CTGF Mediates Smad-Dependent Transforming Growth Factor β Signaling To Regulate Mesenchymal Cell Proliferation during Palate Development

Abstract

Transforming growth factor β (TGF-β) signaling plays crucial functions in the regulation of craniofacial development, including palatogenesis. Here, we have identified connective tissue growth factor (Ctgf) as a downstream target of the TGF-β signaling pathway in palatogenesis. The pattern of Ctgf expression in wild-type embryos suggests that it may be involved in key processes during palate development. We found that Ctgf expression is downregulated in both Wnt1-Cre; Tgfbr2^fl/fl and Osr2-Cre; Smad4^fl/fl palates. In Tgfbr2 mutant embryos, downregulation of Ctgf expression is associated with p38 mitogen-activated protein kinase (MAPK) overactivation, whereas loss of function of Smad4 itself leads to downregulation of Ctgf expression. We also found that CTGF regulates its own expression via TGF-β signaling. Osr2-Cre; Smad4^fl/fl mice exhibit a defect in cell proliferation similar to that of Tgfbr2 mutant mice, as well as cleft palate. We detected no alteration in bone morphogenetic protein (BMP) downstream targets in Smad4 mutant palates, suggesting that the reduction in cell proliferation is due to defective transduction of TGF-β signaling via decreased Ctgf expression. Significantly, an exogenous source of CTGF was able to rescue the cell proliferation defect in both Tgfbr2 and Smad4 mutant palates. Collectively, our data suggest that CTGF regulates proliferation as a mediator of the canonical pathway of TGF-β signaling during palatogenesis.

INTRODUCTION

Palatogenesis is a complex process that involves multiple developmental events, including growth of the palatal primordia from the lateral edges of the maxillary process, reorientation of the palatal shelves from a vertical to a horizontal position, midline fusion of the palatal shelves, and disappearance of the midline epithelial seam (1). These processes encompass cellular activities such as cell proliferation, programmed cell death, and extracellular matrix (ECM) synthesis and remodeling, occurring in tissues that have already been patterned (reviewed in reference 2). Disruption of any of these activities may result in failure of palatal fusion and, therefore, in cleft palate. Craniofacial malformations such as cleft palate are among the most common congenital birth defects in humans (3).

Palatal structures are composed of cranial neural crest (CNC)-derived mesenchyme and pharyngeal ectoderm-derived epithelium. The transforming growth factor β (TGF-β) signaling pathway plays a crucial role in the development of these tissues and their interactions during palatogenesis. Loss-of-function mutation of Tgfb2 or Tgfb3 results in cleft palate. Tgfb2-null mutant mice exhibit anteroposterior cleft of the secondary palate with 23% phenotype penetrance (4). Tgfb3-null mutation results in 100% penetrance of cleft secondary palate (5, 6). The etiology of cleft palate in Tgfb3-null mutant mice is apparently due to a failure of fusion of the palatal shelves, which is caused by a persistent midline epithelium and reduced cell proliferation in the palatal mesenchyme. Interestingly, targeted null mutation of TGF-β receptor II (Tgfbr2) in CNC cells (CNCC) of mice leads to craniofacial malformations, including small mandible, dysmorphic calvaria, and cleft palate (7). The cleft palate in these neural crest-specific Tgfbr2 mutant mice is due to a cell proliferation defect (7). In humans, mutations in TGFBR2 cause Loeys-Dietz syndrome (LDS), which also includes craniofacial anomalies such as cleft palate (8).

TGF-β signals through a complex pathway and regulates downstream gene expression. TGF-β ligands activate the membrane receptor serine/threonine kinase complex composed of TGFβRII and TGFβRI (also known as ALK5). The ligand-receptor complex phosphorylates SMAD2 and SMAD3, which form a transcriptional complex with SMAD4. This complex translocates into the nucleus and binds Smad-binding elements (SBE) in the promoter region of downstream target genes to control their expression (9). Cells can also use noncanonical pathways to transduce TGF-β signals. Those pathways are context dependent and include p38 mitogen-activated protein kinase (MAPK), extracellular signal-regulated kinase (ERK), and Jun N-terminal protein kinase (JNK) (10–12). Although TGF-β signaling has been extensively studied in craniofacial development, the downstream mediators of its function in the developing palate are still unknown. Connective tissue growth factor (CTGF, also known as CCN2) is a potential candidate to mediate TGF-β signaling during palatogenesis.

CTGF is a matricellular protein involved in the regulation of cell survival, proliferation, adhesion, migration, and ECM production (13). Interestingly, TGF-β is able to induce Ctgf expression in diverse cell types, predominantly in fibrotic disorders (14–17), and Ctgf contains a TGF-β response element in its promoter (18). At the protein level, CTGF possesses four domains that interact with growth factors, ECM proteins, low-density lipoprotein (LDL)-receptor related proteins (LRP), and integrins (13, 19–21). Specifically, it contains a von Willebrand type C domain, which is thought to mediate physical interactions with members of the TGF-β superfamily, including bone morphogenetic proteins (BMPs) and TGF-β ligands. Consistent with this, CTGF binds to BMPs and TGF-β ligands, leading to inhibition of BMP and enhancement of TGF-β signaling (19). Moreover, Ctgf deficiency leads to skeletal dysmorphology, as a result of impaired cartilage development, and cleft palate in mice (22) with 100% phenotype penetrance. This observation supports the hypothesis that CTGF exerts an important role during palate development. In this study, we analyze the regulation of Ctgf expression by TGF-β signaling in the developing palate and the impact of CTGF on crucial cellular activities such as mesenchymal cell proliferation.

MATERIALS AND METHODS

Generation of transgenic mice.

