All animals care, housing and experiments were performed in accordance with approved protocols of: (1) the ANU National University Animal Experimentation Ethics Committee, for mice on a C57BL/6 NCrl background; (2) the Garvan Institute of Medical Research/St Vincent’s Hospital Animal Ethics Committee, for mice on a C57BL/6 JAusb background. All experiments conformed to the current guidelines from the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. Within independent experiments, Card11 wild-type and mutant animals were sex- and age-matched.

Card11^M365K mice harbor a germline A to T nucleotide substitution at position 140,889,709 on chromosome 5, resulting in a methionine to lysine M365K substitution in the highly conserved region of the coiled-coil domain of CARD11. Card11^M365K mice were obtained by exome sequencing of first-generation offspring of C57BL/6 mice exposed to N-ethyl-N-nitrosourea (ENU; databases.apf.edu.au/mutations) and bred to homozygosity on a C57BL/6 Ncrl background. Card11^M365K mice were rederived onto a C57BL/6 JAusb background upon transfer from the Australian National University (ANU) Australian Phenomics Facility (APF) to Australian BioResources (ABR; MossVale, Australia).

Card11^loco mice harbor 3 distinct single-nucleotide variants in Card11 introns 2, 10 and 20 that cause a complete loss of CARD11 protein expression (56). Card11^loco mice were also identified by exome sequencing of first-generation offspring of C57BL/6 mice exposed to ENU (databases.apf.edu.au/mutations). The mice were bred to homozygosity and maintained on a C57BL/6 NCrl background.

C57BL/6 NCrl, C57BL/6 JAusb, B6.JSL-Ptprc^aPepc^b (CD45.1) and B6.129S7-Rag1^tm1Mom /J (Rag1^KO/KO ) mice were purchased from ABR.

Single-cell suspensions were prepared from mouse spleen, bone marrow, inguinal lymph nodes, peritoneal cavity and blood. 1-4 x 10⁶ cells in PBS 2% FCS were transferred into appropriate wells of a 96-well U bottom plate. To prevent non-specific antibody binding, cells were incubated with F_c blocking antibody for 20 min at 4°C in the dark. Cells were then incubated with antibodies for 30 min, on ice and in the dark. To fix cells, they were incubated in 10% formalin (Sigma-Aldrich) for 15 min at 4°C, and washed and resuspended in PBS 2% FCS. To stain for intracellular nuclear proteins, cells were fixed and permeabilised using the manufacturer’s instructions and the eBioscience Transcription Factor Staining kit. Stained single-cell suspensions were acquired on the BD LSRFortessa™.

Where appropriate, following extracellular antibody staining, immune populations were sorted by fluorescence-activated cell sorting (FACS) on a FACS Aria III (BD Biosciences).

Antibodies used for flow cytometric study of mouse organs are listed in Table 1 .

To evaluate the effect of the CARD11.M365K substitution, we used a retrovirus gene transfer and culture system to introduce into primary activated B cells the following: CARD11.M365K or as controls wild-type CARD11, BENTA-associated (27–29) CARD11.G123S, CARD11.E134G or empty vector expressing EGFP only.

Briefly, replication-defective retrovirus particles were produced by the Pheonix ecotropic helper-free retrovirus packaging cell line (ATCC; CRL-3214), and transduction efficiency measured by flow cytometric measurement of EGFP expression. C57BL/6 B cells were stimulated with 10µg/mL goat anti-mouse IgM (Jackson ImmunoResearch) and 10µg/mL anti-CD40 (FGK4.5; BioXCell) for 24 hours, followed by spin-infection with retrovirus supernatant containing DOTAP (Roche). The cells were then cultured in fresh RPMI 10µg/mL anti-CD40 for 36 hours, washed with RPMI and resuspended in cRPMI at a density of 10⁶ cells/mL.

The number of live EGFP⁺ cells was determined by hemocytometer counting of trypan blue–negative cells in each culture, and flow cytometric analysis of the same cells.

