Misregulation of SDF1-CXCR4 Signaling Impairs Early Cardiac Neural Crest Cell Migration Leading to Conotruncal DefectsNovelty and Significance
Rationale: Cardiac neural crest cells (NCs) contribute to heart morphogenesis by giving rise to a variety of cell types from mesenchyme of the outflow tract, ventricular septum, and semilunar valves to neurons of the cardiac ganglia and smooth muscles of the great arteries. Failure in cardiac NC development results in outflow and ventricular septation defects commonly observed in congenital heart diseases. Cardiac NCs derive from the vagal neural tube, which also gives rise to enteric NCs that colonize the gut; however, so far, molecular mechanisms segregating these 2 populations and driving cardiac NC migration toward the heart have remained elusive.
Objective: Stromal-derived factor-1 (SDF1) is a chemokine that mediates oriented migration of multiple embryonic cells and mice deficient for Sdf1 or its receptors, Cxcr4 and Cxcr7, exhibit ventricular septum defects, raising the possibility that SDF1 might selectively drive cardiac NC migration toward the heart via a chemotactic mechanism.
Methods and Results: We show in the chick embryo that Sdf1 expression is tightly coordinated with the progression of cardiac NCs expressing Cxcr4. Cxcr4 loss-of-function causes delayed migration and enhanced death of cardiac NCs, whereas Sdf1 misexpression results in their diversion from their normal pathway, indicating that SDF1 acts as a chemoattractant for cardiac NCs. These alterations of SDF1 signaling result in severe cardiovascular defects.
Conclusions: These data identify Sdf1 and its receptor Cxcr4 as candidate genes responsible for cardiac congenital pathologies in human.
- cardiac development
- cardiac neural crest
- conotruncal defects genes
- heart defects, congenital
- receptors, CXCR4
- SDF1 chemokine
Heart formation involves different cell populations originating from distinct embryonic territories.1 The first cardiogenic cells derive from the cardiac crescent known as primary heart field. They form the primitive cardiac tube and ultimately provide the left ventricle, the atrioventricular canal, and part of the atria. The second cardiogenic population appears later in the second heart field. It is incorporated into the primitive heart via the arterial pole and contributes to most of the right ventricle, the conus, the truncus, and also the rest of the atria.2 Cardiac neural crest cells (NCs), originating from the hindbrain, transit through the pharyngeal arches and associate with the second heart field before entering the heart. They play a major role in the outflow tract remodeling and contribute to the septa and valves, allowing separation of the pulmonary trunk from the aorta.3 Ablation of cardiac NCs in chick results in severe heart defects, such as defective outflow septation, outflow misalignment, and mispatterning of the great arteries.4 In addition, cardiac NC ablation disrupts development of the arterial pole by interfering with the second heart field integration into the heart.5 These anomalies mirror those found in congenital heart diseases, including ventricular septal defects (VSD), persistent truncus arteriosus (PTA), double-outlet right ventricle (DORV), tetralogy of Fallot, as well as the DiGeorge and velo-cardial-facial syndromes.6
The prospective cardiac NC territory has been mapped using chick-quail chimeras to the axial region situated between the otic placode and the third somite, thereafter called cardiac level.7,8 This domain overlaps the vagal neural tube that spans the first 7 somites and from which also derive the NC progenitors of the enteric ganglia, vagus nerve, superior cervical ganglion, as well as cartilaginous, supportive, and endocrine tissues of the neck.9 Given the diversity of cell types arising from NCs in this region, one important question concerns how these populations are segregated and selectively driven to their final destinations. Cardiac NCs undergo migration away from the neural tube by following a dorsolateral path under the ectoderm, leading them to the pharyngeal arches. After pausing, they resume migration under the pharyngeal endoderm and, along the aortic arch arteries, populate the cardiac outflow tract where they contribute the aorticopulmonary septum, and ultimately colonize the heart to form the ventricular septum and cardiac ganglia.3 Contrary to cardiac NCs, enteric NCs and progenitors of the peripheral nervous system of the neck follow a ventral route through the somites to reach the pharynx,10 suggesting that cardiac NCs are segregated precociously from the other NC populations by specific guidance cues.
