Connective tissue growth factor is a Smad2 regulated amplifier of transforming growth factor β actions in hepatocytes—But without modulating bone morphogenetic protein 7 signaling

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

In vivo knockdown of connective tissue growth factor (CTGF/CCN2) was recently shown to attenuate the formation of experimental liver fibrosis. The secreted, cysteine-rich growth factor is proposed to adversely modulate the binding of profibrogenic transforming growth factor β (TGF-β) and its natural antagonist bone morphogenetic protein (BMP) to their cognate receptors in several cellular systems, but the functionality of CTGF in modulation of the TGF-β/BMP signaling pathways is still unknown. This study aims at characterizing a potentially differential modulating role of CTGF on TGF-β– and BMP7-dependent transactivation of reporter gene [Ad-(CAGA)12-MLP-luc, Ad-hCTGF-luc, and Ad-(BRE)2-luc reporter gene] expression in rat hepatocytes. In this context, emphasis is also placed on the differential roles of Smad2 and Smad3 in the TGF-β–dependent transactivation of the endogenous CTGF gene and the CTGF gene reporter, as investigated following adenoviral infection of wild-type and dominant negative Smad2/3 or treatment with the specific inhibitor of Smad3 or ALK5-specific (SB-431542) inhibitor. In this analysis, we found (1) a selective transcriptional activation of the CTGF promoter by Smad2 (but not Smad3); (2) the failure of BMP7 to inhibit the transcriptional activation of the Smad3-selective (CAGA)12-luc reporter by TGF-β, as well as the failure of TGF-β to inhibit the transcriptional activation of the Smad5-selective (BRE)2-luc reporter by BMP7; and (3) the sensitization of hepatocytes toward TGF-β type I receptor (ALK5)/Smad2 and Smad3-mediated TGF-β signaling by CTGF, whereas BMP type I receptor (ALK1)/Smad5-mediated BMP7 signaling is not modulated. Conclusion: CTGF acts as a Smad2-dependent sensitizer of TGF-β actions that does not influence BMP7 signaling in hepatocytes. (HEPATOLOGY 2009.)

We and others have shown that hepatocytes substantially synthesize connective tissue growth factor (CTGF/CCN2) during culture and in injured liver and are likely to be the major cellular source of CTGF in the liver.1, 2 CTGF is a 36- to 38-kDa, cysteine-rich, heparin-binding and secreted protein that was first identified in the conditioned medium of human endothelial cells.3 It is now classified as the second of six members of the CCN family of matricellular proteins containing CTGF itself, CYR61, NOV, and others.4 The individual CCN family members share an approximately 40% to 60% sequence similarity and are characterized by a modular structure that consists of four individual structural features.5 Within the CTGF protein, modules II and III are separated by a proteinase-sensitive hinge region.6 The carboxy-terminal domain IV of CTGF binds heparin and can signal through heparan sulfate containing proteoglycans and integrins.7 CTGF promotes adhesion and adhesive signaling and is necessary for optimal adhesion on fibronectin.7, 8

CTGF is suggested as an important amplifier of the profibrogenic action of transforming growth factor β (TGF-β) in a variety of tissues.5 Therefore, CTGF has reached considerable pathophysiological relevance because of its involvement in the pathogenesis of fibrotic diseases, atherosclerosis, skin scarring, and other conditions with excess production of connective tissue.9 Even though the molecular mechanisms of CTGF action are still unknown, its crucial role in fibrogenesis is documented by strong up-regulation in fibrotic liver tissue,4, 10, 11 and even more importantly in recent studies demonstrating that knockdown of CTGF by siRNA (small interfering RNA) technology leads to substantial attenuation of experimental liver fibrosis.12, 13

The strong increase of CTGF expression as observed in fibrotic tissue occurs on the level of transcription and is stimulated by specific growth factors such as TGF-β itself, or endothelin-1, but also by environmental influences such as biomechanical stress and hypoxia.5CTGF gene activation by TGF-β is mediated by a functional Smad-binding element located within the CTGF gene promoter.14 Although Smad2/3 have been identified as intracellular mediators in the TGF-β signaling pathway, little is known about the selective activation of Smad2 versus Smad3. However, Smad2 and Smad3 knockout mouse phenotypes and studies comparing Smad2 and Smad3 activation of TGF-β target genes suggest that Smad2 and Smad3 have distinct roles in TGF-β signaling.15

