Mesenchyme homeobox 1 mediates transforming growth factor-β (TGF-β)–induced smooth muscle cell differentiation from mouse mesenchymal progenitors

Differentiation of smooth muscle cells (SMCs) is critical for proper vasculogenesis and angiogenesis. However, the molecular mechanisms controlling SMC differentiation are not completely understood. During embryogenesis, the transcription factor mesenchyme homeobox 1 (Meox1) is expressed in the early developing somite, which is one of the origins of SMCs. In the present study, we identified Meox1 as a positive regulator of SMC differentiation. We found that transforming growth factor-β (TGF-β) induces Meox1 expression in the initial phase of SMC differentiation of pluripotent murine C3H10T1/2 cells. shRNA-mediated Meox1 knockdown suppressed TGF-β–induced expression of SMC early markers, whereas Meox1 overexpression increased expression of these markers. Mechanistically, Meox1 promoted SMAD family member 3 (Smad3) nuclear retention during the early stage of TGF-β stimulation because Meox1 inhibited protein phosphatase Mg2+/Mn2+-dependent 1A (PPM1A) and thereby prevented PPM1A-mediated Smad3 dephosphorylation. Meox1 appears to promote PPM1A degradation, leading to sustained Smad3 phosphorylation, thus allowing Smad3 to stimulate SMC gene transcription. In vivo, Meox1 knockdown in mouse embryos impaired SMC marker expression in the descending aorta of neonatal mice, indicating that Meox1 is essential for SMC differentiation during embryonic development. In summary, the transcriptional regulator Meox1 controls TGF-β–induced SMC differentiation from mesenchymal progenitor cells by preventing PPM1A-mediated Smad3 dephosphorylation, thereby supporting SMC gene expression.

Smooth muscle cell (SMC) 2 is mainly involved in blood vessel contraction, pressure regulation, and blood flow distribution (1). SMCs in different segments of arteries are derived from a number of precursors such as neural crest, proepicardium, mesothelium, somites, mesoangioblast, etc. For example, smooth muscles located in dorsal aorta originate from somite (2). A number of studies have suggested that several cardiovascular diseases including atherosclerosis are associated with the diversity of SMCs originating from different sources (3). However, the underlying mechanisms remain largely unknown. Several in vitro models are available for exploring mechanisms controlling SMC differentiation from different progenitors. For instance, Monc-1 cells originating from mouse neural crest can be induced into contractile SMC by TGF-␤ treatment (4), mouse pluripotent embryonal carcinoma-derived A404 cells can be induced to express several SMC markers with the treatment of all-trans-retinoic acid followed by puromycin (5), and the mesoderm-derived C3H10T1/2 cells can be induced by TGF-␤ to express SMC markers such as ␣-SMA, calponin, and SM22␣ (6).
Mesenchyme homeobox 1 (Meox1) belongs to a diverged subfamily of homeobox transcription factors (7). Meox1 is expressed in early developing somite and plays an important role in somitogenesis during embryogenesis (8). In adults, Meox1 is expressed in a variety of cells including endothelial cells (9) and vascular SMCs (10). Recent studies show that Meox1 is critical for the specification of endothelial cells in the endotome of the somites, which further give rise to hematopoietic stem cells (11). Because Meox1 is expressed in somite, and somite is one of the SMC precursors, we hypothesized that Meox1 may be involved in SMC differentiation. Indeed, Meox1 was significantly up-regulated in TGF-␤-treated C3H10T1/2 cells. Thus we sought to determine whether Meox1 plays a role in TGF-␤-induced SMC differentiation.
In the present study, we found that Meox1 was up-regulated along with SMC early markers in 10T1/2 cells by TGF-␤ induction. Knockdown or overexpression of Meox1 dramatically altered SMC marker gene expression. Mechanistically, Meox1 prevented nuclear Smad3 from shuttling back to cytoplasm during the early stage of TGF-␤ stimulation. Meox1 appeared to promote Smad3 nuclear retention through inhibiting the expression of the protein phosphatase, Mg 2ϩ /Mn 2ϩ -dependent 1A (PPM1A) as well as promoting its degradation. Most importantly, knockdown of Meox1 impaired SMC differentiation during embryonic development in mice.

