Nonmuscle myosin IIB (NMIIB; heavy chain encoded by MYH10) is essential for cardiac myocyte cytokinesis. The role of NMIIB in other cardiac cells is not known. Here, we show that NMIIB is required in epicardial formation and functions to support myocardial proliferation and coronary vessel development. Ablation of NMIIB in epicardial cells results in disruption of epicardial integrity with a loss of E-cadherin at cell–cell junctions and a focal detachment of epicardial cells from the myocardium. NMIIB-knockout and blebbistatin-treated epicardial explants demonstrate impaired mesenchymal cell maturation during epicardial epithelial–mesenchymal transition. This is manifested by an impaired invasion of collagen gels by the epicardium-derived mesenchymal cells and the reorganization of the cytoskeletal structure. Although there is a marked decrease in the expression of mesenchymal genes, there is no change in Snail (also known as Snai1) or E-cadherin expression. Studies from epicardium-specific NMIIB-knockout mice confirm the importance of NMIIB for epicardial integrity and epicardial functions in promoting cardiac myocyte proliferation and coronary vessel formation during heart development. Our findings provide a novel mechanism linking epicardial formation and epicardial function to the activity of the cytoplasmic motor protein NMIIB.

The development of a functional heart is orchestrated by various lineages of cardiac cells including epicardial, endocardial, myocardial and cardiac non-myocyte interstitial cells (Tirziu et al., 2010; Brade et al., 2013), each of which contains the cytoplasmic contractile protein nonmuscle myosin IIB (NMIIB; heavy chain encoded by MYH10) (Ma and Adelstein, 2012). The epicardium is a single cell layer of mesothelial cells that covers the outer surface of the heart and protects the integrity of the myocardium (Von Gise and Pu, 2012). It is also the source of a number of growth factors that promote myocardial growth during embryonic heart development. Moreover the epicardium serves as the source for many types of cardiac cells following the epicardial epithelial–mesenchymal transition (EMT). These include pericytes, vascular smooth muscle and endothelial cells, which are critical for cardiac coronary vessel formation, as well as interstitial fibroblasts, which play important roles in cardiac architecture. They also may be a source for cardiac myocytes (Von Gise and Pu, 2012).

Decades of research have revealed the important extracellular signals (TGFβ, retinoic acids, PDGF, FGFs etc.), and the downstream Snail and E-cadherin canonical pathway regulating epicardial EMT (Brade et al., 2013; Von Gise and Pu, 2012). The significance of cytoskeletal dynamics in controlling EMT is less frequently addressed. However, there are several lines of evidence that point to a role for NMII in EMT. First, epicardial EMT features an extensive actin cytoskeletal reorganization where epicardial cells lose their epithelial apical-basal polarity and strong cell–cell adhesions, to acquire mesenchymal front-rear polarity and focal adhesions resulting in increased motility and invasiveness (Brade et al., 2013; Von Gise and Pu, 2012). NMII plays important roles in regulating actin-cytoskeletal structures (Heissler and Manstein, 2013; Vicente-Manzanares et al., 2009). Second, Rho activity is also important in epicardial EMT and cardiac coronary vessel formation. Inhibiting Rho signaling blocks epicardial EMT both in vivo and in vitro (Artamonov et al., 2015; Dokic and Dettman, 2006; Lu et al., 2001). Disruption of the Vangl2 and Rho kinase pathway in the developing mouse heart impairs coronary vessel formation (Phillips et al., 2008). NMII is activated by phosphorylation of its associated regulatory light chain (MLC20; also known as MYL12A and MYL12B) by myosin light chain kinase or Rho kinase (which also directly inhibits myosin phosphatase) (Vicente-Manzanares et al., 2009). Third, inhibition of NMII activity affects the behavior of neural crest cells generated during neuroepithelial EMT in vivo in zebrafish brain development (Berndt et al., 2008). Moreover, fourth, NMII activity has been shown to affect stem cell lineage specification (Buxboim et al., 2014; Engler et al., 2006; Kim et al., 2015; Wang et al., 2012). All of these findings suggest that NMII may be involved in regulating epicardial EMT.

NMII is one of the major cellular motor proteins regulating cytoskeletal structure and function by interacting with actin to either generate tension on actin filaments or translocate actin filaments. Three isoforms of NMII have been identified in vertebrates including humans and mice, namely NMIIA, NMIIB and NMIIC based on three different heavy chain (NMHC) genes: Myh9 encoding NMHCIIA, Myh10 encoding NMHCIIB and Myh14 encoding NMHCIIC (Golomb et al., 2004; Berg et al., 2001). Each isoform plays unique as well as overlapping roles during mouse embryonic development partially due to their differences in dynamic motor activities and expression patterns in various tissues (Ma and Adelstein, 2014b).

Compared to NMIIA and NMIIC, NMIIB is relatively enriched in the brain and heart. Mice with a knockout for NMIIB die during embryonic development by embryonic day (E)14.5 with severe congenital cardiac abnormalities. These include a hypoplastic myocardium with reduced proliferative activity of the cardiac myocytes and premature cardiac myocyte bi-nucleation, in addition to cardiac structural abnormalities such as a ventricular septal defect, double outlet of the right ventricle and pulmonary arterial stenosis (Tullio et al., 1997). Our previous studies on NMIIB in the heart primarily focused on cardiac myocytes. Knockout of NMIIB in cardiac myocytes resulted in a failure in cytokinesis (Takeda et al., 2003). Moreover, NMIIB exerts tension to drive contractile ring constriction during cardiac myocyte cytokinesis (Ma et al., 2012). NMIIB is also required to disrupt the cardiac myocyte cell–cell adhesion complex during outflow tract myocardialization, the process necessary for normal alignment of the aorta to the left ventricle (Ma and Adelstein, 2014a), and to maintain the integrity of cardiac intercalated discs in adult hearts (Ma et al., 2009). The roles of NMIIB in other cardiac cells, such as the epicardium, have not yet been studied. The current study seeks to understand the role of NMIIB in epicardial formation and function in vivo during mouse cardiac development.

Abnormal epicardium and coronary vessels in B/B hearts

We have previously shown that NMIIB is required for cardiac myocyte cytokinesis during mouse heart development (Takeda et al., 2003). In addition to its expression in cardiac myocytes, NMIIB is also expressed in epicardial cells (Ma and Adelstein, 2012). We examined the localization of NMIIB in the developing epicardium of freshly isolated hearts from E14.5 mice expressing GFP-tagged NMHCIIB (denoted BGFP) (Bao et al., 2007). Confocal analysis of E14.5 whole mouse hearts shows that NMIIB is concentrated at the cell–cell junctions of the epicardium (Fig. 1A, green). Super-resolution structured illumination microscopy (SIM) analysis further shows paired NMIIB alignment between epicardial cells (Fig. 1B), reminiscent of NMII localization at epithelial cell–cell junctions (Ebrahim et al., 2013) and suggesting a role for NMIIB in regulating epicardial cell–cell adhesion.

Fig. 1.

Localization of NMIIB in epicardium and abnormalities of B/B epicardium. (A,B) Confocal images of freshly isolated E14.5 hearts expressing EGFP–NMHCIIB (BGFP) show localization of NMIIB at cell–cell junctions of the epicardium (A, green). Scale bar: 20 µm. Super-resolution SIM shows paired alignments of NMIIB at the cell–cell junctions (B). (C,D) Whole-mount immunofluorescence confocal images of E13.5 mouse epicardium showing E-cadherin (red) at the epicardial cell–cell junctions in B+/B+ mouse hearts (C). In B/B mouse hearts, E-cadherin is greatly diminished at the cell–cell junctions (D). Nuclei were stained blue with DAPI. Scale bar: 20 µm. (E,F) Biotin permeability assay of E13.5 mouse epicardium showing impaired epicardial integrity in B/B hearts. Biotin was detected with Rhodamine-conjugated streptavidin (red) and shows deep penetrance throughout the entire ventricle in B/B hearts (F, red). Biotin is limited near the epicardial layer in B+/B+ hearts (E, red). Vimentin (green) stains cardiac nonmyocytes. Nuclei were stained blue with DAPI. Arrowheads point to the epicardium. Scale bars: 50 µm.

Fig. 1.

Localization of NMIIB in epicardium and abnormalities of B/B epicardium. (A,B) Confocal images of freshly isolated E14.5 hearts expressing EGFP–NMHCIIB (BGFP) show localization of NMIIB at cell–cell junctions of the epicardium (A, green). Scale bar: 20 µm. Super-resolution SIM shows paired alignments of NMIIB at the cell–cell junctions (B). (C,D) Whole-mount immunofluorescence confocal images of E13.5 mouse epicardium showing E-cadherin (red) at the epicardial cell–cell junctions in B+/B+ mouse hearts (C). In B/B mouse hearts, E-cadherin is greatly diminished at the cell–cell junctions (D). Nuclei were stained blue with DAPI. Scale bar: 20 µm. (E,F) Biotin permeability assay of E13.5 mouse epicardium showing impaired epicardial integrity in B/B hearts. Biotin was detected with Rhodamine-conjugated streptavidin (red) and shows deep penetrance throughout the entire ventricle in B/B hearts (F, red). Biotin is limited near the epicardial layer in B+/B+ hearts (E, red). Vimentin (green) stains cardiac nonmyocytes. Nuclei were stained blue with DAPI. Arrowheads point to the epicardium. Scale bars: 50 µm.