Mating; Tgfbr2^fl/+ male mice with Tgfbr2^fl/fl female mice generated Wnt1-Cre; Tgfbr2^fl/fl mice. Mating Osr2-Cre with R26R mice generated Osr2-Cre; R26R embryos. Osr2-Cre; Smad4^fl/+ male mice were crossed with Smad4^fl/fl female mice to generate Osr2-Cre; Smad4^fl/fl mutant mice. Animal usage was approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Southern California.

Histological analysis.

For histological analysis, samples were fixed in 4% paraformaldehyde (PFA) and processed into paraffin-embedded serial sections using routine procedures. For general morphology, deparaffinized sections were stained with hematoxylin and eosin (H&E) using standard procedures. For X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) staining and detection of β-galactosidase (β-Gal) activity in Osr2-Cre; R26R mice, samples at various stages of embryonic development were fixed in 0.2% glutaraldehyde, passed through a sucrose series, embedded in OCT compound (Tissue-Tek), and sectioned on a cryostat at 10 μm prior to X-Gal staining for lacZ expression. Detection of β-Gal activity in tissue sections was performed as previously described (23).

In situ hybridization.

The expression patterns of Ctgf, Msx1, Bmp4, and Shh in the palate were examined by in situ hybridization using digoxigenin-labeled antisense probes as previously described (12). Negative controls were run in parallel with the experimental reaction. Cryosections or dissected palatal shelves were pretreated with proteinase K (Sigma), refixed in fresh 4% PFA, and then hybridized overnight at 55°C. After hybridization, tissue sections or dissected palatal shelves were washed under high-stringency conditions and preblocked in antibody blocking solution and then incubated with preabsorbed antibody. Color development was performed in nitroblue tetrazolium (NBT)–5-bromo-4-chloro-3′-indolylphosphate (BCIP) color development solution (Sigma). Images were processed using ImageJ. Briefly, images were split into three channels (red, blue, and green). Red channel images were used for further processing. Threshold values were changed to segment and pseudocolor the images using the same parameters for control and mutant samples.

Organ culture of palates and CTGF bead implantation.

Timed-pregnant mice were sacrificed at embryonic day 13.5 (E13.5). The palatal shelves were microdissected and cultured in serum-free chemically defined medium as previously described (7). To test epithelial degeneration and fusion, palatal shelves were put in contact at the midline and cultured for 3 days. Afterwards, they were fixed in 4% PFA and prepared for H&E staining. Heparin beads (Sigma) were used for delivery of CTGF protein as described previously (24). Briefly, the beads were washed in phosphate-buffered saline (PBS) and then incubated for 1 h at room temperature in 20 ng/ml CTGF (Cell Sciences). Control beads were incubated in 0.1% bovine serum albumin (BSA). CTGF- or BSA-containing beads were implanted into palatal explants. After 24 h of incubation, palatal shelves were fixed in 4% PFA and subsequently prepared for immunostaining.

Analysis of cell proliferation and apoptosis.

Cell proliferation was monitored by intraperitoneal BrdU (5-bromo-2-deoxyuridine; Sigma) injection (100 mg/g body weight) at E13.5 and E14.5. One hour after injection, mice were sacrificed and embryos were fixed in 4% PFA and processed. To analyze proliferation in vitro in palatal explants, BrdU was added to the culture medium for 1 h followed by fixation of the palatal shelves in 4% PFA. Serial sections of the specimen were cut at 5-μm intervals. Detection of BrdU-labeled cells was carried out using a BrdU labeling and detection kit according to the manufacturer's protocol (Boehringer Mannheim). BrdU-positive cells and total number of cells within the palatal mesenchyme were counted from five randomly selected sections per sample. Five palate samples were evaluated from each experimental group. Apoptosis assays were performed using caspase-3 immunostaining according to the manufacturer's protocol. Sections were counterstained with hematoxylin or DAPI (4′,6-diamidino-2-phenylindole).

Treatment of primary cultured cells derived from MEPM.

Primary embryonic palatal mesenchyme cells were obtained from E13.5 embryos. Briefly, palatal shelves were dissected and trypsinized for 30 min at 37°C in a CO₂ incubator. After pipetting thoroughly, cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum supplemented with penicillin, streptomycin, l-glutamate, sodium pyruvate, and nonessential amino acids. Under these culture conditions, palatal epithelial cells do not survive or grow. Mouse embryonic palatal mesenchyme (MEPM) cells were treated with CTGF (100 ng/ml), TGF-β1 (10 ng/ml), p38 inhibitor (SB203580 at 10 μM), or a combination (see Results). After 24 h of treatment (or indicated times), cells were harvested and processed for RNA or protein extraction.

qPCR analysis.

Total RNA was isolated from dissected palatal shelves or from MEPM cells using RNeasy minikits (Qiagen). The QuantiTect reverse transcription kit (Qiagen) was used for cDNA synthesis. RNA was quantified and tested for quality by photometric measurement. Quantitative PCR (qPCR) was carried out on the iCycler (Bio-Rad) with gene-specific primers and SYBR green (Bio-Rad). A melting curve was obtained for each PCR product after each run to confirm that the SYBR green signal corresponded to a unique and specific amplicon. Values were normalized to glyceraldehyde3-phosphate dehydrogenase (GAPDH) using the threshold cycle (2^−ΔΔCT) method (25). PCR primers are available upon request.

Western blotting.

The total protein concentration in the palates was determined by comparison with BSA standards. Twenty micrograms of total protein from each sample was loaded in each well of a 12% polyacrylamide gel. Western analysis was carried out as previously described (26). Antibodies used were phospho-SMAD2, phospho-SMAD3, phospho-p38, p38, phospho-ERK, ERK, phospho-JNK, and JNK (Cell Signaling Technologies) and GAPDH or β-actin (Millipore).

