Temporal–spatial ablation of neural crest in the mouse results in cardiovascular defects

Authors


Abstract

Neural crest cells are thought to play a critical role in human conotruncal morphogenesis and dysmorphogenesis. Much of our understanding of the contribution of neural crest to cardiovascular patterning comes from ablation and transplantation experiments in avian species. Although fate mapping experiments in mice suggests a conservation of function, the functional requirement for neural crest in cardiovascular development in mammals has not been formally tested. We used a novel two component genetic system for the temporal–spatial ablation of neural crest in the mouse. Affected embryos displayed a spectrum of cardiovascular outflow tract defects and aortic arch patterning abnormalities. We show that the severity of the cardiovascular phenotype is directly related to the level and extent of neural crest ablation. This is the first report of cardiac neural crest ablation in mammals, and it provides important insight into the role of the mammalian neural crest during cardiovascular development. Developmental Dynamics 237:153–162, 2008. © 2007 Wiley-Liss, Inc.

INTRODUCTION

Conotruncal heart defects account for approximately 50–60% of cardiac malformations diagnosed in the neonatal period (Iserin et al.,1998) and are associated with significantly high morbidity and mortality (Ferencz et al.,1985; Ferencz,1990). This category includes defects such as truncus arteriosus (TA), double outlet right ventricle (DORV), tetralogy of Fallot (TOF) with or without pulmonary atresia, and interrupted aortic arch (IAA).

The morphogenesis of the conotruncus or outflow tract (OFT) is a complex process that involves direct contributions by several cell populations. These include the secondary heart field (from which the myocardial cells of the OFT are derived), the pharyngeal endoderm, mesoderm, and ectoderm, as well as the pharyngeal arch mesenchyme, which is in large part derived from the neural crest. The neural crest is a population of pluripotent cells that originates from most of the craniocaudal length of the neural tube and migrates to various locations during early embryonic development (Kirby,1999). Neural crest contribution to the cardiovascular system was first demonstrated in avian models (Le Lievre and Le Douarin,1975). Margaret Kirby and colleagues localized the area of neural crest that contributes directly to cardiovascular morphogenesis to the region between the mid-otic placode and somite 3 (termed the cardiac neural crest; Kirby et al.,1983,1985) and demonstrated its functional requirement by ablating the premigratory neural crest (Kirby et al.,1983). This ablation results in a very characteristic phenotype, which includes OFT defects and aortic arch patterning abnormalities, as well as abnormalities of the development of the thymus, parathyroid, and thyroid glands (Kirby,1999; Hutson,2003). Similar sets of abnormalities are seen in some human patients born with congenital heart disease, which led to the hypothesis that alterations in the biology of the neural crest were responsible for these defects in humans (Kirby and Bockman,1984; Van Mierop and Kutsche,1986).

Mammalian and avian cardiovascular patterning differ in many important respects. Most notably, chickens have a right-sided aortic arch, bilateral ducti arteriosi, and bilateral brachiocephalic arteries (Brown and Baldwin,2006). Cell labeling and fate mapping experiments in mice (Waldo et al.,1999; Yamauchi et al.,1999; Jiang et al.,2000; Li et al.,2000; Pietri et al.,2003) and rat (Fukiishi and Morriss-Kay,1992) suggest a conservation of function of the neural crest in mammalian cardiovascular development. There are also multiple genetic and teratogenic alterations in mouse and human that result in a similar phenotype to that of chick neural crest ablation (NCA). However, the functional requirement for the mammalian neural crest has not been formally tested by cell ablation. We therefore set out to test the hypotheses that the murine neural crest plays an important role in the remodeling of the cardiac OFT and pharyngeal arch arteries and that its ablation would result in conotruncal heart defects and aortic arch patterning abnormalities.

We used the PuΔTK selector mouse line (Chen et al.,2004) in order to ablate neural crest cells in mouse embryos at critical time points during cardiovascular development. This mouse line expresses a truncated version of the herpes simplex virus-1 thymidine kinase (TK) after Cre-recombination. Ganciclovir (GCV) inhibits DNA synthesis in cells that express TK by competitively inhibiting the incorporation of guanosine into the elongating DNA, resulting in cell death (Chen et al.,2004). To drive expression of Cre recombinase in neural crest, we used the Wnt1-Cre mouse line (Danielian et al.,1998). Using this two-component genetic system, we were able to ablate neural crest at specific levels and to different extents, depending on the timing of GCV administration and the number of doses administered. We show that NCA in the mouse results in a wide spectrum of cardiovascular and aortic arch patterning defects and that the severity of these defects depends on the level and extent of ablation of the neural crest.

