Keywords

9.1 Introduction

The heart consists of many types of cells, including cardiomyocytes, vascular cells, neural cells, and cardiac fibroblasts. Heart function is tightly regulated by the interactions between cardiomyocytes and other type of cells through secreting molecules. Understanding the cell–cell interactions and their molecular mechanisms during heart development might provide new therapeutic and preventive approaches for heart disease.

The relationship between autonomic nerve activity and arrhythmias is well recognized, but the molecular mechanisms determining the innervation densities and the consequence of its disruption were unclear. Compared with other organs, the heart is extensively innervated via the autonomic nervous system, which comprises the sympathetic and parasympathetic nerves. The sympathetic nervous system produces norepinephrine, and increases heart rate, conduction velocity, and myocardial contraction and relaxation. It is well known that sympathetic innervation density, which is abundant in the subepicardium and the central conduction system, is strictly determined within the heart (Ito and Zipes 1994; Crick et al. 1994, 1999a, b). The regional difference in sympathetic innervation influences specific cardiac functions to a different extent, effectively controlling heart rate and myocardial contraction and relaxation. Unbalanced sympathetic innervation might trigger lethal arrhythmia through ion channel modulation in cardiomyocytes (Cao et al. 2000; Dae et al. 1997; Qu and Robinson 2004). Sema3a, a class 3 secreted semaphorin, was cloned as a potent neural chemorepellent and a directional guidance molecule for nerve fibers. We found that Sema3a is expressed in cardiac tissues and determines sympathetic innervation patterning in the developing heart. Disruption of Sema3a expression in mouse hearts lead to fetal arrhythmias and sudden cardiac death. More recently, Nakano et al. reported that a nonsynonymous polymorphism in Sema3a is a risk factor for human unexplained cardiac arrest with ventricular fibrillation. They showed that sympathetic innervation patterning was disrupted in these patients, suggesting proper function of Sema3a is critical for arrhythmia-free hearts in humans.

Congenital heart diseases (CHD) occur in nearly 1 % of all live births and are the major cause of infant mortality and morbidity. Cardiac outflow tract (OFT) defects are estimated to account for approximately 30 % of CHD, but the genetic etiology of most OFT defects remains unknown. Increasing evidence has demonstrated that semaphorin 3C, a member of the semaphorin family of secreted factors, is expressed in the cardiovascular system during development. The semaphorin 3C knockout (KO) mouse revealed impaired migration of cardiac neural crest cells to the developing outflow tract, leading to OFT defects and perinatal death. Gata6 directly regulates semaphorin 3C expression during development, and neural crest-specific Gata6 mutant mice revealed OFT defects, recapitulating the phenotype of the semaphorin 3C KO mouse. Kodo et al. found two mutations in GATA6 in their series of 21 Japanese patients with persistent truncus arteriosus (PTA), a major disease in OFT defects. They demonstrated that both mutations disrupted the transcriptional activity of the GATA6 protein, leading to failure of transactivation of SEMA3C and PLXNA2. The results in human OFT defects are consistent with those observed in Gata6 and Sema3c mutant mice, demonstrating that the GATA6/Sema3c pathway is conserved among species and is critical for development of the cardiovascular system.

In this chapter, we review the function of semaphorin family members, demonstrated as important for cardiovascular development and disease in both mouse and human.

9.2 Cardiac Sympathetic Innervation and Nerve Growth Factor

The heart is innervated by sympathetic, parasympathetic, and sensory nerves derived from neural crest cells. Trunk neural crest cells migrate and form sympathetic ganglia by midgestation, subsequently proliferating and differentiating into mature neurons. The cardiac sympathetic nerves extend from the sympathetic neurons in stellate ganglia, which are located bilateral to the thoracic vertebrae. Sympathetic nerve fibers project from the base of the heart into the myocardium, and are located predominantly in the subepicardium of the ventricle (Fig. 9.1). The central conduction system, which includes the sinoatrial node, atrioventricular node, and His bundle, is abundantly innervated compared with the working myocardium. The regional difference in cardiac sympathetic innervation is highly conserved among mammals.