Epithelial-to-mesenchymal transition (EMT) of adhering hepatocytes into cells with reduced intercellular adhesion, increased motility, and mesenchymal, fibroblast-like properties is gaining more and more importance in the pathogenetic understanding of hepatic fibrogenesis.16, 17 The prototype of the currently most powerful inducer of EMT is TGF-β,18 activating this pathway via induction of Smad2/3 phosphorylation and the Snail transcription factor.18 Currently, bone morphogenetic protein (BMP)-7 is regarded as the most important opposing factor for TGF-β, which not only inhibits EMT but can also induce a mesenchymal-to-epithelial transition.16

Recent reports have provided evidence that up-regulation of CTGF inhibits BMP7 signal transduction in diabetic nephropathy.19 Moreover, by use of co-immunoprecipitation, solid-phase binding assay, surface plasmon resonance analysis, and deletion analysis, it has been demonstrated that CTGF binds to BMP family members (BMP7, BMP4) with high affinity and that the interaction domain is located within the cysteine-rich repeat.19, 20 The finding that this interaction could be partially competed by TGF-β1 further suggests that CTGF serves as an extracellular trapping protein for BMP and TGF-β thus modulating the mutual activity of these cytokines by preventing their binding to respective receptors. When injected into blastomeres of Xenopus laevis, CTGF messenger RNA was able to induce secondary axes which is consistent with inhibition of BMP signaling.20 Of note, the opposite effect, enhancement of receptor binding, was observed for TGF-β. From this, CTGF would act profibrogenic by shifting the balance toward mesenchymal activity during hepatocellular EMT.21 Clarification of this postulation is still pending, however.

The goal of this study was to transfer the results observed in Xenopus laevis to rat hepatocytes to discuss a potential role of CTGF on hepatocellular EMT by differentially modulating TGF-β– and BMP7-dependent target gene expression. In this context, emphasis is also placed on the differential roles of Smad2 and Smad3 in the TGF-β–dependent transactivation of the CTGF promoter.

Abbreviations

BMP, bone morphogenetic protein; CTGF, connective tissue growth factor; EMT, epithelial-to-mesenchymal transition; siRNA, small interfering RNA; SIS3, specific inhibitor of Smad3; TGF-β, transforming growth factor β; SEM, standard error of the mean.

Material and Methods

Reagents and Antibodies.

rhBMP7 (C-12003) was obtained from PeproTech (Hamburg, Germany); rhCTGF (CTGF100, containing only the carboxy-terminal domain IV of CTGF) was obtained from EMP Genetech (Ingolstadt, Germany); rhTGFβ-1 (240-BP) was obtained from R&D Systems (Minneapolis, MN). Proof of bioactivity for the respective cytokines was provided by the manufacturer (R&D, PeproTech) or demonstrated previously.22, 23 The ALK4/5 inhibitor (SB-431542) was obtained from Tocris Bioscience (Ellisville, MO), and specific inhibitor of Smad3 (S0447) was obtained from Sigma Aldrich (St. Louis, MO). Rabbit anti-Smad3 antibody (ab28379) was obtained from Abcam (Cambridge, UK); goat anti-CTGF/CCN2 antibody (L-20, sc-14939) was obtained from Santa Cruz Biotechnology (Santa Cruz, CA); and rabbit anti-pSmad3 (Ser423/425)/ pSmad1 (Ser463/465) antibody (#9514), rabbit anti-pSmad2 (Ser465/467) antibody (#3101), and rabbit anti-Smad2 antibody (#3102) were obtained from Cell Signaling/New England Biolabs (Ipswich, MA).

Animals.

Adult Sprague-Dawley rats (200–250 g) had free access to a standard laboratory chow diet and normal tap water. All animals received care and treatment in compliance with the German Animal Protection Act, which is in accordance with the German Research Council's criteria.

Cell Culture and Preparation of Rat Hepatocytes.