Chemotaxis may account for the precision by which NCs reach their targets. The recent discoveries that the stromal-derived factor-1 (SDF1) chemokine may drive migration of some NC populations,11–13 and that mice deficient for Sdf1 or its receptors, Cxcr4 and Cxcr7, exhibit heart septal defects14–16 suggest that SDF1 signaling may contribute to the guidance of cardiac NCs to the heart. Here, we show in the chick embryo that cardiac NCs express Cxcr4 during their initial migration to the pharyngeal arches, and that Sdf1 exhibits a complementary pattern in the ectoderm along their migration route. Cxcr4 loss-of-function causes delayed migration and enhanced death of cardiac NCs, whereas Sdf1 misexpression results in diversion of cardiac NCs from their normal migration pathway but none of them affect enteric NCs, indicating that SDF1 acts as a chemoattractant specific for cardiac NCs. Finally, alterations of SDF1 signaling during cardiac NC migration lead to severe cardiovascular defects, demonstrating that SDF1 signaling through CXCR4 is required for correct heart morphogenesis. These data open new avenues for identifying genes implicated in cardiac congenital pathologies in human.
Avian embryos, plasmid constructs, procedures for embryo grafting, in ovo electroporation, histological sectioning, in situ hybridization, immunolabeling, and statistical analyses are provided in the Online Data Supplement.
SDF1 Signaling Correlates With Cardiac NC Migration to the Heart
To determine whether SDF1 might constitute a guidance cue for cardiac NCs, we first analyzed in the chick embryo the expression profiles of Sdf1, Cxcr4, and Cxcr7 during their migration to the heart. NCs emerge from the cardiac neural tube at Hamburger and Hamilton17 (HH) stage 10 (ie, at embryonic day [E] 1.5) and can be identified because of HNK1 or Sox10 markers. Sdf1 that was not detectable at the cardiac level before NC migration (Online Figure I) became prominent at onset of migration in the ectoderm situated on either side of the neural tube (Figure 1A and 1C).
Coincidentally, Cxcr4 was expressed specifically in NCs emigrating from the cardiac neural tube and invading the dorsolateral pathway. Simultaneous detection of Sdf1 and Cxcr4 revealed strikingly coordinated spatiotemporal patterns in the ectoderm and underlying NCs throughout migration. As NCs progressed toward the circumpharyngeal region, Sdf1 expression was progressively reduced to the lateral ectoderm, abutting at the front of the Cxcr4-positive NCs and, at stage HH14 (E2), it was restricted to a discrete portion of the ventrolateral ectoderm along pausing NCs (Figure 1B, 1D, and 1E). It is noteworthy that although Sdf1 and Cxcr4 were prominent at the cardiac level, they were also found in other locations, for example, the fore and midbrain, as well as in other NC populations, for example, those emerging from rhombomere-4 (Online Figure I). However, Sdf1 expression in the ectoderm was limited to the cardiac level and was not apparent more caudally beyond the third somite, and vagal NCs migrating ventrally showed no Cxcr4 or Cxcr7 expression (Online Figure I). In contrast to Cxcr4, Cxcr7 expression was never observed in NCs during their subectodermal migration to the pharyngeal arches (Online Figure I).
At E2.5, cardiac NCs resume migration under the pharynx. Because both Sox10 and HNK1 are lost by NCs in the pharyngeal arches after HH14 and that HNK1 becomes instead expressed in the second heart field developing along the endoderm,18 we investigated the expression patterns of Sdf1 and its receptors in avian chimeras after isotopic and isochronic grafting of quail or green-fluorescent–protein-positive (GFP+) transgenic chick neural tubes. NCs were identified using the quail-specific QCPN marker or by their GFP content. When cardiac NC migrated under the pharynx, they progressively lost Cxcr4, and Sdf1 shifted from the ectoderm to the pharyngeal mesoderm (Figure 1F). At E5, cardiac NCs invade the distal part of the outflow tract and, by E8, they reach the ventricular septum (Figure 2A). From the time they invaded the outflow tract until late stages of heart morphogenesis, most cardiac NCs expressed Cxcr7 (Figure 2B–2D; Online Figure II). Interestingly, Sdf1 was found in mesenchymal cells of the conus and truncus closely associated with the NC population (Figure 2B–2D; Online Figure II).
These results show that cardiac NC progression to the heart is associated with SDF1 signaling. Moreover, they identify 2 steps during migration differing in the expression of CXCR4 and CXCR7 receptors and in the SDF1 source. First, NCs express Cxcr4 and migrate as a cohort of cells along the ectodermal epithelium, which expresses Sdf1 in a dynamic fashion until they reach the circumpharyngeal region. Second, as cardiac NCs penetrate the outflow tract, they express Cxcr7 instead of Cxcr4 and migrate as a scattered population mingled with an Sdf1-expressing mesenchyme.
