Functions of Smad Transcription Factors in TGF-Β1–Induced Selectin

Functions of Smad Transcription Factors in TGF- β1−Induced Selectin Ligand Expression on Murine CD4 Th Cells

This information is current as Mark E. Ebel and Geoffrey S. Kansas of September 24, 2021. J Immunol 2016; 197:2627-2634; Prepublished online 19 August 2016; doi: 10.4049/jimmunol.1600723 http://www.jimmunol.org/content/197/7/2627 Downloaded from

References This article cites 55 articles, 24 of which you can access for free at: http://www.jimmunol.org/content/197/7/2627.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 24, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2016 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Functions of Smad Transcription Factors in TGF-b1–Induced Selectin Ligand Expression on Murine CD4 Th Cells

Mark E. Ebel and Geoffrey S. Kansas

Selectins are carbohydrate-binding adhesion molecules that control leukocyte trafﬁc. Induction of selectin ligands on T cells is controlled primarily by cytokines, including TGF-b1, and requires p38a MAPK, but transcriptional mechanisms that underlie cytokine-driven selectin ligand expression are poorly understood. In this study, we show, using mice with conditional deletions of the TGF-b1–responsive transcription factors Smad2, Smad3, or Smad4, that induction of selectin ligands on CD4 cells in response to TGF-b1 requires Smad4 plus either Smad2 or Smad3. Analysis of CD4 cells from mice with only one functional Smad4 allele revealed a sharp gene dosage effect, suggesting the existence of a threshold of TGF-b1 signal strength required for selectin ligand induction. Both Smad4 plus either Smad2 or Smad3 were selectively required for induction of Fut7 and Gcnt1, glycosyltrans-

ferases critical for selectin ligand biosynthesis, but they were not required for St3gal4 or St3gal6 induction. Smad4 plus either Downloaded from Smad2 or Smad3 were also required for induction of Runx transcription factors by TGF-b1. Enforced expression of Runx2, but not Runx1 or Runx3, in Smad2/Smad3 doubly deﬁcient CD4 cells restored selectin ligand expression to wild-type levels. In contrast, enforced expression of Runx1, Runx2, or Runx3 failed to restore differentiation of TGF-b1–dependent Th cell lineages, including Th17, Th9, and induced regulatory T cells. These results show that Smads are directly required for Th cell differentiation independent of Runx induction but only indirectly required via Runx2 for TGF-b1–induced selectin ligand induction on

murine CD4 T cells. The Journal of Immunology, 2016, 197: 2627–2634. http://www.jimmunol.org/

ransforming growth factor b1 is a highly pleiotropic cy- Selectins are carbohydrate-binding adhesion molecules critically tokine with diverse functions in the immune system. Within involved in regulation of leukocyte traffic (15, 16). Glycans that T the T cell compartment, TGF-b1 is critical for the generation function as selectin ligands are formed by the concerted action of of three known types of Th cells: regulatory T cells (Tregs) (1–3), Th17 multiple glycosyltransferases (GTases) and are constitutively cells (4, 5), and Th9 cells (6), which use otherwise distinct cytokine expressed on myeloid cells but absent from naive T cells. Selectin signaling pathways. Defects in TGF-b1 signaling pathways, including ligand expression on activated T cells is controlled largely by cy- either chain of the receptor, associated kinases such as TAK1, or tokines (17–19). We have recently identified a biologically diverse downstream transcription factors such as Smads, lead to defects in set of Th cell–promoting and inflammatory cytokines, including by guest on September 24, 2021 Treg development and consequent fatal autoimmune disease (7–10). TGF-b1, that strongly induce expression of selectin ligands, and Abundant research by many groups has yielded a detailed each of which requires p38a MAPK for induction of selectin li- mechanistic understanding of signaling by TGF superfamily ligands (20). However, a potential role for Smad transcription factors gands, including TGF-b, bone morphogenetic proteins, and acti- in TGF-b1–induced selectin ligand induction has never been vins. Upon binding of TGF-b1 to its heterodimeric receptor assessed, despite the clear importance of the TGF-b1/Smad path- complex, serine phosphorylation of the type I receptor chain GS way to most biological responses to TGF-b1, as well as to regulation domain by the kinase domain of the type II chain leads to binding of numerous aspects of immunity, including Th cell differentiation. and serine phosphorylation of Smad2 and Smad3 by the kinase In this study, we have analyzed the requirements for Smad2, Smad3, domain of the type I chain. This is followed by association of and Smad4 in TGF-b1–induced selectin ligands on murine CD4 these R-Smads with Smad4, translocation of this complex to the T cells and compared them to the functions of these transcription nucleus, association of the activated heteromeric Smad complexes factors in TGF-b1–dependent Th cell differentiation. with one or more coactivators or corepressors, and activation or repression of target genes (11, 12). Additionally, binding of TGF Materials and Methods superfamily ligands to their receptors can trigger non-Smad Mice pathways, including multiple MAPKs (13, 14). Mice on a C57BL/6 background with loxP-flanked (floxed) alleles of Smad2, Smad3, or Smad4 were supplied by Dr. Erwin Bottinger (Mt. Sinai School of Medicine, New York, NY), Dr. Jon Graff (University of Texas Department of Microbiology–Immunology, Feinberg School of Medicine, North- Southwestern Medical Center, Dallas, TX), and Dr. Chen Dong (National western University, Chicago, IL 60611 Institutes of Health, Bethesda, MD), respectively. CD4-cre mice were Received for publication April 25, 2016. Accepted for publication July 27, 2016. obtained from Taconic. Intercrosses were carried out such that, except This work was supported in part by the Chicago Biomedical Consortium with support where noted, all mice were homozygous for the indicated floxed Smad 2 from Searle Funds at The Chicago Community Trust (to G.S.K.). genes and either hemizygous or negative for cre. Cre littermates served as Address correspondence and reprint requests to Dr. Geoffrey S. Kansas, Department controls for all experiments and are referred to as wild-type (WT). Owing of Microbiology–Immunology, Feinberg School of Medicine, Northwestern Univer- to the ultimately lethal pathology associated with deletion of both Smad2 + + + sity, 300 E. Superior Street, Room Ward 6-313, Chicago, IL 60611. E-mail address: and Smad3 in T cells (10, 21), we bred cre Smad2 /fl Smad3 /fl mice to 2 2 [email protected] cre Smad2-fl/fl Smad3-fl/fl mice to generate cre+ and cre Smad2-fl/fl + Abbreviations used in this article: CBF, core-binding factor; GTase, glycosyltransfer- Smad3-fl/fl mice. Cre Smad2-fl/fl Smad3-fl/fl mice were always used ase; iTreg, induced Treg; Treg, regulatory T cell; RV, retrovirus; WT, wild-type. before 6 wk of age, a time point at which in our hands mice exhibited no signs of disease. No such disease was ever noted in mice with Smad4 Copyright Ó 2016 by The American Association of Immunologists, Inc. 0022-1767/16/$30.00 deleted in T cells, consistent with previous reports (22, 23). Mice of both www.jimmunol.org/cgi/doi/10.4049/jimmunol.1600723 2628 Smad TRANSCRIPTION FACTORS IN SELECTIN LIGAND EXPRESSION sexes were used throughout these experiments, and no differences between TGF-b1. Moreover, at a 50% lower dose of TGF-b1(2.5ng/ml), sexes were noted. CD4 cells with only a single functional allele of Smad4 showed T cell culture and analysis virtually no induction of selectin ligands (6.2 6 0.4 versus 5.2 6 0.6 for no TGF-b1; Fig. 2B, 2F), whereas CD4 cells with only a Isolation of splenic CD4 T cells, activation by plate-bound anti-CD3/C28, single functional allele each of Smad2 and Smad3 showed only a each at 1 mg/ml for all experiments, culture in IL-2 with and without 5 ng/ml TGF-b1, analysis of E- and P-selectin ligand expression by flow cytometry slight decrement compared with the higher dose of TGF-b1 using E- and P-selectin/IgM chimeras, retroviral transduction of activated (Fig. 2D, 2H). Collectively, these data suggest that Smad4 associ- CD4 T cells, isolation of RNA, and quantitative RT-PCR were all exactly ation with TGF-b1–activated Smad2 and/or Smad3 represents a as recently described (20). Intracellular staining for IL-17, IL-9, and Foxp3, limiting event in TGF-b1–induced selectin ligand expression and using PE-conjugated (IL-9 and Foxp3) or PerCP-conjugated (IL-17) mAbs from eBioscience, was performed by standard methods following culture could explain why a subset of cells fails to express selectin ligands. under Th17 (TGF-b1 plus IL-6), Th9 (TGF-b1 plus IL-4 plus IL-2), or To understand the molecular basis for the loss of TGF-b1–induced induced Treg (iTreg; TGF-b1 + IL-2) conditions, as described (20). For selectin ligand induction in Smad4- and Smad2/3-deficient CD4 Th17 and Th9, cells were restimulated separately with plate-bound anti-CD3 cells, we examined levels of expression of genes encoding four or PMA/ionomycin prior to addition of GolgiStop, with substantially similar GTases known to be essential for selectin ligand biosynthesis: results. Retroviral expression plasmids and cDNA were obtained from var- ious investigators, except for core-binding factor (CBF) b,whichwasob- Fut7, Gcnt1, St3gal4, and St3gal6 (24–30). Strong induction of all tained from Origene. four of these genes was observed in WT cells in response to TGF- b1, with only a very modest increase due to T cell activation Statistical analysis without TGF-b1 (Fig. 3). The induction of Fut7 and Gcnt1 was

