|Home | About | Journals | Submit | Contact Us | Français|
Myocardin is a muscle lineage–restricted transcriptional coactivator that has been shown to transduce extracellular signals to the nucleus required for SMC differentiation. We now report the discovery of a myocardin/BMP10 (where BMP10 indicates bone morphogenetic protein 10) signaling pathway required for cardiac growth, chamber maturation, and embryonic survival. Myocardin-null (Myocd) embryos and embryos harboring a cardiomyocyte-restricted mutation in the Myocd gene exhibited myocardial hypoplasia, defective atrial and ventricular chamber maturation, heart failure, and embryonic lethality. Cardiac hypoplasia was caused by decreased cardiomyocyte proliferation accompanied by a dramatic increase in programmed cell death. Defective chamber maturation and the block in cardiomyocyte proliferation were caused in part by a block in BMP10 signaling. Myocardin transactivated the Bmp10 gene via binding of a serum response factor–myocardin protein complex to a nonconsensus CArG element in the Bmp10 promoter. Expression of p57kip2, a BMP10-regulated cyclin-dependent kinase inhibitor, was induced in Myocd–/– hearts, while BMP10-activated cardiogenic transcription factors, including NKX2.5 and MEF2c, were repressed. Remarkably, when embryonic Myocd–/– hearts were cultured ex vivo in BMP10-conditioned medium, the defects in cardiomyocyte proliferation and p57kip2 expression were rescued. Taken together, these data identify a heretofore undescribed myocardin/BMP10 signaling pathway that regulates cardiomyocyte proliferation and apoptosis in the embryonic heart.
In vertebrates, the heart is the first organ to form and its function is required for diffusion of nutrients to the embryo and removal of waste (1). Formation of the embryonic heart is a complex process involving a series of sequential morphogenetic events, including cardiogenic induction, cardiomyocyte differentiation and proliferation, specification of noncardiomyocyte lineages, formation and patterning of the primitive heart tube, looping morphogenesis, and finally, chamber maturation (2). These events are governed by an ancient developmental program involving the interplay of cardiogenic transcription factors and growth factors secreted from cardiomyocytes and nonmyocytic cell lineages. Mutations in the genes encoding these cardiogenic transcription factors and growth factors are associated with common forms of congenital heart disease (3, 4).
Cardiogenic transcription factors lie at the core of the developmental pathways regulating specification of cardiogenic precursors and morphogenesis of the heart (1, 5, 6). Serum response factor (SRF) is a MADS (MCM1, AGAMOUS, DEFICIENS, SRF) box transcription factor that is enriched in mesodermal lineages (7). Conditional ablation of the Srf gene in the embryonic mouse heart leads to embryonic lethality attributable to cardiac insufficiency during chamber maturation (8–10). Our group and others have shown that the cardiomyocyte- and SMC-restricted transcriptional coactivator myocardin physically associates with SRF and synergistically activates transcription of a subset of CArG (CC(A/T)6GG) box–containing genes (11). Expression of myocardin is first observed in the cardiac crescent, and subsequently, it is expressed throughout the embryonic myocardium and postnatal heart (12). In addition, myocardin is expressed abundantly in vascular and visceral SMCs. Myocardin-null (Myocd–/–) embryos survive through E10.5, which has been attributed to a block in vascular SMC differentiation (13). Ablation of the Myocd gene in neural crest–derived vascular SMCs revealed that myocardin promotes the contractile vascular SMC phenotype (14). In contrast, much less is understood about the function of myocardin in the embryonic and adult heart. In Xenopus laevis, morpholino-directed knockdown of myocardin mRNA results in inhibition of cardiac development and the absence of expression of cardiac differentiation markers with disruption of cardiac morphological processes (15). Ablation of the Myocd gene in the adult mouse heart leads to the rapid onset of dilated cardiomyopathy and heart failure (16). However, the molecular basis of myocardin function in the embryonic and adult heart remains unclear.
Members of the bone morphogenetic protein (BMP) family of secreted growth factors regulate multiple steps in the cardiogenic program (17, 18). Bmp10, a cardiac-restricted BMP family member, plays a critical role in regulating development of the heart (19, 20). Bmp10 is first expressed in the mouse heart at embryonic day E8.75 and subsequently becomes enriched within the trabecular myocardium between E9.0 and E13.5: a developmental window devoted to cardiac growth and chamber maturation (20). Bmp10–/– embryos survive through E10.5, but succumb from heart failure associated with a normally patterned, but severely hypoplastic, heart (20). Transgenic mice that overexpress Bmp10 during postnatal development demonstrate a 50% reduction in cardiac size (21). Interestingly, a substitution variant of BMP10, Thr326Ile, was identified in patients with hypertensive dilated cardiomyopathy (22). However, it is not clear what mechanisms regulate expression of BMP10 during embryonic or postnatal development.
In this report, we generated and characterized myocardin-null and conditional mutant mice in which the Myocd gene was selectively ablated in cardiomyocytes. Myocd–/– embryos survive until E9.5–E10.0 and exhibit severe defects in cardiac morphogenesis and function. Selective ablation of the Myocd gene in the embryonic heart (Nkx2-5Cre+MyocdF/F) recapitulates many of the defects observed in Myocd–/– embryos, with embryonic lethality observed at E13.5. Myocardial hypoplasia is attributable to a decrease in cardiomyocyte proliferation accompanied by a dramatic increase in cardiomyocyte apoptosis. Surprisingly, the observed defects in cardiac structure and function are attributable, at least in part, to myocardin-induced activation of the Bmp10 gene. Bmp10, in turn, regulates expression of several cardiogenic transcription factors, including Nkx-2.5 and Mef2c and the cyclin-dependent kinase inhibitor p57kip2. Remarkably, when grown ex vivo, defects in cardiomyocyte proliferation in the hearts of E9.5 Myocd–/– embryos are rescued in the presence of Bmp10-conditioned medium. These data demonstrate that myocardin lies upstream of Bmp10 signaling in a cardiogenic program regulating cardiac growth, chamber maturation, and ultimately, embryonic survival.
