Cross-talk between Transforming Growth Factor-β and Estrogen Receptor Signaling through Smad3*

Transforming growth factor-β (TGF-β) plays central roles in embryonic development, organogenesis, and physiologic connective tissue remodeling during wound healing and tissue repair as well as in carcinogenesis. Estrogens have key roles in a variety of biological events, such as the development and maintenance of female reproductive organs and bone and lipid metabolism. Previous studies demonstrated that estrogens suppress TGF-β-induced gene expression, such as type IV collagen in kidney mesangial cells. However, the molecular mechanisms that mediate this antagonistic effect are unknown. To elucidate the mechanisms of cross-talk between TGF-β and estrogen receptor (ER) signaling pathways, we reconstituted TGF-β and ER signaling in human kidney carcinoma cells. Here we demonstrate that TGF-β-induced activation of Sma and MAD-related protein 3 (Smad3) activity, one of the major intracellular transducers of TGF-β signaling, was suppressed by ER, whereas ER-mediated transcriptional activation was enhanced by TGF-β signaling. We provide evidence that this two-way cross-talk between the estrogen and TGF-β signaling pathways was the result of direct physical interactions between Smad3 and ER. These findings have implications for a variety of disease states, such as the pathophysiology of kidney function, atherosclerosis, and breast cancer.

Cellular homeostasis is maintained by the orchestrated functions of a variety of growth regulatory factors, such as peptide and lipophilic hormones. Interactions between growth factor and estrogen signaling pathways at the level of gene expression play important roles in maintaining reproductive homeostasis (1,2).
Transforming growth factor-␤ (TGF-␤) 1 is a member of the TGF-␤ superfamily. In addition to TGF-␤, this superfamily includes the activins/inhibins, bone morphogenetic proteins, and other members, such as Mü llerian inhibiting substance and Lefty (3). These factors regulate growth, differentiation, apoptosis, migration, and secretion of important molecules, including components of the extracellular matrix, adhesion molecules, hormones, and cytokines in a variety of cell types, affecting morphogenesis, tissue repair, tumor suppression, and immunoregulation (3).
TGF-␤ signaling is mediated through transmembrane receptors located at the cell surface (T␤Rs) which are serine/threonine kinases, which in turn use the highly conserved members of the Smad (Sma and MAD-related protein) family of transcription factors to transduce their signals to the nucleus (4). Two of the receptor-regulated Smads, Smad2 and Smad3, transduce signals for TGF-␤. Smad4, on the other hand, acts as a heterodimeric partner for Smad2 and Smad3 for efficient DNA binding and transcriptional activation (5,6). When T␤Rs are activated by the binding of their cognate ligands, Smad2 and Smad3 are phosphorylated by the type I receptor (T␤R-I) serine-threonine kinase. Phosphorylated Smad2 and Smad3 then form stable heterocomplexes with Smad4 which translocate into the nucleus and activate transcription. Smad7, one of the inhibitory Smads, stably interacts with activated T␤R-I and inhibits TGF-␤ signaling (7)(8)(9).
Estrogen receptor (ER) is a ligand-activated transcriptional factor that is a member of the nuclear receptor superfamily (10). Two types of ERs have been identified, ER␣ and ER␤, which appear to have overlapping but distinct roles in mediating estrogen action (11)(12)(13). Estrogens play important roles in the differentiation and development of various organs, maintenance of proper cellular function in a wide variety of tissues, and are also risk factors for breast and endometrial cancer (14). ERs interact with estrogen response elements in the target gene promoters and directly regulate their transcription (10). In addition, ERs interact with other signaling pathways for which DNA binding may not be necessary (15).
