Circulation Research. 2007;100:633-644
doi: 10.1161/01.RES.0000259563.61091.e8
(Circulation Research. 2007;100:633.)
© 2007 American Heart Association, Inc.
Myocardin-Related Transcription Factors
Critical Coactivators Regulating Cardiovascular Development and Adaptation
Michael S. Parmacek
From the University of Pennsylvania Cardiovascular Institute and Department of Medicine, University of Pennsylvania, Philadelphia.
Correspondence to Michael S. Parmacek, MD, 9123 Founders Pavilion, 3400 Spruce St, Philadelphia, PA 19104. E-mail michael.parmacek{at}uphs.upenn.edu
This Review is part of a thematic series on Transcription Factors, which includes the following articles:
Regulation of Vascular Inflammation and Remodeling by ETS Factors
Myocardin-Related Transcription Factors: Critical Coactivators Regulating Cardiovascular Development and Adaptation
Role of Kruppel-Like Transcription Factors in Endothelial Biology
Forkhead Factors in Cardiovascular Biology
Notch Signaling and Angiogenesis
Mukesh Jain Guest Editor
 |
Abstract
|
|---|
The association of transcriptional coactivators with DNA-binding
proteins provides an efficient mechanism to expand and modulate
genetic information encoded within the genome. Myocardin-related
transcription factors (MRTFs), including myocardin, MRTF-A/MKL1/MAL,
and MRTF-B/MKL2, comprise a family of related transcriptional
coactivators that physically associate with the MADS box transcription
factor, serum response factor, and synergistically activate
transcription. MRTFs transduce cytoskeletal signals to the nucleus,
activating a subset of serum response factordependent
genes promoting myogenic differentiation and cytoskeletal organization.
MRTFs are multifunctional proteins that share evolutionarily
conserved domains required for actin-binding, homo- and heterodimerization,
high-order chromatin organization, and transcriptional activation.
Mice harboring loss-of-function mutations in myocardin, MRTF-A,
and MRTF-B, respectively, display distinct phenotypes, including
cell autonomous defects in vascular smooth muscle cell and myoepithelial
cell differentiation and function. This article reviews the
molecular basis of MRTF function with particular focus on the
role MRTFs play in regulating cardiovascular patterning, vascular
smooth muscle cell and cardiomyocyte differentiation and in
the pathogenesis of congenital heart disease and vascular proliferative
syndromes.
Key Words: myocardin myocardin-related transcription factor serum response factor transcription smooth muscle cell cardiovascular development congenital heart disease
 |
Introduction
|
|---|
The association of inducible transcriptional coactivators and/or
corepressors with lineage-restricted transcription factors provides
an efficient mechanism to expand and modulate information encoded
within the genome. Transcriptional coactivators play a critical
role in regulating transcription of genes promoting cellular
differentiation, migration, and proliferation.
1 Cell lineagerestricted
transcription factors direct patterning of the cardiovascular
system through a complex series of steps into the mature heart,
arterial, and venous systems (reviewed elsewhere
2). Patterning
of the cardiovascular system is regulated by the inducible activity
of transcriptional coactivators and repressors. Transcriptional
coactivators respond to specific developmental cues and growth-related
signals modulating the activity of lineage-restricted transcription
factors regulating transcription of sets of genes required for
cell growth, differentiation, and survival. A fundamental property
of the vertebrate cardiovascular system is the capacity to respond
to stress by altering its functional properties to maintain
cardiovascular homeostasis. Transcriptional coactivators and
repressors regulate the response of the cardiovascular system
to vascular injury and hemodynamic stress by transducing extracellular
signals to the nucleus and modulating the biological and physiological
properties of the cell.
This review focuses on the recently discovered family of transcriptional coactivators known as myocardin-related transcription factors (MRTFs), which in metazoan species, include myocardin, MRTF-A, and MRTF-B.3,4 MRTFs physically associate with the MADS box transcription factor, serum response factor (SRF), and synergistically activate transcription of a subset of genes involved in cytoskeletal organization and muscle cell differentiation. This review focuses primarily on the function of MRTFs in the cardiovascular system. Because the majority of MRTF-related research performed to date has been performed in vitro or in mouse models, extrapolating these data to humans must be done with caution. Nevertheless, the genes encoding MRTF family members are conserved in humans, and the patterns of MRTF gene expression in humans generally recapitulates that observed in other vertebrate species. The reader is referred to several excellent general reviews discussing the functions of myocardin and MRTFs.59
 |
SRF: A Critical Modulator of Cardiovascular Growth and Differentiation
|
|---|
SRF is a member of the ancient MADS (
MCM1,
Agamous,
Deficiens,
SRF) family of transcription factors, which share a conserved
57-aa MADS-box domain that mediates homodimerization, association
with other factors, and DNA-binding activity (reviewed elsewhere
8)
(
Figure 1). SRF binds to an A/T-rich sequence (CCWWWWWWGG) that
has been designated as the CArG box.
1012 CArG boxes were
originally identified in transcriptional regulatory elements
controlling expression of a set of growth- or serum-responsive
genes including
c-fos and
egr-1.
13,14 Subsequently, CArG boxes
were identified in transcriptional regulatory elements controlling
expression of a subset of genes encoding myogenic contractile
and cytoskeletal proteins including

-cardiac actin, smooth muscle
(SM)-

-actin,

-skeletal actin, and SM22

.
1519 Microarray
studies and a genome-wide scan uncovered more than 150 genes
that appear to be direct transcriptional targets of SRF.
2024 Consistent with earlier reports, the majority of these SRF-regulated
genes are involved in cell growth and proliferation, cytoskeletal
organization, cell migration, and muscle cell differentiation.

