⁺ T cells, or NK cells were isolated from whole splenocytes using the mouse CD4⁺ T Cell Isolation Kit, mouse CD8a⁺ T Cell Isolation Kit, or NK Cell Isolation Kit (Miltenyi Biotec) according to the manufacturer’s instructions. T cells were stimulated directly after isolation, whereas NK cells were either stimulated directly after isolation or expanded in supplemented RPMI containing 50 ng/ml recombinant mouse IL-15 (210-15; PeproTech) for 6 d prior to stimulation. A total of 1 × 10⁵ (T cells) or 5 × 10⁴ (NK cells) purified cells per sample were stimulated in 96-well plates in supplemented RPMI with or without the indicated concentrations of plate-bound hamster anti-mouse CD3 (145-2C11, 553058; BD Biosciences) and 2 μg/ml hamster anti-mouse CD28 (37.51, 553295; BD Biosciences) for T cells or 5 μg/ml anti-mouse NKG2D (A10, 16-5872-82; Invitrogen) and 50 ng/ml recombinant human IL-2 (589104; BioLegend) for NK cells. After 24 h, culture supernatants were harvested by centrifugation at 1400 rpm for 5 min. Cytokine concentrations in culture supernatants were measured using the mouse IL-2–uncoated ELISA Kit (88-7024-88; Invitrogen), mouse IFN-γ–uncoated ELISA Kit (88-7314-88; Invitrogen), mouse IL-4–uncoated ELISA Kit (88-7044-88; Invitrogen), or mouse IL-13–uncoated ELISA Kit (88-7137-88; Invitrogen) according to the manufacturer’s instructions. For serum ELISAs, serum was isolated from whole blood as described above. The amount of the indicated Ig isotypes in serum samples was measured using the mouse IgE-uncoated ELISA Kit (88-50460-88; Invitrogen), mouse IgM-uncoated ELISA Kit (88-50470-88; Invitrogen), mouse IgG1-uncoated ELISA Kit (88-50410-88; Invitrogen), mouse IgG2a-uncoated ELISA Kit (88-50420-88; Invitrogen), mouse IgG2b-uncoated ELISA Kit (88-50430-88; Invitrogen), or mouse IgG3-uncoated ELISA Kit (88-50440-88; Invitrogen).

Spleens were harvested and processed as described above. Naive CD4⁺ T cells were purified from whole splenocytes by negative selection using the mouse Naive CD4⁺ T Cell Isolation Kit (Miltenyi Biotec) according to the manufacturer’s instructions. A total of 2 × 10⁶ cells per sample were stimulated in 24-well plates in supplemented RPMI with 1 μg/ml plate-bound hamster anti-mouse CD3 (145-2C11, 553058; BD Pharmigen) and 1 μg/ml plate-bound hamster anti-mouse CD28 (37.51, 553295; BD Pharmigen) for 2 d. For Th1 skewing, cultures contained 10 ng/ml mouse IL-12, 10 ng/ml mouse IL-2, and 10 μg/ml anti-IL-4 (Cytobox Th1; Miltenyi Biotec). For Th2 skewing, cultures contained 10 ng/ml mouse IL-4, 10 ng/ml mouse IL-2, and 10 μg/ml anti–IFN-γ (Cytobox Th2; Miltenyi Biotec). On day 2, cells were removed from the stimuli and split 1:4 in media containing fresh skewing cytokines and Abs. On day 4, cells were washed with PBS, counted, and 1 × 10⁵ cells per sample were restimulated with 1 μg/ml plate-bound hamster anti-mouse CD3. After 24 h, culture supernatants were harvested by centrifugation at 1400 rpm for 5 min and the concentrations of IFN-γ, IL-4, and IL-13 were measured by ELISA using the kits described above.

