Characterization and in vivo pharmacological rescue of a Wnt2-Gata6 pathway required for cardiac inflow tract development

Expression of Wnt2 in the posterior pole and inflow tract of the heart

Previous reports have shown that Wnt2 is expressed in early cardiac mesoderm in the cardiac crescent at E7.5 (Monkley et al., 1996). To more fully address the pattern of Wnt2 expression during cardiovascular development, we performed in situ hybridization on whole embryos and tissue sections. From E8.5-E9.5, Wnt2 expression is observed in the posterior cardiac mesoderm surrounding the ventral aspect of the anterior foregut contiguous with the inflow tract of the developing heart (Fig. 1A-E). Expression of Wnt2 is also observed in the atria at E9.5 and E10.5 (Fig. 1E and F). Although expression of Wnt2 is observed in the developing inflow tract at these times, it is not observed in the ventricles or AV canal (Fig. 1A-C, G). Within the inflow tract mesoderm, high levels of Wnt2 are observed in a region overlapping with the dorsal mesocardium and dorsal mesenchymal protrusion (DMP) at E9.5 (Fig. 1D and Supplemental Figure 1). This tissue is contigious with the atrial myocardium including the primary atrial septum and the overlying mesenchymal cap, which in turn helps generate portions of the AV cushions (Fig. 1H and (Snarr et al., 2007a; Snarr et al., 2007b; Wessels et al., 2000). The dorsal mesocardium and DMP overlaps extensively with the posterior portion of the second heart field as shown by Isl1-cre fate mapping (Cai et al., 2003; Snarr et al., 2007a; Sun et al., 2007). Wnt2 expression is not observed in either epicardium or endocardium but is observed in septum transversum mesoderm in a pattern that overlaps with Tbx18 and Wilm's tumor-1 (WT1) at E9.5 and E10.5 (Fig. 1D and Supplemental Figure 2). These data are consistent with a previous report showing that Wnt2 is not expressed in the AV canal or ventricular myocardium at these times (Watanabe et al., 2006). By E10.5, expression of Wnt2 begins to decline and expression in the heart is very low to undetectable after E12.5 (data not shown).

Wnt2 null mutants exhibit defects in cardiac inflow tract and AV canal morphogenesis

To determine the role of Wnt2 signaling in cardiac development, we generated and characterized Wnt2 null embryos. Approximately 85% of Wnt2 null mutants die at birth (Fig. 5A and (Goss et al., 2009)). Histological analysis shows that Wnt2-/- mutants have a thin atrial wall and defects in AV canal development including decreased development or complete loss of the primary atrial septum and a deficiency in development of the superior AV cushion and associated myocardium (Fig. 1I-P, Supplemental Figure 1). This defect is observed as early as E10.5 where there is an early loss of development of the primary atrial septum and the overlying mesenchymal cap which generates portions of the AV cushions (Supplemental Fig. 1). Overall myocardial differentiation in the atria and AV canal is deficient in Wnt2-/- mutants (Fig. 1I-P and Supplemental Figure 1). Together, these defects resemble the human congenital heart syndrome CCAVC. Pulmonary vein development was also significantly defective in Wnt2-/- mutants. At both E14.5 and E18.5, pulmonary veins in Wnt2-/- mutants had a thinner mural wall especially closer to the atria (Fig. 1Q-T). Ventricular and outflow tract development did not appear disrupted in Wnt2-/- mutants although we did observe some thinning of the ventricular wall later in development after E14.5 (Fig. 1P and Supplemental Figure 3). Signals from the epicardium have been shown to play a role in promoting myocardial proliferation and epicardial progenitors are derived from the septum transversum where Wnt2 is expressed (Fig. 1D and E and (Lavine et al., 2005). However, expression of the epicardial markers WT1 and Tbx18 are unchanged in Wnt2-/- mutants suggesting that epicardium is intact in these mutants (Fig. 3M and Supplemental Figure 2). Moreover, Wnt2 expression has been observed at E7.5 in the early cardiac mesoderm (Monkley et al., 1996). Thus, the ventricular thinning at later stages of Wnt2-/- mutants could be secondary to impending heart failure or to a role for Wnt2 in early cardiac mesoderm growth.