Epicardial integrity is maintained by epicardial cell–cell junctions, including adherens and tight junctions. We examined these junctions in mice with global knockout of NMHCIIB (i.e. Myh10–/–; hereafter denoted B/B) at E13.5 by whole-mount immunofluorescence staining for localization of the adherens junction molecule E-cadherin and the tight junction molecule ZO-1 (also known as TJP1). E-cadherin is enriched at the epicardial cell–cell junctions in the wild-type hearts (Fig. 1C, red). In contrast, enrichment for E-cadherin is markedly diminished at epicardial cell–cell junctions in B/B hearts (Fig. 1D, red). The localization of ZO-1 at epicardial cell–cell junctions is observed both in B+/B+ (i.e. wild-type for NMIIB) and B/B hearts (Fig. S1, green). However, instead of forming a uniform continuous sheet of ZO-1-stained tight junctions, as seen in the wild-type epicardium (Fig. S1A, green), B/B epicardium shows irregular ZO-1 staining with an uneven distribution apparent at different confocal z-sections (Fig. S1B–D, green). This uneven distribution of ZO-1 may reflect its relocalization to the lateral sides of the cell due to the altered epicardial apical-basal polarity. z-projections of ZO-1 staining, however, show continuity of the tight junction in B/B hearts. These results are consistent with the disorganization of the B/B epicardium. To further examine whether ablation of NMIIB affects the epicardial barrier function, we carried out a biotin permeability assay by pipetting biotin solution into the thoracic cavity of E13.5 mouse embryos. Detection of biotin (red) in heart sections through immunofluorescence shows that biotin is mostly confined to the epicardial layer in B+/B+ hearts (Fig. 1E, red, arrowheads) but, in contrast, is detected throughout the various layers of the heart in B/B mice (Fig. 1F, red) indicating a compromised barrier function. Thus, NMIIB is important for epicardial integrity because it functions in the maintenance of epicardial cell–cell junctions.

In addition to protecting the heart, the epicardium also plays important roles in coronary vessel formation during heart development. We next examined the coronary vessels in E13.5 B/B hearts through whole-mount staining with antibodies against PECAM1 (CD31, green), a marker for endothelial cells. Whereas the developing coronary vasculature of B+/B+ hearts shows normal ‘chicken wire’ coronary plexuses covering the entire dorsal surface of the ventricles (Fig. 2A, enlarged in 2C), the B/B hearts show abnormal coronary vascular morphogenesis, with markedly reduced coverage by vascular plexuses (Fig. 2B, enlarged in 2D). These phenotypes in B/B hearts are consistent with abnormalities in the epicardium.

Fig. 2.

Abnormal coronary vessel formation in B/B hearts. Confocal z-projections of the dorsal surface of whole-mount E14.5 hearts stained for CD31 to reveal coronal vessels showing that the coronary plexuses cover the whole B+/B+ ventricular surface (A, enlarged in C). The normal ‘chicken wire’ coronary plexus is not obvious in the B/B heart (B, enlarged in D).

Fig. 2.

Abnormal coronary vessel formation in B/B hearts. Confocal z-projections of the dorsal surface of whole-mount E14.5 hearts stained for CD31 to reveal coronal vessels showing that the coronary plexuses cover the whole B+/B+ ventricular surface (A, enlarged in C). The normal ‘chicken wire’ coronary plexus is not obvious in the B/B heart (B, enlarged in D).

NMIIB is important for epicardial function to support myocardial growth

There is extensive cross-talk between the epicardium and myocardium during embryonic heart development. Defects in myocardial development such as those seen in our B/B hearts could conversely also affect epicardial formation. To confirm the importance of epicardial NMIIB in heart development, we generated epicardial-specific NMIIB-knockout mice by crossing NMIIB floxed mice (Bflox/Bflox) with a mouse line expressing Cre recombinase driven by the WT-1 promoter and examined heart development and coronary vessel formation in these mice (BWT-1/BWT-1). Although WT-1-mediated Cre recombinase has been reported to be transiently activated in other cell types (Rudat and Kispert, 2012), the BWT-1/BWT-1 mouse hearts studied here show specific knockout of NMIIB in epicardial cells but not in endocardial and myocardial cells (Fig. S2). Knockout of NMIIB in the epicardium was confirmed by immunofluorescence confocal microscopy using antibodies to NMHCIIB (Fig. S2C,D compared to Fig. S2A,B, red, arrows). The expression of NMIIB in the myocardium (desmin, green, a marker for myocytes) and endocardium (arrowheads) in BWT-1/BWT-1 hearts (Fig. S2C,D) was comparable to that in Bflox/Bflox hearts (Fig. S2A,B).

Knockout of NMIIB does not affect NMIIA expression in BWT-1/BWT-1 epicardial cells (Fig. S3E, red, arrows) compared to what is seen in Bflox/Bflox cells (Fig. S3B, red, arrows). There is also no change in the phosphorylation of serine 19 of the regulatory myosin light chain (pMLC20) in BWT-1/BWT-1 epicardial cells (Fig. S3D, green, arrowheads) compared to what is seen in Bflox/Bflox cells (Fig. S3A, green, arrowheads). Note that the number of NMIIA-positive non-myocytes in compact myocardium is markedly reduced in BWT-1/BWT-1 hearts (Fig. S3F, arrows) compared to that in Bflox/Bflox hearts (Fig. S3C, arrows) indicating a defect in epicardium-derived cell development (see below for details). These results indicate that WT-1 Cre-mediated knockout of NMIIB in mice primarily affects epicardial expression of NMIIB in BWT-1/BWT-1 hearts and not myocardial or endocardial expression.

Histological analysis shows that the compact myocardium is markedly thinner in BWT-1/BWT-1 hearts compared to that in Bflox/Bflox hearts at E15.5 (Fig. 3A,B, brackets). BWT-1/BWT-1 hearts at E12.5 show focal detachment of the epicardium from the myocardium (Fig. 3D, arrows). This detachment is not seen in control hearts (Fig. 3C, Bflox/Bflox). It has been reported that epicardial N-cadherin is required to maintain heterotypic epicardial–myocardial cell–cell interactions (Luo et al., 2006). Immunoconfocal microscopy analysis shows that N-cadherin is expressed in both epicardial and myocardial cells in Bflox/Bflox hearts (Fig. 3E, red). Knockout of NMIIB in epicardial cells does not affect N-cadherin expression in myocardial cells in BWT-1/BWT-1 hearts. However, the expression of N-cadherin in epicardial cells is markedly diminished in BWT-1/BWT-1 hearts (Fig. 3F, red, arrowheads) compared to that seen in the Bflox/Bflox epicardial cells (Fig. 3E, red, arrowheads). The expression of β1-integrin is not affected in BWT-1/BWT-1 epicardium (Fig. S4). These results indicate that epicardial NMIIB is important for epicardial–myocardial attachment by, at least in part, maintaining N-cadherin-mediated epicardial–myocardial cell–cell adhesion. Unlike the B/B heart, no evidence of bi-nucleated cardiac myocytes was found in BWT-1/BWT-1 hearts, suggesting that NMIIB functions normally in BWT-1/BWT-1 cardiac myocyte cytokinesis. BWT-1/BWT-1 embryonic hearts, however, showed a reduced proliferation of cardiac myocytes, as evaluated by the BrdU-labeling assay (Fig. 4A,B, green). BrdU-positive cardiac myocytes in BWT-1/BWT-1 hearts (Fig. 4B, 21±4%; mean±s.d., n=3 mice) are fewer in number than in Bflox/Bflox hearts (Fig. 4A, 28±3%; mean±s.d., n=3 mice, P<0.01). BWT-1/BWT-1 hearts also show an abnormal increase in nuclear accumulation of the Cdk inhibitor p21cip (also known as CDKN1A) in compact cardiac myocytes (Fig. 4D, green) compared to in the Bflox/Bflox hearts (Fig. 4C) indicating an abnormal early exit of cardiac myocytes from the cell cycle in BWT-1/BWT-1 hearts.

Fig. 3.

Epicardial knockout of NMIIB results in a hypoplastic myocardium and focal detachment of epicardium from myocardium. (A–D) H&E-stained heart sections of Bflox/Bflox (A,C) and BWT-1/BWT-1 (B,D) hearts showing reduced thickness of the compact myocardium in BWT-1/BWT-1 hearts (B, bracket) compared to Bflox/Bflox hearts (A, bracket) at E15.5. Detachment of epicardial cells from myocardium is more often seen in BWT-1/BWT-1 hearts at E12.5 (D, arrows) compared to in Bflox/Bflox hearts (C). Scale bars: 200 µm. (E,F) Immunofluorescence confocal microscope images of E13.5 mouse hearts stained with an antibody for N-cadherin (red) showing loss of expression in the epicardium in BWT-1/BWT-1 hearts (F, arrowheads) compared to in Bflox/Bflox hearts (E, arrowheads). Nuclei were stained blue with DAPI. Scale bar: 100 µm.

Fig. 3.

Epicardial knockout of NMIIB results in a hypoplastic myocardium and focal detachment of epicardium from myocardium. (A–D) H&E-stained heart sections of Bflox/Bflox (A,C) and BWT-1/BWT-1 (B,D) hearts showing reduced thickness of the compact myocardium in BWT-1/BWT-1 hearts (B, bracket) compared to Bflox/Bflox hearts (A, bracket) at E15.5. Detachment of epicardial cells from myocardium is more often seen in BWT-1/BWT-1 hearts at E12.5 (D, arrows) compared to in Bflox/Bflox hearts (C). Scale bars: 200 µm. (E,F) Immunofluorescence confocal microscope images of E13.5 mouse hearts stained with an antibody for N-cadherin (red) showing loss of expression in the epicardium in BWT-1/BWT-1 hearts (F, arrowheads) compared to in Bflox/Bflox hearts (E, arrowheads). Nuclei were stained blue with DAPI. Scale bar: 100 µm.

Fig. 4.