Approximately 20 x 10⁶ total splenocytes were incubated for 5 min at room temperature in 1 mL RPMI-1640 (Gibco) containing Cell Trace Violet (CTV; Invitrogen) at a final concentration of 20 μM, followed by three washes in complete RPMI (RPMI-1640 containing 10% heat-inactivated fetal calf serum (HI-FCS), 2% Penicillin-Streptomycin-Glutamine (Gibco), 0.1% 50 mM 2-Mercaptoethanol). CTV-labelled splenocytes were plated at a density of 1 x 10⁶ cells per mL and incubated for 3 to 5 days in complete RPMI alone or containing 10 μg/mL anti-CD3 and 10 μg/mL anti-CD28. Cell divisions were enumerated by flow cytometric measurements of the fluorescence intensity of CTV.

CARD11 mutations M365K and G123S were introduced into the corresponding mouse Card11 sequence using PCR-based site-directed mutagenesis. The coding sequences for Card11 and its variants were fused with the mutant ecDHFR sequence (kindly provided by Dr Wandless, Stanford university) in mammalian expression vector pcDNA3.1+ (66). HEK293 cells were transfected with expression vectors for ecDHFR-CARD11 mutants and reporter plasmids expressing firefly luciferase and Renilla luciferase under NF-kB and thymidine kinase promoters, respectively (pGL4.32 and pGL4.74 from Promega). The expression of CARD11 variants was induced by the addition of 10 mM trimethoprim (TMP). The transfected cells were lysed 5 hr after TMP addition, and luciferase activity was measured by Dual-Luciferase Reporter Assay (Promega).

Sorted Card11^M365K mutant or wild-type naïve CD4 T cells were sorted to high purity by FACS, and cultured in flat bottom 96-well plates coated with 4 μg/mL anti-CD3 (BioLegend), in RPMI1640 (Life technologies) supplemented with 10% heat inactivated FCS (Life technologies), 5×10^-5 M 2-ME, 0.1mM non-essential amino acids, 1mM sodium pyruvate, 10mM HEPES, 100u/mL penicillin, 100ug/mL Streptomycin, 100ug/mL Noromycin (all from Sigma) at a density of 0.5 ×10⁶ cells/mL.

The naïve CD4 T cells were cultured for 4 days in the following polarizing conditions: Th0 (1 μg/mL anti-CD28, 5 μg/mL anti-TGFβ, 5 μg/mL anti-IL-4, 5 μg/mL anti-IFNγ); Th1 (1 μg/mL anti-CD28, 5 μg/mL anti-TGFβ, 5 μg/mL anti-IL-4, 10ng/mL IL-12); Th2 (10ng/mL IL-4, 1 μg/mL anti-CD28, 5 μg/mL anti-TGFβ, 5 μg/mL anti-IFNγ); Th17 (20ng/mL IL-6, 1ng/mL human TGFβ, 5 μg/mL anti-IFNγ, 5 μg/mL anti-IL-4, 1 μg/mL anti-CD28).

After 4 days of culture, cells were stimulated with PMA (50ng/mL) and ionomycin (375ng/mL) for 6 hrs. Brefeldin A (10 μg/mL) was added to each well after 2 hours of stimulation. Cells were harvested, washed and stained with Zombie Aqua Viability dye (BioLegend), fixed with 2% formalin, permeabilized with saponin (0.1%), and stained intracellularly with mAbs directed against TNFα, IFNγ, IL17A, IL-2, IL5, IL-4.

Age- and sex-matched Card11^+/+ Rag1^KO/KO C57BL/6J recipient mice were irradiated with one dose of 425 Rad from an X-ray source (X-RAD 320 Biological Irradiator, PXI). Recipient mice were then intravenously injected with 4 x 10⁶ bone marrow cells consisting of a 1:1 mixture of Card11^+/+ C57BL6 CD45.1⁺ (Ptprc^a/a ) bone marrow cells and CD45.2⁺ (Ptprc^b/b ) bone marrow cells that were Card11^+/+ or Card1^M365K/M365K . 7 weeks were allowed for immune reconstitution before intravenous immunization of recipient mice with 2 x 10⁸ SRBCs. The immunized chimeric mice were sacrificed 7 days post-immunization.

