|Home | About | Journals | Submit | Contact Us | Français|
Cardiogenesis involves the contributions of multiple progenitor pools, including mesoderm-derived cardiac progenitors known as the first and second heart fields. Disruption of genetic pathways regulating individual subsets of cardiac progenitors likely underlies many forms of human cardiac malformations. Hand2 is a member of the basic helix loop helix (bHLH) family of transcription factors and is expressed in numerous cell lineages that contribute to the developing heart. However, the early embryonic lethality of Hand2-null mice has precluded lineage-specific study of its function in myocardial progenitors. Here, we generated and used a floxed allele of Hand2 to ablate its expression in specific cardiac cell populations at defined developmental points. We found that Hand2 expression within the mesoderm-derived second heart field progenitors was required for their survival and deletion in this domain recapitulated the complete Hand2-null phenotype. Loss of Hand2 at later stages of development and in restricted domains of the second heart field revealed a spectrum of cardiac anomalies resembling forms of human congenital heart disease. Molecular analyses of Hand2 mutant cells revealed several genes by which Hand2 may influence expansion of the cardiac progenitors. These findings demonstrate that Hand2 is essential for survival of second heart field progenitors and that the graded loss of Hand2 function in this cardiac progenitor pool can cause a spectrum of congenital heart malformation.
Congenital heart defects (CHDs) represent the most common form of human birth defects and occur in nearly 1% of live births (Hoffman and Kaplan, 2002). The recognition that individual pools of cardiac progenitors contribute to specific regions of the heart suggests that some CHDs may be due to disruption of genetic pathways that control migration, survival, expansion or differentiation of distinct populations of cells that contribute to the heart. Because of the dynamic nature of early embryonic development, it is also likely that the developmental requirement for critical genes within progenitors occurs during specific developmental windows.
Cell lineage analyses have demonstrated that the heart develops from multiple sources of cells (reviewed in Buckingham et al., 2005; Olson, 2006; Srivastava, 2006). Two progenitor cell populations, the first heart field (FHF) and second heart field (SHF) are derived from the lateral plate and splanchnic mesoderm, respectively. A third lineage is derived from cardiac neural crest (CNC) cells. In mice, the FHF forms the crescent shaped heart primordium at embryonic day (E) 7.5. At E8.0, these cells fuse at the ventral midline to form the primary heart tube and later contribute to most of the left ventricle (Buckingham et al., 2005). Meanwhile, SHF cells, initially medial and caudal to the FHF, migrate through the pharyngeal mesoderm into the heart tube from both the anterior and posterior poles as the heart tube breaks symmetry and bends to the right. Molecularly distinct subsets of SHF cells contribute to the outflow tract myocardium, right ventricle and atria. By E10.5, CNC cells migrate from their birthplace along the dorsal aspect of the neural folds into the outflow tract to ultimately septate the outflow into two distinct vessels, where they also differentiate into vascular smooth muscle cells.
Hand2, also known as dHAND, is a member of the basic helix-loop-helix (bHLH) family of transcription factors. Hand2 is expressed in the heart, limb bud, and numerous neural crest derivatives during embryogenesis (Srivastava et al., 1995; Srivastava et al., 1997). In the heart, Hand2 is expressed throughout the entire primary heart tube during early embryonic stages, with dominant expression in the right ventricle and outflow tract as the heart tube loops. Hand2-null (Hand2−/−) mice show severe hypoplasia of the right ventricle and growth retardation from E9.5, with death by E10.5 (Srivastava et al., 1997).
The severe phenotype and early embryonic lethality of Hand2−/− mice has precluded the determination of the precise roles of Hand2 within the individual cell types where it is expressed. Tissue-specific deletion of Hand2 in neural crest cells and in the limb bud has revealed essential roles in each tissue (Galli et al., 2010; Morikawa and Cserjesi, 2008). To determine the function of Hand2 in specific subsets of myocardial progenitors that contribute to the heart as well as the cellular mechanisms underlying the severe hypoplasia of the right ventricle in Hand2−/− mice, we established and examined conditional knockout alleles of Hand2 in specific SHF domains.