Fig. 9.1
figure 1

Inverse expression pattern of Sema3a and sympathetic innervation in developing hearts The cardiac sympathetic nerves (green) extend from the sympathetic neurons in stellate ganglia, which are located bilateral to the thoracic vertebra (left). Cardiac sympathetic innervation shows an epicardial-to-endocardial transmural gradient (right). This patterning is established by the balance between ET-1/NGF and Sema3a expression in the heart. Note that NGF is abundantly expressed in the working myocardium, whereas Sema3a is specifically expressed in the subendocardium (Ieda et al. 2007, 2008)

Full size image

In general, the growth cone behavior of nerves is modulated by coincident signaling between neural chemoattractant and chemorepellent synthesized in the innervated tissue. Nerve growth factor (NGF), a potent neural chemoattractant, is a prototypic member of the neurotrophin family, members of which are critical for the differentiation, survival, and synaptic activity of the peripheral sympathetic and sensory nervous systems (Snider 1994; Lockhart et al. 1997; Brennan et al. 1999). We and others have reported that NGF expression in cardiomyocytes is critical as a sympathetic chemoattractant in normal and diseased hearts. Among several cardiac hypertrophic factors, we found that ET-1 is a key regulator of NGF expression in cardiomyocytes, and that the ET-1/NGF pathway is critical for sympathetic innervation in developing and right ventricular hypertrophy hearts (Ieda et al. 2004; Kimura et al. 2007). Given that ET-1 is strongly induced in myocardial infarction, the ET-1/NGF pathway might also be involved in NGF upregulation and nerve regeneration after myocardial infarction.

9.3 Sema3a Is Critical for Cardiac Sympathetic Innervation Patterning

As already discussed, NGF, a neural chemoattractant, has critical functions in cardiac nerve development. In contrast, the neural chemorepellent that induces growth cone collapse and repels nerve axons was not identified in the heart. Sema3a, a class 3 secreted semaphorin, was cloned as a potent neural chemorepellent and a directional guidance molecule for nerve fibers (Puschel et al. 1995; Tanelian et al. 1997; Kawasaki et al. 2002). However, it was unknown whether cardiomyocytes produce Sema3a, and if so, whether this protein affects sympathetic neural patterning and cardiac performance.

We analyzed the time course and distribution of cardiac sympathetic innervation in developing mice ventricles (Ieda et al. 2007). TH-immunopositive sympathetic nerve endings appeared at embryonic day (E) 15 in the epicardial surface, and gradually increased in the myocardium after postnatal day (P) 7 and 42. In the ventricular myocardium, sympathetic nerves were richer in the subepicardium than the subendocardium, showing an epicardial-to-endocardial gradient (Crick et al. 1994; Ito and Zipes 1994; Chow et al. 1993; Hansson et al. 1998). To identify Sema3a expression and its relationship with innervation patterning in the heart, we analyzed the heterozygous Sema3a knocked-in lacZ mouse (Sema3a lacZ/+). At E12, lacZ expression was strongly detected in the heart, especially in the trabecular component of both ventricles. In E15 hearts, lacZ expression was observed at the subendocardium but not at the subepicardium in the atria and ventricles. At P1 and P42, lacZ expression was reduced in places and highlighted the Purkinje fiber network along the ventricular free wall (Tago et al. 1986; Kupershmidt et al. 1999). Quantitative RT-PCR of Sema3a in developing hearts also revealed Sema3a was detected from E12, and then decreased linearly in contrast to the sympathetic innervation. These results indicated that Sema3a has the opposite time-course and distribution of expression from sympathetic innervation in developing hearts, suggesting Sema3a might regulate cardiac innervation negatively (Fig. 9.1) (Ieda et al. 2008).

To investigate whether Sema3a is critical for cardiac sympathetic nerve development, we analyzed Sema3a-deficient mice (Sema3a −/−) (Behar et al. 1996; Taniguchi et al. 1997). The wild-type (WT) hearts showed a clear epicardial-to-endocardial gradient of sympathetic innervation; in contrast, sympathetic nerve density was reduced in the subepicardium but increased in the subendocardium in Sema3a −/− mice, resulting in disruption of the innervation gradient in Sema3a −/− ventricles. The Sema3a −/− mice also exhibited malformation of stellate ganglia that extended sympathetic nerves to the heart. To address whether the abnormal sympathetic innervation patterning in Sema3a −/− hearts was a secondary effect of the stellate ganglia malformation, we generated transgenic mice overexpressing Sema3a specifically in the heart (SemaTG) (Gulick et al. 1991). SemaTG mice were associated with reduced sympathetic innervation and attenuation of the epicardial-to-endocardial innervation gradient. These results indicated that cardiomyocyte-derived Sema3a has a critical role in cardiac sympathetic innervation by inhibiting neural growth. As cardiomyocyte-derived NGF acts as a chemoattractant, it might be the balance between NGF and Sema3a synthesized in the heart that determines cardiac sympathetic innervation patterning.