View larger version (16K):
[in this window]
[in a new window]
|
Figure 1. Schematic representation of SRF and the myocardin-related family of transcriptional coactivators. The domain structures of the mouse proteins are shown. The numbers to the right indicate the number of amino acids (aa) in each murine protein. Top, The MADS box containing transcription factor SRF. The MADS box regulates homodimerization, cofactor-binding, and DNA-binding activity. Bottom, The myocardin-related family of transcriptional coactivators. The locations of the RPEL, basic (+), glutamine-rich (Q), SAP, LZ, and TAD domains are shown in gray or black rectangles. The RPEL domains regulate actin binding in MRTF-A and MRTF-B. By contrast, in myocardin, these domains have diverged to the point that myocardin does not associate with actin. The basic and glutamine-rich domains are required for SRF-binding activity. The LZ regulates homo- and heterodimerization with other MRTF family members. The mouse myocardin gene encodes 935- and 865-aa protein isoforms generated by alternative splicing. The 935-aa isoform is expressed predominantly in the heart, and the 865-aa isoform is expressed predominantly in SMCs.
|
|
In the cardiovascular system, SRF plays a critical role in regulating vascular smooth muscle cell (SMC) and cardiac myocyte differentiation and in the morphogenetic program regulating development of the heart. Functionally important CArG boxes have been identified in transcriptional regulatory elements controlling expression of sets of myogenic contractile and cytoskeletal proteins (reviewed elsewhere8,25). Of note, in cardiac and skeletal muscle cells, functionally important CArG boxes have been identified in transcriptional regulatory element controlling a relatively limited subset of myofibrillar proteins.26 By contrast, most, but not all, genes encoding SMC-restricted contractile proteins are regulated by SRF. Multiple studies have shown that an SRF-dependent transcriptional regulatory program controls SMC differentiation.16,17 Remarkably, in transgenic mice, when 4 copies of the smooth muscle element-4 nuclear protein binding site, which contains an embedded CArG box, are linked to a minimal promoter transgene expression is restricted to arterial SMCs.17 The molecular mechanism(s) that restrict and distinguish the cell lineage specificity of myogenic CArG elements remains an area of active investigation.
SRF-deficient embryonic stem (ES) cells fail to differentiate and form mesoderm and myogenic genes including cardiac- and SM-
-actin are not expressed in embryoid bodies derived from SRF/ ES cells.2729 SRF/ ES cells display severe defects in cytoskeletal organization and stress fiber formation.27 Targeted mutation of the SRF gene in mice results in embryonic lethality before gastrulation, underscoring the fundamental role SRF plays in cell spreading, adhesion, and migration.28,30 Mice in which the SRF gene was selectively ablated in cardiac myocytes survived only to between embryonic day (E) 10.5 and E13.5. They succumb to cardiac insufficiency manifest by thinning of the compact zone and defective trabeculation of the embryonic heart.29,31 Ablation of the SRF gene reduced cardiomyocyte survival and increased apoptosis was observed in the embryonic heart.29,32 It remains unclear whether the observed apoptosis is a direct result of downregulation of SRF-regulated survival genes, including BCL2 and MCL-1, or an indirect consequence. SRF mutant cardiac myocytes exhibit disorganized sarcomeres, Z-disks, and stress fiber formation.32,33 Conditional ablation of the SRF gene in the adult heart causes dilated cardiomyopathy with progressive heart failure.34 These studies demonstrate that expression of SRF in cardiac myocytes is required for cardiac morphogenesis and strongly suggest that SRF-regulated genes are critical regulators of sarcomeric organization and cardiac contractile function in the adult heart.
Determining the function of SRF in the vasculature and/or in vascular SMCs in vivo has been particularly challenging because promoters that restrict activity of Cre recombinase to SMCs also induce recombination in embryonic cardiac myocytes.35 Not surprisingly, when SM22-Cre transgenic mice were intercrossed to mice containing a floxed SRF allele, informative mutant E10.5 embryos exhibited cardiac defects that were virtually identical to those observed when SRF was selectively ablated in cardiac myocytes.32 However, a marked diminution in the number of SMCs in the embryonic aorta was also observed in these mutant mice. In addition, an absence of actin/intermediate bundles was demonstrated in SRF-deficient SMCs.32 Taken together, these data demonstrate that SRF plays a critical role in the developmental program regulating vascular SMC differentiation and controls expression of most, although not all, genes encoding SMC contractile and cytoskeletal proteins.36
 |
Myocardin: Gene Structure, Expression, and Functional Characterization
|
|---|
The myocardin gene was discovered using a bioinformatics-based
screening strategy designed to identify novel cardiac-specific
genes.
3 Myocardin is a remarkably potent transcriptional coactivator
that binds directly to SRF and activates transcription of a
subset of SRF-regulated genes encoding cytoskeletal and contractile
proteins.
3 The human myocardin gene is encoded on 13 exons,
spanning approximately 92-kb of genomic DNA located at chromosome
17p11.2.
37 The human myocardin gene encodes a protein with a
predicted molecular weight of
Mr 95700.
37 In this regard, it
is noteworthy that the human myocardin mRNA is approximately
9.5-kb in size, including a relatively long 3' untranslated
region, suggesting that posttranscriptional processing may play
a critical role in regulating the expression and biological
function of the myocardin protein.
37 Recently it was discovered
that the mouse myocardin gene encodes 2 alternatively spliced
protein isoforms of 935 and 856 aa, respectively
38 (
Figure 1).
The 2 mouse myocardin protein isoforms are generated via alternative
splicing resulting in the utilization of distinct initiation
codons, leading to truncation of the N terminus of the 856-aa
mouse myocardin protein isoform. Given the tissue-restricted
patterns of expression and distinct functional properties of
these 2 protein isoforms (see below), it will be important to
determine whether 2 myocardin isoforms are expressed in humans.
The myocardin gene is expressed in a precise developmentally regulated, lineage-restricted pattern in the embryo and during postnatal development. Expression of myocardin is restricted to cardiac myocytes and SMCs.3,37 The 935-aa mouse myocardin protein isoform is expressed predominantly in the heart, whereas the 856-aa protein isoform is expressed predominantly in SMC-containing tissues.38 Although both the 935- and 856-aa myocardin isoforms function as potent SRF coactivators, only the 935-aa protein physically associates with myocyte enhancer factor-2 (MEF2) and transactivates MEF2-dependent promoters.38 In this regard, it is noteworthy that MEF2 and SRF are closely related members of the MADS box family of transcription factors. However, SRF and MEF2 bind to distinct DNA sequences and activate distinct sets of genes.38 As such, expression of tissue-restricted myocardin protein isoforms provides a potential mechanism to differentiate the function(s) of myocardin in the heart and SMCs.