Epicardial knockout of NMIIB impairs cardiac myocyte proliferation and epicardial FGF-9 expression. (A,B) Immunofluorescence confocal microscope images of E13.5 hearts stained for BrdU (green) and desmin (red, indicating cardiac myocytes) showing that BrdU-positive cardiac myocytes in BWT-1/BWT-1 hearts (B) are fewer in number than in Bflox/Bflox hearts (A, see text for quantification). (C,D) Immunofluorescence confocal microscope images of E13.5 hearts stained for p21cip (green) and desmin (red) showing a marked increase in p21-positive cardiac myocytes in the compact myocardium in BWT-1/BWT-1 hearts (D, green) compared to in Bflox/Bflox hearts (C, green). (E,F) Immunofluorescence confocal microscope images of E14.5 hearts stained with antibodies for FGF-9 (red) showing diminished FGF-9 expression in the epicardium in BWT-1/BWT-1 hearts (F, arrows) compared to in Bflox/Bflox hearts (E, arrows). Nuclei were stained blue with DAPI. Scale bars: 50 µm (A,B,E,F); 100 µm (C,D).

Fig. 4.

Epicardial knockout of NMIIB impairs cardiac myocyte proliferation and epicardial FGF-9 expression. (A,B) Immunofluorescence confocal microscope images of E13.5 hearts stained for BrdU (green) and desmin (red, indicating cardiac myocytes) showing that BrdU-positive cardiac myocytes in BWT-1/BWT-1 hearts (B) are fewer in number than in Bflox/Bflox hearts (A, see text for quantification). (C,D) Immunofluorescence confocal microscope images of E13.5 hearts stained for p21cip (green) and desmin (red) showing a marked increase in p21-positive cardiac myocytes in the compact myocardium in BWT-1/BWT-1 hearts (D, green) compared to in Bflox/Bflox hearts (C, green). (E,F) Immunofluorescence confocal microscope images of E14.5 hearts stained with antibodies for FGF-9 (red) showing diminished FGF-9 expression in the epicardium in BWT-1/BWT-1 hearts (F, arrows) compared to in Bflox/Bflox hearts (E, arrows). Nuclei were stained blue with DAPI. Scale bars: 50 µm (A,B,E,F); 100 µm (C,D).

We next asked whether knockout of NMIIB affects epicardial function. Epicardial FGF is one of the major signals regulating myocardial growth and coronary vessel formation during embryonic heart development. Both FGF-2 and FGF-9, and the FGF receptor FGFR1 are expressed in the heart during mouse development (Colvin et al., 1999). Since mice with a knockout for FGF-2 undergo normal heart development (Zhou et al., 1998; Schultz et al., 1999) whereas mice with a knockout for FGF-9 show reduced growth of the myocardium (Lavine et al., 2005), we examined epicardial FGF-9 signaling in BWT-1/BWT-1 hearts. FGF-9 expression is enriched in the epicardial cells in the Bflox/Bflox mouse heart (Fig. 4E, arrows), as previously reported. However, this enriched expression of FGF-9 is greatly diminished in BWT-1/BWT-1 epicardial cells (Fig. 4F, arrows). Quantification of the fluorescence intensity shows a 56±7% (mean±s.d., n=5 mice, P<0.001) reduction of FGF-9 expression in the epicardium of the BWT-1/BWT-1 hearts compared to that seen in Bflox/Bflox hearts. These results confirm that epicardial expression of NMIIB is essential in maintaining epicardial integrity and epicardial function in promoting cardiac myocyte proliferation.

NMIIB is important for epicardial function during coronary vessel formation

Whole-mount staining of PECAM1, labeling coronary vessels, showed that the coronary vasculature covered the entire dorsal surface in Bflox/Bflox hearts at E14.5 (Fig. 5A). However, the coronary plexus failed to extend to the apex of BWT-1/BWT-1 hearts (Fig. 5B, ovals). The average coverage of the coronary plexus for Bflox/Bflox hearts is 98±1% and for BWT-1/BWT-1 hearts is 89±5% (mean±s.d., P<0.01, n=4 and n=5 hearts, respectively). Defects in coronary vessel formation are not limited to the surface vasculature. Immunostaining of E15.5 heart sections from BWT-1/BWT-1 and Bflox/Bflox littermates with anti-CD34 antibodies, labeling endothelial cells, shows that the coronary vessels are well developed all over the myocardium in control hearts (Fig. 5C, red, enlarged in D), but significantly fewer coronary vessels are found in BWT-1/BWT-1 hearts (Fig. 5F, red, enlarged in G). Thus, epicardial expression of NMIIB is important for coronary vessel formation. Note that the defects in coronary vessel formation are much less severe in BWT-1/BWT-1 hearts compared to those seen in B/B hearts (Fig. 2). This is not surprising since other cardiac cells such as cardiac myocytes, vascular and endocardial cells also contribute to the normal coronary vessel formation.

Fig. 5.

Epicardial NMIIB knockout impairs coronary vessel formation in BWT-1/BWT-1 mouse hearts. (A,B) Whole-mount PECAM staining of E14.5 Bflox/Bflox (A) and BWT-1/BWT-1 (B) hearts showing defects in coronary vessel growth in BWT-1/BWT-1 hearts as manifested by lack of coronary vascular coverage at the apices of the hearts (B, ovals) compared to in Bflox/Bflox hearts (A). (C,D,F,G) Immunofluorescence confocal images of E15.5 heart sections stained for WT-1 (green, nuclear marker for epicardial cells and EPDCs) and CD34 (red, marker for endothelial cells) showing a marked reduction of WT-1 positive EPDCs in BWT-1/BWT-1 hearts (F, green, enlarged in G, arrows) compared to control littermates (C, green, enlarged in D, arrows). CD34 staining of endothelial cells shows robust development of coronary vessels in the compact myocardium (D, red) in control hearts, but very limited development of coronary vessels in BWT-1/BWT-1 hearts (G, red). Nuclei were stained blue with DAPI. (E,H) Immunofluorescence confocal microscope images of E13.5 mouse hearts stained with antibodies for WT-1 (red, a nuclear marker for epicardial cells and EPDCs) and BrdU (green) showing a marked reduction of EPDCs positive for BrdU in BWT-1/BWT-1 hearts (H, arrow) compared to Bflox/Bflox hearts (E, arrows). Scale bars: 200 µm (C,F); 50 µm (D,E,G,H). (I,J) Quantification of EPDCs in the compact myocardium (arrows in E,H). EPDCs/Total is the percentage of total EPDCs (WT-1 positive) over total cells (DAPI) in the compact myocardium (I). EPDCs/Epi is the the ratio of total EPDCs to total epicardial cells (J). (K) Quantification of WT-1+/BrdU+ EPDCs in the compact myocardium. Results are mean±s.d. (n=3 mice each). P<0.05 (Student's t-test).

Fig. 5.

Epicardial NMIIB knockout impairs coronary vessel formation in BWT-1/BWT-1 mouse hearts. (A,B) Whole-mount PECAM staining of E14.5 Bflox/Bflox (A) and BWT-1/BWT-1 (B) hearts showing defects in coronary vessel growth in BWT-1/BWT-1 hearts as manifested by lack of coronary vascular coverage at the apices of the hearts (B, ovals) compared to in Bflox/Bflox hearts (A). (C,D,F,G) Immunofluorescence confocal images of E15.5 heart sections stained for WT-1 (green, nuclear marker for epicardial cells and EPDCs) and CD34 (red, marker for endothelial cells) showing a marked reduction of WT-1 positive EPDCs in BWT-1/BWT-1 hearts (F, green, enlarged in G, arrows) compared to control littermates (C, green, enlarged in D, arrows). CD34 staining of endothelial cells shows robust development of coronary vessels in the compact myocardium (D, red) in control hearts, but very limited development of coronary vessels in BWT-1/BWT-1 hearts (G, red). Nuclei were stained blue with DAPI. (E,H) Immunofluorescence confocal microscope images of E13.5 mouse hearts stained with antibodies for WT-1 (red, a nuclear marker for epicardial cells and EPDCs) and BrdU (green) showing a marked reduction of EPDCs positive for BrdU in BWT-1/BWT-1 hearts (H, arrow) compared to Bflox/Bflox hearts (E, arrows). Scale bars: 200 µm (C,F); 50 µm (D,E,G,H). (I,J) Quantification of EPDCs in the compact myocardium (arrows in E,H). EPDCs/Total is the percentage of total EPDCs (WT-1 positive) over total cells (DAPI) in the compact myocardium (I). EPDCs/Epi is the the ratio of total EPDCs to total epicardial cells (J). (K) Quantification of WT-1+/BrdU+ EPDCs in the compact myocardium. Results are mean±s.d. (n=3 mice each). P<0.05 (Student's t-test).

To clarify whether knockout of NMIIB in epicardial cells affects subsequent epicardium-derived cell (EPDC) development during mouse heart development, we compared the number of WT-1-positive EPDCs in BWT-1/BWT-1 and control compact myocardium by performing confocal microscopy (Fig. 5C,D,F,G, green). There are markedly fewer WT-1-positive cells in the compact myocardium of BWT-1/BWT-1 hearts (Fig. 5G, arrows) compared to what is seen in Bflox/Bflox hearts (Fig. 5D, arrows) as determined by assessing the percentage of WT-1-positive cells over total cells in the compact myocardium (Fig. 5I, P<0.005, n=3 mice; over 1000 total cells counted for each sample) indicating a reduction of EPDCs. Additionally, BWT-1/BWT-1 hearts have a lower ratio of EPDCs to the number of epicardial cells compared to in Bflox/Bflox hearts (Fig. 5J, P<0.001, n=3 mice). The reduction of EPDCs is associated with decreased proliferation of EPDCs in BWT-1/BWT-1 hearts. As quantified in Fig. 5K, 35.1±8.3% EPDCs are labeled by BrdU in Bflox/Bflox compact myocardium (Fig. 5E, arrows), whereas only 16.9±3.9% of EPDCs are labeled by BrdU in BWT-1/BWT-1 compact myocardium (Fig. 5H, arrow). During embryonic heart development, EPDCs differentiate into pericytes and smooth muscle cells (SMCs) to support coronary vessels. We next examined whether pericytes and SMCs are properly positioned surrounding the coronary vessels in BWT-1/BWT-1 hearts. Both SMCs (Fig. S5A, red, BASM) and pericytes (Fig. S5C, red, NG2) are found surrounding coronary vessels (Fig. S5A,C, green, CD34, arrows) in Bflox/Bflox hearts. Very few pericytes and SMCs are detected surrounding BWT-1/BWT-1 coronary vessels (Fig. S5B,D). Quantification of SMCs in compact myocardium shows that BWT-1/BWT-1 hearts have significantly fewer SMCs (0.132±0.026/µm2) than Bflox/Bflox hearts (0.219±0.019/µm2, P<0.01, mean±s.d., n=4 mice each genotype). Collectively these results indicate that epicardial NMIIB expression is required for EPDC development. Knockout of NMIIB in epicardial cells impairs EPDC development, which contributes to the abnormal coronary vessel formation in BWT-1/BWT-1 and B/B hearts.