9.4 Sema3a Maintains Arrhythmia-Free Hearts Through Sympathetic Innervation Patterning in Mouse

Most Sema3a −/− died within the first postnatal week; only 20 % remained viable until weaning (Taniguchi et al. 1997; Behar et al. 1996). To identify the cause of death and the effects of abnormal sympathetic neural distribution in Sema3a −/− hearts, we performed telemetric electrocardiography and heart rate variability analysis (Shusterman et al. 2002; Saba et al. 2003). In addition to multiple premature ventricular contractions, Sema3a −/− mice developed sinus bradycardia and abrupt sinus arrest as the result of sympathetic neural dysfunction.

The SemaTG mice died suddenly without any symptoms at 10 months of age. Sustained ventricular tachyarrhythmia was induced only in SemaTG mice but not in WT mice after epinephrine administration, and programmed electrical stimulation also revealed that SemaTG mice were highly susceptible to ventricular tachyarrhythmia (Wehrens et al. 2004; Kannankeril et al. 2006). The β-adrenergic receptor density was upregulated and the cAMP response after catecholamine injection was exaggerated in SemaTG ventricles. Action potential duration was significantly prolonged in hypoinnervated SemaTG ventricles, presumably via ion channel modulation. These results suggested that the higher susceptibility of SemaTG mice to ventricular arrhythmia was at least in part caused by catecholamine supersensitivity and action potential duration prolongation, either or both of which might augment triggered activity in cardiomyocytes (Kuo et al. 2001; Opthof et al. 1991; Priori and Corr 1990; Brunet et al. 2004; Costantini et al. 2005). Thus, Sema3a-mediated proper sympathetic innervation patterning is critical to maintain arrhythmia-free hearts (Fig. 9.2) (Ieda et al. 2008).

Fig. 9.2
figure 2

Regulation of cardiac innervation patterning and arrhythmias. Sema3a-mediated proper sympathetic innervation patterning is critical to maintain arrhythmia-free hearts. Sema3a −/− mice exhibited sinus bradycardia, and SemaTG mice were highly susceptible to ventricular tachyarrhythmias (Ieda et al. 2007, 2008)

Full size image

Sympathetic nerves modulate the function of ion channels and trigger various arrhythmias in diseased hearts (Qu and Robinson 2004; Dae et al. 1997). However, the relationship between sympathetic innervation and arrhythmogenicity in structurally normal hearts was unclear. Sema3a −/− mice exhibited sinus bradycardia, abrupt sinus slowing, and stellate ganglia defects. Consistent with our data, right stellectomy induced sinus bradycardia and sudden asystolic death in dogs (Sosunov et al. 2001). Stramba-Badiale et al. also reported that a developmental abnormality in cardiac innervation might be involved role in the genesis of some cases of sudden infant death syndrome (Stramba-Badiale et al. 1992). The SemaTG hearts were also highly susceptible to ventricular arrhythmias, although without contractile dysfunction or structural defects. Given that catecholamine augments systolic function, it is surprising that SemaTG mice showed normal cardiac function. However, patients who underwent heart transplantation and had denervated hearts did not show heart failure, whereas about 10 % of the patients developed sudden cardiac death, presumably because of arrhythmias (Chantranuwat et al. 2004). These results suggested that regulation of cardiac nerves should be a new paradigm in the management of sudden cardiac death.

9.5 The Crosstalk Between Neurotrophin and Semaphorin Signaling in Nerve Development

Because cardiomyocyte-derived NGF acts as a chemoattractant, it is possible that the balance between NGF and Sema3a synthesized in the heart determines cardiac sympathetic innervation patterning. The growth cone behavior of somatic sensory axons is modulated by coincident signaling between NGF and Sema3a (Tang et al. 2004; Kitsukawa et al. 1997), both of which are expressed within the developing spinal cord and influence pathway guidance of sensory axons during development. Sema3a is specifically expressed in the ventral half of the spinal cord and mediates NGF-responsive sensory axons to terminate at the dorsal part of the spinal cord (Puschel et al. 1995; Wright et al. 1995). In addition, targeted inactivation of Sema3a disrupts neural patterning and projections in the spinal cord, further highlighting the importance of Sema3a signaling for the directional guidance of nerve fibers (Taniguchi et al. 1997; Behar et al. 1996).

