An epigenetic regulatory loop controls pro-osteogenic activation by TGF-β1 or bone morphogenetic protein 2 in human aortic valve interstitial cells

Calcific aortic valve disease (CAVD) is common in the elderly population, but pharmacological interventions for managing valvular calcification are unavailable. Transforming growth factor β1 (TGF-β1) and bone morphogenetic protein 2 (BMP-2) induce pro-osteogenic activation of human aortic valve interstitial cells (AVICs) that play an important role in valvular calcification. However, the molecular mechanism underlying pro-osteogenic activation in AVICs is incompletely understood. Here, we investigated an epigenetic regulatory mechanism in human AVIC pro-osteogenic activation induced by TGF-β1 and BMP-2. Microarray and real-time PCR analyses revealed that microRNA (miR)-486 up-regulation and miR-204 down-regulation were characteristic changes in TGF-β1- and BMP-2-stimulated normal AVICs and in AVICs from calcified valves. Both TGF-β1 and BMP-2 down-regulated miR-204 through Smad pathways. Interestingly, an miR-486 antagomir diminished the effect of TGF-β1 and BMP-2 on miR-204 levels and calcium deposit formation. Furthermore, the miR-486 antagomir increased the expression of Smurf2, a Smad inhibitor, in the presence or absence of TGF-β1 or BMP-2 stimulation, whereas a miR-486 mimic reduced Smurf2 expression. Smurf2 knockdown augmented TGF-β1- or BMP-2-induced miR-204 down-regulation and resulted in increased expression of the osteoblastic biomarkers Osx and Runx2. In summary, we found that TGF-β1 and BMP-2 up-regulate miR-486 and down-regulate miR-204 in human AVICs to promote pro-osteogenic activity and that miR-486 inhibits Smurf2 expression to augment the miR-204 down-regulation. We conclude that the miR-486-Smurf2-Smad loop plays an important role in regulating AVIC pro-osteogenic activation in response to TGF-β1 or BMP-2. Targeting this regulatory loop may have therapeutic potential for suppressing aortic valve calcification.

