Research ArticleIMMUNODEFICIENCIES

Pathogenic CARD11 mutations affect B cell development and differentiation through a noncanonical pathway

See allHide authors and affiliations

Science Immunology  29 Nov 2019:
Vol. 4, Issue 41, eaaw5618
DOI: 10.1126/sciimmunol.aaw5618

Calling CARD

CARD11 is a critical player in adaptive immunity through its role in intracellular NF-κB signaling upon antigen detection by T or B cells. CARD11 mutations are associated with different immunodeficiencies that are linked to abnormal NF-κB activation, and now Wei et al. have characterized other CARD11 mutations that affect the AKT-FOXO1 signal axis. Using three different mouse models that mimic different CARD11-associated pathologies, they showed that CARD11 can act as a negative regulator of the AKT-FOXO1 signaling axis in B cells that is independent of NF-κB. This study reveals a noncanonical role for CARD11 that influences B cell development and differentiation and provides insight into immunodeficiencies like BENTA disease.

Abstract

Pathogenic CARD11 mutations cause aberrant nuclear factor κB (NF-κB) activation, which is presumably responsible for multiple immunological disorders. However, whether there is an NF-κB–independent regulatory mechanism contributing to CARD11 mutations related to pathogenesis remains undefined. Using three distinct genetic mouse models, the Card11 knockout (KO) mouse model mimicking primary immunodeficiency, the CARD11 E134G point mutation mouse model representing BENTA (B cell expansion with NF-κB and T cell anergy) disease, and the mouse model bearing oncogenic K215M mutation, we show that CARD11 has a noncanonical function as a negative regulator of the AKT-FOXO1 signal axis, independent of NF-κB activation. Although BENTA disease–related E134G mutant elevates NF-κB activation, we find that E134G mutant mice phenotypically copy Card11 KO mice, in which NF-κB activation is disrupted. Mechanistically, the E134G mutant causes exacerbated AKT activation and reduced FOXO1 protein in B cells similar to that in Card11 KO cells. Moreover, the oncogenic CARD11 mutant K215M reinforces the importance of the noncanonical function of CARD11. In contrast to the E134G mutant, K215M shows a stronger inhibitory effect on AKT activation and more stabilized FOXO1. Likewise, E134G and K215M mutants have converse impacts on B cell development and differentiation. Our results demonstrate that, besides NF-κB, CARD11 also governs the AKT/FOXO1 signaling pathway in B cells. The critical role of CARD11 is further revealed by the effects of pathogenic CARD11 mutants on this noncanonical regulatory function on the AKT-FOXO1 signaling axis.

INTRODUCTION

CARD11 plays an essential role in adaptive immunity because it governs antigen receptor–induced nuclear factor κB (NF-κB) activation by forming a protein complex with B-cell lymphoma/leukemia 10 (BCL-10) and mucosa-associated lymphoid tissue lymphoma translocation protein 1 (MALT1) (13). It is essential for lymphocyte activation and proliferation, as well as specific types of B cell development (48). B cell development starts in the bone marrow (BM) with the generation of immature B cells. Immature B cells then enter the spleen and further differentiate into transitional B cells through the T1, T2, and T3 stages. Transitional B cells lastly mature into marginal zone (MZ) B cells and follicular (FO) B cells, respectively (9). After antigen challenge, activated mature B cells undergo positive selection in germinal centers (GCs), where FO helper T cells facilitate somatic hypermutation, antibody affinity maturation, and class switch recombination (10). Card11 deficiency in mouse models impedes MZ B cell development and compromises mature B cell activation and proliferation after antigen stimulation (47).

In humans, pathogenic mutations in CARD11 cause various immunological diseases. Loss-of-function (LOF) CARD11 mutations or deletions fail to activate the antigen receptor–induced NF-κB signaling pathway in lymphoid cells and thereby compromises adaptive immunity, which leads to severe combined primary immunodeficiency (PID) in patients (1114). In contrast, gain-of-function (GOF) CARD11 mutations with overactivated NF-κB can cause malignant B cell proliferation, such as diffuse large B cell lymphoma (DLBCL). DLBCL is an aggressive non-Hodgkin’s lymphoma, and CARD11 mutations have been found in 10% of the DLBCL cases (15, 16). Germline GOF CARD11 mutations are also reported in BENTA (B cell expansion with NF-κB and T cell anergy) disease, which has elevated NF-κB activity and hyperproliferative B cells (17). It is believed that aberrant NF-κB signaling plays a dominant role in the pathogenesis of these immunological disorders. BENTA disease–associated CARD11 GOF mutants is also associated with immunodeficiency in patients, although the symptoms are less severe compared with PID caused by LOF CARD11 mutants (17, 18). The detailed mechanism remains elusive.

We have generated a point mutation mouse model bearing a GOF BENTA disease–related CARD11 mutant, in which a glycine is substituted for a conserved glutamine residue E134G. Because of the alternative start codon, the nomenclature for the CARD11 E134G mutant corresponds to E127G in a previous report (17). We use this mouse model to better understand the uncharacterized features of CARD11 mutants in immunodeficiency disease in vivo.

RESULTS

Card11 E134G mutant impairs the development of T cell–independent humoral immunity as Card11 KO mice

The Card11 mRNA level in B cells from E134G mutant mice (fig. S1A) is equivalent to that in wild-type (WT) controls. However, because of NF-κB activation-induced CARD11 protein degradation as reported before (19), the E134G mutant has reduced levels of CARD11 (fig. S1B). This phenotype is also consistent with the previous report that CARD11 E134G mutant protein expression level was reduced in transfected human lymphocytes (17). Humoral immunity is mainly mediated by B cells, which originate from fetal liver or adult BM (20). Initially, we evaluated the effect of the CARD11 E134G mutant on the early development of B cells in BM. Compared with littermate WT controls, the precursor B cell (pre-B cell) population in E134G mutant mice is slightly decreased, whereas the immature and transitional B cells are not changed (fig. S1C). Spleens from E134G mutant mice were not enlarged as reported in patients with BENTA disease (fig. S1D). Meanwhile, in the peripheral blood of E134G mutant mice, the transitional B cell population was unchanged (fig. S1E), which does not represent the phenotype found in patients either (17). Those could be due to differences between humans and mouse models or the immunological features in patients, which are absent from our experimental mouse model.

