Egr2 and Egr3 in regulatory T cells cooperatively control systemic autoimmunity through Ltbp3-mediated TGF-β3 production

Transcription

1 Egr2 and Egr3 in regulatory T cells cooperatively control systemic autoimmunity through Ltbp3-mediated TGF-β3 production Kaoru Morita a, Tomohisa Okamura a,b,1, Mariko Inoue a, Toshihiko Komai a, Shuzo Teruya a, Yukiko Iwasaki a, Shuji Sumitomo a, Hirofumi Shoda a, Kazuhiko Yamamoto a,b, and Keishi Fujio a,1 PNAS PLUS a Department of Allergy and Rheumatology, Graduate School of Medicine, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo , Japan; and b Max Planck University of Tokyo Center for Integrative Inflammology, The University of Tokyo, Komaba, Meguro-ku, Tokyo Japan Edited by Shimon Sakaguchi, WPI Immunology Frontier Research Center, Osaka University, Osaka, Japan, and approved November 11, 2016 (received for review July 11, 2016) Systemic lupus erythematosus (SLE) is a prototypical autoimmune disease characterized by multiorgan inflammation induced by autoantibodies. Early growth response gene 2 (Egr2), a transcription factor essential for T-cell anergy induction, controls systemic autoimmunity in mice and humans. We have previously identified a subpopulation of CD4 + regulatory T cells, CD4 + CD25 LAG3 + cells, that characteristically express both Egr2 and LAG3 and control mice model of lupus via TGF-β3 production. However, due to the mild phenotype of lymphocyte-specific Egr2-deficient mice, the presence of an additional regulator has been speculated. Here, we show that Egr2 and Egr3 expressed in T cells cooperatively prevent humoral immune responses by supporting TGF-β3 secretion. T cell-specific Egr2/Egr3 double-deficient (Egr2/3DKO) mice spontaneously developed an early onset lupus-like disease that was more severe than in T cell-specific Egr2-deficient mice. In accordance with the observation that CD4 + CD25 LAG3 + cells from Egr2/3DKO mice completely lost the capacity to produce TGF-β3, the excessive germinal center reaction in Egr2/3DKO mice was suppressed by the adoptive transfer of WT CD4 + CD25 LAG3 + cells or treatment with a TGF-β3 expressing vector. Intriguingly, latent TGF-β binding protein (Ltbp)3 expression maintained by Egr2 and Egr3 was required for TGF-β3 production from CD4 + CD25 LAG3 + cells. Because Egr2 and Egr3 did not demonstrate cell intrinsic suppression of the development of follicular helper T cells, Egr2- and Egr3-dependent TGF-β3 production by CD4 + CD25 LAG3 + cells is critical for controlling excessive B-cell responses. The unique attributes of Egr2/Egr3 in T cells may provide an opportunity for developing novel therapeutics for autoantibodymediated diseases including SLE. Egr2 Egr3 TGF-β3 systemic lupus erythematosus regulatory T cell Antibodies play critical roles in protecting us from infectious threats. Effective humoral immune responses depend on germinal center (GC) reactions. Follicular B cells encounter antigen in the GC and receive T-cell help to differentiate into memory B cells and long-lived plasma cells that produce high-affinity antibodies (1). However, aberrant humoral immune responses against self-antigens lead to the development of autoimmune diseases. Indeed, suggestive clinical symptoms and the detection of autoantibodies in patient sera are essential diagnostic elements. Systemic lupus erythematosus (SLE) is regarded as a prototypic autoimmune disease with loss of immune tolerance to nucleic acid antigens. Antinuclear antibodies (ANAs), which are autoantibodies against nuclear components, including dsdna, are frequently found in patients with SLE (2). The importance of B cells in the pathogenesis of SLE is further confirmed by the fact that anti B cell activating factor (BAFF) monoclonal antibody (Belimumab) was approved by the Food and Drug Administration for treatment of SLE in 2011 (3). BAFF is a transmembrane protein member of the TNF ligand superfamily, and its overexpression is associated with both murine lupus and human SLE. However, as a sizeable proportion of patients with SLE remain refractory to Belimumab treatment, additional approaches for regulating B-cell hyperactivity are needed. Regulatory T-cell (Treg) subsets play a major role in the maintenance of immune homeostasis and the prevention of autoimmunity (4). The most extensively studied Treg subset is CD4 + CD25 + Foxp3 + Tregs (CD25 + Tregs) (5). The CD25 + Tregs exhibit their suppressive function mainly by expressing CTLA4 and CD25 on their cell surface (6). We previously reported CD4 + CD25 Foxp3 LAG3 + Tregs (LAG3 + Tregs), which exert their suppressive activities via IL-10 production in a Foxp3- independent manner (7). In contrast to CD25 + Tregs, high-affinity interactions with peptide/mhc ligands expressed in the thymus are not required for the development of LAG3 + Tregs. Accumulating evidence has shown that CD4 + CD25 + Foxp3 + CXCR5 + Tregs called follicular regulatory T (Tfr) cells directly control GC responses and humoral immunity. However, there is Significance Transcription factors early growth response gene 2 (Egr2) and Egr3 have long been regarded as negative regulators of T-cell activation. Egr2 is also known as a susceptibility gene for systemic lupus erythematosus characterized by dysregulated humoral immune responses to autoantigens. Previously, we reported that Egr2-expressing CD4 + CD25 - LAG3 + regulatory T cells regulate lupus pathogenesis via production of TGF-β3. However, the role of Egr2 and Egr3 in the regulation of humoral immunity is unclear. Here we report that Egr2 and Egr3 regulate germinal center reactions by promoting TGF-β3 production from regulatory T cells. Egr2 and Egr3 induce the expression of latent TGF-β binding protein 3 (Ltbp3), which is required for TGF-β3 secretion. These findings suggest that Egr2 and Egr3 in T cells may be potential novel therapeutic targets for autoantibody-mediated autoimmune diseases. Author contributions: K.M., T.O., K.Y., and K.F. designed research; K.M., T.O., M.I., T.K., S.T., Y.I., S.S., and H.S. performed research; K.M., T.O., M.I., T.K., S.T., Y.I., S.S., H.S., and K.F. contributed new reagents/analytic tools; K.M., T.O., K.Y., and K.F. analyzed data; and K.M., T.O., K.Y., and K.F. wrote the paper. Conflict of interest statement: K.Y. received financial support or fees from AbbVie, Astellas, BMS, Daiichi-Sankyo, Mitsubishi Tanabe, Pfizer, Sanofi, Santen, Takeda, Teijin, Boehringer Ingelheim, Chugai, Eisai, Ono, Taisho Toyama, UCB, ImmunoFuture, Asahi Kasei, and Janssen. K.F. received financial support or fees from Astellas, BMS, Daiichi-Sankyo, Mitsubishi Tanabe, Pfizer, Santen, Takeda, Chugai, Eisai, Taisho Toyama, UCB, and Janssen. The remaining authors declare no competing financial interests. This article is a PNAS Direct Submission. Freely available online through the PNAS open access option. Data deposition: The sequence reported in this paper has been deposited in the GenBank database (accession no. NM_ ). 1 To whom correspondence may be addressed. kfujio-tky@umin.ac.jp or tomohisatky@umin.ac.jp. This article contains supporting information online at /pnas /-/DCSupplemental. IMMUNOLOGY AND INFLAMMATION PNAS Early Edition 1of10