Bioinformatic analysis.

Analysis of the Ctgf promoter region and transcription factor binding sites (TFBS) was conducted using Genomatix MatInspector (Genomatix Software GmbH, Munich, Germany), which utilizes a large library of matrix descriptions for TFBS. For details on the method, see the work of Quandt et al. (27).

Micro-CT scanning and 3D reconstruction.

Control and Osr2-Cre; Smad4^fl/fl newborn mice were sacrificed, and heads were fixed in 4% PFA on a shaker at 4°C overnight. The data set of the skull was acquired using a micro-computed tomography (micro-CT) system (Scanco Medical; V1.2a) with the following parameters: 70 kVp, 0.02-mm-per-slice thickness, resolution at 50 pixels per mm, and 16 bits per pixel. Voxels from each skull were assigned an x, y, and z coordinate in three-dimensional (3D) space using Transform Editor parameters in Avizo 7 software (Visualization Sciences Group Inc., Burlington, MA, USA). Visualization and 3D micro-CT reconstruction of the skull were performed using Isosurface parameters in Avizo 7.

Skeletal staining.

The three-dimensional architecture of the craniofacial skeleton of Osr2-Cre; Smad4^fl/fl and control mice was examined using a modified whole-mount alcian blue-alizarin red S staining protocol. Briefly, control and mutant newborns were fixed in 4% PFA overnight, followed by a 1-hour wash in double-distilled water (ddH₂O) with gentle shaking and postfixation in 70% ethanol. The skin and internal organs were carefully removed from the samples at this step. The skeletons were stained with 0.02% alcian blue 8GX for 2 days. The samples were washed with ethanol-glacial acetic acid (7:3) for 1 h. Afterwards, they were soaked in 100% ethanol overnight and then in ddH₂O for 1 day. Once the cartilage was clearly observed, alizarin red staining was performed overnight. Finally, the samples were treated with a series of KOH-glycerol and stored in glycerol with a crystal of thymol.

Statistical analysis.

The two-tailed Student t test was applied for statistical analysis. For all graphs, data are represented as means ± standard deviations (SD). A P value of less than 0.05 was considered statistically significant.

RESULTS

Ctgf expression pattern suggests specific functions during palatogenesis.

In order to investigate the roles of Ctgf in regulating palate development, we analyzed its expression pattern in mouse embryos from E12.5 to E16.5 by in situ hybridization. Ctgf mRNA expression was detectable at sites of endochondral ossification such as the cranial base. Ctgf was also expressed in the tooth bud and Meckel's cartilage, as previously reported (22, 28, 29). Ctgf was detectable in the developing palate from E13.5, before reorientation of the palatal shelves, to E15.5, when palatal shelves are fused at the midline (Fig. 1B to F). At E13.5, Ctgf was weakly expressed in the palatal epithelium and no signal was detectable in the mesenchyme (Fig. 1B). Subsequently, it was detectable in both the midline epithelium and the mesenchyme at E14.0 and E14.25 (Fig. 1C and D). At E14.5, the expression of Ctgf was no longer detectable in the remaining epithelial cells and it was restricted to the midline palatal mesenchyme (Fig. 1E). At E15.5, Ctgf was detectable in the palatal mesenchyme (Fig. 1F). Expression of Ctgf was not detectable in the palate at E12.5 or E16.5. We confirmed these expression patterns using semiquantitative PCR, with β-actin as an internal control (Fig. 1H).

Fig 1.

Expression pattern of Ctgf during palatogenesis. (A to G) In situ hybridization of Ctgf in the palatal region of control mice at E12.5 to E16.5. Ctgf expression is detectable in Meckel's cartilage at E12.5 (A; arrow). At E13.5 (B), Ctgf is expressed in the palatal epithelium (arrow) and the tooth bud (arrowheads). Expression of Ctgf is detectable in the palatal epithelium (arrowheads) and mesenchyme (arrow) at E14.0 (C) and E14.25 (D). At E14.5 (E) and E15.5 (F), Ctgf expression is restricted to the mesenchyme (arrow). (G) No expression of Ctgf is detectable at E16.5 in the palate. (H) Semiquantitative analysis of Ctgf expression. Bar, 50 μm. Abbreviations: MC, Meckel's cartilage; T, tongue.

Ctgf is a downstream target of the Smad-dependent TGF-β signaling pathway in the developing palate.

Previous studies have demonstrated that TGF-β is a potent inducer of Ctgf expression in diverse cell types (24, 30), but the mechanism(s) involved in TGF-β regulation of Ctgf expression during palatogenesis is unknown. In order to address this issue, we used both in vivo and in vitro approaches. First, we analyzed Wnt1-Cre; Tgfbr2^fl/fl embryos, which lack Tgfbr2 in the neural crest-derived mesenchyme (7). Based on in situ hybridization analysis, Ctgf expression in Tgfbr2 mutant mice was indistinguishable from that of control mice at E13.5 (Fig. 2A and B). This finding suggests that epithelial expression of Ctgf is independent of the activity of TGF-β signaling pathway in the mesenchyme. At E14.5, Ctgf expression was significantly downregulated in the palatal mesenchyme of Wnt1-Cre; Tgfbr2^fl/fl mice, whereas Ctgf epithelial expression persisted (Fig. 2D and E). We confirmed these results by quantifying Ctgf expression in cultured MEPM cells from control and Wnt1-Cre; Tgfbr2^fl/fl embryos (Fig. 2G). For this experiment, we used culture conditions in which palatal epithelial cells do not survive or grow, so that Ctgf expression values correspond to those of mesenchymal cells.