Myocd–/– mice were generated by selectively deleting exon 8 of the Myocd gene (Supplemental Figure 1, A–D; supplemental material available online with this article; doi: 10.1172/JCI63635DS1). This gene-targeting strategy differs from that described in an earlier report resulting in deletion of Myocd exons 8 and 9 (13). Genotype analysis of 824 embryos harvested from staged Myocd+/– × Myocd+/– intercrosses revealed that Myocd–/– embryos survive until E9.5–E10.0 (Table (Table1).1). In situ hybridization analyses performed with a myocardin antisense riboprobe confirmed that myocardin mRNA is expressed abundantly throughout the embryonic heart (Figure (Figure1A).1A). In contrast, myocardin mRNA was markedly attenuated in hearts of E8.5–E9.5 Myocd–/– embryos (Figure (Figure1B).1B). At E9.5 and E10.0, Myocd–/– embryos were readily identifiable by their size, which was consistently smaller than that of WT and heterozygote control littermates (Figure (Figure1,1, C and D). By E10.5, Myocd–/– embryos were not viable and were undergoing various states of reabsorption. Moreover, at or before E10.0, all viable Myocd–/–mutant embryos displayed pericardial effusions indicative of heart failure (Figure (Figure11D).
Inspection of hearts harvested from E10.0 Myocd–/– mutant embryos revealed that these hearts had undergone rightward looping with the future left ventricle directed toward the left side of the pericardial cavity and the bulbus-cordis region located on the right side of the pericardial cavity (Figure (Figure1,1, E–H). 3D high resolution optical mapping of E9.5 control and mutant embryos immunostained with MLC2v antibody to identify cardiac myocytes confirmed that the hearts of Myocd–/– embryos have completed looping morphogenesis (Supplemental Videos 1 and 2). In addition, scanning EM of E9.5 WT and Myocd–/– hearts confirmed that looping morphogenesis had occurred in the Myocd–/– hearts (Figure (Figure1,1, I and K). However, sections cut through the ventricle of E9.5 Myocd–/– embryos revealed gross defects in cardiac growth and chamber maturation characterized by thinning of the ventricular myocardium (Figure (Figure1,1, J and L) and a block in trabecularization (Figure (Figure1,1, J and L, and Figure Figure2,2, T and X). In E9.5 WT embryos, the ventricular compact zone is composed of 3–5 layers of cardiomyocytes, and the process of trabecular invagination has begun (Figure (Figure1J).1J). In contrast, in Myocd–/– hearts, the compact zone is composed of 1–2 layers of cardiomyocytes, and trabecularization is rarely observed (Figure (Figure11L).
Histological analyses of hearts of control and Myocd–/– embryos revealed subtle differences in development of the compact zone myocardium as early as E8.75 (Figure (Figure2,2, A–H). In control embryos, the compact zone consisted of 3–5 layers of tightly packed cardiomyocytes (Figure (Figure2,2, A–D). In contrast, the compact zone of Myocd–/– hearts consisted of only 1–3 layers of irregularly spaced cardiomyocytes (Figure (Figure2,2, E–H). Between E9.0 and E9.5, the loss of compact zone cellularity in Myocd–/– embryos became even more pronounced (Figure (Figure2,2, I–X). In addition, the trabecular myocardium was severely attenuated or absent in the mutant embryos (Figure (Figure2,2, M–P and U–X). Marked changes in development of the atrioventricular canal (AVC) were also observed in E9.0–E9.5 Myocd–/– embryos (Figure (Figure2,2, I–X). At E9.0, expansion of the truncus arteriosus region of outflow tract (OFT) and AVC was readily observed in control embryos (Figure (Figure2,2, I–L). In contrast, the OFT and AVC were underdeveloped in Myocd–/– hearts (Figure (Figure2,2, M–P). Of note, the observed decrease in myocardial cellularity in the hearts of E8.75–E9.0 Myocd–/– embryos occurred prior to changes attributable to a block in vascular SMC differentiation (23).
To examine cardiac function in Myocd–/– embryos, echocardiography was performed in utero on a series of E9.5 control and Myocd–/– embryos. Telemetric ECG recordings revealed severe bradycardia, which is often associated with heart failure during embryonic development, in Myocd–/– embryos compared with WT littermates (63 ± 8 beats/minute in Myocd–/– embryos [n = 10] versus 139 ± 11 beats/minute in WT littermates [n = 12]; P < 0.0001). In WT control embryos, synchronous atrial and ventricular cardiac contractions were readily identified (Supplemental Videos 3 and 4). However, in E9.5 Myocd–/– embryos, a marked decrease in systolic function was observed (Supplemental Videos 5 and 6). In some planes of section, a small pericardial effusion was also apparent (Supplemental Video 5). Comparison of systolic to diastolic ventricular dimensions, or fractional shortening, in 6 E9.5 controls and 6 Myocd–/– embryos revealed a decrease from 37.3% ± 3.5% in control embryos to 18.5% ± 2.3% in Myocd–/– embryos (P < 0.01). Taken together, these data demonstrate that heart failure contributed to the embryonic demise of Myocd–/– embryos.