To generate a conditional allele of Hand2, loxP sites were placed flanking the two introns of Hand2. For selection purposes, a neomycin cassette, flanked by two frt sites, was also placed upstream of the first exon. Homologous recombination and deletion of the Neo cassette by Flip-Frt recombination created the Hand2loxp allele (Fig. 1A). Genotyping was accomplished by digesting DNA with BamHI and ClaI and Southern analysis with a 5' 32P-radiolabeled probe as described (Srivastava et al., 1997). This process produced a 4.5-kb band representing the Hand2loxp allele, a 5.9-kb band representing the null allele, and a 7.5-kb band representing wild type (Fig. 1B). Cre driver (male) and Hand2+/− (female) mice were mated to yield Cre:Hand2+/− mice (male), which were mated with Hand2loxp/loxp mice to generate Cre:Hand2loxp/− conditional knockout mice.
Hand2loxp mice were mated with four different Cre driver lines: Tbx1Cre (Maeda et al 2006), Mef2cCre (Verzi et al., 2007), and Isl1Cre (Cai et al., 2003), which excise the floxed gene in distinct second heart field domains; and Nkx2.5Cre (McFadden et al., 2005), which is active in both the right and left ventricles. Each Cre line was crossed with Hand2+/− mice to obtain Cre:Hand2+/− males. These mice were crossed with Hand2loxp/loxp, Hand2loxp/+, and Hand2loxp/+;ROSALacZ females. All mouse lines were of mixed C57BL6/129SVEJ background. We collected the resulting embryos between E8.5 and E18.5. A summary of cardiac cell types affected by each Cre line as previously published using reporter lines is provided below:
Embryos from timed matings were harvested and fixed overnight in 4% paraformaldehyde/PBS. After fixation, embryos were rinsed in PBS, dehydrated overnight in 70% ethanol, and embedded in paraffin wax. Histological sections were cut and stained with hematoxylin and eosin or used for other analyses, such as TUNEL and cell-proliferation assays.
Embryos from timed matings were harvested and prefixed for 2 hours in 4% paraformaldehyde, 0.25% glutaraldehyde/PBS at 4°C. Beta-galactosidase staining was performed as described (Yamagishi et al., 2003).
TUNEL staining and proliferation assays were performed on transverse sections from E9.0 embryos embedded in paraffin wax. The Cell Death Detection Kit (Roche) was used for TUNEL assays. Proliferation assays used primary anti-phosphohistone H3 antibody (rabbit, diluted 1:200 in 1% BSA/PBS), biotin-labeled goat anti-rabbit IgG, and FITC-strept-avidin. Statistical analysis was performed using the Student’s t-test. A P value of 0.05 or less was considered significant.
E9.0 mouse hearts from wild-type, Hand2−/−, or Nkx2.5Cre:Hand2loxp/− mice were harvested using Trizol reagent (Invitrogen) for total RNA isolation. Total RNA (50 ng) was labeled and hybridized to a mouse mRNA expression microarray (Affymetrix) for analysis. Gene expression values were obtained from Affymetrix CEL files with the GC-RMA package from Bioconductor.
Total RNA was isolated from E9.0 hearts with Trizol reagent (Invitrogen) and 20 ng was reverse transcribed using the Multi Reverse Transcriptase by Random Primer Labeling method (Roche). Quantitative RT-PCR (qRT-PCR) was performed using an ABI7900 with Taqman primer sets for selected cardiac developmental genes, and results were normalized to expression of Gapdh. Statistical analysis was performed using the Student’s t-test. The results shown are representative of more than three independent experiments.
To determine the function of Hand2 in specific domains, we generated a conditional knockout allele of Hand2 (Fig. 1A). Homologous recombination in mouse embryonic stem (ES) cells resulted in ES cells containing loxp sites flanking the Hand2 gene and flp sites surrounding the Neomycin (Neo) resistance cassette 5’ of the first exon. Successful site-specific heterozygous recombination was confirmed by Southern analysis (Fig. 1B). Two distinct targeted ES cell lines were injected into blastocysts and the resulting high percentage chimeras were bred for germline transmission of the targeted allele Heterozygous mice were intercrossed with mice that ubiquitously expressed Flp-recombinase to remove the Neo gene (Hand2loxp/+). Like Hand2+/− mice, Hand2loxp/+ mice were phenotypically normal and fertile. The intercross of Hand2loxp/+ mice generated Hand2loxp/loxp mice, and mice with this genotype approximated Mendelian inheritance at birth, survived to adulthood, and bred normally. Deletion of Hand2 from the Hand2loxp/+ or Hand2loxp/loxp conditional alleles depends on expression and activity of Cre-recombinase. Several domain-specific Cre driver mice that express Cre-recombinase in domains of interest were mated with Hand2+/− mice and further crossed with Hand2loxp/+ or Hand2loxp/loxp mice to obtain mouse embryos with domain-specific ablation of Hand2 gene function.