Endocardial cushions develop through a process of endothelial-mesenchymal transformation (EMT) where the endothelial cells respond to signals from the overlying myocardium to migrate into the extracellular matrix of the cardiac jelly (Nakajima et al., 2000). Wnt2 is expressed in the dorsal mesocardium and DMP and it may signal in a paracrine fashion to the overlying mesenchymal cap of the primary atrial septum and adjacent endocardium to promote proper development of the AV canal through regulation of EMT (Snarr et al., 2007a; Snarr et al., 2007b; Wessels et al., 2000). Thus, we wanted to determine whether EMT was perturbed in Wnt2-/- mutants. AV canal explants from E9.5 heterozygous and Wnt2-/- embryos were cultured in a collagen gel and the number of mesenchymal cells migrating out of the explant were quantitated. These results show that Wnt2-/- mutant explants had a significant decrease in EMT consistent with the observed defects in AV canal development (Fig. 2A-C).

To assess whether cardiac function in Wnt2-/- mutants was compromised due to defective AV canal development, echocardiography was performed at E16.5 on wild-type and Wnt2-/- mutant littermates as well as adult mutants. These studies showed abnormal blood flow across the AV canal in embryonic Wnt2-/- mutants (Fig. 2D and E). Severe AV canal regurgitation was also observed in the small number of Wnt2-/- mutants that survive (Fig. 2F and G). Since we did not observe Wnt2 expression in AV canal cushions and valves, these data suggest that Wnt2 expression in the DMP in the posterior inflow tract region regulates AV canal morphogenesis.

Inflow tract and atrial myocardial defects in Wnt2 null mutants

To characterize the mechanism underlying defective inflow tract development in Wnt2-/- null mutants, we performed transmission electron microscopy (TEM) to assess sarcomeric structure in both atrial and ventricular myocardium. As expected, wild-type littermates exhibited normal sarcomeric structure in both atrial and ventricular myocardium at E12.5 (Fig. 3A and B). Wnt2-/- null ventricular myocytes also exhibited normal sarcomeric structure (Fig. 3D). However, sarcomeric structure was severely compromised observed in atrial myocytes of Wnt2-/- mutants as noted by the dramatic reduction of organized Z-bands (Fig. 3C). In agreement with these findings, expression of important cardiac contractile and cytoskeletal genes were reduced in the atria of Wnt2-/- mutants with cardiac troponin I (cTnI) expression reduced by more than 90% (Fig. 3E).

Given the poor myocardial development observed in the atria of Wnt2-/- mutants, we assessed proliferation in the dorsal mesocardium and DMP region as well as in the atria and ventricles at E9.5 and E11.5. Wnt2-/- mutants showed a considerable reduction in proliferation in the dorsal mesocardium, DMP, and atrial myocardium while ventricular proliferation was unaffected at E9.5 (Fig. 3F, G, J and Supplemental Figure 4). Wnt2-/- mutants continued to show reduced atrial but not ventricular myocardial proliferation at E11.5 (Fig. 3H, I, K and L).

We examined expression of several other key signaling and transcription factors known to be important for myocardial and AV canal development as well as markers of epicardium. Quantitative PCR (Q-PCR) shows that expression of Fgf10 is decreased in Wnt2-/- mutants while expression of Fgf8 is increased at E9.5 (Fig. 3M-Q). This result is consistent with previous studies showing that Fgf10 is a direct target of Wnt/β-catenin signaling (Cohen et al., 2007). Expression of Tbx3, Tbx5, Tbx18, WT1, Bmp2, and Bmp4 was not significantly altered in Wnt2-/- mutants at E9.5 (Fig. 3M and Supplemental Figure 2). Thus, loss of Wnt2 expression leads to disregulation of Fgf ligand expression but unaltered expression of other key factors in cardiac development.

Wnt2 regulates expansion of SHF progenitors through β-catenin dependent signaling