To determine whether the myocardial hypoplasia observed in Myocd–/– hearts resulted from a decrease in cardiomyocyte proliferation and/or an increase in apoptosis, hearts harvested from E8.5–E9.5 control and Myocd–/– embryos were immunostained for markers of cell proliferation and apoptosis. At E9.5, a period of rapid expansion of the compact and trabecular zones (6), WT hearts demonstrated robust cardiomyocyte proliferation. Expression of phospho-histone H3 (pHH3), which correlates with mitotic chromosomal condensation, was observed in 24.0% of cardiomyocytes compared with only 9.6% of cardiomyocytes in Myocd–/– hearts (Figure (Figure3,3, A–C) (P < 0.01). Proliferating cell nuclear antigen (PCNA) was observed in 68% of cardiomyocytes in WT hearts compared with 13% of cardiomyocytes in mutant hearts (Figure (Figure3,3, D–F) (P < 0.01). Expression of the G1/S-specific cyclin, cyclin D2, was observed in 55% of cardiomyocytes in WT hearts, but only 18.8% of mutant cardiomyocytes (Figure (Figure3,3, G–I) (P < 0.01). BrdU, which was injected into pregnant dams 2 hours prior to embryo harvest, was incorporated in 12.5% of cardiomyocytes in WT hearts compared with 2.5% of cardiomyocytes in the mutant hearts (Figure (Figure3,3, J–L) (P < 0.01). Taken together, these data demonstrate a marked decrease in embryonic cardiomyocyte proliferation in E9.5 Myocd–/– mutant hearts.
TUNEL staining of hearts harvested from E8.5–E9.5 WT and Myocd–/– mutant embryos revealed a marked increase in cardiomyocyte apoptosis in Myocd–/– mutant embryos compared with WT littermates (Figure (Figure3,3, M–O). As early as E8.5, a 3-fold increase in cardiomyocyte apoptosis was observed in the mutant hearts (3.2% ± 0.6% versus 1.0% ± 0.2%; P < 0.01). Remarkably, at E9.0, a 12-fold induction in cardiomyocyte apoptosis was observed in Myocd–/– mutant hearts compared with WT hearts (12.6% ± 2.2% versus 1.0% ± 0.1%; P < 0.01). Similarly, at E9.5, a 13-fold induction in cardiomyocyte apoptosis was observed in Myocd–/– hearts compared with WT hearts (13.3% ± 3.5% versus 1.0% ± 0.1%; P < 0.01). Electron microscopic analyses revealed ultrastructural changes indicative of widespread cardiomyocyte apoptosis in E9.5 Myocd–/– hearts, including nuclear chromatin aggregation and fragmentation and cytoplasmic apoptotic body formation (compare Figure Figure3,3, P and Q). In addition, swelling of the mitochondria with breakdown of the mitochondrial membranes was observed in Myocd–/– hearts (compare Figure Figure3,3, R and S). These analyses demonstrate that myocardial hypoplasia in Myocd–/– embryos is attributable to both a decrease in cardiomyocyte proliferation and a dramatic increase in cardiomyocyte apoptosis and that these changes begin at least as early as looping of the primitive heart tube and accelerate during expansion of the compact zone and initiation of trabecularization.
To determine whether the cardiac phenotype observed in Myocd–/– embryos was attributable to the cell-autonomous loss of myocardin function in cardiac myocytes, Nkx2-5Cre+MyocdF/F conditional mutant mice were generated and characterized as harboring a cardiomyocyte-restricted mutation in the Myocd gene. Of note, the Nkx2-5 promoter was utilized in these experiments because Nkx2-5Cre–mediated recombination was observed at E7.5 throughout the cardiac crescent, thereby elucidating the function of myocardin during the earliest stages of cardiogenesis (24). Genotype analysis of E8.5–P21 littermates generated from intercrosses of Nkx2-5Cre+MyocdF/+ mice revealed that Nkx2-5Cre+MyocdF/F mutant mice survive until E13.5–E14.5 (Table (Table2).2). E13.5 Nkx2-5Cre+MyocdF/F mutant embryos demonstrated signs of heart failure, including generalized edema and large pericardial effusions (Figure (Figure4,4, A and B). As anticipated, in E12.5 Nkx2-5Cre+MyocdF/F hearts, myocardin mRNA was markedly attenuated (compare Figure Figure4,4, E and F). Quantitative RT-PCR (qRT-PCR) performed with cardiac mRNA harvested from E12.5 control and conditional mutant embryos revealed a 90% decrease in expression of myocardin mRNA in the conditional mutant hearts. In contrast, comparable levels of myocardin mRNA were observed in visceral SMCs populating the trachea and esophagus of control and mutant embryos (Figure (Figure4,4, E and F). Consistent with these observations, nuclear expression of myocardin was reduced dramatically in Nkx2-5Cre+MyocdF/F mutant hearts compared with control hearts (Figure (Figure4,4, G and H). Breeding of the Myocd conditional mutant mice into a Rosa26 genetic background confirmed that WT (Nkx2-5Cre+MyocdF/+Rosa26+) and myocardin-deficient (Nkx2-5Cre+MyocdF/FRosa26+) cardiomyocytes are indeed derived from Nkx-2.5–expressing cardiomyocytes and/or progenitor cells (Figure (Figure4,4, C and D). Taken together, these data demonstrate that the phenotype observed in Nkx2-5Cre+MyocdF/F conditional mutant embryos is attributable to cardiomyocyte-restricted loss-of-function of myocardin.