8-12 weeks old Card11^M365K mice were sacrificed and single-cell suspensions prepared from their spleens. Splenic CD4 T cells were isolated by incubation with anti-CD4 biotin antibody and positive enrichment by manual magnetic-activated cell sorting (MACS) using LS columns (Miltenyi Biotec). 3-4 x 10⁶ Card11^+/+ or Card11^M365K/M365K CD4 T cells were intravenously transferred into each recipient mouse: either into C57BL6.CD45.1⁺ recipients where donor cells could be isolated based on CD45.1/2 expression, or in an independent experiment into Rag1^KO/KO mice that lack mature B and T cells (67). Recipient mice were treated with intraperitoneal (i.p.) injection of 200 μg anti-mouse PD-1 (clone RMP1-14; BioXCell) or rat IgG2a anti-trinitrophenol isotype control (clone 2A3; BioXCell) at days 0, 2 and 5 post-CD4 T cell transfer. Recipient mice were sacrificed 7 days post-injection, and blood and spleen harvested for analysis.

Statistical analysis of flow cytometric experiments was performed using the GraphPad Prism 6 software (GraphPad, San Diego, USA). A one-tailed unpaired Student’s t-test with Welch’s correction was used for comparisons between two normally distributed groups. An unpaired student’s t-test, corrected for multiple comparisons using the Holm-Sidak method, was used for comparisons of more than two groups. Differences between paired measurements were analyzed by paired t-test. In all graphs presented, the error bars represent the mean and standard deviation. * p < 0.05, ** p < 0.01, *** p < 0.001.

We identified the novel Card11^M365K mouse strain by exome sequencing of first-generation offspring of C57BL/6 mice exposed to the mutagen N-ethyl-N-Nitrosourea (ENU). Card11^M365K mutant mice carry an A to T nucleotide substitution at position 140,889,709 on Chromosome 5, resulting in a methionine to lysine change at amino acid 365 ( Figure 1A ).

To determine the effects in mouse B cells of Card11^M365K mutation relative to known GOF Card11 mutations, we used a retroviral gene transfer and culture system to transduce Card11^M365K into primary activated B cells ( Supplementary Figure 1 ). As controls, B cells were otherwise transduced with an empty vector expressing EGFP only, expressing wild-type Card11 or Card11^G123S , found in patients with BENTA disease, DLBCL and ATL (27–29, 31), or expressing Card11^E134G found in several patients with BENTA disease (27–29). Card11 ^M365K-transduced B cells expressed lower B220 and higher CD86 ( Supplementary Figure 1A ) cell-surface levels compared to control vector-transduced B cells, indicative of increased NF-κB activation in these cells. Card11^M365K -transduced B cells expressed CD86 and B220 at levels intermediate between Card11^E134G - and Card11^G123S -transduced cells, and over a period of four days in culture, Card11^M365K -transduced B cells accumulated in numbers intermediate between Card11^E134G - and Card11^G123S -transduced cells ( Supplementary Figure 1B ). CARD11^M365K was previously shown to enhance NF-κB activity in a luciferase assay (30). To validate this and directly measure the effects of M365K and G123S mutations on NF-κB signalling, we utilized our previously published (65) luciferase reporter method. Following trimethoprim-induced expression, M365K and G123S mutant CARD11 caused a mean 5-fold and 11-fold higher induction of the NF-κB luciferase reporter, respectively, relative to that induced by wild-type CARD11 ( Supplementary Figure 1C ). Collectively, these results indicate that M365K mutation causes CARD11 GOF intermediate between that caused by E134G or G123S.

To test the effects on B cells of Card11^M365K mutation within an otherwise normal gene, we measured survival and proliferation of splenic B cells from Card11^M365K/+ relative to Card11^+/+ mice. As an additional control, we included splenic B cells from homozygous Card11^loco/loco mice harboring 3 distinct single-nucleotide variants that cause a complete loss of CARD11 protein expression (68). Over a period of 5 days in the absence of stimulation, the percentage of live Card11^+/+ versus Card11^M365K/+ B-lymphocytes decreased at the same rate, whilst live Card11^loco/loco B cells decreased in frequency more rapidly ( Supplementary Figure 1D ). CARD11.M365K therefore does not enhance B cell survival in absence of stimulation. Similar results were obtained following stimulation with a 1 μg/mL sub-mitogenic dose of anti-IgM ( Supplementary Figure 1E ).