NF-κB activity is essential for both MZ and FO B cell development (21). Because the CARD11 E134G mutant is a GOF mutant with simultaneously up-regulated NF-κB activation (17), we expect increased MZ B cell maturation, as shown previously in the NF-κB constitutively activated mouse model (22). However, in E134G mutant mice, the MZ B cell numbers were markedly decreased and accompanied by slightly more FO B cells (Fig. 1A) when compared with the WT control. Detailed analysis of transitional B cell development demonstrated that E134G mutant mice have mildly decreased cell populations at the T1 and T3 stages but increased T2 stage cells (Fig. 1B). This suggests that E134G mutants may have a development defect at the transitional B cell stage.

Fig. 1 Card11 E134G mutant impairs the development of T cell–independent humoral immunity as Card11 KO mice.

(A) Top: Flow cytometry analysis of MZ and FO B cell proportion in B220+ CD93 mature B cells from WT and E134G mutant mice spleen. Bottom: MZ and FO B cell absolute numbers in the spleen were counted and are shown below (n = 5). (B) Flow cytometry of WT and E134G mutant B220+ CD93+ transitional B cells at the T1 (IgM+ CD23), T2 (IgM+ CD23), and T3 (IgM CD23) stage. The percentages of transitional T1, T2, and T3 cells are shown below (n = 5). (C) Flow cytometric staining of CD19+ B220mild B1 B cells and CD19+ B220high B2-B cells in WT and E134G mutant mice PC. The absolute cell numbers of B1 and B2 B cells in the PC were counted and are shown below (n = 4). (D) Top: The CD5+ CD43+ B1a and CD5 CD43+ B1b subtypes of B1 B cells were measured by flow cytometry. Bottom: The total cell numbers of B1a and B1b cells were also counted (n = 5). (E) Flow cytometry analysis of CD5+ CD43+ B1a and CD5 CD43+ B1b cells in the 7-day neonatal spleens of WT and E134G mutant mice. The total cell numbers of B1a and B1b cells were also measured and are shown below (n = 4). (F and G) NP-specific antibodies in WT and E134G mutant mice were evaluated by enzyme-linked immunosorbent assay, 12 days after immunization with 50 μg of NP-LPS or NP-Ficoll (n = 4). OD450, optical density at 450 nm. (H) Top: MZ and FO B cells in the spleen of WT and Card11 KO mice were analyzed by flow cytometry. Bottom: MZ and FO B cell percentages in WT and Card11 KO mice spleens were measured (n = 5). (I) Flow cytometry analysis of B1 and B2 B cells in the PC of WT and Card11 KO mice. The percentages of B1 and B2 B cells are shown below (n = 4). (J) B1a and B1b cells in WT and Card11 KO mice PC were measured by flow cytometry. The percentages of B1a and B1b cells are shown below (n = 6). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, paired t test. N.S., not significant. All data are means ± SEM.

Next, we analyzed another type of peripheral B cell, B1 B cells, which are primarily located in the peritoneal cavity (PC). B1 B cells can be further classified into B1a and B1b subtypes based on CD5 expression. Like many other B cell types, B1 B cell development requires NF-κB activity (23). Unexpectedly, B1 B cells, especially B1a cells, were ablated in the PC of E134G mutant mice (Fig. 1, C and D). The early development process of B1 B cells occurs primarily at the fetal and neonatal stages and then decreases in adult mice (24). Therefore, to analyze the early development of B1 B cells, we collected splenocytes from 7-day-old neonatal mice, because most spleen transitional B cells have the potential to develop into B1 B cells. The E134G mutant impeded neonatal transitional B cell development into neonatal CD5+ B1a cells (Fig 1E), accompanied by a substantial decrease at the T1 stage and increase at the T2 stage (fig. S1F).

B cells produce antibodies against pathogen infection in either a T cell–independent (TI) or a T cell–dependent manner. In the absence of help from antigen-specific T cells, B1 B cells and MZ B cells accumulate as a first line of defense against infection, which provides a quick but less specific antibody response (25). Thereby, B1 and MZ B cells are also considered as innate-like B cells (26). We immunized E134G mutant mice with the non–protein-based immunogens NP-LPS (lipopolysaccharide) and NP-Ficoll, which can elicit MZ and B1 cell–mediated TI antibody responses (27). Compared with the WT control mice, E134G mutant mice did not generate immunoglobulin M (IgM) or IgG antibody responses against NP-LPS and NP-Ficoll (Fig. 1, F and G). These results recapitulate the clinical data from patients with BENTA disease, whose vaccination response to TI antigen, such as polysaccharide-based vaccines, was markedly compromised (17). We also analyzed the phenotype in heterozygous mice with only one E134G mutant allele, and we found intermediate defects in the MZ (fig. S1G) and B1 B cell development (fig. S1H), as well as the transitional B cell populations from T1 to T3, compared with the E134G homozygous mutant mice (fig. S1I). We believe that this defect of TI humoral immunity resulting from the E134G mutant contributes to the immunodeficiency symptoms in patients with BENTA disease, and the heterozygous mutation is sufficient to cause a disease phenotype.

Compared with WT B cells, E134G mutant B cells show autonomously activated NF-κB reflected by enhanced inhibitor of nuclear factor κBα (IκBα) phosphorylation even before B cell receptor (BCR) stimulation (fig. S2A). We generated Card11 knockout (KO) mice and compared their B cell development phenotype with E134G mutant mice. As described before (47, 28), in the absence of CARD11, antigen-stimulated NF-κB activation is disrupted, reflected by IκBα phosphorylation as well as compromised IκBα degradation in purified B cells (fig. S2B). Total IκBα is also decreased in E134G mutant MZ and B1 B cells but increased in the Card11-deficient MZ and B1 B cells (fig. S2, C and D). These data confirm the previous findings that E134G is a GOF mutant for the antigen-induced NF-κB pathway (17). However, the MZ and B1 B cell development phenotypes in E134G mutant mice are reminiscent of the similar phenotype reported in Card11-deficient mice (47). Card11-deficient mice have diminished MZ B cells in the spleen (Fig. 1H) and B1 B cells in the PC (Fig. 1, I and J). It has been proposed before that CARD11 regulates the MZ and B1 B cell development in an NF-κB–dependent manner, because mice with constitutively activated NF-κB activity have enhanced MZ and B1 B cell development (22). However, mice carrying the E134G mutation, which causes elevated NF-κB activity, have the same developmental defects as seen in Card11-deficient mice. In conclusion, our data challenge the current model of CARD11 and indicate that the E134G mutant also has an unknown function in CARD11-regulated MZ and B1 B cell development that is independent of the NF-κB pathway.