Consistent with the phenotype of Myocd–/– mutant embryos, the hearts of E13.5–E14.5 Nkx2-5Cre+MyocdF/F conditional mutant embryos demonstrated severe myocardial hypoplasia (Figure (Figure4,4, I–L). In E13.5 Nkx2-5Cre+MyocdF/F embryos, the ventricular compact zone consisted of only 1–4 layers of irregularly spaced cardiomyocytes versus 5–10 layers of densely packed cardiomyocytes in control littermates (Figure (Figure4,4, K and L). In Nkx2-5Cre+MyocdF/F embryos, a large ventricular septal defect (VSD) that is not normally observed at this stage of development was routinely observed (Figure (Figure4,4, I and J). However, defects in aortic and mitral valve formation were not observed in E13.5 Nkx2-5Cre+MyocdF/F embryos (Figure (Figure4,4, I and J, and data not shown). Once again, a significant decrease in cardiomyocyte proliferation was observed in E13.5 conditional mutant embryos (Figure (Figure4,4, M–O). 46% ± 6% of cardiomyocytes in control hearts incorporated BrdU versus 26% ± 3.3% in the conditional mutant hearts (P < 0.01). Even more remarkable was the 11-fold increase in the cardiomyocyte apoptosis in E12.5 Nkx2-5Cre+MyocdF/F hearts compared with control littermates (Figure (Figure4,4, P–R). The apoptotic index observed in conditional mutant hearts was 11.7% ± 1.1% versus 1.0% + 0.1% in control hearts (P < 0.01). These data demonstrate that ablation of the Myocd gene in embryonic cardiomyocytes (or progenitor cells) causes a block in cardiac growth and chamber maturation associated with decreased cardiomyocyte proliferation and a marked increase in cardiomyocyte apoptosis.
To examine the molecular basis of myocardin function in the embryonic heart, qRT-PCR was performed with mRNA harvested from E9.0 WT and Myocd–/– hearts and primers that amplify genes involved in the cardiogenic program (Figure (Figure5A).5A). As anticipated, myocardin mRNA was expressed abundantly in control hearts, but was undetectable in Myocd–/– hearts (Figure (Figure5A).5A). The closely related transcriptional coactivators Mrtf-A/Mkl1 and Mrtf-B/Mkl2 were also expressed at comparable levels in control and Myocd–/– hearts. In addition, the Gata-4 and Srf genes were expressed at comparable levels in control and Myocd–/– hearts (Figure (Figure5A).5A). In contrast, Nkx-2.5, Mef2c, and Tbx5 gene expression were markedly attenuated in the hearts of Myocd–/–embryos compared with WT controls. These data suggested that Bmp10, which regulates Nkx2.5, Mef2c, and Tbx5 gene expression in the embryonic heart, but not Gata4 or Srf gene expression, may also be regulated (directly or indirectly) by myocardin in the embryonic heart. Indeed, qRT-PCR confirmed that Bmp10 mRNA was severely attenuated in the hearts of Myocd–/– embryos compared with WT embryos (Figure (Figure55A).
Immunostaining of sections prepared from E9.5 control and Myocd–/– embryos confirmed that expression of the cardiogenic transcription factors Nkx2.5, Mef2c, and Tbx5 is indeed reduced in the hearts of Myocd–/– embryos compared with controls (Figure (Figure5B).5B). In contrast, Gata4 and Srf were expressed at comparable levels in the hearts of Myocd–/– and control embryos (Figure (Figure5B).5B). This observation was perplexing, as Nkx2-5–/– and Mef2c–/– embryos exhibited defects in looping morphogenesis, while Myocd–/– and Nkx2-5Cre+MyocdF/F embryos did not (25–27). This suggested that expression of Nkx-2.5 and/or Mef2c may be differentially regulated by myocardin (directly or indirectly) during different stages of heart development in the embryo. Consistent with this hypothesis, Nkx-2.5 and Mef2c were expressed abundantly throughout the hearts of E8.5 WT and Myocd–/– embryos (Supplemental Figure 2). Interestingly, Nkx2-5 and Mef2c were also developmentally downregulated in hearts of Nkx2-5Cre+MyocdF/F conditional mutant embryos, but decreased expression of these factors was delayed by approximately 12 to 24 hours compared with Myocd–/– embryos (Supplemental Figure 2). Of note, the temporal alteration in expression of Nkx2-5 and Mef2c once again recapitulated changes observed in hearts of Bmp10–/– embryos (20).
The fetal heart program coordinates expression of a unique set of cardiac-, skeletal-, and smooth muscle-restricted genes encoding myofibrillar isoforms (28). Despite significant alterations in several key cardiogenic transcription factors in Myocd–/– mutant embryos, comparable levels of myosin heavy chain (MF20), myosin light chain 2V (MLC2v), and cardiac troponin T (cTnT) were observed in the hearts of E8.5–E9.5 WT and Myocd–/– embryos (Figure (Figure5B).5B). An important component of the fetal heart program is a subset of SRF/CArG box–dependent genes associated with the SMC contractile phenotype, including SM α-actin (SMA) and SM22α (29). In this regard, it is noteworthy that neither of these genes was expressed in the dorsal aorta of E9.5 Myocd–/– mutants (13). This led us to speculate that the observed changes in cardiac structure and function may relate to a block in SRF-regulated genes that compose the fetal heart program. However, comparable levels of SMA and SM22α were observed in the hearts of Myocd–/– and control littermates (Figure (Figure5B).5B). These data demonstrate that distinct mechanisms have evolved to regulate expression of SRF-regulated genes that are coexpressed in vascular SMCs and the embryonic heart.
The timing of embryonic lethality, block in cardiac chamber maturation, and loss of the Bmp10 gene and target gene expression in Myocd–/– embryos suggested strongly that the transcriptional coactivator myocardin lies upstream and regulates Bmp10 signaling in the embryonic heart. Consistent with this observation, immunostaining and quantitative immunoblot analysis revealed a 90% decrease in Bmp10 protein in E9.5 Myocd–/– hearts compared with control hearts (Figure (Figure6,6, A–E). BMP signaling is transduced to the nucleus via activation of Smads 1, 5, and 8 by MAPK and GSK3 (30). Remarkably, immunostaining of E9.5 control and Myocd–/– hearts with an antibody that recognizes phosphorylated (activated) Smad1/5/8 demonstrated a dramatic reduction in Smad1/5/8 signaling in the hearts of Myocd–/– embryos compared with controls, suggesting generalized repression of BMP signaling in Myocd–/– hearts (Figure (Figure6,6, F–I).