The development and progression of atherosclerosis is significantly reduced in women, and this may at least in part be the result of reduced accumulation of vascular wall extracellular matrix in response to estrogen action (16). Consistent with this hypothesis, estrogen administration reduces collagen type IV deposition in the aorta of hypertensive and hypercholesterolemic animals and reduces collagen synthesis by vascular smooth muscle cells in vitro (17). This has also been associated with a slower rate of progression of renal disease in women (16). In this context, the ability of estradiol to suppress mesangial cell collagen synthesis may result in reduced accumulation of collagen after glomerular injury and thereby limit the development of glomerulosclerosis (18). TGF-␤, on the other hand, plays an important role in mediating progressive renal injury (19). Accumulation of collagen in the glomerular mesan-gium and renal interstitium contributes to the adverse effects of TGF-␤ on the progression of chronic renal disease (19), which is thought to be mediated by stimulation of collagen synthesis in mesangial cells mediated by TGF-␤ (20,21). It was shown previously that estrogens inhibit TGF-␤-induced collagen synthesis in renal mesangial cells (22,23), but how this is achieved is unclear.
In this study, we demonstrate a novel molecular mechanism by which estrogens inhibit TGF-␤ function: there are direct physical and functional interactions between Smad3 and ER␣. These findings provide insights into the cross-regulation between the estrogen and TGF-␤ signaling which may have implications for atherosclerosis and glomerulosclerosis as well as other pathological conditions.
Cell Culture, Transfection, and Luciferase Assays-Human embryonic kidney carcinoma cell line 293T was maintained in DMEM containing 10% FCS and transfected in DMEM containing 1% FCS by the standard calcium precipitation protocol. Human renal mesangial cells were obtained from Clonetics (East Rutherford, NJ) and cultured in MsGM (Clonetics) containing 5% FCS according to the manufacturer's instructions. Before stimulation, cells were cultured for 12 h in MsGM containing 1% FCS followed by treatment with TGF-␤ and/or E 2 . Human breast cancer cell line MCF-7 was a kind gift from the Cell Resource Center for Biomedical Research (Tohoku University, Sendai, Japan) maintained in DMEM containing 10% FCS (30). Before stimulation, the cells were cultured for 24 h in DMEM containing 1% FCS followed by treatment with TGF-␤ and/or E 2 (30,31). MCF-7 cells (2-2.5 ϫ 10 5 in a 6-cm dish) were transfected by using LipofectAMINE PLUS (Life Technologies, Inc., Carlsbad, CA) following the manufacturer's instructions. Luciferase assay was performed as described (29). The cells were harvested 48 h after transfection and lysed in 100 l of PicaGene Reporter Lysis Buffer (Toyo Ink, Tokyo, Japan) and assayed for luciferase and ␤-galactosidase activities according to the manufacturer's instructions. Luciferase activities were normalized to the ␤-galactosidase activities. Three or more independent experiments were carried out for each assay.
Immunoprecipitation and Immunoblotting-The immunoprecipitation and Western blotting assays were performed as described previously (29). Cells were harvested and lysed in lysis buffer (50 mM Tris-HCl, pH 7.4, 0.15 M NaCl, containing 0.5% Nonidet P-40, 1 M sodium orthovanadate, 1 M phenylmethylsulfonyl fluoride, and 10 g/ml each of aprotinin, pepstatin, and leupeptin). The immunoprecipitates from cell lysates were resolved on 5-20% SDS-polyacrylamide gel electrophoresis and transferred to Immobilon filter (Millipore; Bedford, MA). The filters were then immunoblotted with each antibody. Immunoreactive proteins were visualized using an enhanced chemiluminescence detection system (Amersham Pharmacia Biotech).
Northern Blot Analysis-293T cells were maintained as described above. Human renal mesangial cells were maintained according to the manufacturer's instructions. After 12 h of incubation in 1% FCS, human renal mesangial cells were treated with 30 units/ml TGF-␤ and/or 10 Ϫ8 M E 2 for 24 h. Total RNAs were prepared by using Iso-Gen (Nippon Gene, Tokyo) and used in Northern analysis according to the established procedures. A nylon membrane (Hybond N ϩ , Amersham Pharmacia Biotech) and radiolabeled cDNA probes were used, where indicated.

