TGF-β1 prevents the development of autoimmune disease by restraining the development of autoreactive Th1 cells. TGF-β1 inhibits Th1 development in part by suppressing the expression of T-bet, an IFN-γ-induced transcription factor that promotes Th1 differentiation, but how TGF-β1 suppresses T-bet is not known. In this study we show that TGF-β1 suppresses IFN-γ-induced T-bet expression through the hemopoietic protein tyrosine phosphatase (PTP) Src homology region 2 domain-containing phosphatase-1 (Shp-1). In murine CD4+ T cells, IFN-γ rapidly induced the expression of T-bet as well as of IFN regulatory factor-1, another transcription factor important for Th1 development. TGF-β1 antagonized the effects of IFN-γ, inhibiting IFN-γ’s induction of both Th1 transcription factors. In the presence of IFN-γ, TGF-β1 rapidly induced in Th cells the synthesis of the PTP Shp-1, but did not induce Shp-2 or several members of the suppressor of cytokine signaling family of Jak-Stat inhibitors. We tested the requirement for Shp-1 by using T cells from the Shp-1-deficient mev/mev mouse strain. Shp-1 was required for TGF-β1’s suppressive effects, because its suppression of T-bet and IFN regulatory factor-1 was completely abrogated in mev/mev CD4+ T cells. Receptor-proximal responses to IFN-γ, such as the induction of Jak-Stat phosphorylation, were inhibited by TGF-β1 in wild-type T cells, but not in mev/mev T cells. Consistent with a direct role for Shp-1, TGF-β1’s inhibition of IFN-γ-induced Stat1 phosphorylation was sensitive to the general PTP inhibitor pervanadate. Together, these data show that TGF-β1 suppresses IFN-γ signaling and transcriptional responses in CD4+ T cells through the PTP Shp-1.
During their initial priming with Ag, CD4+ T cells integrate signals received from cytokines in the microenvironment to activate subsequent effector cell development pathways that functionally shape the immune response. IFN-γ is a pleiotropic member of the hemopoietic cytokine family that promotes Th1 effector cell development through direct actions on CD4+ T cells (1). The binding of IFN-γ to its receptor activates the tyrosine phosphorylation of Jak1 and Stat1, and through nuclear translocation and transcriptional activity of Stat1 homodimers, this pathway stimulates the expression of genes encoding proteins important for Th1 development, such as the transcription factors T-bet and IFN regulatory factor-1 (IRF-1)3 (2, 3).
TGF-β1 is the signature cytokine of an evolutionarily distinct family of cytokines that signals through Ser/Thr kinases and Smad proteins. TGF-β1 has potent immunoregulatory activity (4) and suppresses immunity in part by repressing the activation, differentiation, and effector functions of Th cells (5). Mice deficient in TGF-β1 spontaneously develop multifocal CD4+ T cell-mediated autoimmunity (6). CD4+ T cells from TGF-β1−/− mice are strongly skewed to the Th1 phenotype (7). TGF-β1 inhibits Th1 development in large part by suppressing T-bet expression (8, 9). The mechanism by which TGF-β1 inhibits T-bet expression has not yet been determined.
TGF-β1 and IFN-γ occupy a mutually antagonistic relationship (10). BALB/c TGF-β1−/− mice spontaneously develop CD4+ T cell-mediated necroinflammatory liver disease, with overexpression of IFN-γ in liver. IFN-γ is required for disease, because BALB/c-TGF-β1−/−IFN-γ−/− double-knockout mice are protected (7, 11). Tissues isolated from TGF-β1−/− mice show chronic activation of the IFN-γ signaling pathway (12), indicating that TGF-β1 has an important role in vivo in inhibiting IFN-γ signaling. Unraveling the mechanisms underlying the antagonistic relationship between TGF-β1 and IFN-γ is important for understanding the delicate balance between tolerance and autoimmunity. IFN-γ can inhibit cellular responses to TGF-β1 through the induction of expression of Smad7, an inhibitory member of the Smad family that interferes with TGF-β1-induced Smad3 activation (13). IFN-γ can also inhibit TGF-β1 responses through Stat1-mediated sequestration of the nuclear coactivator p300/CREB binding protein (14), preventing its association with Smads and blocking Smad transcriptional activity. In contrast, very little is known about the mechanisms by which TGF-β1 interferes with cellular responses to IFN-γ.