Comparison between groups was by the multiple t test within Prism 6. almost completely lost in the absence of either Smad4 (Fig. 3A) Downloaded from or both Smad2 and Smad3 (Fig. 3B). In contrast, induction of Results St3Gal4 or St3gal6 was not affected by Smad deletion. Thus, We first determined whether Smad2, Smad3, or Smad4 was required Smad transcription factors control induction of selectin ligands in for induction of selectin ligands by TGF-b1. CD4 cells were isolated response to TGF-b1 via direct or indirect control of Fut7 and from mice with conditional deletion of each of these transcription Gcnt1 expression. 2 factors in T cells, along with WT (cre ) control littermates, and Elsewhere, we show that Runx transcription factors are involved http://www.jimmunol.org/ activated and cultured in the presence or absence of 5 ng/ml in TGF-b1–induced selectin ligand induction via induction of TGF-b1, which we found in preliminary experiments to yield the Fut7 and Gcnt1 (M.E. Ebel and G.S. Kansas, manuscript in maximal induction of selectin ligands (data not shown). The results preparation). Induction of Runx transcription factors is observed showed that deletion of Smad4 completely blocked induction of in many different cell types in response to TGF superfamily li- selectin ligands in response to TGF-b1(Fig.1A,1E),whereas gands and is often dependent on Smad transcription factors (31). defects observed with either the Smad2- or Smad3-deficient CD4 We therefore examined expression of Runx1, Runx2, Runx3, and cells were very modest (Fig. 1B, 1C, 1F, 1G). These results sug- their obligate partner CBFb in WT CD4 cells and in CD4 cells gested that, as is the case for many responses to TGF-b1, either deleted of either Smad4 or both Smad2 and Smad3. We found Smad2 or Smad3 was required, but not both. This was confirmed by significant induction of all three Runx genes and CBFb by TGF- by guest on September 24, 2021 analysis of CD4 cells deficient in both Smad2 and Smad3 (Fig. 1D, b1 (Fig. 4). As with the GTases, very little induction was seen in 1H), which showed a complete loss of selectin ligand induction, as the absence of TGF-b1. This induction was severely compromised was seen with the Smad4-deficient CD4 cells. In all cases, defects in in the absence of either Smad4 (Fig. 4A) or both Smad2 and expression of selectin ligands by Smad-deficient CD4 cells was not Smad3 (Fig. 4B), closely resembling the pattern of inhibition of due to alterations in kinetics. Thus, induction of selectin ligands on Fut7 and Gcnt1 expression. These results show that all three Runx CD4 cells requires Smad4 plus either Smad2 or Smad3. genes and CBFb are induced by TGF-b1, and that Smad tran- In all cases, expression of selectin ligands on activated T cells scription factor signaling is essential for induction of Runx gene exhibited a clear bimodal distribution, with distinct positive and expression in CD4 T cells in response to TGF-b1. In combination negative subsets (Fig. 1I) (20), and this is also observed with other with the finding that inhibition of Runx activity also blocks in- cytokines (20). The basis for this pattern of expression has never duction of Fut7, Gcnt1, and selectin ligand expression (M.E. Ebel been determined, but could potentially be explained by a threshold and G.S. Kansas, manuscript in preparation), these results sug- of TGF-b1 signal strength required for induction of selectin li- gested that induction of Runx may be central to the function of gands. To approach this question, we analyzed selectin ligand Smad transcription factors in selectin ligand induction. expression on CD4 cells that had only one functional allele of To test this, and to identify which Runx proteins were involved in Smad4, or only one functional allele each of Smad2 and Smad3. selectin ligand induction in response to TGF-b1, we performed CD4 cells with only a single functional allele of Smad4 and reconstitution experiments with Smad2/Smad3 doubly deficient cultured with the optimal dose of TGF-b1 showed a significant CD4 cells, using retroviruses (RV) expressing cDNA encoding decrease compared with WT in expression of selectin ligands each Runx protein, plus a second RV encoding CBFb, which we (mean 42.5 6 2.2 versus 27.6 6 1.9% for E-selectin ligands and found in preliminary experiments to be essential for transcrip- 48.6 6 2 versus 32.2 6 2.9% for P-selectin ligands at day 10, both tional activity of transduced Runx genes (data not shown). Neg- p , 0.01) (Fig. 2A, 2B, 2E, 2F). CD4 cells with only a single ative controls consisted of dual RV containing no cDNA. Cells functional allele each of Smad2 and Smad3 and cultured with the were activated in the absence of TGF-b1, doubly transduced with optimal dose of TGF-b1 showed a similar decrease (53 6 3.6 RV, and then cultured either with or without TGF-b1 and analyzed versus 33 6 1% for E-selectin ligands and 53.8 6 2.8 versus 35.4 6 for selectin ligand expression on cells expressing both RV mark- 2.7% for P-selectin ligands at day 10, both p , 0.01) (Fig. 2C, 2D, ers. We found that enforced expression of Runx2 plus CBFb al- 2G, 2H). Notably, deletion of one Smad2 and one Smad3 allele most completely restored expression of both E- and P-selectin had a greater effect than deletion of two Smad2 alleles or two ligands (Fig. 5A, 5B). In contrast, enforced expression of Runx3 Smad3 alleles (compare with Fig. 1), suggesting that both Smad2 plus CBFb had no ability to restore selectin ligand expression in and Smad3 participate in gene regulation in this system. These Smad-deficient cells and actually modestly inhibited expression of results identify a clear effect of gene dosage on this response to E- and P-selectin ligands in WT cells. Enforced expression of The Journal of Immunology 2629 Downloaded from http://www.jimmunol.org/