RESULTS

Neural Crest Cells Are Temporally–Spatially Ablated in PuΔTK:Wnt1-Cre Embryos Exposed to GCV In Utero

To determine whether exposure of PuΔTK:Wnt1-Cre embryos to GCV would result in NCA and the optimal timing of injection to specifically target cardiac neural crest, we performed Lysotracker Red (LTR) staining on these embryos after completing specific GCV injection schemes (see the Experimental Procedures section; Fig. 1). LTR accumulates within dead and dying cells, allowing us to identify areas of cell death in whole-mount embryos, and has been used previously to study apoptosis in the embryonic mouse (Zucker et al.,1998; Watanabe et al.,2002).

Figure 1.

Neural crest ablation in PuΔTK:Wnt1-Cre embryos. A,B,E,H: Wnt1-Cre:R26R fate-mapped embryos define neural crest streams. C,D,F,G,I,J: LTR stained PuΔTK:Wnt1-Cre embryos. C. Dorsal view of cell death in an embryonic day (E) 9.0 PuΔTK:Wnt1-Cre embryo that received a single ganciclovir (GCV) dose at E7.5. C,F: Cardiac neural crest is between the white arrows. Ablated neural crest cells (white arrow head in C) are cephalic to the cardiac neural crest. Cell death corresponds with neural crest present at the time of injection (E7.5, A). D: Left lateral view of the embryo in C. Ablated neural crest cells are visible in the first pharyngeal arch (PA1). F: Dorsal view of E9.5 PuΔTK:Wnt1-Cre embryo after GCV injections at E7.5 and E8.5. The cardiac neural crest displays extensive cell death. G: Right lateral view of embryo in F, showing streams of ablated cells in the proximal and caudal (asterisk) pharyngeal arches; PA2, second pharyngeal arch. I: Right lateral view of E10.5 PuΔTK:Wnt1-Cre embryo after receiving GCV at E7.5, E8.5, and E9.5. J: Close-up of same embryo as in I, showing extensive cell death in pharyngeal arches.

We found that PuΔTK:Wnt1-Cre embryos that were exposed to GCV at or before 7.0 days post coitum (dpc) did not experience NCA, and were indistinguishable from wild-type (WT) littermates (data not shown). Wnt1 and the Wnt1-Cre transgene expression starts at the 1 somite stage (approximately 7.5 dpc) at the midbrain level and progresses caudally (Fig. 1A), so that, by 8.5 dpc, the level of expression includes the neural crest in its entirety (Fig. 1B; Echelard et al.,1994). When a single GCV dose was given at 7.5 dpc (single-injection group), and embryos were harvested at 9 dpc, neural crest cell ablation was evident. Ablation was limited to the most cephalic portion of the neural tube and the midbrain (Fig. 1C), as well as migrating cells at the level of the first pharyngeal arch (Fig. 1D). The single 7.5 dpc injection scheme did not include the cardiac neural crest, which is located between the mid-otic placode and the third somite (between white arrows in Fig. 1C,F,G). This finding is consistent with a later generation of the cardiac neural crest at 8–8.5 dpc.

It is important to note that the single injection at embryonic day (E) 7.5 limits the ablation to neural crest cephalic to the cardiac neural crest because the Cre-transgene has activated the PuΔTK allele only in the most cranial neural crest precursors. GCV injections at E8.5 and beyond can potentially ablate neural crest at all axial levels, because once Cre-induced recombination occurs in neural crest precursors, all of the daughter cells express TK and will be susceptible to GCV while undergoing mitosis. When a second dose of GCV was administered at 8.5 dpc (two-injection group) neural crest ablation was observed at 9.5 dpc extending down almost the full length of the neural tube, including the cardiac neural crest (Fig. 1F). The two-injection scheme also resulted in ablation of neural crest cells that had migrated to the first three pharyngeal arches by 9.5 dpc (Fig. 1G). When daily GCV injections were given on dpc 7, 8, and 9 (three-injection group), the extent of ablation of migrating neural crest cells was increased, affecting higher numbers of cells in the caudal pharyngeal arches (Fig. 1I, J).

When comparing the pattern of cell death observed using LTR staining with the location of neural crest cells as evidenced by LacZ-stained Wnt1Cre:R26R WT embryos (neural crest fate-mapped embryos; Fig. 1A,B,E,H), it is evident that the pattern of cell death very closely correlates with the location of neural crest cells at the time of injection and lags slightly behind their location at the time of harvest (Fig. 1). Comparison of LTR staining in Wnt1Cre::PuΔTK embryos to simultaneously stained WT littermates confirmed that cell death was restricted to areas of induced cell death in the neural crest migration streams as well as areas with known high levels of apoptosis such as the otic vesicle and OFT myocardium (data not shown). In all cases, cell death was consistent both with the cephalocaudal activation of the Wnt1 gene (Fig. 1A,B,E,H; Echelard et al.,1994), and the relatively short half life of GCV (Crumpacker,1996). The phenotypes observed were limited to known functions of neural crest, and WT littermates were never affected. These observations demonstrate that the cell death induced was highly specific for the neural crest cell population and that the GCV treatment did not produce generalized cell toxicity.