In this study we ask whether and how TGF-β1 inhibits IFN-γ’s induction of T-bet and IRF-1 during the priming of CD4+ T cells. Both T-bet and IRF-1 are necessary for the development of Th1 cells. We chose to study CD4+ T cells at initial priming, because TGF-β1 and IFN-γ exert their powerful effects on immune responses and Th1 development largely through influencing T cell responses early in the course of the adaptive immune response. We found that TGF-β1 indeed inhibits IFN-γ’s induction of IRF-1 and T-bet gene expression in CD4+ T cells. These effects of TGF-β1 require new protein synthesis, indicating that one or more factors must be synthesized to mediate TGF-β1’s effects. Therefore, we examined whether TGF-β1 affects the expression levels in T cells of genes within two families of proteins, the suppressor of cytokine signaling (SOCS) family and the protein tyrosine phosphatase (PTP) family. Members of these families suppress IFN-γ signaling through a variety of mechanisms (15) and would be reasonable candidates to mediate the effects of TGF-β1. We found that TGF-β1 in conjunction with IFN-γ rapidly induced the expression of the PTP Src homology region 2 domain-containing phosphatase-1 (Shp-1). Using T cells from Shp-1-deficient mev/mev mice, we found that Shp-1 is required for TGF-β1’s inhibition of IFN-γ’s induction of T-bet and IRF-1. Additional analyses showed that Shp-1 is also required for TGF-β1’s inhibition of certain receptor-proximal signaling events induced by IFN-γ, including the tyrosine phosphorylation of Jak1 and Stat1. Thus, in CD4+ T cells, TGF-β1 inhibits IFN-γ-induced Jak-Stat phosphorylation and the induction of the Th1 transcription factors, T-bet and IRF-1, through the PTP Shp-1.
Materials and Methods
Cycloheximide (CHX; Sigma-Aldrich) was prepared in absolute ethanol and used at a final concentration of 0.5 mM. Pervanadate (PV; 10 mM) was prepared by incubating 10 mM sodium orthovanadate (Sigma-Aldrich) with 10 mM hydrogen peroxide for 20 min at room temperature, then was used immediately at a final concentration of 25 μM. Recombinant IFN-γ and TGF-β1 were purchased from PeproTech and R&D Systems, respectively.
BALB/cJ mice and BALB/cJ-IFN-γ−/− mice were purchased from The Jackson Laboratory and maintained under pathogen-free conditions in a Dartmouth Medical School animal care facility. The mev mutation has been extensively backcrossed to the BALB/cByJ background, and mev/mev and littermate control (+/mev and +/+) mice were bred from heterozygous breeders at The Jackson Laboratory.
Dissected murine spleens and lymph nodes were disrupted, and cell suspensions were centrifuged. Collected cells were resuspended in RPMI 1640 (Mediatech), and mononuclear cells were isolated over Ficoll (Histopaque 1119; Sigma-Aldrich). CD4+ cells were positively selected using magnetic beads, resuspended in medium containing 10% FBS, and plated at 1 × 106 cells/well on immobilized anti-CD3 (10 μg/ml) mAb and soluble anti-CD28 mAb (1 μg/ml). After overnight culture, cells were washed with serum-free RPMI 1640 medium and rested for 2 h, followed by treatment with IFN-γ (50 ng/ml) and/or TGF-β1 (5 ng/ml) for various times as indicated in the figure legends.
After cytokine treatment, cells were lysed in sample buffer containing SDS and 2-ME. Boiled samples were run on a 10% polyacrylamide gel, transferred to nitrocellulose membrane, and probed with Abs. Abs recognizing IRF-1, phosphotyrosine-Stat1 (Y701), Stat1 (p84/p91), Jak1, Shp-1, or tubulin were purchased from Santa Cruz Biotechnology. Abs recognizing phosphoserine-Stat1 (S727) and phosphotyrosine-Jak1 (Y1022/Y1023) were purchased from Upstate Biologicals and Cell Signaling Technology, respectively. Detection was performed using labeled appropriate secondary Ab and ECL. After detection of phosphorylated Jak1 or phosphorylated Stat1, in some experiments (Figs. 5,C, 6, and 7) heat was used to strip Abs off the filter, and after confirming the efficacy of stripping, filters were reprobed with the corresponding Ab to total Jak1 or total Stat1 as appropriate. In other experiments (Figs. 5, A and B, and 8,A), phosphorylated Jak1 (or phosphorylated Stat1) and total Jak1 (or total Stat1) were determined in separate gels loaded with equal aliquots of the cell lysates. For Western blots involving IRF-1 (Fig. 1,A) or Shp-1 (Fig. 2 A), equal aliquots of lysates were run on replicate gels that were probed with anti-Shp-1 and anti-tubulin, respectively. Densitometric analysis of Western blots was performed using ImageJ software (version 1.33u), downloaded from the National Institutes of Health.