FIGURE 1. Effects of Smad2, Smad3, and Smad4 deletion on TGF-b1–induced selectin ligand expression. CD4 T cells from cre2 (WT) and cre+ mice with loxP-flanked Smad4, Smad2, Smad3, or both Smad2 and Smad3 were isolated and cultured with IL-2 plus either 5 ng/ml TGF-b1 or no TGF-b1 and analyzed by flow cytometry with selectin chimeras every 2 d. Results are gated on all live cells and are means 6 SD for three independent experiments. by guest on September 24, 2021 Genotypes are indicated at the top. Open symbols, no TGF-b1; filled symbols, + TGF-b1; circles are cre2 (WT); squares are cre+ (cKO). (A–D) E-selectin ligands; (E–H) P-selectin ligands. (I) Representative dot plots showing forward scatter (x-axis) versus staining for E-selectin ligands (top row) or P-selectin ligands (bottom row) (y-axis), showing clear bimodality of selectin ligand expression. *p , 0.01. cKO, conditional knockout.

Runx1 plus CBFb partially restored E-selectin ligand expression of TGF-b1 (Fig. 6D–F), confirming the requirement for TGF-b1 but had no effect on P-selectin ligand expression. Notably, no signaling. Hence, differentiation of Th17, Th9, or iTregs directly selectin ligand expression was induced in the absence of TGF-b1, requires expression and function of Smad2/Smad3 transcription regardless of Runx expression, in either WT or Smad-deficient factors. These results further distinguish the requirements for Th cells (Fig. 5C, 5D). Taken together, these results show that cell differentiation from those for selectin ligand expression on Runx2 (plus CBFb) is sufficient in the absence of Smad2/Smad3 these classes of Th cells. expression to confer TGF-b1–induced selectin ligand expression. Finally, to investigate further the relationship between induction of Both Smad and Runx transcription factors are essential for selectin ligands by TGF-b1 and Th cell differentiation induced by differentiation and function of Th17 and iTreg (32–41), and at least TGF-b1, we analyzed the relationship between selectin ligand ex- Smads are required for Th9 differentiation (6, 42, 43) (a repression and expression of either IL-17, IL-9, or Foxp3 in Th17, quirement for Runx in Th9 differentiation has not been tested to Th9, or iTreg cultures, respectively. As previously found for Th1 our knowledge), but whether Runx expression in the absence of and Th2 cells (44–46), we found no correlation between selectin TGF-b1–triggered Smad signaling is sufficient for successful ligand expression and any of these Th cell lineage markers (Fig. 7). differentiation or, alternatively, whether Smad transcription factors are essential apart from their function in Runx induction has Discussion not been addressed. We therefore performed experiments analo- Signaling via TGF-b1 is of critical importance for numerous aspects gous to those above, doubly transducing activated CD4 T cells of T cell function, especially the induction of several classes of Th with RV encoding Runx1, Runx2, or Runx3, plus RV encoding cells, including Tregs, Th17 cells, and Th9 cells. Loss-of-function CBFb, or dual control RV, followed by culture under Th17, Th9, mutations in any of several mediators of TGF-b1 signaling, in- or iTreg conditions, and we analyzed the cells to determine cluding the receptor for TGF-b1 or the receptor-proximal whether expression of IL-17, IL-9, or Foxp3, respectively, was MAPKKK TAK1, leads to a failure of Treg development and the restored under the appropriate culture conditions. In clear contrast death of mice due to intense multifocal inflammatory disease within to the results with selectin ligand expression, no induction of IL- ∼6–8 wk (7, 10). Loss of both Smad2 and Smad3 in conventional 17 (Fig. 6A), IL-9 (Fig. 6B), or Foxp3 (Fig. 6C) was seen over T cells leads them to be refractory to TGF-b1–mediated inhibition, background levels in Smad2/Smad3-deficient cells. Additionally, leading to similar pathology (10, 21). In addition to a role in Th cell no induction of these markers was seen in WT cells in the absence differentiation, TGF-b1 is a potent inducer of E- and P-selectin 2630 Smad TRANSCRIPTION FACTORS IN SELECTIN LIGAND EXPRESSION

