Myocardin-related transcription factor B is required in cardiac neural crest for smooth muscle differentiation and cardiovascular development

Abstract

Members of the myocardin-related family of transcription factors play critical roles in regulating vascular smooth muscle and cardiac differentiation. To examine the function of myocardin-related transcription factor (MRTF)-B, mice were generated from ES cells harboring a conditional insertional mutation, or gene trap, of the MRTF-B gene. Expression of the MRTF-B mutant allele results in a fusion protein consisting of the N terminus of MRTF-B fused to β-galactosidase, which is functionally null. Homozygous MRTF-B gene trap mice (MRTF-B^–/–) die between embryonic day (E)17.5 and postnatal day 1 from cardiac outflow tract defects. MRTF-B is expressed in the premigratory neural crest, in rhombomeres 3 and 5, and in the neural crest-derived mesenchyme surrounding the aortic arch arteries. Consistent with the pattern of expression, E10.5 and E11.5 MRTF-B^–/– mutants exhibit deformation of aortic arch arteries 3, 4, and 6 and severe attenuation of smooth muscle cell differentiation in the arch arteries and the aorticopulmonary septum, despite normal migration and initial patterning of cardiac neural crest cells. Remarkably, the observed pathology was rescued and viable mice generated by intercrossing MRTF-B mutants with mice expressing Cre recombinase under the transcriptional control of the neural crest-restricted Wnt-1 promoter, which results in restoration of normal MRTF-B expression in the neural crest. Taken together, these studies reveal that MRTF-B plays a critical role in regulating differentiation of cardiac neural crest cells into smooth muscle and demonstrate that neural crest-derived smooth muscle differentiation is specifically required for normal cardiovascular morphogenesis.

The critical role that the cardiac neural crest plays in the morphogenetic program regulating cardiac outflow tract development and formation of the great arteries has been revealed through classical anatomic neural crest ablation experiments in combination with quail–chick cell transplantation studies (1, 2). Recently, the application of genetic Cre-Lox fate-mapping techniques and the generation of mouse models recapitulating human congenital heart disease have begun to define the molecular basis of cardiac outflow tract development (3, 4). Neural crest cells arising in the dorsal neural tube between the midotic placode and the third somite contribute to portions of the heart and great vessels leading to their designation as the cardiac neural crest (1). In response to poorly understood developmental cues, these cells migrate ventrally, populating the aortic arch arteries and cardiac outflow tract. The neural crest derivatives then differentiate into smooth muscle cells (SMCs) forming the tunica media of the great arteries and the ascending aorta (1). Concomitant with this migration, two advancing columns of neural crest cells invade the cardiac outflow tract where they fuse, dividing the single great vessel, the truncus arteriosus, forming the aorticopulmonary septum and establishing the aorta and pulmonary arteries (1). Alterations in this complex developmental program lead to a spectrum of pathologies involving the cardiac outflow tract and great arteries that are associated with common forms of congenital heart disease observed in humans (5).

A critical step in the formation of the cardiac outflow tract and of the great arteries is the differentiation of neural crest cells to specified SMCs. Relatively little is currently understood about the molecular mechanisms regulating differentiation of SMCs from neural crest progenitors. By contrast, SMCs arising from lateral mesoderm and local mesenchyme are regulated by a serum response factor (SRF)-dependent transcriptional regulatory program (6). SRF activates SMC transcription regulatory elements by physically associating with the SMC- and cardiac-restricted transcriptional coactivator myocardin (7–9). Mice harboring a myocardin null mutation die at embryonic day (E)10.5 and show no evidence of vascular SMC differentiation in the dorsal aorta (10). It remains unclear whether identical, overlapping, or distinct transcriptional programs regulate differentiation of SMCs arising from the lateral mesoderm and cardiac neural crest. However, the finding that common forms of congenital heart disease involving neural crest-derived structures are generally not associated with other vascular malformations suggests that the developmental programs regulating these processes are distinct.