The PuΔTK cassette is targeted into Hprt locus on the X chromosome. Random X chromosome inactivation in females results in a mosaic pattern of partial ablation (Chen et al.,2004). Male embryos injected on E7.5 and E8.5 and harvested at E9.5 displayed more extensive NCA and a more severe craniofacial phenotype than female littermates (compare Fig. 2B with 2C). To determine to what extent gender affects NCA in the pharyngeal arches and conotruncus during remodeling, we stained E10.5 PuΔTK:Wnt1Cre:R26R embryos from the three-injection group with X-gal and compared the staining pattern to Wnt1-Cre:R26R littermates (Fig. 2D–L). In the WT embryos, the pharyngeal arch mesenchyme was derived from NC fate mapped cells (Fig. 2D,G). Female embryos displayed a mosaic pattern with reduced labeled cells (Fig. 2E,H), whereas the male ablated embryos (Fig. 2F,I) had very few fate mapped cells of normal appearance in the pharyngeal arch mesenchyme. Residual punctate blue staining (Fig. 2F,I) represents dead cells or cell fragments. The cellularity of the OFT cushion mesenchyme is decreased in the NCA embryos as compared to WT (Fig. 2J–L). The WT embryos had more Lac-Z–positive cells in the cushion mesenchyme at E10.5 (Fig. 2J) than the female NCA embryos (Fig. 2K), whereas the NCA male embryos have very few Lac-Z–positive cells (Fig. 2L).

Figure 2.

Sex differences and efficiency of neural crest ablation. AC: Right lateral view of cell death in 9.5 days post coitum (dpc) embryos wild-type (WT; A), or PuΔTK female (B), or PuΔTK male (C) exposed to ganciclovir (GCV) at embryonic day (E) 7.5 and E8.5 and stained with Lysotracker Red (LTR). There is a higher amount of cell death in the male than the female, as evidenced both by a higher level of LTR staining and a more severe craniofacial phenotype. DF: Eosin stain of frontal sections of 10.5 dpc WT (D) and neural crest ablated female (E) and male (F) fate mapped embryos (Wnt1-Cre: R26R showing second, third, and fourth pharyngeal arches [II, III, and IV], respectively). G–I. Higher magnification of left third pharyngeal arch from D to F (respectively). J–L: Eosin stain of frontal sections through the outflow tract of embryos in A–C (respectively). Ca, caudal; Cr, cranial; L, left; R, right. There are fewer fate mapped neural crest cells(blue stain) in the pharyngeal arches (H) and outflow tract (K) of the female neural crest ablated embryo. Significant depletion of neural crest cells is observed in the male embryo (I,L) with residual dead cells observed as punctate staining.

Ablation of Neural Crest in the Mouse Embryo Results in a Wide Spectrum of Cardiovascular Defects

To evaluate the cardiovascular phenotypes associated with cardiac neural crest ablation, we examined PuΔTK:Wnt1-Cre embryos from all three injection groups by whole-mount and histological examination, and compared them to their WT littermates. A wide spectrum of cardiovascular malformations was evident in the ablated embryos. Subtle defects in OFT elongation and looping were noticed in the ablated embryos as early as E9.5. At this age, the OFT and right ventricle are underdeveloped and assumed a more dorsal position as compared to WT littermates (Fig. 2). At E13.5 or older, malformations included ventricular septal defects (VSDs) with normally aligned great arteries, VSDs associated with varying degrees of dextroposition of the aorta (DPA), DORV, and TA. Aortic arch patterning defects commonly seen included: right aortic arch (RAoA), aberrant subclavian artery (ASCA), and interrupted aortic arch (IAA). PuΔTK:Wnt1-Cre embryos and WT littermates in the three-injection group were represented at the expected 1:1 ratio until 11.5 dpc. PuΔTK:Wnt1-Cre embryos in this group are underrepresented by 7–17% in litters harvested beyond 12.5 dpc. The cause of embryonic death in this subset of embryos appears to be congestive heart failure, as demonstrated by findings such as generalized edema, enlarged heart, pericardial effusion, and small size for gestational age, which were evident as early as 10.5 dpc. In addition to cardiovascular defects, the thymus also formed abnormally with defects including hypoplasia, single lobe, asymmetric lobes, and absent thymus (data not shown).