Loss of CXCR4 Signaling Leads to Defective Cardiac NC Migration
To assess the role of SDF1 signals in cardiac NC migration, we performed CXCR4 knockdown experiments. A vector encoding an miRNA directed against Cxcr4 (miRNA-Cxcr4) was electroporated into the cardiac neural tube of HH9-chick embryos. Efficiency of the miRNA-Cxcr4 to reduce CXCR4 protein level was evidenced in both the electroporated neural tube and cardiac NCs 24 hours postelectroporation (hpe) and could be reversed by cotransfection with a wild-type form of Cxcr4 (Online Figure III). Although initial lateral migration of cardiac NCs was not strongly affected in miRNA-Cxcr4-transfected embryos (Figure 3A–3C), a severe reduction in the number of NCs populating the circumpharyngeal region and migrating into the outflow tract was observed 48 hpe (Figure 3D–3J). Quantitation of the number of GFP-positive NCs revealed a 65% decrease in migrating cardiac NCs after 36 hpe (Figure 3K), which could be attributed essentially to apoptosis, as judged on the detection of activated caspase-3 (Figure 3H, 3J, and 3L). In contrast, NCs following the ventral pathway were not affected by miRNA-Cxcr4 because spinal ganglia formed and enteric NCs succeeded in colonizing the foregut (Online Figure IV).
Because SDF1 signaling is known to regulate both chemotactic migration and cell survival,19 we overexpressed a construct encoding a dominant-negative form of Cxcr4 (DN-Cxcr4), in which the C-tail of the molecule was deleted. CXCR4 C-tail has been shown to be dispensable for SDF1-dependent survival but critical for chemotaxis,20,21 allowing it to discriminate between SDF1 activities during cardiac NC migration. HNK1 immunolabeling revealed that 15 hpe, NCs expressing the DN-Cxcr4 were lagging behind the nonelectroporated NCs (Figure 4A and 4B). Delay in migration was even more pronounced 24 hpe, with most DN-Cxcr4–expressing NCs remaining near the neural tube or migrating in the ventral route through the paraxial mesoderm (Figure 4C and 4D). Quantitation of migrating GFP-positive NCs in the dorsal and circumpharyngeal quadrants of the embryo showed that DN-Cxcr4–expressing cardiac NCs mainly remained dorsally near the neural tube, whereas their number was decreased by half in the circumpharyngeal region (Figure 4E and 4F). As expected, apoptosis was not increased in DN-Cxcr4–electroporated embryos (not shown), confirming that decrease in NCs in the circumpharyngeal area resulted from reduced or misrouted migration. To trace NC fate later during development, the DN-Cxcr4 construct was coelectroporated with a vector encoding red-fluorescent protein known to be more stable than GFP. Almost no cardiac NCs were observed 72 hpe in the outflow tract of DN-Cxcr4–electroporated embryos, at a time when they normally massively populate this region (Figure 4G–4K). In contrast, DN-Cxcr4–expressing NCs were observed colonizing normally the foregut (Figure 4L). These results indicate that SDF1 signaling drives cardiac NC migration from the neural tube toward the circumpharyngeal region via CXCR4. Cxcr4 loss-of-function by miRNA-Cxcr4 resulted in cardiac NC apoptosis, whereas CXCR4 C-terminal-tail truncation provoked NC misrouting, arguing for a dual role of CXCR4 in controlling both NC survival and migration and leading, in both cases, to cardiac NC deprivation in the outflow tract.
SDF1 Acts as a Chemoattractant for Cardiac NCs
To further determine whether SDF1 could exert a chemotactic effect, we analyzed the migratory response of cardiac NCs confronted to an ectopic source of SDF1. An Sdf1 construct was electroporated into the cardiac neural tube of HH9-chick embryos to induce Sdf1 misexpression close to the NC-emergence site. In these embryos, NCs failed to migrate laterally, and 24 hpe, they formed large cell clusters accumulated near the Sdf1-producing neural tube (Figure 5A and 5D). Interestingly, in high-Sdf1–expressing specimens, NCs originating from the contralateral side were prevented from migrating and accumulated close to the neural tube, showing that excess of secreted SDF1 from one side of the neural tube was sufficient to attract all CXCR4-positive NCs exiting the neural tube. This resulted in the complete absence of NCs in the circumpharyngeal region in both sides of the embryo (Figure 5E and 5F). Forty-eight hpe, NCs were still aggregated near the neural tube in Sdf1-misexpressing embryos, but most of them were apoptotic (Figure 5H and 5I). Sdf1 misexpression in the vagal neural tube produced no apparent effect on ventral NC migration as well as on enteric NC penetration into the foregut (Online Figure IV).