Two myocardin-related transcription factors (MRTFs), designated MRTF-A and -B, have been described (11). Like myocardin, these factors physically associate with SRF and activate SMC-restricted transcription. MRTF-A is expressed by multiple cell lineages and is a remarkably potent transcriptional coactivator of some SRF-dependent genes (11–15). In response to RhoA-mediated and cytoskeletal signals, MRTF-A translocates from the cytoplasm to the nucleus and activates gene expression (12–14). By contrast, relatively little is currently understood about the function of MRTF-B, which is expressed in a unique developmentally regulated lineage-restricted fashion, including expression in some SMCs and the heart (11). Compared with myocardin and MRTF-A, the full length MRTF-B protein is a relatively weak transcriptional coactivator, although it contains a powerful transcriptional activation domain (11). Of note, complete inhibition of RhoA-inducible SRF-mediated gene expression requires blockade of both MRTF-A and -B, suggesting some redundancy of function between MRTF-A and -B (15).

Materials and Methods

Generation and Characterization of MRTF-B Gene Trap Mice. ES cells containing an insertional mutation or gene trap in the MRTF-B gene (clone RRJ478) were provided by BayGenomics, which can be accessed at http://baygenomics.ucsf.edu. The gene-trap vector (pGT0 lxf) contains a splice-acceptor sequence flanked by loxP sites (floxed) subcloned 5′ of the βgeo reporter cassette encoding a fusion protein consisting of the bacterial lacZ gene fused to neomycin phosphotransferase II (16). DNA sequence analysis revealed that the floxed splice acceptor and β-geo cassette integrated 191 bp 3′ of MRTF-B exon 10 generating truncated MRTF-B lacking a transcriptional activation domain (Fig. 1A). Clone RRJ478 ES cells were injected into blastocysts from C57BL/6J mice, and mutant mice were generated and characterized as described (17). To rescue MRTF-B^–/– mice specifically in neural crest derivatives, MRTF-B^+/– mice were intercrossed with Wnt-1-Cre transgenic mice kindly provided by A. McMahon (Harvard University, Boston) (3).

Fig. 1.

Characterization of MRTF-B gene trap mice. (A) Schematic representation of the MRTF-B gene trap. (Top) The gene trap vector contains a splice acceptor (SA) sequence flanked by loxP sites (triangles) subcloned 5′ of the βgeo cassette and a polyadenylation sequence (pA). (Middle) A partial restriction map of the mouse MRTF-B gene showing exon 10 and 11 sequences (rectangles). PCR primers are shown (arrows). (Bottom) A partial restriction map of the MRTF-B mutated allele. (B) Southern blot analysis of DNA prepared from the offspring of MRTF-B^+/– × MRTF-B^+/– mating. The positions of the wild-type (9.8-kb) and mutant (6.5-kb) alleles are indicated to the left. (C) Northern blot analyses of MRTF-B gene expression in wild-type (+/+), heterozygous (+/–), and null (–/–) MRTF-B embryos. RNA was harvested from E12.5 embryos. The wild-type (9.5-kb) and mutant (7.5-kb) transcripts are shown to the left. (D) Real-time RT-PCR of MRTF-B gene expression in wild-type (+/+), heterozygous (+/–), and null (–/–) MRTF-B embryos. (E) MRTF-A- and –B-induced transactivation of the SM22α promoter in NIH 3T3 cells. NIH 3T3 cells were cotransfected with the indicated amounts (in micrograms) of expression plasmids and p-441SM22.luc. The data are presented as relative luciferase activities ± SEM. (F) Forced expression of the MRTF-BΔ731 mutant protein does not repress myocardin-induced transactivation of the SM22α promoter in NIH 3T3 cells. NIH 3T3 cells were cotransfected with the indicated amounts (μg) of expression plasmid and p-441SM22.luc.

Northern Blot and Real-Time RT-PCR Analyses. Northern blot analysis was performed as described (9). The relative intensity of the 9.5-kb MRTF-B and 7.5-kb mutant MRTF-B transcripts was quantified by using a Molecular Dynamics PhosphorImager. Real-time RT-PCR was performed and quantified as described (9).

Plasmids and Transient Cotransfection Analyses. The pcDNA-MRTF-B and pcDNA-MRTF-BΔ731 expression plasmids encode full-length mouse MRTF-B and the N-terminal 731 amino acids of mouse MRTF-B (NM_153588), respectively. pcDNA3.1-myocardin, pcDNA3-MRTF-A, and the pcDNA-MyocardinΔ585 expression plasmids and the p-441SM22.luc, pPI-Act.luc, and pPI-Myo.luc luciferase reporter plasmids were described previously (9, 13). NIH 3T3 cells were cotransfected with 200 ng of the indicated luciferase reporter plasmid, 100–500 ng of the indicated expression plasmid, and 10 ng of the phRL-TK (-Int) reference plasmid (Promega), as described (18).

