CARD11 gain-of-function mutation drives cell-autonomous accumulation of PD-1+ ICOShigh activated T cells, T-follicular, T-regulatory and T-follicular regulatory cells

Introduction Germline CARD11 gain-of-function (GOF) mutations cause B cell Expansion with NF-κB and T cell Anergy (BENTA) disease, whilst somatic GOF CARD11 mutations recur in diffuse large B cell lymphoma (DLBCL) and in up to 30% of the peripheral T cell lymphomas (PTCL) adult T cell leukemia/lymphoma (ATL), cutaneous T cell lymphoma (CTCL) and Sezary Syndrome. Despite their frequent acquisition by PTCL, the T cell-intrinsic effects of CARD11 GOF mutations are poorly understood. Methods Here, we studied B and T lymphocytes in mice with a germline Nethyl-N-nitrosourea (ENU)-induced Card11M365K mutation identical to a mutation identified in DLBCL and modifying a conserved region of the CARD11 coiled-coil domain recurrently mutated in DLBCL and PTCL. Results and discussion Our results demonstrate that CARD11.M365K is a GOF protein that increases B and T lymphocyte activation and proliferation following antigen receptor stimulation. Germline Card11M365K mutation was insufficient alone to cause B or T-lymphoma, but increased accumulation of germinal center (GC) B cells in unimmunized and immunized mice. Card11M365K mutation caused cell-intrinsic over-accumulation of activated T cells, T regulatory (TREG), T follicular (TFH) and T follicular regulatory (TFR) cells expressing increased levels of ICOS, CTLA-4 and PD-1 checkpoint molecules. Our results reveal CARD11 as an important, cell-autonomous positive regulator of TFH, TREG and TFR cells. They highlight T cell-intrinsic effects of a GOF mutation in the CARD11 gene, which is recurrently mutated in T cell malignancies that are often aggressive and associated with variable clinical outcomes.

The study of CARD11 mutations in B-and T-lymphomas is complicated by the many genomic alterations acquired by these cancer cells (38,51). Previous studies have used mouse models to reveal B cell-intrinsic effects of GOF CARD11.L232LI (52), CARD11.L251P (53), CARD11.K215M or CARD11.E134G (54). These studies reported variable effects on B cells of different Card11 mutations, and the wide spectrum of GOF CARD11 mutations have diverse biochemical effects (55). The graded, variable effects of hypomorphic Card11 mutation within distinct T cell populations could not be predicted from knockout studies (16). The effects of hypermorphic CARD11 mutations are thus similarly hard to predict a priori. Collectively, the above observations highlight open questions regarding qualitative differences in NF-kB activation by CARD11 mutations, and possible discontinuity in the graded effects of CARD11 GOF mutations within different cell types. Crucially, to our knowledge, no studies have reported T cell-intrinsic effects of CARD11 GOF, despite the striking recurrence of somatic CARD11 GOF mutations in PTCL.
Here, we addressed these open questions by analyzing B and T lymphocytes in mice with a germline Card11 M365K mutation identical to CARD11 M365K previously identified in DLBCL (30)   and modifying a conserved region of the CC domain recurrently mutated in B-lymphomas (31) and T-lymphomas (38,39,(44)(45)(46) ( Figure 1A). CARD11.M365K increased activation and proliferation of B and T lymphocytes following antigen-receptor stimulation. Card11 M365K mice had increased numbers of GC B cells before and at multiple timepoints during a T cell-dependent response to immunization. Card11 M365K mutation was insufficient to cause lymphoma, or B cell lymphocytosis as observed in individuals with BENTA disease. However, Card11 M365K mutation caused mutant allele dose-dependent, cell-autonomous accumulation of T follicular (T FH ), T regulatory (T REG ) and T follicular regulatory (T FR ) cells over-expressing stimulatory and inhibitory checkpoint molecules. Our findings add to our understanding of CARD11 as a critical signaling protein in lymphocytes. They reveal T FH , T REG and T FR cells as T cell populations particularly sensitive to CARD11 signaling, and help to explain the recurrence of somatic GOF CARD11 mutations in aggressive human T-lymphomas arising from CD4, T REG and T FH cells.
2 Materials and methods

Mice
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 firstgeneration offspring of C57BL/6 mice exposed to N-ethyl-Nnitrosourea (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).

