Smad2 Protects against TGF-β/Smad3-Mediated Renal Fibrosis

Abstract

Smad2 and Smad3 interact and mediate TGF-β signaling. Although Smad3 promotes fibrosis, the role of Smad2 in fibrogenesis is largely unknown. In this study, conditional deletion of Smad2 from the kidney tubular epithelial cells markedly enhanced fibrosis in response to unilateral ureteral obstruction. In vitro, Smad2 knockdown in tubular epithelial cells increased expression of collagen I, collagen III, and TIMP-1 and decreased expression of the matrix-degrading enzyme MMP-2 in response to TGF-β1 compared with similarly treated wild-type cells. We obtained similar results in Smad2-knockout fibroblasts. Mechanistically, Smad2 deletion promoted fibrosis through enhanced TGF-β/Smad3 signaling, evidenced by greater Smad3 phosphorylation, nuclear translocation, promoter activity, and binding of Smad3 to a collagen promoter (COL1A2). Moreover, deletion of Smad2 increased autoinduction of TGF-β1. Conversely, overexpression of Smad2 attenuated TGF-β1–induced Smad3 phosphorylation and collagen I matrix expression in tubular epithelial cells. In conclusion, in contrast to Smad3, Smad2 protects against TGF-β–mediated fibrosis by counteracting TGF-β/Smad3 signaling.

TGF-β/Smad signaling has been shown to play a critical role in renal fibrosis.^1–4 It is now clear that TGF-β1 signals through the heteromeric complex of TGF-β type I receptor and TGF-β type II receptor to activate two key downstream mediators, Smad2 and Smad3, to exert its biological activities such as cell growth, differentiation, extracellular matrix (ECM) production, and apoptosis.⁵ Although it is known that Smad2 and Smad3 physically interact and are structurally similar with >90% homology in their amino acid sequences,⁶ the distinct functions of these two genes in embryonic development has been noted. Genetic deletion of Smad2 in mice results in embryonic lethality at an early stage of development, whereas mice null for Smad3 survive with impaired immunity.^7,8 Although the functional role of Smad3 in cell growth, differentiation, apoptosis, tissue repair and fibrosis, and immune responses has been well studied,^8–10 the pathophysiologic role of Smad2 in these processes remains largely unclear. This may be attributed to the unavailability of Smad2 knockout (KO) mice for such studies as a result of the early embryonic lethality.

In the context of tissue repair and fibrosis, Smad3 has been studied extensively, but little attention has been paid to the functional role of Smad2. It is now well accepted that Smad3 is a key mediator of TGF-β signaling in ECM production and tissue fibrosis. This may be associated with the finding that Smad3-binding elements are found in most collagen promoters^11–13; therefore, TGF-β1–induced collagen matrix expression is Smad3 dependent. Emerging evidence has shown that Smad3 plays an important role in tissue repair and fibrosis including wound healing,¹⁴ epithelial-to-mesenchymal transition (EMT),¹⁵ and tissue scar formation under various disease conditions in skin,¹³ lung,¹⁶ heart,¹⁷ kidney,¹⁵ and liver¹⁸; however, the functional importance of Smad2 and the interaction between Smad2 and 3 in the profibrotic response to TGF-β1 remain largely unclear. Thus, this study investigated the functional role of Smad2 in ECM production and renal fibrosis in vivo in a mouse model of unilateral ureteral obstruction (UUO) and in vitro in tubular epithelial cells (TECs) with knockdown or overexpression of Smad2 and in mouse embryonic fibroblasts (MEFs) lacking Smad2. The mechanisms of Smad2 in regulating fibrosis in response to TGF-β1 were explored.

Results

Generation and Characterization of Conditional Smad2 KO Mice, Smad2 KO MEFs, and Stable Smad2 Knockdown TECs