2 little evidence regarding the contributions of other Treg populations to humoral immune tolerance. Recently, we have reported that LAG3 + Tregs (7) regulate humoral immunity and lupus disease in MRL-Fas lpr/lpr (MRL/lpr) mice via TGF-β3 production (8). Although TGF-β1 is well known for its antiinflammatory effects (9), we have previously revealed the regulatory activity of TGF-β3 on humoral immune responses. LAG3 + Tregs, which characteristically express the transcription factor early growth response gene 2 (Egr2), were identified as Foxp3-independent Tregs that produce high amounts of IL-10, and forced expression of Egr2 in naïve T cells induced IL-10 and LAG3 expression (7). Furthermore, we and our collaborators have also shown that polymorphisms in EGR2 influence SLE susceptibility (10). Intriguingly, lymphocytespecific Egr2-deficient mice develop a mild lupus-like autoimmune phenotype (11). These studies suggest that the expression of Egr2 in LAG3 + Tregs contributes to the control of SLE pathogenesis. Egr2, a member of the Egr family, is a C 2 H 2 -type zinc finger transcription factor that was first identified as a major regulator of myelination and hindbrain development (12, 13). Egr2 deficiency results in perinatal or neonatal death due to respiratory or feeding deficits (12). Recent studies have focused on the role of Egr2 in immune responses and revealed that Egr2 is essential for full induction of T-cell clonal anergy (14, 15). Egr2 has long been regarded as a negative regulator of T-cell activation by promoting expression of the E3 ubiquitin ligase Cbl-b and the cyclin-dependent kinase inhibitor p21 cip1 and p27 kip, which also contribute to T-cell anergy induction (11, 15, 16). However, because CD2-Cre driven lymphocyte-specific Egr2-deficient mice demonstrated only a mild form of systemic autoimmunity with limited anti-dsdna antibody production (11), the presence of additional regulators that control autoimmunity has been speculated. Among the four Egr family members (Egr1 4), it is thought that Egr3 is able to partially compensate for Egr2 (11), although a systemic deletion of Egr3 induces only gait ataxia in mice due to the lack of muscle spindles (17). Actually, Egr2 and Egr3 deletion in both T cells and B cells causes a more severe early-onset systemic autoimmune syndrome, compared with deletion of Egr2 alone using a CD2-Cre driver (18, 19). The Egr2/Egr3 double-deficient mice showed enhanced effector T-cell differentiation due to the reduction of suppressor of cytokines signaling 1 (SOCS1) and SOCS3 and induction of Batf (18). Whereas there has been no report of T cell- or T cell/b cellspecific Egr3-deficient mice that develop spontaneous systemic autoimmunity (18, 20), these results indicated a compensatory role of Egr3 for Egr2-mediated control of systemic autoimmunity. Nevertheless, both Egr2 and Egr3 expressed in B cells may modulate systemic autoimmunity in CD2-Cre driven lymphocyte-specific Egr2/Egr3 double-deficient mice, because Egr2 expressed in B cells regulates the development of B cells (21) and Egr3 is preferentially expressed in follicular B cells and marginal zone B cells, among various B-cell populations (22). Therefore, it remains elusive whether and how Egr2 and Egr3 expression on T cells solely regulates humoral immune responses. In this report, we identify a previously unknown role of Egr2 and Egr3 in T cells in the regulation of humoral immunity. To elucidate the effect of both Egr2 and Egr3 in T cells, we generated T cellspecific Egr2/Egr3 double-deficient mice. The double-deficient mice developed an earlier onset lupus-like syndrome compared with T cell-specific Egr2 single-deficient mice. The phenotype in T cell-specific Egr2/Egr3 double-deficient mice is attributed to insufficient production of TGF-β3 fromlag3 + Tregs, which was associated with reduced expression of latent TGF-β binding protein (Ltbp)3 required for the assembly and secretion of TGF-β3 (9). Results Egr2/3 Double Conditional KO Mice Develop a More Severe Lupus-Like Autoimmune Disease than Egr2 Conditional Single KO Mice. We previously established Egr2 fl/fl Cd4-Cre + [Egr2 conditional single knockout (CKO)] mice, and T cell-specific Egr2-deficiency led to the development of a mild form of systemic autoimmunity at nearly 1yofage(8).ToinvestigatetheroleofEgr2andEgr3in T cells, we constructed a mouse strain in which both Egr2 and Egr3 were deleted specifically in T cells. First, we established loxpflanked alleles encoding Egr3 (Egr3 fl/fl )mice(fig. S1A). Egr3 fl/fl mice and Egr2 floxed (Egr2 fl/fl ) mice (23) were then crossed with mice transgenic for Cd4-Cre + mice to obtain Egr3 fl/fl Egr2 fl/fl Cd4- Cre + [Egr2 and Egr3 double conditional knockout (Egr2/3DKO)] mice. When efficiency of the Cre-mediated recombination in T cells was evaluated in initial breedings (Fig. S1 B D), Egr2 and Egr3 expression was abrogated in naïve T cells, but not in B cells, derived from Egr2/3DKO mice. Egr2/3DKO mice exhibited a statistically significant decrease in survival compared with WT and Egr2CKO mice (Fig. 1A). Egr2/3DKO mice generated higher concentrations of anti-dsdna autoantibodies in their serum and progressive proteinuria (Fig. 1 B and C), which are the hallmark features of SLE in humans (24). Antibodies to dsdna were also detected with the Crithidia luciliae immunofluorescence tests (Fig. S2A). Egr2/3DKO mice also had inflammatory dermatitis, which was not observed in WT and Egr2CKO mice. Beginning at 16 wk of age, Egr2/3DKO mice developed skin inflammation on their backs that was accompanied by hair loss (Fig. 1D). Skin sections from Egr2/3DKO mice at 16 wk of age showed extensive inflammatory cell infiltration, acanthosis, and degeneration of the basal layer (Fig. S2B), which are often observed in SLE in humans. Consistent with proteinuria progression, kidney sections from Egr2/3DKO mice showed hyperplasia of mesangial cells at 16 wk of age (Fig. 1 E and F) and massive inflammatory cell infiltration and glomerular crescent formation at 36 wk of age (Fig. S2 C and D). When we evaluated glomerular deposition of immune complexes in the kidney, one of the distinctive histologic findings in SLE, immunohistochemical analyses of kidney sections revealed considerable deposition of IgG in the glomeruli of Egr2/3DKO mice (Fig. 1E). In addition to the lupuslike organ damage, Egr2/3DKO mice also exhibited massive organ inflammation, including liver, stomach, salivary gland, lung, and pancreas (Fig. S2E). This multiorgan inflammation contrasts sharply with the lung-restricted cellular infiltrate in Egr2CKO mice. Although Egr2CKO mice also demonstrated a lupus-like disease, their phenotypes were less severe than those of Egr2/ 3DKO mice. Consistent with our previous report (7), the expression of Egr2 mrna was much higher in LAG3 + Tregs than in CD4 + CD25 CD44 low CD62L high naïve T cells and CD4 + CD25 + Tregs (Fig. 1G). It was reported that, among the Egr family, Egr3 has a similar role in establishing T-cell anergy as Egr2 (25), and potentially compensates for Egr2 in Egr2-deficient conditions (18). Intriguingly, LAG3 + Tregs from WT mice did not express high levels of Egr3; however, Egr2 deficiency heightened the expression of Egr3 mrna, especially in LAG3 + Tregs, presumably due to a compensatory effect (Fig. 1H). Loss of Egr3 did not change the expression of Egr2 mrna (Fig. 1G). Together, these results suggested that genetic absence of both Egr2 and Egr3 in T cells, including LAG3 + Tregs, leads to the spontaneous autoimmune disease resembling SLE and that Egr3 partially compensated for Egr2 deficiency. Excessive Development of Follicular Helper T Cells and GC B Cells in Egr2/3DKO Mice. To investigate the pathogenicity of CD4 + T cells under Egr2 and Egr3 deficiency, we examined CD4 + T-cell profiles of WT, Egr2CKO, and Egr2/3DKO mice. Egr2 has been found to be involved in the positive selection of thymocytes by up-regulating the survival molecule Bcl-2 and IL-7 (26, 27). Consistent with previous reports, we found a small reduction in the frequency of CD4 + thymocytes in Egr2/3DKO mice compared with WT mice (Fig. S3A). However, the frequency and numbers of CD4 + thymocytes were not reduced in Egr2/3DKO mice (Fig. S3B). In contrast, Egr2/3DKO mice had severe splenomegaly and increased numbers of lymphocytes compared 2of10 Morita et al.