However, the interaction of NGF and Sema3A signaling is further complicated as the low-affinity NGF receptor p75 can partner with the Sema3A receptors neuropilin-1 and plexin A4 to attenuate their combined ability to repel growing axons (Ben-Zvi et al. 2007). Lorentz et al. found that the subendocardium innervation of adult p75 knockout ventricles was disrupted, with the left ventricle essentially devoid of sympathetic nerve fibers, whereas innervation density of the subepicardium was normal (Lorentz et al. 2010). This neural patterning defect is similar to that seen in mice overexpressing Sema3A, such that the sympathetic axons lacking p75 are highly sensitive to Sema3A-mediated inhibition of neurite outgrowth. The heterogeneous innervation was also associated with altered cardiac β1-adrenergic receptor expression and sensitivity, and a significant increase in spontaneous ventricular arrhythmias, which were also observed in Sema3A-overexpressing mice. These results suggested that as sympathetic neurons are innervating the heart, p75, a receptor for neurotrophins, acts to blunt the repulsive effects of Sema3A from the subendocardium, thereby allowing axonal arborization and orchestrating a highly organized epicardial-to-endocardial innervation gradient in the ventricle (Carter et al. 2010).

9.6 A Nonsynonymous Polymorphism in Sema3a Is a Risk Factor for Human Unexplained Cardiac Arrest with Ventricular Fibrillation

Unexplained cardiac arrest (UCA) with documented ventricular fibrillation is defined as spontaneous ventricular fibrillation (VF) that is not associated with known structural or electrical heart diseases and is one of the major causes of sudden cardiac death. The relationship between abnormal autonomic nerve activity and lethal ventricular arrhythmias is well known, but the molecular mechanism determining the innervation densities in patients with lethal arrhythmia was not clarified because of the high mortality rate and difficulty of diagnosis. Nakano et al. reported that UCA patients with spontaneous VF had a high incidence of a polymorphism in the semaphorin 3A gene (Nakano et al. 2013). A nonsynonymous polymorphism (I334V, rs138694505A>G) in exon 10 of the SEMA3A gene identified through resequencing was significantly associated with UCA (combined P = 0.0004, OR 3.08, 95% CI 1.67–5.7). Overall, 15.7 % of UCA patients carried the risk genotype G, whereas only 5.6 % of controls did so. In patients with SEMA3AI334V, VF predominantly occurred at rest during the night. They showed sinus bradycardia and sinus dysfunction, and their RR intervals on electrocardiography tended to be longer than those in patients without SEMA3AI334V (P = 0.039), which is consistent with the Sema3A KO mouse phenotype (Ieda et al. 2007). Immunofluorescence staining of cardiac biopsy specimens revealed that sympathetic nerves were absent in the subendocardial layer in normal hearts and in patients without SEMA3AI334V. In patients with SEMA3AI334V, in contrast, sympathetic nerves were distributed in the subendocardial layer. This finding was consistently observed in patients with SEMA3AI334V (n = 4) but not in those without SEMA3AI334V (n = 8), suggesting disruption of the epicardial-to-endocardial transmural sympathetic innervation patterning in the heart of UCA patients with SEMA3AI334V. On the other hand, NGF, a neural attractant factor, was similarly expressed in the subendocardial layer in patients with and without SEMA3AI334V. The dorsal root ganglion (DRG) repulsion assay revealed that the axon-repelling and axon-collapsing activities of mutant SEMA3AI334V genes were significantly weaker than those of wild-type SEMA3A genes. These results indicate that SEMA3AI334V disrupts the SEMA3A function of inhibiting neural growth and impaired appropriate innervation patterning, and that SEMA3AI334V is a risk factor for human UCA and contributes to the pathogenesis of UCA.

9.7 The Gata6/Sema3c Pathway Is Critical for Cardiovascular Morphogenesis in the Mouse

The cardiac neural crest plays a critical role in patterning the vertebrate vascular system (Creazzo et al. 1998). During early embryogenesis (E8.0–9.0 in the mouse), the cardiac neural crest arises from the dorsal neural tube and migrates ventrally to populate the aortic arch arteries and cardiac outflow tract. These neural crest-derived cells respond to poorly characterized developmental cues to differentiate into vascular smooth muscle cells (VSMCs) and contribute to vascular formation during development. Semaphorin 3C is a member of the semaphorin family of secreted factors that are known to be important in regulating neuronal as well as vascular patterning. During murine embryonic development, semaphorin 3C is expressed in the mesenchyme surrounding the developing aortic arch arteries and in neural crest-derived SMCs. These cells also express neuropilin 1 and plexin-A2, the components of a potential heterodimeric semaphorin 3C receptor. Targeted inactivation of semaphorin 3C causes impaired migration of neural crest cells to the developing cardiac outflow tract, resulting in interruption of the aortic arch and persistent truncus arteriosus (Feiner et al. 2001; Brown et al. 2001). It has been proposed that semaphorin 3C acts as a guidance molecule, regulating migration of neural crest cells that express semaphorin receptors such as plexin-A2, but the molecular mechanisms regulating Sema3c expression and disruption of this pathway in human hearts were not clear until recently. Toyofuku et al. reported that the coordination of repulsive cues provided by Sema6A/Sema6B through plexin-A2 paired with the attractive cue by Sema3C through plexin-D1 is required for the precise navigation of migrating cardiac neural crest cells to the cardiac outflow tract (Toyofuku et al. 2008).