Because mice with constitutive deletion of Smad2 are embryonic lethal,⁷ we generated conditional Smad2 KO mice (Smad2ff/KspCre) by crossing the Smad2 flox/flox mouse (Smad2ff) with the kidney-specific (Cadherin-16 promoter) Cre mouse (KspCre) specifically to delete Smad2 from kidney TECs. As shown in Figure 1, PCR detected that transgenic expression of Cre recombinase (235 bp) in the Smad2ff mouse resulted in a substantial deletion of the floxed Smad2 gene (451 bp) by the Cre recombination as identified by the deleted allele (592 bp). This observation was further confirmed at the mRNA level by real-time PCR that up to 85% of Smad2 mRNA was deleted from both normal and diseased kidneys of conditional Smad2 KO mice when compared with the Smad2ff mice (Figure 1B). Interestingly, conditional deletion of Smad2 from the kidney also largely prevented a marked increase in Smad2 mRNA expression in the UUO kidney as detected in Smad2ff mice (Figure 1B). Similar findings were also demonstrated at the protein level by Western blot analysis (Figure 1C). Immunohistochemically, Smad2 was highly expressed in all kidney cell types in the Smad2ff mouse but absent from most TECs in Smad2ff/KspCre mice, despite that high levels of Smad2 remained in glomerular and vascular cells (Figure 1D) as a result of active KspCre in kidney TECs only. Confirming Smad2 KO MEFs and Smad2 knockdown TEC, Western blot analysis showed that there was no Smad2 protein detectable in Smad2 KO MEFs (Figure 1E) and TECs (NRK52E) that stably express Smad2 Small Interfering RNA exhibited a partial deletion of Smad2 (Figure 1F).

Figure 1.

Characterization of conditional Smad2 KO mice, Smad2 KO MEFs, and Smad2 knockdown TECs. (A) PCR detects that the Cre recombination (235 bp) results in the Smad2 floxed gene (451 bp) being deleted from the kidney (592 bp) in the conditional Smad2 KO mice (Smad2ff/KspCre). (B and C) Real-time PCR and Western blot analysis show a substantial deletion of Smad2 mRNA and protein from the kidney in S2ff/CreKsp mice under normal and diseased conditions. (D) Immunohistochemistry reveals that Smad2 (brown) is highly expressed by all kidney cell types in Smad2ff mice but is specifically deleted from most of kidney TECs. Note that expression of Smad2 in glomeruli (g) and vascular cells remains high in a Smad2ff/KspCre mouse. (E) Western blot analysis shows a deletion of Smad2 expression in Smad2 KO MEFs. (F) Western blot analysis shows a Smad2 being knocked down from TECs (NRK52E). Results represent groups of eight mice (A through D). Real-time PCR data are expressed as means ± SEM for groups of eight mice. ***P < 0.001 versus normal; ^###P < 0.001 versus diseased Smad2ff (UUO) mice. Magnification, ×400 in D.

Deletion of Smad2 Enhances Renal Fibrosis in a Mouse Model of UUO and In Vitro in TECs and in MEFs

We first examined the functional importance of Smad2 in progressive renal fibrosis in a mouse model of UUO induced in conditional Smad2 KO mice (Smad2ff/KspCre) in which Smad2 was specifically deleted from TECs by the Cre/LoxP technology (Figure 1, A through D). Unexpectedly, histology stained with Masson trichrome detected development of moderate to severe tubulointerstitial fibrosis in Smad2ff mice at 10 days after UUO; this was further increased in the UUO kidney of conditional Smad2 KO mice (Figure 2A). Immunohistochemistry, real-time PCR, and Western blot analysis also revealed that progressive tubulointerstitial fibrosis with accumulation of collagens I and III and a corresponding increase in mRNA expression in the UUO kidney of Smad2ff mice was largely enhanced in conditional Smad2 KO mice, resulting in a two- to threefold increase in collagen I and III mRNA and a 1.5-fold increase in protein expression in Smad2ff/KspCre mice when compared with the Smad2ff mice with UUO (Figures 2, B through D, and 3).

Figure 2.

Disruption of Smad2 from the kidney promotes tubulointerstitial fibrosis and collagen I expression in a mouse model of UUO at day 10. (A) Masson's trichrome staining paraffin sections and quantitative analysis. Note that compared with the UUO kidney from a littermate Smad2 flox/flox mouse (S2ff), interstitial collagen matrix accumulation (green) is largely increased in the UUO kidney of conditional Smad2 KO mice (S2ff/KspCre). (B) Collagen I immunohistochemical staining. (C) Collagen I mRNA expression by real-time PCR analysis. (D) Collagen I expression by Western blot analysis. Each lane represents one mouse UUO kidney. Data are means ± SEM for groups of eight mice. *P < 0.05, ***P < 0.001 versus normal; ^###P < 0.001 versus injured Smad2ff (UUO) mice. Magnifications: ×100 in A; ×200 in B.

Figure 3.

Disruption of Smad2 from the kidney promotes tubulointerstitial fibrosis as identified by collagen III expression in a mouse model of UUO at day 10. (A) Immunohistochemistry. (B) Real-time PCR. (C) Western blot. Data are means ± SEM for groups of eight mice. *P < 0.05, ***P < 0.001 versus normal; ^###P < 0.001 versus injured Smad2ff (UUO) mice. S2ff, Smad2 flox/flox mouse; S2ff/KspCre, conditional Smad2 KO mice, Col., collagen. Magnification, ×200 in A.