Card11 E134G mutant exacerbates AKT activation and compromises B cell development through the FoxO1 pathway

To clarify whether the E134G mutant causes a cell-intrinsic effect on MZ and B1 B cell development, we generated BM chimeric mice by mixing BM cells from E134G mutant mice with BM cells from WT control mice that transgenically expressed green fluorescent protein (GFP) protein and reintroduced this BM mixture into sublethally irradiated recipient mice (29). GFP was used as a marker to distinguish WT cells from E134G mutant cells in chimeras, and GFP expression did not change B cell development potential in the spleen and PC (fig. S3, A to C). In the spleens of chimeric mice, MZ B cells with the E134G mutation were markedly decreased relative to WT (GFP+) MZ B cells (Fig. 2A, left). WT and E134G mutant FO B cells were present at similar frequencies in recipient mice (Fig. 2A, right). At the transitional B cell stage in chimeras, the E134G mutant disrupted transitional B cell development as shown by decreased T1 stage cells and increased T2 stage population (fig. S3D). WT B1 B cells are a significant fraction of cells in the PC of chimeric mice, whereas B1 cells from E134G mutant mice are negligible. However, the B2 cells in the PC are comparable between WT and E134G mutant cells (Fig. 2B). These results suggest that the E134G mutation exerts its effects on MZ and B1 B cell development in a cell-intrinsic manner.

Fig. 2 Card11 E134G mutant exacerbates AKT activation and compromises B cell development through the FoxO1 pathway.

(A) In WT (GFP+) and E134G (GFP) mutant BM mixed chimeras, flow cytometry analysis of WT (GFP+) MZ and FO B cells, as well as E134G (GFP) mutant MZ and FO B cells. The percentages of each group were measured and are shown below (n = 5). (B) B1 and B2 B cells in the PC of chimeras were analyzed by flow cytometry, and the percentages of B1 and B2 B cells are shown below (n = 4). (C) GSEA analysis of the FoxO pathway genes over RNA-seq data obtained from WT and E134G mutant B cells. FDR, false discovery rate. (D) Resting B cells from WT and E134G mutant mice were stimulated with anti-IgM (20 μg/ml) at different time points, and the phosphorylation of AKT308, AKT473, FOXO1, and S6 was measured by Western blot. (E) Phosphorylation of AKT at T308 residue in the MZ and FO B cells of WT and E134G mutant mice was measured by flow cytometry (n = 4). (F) Total FOXO1 protein level in MZ and FO B cells from WT and E134G mutant mice was measured by flow cytometry (n = 4). (G) qRT-PCR analysis of FOXO1 target genes expression in splenic B cells from WT and E134G mutant mice. (H) WT and EG BM hematopoietic stem cells were transduced with empty retrovirus or FOXO1 retrovirus, respectively, and then applied to BM chimeras. Six weeks later, GFP-positive MZ B cells in the spleens of recipient mice were analyzed by flow cytometry. MSCV, murine stem cell virus. (I) The percentage of MZ B cells among GFP-positive splenic B cells in retroviral transduced BM chimeras is shown on the right. (J) Resting B cells from WT and Card11 KO mice were stimulated with anti-IgM (20 μg/ml) at different time points, and the phosphorylation of AKT308, AKT473, FOXO1, and total AKT and FOXO1 was measured by immunoblot. Immunoblot quantification was normalized to actin and is shown below. (K) AKT phosphorylation in MZ and FO B cells from WT and Card11 KO mice were measured by flow cytometry (n = 4). (L) Total FOXO1 protein level in MZ and FO B cells from WT and Card11 KO mice was measured by flow cytometry (n = 5). (M) qRT-PCR analysis of FOXO1 target gene expression in WT and Card11 KO splenic B cells. *P < 0.05, **P < 0.01, and ***P < 0.001, paired t test. All data are means ± SEM.

To illuminate the underlying mechanism by which CARD11 regulates MZ and B1 B cell development, we used transcriptome analysis of E134G mutant B cells by RNA sequencing (RNA-seq). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis revealed that FoxO signaling was the most markedly altered pathway in E134G mutant B cells (fig S4A). Gene set enrichment analysis (GSEA) also confirmed that the FoxO gene signature is down-regulated in E134G mutant transitional B cells (Fig. 2C). In B cells, phosphatidylinositol 3-kinase (PI3K) and mammalian target of rapamycin complex 2 (mTORC2) can activate AKT by phosphorylating AKT at T308 and S473 residues, respectively, and activated AKT phosphorylates FOXO1 and excludes FOXO1 from nuclear (30). FOXO1 is a critical transcription factor regulating B cell development (31). We purified resting B cells from WT and E134G mutant spleens and stimulated them with goat anti-mouse IgM to activate BCR downstream signaling. Compared with WT, phospho-AKT308, phospho-AKT473, and FOXO1 phosphorylation in E134G mutant B cells all increased after BCR stimulation (Fig. 2D). However, the activation of a downstream target of mTORC1 S6 phosphorylation, as well as upstream regulators of AKT, such as SYK and PDK1, was not changed in E134G mutant B cells (fig. S4B). We also measured AKT phosphorylation and total FOXO1 protein levels by flow cytometry in different subtypes of B cells. Both remaining MZ and FO B cells from E134G mutants have greater AKT phosphorylation (Fig. 2E) and reduced total FOXO1 expression levels in MZ B cells (Fig. 2F). We evaluated FOXO1-regulated gene expression in both WT and E134G mutant resting B cells by reverse transcription polymerase chain reaction (RT-PCR). Similar to data from RNA-seq transcriptome analysis, most FOXO1-regulated genes, especially cell cycle–related gene expression, such as cyclin D2 (CCND2), are down-regulated in E134G mutant B cells. Regulatory cytokine Il-10 and terminal differentiation regulator gene Prdm1 are markedly reduced as well (Fig. 2G). Protein levels of IL-10 in E134G mutant B cells were decreased (fig. S4C), and the same reduction trend of cell cycle protein CCND2 was also observed (fig. S4D). The E134G mutation was not associated with a change in cell viability of MZ and FO B cells (fig. S4E), as well as of transitional B cells (fig. S4F), as measured by intracellular cleavage of caspase 3. However, DNA cell cycle analysis confirmed that both transitional and mature B cells from the E134G mutant result in fewer cells entering S phase when compared with WT (fig. S4, G and H). E134G heterozygous mutants also had elevated levels of AKT and FOXO1 phosphorylation but at an intermediate level between WT and homozygous mutant cells (fig. S5A). The same trend was also observed for FOXO1 target gene expression in mature B cells (fig. S5B). These data indicate that a single allele of E134G mutation is sufficient to influence the AKT/FOX1 signaling axis and contribute to BENTA disease pathogenesis.