Meox1 expression was up-regulated in TGF-␤-induced SMC differentiation
TGF-␤-treated 10T1/2 cells are a widely used model for studying SMC differentiation. We confirmed that TGF-␤ induced the expression of SMC early differentiation markers ␣-SMA, calponin, and SM22␣ in 10T1/2 cells at both the mRNA (Fig. 1, A-D) and protein levels ( Fig. 1, E-I). TGF-␤ induced Meox1 expression along with the SMC marker expression. Importantly, Meox1 expression was induced as early as 2 h following TGF-␤ stimulation, suggesting that Meox1 may be involved in TGF-␤-induced SMC differentiation.

Meox1 was essential for TGF-␤-induced SMC differentiation
To test if Meox1 plays a role in TGF-␤-induced SMC differentiation, we knocked down Meox1 expression using its specific shRNA delivered via adenoviral vector prior to the TGF-␤ induction. As shown in Fig. 2A, Meox1 and SMC markers were induced simultaneously by TGF-␤, and Meox1 expression was effectively blocked by its shRNA (Fig. 2, A and B). Importantly, knockdown of Meox1 blocked TGF-␤-induced expression of SMC early markers ␣-SMA, calponin, and SM22␣ by 47, 84, and 78%, respectively ( Fig.  2, C-E). Meox1 reduction appeared not to alter the basal level expression of SMC markers. These results indicated that Meox1 was essential for TGF-␤-induced SMC differentiation.

Meox1 is sufficient for SMC differentiation of 10T1/2 cells
To further investigate the role of Meox1 in SMC differentiation, we tested if Meox1 alone can induce SMC marker gene expression. Thus, we forcefully expressed Meox1 in 10T1/2 cells via adenoviral transduction of Meox1 cDNA and detected the expression of SMC early differentiation markers. As shown in Fig. 3, Meox1 was robustly expressed in 10T1/2 cells with a 5-fold increase compared with the control vector transduction, and notably, overexpression of Meox1 promoted the protein expression of ␣-SMA and calponin, indicating that Meox1 was able to stimulate 10T1/2 cells to differentiate to a SMC lineage.
In addition to SMC marker gene expression, Meox1 also affected SMC morphology. TGF-␤ induced a spindle-shaped SMC morphology. However, knockdown of Meox1 reversed the cell morphology to its original appearance (Fig. 3E), suggesting that Meox1 may be essential for the progenitor cells to differentiate to mature SMCs.

Meox1 regulated SMC marker gene transcription and mediated Smad3 function
Transcriptional activation of SMC genes is one of the molecular mechanisms controlling SMC differentiation (12)(13)(14). Therefore, we sought to test if Meox1 regulates SMC differentiation by mediating SMC marker gene transcription. We first detected whether Meox1 regulates SMC marker expression at the mRNA level and found that knockdown of Meox1 by its shRNA dramatically attenuated TGF-␤-induced mRNA expression of ␣-SMA and SM22␣ without changing their levels at the basal state (Fig. 4, A-D), suggesting that Meox1 may be

Meox1 in SMC differentiation
involved in SMC marker gene transcription induced by TGF-␤. Indeed, knockdown of Meox1 inhibited TGF-␤ induction of ␣-SMA and SM22␣ promoter activities, as shown by SMC promoter-reporter luciferase assays (Fig. 4, E and F). Remarkably, Meox1 shRNA reduced ␣-SMA and SM22␣ promoter activities by 58.6 and 43%, respectively, demonstrating that Meox1 was a critical regulator for SMC gene transcription.
We and others have reported previously that Smad3 is the key factor mediating TGF-␤ function in SMC differentiation of 10T1/2 cells (6,15,16). Smad3 mainly regulates the transcription of SMC differentiation marker genes in the initial stage of SMC differentiation (15). Therefore, we sought to determine whether Meox1 is involved in Smad3-mediated early SMC marker gene transcription. As shown in Fig. 4, G and H, forced expression of Smad3 markedly up-regulated SMC marker ␣-SMA and SM22␣ promoter activities. However, knockdown of Meox1 blocked Smad3 function in increasing the activities of both ␣-SMA and SM22␣ promoters. Interestingly, Smad3 increased the promoter activities even in the basal state, i.e. without TGF-␤ treatment, whereas knockdown of Meox1 completely blocked the Smad3-mediated promoter activities, suggesting that in addition to TGF-␤-treated cells, Meox1 was also important for Smad3 activity in the quiescent SMC.