To confirm the results found in the UUO kidney in conditional Smad2 KO mice, we examined collagens I and III expression in Smad2 wild-type (WT) and KO MEFs and in TECs with knockdown for Smad2 in vitro. Real-time PCR and Western blot analysis showed that addition of TGF-β1 (2 ng/ml) stimulated both collagens I and III mRNA and protein expression in Smad2 WT MEFs in a time- and dosage-dependent manner (Figures 4, A and B, and 5). Loss of Smad2 promoted further TGF-β1–induced collagen matrix expression, resulting in a two- to threefold increase in collagens I and III mRNA expression when compared with Smad2 WT MEFs (Figure 4, A and B). This observation was also detected at the protein level by Western blot analysis (Figure 5). Similarly, knockdown of Smad2 in TECs (NRK52E) significantly enhanced TGF-β1–induced (2 ng/ml) collagens I and III mRNA expression (Figure 4C).

Figure 4.

Real-time PCR reveals that disruption of Smad2 in MEFs or knockdown of Smad2 from TECs enhances TGF-β1–induced collagens I and III mRNA expression. (A) Deletion of Smad2 enhances collagen I and collagen III mRNA in a time-dependent manner in response to TGF-β1 (2 ng/ml). (B) Deletion of Smad2 enhances TGF-β1–induced (2 ng/ml) collagen I and collagen III mRNA at a peak time (3 hours) in a dosage-dependent manner. (C) Knockdown of Smad2 from TECs promotes collagen I and collagen III mRNA expression induced by TGF-β1 (2 ng/ml) at 3 hours. Data are means ± SEM for four independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 versus time or dosage 0; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus Smad2 WT MEFs (A and B) or Smad2 knockdown TECs (C). VC, empty vector control; S2Kd, Smad2 knockdown.

Figure 5.

Western blot analysis shows that deletion of Smad2 in MEFs enhances TGF-β1–induced collagens I and III protein expression in a time- and dosage-dependent manner. (A) TGF-β1 (2 ng/ml) induces collagen I (Col. I) and collagen III (Col. III) protein expression in a time-dependent manner. (B) TGF-β1 induces collagen I (Col. I) and collagen III (Col. III) protein expression at 24 hours in a dosage-dependent manner. Data are means ± SEM for four independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 versus time or dosage 0; ^#P < 0.05, ##P < 0.01, ^###P < 0.001 versus Smad2 WT MEFs.

Deletion of Smad2 Prevents Collagen Matrix Protein from Degradation In Vivo and In Vitro

After determining the protective role of Smad2 in collagen matrix synthesis in vivo and in vitro, we investigated whether Smad2 has a role in collagen matrix protein degradation. In UUO kidneys, tubulointerstitial fibrosis in Smad2ff mice was associated with a marked upregulation of tissue inhibitor of metalloproteinase 1 (TIMP-1), although matrix metalloproteinase 2 (MMP-2) mRNA was also increased (Figure 6A); however, conditional deletion of Smad2 from the kidney resulted in a further increase in TIMP-1 expression while suppressing expression of MMP-2 mRNA (Figure 6A). A similar result was also found in vitro that knockdown of Smad2 in TECs significantly attenuated TGF-β1–induced MMP-2 expression but enhanced further upregulation of TIMP-1 mRNA expression (Figure 6B).

Figure 6.

Real-time PCR detects that disruption of Smad2 impairs collagen matrix degradation in a mouse model of UUO at day 10 and in vitro. (A) Real-time PCR shows that compared with the Smad2ff mice, conditional deletion of Smad2 from the kidney results in further inhibition of MMP-2 but significantly increases TIMP-1 expression. (B) In vitro, knockdown of Smad2 from TECs (NRK52E) results in a significant inhibition of TGF-β1–induced (2 ng/ml) MMP-2 while further increasing TIMP-1 mRNA expression at 3 hours after TGF-β1 stimulation. (C) Real-time PCR reveals that compared with the Smad2 WT MEFs, MEFs lacking Smad2 show a further suppression of MMP-2 while upregulating TIMP-1 mRNA expression in response to TGF-β1 in a time-dependent (3 hours) and a dosage-dependent (2 ng/ml) manner. Data are means ± SEM for four independent experiments in vitro and groups of eight mice in vivo. *P < 0.05, **P < 0.01, ***P < 0.001 versus time (dosage) 0 or normal; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus Smad2WT MEFs or injured Smad2ff (UUO) mice.