Fig 2.

Ctgf expression is regulated by the canonical TGF-β pathway. (A to F) In situ hybridization of Ctgf in control, Wnt1-Cre; Tgfbr2^fl/fl, and Osr2-Cre; Smad4^fl/fl mice at E13.5 and E14.5. The inset in panel D shows Ctgf expression in control embryos at E14.5 in which palatal shelves are not in contact at the midline. Arrowheads indicate expression of Ctgf in the palatal epithelium, the arrow points to Ctgf expression in the palatal mesenchyme, and dashed lines delineate the tooth buds. (G) qPCR analysis of Ctgf in the palates of Wnt1-Cre; Tgfbr2^fl/fl mice (Tgfbr2 CKO) compared with control mice at E14.5. (H) qPCR analysis of Ctgf in the palates of Osr2-Cre; Smad4^fl/fl mice (Smad4 CKO) compared with control mice at E14.5. (I) qPCR analysis of Ctgf expression in control and Tgfbr2 CKO MEPM cells untreated (0′) or treated with TGF-β1 for indicated time periods (5 to 60 min). *, P < 0.05. Error bars, SD. Abbreviations: T, tongue; CKO, conditional knockout.

To confirm specific Ctgf induction by TGF-β signaling in the developing palate, we treated control and Tgfbr2 mutant MEPM cells with TGF-β1 ligand for 5 to 60 min and compared them with untreated samples. Quantitative PCR analysis showed a continuous increase in Ctgf expression in control cells treated with TGF-β1 ligand over time (Fig. 2I). After 5 min of treatment, a 1.2-fold increase was already detectable, which suggests a direct regulation of Ctgf expression by TGF-β signaling. Ctgf expression increased more than 100% in MEPM cells after 1 h of treatment compared to untreated cells (Fig. 2I). In an independent experiment shown in Fig. 3B, treatment of control MEPM cells for 24 h with TGF-β1 caused an 8-fold increase in Ctgf expression. No significant change in Ctgf expression was detectable in Tgfbr2 mutant MEPM cells treated with TGF-β1 ligand for different periods of time (Fig. 2I), suggesting that a functional TGFBRII receptor is necessary for the induction of Ctgf expression by TGF-β1 ligand. In addition, we treated MEPM cells with cycloheximide to determine whether the induction of Ctgf by TGF-β1 can occur without new protein synthesis. The results suggest that protein synthesis is not required for Ctgf upregulation after exposure to TGF-β1 ligand (data not shown). Thus, Ctgf is likely a direct downstream target of TGF-β signaling in the palatal mesenchyme.

Fig 3.

Upregulation of p38 MAPK in Wnt1-Cre; Tgfbr2^fl/fl mice has an inhibitory effect on Ctgf expression. (A) Bioinformatics analysis of the murine Ctgf promoter region. Predicted binding sites for transcription factors related to the canonical and noncanonical pathways (p38 MAPK, ERK, and JNK) of TGF-β signaling are depicted in different colors. (B) qPCR analysis of Ctgf expression in primary MEPM cells from control embryos untreated (control) or after treatment with p38 MAPK inhibitor (p38 inh), TGF-β1, or p38 MAPK inhibitor combined with TGF-β1 (TGF-β1 + p38 inh). (C) qPCR analysis of Ctgf expression in primary MEPM cells from control and Wnt1-Cre; Tgfbr2^fl/fl (Tgfbr2 CKO) embryos treated with p38 MAPK inhibitor (p38 inh) or untreated. *, P < 0.05. CKO, conditional knockout.

To determine whether Ctgf expression induced by TGF-β in the palate is dependent on the canonical Smad-dependent pathway, we examined Ctgf expression in Osr2-Cre; Smad4^fl/fl palates by in situ hybridization. In these conditional mutants, loss of Smad4 is restricted to mesenchymal cells in the craniofacial region (see below for details on these mice). Ctgf was expressed in the epithelium in both control and Osr2-Cre; Smad4^fl/fl mice at E13.5 (Fig. 2A and C); however, we detected reduced expression of Ctgf in the palatal mesenchyme of E14.5 Osr2-Cre; Smad4^fl/fl mice compared with control (Fig. 2D and F). These findings were confirmed by qPCR analysis (Fig. 2H). Taken together, our results indicate that Smad-dependent TGF-β signaling is required for Ctgf expression in the palatal mesenchyme.

We performed a bioinformatics analysis of the Ctgf promoter region, which revealed the presence of not only one potential Smad-binding site but also various sites for transcription factors used by TGF-β noncanonical pathways (Fig. 3A) (31, 32). Noncanonical TGF-β pathways, including p38 MAPK, ERK, and JNK, are also involved in the regulation of the expression of downstream targets in various cell types (33). We focused our analysis on p38 MAPK because its activity is upregulated in Wnt1-Cre; Tgfbr2^fl/fl palates (10), in which Ctgf is downregulated. To determine whether p38 MAPK is involved in the regulation of Ctgf expression by TGF-β, we treated MEPM cells from control embryos with an inhibitor of p38 MAPK (SB203580), with TGF-β1 alone, or with TGF-β1 in combination with p38 MAPK inhibitor. qPCR analyses showed that the expression of Ctgf in MEPM cells treated with p38 MAPK inhibitor alone was slightly downregulated compared with that in untreated cells (Fig. 3B). This downregulation might be due to the disruption of TGF-β signal transduction but could also be due to blockage of other signaling pathways that use p38 MAPK as an intracellular transducer, such as fibroblast growth factor (FGF), BMP, or epidermal growth factor (EGF). Therefore, we compared Ctgf expression levels in MEPM cells treated with TGF-β1 alone and with TGF-β1 plus p38 MAPK inhibitor. Interestingly, Ctgf relative expression levels were upregulated under these two conditions compared to those in untreated cells and were not significantly different from each other. These findings suggest that induction of Ctgf expression by TGF-β is independent of p38 MAPK and that regulation of Ctgf expression depends in part on other signaling pathway mediators (Fig. 3B).