Induction of atopic dermatitis-like skin inflammation via topical application of MC903 was performed as previously described (42). MC903 (Tocris 2700) was used at a working concentration of 45 μM. Mice were anesthetized via isoflurane inhalation and 12.5 μl MC903 or 100% ethanol was pipetted onto each side of an unpunched ear once daily in the following treatment sequence: 5 d of treatment, 2 d of no treatment, 5 d of treatment, 2 d of no treatment, and then 2 d of treatment. Mice were sacrificed 24 h after the last treatment and 6-mm punch biopsies of the treated ear tissue were harvested along with the draining lymph node. The 6-mm punch biopsies were bisected, and one half was fixed in 10% formalin and embedded in paraffin. Skin cross sections were prepared and stained with H&E by the Johns Hopkins Reference Histology Laboratory. Images were taken at 20× magnification on an Olympus BX51 microscope with an Olympus DP70 color camera and ImageJ software was used to measure epidermal thickness. The other half of the punch biopsy was homogenized in T-PER Tissue Protein Extraction Reagent (Thermo Fisher Scientific) containing protease inhibitor mixture (P8340; Sigma-Aldrich). The protein concentration in the homogenate was determined using a Pierce BCA Protein Assay Kit (23227; Thermo Scientific Fisher), and 40 μg of protein per sample was used to measure the concentration of IFN-γ and IL-4 in the homogenates by ELISA using the kits described above. The draining lymph node was processed as described above for the spleen. Cells were washed with FACS buffer, then stained in FACS buffer for 30 min on ice with CD11b-BV421 (M1/70; BD Biosciences), CD4-PerCP-Cy5.5 (RM 4-5, Invitrogen), Siglec F-PE (E50-2440; BD Biosciences), Ly6C-FITC (AL-21; BD Biosciences), and Ly6G-allophycocyanin (1A8; BD Biosciences). Flow cytometry was performed on a BD FACSCelesta Flow Cytometer and data were analyzed using FlowJo software (TreeStar).

Spleens were harvested and processed as described above. NK cells were purified from whole splenocytes by negative selection using the mouse NK Cell Isolation Kit (Miltenyi Biotec) according to the manufacturer’s instructions. Purity of isolated NK cells was assessed via flow cytometry by staining for NK1.1-PE (PK136; BD Biosciences) and CD49b (DX5; BioLegend). The YAC-1 lysis assay was performed as described previously (43). YAC-1 cells were loaded with 500 nM CellTrace CFSE (Invitrogen) according to the manufacturer’s instructions. Equivalent numbers of NK1.1⁺CD49b⁺ NK cells were cocultured in 96-well plates with CFSE-loaded YAC-1 cells in the indicated E:T ratios for 4 h. Cells were then stained with Annexin V and analyzed via flow cytometry. YAC-1 lysis was calculated as the percentage of CFSE⁺Annexin V⁺ cells and is reported as the percentage of specific lysis.

To investigate the impact of a dominant negative CARD11 allele on the immune system, we generated mice bearing the CARD11 R30W allele using CRISPR/Cas9 genome editing in single-cell zygotes (41). CARD11^R30W/+ mice were born in Mendelian ratios, appeared healthy, and had normal numbers of splenocytes and thymocytes (Supplemental Fig. 1A, 1B). Our initial examination of cell populations in CARD11^R30W/+ mice also showed normal numbers of splenic B cells, CD4⁺ T cells, CD8⁺ T cells, and NK cells (Supplemental Fig. 1C–F). CARD11^R30W/+ mice exhibited no significant differences from WT in thymic double-negative, double-positive, and single-positive T cells, suggesting normal T cell development (Supplemental Fig. 1G). Patients heterozygous for the dominant negative CARD11 R30W mutation show enhanced susceptibility to a variety of infections and reduced ability to clear their infections (38, 39). To determine whether the CARD11 R30W mutation is sufficient to cause immunodeficiency in the absence of other modifier genes or environmental effects, we immunized WT and CARD11^R30W/+ mice with model Ags and monitored the levels of Ag-specific serum Abs over the course of 2 wk. When immunized with the T-independent Ag DNP-Ficoll, WT mice exhibited a robust response, producing large quantities of DNP-specific IgM and IgG3 at both day 7 and day 14 postinjection. In contrast, CARD11^R30W/+ mice failed to produce detectable levels of both Ag-specific isotypes throughout the course of the experiment (Fig. 1A). We also immunized mice with DNP-CGG, a protein Ag that requires T cell help for the production of Ag-specific Abs. DNP-specific IgM, IgG1, and IgG2a were found in the serum of immunized WT mice 7 d after immunization, and the levels of all three isotypes had increased further on day 14. CARD11^R30W/+ mice were able to produce some DNP-specific Abs against this Ag, but their response varied by isotype and was generally weaker or delayed relative to WT mice (Fig. 1B). For example, the DNP-IgM response in CARD11^R30W/+ mice was delayed compared with the response in WT mice, with lower levels of DNP-IgM detected on day 7 postinjection but levels similar to WT detected on day 14. The same delayed response was observed in CARD11^R30W/+ mice when measuring DNP-IgG1, whereas the DNP-IgG2a response was clearly reduced at both time points. These results indicate that a heterozygous dominant negative CARD11 R30W mutation is sufficient to cause immunodeficiency in the absence of other mutations or environmental factors.