Hepatocytes were isolated from male Sprague-Dawley rats (180–250 g) using the two-step collagenase method of Seglen24 with some modification.25 Cells were isolated and cultured in the complete absence of TGF-β, BMP7, or CTGF. The viability of the final parenchymal cell suspension, estimated by trypan blue exclusion, was around 85%, and cell recovery was 20 to 50 × 107 cells/liver. Contamination with other nonparenchymal cells was <2%. Cells were seeded in 2 mL GIBCO HepatoZYME-SFM (Invitrogen, Carlsbad, CA) without fetal bovine serum on type I collagen (BD Bioscience, Franklin Lakes, NJ)–coated plastic dishes (BD Bioscience) at a density of 5.4 × 104 cells/cm2. They were cultured at 37°C in a humidified atmosphere of 5% CO2 and 95% air supplemented with 4 mM L-glutamine, penicillin (100 IU/mL), and streptomycin (100 mg/mL) (all from Cambrex, East Rutherford, NJ). The first change of the medium was 1 hour after seeding, and unattached hepatocytes were washed off with Dulbecco's modified Eagle's medium (Cambrex) and then cultured for various times in GIBCO HepatoZYME-SFM supplemented as described above.

Gene Silencing of Smad2 and Smad3.

Hepatocytes were transfected with 300 ng siRNAs using the HiPerFect transfection reagent and HP Genome Wide siRNAs directed against rat Smad2 (SI00261667, SI00261688, SI03062640, SI03108140) and Smad3 (SI00255983, SI00256004, SI03056515, SI03090437) or the fluorescein-labeled AllStars negative control siRNA (#1027282) obtained from Qiagen (Hilden, Germany). Transfection rates were controlled via visual examination under fluorescence microscopy. Four combinations of each two siRNAs were tested for their ability to repress Smad2 and Smad3 expression using western blot analysis and reverse-transcription polymerase chain reaction. Studies were continued with those combinations resulting in the strongest repression.

Adenoviral Expression Vectors.

The generation of adenoviruses expressing wild-type and dominant negative Smad2 and Smad3 with serine-to-alanine mutations in the phosphorylation site SSXS (to AAXA) has been described.26 For generation of the Ad-hCTGF-luc reporter adenovirus, the ≈2.5-kbp ClaI fragment of vector pGL3-Basic-hCTGF-luc (described by Gressner et al.1) harboring a fusion of human CTGF gene promoter (containing the minimal CTGF promoter element necessary) and sufficient to confer TGF-β responsiveness27–29 and the luciferase reporter gene was cloned into the ClaI site of vector pΔE1sp1A30 resulting in the generation of pΔE1sp1A-hCTGF-luc. The integrity of cloning boundaries was verified by restriction analysis and sequencing with the flanking primers 5′-GCG TAA CCG AGT AAG AAT TTG-3′ and 5′-GGC GAC CAT CAA TGC TGG AG-3′. The integration of the reporter cassette from pΔE1sp1A-hCTGF-luc into the adenoviral backbone vector pJM1731 was performed by in vitro homologous recombination in the human embryo kidney cell line 293 using a protocol described previously.32 For generation of the adenoviral reporter vectors Ad-(BRE)2-luc and Ad-(CAGA)12-MLP luciferase reporter viruses, the vectors pGL3-Basic-(CAGA)12MLP-luc33 and pGL3-Basic-(BRE)2-luc34 were digested with BamHI and KpnI, filled in by Klenow enzyme, and cloned into the blunted EcoRI site of pENTR4 (Invitrogen) and integrated into the pAd/Block-it-Dest vector (Invitrogen) using the LR recombination reaction kit. Ad-βGal was used as a control virus throughout the study. (CAGA)12-MLP-luc represents an established Smad3 responsive reporter gene construct Smad5 responsive (BRE)2-luc resembles a promoter sequence from the classical BMP7 target Id-1.34–37

Adenoviral Transfection.

Hepatocytes were infected with 1 × 108 pfu/mL of E1-deleted adenoviral vectors expressing rat wild-type Smad2, wild-type Smad3, dominant negative Smad2, dominant negative Smad3, and β-galactosidase (Ad-β-Gal, Ad-wt-Smad2, Ad-dn-Smad2, Ad-wt-Smad3, or Ad-dn-Smad3). Eighteen hours after infection, the medium was changed to fresh GIBCO HepatoZYME-SFM medium, and the cells were incubated for another 6 hours before stimulation. Recombinant adenoviruses were delivered to hepatocyte with an efficiency >95% (confirmed by staining for Ad-β-Gal, data not shown). For coinfection experiments with Ad-CTGF-luc (1 × 108 pfu/mL), cells were infected simultaneously with two different viruses.