Misalignment of the great arteries was a common defect observed in NCA embryos. Normally by 14.5 dpc, the aortic valve has come into direct fibrous continuity with the mitral valve (Fig. 3B) and has connected with the left ventricle (Fig. 3C). NCA embryos displayed varying degrees of dextroposition of the aorta, resulting in either straddling of the aortic valve over the interventricular septum (Fig. 3E; which we call VSD with dextroposition of the aorta, abbreviated as DPA) or, if severe enough, in the aortic valve being connected only to the right ventricle (Fig. 3F; which we call DORV). These malformations represent different points in the same spectrum. The type of DORV seen was always with the aorta to the right of the pulmonary artery, with lack of semilunar valve to atrioventricular valve fibrous continuity (bilateral conus), and the VSD was either uncommitted (not directly under either semilunar valve) or subaortic (directly under the aortic valve). DORV of the Taussig-Bing type (with VSD directly under the pulmonary valve) was not encountered in this series. Within the group classified as DORV, there were also varying degrees of dextroposition of the aorta. On one side of the spectrum were embryos with side-by-side great arteries (Fig. 4G; n = 4 of 14, or 29%) and on the other, embryos with the aorta clearly posterior and to the right of the pulmonary artery (Fig. 3D,F; n = 10 of 14, or 71%). Some of the hearts with DORV exhibited significant stenosis or hypoplasia of the pulmonary OFT and pulmonary valve (Fig. 4H,I; n = 4 of 14 or 29%). Three of these hearts with DORV with pulmonary stenosis had the aortic valve positioned posterior and to the right of the pulmonary valve, reminiscent of cases of TOF. TOF is defined as the combination of subpulmonary infundibular stenosis, conoventricular septal defect, overriding aorta with aortic valve to mitral valve fibrous continuity, and right ventricular hypertrophy (Van Praagh et al.,1982). In all of our specimens, the aortic valve was separated from the mitral valve by a band of muscle (lack of aortic valve to mitral valve fibrous continuity), making the diagnosis of DORV more adequate. However, since these embryos were harvested at E14.5, it is conceivable that, if allowed to continue developing, some of the cases of DPA might actually develop into TOF, so we cannot firmly conclude that TOF is not part of the spectrum of malformations related to NCA.

Figure 3.

Ventricular septal defect (VSD) with dextroposition of the aorta. A–C: Wild-type (WT) 14.5 days post coitum (dpc) hearts. A: Whole-mount right lateral view after removal of the right atrium, showing proximity of the aortic valve (AoV) with the tricuspid valve (TV). B: Hematoxylin and eosin (H&E) stain of transverse section showing AoV to mitral valve (MV) fibrous continuity. C: H&E of transverse section showing characteristic angle formed between the long axis of the interventricular septum (IVS) and the long axis of the proximal ascending aorta (gray lines). D–F: Three different PuΔTK:Wnt1-Cre 14.5 dpc embryos with varying degrees of dextroposition of the aorta. D: Whole-mount right lateral view after removal of the right atrium, showing wider distance between AoV and TV (as compared to A). E: H&E of transverse section of heart specimen with VSD and dextroposition of the aorta, showing myocardium (black arrow head) interposed between the MV and the AoV, resulting in lack of AoV to MV fibrous continuity. Note that the aortic valve also overrides the IVS. F: H&E of transverse section showing the AoV in direct relation to the RV. Note that in E and F the long axis of the proximal ascending aorta is approximately parallel to that of the IVS (gray lines). PV, pulmonary valve.

Figure 4.

Double outlet right ventricle (DORV) and truncus arteriosus (TA) in neural crest ablated embryos. AC: Ink injections of embryonic day (E) 13.5, E14.5, and E15.5 wild-type (WT) embryos presented for comparison. D: X-gal stain of neural crest ablation (NCA) -Sm22α-LacZ E13.5 embryo showing TA type A4, interrupted aortic arch type C, and an aberrant right subclavian artery (ARSC) traversing behind the trachea (t). E: E14.5 NCA embryo with similar phenotype as D, showing the shortening of the myocardial OFT (asterisk) between E13.5 and E14.5. F: Ink injection of E15.5 NCA embryo with TA type A2 with right aortic arch. G–I: E14.5 NCA embryos with DORV. G: Side by side great arteries of equal dimensions. H: The aorta (Ao) is larger, the right ventricular outflow tract (RVOT) is hypoplastic. I: Frontal section through semilunar valves of NCA-R26R stained with X-gal. The pulmonary valve (PV) is significantly smaller than the aortic valve (AoV). Note that surviving neural crest cells (blue cells) form the septum between the semilunar valves. A, anterior; L, left; P, posterior; R, right; LV, left ventricle; RV, right ventricle; AsAo, ascending aorta; BCA, brachiocephalic artery; PA, pulmonary artery; LCC, left common carotid; DA, ductus arteriosus; RCC, right common carotid; LSCA, left subclavian artery; RSCA, right subclavian artery.