3 IMMUNOLOGY AND INFLAMMATION PNAS PLUS Fig. 1. Egr2/3DKO mice develop a more severe lupus-like autoimmune disease than Egr2CKO mice. (A) Survival rates of WT, Egr2 fl/fl Cd4-Cre + (Egr2CKO), and Egr2 fl/fl Egr3 fl/fl Cd4-Cre + (Egr2/3DKO) mice at the indicated time periods (n = 20 per group). P = (log-rank test). (B) Titers of anti-dsdna antibody from serum of WT, Egr2CKO, and Egr2/3DKO mice at 16 wk of age (n = 10 per group). *P < 0.05 (Bonferroni posttest). (C) Proteinuria progression of WT, Egr2CKO, and Egr2/3DKO mice (n = 10 per group). *P < 0.05 (Mann Whitney u test). (D) Skin inflammation in Egr2/3DKO mice. A representative macroscopic view (Top) and H&E staining (Bottom) of the back skin from WT, Egr2CKO, and Egr2/3DKO mice at 16 wk of age. (Scale bars, 500 μm.) (E) Histopathological analysis of kidneys from WT, Egr2CKO, and Egr2/3DKO mice at 16 wk of age. H&E staining (Top), periodic acid-schiff staining (PAS; Middle), and anti-igg immunofluorescent staining (IgG; Bottom). (Scale bars, 50 μm.) (F) Histopathological scoring of extent and severity of renal disease from mice as in E (n = 10 per group). (G and H) Quantitative RT-PCR (qrt-pcr) analysis of the expression of Egr2 (G) and Egr3 (H) mrna in T-cell subsets from WT, Egr2CKO, and Egr2/3DKO mice. Results are presented relative to expression of Actb mrna encoding β-actin. n.d., not detected (n = 3 per group). *P < 0.05 (Bonferroni posttest). Data in G and H are representative of three independent experiments. The mean ± SD are indicated. with WT and Egr2CKO mice (Fig. 2A). Analysis of peripheral CD4 + T cells from spleens showed that Egr2/3DKO mice had a larger population of CD4 + CD25 CD44 high CD62L low effector/ memory T cells than WT and Egr2CKO mice (Fig. 2B). It was demonstrated that Egr2 and Egr3 negatively regulate production of Th1 and Th17 cytokines including IFN-γ and IL-17 through a SOCS1/3-dependent manner (18). Consistent with these previous reports, IFN-γ and IL-17 production was increased in Th1 and Th17 condition from aged Egr2CKO mice (Fig. S4A). However, the additional deficiency of Egr3 in Egr2CKO mice did not cause the enhanced differentiation of Th1 and Th17 cells in vitro (Fig. S4 A and B), indicating that Egr3 does not compensate the Egr2 role in Th1 and Th17 differentiation intrinsically. Although freshly isolated total CD4 + T cells from aged Egr2CKO mice produced larger amount of IFN-γ and IL-17 compared with those from WT mice, these cytokine levels were not increased in Egr2/3DKO mice compared with Egr2CKO mice (Fig. S4C). These findings suggest that the more severe autoimmune syndrome in Egr2/3DKO mice than Egr2CKO mice is not explained by enhanced response of Th1 and Th17 cells. As excessive numbers of follicular helper T (Tfh) cells have recently been reported to have harmful effects in autoimmune diseases both in mice and humans (28 31), we next examined the commitment of Egr2 and Egr3 to the differentiation of Tfh cells in Egr2CKO and Egr2/3DKO mice. In clear contrast with the few splenic CXCR5 + PD-1 + Tfh cells observed in unimmunized WT mice, Egr2CKO mice exhibited an increased frequency and number of Tfh cells from 6 wk of age (Fig. 2 C and D and Fig. S3C). Furthermore, Egr2/3DKO mice displayed an augmented spontaneous development of Tfh cells compared with Egr2CKO mice. Excessive development of Tfh cells was also observed in cervical, lumbar, and inguinal lymph nodes of Egr2CKO and Egr2/3DKO mice compared with those of WT mice (Fig. S3E). GC B cells are required for the development of GCs, as well as Tfh cells. In GC, GC B cells undergo somatic hypermutation (SHM) (32, 33) and Ig class switching (34), and differentiate into either memory cells or plasma cells that confer lasting humoral immune responses. SHM results in stochastic changes in antibody affinity and specificity and may inadvertently generate autoreactive B cells. It is well known that the development of Tfh cells and GC B cells are mutually dependent on each other (35). Consistent with uncontrolled Tfh cell accumulation in Egr2CKO and Egr2/3DKO mice, we also observed spontaneous accumulation of GL7 + Fas + GC B cells in the spleens of Egr2CKO and Egr2/3DKO mice (Fig. 2E). Both the frequency and number of GC B cells from Egr2/3DKO mice were higher than those from WT and Egr2CKO mice (Fig. 2F and Fig. S3D). Similar to Tfh cells, excessive GC B-cell formation was also observed in the peripheral lymph nodes of Egr2CKO and Egr2/3DKO mice (Fig. S3F). Collectively, these data indicate that the expression of both Morita et al. PNAS Early Edition 3of10