It is also noteworthy that Bmp10 has been shown to suppress the expression and activity of p57kip2, a negative regulator of the cell cycle, which could explain, at least in part, the block in cardiomyocyte proliferation observed in Myocd–/– hearts (20). Consistent with this hypothesis, expression of p57kip2 was induced in the ventricular compact zone and trabecular myocardium and OFT of E9.0 Myocd–/– embryos compared with WT embryos (Figure (Figure6,6, J–M). Quantitative analyses of p57kip2 antibody–stained sections demonstrated greater than a 2-fold increase in p57kip2-positive nuclei in the hearts of Myocd–/– embryos compared with control embryos (P < 0.01). Taken together, these data demonstrate a block in BMP signaling in myocardin-null hearts manifested as (a) downregulation in Bmp10 gene expression, (b) downregulation of phosphorylated Smad1/5/8 signaling, (c) downregulation of the Bmp10-regulated cardiogenic transcription factors Nkx2.5 and Mef2c, and finally, (d) increased expression of the cell cycle inhibitor p57kip2.
As schematically depicted in Figure Figure7A,7A, a search of GenBank with the TRANSFAC database revealed that the mouse Bmp10 promoter contains a near consensus (9 of 10 nucleotide identity) SRF-binding site, or CArG motif, at position –515 to –524. In addition, a second near-consensus CArG box (CCTAAACTGG) is located at bp –1163 in the mouse Bmp10 promoter. Near-consensus CArG boxes, each containing a 1-bp substitution mutation in the central (AT)6 nucleotide sequence, were also identified in the rat, human, and chicken BMP10 promoters (Figure (Figure7A).7A). EMSA performed with nuclear extracts prepared from Cos7 cells transfected with an expression plasmid encoding the 935–amino acid cardiac-restricted isoform of myocardin or SRF revealed a low-mobility nuclear protein complex that bound specifically to the nonconsensus CArG box motif in the Bmp10 promoter (Figure (Figure7B,7B, lanes 1–5 and 9). This nuclear protein complex was markedly attenuated when increasing amounts of unlabeled nonconsensus CArG box oligonucleotides were added to the incubation reaction (Figure (Figure7B,7B, lanes 6, 7). Moreover, addition of SRF antibody to the binding reaction super-shifted this nuclear protein complex, demonstrating that this nuclear protein complex contains Srf or an antigenically related protein (Figure (Figure7B,7B, lanes 8 and 10). Similarly EMSAs performed with the nonconsensus CArG box motif in the human BMP10 promoter demonstrated specific binding of a myocardin-SRF protein complex to the human BMP10 promoter (data not shown).
To determine whether this CArG motif is required for transcriptional activity of the 795-bp mouse Bmp10 promoter, a series of transient cotransfection analyses were performed. Cotransfection of Cos7 cells with the pcDNA.MyocdL expression plasmid, encoding the 935–amino acid isoform of myocardin, and the pBmp10.luc reporter plasmid demonstrated a 38-fold increase in luciferase activity compared with cells cotransfected with the pcDNA3 control plasmid and pBmp10.luc (Figure (Figure7C,7C, lanes 2 and 5). Similarly, cotransfection with the pcDNA.MyocdS expression plasmid, encoding the 856-amino acid smooth muscle–restricted isoform of myocardin, which binds Srf, but not Mef2c (31), and pBmp10.luc resulted in a 35-fold increase in luciferase activity compared with cells cotransfected with pcDNA3 and pBmp10.luc (Figure (Figure7C,7C, lanes 2 and 6). In contrast, cotransfection of Cos7 cells with the pμBmp10.luc reporter plasmid, which contains the Bmp10 promoter with a mutation that abolishes binding of SRF/myocardin to the embedded CArG motif (Figure (Figure7A),7A), and pcDNA.MyocdL or pcDNA.MyocdS, respectively, failed to increase luciferase activity above levels observed in cells cotransfected with pcDNA3 and to pBmp10.luc or pμBMP10.luc, respectively (Figure (Figure7C,7C, lanes 2, 7, and 8).
In addition, ChIP assays performed with crosslinked DNA harvested from E9.5 hearts confirmed that an SRF-myocardin complex binds to the nonconsensus CArG element in the mouse Bmp10 promoter in vivo. As anticipated, when chromatin was immunoprecipitated with an anti-SRF antibody and the input DNA amplified with primers flanking the consensus CArG box in the c-fos promoter, 18-fold enrichment of the input DNA was observed (Figure (Figure7D).7D). Similarly, when the crosslinked cardiac chromatin was amplified with PCR primers flanking the nonconsensus CArG box in the mouse Bmp10 promoter, 28-fold enrichment of the input DNA was observed (Figure (Figure7D).7D). Moreover, 20-fold enrichment was observed when crosslinked cardiac chromatin was immunoprecipitated with anti-myocardin antibody and DNA was amplified with primers flanking the CArG box in the Bmp10 promoter (Figure (Figure7D).7D). Taken together, these data demonstrate that a nuclear protein complex containing myocardin and Srf binds to a nonconsensus and conserved CArG box in the Bmp10 promoter transactivating the Bmp10 gene.