Meox1 was essential for Smad3 nuclear location
Because Smad3 is continuously shuttling back and forth between cytoplasm and nuclei with or without TGF-␤ treatment (17) and Smad3 nuclear translocation is required for its transcription activity in SMC differentiation, we tested if Meox1 affects Smad3 nuclear localization. Immunostaining showed that TGF-␤ treatment for 30 min caused the majority of Smad3 translocated into the nuclei of 10T1/2 cells (Fig. 5A). However, knockdown of Meox1 by its shRNA caused less Smad3 located in the nuclei (Fig. 5A). Because Meox1 is a nuclear protein, these data suggest that Meox1 may be important for Smad3 nuclear retention. To confirm the Meox1 function in Smad3 nuclear location, we detected cytoplasmic and nuclear Smad3 levels in TGF-␤-treated cells when the Meox1 expression was manipulated. As shown in Fig. 5, B and C, TGF-␤ induction for 2 h caused a significantly higher level of nuclear Smad3 compared with the untreated cells along with a reduction of cytoplasmic Smad3. Knockdown of Meox1, however, significantly reduced the nuclear Smad3 level by 48.7% compared with TGF-␤-treated cells with intact Meox1 (Fig. 5, B and C). These data further demonstrated that Meox1 was essential for Smad3 nuclear retention.

Meox1 in SMC differentiation
tion leads to the nuclear location whereas dephosphorylation of the protein causes Smad3 export from the nuclei (18). Thus, the phosphorylated Smad3 reflects the nuclear Smad3 protein level. Time-dependent study showed that Smad3 was phosphorylated 10 min after TGF-␤ induction, and the phospho-Smad3 reached the highest level at 10 -30 min but gradually decreased after 2 h of TGF-␤ treatment (Fig. 6, A and B). However, knockdown of Meox1 by shRNA significantly or completely diminished the phospho-Smad3 level during TGF-␤ induction. Because Meox1 is not located in the cytoplasm, it is unlikely that Meox1 blocks TGF-␤-induced Smad3 phosphorylation process. Instead, we hypothesized that Meox1 may be essential in preserving the phosphorylation of nuclear Smad3, which can be removed by PPM1A (19). PPM1A is a phosphatase belonging to the protein serine/threonine phosphatase family, which specifically removes phosphates on serine/threonine residues (20,21). Among 49 family members, PPM1A is the only phosphatase that dephosphorylates Smad2/3 on C-terminal and terminates TGF-␤ signaling (19). PPM1A-mediated Smad3 dephosphorylation is known to reduce Smad3 phosphorylation and thus facilitate Smad3 nuclear export. Knockdown of Meox1 significantly enhanced the expression of PPM1A in both vehicle and TGF-␤-treated cells (Fig. 6, C and D). To determine how Meox1 inhibited the PPM1A level, we tested if Meox1 affected PPM1A protein stability when the protein synthesis was blocked by cycloheximide. Time-dependent study indicated that knockdown of Meox1 increased the stability of PPM1A protein as shown by the reduced degradation rate or prolonged half-life of PPM1A as compared with the control vector-treated cells (Fig. 6, E and F). These results indicated that Meox1 may decrease PPM1A level by increasing its degradation. To further confirm the role of Meox1 in PPM1A expression, we overexpressed Meox1 in 10T1/2 cells and found that forced expression of Meox1 could inhibit PPM1A expression (Fig. 6, G and H). Taken together, these data support a novel concept that Meox1 confines PPM1A level to preserve Smad3 phosphorylation and thus facilitates Smad3 nuclear location.

Knockdown of Meox1 impeded SMC differentiation in vivo
To test if Meox1 is important for SMC differentiation in vivo, we administered adenoviral vector-expressing control and Meox1 shRNA intraplacentally into C57BL/6J mouse embryos (E12.5) and collected the descending aortas of mice that were born on postnatal day 1 (P1). As shown in Fig. 7A, Meox1 expression in the descending aorta was significantly knocked down by its shRNA, which was confirmed by immunostaining of Meox1 in the descending aorta frozen section (Fig. 7B). Importantly, knockdown of Meox1 significantly inhibited the

Meox1 in SMC differentiation
expression of SMC markers ␣-SMA and calponin, as shown by the reduction of both their mRNA and protein expression in the descending aorta (Fig. 7, A-E). In addition, the SMC numbers were also reduced when the Meox1 was knocked down (Fig.  7F). These data indicated that Meox1 played an important role in SMC differentiation during embryonic development.