Next, we treated Wnt1-Cre; Tgfbr2^fl/fl MEPM cells with p38 MAPK inhibitor to determine whether a decrease of the elevated p38 MAPK activity could rescue Ctgf expression. Indeed, treatment with p38 MAPK inhibitor resulted in a significant upregulation of Ctgf expression levels compared to those in untreated Wnt1-Cre; Tgfbr2^fl/fl cells (Fig. 3C). This result indicates that elevated p38 MAPK activity inhibits Ctgf expression in Wnt1-Cre; Tgfbr2^fl/fl palates, which might be part of the pathological mechanism leading to defective palatal mesenchymal cell proliferation.

CTGF regulates its own expression via the TGF-β signaling pathway.

Substantial evidence suggests that CTGF enhances TGF-β pathway transduction through binding to TGF-β ligands (19). Therefore, we investigated whether CTGF might be upstream of TGF-β signaling and modulate its own expression in the developing palate. We treated Wnt1-Cre; Tgfbr2^fl/fl and control MEPM cells with CTGF recombinant protein for 24 h and performed qPCR analyses. In control cells, Ctgf expression was upregulated after treatment with CTGF (Fig. 4A), consistent with a positive feedback mechanism. In contrast, CTGF treatment of Wnt1-Cre; Tgfbr2^fl/fl cells failed to result in any effect on Ctgf expression compared to that in untreated Wnt1-Cre; Tgfbr2^fl/fl cells. Thus, TGF-β signaling appears to be required by CTGF to regulate its own expression. Furthermore, we examined whether other downstream targets of TGF-β signaling might also be upregulated in response to CTGF treatment. Addition of CTGF to control cells resulted in induction of expression of PAI-1 (Fig. 4B), a previously identified TGF-β downstream target, suggesting a general enhancement of the TGF-β signaling pathway by CTGF. To test this hypothesis, we evaluated phosphorylation levels of SMAD2 and SMAD3 in CTGF-treated MEPM cells versus untreated cells by Western blotting and found that phosphorylation of SMAD2 but not of SMAD3 was upregulated (Fig. 4C). These findings indicate that CTGF enhances specific aspects of TGF-β signaling in the palatal mesenchyme.

Fig 4.

CTGF regulates its own expression through the TGF-β signaling pathway. (A) qPCR analysis of Ctgf expression in primary MEPM cells from control and Wnt1-Cre; Tgfbr2^fl/fl (Tgfbr2 CKO) embryos treated with CTGF recombinant protein compared with untreated cells. (B) qPCR analysis of PAI-1 expression in primary MEPM cells from control and Tgfbr2 CKO embryos treated with CTGF compared with untreated cells. (C) Immunoblotting analysis of phosphorylated SMAD2 (P-SMAD2) and SMAD3 (P-SMAD3) in control MEPM treated with CTGF compared with untreated control MEPM cells. *, P < 0.05. CKO, conditional knockout.

Conditional inactivation of Smad4 disturbs palatal mesenchymal cell proliferation.

As described above, Ctgf is downregulated in Osr2-Cre; Smad4^fl/fl mice. We analyzed these mice in greater detail in order to understand the function of Smad4 in palate development and its connection with Ctgf expression in the context of TGF-β signaling. Smad4 mutation in these mice is restricted to the CNCC-derived mesenchyme as revealed by LacZ staining in Osr2-Cre; R26R mice. β-Gal expression is detectable from E10.5 in the first branchial arch mesenchyme and later in its derivatives. In the palate, β-Gal expression is homogeneously distributed in the mesenchyme (Fig. 5). Osr2-Cre; Smad4^fl/fl mice die a few hours after birth (Fig. 6A) and display complete cleft palate (Fig. 6B to G). Palatal clefting was first detectable at E14.5 (Fig. 6L and M). We found that the palatal shelves in Osr2-Cre; Smad4^fl/fl mice are able to reorient normally from a vertical to a horizontal position but cannot establish contact in the midline (Fig. 6J to O).

Fig 5.

Expression patterns of Osr2 during palatogenesis. LacZ staining in Osr2-Cre; R26R embryos from E12.5 to E14.5 and newborn (NB). Right panels are higher magnifications of the regions boxed in the left panels.

Fig 6.

Disruption of Smad4 in cranial neural crest cells leads to cleft palate. (A) Control and Osr2-Cre; Smad4^fl/fl newborn mice. Note the absence of milk in the stomach of the Smad4 CKO mouse (arrow). CKO, conditional knockout. (B and C) Oral view of the palates of newborn control and Osr2-Cre; Smad4^fl/fl mice. Arrows indicate the midline of palatal fusion (B) or the edge of the palatal shelf (C). (D and E) Oral view of 3D reconstructions of micro-CT scans of control and Osr2-Cre; Smad4^fl/fl skulls. The palate and related structures are depicted in different colors: premaxilla, yellow; maxillary bone, green; palatine bone, red; maxillary palatine process, purple; and palatine process of the palatine bone, pink. (F and G) Alcian blue and alizarin red staining of control and Osr2-Cre; Smad4^fl/fl palates. (H to O) H&E staining of coronal sections of E12.5 (H and I), E13.5 (J and K), E14.5 (L and M), and newborn (N and O) palates. Arrowhead, palatal shelf or the palate at newborn stage. Abbreviations: M, maxillary bone; NB, newborn; P, palatine bone; PM, premaxilla; pppl, palatal process of the palatine; ps, presphenoid.