Wnt2 is expressed at significant levels in the posterior dorsal mesocardium and DMP region of the inflow tract which generates portions of the AV canal as well as inflow tract myocardium and overlaps extensively with posterior Isl1+ SHF progenitors (Snarr et al., 2007a; Snarr et al., 2007b; Wessels et al., 2000). To determine whether Wnt2 regulates the β-catenin dependant canonical Wnt pathway in the posterior inflow tract and AV canal, we generated Wnt2-/-:TOPGAL compound mutants. Previous reports have shown significant TOPGAL activity in the developing AV canal region in early cardiac development (Gitler et al., 2003). Wild-type TOPGAL animals revealed strong lacZ staining in the AV canal including both cushion mesenchyme and adjacent myocardium at E11.5 (Fig. 4A). However, Wnt2-/- mutants showed a dramatic down-regulation of lacZ staining in the AV canal region (Fig. 4B). To further assess Wnt signaling activity in Wnt2-/- mutants, we examined expression of the Wnt target genes axin2 and Lef1 at E9.5. Expression of axin2 and Lef1 were also significantly reduced in the inflow tract of Wnt2-/- mutants indicating a decrease in Wnt/β-catenin signaling (Fig. 4E and F). High levels of β-catenin protein expression as well as nuclear accumulation of β-catenin have been shown to reflect active Wnt staining in vivo (Zhang et al., 2008). High levels of β-catenin expression are observed both in the inflow tract mesoderm including both myocardium and endocardium as well as in the foregut endoderm of wild-type embryos while β-catenin expression was reduced in the inflow tract myocardium and endocardium of Wnt2-/- mutants at E9.5 (Fig. 4C and D). β-catenin expression in foregut endoderm of Wnt2-/- mutants was less affected (Fig. 4D). This is supported by decreased nuclear β-catenin expression in the dorsal mesocardium and DMP region of Wnt2-/- mutants at E9.5 in comparison to wild-type littermates (Fig. 4H-K).

We and others have shown that Wnt/β-catenin signaling regulates expansion of SHF progenitors expressing the Isl1 transcription factor in the anterior portion of the developing heart (Ai et al., 2007; Cohen et al., 2007; Kwon et al., 2007; Lin et al., 2007; Qyang et al., 2007). Isl1 has also been shown to be a direct transcriptional target of Wnt/β-catenin signaling (Lin et al., 2007). Isl1 expression domain encompasses the majority of the mesoderm within the posterior pole of the heart including the DMP and Isl1-cre fate mapping has shown that most of the inflow tract of the heart is derived from Isl1+ progenitors (Cai et al., 2003; Galli et al., 2008; Snarr et al., 2007a; Sun et al., 2007). Moreover, the dorsal mesocardium and DMP region is the likely point of entry for Isl1+ SHF progenitors as they migrate into the posterior pole of the heart (Snarr et al., 2007a). However, the role for Wnt/β-catenin signaling in posterior SHF development is less clear. Isl1 gene expression was reduced in the posterior inflow tract region of Wnt2-/- mutants consistent with an effect on both Isl1 expression and number of Isl1+ SHF progenitors at E9.5 (Fig. 4G). Isl1 immunostaining confirmed that the number of Isl1+ SHF progenitors is significantly reduced in the dorsal mesocardium and DMP of Wnt2-/- mutants at E9.5 and E10.5 (Fig. 4L-N). Moreover, the posterior dorsal mesocardium of the heart including the DMP is hypoplastic suggesting that Isl1+ posterior progenitors in addition to the adjacent mesoderm requires Wnt2 for proper proliferation and development.

To demonstrate that reduced Isl1+ SHF progenitor number correlated, in part, with a loss of β-catenin dependent Wnt signaling we generated Ctnnb1^flox/flox:Nkx2.5^cre/+ mutants to delete β-catenin throughout early FHF and SHF progenitors (Moses et al., 2001). Loss of β-catenin expression in early cardiac progenitors resulted in a significant decrease in Isl1+ progenitors in the posterior inflow tract mesoderm (Fig. 4O-Q). These data indicate that Wnt2 regulates the expansion of Isl1+ SHF progenitors in the posterior pole of the heart in a region overlapping with the DMP, which likely accounts for the defects in myocardial and AV cushion development in the inflow tract region of these mutants and leads to the CCAVC phenotype.

Transient pharmacological activation of Wnt signaling leads to rescue of cardiac defects in Wnt2 mutants