Luciferase Gene Reporter Assays and Western Blotting.

Cells were cultured in black 96-well plates and infected with 1 × 108 virons/mL of respective reporter virus [Ad(CAGA)12-MLP-luc, Ad(BRE)2-luc, or Ad-hCTGF-luc reporter virus]. After specific treatment, the luciferase activity was measured as described.1 Preparations of cytoplasmatic cell extracts, determination of protein concentrations, and western blot analysis were essentially performed as described.1

Statistical Analysis.

For statistical analysis, SPSS 16.0 (SPSS, Chicago, IL) was used, applying two-tailed unpaired Student's t test with statistical significance set at P < 0.05.

Results

Differential Effect of BMP7 and TGF-β on Hepatocellular (CAGA)12- and (BRE)2-Luciferase Reporter Gene Transactivation in Hepatocytes, and their Interaction.

As already shown in other cell systems, TGF-β efficiently increased the activity of (CAGA)12-MLP-luc representing an established Smad3-responsive reporter gene construct in cultured rat hepatocytes (1 ng/mL; 5.2-fold ± 1.3) (Fig. 1A).33 Coincubation with 100 ng/mL recombinant BMP7 did not reduce TGF-β–dependent and TGF-β–independent transcriptional activation of the (CAGA)12-MLP-luc reporter. As expected, BMP7 significantly increased the transcriptional activation of the BMP-responsive reporter (BRE)2-luc, which resembles a promoter sequence from the classical BMP7 target Id-134–37 (100 ng/mL; 5.3-fold ± 1.0) (Fig. 1B). The rise in (BRE)2-luc activity that was induced by BMP7 was, however, not significantly inhibited by TGF-β.

Figure 1.

Differential effect of BMP7 and TGF-β on (CAGA)9-MLP-luc and (BRE)2-MLP-luc reporter plasmid transactivation in hepatocytes. Primary rat hepatocytes were transiently infected with (A) Ad-(CAGA)12-MLP-luc or (B) Ad-(BRE)2-luc and stimulated (24 hours) with TGF-β1 (1 ng/mL), BMP7 (100 ng/mL), or a combination of both cytokines or were left untreated. The cells were extracted and the luciferase activities determined. All data represent the mean ± standard error of the mean (SEM) of at least three independent experiments performed in triplicate. **P < 0.01 versus untreated control.

In contrast, 100 ng/mL BMP7 prevented the stimulation of hCTGF-luc activity (100 pg/mL, 0.17-fold ± 0.08; 1 ng/mL, 0.83-fold ± 0.07; 5 ng/mL, 1.31-fold ± 0.11) (Fig. 2A) and CTGF protein synthesis (Fig. 2B, lower western blot and quantification) in primary hepatocytes; both were induced by increasing dosages of TGF-β.

Figure 2.

Opposing activities of BMP7 and TGF-β on hCTGF-luc reporter plasmid transactivation in hepatocytes. (A) Primary rat hepatocytes were infected with Ad-hCTGF-luc and stimulated (24 hours) with increasing concentrations of TGF-β1 with or without addition of 100 ng/mL BMP7. The cells were extracted and the luciferase activities were determined. All data represent the mean ± SEM of at least three independent experiments performed in triplicate. *P < 0.05, **P < 0.01 versus the respective TGF-β-treated equivalent. (B) Primary rat hepatocytes were stimulated (24 hours) with increasing concentrations of TGF-β1 with (lower panel) or without (upper panel) 100 ng/mL BMP7. Respective lysates (20 μg) were analyzed for CTGF protein expression via western blot analysis. β-Actin served as a loading control. Blots were quantified relative to β-actin and values are shown as x-fold of untreated control. Quantifications represent the mean ± SEM of four independent cultures. *P < 0.05.

However, constitutive CTGF promoter activation and protein expression in hepatocyte was not inhibited by BMP7, as exemplified in Fig. 2B (lower western blot).

In order to further elucidate the differential response of (CAGA)12-luc and hCTGF-luc activity toward TGF-β and/or BMP7, it has to be considered that the promoter sequence of hCTGF-luc may be transcriptionally activated by both Smad2 and Smad3, whereas (CAGA)12-luc-luc specifically responds to Smad3 only.