Previous studies have shown that TGF-β signaling is crucial for regulating midline epithelial cell disappearance during palatal fusion (6). To determine whether this process is altered in Osr2-Cre; Smad4^fl/fl mice, we analyzed palatal fusion in a palatal shelf organ culture model. At E13.5, the developing palatal shelves were pointing downwards on both sides of the tongue. Each isolated pair of palatal shelves was dissected and placed in culture with the two segments just touching at the medial edge and kept in the original anterior-posterior orientation. During the 3-day culture period, both control and Osr2-Cre; Smad4^fl/fl mutant palatal explants were able to fuse completely, with normal degeneration of the midline epithelium (Fig. 7A and B). Moreover, osteoid-like structures were present in the cultured palatal shelves from both control and Osr2-Cre; Smad4^fl/fl embryos, suggesting that palatal bone formation was initiated in vitro.

Fig 7.

Cleft palate in Osr2-Cre; Smad4^fl/fl mice is due to decreased mesenchymal cell proliferation. (A and B) In vitro palatal fusion analysis of control and Osr2-Cre; Smad4^fl/fl palatal explants. Arrowheads point to palatal midline, where no epithelial cells remain in either control or Osr2-Cre; Smad4^fl/fl explants. Two arrows point to osteoid-like structures. (C to F) Active caspase-3 immunostaining assays of control and Osr2-Cre; Smad4^fl/fl palates at E13.5 and E14.5. (G to J) BrdU incorporation experiments of control and Osr2-Cre; Smad4^fl/fl palates at E13.5 and E14.5. (K) Quantitation of the percentage of BrdU-labeled nuclei in control and Osr2-Cre; Smad4^fl/fl (Smad4 CKO) palates from panels G to J. *, P < 0.05. Error bars, SD. CKO, conditional knockout.

To explore the mechanism responsible for cleft palate in Osr2-Cre; Smad4^fl/fl embryos, we investigated whether there was an increase in programmed cell death, a decrease in cell proliferation, or a combination of both in the CNCC-derived palatal mesenchyme. Cell death was evaluated by caspase-3 immunostaining. We detected no difference in the number of apoptotic cells between Osr2-Cre; Smad4^fl/fl and control palates at E13.5 and E14.5 (Fig. 7C to F). Cell proliferation activity within the palatal mesenchyme, as measured by BrdU incorporation, was indistinguishable in control and Osr2-Cre; Smad4^fl/fl embryos at E13.5 (Fig. 7G, H, and K). In contrast, there was a significant decrease in the number of proliferating cells in the palatal mesenchyme of Osr2-Cre; Smad4^fl/fl embryos compared with controls at E14.5 (Fig. 7I to K). Interestingly, proliferation is also reduced in Wnt1-Cre; Tgfbr2^fl/fl palates starting at E14.5 (10). The decrease in proliferation likely contributes to the small size of the palatal shelves in newborn Smad4 mutant mice. Our findings indicate that the canonical Smad-dependent TGF-β pathway controls cell cycle progression in the palatal mesenchyme around the time that palatal fusion should be occurring.

TGF-β signaling, but not BMP signaling, is altered in Osr2-Cre; Smad4^fl/fl embryos.

TGF-β and BMP are both involved in the regulation of mesenchymal cell proliferation during palate development (7, 34), and both pathways use SMAD4 as a common mediator of signal propagation. The downregulation of Ctgf in Osr2-Cre; Smad4^fl/fl palates at E14.5 (Fig. 2) is consistent with alteration of TGF-β pathway transduction. In order to determine whether BMP signaling pathways were also affected in Osr2-Cre; Smad4^fl/fl palates, we evaluated the expression of BMP downstream targets Msx1 and Bmp4 by in situ hybridization and quantitative PCR. We detected no change in the expression pattern or level of these genes in Osr2-Cre; Smad4^fl/fl palates compared to controls at E14.5 (Fig. 8A to D and G). BMP4 exerts its functions in part through induction of Shh expression in the palatal epithelium (34). We analyzed Shh expression by whole-mount in situ hybridization but detected no difference in Shh expression in control and Osr2-Cre; Smad4^fl/fl mice at E14.5 (Fig. 8E and F).

Fig 8.

BMP transduction, epithelial-mesenchymal interactions, and noncanonical TGF-β/BMP pathways are not altered in Osr2-Cre; Smad4^fl/fl palates. (A to F) In situ hybridization of Msx1 (A and B), Bmp4 (C and D), and Shh (E and F) in Osr2-Cre; Smad4^fl/fl and control mice at E14.5. Arrows indicate positive signal. (G) qPCR analysis of Msx1 and Bmp4 expression in control and Osr2-Cre; Smad4^fl/fl (Smad4 CKO) palates. (H) Immunoblotting analysis of the nonphosphorylated and phosphorylated (P-) forms of p38 MAPK, ERK, and JNK in control and Smad4 CKO MEPM cells. CKO, conditional knockout.

To examine whether Ctgf downregulation in palates lacking Smad4 was caused by overactivation of noncanonical TGF-β pathways, we analyzed expression levels of phosphorylated and nonphosphorylated forms of p38 MAPK, ERK, and JNK by Western blotting at E14.5. We detected no significant differences between Osr2-Cre; Smad4^fl/fl and control palates (Fig. 8H). Taken together, our results suggest that the defect in mesenchymal cell proliferation in Osr2-Cre; Smad4^fl/fl palates is due to defective transduction of TGF-β, but not BMP, signals via the canonical pathway.