To measure proliferation following stimulation, we labelled splenic B cells with Cell Trace Violet (CTV). Relative to Card11^+/+ B cells, Card11^M365K/+ cells increased in size faster and Card11^loco/loco cells more slowly, following stimulation with different concentrations of anti-IgM ( Supplementary Figure 1E ). B cells stimulated with 10 μg/mL anti-IgM divided up to 5 times and a mildly increased percentage of Card11^M365K/M365K B cells divided 3 or more times relative to Card11^+/+ cells, whereas 80% of Card11^loco/loco B cells failed to divide at all ( Supplementary Figure 1F ). The mean percentage of divided cells was 70% for Card11^M365K/M365K, 62% for Card11^M365K/+ and 57% for Card11^+/+ B cells. By contrast, only 27% of Card11^loco/loco B cells had divided ( Supplementary Figure 1G ). Given the small number of WT CD4 T cells assessed, we were unable to conclude that these effects were statistically significant.

CARD11.M365K is thus a mild GOF protein that increases BCR-stimulated activation, survival and to a small extent proliferation.

To determine the effects of Card11^M365K mutation in vivo, we analyzed Card11^M365K mice on a C57BL/6 JAusb or C57BL/6 Ncrl background. All results presented herein were consistent between backgrounds and unless specified otherwise, all figures present data from C57BL/6 JAusb mice. Following inter-cross of heterozygous mutant mice, Card11^M365K/+ and Card11^M365K/M365K mice were detected at expected Mendelian frequencies at time of weaning and genotyping ( Figure 1B ). Heterozygous and homozygous mutant mice developed no obvious pathologies and had comparable weight and survival to wild-type mice over a period of up to 50 weeks ( Figure 1B ). Germline Card11^M365K mutation therefore appears insufficient to cause overt pathology in mice.

Given the recurrence of somatic CARD11 GOF mutations in B lymphomas (31), and the effects of germline CARD11 GOF mutations on B cells in mice and humans (22–29, 52–54), we assessed B cell populations in the bone marrow, spleen and lymph nodes of wild-type and Card11^M365K mutant mice. Card11^+/+ , Card11^M365K/+ and Card11^M365K/M365K mice had comparable percentages of Lin^neg Sca-1^pos c-Kit^pos (LSK) stem cells in the bone marrow ( Supplementary Figure 2A ). 5-7 week-old Card11^M365K mutant mice had a reduced percentage of CD5^pos CD11b^pos CD23^low CD43^high peritoneal cavity B1a cells. Interestingly, this difference waned with age ( Supplementary Figure 2B ). Card11^M365K mutant mice also had comparable percentages of bone marrow leukocytes of the B220^pos B-lineage, and within these of IgM^neg IgD^neg precursor, IgM^pos IgD^int immature or IgM^low IgD^pos mature recirculating B cells ( Supplementary Figure 2C ), and comparable percentages of precursor B cells with a CD43^high CD24^neg pre-pro-, CD43^int CD24^int pro- or CD43^low CD24^pos pre-B phenotype ( Supplementary Figure 2C ).

Notably, Card11^M365K mutant mice had increased cellularity and increased percentage of B leukocytes in the spleen and inguinal lymph nodes ( Figure 1C , Supplementary Figure 2D ). Card11^M365K mutant mice had normal numbers of CD93⁺ transitional and CD93^- mature B cell populations ( Supplementary Figure 2E ), but though unimmunized, had an increased percentage and number of germinal center (GC) B cells in both spleen and lymph nodes ( Figure 1D ). We therefore studied the effect of Card11^M365K mutation on T cell-dependent GC responses, by immunizing Card11^M365K mice with sheep red blood cells (SRBCs) and sacrificing them 5, 7, 12 or 15 days later. Relative to wild-type controls, Card11^M365K/M365K mice had increased numbers of B220^pos CD38^low CD95^pos GC B cells at days 7, 12 and 15 post-immunization ( Figure 1E ).