To confirm these findings seen in the conditional Smad2 KO kidney and knockdown TEC, we performed studies in Smad2 KO MEFs. As shown in Figure 6C, in Smad2 WT MEFs, addition of TGF-β1 inhibited collagen matrix degradation as evidenced by decreased MMP-2 and upregulation of TIMP-1 mRNA expression in a time- and a dosage-dependent manner. Strikingly, deletion of Smad2 resulted in a further inhibition of collagen matrix degradation in response to TGF-β1 by a further decrease in MMP-2 while enhancing TIMP-1 mRNA expression when compared with Smad2 WT MEFs (Figure 6C). Taken together, loss of Smad2 promoted renal fibrosis by enhancing the inhibitory effect of TGF-β1 on collagen matrix protein degradation in vivo and in vitro.

Enhanced TGF-β/Smad3 Signaling Is a Mechanism by Which Disruption of Smad2 Promotes Collagen Matrix Expression and Tissue Fibrosis In Vivo and In Vitro

We further examined in vivo and in vitro the mechanisms whereby deletion of Smad2 enhances collagen matrix expression. Because it is well established that TGF-β1 activates Smad3 to mediate fibrosis,¹⁹ we determined whether deletion of Smad2 promotes collagen matrix production by enhancing Smad3 signaling. As shown in Figure 7A, compared with the UUO kidney from Smad2 ff mice, conditional deletion of Smad2 from the kidney largely enhanced Smad3 signaling as identified by a marked phosphorylated Smad3 (phospho-Smad3) nuclear location in the fibrotic tissue of tubulointerstitium after UUO. Similarly, Western blot analysis revealed that higher levels of phospho-Smad3 were detected in the UUO kidney of conditional Smad2 KO mice compared with the Smad2ff mice (Figure 7B).

Figure 7.

Deletion of Smad2 enhances Smad3 phosphorylation in the UUO kidney and in vitro in response to TGF-β1. (A and B) Immunohistochemistry and Western blot analysis show that compared with the UUO kidney of Smad2ff mice, numbers of nucleated phospho-Smad3–positive cells and phospho-Smad3 protein are markedly increased in the UUO kidney of conditional Smad2 KO mice. (C) Knockdown of Smad2 from TECs (NRK52E) results in a significant increase in phospho-Smad3 at 30 minutes (peak time) after TGF-β1 stimulation. (D) Western blot analysis shows that compared with Smad2 WT MEFs, addition of TGF-β1 (2 ng/ml) largely enhances Smad3 phosphorylation in Smad2 KO MEFs in a time-dependent manner when compared with the Smad2 WT MEFs. Data are means ± SEM for groups of eight mice in vivo and four independent experiments in vitro. *P < 0.05, **P < 0.01, ***P < 0.001 versus time (dosage) 0 or normal; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus Smad2WT MEFs or injured Smad2ff (UUO) mice.

The finding that conditional deletion of Smad2 sustained Smad3 signaling in the UUO kidney was further examined in vitro in a stable Smad2 knockdown TEC line (NRK5sE). Western blot analysis showed that knockdown of Smad2 from TECs with small interfering RNA significantly enhanced TGF-β1–induced phosphorylation of Smad3 without alteration of total Smad3 protein (Figure 7C). This finding was further confirmed in Smad2 KO MEFs. Indeed, addition of TGF-β1 was able to activate Smad3 significantly by phosphorylation in Smad2 WT MEFs in a time-dependent manner, being significant at 15 minutes and peaking over 30 to 60 minutes, which was substantially enhanced in Smad2 KO MEFs, resulting in up to a sixfold increase in phospho-Smad3 (Figure 7D). Again, no changes in levels of total Smad3 protein were noted in MEFs null for Smad2 (Figure 7D).

Enhanced Smad3 signaling in response to TGF-β1 in Smad2 KO MEFs was further examined by phospho-Smad3 nuclear translocation, Smad3-responsive promoter assay, and chromatin immunoprecipitation (ChIP) assay. As shown in Figure 8, A and B, immunofluorescence revealed that disruption of Smad2 enhanced TGF-β1–induced phospho-Smad3 nuclear translocation. Promoter assays also demonstrated that MEFs lacking Smad2 enhanced largely TGF-β1–induced Smad3-dependent promoter activity, resulting in a fourfold increase in p (CAGA)₁₂-luciferase activities when compared with the Smad2 WT MEFs (Figure 8C). Moreover, ChIP assay revealed that Smad3 was capable of binding to the collagen I promoter (COL1A2), and deletion of Smad2 largely increased the binding of Smad3 to the collagen I promoter in response to TGF-β (Figure 8, D and E).