RNA isolation and RT-PCR
Cells were harvested, and total RNA was isolated using an RNeasy kit (Qiagen). cDNA was generated from 1 μg of total RNA using the Omniscript kit (Qiagen) and was subjected to conventional PCR or real-time quantitative PCR using SYBR Green and/or the TaqMan system (Applied Biosystems). Sequences for primer pairs are as follows: IRF-1, 5′-TTAGC CCGGACACTTTCTCTGATGG-3′ and 5′-GTCCCCTCGAGGGCTGTC AATCT-3′; T-bet, 5′-CAACAACCCCTTTGCCAAAG-3′ and 5′-TCCCC CAAGCAGTTGACAGT-3′; probe, FAM-5′-CCGGGAGAACTTTGAG TCCATGTACGC-3′-TAMRA; SOCS-1, 5′-CTCGAGTAGGATGGTAG CACGCAA-3′ and 5′-CATCTTCACGCTGAGCGCTGAGCGCGAAGA A-3′; SOCS-2, 5′-GACCAGCTGTCTGGGACGTGTTGA-3′ and 5′-GA GAGAGAAATACTTATACCTGGAAT-3′; SOCS-3, 5′-TGCGCCATGG TCACCCACAGCAAGTTT-3′ and 5′-GCTCCTTAAAGTGGAGCATC ATACTGA-3′; Shp-1, 5′-AGTTGAATTCATGCTGTCCCGTGGGTGGTT TCAC-3′ and 5′-GGCCGAATTCCCCCGTCTTCTTGAAATGCTCTAC-3′; Shp-2, 5′-GGTTGAATTCATGACATCGCGGAGATGGTTTCAC-3′ and 5′-GGCCGAATTCGGATTCTTCTTATAATGTTCCACA-3′; and β-actin, 5′-GTGGGCCGCTCTAGGCACCAA-3′ and 5′-TCTTTGCCAATAG TGATGACTTGGC-3′.
For most amplicons, amplification was performed for 40 cycles at 94°C for 15 s, 62°C for 45 s, and 72°C for 15 s. For Shp-1 and Shp-2 amplicons, amplification was performed for 40 cycles at 94°C for 30 s, 50°C for 30 s, and 72°C for 45 s. All real-time PCR applications involving SYBR Green included a melting curve check analysis.
First, we determined whether TGF-β1 inhibits Th cell responses to IFN-γ. We examined the expression of IRF-1, a well-characterized immediate response protein strongly induced by IFN-γ (16) and required for Th1 development (17). Murine CD4+ T cells were isolated to high purity, then stimulated in vitro with a combination of anti-CD3 and anti-CD28 and with IFN-γ and TGF-β1, separately and in combination. IRF-1 protein expression was assessed by Western blot. In cells stimulated in the presence of IFN-γ, IRF-1 protein was strongly induced compared with expression in cells stimulated in the absence of IFN-γ (Fig. 1,A). Inclusion of TGF-β1 completely inhibited IFN-γ-induced IRF-1 expression. A low level of IRF-1 expression even without specific addition of IFN-γ was observed in some experiments. This was due to autocrine IFN-γ, because inclusion of a neutralizing anti-IFN-γ Ab suppressed expression (Fig. 1 A, lane 3). Therefore, in several subsequent experiments, we used CD4+ T cells from IFN-γ−/− mice to avoid potentially confounding effects of endogenous IFN-γ.