Last, we introduced ectopic FOXO1 expression in WT and E134G mutant BM hematopoietic stem cells through retroviral transduction, and transduced cells were applied to BM chimeras. After 6 weeks, higher levels of FOXO1 not only enhanced WT MZ B cell development in chimeras but also rescued the development defect of E134G mutant MZ B cells (Fig. 2, H and I), which confirmed that the E134G mutant disrupted the MZ B cell development through the FOXO1 pathway.

Stimulation of Card11 KO B cells with anti-IgM was associated with a similar AKT signaling reinforcement as seen in E134G mutant B cells, and phosphorylation of AKT and FOXO1 phosphorylation both increased in Card11 KO B cells (Fig. 2J). Ex vivo flow cytometry analysis demonstrated that AKT phosphorylation at T308 was enhanced (Fig. 2K), and the total FOXO1 protein level was reduced in Card11 KO MZ and FO cells (Fig. 2L) compared with WT. Card11-deficient B cells also show down-regulation of FOXO1 target gene expression, like Ccnb1, Il-10, and Prdm1 (Fig. 2M), and similar to that observed in E134G mutant B cells.

These data indicate that CARD11 has a noncanonical role as a negative regulator for the AKT signal pathway activation, which is irrelevant to NF-κB activation upon antigen receptor stimulation. The absence of WT CARD11 exacerbates AKT activation, facilitates downstream FOXO1 phosphorylation and degradation, and attenuates FOXO1 target genes expression. Although the CARD11 E134G mutant has an increased potential to activate NF-κB, it loses its negative regulatory function, which leads to enhanced AKT activation as seen in Card11 KO cells.

Card11 K215M mutant accelerates B cell development by suppressing AKT activation

It has been reported that various regulatory proteins, including AKT, can bind CARD11 and contribute to CARD11-mediated NF-κB activation (32). CARD11 contains several distinctive protein domains, including CARD, coiled-coil (C-C), linker, PDZ, and MGUK domain. By applying different CARD11 fragments, we confirmed that the C-C domain of CARD11 is critical for AKT binding (fig. S6A). The C-C domain is also a hotspot for pathogenic CARD11 mutations (15, 33). Besides BENTA disease, most of the DLBCL-related CARD11 mutations are located on the C-C domain. DLBCL can be classified into activated B cell–like (ABC) type and germinal center B cell (GCB)–like type (34). Most oncogenic CARD11 mutations lead to elevated NF-κB activation and show a preference toward the ABC-type DLBCL (35, 36). There are a few CARD11 mutants being reported in GCB-type DLBCL, like K215M and L232LI (fig. S6B). We tested the individual AKT binding ability of several CARD11 oncogenic mutants, including G123S, K215M, and L232LI, along with WT CARD11 and the E134G mutant. Co-immunoprecipitation data demonstrated that CARD11 mutants have various binding abilities with AKT. The E134G mutant and G123S mutant have reduced binding activity with AKT compared with WT CARD11 (fig. S6C). Whereas the GCB-type CARD11 mutant K215M shows enhanced AKT binding compared with WT CARD11, the L232LI mutant has a similar binding affinity with WT CARD11 (fig. S6C). Next, we transduced WT CARD11 as well as various CARD11 mutants into two human B cell lymphoma cell lines, OCI-Ly10 (ABC-DLBCL type) and OCI-Ly19 (GCB-DLBCL type), and measured their impact on AKT phosphorylation. We found that E134G and G123S mutants can enhance AKT phosphorylation in the OCI-Ly10 cell line, whereas the K215M and L232LI mutants suppress AKT activation (fig. S4D). In OCI-Ly19 cells, E134G and G123S mutants also increase AKT activation, but K215M and L232LI mutants mildly suppress AKT phosphorylation (fig. S6E). The CC deletion form of CARD11 has little effect on AKT activation in either DLBCL cell line. It has been reported that AKT K63 ubiquitination facilitates AKT membrane recruitment and further activation (37, 38), and we investigated whether the CARD11 mutants affect AKT K63 ubiquitination in transfected 293T cells. In the presence of WT CARD11, AKT K63 ubiquitination was suppressed, indicating that WT CARD11 is a negative regulator of AKT activation, and explains the stronger AKT activation in Card11 KO B cells. In cells transfected with CARD11 mutants, E134G enhanced the AKT K63 ubiquitination, whereas K215M and L323LI mutants suppressed K63 ubiquitination when compared with the WT control (fig. S6F). The CC deletion form of CARD11, which is unable to bind with AKT, also enhanced AKT ubiquitination. The altered AKT K63 ubiquitination caused by various WT and CARD11 mutants corresponds to their different modulation effects on AKT activation in stable cell lines. Thereby, CARD11 and its mutants adjust AKT activity through regulating the AKT K63 ubiquitination pathway. Because L232LI mutant, which reduces AKT ubiquitination and activation compared with WT control, has the same binding affinity with AKT as WT CARD11, it also indicated that the interaction intensity between AKT and different CARD11 mutants may not be critical for the regulation of AKT activity.

Together with the suppressive effect of the K215M mutant on AKT activation in OCI-Ly19 cells and K63 ubiquitination in transfected 293T cells, the data imply that the K215M mutant has a converse effect on AKT activation compared with the E134G mutant. We generated a genetic mouse model carrying the K215M mutation to illuminate the in vivo function of the GCB-type CARD11 mutant (fig. S7A). K215M mutation does not influence Card11 gene expression at either the mRNA or protein level (fig. S7B). K215M mutant mice do show a converse phenotype to that observed in E134G mutant mice. Compared with WT control, K215M mutant mice have more pre-B cells but less mature B cells in the BM (fig. S7C) and markedly enlarged spleens (fig. S7D). Although the percentage of MZ B cells in the K215M mutant mice spleen is equivalent to WT control, the absolute number of MZ B cell numbers is twice as high as that of WT MZ B cells (Fig. 3A). In the PC of K215M mutant mice, B1 B cell numbers (Fig. 3B), both B1a and B1b cells (Fig. 3C), are markedly increased too. Transitional B cell analysis indicates a slightly decreased T2 stage percentage during development in K215M mutant mice (fig. S7E), which is also converse to the T2 stage phenotype in E134G mutant mice. We used the BM chimera model by mixing K215M mutant BM cells with GFP-expressing WT control BM cells and reintroduced them into sublethally irradiated recipient mice. In contrast to the phenotype observed in WT and E134G mutant mixed, K215M mutant cells showed a higher development privilege than WT cells. K215M mutant MZ and B1 B cells were the dominant population in the spleen (Fig. 3D) and the PC (Fig. 3E) of chimeras, respectively. These results confirmed that the K215M mutant also affects B cell development in a cell-intrinsic manner.