As schematically depicted in Figure 1, the N terminus of myocardin contains 2 conserved RPEL domains (Figure 1, gray rectangles), which in MRTF-A and -B facilitate association with monomeric G-actin.9,39 However, the RPEL domains in myocardin have diverged from the consensus RPxxxEL sequence to the point where myocardin does not bind G-actin, and myocardin is localized exclusively in the cell nucleus.39 Therefore the capacity of myocardin to transduce Rho/actin signals may be dependent on its capacity to heterodimerize with other MRTFs through its conserved leucine zipper (LZ) domain (Figure 1, black rectangle).39,40 Myocardin binding to SRF is mediated by a 7-residue sequence called the B1 domain that is located between the conserved basic (Figure 1, black rectangle) and glutamine-rich domains (Figure 1, dark gray rectangle).3,41 Myocardin and other MRTFs contain a conserved 35-aa SAP domain (Figure 1, dark gray rectangle), named for the related factors SAF-A/B, Acinus, and PIAS. SAP domains have been shown to regulate nuclear organization, chromosomal dynamics, and apoptosis.42 However, the function of the myocardin SAP domain in vivo remains unclear. Myocardin deletion mutants lacking the SAP domain physically associate with SRF and synergistically activate transcription of some, but not all, SRF-regulated myogenic genes.3 The C terminus of myocardin contains a powerful transcriptional activation domain (TAD) (Figure 1, light gray box) that functions with heterologous promoters but shares only low-level sequence identity with the TADs in MRTF-A and MRTF-B.4
 |
The Myocardin Family of SRF Coactivators
|
|---|
Two related MRTFs, MRTF-A and MRTF-B, have been identified in
mammalian species
4 (
Figure 1). The family of myocardin-related
transcription factors (MRTFs) evolved from a single ancestral
gene in
Drosophila, designated as
DMRTF, that is most closely
related to MRTF-A
43 (
Figure 1). Like its mammalian counterparts,
DMRTF physically associates with
Drosophila SRF and synergistically
activates transcription.
43 Fly mutants lacking
DMRTF exhibit
severe defects in cytoskeletal organization and cell migration.
43,44 Myocardin, MRTF-A, and MRTF-B share conserved RPEL, basic, glutamine-rich,
SAP, and LZ domains that facilitate their function as SRF coactivators
4 (
Figure 1). Myocardin, MRTF-A, and MRTF-B also share conserved
B1 domains, located between the basic and glutamine-rich domains,
and B2 domains, located between RPEL motifs 2 and 3, that are
involved in the nuclear import of MRTF-A and -B.
9 Mouse MRTF-A
and MRTF-B are approximately 35% identical to myocardin but
share >60% amino acid sequence identity across these conserved
functional domains.
4
The myocardin gene is expressed in a precise developmentally regulated pattern in the heart as well as visceral and vascular SMCs during embryonic development.37 At E9.5, the myocardin gene is expressed abundantly in the primitive heart.3,37 However, despite the fact that vascular SMCs are observed in the dorsal aorta of the mouse at E9.5, at the level of sensitivity afforded by in situ hybridization analyses, myocardin mRNA is not detectable in the dorsal aorta until E11.5, suggesting that myocardin is not absolutely required for the specification and differentiation of vascular SMCs. By contrast, before, or coincident with, the differentiation of visceral SMCs from surrounding mesenchyme, myocardin gene expression is detected in presumptive SMCs.37 Subsequently, during late fetal and postnatal development, the myocardin gene is expressed abundantly in cardiac myocytes and visceral and vascular SMCs.37
The human homolog of the MRTF-A gene had previously been designated as MKL1 and/or MAL, because it was originally identified at a chromosomal translocation breakpoint associated with acute megakaryoblastic leukemia in infants and children.4,4547 The human MRTF-A gene is located at chromosome 22q13.2. MRTF-A is the most ubiquitously expressed mammalian MRTF. MRTF-A gene expression is observed in multiple cell lineages including undifferentiated ES cells and fibroblasts.4,40 MRTF-A is coexpressed with myocardin in the human heart and aorta.40 During embryonic development, MRTF-A is enriched in mesenchymal cells, muscle cells and epithelial cells.5 Like myocardin, MRTF-A is a remarkably potent SRF coactivator.4,40 However, in contrast to myocardin, which is localized predominantly in the nucleus of serum-starved fibroblasts, MRTF-A (and MRTF-B) is localized in the cytoplasm and translocates to the nucleus in response to serum stimulation and other signals that promote actin polymerization (see below).39
MRTF-B is expressed in a unique developmentally regulated, cell lineagerestricted pattern. The human MRTF-B gene is located at chromosome 16p13.12. In the E8.0 mouse embryo, MRTF-B gene expression is observed in the primitive heart and in rhombomeres 3 and 5, from which the cardiac neural crest is derived.48 By E9.5, MRTF-B gene expression is observed in the embryonic dorsal aorta before expression of myocardin.48 At E11.5, coincident with heart and cardiac outflow tract patterning, the MRTF-B gene is expressed in the cardiac neural crest cells that give rise to SMCs populating the cardiac outflow tract and aortic arch arteries.48 During late fetal development, MRTF-B is expressed in the epithelial cells of the lung, kidney, and testes.4 In the adult mouse, the highest levels of MRTF-B mRNA are observed in the heart and brain, although MRTF-B mRNA is also detectable in the lung, liver, kidney, and testes.4 In vitro structure/function analyses suggest that MRTF-B is not as potent a transcriptional coactivator as myocardin or MRTF-A.4 However, these in vitro studies are potentially misleading because MRTF-B contains a powerful TAD, suggesting that molecular mechanisms may have evolved to regulate the transcriptional activity of MRTF-B in vivo.4
 |
MRTFs: Transduction of Rho/Actin Cytoskeletal Signaling
|
|---|
The cytoskeleton is a dynamic structure regulating cell shape
and transducing signals to the nucleus influencing cellular
morphology, organization, growth and migration. SRF plays a
critical role in transducing cytoskeletal signals to the nucleus
and in regulating expression of genes encoding cytoskeletal
proteins.
21,49,50 As shown in
Figure 2, Rho GTPases regulate
cytoskeletal actin polymerization influencing SRF-dependent
transcription.
49,51 Activation of Rho induces actin polymerization
via the ROCK/LIM kinase (LIMK)/cofilin pathway, stabilizing
F-actin and mDia1 and promoting the assembly of monomeric G-actin
into F-actin filaments.
49,52 MRTFs play a critical role in transducing
Rho/actin signaling from the cytoplasm to the nucleus and activating
SRF-dependent transcription. RhoA signaling promotes actin polymerization
(decreasing the concentration of free G-actin), resulting in
the translocation of MRTF-A from the cytoplasm to the nucleus.
39,40,53 The B1domain, located between the basic and glutamine-rich domains,
and the B2 domain, located between RPEL motifs 2 and 3, play
essential roles in the nuclear import of MRTF-A.
39 Nuclear accumulation
of MRTF-A can be inhibited by forced expression of nonpolymerizing
actin mutants.
39 As such, G-actin itself may act as a shuttle
for MRTFs between the nucleus and the cytoplasm and the concentration
of free G-actin or a subpopulation of G-actin may directly influence
the association and/or disassociation of MRTFs and SRF.
39 Alternatively,
nuclear import of MRTF-A (and MRTF-B) may be regulated by a
critical cofactor bound to the MRTF/actin complex. Rho/actin-induced
nuclear import of MRTFs is also regulated by STARS (
striated
muscle
activator of
Rho
signaling) protein, which binds to F-actin,
facilitating translocation of MRTF-A and -B to the nucleus.
54