We ablated Hand2 with the Nkx2.5-enhancer-driven Cre transgenic mouse that results in Cre-recombinase activity in the right and left ventricular chambers after E8.5, but not in the earlier cardiac progenitors prior to their differentiation into cardiac cells (McFadden et al., 2005). Nkx2.5Cre:Hand2loxp/− mice formed both right and left ventricles (Fig. 2B,F), although they were slightly smaller than those of wild-type embryos at E9.5 (Fig. 2A,E) and the mutant mice died by E12.5. Thus, deletion of Hand2 after initial specification of cardiac progenitors appeared to affect the expansion of cardiomyocytes in the embryonic heart, but to a much lesser degree than observed in the complete Hand2-null state.
Hand2 expression is high in the early cardiac progenitors of the FHF and SHF and begins to decline by E9.5 (Srivastava et al., 1995; Srivastava et al., 1997), suggesting that it might have earlier functions in cardiac progenitor cells. To investigate this, we used the Islet1-Cre mouse that expresses Cre-recombinase in all of the early SHF cells, prior to the onset of sarcomeric gene expression (Cai et al., 2003). Islet1-Cre:Hand2loxp/− embryonic hearts died by E10.5 with severe hypoplasia of the right ventricle and a shortened outflow tract (Fig. 2C,G and Fig. 3A–D), phenocopying complete Hand2−/− hearts (Fig. 2D,H). These results suggest that Hand2 expression and function in the early SHF progenitors is essential for their contribution to the developing heart and that the Hand2−/− cardiac defect (Srivastava et al., 1997) largely reflects its function in the SHF.
To investigate the cause for right ventricular hypoplasia in Islet1Cre:Hand2loxp/− embryos, we performed lineage analyses of SHF cardiac progenitor cells by crossing this mouse into the Rosa26lacZ background (Zambrowicz et al., 1997). In this system, Cre-mediated recombination of the Rosa26lacZ allele resulted in β-galactosidase (β-gal) production from the recombined, constitutively expressed Rosa26 locus. Thus, β-gal staining (blue) marked derivatives of the SHF that once expressed Islet1. At E9.5, blue stain was detected in the pharyngeal region and outflow tract of Islet1Cre:Hand2loxp/+;Rosa26LacZ embryos, indicating that the SHF cardiac progenitor cells were migrating, as expected, from the pharyngeal mesoderm into the outflow tract (Fig. 3E). In contrast, fewer LacZ-positive cells were found in Islet1Cre:Hand2loxp/−;Rosa26LacZ embryos at E9.5 (Fig. 3F). The reduction was especially pronounced in the pharyngeal region of these embryos, suggesting that the decrease in progenitor cell number occurred before their migration into the outflow tract.
To determine if the decreased number of progenitor cells from the SHF resulted from abnormal cell death or decreased proliferation, we performed TUNEL and proliferation assays on embryos collected at E9.0. More cell death was detected in the pharyngeal mesoderm of both Islet1-Cre:Hand2loxp/− and Hand2−/− mutants than in wild type embryos (Fig. 3G), while the number of TUNEL positive cells in the outflow tract was only significantly greater in Hand2−/− embryos (Fig. 3H), likely reflecting the influence of Hand2 in neural crest derivatives in the outflow tract as previously reported (Thomas et al., 1998). Immunohistochemical analyses using anti-Ph3 antibody revealed no significant difference in cell proliferation among wild-type, Islet1Cre:Hand2loxp/− and Hand2−/− embryos (data not shown). These results suggest that Hand2 is required within SHF progenitor cells to ensure their survival and that the severe right ventricular hypoplasia that occurs from loss of Hand2 is a result of enhanced apoptosis of SHF cells within the pharyngeal mesoderm before they contribute to the heart.