Flow cytometry
Single-cell suspensions were prepared from mouse spleen, bone marrow, inguinal lymph nodes, peritoneal cavity and blood. 1-4 x 10 6 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 cytometry
Antibodies used for flow cytometric study of mouse organs are listed in Table 1.

Retroviral gene transfer system
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 10 6 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.

T cell proliferation assays
Approximately 20 x 10 6 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 mM, 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 10 6 cells per mL and incubated for 3 to 5 days in complete RPMI alone or containing 10 mg/mL anti-CD3 and 10 mg/ mL anti-CD28. Cell divisions were enumerated by flow cytometric measurements of the fluorescence intensity of CTV.

2.6
In vitro analysis of CARD11 M365K and CARD11 G123S function CARD11 mutations M365K and G123S were introduced into the corresponding mouse Card11 sequence using PCR-based sitedirected 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).
2.9 CD4 T cell adoptive transfer and anti-PD-1 treatment 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 10 6 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 mg 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
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.

CARD11.M365K is a GOF protein that increases BCR-induced activation and proliferation in vitro
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)(28)(29)31), or expressing Card11 E134G found in several patients with BENTA disease (27)(28)(29). Card11 M365K -transduced B cells expressed lower B220 and higher CD86 (Supplementary Figure 1A Figure 1B). CARD11 M365K was previously shown to enhance NF-kB activity in a luciferase assay (30). To validate this and directly measure the effects of M365K and G123S mutations on NF-kB 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-kB 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 mg/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 mg/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 BC R-stimulated activation, survival and to a small extent proliferation.

Germline Card11 M365K mutation causes accumulation of germinal center B cells
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.
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 CD93mature 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 postimmunization ( 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 cellautonomous advantage to GC B cells, 7 days post-SRBC immunization in this model.

Germline Card11 M365K mutation causes accumulation of activated CD8 and CD4 T cells, T FH , T FR and T REG cells
Based on the recurrence of somatic CARD11 GOF mutations in PTCL (38,39,(43)(44)(45)(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 TCRb + 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 TCRb + 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 TCRb + 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 FHlike and T REG cells, which were by contrast significantly reduced in Card11 loco/loco mice (Supplementary Figures 3D, E).
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.

Card11 M365K mutation provides a cellautonomous advantage to activated CD8 and CD4 T cells, to T FH , T REG and T FR cells
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 cellswhereas 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 cellintrinsic 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 M 3 6 5 K / M 3 6 5 K 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 wildtype and mutant mice by fluorescence-activated cell sorting (FACS), and incubated them for 4 days in conditions that skew towards T helper 0 (T H 0), T H 1, T H 2 or T H 17 differentiation. At day 4, we observed increased frequencies of Card11 M365K/M365K relative to Card11 +/+ IL-4 + and IL-5 + T H 2-like cells (Supplementary Figure 4F). These results indicate that weak CARD11 GOF may skew naïve CD4 T cells towards T H 2 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 10 6 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