To test whether CARD11.M365K drives GC B cell accumulation cell-autonomously or rather secondary to dysregulation of T cells or other hematopoietic cells, we generated mixed chimeras wherein a fraction of all hematopoietic cells had mutant Card11^M365K/M365K and the remainder had wild-type Card11. Card11^+/+ Rag1^KO/KO mice were irradiated and transplanted with an equal mixture of Card11^M365K/M365K Ptprc^b/b and control Card11^+/+ Ptprc^a/a bone marrow. As an additional control, another set of mixed chimeras received an equal mixture of Card11^+/+ Ptprc^a/b and Card11^+/+ Ptprc^a/a bone marrow. All chimeras were immunized with sheep red blood cells (SRBCs) and sacrificed 7 days later. Flow cytometric analysis revealed no significant difference in frequency of B cells or of germinal center B cells of Card11^+/+ versus Card11^M365K/M365K donor origin ( Figure 1F ). Card11^M365K/M365K thus provides no striking cell-autonomous advantage to GC B cells, 7 days post-SRBC immunization in this model.

Based on the recurrence of somatic CARD11 GOF mutations in PTCL (38, 39, 43–46), we hypothesized that Card11^M365K mutation would dysregulate T cells. Following flow cytometric analysis of T cell populations, we observed a mutant allele gene dose-dependent increase in percentage (but not total number) of CD62L^- CD44⁺ effector memory CD8 and CD4 T cells in the spleen and lymph nodes of Card11^M365K mutant relative to wild-type mice ( Figure 2A ). The increased fraction of effector memory CD8 T cells was not accompanied by changes in their surface expression of CX3CR1, KLRG1, NKG2D or by changes in their granularity, whereas by contrast an increased fraction of Card11-mutant effector memory CD8 T cells expressed high levels of CD69 ( Supplementary Figure 3A ). We observed no change in fraction of CD8 T cells expressing the cytotoxic effector molecule granzyme B ( Supplementary Figure 3B ). Unimmunized Card11^M365K/+ and to a greater extent Card11^M365K/M365K mice had an accumulation of TCRβ⁺ CD3⁺ CD4⁺ CXCR5^high PD-1^high T follicular helper (T_FH)-like cells ( Figure 2B ). These accumulating Card11-mutant T_FH-like cells expressed homogenously higher cell-surface levels of ICOS and some but not all expressed higher cell-surface levels of PD-1, relative to wild-type cells ( Figure 2C ). To test whether germline Card11^M365K mutation also increases accumulation of T_FH cells during T cell-dependent responses, we analyzed the same mice described earlier at 5, 7, 12 and 15 days post-immunization with SRBCs. Card11^M365K/M365K mice had an increased frequency and total number of splenic T_FH cells at days 7, 12 and 15 ( Figure 2D ). Interestingly, they also had a significant accumulation of TCRβ⁺ CD3⁺ CD4⁺ CXCR5⁺ PD-1⁺ CD25⁺ FoxP3⁺ T follicular regulatory (T_FR) cells ( Figure 2E ). As expected, the T_FH cells were Bcl-6^high, ICOS^high and CD44⁺ and the T_FR cells were FoxP3⁺, Bcl-6⁺, ICOS^high, CD44⁺ and Blimp-1^high ( Figure 2F ). Similar to our observations in unimmunized mice, the accumulating Card11^M365K/M365K T_FH and T_FR cells expressed homogeneously higher levels of ICOS relative to their Card11^+/+ counterparts.