To investigate the basis for immunodeficiency in mice bearing the CARD11 R30W mutation, we assessed T cell activation in response to stimulation with anti-CD3/ anti-CD28. We isolated splenic CD4⁺ T cells from WT, CARD11^R30W/+, and CARD11^R30W/R30W mice, stimulated with increasing amounts of anti-CD3 while holding the concentration of anti-CD28 constant, and measured IL-2 production by ELISA. We found that although WT T cells showed strong secretion of IL-2 even at low concentrations of anti-CD3, CARD11^R30W/+ T cells failed to produce any detectable IL-2 even at the highest concentration of anti-CD3 tested (Fig. 2A). We observed the same profound defect when measuring IFN-γ production by CARD11^R30W/+ CD8⁺ T cells stimulated in the same manner (Fig. 2B). We also measured the upregulation of the high-affinity IL-2 receptor subunit CD25 upon stimulation in purified CD4⁺ T cells via flow cytometry. CD25 was readily upregulated on the surface of WT T cells, and its signal increased steadily with the concentration of anti-CD3 used for stimulation. CARD11^R30W/+ T cells were able to upregulate CD25 slightly at higher concentrations of anti-CD3, but the signal was consistently ∼2-fold lower than in WT T cells. Furthermore, CARD11^R30W/R30W T cells showed a complete defect in CD25 upregulation (Fig. 2C, 2D). These data indicate that even one copy of the CARD11 R30W mutation confers severe defects in T cell activation.

We next examined the effect of the CARD11 R30W mutation on the CARD11-associated signaling pathways underlying T cell activation. We isolated and stimulated splenic CD4⁺ T cells, then probed key steps of multiple signaling pathways via Western blotting. Signaling to NF-κB was severely impaired in CARD11^R30W/+ T cells relative to WT T cells, as indicated by a 4.7-fold decrease in peak signal-induced IKKα/β phosphorylation, a 5.5-fold decrease in peak IκBα phosphorylation, and reduced IκBα degradation (Fig. 2E, 2F). We also observed a 2.2-fold and 2.5-fold reduction in peak MALT1-mediated cleavage of its substrates HOIL-1 and CYLD, respectively (Fig. 2G, 2H), and completely defective signal-inducible binding of Bcl10 to CARD11 (Fig. 2I, 2J). T cells from heterozygous mice express both WT and mutant CARD11 proteins that assemble into mixed oligomers of undetermined stoichiometry. These results indicate that the presence of the CARD11 R30W mutant in mixed oligomers impairs the ability of CARD11 to interact with and activate its cofactors during induced signaling in primary T cells.

Signaling through JNK2 was also disrupted in CARD11^R30W/+ T cells, as indicated by a 5.9-fold decrease in peak phosphorylation of JNK2 compared with WT T cells (Fig. 2K, 2L). However, accumulation of the transcription factor c-Jun was largely unaffected by the CARD11 R30W mutation, likely because JNK1 can still be phosphorylated and lead to c-Jun accumulation independently of CARD11 (44). In patient studies, dominant negative CARD11 alleles have been associated with defective mTOR signaling in some cases, but not others. Unlike the NF-κB and JNK pathways, we found only a modest 1.6-fold reduction in peak phosphorylation of the ribosomal protein S6 in CARD11^R30W/+ T cells when measuring signaling through mTORC1 (Fig. 2M, 2N). The profound signaling defects observed in CARD11^R30W/+ mice provide an explanation for the observed deficits in T cell activation and in turn, for aspects of their immunodeficiency.