Discussion
The findings here reveal that gain-of-function mutation of a conserved CARD11 residue, located in the coiled-coil domain recurrently mutated in B-and T-lymphomas, caused cellautonomous accumulation of effector CD8 and CD4 T cells, and particularly of T FH , T REG and T FR cells that are critical to coordinating and regulating adaptive immune responses. Germline Card11 M365K GOF caused accumulation of spontaneous GC B cells and increased GC response to immune challenge but caused no dramatic B cell lymphocytosis as observed in individuals with BENTA disease. Our results demonstrate that CARD11 GOF perturbs T cells by increasing their activation and proliferation downstream of TCR or co-stimulatory receptor signaling. By revealing that GOF CARD11 drives aberrant expression of checkpoint molecules including PD-1 and ICOS, a known positive regulator of T FH , T REG and T FR cells, the findings here indicate that GOF CARD11 mutations perturb T lymphocytes by dysregulating not only TCR-NFkB signaling but also co-stimulatory signaling. These results highlight likely effects of acquired GOF CARD11 mutations that are strikingly recurrent in aggressive human PTCL derived from effector, follicular and regulatory CD4 T cells.
The mild increase in B cell numbers in Card11 M365K mutant mice contrasts with lethal B cell lymphoproliferation upon B cellconditional Card11 L232LI mutation (Card11 L225LI in the original publication (52);). This latter phenotype is also absent from mice with a Card11 E134G or Card11 K215M (54) or Card11 L251P mutation (53). Card11 K215M creates a cell-intrinsic advantage whereas Card11 E134G creates a cell-intrinsic disadvantage for GC B cells (54), and Card11 L251P acts primarily to alter GC kinetics (53). The variable effects on B cells of different CARD11 mutations may relate to qualitative differences in their effect on NF-kB signaling activity (31,55). In a luciferase reporter system in 293T cells, CARD11 p.M365K increased the transcription of a NF-kB target gene to levels above wild-type CARD11, but below CARD11 p.L251P (30). Our results, based on a NF-kB luciferase assay, B220 and CD86 expression and survival of B cells transduced with CARD11 WT , CARD11 G123S , CARD11 E134G and CARD11 M365K , indicate that CARD11 M365K leads to weak GOF intermediate between that caused by CARD11 E134G and CARD11 G123S .
The Card11 M365K mutant mouse strain was generated by ENU mutagenesis, which enabled us to study the effects of Card11 GOF mutation in an otherwise normal gene, as opposed to expression of mutant Card11 cDNA from a heterologous promoter and locus (52,53,76). In this context, CARD11 M365K was insufficient to cause B cell malignancy or the striking B cell lymphocytosis seen in individuals with BENTA disease. This contrast may relate to differing CARD11 mutations, as discussed above and given that CARD11 M365K has not been identified in the germline of children with BENTA disease. Alternatively, the contrast may relate to the specific-pathogen free environment of the Card11 M365K mice or to differences between human and mouse lymphocytes. Like germline GOF CARD11 mutations, BTK, NFKB1 or NFKB2 deficiency have different consequences in humans relative to mice. In humans, they cause profound loss of transitional and mature naïve B cells (77)(78)(79)(80) but in mice they cause a less drastic decrease (81)(82)(83)(84), indicating that human B cells may be more strongly dependent on BTK-NF-kB signaling. Several hypotheses may explain the profound increase in transitional and mature B cells caused by germline GOF CARD11 mutations in humans but not mice. (i) Card11 mRNA increases 10fold between pre-B cells and immature IgM + B cells and transitional B cells in mice (Immgen Database). It is possible that CARD11 mRNA and protein are more strongly expressed in human relative to mouse naïve B cells, beyond a threshold where GOF in the protein dysregulates proliferation and survival. Nevertheless, previous publications (52,54,76) and our in vitro data indicate that GOF CARD11 can provide a cell-intrinsic advantage to mouse B cells. (ii) CARD11 mRNA or protein may be down-regulated in mouse B cells as an adaptive response to GOF CARD11 signaling that does not function in human B cells. (iii) Human B cells may be less able to induce counter-regulatory processes acting downstream or upstream from CARD11 (i.e. induction of NFKBIA or TNAIP3). Future studies comparing CARD11, NFKBIA and TNFAIP3 protein levels in normal and CARD11-mutant human and mouse transitional and mature B cells may help to distinguish between these alternatives.