Because Sdf1 exhibits a dynamic spatiotemporal pattern in the ectoderm in relation with NC progression to the circumpharyngeal region, we analyzed the effect of altering this dynamics on cardiac NC migration. Sdf1 was electroporated in the lateral ectoderm to induce its strong, continuous expression from the neural tube up to the circumpharyngeal ridge. Under these conditions, NCs remained accumulated under the Sdf1-expressing ectoderm and failed to reach the circumpharyngeal region (Figure 5G). Thus, these results indicate that SDF1 drives cardiac NC progression under the ectoderm by a chemotactic mechanism requiring precise spatiotemporal regulation of its expression.
Cross-Regulation of SDF1 and CXCR4 Expression During Cardiac NC Migration
To gain insight into the mechanisms controlling spatiotemporal regulation of SDF1 activity during cardiac NC migration, we analyzed the Sdf1 and Cxcr4 expression profiles in the contexts of Cxcr4 loss-of-function and Sdf1 misexpression. In embryos in which Sdf1 was ectopically expressed in the neural tube, CXCR4 was downregulated both in the producing neural tube cells and in cardiac NCs accumulated near the source of Sdf1 (Figure 6A and 6B), indicating that an excessive or prolonged exposure to SDF1 leads to CXCR4 repression. Interestingly, in embryos in which Sdf1 was overexpressed in the neural tube and in DN-Cxcr4–expressing embryos in which NCs failed to populate the pharyngeal arches, a nearly complete disappearance of Sdf1 expression was observed in the lateral ectoderm facing the pharynx (Figure 6C–6E), suggesting that maintenance of Sdf1 expression in the ectoderm depends on cardiac NC–derived signals.
Misregulation of SDF1 Signaling in Cardiac NCs Causes Heart Defects
Finally, we analyzed the consequences of NC migratory defects caused by alterations in SDF1 signaling on heart morphogenesis and the patterning of great arteries. To target all cardiac NCs, neural tubes were electroporated bilaterally. Hearts were collected and their anatomy and histology examined after completion of cardiac morphogenesis at E9.5. The majority (>80%) of control embryos electroporated with GFP or with a control miRNA exhibited normal hearts (Figure 7A and 7F). In contrast, in embryos electroporated with miRNA-Cxcr4, although the gross anatomy of the heart seemed normal in many cases (not shown), histological analyses revealed large cardiovascular defects, notably VSD in 67% of the embryos and, less frequently, hearts with DORV and transpositions of the great arteries (GAT) (Figure 7B and 7F). Heart defects were even more severe in DN-Cxcr4 embryos, 45% of them presented DORV in which both the aorta and pulmonary trunk exited from the right ventricle (Figure 7C and 7F). This misalignment of the great arteries was systematically accompanied with VSD, which otherwise were observed in 55% of the embryos. Sdf1 misexpression in the neural tube also led to severe heart anomalies. Indeed, 29% of them showed a persistent truncus arteriosus in which the outflow tract was not divided into aorta and pulmonary trunk, giving rise systematically to VSD as a consequence of a missing aorticopulmonary septation (Figure 7D and 7F). VSD were otherwise observed in 43% of the hearts, DORV and transposition of great arteries in 14% each. Therefore, these results show that SDF1 signaling defects in cardiac NCs leads to a large spectrum of cardiac anomalies that phenocopy congenital heart defects, underlining the importance of SDF1 signaling in heart morphogenesis. We also analyzed gut innervation in embryos, in which the SDF1 pathway was altered in cardiac NCs. Immunodetection of Tuj1, a pan-neuronal marker, revealed that, in all miRNA-Cxcr4 and DN-Cxcr4 embryos and in the majority of the Sdf1-misexpressing embryos, gut innervation was normal (Online Figure IV). However, in ≈20% of Sdf1-misexpressing embryos, the hindgut was devoid of enteric innervation, immunolabeling being restricted to the nerve of Remak originating from the sacral neural tube (Online Figure IV). A likely explanation for the occurrence of aganglionic hindgut is that cardiac NC accumulation provoked by Sdf1 misexpression prevented enteric NC migration into the ventral pathway by spatial hindrance, thereby reducing the number of cells colonizing the gut.