Immunohistochemistry and in Situ Hybridization Analyses. Immunohistochemistry and in situ hybridization protocols are available at www.uphs.upenn.edu/mcrc/histology/histologyhome.html (19). Antibodies included: anti-MF20 monoclonal antibody (Developmental Studies Hybridoma Bank, Iowa City, IA); anti-smooth muscle α-actin monoclonal antibody (Dako); neurofilament 2H3 monoclonal antibody (Developmental Studies Hybridoma Bank); anti-Krox20 polyclonal antibody (Covance Research Products, Denver, PA). The plexinA2 antisense cRNA probe was described previously (20).

Intracardiac Ink Injections. India ink was injected intracardially or into the descending aorta of mouse embryos by using a pulled glass pipette and immediately photographed as described (21).

Results

Generation of MRTF-B Gene Trap Mice. Mice were generated from ES cells containing an insertional mutation or “gene trap” in intron 10 of the mouse MRTF-B gene. As shown in Fig. 1A, gene trap vectors were designed to generate spliced fusion transcripts between a reporter gene (βgeo) and the endogenous gene (MRTF-B) present at the site of integration (16, 22). In this gene trap vector, the splice acceptor is flanked by LoxP sites (floxed) to facilitate Cre-mediated cell lineage-restricted rescue of the mutant allele (16). Southern blot analyses confirmed transmission of the targeted allele containing the MRTF-B gene trap through the germline (Fig. 1B). Northern blot analyses demonstrated expression of both the wild-type MRTF-B transcript (MRTF-B^+/+) and the mutated MRTF-B-lacZ fusion transcript (MRTF-B^+/–) (Fig. 1C). The native 9.5-kb MRTF-B transcript was also detectable in MRTF-B^–/– mice demonstrating low-level alternative splicing around the insertional gene trap mutation (Fig. 1C). Quantification of the hybridization signals revealed wild-type MRTF-B mRNA represented 5% of total MRTF-B transcripts in MRTF-B^–/– embryos. This finding was confirmed by real-time RT-PCR (Fig. 1D). Thus, the phenotype of MRTF-B^–/– mice may represent a weak hypomorphic rather than true null phenotype.

Consistent with prior studies (11), inductions in luciferase activity were observed in NIH 3T3 cells cotransfected with expression plasmids encoding MRTF-A and -B, respectively, and the SMC-restricted p-441SM22.luc reporter plasmid (Fig. 1E). In contrast, forced expression of the mutant MRTFBΔ731 protein failed to transactivate the SMC-restricted SM22α promoter (Fig. 1E). As expected, forced expression of a dominant negative myocardin mutant protein (MYCDΔ585) suppressed myocardin-induced SM22α promoter activity (Fig. 1F). By contrast, forced expression of MRTFBΔ731 failed to suppress myocardin- or MRTF-A-induced SM22α promoter activity (Fig. 1F and data not shown). These data demonstrate that the mutant MRTF-B allele is functionally null as a transcriptional coactivator and does not function as a dominant negative to repress activity of myocardin or MRTF-A.

Characterization of MRTF-B-Deficient Embryos. Intercrossing MRTF-B^+/– heterozygotes failed to generate the anticipated Mendelian pattern of inheritance (Table 1). Of 281 postnatal day (P)1 offspring analyzed, only 4 (1.4%) null (MRTF-B^–/–) mice were identified. At birth (P0), 6% of 294 pups were genotyped as null, 65% were heterozygotes (MRTF-B^+/–), and 30% were wild-type (MRTF-B^+/+). Through E16.5, the percentage of MRTF-B null mice varied between 21% and 23%, suggesting that MRTF-B null mice are generally viable until late in gestation. However, at E17.5 and E18.5, only 16% of embryos were MRTF-B null. These data demonstrate that, in contrast to myocardin null embryos that die at E10.5 (10), MRTF-B null embryos develop a vasculature that is capable of sustaining an embryo through late gestation, and that perinatal demise initiates at about E16.5.

Table 1. Genotype distribution of MRTF-B embryonic and perinatal lethality.