Differential Regulation of Hepatocellular CTGF Expression by Smad2 and Smad3.

The results presented above suggest a predominant role of Smad2 (compared with Smad3) in the transactivation of the CTGF promoter in hepatocytes. This is of particular interest insofar as both Smads were regarded as being largely interchangeable in terms of mediating TGF-β signaling to the CTGF gene promoter.5

Therefore, to determine the differential role of Smad2 and Smad3 in mediating TGF-β responsiveness in CTGF expression, rat hepatocytes were comparatively infected with different concentrations (0.5 − 3 × 108 pfu/mL) with Ad-wt-Smad2, Ad-dn-Smad2, Ad-wt-Smad3, or Ad-dn-Smad3 and Ad-βGal as a control. The expression of transgenes (Smad2, Smad3) and CTGF were analyzed 24 hours after infection (Fig. 3A,B). A potential cross-talk between these Smads as far as expression was excluded previously.26

Figure 3.

Impact of Ad-wt-Smad2, Ad-dn-Smad2, Ad-wt-Smad3, and Ad-dn-Smad3 on hepatocellular CTGF synthesis. Primary rat hepatocytes were infected with the indicated adenoviruses immediately after isolation and plating and treated with 1 ng/mL TGF-β1 for 24 hours after infection. Protein lysates were probed for CTGF and (A) Smad2 or (B) Smad3 via western blot analysis. β-Actin served as a control for equal gel loading. Representative blots were quantified relative to β-actin. Quantifications represent the mean ± SEM of three independent cultures. *P < 0.05, **P < 0.01 versus control cells infected with Ad-βGal. (C) Primary rat hepatocytes were coinfected with Ad-hCTGF-luc and with the indicated adenoviruses immediately after plating and selectively treated with 1 ng/mL TGF-β1. Control cells were infected with Ad-βGal. The cells were extracted 24 hours after infection and addition of TGF-β1, and the luciferase activities were determined. The mean values of triplicate determinations from three different cultures are shown. All data represent the mean ± SEM of at least three independent experiments performed in triplicate. **P < 0.01 versus control cells infected with Ad-βGal.

Ad-wt-Smad2 caused a significant increase in TGF-β-induced CTGF protein synthesis, whereas overexpression of the dominant negative Smad2 resulted in a marked decrease of CTGF protein land promoter activity (Fig. 3A,C). However, infection with Ad-wt-Smad3 or Ad-dn-Smad3 had no influence on CTGF synthesis (Fig. 3B,C).

Consistent with these findings, knockdown studies using siRNAs targeting Smad2 and Smad3 genes clearly showed that in the presence of TGF-β, knockdown of Smad2 (Fig. 4A), but not of Smad3 (Fig. 4B) or scramble control siRNA, strongly reduced CTGF protein expression 24 hours after transfection.

Figure 4.

Differential inhibition of TGF-β-induced CTGF expression in hepatocytes in response to selective Smad2 and Smad3 gene knockdown by RNA interference. Hepatocytes were transfected with pSilencer siRNAs targeting rat Smad2 or Smad3 gene expression and selectively treated with 1 ng/mL TGF-β1 for 24 hours after transfection. Three different combinations of each of the two siRNAs were transfected as follows. Smad2: (1) SI00261667, SI00261688; (2) SI00261688, SI03108140; (3) SI03062640, SI03108140. Smad3: (1) SI00255983, SI03056515; (2) SI00256004, SI03090437; (3) SI00255983, SI03090437. The efficiency of Smad2 and Smad3 knockdown was furthermore confirmed via reverse-transcription polymerase chain reaction (data not shown). Protein lysates were probed with antibodies specific for (A,B) CTGF and (A) Smad2 or (B) Smad3 via western blot analysis. β-Actin served as a loading control. Blots were quantified relative to β-actin. Representative blots are depicted. Quantifications represent the mean ± SEM of three independent cultures. **P < 0.01 versus control cells transfected with scramble siRNA.