In NCA embryos of the three-injection series, a commonly observed cardiovascular defect was TA, in which one great artery arises from the base of the heart and gives origin to the coronary, pulmonary, and systemic arteries (Abbott,1927; Lev and Saphir,1942; Collett,1949; Van Praagh and Van Praagh,1965; Fig. 5D–F). The type of TA most commonly seen in this study was type A4 (n = 29 of 40 or 73%) of the Van Praagh classification (Van Praagh and Van Praagh,1965). This type of TA is characterized by a small ascending aortic component, a large ductus arteriosus, and underdevelopment of the arch, varying from arch hypoplasia to interruption. In this study, all cases of type A4 TA were associated with interrupted aortic arch (IAA; Fig. 4D,E). Two different types of IAA were seen: type B (interruption between the take-off of the left common carotid artery and the left subclavian artery; 3 of 29, or 10%) and type C (interruption between the brachiocephalic artery and the left common carotid artery (Fig. 4D,E; 26 of 29, or 90%). Other types of TA seen in this study included A1 and A2. TA type A2 (4 of 40, or 10%; Fig. 4F), is characterized by an absent ductus arteriosus and a large ascending aorta and aortic arch components (Van Praagh and Van Praagh,1965). TA type A1 was seen in 17% of cases of TA (7 of 40). This type of TA is characterized by a partially formed aorticopulmonary septum separating the common trunk into aortic and pulmonary components distally. All of these were characterized by a much larger aortic component, and 6 of the 7 had absence of the ductus arteriosus. The embryos harvested at E13.5 and E14.5 have an elongated muscular OFT. This OFT seems to recede between E14.5 and E15.5, consistent with the pattern of OFT shortening seen in WT murine embryos (Barbosky et al.,2006; compare Fig. 4A–C with 4D–F).

Figure 5.

Severity of cardiovascular phenotype is related to extent of neural crest ablation. A–C: Pie charts illustrating the distribution of cardiac defects by injection scheme. A: Group of embryos that received a single ganciclovir (GCV) injection at embryonic day (E) 7.5 (single injection group; n = 21). B: Group of embryos that received GCV injections at E7.5 and E8.5 (two-injection group; n = 25). C: Group of embryos that received GCV injections at E7.5, E8.5, and E9.5 (three-injection group; n = 20). The penetrance and severity of the defects increases with increasing number of injections. D,E: 100% stacked bar charts illustrating comparison of male vs. female phenotypes. In the two- and three-injection groups (2-inj and 3-inj, respectively), the males tend to have a more severe and penetrant cardiovascular phenotype. D: Cardiac phenotype by injection scheme and sex (2-inj group, female n = 8 and male n = 17; 3-inj group, female n = 9, male n = 11). E: Aortic arch defects by injection scheme and sex (2-inj group, female n = 9 and male n = 18; 3-inj group, female n = 11, male n = 14). VSD, ventricular septal defect; DPA, dextroposition of the aorta; DORV, double outlet right ventricle; TA, truncus arteriosus; N1, normal aortic arch anatomy; RAoA, right aortic arch; ASCA, aberrant subclavian artery; IAA, interrupted aortic arch.

Severity of Cardiovascular Phenotype Is Dependent Upon the Timing and the Extent of NCA

To determine whether the segmental level and extent of NCA had an effect on the resultant cardiovascular phenotype, we compared the defects seen in each of the three injection groups in female and male littermates. The severity of cardiovascular phenotypes correlated directly with the extent of GCV exposure (number of injections) and gender of the embryo observed (Fig. 5). PuΔTK:Wnt1-Cre embryos from the 7.5 dpc single-injection group, affecting only cranial neural crest, had a normal cardiovascular phenotype 52% of the time (n = 21). The defects that were observed in this group consisted of VSDs and malalignment defects (Fig. 5A). Embryos in the two-injection group exposed to GCV on 7.5 and 8.5 dpc, affecting both cranial and cardiac neural crest (Fig. 5B), exhibited a more severe cardiovascular phenotype, including TA in 8% (n = 25). The three-injection group of embryos was exposed to GCV on 7.5, 8.5, and 9.5 dpc (Fig. 5C) and exhibited the most severe phenotype, with 75% having TA, 15% DORV, and 10% DPA (n = 20), consistent with significant ablation of the cardiac neural crest. None of the ablated embryos in the three-injection group had a normal cardiovascular phenotype. As previously discussed, female embryos undergo partial neural crest ablation, and this finding was also evident when comparing their phenotypes to those of male littermates (Fig. 5D,E). In the two-injection group, only the male embryos exhibited TA and aortic arch anomalies. Most noteworthy, in the three-injection group, the male embryos exhibited TA and aortic arch anomalies 100% of the time (n = 11).