Genotype	E12.5 (53)	E13.5 (40)	E14.5 (66)	E16.5 (142)	E17.5 (55)	E18.5 (113)	P0 (294)	P1 (281)
+/+	17	12	17	39	14	36	87	87
+/-	24	19	35	72	32	58	190	190
-/-	12 (23%)	9 (23%)	14 (21%)	31 (22%)	9 (16%)	18 (16%)	17 (6%)	4 (1.4%)

To determine the cause of demise of MRTF-B null embryos, MRTF-B^–/– embryos were characterized at E16.5 and E18.5 (Table 2). The embryos did not appear to be growth retarded, although 6 of 16 MRTF-B null embryos appeared to be grossly edematous. Inspection of the heart and great arteries revealed cardiac outflow tract defects in all 16 MRTF-B^–/– embryos analyzed (Fig. 2). Eight of 16 embryos exhibited double-outlet right ventricle (Fig. 2 G–I), and 7 of 16 embryos exhibited persistent truncus arteriosus (Fig. 2 J–L). In addition, two mutant embryos exhibited an interrupted aortic arch (arrow, Fig. 2B), and one exhibited a right-sided aortic arch (arrow, Fig. 2C). Large ventricular septal defects were observed in all MRTF-B null embryos but were never observed in wild-type embryos after E14.5 (Fig. 2 F and I). Complex structural heart disease in association with signs of congestive heart failure (edema) suggests that prenatal lethality in some embryos is due to cardiovascular insufficiency. Structural heart defects such as those observed in MRTF-B null pups are sufficient to account for postnatal lethality of the remaining animals.

Table 2. Summary of phenotypes in MRTF-B-deficient embryos.

Ventricular septal defect	16/16
Persistent truncus ateriosus	7/16
Double-outlet right ventricle	8/16
Right-sided aortic arch	1/16
Interrupted aortic arch	2/16
Perivascular hemorrhage and edema	6/16

Fig. 2.

Cardiovascular abnormalities in MRTF-B null mice. (A) Wild-type E18.5 embryo demonstrating left-sided aortic arch (arrow). (B) MRTF-B^–/– embryo with interrupted aortic arch (arrow). (C) MRTF-B^–/– embryo with right-sided aortic arch (arrow). (D–F) Rostral (D) to caudal (F) sections demonstrating the aorta (Ao) arising from the LV and the pulmonary artery (PA) arising from the RV in a control E16.5 embryo. (G–I) Double-outlet right ventricle in an E16.5 MRTF-B^–/– mutant embryo shown rostral (G) to caudal (I). Both the aorta (Ao) and pulmonary artery (PA) arise from the RV, and there is an obligate ventricular septal defect (VSD). (J–L) Truncus arteriosus defect in an E16.5 MRTF-B^–/– mutant embryo shown rostral (J) to caudal (L).

MRTF-B Expression in the Cardiac Neural Crest and Aortic Arch Artery Malformations. The pattern of MRTF-B gene expression was mapped by characterizing the pattern of lacZ staining in MRTF-B^+/– embryos. Surprisingly, at E8.5, intense β-galactosidase activity was observed in the dorsal neural folds at the level of rhombomeres 3 and 5 (Fig. 3A). This lacZ activity colocalized with cells expressing Krox20, a marker of neural crest (Fig. 3B) (23). Intense blue staining was also observed throughout the heart at E8.5 and E9.5 (Fig. 3A). LacZ expression surrounding the aortic arch arteries and the thoracic aorta was still evident at E10.5 (Fig. 3C). At E11.5, lacZ-positive MRTF-B-expressing cells populated branchial arch arteries 3, 4, and 6 (Fig. 3D, arrows). MRTF-B staining colocalized with expression of SM-α-actin in the aortic arch arteries (Fig. 3 E and F). The finding that MRTF-B is expressed at high levels in the dorsal neural tube and subsequently in the aortic arch arteries suggested that expression of MRTF-B in the cardiac neural crest could be responsible for the outflow tract pathology observed in MRTF-B null embryos.

Fig. 3.