To determine whether inhibitory effects of TGF-β1 were observed at the mRNA level for IRF-1, IFN-γ−/− CD4+ T cells were stimulated overnight with anti-CD3/28, washed, rested for 2 h, and incubated in IFN-γ, without or with TGF-β1. IFN-γ strongly induced IRF-1 mRNA expression (Fig. 1,B); a time-course analysis (not shown) indicated peak expression 3 h after the addition of IFN-γ, similar to published results using human macrophages (16). As observed for IRF-1 protein, the induction of IRF-1 mRNA expression was strongly suppressed by TGF-β1. We next examined the expression of T-bet, because TGF-β1 suppresses Th1 development largely through suppressing the expression of this Th1 transcription factor. T-bet mRNA expression was similar to that observed for IRF-1. That is, IFN-γ strongly induced T-bet mRNA expression at 3 h, and this effect was completely suppressed by TGF-β1 (Fig. 1 B).
IFN-γ induces the expression of immediate target genes through Jak-Stat signaling. Because the components of this pathway are constitutively present, no new protein synthesis is required for signaling through this pathway, and IFN-γ signaling to IRF-1 induction is intact even in the presence of protein synthesis inhibitors, such as CHX (16). Therefore, we asked whether the suppressive effects of TGF-β1 on IFN-γ’s induction of expression of IRF-1 and T-bet are direct or indirect, requiring the synthesis of one or more intermediary proteins. To test this, we repeated these experiments using CHX. In the presence of CHX, the baseline expression levels of both IRF-1 and T-bet were higher than in its absence, as expected (16). As previously reported (16), strong induction of expression by IFN-γ was observed even in the presence of CHX, indicating that IFN-γ can induce the expression of IRF-1 and T-bet through a pathway that does not require new protein synthesis. However, CHX completely abrogated suppression by TGF-β1 (Fig. 1 B). This indicates that new protein synthesis is required to mediate TGF-β1’s suppressive effects on IRF-1 and T-bet gene expression.
Accordingly, we looked for candidates induced by TGF-β1 that could mediate its effects on IFN-γ signaling. Cytokine signaling can be inhibited by three principal mechanisms, involving members of the PTP, the SOCS, or the protein inhibitor of activated Stat families of proteins (18). Shp-1 is a PTP expressed primarily in hemopoietic cells (19) that inhibits IFN-γ signaling (20). We tested whether TGF-β1 induces the expression of Shp-1 in CD4+ T cells. T cells were stimulated overnight with anti-CD3/anti-CD28, washed, and rested for 2 h, then incubated with IFN-γ, TGF-β1, or both cytokines for 0.5–6 h. Cells were lysed, and Shp-1 expression was analyzed by Western blot (Fig. 2,A). The quantitative analysis of this blot by densitometry is shown in Fig. 2,B. Shp-1 is present in T cells at baseline, that is, without specific addition of either IFN-γ or TGF-β1 (Fig. 2 A, left-most lane). The addition of either IFN-γ alone or TGF-β1 alone modestly induced Shp-1 expression. However, induction was clearly greatest with the two cytokines combined. In the presence of both IFN-γ and TGF-β1, Shp-1 expression was augmented as early as 0.5 h, and the level of Shp-1 protein progressively increased with time, reaching >30-fold by 6 h.
To analyze the effects of IFN-γ and TGF-β1, we assessed Shp-1 mRNA levels. T cells were stimulated overnight with anti-CD3/anti-CD28, washed, rested, and incubated with cytokines for 3 h, and Shp-1 mRNA was measured by real-time PCR. IFN-γ alone had little effect on Shp-1 mRNA expression levels, whereas TGF-β1 alone modestly induced Shp-1 mRNA expression levels. The strongest induction of Shp-1 mRNA was observed with the combination of IFN-γ and TGF-β1. In additional studies, we observed that the induction of Shp-1 mRNA by the combination of IFN-γ and TGF-β1 was observed even in the presence of CHX, indicating that Shp-1 mRNA induction is direct and not through the synthesis of one or more intermediary signaling proteins (data not shown). We did not assess the induction of Shp-1 protein under these conditions, because CHX interferes with this analysis.
We next determined the degree of specificity of the induction of Shp-1 expression. TGF-β1 did not induce detectable mRNA expression of the very closely related PTP family member Shp-2. Indeed, expression of Shp-2 in murine primary CD4+ T cells was not observed, although its expression was easily detected in the Jurkat T cell line (Fig. 3,A). Expression of mRNAs encoding SOCS-1, SOCS-2, or SOCS-3 was observed in primary murine CD4+ T cells, but expression levels were little affected by IFN-γ or IFN-γ plus TGF-β1 compared with medium alone (Fig. 3 B).