To assess the effect of the heterozygous CARD11 R30W mutation on B cell function, we isolated splenic B cells from WT, CARD11^R30W/+, and CARD11^R30W/R30W mice, stimulated them with increasing concentrations of anti-IgM, and measured upregulation of the activation marker CD86 by flow cytometry after 24 h. We detected no difference in surface IgM expression between WT and CARD11^R30W/+ B cells (Supplemental Fig. 2). The level of CD86 detected on the surface of WT B cells increased steadily with the concentration of anti-IgM used for stimulation. While CARD11^R30W/+ B cells exhibited slightly reduced upregulation of CD86 at higher concentrations of anti-IgM, this phenotype was not as severe as the defect in CD25 upregulation observed in CARD11^R30W/+ T cells upon stimulation. In addition, CARD11^R30W/R30W B cells exhibited a slight reduction in CD86 upregulation that was similar to the heterozygous B cells (Fig. 3A, 3B). The effect of the heterozygous CARD11 R30W mutation on CARD11-associated signaling pathways was also weaker in B cells than in CD4⁺ T cells. In the NF-κB pathway, peak phosphorylation of IκBα and phosphorylation of IKKα/β were reduced by 1.4-fold and 2.0-fold respectively in CARD11^R30W/+ B cells relative to WT B cells (Fig. 3C, 3D) with no significant effect on IκBα degradation. Peak MALT1-mediated proteolysis of HOIL-1 and CYLD was not significantly reduced in CARD11^R30W/+ B cells (Fig. 3E, 3F), but inducible Bcl10 binding to CARD11 was completely abrogated (Fig. 3G, 3H). Peak phosphorylation of JNK2 was only decreased by 1.2-fold in CARD11^R30W/+ B cells after 15 min of stimulation but was reduced by 6.3-fold after 30 min of stimulation and 16-fold after 60 min of stimulation. c-Jun accumulation remained largely stable, as in T cells (Fig. 3I, 3J). Signaling through mTORC1 was not reduced in CARD11^R30W/+ B cells, as indicated by levels of S6 phosphorylation comparable to WT (Fig. 3K, 3L). These data indicate that the branching pathways downstream of CARD11 are differentially affected by the presence of CARD11 R30W in mixed CARD11 oligomers in B and T lymphocytes.

Given their diminished production of Ag-specific Abs, we next sought to determine whether CARD11^R30W/+ mice would display altered basal Ig profiles. Of the four patients heterozygous for the CARD11 R30W allele who had their Igs measured, two were found to have low serum IgM or IgG, and this phenotype is also common in patients bearing other dominant negative CARD11 alleles (37–39). We measured basal levels of IgM, IgG2b, and IgG3 in the serum of WT and CARD11^R30W/+ mice by ELISA and found that the levels of IgG2b and IgG3 in CARD11^R30W/+ serum samples were normal, but the level of IgM was reduced 4.8-fold relative to WT mice (Fig. 4A). To investigate the basis for the observed decrease in basal serum IgM levels, we used flow cytometry to examine the B1a and B1b B cell populations in the peritoneal cavity of WT and CARD11^R30W/+ mice because these “innate-like” B cells are thought to be a major source of secreted IgM in mice (45, 46). We discovered that CARD11^R30W/+ mice had 3.9-fold fewer peritoneal CD5⁺IgM⁺ B1a cells on average than their WT counterparts, while CD5^–IgM⁺ B1b cell numbers were similar in both genotypes (Fig. 4B, 4C). Because marginal zone B cells are another important source of secreted IgM (47, 48), we also assessed marginal zone and follicular B cell populations in the spleens of WT and CARD11^R30W/+ mice and found that CARD11^R30W/+ mice had 2.0-fold fewer splenic marginal zone B cells than WT mice (Fig. 4D, 4E). Follicular B cell numbers, however, were not significantly different in WT and CARD11^R30W/+ mice (Fig. 4F, 4G). These results provide an explanation for the reduced IgM in CARD11^R30W/+ mice and indicate that a dominant negative CARD11 allele can impact both peritoneal B1a cell and splenic marginal zone B cell numbers.