CTGF rescues the defective mesenchymal proliferation of Wnt1-Cre; Tgfbr2^fl/fl and Osr2-Cre; Smad4^fl/fl palates.

Numerous studies have reported that CTGF promotes proliferation in various cell types in adult tissues (35–37, 62). Given our finding that Ctgf is a downstream target of the TGF-β canonical pathway in the developing palate, we hypothesized that CTGF might be able to rescue the proliferative defect observed in both Wnt1-Cre; Tgfbr2^fl/fl and Osr2-Cre; Smad4^fl/fl palates. We implanted CTGF- and BSA-embedded beads in E13.5 control, Wnt1-Cre; Tgfbr2^fl/fl, and Osr2-Cre; Smad4^fl/fl palatal shelves in vitro and incubated them for 24 h. There was no difference in the number of proliferating cells between control palatal explants treated with CTGF-embedded beads and those treated with BSA-embedded beads (Fig. 9A, B, and G). However, more BrdU-positive nuclei were detectable close to CTGF-embedded beads than close to BSA beads in both Wnt1-Cre; Tgfbr2^fl/fl and Osr2-Cre; Smad4^fl/fl palates (Fig. 9C to G). Statistical analysis showed that exogenous CTGF was able to increase cell proliferation significantly in Wnt1-Cre; Tgfbr2^fl/fl and Osr2-Cre; Smad4^fl/fl palatal mesenchyme. These results are consistent with Ctgf acting as a downstream mediator of the canonical pathway of TGF-β signaling to regulate cell proliferation during palate development.

Fig 9.

CTGF rescues the proliferative defect in Wnt1-Cre; Tgfbr2^fl/fl and Osr2-Cre; Smad4^fl/fl palates. (A to F) BrdU incorporation experiments after treatment of control, Wnt1-Cre; Tgfbr2^fl/fl, and Osr2-Cre; Smad4^fl/fl palate explants with CTGF- or BSA-embedded beads. Arrowheads indicate BrdU-positive cells. Dashed lines outline the beads. (G) Quantitation of the percentage of BrdU-labeled nuclei in control, Wnt1-Cre; Tgfbr2^fl/fl (Tgfbr2 CKO), and Osr2-Cre; Smad4^fl/fl (Smad4 CKO) palate explants from panels A to F. *, P < 0.05. Error bars, SD. CKO, conditional knockout.

DISCUSSION

Regulation of Ctgf expression.

In this study, using both in vivo and in vitro approaches, we have found that Ctgf is a downstream target of TGF-β signaling that regulates proliferation in mesenchymal cells during palatogenesis. Recently, we demonstrated that Fgf9 and Pitx2 are also downstream elements involved in the mediation of the TGF-β signaling pathway during palate development (26). The expression patterns of Ctgf indicate that it might be involved in several different steps during palate development, including palatal reorientation, epithelial fusion, and degeneration, as well as mesenchymal proliferation. We found that Ctgf expression in the palatal mesenchyme is induced by TGF-β1 ligand in vitro after 5 min of exposure and increases over time, which suggests a direct transcriptional regulation by TGF-β-related transcription factors. Additionally, Ctgf is downregulated in Wnt1-Cre; Tgfbr2^fl/fl mice, which have defective mesenchymal proliferation. We also found that expression of Ctgf in the epithelium of the developing palate is independent of mesenchymal TGF-β signaling at both E13.5 and E14.5. Persistence of Ctgf expression in the epithelium of Wnt1-Cre; Tgfbr2^fl/fl mice at E14.5 is reminiscent of an E13.5 wild-type palate. Overall, these findings suggest a cell-autonomous mechanism for the regulation of Ctgf expression by TGF-β in the palatal mesenchyme. We cannot exclude the possibility that other signaling pathways also control the expression of Ctgf in the epithelium.

We investigated whether TGF-β induces Ctgf expression through the Smad-dependent pathway or through activation of a noncanonical pathway in the palatal mesenchyme. In vitro studies have shown that Smads are essential for the induction of expression of ECM proteins, including CTGF (63). However, other conflicting results suggest that MAPKs stimulate or repress Ctgf expression depending on the cell type (39–44). Our analysis of Osr2-Cre; Smad4^fl/fl embryos demonstrated that Ctgf expression in the palatal mesenchyme is Smad dependent. Our recent studies showed that loss of Tgfbr2 in CNCC-derived palatal mesenchyme results in an overactivation of an alternative TGF-β signaling cascade that involves the upregulation of p38 MAPK (10). Inhibition of p38 MAPK in the palatal mesenchyme of Wnt1-Cre; Tgfbr2^fl/fl mice resulted in an increase of Ctgf expression and rescued cell proliferation, consistent with an inhibitory effect of p38 MAPK on Ctgf expression. Thus, Ctgf downregulation in Tgfbr2 mutant mice might be an important aspect of the pathological mechanism leading to decreased cell proliferation and cleft palate. We have previously reported a similar effect of p38 MAPK on Fgf9 and Pitx2 expression in Tgfbr2 mutant mice (26). However, p38 MAPK is not upregulated in Osr2-Cre; Smad4^fl/fl mice, but Ctgf is still downregulated, suggesting that Ctgf downregulation in Wnt1-Cre; Tgfbr2^fl/fl mice might be the result of a combined effect of reduction of Smad activity and overactivation of p38 MAPK.

Interestingly, we also found that CTGF is able to regulate its own expression in mesenchymal cells during palatogenesis and that this effect is TGF-β dependent. Addition of CTGF to MEPM cells produced an increase in phosphorylation of SMAD2 but not in SMAD3. This finding indicates that CTGF acts as a modulator/enhancer of specific aspects of the TGF-β signaling pathway. These results also suggest that Ctgf expression induced by itself might depend primarily on SMAD2 activation rather than that of SMAD3 in palatal mesenchymal cells. This mechanism might be stage and tissue dependent.