NMIIB is required for mesenchymal maturation during epicardial EMT

To further understand the role of NMIIB in EPDC development during epicardial EMT, we employed a three-dimensional collagen gel invasion assay of epicardial explants prepared from E11.5 mouse hearts. Following 3 days of stimulation with 10% fetal bovine serum (FBS) to induce EMT, the explants were stained with phalloidin to show F-actin (Fig. 6A,B, red). In wild-type explants, EPDCs generated following epicardial EMT migrate into the collagen gel (similar to mesenchymal cells) and form vessel-like projections (Fig. 6A). In contrast, B/B explants show a marked reduction of the number of cells migrating into the collagen gel (Fig. 6B), consistent with impaired epicardial EMT. Note that the total number of cells present in the explants is comparable between B+/B+ and B/B explants. The lower part of the panels of Fig. 6A,B are x-z plots of the explants showing cell nuclei stained with DAPI. Again, B/B explants show significantly fewer cells migrating downward into the gel compared to for B+/B+ explants. The average percentages of cells migrating into the gel are 34±1% and 14±3% for B+/B+ and B/B explants, respectively (mean±s.d., n=3 each genotype, P<0.001).

Fig. 6.

Ablation of NMIIB or inhibition of NMII activity impairs epicardial EMT. (A,B) Immunofluorescence confocal images of epicardial explants grown on collagen gels showing long F-actin filaments (phalloidin) invading into the collagen gel in control explants (A, red, 3D projection). In contrast, a markedly reduced in number of much shorter and more irregular F-actin filaments are observed in B/B explants (B, red; 3D projection). At the bottom of the A,B panels are x-z views showing decreased cell numbers migrating down into the gel in B/B hearts compared to the controls. DAPI stains nuclei (white). Scale bar: 200 µm. (C,D) Epicardial explants on the surface of collagen gels, stained for SMA (green, a marker for mesenchymal cells) show decreased SMA staining in B/B hearts (D) compared to in controls (C). Scale bar: 200 µm. (E,F) Confocal images of epicardial explants following 3 day 10% FBS treatment to induce EMT on a collagen surface stained for F-actin (red, phalloidin) showing a mesenchymal phenotype distribution of actin stress fibers in the cells in control explants (E). Many of the cells in B/B explants show actin fibers surrounding the cell periphery, which is similar to what is found in epithelial cells (F, asterisks). Scale bar: 20 µm. (G–I) Confocal images of epicardial explants grown on 3D collagen gels stained for SMA (green, marker for EMT). H and I show that there is a marked decrease in SMA staining in explants treated with blebbistatin (a small molecule inhibitor of NMII ATPase activity, H) or Y27632 (a Rho kinase inhibitor which inhibits MLC20 phosphorylation, I) compared to what is seen in control explants (G). Explants were cultured for 3 days in DMEM with 10% FBS to induce EMT. Nuclei were stained blue with DAPI. Scale bar: 20 µm.

Fig. 6.

Ablation of NMIIB or inhibition of NMII activity impairs epicardial EMT. (A,B) Immunofluorescence confocal images of epicardial explants grown on collagen gels showing long F-actin filaments (phalloidin) invading into the collagen gel in control explants (A, red, 3D projection). In contrast, a markedly reduced in number of much shorter and more irregular F-actin filaments are observed in B/B explants (B, red; 3D projection). At the bottom of the A,B panels are x-z views showing decreased cell numbers migrating down into the gel in B/B hearts compared to the controls. DAPI stains nuclei (white). Scale bar: 200 µm. (C,D) Epicardial explants on the surface of collagen gels, stained for SMA (green, a marker for mesenchymal cells) show decreased SMA staining in B/B hearts (D) compared to in controls (C). Scale bar: 200 µm. (E,F) Confocal images of epicardial explants following 3 day 10% FBS treatment to induce EMT on a collagen surface stained for F-actin (red, phalloidin) showing a mesenchymal phenotype distribution of actin stress fibers in the cells in control explants (E). Many of the cells in B/B explants show actin fibers surrounding the cell periphery, which is similar to what is found in epithelial cells (F, asterisks). Scale bar: 20 µm. (G–I) Confocal images of epicardial explants grown on 3D collagen gels stained for SMA (green, marker for EMT). H and I show that there is a marked decrease in SMA staining in explants treated with blebbistatin (a small molecule inhibitor of NMII ATPase activity, H) or Y27632 (a Rho kinase inhibitor which inhibits MLC20 phosphorylation, I) compared to what is seen in control explants (G). Explants were cultured for 3 days in DMEM with 10% FBS to induce EMT. Nuclei were stained blue with DAPI. Scale bar: 20 µm.

The function of NMIIB in regulating cell migration is well documented (Vicente-Manzanares et al., 2009). Defects in B/B epicardial EMT could simply be the result of a slowed cell migration and/or a decreased EPDC proliferation. To test whether ablation of NMIIB impairs the epicardial EMT program, we first examined the expression of smooth muscle α-actin (SMA, a mesenchymal marker for EMT) in explants by undertaking immunofluorescence confocal microscopy. B/B epicardial explants show reduced expression of SMA (Fig. 6D, green) compared to that seen in wild-type littermates (Fig. 6C, green). Phalloidin staining of F-actin shows that most of the epicardial cells from wild-type explants develop extensive actin stress fibers throughout the cell (Fig. 6E, red) manifesting a mesenchymal phenotype. Many B/B cells, however, show cortical stress fibers that are similar to those in cobblestone-like epithelial cells (Fig. 6F, asterisks). These results are consistent with defects in B/B epicardial EMT. To confirm the defect in epicardial EMT and dissect the molecular mechanism of the defect, we examined the expression of genes known to be important for EMT as well as the expression of genes encoding epithelial and mesenchymal markers in wild-type and B/B epicardial explants by performing quantitative real-time PCR (qRT-PCR). Compared to the wild-type explants, B/B epicardial explants show significantly reduced expression of genes for mesenchymal markers including Cdh6 (K-cadherin), Acta2 (smooth muscle α-actin, SMA), Tagln (SM22α) and Postn (Table 1, P<0.05, n=5). These results again support abnormal epicardial EMT in B/B epicardial explants. However, no statistical difference in the expression of the important transcription factors for EMT such as Snail and Slug (also known as Snai2) is observed in B/B epicardial explants (Table 1). There are also no differences in the expression of genes for epithelial markers such as Cdh1 (E-cadherin), Krt14 and Tcf21. Table S1 is a list of the genes which are thought to be important for EMT that were tested in our study and were unchanged with the exception of those noted in Table 1.

Table 1.

qRT-PCR comparison of genes related to epicardial EMT between B/B and B+/B+ epicardial explants

qRT-PCR comparison of genes related to epicardial EMT between B−/B− and B+/B+ epicardial explants
qRT-PCR comparison of genes related to epicardial EMT between B−/B− and B+/B+ epicardial explants

To further test the importance of NMII function in mesenchymal gene expression during epicardial EMT, we prepared epicardial explants from wild-type mouse hearts, and induced EMT in the presence or absence of the NMII inhibitor blebbistatin or the Rho kinase inhibitor Y27632. Both inhibitors block epicardial EMT, as with knockout of NMIIB, as manifested by a significant reduction in SMA expression (Fig. 6H,I, green) compared to that seen in the control (Fig. 6G, green). qRT-PCR analysis confirms marked reductions in the expression of the mesenchymal genes Acta2 (0.65-fold), Cdh6 (0.51-fold), Tagln (0.33-fold) and Postn (0.27-fold) with blebbistatin treatment compared to that in the control, demonstrating a requirement for active NMII during EMT.

To understand which signaling pathways might be affected when NMII activity is inhibited by blebbistatin during epicardial EMT, we performed an expression profile study. RNAseq analysis showed a major reduction in the expression of mesenchymal genes, including those identified by the qRT-PCR above, such as Acta2, Postn and Tagln (Table S2). Again Snail and Slug were not found to be differentially expressed. Table S2 is a list of 21 genes showing a reduction of multimodal fold change greater than two. A total of 15 genes from the list have been reported either to promote EMT or to be upregulated during EMT, and two of the genes have been shown to suppress EMT. This result strongly suggests that inhibition of NMII activity impairs epicardial EMT. An interpretative phenomenological analysis (IPA) signaling pathway analysis identified the top three downregulated downstream signaling pathways as being those mediated by FAK1 (also known as PTK2), VEGF and TGFβ. These findings are consistent with the phenotypes observed in our NMIIB-knockout mice, which include epicaridial detachment from the myocardium, impaired coronary vessel formation and defects in epicardial EMT.

In addition to actin filament reorganization, intermediate filament reorganization is also important during EMT. An increased expression of the intermediate filament vimentin is another indication of mesenchymal acquisition during EMT (Morabito et al., 2001; Vuoriluoto et al., 2011). The defect in mesenchymal gene expression is further demonstrated in the BWT-1/BWT-1 compact myocardium in vivo as evaluated by vimentin expression. Vimentin is expressed in cardiac non-myocytes including epicardial, mesenchymal and endocardial cells, but not in cardiac myocytes (Fig. S6A,E, green). Vimentin expression is markedly reduced in mesenchymal cells in the BWT-1/BWT-1 compact myocardium (Fig. S6C,G, CM, green) compared to that seen in the Bflox/Bflox compact myocardium (Fig. S6A,E, CM, green). No difference in vimentin expression is observed in endocardial cells between BWT-1/BWT-1 and Bflox/Bflox hearts. Fig. S6I shows the quantification of the staining intensity. Taken together, these results indicate that knockout of epicardial NMIIB does not affect the initiation of epicardial EMT, but impairs EMT progression by blocking mesenchymal gene expression.