Calcific aortic valve disease (CAVD) 2 is one of the leading cardiovascular diseases in the United States (1,2). CAVD is an active process involving chronic valvular inflammation and calcification (3). Severe aortic valve calcification causes morbidity and results in the second most common cardiovascular surgery performed (4). The growing prevalence of this disease associated with prolongation of the human life span and unavailability of pharmacological interventions for prevention of this disease progression emphasize the importance of understanding of the pathobiological mechanisms of this disease. Particularly, it is critical to elucidate the cellular and molecular mechanisms by which aortic valve leaflets become calcified.
TGF-␤1 and bone morphogenetic protein 2 (BMP-2) are recognized as important pro-osteogenic factors involved in vascular and aortic valve calcification (5,6). Previous immunohistochemical studies revealed the presence of higher levels of TGF-␤1 and BMP-2 in calcified human aortic valve leaflets compared with non-calcified aortic valve leaflets (7). Both TGF-␤1 and BMP-2 have been shown to up-regulate alkaline phosphatase expression and activity in human aortic valve interstitial cells (AVICs), the primary cells involved in aortic valve calcification (8). Thus, the pro-osteogenic effects of TGF-␤1 and BMP-2 on AVICs may play an important role in aortic valve calcification and CAVD progression. Investigation of the molecular mechanism by which TGF-␤1 and BMP-2 induce pro-osteogenic changes in human AVICs will improve understanding of the mechanism underlying CAVD progression.
MicroRNAs (miRs) are a group of small non-coding, 21-to 23-nt-long RNA molecules (9). miRs regulate the expression of protein-coding genes and thus play important roles in the epigenetic regulation of gene expression. Emerging evidence demonstrates that miRs modulate a wide range of developmental and physiological processes, including cell proliferation and differentiation (10). Recently, miRs are implicated in the pathophysiology of various cardiovascular diseases and have become intriguing targets for therapeutic interventions (11). In addition, several studies found that altered miR expression is associated with aortic valve calcification (12,13). However, the mechanistic role of altered miRs in AVIC pro-osteogenic acti-vation by TGF-␤1 and BMP-2 remains unclear. Studies in this line will provide insight into the mechanism underlying CAVD pathogenesis and may identify therapeutic targets for suppression of CAVD progression.
Clinicopathological and conventional diagnostic imaging studies of calcified human aortic valves demonstrate that AVICs express osteopontin, osteocalcin, and osteogenic transcription factors such as Osterix (Osx) and Runt-related transcription factor 2 (Runx2)/Cbfal (14). Runx2 is required for osteoblast differentiation (15). Osx is identified as a BMP2specific transcription factor involved in bone formation (16). Importantly, Osx knock-out mice display severe impairment of bone formation and osteoblastogenesis, indicating that this transcription factor is essential for osteoblast differentiation (16,17). The significance of Runx2 and Osx in CAVD is highlighted by the observation that the expression of these two osteogenic transcription factors is elevated in AVICs from calcified human aortic valves (18).
Smad signaling plays a critical role in mediating the effect of TGF-␤1 and BMP-2. Smad1 is required for induction of Osx by BMP-2 (19). We have observed that Smad1 and Smad3 are involved in mediating the expression of Runx2 and alkaline phosphatase in human AVICs induced by biglycan that up-regulates the production and release of BMP-2 and TGF-␤1 (20). In contrast, Smurf2 is a negative regulator of Smad signaling and plays a role in the control of cellular response to TGF-␤ and BMP (21,22). Currently, the knowledge of miR modulation by TGF-␤1 and BMP-2 in human AVICs is limited. The roles of the Smad pathways and Smurf2 in regulation of AVIC expression of miRs remain unclear.
We hypothesized that TGF-␤1 and BMP-2 induce pro-osteogenic activation in human AVICs via a miR-osteogenic transcription factor axis. The purpose of this study was to determine: 1) the effects of TGF-␤1 and BMP-2 on the expression of miRs in human AVICs, 2) the role of Smad pathways in the mechanism by which TGF-␤1 and BMP-2 exert their effects on miR expression, 3) the role of miRs in modulation of the expression of osteogenic transcription factors and pro-osteogenic activity in human AVICs, and 4) the effect of miRs on Smurf2 expression and the role of Smurf2 in modulation of AVIC response to TGF-␤1 and BMP-2.

TGF-␤1 and BMP-2 up-regulate miR-486 and down-regulate miR-204 in human AVICs
Microarray analysis revealed that increased levels of miR-486 and decreased levels of miR-204 were the common changes in normal human AVICs exposed to TGF-␤1 and BMP-2 (supplemental Table S1). TGF-␤1 and BMP-2 reduced miR-204 levels by 75 and 63%, respectively, and these two pro-osteogenic factors increased miR-486 levels by 68 and 72%, respectively. Realtime PCR data confirmed significant changes in miR-204 and miR-486 levels in cells exposed to TGF-␤1 or BMP-2, and showed that altered levels of these two miRs were comparable with those in AVICs from calcified human aortic valves (Fig. 1). Thus, both TGF-␤1 and BMP-2 reduce miR-204 and increase miR-486 in human AVICs.