Fig. 3 Card11 K215M mutant accelerates B cell development by suppressing AKT activation.

(A) Top: Flow cytometry analysis of MZ and FO B cell proportion in B220+ CD93 mature B cells in WT and K215M mutant mice spleen. Bottom: MZ and FO B cell absolute cell numbers in mice spleen were counted and are shown below (n = 5). (B) Flow cytometric staining of CD19+ B220mild B1 B cells and CD19+ B220high B2 cells in the PC of WT and K215M mutant mice (top). The absolute cell numbers of B1 and B2 B cells in the PC were counted and are shown below (n = 5). (C) Top: The CD5+ CD43+ B1a and CD5 CD43+ B1b cells in the PC of WT and K215M mutant mice were also measured by flow cytometry. Bottom: The absolute cell numbers of B1a and B1b cells were counted (n = 5). (D) In WT (GFP+) and K215M (GFP) mutant BM chimeras, WT and K215M mutant MZ and FO B cells were analyzed by flow cytometry, and the percentages of MZ and FO B cells are shown below (n = 5). (E) B1 and B2 B cells in the PC of WT and K215M mutant BM mixture chimeras were analyzed by flow cytometry, and the percentages of B1 and B2 B cells are shown below (n = 5). (F) Splenic resting B cells from WT and K215M mutant mice were stimulated with anti-IgM (20 μg/ml) at different time points, and the phosphorylation of AKT308, AKT473, FOXO1, and total AKT and FOXO1 was measured by Western blot. Immunoblot quantification was normalized to actin and is shown below. (G) AKT phosphorylation in MZ and FO B cells in the spleen of WT and K215M mutant mice was analyzed by flow cytometry (n = 4). (H) Total FOXO1 protein level in MZ and FO B cells in the spleen of WT and K215M mutant mice was analyzed by flow cytometry (n = 5). (I) qRT-PCR analysis of FOXO1 target gene expression in the splenic B cells from WT and K215M mutant mice. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, paired t test. All data are means ± SEM.

Immunoblot results demonstrated that K215M mutant B cells show elevated NF-κB activation upon BCR stimulation (fig. S7F) and markedly reduced AKT phosphorylation at T308 and S473 sites (Fig. 3F). FOXO1 phosphorylation by AKT was also decreased (Fig. 3F). Ex vivo flow cytometry analysis of WT and K215M mutant mice demonstrated that AKT phosphorylation at T308 was mitigated in both K215M mutant MZ and FO B cells (Fig. 3G), which was accompanied by elevated FOXO1 protein levels (Fig. 3H). We purified K215M mutant B cells and applied transcriptome comparison analysis by RNA-seq in the same manner used for E134G mutant B cells. The K215M mutant B cells were also enriched for the FoxO pathway gene signature, but in contrast to the E134G mutant (fig. S4A), K215M mutant cells had more up-regulation of FoxO pathway gene expression (fig. S7G). Quantitative RT-PCR (qRT-PCR) results confirmed that the expression levels of FOXO1-regulated cell cycle–related genes Chek1, Ccnd1, and Ccnd2, as well as Il-10 and Prdm1, are all increased in K215M mutant B cells (Fig. 3I). The protein level of IL-10 (fig. S7H) and CCND2 (fig. S7I) was also elevated in K215M mutant B cells. The K215M mutant does not influence the cell viability (fig. S7J), but it changed the cell cycle by promoting S phase entry in B cells (fig. S7K). In heterozygous K215M mutant mice, MZ B and B1 B cell development was also intermediately enhanced compared with WT control and homozygous mutants (fig. S8, A and B). Heterozygous K215M mutant B cells show moderate changes of AKT and FOXO1 phosphorylation (fig. S8C) and FOXO1 target gene expression (fig. S8D). These data indicate that the K215M mutant executes a converse function compared with the E134G mutant and acquires enhanced suppression effect on AKT activation, which leads to more aggressive B cell development phenotype than WT control.

E134G and K215M mutants have reverse impacts on GC B cell response

GCs are the specific histological structures dedicated to the further differentiation of mature B cells. Malfunctions in the GC are associated with multiple B cell–related disorders, especially B cell malignancy (39). Recent studies have proposed that the PI3K/AKT/FOXO1 signaling axis plays a predominant role in the GC B cell responses (4042), but the role of this highly dynamic signaling network in GC B cells remains elusive. The noncanonical function of CARD11 in the AKT1/FOXO1 signal axis may be critical for GC B cell responses. In the absence of CARD11, B cells are unable to be fully activated, because CARD11-dependent NF-κB activation in antigen-stimulated B cells is disrupted (47). CARD11 E134G and K215M mutants provide us useful platforms to analyze the GC reaction regulated by the noncanonical function of CARD11.

After immunization with the model antigen sheep red blood cells (SRBCs), we found that the percentage and the absolute number of GC B cells in the spleen of E134G mice were significantly reduced (Fig. 4A). A similar reduction was observed in gut GC responses in the mesenteric lymph node (mLN) (Fig. 4B) and Peyer’s patch (PP) (Fig. 4C) of E134G mutant mice. In immunized K215M mutant mice, GC B cells increased slightly (fig. S9, A and C). These data indicate that the other lymphoid lineage cells might also influence the GC B cell response in K215M mutant mice. Mixed BM chimera experiments confirmed that such GC defects were B cell intrinsic, and we found that the GC compartment associated with E134G origin was severely shrank when compared with WT control, whereas the E134G FO B cell compartment remained normal (Fig. 4D). In contrast to E134G mice, K215M mutant GC B cells were overwhelming in the mixed BM chimeras when compared with WT cells (Fig. 4E). Consistent with an increased number of total B cells in K215M mice, we observed that the FO B cell compartment in K215M mixed BM chimeras was also slightly increased, although not as markedly as the GC B cell compartment. E134G mutant GC B cells retained a higher level of AKT phosphorylation than WT cells (Fig. 4F), and the total level of FOXO1 protein was lower (Fig. 4G). AKT phosphorylation was reduced in K215M mutant GC B cells (Fig. 4H), and total FOXO1 protein level was elevated (Fig. 4I) compared with WT control, and FOXO1 target gene expression was up-regulated in K215M mutant GC B cells, including Ccnd1 (Fig. 4J).