View larger version (25K):
[in this window]
[in a new window]
|
Figure 2. Rho signaling and actin dynamics regulate MRTF-mediated transcription of SRF-dependent genes encoding contractile and cytoskeletal proteins. Activation of Rho signaling promotes the assembly of F-actin from monomeric G-actin via Rho effectors ROCK-LIMK and mDia. The concentration of free G-actin itself or the concentration of a G-actin subpopulation promotes the nuclear accumulation of MRTF-A (and MRTF-B). MRTFs associate with SRF and activate transcription of genes encoding contractile and cytoskeletal proteins. This figure was adapted from.39
|
|
In the nucleus, MRTFs physically associate with SRF, facilitating the binding of SRF to single or dual CArG boxes, activating transcription of genes encoding cytoskeletal and myogenic proteins (Figure 2).39,40,53,55,56 MRTF-A and -B heterodimerize in vitro with myocardin via conserved LZ motifs.40 However, it remains unclear whether myocardin, MRTF-A, and/or MRTF-B heterodimerize in vivo when they are coexpressed. In this regard, it is noteworthy that in SMC myocardin is constitutively nuclear. As such, it is possible that Rho/actin signals are transduced to myocardin via MRTF-A (and/or MRTF-B) via the capacity of these cofactors to heterodimerize with myocardin and form multiprotein complexes with SRF. Hydrophobic residues within the MRTF-A B1 region are essential for the association of MRTF-A with the hydrophobic groove and pocket region of the SRF DNA-binding domain.41 This is the same region of SRF that binds ternary complex factors (TCFs), and competition for binding to this common region of SRF between TCFs and MRTFs has been observed.41,57
 |
Myocardin: Cardiomyocyte Differentiation and Hypertrophy
|
|---|
Myocardin is expressed early in the morphogenetic program regulating
cardiac development, and myocardin appears to play an important
role in activating a set of SRF-regulated cardiac-specific proteins.
3,37 Expression of a dominant-negative myocardin mutant protein or
exposure of
Xenopus embryos to morpholino oligonucleotides that
target myocardin mRNA blocks expression of genes encoding some
cardiac-specific markers and disrupts heart tube formation.
58 In addition, forced expression of myocardin in
Xenopus animal
cap explants induces transcription of endogenous cardiac (and
SMC) genes. However, expression of myocardin in animal caps
fails to activate the full repertoire of cardiac-restricted
genes, including the upstream transcription factors GATA4 and
Nkx2.5.
58 Similarly, forced expression of myocardin in ES cells
and fibroblasts activates CArG boxdependent cardiac and
SMC-restricted genes but does not specify the cardiac myocyte
cell fate.
59,60 Conversely, embryoid bodies derived from
MYOCD/ ES cells express genes encoding multiple cardiac-specific markers.
60 Furthermore, at least through E10.5, the hearts of myocardin-null
embryos appear morphologically normal.
61 Taken together, these
data demonstrate that myocardin functions as an important SRF
cofactor activating a relatively limited subset of cardiac-restricted
genes but that myocardin is not absolutely required for cardiac
myocyte specification and/or differentiation.
Emerging evidence suggests that myocardin may play an important role in regulating the hypertrophic response of the heart. This is not surprising, as transcriptional coactivators have the capacity to modulate gene transcription in response to inducible stimuli. Hypertrophic stimuli induce expression and transcriptional activity of myocardin in the heart.62 In addition, adenoviral-mediated overexpression of myocardin stimulates cardiac myocyte hypertrophy in cultured neonatal rate cardiac myocytes.63 Cardiac myocyte hypertrophy is controlled primarily at the level of transcription and is inhibited by glycogen synthase kinase 3ß (GSK3ß).64,65 GSK3ß phosphorylates multiple serine residues in myocardin, inhibiting myocardin-induced atrial natriuretic factor transcription.63 Although myocardin is one of many proteins phosphorylated by GSK3ß in the heart (others include GATA-4, ß-catenin, c-myc, and CREB), these data suggest that myocardin may be involved in regulating the response of the heart to hemodynamic stress. Intriguingly, in a relatively small sample of explanted failing human hearts, myocardin mRNA was upregulated relative to the level of myocardin mRNA observed in nonfailing donor hearts.66 Ongoing experiments characterizing mice containing cardiac-selective ablation of the myocardin gene may help to elucidate the role of myocardin during cardiogenesis and in the postnatal heart. In addition, studies assessing myocardin mRNA and protein expression and activity in the hypertrophied and failing human heart in well-matched patient and control populations may provide important clinical correlations to these basic and translational observations.
 |
Myocardin: A Critical Coactivator Stimulating SMC Differentiation
|
|---|
Multiple studies have shown that a SRF-dependent transcriptional
regulatory program regulates SMC differentiation (reviewed elsewhere
8,25).
However, until the discovery of myocardin, it remained unclear
how SRF, which is expressed in multiple cell lineages, could
selectively activate the set of genes encoding SMC differentiation
markers. The discovery of myocardin led our group and others
to propose a molecular model wherein binding of the lineage-restricted
coactivator myocardin to SRF synergistically activates a subset
of SRF-regulated genes, stimulating SMC differentiation.
37,67,68 Consistent with this model, forced expression of myocardin transactivates
multiple SMC-restricted transcriptional regulatory elements
including the SM22