These findings suggested a model wherein myocardin activates transcription and expression of BMP10, which in turn suppresses p57kip2 expression, thereby inducing cardiomyocyte proliferation in the embryonic heart. To test this hypothesis, NIH3T3 cells were stably transduced with control lentivirus encoding EGFP or with a lentivirus encoding mouse Bmp10, generating control and Bmp10-conditioned medium, respectively. Immunoblot analysis confirmed that stably transduced NIH3T3 cells secreted abundant Bmp10 (data not shown). To determine whether Bmp10-conditioned medium rescues the defect in cell proliferation observed in Myocd–/– mutant hearts, hearts were isolated from E9.5 WT control and Myocd–/– embryos and grown ex vivo for 48 hours in the presence of Bmp10-conditioned medium (n = 5) or control conditioned medium (n = 5). As anticipated, WT hearts were significantly larger than Myocd–/– hearts and consistently beat more rapidly and vigorously over the course of 48 hours in culture (Supplemental Videos 7–10 and Figure Figure8,8, A–F). In control hearts, following 24 hours of incubation with BrdU, 88.2% ± 0.6% of cardiac myocytes incorporated BrdU (Figure (Figure8G).8G). In contrast, only 56.1% ± 0.9% of ventricular cardiomyocytes incorporated BrdU in Myocd–/– hearts (Figure (Figure8H)8H) (P < 0.01 versus control hearts). However, when grown in the presence of Bmp10-conditioned medium, 85.2% ± 0.7% of ventricular cardiomyocytes in Myocd–/– hearts incorporated BrdU (Figure (Figure8I)8I) (P < 0.001 versus control medium). Consistent with these findings, 22.7% ± 0.3% of ventricular cardiomyocytes expressed the cell proliferation marker pHH3 (Figure (Figure8J).8J). Of note, centronuclear orange staining indicative of prophase/metaphase and speckled nuclear staining indicative of anaphase were observed with this mitotic marker. In contrast, 5.9% ± 0.1% of ventricular cardiomyocytes expressed pHH3 in Myocd–/– hearts (Figure (Figure8K)8K) (P < 0.001 versus WT hearts). However, when grown in the presence of BMP10-conditioned medium, 13.6% ± 0.1% of Myocd–/– ventricular cardiomyocytes were pHH3 positive (Figure (Figure8L)8L) (P < 0.01 versus control medium). Moreover, consistent with the changes observed in vivo, 7.9% ± 0.7% of cardiomyocytes in control hearts expressed the cell cycle inhibitor p57kip2 (Figure (Figure8M),8M), while 35.3% ± 0.9% of ventricular cardiomyocytes in Myocd–/– hearts expressed p57kip2 (Figure (Figure8N).8N). However, when grown in the presence of Bmp10-conditioned medium, only 9.9% ± 0.8% of ventricular cardiomyocytes expressed p57kip2 (Figure (Figure8O)8O) (P < 0.001 versus control medium). In contrast, no significant difference was observed in apoptosis as detected by TUNEL staining between Myocd–/– hearts grown in control and Bmp10-conditioned medium (data not shown).
The molecular programs regulating development and morphogenesis of the heart are dependent upon the capacity of cardiomyocytes to transduce and respond to developmental cues and environmental signals. Analyses of genetically engineered mice harboring null and conditional loss-of-function mutations in the Myocd gene have shown that myocardin plays a critical role in regulating vascular SMC differentiation and patterning of the vascular system (13, 14). However, despite the finding that myocardin is expressed early and abundantly in the embryonic heart, its role, if any, in promoting cardiomyocyte differentiation and cardiac morphogenesis has remained unclear. The data presented in this report demonstrate unequivocally that in the embryonic heart, myocardin is not required for cardiac myocyte specification and/or looping morphogenesis, but is required for cardiac growth and chamber maturation, maintenance of cardiac function, and ultimately, embryonic survival. Surprisingly, the basis of myocardin function is not predicted by its activity in vascular SMC differentiation, but rather via the capacity to regulate cardiomyocyte proliferation and apoptosis in the embryonic heart, acting at least in part via activation of the Bmp10 gene.
As schematically shown in Figure Figure9,9, in the embryonic heart, myocardin is poised to respond to developmental cues, growth factors, and hemodynamic signals influencing the balance of cardiomyocyte proliferation and apoptosis required for atrial and ventricular chamber maturation. After completion of looping morphogenesis, coincident with the onset of rapid growth associated with chamber maturation, myocardin acting via induction of Bmp10 gene expression represses activity of the cell cycle inhibitor p57kip2, stimulating cardiomyocyte proliferation. During this developmental window, myocardin is also required for maintenance of a subset of key cardiogenic factors, including Nkx2-5 and Mef2c (20). BMP-induced activation of Nkx2-5 gene transcription occurs via binding of Smad1/4 to the Nkx2-5 transcriptional enhancer (32). Consistent with these observations, conditional ablation of the Nkx2-5 gene at E12.5 leads to a block in chamber maturation, recapitulating the phenotype observed in Myocd conditional mutant embryos (33). Moreover, in the primitive heart, at least as early as E8.5, myocardin represses cardiomyocyte apoptosis, preserving cardiac function. Taken together, these data demonstrate that as a transcriptional coactivator, myocardin transduces signals influencing cardiomyocyte proliferation, structural organization of the cardiomyocyte, and programmed cell death required for the late stages of cardiac morphogenesis needed for atrial and ventricular chamber maturation.
The demonstration that cardiogenic factors including Nkx2-5 and Mef2c are downregulated in the hearts of E9.5 Myocd–/– embryos was initially perplexing because Myocd–/– embryos progress through the early stages of cardiogenesis, including formation of the primitive heart tube and looping morphogenesis, while Nkx2-5–/– and Mef2c–/– embryos exhibit a block in looping morphogenesis (25–27). What explains the divergent phenotype of Myocd–/– versus Nkx2.5–/– and Mef2c–/– mutant embryos? Surprisingly, at E8.0–E8.5, prior to initiation of cardiac looping, Nkx-2.5 and Mef2c are expressed at comparable levels in control and Myocd–/– embryos (Supplemental Figure 2). Similarly, comparable levels of Nkx-2.5 and Mef2c were observed at E8.5 in control and Bmp10–/– embryos (20). These data define a developmental window when myocardin (and Bmp10) regulates expression of the cardiogenic transcription factors Nkx-2-5 and Mef2c. This window occurs after formation of the primitive heart tube and looping morphogenesis (E8.0–E8.5 in the mouse) coincident with chamber maturation of the embryonic heart (E9.5). This observation highlights the role of myocardin as a transcriptional coactivator that is capable of transducing and responding to specific developmental cues in the embryonic heart.