FIGURE 2. Effects of Smad2, Smad3, and Smad4 gene dosage on TGF-b12induced selectin ligand expression. CD4 T cells from cre2 (WT) and cre+ mice with only a single loxP-flanked allele of either Smad4, or both Smad2 and Smad3, as indicated at the top, were isolated and cultured with IL-2 plus either 5 ng/ml TGF-b1(d), 2.5 ng/ml TGF-b1(N), or no TGF-b1(n) and analyzed by flow cytometry with selectin chimeras every 2 d. Results are gated on Downloaded from all live cells and are mean 6 SD of three independent experiments. (A–D) E-selectin ligands; (E–H) P-selectin ligands. *p , 0.01. ligands in CD4 T cells, both in humans and mice (19, 20). We have duction of selectin ligands, whereas deletion of either Smad4 or recently shown that this induction of selectin ligands requires p38a both Smad2 and Smad3 abrogated the induction of selectin li- MAPK (20), one of several non-Smad pathways activated by TGF- gands in response to TGF-b1. Further analysis revealed that Smad b1 (13, 14). The ability of Tregs, Th17 cells, and Th9 cells to gain deletion blocked induction of both Fut7, which is essential for http://www.jimmunol.org/ access to inflammatory sites, particularly the skin and gut, likely expression of both E- and P-selectin ligands (24–26), and Gcnt1, depends on this ability of TGF-b1 to induce selectin ligands. which is critical for expression of P-selectin ligands (27, 28, 47), However, it has previously been unclear whether Smad transcription on CD4 T cells. Expression of two other enzymes that are also factors were involved in selectin ligand induction in response to involved in selectin ligand biosynthesis, St3gal3 and St3gal6, was TGF-b1, and if so, via which mechanisms. Additionally, it has been also induced by TGF-b1 but was not blocked by Smad deletion. unclear how the requirements for these TGF-b1–activated tran- Elsewhere, we show that Runx transcription factors are involved scription factors in selectin ligand induction compare with those for in selectin ligand induction in response to TGF-b1, and that differentiation of TGF-b1–dependent Th cells. In this study, we inhibition of Runx activity also blocks induction of Fut7 and have analyzed CD4 cells from mice with conditional deletion of Gcnt1 (M.E. Ebel and G.S. Kansas, manuscript in preparation). by guest on September 24, 2021 Smad4, Smad2, Smad3, or both Smad2 and Smad3 to identify these We therefore determined whether induction of Runx transcription requirements and the mechanisms underlying them. factors requires Smads. We found that TGF-b1 induced expres- We found that induction of selectin ligands by TGF-b1 requires sion of all three Runx genes as well as CBFb, a non-DNA–binding both Smad4 plus either Smad2 or Smad3, the canonical pattern of obligate partner for Runx proteins that is required for Runx Smad requirements seen for many TGF-b1–responsive genes. transcriptional activity (48, 49). Induction of all three Runx genes Deletion of either Smad2 or Smad3 alone had little effect on in- and CBFb was blocked by deletion of either Smad4 or both

FIGURE 3. Smad requirements for induction of GTase genes involved in TGF-b1–induced selectin ligand biosynthesis. Cells from cultures as in Fig. 1 were isolated on day 10, and total mRNA was prepared and used for quantitative RT-PCR with SYBR Green using primers for Fut7, Gcnt1, St3gal4, St3gal6,orHprt. Results are presented as the ratio of each gene to the housekeeping gene Hprt, using mice deﬁcient in either (A) Smad4 or (B) Smad2/ Smad3. Results represent mean 6 SD for three independent experiments. *p , 0.01 and different from all other groups in the same panel. **p , 0.01 for either WT or knockout (KO) naive CD4 T cells versus either WT or KO CD4 T cells cultured without TGF-b1. #p , 0.01 versus naive or versus IL-2 only but not different from each other. cKO, conditional knockout. The Journal of Immunology 2631

FIGURE 4. Smad requirements for induction of Runx genes in response to TGF-b1. Identical analysis as in Fig. 3, except that here, speciﬁc primers were for Runx1, Runx2, Runx3, and Cbfb, again using mice deﬁcient in either (A) Smad4 or (B) Smad2/Smad3. Results are mean 6 SD for three independent , , experiments. *p 0.01 and different from all other groups in the same panel. **p 0.01 for either WT or knockout (KO) naive CD4 T cells versus either Downloaded from WT or KO CD4 T cells cultured without TGF-b1. cKO, conditional knockout.

Smad2 and Smad3, paralleling the pattern of inhibition of Fut7, required concomitant TGF-b signaling, as no selectin ligand in- Gcnt1, and selectin ligands by Smad deletion. Taken together, duction was observed without TGF-b1. This suggests that one or these ﬁndings suggested that Smad transcription factors control more additional TGF-b1–triggered signaling pathways are re-

expression of selectin ligands via Runx-mediated induction of quired for selectin ligand induction. Our recent finding that de- http://www.jimmunol.org/ Fut7 and Gcnt1. We therefore reconstituted Smad2/Smad3 doubly letion of p38a MAPK blocks TGF-b1–induced expression of Fut7 deficient cells with each Runx protein along with CBFb and de- and Gcnt1 and selectin ligands without affecting levels of ST3gal4 termined whether this rescued selectin ligand expression. We or St3gal6 (20), closely paralleling the effects of deletion of either found that enforced expression of Runx2 (plus CBFb), but not Smad4 or Smad2/Smad3, makes p38a MAPK a good candidate either Runx1 (plus CBFb) or Runx3 (plus CBFb) was sufficient to for one such pathway. Consistent with this idea, Runx2 is a well- restore selectin ligand expression in Smad2/Smad3-deficient cells. defined target of p38 in mesenchymal cells, where phosphoryla- Thus, enforced expression of Runx2/CBFb, in the context of TGF- tion of Runx2 by p38 is essential for bone formation (50, 51). It b1 signaling, is sufficient for induction of selectin ligands in the will be important to determine whether this pathway is also op- absence of Smad2/Smad3. These findings show that, surprisingly, erative in T cells. Additionally, both Smad2 and Smad3 are well- by guest on September 24, 2021 Smad transcription factors are not directly required for induction known targets of p38 (and ERK MAPK), and phosphorylation by of Fut7 or Gcnt1 in response to TGF-b1, but rather are indirectly these kinases can greatly increase or modulate the transcriptional involved via induction of Runx2 and CBFb. potential of Smad2 and Smad3. It will therefore also be important It is notable that reconstitution of selectin ligands in Smad2/ to determine whether phosphorylation of Smad2 and/or Smad3 by Smad3-deficient cells via enforced expression of Runx2/CBFb p38a MAPK underlies the parallel requirements for Smads and