The data above were furthermore confirmed by treatment of hepatocytes with the small molecule inhibitors SB-431542 and specific inhibitor of Smad3 (SIS3) (Fig. 5A,B). SB-431542 was originally identified as an inhibitor of the activin receptor-like kinase (ALK)5 (TGF-β type I receptor), thus unselectively preventing Smad3 and Smad2 phosphorylation.38 In contrast, SIS3 has been characterized as a potent and highly selective inhibitor of Smad3 function that attenuates the TGF-β–induced phosphorylation of Smad3 and the interaction of Smad3 with Smad4, whereas the phosphorylation of Smad2 remains unaffected.39, 40 The presented data confirm our previous observation41 that SB-431542 is highly effective in suppressing TGF-β-induced and spontaneous CTGF synthesis in hepatocytes (Fig. 5A), whereas treatment with SIS3 had only a marginal effect (Fig. 5B).

Figure 5.

Effect of ALK5 (TGF-β type I receptor) and selective Smad3 inhibition on TGF-β–induced CTGF expression in hepatocytes. (A) Upper panel: western blot of CTGF of hepatocytes cultured for 24 hours with or without addition of 1 ng/mL TGF-β1. Some cultures received the ALK4/5 inhibitor (SB-431542, 5 and 10 μM). Lysates were probed for CTGF and β-actin expression. A representative blot of four independent experiments is shown. Lower panel: As positive controls, western blots for p-Smad2 and p-Smad3 were performed using extracts of hepatocytes that were cultured for 24 hours in the presence of 1 ng/mL TGF-β1 with or without addition of 10 μM SB-431542. β-Actin expression served as a loading control. A representative blots that is shown in the upper panel was quantified for CTGF expression relative to β-actin. Bottom graphs represent the mean ± SEM of quantifications from four independent cultures. **P < 0.01 versus the respective controls not treated with SB-431542. (B) Western blot of CTGF in hepatocytes cultured for 24 hours in the absence or presence of 1 ng/mL TGF-β1. Some cultures received SIS3 at concentrations of 5 or 10 μM. Protein lysates were probed for CTGF or Smad2/3 and p-Smad1/2/3 expression via western blot analysis. Blots of CTGF and p-Smad2 were quantified relative to β-actin. Quantifications represent the mean ± SEM of four independent cultures.

Effect of CTGF on TGF-β–Dependent Transactivation of the (CAGA)12-MLP-luc and hCTGF-luc Reporter Genes as Well as on BMP7-Induced Stimulation of (BRE)2-luc Activity.

We next tested the impact of exogenously added CTGF on TGF-β–dependent transactivation of the (CAGA)12-MLP-luc reporter (Fig. 6A). In this set of experiments, we found that CTGF dose-dependently enhanced TGF-β/Smad3-stimulated (CAGA)12-MLP-luc activity. The effect was consistently observed in multiple cell cultures (100pM, 1.08-fold ± 0.06; 200pM, 1.7-fold ± 0.14). Based on the fact that hepatocellular CTGF expression is strongly stimulated by TGF-β,1, 41, 42 it was not surprising that exposition of hepatocyte to CTGF also lead to a sensitization toward TGF-β–induced activation of hCTGF-luc, thus amplifying the TGF-β signal to the CTGF promoter in an autostimulatory manner during conditions of increased TGF-β activity (100pM, 1.5-fold ± 0.08; 200pM, 1.8-fold ± 0.11) (Fig. 6B). These data were confirmed via western blot analysis, showing an enhancement of TGF-β–induced Smad2 and Smad3 phosphorylation in the presence of exogenously added CTGF (Fig. 6C). However, CTGF had no impact on BMP7/Smad5-dependent target gene expression in hepatocyte as exemplified by (BRE)2-luc activity (Fig. 6D).

Figure 6.

Influence of CTGF on BMP7- and TGF-β signaling in hepatocytes. Primary rat hepatocytes were infected with (A) Ad-(CAGA)12-MLP-luc, (B) Ad-hCTGF-luc or (D) Ad(BRE)2-luc and stimulated (24 hours) with increasing concentrations of CTGF with or without (A, B) 1 ng/mL TGF-β1 or (D) 100 ng/mL BMP7. The cells were extracted and the luciferase activities were determined. All data represent the mean ± SEM of four independent experiments performed in triplicate. **P < 0.01 versus control. (C) Western blot of pSmad2/3 in hepatocytes cultured for 24 hours in the absence or presence of 0.25 ng/mL TGF-β1. Some cultures received rhCTGF at the indicated concentrations. Protein lysates were probed for pSmad2/3 expression via western blot analysis. Blots were quantified relative to β-actin, and values are shown as x-fold of untreated control. Quantifications represent the mean ± SEM of three independent cultures. A representative blot is shown.