Surprisingly, some, but not all, of myocardin-regulated functions in the embryonic heart are conserved in the heart during postnatal development. Myocd gene ablation in the adult heart is accompanied by dissolution of sarcomeric organization, disruption of the intercalated disc, and cell-autonomous loss of cardiomyocytes via apoptosis (16). The structural defects are attributable, at least in part, to a block in SRF-dependent genes encoding a subset of myofibrillar and structural proteins, including α-cardiac actin, MLC2v, desmin, and connexin 43 (16). Interestingly, each of these proteins is observed in the hearts of E9.5 Myocd–/– embryos (J. Huang, unpublished observations), suggesting that MRTF-A/MKL1 and/or MRTF-B/MKL2 may subserve a partially redundant function with myocardin in the embryonic heart: a capacity that is lost during postnatal development. In this regard, it is noteworthy that the structure of the sarcomere and the repertoire of genes encoding myofibrillar proteins are distinct in the embryonic and adult heart. In contrast, throughout development, myocardin represses programmed cell death in the cardiomyocyte.
The molecular basis of myocardin function in the embryonic heart was not anticipated by a decade of studies that defined its functions in SMCs and the vasculature (13, 14). Ectopic expression of myocardin in X. laevis embryos activates some, but not all, SRF-regulated genes encoding contractile proteins expressed in the fetal heart (15). The fetal heart program includes the transient expression of a select group of SRF-regulated genes encoding SMC-restricted contractile proteins, including SMA and SM22α, which are transiently expressed in the embryonic heart and reexpressed in the adult heart in response to hemodynamic stress. At E9.5, SM22α and SMA are not expressed in vascular SMCs populating the dorsal aorta of Myocd–/– embryos (13). This led us to postulate that the defects observed in the hearts of Myocd–/– embryos were caused, at least in part, by a block in expression of the subset of SRF-dependent genes encoding SMC-restricted contractile proteins. Interestingly, this is not the case, as comparable levels of SM22α and SMA were observed in the hearts of Myocd–/– embryos and control littermates. Taken together, these data demonstrate that distinct transcriptional mechanisms have evolved to control the expression of SRF-dependent genes encoding SMC contractile proteins in the embryonic heart and vasculature, respectively.
For many years it has been recognized that a subset of BMP growth factors plays critical roles in the morphogenetic program regulating development of the heart (17, 18). However, the mechanisms regulating expression of BMPs in the heart remain poorly understood. The findings that (a) Bmp10 gene expression is markedly downregulated in the hearts of Myocd–/– and cardiomyocyte-restricted conditional mutant embryos, (b) forced expression of myocardin transactivates the Bmp10 promoter via binding of a myocardin/SRF protein complex to a conserved CArG box, (c) expression of phosphorylated Smad1/5/8 is repressed in Myocd–/– hearts at E9.5, (d) Bmp10-regulated genes including Nkx2-5 and Mef2c are suppressed in the hearts of Myocd–/– and Bmp10–/– embryos, (e) p57kip2 is induced in the hearts of Myocd–/– and Bmp10–/– embryos, (f) Bmp10-conditioned medium rescues the proliferative defect in Myocd–/– mutant hearts and suppresses p57kip2, and most importantly, (g) shared morphogenetic defects are observed in the hearts of Myocd–/–, Nkx2-5Cre+MyocdF/F, and Bmp10–/– embryos strongly support the conclusion that myocardin (and SRF) regulates transcription of the Bmp10 gene in the embryonic heart. This conclusion was not anticipated by a decade of research into the function of myocardin, which had not suggested a role for myocardin in the regulation of BMP-related growth factors.
The demonstration of apoptosis in the hearts of Myocd–/– and Nkx2-5Cre+MyocdF/F mutant embryos reveals a conserved function for myocardin throughout development of the heart. Consistent with this observation, conditional ablation of Srf in the embryonic heart is associated with increased cardiomyocyte apoptosis and heart failure (8). In this regard, it is noteworthy that Bmp10 did not rescue apoptosis in Myocd–/– mutant hearts, suggesting strongly that myocardin-induced repression of programmed cell death is not linked to cell proliferation and is independent of BMP10 signaling. This primal function of myocardin (and SRF) appears to be cell autonomous and not dependent upon hemodynamic signals because the loss of myocardin in primary neonatal cardiomyocytes in cell culture rapidly induces apoptosis, acting via the intrinsic and extrinsic apoptotic pathways (16). This observation raises a fundamental question: what is the evolutionary advantage of a transcriptional coactivator required for cardiomyocyte survival? Are there circumstances in the embryonic heart and/or when the heart adapts to hemodynamic stress in which cardiomyocyte apoptosis is beneficial? Does myocardin play a unique role in the embryonic heart in response to hemodynamic stress? The answers to these questions will provide new insights into understanding the molecular basis of cardiac morphogenesis and may be relevant to understanding the pathogenesis of heart failure and cardiomyopathy.
To generate Myocd+/– mutant mice, homozygous myocardin conditional mutant mice (MyocdF/F) (16) were intercrossed with CMV-Cre (BALB/c-Tg(CMV-cre)1Cgn/J) transgenic mice (34). Southern blot analyses confirmed that the Myocd targeted allele was deleted and passed on through the germline (Supplemental Figure 1). Deletion of Myocd exon 8 creates an out-of-frame mutation and N-terminal truncation of the myocardin protein, which is effectively null (16). To delete Myocd selectively in embryonic cardiomyocytes, MyocdF/F mice were interbred with Nkx2-5Cre transgenic mice (24).
Immunohistochemistry analyses were performed with the protocols reported previously (14). Antibodies utilized for immunohistochemistry studies are included in Supplemental Table 1. Quantification of apoptosis in the embryonic heart was assessed by TUNEL staining, as described previously (16). Histological sections were counterstained with MF20 antibody (Hybridoma Bank, University of Iowa, Iowa City, Iowa, USA) to identify cardiomyocytes.