Discussion
The present studies have demonstrated that Meox1 plays an important role in regulating TGF-␤-induced SMC differentiation. Meox1 is induced along with the expression of early SMC markers. Knockdown of Meox1 inhibits SMC marker gene expression whereas forced expression of Meox1 enhances the marker expression. The essential roles of Meox1 in TGF-␤induced SMC marker mRNA expression and promoter activities suggest that Meox1 regulates SMC differentiation by controlling SMC gene transcription. Most importantly, Meox1 is required for SMC differentiation during the embryonic development in vivo. It appears that Meox1 is involved in Smad3 regulation of SMC genes because knockdown of Meox1 diminishes Smad3-mediated SMC gene promoter activities. Mechanistically, Meox1 promotes PPM1A degradation or reduces its expression, which permits Smad3 to preserve its phosphorylation status in the nuclei of 10T1/2 cells upon TGF-␤ treatment, and thus allows Smad3 to activate SMC gene transcription, leading to the initiation of SMC differentiation (Fig. 8).
TGF-␤/Smad signaling plays a critical role in SMC differentiation (22,23). In fact, Smad proteins are involved in progenitor-specific regulation of SMC differentiation, i.e. Smad2 is important for neural crest cell differentiation to SMC whereas Smad3 is essential for the SMC differentiation from mesenchymal progenitors (24). Although PPM1A dephosphorylates and promotes nuclear export of both the TGF-␤-activated Smad2 and Smad3 (19), we only focus on the effect of PPM1A on Smad3 phosphorylation because Smad2 appears not to be involved in TGF-␤-induced SMC differentiation from mesenchymal progenitors (25). Smad3 activity may be regulated in multiple levels such as expression, phosphorylation, nuclear translocation, and interactions with other transcription factor or Smad-binding elements on the target gene promoters. Our results suggest that Meox1 regulates Smad3 phosphorylation status in the nuclei rather than participating in the TGF-␤ receptor-mediated Smad3 phosphorylation in the cytoplasm, largely because of the nuclear location of Meox1. Meox1 appears to be a novel suppressor for PPM1A expression because knockdown of Meox1 promotes the expression of PPM1A, even in the basal state prior to TGF-␤ stimulation. In addition, Meox1 promotes PPM1A degradation because knockdown of Meox1 causes an enhanced accumulation of PPM1A when the protein synthesis is blocked.
In addition to serving as a PPM1A inhibitor, Meox1 is likely to also function as a transcription factor directly regulating SMC

Meox1 in SMC differentiation
gene promoter. This notion is supported by the fact that Meox1 alone is able to enhance the activities of exogenously introduced SMC promoters in quiescent SMCs where Smad3 is mainly located in the cytoplasm. These results are consistent with a previous report showing that Meox1 may be involved in gene transcription because it can specifically bind the DNA sequences recognized by Hoxa2 on its target genes (26). A future identification of the Meox1-binding element on SMC marker promoters and the in vivo function may provide more detailed insights into the regulatory mechanism governing Meox1 function in SMC differentiation.
A recent study has shown that Meox1 up-regulated by PDGF-BB inhibits ␣-SMA and calponin expression in cultured SMCs (35), which is opposite to the role of Meox1 identified in this study. The discrepancy could be because of the different environments and cellular contexts involved in these two studies. TGF-␤-treated progenitors mimic the development program whereas PDGF-BB-treated SMCs imitate the injury-induced vascular pathology, which may lead to completely different intracellular responses and thus may produce opposite outcome even to the same gene. Indeed, TGF-␤-induced Meox1 promotes Smad3 activity via inhibiting PPM1A, resulting in the increased SMC marker expression. However, PDGF-BB-induced Meox1 activates FAK-ERK1/2 signaling and induces autophagy, causing down-regulation of SMC markers (35). These results suggest that the functional outcome of a gene may not only depend on its target, but also is influenced largely by the intracellular environment especially its associated factors.
Taken together, our studies have identified Meox1 as a novel regulator for TGF-␤-induced SMC differentiation from mesenchymal progenitors. Meox1 regulates SMC marker gene expression by preserving Smad3 phosphorylation in cell nuclei through promoting PPM1A degradation and possibly also by blocking PPM1A expression, which prevents Smad3 dephosphorylation. Meox1 may also serve as a transcription factor directly regulating SMC gene promoter activation. Because Meox1 is involved in early somite development and somite is one of the progenitors for vascular SMCs, Meox1-mediated SMC differentiation is likely to be important for embryonic vasculature development.

Experimental procedures
Cell culture and transfection C3H10T1/2 (10T1/2) cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 5% L-glutamine. Cells were starved in serum-free medium for 48 h followed by incubating with TGF-␤ (5 ng/ml) for various times as needed. For transfection of plasmid DNA, cells were plated on 12-well plates or 6-cm dishes with 70 -80% confluence. 24 h later, transfection was carried out using Lipofectamine LTX reagents (Life Technologies) according to the manufacturer's instruction.