In addition to their immunodeficiency, patients heterozygous for the CARD11 R30W mutation also exhibit elevated serum IgE and atopy, manifested as atopic dermatitis, asthma, and food and environmental allergies (38, 39). To determine whether the CARD11 R30W mutation is sufficient to cause these phenotypes in the absence of modifier genes and environmental exposures, we measured basal IgE levels in the serum of WT and CARD11^R30W/+ mice by ELISA. In CARD11^R30W/+ mice less than 10 wk old, serum IgE was 2.0-fold higher than WT levels on average, but this 2.0-fold increase was primarily due to the presence of two mice of 16 assayed that showed 10- to 15-fold increases over WT IgE values (Fig. 4H). This prompted us to examine IgE levels in older mice as well. We found that among CARD11^R30W/+ mice between 10 and 20 wk of age, a larger percentage had elevated serum IgE, leading to a 2.2-fold average increase over WT values (Fig. 4I). Furthermore, we detected massively elevated serum IgE in 50% of CARD11^R30W/+ mice older than 20 wk of age, with the average IgE measurement ∼13-fold higher than in WT mice (Fig. 4J). The increases in serum IgE occurred in equal proportions in male and female mice, indicating a lack of sex-dependence. These results show that one copy of the CARD11 R30W mutation is sufficient to cause elevated serum IgE that becomes progressively more pronounced with age, but with only 50% penetrance. Interestingly, although 50% of CARD11^R30W/+ mice developed elevated serum IgE levels as they aged, none of these mice developed atopic dermatitis.

Previous studies have examined the effect of a homozygous hypomorphic CARD11 mutation, L298Q, in the unmodulated mouse strain (7, 49). Unmodulated mice also exhibit elevated serum IgE, but they develop spontaneous atopic dermatitis as they age. These phenotypes in unmodulated mice have been attributed to a substantially reduced Treg population that leads to the progressive expansion of Th2 cells, which secrete cytokines that promote B cell class switching and in turn lead to elevated IgE levels and atopic dermatitis (49). To assess whether there is a similar basis for the development of elevated IgE in CARD11^R30W/+ mice, we assessed the Treg population in the spleen and thymus of WT, CARD11^R30W/+, and CARD11^R30W/R30W mice. We found that in the spleen and the thymus, CARD11^R30W/+ mice had 4.4-fold and 3.9-fold fewer Tregs than WT mice, respectively (Fig. 5A–D). CARD11^R30W/R30W mice almost completely lacked Tregs, with a 63-fold reduction in Treg numbers in the spleen and a 96-fold reduction in the thymus relative to WT mice (Fig. 5A–D).

To determine whether the reduction in Treg numbers in CARD11^R30W/+ mice might lead to the expansion of Th2 cells, we isolated CD4⁺ T cells from WT and CARD11^R30W/+ mice younger than 10 wk old, stimulated them, and measured cytokine production by ELISA, using IFN-γ as a signature Th1 cytokine and IL-4 and IL-13 as signature Th2 cytokines. CARD11^R30W/+ T cells exhibited substantial defects in production of IFN-γ, IL-4, and IL-13 relative to WT T cells, indicating that the isolated splenic T cells had not obviously skewed to a Th2 state in CARD11^R30W/+ mice (Fig. 5E). Similarly, when we skewed naive CD4⁺ T cells to Th1 and Th2 subtypes in culture, we observed defective production of IFN-γ, IL-4, and IL-13 in CARD11^R30W/+ Th1 and Th2 cells compared with WT cells (Fig. 5F). Finally, we assessed whether Th2 cells were more prevalent in CARD11^R30W/+ mice that had already developed elevated IgE. We isolated CD4⁺ T cells from WT mice, CARD11^R30W/+ mice that had aged with normal IgE, and CARD11^R30W/+ mice that had aged with high IgE, stimulated them, and measured Th1 and Th2 cytokine production by ELISA. Once again, we found that IFN-γ, IL-4, and IL-13 production was suboptimal in isolated CD4⁺ T cells from CARD11^R30W/+ mice as compared with those from WT mice, regardless of whether the mouse had concomitant high serum IgE (Fig. 5G). Together, these results indicate that CARD11^R30W/+ mice exhibit deficits in both Th1 and Th2 cytokine production and do not show a propensity to skew toward a Th2 state such as what has been observed in the unmodulated mouse.