NMII-mediated regulation of actin filament formation is important for mesenchymal maturation in epicardial EMT

EMT features extensive actin cytoskeletal reorganization, and NMII plays a key role regulating cytoskeletal structures. Corresponding to the reduced mesenchymal gene expression, B/B epicardial explants show a less mature mesenchymal phenotype compared to wild-type explants, as evaluated by focal adhesion and F-actin formation. As shown in Fig. 7A,B, B/B explant epicardial cells cultured on a gelatin-coated glass surface form significantly smaller and thinner focal adhesion complexes compared to those in wild-type cells, as demonstrated by vinculin staining (Fig. 7A,B, green, arrows). The average size of the mature focal adhesions is 5.91±1.13 µm2 and 4.19±1.21 µm2 for B+/B+ and B/B cells, respectively (mean±s.d., P<0.05, n=3 each genotype). There is also a marked reduction in the thickness of actin stress fibers anchoring the focal adhesion in B/B cells (Fig. 7B, red) compared to those in wild-type cells (Fig. 7A, red). The average thicknesses of stress fibers are 0.89±0.03 µm and 0.54±0.03 µm for wild-type and B/B cells, respectively (n=3 each genotype, P<0.001).

Fig. 7.

Altering actin cytoskeletal dynamics impairs epicardial EMT. (A,B) Confocal images of epicardial explants on a gelatin-coated glass surface stained for vinculin (green, marker for focal adhesion) and F-actin (red, phalloidin) showing that epicardial explants from B/B hearts develop smaller focal adhesions with thinner actin filaments. Nuclei were stained blue with DAPI. Arrows indicate focal adhesions. Scale bar: 20 µm. (C–F) Confocal images of epicardial explants on collagen gels stained for SMA (green, a marker for mesenchymal cells) and F-actin (red, phalloidin) show a marked reduction of SMA staining in epicardial explants treated with either the F-actin disrupting agent, latrunculin-A (LatA, D, green) or the F-actin stabilizer, jasplakinolide (JASP; F, green) compared to the control explants treated with only DMSO (C,E, green). Decreased F-actin staining (D, red) and an increased F-actin staining (F, red) confirm the effects of latrunculin-A and jasplakinolide on F-actin, when compared to what is seen in the explants without treatment (C,E, red). Note that the difference in F-actin staining between controls (C,E, red) is due to the difference in gain settings during confocal imaging. The saturation level of red signal in C and D is set based on sample in C; the saturation level in panel E and F is set based on sample in F. Nuclei were stained blue with DAPI. Scale bar: 200 µm.

Fig. 7.

Altering actin cytoskeletal dynamics impairs epicardial EMT. (A,B) Confocal images of epicardial explants on a gelatin-coated glass surface stained for vinculin (green, marker for focal adhesion) and F-actin (red, phalloidin) showing that epicardial explants from B/B hearts develop smaller focal adhesions with thinner actin filaments. Nuclei were stained blue with DAPI. Arrows indicate focal adhesions. Scale bar: 20 µm. (C–F) Confocal images of epicardial explants on collagen gels stained for SMA (green, a marker for mesenchymal cells) and F-actin (red, phalloidin) show a marked reduction of SMA staining in epicardial explants treated with either the F-actin disrupting agent, latrunculin-A (LatA, D, green) or the F-actin stabilizer, jasplakinolide (JASP; F, green) compared to the control explants treated with only DMSO (C,E, green). Decreased F-actin staining (D, red) and an increased F-actin staining (F, red) confirm the effects of latrunculin-A and jasplakinolide on F-actin, when compared to what is seen in the explants without treatment (C,E, red). Note that the difference in F-actin staining between controls (C,E, red) is due to the difference in gain settings during confocal imaging. The saturation level of red signal in C and D is set based on sample in C; the saturation level in panel E and F is set based on sample in F. Nuclei were stained blue with DAPI. Scale bar: 200 µm.

We thus hypothesized that NMIIB regulation of the actin cytoskeleton plays an active role during epicardial EMT. To test this hypothesis, we first examined whether NMIIB is actively reorganized in explants during epicardial EMT, in a similar manner to what is seen for actin. As shown in Fig. S7, NMIIB forms no obvious filaments and is uniformly distributed in epicardial cells before EMT (Fig. S7A, green), while F-actin is largely concentrated at the cell–cell boundaries similar to in epithelial cells (Fig. S7C, red). At 2 h after stimulation with TGFβ to induce EMT, NMIIB filaments markedly increase (Fig. S7B, green) as do F-actin filaments (Fig. S7D, red) throughout the cells. We then tested whether altering actin filaments could affect epicardial EMT. Epicardial explants were treated with latrunculin-A to disassemble actin filaments or jasplakinolide to stabilize actin filaments. As seen in Fig. 7C–F, applying either latrunculin-A or jasplakinolide to explants inhibits epicardial EMT, as shown by a marked reduction of the numbers of cells positive for SMA (a marker for mesenchymal cells) and migrating from explants (Fig. 7D,F, green) compared to what is seen for the control explants (Fig. 7C,E). The average number of total SMA-positive cells migrating from explants for the entire panel are: 598±151 (DMSO, control), 249±71 (10 nM latrunculin-A, P<0.05 vs control, n=3), and 161±35 (10 nM jasplakinolide, P<0.005 vs control, n=3). qRT-PCR analysis shows a reduction in expression for mesenchymal genes including Postn (0.7-fold), Tagln (0.34-fold) and Vim (0.68-fold) in explants treated with latrunculin, but not in explants treated with jasplakinolide. Thus, the assembly of actin filaments is important for mesenchymal maturation during epicardial EMT. These results show how NMII-mediated regulation of the cytoskeletal structure can directly affect epicardial EMT.

Loss of the cytoskeletal motor protein NMIIB in epicardial cells impairs epicardial integrity resulting in a compromised epicardial barrier function and attachment of epicardial cells to the myocardium. Functional consequences of impaired epicardial formation include: (1) a failure of myocardial cell proliferation contributing to a hypoplastic myocardium; and (2) an impairment of EPDC proliferation and maturation during epicardial EMT leading to defects in coronary vessel formation. In this study, we provide evidence indicating a novel function for NMIIB in regulating mesenchymal maturation during epicardial EMT through coordinating actomyosin cytoskeletal dynamics.

NMIIB function in myocardial growth

NMIIB is the major NMII isoform expressed in cardiac myocytes throughout embryonic heart development (Ma and Adelstein, 2014b), although the expression of NMIIB is markedly reduced in cardiac myocytes in the first week after birth. However, there is no increase in NMIIA or NMIIC, the former being absent and the later remaining low after birth. Moreover, Nkx2.5-Cre-mediated knockout of NMIIA in cardiac myocytes shows no effect on mouse heart development (Conti et al., 2015) nor does knockout of NMIIC (Ma et al., 2010). The temporal pattern of NMIIB expression in developing cardiac myocytes is consistent with its role in regulating cardiac myocyte proliferation. Previous findings from B/B and NMIIB-R709C mutant (BR709C/BR709C) mouse hearts demonstrated a major role for NMIIB in driving actomyosin contractile ring constriction during cardiac myocyte cell division by exerting tension on actin filaments (Ma et al., 2012). Knockout of NMIIB, but not NMIIA or NMIIC, in cardiac myocytes results in a failure of cardiac myocyte cytokinesis, causing premature bi-nucleation and a reduction of cardiac myocytes in B/B hearts (Takeda et al., 2003). Although NMIIB is also expressed in cardiac non-myocytes (epicardial, endocardial and interstitial fibroblast cells), knockout of NMIIB does not affect cytokinesis in these cells. This is most likely due to the function of NMIIA, which is also present. Of note, expression of NMIIA alone in epicardial cells is sufficient for epicardial cell cytokinesis, but it cannot support the normal formation of a functional epicardium. Epicardial cells require NMIIB to maintain epicardial integrity and epicardial function in order to promote cardiac myocyte proliferation. Epicardium-specific ablation of NMIIB in mouse hearts resulted in decreased cardiac myocyte proliferation, which consequently led to a hypoplastic myocardium. It is likely that the compromised epicardial function and physical separation of epicardium from myocardium together lead to the loss of proliferative signaling for cardiac myocytes. The specific requirement for NMIIB function in epicardium formation is surprising, since epithelial integrity generally requires functional NMIIA (Ivanov et al., 2007; Naydenov et al., 2016; Smutny et al., 2010). Thus, NMIIB regulates cardiac myocyte development directly and indirectly: (1) by directly generating tension during cardiac myocyte division to drive contractile ring constriction, and (2) indirectly in forming a functional epicardium capable of stimulating cardiac myocyte proliferation.

NMIIB is required for epicardium-derived cell development during the EMT program

The dynamics underlying changes in NMII reorganization are just beginning to be understood (Beach et al., 2017; Hu et al., 2017). EMT, the process by which static, well-ordered and apical-basal polarized epithelial cells transform into dynamically migrating mesenchymal cells associated with extensive reorganization of the actomyosin cytoskeleton is of major interest and is important for cardiac development. One of the novel findings from this study is the role of NMIIB in regulating EPDC proliferation and maturation following epicardial EMT. Following EMT, EPDCs migrate, proliferate and differentiate into pericytes, smooth muscle cells and cardiac fibroblasts to support coronary vessel formation. In this study, we provide evidence that knockout of NMIIB specifically in mouse epicardial cells results in a marked reduction of EPDC proliferation, resulting in fewer EPDCs being found in the compact myocardium. It is likely that EPDCs also receive proliferative signals from the epicardium, similar to for the cardiac myocytes as discussed above. Defects in epicardial function and attachment of the epicardium to the myocardium may block the epicardial signaling for EDPC proliferation in BWT-1/BWT-1 hearts.