Morphology of the Truncal Valve

To evaluate the morphology of the truncal valve in the cases of TA, we histologically evaluated 12 truncal valves from embryos ranging in gestational age from E12.5 to E15.5. All of the truncal valves examined had three cusps of approximately the same size (Fig. 6E,G,H), and 3 of the 12 embryos had a fourth cusp that was markedly hypoplastic (Fig. 6G). It was evident that the OFT as a whole failed to expand. Failure of OFT expansion resulted in one of the sides of the truncal valve (aortic side in 7/12 specimens, and pulmonary side in 5/12 specimens) and its corresponding half of the unseptated infundibulum being hypoplastic (Fig. 6G) or atretic (Fig. 6E,F,H). The majority of the truncal valve leaflets had a normal appearance when compared to semilunar valves of WT littermates (compare Fig. 6D and H). Consistent with previous reports, when we fate-mapped neural crest cells in E14.5 WT Wnt1-Cre:R26R embryos and their NCA littermates, we found that, in the WT embryos, there is extensive contribution of neural crest cells to the semilunar valve leaflets, especially the septal leaflets (Fig. 6I; Waldo et al.,1999; Jiang et al.,2000; Nakamura et al.,2006). The leaflets of the NCA embryos had a significantly decreased number of neural crest fate-mapped cells in the case of the female (Fig. 6J), and virtually no neural crest fate-mapped cells in the case of the male (Fig. 6K), indicating that neural crest are not essential to valve leaflet formation.

Figure 6.

Neural crest ablated embryos have failure of expansion of the outflow tract and a tricuspid truncal valve. Hematoxylin and eosin (H&E) stains of frontal sections through the outflow tract (OFT) of wild-type (WT; A–D) and neural crest ablation (NCA; E–H) embryos. A,B,E,F: Embryonic day (E) 12.5 embryos. A: AoV is posterior and to the right of the pulmonary valve (PV), both with three cusps, one formed by the corresponding intercalated cushion (asterisk), and the other two by the ipsilateral half of the truncal cushions (unmarked). E. There is only one intercalated cushion (asterisk; left sided). B: Subvalvar section of embryo in A showing adequate expansion of the left ventricular OFT (LVOT) and the right ventricular OFT (RVOT). MV, mitral valve. E,F: Subvalvar section through conus of same embryo as E showing the right side of the conus is obliterated. C,G: E13.5 embryos, showing lack of expansion in NCA OFT as a whole (G). The left-sided intercalated cushion is exceedingly small (black arrow), and the left half of each truncal cushion is hypoplastic. D,H: E14.5 embryos, showing the truncal valve of normal appearance as compared to a semilunar valve of a WT embryo of the same age (D). I–K: Eosin stain of a transverse section through the OFT of E14.5 neural crest fate-mapped embryos. I: Pulmonary valve of wild-type (WT) embryo. There is extensive contribution of neural crest cells (blue cells) to the valve leaflets, especially the septal leaflets. J: The truncal valve of an NCA female embryo shows the presence of some neural crest fate-mapped cells, but the number is greatly reduced as compared to WT. K: Truncal valve of NCA male embryo shows virtually no blue cells in the valve leaflets. A, anterior; L, left; P, posterior; R, right.

DISCUSSION

In the present study, we demonstrate that a novel two component genetic system for the temporal–spatial ablation of neural crest in the mouse provides a useful model for the evaluation of neural crest cell function in vivo. Neural crest cell ablation using this model results in a wide spectrum of cardiac OFT abnormalities and aortic arch defects, the severity of which was determined by the timing and number of GCV injections administered. The ability to perform time-sensitive neural crest ablation is a unique characteristic of this system that allowed us to correlate minor alignment defects with ablation of cranial neural crest and severe patterning and septation defects with ablation of the cardiac neural crest. The data presented here provide a rough benchmark that may be used to estimate the extent of neural crest perturbation in other models of neural crest dysfunction. These data represent the cardiovascular manifestations of severe depletion of the cardiac neural crest. This system is amenable to future experiments to address more narrowly defined questions of the functions of cardiac neural crest cell subpopulations.