MRTF-B is expressed in the cardiac neural crest and aortic arch arteries during embryonic development. (A) An E8.5 MRTF-B^+/– embryo demonstrating expression of the MRTF-B-lacZ fusion protein (arrows) in rhombomeres 3 and 5 of the dorsal neural folds. (B) An E8.5 embryo demonstrating Krox20 expression (arrows) in rhombomeres 3 and 5. (C) An E9.5 MRTF-B^+/– embryo demonstrating lacZ-positive cells populating the aorta (arrows), aortic arch arteries, cardiac outflow tract, and heart. (D) An E11.5 MRTF-B^+/– embryo demonstrating lacZ expression in the cells populating the mesenchyme surrounding aortic arch arteries 3, 4, and 6 (arrows). (E and F) Higher-power view of D demonstrating colocalization of β-galactosidase activity (blue in E) and SM-α-actin (red in F) in the cells populating the third left aortic arch artery. (E Inset) Colocalization of lacZ (blue) and SM-α-actin (orange) expression.

To examine the timing and origin of the cardiac outflow tract defects, India ink was injected into the heart and dorsal aorta of wild-type (Fig. 4 A and C) and MRTF-B^–/– (Fig. 4 B and D) embryos harvested at E10.5 and E11.5. In wild-type embryos, three right-sided and three left-sided arch arteries were clearly delineated (arrows, Fig. 4 A and C). By contrast, generally only one or two right- and left-sided aortic arch arteries were observed in MRTF-B-deficient embryos (arrows, Fig. 4 B and D). Several different abnormal arch artery patterns were observed in MRTF-B null embryos. In some cases, the left fourth arch artery, which gives rise to the definitive aortic arch, was absent, consistent with our observation of interrupted aortic arch phenotypes in some MRTF-B null animals. In other embryos, one or both sixth arch arteries, which normally give rise to the pulmonary trunk and ductus arteriosus, were not visualized. These data suggest that abnormal aortic arch artery patterning is evident as early as E10.5 in a manner consistent with subsequent structural heart disease seen in newborn MRTF-B-deficient pups.

Fig. 4.

Cardiac neural crest cells migrate appropriately but exhibit a block in SMC differentiation in MRTF-B^–/– embryos. (A–D) Intracardiac ink injection to visualize the pharyngeal arch arteries in E10.5 (A and B) and 11.5 (C and D) wild-type (A and C) and MRTF-B^–/– (B and D) embryos. Note regression of aortic arch arteries 4 and 6 in MRTF-B^–/– mutant embryos (B and D). (E–H) Coronal sections of E11.5 wild-type (E and F) and MRTF-B^–/– (G and H) embryos demonstrate expression of SM-α-actin (E) and plexinA2 (F) in the mesenchyme surrounding the aortic arch arteries in wild-type embryos. By contrast, in E11.5 MRTF-B-deficient embryos (G and H), expression of SM-α-actin is markedly down-regulated in the aortic arch arteries (G, arrows). In this representative embryo, expression of SM-α-actin in the fourth and sixth right arch arteries was barely detectable (G). Expression of plexinA2 was readily detectable in the pharyngeal mesenchyme surrounding the aortic arch arteries (H). PlexinA2 gene expression in wild-type and MRTF-B^–/– embryos in the pharyngeal mesenchyme was comparable (compare F and H). (I–L) Sagittal sections cut at the level of the aorticopulmonary septum of E11.5 wild-type (I and J) and MRTF-B^–/– (K and L) embryos demonstrating coexpression of SM-α-actin (I) and plexinA2 (J) in the nascent aorticopulmonary septum of wild-type embryos (arrows, I and J). By contrast, in MRTF-B^–/– embryos, the aorticopulmonary septum fails to develop, and expression of SM-α-actin is severely down-regulated at the level of the outflow tract (arrows in K). PlexinA2-expressing cells (arrows in L) are observed, demonstrating that neural crest cells migrated appropriately to the cardiac outflow tract.

At E11.5, SM-α-actin, an early SMC marker, is expressed strongly in the aortic arch arteries of wild-type mice (Fig. 4E, arrows). Consistent with their neural crest origin, these cells (arrows) also coexpress the neural crest marker, plexinA2 (Fig. 4F). By contrast, E11.5 MRTF-B-deficient embryos exhibited multiple defects in aortic arch patterning, including severely dilated aortic arch arteries, absent arch arteries, and dilated aortic sac (Fig. 4G). Moreover, SM-α-actin was absent or severely attenuated in the aortic arch arteries of MRTF-B null embryos (arrow, Fig. 4G). No obvious difference in the expression pattern of plexinA2 was observed in the mesenchyme surrounding aortic arch arteries of wild-type and MRTF-B null embryos (compare Fig. 4 F and H).