4 Fig. 2. The deficiency of Egr2 and Egr3 in T cells leads to excessive Tfh and GC B-cell formation. (A) Representative macroscopic view (Left), weight (Middle), and total cellularity (Right) of the spleens from WT, Egr2CKO, and Egr2/3DKO mice at 12 wk of age (n = 6 per group). *P < 0.05 (Bonferroni posttest). (B)Flow cytometry analysis of the expression of CD44 and CD62L in splenic CD4 + CD25 TcellsfrommiceasinA (Left). Numbers adjacent to outlined areas indicate percentofcd44 high CD62L low memory CD4 + T cells. Graph indicates frequency (among total CD4 + CD25 T cells) of memory CD4 + T cells from mice as in A (Right, n = 6 per group). *P < 0.05 (Bonferroni posttest). (C) Flow cytometry analysis of the expression of PD-1 and CXCR5 in splenic CD4 + CD25 T cells from WT, Egr2CKO, and Egr2/3DKO mice at 18 wk of age. Numbers adjacent to outlined areas indicate percent of PD-1 + CXCR5 + Tfh cells. (D) Frequency (among total CD4 + CD25 T cells) of Tfh cells in the spleens of mice from WT, Egr2CKO, and Egr2/3DKO mice at different weeks of age (n = 6pergroup).*P < 0.05 (Bonferroni posttest). (E) Flow cytometry analysis of the expression of GL7 and Fas in splenic B220 + B cells from mice as in C. Numbers adjacent to outlined areas indicate percent of GL7 + Fas + GCBcells.(F) Frequency (among total B220 + B cells) of GC B cells in the spleens of mice as in D (n = 6pergroup).*P < 0.05 (Bonferroni posttest). Data are representative of three independent experiments. The mean ± SD are indicated. Egr2 and Egr3 in T cells is necessary for the regulation of Tfh and GC B-cell differentiation. Dysfunction of LAG3 + Tregs Is the Cause of Aberrant GC Responses in Egr2/3DKO Mice. We have previously reported that LAG3 + Tregs regulate humoral immune responses by suppressing B-cell proliferation and antibody production in an Egr2-dependent manner (8). As described above, Egr3 up-regulation in Egr2CKO LAG3 + Tregs suggests a compensatory role of Egr3 for Egr2 in Egr2CKO LAG3 + Tregs (Fig. 1 G and H). Therefore, we addressed whether the uncontrollable Tfh and GC B-cell formation in Egr2/3DKO mice reflects the absence or malfunction of LAG3 + Tregs. Unexpectedly, LAG3 + Tregs were present in both Egr2CKO and Egr2/3DKO mice (Fig. 3A), and the frequency and number of LAG3 + Tregs were higher in Egr2CKO and Egr2/3DKO mice. Next, we assessed whether deletion of both Egr2 and Egr3 diminishes the regulatory function of LAG3 + Tregs against B-cell antibody responses. In an in vitro T-cell B-cell coculture assay, anti CD3-stimulated LAG3 + Tregs from WT and Egr2/3DKO mice were cultured with WT B cells in the presence of anti-cd40 and recombinant IL-4 (ril-4) for 3 and 7 d. Consistent with our previous report (8), LAG3 + Tregs from WT mice effectively suppressed B-cell proliferation, survival, and IgG antibody production. In contrast, LAG3 + Tregs from Egr2/3DKO mice lost their suppressive capacity (Fig. 3 B D). We have previously defined LAG3 + Tregs as those that do not express CD25 and Foxp3 (7). In Egr2CKO and Egr2/3DKO mice, LAG3 + Tregs were still negative for both Foxp3 and CD25 expression (Fig. S5 A and B). The ratio of CD25 + Tregs in Egr2CKO and Egr2/3DKO mice was also unchanged compared with WT mice (Fig. S5C). Unlike LAG3 + Tregs from Egr2/3DKO mice, CD25 + Tregs from Egr2/ 3DKO mice still exhibit suppressive effects on B-cells, which was comparable to that of WT mice (Fig. S5 D and E), indicating that Egr3 in concert with Egr2 control the regulatory function of LAG3 + Tregs, but not CD25 + Tregs, on humoral immune responses. Next, we investigated whether malfunction of Egr2- and Egr3-deficient LAG3 + Tregs leads to excessive accumulation of Tfh and GC B cells in Egr2/3DKO mice. We transferred splenic LAG3 + Tregs from WT mice into Egr2/3DKO mice twice and analyzed their effect on Tfh and GC B-cell formation. Transfer of Egr2- and Egr3-sufficient LAG3 + Tregs effectively suppressed excess Tfh cell and GC B-cell formation (Fig. 3 E and F). As LAG3 + Tregs control humoral immunity via TGF-β3, we investigated whether Egr2 and Egr3 regulate TGF-β3 production through LAG3 + Tregs. At 3 d after T-cell receptor (TCR) stimulation, LAG3 + Tregs from WT mice secreted about 10 ng of TGFβ3 protein,whichwas 50-fold higher than TGF-β1 protein produced by CD25 + Tregs (Fig. 3 G and H). In contrast, LAG3 + Tregs from Egr2CKO mice secreted lower levels of TGF-β3 proteinthan WT LAG3 + Tregs. Moreover, we did not detect any TGF-β3 secretion in the supernatants of LAG3 + Tregs from Egr2/3DKO mice. Although it was reported that TGF-β1 could partially compensate the function of TGF-β3 (36), we did not detect TGF-β1 production from either WT LAG3 + Tregs, Egr2CKO, or Egr2/ 3DKO LAG3 + Tregs (Fig. 3H). These results suggest that the aberrant development of Tfh and GC B cells in Egr2/3DKO mice might be caused by the defective regulatory function of LAG3 + Tregs to produce TGF-β3. As Egr2 and Egr3 have been reported to be transiently upregulated in response to TCR stimulation (15, 25), we also explored the possibility that Egr2 and Egr3 intrinsically modulate the differentiation of Tfh cells. To evaluate antigen-specific Tfh responses, naïve CD4 + T cells from WT OT-II, Egr2CKO OT-II, or Egr2/3DKO OT-II mice that express TCR specific for the ovalbumin (OVA) peptide in the context of I-A b were adoptively transferred into CD C57BL/6 (B6) recipient mice, followed by immunization of the recipients with OVA conjugated to 4-hydroxy-2-nitrophenylacetyl (NP-OVA) in complete Freund s adjuvant (CFA). At 7 d after the immunization, the frequency of CD4 + CXCR5 + PD-1 + Tfh cells and the expression of Bcl-6 protein in Egr2-, or Egr2- and Egr3-deficient CD T cells were similar to those in their WT counterparts (Fig. 4 A and B). The frequency of GC B cells and the levels of NP-specific IgG were also similar in all groups of recipient mice (Fig. 4 C and D). Thus, Egr2 and Egr3 had little effect on the differentiation of Tfh cells, supporting that Egr2 and Egr3 expression in LAG3 + Tregs controls aberrant GC responses. 4of10 Morita et al.