Figure 8.

Deletion of Smad2 enhances Smad3 signaling in MEFs. (A) Immunofluorescence detects that MEFs lacking Smad2 substantially enhance phosphorylated Smad3 nuclear translocation at 30 minutes after TGF-β1 (2 ng/ml) stimulation when compared with the Smad2 WT MEFs. (B) Quantitative analysis of phospho-Smad3 nuclear translocation in response to TGF-β1 (2 ng/ml). (C) Smad3 promoter activity assay. (D) ChIP assay for binding of Smad3 to the COL1A2 promoter. (E) Quantitative real-time PCR analysis of the binding of Smad3 to COL1A2. Data are means ± SEM for four independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 versus time 0 or empty vector control; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus Smad2 WT MEFs.

Deletion of Smad2 Enhances the Autoinduction of TGF-β1 In Vitro and In Vivo

Interestingly, immunohistochemistry and real-time PCR demonstrated that conditional deletion of Smad2 from the kidney largely enhanced both TGF-β1 and CTGF expression in the fibrotic tubulointerstitium (Figure 9, A and B). To test whether increased TGF-β1 and CTGF expression is also TGF-β dependent, we stimulated TECs with Smad2 knockdown with TGF-β1 (2 ng/ml). Results showed that knockdown of Smad2 significantly enhanced TGF-β1 and CTGF mRNA expression in response to TGF-β1 stimulation (Figure 9C). This observation was further confirmed in Smad2 KO MEFs. Compared with Smad2 WT MEFs, addition of TGF-β1 stimulated much higher levels of TGF-β1 and CTGF mRNA expression in Smad2 KO MEFs in a time- and a dosage-dependent manner (Figure 9D).

Figure 9.

Disruption of Smad2 increases autoinduction of TGF-β1 and upregulation of CTGF in the diseased kidney after UUO at day 10 and in vitro. (A and B) Both immunohistochemistry and real-time PCR analyses show that compared with the UUO Smad2ff mice (iii and v), conditional Smad2 KO (Smad2ff/KspCre) significantly enhances TGF-β1 (A) and CTGF (B) expression (iv and v). (C) NRK52E TECs with Smad2 knockdown exhibit an enhancement of TGF-β1 and CTGF mRNA expression in response to TGF-β1(2 ng/ml) stimulation at 3 hours. (D) Deletion of Smad2 in MEFs enhances the mRNA levels of TGF-β1 and CTGF in response to TGF-β1 stimulation in a time-dependent (6 hours) and a dosage-dependent (2 ng/ml TGF-β1) manner. Data are means ± SEM for groups of eight mice in vivo (A and B) or for four independent experiments in vitro (C and D). *P < 0.05, **P < 0.01, ***P < 0.001 versus time (dosage) 0 or normal; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus Smad2 WT MEFs (A) or injured Smad2ff (UUO) mice. VC, empty vector control; S2Kd, Smad2 knockdown.

Overexpression of Smad2 Attenuates TGF-β1–Induced Collagen Matrix Expression and Degradation in TECs

To confirm the protective role of Smad2 in TGF-β–induced renal fibrosis, we performed studies by overexpressing Smad2 in TECs. As shown in Figure 10A, real-time PCR and Western blot analysis detected that transient transfection of TECs (NRK52E) with Smad2 WT plasmid resulted in an increase in Smad2 mRNA and protein expression. TECs overexpressing Smad2 were able to attenuate significantly TGF-β1–induced Smad3 phosphorylation at 30 minutes (Figure 10B); this was associated with a significant inhibition of TGF-β1–induced fibrosis responses including expression of profibrotic growth factors (TGF-β1 and CTGF) and collagen matrix (collagens I and III) and an increase in collagen matrix degradation by upregulating MMP-2 while suppressing TIMP-1 mRNA expression (Figure 10C).

Figure 10.

Overexpression of Smad2 attenuates the fibrotic effect of TGF-β1 in TECs. (A) Characterization of Smad2-overexpressing TECs. (B) Western blot analysis shows that compared with empty vector control (VC), overexpression of Smad2 (S2Over) inhibits TGF-β1–induced (2 ng/ml) phosphorylation of Smad3 in TECs (NRK52E). (C) Real-time PCR shows that overexpression of Smad2 in TECs attenuates the fibrosis response to TGF-β1 (2 ng/ml, 3 hours), including a significant inhibition of TGF-β1, CTGF, collagens I and III, and TIMP mRNA expression, while increasing MMP-2 expression. Data are means ± SEM for four independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 versus time 0 or empty vector control; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 versus empty vector control (VC).