The discovery of distinct subpopulations of SHF progenitors raises the possibility that some forms of CHD involving the outflow tract or right ventricle are caused by loss of some or all SHF cells during cardiogenesis. Having demonstrated that Hand2 was essential for survival of SHF progenitors, we investigated whether loss of Hand2 in smaller subsets of SHF could result in cardiac anomalies more similar to CHDs in humans. Transgenic mice with Cre-recombinase under control of the SHF-specific enhancer of Mef2c (Mef2c-Cre) express Cre-recombinase in a distinct subset of SHF-derived cells that is narrower than that of the Islet1-Cre driver mice, encompassing the outflow tract and right ventricle, including the ventricular septum (Verzi et al., 2007). Mef2c-Cre:Hand2loxp/− embryos had no gross growth retardation until E12.5, but died around E13.5. By morphological analyses, the right ventricle was smaller in the Mef2cCre:Hand2loxp/− mutant embryos than in wild type embryos at E9.5 (Fig. 4A–D), but larger than Hand2−/− or Islet1Cre:Hand2loxp/− right ventricles. Analyses of these mutants at E12.5 revealed a hypoplastic right ventricle with relatively thin myocardium (Fig. 4E–H). Mef2cCre:Hand2loxp/− mutant embryos lacked the anlage of a tricuspid valve between the right atrium and ventricle in any level of section, and displayed ventricular septal defects (Fig. 4H). This anatomy is reminiscent of a type of human CHD known as tricuspid atresia. At this stage, the distal outflow tract was divided into two vessels in the wild type (Fig. 4I), but an immature truncus arteriosus was noted in Mef2cCre:Hand2loxp/− mutants (Fig. 4J), probably as a result of delayed remodeling of the outflow tract and aortic arches.
We previously showed that the transcription factor Tbx1 is expressed in cells whose descendants contribute mainly to the distal right ventricular outflow tract and pulmonary artery (Maeda et al., 2006). We used an enhancer of Tbx1 driving Cre-recombinase (Tbx1-Cre) to delete Hand2 in this narrower subset of SHF-derived cells (Maeda et al, 2006). Analyses of Tbx1Cre:Hand2loxp/− mutants at E10.5 revealed a shortened outflow tract, while the right ventricle appeared to be developing normally (Fig. 5A–D). By E13.5, Tbx1Cre:Hand2loxp/− embryos had smaller right ventricles and outflow tracts than wild-type mice (Fig. 5E,F) and died by E15.5.
We used the Rosa26LacZ mouse to lineage trace the consequence of Hand2 deletion in Tbx1-expressing cells. LacZ staining of E12.5 Tbx1Cre:Hand2loxp/−;Rosa26LacZ mutant embryos highlighted the narrowing of the outflow tract and hypoplasia of the outlet, or subpulmonary conus, of the right ventricle compared to Tbx1Cre:Hand2loxp/+;RosaLacZ littermates (Fig. 5G–J). Transverse sections of embryos at E14.5 further demonstrated that the right ventricular cavity of Tbx1Cre:Hand2loxp/− mutants was smaller than wild type, although the right ventricular muscle wall thickness was comparable (Fig. 5K,L). The decreased chamber volume was likely secondary to hypoplasia of the infundibular components, as the pulmonary valve appeared normal (Fig. 5M,N). Notably, the relationships of the great vessels to the ventricular chambers and the ventricular septum were normal in all of the mutant embryos. Hypoplasia of the infundibular components (outlet portion) of the right ventricle is often observed in many forms of human CHD and may reflect loss of the distal SHF progenitors.
To elucidate the gene expression alterations that may underlie the severe right ventricular hypoplasia in Hand2−/− mice and Islet1Cre:Hand2loxp/− mice, we performed an mRNA expression array analysis that compared E9.5 wild-type, Hand2−/−, and Nkx2.5Cre:Hand2loxp/− embryonic hearts. Since severe right ventricular hypoplasia was not demonstrated in Nkx2.5Cre:Hand2loxp/− animals, we hypothesized that genes dysregulated in Hand2−/− hearts, but not in Nkx2.5Cre:Hand2loxp/− hearts, may be implicated in the defects specifically resulting from ablation of Hand2 in the SHF.