The lack of spontaneous atopic dermatitis in CARD11^R30W/+ mice prompted us to test whether these mice might exhibit an enhanced inflammatory response in a model of induced atopic dermatitis. If so, the behavior could reveal the possibility that a dominant negative CARD11 allele could predispose mice toward an atopic response in the skin. Therefore, we assayed the response of CARD11^R30W/+ mice to topical application of the vitamin D analogue MC903, which triggers the overexpression of the cytokine thymic stromal lymphopoietin (TSLP) in keratinocytes (42, 50). TSLP acts as an alarmin, initiating a Th2-dominant immune response that leads to the development of atopic dermatitis-like symptoms (51). We applied MC903 or 100% ethanol vehicle to the ears of WT and CARD11^R30W/+ mice over the course of 16 d, then harvested tissue from the treated ears to assess inflammation. Visible signs of inflammation such as redness, scaling, and increased vascularization began to appear in mouse ears treated with MC903 after ∼5–7 d of treatment and grew progressively worse as the treatment course continued, whereas mouse ears treated with ethanol alone appeared grossly normal throughout the treatment course. Tissue biopsies of ears treated with MC903 weighed substantially more than biopsies of ear tissue treated with ethanol, indicating swelling, but tissue from WT and CARD11^R30W/+ mice treated with MC903 had similar weights (Fig. 6A). H&E staining of the ear tissue showed that MC903 treatment caused massive infiltration of immune cells into the dermis and thickening of the epidermis, but epidermal thickening occurred to a similar degree in both WT and CARD11^R30W/+ mice (Fig. 6B, 6C). Likewise, ELISAs on ear tissue lysates for Th1 and Th2 cytokines showed that MC903 treatment caused a dramatic increase in both IFN-γ and IL-4 levels compared with ethanol treatment, and CARD11^R30W/+ cytokine levels were comparable to WT levels (Fig. 6D). We also harvested the draining lymph node from mice treated with MC903 and ethanol and found that lymph nodes from MC903-treated mice were also swollen, containing an average of 14-fold more total cells than lymph nodes from ethanol-treated mice. CARD11^R30W/+ lymph nodes treated with MC903 were slightly smaller than their WT counterparts, but this difference was not statistically significant (Fig. 6E). We used flow cytometry to measure the infiltration of CD4⁺ T cells, eosinophils, monocytes, and neutrophils into the draining lymph node and found that while MC903 treatment led to a large influx of immune cells into the lymph node, there were no significant differences in the numbers of any immune cell subset in the lymph nodes of treated WT or CARD11^R30W/+ mice (Fig. 6F). Taken together, these results indicate that the CARD11 R30W allele does not lead to an enhanced inflammatory response to TSLP release in the skin in the MC903 model of induced atopic dermatitis.

Given the stark signaling and functional deficits observed in CARD^R30W/+ B cells and T cells, we next sought to determine whether the CARD11 R30W mutation impacts NK cell function. CARD11 is expressed in NK cells and has been shown to coordinate NF-κB activation downstream of their activating receptors (52, 53), but the possibility that a dominant negative CARD11 allele could influence NK cell function has not been addressed. To assess the effect of the heterozygous CARD11 R30W mutation on inflammatory cytokine secretion, we isolated splenic NK cells from WT, CARD11^R30W/+, and CARD11^R30W/R30W mice, stimulated with anti-NKG2D and IL-2 for 24 h, and measured the production of IFN-γ by ELISA. We found that the amount of secreted IFN-γ decreased as the number of copies of the CARD11 R30W mutation increased. CARD11^R30W/+ NK cells exhibited a 1.7-fold reduction in IFN-γ production relative to WT NK cells, while CARD11^R30W/R30W NK cells exhibited a 5.0-fold reduction (Fig. 7A). We observed similar decreases in IFN-γ secretion by NK cells harboring the CARD11 R30W mutation when using NK cells that were expanded in culture in the presence of IL-15 for 6 d prior to stimulation (Fig. 7B). To assess how the CARD11 R30W mutation affects NK cell cytotoxicity, we employed an NK cell killing assay using the YAC-1 lymphoma cell line as target (43, 54). We cocultured CFSE-loaded YAC-1 cells with equal numbers of freshly isolated splenic NK cells (NK1.1⁺CD49b⁺) from WT, CARD11^R30W/+, or CARD11^R30W/R30W mice in a range of E:T ratios for 4 h, then used Annexin V to stain for dead cells and measured the number of CFSE⁺Annexin V⁺ cells via flow cytometry. NK cells from CARD11^R30W/+ and CARD11^R30W/R30W mice lysed the target YAC-1 cells just as efficiently as WT NK cells across the full spectrum of E:T ratios (Fig. 7C), indicating that the CARD11 R30W mutation specifically impairs cytokine secretion downstream of NKG2D without affecting cytotoxicity.