FIGURE 5. Enforced expression of Runx2/CBFb restores TGF-b1–induced selectin ligand expression in Smad2/Smad3-deficient CD4 T cells. WT or Smad2/ Smad3-deficient CD4 cells were activated with plate-bound anti-CD3/CD28 for ∼40 h in the absence of any cytokines, washed, and cotransduced with RV expressing Runx1, Runx2, or Runx3 plus a second RV expressing CBFb. Cells were then cultured for an additional 8 d with IL-2 either with (A and B,top)or without (C and D, bottom) 5 ng/ml TGF-b1 and then analyzed for selectin ligand expression on live cells expressing both RV markers. Results presented are from day 10 and represent mean 6 SD from three independent experiments. #p , 0.05, *p , 0.01 for the indicated comparisons. cKO, conditional knockout. 2632 Smad TRANSCRIPTION FACTORS IN SELECTIN LIGAND EXPRESSION Downloaded from FIGURE 6. Smad2 and/or Smad3 are directly required for differentiation of TGF-b1–dependent Th cell lineages. WT or Smad2/Smad3-deficient CD4 cells were activated with plate-bound anti-CD3/CD28 for ∼40 h in the absence of any cytokines, washed, and cotransduced with RV expressing Runx1, Runx2, or Runx3 plus a second RV expressing CBFb and then cultured for an additional 8 d with (A) TGF-b1 plus IL-6, (B) TGF-b1 plus IL-4 plus IL-2, or (C) TGF-b1 plus IL-2 to induce Th17, Th9, or iTregs, respectively, or with no polarizing cytokines (D–F). On day 10, cells were harvested and analyzed by intracellular FACS for (A and D) IL-17A, (B and E) IL-9, or (C and F) Foxp3. Cells being analyzed for IL-17A or IL-9 were restimulated with plate-bound anti-CD3 for 4 h prior to analysis; substantially identical results were obtained with PMA/ionomycin (data not shown). Results presented are from day10 and are mean 6 SD for three independent experiments. *p , 0.01 for pairwise comparison for a single RV combination. cKO, conditional knockout. http://www.jimmunol.org/ p38a in induction of Runx2 and selectin ligands and/or whether in selectin ligand expression (52) and is likely required for ex- p38a MAPK plays a Smad-independent function. pression of numerous other genes as well (53), likely via changes Although signals emanating from the TCR and the TbR both in chromatin conformation and accessibility. T cell activation may induce expression of selectin ligands, the levels of selectin ligands also be required to induce full cellular responsiveness to TGF-b1, seen in the absence of TGF-b1 are substantially lower than those despite constitutive expression of receptors for TGF-b1on in the presence of the cytokine, and this is reflected in the levels of CD4 cells, because activation of p38 MAPK, as assessed by the mRNA encoding GTases required for selectin ligand biosyn- phosphorylation, is not observed on naive cells exposed to TGF- thesis as well as in the degree of induction of Runx genes. Indeed, b1 (20). by guest on September 24, 2021 both of these sets of genes are expressed at only slightly higher Previous studies have uncovered a critical role for Runx3 in Th1 levels in activated T cells (without TGF-b1) than in naive T cells, development and repression of Th2 development, as well as a role but they are expressed at substantially higher levels following for Runx1 in both iTregs and Th17 development, but a role for culture with TGF-b1. Nonetheless, signals from the TCR make Runx2 in T cells has not previously been described. Both Runx1 essential, nonredundant, independent contributions to induction of and Runx2 were strongly induced by TGF-b1, with Runx3 in- selectin ligand expression. Cell cycling initiated and driven via the duction being somewhat lower. The results presented in the pre- TCR is thought to be essential for expression of GTases involved sent study suggest distinct functions for Runx1 and Runx2 in Th

FIGURE 7. Lack of correlation between expression of selectin ligands and IL-17, IL-9, or Foxp3. WT CD4 cells were cultured under Th17, Th9, or iTreg conditions as in Fig. 6 for 8 d, and cells were harvested and analyzed by intracellular FACS for (A and D) IL-17A, (B and E) IL-9, or (C and F) Foxp3. Cells being analyzed for IL-17A or IL-9 were restimulated with plate-bound anti-CD3 for 4 h prior to analysis; substantially identical results were obtained with PMA/ionomycin (data not shown). One representative experiment out of three is shown. The Journal of Immunology 2633 cells: Runx1 functions in the differentiation of Tregs via induction important participants in induction of selectin ligands and, in and stabilization of Foxp3 and in the induction of Th17 cells via combination with the results of Smad4 haploinsufficiency, sug- induction of retinoic acid–related orphan receptor gt, whereas gests that heteromeric Smad2/Smad3/Smad4 complexes may be Runx2 functions in the induction of selectin ligands for these two the functionally relevant transcriptional regulator for Fut7, Gcnt1, Th cell lineages and presumably Th9 cells as well. To our or both. knowledge, this is the first function of Runx2 to be identified in In summary, we have defined a pathway from TGF-b1re- T cells. ceptors → Smad2/Smad3 and Smad4 → Runx2 and CBFb → Previous work has uncovered differential requirements for Smad Fut7 and Gcnt1, which operates to regulate expression of transcription factors in differentiation of TGF-b1–dependent Th selectin ligands on activated murine CD4 cells. This pathway cell lineages. Smad2 is a positive regulator of Th9 and Th17 likely accounts for selectin ligand expression on Th17, Th9, and differentiation, whereas Smad3 is a negative regulator of Th17 iTregs and may represent an attractive target for intervention in development (32–34, 54). Smad3 is also a positive regulator of disease states characterized by disorders in inflammatory T cells Th9 development, but mechanisms by which Smad2 and Smad3 of these classes. induce Th9 cells appear to be quite distinct (42, 55). Deletion of both Smad2 and Smad3 strongly blocks iTreg development, but Acknowledgments not tTreg development (10). Smad4 appears to be necessary for We are grateful to Dr. Erwin Bottinger (Mt. Sinai School of Medicine, New iTreg and Th9 differentiation, although Smad4 deletion leads to York, NY), Dr. Jon Graff (University of Texas Southwestern Medical Center, only an ∼50% reduction of Foxp3+ or IL-9+ cells, respectively, Dallas, TX), and Dr. Chen Dong (National Institutes of Health, Bethesda, and Smad4 deletion does not affect Th17 differentiation (22, 42). MD), respectively, for mice on a C57BL/6 background with loxP-flanked Downloaded from In contrast, deletion of Smad4 completely blocked TGF-b1– (floxed) alleles of Smad2, Smad3 or Smad4. We also thank Dr. Takeshi Egawa dependent induction of selectin ligands. Conversely, deletion of (Washington University, St. Louis, MO) for retroviral plasmids encoding Smad2 blocks Th17 development by ∼60% (32–34) but has very Runx1 and Runx3, and Dr. Gerard Karsenty for Runx2 cDNA. little effect on selectin ligand induction. Similarly, deletion of Smad3 enhances Th17 development but has little effect on selectin Disclosures