Based on our expression data as well as previous studies, Wnt2 is expressed early in cardiac mesoderm at approximately E7.5-E8.0 while expression later between E8.5-E10.5 is confined to the cardiac inflow tract including the dorsal mesocardium, DMP, and developing atria (Fig. 1, Supplemental Figure 1 and (Monkley et al., 1996; Watanabe et al., 2006)). Therefore, we wanted to determine the temporal requirement of Wnt2 action in cardiac inflow tract development by activating Wnt/β-catenin signaling in a temporally restricted manner. Since previous reports using cre activated stabilization of β-catenin protein shows that persistent Wnt/β-catenin activation during cardiac development leads to severe cardiac defects and early embryonic demise (Cohen et al., 2007; Kwon et al., 2007), we chose a pharmacological approach using LiCl, an inhibitor of Gsk-3β and strong activator of Wnt/β-catenin signaling (Chu et al., 2004; Cohen et al., 2007; Klein and Melton, 1996; Nakamura et al., 2003). To determine when Wnt2 was required for cardiac development, we injected LiCl once a day in pregnant females from heterozygous intercrosses during three stages of development: E6-E8, E8-E10, and E11-E13. Since approximately 85% of Wnt2 mutants die at birth, we first asked whether LiCl treatment would increase survival of Wnt2-/- mutants. Wnt2-/- mutants treated with LiCl from E8-E10 had a significant increase in survival from 15% to 72% (Fig. 5A and B). All E8-E10 LiCl rescued Wnt2-/- mutants survived to at least 3 months of age and were normal as assessed by histology and echocardiography (data not shown). The lung hypoplasia previously reported in these mice was also rescued (data not shown and (Goss et al., 2009). Histological analysis of E18.5 hearts from NaCl (control) and LiCl E8-E10 treated Wnt2-/- mutants revealed a restoration of AV canal and inflow tract morphology reversing the CCAVC phenotype (Fig. 5C-F). Wnt2-/- mutants treated with LiCl from E6-E8 also exhibited an increase in survival up to 44% (Supplemental Figure 5). LiCl treatment between E6-E8 did not rescue the CCAVC phenotype but did rescue the thinned ventricular myocardium observed in many Wnt2-/- mutants (Fig. 5H). This correlates with previous studies showing low level expression of Wnt2 in the early cardiac crescent which suggests that Wnt2 acts at this earlier stage to promote general myocardial growth and loss of Wnt2 expression at this earlier stage may lead to the ventricular thinning we observe in some Wnt2-/- mutants in late development (Fig. 1P and (Monkley et al., 1996). LiCl also increased myocardial mass in wild-type embryos although this resolved itself after birth (Fig. 5E and data not shown). Echocardiograms at E16.5 and in adults showed normal AV canal function in E8-E10 LiCl treated Wnt2-/- mutants (Fig. 5I and J and data not shown). Wnt2-/- mutants treated with LiCl between E11-E13 did not show any significant increase in survival nor rescue of the CCAVC phenotype suggesting a lack of important Wnt2 activity after E11 (Supplemental Figure 5 and data not shown). Together, these data show that the E8-E10 LiCl rescued Wnt2-/- mutants are phenotypically indistinguishable from wild-type littermates and indicate that Wnt2 function is restricted to cardiac development before E11.

To determine whether the loss in inflow tract myocardial proliferation observed in Wnt2-/- mutants was rescued with LiCl activation of Wnt signaling, we performed MF20 and Ki67 immunostaining. These data show that DMP and atrial myocardial proliferation was largely rescued in E8-E10 LiCl treated Wnt2-/- mutants compared to controls at E11.5 (Fig. 5K-W). Proliferation in the AV canal, which is reduced in Wnt2-/- mutants, is also rescued by E8-E10 LiCl treatment (Fig. 5W). Together, these data reveal a significant reversal of the Wnt2-/- mutant CCAVC phenotype and perinatal lethality in a time dependent fashion by the Gsk-3β inhibitor LiCl. Moreover, these data indicate that Wnt2 acts primarily between E8-E10 of development to regulate cardiac inflow tract development.

Wnt2 regulates a transcriptional pathway important for cardiac progenitor development

To characterize the underlying molecular pathway that integrates Wnt2 signaling during cardiac inflow tract development, we performed microarray analysis on wild-type and Wnt2-/- mutant hearts at E12.5. These studies revealed a set of important transcription factors down-regulated in Wnt2-/- mutant hearts including Gata6, Sall1, and Sall3 (Fig. 6A). Decreased expression of Gata6, Sall1, and Sall3 was confirmed by Q-PCR (Fig. 6B). Moreover, expression of Gata6 in the posterior pole of the heart was decreased in Wnt2-/- mutants as assessed by in situ hybridization (Fig. 6C). Gata4/6 are important modulators of cardiac myocardial differentiation and proliferation (Rojas et al., 2008; Zeisberg et al., 2005). Sall genes are known to regulate important aspects of heart, kidney, and neural development as well as play important roles in stem/progenitor cells including embryonic stem cells (Harvey and Logan, 2006; Kohlhase et al., 2003; Koshiba-Takeuchi et al., 2006; Sweetman and Munsterberg, 2006; Warren et al., 2007).