Discussion

Although it is widely recognized that TGF-β1 acts by stimulating its downstream mediators Smad⅔ to exert its biological activities and Smad3 is a key mediator of TGF-β–induced ECM production and fibrosis including EMT,^19,20 the pathogenic importance of Smad2 in the context of renal fibrosis remains largely unclear. In this study, we found that there are distinct roles for Smad2 and Smad3 in ECM expression and tissue fibrosis under pathophysiologic conditions. TGF-β1 activated Smad2 to regulate Smad3 signaling negatively, thereby playing a protective role in Smad3-dependent ECM production and tissue fibrosis. This conclusion was supported by the findings that conditional deletion of Smad2 from the kidney or from TECs promoted TGF-β/Smad3 signaling and severe tubulointerstitial fibrosis in a mouse model of UUO and in vitro. These findings were further demonstrated in Smad2 KO MEFs. MEFs lacking Smad2 substantially enhanced TGF-β1/Smad3 signaling including phosphorylation, nuclear translocation, promoter activities, and the binding of Smad3 to the collagen I promoter (COL1A2), resulting in enhanced collagen production. Thus, although it is commonly believed that Smad2 and Smad3 both are key downstream mediators of TGF-β1, activation of Smad2 and Smad3 may play distinct roles in terms of collagen matrix expression and tissue fibrosis. This novel finding may also well explain the previous observations that deletion of Smad3, not Smad2, inhibits vascular and renal fibrosis in response to angiotensin II and TGF-β1^21–24 and supported the notion that deletion of Smad2 promotes TGF-β–induced EMT in both renal TECs and Smad2 KO hepatocytes.^25,26

There are at least two possible mechanisms by which Smad2 may exert its protective role in negatively regulating TGF-β–mediated ECM production and tissue fibrosis. First, although Smad2 interacts with Smad3 physically, Smad2 may function to inhibit competitively the phosphorylation of Smad3 in response to TGF-β1. This possibility was supported by the finding that deletion of Smad2 enhanced TGF-β1–induced phosphorylation of Smad3 in both a time- and dosage-dependent manner. It is now clear that after binding to receptors, TGF-β1 induces phosphorylation of both Smad2 and Smad3 in a time- and a dosage-dependent manner. It is possible that as a result of their physical interaction, Smad2 and Smad3 may compete for phosphorylation in response to TGF-β1 stimulation. Thus, loss of Smad2 results in enhanced phosphorylation of Smad3, thereby promoting TGF-β/Smad3 signaling and the fibrotic response to TGF-β1 in vitro and in vivo. Second, it is also possible that the physical interaction of Smad2 and Smad3, together with Smad4, may affect phospho-Smad3 nuclear translocation and subsequent binding to its target genes. This possibility was supported by the findings that MEFs lacking Smad2 show enhanced nuclear translocation of phospho-Smad3, increased Smad3-dependent promoter activity, and promoted the binding of Smad3 to the collagen I promoter (COL1A2) in response to TGF-β1, thereby resulting in enhanced expression of Smad3-dependent target genes such as CTGF, collagen I, and collagen III. The finding in vivo that conditional deletion of Smad2 in renal TECs enhanced Smad3 signaling and increased the severity of tubulointerstitial fibrosis also supports this hypothesis. Thus, although Smad2 and Smad3 bind together physically and work in a nonredundant manner in embryonic development, they may work in a reciprocal manner in ECM production and tissues fibrosis in response to TGF-β1 under pathophysiologic conditions.

The protective role of Smad2 in ECM production and fibrosis may also reflect the distinct roles of Smad2 and Smad3 in ECM degradation. Indeed, the balance between MMPs and their inhibitors such as TIMP-1 and TIMP-2 has been shown to play a pivotal role in ECM remodeling. The profibrotic effect of TGF-β1 on ECM homeostasis may be due, in part, to its ability to alter this balance via its downstream signaling pathway. It is noted that TGF-β1–induced expression of TIMP-1 and repression of MMP-1 are Smad3 dependent,^27,28 whereas induction of MMP-2 is critically dependent on Smad2.²⁹ This study added new information that Smad2 and Smad3 worked in a reciprocal manner in regulating matrix protein degradation. Deletion of Smad2 inhibited matrix protein degradation by enhancing TGF-β/Smad3-dependent TIMP-1 but suppressing MMP-2 expression in response to TGF-β1.