TGF-β1’s inhibition of IFN-γ’s induction of IRF-1 and of T-bet expression requires new protein synthesis, and Shp-1 is unique among the five genes examined in this study in its induction by TGF-β1 (with IFN-γ). We therefore hypothesized that Shp-1 mediates the effects of TGF-β1 on IFN-γ’s induction of T-bet and IRF-1 expression. To rigorously test this hypothesis, we isolated CD4+ T cells from BALB/c mice homozygous for the viable motheaten (Hcphme-v, abbreviated mev) defect. The mev/mev mice lack almost all Shp-1 activity due to a point mutation in the Shp-1 protein (21). In both littermate control and mev/mev CD4+ T cells, IFN-γ strongly induced the expression of T-bet mRNA and IRF-1 mRNA, although the induction of T-bet was somewhat less robust in mev/mev T cells than in wild-type T cells. In Shp-1-replete littermate control cells, TGF-β1 suppressed IFN-γ’s induction of these two mRNAs, as expected. In Shp-1-deficient mev/mev T cells, however, TGF-β1 had no effect on IFN-γ’s induction of IRF-1 and T-bet mRNAs (Fig. 4). Thus, Shp-1 activity is necessary for TGF-β1’s inhibition of IFN-γ’s induction of T-bet and IRF-1.
Early cellular responses to IFN-γ include the tyrosine phosphorylations of Jak1 and Stat1 (22). We therefore considered whether TGF-β1 inhibits these receptor-proximal effects of IFN-γ in CD4+ T cells. Primary murine CD4+ T cells were stimulated overnight in anti-CD3/CD28, rested for 2 h, and then stimulated with IFN-γ without or with TGF-β1 for 15–60 min. Cells were lysed, and phospho-Jak1 and phospho-Stat1 were assessed by Western blot. In response to IFN-γ, Jak1 was rapidly and strongly tyrosine phosphorylated; in the presence of TGF-β1, this Jak1 phosphorylation was completely abrogated (Fig. 5,A). Western blot for total Jak1 showed that the effects of IFN-γ and TGF-β1 were specific to the phosphorylated form and were not due to changes in protein expression levels. A similar result was found for Stat1, with IFN-γ inducing, and TGF-β1 suppressing, its tyrosine phosphorylation (Fig. 5,B). IFN-γ also induces the serine phosphorylation of Stat1 on residue 727 (23). TGF-β1 did not block this response (Fig. 5 C), however, indicating that TGF-β1 does not block all forms of cellular signaling in response to IFN-γ, but, rather, shows a degree of selectivity.
We asked whether Shp-1 mediates TGF-β1’s suppression of IFN-γ-induced receptor-proximal tyrosine phosphorylation events. T cells were isolated from mev/mev mice and littermate control mice, and the tyrosine phosphorylation of Jak1 and Stat1 was assessed as before. In both control and mev/mev T cells, IFN-γ induced the tyrosine phosphorylation of Jak1. In control T cells, TGF-β1 inhibited IFN-γ-induced Jak1 tyrosine phosphorylation at the time points (15, 30, and 60 min) examined (Fig. 6,A). In T cells from mev/mev mice, however, TGF-β1 did not inhibit IFN-γ-induced Jak1 tyrosine phosphorylation at 15 and 30 min; partial inhibition at 60 min was sometimes observed. Similar results were obtained for Stat1; that is, IFN-γ induced Stat1 tyrosine phosphorylation in both control cells and mev/mev cells, and TGF-β1 inhibited Stat1 phosphorylation in Shp-1-intact control T cells, but not in Shp-1-deficient mev/mev T cells (Fig. 6 B). Thus, Shp-1 is required for TGF-β1 to inhibit IFN-γ’s induction of tyrosine phosphorylation of Jak1 and of Stat1.
PV is a pharmacologic inhibitor of PTPs. To confirm that TGF-β1’s inhibition of Stat1 tyrosine phosphorylation required PTP activity, CD4+ T cells from wild-type control mice were stimulated with IFN-γ or IFN-γ plus TGF-β1 in the absence or the presence of PV. Under these conditions, IFN-γ induced Stat1 phosphorylation, as expected, but the ability of TGF-β1 to inhibit Stat1 tyrosine phosphorylation was abrogated (Fig. 7). This confirms that PTP activity is necessary for TGF-β1’s inhibition of IFN-γ-induced receptor proximal tyrosine phosphorylation responses.