Unimmunized Card11^M365K/+ and to a greater extent Card11^M365K/M365K mice had a mutant allele dose-dependent increase accumulation of T_REG cells ( Figure 3A ), of phenotype TCRβ⁺ CD3⁺ CD4⁺ CD25⁺ FoxP3⁺ and having first excluded CXCR5^high PD-1^high T_FH-like or T_FR-like cells. The accumulating Card11-mutant T_REGS expressed homogeneously higher levels of ICOS but also of CTLA-4, and higher levels of CD69 and CD44 ( Figure 3B ), and the Card11-mutant mice had a significant accumulation of T_REG cells with a CD62L^- CD44⁺ effector memory-like phenotype ( Figure 3C ). Similarly, Card11^M365K mice on a C57BL/6 Ncrl background had a significant increase in percentage and total number per spleen of CD44^high and PD-1^high CD4 and CD8 T cells ( Supplementary Figures 3A-C ) and of T_FH-like and T_REG cells, which were by contrast significantly reduced in Card11^loco/loco mice ( Supplementary Figures 3D, E ).

Given the above findings, we tested whether germline Card11^M365K mutation alters early T cell development in the thymus. Thymus cellularity was similar in Card11^M365K/+ but mildly decreased in Card11^M365K/M365K relative to Card11^+/+ mice ( Figure 4A ). Card11-mutant mice had a significantly increased percentage of CD25⁺ FoxP3⁺ T_REGS among CD4 single-positive (SP) cells, but no change in total number of thymic T_REGS, relative to wild-type mice ( Figure 4B ). Cell-surface Neuropilin-1 (NRP1), CCR6 and CD24 were used to identify peripherally induced versus newly developed or recirculating thymus-derived T_REGS (68–70). We observed no change in percentage (or total number) of thymus-derived NRP1⁺, thymus-derived nascent CCR6^- CD24⁺ or recirculating CCR6⁺ CD24^- T_REGS in Card11-mutant relative to wildtype mice ( Figure 4C ). Card11 wild-type and mutant mice also had comparable frequencies and numbers of CD4^- CD8^- double-negative (DN), CD4⁺ CD8⁺ double-positive (DP), CD4⁺ single-positive (SP) and CD8⁺ SP thymocytes ( Figure 4D ), and of CD44⁺ CD25^- DN1, CD44⁺ CD25⁺ DN2, CD44^- CD25⁺ DN3 and CD44^- CD25^- DN4 early progenitors ( Figure 4E ). Thymic T cell development thus appears overtly normal in Card11^M365K mutant mice.

Collectively, the above results demonstrate that germline CARD11 gain-of-function causes over-accumulation in the periphery of activated CD8 and CD4 T cells, T_FH, T_FR and T_REG cells expressing increased levels of activating and inhibitory checkpoint molecules.

To test whether Card11^M365K acts cell-autonomously to dysregulate CD8 and CD4 T cells, we analyzed mixed chimeric mice containing Card11^+/+ CD45.1⁺ bone marrow-derived hematopoietic cells and CD45.2⁺ bone marrow-derived hematopoietic cells that were either Card11^+/+ or Card11^M365K/M365K . Within individual chimeric mice, there was a significant increase in frequency of Ptprc^b/b Card11^M365K/M365K relative to Ptprc^a/a Card11^+/+ CD4 effector memory (EM), T_FH, T_REG and CD8 EM cells – whereas no such difference was observed between Ptprc^b/b Card11^+/+ and Ptprc^a/a Card11^+/+ cells ( Figure 5A ). Card11^M365K/M365K thus provides a cell-intrinsic advantage to effector CD8 and CD4 T cells, T_FH and T_REG cells. The accumulating Card11^M365K/M365K T_REGS had a significant cell-intrinsic increase in levels of cell-surface ICOS and CD44 and of intracellular CTLA-4 ( Figures 5B, C ). Within T_REG cells, and reminiscent of observations in germline Card11-mutant mice, Card11^M365K/M365K mutation caused significant cell-autonomous accumulation of CD62L^- CD44⁺ T_REGS ( Figure 5D ).