5 IMMUNOLOGY AND INFLAMMATION PNAS PLUS Fig. 3. Dysregulated function of LAG3 + Tregs is responsible for the excessive Tfh and GC B-cell formation in Egr2/3DKO mice. (A) Flow cytometry analysis of splenic CD4 + CD25 T cells from WT, Egr2CKO, and Egr2/3DKO mice at 18 wk of age (Left). Numbers adjacent to outlined areas indicate percent of CD4 + LAG3 + CD45RB low T cells (LAG3 + Tregs). Graph indicates frequency (among total CD4 + CD25 T cells) of LAG3 + Tregs (Right, n = 8 per group). *P < 0.05 (Bonferroni posttest). (B) Flow cytometry analysis of CFSE-labeled B-cell proliferation. Each anti CD3-stimulated T-cell subset was cocultured with WT B cells stimulated with anti-cd40 mab and ril-4for4d(n = 4 per group). (C) Viability of cocultured B cells as in B was assessed by 7-amino-actinomycin D (7-AAD) (n = 4 per group). (D) Quantification of total IgG production in the culture supernatants of cocultured B cells (as in B)onday7(n = 4 per group). *P < 0.05 (Bonferroni posttest). (E and F) Flow cytometry analysis of splenic Tfh and GC B cells in Egr2/3DKO mice. Splenic LAG3 + Tregs from WT mice were transferred into Egr2/3DKO mice at 2 and 4 wk of age. Four weeks after the last cell transfer, frequency (among total CD4 + CD25 T cells) of Tfh cells (E) and frequency (among total B220 + Bcells)ofGCBcells(F) inthespleensofegr2/ 3DKO mice were examined (n = 6 per group). *P < 0.05 (Bonferroni posttest). (G and H) TGF-β3 (G) andtgf-β1 (H) protein levels in the culture supernatants of stimulated LAG3 + Tregs from WT, Egr2CKO, and Egr2/3DKO mice on day 3 (n = 4 per group). *P < 0.05 (Bonferroni posttest). Data are representative of three (A and E H) ortwo(b D) independent experiments. The mean ± SD are indicated. TGF-β3 Negatively Regulates GC B-Cell Differentiation. Although we have previously shown that TGF-β3 is responsible for the suppressive activity of LAG3 + Tregs on B-cell functions (8), TGF-β1 has long been known as the inhibitory cytokine for B-cell responses through inhibiting the activation of Syk and phospholipase C-γ2, as well as Stat6 phosphorylation (37). TGF-β3 shares 72%amino acid identity with TGF-β1, and TGF-β3 binds to the TGF-β receptor II in the same way as TGF-β1 does (38, 39). Restrained differentiation of Tfh and GC B cells by TGF-β signaling was demonstrated using TGF-β receptor II KO mice (40), indicating that TGF-β is necessary for the control of humoral immunity. To investigate the role of TGF-β3 in Tfh and GC B-cell formation, we first analyzed the suppressive function of TGF-β3 in B cells after in vitro activation under conditions that promote the differentiation of GCphenotype B cells. Although B cells stimulated with anti-cd40, ril-4, and rbaff acquired GL7 and Fas expression, which is the typical phenotype of GC B cells (41), the addition of TGF-β3 or TGF-β1 suppressed the acquisition of GL7 + Fas + GC phenotype and proliferation of B cells (Fig. 5 A and B). Moreover, treatment with TGF-β3 or TGF-β1 effectively inhibited the expression of Bcl6 mrna, which is essential for the differentiation of GC B cells (Fig. 5C) and the production of IgG, IgA, and IgM in the culture supernatants at day 7 (Fig. 5D). Notably, TGF-β3 and TGF-β1 did not suppress the expression of Bcl6 mrna in in vitro induced Tfh cells (Fig. 5E). Thus, TGF-β3 acted as a negative regulator of the differentiation of GC B cells as well as TGF-β1. These results also suggest that B cells are the primary targets of TGF-β mediated suppression. We next examined whether excessive Tfh and GC B-cell formation in Egr2/3DKO mice was rescued by exogenous TGF-β3. We injected a TGF-β3 expression plasmid (pcaggs-tgfb3)ora control pcaggs plasmid into Egr2/3DKO mice three times and analyzed Tfh and GC B formation at 4 wk after the final injection. Although the excessive Tfh formation was not improved by the injection of pcaggs-tgfb3, the frequency of GC B cells was markedly reduced in Egr2/3DKO mice injected with pcaggs- Tgfb3, compared with those injected with control pcaggs (Fig. 5 F and G and Fig. S6). This result suggests that loss of TGF-β3 production from Egr2- and Egr3-deficient LAG3 + Tregs was responsible for the cause of unrestrained GC B-cell formations in Egr2/3 DKO mice. Moreover, in accordance with suppression of GC B cells, the injection of pcaggs-tgfb3 improved proteinuria Morita et al. PNAS Early Edition 5of10

6 Fig. 4. Egr2 and Egr3 deficiency do not augment Tfh cell differentiation intrinsically. (A) Flow cytometry analysis of the expression of PD-1 and CXCR5 in donor CD OT-II CD4 + T cells obtained from draining lymph nodes (dlns) of CD recipient WT mice given adoptive transfer of naïve WT, Egr2CKO, or Egr2/3DKO CD OT-II CD4 + T cells, followed by immunization of NP 13 -OVA in CFA (Left). Numbers adjacent to outlined areas indicate percent of PD-1 + CXCR5 + Tfh cells. Graph indicates frequency (among total CD4 + CD25 CD T cells) of PD-1 + CXCR5 + Tfh cells from dlns (Right, n = 6 per group). (B) Flow cytometry analysis of the expression of Bcl-6 and CXCR5 in donor CD OT-II CD4 + T cells obtained from dlns as in A (Left). Numbers adjacent to outlined areas indicate percent of Bcl-6 + CXCR5 + Tfh cells. Graph indicates frequency (among total CD4 + CD25 CD T cells) of Bcl-6 + CXCR5 + Tfh cells from dlns as in A (Right, n = 6 per group). (C) Flow cytometry analysis of the expression of GL7 and Fas in B220 + B cells obtained from dlns as in A (Left). Numbers adjacent to outlined areas indicate percent of GL7 + Fas + GC B cells. Graph indicates frequency (among total B220 + B cells) of GC B cells from dlns as in A (Right, n = 6 per group). (D) Quantification of NP-specific IgG in the serum of the recipient mice as in A (n = 6 per group). Data are representative of two independent experiments. The mean ± SD are indicated. progression and titer of anti-dsdna autoantibodies (Fig. 5 H and I). Collectively, these results demonstrate that TGF-β3 production controlled by Egr2 and Egr3 in LAG3 + Tregsisnecessaryforthe regulation of aberrant GC B-cell differentiation and related to disease pathogenesis of Egr2/3DKO mice. Egr2 and Egr3 Regulate the Secretion of TGF-β3 Protein Through an Ltbp3-Dependent Manner. We next explored the molecular basis for the requirement of Egr2 and Egr3 for TGF-β3 production from LAG3 + Tregs. Surprisingly, although the secretion of TGF-β3 from LAG3 + Tregs was strongly regulated by Egr2 and Egr3, the expression of Tgfb3 mrna was not altered in Egr2- and Egr3-deficient LAG3 + Tregs, irrespective of the presence of TCR stimulation (Fig. 6 A and B). This result indicates that the expression of Tgfb3 mrna is not directly regulated by Egr2 or Egr3. TGF-β3 