Discussion

In this study, we demonstrate three results on the reciprocal interaction of TGF-β, BMP7, and CTGF on selected hepatocellular target gene activation and translation: (1) the selective transcriptional activation of the CTGF promoter by Smad2 (but not Smad3); (2) the failure of BMP7 to inhibit the transcriptional activation of Smad3-selective (CAGA)12-luc reporter by TGF-β, as well as the failure of TGF-β to inhibit the transcriptional activation of the Smad5-selective (BRE)2-luc reporter by BMP7 in hepatocytes; and (3) the sensitization of hepatocytes toward type 1 TGF-β (ALK5) receptor/Smad2- and Smad3-mediated TGF-β signaling by CTGF while type 1 BMP (ALK1) receptor/Smad5-mediated BMP7 signaling is not quantitatively inhibited (Fig. 7).

Figure 7.

A simplified and schematic overview of the proposed Smad2-dependent regulation of CTGF in hepatocytes and of its interaction with TGF-β and BMP7 signaling pathways. Hepatocellular CTGF is a Smad2-regulated target gene of TGF-β. The extracellular presence of CTGF in turn amplifies TGF-β actions of both Smad2 (CTGF) and Smad3-regulated pathways. No effect of CTGF on BMP7 signaling is observed. However, BMP7 inhibits CTGF expression while not interfering with upstream (Smad3 or TβRI/ ALK5-mediated) TGF-β signaling. Arrows indicate positive feedback; anchors symbolize an inhibition. TβRI, TGF-β type I receptor.

Considering the predominant role of Smad2 (when compared with Smad3) in transmitting TGF-β–dependent (and TGF-β–independent) signals41, 42 to the hepatocellular CTGF promoter, the differential response of (CAGA)12-MLP-luc and hCTGF-luc toward exogenously added BMP7 suggests that despite different observations in L6E9 cells,36 BMP7 signaling interferes with the activation of the profibrogenic Smad2 but not with Smad3 in hepatocytes. Thus, an interaction of both cytokines at their respective receptors on the hepatocyte membrane seems unlikely, because a disturbance of TGF-β binding to its type II or type I (ALK5) receptor by BMP7 would very likely also result in a disturbance of Smad3 signaling, which we did not observe, as seen by a steady TGF-β–induced (CAGA)12-MLP-luc activity in the presence of BMP7. Conversely, TGF-β does not inhibit BMP7-induced activation of the ALK1 receptor, because this would inevitably lead to reduced BMP7-stimulated (BRE)2-luc activity by Smad5 in the presence of TGF-β, which was also not observed.

Thus, mutual inhibition of BMP7 and TGF-β signaling in hepatocytes seems to be a selective process, involving downstream Smad2 signaling but not Smad3/5 pathways. Based on this, it may be concluded that both TGF-β type I (ALK5) receptor-phosphorylated Smads are not interchangeable in terms of transcriptional regulation of the CTGF gene promoter, which has already been proposed by our group.41–43 The (CAGA)12-MLP-luc reporter gene construct only contains a Smad3-specific (but not Smad2-specific) binding site,33, 44 whereas the promoter sequence of hCTGF-luc contains a Smad binding element that may be activated by both Smad2 and Smad3.27–29

Therefore, based on the differential responses of Smad3-specific (CAGA)12-luc33, 44 and Smad2/3-responsive hCTGF-luc gene reporters27–29 toward TGF-β/BMP7 signaling, an exclusive role of Smad2 in the transcriptional activation of the hepatocellular CTGF promoter was again suggested. This assumption could finally be confirmed by three different approaches with adenoviral transfection of wt-Smad2/3 and dn-Smad2/3, siRNA-mediated gene silencing of Smad2/3, and selective inhibition of the ALK5 receptor and Smad3. More and more data are being published addressing essential differences between TGF-β–mediated signaling pathways in terms of Smad3 as the profibrogenic Smad in the liver,15 with particular focus on the morphological and functional maturation of hepatic myofibroblasts,26, 45–47 but also in several other organ systems,48–51 while a potential relevance of Smad2 the molecular pathogenesis of fibrosis has long been neglected.