We selected several key regulators of cardiac development that appeared to be dysregulated by microarray (Fig. 6A) and performed qPCR analyses to validate the array results. Interestingly, expression levels of Gata4, Has2 and Bmp5 were significantly downregulated in Hand2−/− hearts, but not in Nkx2.5Cre:Hand2loxp/− hearts, by array (Fig. 6B) and qPCR (Fig. 6C–E). Decreased levels of Has2 in the Hand2−/− hearts may reflect the smaller right ventricle since Has2 is more enriched in this domain; however, the downregulation of Gata4 and Bmp5, among other dysregulated genes, may contribute to the severe right ventricle hypoplasia observed in Hand2−/− or Islet1Cre:Hand2loxp/− mice. Downregulation of Hand2 was also confirmed in both Hand2−/− and Nkx2.5Cre:Hand2loxp/− hearts by qPCR (Fig 6F).
We sought to dissect the roles of Hand2 in the various cardiac progenitor cells where it is normally expressed and to gain an understanding of the process by which Hand2 controls heart development. We created a floxed Hand2 allele, which we used to delete Hand2 in specific domains of the first or second heart fields. Through analyses of these mutant mice, we found that Hand2 was required in SHF progenitors for their early survival. Temporal- or spatially-restricted ablation of Hand2 function in the SHF revealed its necessity for normal development of the outflow tract, tricuspid valve, and ventricular septum. Interestingly, the abnormalities seen in mice with conditional ablation of Hand2 in discrete sub-domains shared some resemblance to several human congenital heart malformations. Gene expression analyses of these mutant mice revealed potential effectors of Hand2 in right ventricular and general myocardial development.
In this study, we found that loss of Hand2 in the broadest SHF domain phenocopies the ventricular defect observed in global Hand2 knockout, while its ablation in a subset of SHF and FHF is not sufficient to phenocopy the Hand2−/− mutant hearts. We conclude that the SHF is the critical region of Hand2 expression required for proper development of the right ventricle and that this requirement is prior to the migration of SHF cells into the heart. Intriguingly, loss of Hand2 in the SHF led to enhanced apoptosis of cardiac progenitor cells in the pharyngeal mesoderm before they participated in the process of cardiac development, resulting in severe hypoplasia of the right ventricle. Thus, it is likely that Hand2 is essential for survival of undifferentiated cardiac progenitor cells in the SHF. Deletion of Hand2 after cardiac differentiation had begun using the ventricular-specific Nkx2.5-Cre transgenic mouse resulted in a later defect of ventricular expansion, similar to that observed upon deletion of Hand2 in the domain of the myocardial sarcomeric gene, cardiac Troponin T (Morikawa and Cserjesi, 2008).
Deleting Hand2 in a subset of SHF cells resulted in various degrees of hypoplastic right heart syndrome that correlated with the breadth of the SHF cell population that was targeted. For example, the Tbx1Cre:Hand2loxp/− phenotype was similar to severe right ventricular outflow tract obstruction that retards right ventricular chamber growth. In contrast, Mef2cCre:Hand2loxp/− mice developed small, thin-walled right ventricles and tricuspid valve atresia, likely reflecting the broader deletion of Hand2 in cells that contribute to most of the right ventricular chamber.
While homozygous deletion of Hand2 has not been described in humans, several reports indicate that Hand2 heterozygous deletion is part of 4q syndrome, which involves developmental craniofacial, musculoskeletal, and cardiac defects. 4q33, the Hand2 locus (Natarajan et al., 2001), has been identified as the critical region of 4q syndrome (Towns et al., 1979; Yu et al., 1981), and CHDs were diagnosed in half of 4q syndrome patients. The defects included pulmonary valve stenosis or atresia, patent ductus arteriosus, tetralogy of Fallot, tricuspid valve atresia, and coarctation of the aorta. (Keeling et al., 2001; Strahle and Middlemiss, 2007). Some of these may reflect Hand2’s function in the neural crest, as neural crest-specific ablation of Hand2 causes similar loss of neural crest derivatives in the heart (Morikawa and Cserjesi, 2008) and deletion of the pharyngeal arch-specific Hand2 enhancer causes craniofacial defects (Yanagisawa et al., 2003). The overlapping phenotypes of Hand2 SHF ablation in mice and 4q syndrome in humans suggests a potential role for Hand2 in human CHDs, and points to the relevance of abnormal SHF progenitor cell development to human heart disease.