Smad3 binds DNA sequences directly through its MH1 domain, whereas Smad2 has no DNA-binding sequences and activates transcription indirectly by binding to transcriptional co-activators or repressors.^6,9 Sequence analysis reveals that Smad3-binding sequences are located in the promoter regions of COL1A2, COL2A1, COL3A1, COL5A1, COL6A1, and COL6A3, which may contribute to the Smad3-dependent mechanism of TGF-β–induced ECM expression.^11–13 Similarly, CTGF and TIMP-1 are target genes of Smad3.^28,30 In this study, we provided clear evidence that Smad3 is capable of binding to the COL1A2 promoter as demonstrated by a ChIP assay. More important, deletion of Smad2 in MEFs largely enhanced the binding of Smad3 to COL1A2. Thus, enhanced binding of Smad3 to its target genes may be a mechanism by which deletion of Smad2 promotes TGF-β–induced collagen matrix expression and tissue fibrosis.

Enhanced TGF-β1 autoinduction may be another mechanism whereby deletion of Smad2 promotes ECM production and tissue fibrosis. It has been clearly shown that the autoinduction of TGF-β is Smad3 dependent.²⁹ In this study, MEFs and mice lacking Smad2 revealed a marked increase in endogenous TGF-β1 expression in vitro and in vivo, suggesting a critical role for Smad2 in negatively regulating the amplification loop of TGF-β–mediated fibrosis in response of TGF-β1.

In summary, this study provides novel evidence for a protective role of Smad2 in ECM production and renal fibrosis. Although Smad2 and Smad3 interact physically, Smad2 may function to inhibit Smad3-mediated ECM production and fibrosis by competitively inhibiting Smad3 phosphorylation, nuclear translocation, transcriptional activity, and the binding to the target genes. Results from this study implicate that targeting Smad3 may be a specific therapeutic approach for the treatment of diseases with progressive tissue fibrosis.

Concise Methods

Generation of Smad2 Conditional KO Mice

Mice with conditional deletion of Smad2 from the kidney were generated by mating the Smad2ff mouse (C57B/L6) with the KspCre (C57B/L6) mouse. The generation and characterization of the Smad2 flox mouse and the KspCre transgenic mouse were described previously.^26,31,32 The genotypes of the Smad2ff/KspCre mouse were identified by PCR on DNA that was obtained from tails of adult mice with the primers as described previously.²² Conditional disruption of the Smad2 gene from the kidney TECs was confirmed by PCR, quantitative real-time PCR, Western blotting, and immunohistochemistry (Figure 1).

Establishment of UUO Model

The UUO kidney disease model was established in groups of eight Smad2ff and S2ff/KspCre mice (both genders, 8 weeks of age, 22 to 25 g body wt) by left ureter ligation as described previously.^33,34 Eight normal Smad2ff and S2ff/KspCre mice were used as controls. All mice were generated from the genetically identical littermates, and all animals were killed on day 10 after left ureter ligation. Kidney tissue samples were collected for immunohistochemistry, Western blot, and real-time reverse transcriptase–PCR as described previously.³⁴

Cell Culture

Characterized Smad2 WT and KO MEFs were used for this study.²⁹ The stable cell lines with Smad2 gene knockdown in NRK52E were generated and characterized as described previously.³⁵ To force overexpression of Smad2 in TECs, a pcDNA Smad2-expressing plasmid or control pcDNA empty vector (a gift from Dr. Rik Derynck, University of California, San Francisco) were transfected into NRK52E by using Lipofectamine LTX (Invitrogen). Characterization of Smad2-overexpressing NRK52E was determined by real-time PCR with forward primer 5′-CTTCACAGACCCATCAAACTCGGA and reverse primer 5′-GCACTATCACTTAGGCACTCAGCA and by Western blot with anti-Smad2 antibody (Zymed Laboratories, San Francisco, CA). A recombinant human TGF-β1 (R&D Systems, Minneapolis, MN) at concentrations of 2 ng/ml was added to the cell culture for periods of 0, 15, and 30 minutes and 1, 3, 12, and 24 hours for detection of TGF-β/Smad3 signaling and fibrosis response. In addition, TGF-β1 at dosages of 1.0, 2.0, and 4.0 ng/ml was applied for a dosage-dependent assay. At least four independent experiments were performed throughout the study.