Fig. 4 E134G and K215M mutants have reverse impacts on GC B cell response.

(A to C) WT and E134G mice were immunized with SRBCs, and GC B cells in the spleen (A), mLN (B), and PP (C) were analyzed by flow cytometry 7 days later. Data were representative of three experiments. (D) Mixed BM chimeras generated with ~50% WT or E134G mutant cells plus ~50% CD45.1+ WT cells were immunized with SRBCs, and the percentages of CD45.2+ FO B cells and GC B cells were analyzed by flow cytometry. Data were pooled from two experiments. (E) Mixed BM chimeras generated with ~20% WT or K215M mutant cells plus ~80% CD45.1+ WT cells were immunized with SRBCs, and the percentages of CD45.2+ FO B cells and GC B were analyzed by flow cytometry. Data were pooled from two experiments. (F) Splenocytes from SRBC-immunized mice were fixed during isolation and intracellularly stained for p-AKT(308) in GC B cells. Data were representative of two experiments and shown as means ± SEM. (G) Flow cytometry analysis of intracellular FOXO1 in GC B cells from SRBC-immunized mice. Data were representative of two experiments and shown as means ± SEM. (H) Flow cytometry analysis of p-AKT(308) in mLN GC B cells from immunized K215M mutant mice. Data were representative of two experiments and shown as means ± SEM. (I) Flow cytometry analysis of intracellular FOXO1 in mLN GC B cells from K215M mutant mice. Data were representative of two experiments and shown as means ± SEM. (J) qRT-PCR analysis of mRNA expression of FOXO1 target genes with sorted WT and K215M GC B cells from immunized mice mLN. Data were representative of two experiments and shown as means ± SEM. (K and L) Plasma cell differentiation in the BM of WT and E134G mutant mice (K), as well as WT and K215M mutant mice (L), was analyzed by flow cytometry. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, paired t test. All data are means ± SEM.

GCs can be divided into two functionally distinct compartments: dark zone (DZ) and light zone (LZ). DZ B cells have elevated FOXO1 levels and are highly proliferative (41, 42). In line with down-regulated FOXO1 in E134G mutant B cells, we found that the DZ compartment was reduced in E134G mutant GCs accompanied by an increased portion of LZ B cells compared with WT controls (fig. S9D). Immunized mice were injected with DNA analog EdU (5-ethynyl-2′-deoxyuridine) to monitor the cell proliferation rate. We found that E134G mutant GC B cells have a reduced proliferation rate (fig. S9E), but K215M mutant gave rise to more EdU-positive GC B cells (fig. S9F), confirming accelerated GC B cell proliferation. Neither E134G nor K215M mutants showed changes in GC B cell viability based on intracellular cleaved caspase 3 (fig. S9, G and H). Cell cycle analysis demonstrated that E134G mutant GC B cells have fewer cells at S phase compared with WT control (fig. S9I), and conversely, the K215M mutation promotes GC B cell entry into S phase and proliferation (fig. S9J).

After GC reactions, B cells undergo terminal differentiation into antibody-secreting plasma cells. In accordance with down-regulated Prdm1 expression, SRBC-immunized E134G mutant mice showed markedly less plasma cell differentiation in the BM (Fig. 4K). However, the plasma cell differentiation in the K215M mutant mice was markedly increased (Fig. 4L). These data suggest that CARD11 E134G and K215M mutants have converse effects on GC B cell differentiation and plasma cell terminal differentiation.

DISCUSSION

Pathogenic mutations on CARD11 are frequently reported in multiple B cell–related immunological disorders (15, 17, 43). To illuminate the physiological function of distinctive pathogenic CARD11 mutants in vivo, we established three CARD11 mutation mouse models: the Card11 KO mouse model and CARD11 models with the E134G point mutation or K215M point mutation. Systemic comparisons demonstrated that CARD11 has a noncanonical function as a negative regulator of AKT activation, which is indispensable for B cell development and B cell differentiation. Although both E134G and K215M are GOF mutants for the NF-κB pathway, E134G and K215M mutants have a different regulatory effect on the AKT/FOXO1 signal axis. The BENTA disease–associated E134G mutation is associated with hyperactivation of the AKT pathway and defective B cell development and differentiation, whereas the oncogenic K215M mutation causes attenuated AKT activation and a converse B cell phenotype. Heterozygous mice with a mutation in one allele of E134G or K215M were prone to abnormal B cell development, and the homozygous mutations gave rise to more pronounced phenotypic changes.

Previous studies indicated that PI3K is indispensable for the MZ B cell development (44), and mice with constitutive activation of PI3K have elevated MZ B cell populations as seen in Foxo1-deficient mice (45). However, genetic deletions of Foxo1 at different B cell development stages caused distinctive MZ B cell phenotypes such that early deletion of Foxo1 by Cd19-Cre at the pro-B cell stage causes increased MZ B cell populations (46), but Cd21-Cre–mediated Foxo1 deficiency at the transitional B cell stage reduces both MZ and B1 B cells (31). In a BCR-deficient mouse model, constitutively activated PI3K (p110α*CA) at the transitional B cell stage can restore MZ B cell development, but deletion of Foxo1 at the transitional B cell stage failed to return MZ B cell populations back to a normal range (47). These results indicate that the transitional B cell stage involves a specific regulatory mechanism on PI3K-to-FOXO1 axis to sustain a normal MZ B cell development, which is downstream of PI3K but upstream of FOXO1. We measured Card11 expression level at different B cell development stages and found that Card11 expression is kept low at the pre-B cell stage in the BM but increased markedly at the transitional B cell stage (fig. S10). It indicates that CARD11 precisely controls AKT activation and FOXO1 activity starting from a specific time point during B cell development. Another critical downstream target of AKT is mTORC1, which is also crucial for B cell development and differentiation. However, in CARD11 mutant B cells, we did not observe changes in mTORC1 activity. Other regulatory pathways like extracellular signal–regulated kinase and AMP (adenosine 5′-monophosphate)–activated protein kinase or amino acid nutrients may control or contribute to mTORC1 activation.