Expression of miR-486 antagomir attenuates miR-204 downregulation and calcium deposit formation induced by TGF-␤1 or BMP-2
miR-204 is a negative regulator of osteoblastic differentiation and miR-204 down-regulation plays a role in AVIC and vascular smooth muscle cell calcification in vitro (23)(24)(25). To examine whether miR-486 is involved in miR-204 down-regulation induced by TGF-␤1 and BMP-2, we determined the effect of lentiviral expression of miR-486 antagomir. We observed that expression of the miR-486 antagomir attenuated the negative effect of TGF-␤1 and BMP-2 on miR-204 expression ( Fig. 2A). Furthermore, we observed that attenuation of miR-204 downregulation by the miR-486 antagomir correlated with a marked reduction of calcium deposit formation following prolonged stimulation with TGF-␤1 or BMP-2 (Fig. 2B). Therefore, upregulation of miR-486 by TGF-␤1 and BMP-2 contributes to the mechanism by which these two pro-osteogenic factors down-regulate miR-204 and promote pro-osteogenic activity in human AVICs.
In silico analysis (UCSC Genome Bioinformatics) indicated that a potential Smads-binding site (CAGAC) might be present in the miR-204 promoter region. To elucidate the mechanism by which Smads mediate the effect of TGF-␤1 and BMP-2 on miR-204 expression, a ChIP assay was performed to examine the interaction of Smads with the miR-204 promoter region. Antibody-mediated immunoprecipitation of fragmented chromatin from formaldehyde cross-linked cell lysates with a subsequent PCR amplification revealed binding of the Smad1 and Smad3 to the miR-204 promoter region from Ϫ1056 to Ϫ1223
To determine the role of Smurf2 in modulating miR-204 expression in human AVICs exposed to TGF-␤1 and BMP-2, we treated cells with Smurf2 shRNA. As shown in Fig. 5A, Smurf2 shRNA effectively knocked down Smurf2 protein.
Knockdown of Smurf2 further reduced miR-204 levels in cells exposed to TGF-␤1 or BMP-2, whereas treatment with lentiviral control shRNA had no effect (Fig. 5A).
In investigation of the role of miR-204 down-regulation in Osx and Runx2 expression induced by TGF-␤1 and BMP-2, we applied miR-204 mimic or miR-204 antagomir to AVICs exposed to TGF-␤1 or BMP-2. As shown in Fig. 5B, miR-204 mimic reduced Osx and Runx2 expression induced by either TGF-␤1 or BMP-2. Conversely, miR-204 antagomir enhanced the effect of TGF-␤1 and BMP-2 on the expression of these two osteogenic transcription factors. However, control miRNA had no effect.
To understand whether Smurf2 functions as a negative regulator in the expression of Osx and Runx2 induced by TGF-␤1 and BMP-2, we treated AVICs expressing Smurf2 shRNA with TGF-␤1 or BMP-2 and examined Osx and Runx2 levels. Smurf2 Figure 2. Expression of miR-486 antagomir attenuates miR-204 down-regulation and calcium deposit formation in human AVICs exposed to TGF-␤1 or BMP-2. A, real-time quantitative RT-PCR analysis shows that lentiviral expression of miR-486 antagomir markedly attenuates the negative effect of TGF-␤1 and BMP-2 on miR-204 levels. n ϭ 5 separate experiments using distinct cell isolates; *, p Ͻ 0.05 versus untreated control; #, p Ͻ 0.05 versus TGF-␤1 or BMP-2 with or without irrelevant miR (miR-C). B, normal AVICs were untreated or treated with TGF-␤1 or BMP-2 in conditioning medium for 14 days. Treatment with either TGF-␤1 or BMP-2 increased the formation of calcium deposits. Lentiviral expression of miR-486 antagomir reduced the pro-osteogenic activity in cells treated with TGF-␤1 or BMP-2, whereas lentiviral expression of irrelevant miR (miR-C) had no effect. n ϭ 5 separate experiments using distinct cell isolates; *, p Ͻ 0.05 versus untreated control; #, p Ͻ 0.05 versus TGF-␤1 alone or BMP-2 alone; †, p Ͻ 0.05 versus stimulant ϩ miR-C.

The regulatory loop of AVIC pro-osteogenic activity
knockdown enhanced the expression of Osx and Runx2 following exposure to TGF-␤1 or BMP-2 (Fig. 5C). Together, the results show that reduction of the Smurf2 level exaggerates miR-204 down-regulation to enhance Osx and Runx2 expression following TGF-␤1 and BMP-2 stimulation, and suggest that the effect of miR-486 on Smurf2 levels accounts for its impact on miR-204 down-regulation and the pro-osteogenic activity in human AVICs exposed to TGF-␤1 or BMP-2.