Suppression by TGF-β1 of IRF-1 and T-bet gene expression requires new protein synthesis. TGF-β1 suppresses IFN-γ induced receptor-proximal tyrosine phosphorylation events within 15 min, a response so rapid, however, as to be unlikely to require new protein synthesis. To test the requirement for new protein synthesis, we examined the effects of TGF-β1 on IFN-γ-induced Stat1 phosphorylation in CD4+ T cells in the presence and the absence of CHX (Fig. 8). Fig. 8,A shows a representative Western blot, and Fig. 8 B shows the combined densitometric data from four separate experiments. At 15 min, IFN-γ induced Stat1 tyrosine phosphorylation regardless of whether protein synthesis was inhibited by CHX. This was expected, because IFN-γ induces the phosphorylation of Jak1 and Stat1 through receptor clustering and Jak1 activation, and not through the synthesis of an intermediate factor. At 15 min, TGF-β1 inhibited IFN-γ-induced Stat1 phosphorylation regardless of whether protein synthesis was inhibited by CHX. At later time points, the results were different. In the absence of CHX, at 30 and 60 min, IFN-γ treatment induced, and TGF-β1 inhibited, Stat1 phosphorylation. In the presence of CHX, IFN-γ treatment induced Stat1 phosphorylation, albeit to a lesser degree at the later time points. Interestingly, with protein synthesis inhibited, TGF-β1 progressively lost the ability to inhibit IFN-γ-induced Stat1 phosphorylation. In the presence of CHX, at 30 min, TGF-β1 only modestly inhibited IFN-γ-induced Stat1 phosphorylation, and by 60 min, TGF-β1 did not inhibit (and possibly even enhanced) Stat1 phosphorylation. Similar results were obtained for Jak1 phosphorylation (data not shown). Thus, early (i.e., at 15 min) TGF-β1 uses existing cellular components to suppress receptor-proximal IFN-γ signaling through the Jak-Stat pathway. After ∼30 min, however, the synthesis of one or more proteins is required for TGF-β1 to sustain its inhibition of receptor-proximal IFN-γ signaling through the Jak-Stat pathway.
From these data, we would predict that the combination of IFN-γ and TGF-β1 does not enhance Shp-1 protein expression levels within the first 15 min of cytokine exposure. We tested this prediction directly. Indeed, within the first 15 min of cytokine treatment, Shp-1 protein levels were not enhanced compared with those in medium only and were perhaps slightly diminished at 15 min. As a positive control, IFN-γ and TGF-β1 strongly induced Shp-1 expression at 3 and 6 h, as expected (Fig. 8 C).
The key finding in this study is the defining of a novel pathway of regulation in Th cells. In collaboration with IFN-γ, TGF-β1 induces expression of the PTP Shp-1 in CD4+ T cells, and functional Shp-1 is directly involved in TGF-β1’s suppression of IFN-γ responses. Shp-1 is required both for TGF-β1’s inhibition of several IFN-γ-induced receptor-proximal signaling events, such as the tyrosine phosphorylation of Jak1 and Stat1, and for TGF-β1’s inhibition of IFN-γ’s induction of expression of the Th1 transcription factors T-bet and IRF-1. TGF-β1’s inhibition of IFN-γ-induced Jak1 and Stat1 phosphorylation at very early time points (within 15 min of cytokine exposure) does not require new protein synthesis, but, nevertheless, requires Shp-1, because TGF-β1 did not inhibit IFN-γ-induced Stat1 tyrosine phosphorylation in Shp-1-deficient mev/mev T cells. The role of Shp-1 in this early phase is supported by the observation that PV, a general inhibitor of PTPs, abrogates the ability of TGF-β1 to inhibit IFN-γ-induced Stat1 phosphorylation.
Shp-1 is expressed in T cells even before exposure to TGF-β1, suggesting that there is a mechanism by which TGF-β1 activates or engages a pre-existing pool of Shp-1 to effectively dephosphorylate tyrosine residues on Jak1 and/or Stat1 and thereby rapidly antagonize IFN-γ signaling through the Jak-Stat signaling pathway. One report showed that exposure of rat T cells to TGF-β1 rapidly induces the tyrosine phosphorylation of Shp-1, but the functional significance of this biochemical modification was not assessed (24). In our experiments, however, we have not detected the phosphorylation of Shp-1 after exposure of T cells to TGF-β1 (data not shown). Thus, the mechanism by which TGF-β1 engages the basal pool of Shp-1 to inhibit IFN-γ-Jak1-Stat1 signaling remains to be determined.