Renal RNA Extraction and Real-Time PCR Examination

Total RNA was isolated from the cultured cells and kidney tissues using RNeasy Isolation Kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. Real-time PCR was performed using Bio-Rad iQ SYBR Green supermix with Opticon2 (Bio-Rad, Hercules, CA) as described previously.³⁴ The primers used in this study, including mouse collagen I, collagen III, CTGF, TGF-β1, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were described previously.^34,36 Other primers included the following: Smad2 forward 5′-ATGTCGTCCATCTTGCCATTC and reverse 5′-AACCGTCCTGTTTTCTTTAGCTT; MMP-2 forward 5′- AACTACGATGATGACCGGAAGTG and reverse 5′-TGGCATGGCCGAACTCA; and TIMP-1 forward 5′-GCATGGACATTTATTCTCCACTGT and reverse 5′-TCTCTAGGAGCCCCGATCTG. The ratio for the mRNA interested was normalized with GAPDH and expressed as the mean ± SEM.

Immunohistochemistry

Immunohistochemistry was performed in paraffin sections using a microwave-based antigen retrieval technique.³⁶ The antibodies used in this study included phospho-Smad3 (Rockland Immunochemicals, Gilbertsville, PA), CTGF, TGF-β1 (Santa Cruz Biotechnology), and collagen I and collagen III (Southern Tech, Birmingham, AL). After immunostaining, sections were counterstained with hematoxylin.

Western Blot Analysis

Protein from kidney tissues and cultured cells was extracted with RIPA lysis buffer, and Western blot analysis was performed as described previously.³⁷ After blocking nonspecific binding with 5% BSA, membranes were then incubated overnight at 4°C with the primary antibody against phospho-Smad3, total Smad3, collagens I and III, and GAPDH (Chemicon, Temecula, CA), followed by IRDye800-conjugated secondary antibody (Rockland Immunochemicals). Signals were detected using the LiCor/Odyssey infrared image system (LI-COR Biosciences, Lincoln, NE). Signal intensities of each Western blot were quantified by using the LiCor/Odyssey followed by analysis with Image J software (National Institutes of Health).

Phospho-Smad3 Nuclear Translocation Analysis

Smad2 WT and KO MEFs were cultured in eight-chamber glass slides in the presence or absence of TGF-β1 (2 ng/ml) for 0, 15, and 30 minutes and 1, 3, 12, and 24 hours. Cells were then fixed in 2% paraformaldehyde and incubated with the anti–phospho-Smad3 primary antibody at 4°C overnight, followed by 1 hour of incubation with goat anti-rabbit IgG-rhodamine (Millipore, Billerica, MA). Cells were counterstained with DAPI and analyzed under fluorescence microscope (Zeiss-spot; Carl Zeiss MicroImaging GmbH, Göttingen, Germany). The positive cells were counted from 500 nucleated cells and expressed as percentage of positive cells scored.

Smad3-Dependent Promoter Assay

Smad2 WT and KO MEFs were transiently transfected with a Smad3/4-responsive promoter p(CAGA)12-Luc as described previously.³⁸ PGL₃ basic plasmid was co-transfected into the cells as control. The transfection procedure was performed using jetPEI (Polyplus-Transfection, New York, NY) according to the manufacturer's instructions. After transfection and serum starving for 24 hours, TGF-β1 (2 ng/ml) was added to the cells for 24 hours. The luciferase activities of p(CAGA)₁₂ were analyzed by luciferase reporter gene assay kit according to the manufacturer's instructions (Roche Biochemical, Indianapolis, IN). The Smad3-responsive promoter activity was reported as the luciferase activity normalized to protein concentration measured with Lowry protein assay (Bio-Rad). Three independent experiments were performed throughout the study.

ChIP Analysis

ChIP was performed by transcription factor ChIP kit according to the manufacturer's instructions. In brief, Smad2 WT and KO MEFs pretreated with TGF-β1 were cross-linked with 1% formaldehyde for 10 minutes at 37°C, quenched with glycine, and then sonicated using a Bioruptor (Diagenode, Liege, Belgium) to generate 300- to 600-bp DNA fragments. Immunoprecipitation was performed with the antibody against Smad3 (Upstate), and IgG was used as a control. Precipitated DNAs were detected by PCR using specific primers to detect the binding of Smad3 to COL1A2: Forward 5′-AGGCAGGTCTGGGCTTTATT and reverse 5′-CGTATCCACAAAGCTGAGCA. The reaction mix consisted of a 1.5-QL template, 5.0 QL of 2× Bio-Rad iQ SYBR Green supermix, 0.3 μl of forward primer, 0.3 μl of reverse primer, and 2.9 μl of ddH₂O. qPCR was performed as described already using quantitative real-time PCR.

Statistical Analysis

Data obtained from this study are expressed as the mean ± SEM from at least three independent experiments or groups of eight mice. Statistical analyses were performed using one-way ANOVA followed by Newman-Keuls post hoc test (Prism 4.0; GraphPad Software, San Diego, CA).

Disclosures

None.