T cells from human BENTA patients carrying CARD11 GOF mutations typically proliferate less than healthy controls in response to anti-CD3/CD28 stimulation, a difference linked to a mildly anergic phenotype associated with poor IL-2 expression by CARD11-mutant human T cells (29). By contrast, Card11 M365K mutant mouse T cells had a mild proliferative advantage relative to Card11 wild-type mouse T cells. In addition to the considerations discussed above, it is possible that secondary effects that are visible in humans over time may not be visible in mice at 8-12 weeks of age. (i) These secondary effects may be pathological, as seen in CTLA4 deficiency in humans, which results in loss of B cells even though B cells mostly lack CTLA4 expression. Affected patients have relatively normal B cell numbers prior to developing pathology but start losing B cells when they develop the syndrome (85). (ii) These secondary effects may be compensatory, as seen in transgenic B cells expressing chimeric IgMG receptors containing the IgG tail segment. These cells adopt a gene expression profile of anergy, but this occurs secondary to their down-regulation of cell-surface receptor (86).
The skewing of towards Th2 differentiation of Card11 M365K/ M365K CD4 T cells is interesting, given that Th2 skewing occurs in humans with loss-of-function or dominant negative CARD11 m u t a t i o n s . W e c a n n o t e x c l u d e t h e p o s s i b i l i t y t h a t CARD11.M365K results in "blended" GOF and LOF effects, as previously observed in BENTA disease (26). Mice homozygous for the hypomorphic Card11 unmodulated mutation develop penetrant, spontaneous atopy and dermatitis with age (10), caused by partial reduction in effector T cell accumulation but also partial T REG deficiency leading to progressive, selective T H 2 accumulation and subsequent IgE production (16). In that context, hypomorphic Card11 mutation produces outcomes that could not be predicted from null alleles, through unequal titration of opposing effects within different T cell subsets (16). Similarly, unequal effects of hypermorphic mutations in different lymphocyte populations may contribute to the variable B and T cell pathologies in humans and mice with germline and somatic CARD11 GOF mutations.
In addition to cell-intrinsic effects, B cell homeostasis may be perturbed by CARD11 GOF within CD4 T cells. Previous publications studied Card11 L251P (53) and Card11 L232LI (52) expressed in B cells only, whilst T cell populations from germline Card11 E134G and Card11 K215M mutant mice were not reported (54). In Card11 M365K mutant mice, the accumulation of splenic T FH cells at days 7, 12 and 15, but not at day 5 post-immunization, correlated with accumulation of splenic GC B cells at days 7, 12 and 15, but not at day 5. Card11 M365K mutant T FH cells expressed homogeneously increased levels of cell-surface ICOS, and both T FH accumulation and increased ICOS expression (87) are known to drive GC B cell accumulation. Card11 M365K/M365K mutation caused cellautonomous accumulation of ICOS high T FH cells, but also of T FR cells that can act to suppress the GC response (88)(89)(90). The relative, and possibly graded, effects of Card11 mutation in T FH versus T FR cells, and in turn on B cell homeostasis, are difficult to distinguish without T FH or T FR -specific CARD11 GOF models. Nevertheless, our data raise the possibility that CARD11 GOF CD4 T cells may perturb B cells in individuals with germline or somatic CARD11 GOF mutations. Individuals with BENTA disease have normal numbers of circulating CD4 and CD8 T cells (25,29), but to our knowledge no detailed T cell immunophenotyping has been reported. Future studies should assess T cell populations in humans and mouse models with different germline or T cellrestricted CARD11 mutations.
With regards to T cell lymphoma, our results reveal likely cellintrinsic effects of the somatic GOF CARD11 mutations that recur in up to 30% of ATL (38,39), CTCL and Sezary Syndrome (43-46) and at lesser frequency in angioimmunoblastic T cell lymphoma (AITL) (91). CARD11 and PRKCB mutations are positively correlated in ATL (38), suggesting that NF-kB activating mutations may synergize in driving ATL. The striking recurrence of mutations modifying the TCR/NF-kB pathway highlights its importance in PTCLs including ATL (38,39,(56)(57)(58) and CTCL or Sezary syndrome (43-46, 59-61, 92, 93). One limitation of our study is that CARD11 M365K has not been identified in PTCL or CTCL. Nevertheless, CARD11 