Estrogens Inhibit TGF-␤ Signaling in 293T
Cells-To examine the molecular basis of the cross-talk between TGF-␤ and ER signaling pathways, we performed the transient transfection assay using human embryonic kidney carcinoma cell line 293T. The TGF-␤-mediated transcriptional responses were measured by p3TP-LUC, which is one of the standard reporters for assessing TGF-␤ activity (25). In these experiments, a constitutively active form of TGF-␤ type I receptor, T␤R-I(T204D) (25), was used which stimulated p3TP-LUC more effectively than TGF-␤ plus wild-type T␤R-I (31).
When 293T cells were transfected with p3TP-LUC together with an expression vector for T␤R-I(T204D), LUC expression was increased by 5-6-fold (Fig. 1A). We then examined the effect of estrogens on TGF-␤ signaling in this model system. 293T cells were transfected with an expression vector for ER␣, T␤R-I(T204D), and p3TP-LUC and were either left untreated or stimulated with E 2 . As shown in Fig. 1A, E 2 suppressed T␤R-I(T204D)-induced p3TP-LUC activity in a dose-dependent manner. These results indicate that the inhibitory effects of E 2 on T␤R-I(T204D)-induced transcriptional activity can be reconstituted in 293T cells similar to those observed in kidney cells, such as mesangial cells.
We then assessed the reverse situation for the possible effect of TGF-␤ signaling on ER activity using the reporter Vit-LUC gene in which two copies of an estrogen response element drive expression of the LUC gene. In the presence of ER␣, E 2 treatment resulted in a 50-fold increase in Vit-LUC activity (Fig.  1A). Surprisingly, this activation was augmented by T␤R-I(T204D) expression in a dose-dependent manner, although T␤R-I(T204D) alone did not affect basal reporter activity. These results suggest that in contrast with the inhibitory effects of ER␣ on TGF-␤ signaling, activation of the TGF-␤ pathway has stimulatory roles for estrogen signaling.
To examine the cross-talk between TGF-␤ and ER signaling in greater detail, we used a constitutively active form of ER␣, ER␣-L536P (26). 293T cells were transfected with p3TP-LUC, an expression vector for ER␣-L536P and/or increasing amounts of an expression vector for T␤R-I(T204D), and LUC activities were measured. As shown in Fig. 1B, T␤R-I(T204D)-induced p3TP-LUC activity was inhibited by ER␣-L536P in a dose-dependent manner. In contrast, expression of T␤R-I(T204D) resulted in enhancement of ER␣-L536P-induced Vit-LUC activation. These results support the data in Fig. 1A and clearly show the existence of cross-talk between TGF-␤ and ER signaling in 293T cells.
To assess the specificity of the estrogen effects on TGF-␤ signaling, we utilized the anti-estrogen tamoxifen, which inhibits E 2 binding to ER␣. p3TP-LUC and T␤R-I(T204D) were transfected into 293T cells with an expression vector for ER␣, and the cells were either left untreated or treated with E 2 in the presence or absence of tamoxifen. The inhibitory effect of ER␣ in the presence of E 2 or ER␣-L536P on TGF-␤ signaling was reversed by tamoxifen (Fig. 1C). These results indicate that the inhibitory effect of E 2 on TGF-␤ signaling in 293T cells is mediated directly by ER␣.
We next examined if the other major ER isoform, ER␤ (11,12), has similar inhibitory effects on TGF-␤ signaling in an analogous experiment. As shown in Fig. 1D, T␤R-I(T204D)induced p3TP-LUC activity was inhibited by ER␤ in the presence of E 2 , similar to that observed with ER␣, and this inhibitory effect was reversed by tamoxifen. In contrast, ER␤induced Vit-LUC activation was augmented by T␤R-I(T204D), similar to that observed for ER␣. These data suggest that both ER isotypes may be involved in the interference of estrogen signaling with the TGF-␤ pathway.