promoter, SM myosin heavy chain (SM-MyHC)
promoter/enhancer, SM-

-actin promoter/enhancer, SM myosin light
chain kinase promoter, and the smoothelin-A promoter.
3,37,67,69,70 In addition, forced expression of a dominant-negative myocardin
mutant protein or small interfering RNAinduced myocardin
knockdown significantly reduced SM22

promoter activity in SMCs.
37 Remarkably, forced expression of myocardin in embryonic stem
(ES) cells transactivates the endogenous SM22

promoter and induces
expression of multiple endogenous genes including SM22

, SM-MyHC,
and SM-

-actin.
37,67 Most importantly, myocardin-null embryos
survive only to E10.5 and show no evidence of vascular SMC differentiation.
61
The finding that forced expression of myocardin induces expression of multiple SMC-restricted genes in non-SMCs raised the question of whether myocardin functions as a master regulator of the SMC lineage, acting in a manner analogous to that of MyoD in skeletal muscle specification.59 However, this does not appear to be the case. Forced expression of myocardin in A404 SMC precursor cells does not activate the full repertoire of genes encoding SMC-restricted proteins.71 Smoothelin-B, aortic carboxypeptidase-like protein (ACLP), and focal adhesion kinaserelated nonkinase, whose promoters lack functionally important CArG boxes, are not expressed in A404 cells expressing myocardin.71 Conversely, myocardin-null ES cells contribute to the vascular SMC lineage in MYOCD//C57BL6 chimeric mice, demonstrating that myocardin is not required in a cell autonomous manner for SMC specification or differentiation.60 Taken together, these data demonstrate that myocardin functions as a critical transcriptional coactivator in the SRF-dependent transcriptional regulatory program regulating SMC differentiation. It remains possible that MRTF-A and/or MRTF-B subserve partially redundant functions with myocardin in promoting specification and/or differentiation of the SMC lineage. Consistent with this finding, forced expression of either MRTF-A or MRTF-B in undifferentiated ES cells also activates endogenous CArG box-dependent SMC genes including SM-
-actin, SM-myosin heavy chain, and SM22
.40
Multiple signaling pathways modulate the capacity of myocardin to stimulate SMC differentiation. RhoA-induced actin polymerization induces SRF-dependent transcription of the SM22
and SM-
-actin promoters.72 Conversely, inhibition of stress fiber formation or repression of actin polymerization suppresses transcription of genes encoding of SMC-restricted contractile and cytoskeletal proteins.72 Consistent with these findings, forced expression of MRTF-A transactivates multiple SMC-restricted transcriptional regulatory elements, whereas a dominant-negative MRTF-A mutant protein represses SMC-restricted transcriptional activity.40 In most cell lineages, MRTF-A is observed predominantly in the cytoplasm and translocates to the nucleus in response to serum or Rho/actin signaling. By contrast, in SMCs, MRTF-A is observed exclusively in the nucleus.40 Agents that disrupt actin polymerization and/or RhoA signaling cause MRTF-A to translocate to the cytoplasm, demonstrating that an active actin/Rho MRTF-mediated signaling process reinforces the contractile, differentiated SMC phenotype.40 The molecular basis of the basal Rho/actin signal has not been characterized but may be related to the extensive rib-like array of cytoskeletal elements observed in SMCs.
Insulin-like growth factor (IGF)-1 stimulates SMC differentiation, promoting the contractile SMC phenotype. In SMCs, the IGF receptor signals through phosphoinositide-3-kinase (PI3K) to Akt.73 Akt stimulates phosphorylation of multiple proteins including Foxo4, a forkhead factor that represses transcription. Foxo4 physically associates with myocardin, repressing transcription of genes encoding SMC contractile genes. Phosphorylation of Foxo4 stimulates its nuclear export, thereby potentiating myocardin-induced transcription of SMC cytoskeletal and contractile genes and promoting the contractile SMC phenotype observed in quiescent SMCs.74 Consistent with this, small interfering RNAmediated knockdown of Foxo4 mRNA in A7r5 SMCs enhances myocardin-induced SMC differentiation, upregulating expression of SM22
, SM-MyHC, and SM-calponin gene expression.74 Taken together, these data demonstrate that the SMC contractile phenotype is dependent on a basal signal emanating from the cytoskeleton, which is transduced to the nucleus via MRTFs activating expression of SRF-dependent genes encoding contractile and cytoskeletal proteins. This signal is reinforced by IGF-1/Akt signaling, which phosphorylates Foxo4 stimulating its nuclear export.
SMC differentiation is accompanied by alterations in chromatin structure, affecting transcriptional regulatory elements controlling genes encoding SMC contractile proteins.7579 The binding of SRF to SMC CArG boxes is associated with specific alterations in chromatin structure including the methylation and acetylation of histones.76,79 Myocardin-induced transcription is modulated by the acetylation of nucleosomal histones flanking SRF-binding sites in SMC-restricted genes.78 The C-terminal TAD of myocardin physically associates with histone acetyltransferases (HATs), including p300, enhancing its capacity to activate transcription of SMC-restricted genes.78 By contrast, class II histone deacetylases (HDACs) suppress myocardin-induced transcription of SMC-restricted genes.78 The regulation of SMC genes is also modulated by the recently discovered transcriptional repressor PRISM/PRDM.80 PRISM interacts with class I HDACs and the G9a methyltransferase repressing a subset of genes encoding SMC contractile proteins. In addition, PRISM stimulates genes associated with the SMC synthetic phenotype.80 Taken together, these data suggest MRTF-induced transcription of SMC genes is regulated via transcriptional as well as epigenetic mechanisms influenced by the capacity of myocardin to differentially associate with HATs versus HDACs.
 |
Myocardin and MRTFs: Modulation of SMC Phenotype
|
|---|
SMCs modulate their phenotype in response to vascular injury
from a quiescent cell expressing high levels of genes encoding
contractile and cytoskeletal proteins to a proliferative cell
expressing high levels of genes encoding cytokines, growth factors,
and extracellular matrix (reviewed elsewhere
8183). As
shown in
Figure 3, SRF integrates multiple signals influencing
SMC phenotype via its capacity to differentially associate with
MRTFs and TCFs in the
Ets family of transcription factors.
57,84 Pharmacological agents that inhibit actin polymerization and/or
forced expression of nonpolymerizing actin mutant proteins stimulates
the nuclear export of MRTF-A.
40 This, in turn, leads to the
dissociation of MRTF-SRF complexes and downregulation of genes
encoding SMC contractile and cytoskeletal proteins.
40 MRTF-A
and TCFs compete for binding to a common surface of SRF.
41 The
TCF Elk-1 displaces myocardin bound to SRF on the SMC-restricted
SM22

and SM-

-actin promoters (
Figure 3).
57 Consistent with this
in vitro finding, forced expression of Elk-1 inhibits transcription
of SMC contractile genes in vivo.
57,85 Moreover, forced expression
of TCFs Elk-1, SAP-1, or SAP-2 blocks myocardin-induced SMC
differentiation.
57,85