Our data not only clarify a mechanism that contributes to BENTA disease but also explain the onset of B cell malignancy (39). In ABC-type DLBCL, malignant B cells originate from activated B cells, which are prohibited for the terminal differentiation into plasma cells (35). Loss of the tumor suppressor gene PRDM1 is found in most of the ABC-DLBCL cases, because it plays an indispensable role in plasma cell differentiation. B cells from patients with BENTA disease fail to differentiate into plasma cells (48) because of down-regulated PRDM1, which is also confirmed in our mouse model. Moreover, the CARD11 G123S mutation was found in both patients with BENTA disease and those with ABC-DLBCL cells. Therefore, CARD11 mutations correlated with ABC-type DLBCL and BENTA may share a similar molecular mechanism for disease pathogenesis. GCB-type DLBCL cells are mostly GCB-like and have a different molecular signature compared with ABC-DLBCL cells. Another well-known GCB-type CARD11 oncogenic mutation is L232LI, which induced the strongest level of NF-κB activation (15) and caused B cell malignancy and early postnatal lethality in a mouse model (49). The L232LI mutant also suppressed AKT activation in antigen receptor–stimulated B cells (49), which is similar to the K215M mutant. Patient samples with the K215M mutation have been accompanied by another oncogenic mutation, F130I (15). Although the K215M mutant mildly up-regulates NF-κB activity, the coincident F103I mutation alone or the F130I/K215M double mutation can induce NF-κB strongly. These data suggest that, together with NF-κB, down-regulated AKT activity and subsequent enhancement of FOXO1 pathway may be crucial for GCB-DLBCL pathogenesis. Recently, it has been reported that FOXO1 promotes B cell lymphomagenesis in GC (50), which is consistent with our findings that the GCB-DLBCL–related K215M mutant causes malignant proliferation and differentiation of GC B cells through elevated FOXO1 activity. In conclusion, the noncanonical regulatory function of CARD11 may determine the type of DLBCL to which malignant B cells transform. The dysregulated AKT-FOXO1 signal axis can also be used as a new standard to better classify pathogenic CARD11 mutants from multiple immunological disorders, which can help improve diagnostics and precision medicine development.

MATERIALS AND METHODS

Study design

The goal of this study was to investigate the role of CARD11 in B cells beyond NF-kB signaling. The study was performed by using three different CARD11 mouse models, including KO, E134G, and K215M mutants. B cell phenotyping and functional analyses were performed by flow cytometry and Western blot. RNA-seq analysis combined with in vitro protein analysis was used to determine the changes in gene expression and signal pathway enrichment. Sample sizes were based on previous experiments and the availability of genetically engineered donors. The number of replicates is given in each figure.

Generation of Card11 E134G and K215M KI mice

CRISPR-Cas9 technology was used to generate the Card11 E134G and K215M mice. Briefly, the Cas9 mRNA and guide RNA (gRNA), together with the synthesized oligo donor DNA that contains the mutation sites (401AG to GC in exon 5 for E134G or 644AA to TG in exon 5 for K215M) as the homologous recombination templates, were microinjected into the C57BL/6 fertilized eggs. The C57BL/6 background F0 generation was genotyped by PCR analysis of tail DNAs using allele-specific primers, and the mutant mice were confirmed by sequencing of PCR products. The gRNA sequences are CACACACTTCCTGATGAACG for E134G and GCTCAGTGAGGAGAAAAACA for K215M. Genotyping primers were as follows: forward, 5′-CCACAGTCAGGAAACCGAGAGG-3′; reverse, 5′-GCTGGGGAGGGCTGTGT-3′. The E134G and K215M gene site had been labeled as E127G and K208M in an earlier reference (17).

Generation of Card11 KO mice

The gRNA to mouse Card11 gene and Cas9 mRNA were coinjected into fertilized mouse eggs to generate targeted KO offspring. F0 founder animals were identified by PCR followed by sequence analysis, which was bred to WT mice to test germline transmission and F1 animal generation. Genotyping primers were as follows: forward, 5′-CCTGCATAGA AAGAGCATATAC-3′; reverse, 5′-CACCTAATAGGTGCTTTCCATTC-3′. KO allele size: 730 base pairs (bp); WT allele size: 7714 bp.

Flow cytometry analysis

Single-cell suspensions obtained from mouse BM, spleen, and PC were treated with RBC lysis buffer (eBioscience). For surface and intracellular staining, cells were blocked with Fc receptor antibody (catalog no. 101310, BioLegend) for 10 min at 4°C before staining with fluorochrome-conjugated surface antibodies: B220 (clone: RA3-6B2, BioLegend), CD93 (clone: AA4.1, eBioscience), CD23 (clone: B3B4, BioLegend), CD21 (clone: 7E9, BioLegend), CD43 (clone: S11, BioLegend), CD5 (clone: 53-7.3, BioLegend), CD24 (clone: M1/69, eBioscience), IgM (clone: RMM-1, BioLegend), IgD (clone: 11-26c.2a, BioLegend), CD19 (clone: 6D5, BioLegend), GL7 (clone: GL7, BioLegend), CD95 (clone: Jo2, BD Biosciences), CD138 (clone:281-2, BioLegend), CXCR4 (clone: L276F12, BioLegend), CD86 (clone: GL-1, BioLegend), CD45.1 (clone: A20, BioLegend), and CD45.2 (clone: 104, BioLegend). After surface staining, cells were fixed with fixation/permeabilization buffer (catalog no. 00-5123-43, eBioscience) and then permeabilized and stained in permeabilization buffer (catalog no. 00-8333-56, eBioscience). The following antibodies were used for intracellular or cell cycle staining: Ki-67(clone: 16A8, BioLegend), phospho-S6 (pS240; clone: N4-41, BD Biosciences), phospho-AKT (pS473; clone: M89-61, BD Biosciences), phospho-AKT (pT308; catalog no. 558275, BD Biosciences), phospho-AKT (pT308; catalog no. 13038S, Cell Signaling Technology), FOXO1 (catalog no. 2880S, Cell Signaling Technology), cleaved caspase 3 (catalog no. 8788S, Cell Signaling Technology), phycoerythrin goat anti-rabbit IgG (H + L) (catalog no. 111-116-144, Jackson ImmunoResearch), and 4′,6-diamidino-2-phenylindole (catalog no. D1306, Invitrogen). For EdU staining, mice were intravenously injected with 1 mg of EdU (Cell-Light EdU Apollo643 In Vitro Flow Cytometry kit, RiboBio) 30 min before they were euthanized. Flow cytometric analysis was performed on a Gallios flow cytometer (Beckman Coulter), and data were analyzed with FlowJo software (Tree Star).