undergoes complex processing steps intracellularly before its secretion from a cell membrane (42, 43). After translation, TGF-β3 precursor protein is cut by furin convertase and forms a small latent complex (SLC) that includes mature TGF-β3 and latency associated peptide (LAP). SLCs are usually associated with Ltbp and secreted outside of the membrane as a large latent complex (LLC). Given that Ltbp has been reported to require binding to TGF-β for efficient secretion (44, 45), we hypothesized that the secretion of TGF-β3 is also dependent on binding to Ltbp. Among the Ltbp family that consists of four members from Ltbp1 4 (46), LAG3 + Tregs characteristically express Ltbp3 mrna (Fig. 6C and Fig. S7A). Furthermore, the expression of Ltbp3 mrna was significantly down-regulated in Egr2/3DKO LAG3 + Tregs compared with WT and Egr2CKO LAG3 + Tregs (Fig. 6C). To investigate whether Ltbp3 is essential for TGF-β3 to be secreted, we transfected LAG3 + Tregs with sirna specifically designed to downregulate the expression of Ltbp3 and analyzed the secretion of TGF-β3 protein after TCR stimulation. Although the expression of Tgfb3 mrna was unchanged, the secretion of TGF-β3 protein was strongly down-regulated in Ltbp3 sirna-transfected LAG3 + Tregs (Fig. 6D). The expression of Ltbp1, 2, and 4 was not suppressed by Ltbp3 sirna (Fig. S7B). These results demonstrate that Egr2 and Egr3 positively regulate the secretion of TGF-β3 by inducing Ltbp3 expression. To elucidate whether overexpression of Ltbp3 rescues the loss of TGF-β3 secretion in Egr2/3DKO T cells, we next constructed pmig-ltbp3 retroviral vector. Because we previously reported that IL-27 treated T cells express Egr2 and produce TGF-β3 (8, 47), we retrovirally transduced WT or Egr2/3DKO T cells with pmig-mock or pmig- Ltbp3, followed by treatment with IL-27, and determined the production of TGF-β3 in the culture supernatants. As shown in Fig. 6E, transduction of pmig-ltbp3 significantly increased the production of TGF-β3 inil-27 treated Egr2/3DKO T cells. Moreover, IL-27 treated Egr2/3DKO T cells retrovirally transduced with pmig-ltbp3 significantly suppressed B-cell proliferation and antibody production compared with those transduced with pmig-mock (Fig. 6 F and G). These results demonstrated that Ltbp3 expression in LAG3 + Tregs plays a critical role in controlling humoral immunity via TGF-β3 secretion. Discussion Although it is known that dysregulated adaptive immunity is associated with the pathogenesis of SLE, the underlying distinct molecular mechanisms have been elusive. We demonstrated in this report that Egr2 and Egr3 in T cells are central molecules for the maintenance of humoral immune tolerance in the steady state. In addition to the previously observed linkage between Egr2 and systemic autoimmunity in mice and humans (10, 11), our studies indicate that the combination of Egr2 and Egr3 in T cells is a promising candidate target involved in the regulation of humoral immunity. The absence of both Egr2 and Egr3 in T cells led to earlier onset of a lupus-like syndrome compared with Egr2CKO mice, indicating that a compensatory function of Egr3 for Egr2 in T cells is necessary for the prevention of lupus pathogenesis. Intriguingly, Egr2/3DKO mice not only developed a lupus-like phenotype characterized by marked high titers of anti-dsdna accompanied by severe glomerulonephritis resembling SLE, but also showed a severe systemic autoimmune syndrome with lymphocytic infiltration in multiple organs, such as liver and pancreas, that are rarely impaired in human lupus (48). These findings suggest that Egr2 and Egr3 expression in T cells is required for the immune system to avoid attacking self-tissue affected by a wide range of autoimmune disorders that are not limited to lupus. SLE is a B-cell mediated autoimmune disease characterized by the loss of tolerance to nucleic acid antigens and is regulated by a variety of mechanisms including aberrant Tfh cell differentiation (49). Our observation that Bcl-6 expression in CD4 + TcellsinTfh 6of10 Morita et al.

7 IMMUNOLOGY AND INFLAMMATION PNAS PLUS Fig. 5. TGF-β3 mediated suppression of B-cell differentiation. (A) Flow cytometry analysis of the expression of GL7 and Fas in CFSE-labeled B cells stimulated without cytokines (no stim.) or with ril-4, anti-cd40 mab, and BAFF in the presence or absence of rtgf-β1 or-β3(left). Numbers adjacent to outlined areas indicate percent of GL7 + Fas + B cells. Graph indicates frequency of GL7 + Fas + B cells (Right, n = 4 per group). *P < 0.05 (Bonferroni posttest). (B) Flow cytometry analysis of CFSE-labeled B-cell proliferation stimulated as in A (Left). Graph indicates frequency of undivided B cells (Right, n = 4 per group). *P < 0.05 (Bonferroni posttest). (C) qrt-pcr analysis of the expression of Bcl6 and Aicda mrna in B cells stimulated as in A (n = 4 per group). *P < 0.05 (Bonferroni posttest). (D) Quantification of IgG, IgA, and IgM production in the culture supernatants of B cells stimulated as in A (n = 4 per group). *P < 0.05 (Bonferroni posttest). (E) qrt-pcr analysis of the expression of Bcl6 mrna in T cells stimulated with ril-6 and ril-21 (Tfh cell condition) in the presence or absence of rtgf-β1 or-β3 for4d(n = 3 per group). *P < 0.05 (Bonferroni posttest). (F and G) Analysis of splenic Tfh and GC B cells in Egr2/3DKO mice after i.v. injection with pcaggs-control or pcaggs-tgfb3 plasmid vector. Frequency (among total CD4 + CD25 T cells) of Tfh cells (F) and frequency (among total B220 + Bcells)ofGCBcells(G) in the spleens of Egr2/3DKO mice were examined at 4 wk after last injection (n = 6 per group). *P < 0.05 (Bonferroni posttest). (H) Quantification of serum anti-dsdna antibodies in Egr2/3DKO mice treated as in F (n = 6per group). *P < 0.05 (Student s t test). (I) Proteinuria progression in Egr2/3DKO mice treated as in F (n = 6 per group). *P < 0.05 (Mann Whitney u test). Data are representative of two independent experiments. The mean ± SD are indicated. cell-polarizing conditions was not suppressed by TGF-β3 does not exclude the suppressive activity of TGF-β in longer duration because Tfh cells overly accumulate in the absence of TGF-β signaling in T cells (40). Ogbe et al. reported that Egr2 and Egr3 directly regulate Bcl-6 expression and are essential for the late stage of Tfh cell differentiation during viral infection (50). They also revealed that the key molecules involved in the regulation of Tfh cell chemotaxis, such as expression levels of Cxcr5, Icos, and Sh2d1a, are not affected by Egr2/Egr3 deficiency in lymphocytes. Similarly, we revealed that absence of Egr2/Egr3 does not alter Tfh cell