Our data suggest that the severity of cardiovascular phenotypes observed is dependent on the number of neural crest cells ablated. Wnt1-Cre activity on day 8.5 should activate the PuΔTK transgene in all cardiac neural crest populations. Thus, we observed cardiovascular phenotypes reminiscent of results from chicken ablation studies with GCV exposure at 8.5 dpc. GCV injection on E9.5 results in increased cell death and increased penetrance and severity of the cardiovascular defects observed in the two-injection series. The mosaic ablation seen as a result of the X-linked nature of the transgene provides further support for the hypothesis that severe cardiovascular phenotypes result from higher ablation efficiency. All of the males in the three-injection group had TA and aortic arch patterning abnormalities, while only 44% of the female embryos from the same litters exhibited these phenotypes. Since X-inactivation should be a random event within the neural crest population, approximately 50% of the neural crest cells in females should have an active PuΔTK transgene. Neural crest fate mapping confirmed higher levels of neural crest cell death in male embryos. A similar conclusion regarding the relation of the extent of neural crest ablated to the severity of the cardiovascular phenotype has been reached in the chick model. When single somite-length pieces of premigratory neural crest were removed unilaterally, the spectrum of cardiovascular defects observed included a high incidence of DORV and VSDs, but TA was only rarely seen (Besson et al.,1986), whereas excision of the entire cardiac neural crest results in TA 90% of the time in the chick NCA model (Kirby,1999). This finding is also consistent with mouse models of neural crest dysfunction that result in TA, despite neural crest cells migrating normally, but in reduced numbers and lacking proper positional identity (Epstein et al.,2000; Kochilas et al.,2002). These data taken together suggest that, while positional identity and other cell intrinsic and extrinsic factors are important in the normal function of the cardiac neural crest, alteration of cell number is sufficient to generate the spectrum of commonly observed cardiovascular defects attributed to neural crest dysfunction.

Cardiovascular Phenotype in Murine and Avian Models of Neural Crest Ablation

In the present study, some NCA embryos die in utero after E12.5 with heart failure. Myocardial dysfunction in the chick after NCA has been found to be secondary to depressed L-type calcium current, excitation contraction coupling, and calcium sensitivity of the contractile apparatus (Creazzo et al.,1998). A similar myocardial dysfunction has been documented in NCA zebrafish (Li et al.,2003) and in the Pax3 mutant Splotch mouse (Conway et al.,1997a). Splotch mice have many phenotypic features that resemble the chick model of NCA and also die in utero mid-gestation around day 14 secondary to heart failure (Conway et al.,1997b). Thus, neural crest-mediated myocardial dysfunction is the likely cause of death due to heart failure in NCA mouse embryos.

Although the cardiovascular phenotype of the NCA mouse embryos described here correlates well with that described for NCA chick embryos, we also observed several differences between the phenotypes observed in the chick and mouse. We did not observe any inflow tract malformations in the present study. Inflow defects such as tricuspid atresia, straddling tricuspid valve, and double inlet left ventricle have been reported after cardiac neural crest ablation in the chick (Besson et al.,1986; Nishibatake et al.,1987). Another important difference is that we see cases of pulmonary valvar and infundibular stenosis (Fig. 5H,I). It has been recently observed that this is not part of the chick NCA phenotype, pulmonary stenosis being considered more characteristic of defects in secondary heart field function (Hutson,2007). Whether these phenotypes represent true species-related differences or an inherent characteristic of our system remains to be determined.

Potential Nonautonomous Effects of NCA

The neural crest ablated after a single GCV injection at E7.5 is cranial to the cardiac neural crest (Fig. 1C) and does not contribute directly to the morphogenesis of the cardiovascular system (Kirby et al.,1983). That we see VSDs and misalignment defects in these embryos suggests that this region of the NC contributes to OFT morphogenesis in some indirect fashion. Similarly, ablation of NC cranial to the otic placode in chick embryos has been found to result in similar cardiovascular anomalies in 40–60% of the embryos (Kirby et al.,1985; Nishibatake et al.,1987). Early alterations in looping and OFT elongation (before the neural crest is known to colonize the OFT) documented in the chick NCA model and observed in the present study are also consistent with such an indirect effect. Recent studies in the chick have shown that ablation of the cardiac neural crest results in failure of the myocardial component of the secondary heart field to migrate into the OFT (Yelbuz et al.,2002,2003; Waldo et al.,2005), suggesting that in addition to the direct role the cardiac neural crest plays in OFT septation, it also plays an indirect role in OFT morphogenesis by modulating secondary heart field function. Interestingly, in the present study, when partial cardiac neural crest ablation took place (two-injection group, and female embryos in three-injection group), we also observed a wide spectrum of OFT misalignment defects reminiscent of those seen when secondary heart field function is disturbed (Vitelli et al.,2002; Brown et al.,2004; Xu et al.,2004).