We then examined the cross-talk between TGF-␤ signaling and retinoic acid, 1␣,25-dihydroxy vitamin D 3 (VDR), glucocorticoid, and androgen receptors in 293T cells because recent studies demonstrated that some members of the steroid receptor family interacted with TGF-␤ signaling via Smads (31)(32)(33)(34). However, none of these nuclear receptors had an inhibitory effect on Smad3 activation by T␤R-I(T204D) in 293T cells in the presence of their respective ligands (data not shown). These data suggest that inhibition of TGF-␤ signaling in 293T cells is not a general phenomenon of nuclear receptors but rather is highly specific for ER.
Smad3 Mediates the Interaction of Estrogen and TGF-␤ Signaling Pathways-We next examined whether the cross-talk between TGF-␤ and ER signaling is mediated by the Smads, downstream signal transducers of the TGF-␤ superfamily (4). When Smad3 was expressed in 293T cells, p3TP-LUC activity was induced in a dose-dependent fashion, and coexpression of ER␣-L536P inhibited this activation ( Fig. 2A). Conversely, ER␣-L536P-induced Vit-LUC activation was augmented by the  expression of Smad3 ( Fig. 2A). Moreover, enhancement of both ER␣-and ER␣-L536P-induced Vit-LUC activity by T␤R-I(T204D) was inhibited by the expression of a dominant negative form of Smad3, Smad3DE (24) (Fig. 2B), or an inhibitory Smad, Smad7 (Fig. 2C). These data suggest that the cross-talk between TGF-␤ and ER in 293T cells is mediated by Smad3.
Physical Interactions between ER and Smads in Vivo-One of the possible mechanisms that is consistent with the data described above is that there are direct physical interactions between ER␣ and Smad3. We tested this possibility by coimmunoprecipitation experiments. 293T cells were transfected with expression vectors encoding ER␣-L536P or wild-type ER␣ together with FLAG-tagged Smad3 and T␤R-I(T204D), left untreated or treated with E 2 , and the cells were lysed and subjected to immunoprecipitation with an anti-FLAG antibody. Immunoprecipitates were then used in Western analysis with an antiserum against ER␣. As shown in Fig. 3A, ER␣-L536P and Smad3 were found to be in a complex in 293T cells. Consistent with the fact that ER␣ inhibits TGF-␤ signaling only in the presence of E 2 , ER␣-Smad3 interactions were only detected in E 2 -treated cells (Fig. 3A).
We next tested whether TGF-␤ affects ER␣-Smad3 interactions. 293T cells were transfected with ER␣-L536P together with FLAG-tagged Smad3 with or without T␤R-I(T204D), and immunoprecipitation and Western analysis were carried out as above. As shown in Fig. 3B, ER␣-L536P interacted with Smad3 only in the presence of T␤R-I(T204D), suggesting that stimulation of the TGF-␤ signaling pathway is a prerequisite for ER␣-Smad3 interactions.
We also tested whether ERs interacted with any of the other Smads. 293T cells were transfected with ER␣-L536P or ER␤ together with FLAG-tagged Smad2, Smad3, or Smad4 in the presence of T␤R-I(T204D). As shown in Fig. 3C, both ER␣ and ER␤ interacted with all three Smads tested; however, interaction with Smad2 appeared to be weaker compared with that observed with Smad3 or Smad4. These results suggest that multiple Smad members may be involved in the cross-talk between ER and TGF-␤ signaling.
MH2 Domain of Smad3 Is Required for Efficient Physical and Functional Interactions with ER␣-We next determined the domains of Smad3 which mediate interactions with ER␣, using either N-or C-terminal deletion mutants of Smad3 (24) (Fig. 4A). Expression vectors encoding ER␣-L536P and/or FLAG-tagged full-length Smad3 or one of its four deletion mutants were transiently transfected into 293T cells, together with T␤R-I(T204D). Cells were lysed and subjected to immunoprecipitation with an anti-ER␣ antibody. Immunoprecipitates were then used in Western blot analysis with an anti-FLAG antibody. As shown in Fig. 4B, whereas the full-length Smad3 interacted with ER␣-L536P, the C-terminal deletion mutants lacking the MH2 or LϩMH2 domains were unable to bind ER␣. In contrast, the N-terminal mutants in which MH1 or MH1ϩL domains are deleted retained interactions with ER␣. These results indicate that efficient ER␣-Smad3 interactions require the MH2 domain of Smad3.