development. These results suggested that Tfh cell intrinsic expression of Egr2/Egr3 does not explain the enhanced humoral immunity observed in Egr2/3DKO mice. The fact that treatment with a TGF-β3 expression vector ameliorated only GC B development in Egr2/ 3DKO mice indicated that a primary direct target of suppression mediated by Egr2/Egr3 TGF-β3 axis may be B cells. To date, a number of studies have demonstrated that Treg subsets maintain immunological self-tolerance. Foxp3 + CD25 + Tregs have been the most intensively studied based on the accumulating evidence for their therapeutic effects to prevent various autoimmune diseases. It has been reported that some specialized CD25 + Treg subsets, such as Tfr cells (51) and CD4 + CD25 + CD69 Tregs (52), play a major role in the regulation of humoral immunity. However, contradictory results have been reported for the numbers and function of Treg in SLE patients, and the precise roles of CD25 + Tregs in SLE remain elusive (53), because the clinical manifestations of immunodysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome caused by a mutation in the FOXP3 gene and SLE differ considerably (54, 55). In addition to CD25 + Treg, Egr2-expressing LAG3 + Tregs may also contribute to the control of systemic autoimmunity. Egr2 and Egr3 appeared to exhibit regulatory activity in LAG3 + Tregs, because both Egr2- and Egr3-deficient CD25 + Tregs retain in vitro suppressive activity (18) and the adoptive transfer of WT LAG3 + Tregs into Egr2/ 3DKO mice effectively suppressed the aberrant development of both Tfh and GC B cells. In the present study, we observed that TGF-β3 suppressed B-cell activation and differentiation as efficaciously as TGF-β1, suggesting that TGF-β3 is a potent regulator of humoral immunity. It was recently identified that TGF-β1 plays an integral role in maintaining immune tolerance (9). As for regulation of humoral immunity, TGF-β1 controls B-cell activation by inhibiting both Ig synthesis and the switch from the membrane form to the secreted forms of μ- andγ-mrna (56). CD4 + CD25 + CD69 Treg-mediated Morita et al. PNAS Early Edition 7of10

8 Fig. 6. Egr2 and Egr3 regulate TGF-β3 secretion through induction of Ltbp3. (A) qrt-pcr analysis of the expression of Tgfb3 mrna in freshly isolated T-cell subsets (n = 3 per group). (B) qrt-pcr analysis of the expression of Tgfb3 mrna in T-cell subsets stimulated for 3 d with or without anti-cd3 and anti-cd28 mab (n = 3 per group). (C) qrt-pcr analysis of the expression of Ltbp3 mrna in T-cell subsets stimulated as in B (n = 3 per group). *P < 0.05 (Bonferroni posttest). (D) qrt-pcr analysis of the expression of Ltbp3 (Left) and Tgfb3 (Middle) mrna, and quantification of TGF-β3 protein levels (Right) in the culture supernatants of LAG3 + Tregs transfected with control sirna or Ltbp3 sirna (n = 4 per group). *P < 0.05 (Student s t test). (E) Analysis of TGF-β3 production in the culture supernatants of IL-27 treated T cells from WT and Egr2/3DKO mice, transduced with pmig-mock or pmig-ltbp3 vector. After CD4 + T cells were retrovirally transduced with pmig-mock or pmig-ltbp3 vector, the cells were treated with IL-27 for 2 d and TGF-β3 production was determined by ELISA (n = 4 per group). *P < 0.05 (Student s t test). (F) In vitro suppression of B cells by Ltbp3-transduced Egr2 and Egr3-deficient T cells. CFSE-labeled B cells were cocultured with pmig-mock or pmig-ltbp3 vector transduced CD4 + T cells that were treated with IL-27 after transduction. B-cell proliferation was assessed by CFSE after 4 d of culture (n = 4 per group). (G) Quantification of total IgG production in the culture supernatants of cocultured B cells as in F on day 7 (n = 4 per group). *P < 0.05 (Bonferroni posttest). Data are representative of three (A C) ortwo(d G) independent experiments. The mean ± SD are indicated. suppression of B-cell antibody production is mediated at least partially through expression of TGF-β1 (52). On the other hand, the immune suppressive role of TGF-β3 has not been extensively evaluated, and a mechanism for the production of TGF-β3 has not been clarified. We found that double deficiency of both Egr2 and Egr3 in T cells affects TGF-β3 production and Ltbp3 expression, and not Tgfb3 mrna expression, in LAG3 + Tregs. Among the four isoforms of Ltbp, Ltbp1, and Ltbp3 can associate efficiently with pro TGF-β1, -β2, and -β3, whereas Ltbp2 reportedly does not (57). On the other hand, Ltbp4 binds to TGF-β1 LAP more weakly than Ltbp1 and Ltbp3, indicating that Ltbp1 and Ltbp3 might be the primary proteins responsible for binding to the TGF-β LAP complex SLC (58). Several lines of evidence indicate that the Ltbps, which are not necessary for latency (59), play a major role in the secretion (44), extracellular membrane localization (17), and activation (60, 61) of latent TGF-β. However, it was reported that, unlike Ltbp1, Ltbp3 is not suitable for integrin-mediated latent TGF-β activation, and this isoform-specific function is most likely related to the great sequence divergence of their hinge domains (62). In this study, we revealed a previously unrecognized function for Ltbp3 as an efficient regulator of TGF-β3 secretion from LAG3 + Tregs. These results suggest that Ltbp3 may play a critical role in regulating TGF-β3 mediated immune tolerance and define Ltbp3 as a potential therapeutic target. Further investigation is needed to confirm whether Ltbp3 is a direct or indirect target of Egr2 and Egr3. As described above, our study suggests the importance of Egr2 and Egr3 in T cells for the maintenance of humoral immune tolerance. Although Li et al. reported that mice in which Egr2 and Egr3 are deleted specifically in both T and B cells die within 8 mo (18), three-quarters of T-cell specific Egr2 and Egr3 doubledeficient (Egr2/3DKO) mice were still alive after 1 y of age (Fig. 1A), suggesting the importance of Egr2 and Egr3 in B cells in the immune system. However, there have been no reports of B-cell specific Egr2/Egr3 double-deficient mice. Detailed analyses of the role of Egr2 and Egr3 in B cells may provide novel therapeutic targets on B cells for autoantibody-mediated autoimmune diseases. In summary, our findings have provided insight into the mechanism of how Egr2 and Egr3 in CD4 + T cells regulate humoral immunity and established the combination of Egr2 and Egr3 as a pivotal regulator of TGF-β3 secretion from LAG3 + Tregs. Understanding the molecular basis of TGF-β3 secretion will be key to understanding lupus pathogenesis and will provide new therapeutic avenues for manipulating the excessive humoral immune responses using LAG3 + Tregs. 8of10 Morita et al.