Although NMIIB knockout affects neither activation of canonical Snail/E-cadherin signaling nor the decreased expression of epithelial genes during epicardial EMT, the increased expression of mesenchymal genes such as Acta2 (smooth muscle actin) and Tagln (SM22α) is markedly impaired during epicardial EMT. While the defects in NMIIB-knockout epicardium during EMT are not secondary to the insufficiency of canonical Snail/E-cadherin signaling, NMIIB is required for full acquisition of the mesenchymal phenotype by EPDCs following epicardial EMT. Our finding of decreased periostin expression during B/B epicardial EMT is also consistent with a lack of EPDC maturation, since mice with a knockout for periostin show a defect in mesenchymal differentiation of the cardiac cushions into fibroblast tissue during cardiac valve formation (Snider et al., 2008). This is also consistent with a previous report showing a NMIIC to NMIIB isoform switch, and a reduced mesenchymal cell invasion following siRNA knockdown of NMIIB during EMT in mouse mammary epithelial cells (NMuMG) in response to TGFβ (Beach et al., 2011). Recently, NMIIB was shown to generate tension for nuclear translocation during migration in 3D collagen gels (Thomas et al., 2015).

EMT features an extensive reorganization of cytoskeletal structures, including formation of intracellular actin stress-fibers and formation of focal adhesion complexes in mesenchymal cells, which are required for cell morphological changes and cell migration during EMT (Lamouille et al., 2014). Increased expression of the ERM (ezrin/radixin/moesin) protein moesin also contributes to actin cytoskeleton remodeling and morphological changes during EMT (Haynes et al., 2011). It was previously unclear whether actin cytoskeleton remodeling is required for EMT. Shankar et al. were among the first to propose that actin dynamics controls EMT in metastatic cancer cells. This hypothesis was based on their studies with cultured metastatic cancer cell lines showing that knockdown of pseudopodia-enriched proteins (Shankar et al., 2010) or cytochalasin D-induced disruption of actin filaments (Shankar and Nabi, 2015) decreased pseudopod formation and tumor cell invasion. This was associated with reduced actin filament stability, the loss of N-cadherin and vimentin expression, and the regaining of E-cadherin expression. These findings suggest that actin stability plays a role in maintaining the mesenchymal properties of metastatic cancer cells in culture. The results presented here provide direct evidence showing that NMIIB-mediated actin filament formation is actively involved in maturation of the mesenchymal phenotype during epicardial EMT in explants and in vivo during mouse heart development. NMII regulates both actin stress fiber formation and focal adhesion formation (le Duc et al., 2010; Pasapera et al., 2010). Activation of NMII promotes actin stress fiber formation, while NMII tension on actin stress fibers promotes maturation of nascent focal adhesion complexes. Moreover, we demonstrate that ablation of NMIIB or inhibition of NMIIB activity impairs mesenchymal gene expression during epicardial EMT associated with defects in actin stress fiber formation and focal adhesion maturation. Since interference with actin dynamics directly by destabilizing actin stress fibers blocks mesenchymal phenotype maturation during epicardial EMT, we suggest that NMIIB regulates epicardial EMT in vivo and in vitro by altering actin dynamics.

Taken together, our results support the idea that epicardial EMT may be divided into two steps: one that does not rely on NMIIB, including the initial induction of EMT Snail signaling and loss of the epithelial phenotype, and a second step that requires NMIIB activity to acquire the mesenchymal phenotype, such as changes in cell polarity and cytoskeletal reorganization that is needed for cell migration, and an increase in mesenchymal gene expression. While these two programs may not be completely independent of each other, both are ultimately necessary for completion of epicardial EMT.

Animals

B/B, Bflox/Bflox, and BGFP/BGFP mice were generated as previously described (Bao et al., 2007; Ma et al., 2009; Tullio et al., 1997) and are available through the Mutant Mouse Regional Resource Centers (MMRRC, #16991and #37053). WT-1 Cre mice were generously provided by Dr William Pu (Boston Children's Hospital, Boston, MA). All procedures were conducted using an approved animal protocol (H0053R3) in accordance with National Heart, Lung, and Blood Institute Animal Care and Use Committee guidelines.

Histology and immunofluorescence staining

The mouse embryos or hearts were collected in PBS and directly immersed in 4% paraformaldehyde (PFA) in PBS (pH 7.4) overnight. Paraffin sections at a thickness of 5 µm were prepared by Histoserv, Inc. (Germantown, MD). Primary antibodies for immunostaining were incubated at 4°C overnight following antigen retrieval in 10 mM citrate buffer (pH 6). The following primary antibodies were used in this study: polyclonal antibodies against pMLC20 (1:400, Cell Signaling Technology), NG2 (1:400, Millipore Sigma), NMHCIIB (1:3000, Covance), vimentin Alexa Fluor 488 conjugate (1:200, Cell Signaling Technology), bovine aorta smooth muscle myosin heavy chain (BASM, 1:200; Kelley and Adelstein, 1990) and WT-1 (1:100, Neomarker); and monoclonal antibodies against CD34 (1:500, GeneTex), E-cadherin (1:500, BD Biosciences, San Jose, CA), desmin (1:100, DakoCytomation, Denmark), FGF-9 (1:100, Santa Cruz Biotechnology, Santa Cruz Biotechnology, CA), β1 integrin (1:300, BD Transduction), N-cadherin (1:200, Invitrogen), NMHCIIA (1:300, Abcam), PECAM1 (CD31, 1:200, BD Pharmingen), smooth muscle α-actin (1:2000, Sigma), vinculin (1:1000, Sigma) and ZO-1 (1:200, Invitrogen). Fluorescence secondary antibodies used were: Alexa Fluor 488-conjugated goat anti-rabbit-IgG or Alexa Fluor 594-conjugated goat anti-mouse-IgG (1:250, Invitrogen, Carlsbad, CA). The slides were mounted with Prolong Gold antifade mounting medium (Invitrogen) and observed under a confocal microscope. The confocal images were collected using a Zeiss LSM 510-META. In all cases, when possible, comparison was made among littermates. For each genotype, we analyzed at least three to five mice.

Biotin epicardial permeability assay

EZ-Link Sulfo-NHS-LC-Biotin (Thermo Scientific) solution was freshly prepared at 10 mg/ml in 1× PBS. Mouse embryos were dissected and placed in PBS. An opening in the chest wall, which broke the pericardium, was made using forceps, and 10 µl of biotin was pipetted into the opening and allowed to perfuse for 15 min. Embryos were then fixed in 4% PFA overnight, paraffin embedded and sectioned. Following deparaffinization, antigen retrieval, and blocking in 10% goat serum, sections were incubated with Alexa Fluor 488-conjugated vimentin (1:200, Cell Signaling Technology), to visualize cardiac non-myocytes, and Rhodamine-conjugated streptavidin, to detect biotin, for 30 min prior to confocal imaging.

Epicardial explants

3D epicardial explants were prepared from E11.5 mouse hearts with the atria and outflow tract removed. The hearts were placed on a 1% collagen gel (prepared from rat tail collagen I in MEM) with the dorsal surface against the gel. The epicardial cells were allowed to grow out over the gel for 2 days with only one drop of Dulbecco's modified Eagle's medium (DMEM) over the hearts. The hearts were then removed, and the explants were cultured for another 3 days in 10% FBS in DMEM with pencillin-streptomycin to induce EMT. To study the effect of various chemicals, as indicated in the text, on epicardial explants, the explants were prepared from E11.5 wild-type (C57BL6, Jackson Lab) mouse hearts and then the explants were randomly divided into two groups. Following removal of hearts from explants at day 2, each group of explants was further cultured in medium with various chemicals or the solvent (DMSO) as control. 2D epicardial explants were prepared by placing E11.5 hearts on gelatin-coated coverslips with 10% FBS in DMEM with pencillin-streptomycin for 2 days and then removing the hearts. The explants were grown for 2 additional days. For confocal analysis, the explants were fixed in 4% PFA, permeabilized with Triton X-100 and then incubated with various antibodies of interest.

RNAseq and qRT-PCR Analysis

RNA was extracted from 3D epicardial explants after 72 h incubation with 10% FBS in DMEM using the RNeasy Mini plus Kit (Qiagen) and quantified by Qubit Fluorometer. RNA-seq libraries were constructed using Ovation RNA-Seq System V2 kit (NuGen, Inc., San Carlos, CA) and Nextera XT DNA library preparation kit (Illumina, Inc., San Diego, CA) by following the manufacturer's instructions. After the final amplification step, the PCR products were separated on 2% agarose gel to excise 250–450 bp fragments. The resulting barcoded RNA-seq libraries were then pooled and subjected to 2×50 bp paired-end sequencing using the Illumina HiSeq3000 platform (Illumina, San Diego, CA, USA). Raw sequencing data were demultiplexed and converted into a FASTQ format. Real-time PCR was performed using an ABI 7500 real-time PCR instrument following the manufacturer's instruction. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression level was used as a control for normalization.

Quantification and statistical analyses

The BrdU labeling index of cardiac myocytes was calculated as a percentage of the number of BrdU- and desmin-positive cells over the total number of desmin-positive cells. Embryonic heart sections were stained with antibodies against BrdU and desmin. Nuclei were stained by DAPI. Confocal images were captured with a 40× objective. The images were analyzed with IDL Software (programmed by Christian A. Combs, NHLBI). The index was scored from three different embryos per each genotype. For each mouse, more than 1000 total cells of the compact myocardium were counted.

The percentage of epicardial cells that dissociated from myocardium over the total epicardial cells was calculated from H&E images with ImageJ software. Hearts from three mice per genotype were quantified and more than 1000 total epicardial cells were counted per mouse.