EM and scanning EM studies were performed as described previously (14). Whole-mount immunostaining was performed with monoclonal anti-MLC2v primary antibody (AXXORA) and HRP-conjugated IgG secondary antibody (Cell Signaling), as described (35). High-resolution optical mapping studies were performed on an Olympus BX51WI microscope equipped with a high-speed CMOS camera (MiCAM ULTIMA L; SciMedia Ltd.), as described (36).
Cardiac morphometric and echocardiographic measurements were performed as described previously (16). To measure systolic to diastolic ventricular dimension and calculate fractional shortening in E9.5 embryos, the outer edge of the common ventricle was traced during diastole and systole. The fraction shortening (%) was defined as (diastolic area – systolic area)/diastolic area.
DNA was isolated from the yolk sac of embryos or tail biopsies of weaned mice and genotype determined by PCR, as described previously (14). Real-time qRT-PCR was performed using the DNA Engine Opticon 2 Real Time Detection System (Applied Biosystems), as described previously (14). Expression of myocardin, Srf, Nkx-2.5, Mef2c, Mkl1, Mkl2, Tbx5, Gata4, Mlc2v, β-myosin heavy chain, Bmp10, and GAPDH genes was assessed with complementary primer pairs as described (14).
A search for conserved CArG motifs in Bmp10 promoter across species was performed by searching GenBank using the TRASFAC database with the identify threshold set to greater than 85%. Transient cotransfection analyses of C3H10T1/2 and Cos7 cells were performed as described previously (37). The pcDNA3.1-MyocdL expression plasmid encodes the 935–amino acid mouse myocardin protein, and the pcDNA3.1-MyocdS expression plasmid encodes the 856–amino acid mouse myocardin protein isoform; these were reported previously (14). Mammalian expression plasmids encoding mouse myocardin and SRF were generated in pcDNA3 as described (14). The 795-bp mouse Bmp10 promoter was generated via PCR and subcloned into SmaI/BglII-digested pGL3-Basic (Promega) luciferase reporter vector generating the pBmp10.luc reporter plasmid. The pμBmp10.luc promoter is identical to pBmp10.luc, but contains a 2-bp mutation in the mouse Bmp10 CArG box (CC to GG), which was generated using the QuikChange Site-Directed Mutagenesis kit (Stratagene). Transient cotransfections were performed using FuGENE 6 (Roche Applied Science), as described previously (14). Transcriptional activity was quantified by measuring firefly luciferase activity relative to the internal control TK-Renilla luciferase activity using the Dual Luciferase Assay System (Promega), as described (14). A minimum of 6 independent transfections were performed, and all assays were performed in triplicate. Results are reported as the relative luciferase activity ± SEM.
ChIP assays were performed using reagents and protocols from Upstate Biotechnology. Hearts harvested from E9.5 WT mouse embryos were homogenized and fixed with formaldehyde to crosslink chromatin. The crosslinked chromatin was immunoprecipitated with either an anti-SRF antibody (G20X; Santa Cruz Biotechnology), polyclonal anti-myocardin antibody (37), or preimmune rabbit sera (negative control). PCR primers to quantify enrichment of input DNA included the following: c-fos CArG: 5′-CGGTTCCCCCCCTGCGCTGCACCCTCAGAG-3′; 5′-AGAACAACAGGGACCGGCCGTGGAAACCTG-3′; Bmp10 CArG: 5′-ATGACCGTGACATTCATTCTTTCCTG-3′; 5′-TTTGCACGTGTGTATGACAGATGTTAAG-3′. Data are expressed as fold enrichment (arbitrary units) normalized to input DNA ± SEM.
Western blot analyses were performed as described previously (14). The hybridization signal was quantified using Image Quant 5.0 software, as described by the manufacturer (Molecular Dynamics Inc.).
Nuclear protein extracts were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific). Binding reactions were performed using the Light Shift Chemiluminescent EMSA Kit (Thermo Scientific) according to the manufacturer’s protocol. Biotin-labeled primers included the following: mouse Bmp10 CArG 5′-TTTATGCCATCTTTGGACATGT-3′; mouse Bmp10 mutCArG 5′-TTTATGGGATCTTTGGACATGT-3′; human BMP10 CArG 5′-GTCTTTCCTAATCTGGACACAA-3′; human BMP10 CArG 5′-TTGTGTCCAGATTAGGAAAGAC-3′.
The mouse Bmp10 cDNA was generated via PCR and subcloned into the pSMPUW-IRES-GFP lentiviral expression vector. HEK293T cells were transfected for Bmp10 lentiviral expression and EGFP control virus production and purification according to the manufacturer’s instructions (Cell Biolabs). NIH3T3 cells were transduced with lentiviral supernatants and cultured for 2 days in DMEM containing 10% FBS as described (20). Embryonic hearts were dissected from E9.5 embryos and cultured ex vivo for 48 hours as described (20). At 24 hours after isolation, BrdU was added to the control or Bmp10 conditioned medium and hearts were harvested 24 hours later for histological analyses.
All measurement data are expressed as mean ± SEM. The statistical significance of differences between groups was determined by 2-tailed Student’s t test. Differences were considered significant at a P value of less than 0.05.
All animal experimentation was performed under protocols approved by the University of Pennsylvania IACUC and in accordance with NIH guidelines.
We thank Jon Epstein, Mark Kahn, and Ed Morrisey for their advice and helpful comments. We thank Kurt Engleka for assistance with 3D optical mapping. This work was supported in part by NIH R01-HL102968 and R01-HL094520 and the Commonwealth of Pennsylvania.
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article: J Clin Invest. 2012;122(10):3678–3691. doi:10.1172/JCI63635.