We then tested the Smad3 mutants used in the above analysis for their ability to activate the p3TP-LUC reporter. As shown in Fig. 4C, whereas the full-length Smad3 and the N-terminal deletion mutants increased T␤R-I(T204D)-induced p3TP-LUC activation, the C-terminal mutants in which the MH2 is deleted were inactive. These results, consistent with the physical interaction data presented above and previous findings on TGF-␤ signaling (9,35,36), suggest that the MH2 domain is most critical for Smad3 activity.
We then determined the domain(s) of Smad3 which is involved in the inhibition of TGF-␤ signaling by ER␣. As shown in Fig. 4C, expression of ER␣-L536P significantly reduced p3TP-LUC activation. Expression of full-length Smad3, or its N-terminal deletion mutants, but not the C-terminal deletion mutants lacking the MH2 domain, largely reversed the ER␣mediated inhibition of p3TP-LUC expression. These results suggest that the MH2 domain of Smad3 mediates the inhibition of TGF-␤ signaling by ER␣.
It was also of interest to determine whether the stimulatory action of Smad3 on ER␣ is mediated by the MH2 domain. To that end, the Vit-LUC reporter was transfected into 293T cells with T␤R-I(T204D) and ER␣ with or without Smad3 or its deletion mutants. As shown in Fig. 4D, ER␣-mediated activation of Vit-LUC was augmented by T␤R-I(T204D) ϳ6-fold in the presence of full-length Smad3. Similar levels of activation were observed by T␤R-I(T204D) and the N-terminal mutants of  ). B, 1 ϫ 10 7 cells were transfected with 7.5 g of ER␣-L536P and/or 7.5 g of FLAG-tagged Smad3 in the presence or absence of 2 g of hemagglutinin-tagged T␤R-I(T204D). Cell lysates were then immunoprecipitated with an anti-ER␣ antibody and immunoblotted either with anti-FLAG antibody (top panel) or anti-ER␣ antibody (upper middle panel). Total cell lysates (1%) were blotted with anti-FLAG antibody or anti-hemagglutinin antibody as indicated (lower middle and bottom panels, respectively). C, 1 ϫ 10 7 cells were transfected with 7.5 g of ER␣-L536P (lanes 1-4) or 7.5 g of ER␤ (lanes 5-8) and/or 7.5 g of FLAG-tagged Smad2, Smad3, and Smad4 together with 2 g of T␤R-I(T204D). Cell lysates were then immunoprecipitated with anti-ER␣ or anti-ER␤ antibody and immunoblotted with anti-FLAG antibody (top panels) or anti-ER␣/anti-ER␤ antibody (middle panels). Total cell lysates (1%) were blotted with anti-FLAG antibody as indicated (bottom panels).
Smad3 but not the C-terminal mutants. Consistent with the in vivo interaction data presented above, these results suggest that the MH2 domain of Smad3 is required for the cross-talk between ER␣ and TGF-␤ signaling in both directions.
Estrogens Inhibit TGF-␤-induced Smad3 Activation in MCF-7 and Renal Mesangial Cells-To examine the cross-talk between TGF-␤ and estrogen signaling pathways under more physiological conditions through endogenous proteins, we first utilized a TGF-␤-responsive, ER-positive breast cancer cell line, MCF-7 (30), and the transient transfection assay. MCF-7 cells were transfected with p3TP-LUC and treated with TGF-␤ and/or E 2 , and LUC activities were determined. As shown in Fig. 5A, TGF-␤ stimulated p3TP-LUC activity, whereas E 2 alone did not have an affect. When cells were treated with both TGF-␤ and E 2 , p3TP-LUC activation was decreased by 40 -50% compared with the activation by TGF-␤ alone. In contrast, when MCF-7 cells were transfected with Vit-LUC and treated with TGF-␤ and/or E 2 , TGF-␤ further stimulated E 2 -induced Vit-LUC activity by 30 -40%, similar to that observed in 293T cells (Fig. 1).