In addition at E11.5, SM-α-actin-positive cells may be observed forming the nascent septum that divides the truncus arteriosus into the aorta and pulmonary artery (Fig. 4I). These cells express the plexinA2 gene, confirming their origin as neural crest (Fig. 4J). By contrast, in MRTF-B null embryos, the aorticopulmonary septum was either absent or severely diminished (arrows), and SM-α-actin-positive cells were rarely observed (arrows, Fig. 4K). However, two columns of plexinA2-expressing cells were readily observed populating the cardiac outflow tract at the level of the aorticopulmonary septum (arrows, Fig. 4L), demonstrating that the cardiac neural crest cells migrated appropriately to the cardiac outflow tract. Taken together, these data reveal a specific block in the capacity of cardiac neural crest derivatives to undergo SMC differentiation in MRTF-B-deficient mice.

Neural Crest-Specific Rescue of MRTF-B Null Embryos. To confirm that the outflow tract defects and mortality observed in MRTF-B-deficient mice resulted from a cell autonomous defect of cardiac neural crest, MRTF-B^+/– mice were intercrossed with mice expressing Cre under the transcriptional control of the neural crest-restricted Wnt-1 promoter (Fig. 5A) (3). To confirm that Wnt1-Cre (Wnt1-Cre⁺) mice efficiently excised the conditional MRTF-B gene trap, expression of β-galactosidase in MRTF-B^–/–/Wnt1-Cre⁺ and MRTF-B^–/–/Wnt1-Cre^– mice was compared. In E11.5 embryos stained for extended periods of time, β-galactosidase expression was virtually identical in Cre⁺ and Cre^– embryos, except in the neural crest-derived regions of mandible, pharyngeal arches, and cardiac outflow tract (arrows, Fig. 5 B and C). Most importantly, of 74 live-born offspring generated from MRTF-B^+/–/Wnt1-Cre⁺ × MRTF-B^+/–/Wnt-1-Cre⁺ intercrosses, 15 (20.2%) MRTF-B^–/–/Wnt1-Cre⁺ offspring were observed at P1 compared with two of 74 (2.7%) offspring of the MRTF-B^–/–/Wnt1-Cre^– genotype. All MRTF-B^–/–/Wnt1-Cre⁺ E16.5–18.5 embryos and P14 mice appeared phenotypically normal, and outflow tract defects and ventricular septal defects were not observed. Furthermore, 100% (15/15) of the MRTF-B^–/–/Wnt1-Cre⁺ mice were alive and well 12 weeks after birth, whereas only two MRTF-B^–/–/Wnt1-Cre^– littermates survived past P1. Hence, cardiac outflow tract defects in MRTF-B null mice are efficiently rescued by neural crest-restricted expression of Cre, supporting the conclusion that expression of MRTF-B in the cardiac neural crest is required for differentiation of neural crest derivatives to SMCs and for cardiac outflow tract patterning.

Fig. 5.

Neural crest-restricted rescue of MRTF-B gene trap mice. (A) Schematic representation of the MRTF-B gene trap rescue strategy. In MRTF-B^–/–/Cre⁺ mice, the splice acceptor sequence (SA) is deleted, regenerating the native MRTF-B transcript specifically in neural crest cells. (B and C) LacZ expression (blue staining) in MRTF-B^–/–/Cre^– (B) and MRTF-B^–/–/Cre⁺ (C) E11.5 embryos. Note the failure of the neural crest-derived branchial arch region of the Cre⁺ embryo to stain blue (arrows).

Discussion

The observed phenotype of MRTF-B-deficient embryos demonstrates that, despite the high-level sequence identity between MRTFs and the capacity of each MRTF to transactivate multiple SMC-restricted transcriptional elements (11), the function of MRTF-B in the cardiac neural crest is unique and nonredundant with other MRTF family members. MRTF-B-deficient embryos survived only through late gestation and demonstrated cell autonomous defects in neural crest differentiation into SMCs cells accompanied by cardiac outflow tract defects. The conclusion that the primary defect in MRTF-B-deficient mice is a cell autonomous defect in cardiac neural crest derivatives rather than a primary defect in cardiac morphogenesis and/or myocyte differentiation is substantiated by the finding of normal heart morphology in rescued MRTF-B^–/–/Wnt-1-Cre⁺ mice. As such, these data serve to identify MRTF-B as a candidate gene for common forms of congenital heart disease, including persistent truncus arteriosus, interrupted aortic arch, and double-outlet right ventricle.