Lepore et al. demonstrated that GATA-6 regulates semaphorin 3C expression in the neural crest cells, and this pathway regulates morphogenetic patterning of the aortic arch and cardiac outflow tract during development in the mouse (Lepore et al. 2006). Conditional inactivation of GATA-6 in vascular smooth muscle cells resulted in perinatal mortality from a spectrum of cardiovascular defects, including interrupted aortic arch and persistent truncus arteriosus. Inactivation of GATA-6 in neural crest cells recapitulated these abnormalities, demonstrating a cell-autonomous requirement for GATA-6 in neural crest-derived SMCs. Importantly, deletion of GATA-6 in neural crest cells resulted in an absence of detectable semaphorin 3C- or plexin-A2-expressing cardiac neural crest cells within the conotruncal endocardial cushions that contribute to aorticopulmonary septation. These results are consistent with the phenotypes observed in semaphorin 3C-deficient mice (Feiner et al. 2001; Brown et al. 2001), the impairment of neural crest cell migration into the cardiac outflow tract.

9.8 Gata6 Mutations Cause Human Congenital Heart Disease by Disrupting Sema3c/Plexin-A2 Signaling

Congenital heart diseases (CHD) occur in nearly 1 % of all live births and are the major cause of infant mortality and morbidity (Hoffman and Kaplan 2002). Although an improved understanding of the genetic causes of CHD would provide insight into the underlying pathobiology, the genetic etiology of most CHD remains unknown. Cardiac outflow tract (OFT) defects are estimated to account for approximately 30 % of CHD and usually require an intervention during the first year of life. Kodo et al. reported the identification and characterization of two mutations of GATA6 DNA (9.5 %) in their series of 21 Japanese patients with persistent truncus arteriosus (PTA), a form of OFT defect (Kodo et al. 2009). Both mutations disrupted the transcriptional activity of the GATA6 protein on downstream target genes involved in the development of the outflow tract. They also confirmed that the expression of SEMA3C and PLXNA2 in the developing OFT was regulated directly through the consensus GATA binding sites conserved between human and mouse, in vitro and in vivo. Mutant GATA6 proteins failed to transactivate SEMA3C and PLXNA2, and mutation of the GATA sites on enhancer elements of Sema3c and Plxna2 abolished their activity, suggesting that mutations of GATA6 cause specific forms of human CHDs. These results are consistent with the mouse phenotype observed in semaphorin 3C-deficient mice and conditional inactivation of GATA-6 in neural crest cells and VSMCs, as previously discussed.

9.9 Conclusions

The heart consists of many types of cells, including cardiomyocytes, vascular cells, and neural cells, which tightly regulate heart development and cardiac function. Semaphorin family members are critical in the interactions between cardiomyocytes and other types of cells (Fig. 9.3). Cardiac nerves are highly plastic, and innervation patterning is strictly controlled by the balance between NGF and Sema3a synthesized in the heart. Sema3a inhibits neural growth and establishes proper innervation patterning in the heart. The disruption of sympathetic innervation patterning by Sema3a misexpression may lead to sudden cardiac death and fetal arrhythmias in both mouse and human (Ieda et al. 2007; Nakano et al. 2013). Cardiac outflow tract formation is organized by the interaction of migrating cardiac neural crest cells and surrounding tissues. The GATA6/Sema3c signaling pathway is conserved among mammals, and its disruption results in congenital heart diseases in humans. Understanding of the divergent effects of semaphorin family members on cardiac innervation and cardiovascular patterning might thus represent a novel step toward potential therapies for heart diseases.

Fig. 9.3
figure 3

Summary of cardiac defects mediated by Sema3a and Sema3c signaling. Sema3a is important for cardiac sympathetic innervation and maintains arrhythmia-free hearts, and Gata6/sema3c is critical for proper cardiac outflow tract formation (Ref. indicates references in the text)

Full size image