Given the above findings, we tested whether Card11^M365K mutation increases T cell responses to TCR, CD28 or high-affinity IL-2 receptor stimulation, which engage pathways crucial to the differentiation, survival and proliferation of effector T cells and T_REGS. Following CTV labelling and 3 days of stimulation with anti-CD3 and anti-CD28 in vitro, the mean percentage of divided cells was 74% for Card11^M365K/M365K , 68% for Card11^M365K/+ and 58% for Card11^+/+ CD4 T cells, and 94%, 90% and 87% for CD8 T cells of the respective genotypes. By contrast, only 12% and 24% for Card11^loco/loco CD4 and CD8 T cells, respectively ( Supplementary Figure 4A ). Card11^M365K mutation also increased and Card11^loco/loco mutation decreased the size and cell-surface CD25 and PD-1 levels of stimulated CD4 and CD8 T cells ( Supplementary Figure 4B, C ). Given the small number of WT CD4 T cells assessed, we were unable to conclude that these effects were statistically significant. Our flow cytometric analysis of Ki67 expression revealed an increased fraction of Ki67+ T cells within the spleens of Card11^M365K/M365K mutant mice ( Supplementary Figure 4D ). Thus, GOF CARD11.M365K caused increased activation and a mild increase in proliferation of CD4 and CD8 T cells following TCR stimulation and CD28 co-stimulation. By contrast, Card11^M365K/M365K mutation had no effect on STAT5 phosphorylation following IL-2 stimulation ( Supplementary Figure 4E ). To determine the effects of CARD11 GOF on CD4 T cell differentiation in vitro, we purified naïve CD4 T cells from wild-type and mutant mice by fluorescence-activated cell sorting (FACS), and incubated them for 4 days in conditions that skew towards T helper 0 (T_H0), T_H1, T_H2 or T_H17 differentiation. At day 4, we observed increased frequencies of Card11^M365K/M365K relative to Card11^+/+ IL-4⁺ and IL-5⁺ T_H2-like cells ( Supplementary Figure 4F ). These results indicate that weak CARD11 GOF may skew naïve CD4 T cells towards T_H2 differentiation in response to TCR and cytokine stimulation.

Collectively, these results demonstrate that GOF CARD11 increases T cell activation and proliferation following TCR and CD28 stimulation and provides a cell-intrinsic advantage to activated CD8 and CD4 T cells, T_FH, T_REG and T_FR cells expressing increased levels of checkpoint molecules ICOS and PD-1.

Notably, PD-1 acts as a tumor suppressor in CD4 T cells (71), but PD-1 checkpoint therapy significantly worsens disease progression in some (72) but not all (73) individuals with ATL. ATL, which are thought to arise from effector and/or FoxP3⁺ CD4 T cells (74, 75), harbor recurrent somatic GOF CARD11 mutations (38, 39). To test whether PD-1 restrains the over-accumulation of Card11^M365K mutant CD4 T cells in vivo, we adapted a workflow used by Wartewig et al. to demonstrate that PD-1 inhibition synergises with ITK-SYK fusion to cause lethal CD4 T cell lymphoproliferation (71). We adoptively transferred 4 x 10⁶ Ptprc^b/b CD4 T cells that were either Card11^+/+ or Card11^M365K/M365K , into Ptprc^a/a Card11^+/+ C57BL/6 recipient mice. We injected the recipient mice with anti-PD-1 monoclonal antibody (mAb) or isotype control mAb at days 1, 3, 5 and 6, and sacrificed them at day 7 post-adoptive transfer ( Supplementary Figure 5A ). Consistent with our mixed chimera results, a higher fraction of Card11-mutant relative to wild-type CD4 T cells had an effector memory phenotype ( Supplementary Figure 5B ). Relative to control mAb-treated mice, anti-PD-1-treated mice had similar total numbers of cells per spleen ( Supplementary Figure 5C ). Amongst mice that received Card11^M365K/M365K CD4 T cells, anti-PD-1 treatment increased the total number of donor-derived CD4 T cells, and resulted in a trend towards increased number of donor-derived effector memory CD4 T cells ( Supplementary Figure 5D, E ). These results indicate that PD-1 inhibition is insufficient to cause lymphoma or lethal lymphoproliferation of Card11^M365K/M365K CD4 T cells, but that PD-1 may play a role in restraining the accumulation of Card11^M365K/M365K CD4 T cells in vivo.