In parallel with these transfection studies, coimmunoprecipitation experiments were performed using cell lysates obtained from MCF-7 cells that were either left untreated or were treated with TGF-␤ and E 2 . Similar to the results obtained in transfected 293T cells (Fig. 3), we found that ER␣ that was immunoprecipitated from MCF-7 cells was in a complex with Smad3, and this interaction was dependent on the presence of TGF-␤ and E 2 (Fig. 5B).
We then examined whether estrogens have similar effects on TGF-␤-induced transcriptional activation of endogenous genes. To that end, we first used 293T cells that were transfected with T␤R-I/Smad3 to activate the TGF-␤ pathway. The steady-state mRNA levels of a TGF-␤ target gene, ␣2(I)-collagen (COL1A2), which codes for a major structural component of the extracellular matrix (37), was determined by Northern analysis in the presence or absence of the constitutively active ER form ER␣-P563L.
As shown in Fig. 5C, expression of T␤R-I/Smad3 in 293T cells induced endogenous COL1A2 expression by ϳ3-fold, and this activation was abolished in the presence of ER␣-L536P, suggesting that ER␣ can directly target TGF-␤ regulation of endogenous genes.
We next performed a similar experiment in renal mesangial cells but without ectopic expression of any factors. Cells were either left untreated or treated with TGF-␤ and/or E 2 , and COL1A2 expression was monitored by Northern analysis. As shown in Fig. 5D, TGF-␤ treatment induced COL1A2 expression by ϳ3.5-fold, and this activation was completely abolished in the presence of E 2 (Fig. 5C). These data show that E 2 inhibits TGF-␤-target gene transcription in renal mesangial cells. DISCUSSION Perturbations in the TGF-␤ signaling have been associated with a variety of clinical disorders including some cancers, renal disease, and vascular disease (3,19,38,39). Although mutations in various components of the TGF-␤ pathway may account for some of these abnormalities, it has been shown that TGF-␤ signaling can be strongly affected by interactions with other molecules in the cell. Recent studies have documented the interaction of a large number of intracellular proteins with the effector molecules Smads to influence TGF-␤ signaling (40). Whereas some of these proteins have been found to cooperate functionally with and activate Smads, others were found to repress Smad activity.
One class of proteins that interact with the TGF-␤ signaling pathway is the steroid receptor family, which can influence TGF-␤ signaling both positively and negatively to impact a variety of physiological and pathological processes (22,23,41,42). Smad3 enhanced VDR transcriptional activity by its physical interaction of ligand-induced VDR and the coactivators SRC-1/TIF2 (31); however, the effect of VDRs on Smad3 activity was not assessed. In other studies, androgen receptors were found to stimulate TGF-␤ signaling via direct binding to Smad3 (33), whereas Smad3 repressed androgen receptor-mediated transcription (34). In contrast, interaction between glucocorticoid receptors and Smad3 suppressed TGF-␤ signaling in hepatoma cells, but Smad3 did not affect glucocorticoid receptordependent transcriptional activation (32). Thus, there are divergent results in the interaction of individual steroid receptors and TGF-␤ signaling pathways and the molecular mechanisms that may be involved.