ligand induction. These findings collectively illustrate that Smad The authors have no financial conflicts of interest. http://www.jimmunol.org/ requirements for selectin ligand induction are quantitatively and qualitatively distinct from those for differentiation of TGF-b1– References dependent Th cell lineages. 1. Marie, J. C., J. J. Letterio, M. Gavin, and A. Y. Rudensky. 2005. TGF-b1 Expression of selectin ligands on activated T cells exhibits a maintains suppressor function and Foxp3 expression in CD4+CD25+ regulatory bimodal pattern, with clear positive and negative subsets, and this is T cells. J. Exp. Med. 201: 1061–1067. 2. Chen, W., W. Jin, N. Hardegen, K. J. Lei, L. Li, N. Marinos, G. McGrady, and independent of the nature of the stimulus (Fig. 1) (20). This het- S. M. Wahl. 2003. Conversion of peripheral CD4+CD252 naive T cells to CD4+ erogeneity may be important for diversification of the immune CD25+ regulatory T cells by TGF-b induction of transcription factor Foxp3. response. This bimodal expression pattern is not a consequence of J. Exp. Med. 198: 1875–1886. 3. Liu, Y., P. Zhang, J. Li, A. B. Kulkarni, S. Perruche, and W. Chen. 2008. A differences in early T cell activation, as T cells analyzed at early critical function for TGF-b signaling in the development of natural CD4+CD25+ by guest on September 24, 2021 time points uniformly express both CD69 and CD25 (data not Foxp3+ regulatory T cells. Nat. Immunol. 9: 632–640. 4. Harrington, L. E., R. D. Hatton, P. R. Mangan, H. Turner, T. L. Murphy, shown). Additionally, bimodality/heterogeneity of expression is K. M. Murphy, and C. T. Weaver. 2005. Interleukin 17-producing CD4+ effector observed for effector cytokines and Foxp3 as well, and no cor- T cells develop via a lineage distinct from the T helper type 1 and 2 lineages. relation between selectin ligand expression and expression of IL- Nat. Immunol. 6: 1123–1132. 5. Mangan, P. R., L. E. Harrington, D. B. O’Quinn, W. S. Helms, D. C. Bullard, 17, IL-9, or Foxp3 was observed, consistent with previous studies C. O. Elson, R. D. Hatton, S. M. Wahl, T. R. Schoeb, and C. T. Weaver. 2006. of Th1 and Th2 cells (44–46). Although there are likely to be Transforming growth factor-b induces development of the TH17 lineage. Nature multiple mechanisms underlying this heterogeneity, one possible 441: 231–234. 6. Dardalhon, V., A. Awasthi, H. Kwon, G. Galileos, W. Gao, R. A. Sobel, mechanism relates to the number of activated heteromeric Smad M. Mitsdoerffer, T. B. Strom, W. Elyaman, I. C. Ho, et al. 2008. IL-4 inhibits complexes formed in response to TGF-b1. Data presented in the TGF-b-induced Foxp3+ T cells and, together with TGF-b, generates IL-9+ IL- 10+ Foxp32 effector T cells. Nat. Immunol. 9: 1347–1355. present study show that when one allele of Smad4 was deleted and 7. Marie, J. C., D. Liggitt, and A. Y. Rudensky. 2006. Cellular mechanisms of fatal cells were activated with a suboptimal concentration of TGF-b1 early-onset autoimmunity in mice with the T cell-specific targeting of trans- (2.5 ng/ml), the response was lost. In contrast, CD4 cells with only forming growth factor-beta receptor. Immunity 25: 441–454. 8. Shull, M. M. I., I. Ormsby, A. B. Kier, S. Pawlowski, R. J. Diebold, M. Yin, one functional Smad2 allele and one functional Smad3 allele and R. Allen, C. Sidman, G. Proetzel, D. Calvin, et al. 1992. Targeted disruption of activated in the presence of 2.5 ng/ml TGF-b1 showed only a the mouse transforming growth factor-b1 gene results in multifocal inflamma- partial loss of selectin ligand induction compared with the higher tory disease. Nature 359: 693–699. 9. Kulkarni, A. B., C. G. Huh, D. Becker, A. Geiser, M. Lyght, K. C. Flanders, dose. These results imply that a minimum number of Smad4- A. B. Roberts, M. B. Sporn, J. M. Ward, and S. Karlsson. 1993. Transforming containing complexes are necessary for induction of Runx2 growth factor beta 1 null mutation in mice causes excessive inflammatory response and early death. Proc. Natl. Acad. Sci. USA 90: 770–774. (and/or other gene) expression upstream of Fut7 and Gcnt1 in- 10. Gu, A. D., Y. Wang, L. Lin, S. S. Zhang, and Y. Y. Wan. 2012. Requirements of duction and selectin ligand expression. It is possible that in cells transcription factor Smad-dependent and -independent TGF-b signaling to with limiting numbers of Smad4-containing complexes, genes control discrete T-cell functions. Proc. Natl. Acad. Sci. USA 109: 905–910. 11. Shi, Y., and J. Massague´. 2003. Mechanisms of TGF-b signaling from cell other than Runx2 may bind Smad4 complexes with higher affinity, membrane to the nucleus. Cell 113: 685–700. thereby preventing transcriptional activation of genes with lower 12. Derynck, R., Y. Zhang, and X.-H. Feng. 1998. Smads: transcriptional activators affinity Smad4 complex binding sites. This could explain why of TGF-b responses. Cell 95: 737–740. 13. Derynck, R., and Y. E. Zhang. 2003. Smad-dependent and Smad-independent lower doses of either Smad genes or TGF-b1 result in discrete pathways in TGF-beta family signalling. Nature 425: 577–584. bimodal expression of selectin ligands (or no expression) rather 14. Mu, Y., S. K. Gudey, and M. Landstro¨m. 2012. Non-Smad signaling pathways. Cell Tissue Res. 347: 11–20. than a reduction in the surface density, as reflected in mean 15. Ley, K., and G. S. Kansas. 2004. Selectins in T-cell recruitment to non-lymphoid fluorescence intensity. Additionally, we found that the loss of one tissues and sites of inflammation. Nat. Rev. Immunol. 4: 325–335. allele each of Smad2 and Smad3 led to a greater loss of selectin 16. Lasky, L. A. 1992. Selectins: interpreters of cell-specific carbohydrate information during inflammation. Science 258: 964–969. ligand expression than the loss of both alleles of Smad2 or both 17. Austrup, F., D. Vestweber, E. Borges, M. Lo¨hning, R. Braüer, U. Herz, H. Renz, alleles of Smad3. This implies that both Smad2 and Smad3 are R. Hallmann, A. Scheffold, A. Radbruch, and A. Hamann. 1997. P- and 2634 Smad TRANSCRIPTION FACTORS IN SELECTIN LIGAND EXPRESSION