Discussion
TGF-␤1 and BMP-2 are pro-osteogenic mediators in the development and progression of CAVD. In this regard, high levels of these two mediators have been found in the aortic valves explanted from patients with CAVD (8,26). Furthermore, a number of studies demonstrate that TGF-␤1 and BMP-2 induce the osteogenic responses in human AVICs (8,(27)(28)(29). Previous studies have shown that several miRs, including miR-322, regulate osteoblast differentiation induced by TGF-␤1 and BMP-2 (30). More importantly, miR-141 is involved in porcine aortic valve cell calcification induced by BMP-2 (31). As TGF-␤1 and BMP-2 have similar pro-osteogenic effects on human AVICs, it is likely that these two pro-osteogenic mediators modulate the expression of a specific group of miRs to exert their effects on AVIC osteogenic responses. Interestingly, our microarray data show that only miR-486 and miR-204 levels are altered in human AVICs by both TGF-␤1 and BMP-2. Furthermore, either of these two pro-osteogenic mediators down-regulates miR-204 expression and up-regulates miR-486 expression to levels comparable with those seen in AVICs from calcified human aortic valves.
miR-204 is a known negative regulator of osteoblastic differentiation. Currently, it is unclear how TGF-␤1 and BMP-2 down-regulate miR-204 expression in human AVICs. In addition, it is unknown whether miR-486 up-regulation plays a role in mediating osteogenic responses to TGF-␤1 and BMP-2 in human AVICs and whether miR-486 and miR-204 interact in A, human AVICs were treated with Smad1 siRNA, Smad3 siRNA, or scrambled siRNA (50 nM each) for 48 h before being exposed to TGF-␤1 or BMP-2 for 24 h. Smad1 siRNA reduced Smad1 levels, but had no effect on Smad3 levels. Smad3 siRNA reduced Smad3 levels without an influence on Smad1 levels. Real-time quantitative RT-PCR analysis show that knockdown of Smad3 and Smad1 attenuates miR-204 down-regulation by TGF-␤1 and BMP-2, respectively. n ϭ 5 separate experiments using distinct cell isolates; *, p Ͻ 0.05 versus untreated control; #, p Ͻ 0.05 versus TGF-␤1 alone or BMP-2 alone; ‡, p Ͻ 0.05 versus stimulant ϩ scrambled siRNA. B, AVICs were stimulated with TGF-␤1 or BMP-2 for 8 h and harvested for ChIP assay. ChIP DNA was amplified using specific primers for the miR-204 gene promoter after immunoprecipitation with antibodies against Smad1 (Ab-Smad1), Smad3 (Ab-Smad3), non-immune IgG (negative control), or RNA polymerase II (positive control). Input chromatin isolated before the immunoprecipitation was used to control for equal amounts of input DNA. Results of ChIP assay show that Smad1 and Smad3 can bind to the promoter region of miR-204 gene.
mediating the osteogenic responses. The findings of the present study provide mechanistic information.

Smads binds to miR-204 gene promotor to modulate the expression of this negative regulator of pro-osteogenic activity
Both TGF-␤1 and BMP-2 utilize the Smad signaling pathways to exert their effects on human AVICs (20). Eight mammalian Smad proteins have been identified to date, and the Smad family of proteins have distinct roles in TGF-␤1 and BMP-2 signaling (32). In this regard, Smad3 mediates the effect of TGF-␤1, whereas Smad1 mediates the effect of BMP-2 in the up-regulation of pro-osteogenic activity in human AVICs (20). Using a ChIP approach, we found Smad1 and Smad3 can bind to the promoter region of miR-204. Knockdown of Smad3, not Smad1, reduces the effect of TGF-␤1 in the down-regulation of miR-204 expression, whereas knockdown of Smad1 reduces the negative effect of BMP-2 on miR-204 expression. Thus, TGF-␤1 and BMP-2 down-regulate miR-204 expression in human AVICs through the classical Smad signaling pathways, and Smad3 and Smad1 appear to mediate the effects of TGF-␤1 and BMP-2, respectively, on miR-204 expression through modulation of gene transcription. miR-204 modulates osteoblastic differentiation (23), and down-regulation of miR-204 promotes in vitro calcification in vascular smooth muscle cells and human AVICs (24,25). To confirm the role of miR-204 down-regulation in mediating the pro-osteogenic effects of TGF-␤1 and BMP-2, we evaluated the expression of osteogenic transcription factors Osx and Runx2 in human AVICs. The results show that TGF-␤1 and BMP-2 increase the levels of both Osx and Runx2. Two lines of evidence indicate that miR-204 down-regulation plays a role in promoting the expression of Osx and Runx2 in human AVICs. First, miR-204 mimic suppresses and miR-204 antagomir increases cellular levels of Osx and Runx2 in human AVICs exposed to TGF-␤1 or BMP-2. Second, treatment of human AVICs with miR-204 antagomir in the absence of TGF-␤1 or BMP-2 causes an increase in the levels of Osx and Runx2. Thus, modulation of miR-204 levels may affect aortic valve pro-osteogenic activity. miR-141 has been shown to mediate BMP-2-induced osteogenic responses in porcine aortic valve cells (31). However, neither TGF-␤1 nor BMP-2 alters miR-141 levels in human AVICs. Our work demonstrates a significant down-regulation of miR-204 by either TGF-␤1 or BMP-2 in human AVICs and an important role of miR-204 down-regulation in mediating the osteogenic responses to TGF-␤1 and BMP-2. The difference between our observations and the previously reported finding could be due to a species-related difference in AVIC response to pro-osteogenic mediators.