M365K modifies a conserved region of the CC domain recurrently mutated in PTCL (Figure 1). ATL, CTCL and AITL are thought to arise from activated, T FH -like and/ or T REG -like CD4 T cells (50, 94-96), and Card11 M365K mutation causes cell-autonomous accumulation of activated, T FH , T REG and T FR CD4 T cells. In addition, Card11 M365K/M365K caused overexpression of stimulatory and inhibitory receptors ICOS, CTLA-4 and PD-1, and increased activation, proliferation and PD-1 expression by mutant T cells following TCR and CD28 stimulation. Activating CD28 mutations recur in 10-11% of AITL (97,98), and in-frame fusions involving CD28, CTLA4 and ICOS recur in 7% of ATL along with CD28 focal gains and missense mutations, all of which result in continuous or prolonged costimulatory signaling (38). When expressed in mice on a Tet2 -/-background, the RHOA G17V mutation identified in 70% of AITL (99)(100)(101) results in T cell lymphomas that partially require ICOS and PI3K signaling for their proliferation and survival (102). The cell-intrinsic increase of ICOS and CTLA-4 expression on Card11 M365K mutant CD4 and T REG cells indicates that CARD11 GOF may contribute to CD4 T cell dysregulation not just via TCR-NFkB but also via PI3K signaling. ICOS expression increases the accumulation of T FH cells but also of T REG and T FR cells (103), such that ICOS over-expression on expanded CARD11 M365K mutant T FH , T REG and T FR cells may further their accumulation. By contrast, increased CTLA-4 on the surface of these Card11-mutant cells may limit their accumulation (104).
In addition to ICOS and CTLA-4, Card11 M365K mutation increased PD-1 expression by CD4 T cells, in vivo and following TCR/CD28 stimulation ex vivo. Parallel observations could be drawn by future studies testing the association of CARD11 mutations with increased PD-1 or ICOS expression on human Tlymphoma cells. PDCD1 (encoding PD-1) is increased in CD4 malignancies with gene signatures of dysregulated TCR signaling (71). PD-1 acts as a tumor suppressor in CD4 T cells and PDCD1 alterations, most commonly focal deletions, recur in 10-20% of CTCL, 36% of Sezary syndrome and 26% of ATL (71). Consistent with the effects of PD-1 in inhibiting TCR signaling and also CD28 co-stimulation (105), PD-1 inhibition mildly increased Card11 M365K/M365K CD4 T cell accumulation in vivo, but was nevertheless insufficient to cause CD4 lymphoma or lymphoproliferation. This contrasts with the lethal lymphoproliferation of CD4 T cells expressing an ITK-SYK fusion upon their exposure to anti-PD-1 monoclonal antibody (71). This dichotomy may point to a threshold of CARD11 or NF-kB GOF required for synergy with PD-1 LOF to drive CD4 lymphoproliferation. The acquisition of different somatic driver gene mutations (e.g. ITK-SYK fusion versus intermediate CARD11 GOF mutation) may explain why PD-1 inhibition accelerates disease progression in some (72) but not all (73) cases of ATL.
Unlike other PTCL, ATL requires HTLV-1 infection (41,106,107). The variable, often long, latent phase between HTLV-1 infection and ATL diagnosis implicated additional environmental or genetic events in ATL pathogenesis, and led to the discovery of TCR/NF-kB pathway genes and CARD11 as recurrently mutated in ATL (38,39). Notably, HTLV-1 viral proteins TAX and HBZ increase NF-kB activation and survival (108,109), and HTLV-1 has tropism for FoxP3 + CD4 T cells (109)(110)(111). Given that Card11 M365K mutation increases NF-kB activation and creates a cell-intrinsic advantage for T REGS , the cell-intrinsic effects of CARD11 mutations and of HTLV-1 infection may cooperate in driving ATL pathogenesis.
There is a striking paucity of information on the cell-intrinsic effects of somatic CARD11 GOF mutations in PTCL, which are heterogeneous and often aggressive malignancies associated with poor clinical outcomes (112). The above findings reveal cellintrinsic effects of a CARD11 GOF protein within T cells. They highlight the need to study T cells in humans with germline CARD11 GOF mutations and BENTA disease, and in mouse models with PTCL hotspot CARD11 mutations. Our findings further highlight the crucial role played by CARD11 in lymphocytes and the possible therapeutic utility of developing small molecule inhibitors targeting CARD11.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

Ethics statement
The animal study was reviewed and approved by Garvan Institute of Medical Research/St Vincent's Hospital Animal Ethics Committee; ANU National University Animal Experimentation Ethics Committee.