Canonical TGF-β signaling in palate development.

The regulation of palate development depends on a network of factors, including growth factors, ECM proteins, cell adhesion molecules, and transcription factors (2, 61). Disturbance of this tightly controlled network may result in failure of the multistep process leading to the formation of a normal palate. TGF-β signaling is critical in palatogenesis because mutations in diverse members of the pathway lead to severe malformations in palatal structures. This current study is part of our general strategy to identify TGF-β downstream signaling molecules that are crucial for palatogenesis.

Our analysis of two conditional mutant strains, the Wnt1-Cre; Tgfbr2^fl/fl and Osr2-Cre; Smad4^fl/fl strains, provides us with the opportunity to examine two components of the complex TGF-β signaling pathway in palate development. Mutation of Smad4 in Osr2-expressing cells recapitulates the defect in cell proliferation observed in the developing palates of mice with neural-crest conditional inactivation of Tgfbr2 (7), consistent with the role of SMAD4 in mediating TGF-β signals. The defective mesenchymal proliferation in Osr2-Cre; Smad4^fl/fl mice is associated with severe complete cleft palate. SMAD4 not only is a mediator of TGF-β but also transduces BMP signals. However, in this study we demonstrated that the expression of BMP downstream elements, such as Msx1 and Bmp4, is not altered in the developing palate of Smad4 conditional mutants. We also showed that noncanonical BMP pathways are not overactivated in Osr2-Cre; Smad4^fl/fl mice. These results suggest that a more complex mechanism than the one previously reported is used to transduce BMP signals and, more importantly, that the phenotype observed in Smad4 mutant mice is due to altered TGF-β signaling.

Osr2-Cre; Smad4^fl/fl mice have the potential to become a great model to analyze the function of Smad4 throughout the course of craniofacial development. Because Osr2 is expressed in the CNCC-derived mesenchyme of the first branchial arch beginning only at E10.5, Osr2-Cre; Smad4^fl/fl mice overcome the early embryonic lethality displayed by both Smad4-null mice and Wnt1-Cre; Smad4^fl/fl mice (38, 45). In humans, SMAD4 mutations seem to be associated with cancer rather than with developmental pathologies (46), perhaps due to embryonic lethality such as that in conventional Smad4 mutant mice (38). Nevertheless, heterozygous mutations of human TGFBR1 or TGFBR2 are associated with Loeys-Dietz syndrome. Patients with this syndrome display craniofacial malformations such as cleft palate, craniosynostosis, hypertelorism, skeletal defects, and vascular problems (33, 47).

On the other hand, gain of function of TGF-β also causes pathological conditions in humans. Overexpression of TGF-β ligands in adult organs has been associated with fibrotic processes, which involve exaggerated ECM production (30, 48–50). One of the activities of TGF-β in adult tissues is the induction of ECM proteins (51), and among them, CTGF appears to be a crucial mediator of collagen production (52, 53). Although evidence of Ctgf induction by TGF-β was found in fibrotic disorders (reviewed in reference 54), here we have described CTGF as a mediator of TGF-β activity during palate development.

Role of CTGF in mesenchymal proliferation in the developing palate.

CTGF is a member of the CCN family of matricellular proteins and as such is a contextual modulator of a diverse array of cellular functions (54, 55). CTGF has been implicated in complex biological processes such as angiogenesis, chondrogenesis, and osteogenesis, as well as fibrosis and tumorigenesis (22, 54). Specifically, CTGF induces proliferation in diverse cell types, including chondrocytes (56, 57), osteoblasts (51), human bone marrow stromal cells (58), lung fibroblasts (59), and beta cells (60), among others.

Our in vitro rescue experiments showed that CTGF is able to restore cell proliferation in both Wnt1-Cre; Tgfbr2^fl/fl and Osr2-Cre; Smad4^fl/fl palates. This result, together with the demonstration that Ctgf is a downstream target of TGF-β signaling, suggests that CTGF acts as mediator of the canonical pathway of TGF-β signaling to regulate cell proliferation during palatogenesis. Although we demonstrated that CTGF requires a functional TGF-β pathway to regulate its own expression, the mechanism of induction of proliferation by CTGF must be different. In this case, CTGF must use a pathway other than TGF-β because Wnt1-Cre; Tgfbr2^fl/fl and Osr2-Cre; Smad4^fl/fl mesenchymal cells are able to proliferate after stimulation with exogenous CTGF. Accordingly, previous studies have shown that CTGF regulates proliferation of myofibroblasts through activation of p44/42 MAPK/extracellular-signal regulated kinase (ERK) (35) and of human lung fibroblasts through ERK (57) and phosphatidylinositol 3-kinase (PI3K)/protein kinase B (PKB) (17).

Our findings are consistent with a previous study in which we demonstrated that CTGF acts downstream of TGF-β signaling to regulate mandibular morphogenesis (24). In addition, Ctgf-null mice have cleft palate, micrognathia, and cranial base defects (reference 22 and our unpublished results). These phenotypes are similar to those observed in Wnt1-Cre; Tgfbr2^fl/fl mice. Therefore, a TGF-β/CTGF signaling cascade appears to be a well-conserved signaling mechanism in regulating craniofacial development. Altered Ctgf expression might be used as an indicator for loss of TGF-β signaling in diagnosis. Modulation of CTGF signaling in TGF-β mutant models may offer a potential opportunity for prevention or rescue of congenital birth defects.