miR-486 modulates Smurf2 expression to augment miR-204 down-regulation
Interestingly, the results of this study show that up-regulation of miR-486 also plays a role in mediating the pro-osteogenic activity in human AVICs exposed to TGF-␤1 or BMP-2 because lentiviral expression of miR-486 antagomir markedly attenuates calcium deposition induced by TGF-␤1 or BMP-2. To our knowledge, such a role for miR-486 has not been reported. In correlation to this effect, lentiviral expression of the miR-486 antagomir preserves cellular levels of miR-204 in human AVICs exposed to TGF-␤1 or BMP-2. Like all RNAs, miRs are products of transcription. It is not surprising that the expression of some miRs are regulated by other miRs. An example of this type of miR-modulated miR expression is the crossregulation between members of the "myomiR" family of miRs that are encoded within myosin heavy chain genes (33,34). Our data show an interaction of miR-486 with miR-204 and suggest that up-regulation of miR-486 by TGF-␤1 and BMP-2 plays a role in the mechanism by which they down-regulate miR-204.
It is likely that miR-486 contributes to the mechanism of elevated pro-osteogenic activity in human AVICs stimulated by TGF-␤1 or BMP-2 through its negative impact on miR-204 expression. This raises the question of how miR-486 exerts its effect on miR-204 expression. The results of this study show that the miR-486 mimic decreases and miR-486 antagomir increases Smurf2 levels in human AVICs. TGF-␤1 and BMP-2 reduce Smurf2 levels and antagonizing miR-486 elevates Smurf2 levels in human AVICs exposed to TGF-␤1 and BMP-2. Therefore, miR-486 negatively regulates Smurf2 expression in human AVICs. Because Smurf2 is an intrinsic inhibitor of Smad signaling (22), it is reasonable to propose that up-regulation of miR-486 reduces Smurf2 expression to augment the osteogenic responses to TGF-␤1 and BMP-2. Indeed, our data show that knockdown of Smurf2 exaggerates the induction of Osx and Runx2 expression by TGF-␤1 and BMP-2.
Overall, this study found miR-204 down-regulation and miR-486 up-regulation as a common feature of AVICs from calcified human aortic valves and normal AVICs stimulated by proosteogenic mediators TGF-␤1 and BMP-2. Furthermore, we identified a miR-486-Smurf2-Smad loop in modulation of miR-204 expression and pro-osteogenic activity in human AVICs exposed to TGF-␤1 and BMP-2. Targeting this regulatory loop may have a therapeutic potential for suppression of aortic valve calcification.

Conclusions
TGF-␤1 and BMP-2 promote pro-osteogenic activity in human AVICs through up-regulation of miR-486 and downregulation of miR-204. The Smad pathways mediate the effects of TGF-␤1 and BMP-2 in down-regulation of the expression of miR-204, a negative regulator of osteogenic responses. miR-486 down-regulates Smurf2 levels to enhance the negative effect of TGF-␤1 and BMP-2 on miR-204 expression. The miR-486-Smurf2-Smad loop plays an important role in regulating AVIC osteogenic responses to TGF-␤1 and BMP-2 (Fig. 6).