We next tested whether LiCl treatment would rescue the decreased expression of these Wnt2 regulated factors. Q-PCR of E10.5 hearts from Wnt2+/- and Wnt2-/- mutants treated with LiCl from E8-E10 showed that LiCl rescued expression of Gata6, Sall3, and Isl1 in Wnt2-/- mutants (Fig. 6D). Given the critical role of Gata factors in cardiac development and progenitor specification (Zhao et al., 2008), we further characterized the regulation of Gata6 by Wnt2. Previous studies have identified an upstream regulatory region of the mouse Gata6 gene that can direct expression of lacZ to certain regions of the heart including the atria (Molkentin et al., 2000). We scanned the sequence of this enhancer as well as regions immediately flanking it and found six potential Lef/Tcf DNA binding sites (Fig. 6E). We performed chromatin immunoprecipitation (ChIP) analysis using both β-catenin and Tcf3/4 antibodies to assess association of these Wnt transcriptional effectors with the Lef/Tcf sites in vivo using chromatin from E9.5 hearts. These studies revealed that β-catenin and Tcf3/4 were associated with sites 3 and site 6 (Fig. 6F and G). Luciferase reporters containing different portions of this enhancer region revealed that of these two sites, site 3 was required for trans-activation by activated β-catenin and Wnt2 (Fig. 6H). To determine whether exogenous Wnt can activate Gata6 expression in the cardiac inflow tract, posterior pole cardiac explants were cultured in the presence or absence of Wnt3a. These data show that exogenous Wnt3a increases expression of both Gata6 and Isl1, a known target of Wnt signaling, in posterior cardiac explants (Fig. 6I). Together, these data indicate that Gata6 is a direct target of Wnt2 during cardiac development.

Wnt2 and Gata6 interact genetically to regulate inflow tract development in the heart

Although a role for Gata6 in cardiac neural crest dependent outflow tract development has been reported (Lepore et al., 2006), less is understood about its role in early cardiac development. Therefore, we inactivated Gata6 in early cardiac progenitors using the Nkx2.5^cre/+ line (Moses et al., 2001). Approximately 50% of Nkx2.5^cre/+:Gata6^flox/flox mutants have distinct defects in AV canal development leading to a phenotype similar to that observed in Wnt2-/- mutants including CCAVC (Fig. 7A and B). The affected Nkx2.5^cre/+:Gata6^flox/flox mutants have more severe ventricular septal defects than Wnt2-/- mutants (Fig. 7B), demonstrating that Gata6 plays an important role in early myocardial development as has been reported (Zhao et al., 2008). To determine whether Wnt2 and Gata6 genetically interacted, we generated Wnt2/Gata6 compound mutants. The results from these studies revealed that Wnt2/Gata6 double heterozygous (Wnt2+/-:Gata6+/-) mutants are underrepresented at weaning (Fig. 7C). Histological analysis of Wnt2+/-:Gata6+/- animals shows defects similar to but less severe than Wnt2-/- null mutants including a thin atrial myocardial wall and decreased AV canal development including an atrial septal defect (Fig. 7D-F). To determine whether Wnt2 and Gata6 cooperated to regulate the same molecular pathway as Wnt2 alone, expression of genes found to be down-regulated by loss of Wnt2 were assessed in Wnt2+/-:Gata6+/- mutants. These data show that the transcriptional pathway downstream of Wnt2 is cooperatively decreased in Wnt2+/-:Gata6+/- mutants (Fig. 7G). The loss of atrial and AV canal myocardial development in Wnt2+/-:Gata6+/- mutants is illustrated by in situ hybridization with a Mlc2a probe (Fig. 7H and I). Moreover, atrial myocardial proliferation was decreased in Wnt2+/-:Gata6+/- mutants in comparison to Wnt2 and Gata6 single heterozygous mutants (Fig. 7J and K). Together with previous data showing that Wnt2 is a direct target of Gata6 (Alexandrovich et al., 2006), these data support a Wnt2-Gata6 feed-forward mechanism required for development of the cardiac inflow tract region including the DMP (Fig. 7L).