Epicardial explants were quantified from three experiments per genotype (or treatment) with ImageJ software. For the 3D collagen gel migration assay, ∼200 total cells were counted for each explant. The focal adhesion size and actin filament thickness were calculated as an average from three different explants per genotype; 10 cells were measured for each explant. On average 15 to 20 focal adhesions and actin-filaments were measured for each cell.

Data are expressed as mean±s.d. The Student's t-test was performed to compare two means.

We thank Dr Mary Anne Conti for her significant contributions to this manuscript. Dr Sachiyo Kawamoto and members of the Laboratory of Molecular Cardiology also provided critical comments on the manuscript. Special thanks for Dr Keekwang Kim for his help with qRT-PCR analysis. We also thank Drs Chengyu Liu and Yubin Du [National Heart, Lung, and Blood Institute (NHLBI) Transgenic Core], Drs Christian A. Combs and Daniela Malide (NHLBI Light Microscopy Core). Antoine Smith and Dalton Saunders provided technical assistance.

Author contributions

Conceptualization: X.M., R.S.A.; Methodology: X.M.; Formal analysis: X.M., Y.Y.; Investigation: R.S.A.; Data curation: X.M., D.C.S., Y.W.; Writing - original draft: X.M., D.C.S., R.S.A.; Writing - review & editing: X.M., D.C.S., R.S.A.; Supervision: R.S.A.; Funding acquisition: R.S.A.

Funding

This research was supported by the National Institutes of Health, National Heart, Lung, and Blood Institute (HL-004228). Deposited in PMC for release after 12 months.

Data availability

RNAseq data reported in this paper has been deposited in the Gene Expression Omnibus under (GEO) accession number GSE101701 (https:////www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE101701).