In an effort to examine gene expression defects in the SHF of Hand2−/− mice, we found that expression of the zinc finger transcription factor, Gata4, was reduced in Hand2−/− hearts, but not in Nkx2.5Cre:Hand2loxp/−. Interestingly, Gata4 expression was also unchanged in mice containing a specific mutation in the Hand2 DNA-binding domain, and these mice lived to a similar stage as the Nkx2.5Cre:Hand2loxp/− mice (Liu et al., 2009). Gata4 is normally expressed in the SHF and directly regulates Hand2 during right ventricular development (McFadden et al., 2000), and Gata4 conditional knockout mice demonstrate severe hypoplasia of the right ventricle with reduced expression of Hand2 (Zeisberg et al., 2005). Thus, there may be a reinforcing loop between Hand2 and Gata4 that amplifies expression of both genes early during SHF and subsequent right ventricular development.
In addition to Gata4, the expression of Has2 and Bmp5 was reduced in Hand2−/− hearts compared to wild type and Nkx2.5Cre:Hand2loxp/−. Has2 encodes hyaluronan synthase-2, which is essential for hyaluronic acid secretion into the extracellular matrix between the early endocardium and myocardium, known as the cardiac jelly (Camenisch et al., 2000; Mjaatvedt et al., 1987). The cardiac jelly mediates reciprocal signaling between the endocardium and myocardium that results in myocardial cell proliferation and maturation. Reduced expression of Has2 is associated with hypoplastic right ventricle (Mjaatvedt et al., 1987) and may also contribute to the abnormalities in Hand2−/− or Islet1Cre:Hand2loxp/− mice. However, it is also possible that the decrease in Has2 and other genes uniquely expressed in the right ventricle simply reflects the lack of right ventricular cells in the developing heart,
The bone morphogenic protein (BMP) family of signaling molecules mediates the simultaneous activation of Smad and Wnt signal transduction in the cardiogenic pathway, including induction of the earliest markers of cardiac differentiation, such as Gata4, Nkx2.5, and Mef2c (Pal and Khanna, 2006). In chick embryos, Bmp5 is expressed in the outflow tract myocardium and may play a role in the process of recruitment of cardiomyocytes to the outflow tract (Somi et al., 2004). Indeed, loss of Bmp5 and Bmp7 resulted in outflow tract defects (Solloway and Robertson, 1999), suggesting that reduced Bmp5-mediated signaling could contribute to the decreased numbers of SHF progenitors observed upon reduction of Hand2 dosage. It remains to be determined whether Gata4, Has2 or Bmp5 is directly or indirectly regulated by Hand2, but the regulation of Gata4 appears to be DNA-independent (Liu et al., 2009).
By establishing a conditional knockout allele of Hand2 utilizing a variety of unique Cre-driver mouse lines, we have revealed some tissue-specific roles of the Hand2 gene during development of early cardiac progenitors. More importantly, the loss of Hand2 in specific sub-domains of the SHF supports the idea that some forms of CHD might be a result of insufficient cells derived from discrete pools of progenitor cells. Our identification of genes that are specifically dysregulated when early SHF cells lack Hand2 and are unaffected in the later knockout offers an opportunity to focus on pathways uniquely important for survival and expansion of cells populating the right ventricle. Further investigation of the function of these factors may contribute to identification of the underlying causes of an array of CHDs affecting the human population.
We thank members of the Gladstone Histology Core for embryo sections (J. Fish and C. Miller); the Gladstone Genomics Core for microarray experiments (L. Ta and Y. Hiao); Gladstone Bioinformatics Core for data analysis (A. Holloway); and Bethany Taylor for manuscript and figure preparation. We also thank members of the Srivastava lab for helpful discussion. D.S. was supported by grants from NHLBI/NIH and the California Institute for Regenerative Medicine and the Younger Family Foundation. H.Y. was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan. This work was also supported by NIH/NCRR grant C06 RR018928 to the Gladstone Institutes.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.