View larger version (34K):
[in this window]
[in a new window]
|
Figure 3. Modulation of vascular SMC phenotype in response to growth factor stimulation and/or vascular injury. In response to growth factor stimulation and/or vascular injury, F-actin polymerization is inhibited and/or F-actin disassembles, resulting in an increased concentration of free G-actin. This signal is transduced to the nucleus resulting in the nuclear export of MRTF-A (and MRTF-B) and disrupting SRF/MRTF complexes, resulting in the downregulation of genes encoding SMC contractile proteins. Growth factor stimulation also induces a mitogen-activated protein kinase signaling cascade, activating extracellular signal-regulated kinase 1/2 (ERK1/2) which phosphorylates TCFs. Phosphorylated TCFs displace MRTFs from SRF downregulating transcription of SMC contractile and cytoskeletal genes. Concomitant with this, SRF-TCF complexes bind to CArG boxes and adjacent ETS binding sites activating transcription of a subset of SRF-regulated growth responsive genes. Activation of Notch receptors and effectors in the Hairy-related transcription (HRT), including HRT2, repress MRTF-mediated transcription of genes encoding SMC contractile proteins. PDGF induces transcription of the transcriptional repressor KLF-4, which binds to GC-rich elements located in transcriptional regulatory elements controlling expression of SMC contractile genes. Foxo4 acts as a transcriptional repressor opposing myocardin-induced SMC differentiation. Foxo4 physically associates with myocardin and forms ternary complexes with myocardin and SRF bound to DNA.
|
|
Concomitant with the downregulation of genes encoding SMC contractile proteins, TCFs induce the expression of SRF-regulated growth-related genes including c-fos and egr-186 (Figure 3). Platelet-derived growth factor (PDGF) stimulates a mitogen-activated protein kinase signaling cascade, leading to phosphorylation of the TCF Elk-1 increasing its affinity for SRF. TCFs associate with SRF through a conserved B-box domain.87 The association of TCF with SRF is stabilized by formation of ternary complexes requiring specific contacts between TCF and ETS binding sites located adjacent to CArG boxes (Figure 3). The formation of stable ternary complexes induces transcription of a subset of growth-responsive genes observed in synthetic, proliferating SMCs. Although this binary system serves as a useful working model, it fails to account for the myriad of possible SMC phenotypes observed in during embryonic angiogenesis and in vascular proliferative syndromes.
During embryonic angiogenesis and following vascular injury, SMC phenotype is "fine tuned" by additional signaling pathways that modulate MRTF-induced SMC differentiation (Figure 3). Notch receptors and effectors in the Hairy-related transcription (HRT) family play essential roles in vascular patterning and the arterial response to injury (reviewed elsewhere88). Activation of Notch receptors by their endogenous ligand, Jagged 1, induces translocation of the Notch intracellular domain (ICD) to the nucleus where it inhibits myocardin-induced expression of SMC genes.89 Consistent with this finding, forced expression of the Notch effector HRT2 in SMCs represses transcription of multiple SMC-restricted genes.89,90 In addition to phosphorylating TCFs (see above), PDGF stimulates expression of Kruppel-like transcription factor (KLF)-4 in SMCs.91 KLF-4, which is normally not expressed in quiescent SMCs, is rapidly upregulated following vascular injury.92 KLF-4 functions as a transcriptional repressor antagonizing myocardin-induced activation of SMC genes as well as expression of myocardin.92 KLF-4 binds to a G/C-rich cis-acting sequence located in transcriptional elements regulating expression of multiple genes encoding SMC proteins.93 Both PDGF-BB and KLF-4 inhibit SRF binding to CArG boxes downregulating transcription of SMC contractile genes.92 These data are consistent with a model wherein SMC phenotype is regulated primarily by the mutually exclusive binding of MRTFs and TCFs to SRF. However, SMC phenotype is fine tuned by multiple additional signaling pathways that directly and indirectly modulate the activity of MRTFs and TCFs, ultimately converging on SRF bound to CArG boxes regulating expression of contractile and/or growth-responsive genes. Ultimately, it is the sum of these combinatorial signals that modulate SMC phenotype during embryonic angiogenesis and in response to vessel wall injury.
 |
MRTFs: Loss-of-Function Mutant Mice
|
|---|
Genetically engineered mice have been generated containing loss-of-function
mutations in myocardin,
MRTF-A, and
MRTF-B, respectively.
48,61,9496 Despite the conserved domain structure and overlapping patterns
of expression of MRTFs in the cardiovascular system,
4 mice harboring
single genenull mutations in myocardin,
MRTF-A, and
MRTF-B display distinct phenotypes. Myocardin-null mice survive only
to embryonic day 10.5 (E10.5) and exhibit obvious vascular defects
including an underdeveloped aorta lacking detectable SMCs.
61 This defect is probably attributable to a block in differentiation