Materials
Antibodies against human Osx were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Antibodies against human Runx2, Smad1, Smad3, and Smurf2 were purchased from Cell Signaling, Inc. (Beverly, MA). miR mimics and antagomirs, control miR, HiPerFect transfection reagent, other transfection-related reagents, and EndoFree Plasmid Maxi Kit were purchased from Qiagen (Valencia, CA). Lentivirus vector expressing miR mimic or antagomir and Block-it lentiviral pol II miR RNAi expression plasmids were purchased from Invitrogen. Lenti-Smurf2 shRNA and lenti-non-target shRNA were purchased from the Functional Genomics Facility of the University of Colorado. Lipofectamine 2000 was pur- Figure 5. Smurf2 knockdown augments miR-204 down-regulation induced by TGF-␤1 or BMP-2 and leads to enhanced expression of osteoblastic biomarkers. A, AVICs were untreated or infected with lentivirus that expresses Smurf2 shRNA. A representative immunoblot shows that Smurf2 shRNA effectively knocks down Smurf2 protein. Knockdown of Smurf2 further reduced miR-204 levels in AVICs exposed to TGF-␤1 or BMP-2. n ϭ 5 separate experiments using distinct cell isolates. *, p Ͻ 0.05 versus TGF-␤1 alone or BMP-2 alone. B, representative immunoblots and densitometric data show that the miR-204 mimic suppresses and miR-204 antagomir enhances the expression of Runx2 and Osx at 24 h following treatment with TGF-␤1 or BMP-2. n ϭ 6 separate experiments using distinct cell isolates; *, p Ͻ 0.05 versus control; #, p Ͻ 0.05 versus TGF-␤1 alone or BMP-2 alone; †, p Ͻ 0.05 versus stimulant ϩ miR-C (irrelevant oligonucleotide). C, representative immunoblots and densitometric data show that Smurf2 knockdown enhances the expression of Runx2 and Osx in AVICs exposed to TGF␤-1 or BMP-2. n ϭ 5 separate experiments using distinct cell isolates; *, p Ͻ 0.05 versus control; #, p Ͻ 0.05 versus TGF-␤1 alone or BMP-2 alone; ‡, p Ͻ 0.05 versus stimulant ϩ control shRNA.

The regulatory loop of AVIC pro-osteogenic activity
chased from Life Technologies, Inc. ChromaFlash TM Chromatin Extraction Kit and ChromaFlash TM One-Step ChIP Kit were purchased from Epigentek Group Inc. (Farmingdale, NY). TransDux transduction reagent was purchased from System Biosciences (Mountain View, CA). Medium 199 was purchased from Lonza (Walkersville, MD). All other chemicals and reagents were from Sigma.

Isolation, culture, and treatment of human AVICs
Normal aortic valve leaflets were collected from explanted hearts of 6 patients (4 males and 2 females, age 59.0 Ϯ 8.1 years) undergoing heart transplantation due to late stage cardiomyopathy. These valve leaflets were thin and had no histological abnormality. Calcified aortic valve leaflets were from 6 patients (4 males and 2 females, age 63.0 Ϯ 6.9 years) undergoing aortic valve replacement surgery due to CAVD. This study was approved by the University of Colorado Denver Institutional Review Board. All patients gave informed consent for the use of their aortic valves for this study.
AVICs were isolated and cultured using a previously described method (28,35,36). Briefly, a high concentration of collagenase (2.5 mg/ml) was used to remove endothelial cells, and the remaining tissue was treated with a low concentration of collagenase (0.8 mg/ml) to free the interstitial cells. Cells were collected by centrifugation and cultured in M199 growth medium supplemented with 10% fetal bovine serum, penicillin G (100 units/ml), streptomycin (100 mg/ml), and amphotericin B (0.25 g/ml). Cells of passage 3 to 6 were used for the experiments.
Lentiviral expression of miR-486 antagomir was applied to determine its effect on calcium deposition induced by TGF-␤1 (0.005 g/ml) or BMP-2 (0.100 g/ml). AVICs were treated with lentivirus expressing miR-486 antagomir for 72 h and then stimulated with TGF-␤1 or BMP-2 in a conditioning medium (growth medium supplemented with 10 mmol/liter of ␤-glycerophosphate, 10 nmol/liter of vitamin D 3 , 10 nmol/liter of dexamethasone, and 8 mmol/liter of CaCl 2 ) for 14 days. Alizarin Red staining was applied to evaluate calcium deposition.
To determine the role of miR-204 in modulating the effects of TGF-␤1 and BMP-2 on AVIC expression of Osx and Runx2, cells were transfected with miR-204 mimic (5 nM), miR-204 antagomir (50 nM), or control miR (50 nM) with or without stimulation with TGF-␤1 or BMP-2.
Gene knockdown was applied to determine the role of the Smad-dependent pathway in mediating the effect of TGF-␤1 and BMP-2 on AVIC expression of miR-204. Cells were pretreated with Smad1 siRNA (50 nM), Smad3 siRNA (50 nM), or scrambled siRNA (50 nM) for 48 h, and then stimulated with TGF-␤1 or BMP-2 for 24 h.
To determine the role of Smurf2 in modulating the effects of TGF-␤1 and BMP-2 on AVIC expression of Osx and Runx2, cells were treated with lentiviral Smurf2 shRNA, or lentiviral control shRNA. Then cells were stimulated with TGF-␤1 or BMP-2.