Construction of adenoviral vectors
Mouse Meox1 cDNA was inserted into the XhoI site of pShuttle-IRES-hrGFP-1 (Agilent Technologies) and was confirmed by sequencing. For Meox1 shRNA adenoviral vector,  Upon TGF-␤ stimulation, Smad3 is phosphorylated and shuttled into nucleus where Smad3 binds to Smad-binding element (SBE) in SMC marker gene promoters to initiate SMC marker transcription and subsequent SMC differentiation. As a mild activator for SMC genes, Smad3 facilitates SMC differentiation in early stage of TGF-␤ treatment through a temperate and precise regulation. In the late stages, nuclear Smad3 is exported to cytoplasm because of the dephosphorylation by PPM1A. In the initial stage of SMC differentiation, Meox1 confines PPM1A level through promoting its degradation, which enhances nuclear Smad3 retention and thus safeguards SMC marker gene transcription.

Viral inoculation of mouse embryo
C57BL/6J mice were housed under conventional conditions in the animal care facilities and received humane care in compliance with the Principles of Laboratory Animal Care formulated by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals. Animal procedures were approved by the Institutional Animal Care and Use Committee of the University of Georgia. 3-monthold mice were mated to produce embryos for viral inoculation. Females were checked to see whether there was a vaginal plug every morning. The day that the plug was found was considered as E0.5. Pregnant mice with E12.5 embryos were intraplacentally administered 3 l 5.3 ϫ 10 5 TCID 50 /ml adenovirus (where TCID 50 is tissue culture infective dose) expressing Meox1 shRNA into E12.5 embryos; scramble shRNA-expressing adenovirus was injected as control. To improve retention of the pregnancy, the viral injection of the two embryos next to the ovaries and the upper vagina was avoided. Injected embryos were placed back in pregnant dams and allowed to develop to the full term (28,29). The descending aorta of postnatal mice were collected day 1 after the mice were born.

Quantitative reverse transcription PCR (qPCR)
Total RNA was extracted from cells using TRIzol Reagent (Life Technologies) followed by a reverse transcription using iScript cDNA Synthesis Kit (Bio-Rad) according to the manufacturer's instructions. qPCR was performed using Mx3005P qPCR machine with SYBR Green Master Mix (Agilent Technologies). Mouse Meox1 primers used were 5Ј-GGA AGG AGA GGA CAG CCT TC-3Ј (forward) and 5Ј-CCC TTC ACA CGT TTC CAC TT-3Ј (reverse). Smooth muscle marker primers were described previously (30).

Immunofluorescent staining
10T1/2 cells were seeded on sterile coverslips and transfected with Meox1 or control (Ctrl) shRNA in complete medium for 1 day and then starved for 2 days before TGF-␤ treatment for 30 min and 2 h. Cells were washed with PBS three times before fixing in 4% paraformaldehyde for 5 min. For immunostaining of aorta tissues, 10 m fresh frozen sections were air-dried for 30 min at room temperature followed by 10 min of fixation with 4% paraformaldehyde. The fixed cells or tissue sections were washed by PBS three times followed by incubation with PBS containing 0.1% Triton X-100 for 10 min. After washing with PBS three times, 5% goat serum in PBS was used to block cells for 30 min. Anti-Meox1 (Abcam, ab105349), anti-Smad3 (Cell Signaling Technology, 9523s), anti-␣-SMA (Sigma-Aldrich, A2547), or anti-calponin (Abcam, ab46794) primary antibodies were diluted at 1:50 or 1:100 and the cells incubated at 4°C overnight. Cells and sections were then incubated with FITC-or TRITC-conjugated secondary antibodies (1:150 and 1:50, respectively) for 30 min followed by PBS washing for three times. Stained cells and sections were observed and imaged with Nikon Eclipse 90i microscope, and images were captured with Nikon 12.7MP digital Sight DS-Ri1 color camera as described previously (31).

Promoter-reporter luciferase assay
␣-SMA or SM22␣ luciferase promoter constructs were transfected into 10T1/2 cells with or without other plasmids using Lipofectamine LTX (Life Technologies) as described previously (34). 24 h after the transfection, 10T1/2 cells were starved for 2 days before TGF-␤ treatment for 8 h. Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega).

Statistical analysis
All values are expressed as mean Ϯ S.E. Data were evaluated with a two-tailed, unpaired, Student's t test or compared by one-way analysis of variance (ANOVA) followed by Fisher exact test. A p value Ͻ0.05 was considered statistically significant.