of SMCs, as normal endothelial cell differentiation and patterning
of the vasculature is observed in
MYOCD/ embryos
at least through E9.5.
61 However, the molecular basis of the
observed block in vascular SMC differentiation is not obvious
because, at least at the level of sensitivity afforded by in
situ hybridization, myocardin mRNA is not detectable in aortic
SMCs until E11.5.
37 Although it is possible that low levels
of myocardin are expressed in vascular SMCs before E9.5, which
may explain the observed phenotype, it is noteworthy that
MYOCD/ embryos also exhibit a defect in the extraembryonic yolk sac
vasculature that may have contributed to fetal demise.
5,61 Consistent
with this possibility, E10.5 myocardin-null embryos show overall
developmental delay and they are 50% to 60% the size of their
control littermates.
61
The early lethality observed in myocardin-null embryos also precludes assessment of the function of myocardin in the heart during late embryonic and postnatal development. Despite that fact that myocardin is expressed abundantly in the embryonic heart as early as E8.037 in myocardin-null embryos, cardiac looping and chamber formation appear normal at least through E9.5.61 It is possible that expression of MRTF-A and/or MRTF-B compensates for the loss of myocardin in the embryonic heart (and vasculature), as all 3 MRTFs are expressed in the primitive heart tube (and in aortic SMCs).4 Consistent with this hypothesis, complete inhibition of RhoA-inducible SRF-mediated gene expression requires blockade of both MRTF-A and MRTF-B, demonstrating that in some cellular contexts redundancy between these closely related cofactors does exist.97 The generation of mice containing compound-conditional loss-of-function mutations in myocardin, MRTF-A, and/or MRTF-B will address questions of redundancy between MRTFs and may help to elucidate the function of myocardin in the heart and vasculature during embryonic and postnatal development.
MRTF-A is the most ubiquitously expressed of the MRTFs and plays a critical role in transducing Rho/actin signals to the nucleus.39 It is therefore surprising that mice harboring null mutations in MRTF-A are viable, fertile, and born with an equal male to female ratio.94,95 Li et al reported that MRTF-A/ mice are born in the anticipated Mendelian ratio,95 whereas Sun et al reported that MRTF-A/ null mice are born at less than the anticipated Mendelian frequency. The fetal loss was attributed to dilated cardiomyopathy accompanied by heart failure, which was observed in 35% of MRTF-A/ embryos.94 Gross and histological examination of MRTF-A/ liveborn pups failed to reveal obvious abnormalities. However, pups born to MRTF-A/ mutant dams fail to thrive and die between postnatal day 14 (P14) and P20.94,95 MRTF-A/ dams exhibit a defect in maternal lactation correlated with milk accumulation and the premature onset of mammary gland involution. During early lactation phase in MRTF-A/ dams, a defect in mammary gland myoepithelial cell differentiation was observed that was manifested by severely attenuated, or absent, expression of genes encoding SMC-restricted contractile proteins including SM-
-actin, SM-myosin heavy chain, calponin 1, and tropomyosin 2.94,95 Later in the lactation cycle, massive apoptosis of myoepithelial cells surrounding alveolar lumens was observed in MRTF-A/ mutant dams.95 Taken together, these data demonstrate that MRTF-A is required for differentiation and survival of myoepithelial cells during lactation cycles. As the decrease in expression of SMC differentiation markers in mammary myoepithelial cells preceded the observed apoptosis, these data suggest that the processes of myoepithelial cell differentiation and myoepithelial cell survival may be linked.
Generation of MRTF-B loss-of-function mutant mice revealed an unanticipated cell autonomous defect in the differentiation of SMCs from the cardiac neural crest.48,96 MRTF-Bdeficient embryos displayed a spectrum of cardiac outflow tract defects recapitulating forms of congenital heart disease commonly observed in humans (Figure 4B through 4D).48,96 Li et al used a MRTF-B gene trap strategy to generate MRTF-B mutant mice with a hypomorphic phenotype.48 Homozygous MRTF-B gene trap mutant mice died between E17.5 and P1, exhibiting a spectrum of cardiac outflow tract defects including persistent truncus arteriosus, double-outlet right ventricle, right-sided aortic arch, and interrupted aortic arch48 (Figure 4B through 4D). Oh et al used a MRTF-B gene targeting strategy to generate mice with a null phenotype.96 MRTF-Bnull mice died at mid-gestation (E13.5-E14.5) and displayed a nearly identical spectrum of cardiac outflow tract defects.96 It remains unclear why the MRTF-B gene trap mutant mice survived to between E17.5 and birth, whereas the MRTF-Bnull embryos survived only to E14.5. This probably resulted from the 5% residual wild-type MRTF-B gene expression demonstrated in the gene trap mutant mice, suggesting that MRTF-B gene dosage plays a critical role in regulating cardiac outflow tract development and embryonic survival.48