It is likely that the respective functions of MRTF-B and myocardin during embryonic development are in large part related to differences in their temporal and spatial patterns of expression. MRTF-B, but not myocardin, is expressed in the cardiac neural crest as early as E8.5, and the MRTF-B gene continues to be expressed in the migrating cardiac neural crest at least through E10.5. In addition, MRTF-B is expressed in the embryonic heart and lateral mesoderm-derived vasculature but, other than secondary heart defects that presumably were related to the outflow tract defects, these tissues appeared grossly normal in MRTF-B-deficient mice. Therefore, MRTF-B is not required for expression of differentiated SMC markers in the heart and peripheral vasculature during embryonic development, although it may mediate redundant functions with MRTF-A and/or myocardin in these tissues. By contrast, despite the fact that myocardin and MRTF-B are coexpressed in embryonic vascular SMCs, myocardin null mice die at E10.5, and differentiated SMCs are not observed in the dorsal aorta (12). Thus, MRTF-B cannot compensate for loss of myocardin at this early stage of embryonic angiogenesis in these cells.

Formation of the cardiac outflow tract and great arteries is controlled by multiple distinct signals originating in the pharyngeal endoderm and local mesoderm that converge upon the cardiac neural crest cells acting in a combinatorial fashion (1, 2, 24). Neural crest-specific inactivation of the type 1 bone morphogenetic protein receptor Alk2 causes a failure of outflow tract septation with deficient differentiation of neural crest-derived SMCs (21). Inactivation of Semaphorin 3C, a secreted class 3 semaphorin, results in interruption of the aortic arch and persistent truncus arteriosus (25). Class 3 semaphorins bind to heterodimeric receptors composed of neuropilin and plexin subunits. Neuropilin 1-deficient mice display persistent truncus arteriosus, and both neuropilin 1 and plexinA2 are expressed by migrating cardiac neural crest cells (20, 26). However, interpretation of these data is complicated, because neuropilin receptors also bind to VEGF. Selective ablation of the VEGF-165 isoform, which signals via neuropilin-containing receptors, results in a spectrum of outflow tract defects recapitulating DiGeorge syndrome (27). Although neural crest-derived SMC differentiation has been noted to be abnormal in several mouse models with outflow tract defects, it has not been clear whether this represents the primary cause of the morphogenetic abnormalities. However, loss of MRTF-B is likely to represent a specific loss of SMC differentiation due to deficient activation of SMC structural gene promoters in neural crest cells. Hence, the development of typical forms of congenital heart disease in MRTF-B mutants provides compelling evidence that disruption of SMC differentiation is sufficient to produce cardiac outflow tract defects.

Analysis of the MRTF-B-deficient mice suggests a molecular model in which MRTF-B sits at a convergence point in cardiac neural crest, integrating multiple extracellular signals and developmental cues ultimately promoting vascular SMC differentiation. This hypothesis is supported by the finding that MRTF-B is expressed at high levels in the cardiac neural crest, and MRTF-B-deficient embryos recapitulate the outflow tract defects observed in bone morphogenetic protein/semaphorin/VEGF mutant mice (21, 26, 27). In this regard, it is noteworthy that each of these growth factors is secreted by cells in the pharyngeal endoderm and mesoderm surrounding the aortic arch arteries and outflow tract (27). We postulate that the appropriate combination of signals promotes the translocation of MRTF-B from the cytoplasm to the nucleus, where it physically associates with SRF-activating transcription of a set of genes promoting SMC differentiation (14). The demonstration that forced expression of MRTF-B and SRF transactivates multiple SMC-transcriptional regulatory elements in non-SMCs and that both MRTF-B and SRF are required to activate endogenous SMC genes in SRF-deficient ES cells (J.L., unpublished observation) strongly supports the contention that MRTF-B acts as a critical transcriptional coactivator of SRF in the cardiac neural crest. As such, MRTF-B mediates a cell autonomous function in the cardiac neural crest integrating multiple signals that regulate the process of neural crest differentiation into SMCs and the morphogenetic program controlling formation of the cardiac outflow tract and great arteries.