Animals

Generation and genotyping of the Wnt2, Gata6^flox/flox, Ctnnb1^flox/flox, Nkx2.5^cre/+, and TOPGAL mouse lines has been reported (Brault et al., 2001; DasGupta and Fuchs, 1999; Goss et al., 2009; Lepore et al., 2006; Moses et al., 2001). Mice were kept on a mixed C57BL/6:129SVJ background. For the LiCl rescue studies, pregnant mice were injected interperitoneally with 200 mg/Kg once a day at the indicated times. All animal procedures were performed in accordance with the Institute for Animal Care and Use Committee at the University of Pennsylvania

Histology

Embryos and tissues were fixed in 4% paraformaldehyde for 24 hours, dehydrated in increasing concentrations of ethanol, and embedded in paraffin for tissue sectioning. In situ hybridization was performed as described (Morrisey et al., 1997a; Morrisey et al., 1997b). Immunohistochemistry was performed using the following antibodies: mouse anti-rat Isl-1 (University of Iowa Hybridoma Bank), rat anti-Ki67 (clone TEC-3, Dako), β-catenin (BD Biosciences, 1:50), and MF20 (University of Iowa Hybridoma Bank). For scanning electron microscopy, hearts from E12.5 embryos were fixed in 2% gluteraldehyde and processed as described (Shu et al., 2002). LacZ histochemical staining of embryos was performed as described (Shu et al., 2002).

EMT and posterior cardiac pole explant assays

AV canal regions of E9.5 hearts were dissected from the indicated genotypes and cultured for 48 hours on collagen/fibronectin-coated tissue culture plate as described (Wang et al., 2005). Photographs were taken using phase contrast microscopy. The posterior cardiac pole was dissected at E8.5 and cultured on collagen coated tissue culture plates for 48 hours as previously described (Cohen et al., 2007). Wnt3a (R&D Systems) was added at a concentration of 200 ng/ml.

Gene expression analysis

Total RNA was isolated from hearts at the indcated ages using Trizol reagent, reverse transcribed using SuperScript First Strand Synthesis System (Invitrogen), and used in quantitative real time PCR analysis using the oligonucleotides listed in Supplemental Table 1. Microarray analysis was performed with a Mouse Genome 430 2.0 array (Affymetrix, Santa Clara, CA). Total RNA isolated from E12.5 hearts from wild-type and Wnt2-/- null mutant littermates using three samples of each genotype. Data was analyzed using Microarray Suite 5.0 (MASS, Affymetrix) and Significance Analysis of Microarray (SAM). Genes with 1.5-fold and greater changes than the experimental mean and carrying out ANOVA with a p value less than 0.01 were considered significant.

Cell transfection and chromatin immunoprecipitation (ChIP) assays

The upstream enhancer region of mouse Gata6 has been described preciously (Molkentin et al., 2000). The region between -4kb and -1.5 kb was subcloned into the pGL3 promoter luciferase reporter vector to generate pGL3Gata6.luc vector. This plasmid was transfected with an active form of β-catenin (β-catenin^41A/45A) into HEK293 cells using Fugene 6 and luciferase activities were determined using a commercially available kit (Promega). For ChIP assays, chromatin was extracted from 25 wild-type embryonic hearts at E9.5 using a ChIP assay kit (Upstate Biotechnology). Chromatin was immunoprecipated with either the β-catenin mAb (BD Biosciences), a Tcf3/4 antibody (Abcam ab12065), or a nonimmune control antibody mouse IgG (Sigma). Purified chromatin was subjected to PCR using the oligonucleotides listed in Supplemental Table 1.

Echocardiography

Pregnant mice were lightly anesthetized with 1-2% isoflurane and embryonic echocardiography was performed using a high-resolution Vevo 770 micro-ultrasound system (VisualSonics Inc.) equipped with a RMV-704 transducer with a center frequency of 40 MHz. For adult mouse echocardiography, a RMV-707 transducer with a center frequency of 30 MHz was used. Cardiac four-chamber view of the embryo was acquired by rotating the transducer about 45 degree counterclockwise from the longitudinal axis of the embryo body. End-diastolic and end-systolic left ventricular internal diameters (LVIDd, LVIDs) were measured from the left ventricular short axis view with 2D orientated M-mode imaging. Left ventricular systolic function was estimated by fractional shortening (FS, %), which was derived using the formula: (LVIDd-LVIDs) / LVIDd × 100, and ejection fraction (EF, %), which was calculated using the end-systolic and end-diastolic volumes as described (Stypmann et al., 2009). Doppler echocardiography was performed to detect flow velocities.