E-selectin mediate recruitment of T-helper-1 but not T-helper-2 cells into role of the Runx1-Cbfb transcription complex for in vivo-suppressive function of inflammed tissues. Nature 385: 81–83. FoxP3+ regulatory T cells. Immunity 31: 609–620. 18. Wagers, A. J., C. M. Waters, L. M. Stoolman, and G. S. Kansas. 1998. Inter- 37. Ono, M., H. Yaguchi, N. Ohkura, I. Kitabayashi, Y. Nagamura, T. Nomura, leukin 12 and interleukin 4 control T cell adhesion to endothelial selectins Y. Miyachi, T. Tsukada, and S. Sakaguchi. 2007. Foxp3 controls regulatory through opposite effects on a1,3-fucosyltransferase VII gene expression. J. Exp. T-cell function by interacting with AML1/Runx1. Nature 446: 685–689. Med. 188: 2225–2231. 38. Rudra, D., T. Egawa, M. M. Chong, P. Treuting, D. R. Littman, and 19. Wagers, A. J., and G. S. Kansas. 2000. Potent induction of a1,3-fucosyl- A. Y. Rudensky. 2009. Runx-CBFb complexes control expression of the tran- transferase VII by TGF-b1 through a p38 mitogen-activated protein kinase- scription factor Foxp3 in regulatory T cells. Nat. Immunol. 10: 1170–1177. dependent pathway. J. Immunol. 165: 5011–5016. 39. Zhang, F., G. Meng, and W. Strober. 2008. Interactions among the transcription 20. Ebel, M. E., O. Awe, M. H. Kaplan, and G. S. Kansas. 2015. Diverse inflam- factors Runx1, RORgt and Foxp3 regulate the differentiation of interleukin 17- matory cytokines induce selectin ligand expression on murine CD4 T cells via producing T cells. Nat. Immunol. 9: 1297–1306. p38a MAPK. J. Immunol. 194: 5781–5788. 40. Bruno, L., L. Mazzarella, M. Hoogenkamp, A. Hertweck, B. S. Cobb, S. Sauer, 21. Takimoto, T., Y. Wakabayashi, T. Sekiya, N. Inoue, R. Morita, K. Ichiyama, S. Hadjur, M. Leleu, Y. Naoe, J. C. Telfer, et al. 2009. Runx proteins regulate R. Takahashi, M. Asakawa, G. Muto, T. Mori, et al. 2010. Smad2 and Smad3 are Foxp3 expression. J. Exp. Med. 206: 2329–2337. redundantly essential for the TGF-b-mediated regulation of regulatory T plas- 41. Klunker, S., M. M. Chong, P. Y. Mantel, O. Palomares, C. Bassin, M. Ziegler, ticity and Th1 development. J. Immunol. 185: 842–855. B. Ruckert,€ F. Meiler, M. Akdis, D. R. Littman, and C. A. Akdis. 2009. Tran- 22. Yang, X. O., R. Nurieva, G. J. Martinez, H. S. Kang, Y. Chung, B. P. Pappu, scription factors RUNX1 and RUNX3 in the induction and suppressive function B. Shah, S. H. Chang, K. S. Schluns, S. S. Watowich, et al. 2008. Molecular of Foxp3+ inducible regulatory T cells. J. Exp. Med. 206: 2701–2715. antagonism and plasticity of regulatory and inflammatory T cell programs. Im- 42. Wang, A., D. Pan, Y. H. Lee, G. J. Martinez, X. H. Feng, and C. Dong. 2013. munity 29: 44–56. Cutting edge: Smad2 and Smad4 regulate TGF-b-mediated Il9 gene expression 23. Gu, A. D., S. Zhang, Y. Wang, H. Xiong, T. A. Curtis, and Y. Y. Wan. 2015. A via EZH2 displacement. J. Immunol. 191: 4908–4912. critical role for transcription factor Smad4 in T cell function that is independent 43. Tamiya, T., K. Ichiyama, H. Kotani, T. Fukaya, T. Sekiya, T. Shichita, K. Honma, of transforming growth factor b receptor signaling. Immunity 42: 68–79. K. Yui, T. Matsuyama, T. Nakao, et al. 2013. Smad2/3 and IRF4 play a coop- 24. Maly´, P., A. Thall, B. Petryniak, C. E. Rogers, P. L. Smith, R. M. Marks, erative role in IL-9-producing T cell induction. J. Immunol. 191: 2360–2371. R. J. Kelly, K. M. Gersten, G. Cheng, T. L. Saunders, et al. 1996. The a(1,3) 44. Teraki, Y., and L. J. Picker. 1997. Independent regulation of fucosyltransferase Fuc-TVII controls leukocyte trafficking through an essential cutaneous lymphocyte-associated antigen expression and cytokine synthesis Downloaded from role in L-, E-, and P-selectin ligand biosynthesis. Cell 86: 643–653. phenotype during human CD4+ memory T cell differentiation. J. Immunol. 159: 25. Knibbs, R. N., R. A. Craig, P. Ma´ly, P. L. Smith, F. M. Wolber, N. E. Faulkner, 6018–6029. J. B. Lowe, and L. M. Stoolman. 1998. a(1,3)-Fucosyltransferase VII-dependent 45. Kretschmer, U., K. Bonhagen, G. F. Debes, H. W. Mittrucker,€ K. J. Erb, synthesis of P- and E-selectin ligands on cultured T lymphoblasts. J. Immunol. O. Liesenfeld, D. Zaiss, T. Kamradt, U. Syrbe, and A. Hamann. 2004. Expression 161: 6305–6315. of selectin ligands on murine effector and IL-10-producing CD4+ T cells from 26. Homeister, J. W., A. D. Thall, B. Petryniak, P. Maly´, C. E. Rogers, P. L. Smith, non-infected and infected tissues. Eur. J. Immunol. 34: 3070–3081. R. J. Kelly, K. M. Gersten, S. W. Askari, G. Cheng, et al. 2001. The a(1,3) 46. Syrbe, U., U. Hoffmann, K. Schlawe, O. Liesenfeld, K. Erb, and A. Hamann.