To our knowledge, this is the first report to show that Smad2 but not Smad3 is responsible in mediating TGF-β signals to the hepatocellular CTGF promoter. Based on this, and on the crucial role of CTGF in fibrogenesis as documented by strong up-regulation in fibrotic liver tissue,4, 10, 11 or in vivo knockdown experiments,12, 13 our data suggest that Smad2 is also profibrogenic by mediating TGF-β signals to the CTGF promoter in hepatocytes, which we have identified to be the major cellular source of CTGF in the liver.1, 2 This might have implications for future therapeutic approaches in patients with chronic fibrogenic liver disease. It is interesting that, in contrast to our results in hepatocytes, it has been shown that TGF-β cannot induce CTGF protein in Smad3-deficient embryonic fibroblasts,52 emphasizing the context-specific nature of CTGF gene regulation. Therefore, future studies to comparatively analyze the regulation of CTGF expression in cells that were taken from Smad2- or Smad3-deficient mice would be of great interest. Although Smad2-deficient embryos fail to establish an anterior–posterior axis and die around day 7.5 of gestation,53 cells taken from these mice should be helpful in defining the precise role of Smad2 in organ-specific regulation of CTGF gene expression.

Other parts of this experimental study were motivated by an earlier, well-respected report showing that CTGF can antagonize BMP4 activity in Xenopus laevis by preventing its binding to BMP receptors and that it has the opposite effect (enhancement of receptor binding) on the TGF-β type I receptor.20 We therefore aimed at investigating the possibility of transferring these results to hepatocytes in order to discuss a potential role of CTGF on hepatocellular EMT by differentially modulating TGF-β– and BMP7-dependent target gene expression.

In contrast to the observations found in Xenopus laevis and results established in the diabetic kidney,19 we found no evidence for an interaction of CTGF with BMP7, its receptor, or the Smad5 signaling pathway. However, our data show that CTGF sensitizes hepatocytes toward TGF-β–dependent activation of Smad2 and Smad3 signaling, suggesting a selective extracellular enhancement of TGF-β–dependent receptor activity rather than downstream modulation of the TGF-β signaling pathway. Our proposed extracellular mode of action of CTGF is supported by the fact that CTGF enhances hCTGF-luc activity in the presence of TGF-β, whereas basal hCTGF-luc activity, which has been shown to be at least partially regulated by intracellular signaling of TGF-β and activin A,41, 43 is not affected. Therefore, a TGF-β–independent, entirely autostimulatory mechanism of CTGF expression in hepatocytes has to be excluded.

Previous results from other, nonhepatic cell systems support our observations that CTGF acts as a cofactor potentiating TGF-β signaling in hepatocytes. For example, by using embryonic fibroblasts isolated from CTGF−/− mice, Shi-wen et al.54 reported that CTGF is a necessary cofactor for a subset of responses to TGF-β, such as activation of the adhesive focal adhesion kinase/Akt/phosphatidylinositol 3-kinase cascade, activation of focal adhesion kinase/Akt-dependent genes, and adhesion to matrix. These findings are consistent with the findings of a Japanese group investigating the in vivo effects of subcutaneous injections of TGF-β and CTGF into newborn mice and show that interaction of these growth factors is required for persistent formation of fibrotic skin tissue, with TGF- β causing the induction and CTGF needed for maintenance of skin fibrosis.55

In conclusion, we could identify the CTGF gene as a Smad2-regulated sensitizer of TGF-β actions in hepatocytes that does not interfere with the BMP7 signaling pathway. Future studies on the detailed mechanisms of CTGF-mediated enhancement of TGF-β actions in hepatocytes as well as on the expression of relevant BMP7 receptors on hepatocytes are still required, and it is our hope that the present study will initiate further investigations on this topic.

Acknowledgements

The authors thank R. G. Wells for providing adenovirus-expressing wild-type and dominant negative Smad2 and Smad3. The authors furthermore appreciate the cooperation of C. H. Heldin (Ludwig Institute for Cancer Research, Sweden) in sending vectors pGL3-Basic-(CAGA)12MLP-luc and pGL3-Basic-(BRE)2-luc and Wanda N. Vreden for cloning of the adenoviral constructs Ad-(CAGA)12-MLP-luc and Ad-(BRE)2-luc.

Ancillary

Advertisement