As far as we are aware, this study constitutes the first demonstration that TGF-β1 interferes with early IFN-γ signaling events in any system. Previous studies of astrocytes found no effect of TGF-β1 on early IFN-γ signaling events, including Jak-Stat activation (25). In astrocytes, TGF-β1 inhibits the induction of class II mRNA not by inhibiting receptor-proximal signaling events, but by suppressing the transcriptional induction of the gene encoding the coactivator molecule CIITA (26, 27). The simplest explanation for the differences from this report is that the mechanism by which TGF-β1 inhibits IFN-γ responses shows cell type specificity. Because Shp-1 expression is restricted to hemopoietic cells, it follows that in nonhemopoietic cells, TGF-β1 must use a different mechanism to inhibit IFN-γ responses. It is known that IFN-γ can inhibit TGF-β1 signaling by inducing the expression of the inhibitory Smad7, which negatively regulates TGF-β1/Smad signaling (13). IFN-γ can also inhibit TGF-β1 signaling in the nucleus, through Stat-mediated sequestration of the coactivator molecule CREB binding protein/p300 (14). Thus, these two antagonistic and evolutionarily distinct signaling pathways can inhibit one another through multiple distinct mechanisms involving effects both on cytoplasmic signaling (e.g., Shp-1 and Smad7) and on nuclear gene expression events (e.g., competition for coactivator molecules). Given the importance of striking the appropriate balance between effective immunity to pathogens and avoidance of autoimmunity, it is not surprising to find multiple mechanisms of cross-regulation between these two important cytokines.
The current study constitutes the first evidence directly linking TGF-β1, the specific PTP Shp-1, and inhibition of IFN-γ signaling. Interactions between TGF-β1 and PTPs and between TGF-β1 and Jak-Stat signaling have been suggested by previous studies in other contexts. TGF-β1 induces tyrosine phosphatase activity in keratinocytes (28). PV can partially abrogate the ability of TGF-β1 to inhibit TCR signaling in T cells (24) and thrombopoietin signaling in myeloid cell lines (29). In these studies, however, specific PTPs were neither identified nor mechanistically linked to these effects of TGF-β1. Several previous studies have shown that TGF-β1 can inhibit Jak-Stat signaling in T cells in response to a number of other hemopoietic cytokines, including IL-6 (30), IL-2 (31, 32), and IL-12 (33), although at least one contradictory report showed no inhibition by TGF-β1 of IL-2- and IL-12-induced Jak-Stat signaling (34). The mechanism(s) by which TGF-β1 inhibits Jak-Stat signaling from these other cytokine receptors is not known, but it is tempting to speculate that Shp-1 is involved.
The pathway by which TGF-β1 inhibits the early IFN-γ-induced Jak-Stat tyrosine phosphorylation events is not sufficient to sustain the inhibition of IFN-γ-induced, receptor-proximal signaling events or to inhibit IFN-γ’s induction of gene expression. These effects of TGF-β1 require new protein synthesis, because the presence of the protein synthesis inhibitor CHX interferes with TGF-β1’s ability to sustain past 15–30 min its inhibition of IFN-γ-induced Stat1 phosphorylation or to inhibit IFN-γ-induced IRF-1 and T-bet mRNA induction. In conjunction with IFN-γ, TGF-β1 induced Shp-1 protein expression in CD4+ T cells. Moreover, in Shp-1-deficient mev/mev CD4+ T cells, TGF-β1 failed to inhibit IFN-γ’s induction of IRF-1 and T-bet. We speculate that inhibition of IRF-1 and T-bet expression requires the synthesis of additional Shp-1 protein, over and above that required to inhibit the earliest receptor-proximal IFN-γ Jak-Stat tyrosine phosphorylation events. Thus, the Shp-1 present at baseline appears to be insufficient for sustained inhibition of the IFN-γ pathway, perhaps because Shp-1 is degraded or somehow inactivated soon after its activation or recruitment by TGF-β1. An alternative model is that TGF-β1 induces the synthesis of one or more proteins, in addition to Shp-1, that work in concert with Shp-1 to sustain the inhibition of Jak-Stat signaling and to enforce the inhibition of IRF-1 and T-bet expression. Our experiments do not distinguish between these possibilities.