View larger version (113K):
[in this window]
[in a new window]
|
Figure 4. MRTF-B is expressed in the cardiac neural crest, and MRTF-B loss-of-function mutant mice display a spectrum of cardiac outflow tract defects. A, Heterozygous E8.5 MRTF-B gene trap mutant embryo demonstrating expression of the MRTF-B/LacZ fusion protein (arrows) in rhombomeres 3 and 5 of the dorsal neural folds. Note also expression of the MRTF-B/LacZ fusion protein in the heart (H). B, Wild-type E18.5 embryo demonstrating a left-sided aortic arch (Ao) and intact ductus arteriosus (DA). C, Homozygous MRTF-B gene trap mutant embryo with an interrupted aortic arch (IAA). All homozygous MRTF-B mutant embryos exhibited either a persistent truncus arteriosus (PTA) or double-outlet right ventricle (DORV) and an obligatory ventricular septal defect (VSD). D, Homozygous MRTF-B gene trap mutant embryo with a right-sided aortic arch (RSAA). This figure was adapted and reproduced with permission from.48
|
|
The spectrum of cardiac outflow tract defects observed in MRTF-B loss-of-function mutants closely resembled the cardiac outflow tract defects observed in mice subjected to selective neural crest ablation.98,99 Consistent with this observation, in E8.0 MRTF-B gene trap mice, expression of the MRTF-B/LacZ fusion protein was observed in rhombomeres 3 and 5 of the dorsal neural folds, which colocalizes with the cardiac neural crest48 (Figure 4A, arrows). Subsequently, between E9.5 and E11.5, MRTF-Bexpressing cells migrate ventrally to branchial arch arteries 3, 4, and 6 and give rise to SMCs populating the cardiac outflow tract and great arteries.48 Analyses of E10.5 to 11.5 MRTF-B mutant embryos revealed that plexinA2-positive cardiac neural crest cells migrated to, and populated, the cardiac outflow tract and aortic arch arteries, effectively ruling out a significant migratory defect.48 India ink injections into the heart and dorsal aorta of E10.5 to 11.5 MRTF-B mutant embryos revealed only 1 or 2 right- and left-sided aortic arch arteries compared with 3 right-sided and left-sided arch arteries normally observed.48,96 However, expression of SMC markers, including SM-
-actin and SM22
, was absent, or severely attenuated, in the presumptive SMCs surrounding the aortic arch arteries and the aorticopulmonary septum.48,96 Taken together, these data reveal a cell autonomous block in the differentiation of cardiac neural crest cells into vascular SMCs in MRTF-B loss-of-function mutant mice. These data demonstrate that cardiac neural crest differentiation is required for cardiac outflow tract patterning. Moreover, these data suggest that MRTF-B, and genes regulated by MRTF-B, serve as candidate loci for mutations causing common forms of congenital heart disease observed in humans.
The distinct phenotypes of myocardin, MRTF-A, and MRTF-B loss-of-function mutant mice demonstrates that despite the high-level sequence identity and overlapping patterns of expression of these 3 related transcriptional coactivators, molecular mechanisms have evolved to distinguish their functions in vivo. Differences in the temporal and spatial patterns of MRTF expression in the embryo probably accounts for some of these differences. For example, MRTF-A is expressed in myoepithelial cells, whereas myocardin is not.95 However, expression of MRTF-B is observed in MRTF-A/ myoepithelial cells, demonstrating that these 2 related factors are not functionally redundant in this cellular context. Similarly, MRTF-A and subsequently myocardin (at E12.5) are coexpressed with MRTF-B in embryonic vascular SMCs populating the proximal aorta and great arteries, strongly suggesting that MRTFs mediate nonredundant functions in neural crestderived SMCs. As such, structural differences between myocardin, MRTF-A and MRTF-B most likely exist and distinguish their function(s) in tissues when they are coexpressed. The recent identification of 2 tissue-restricted murine myocardin isoforms adds an additional level of complexity to the MRTF family that may also explain functional differences between family members.38 Further experiments examining the structure/function relationships among myocardin, MRTF-A, and MRTF-B should provide important insights into the molecular mechanisms regulating cardiovascular patterning and myogenic differentiation.
 |
Myocardin and MRTFs: Summary and Outstanding Questions
|
|---|
Cell lineagespecific transcription factors that bind
to specific
cis-acting elements embedded within the genome play
critical roles in regulating the processes of cell lineage specification
and cellular differentiation. Forced expression of the bHLH
(basic helixloophelix) skeletal musclespecific
transcription factor, MyoD, in nonmuscle cells activates the
molecular program specifying the skeletal muscle cell fate (reviewed
elsewhere
100). In the heart, cardiac myocyte specification and
differentiation is controlled by the precisely orchestrated
expression of multiple cardiac-restricted transcription factors
including GATA-4, Nkx2.5, and MEF2C in response to specific
growth factors and developmental cues (reviewed elsewhere
2).
In contrast to the striated muscle cell lineages, SMCs maintain
the capacity to proliferate and modulate their phenotype during
postnatal development. Therefore, it is not surprising that
dominant-acting SMC lineage-specific transcription factors that
"lock-in" a terminally differentiated SMC phenotype have not
been identified. Instead, the reversible association of the
MADS box transcription factor SRF with MRTFs promoting SMC differentiation
and TCFs promoting SMC growth provides the necessary plasticity
required to regulate vascular tone and respond to hemodynamic
stress and/or vascular injury.
Transcriptional coactivators and corepressors have evolved to facilitate the rapid induction or repression of sets of genes in response to specific external stimuli. MRTFs transduce extracellular signals through the cytoskeleton that promote SMC differentiation and modulate SMC phenotype. It is noteworthy that in vascular SMCs, MRTF-A, and MRTF-B are localized exclusively in the nucleus, although in most other cell lineages, MRTF-A and -B are localized in the cytoplasm and translocate to the nucleus only in response to Rho/actin signaling.40 Agents that disrupt actin polymerization and block RhoA-signaling drive MRTF-A from the SMC nucleus, demonstrating that a "default" or basal Rho/actin signal promoting the contractile SMC phenotype exists in SMCs. In the vasculature, SMCs are programmed to regulate vascular tone/contractile function through the association of SRF with MRTFs. At the same time, SMCs are poised to respond to vascular stress or injury by responding to extracellular signals transduced through the cytoskeleton inhibiting SRF/MRTF-induced transcription of genes encoding SMC contractile proteins and promoting SRF/TCFmediated activation of growth responsive genes.
In less than 5 years, the seminal discovery of myocardin has led to fundamentally important insights into the molecular programs regulating SMC differentiation and cardiovascular development. However, the precise function(s) of myocardin, MRTF-A, and MRTF-B, respectively, in the heart and vasculature remains to be determined. Experiments performed to date, including the generation of mice harboring single-gene loss-of-function mutations in myocardin, MRTF-A, and MRTF-B, respectively, have raised as many questions as they have answered. Do MRTFs subserve partially, or completely, redundant functions in tissues when they are coexpressed? What mechanisms distinguish the function(s) of these 3 closely related transcriptional coactivators in tissues when they are coexpressed? What mechanisms regulate the association and disassociation of MRTFs and monomeric actin and/or specific MRTFs and SRF? What are the functions of MRTFs in the heart and vasculature during postnatal development? Do mutations in the genes encoding myocardin, MRTF-A, MRTF-B, and/or their transcriptional targets cause common forms of congenital heart disease observed in humans, and what is the role of these cofactors in vascular proliferative syndromes including atherosclerosis? The molecular reagents and animal models required to examine these important outstanding questions have already been, or currently are, being generated and these important questions will be addressed in the very near future. These studies promise to provide important new insights into the molecular programs regulating cardiovascular growth, differentiation, and adaptation of the cardiovascular system to hemodynamic stress.
 |
Acknowledgments
|
|---|
The author thanks Lisa Gottschalk for expert preparation of
graphic illustrations; and Jon Epstein, Ed Morrisey, Jian Li,
and John Huang for helpful comments and suggestions.
Sources of Funding
This work was supported by NIH grant PO1-HL075215 and the Commonwealth of Pennsylvania.
Disclosures
None.
 |
Footnotes
|
|---|
Original received November 29, 2006; revision received January
4, 2007; accepted January 24, 2007.
 |
References
|
|---|