fucosyltransferases FucT-IV and FucT-VII exert collaborative control over 2006. Microenvironment-dependent requirement of STAT4 for the induction of http://www.jimmunol.org/ selectin-dependent leukocyte recruitment and lymphocyte homing. Immunity 15: P-selectin ligands and effector cytokines on CD4+ T cells in healthy and parasite- 115–126. infected mice. J. Immunol. 177: 7673–7679. 27. Ellies, L. G., S. Tsuboi, B. Petryniak, J. B. Lowe, M. Fukuda, and J. D. Marth. 47. Sperandio,M.,A.Thatte,D.Foy,L.G.Ellies,J.D.Marth,andK.Ley.2001.Severe 1998. Core 2 oligosaccharide biosynthesis distinguishes between selectin ligands impairment of leukocyte rolling in venules of core 2 glucosaminyltransferase- essential for leukocyte homing and inflammation. Immunity 9: 881–890. deficient mice. Blood 97: 3812–3819. 28. Snapp, K. R., C. E. Heitzig, L. G. Ellies, J. D. Marth, and G. S. Kansas. 2001. 48. Wang, Q., T. Stacy, J. D. Miller, A. F. Lewis, T. L. Gu, X. Huang, Differential requirements for the O-linked branching enzyme core 2 b1-6-N- J. H. Bushweller, J. C. Bories, F. W. Alt, G. Ryan, et al. 1996. The CBFb subunit glucosaminyltransferase in biosynthesis of ligands for E-selectin and P-selectin. is essential for CBFa2 (AML1) function in vivo. Cell 87: 697–708. Blood 97: 3806–3811. 49. Miller, J., A. Horner, T. Stacy, C. Lowrey, J. B. Lian, G. Stein, G. H. Nuckolls, 29. Ellies, L. G., M. Sperandio, G. H. Underhill, J. Yousif, M. Smith, J. J. Priatel, and N. A. Speck. 2002. The core-binding factor b subunit is required for bone G. S. Kansas, K. Ley, and J. D. Marth. 2002. Sialyltransferase specificity in formation and hematopoietic maturation. Nat. Genet. 32: 645–649.

selectin ligand formation. Blood 100: 3618–3625. 50. Greenblatt, M. B., J. H. Shim, W. Zou, D. Sitara, M. Schweitzer, D. Hu, by guest on September 24, 2021 30. Yang, W. H., C. Nussbaum, P. K. Grewal, J. D. Marth, and M. Sperandio. 2012. S. Lotinun, Y. Sano, R. Baron, J. M. Park, et al. 2010. The p38 MAPK pathway Coordinated roles of ST3Gal-VI and ST3Gal-IV sialyltransferases in the syn- is essential for skeletogenesis and bone homeostasis in mice. J. Clin. Invest. 120: thesis of selectin ligands. Blood 120: 1015–1026. 2457–2473. 31. Ito, Y., and K. Miyazono. 2003. RUNX transcription factors as key targets of 51. Greenblatt, M. B., J. H. Shim, and L. H. Glimcher. 2013. Mitogen-activated TGF-b superfamily signaling. Curr. Opin. Genet. Dev. 13: 43–47. protein kinase pathways in osteoblasts. Annu. Rev. Cell Dev. Biol. 29: 63–79. 32. Martinez, G. J., Z. Zhang, J. M. Reynolds, S. Tanaka, Y. Chung, T. Liu, 52. Syrbe, U., S. Jennrich, A. Schottelius, A. Richter, A. Radbruch, and A. Hamann. E. Robertson, X. Lin, X. H. Feng, and C. Dong. 2010. Smad2 positively reg- 2004. Differential regulation of P-selectin ligand expression in naive versus ulates the generation of Th17 cells. J. Biol. Chem. 285: 29039–29043. memory CD4+ T cells: evidence for epigenetic regulation of involved glyco- 33. Martinez, G. J., Z. Zhang, Y. Chung, J. M. Reynolds, X. Lin, A. M. Jetten, syltransferase genes. Blood 104: 3243–3248. X. H. Feng, and C. Dong. 2009. Smad3 differentially regulates the induction of 53. Bird, J. J., D. R. Brown, A. C. Mullen, N. H. Moskowitz, M. A. Mahowald, regulatory and inﬂammatory T cell differentiation. J. Biol. Chem. 284: 35283– J. R. Sider, T. F. Gajewski, C. R. Wang, and S. L. Reiner. 1998. Helper T cell 35286. differentiation is controlled by the cell cycle. Immunity 9: 229–237. 34. Malhotra, N., E. Robertson, and J. Kang. 2010. SMAD2 is essential for TGFb- 54. Yoon, J. H., K. Sudo, M. Kuroda, M. Kato, I. K. Lee, J. S. Han, S. Nakae, mediated Th17 cell generation. J. Biol. Chem. 285: 29044–29048. T. Imamura, J. Kim, J. H. Ju, et al. 2015. Phosphorylation status determines the 35. Lazarevic, V., X. Chen, J. H. Shim, E. S. Hwang, E. Jang, A. N. Bolm, opposing functions of Smad2/Smad3 as STAT3 cofactors in TH17 differentiation. M. Oukka, V. K. Kuchroo, and L. H. Glimcher. 2011. T-bet represses TH17 Nat. Commun. 6: 7600. differentiation by preventing Runx1-mediated activation of the gene encoding 55. Elyaman, W., R. Bassil, E. M. Bradshaw, W. Orent, Y. Lahoud, B. Zhu, RORgt. Nat. Immunol. 12: 96–104. F. Radtke, H. Yagita, and S. J. Khoury. 2012. Notch receptors and Smad3 sig- 36. Kitoh, A., M. Ono, Y. Naoe, N. Ohkura, T. Yamaguchi, H. Yaguchi, naling cooperate in the induction of interleukin-9-producing T cells. Immunity I. Kitabayashi, T. Tsukada, T. Nomura, Y. Miyachi, et al. 2009. Indispensable 36: 623–634.