In either case, Shp-1 is clearly required for the pathway by which TGF-β1 suppresses IFN-γ-induced Jak-Stat signaling and T-bet and IRF-1 expression. This suggests that Shp-1 may be integral to the mechanism by which TGF-β1 inhibits type 1 responses in T cells. Given the critical role of TGF-β1 in inhibiting the spontaneous development of Th1-mediated autoimmunity, it may be predicted that Shp-1 is a negative regulator of Th1 development. Indeed, it was recently shown that Shp-1 antagonizes Th1 development (35), but the mechanism of action of Shp-1 was not determined. Our study not only links TGF-β1 to Shp-1, but also suggests that the mechanism by which Shp-1 inhibits Th1 development is through regulation of expression of the Th1 transcription factors T-bet and IRF-1.
Shp-1 induction by TGF-β1 and IFN-γ: mechanisms and implications
How might TGF-β1 and IFN-γ synergize to induce Shp-1? Our observation that TGF-β1 rapidly suppresses IFN-γ-induced Jak-Stat signaling presents, at first inspection, a possible conundrum, in that TGF-β1 rapidly inhibits an important signaling pathway (Jak-Stat) initiated by the same cytokine (IFN-γ) with which it cooperates to augment Shp-1 expression. However, not all signaling initiated by IFN-γ is through the tyrosine phosphorylation of Jak1 and Stat1. Indeed, we found that the phosphorylation of Stat1 on serine 727 is induced by IFN-γ, but is not inhibited by TGF-β1, providing direct evidence in our system that at least one IFN-γ-activated signaling pathway remains intact in the presence of TGF-β1. Over the past several years, evidence has been accumulating that IFN-γ signals through Stat1-independent pathways as well as Stat1-dependent pathways (36). A microarray expression study using Stat1−/− mouse embryo fibroblasts established that IFN-γ induces the expression of a large subset of genes through Stat1-independent pathways (37). We speculate that the contribution of IFN-γ to the induction of Shp-1 expression is Stat1 independent, perhaps involving the MAPK pathway. In an analogous fashion, TGF-β1 signals through both Smad-dependent and Smad-independent pathways (38). We have generated preliminary data that TGF-β1 induces Shp-1 and suppresses IFN-γ responses in T cells from Smad3−/− mice (I.-K. Park, unpublished observations), indicating that TGF-β1’s contribution to the induction of Shp-1 expression is mediated by Smad3-independent pathways.
Shp-1 is a multifunctional PTP that regulates several key steps in TCR signaling, with potential target substrates including Vav, Lck, PI3K, and ZAP70 (reviewed in Ref.39). The net effect of these activities of Shp-1 is to raise the threshold for TCR signaling (40). In addition, Shp-1 can bind and dephosphorylate the β-chain of the activated IL-2R (41). Thus, high expression of Shp-1 could be expected to have a number of effects on a variety of T cell responses and functions. In this context, therefore, it may be important that strong induction of Shp-1 expression is observed only when both IFN-γ and TGF-β1 are added, with no or only modest induction when either cytokine is added individually. This indicates remarkable synergy and specificity in the regulation of expression of this PTP by these two evolutionarily distinct cytokines. The induction of Shp-1 expression by TGF-β1 in the context of IFN-γ signaling may be important for the attenuation of Th1 immune responses and for other immunoregulatory effects mediated by TGF-β1. Interestingly, a recent report shows that suppression of CD4 T cell division by an immunoregulatory T cell population involves both IFN-γ and TGF-β1 (42); it should be interesting to determine whether Shp-1 transduces the immunosuppressive effects of these two cytokines in that system.
We thank R. J. Noelle (Dartmouth Medical School) and T. F. Pearson (Dartmouth Medical School) for critical review of the manuscript, and Brent Berwin (Dartmouth Medical School) for helpful discussions.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grants AI053056 (to J.D.G.) and CA20408 (to L.D.S.) and Grant P20RR16437 from the COBRE Program of the National Center for Research Resources (to J.D.G.).
Abbreviations used in this paper: IRF-1, IFN regulatory factor-1; CHX, cycloheximide; PTP, protein tyrosine phosphatase; PV, pervanadate; Shp-1, Src homology region 2 domain-containing phosphatase-1; SOCS, suppressor of cytokine signaling.