The defect we observed in cytokine production by CARD11^R30W/+ NK cells prompted us to assess the effect of the CARD11 R30W allele on signaling to NF-κB in these cells. We isolated splenic NK cells from WT and CARD11^R30W/+ mice, expanded them in culture in the presence of IL-15 for 6 d, stimulated them with PMA and ionomycin to mimic signaling downstream of NK receptors, including NKG2D, that couple to ITAM-containing NK adapters (52), and evaluated activation of the NF-κB pathway by Western blotting. Peak phosphorylation of IKKα/β was reduced by 11-fold in CARD11^R30W/+ NK cells, whereas peak IκBα phosphorylation was decreased by 14-fold and IκBα degradation was markedly impaired (Fig. 7D, 7E). Together, these results indicate that the CARD11 R30W allele severely impairs signaling to NF-κB in NK cells, leading to reduced inflammatory cytokine production without affecting cytotoxicity.

At least three forms of primary immunodeficiency are caused by germline mutations in the CARD11 gene including CARD11 deficiency, caused by homozygous LOF mutations, BENTA, caused by heterozygous GOF mutations, and CADINS, caused by heterozygous LOF dominant negative mutations. While all three syndromes involve an enhanced susceptibility to infection, CADINS uniquely involves a high incidence of atopy accompanied by a range of clinical presentations that can differ from patient to patient (37–40, 55). In an effort to study the impact of a dominant negative LOF CARD11 allele in the absence of the modifier genes, environmental exposures, and treatment regimens that could influence the presentation of CADINS patients, we generated and characterized mice heterozygous for the CARD11 R30W mutation identified in one of the initial CADINS families (38). Our results reveal that a single dominant negative CARD11 allele is sufficient to cause pathway-specific defects in the T, B, and NK cell lineages, leading to defective immune responses and an age-dependent dramatic increase in serum IgE without the spontaneous atopic dermatitis and Th2 expansion observed in other mice with diminished CARD11 function.

CARD11 signals as an oligomer of undetermined stoichiometry. Previous studies have characterized the effects of diminished CARD11 function in CARD11-deficient mice, which express no CARD11 oligomer at all (5, 6, 52, 53, 56), or in homozygous unmodulated mice, which express CARD11 oligomers composed solely of hypomorphic mutant subunits that retain some signaling potential (7, 49). Both CARD11-deficient and unmodulated mice have defects in Ag-induced T cell and B cell activation and in TCR and BCR signaling to NF-κB and JNK. Both strains also display reduced Treg numbers, but in contrast to CARD11-deficient mice, unmodulated mice retain enough CARD11 signaling in lymphocytes to allow a scenario in which reduced Treg function promotes a skewing of T cells to a Th2 phenotype that induces class switching in B cells, leading to hyper IgE production and atopic dermatitis (49). The CARD11^R30W/+ mice we have generated have provided the opportunity to examine the phenotypic consequences of mixed CARD11 oligomers composed of WT subunits and mutant subunits that are LOF and act in a dominant negative manner to reduce signaling from the WT subunits in the same oligomer.

Our data show that such mixed oligomers have different effects depending on the cellular context and the pathway downstream of CARD11, leading to an overall phenotype that appears different from CARD11-deficient and unmodulated mice and that expands our understanding of CARD11 signaling. We observed a greater apparent effect of mixed CARD11 oligomers on Ag receptor signaling to NF-κB in CD4⁺ T cells than in B cells in the first 30 min of signaling, despite the fact that in both cell types the coexpression of WT and mutant CARD11 is sufficient to prevent the inducible formation of the CBM complex as reflected in Bcl10 recruitment to CARD11. The data suggest that in this early phase of signaling, Ag-mediated activation of NF-κB is much less dependent on CBM complex formation in B cells than it is in CD4⁺ T cells, raising the likelihood that BCR signaling activates NF-κB through previously unrecognized mechanisms. We also observed greater effects of mixed WT:R30W CARD11 oligomers on signaling to JNK in T cells than in B cells, and mTOR signaling was only mildly reduced by the mixed oligomers in T cells and not reduced at all in B cells. The data highlight the need for further mechanistic study to fully understand the differences in Ag receptor signaling steps in T and B cells and uncover what aspects of the T and B cell environments make signaling pathways downstream of CARD11 more susceptible to mixed CARD11 oligomers in T cells than in B cells.