Immunizations and measurements of antibody responses

For a TI immune response, both WT and mutant mice were intraperitoneally injected with 20 μg of NP-LPS (0111:B4) or with 50 μg of NP-Ficoll (Biosearch Technologies) in 100 μl of phosphate-buffered saline. Specific classes or isotypes were then detected using the series of an isotype-specific second antibody of the SBA Clonotyping System (Southern Biotech), as previously described (27). Data for antigen-specific antibodies are shown after subtraction of low absorbance (A) values from pre-immune controls analyzed together with the immune sera and were separately determined to match values yielded by titration. For induction of GCs, mice were intraperitoneally immunized with 2 × 108 SRBCs or 100 μg of NP-KLH (keyhole limpet hemocyanin) precipitated with aluminum hydroxide gel adjuvant (Accurate Chemical and Scientific).

Gene expression profiling

Single-cell suspensions from spleens of WT and CARD11 mutant mice were applied to total B cell isolation by using the B Cell Isolation Kit from STEMCELL Technologies. CD93+ B cells were then positively selected by using overwhelming CD93+ beads from Miltenyi Biotec. Then, total RNA was isolated from CD93+ and CD93 B cells, respectively. Three biological replicates were mixed and provided to the Huada for library construction and sequencing done by BGI Co. Ltd. (Beijing). Briefly, libraries were constructed from polyadenylated RNAs and sequenced with an Illumina HiSeq 2500 on an SR-50 run aiming for 30 million reads per sample. Reads were aligned to the mm10 mouse transcriptome using TopHat, and differential gene expression was determined using Cuffdiff. GSEA was performed using software available from the Broad Institute (www.broadinstitute.org/gsea), which was performed by using MSigDB. C2: curated gene sets and PID-FOXO pathway (51) was ranked no. 18 among all the 1329 signatures.

Mixed BM chimeras

BM cells were obtained from CARD11 mutant mice and C57BL/6-Tg [human ubiqutin C promoter (UBC)–GFP] (a gift from Y. Xiao, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences) or B6.SJL-Ptprca Pepcb/BoyJ (from the Jackson Laboratory) as the WT control. WT C57/BL6 recipient mice were irradiated using 10 gray, followed by the intravenous injection of 2 × 106 GFP+ BM cells or BoyJ CD45.1+ cells and 2 × 106 CARD11 mutant cells. Mice were euthanized 6 to 8 weeks after reconstitution. Spleen and PC lymphocytes were analyzed for the presence of MZ/FO, B1/B2 B cells, and GCBs, respectively.

B cell isolation, stimulation, and immunoblot analysis

Splenic B cells were purified by using the B Cell Isolation Kit from STEMCELL Technologies. Then, CD93+ transitional B cells were isolated by using CD93+ beads from Miltenyi Biotec. For measurements of the induction of BCR downstream signaling pathway, purified B cells were stimulated with F(ab)2 goat anti-IgM (20 μg/ml) for the indicated time point. Proteins in the whole B cell lysates were separated by SDS–polyacrylamide gel electrophoresis, transferred onto nitrocellulose membranes (Millipore), and then blotted with rabbit antibodies against phospho-AKT (T308), phospho-AKT (S473), phospho-FOXO1, phospho-P65, phospho-S6 (240), total AKT and S6, and mouse anti-actin antibodies followed by the appropriate horseradish peroxidase–conjugated, species-specific secondary anti-IgG antibodies (Cell Signaling Technology). Proteins were visualized and quantitated by the Bio-Rad ChemiDoc XRS System.

Retroviral transduced BM chimeras

Mouse FoxO1 complementary DNA was cloned into an MSCV-IRES2-EGFP retroviral vector. Retrovirus was made by transfection of 293T cells. BM cells were collected from the femurs and tibiae of 6- to 8-week-old donor mice. After red blood cell lysis, hematopoietic stem cells were enriched by using the Lineage Cell Depletion Kit (Miltenyi Biotec). Hematopoietic stem cells were cultured in chemically defined serum-free medium X-VIVO 10 with gentamicin (Lonza) supplemented with l-glutamine, β-mercaptoethanol (50 mM), mouse recombinant stem cell factor (50 ng/ml), IL-6 (20 ng/ml), IL-3 (10 ng/ml), FLT-3L (10 ng/ml), and IL-7 (10 ng/ml) (PeproTech) for 24 hours. Then, cocultured cells were transduced by spin infection with retroviral supernatant. Cells were incubated for another 24 hours before intravenous tail injection into WT recipient mice, which were irradiated at 9 Gy at least 12 hours before adoptive transfer. Six weeks after transfer, chimeras were analyzed by flow cytometry.

Statistics

Two-tailed t tests were used for two–data group comparison, and multiple t test statistics significance uses the Holm-Sidak method for multiple comparisons. Statistical tests were run using GraphPad software (v8.01). SEM was reported for all experiments, and P < 0.05 was considered statistically significant.

SUPPLEMENTARY MATERIALS

immunology.sciencemag.org/cgi/content/full/4/41/eaaw5618/DC1

Fig. S1. Evaluation of Card11 E134G mutant mice.

Fig. S2. NF-κB activation in E134G mutant and Card11 KO mice.

Fig. S3. BM chimeric mouse model by using GFP+ WT mice.

Fig. S4. BCR downstream signaling pathways, B cell apoptosis, and cell cycle analysis in Card11 E134G mice.

Fig. S5. AKT/FOXO1 signal axis in WT, heterozygous, and homozygous E134G mutant mice.

Fig. S6. CARD11 interacts with AKT and regulates AKT K63 ubiquitination.

Fig. S7. B cell development in Card11 K215M mutant mice.

Fig. S8. Phenotype analysis of WT, heterozygous, and homozygous K215M mutant mice.

Fig. S9. GCB response analysis in Card11 E134G and K215M mutant mice.

Fig. S10. Dynamic expression of Card11 during B cell development.

Table S1. Raw data in Excel spreadsheet.

REFERENCES AND NOTES

Acknowledgments: We thank M. Lenardo, L. Zheng, S. Muljo, and C. Kanellopoulou from LISB/NIAID and Z.-g. Liu from NCI for comments and suggestions. Funding: This study was supported by the National Key R&D Program of China (2018YFA0902703 and 2018YFA0800602), Research Funding via the National Natural Science Foundation of China (nos. 31771576, 81830078, and 31570881), and the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA12010311 to W.L.). Author contributions: Z.W., J.C., Y. Zhang, Y.H., P.J., Y.D., Q.Z., and Xuelei Wang performed experiments. Y. Zang and N.L. performed data and statistical analysis. J.Q., Xiaoming Wang, and W.L. designed the study. Xiaoming Wang and W.L. wrote the paper. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The RNA-seq data for this study have been deposited in the NCBI Gene Expression Omnibus database (GSE136588). Other materials are available through a simple academic material transfer agreement, where necessary for institutional transfer.
View Abstract

Navigate This Article