Lentiviral transduction
Block-it lentiviral pol II miR RNAi expression plasmids were amplified using standard bacterial transformation and purified using EndoFree Plasmid Maxi Kit. Lentivirus expressing miR-486 mimic and antagomir and non-target miRNA was generated by Lipofectamine 2000 co-transfection of 293T-cells. After 48 h, lentiviral supernatants were collected and concentrated. AVICs were infected with lentivirus expressing miRNA mimic and co-transfected with TransDux transduction reagent.

Immunoblotting
Immunoblotting was performed, as described previously (37), to analyze protein levels of Smad1, Smad3, Osx, Runx2, and Smurf2. ␤-Actin levels were examined for normalization of loading. In brief, whole cell lysates were prepared with a sample buffer (100 mmol/liter of Tris-HCl, pH 6.8, 2% SDS, 0.02% bromphenol blue, and 10% glycerol) after treatment of cells. Proteins in cell lysate were fractioned by 4 -20% SDS-PAGE and subsequently transferred to PVDF membranes. After being blocked with 5% skim milk solution, membranes were incubated with primary antibodies, followed by peroxidase-linked secondary antibodies specific to the primary antibodies. Protein bands were visualized with ChemiDoc TM MP imaging System (Bio-Rad). Band density was analyzed using the ImageJ software (Wayne Rasband, National Institutes of Health, Bethesda, MD).

Chromatin immunoprecipitation
Chromatin was prepared from human AVICs and ChIP was performed using ChromaFlash TM Chromatin Extraction Kit and ChIP Assay Kit according to the manufacturer's instruction (Epigentek Group Inc.). The sheared chromatin were immunoprecipitated with antibodies against Smad1, Smad3, non-immune IgG (negative control), or RNA polymerase II (positive control) at room temperature for 90 -120 min. DNA-protein complexes were collected and eluted. ChIP DNA was used as a template for PCR. Primers specific for detection of Smadsbinding site (amplifying a region from Ϫ1223 to Ϫ1056 of the miR-204 promoter) were as follows: 5Ј-TGATTGGGCTGAA-CAGATTG-3Ј; 5Ј-TGCACAGGGTCAGAATCATC-3Ј. PCR products were separated on a 1% agarose gel. Input chromatin isolated before the immunoprecipitation was used to control for equal amounts of input DNA (39).

Staining for calcium deposits
Alizarin Red staining was performed as described previously (28,40). Briefly, cell monolayers and aortic valves were washed twice with phosphate-buffered saline (PBS) and fixed for 15 min in 4% paraformaldehyde. Specimens were incubated with 0.2% Alizarin Red solution (pH 4.2) to examine calcium deposits. Specimens were washed with distilled water. Alizarin Red stains were examined and photographed with a Nikon Eclipse TS100 microscope (Tokyo, Japan). To quantitatively analyze Alizarin Red staining, Alizarin Red stains were bleached with 10% acetic acid at 85°C. Supernatant was spectrophotometrically analyzed at 450 nm (41).

Statistical analysis
All results are expressed as mean Ϯ S.D. Comparisons between groups were performed using SPSS 13.0 software with one-way analysis of variance with the post hoc Bonferroni/ Dunn test. Differences were confirmed with the Kruskal-Wallis H test. A difference was considered significant at p Յ 0.05.