9 Materials and Methods Mice. C57BL/6 (B6) mice were purchased from Japan SLC. B6 mice congenic for the CD45 locus (B6-CD ) were purchased from Sankyo Lab Service. TCR transgenic OT-II mice (specific for the chicken ovalbumin peptide (amino acid residues ) in the context of MHC class II I-A b ) were purchased from The Jackson Laboratory. Egr2 fl/fl mice were provided by Patrick Charnay (INSERM) (23). CD4-Cre transgenic mice (line 4196), originally generated by C. B. Wilson and colleagues, were purchased from Taconic. Egr2 fl/fl mice were crossed with CD4-Cre transgenic mice to generate Egr2CKO mice (Egr2 fl/fl Cd4-Cre + ). Egr2/3DKO mice (Egr2 fl/fl Egr3 fl/fl Cd4-Cre + ) were generated by crossing Egr2CKO mice with Egr3 fl/fl mice. All animal experiments were approved by the ethics committee of the University of Tokyo Institutional Animal Care and Use Committee. Generation of Egr3 Floxed Mice. The targeting vector was constructed by inserting a 1.4-kb fragment containing exon 2 of the Egr3 gene flanked by loxp sites, 3.0 kb of a 3 sequence, 3.9 kb of a 5 sequence, and a neomycinresistance gene flanked by Frt sites into a pbluescriptii SK (+) vector. The targeting vector was linearized and transfected into B6 embryonic stem cells by electroporation. Recombinant ES clones were selected in medium supplemented with G418. Thymidine kinase was used as a counterselection. G418-resistant clones were screened for homologous recombination by PCR and Southern blot analysis. The positive clones were microinjected into blastocytes derived from BALB/c mice and transferred to surrogate mothers. Mating of chimeric male mice to B6 female mice resulted in the transmission of the floxed allele to the germline. The neomycin selection cassette flanked by Frt sites was excised in vivo by crossing the C57BL/6-Tg (CAG-FLP) mice. For detection of floxed alleles, genomic DNA obtained from B cells and T cells was assessed by PCR. PCR primer pairs were as follows: forward CGAGGACAAAAGCGTCGAAGCTC and reverse GATCAAGGCGATCCTAACTGAAC. Reagents, Antibodies, and Media. Purified and conjugated antibodies were purchased from BD Bioscience, ebiosciences, or Biolegend and recombinant cytokines were purchased from Miltenyi Biotec, R&D, and Biolegend. See SI Materials and Methods for details. Flow Cytometry and Cell Sorting. These procedures are described in SI Materials and Methods. Histopathological Examination. Histopathologic examination of WT, Egr2CKO, and Egr2/3DKO mice was done at 16 or 36 wk of age. See SI Materials and Methods for details. Renal pathology was graded as described in SI Materials and Methods. B- and T-Cell Isolation and Proliferation. These procedures are described in SI Materials and Methods. T Cell B Cell Coculture Assay. These procedures are described in SI Materials and Methods. Transfer of LAG3 + Tregs into Egr2/3DKO Mice. FACS-sorted LAG3 + Tregs ( cells) from WT mice were i.v. transfused into Egr2/3DKO mice at 2 and 4 wk of age. At 4 wk after the transfer, the mice were killed and splenic Tfh and GC B-cell formation was analyzed by flow cytometry. Adoptive Transfer of Naïve T Cells from OT-II Mice. These procedures are described in SI Materials and Methods. Injection of TGF-β3 Expressing Plasmid Vector into Egr2/3DKO Mice. Construction of TGF-β3 expressing pcaggs vector (pcaggs-tgfb3) was previously described (8). Egr2/3DKO mice were injected i.v. with 100 μg of pcaggs- Tgfb3 or control pcaggs in sterile PBS at 4, 6, and 8 wk of age. At 4 wk after the last injection, the mice were killed and splenic Tfh and GC B-cell formation was analyzed by flow cytometry. Transfection of sirna. LAG3 + Tregs ( cells per well) were transfected with sirna in Accell sirna delivery media (GE Healthcare) according to the manufacturer s protocol. sirna targeting Ltbp3 and control sirna were used at a concentration of 1 μm. At 48 h after the transfections, the cells were transferred to 96-well flat-bottomed plates coated with anti-cd3 mab (2 μg/ml) and anti-cd28 mab (2 μg/ml) in RPMI-1640 medium as described above. The cells were incubated for 72 h and then gene expression and TGFβ3 protein production were analyzed. Retroviral Transduction into IL-27 Treated T cells. These procedures are described in SI Materials and Methods. RNA Isolation, cdna Synthesis, and Quantitative Real-Time PCR. These procedures are described in SI Materials and Methods. Statistical Analysis. Survival rates were analyzed with the log-rank test. Quantitative proteinuria progression was analyzed with the Mann Whitney u test. For the comparison of more than three groups, a one-way analysis of variance followed by a Bonferroni multiple comparison test was performed. All other statistical differences were determined using the two-tailed Student s t test. All data are presented as mean and SD. Statistical significance was defined as a P value of <0.05 for all tests. ACKNOWLEDGMENTS. We thank K. Watada, J. Takezawa, and R. Yamagata for excellent technical assistance; Dr. P. Charnay (INSERM) for providing Egr2 floxed mice; and Dr. J. Miyazaki (Osaka University Medical School) for kindly providing the pcaggs vector. This work was supported by Grant-in-Aid for Japan Society for the Promotion of Science Fellows Grant and Scientific Research Grant 16K PNAS PLUS IMMUNOLOGY AND INFLAMMATION 1. Zhang Y, Garcia-Ibanez L, Toellner KM (2016) Regulation of germinal center B-cell differentiation. Immunol Rev 270(1): Pisetsky DS (2016) Anti-DNA antibodies: Quintessential biomarkers of SLE. Nat Rev Rheumatol 12(2): Stohl W (2013) Future prospects in biologic therapy for systemic lupus erythematosus. Nat Rev Rheumatol 9(12): Abbas AK, et al. (2013) Regulatory T cells: Recommendations to simplify the nomenclature. Nat Immunol 14(4): Li MO, Rudensky AY (2016) T cell receptor signalling in the control of regulatory T cell differentiation and function. Nat Rev Immunol 16(4): Sakaguchi S, Wing K, Onishi Y, Prieto-Martin P, Yamaguchi T (2009) Regulatory T cells: How do they suppress immune responses? Int Immunol 21(10): Okamura T, et al. (2009) CD4+CD25-LAG3+ regulatory T cells controlled by the transcription factor Egr-2. Proc Natl Acad Sci USA 106(33): Okamura T, et al. (2015) TGF-β3-expressing CD4+CD25( )LAG3+ regulatory T cells control humoral immune responses. Nat Commun 6: Okamura T, et al. (2015) Role of TGF-β3 in the regulation of immune responses. Clin Exp Rheumatol 33(4, Suppl 92):S63 S Myouzen K, et al. (2010) Regulatory polymorphisms in EGR2 are associated with susceptibility to systemic lupus erythematosus. Hum Mol Genet 19(11): Zhu B, et al. (2008) Early growth response gene 2 (Egr-2) controls the self-tolerance of T cells and prevents the development of lupuslike autoimmune disease. J Exp Med 205(10): Topilko P, et al. (1994) Krox-20 controls myelination in the peripheral nervous system. Nature 371(6500): Schneider-Maunoury S, et al. (1993) Disruption of Krox-20 results in alteration of rhombomeres 3 and 5 in the developing hindbrain. Cell 75(6): Harris JE, et al. (2004) Early growth response gene-2, a zinc-finger transcription factor, is required for full induction of clonal anergy in CD4+ T cells. J Immunol 173(12): Safford M, et al. (2005) Egr-2 and Egr-3 are negative regulators of T cell activation. Nat Immunol 6(5): Anderson PO, et al. (2006) Persistent antigenic stimulation alters the transcription program in T cells, resulting in antigen-specific tolerance. Eur J Immunol 36(6): Tourtellotte WG, Milbrandt J (1998) Sensory ataxia and muscle spindle agenesis in mice lacking the transcription factor Egr3. Nat Genet 20(1): Li S, et al. (2012) The transcription factors Egr2 and Egr3 are essential for the control of inflammation and antigen-induced proliferation of B and T cells. Immunity 37(4): Zheng Y, Zha Y, Driessens G, Locke F, Gajewski TF (2012) Transcriptional regulator early growth response gene 2 (Egr2) is required for T cell anergy in vitro and in vivo. J Exp Med 209(12): Sumitomo S, et al. (2013) Transcription factor early growth response 3 is associated with the TGF-β1 expression and the regulatory activity of CD4-positive T cells in vivo. J Immunol 191(5): Li S, et al. (2011) Early growth response gene-2 (Egr-2) regulates the development of B and T cells. PLoS One 6(4):e Shi W, et al. (2015) Transcriptional profiling of mouse B cell terminal differentiation defines a signature for antibody-secreting plasma cells. Nat Immunol 16(6): Gianchecchi E, Delfino DV, Fierabracci A (2013) Recent insights into the role of the PD- 1/PD-L1 pathway in immunological tolerance and autoimmunity. Autoimmun Rev 12(11): Tsokos GC (2011) Systemic lupus erythematosus. N Engl J Med 365(22): Collins S, et al. (2008) Opposing regulation of T cell function by Egr-1/NAB2 and Egr-2/ Egr-3. Eur J Immunol 38(2): Morita et al. PNAS Early Edition 9of10