In this study, we have shown that ERs suppress TGF-␤ signaling by associating with, and acting as a transcriptional corepressor for, Smad3. The inhibition of Smad3 activity by ERs provides a molecular mechanism for the opposing effects of estrogens and TGF-␤ signaling in some disease states, such as renal injury, atherosclerosis, and breast cancer (22,23). Reversal of the suppression of Smad3 activation by ER␣/E 2 by the anti-estrogen tamoxifen indicated that this effect is mediated directly by ER; both ER␣ and ER␤ had this activity. These A, MCF-7 cells were grown in a 6-cm dish and transfected with 1 g of p3TP-LUC or Vit-LUC and then stimulated with 100 units/ml TGF-␤ and/or 10 Ϫ7 M E 2 as indicated. 48 h after transfection, cells were stimulated for an additional 12 h. Cells were harvested, and relative luciferase activities were measured. The results are presented as fold induction of luciferase activity from triplicate experiments, and the error bars represent the standard deviations. B, 5 ϫ 10 7 MCF-7 cells were maintained in DMEM containing 1% FCS for 12 h before stimulation. After 1 h of stimulation with or without 30 units/ml TGF-␤ and 10 Ϫ8 M E 2 , cells were lysed, immunoprecipitated (IP), and immunoblotted (Blot) with control IgG, anti-ER␣, or anti-Smad3 antibody as indicated. Total cell lysates (1%) were blotted with anti-ER␣ or anti-Smad3 antibody. C, effect of ER␣-L536P on COL1A2 expression in the T␤R-I/Smad3-transfected 293T cells. 293T cells were transfected with 10 g of Smad3 and 5 g of T␤R-I(T204D) and/or 7.5 g of ER␣-L536P as indicated. 20 g of total RNA isolated from these cells was used in Northern blot analysis of COL1A2 expression. The same blot was probed with glyceraldehyde-3-phosphate dehydrogenase (G3PDH) cDNA as control (lower panel). Relative intensities (Rel. Int.) of the bands shown below the autoradiograms were determined by densitometric analysis. D, human renal mesangial cells (HRMC) were either left untreated or treated with 30 units/ml TGF-␤ and/or 10 Ϫ8 M E 2 . COL1A2 expression was monitored by Northern blot analysis of 15 g of total RNA for each treatment. The same blot was probed with glyceraldehyde-3-phosphate dehydrogenase cDNA as control (lower panel). Relative intensities of the bands shown below the autoradiograms were determined by densitometric analysis. findings are consistent with the reported therapeutic effects of anti-estrogens, such as tamoxifen, through local boosting of TGF-␤ signaling (43).
Conversely, ER transcriptional activity was increased by activation of the TGF-␤ signaling pathway. This is similar to that found for the VDR but opposite of that found for the androgen receptor (34), whereas there was no effect of TGF-␤ signaling on the glucocorticoid receptor (32). The physiological significance of TGF-␤ signaling-induced ER activity remains to be established. However, activation of ER by the TGF-␤ pathway can establish a feedback loop where estrogen signaling would be accentuated by the TGF-␤ signaling itself which in turn would be inhibited more quickly and effectively. Upon inhibition of the TGF-␤ signaling, ER activity would return to normal levels again. Further work will be required to assess the possible significance of this two-way cross-talk and the feedback loop that may be established in the pathophysiology of disease states.
Several additional molecular mechanisms have been proposed for the inactivation of Smad/TGF-␤ signaling pathway. For example, TGF-␤ signaling is inhibited by interferon-␥ (44) and tumor necrosis factor-␣ (45), which induce the expression of an inhibitory Smad, Smad7. In addition, the zinc finger protein Evi-1 interacts with Smad3 and represses its DNA binding activity (46), whereas the nuclear Ski and SnoN oncoproteins have been reported to inhibit TGF-␤ signaling by recruitment of the transcriptional repressor N-CoR to TGF-␤responsive promoters through interaction with Smad proteins (47)(48)(49). Smad2/3 interacts with Ski through its C-terminal MH2 domain in a TGF-␤-dependent manner, similar to ER␣. Further studies will determine whether any of these mechanisms also plays some role in the ER-mediated repression TGF-␤ signaling.
In this study, we demonstrated that Smads are important regulators of ER function and thereby may have critical roles in the progression of diseases, such as breast cancer or kidney disease. More detailed understanding of the cross-talk between Smads and ER is therefore important because this new information may provide new therapeutic approaches for these and other pathological conditions.