While early CARD11 signaling to NF-κB, JNK and mTOR is largely intact in CARD11^R30W/+ B cells, the mixed CARD11 oligomers clearly impact longer-term consequences of Ag receptor signaling in B cells. CARD11^R30W/+ mice exhibit a severely defective immune response to T-independent Ag and a delayed response to T-dependent Ag that generates lower amounts of Ag-specific serum IgG2a. The severely defective T-independent response likely reflects a more pronounced effect of mixed CARD11 oligomers in B1a and marginal zone B cells, which mediate this response (46, 57). Consistent with this, we observed significant selective reductions in these B cells populations in CARD11^R30W/+ mice.

Previous studies in CARD11-deficient mice have demonstrated a required role for CARD11 in NK cell cytokine secretion downstream of activating receptor engagement (52, 53). However, it was unclear whether a dominant negative CARD11 mutant assembled into mixed WT:mutant oligomers would impact NK cell function, especially because no studies of NK cells from CADINS patients have been reported. To our knowledge, our data reveal for the first time that mixed CARD11 oligomers disrupt NK cell function. CARD11^R30W/+ NK cells display a substantially reduced capacity to produce IFN-γ upon stimulation with anti-NKG2D but show no reduced capacity to lyse YAC-1 target cells in vitro, although YAC-1 killing by murine NK cells is partially dependent on NKG2D triggering (54). This specificity is consistent with previous studies in CARD11-deficient mice that demonstrated a specific requirement for CARD11 in NKG2D-mediated cytokine signaling, but not in cytotoxicity, attributable to a role for CARD11 downstream of the DAP12 ITAM-containing adapter (52). We also observed that mixed WT:R30W CARD11 oligomers severely perturb signaling to NF-κB induced by PMA/ionomycin treatment, which mimics PKC-dependent signaling downstream of ITAM-coupled receptors on NK cells. Because human NK cells are also activated by NKG2D triggering and by other receptors that signal through ITAM-containing adapters, the data suggest that NK cell defects could contribute to the high susceptibility of patients with dominant negative CARD11 alleles to viral infections and that NK function should be examined in such patients.

The development of atopy in patients with dominant negative CARD11 mutations is not well understood. While their diminished lymphocyte function is sufficient to cause immunodeficiency, the residual immune function present appears to be abnormally skewed, yielding atopic symptoms in nearly 90% of patients (39). It remains mysterious to what extent the atopy results from cell-intrinsic defects in lymphocyte development and behavior, whether the high IgE observed in patients is sufficient to induce atopy, whether modifier genes contribute, and whether environmental exposure to Ag or other factors play a role. Our data indicate that heterozygous presence of the CARD11 R30W mutation is sufficient to cause elevated serum IgE in an age-dependent manner with 50% penetrance. However, unlike in unmodulated mice (7, 49), the elevated IgE is not accompanied by atopic dermatitis. In addition, despite having reduced Treg numbers, CARD11^R30W/+ mice do not display an obvious skewing of CD4⁺ T cells to a Th2 state, either before or after the development of elevated IgE levels. Clearly high IgE is not sufficient to induce atopic symptoms. It is likely that exposure to Ags or pathogens in the environment are required for the induction of atopy in mice or patients heterozygous for a dominant negative CARD11 variant, and that a Th2 response triggered by factors other than reduced Treg activity plays a role. Consistent with this notion, most CADINS patients do not exhibit reductions in Treg numbers or suppressive capability (37, 39). The response of CARD11^R30W/+ mice to application of the vitamin D analogue MC903 suggests that an enhanced response to TSLP release in the skin is not likely to be the basis for the atopic dermatitis observed in CADINS patients. It remains unclear why the development of high serum IgE in CARD11^R30W/+ mice is age-dependent and 50% penetrant. We did not observe any correlations between the development of high serum IgE and sex or cage co-occupancy, suggesting that the microbiome may not contribute to high IgE induction. The age-dependence and incomplete penetrance that we observed in CARD11^R30W/+ mice emphasizes that even in a controlled genetic and environmental setting, there is individual-to-individual variability in how a dominant negative CARD11 allele can predispose to an abnormal immunological response.

An increasing number of germline mutations have been recognized as the basis for primary immunodeficiency syndromes that are associated with IgE-related allergic disease (58). The CARD11^R30W/+ mice will be a useful tool to unravel the complex mechanisms by which incomplete immune system dampening elicits these syndromes.

We thank Chip Hawkins for performing zygote injections, Nathan Archer and Mollie Meffert for advice, and Nicole M. Carter and Tyler A. Jones for critical reading of the manuscript.

The authors have no financial conflicts of interest.