Our understanding of neural crest contribution to cardiovascular development has increased greatly in the past 30 years. The primary challenges now facing the field involve deciphering the complex reciprocal signaling events between the cardiac neural crest and the myriad of cell populations with which they interact, and in deciphering the pathway relationships between the ever expanding list of genes with cardiac neural crest associated phenotypes. We believe the model presented here will allow us to evaluate alterations in gene expression in the secondary heart field and OFT in the setting of varying degrees of NCA. These ongoing and future experiments will provide further insight into the mechanisms by which the neural crest affects cardiovascular development, and the ways in which it interacts with the different cell populations involved in this complex process.

EXPERIMENTAL PROCEDURES

Transgenic Mouse Lines and Genotyping

We used the PuΔTK selector mouse line, which was obtained from Dr. Allan Bradley (Chen et al.,2004). Briefly, the transgene was constructed so that the PuΔTK cassette was downstream of the PGK promoter and a floxed neomycin cassette was located in between them. The construct was targeted into exon 3 of the Hprt gene, which is located on the X chromosome.

We also used the Wnt1-Cre transgenic line (Danielian et al.,1998), and the R26R conditional reporter line (Soriano,1999). The Sm22α-LacZ knockin line (Zhang et al.,2001) was used with the purpose of facilitating visualization of the arterial vessels. The mice were genotyped according to published protocols, except for the PuΔTK selector mouse line, for which we developed the following genotyping primers (from 5′ to 3′): Hprtexon3rev, CTTACACAGTAGCTC TTCAGTCTG; Hprtexon3for, TTTCTATAGGACTGAAAGACTTGC; and PgkNeoPromRev, GTGCTACTTCCATTTGTCACGTCC, which yield a WT band of 187 bp, and a target band of 402 bp. Since the PuΔTK transgene is X-linked we also genotyped embryos for sex according to published protocols (Juriloff et al.,2006).

Mice were maintained in a mixed background, and a 0600 to 1800 light–dark cycle was used. Noon of the day of observation of a vaginal plug was defined as E0.5. All procedures and animal experiments were approved by the Vanderbilt University Institutional Animal Care and Use Committee.

GCV Administration

GCV (Roche, CA) was administered daily at the desired time points at a dose of 100 mg/kg by intraperitoneal injection. Three injection schemes were used: single injection at 7.5 dpc (single-injection group); injections at 7.5 and 8.5 dpc (two-injection group); injections at 7.5, 8.5, and 9.5 dpc (three-injection group).

X-gal Staining and Histology

Embryos were isolated, fixed in 2 % paraformaldehyde (PFA) for 2 hr, and X-gal stained in whole-mount by standard procedures. In embryos older the E10.5, the chest wall was removed and the cardiovascular structures were completely exposed prior to fixation. They were photographed using a dissecting stereoscope and an Olympus Q color camera. After overnight fixation in 2% PFA at 4°C, they were processed for histology by standard procedures and sectioned at 10 μm thickness. All sections were counterstained with eosin, except for those of embryos that were not X-gal stained, which were stained with hematoxylin and eosin (H&E).

Ink Injections

Embryos were harvested and isolated between 14.0 and 15.5 dpc. The chest wall was opened and India ink (Fount India, Pelikan) diluted 1:50 in phosphate buffered saline (PBS) was injected into the left ventricle using 30-gauge needle and insulin syringe. Embryos were then fixed in 2% PFA overnight and stored in PBS at 4°C.

LTR Staining

Embryos were harvested at the desired gestational ages (between 8.0 and 11.5). After removing membranes for genotyping, embryos were immediately placed in 5 μM Lysotracker Red solution in PBS (1 mM stock solution in dimethyl sulfoxide, Molecular Probes, Eugene, OR), and incubated at 37°C for 20 min. Then they were washed in PBS three times over 20 min and fixed in 2% PFA on ice for 2 hr. Embryos were then observed and photographed under fluorescent dissecting stereoscope with a rhodamine filter, and subsequently post-fixed in 2% PFA overnight at 4°C. At least three NCA embryos and three WT embryos were processed for each time point.

Acknowledgements

We thank Drs. Scott Baldwin, Joey Barnett, and Deborah Lang for their critical review of the manuscript and their valuable comments, and Dr. Bruce Appel for making the fluorescent stereoscope available to us. We thank Drs. Bin Zhou and Allan Bradley for providing us with the PuΔTK mice. D.P. was funded by the NIH/NHLBI., C.P.P. was funded by an American Heart Association Scientist Development Award and a March of Dimes Basil O'Connor Award.

Ancillary