Artamonov
,
M. V.
,
Jin
,
L.
,
Franke
,
A. S.
,
Momotani
,
K.
,
Ho
,
R.
,
Dong
,
X. R.
,
Majesky
,
M. W.
and
Somlyo
,
A. V.
(
2015
).
Signaling pathways that control rho kinase activity maintain the embryonic epicardial progenitor state
.
J. Biol. Chem.
290
,
10353
-
10367
.
Bao
,
J.
,
Ma
,
X.
,
Liu
,
C.
and
Adelstein
,
R. S.
(
2007
).
Replacement of nonmuscle myosin II-B with II-A rescues brain but not cardiac defects in mice
.
J. Biol. Chem.
282
,
22102
-
22111
.
Beach
,
J. R.
,
Hussey
,
G. S.
,
Miller
,
T. E.
,
Chaudhury
,
A.
,
Patel
,
P.
,
Monslow
,
J.
,
Zheng
,
Q.
,
Keri
,
R. A.
,
Reizes
,
O.
,
Bresnick
,
A. R.
, et al. 
(
2011
).
Myosin II isoform switching mediates invasiveness after TGF-beta-induced epithelial-mesenchymal transition
.
Proc. Natl. Acad. Sci. USA
108
,
17991
-
17996
.
Beach
,
J. R.
,
Bruun
,
K. S.
,
Shao
,
L.
,
Li
,
D.
,
Swider
,
Z.
,
Remmert
,
K.
,
Zhang
,
Y.
,
Conti
,
M. A.
,
Adelstein
,
R. S.
,
Rusan
,
N. M.
, et al. 
(
2017
).
Actin dynamics and competition for myosin monomer govern the sequential amplification of myosin filaments
.
Nat. Cell Biol.
19
,
85
-
93
.
Berg
,
J. S.
,
Powell
,
B. C.
and
Cheney
,
R. E.
(
2001
).
A millennial myosin census
.
Mol. Biol. Cell
12
,
780
-
794
.
Berndt
,
J. D.
,
Clay
,
M. R.
,
Langenberg
,
T.
and
Halloran
,
M. C.
(
2008
).
Rho-kinase and myosin II affect dynamic neural crest cell behaviors during epithelial to mesenchymal transition in vivo
.
Dev. Biol.
324
,
236
-
244
.
Brade
,
T.
,
Pane
,
L. S.
,
Moretti
,
A.
,
Chien
,
K. R.
and
Laugwitz
,
K.-L.
(
2013
).
Embryonic heart progenitors and cardiogenesis
.
Cold Spring Harb. Perspect. Med.
3
,
a013847
.
Buxboim
,
A.
,
Swift
,
J.
,
Irianto
,
J.
,
Spinler
,
K. R.
,
Dingal
,
P. C. D. P.
,
Athirasala
,
A.
,
Kao
,
Y.-R. C.
,
Cho
,
S.
,
Harada
,
T.
,
Shin
,
J.-W.
, et al. 
(
2014
).
Matrix elasticity regulates lamin-A,C phosphorylation and turnover with feedback to actomyosin
.
Curr. Biol.
24
,
1909
-
1917
.
Colvin
,
J. S.
,
Feldman
,
B.
,
Nadeau
,
J. H.
,
Goldfarb
,
M.
and
Ornitz
,
D. M.
(
1999
).
Genomic organization and embryonic expression of the mouse fibroblast growth factor 9 gene
.
Dev. Dyn.
216
,
72
-
88
.
Conti
,
M. A.
,
Saleh
,
A. D.
,
Brinster
,
L. R.
,
Cheng
,
H.
,
Chen
,
Z.
,
Cornelius
,
S.
,
Liu
,
C.
,
Ma
,
X.
,
Van Waes
,
C.
and
Adelstein
,
R. S.
(
2015
).
Conditional deletion of nonmuscle myosin II-A in mouse tongue epithelium results in squamous cell carcinoma
.
Sci. Rep.
5
,
14068
.
Dokic
,
D.
and
Dettman
,
R. W.
(
2006
).
VCAM-1 inhibits TGFbeta stimulated epithelial-mesenchymal transformation by modulating Rho activity and stabilizing intercellular adhesion in epicardial mesothelial cells
.
Dev. Biol.
299
,
489
-
504
.
Ebrahim
,
S.
,
Fujita
,
T.
,
Millis
,
B. A.
,
Kozin
,
E.
,
Ma
,
X.
,
Kawamoto
,
S.
,
Baird
,
M. A.
,
Davidson
,
M.
,
Yonemura
,
S.
,
Hisa
,
Y.
, et al. 
(
2013
).
NMII forms a contractile transcellular sarcomeric network to regulate apical cell junctions and tissue geometry
.
Curr. Biol.
23
,
731
-
736
.
Engler
,
A. J.
,
Sen
,
S.
,
Sweeney
,
H. L.
and
Discher
,
D. E.
(
2006
).
Matrix elasticity directs stem cell lineage specification
.
Cell
126
,
677
-
689
.
Golomb
,
E.
,
Ma
,
X.
,
Jana
,
S. S.
,
Preston
,
Y. A.
,
Kawamoto
,
S.
,
Shoham
,
N. G.
,
Goldin
,
E.
,
Conti
,
M. A.
,
Sellers
,
J. R.
and
Adelstein
,
R. S.
(
2004
).
Identification and characterization of nonmuscle myosin II-C, a new member of the myosin II family
.
J. Biol. Chem.
279
,
2800
-
2808
.
Haynes
,
J.
,
Srivastava
,
J.
,
Madson
,
N.
,
Wittmann
,
T.
and
Barber
,
D. L.
(
2011
).
Dynamic actin remodeling during epithelial-mesenchymal transition depends on increased moesin expression
.
Mol. Biol. Cell
22
,
4750
-
4764
.
Heissler
,
S. M.
and
Manstein
,
D. J.
(
2013
).
Nonmuscle myosin-2: mix and match
.
Cell Mol. Life Sci.
70
,
1
-
21
.
Hu
,
S.
,
Dasbiswas
,
K.
,
Guo
,
Z.
,
Tee
,
Y.-H.
,
Thiagarajan
,
V.
,
Hersen
,
P.
,
Chew
,
T.-L.
,
Safran
,
S. A.
,
Zaidel-Bar
,
R.
and
Bershadsky
,
A. D.
(
2017
).
Long-range self-organization of cytoskeletal myosin II filament stacks
.
Nat. Cell Biol.
19
,
133
-
141
.
Ivanov
,
A. I.
,
Bachar
,
M.
,
Babbin
,
B. A.
,
Adelstein
,
R. S.
,
Nusrat
,
A.
and
Parkos
,
C. A.
(
2007
).
A unique role for nonmuscle myosin heavy chain IIA in regulation of epithelial apical junctions
.
PLoS ONE
2
,
e658
.
Kelley
,
C. A.
and
Adelstein
,
R. S.
(
1990
).
The 204-kDa smooth muscle myosin heavy chain is phosphorylated in intact cells by casein kinase II on a serine near the carboxyl terminus
.
J. Biol. Chem.
265
,
17876
-
17882
.
Kim
,
K.
,
Ossipova
,
O.
and
Sokol
,
S. Y.
(
2015
).
Neural crest specification by inhibition of the ROCK/Myosin II pathway
.
Stem Cells
33
,
674
-
685
.
Lamouille
,
S.
,
Xu
,
J.
and
Derynck
,
R.
(
2014
).
Molecular mechanisms of epithelial-mesenchymal transition
.
Nat. Rev. Mol. Cell Biol.
15
,
178
-
196
.
Lavine
,
K. J.
,
Yu
,
K.
,
White
,
A. C.
,
Zhang
,
X.
,
Smith
,
C.
,
Partanen
,
J.
and
Ornitz
,
D. M.
(
2005
).
Endocardial and epicardial derived FGF signals regulate myocardial proliferation and differentiation in vivo
.
Dev Cell
8
,
85
-
95
.
le Duc
,
Q.
,
Shi
,
Q.
,
Blonk
,
I.
,
Sonnenberg
,
A.
,
Wang
,
N.
,
Leckband
,
D.
and
De Rooij
,
J.
(
2010
).
Vinculin potentiates E-cadherin mechanosensing and is recruited to actin-anchored sites within adherens junctions in a myosin II-dependent manner
.
J. Cell Biol.
189
,
1107
-
1115
.
Lu
,
J.
,
Landerholm
,
T. E.
,
Wei
,
J. S.
,
Dong
,
X.-R.
,
Wu
,
S.-P.
,
Liu
,
X.
,
Nagata
,
K.-i
,
Inagaki
,
M.
and
Majesky
,
M. W.
(
2001
).
Coronary smooth muscle differentiation from proepicardial cells requires rhoA-mediated actin reorganization and p160 rho-kinase activity
.
Dev. Biol.
240
,
404
-
418
.
Luo
,
Y.
,
High
,
F. A.
,
Epstein
,
J. A.
and
Radice
,
G. L.
(
2006
).
N-cadherin is required for neural crest remodeling of the cardiac outflow tract
.
Dev. Biol.
299
,
517
-
528
.
Ma
,
X.
and
Adelstein
,
R. S.
(
2012
).
In vivo studies on nonmuscle myosin II expression and function in heart development
.
Front Biosci. (Landmark Ed)
17
,
545
-
555
.
Ma
,
X.
and
Adelstein
,
R. S.
(
2014a
).
A point mutation in Myh10 causes major defects in heart development and body wall closure
.
Circ. Cardiovasc. Genet.
7
,
257
-
265
.
Ma
,
X.
and
Adelstein
,
R. S.
(
2014b
).
The role of vertebrate nonmuscle Myosin II in development and human disease
.
Bioarchitecture
4
,
88
-
102
.
Ma
,
X.
,
Takeda
,
K.
,
Singh
,
A.
,
Yu
,
Z.-X.
,
Zerfas
,
P.
,
Blount
,
A.
,
Liu
,
C.
,
Towbin
,
J. A.
,
Schneider
,
M. D.
,
Adelstein
,
R. S.
, et al. 
(
2009
).
Conditional ablation of nonmuscle myosin II-B delineates heart defects in adult mice
.
Circ. Res.
105
,
1102
-
1109
.
Ma
,
X.
,
Jana
,
S. S.
,
Conti
,
M. A.
,
Kawamoto
,
S.
,
Claycomb
,
W. C.
and
Adelstein
,
R. S.
(
2010
).
Ablation of nonmuscle myosin II-B and II-C reveals a role for nonmuscle myosin II in cardiac myocyte karyokinesis
.
Mol. Biol. Cell
21
,
3952
-
3962
.
Ma
,
X.
,
Kovacs
,
M.
,
Conti
,
M. A.
,
Wang
,
A.
,
Zhang
,
Y.
,
Sellers
,
J. R.
and
Adelstein
,
R. S.
(
2012
).
Nonmuscle myosin II exerts tension but does not translocate actin in vertebrate cytokinesis
.
Proc. Natl. Acad. Sci. USA
109
,
4509
-
4514
.
Morabito
,
C. J.
,
Dettman
,
R. W.
,
Kattan
,
J.
,
Collier
,
J. M.
and
Bristow
,
J.
(
2001
).
Positive and negative regulation of epicardial-mesenchymal transformation during avian heart development
.
Dev. Biol.
234
,
204
-
215
.
Naydenov
,
N. G.
,
Feygin
,
A.
,
Wang
,
D.
,
Kuemmerle
,
J. F.
,
Harris
,
G.
,
Conti
,
M. A.
,
Adelstein
,
R. S.
and
Ivanov
,
A. I.
(
2016
).
Nonmuscle myosin IIA regulates intestinal epithelial barrier in vivo and plays a protective role during experimental colitis
.
Sci. Rep.
6
,
24161
.
Pasapera
,
A. M.
,
Schneider
,
I. C.
,
Rericha
,
E.
,
Schlaepfer
,
D. D.
and
Waterman
,
C. M.
(
2010
).
Myosin II activity regulates vinculin recruitment to focal adhesions through FAK-mediated paxillin phosphorylation
.
J. Cell Biol.
188
,
877
-
890
.
Phillips
,
H. M.
,
Hildreth
,
V.
,
Peat
,
J. D.
,
Murdoch
,
J. N.
,
Kobayashi
,
K.
,
Chaudhry
,
B.
and
Henderson
,
D. J.
(
2008
).
Non-cell-autonomous roles for the planar cell polarity gene Vangl2 in development of the coronary circulation
.
Circ. Res.
102
,
615
-
623
.
Rudat
,
C.
and
Kispert
,
A.
(
2012
).
Wt1 and epicardial fate mapping
.
Circ. Res.
111
,
165
-
169
.
Schultz
,
J. E.
,
Witt
,
S. A.
,
Nieman
,
M. L.
,
Reiser
,
P. J.
,
Engle
,
S. J.
,
Zhou
,
M.
,
Pawlowski
,
S. A.
,
Lorenz
,
J. N.
,
Kimball
,
T. R.
and
Doetschman
,
T.
(
1999
).
Fibroblast growth factor-2 mediates pressure-induced hypertrophic response
.
J. Clin. Invest.
104
,
709
-
719
.
Shankar
,
J.
and
Nabi
,
I. R.
(
2015
).
Actin cytoskeleton regulation of epithelial mesenchymal transition in metastatic cancer cells
.
PLoS ONE
10
,
e0119954
.
Shankar
,
J.
,
Messenberg
,
A.
,
Chan
,
J.
,
Underhill
,
T. M.
,
Foster
,
L. J.
and
Nabi
,
I. R.
(
2010
).
Pseudopodial actin dynamics control epithelial-mesenchymal transition in metastatic cancer cells
.
Cancer Res.
70
,
3780
-
3790
.
Smutny
,
M.
,
Cox
,
H. L.
,
Leerberg
,
J. M.
,
Kovacs
,
E. M.
,
Conti
,
M. A.
,
Ferguson
,
C.
,
Hamilton
,
N. A.
,
Parton
,
R. G.
,
Adelstein
,
R. S.
and
Yap
,
A. S.
(
2010
).
Myosin II isoforms identify distinct functional modules that support integrity of the epithelial zonula adherens
.
Nat. Cell Biol.
12
,
696
-
702
.
Snider
,
P.
,
Hinton
,
R. B.
,
Moreno-Rodriguez
,
R. A.
,
Wang
,
J.
,
Rogers
,
R.
,
Lindsley
,
A.
,
Li
,
F.
,
Ingram
,
D. A.
,
Menick
,
D.
,
Field
,
L.
, et al. 
(
2008
).
Periostin is required for maturation and extracellular matrix stabilization of noncardiomyocyte lineages of the heart
.
Circ. Res.
102
,
752
-
760
.
Takeda
,
K.
,
Kishi
,
H.
,
Ma
,
X.
,
Yu
,
Z. X.
and
Adelstein
,
R. S.
(
2003
).
Ablation and mutation of nonmuscle myosin heavy chain II-B results in a defect in cardiac myocyte cytokinesis
.
Circ. Res.
93
,
330
-
337
.
Thomas
,
D. G.
,
Yenepalli
,
A.
,
Denais
,
C. M.
,
Rape
,
A.
,
Beach
,
J. R.
,
Wang
,
Y. L.
,
Schiemann
,
W. P.
,
Baskaran
,
H.
,
Lammerding
,
J.
and
Egelhoff
,
T. T.
(
2015
).
Non-muscle myosin IIB is critical for nuclear translocation during 3D invasion
.
J. Cell Biol.
210
,
583
-
594
.
Tirziu
,
D.
,
Giordano
,
F. J.
and
Simons
,
M.
(
2010
).
Cell communications in the heart
.
Circulation
122
,
928
-
937
.
Tullio
,
A. N.
,
Accili
,
D.
,
Ferrans
,
V. J.
,
Yu
,
Z.-X.
,
Takeda
,
K.
,
Grinberg
,
A.
,
Westphal
,
H.
,
Preston
,
Y. A.
and
Adelstein
,
R. S.
(
1997
).
Nonmuscle myosin II-B is required for normal development of the mouse heart
.
Proc. Natl. Acad. Sci. USA
94
,
12407
-
12412
.
Vicente-Manzanares
,
M.
,
Ma
,
X.
,
Adelstein
,
R. S.
and
Horwitz
,
A. R.
(
2009
).
Non-muscle myosin II takes centre stage in cell adhesion and migration
.
Nat. Rev. Mol. Cell Biol.
10
,
778
-
790
.
Von Gise
,
A.
and
Pu
,
W. T.
(
2012
).
Endocardial and epicardial epithelial to mesenchymal transitions in heart development and disease
.
Circ. Res.
110
,
1628
-
1645
.
Vuoriluoto
,
K.
,
Haugen
,
H.
,
Kiviluoto
,
S.
,
Mpindi
,
J. P.
,
Nevo
,
J.
,
Gjerdrum
,
C.
,
Tiron
,
C.
,
Lorens
,
J. B.
and
Ivaska
,
J.
(
2011
).
Vimentin regulates EMT induction by Slug and oncogenic H-Ras and migration by governing Axl expression in breast cancer
.
Oncogene
30
,
1436
-
1448
.
Wang
,
H.
,
Rusielewicz
,
T.
,
Tewari
,
A.
,
Leitman
,
E. M.
,
Einheber
,
S.
and
Melendez-Vasquez
,
C. V.
(
2012
).
Myosin II is a negative regulator of oligodendrocyte morphological differentiation
.
J. Neurosci. Res.
90
,
1547
-
1556
.
Zhou
,
M.
,
Sutliff
,
R. L.
,
Paul
,
R. J.
,
Lorenz
,
J. N.
,
Hoying
,
J. B.
,
Haudenschild
,
C. C.
,
Yin
,
M.
,
Coffin
,
J. D.
,
Kong
,
L.
,
Kranias
,
E. G.
, et al. 
(
1998
).
Fibroblast growth factor 2 control of vascular tone
.
Nat. Med.
4
,
201
-
207
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information