Chemotactic Guidance of Cardiac NCs to the Heart
Our study demonstrates that cardiac NCs express CXCR4 and are guided during their early migration toward the pharyngeal arches by a chemotactic mechanism elicited by Sdf1 released by the ectoderm. Studies in chick revealed that other chemoattractants may also be implicated in the guidance of cardiac NCs. Fibroblast growth factor (FGF)-8 is expressed in the pharyngeal region and a dominant-negative form of fibroblast growth factor receptor (FGFR)-1 receptor expressed in cardiac NCs provokes a delay in their migration combined with increased cell death.22 Semaphorin-3C is expressed in the outflow tract, and loss-of-function of its receptor Plexin-D1 results in an NC migration default.23 Therefore, the ability of cardiac NCs to reach the heart would rely on different chemoattractants operating in a coordinated manner throughout NC navigation. SDF1 and FGF8 signaling would guide cardiac NCs in the lateral path toward the circumpharyngeal region, whereas Semaphorin-3C would drive them to the outflow tract. It will be of interest to test whether SDF1 and FGF8 signaling regulate each other to coordinate initial migration of cardiac NCs.
Sdf1 Drives Cardiac NC Migration and Survival Through the CXCR4 Receptor
Although SDF1 signaling has been shown to regulate both chemotactic migration and cell survival in various cell types,19 it is not clear which pathways mediate these functions. It has been suggested that CXCR4 is involved in chemotaxis and CXCR7 in survival,24 but other studies also proposed that both processes may be controlled by CXCR4.25 In addition, biochemical studies showed that the CXCR4 C-terminal-tail is dispensable for activation of the Jak2/STAT3 pathway involved in SDF1-dependent survival20 or for integrin-mediated adhesion,21 but is critical for chemotaxis.20 Our results show that CXCR4, but not CXCR7, is expressed in cardiac NCs during their initial migration and that CXCR4 downregulation by miRNA leads to their apoptosis, whereas a dominant-negative form of CXCR4 lacking its C-terminal domain induces their misrouting without causing cell death, arguing for a role of CXCR4 in controlling both events in cardiac NCs. Surprisingly, CXCR4 knockdown by miRNA-Cxcr4 did not prevent initial NC migration, although it was expected to affect both survival and migration. A likely explanation is that because the miRNA induced CXCR4 downregulation only after 24 hpe, NCs were capable of initiating migration under the ectoderm but failed to survive as they reached the circumpharyngeal region, that is, precisely at the time when miRNA-Cxcr4 became fully efficient.
In contrast to Cxcr4, Cxcr7 was not expressed in cardiac NCs during their initial migration, emphasizing the exclusive role of CXCR4 in mediating the SDF1-guidance cue during this step. However, when NCs invaded the aorticopulmonary trunk, Cxcr4 was replaced by Cxcr7. In this region, non–NC-derived mesenchymal cells express Sdf1, suggesting a role for SDF1 signaling transduced exclusively by CXCR7 at this step. Recent studies showed that CXCR7 may promote cell migration26 or proliferation27 independently of CXCR4. During heart development, the great arteries derived from cardiac NCs undergo extensive lengthening during septation of the outflow tract.8 One possibility is that SDF1 signaling would be implicated in the regulation of cardiac NC proliferation via CXCR7 during this remodeling step.
Cross-Regulation of Sdf1 and CXCR4 Expression During Cardiac NC Migration
An important question concerning chemotaxis is how the ligand is dynamically regulated to create a gradient along the progressing cell population. During germ cell migration in zebrafish, Sdf1 activity is regulated by CXCR7 expressed in the neighboring somatic cells and acting as SDF1 sequester.28 Alternatively, in the lateral line, CXCR7 regulates Cxcr4 transcription in the migrating population to maintain its polarized expression at the front.29 Such mechanisms could not account for Sdf1 regulation in the ectoderm during initial cardiac NC migration as Cxcr7 was not detected. Our results suggest instead an alternative process involving both Sdf1 and Cxcr4 cross-regulating each other. Indeed, we found that during NC migration, Sdf1 is expressed initially in the ectoderm up to the neural tube and becomes gradually repressed medially during migration. This suggests that NCs themselves provide the cue for repressing Sdf1 expression during their progression. However, we also showed that when cardiac NCs fail to reach the pharyngeal arches in Sdf1 or DN-Cxcr4–electroporated embryos, Sdf1 is downregulated in the ectoderm, suggesting that its expression requires positive signals emanating from NCs. In addition, when Sdf1 was overexpressed, it induced CXCR4 downregulation. A plausible model accounting for these observations would be that cardiac NCs maintain Sdf1 expression in the ectoderm during the course of their migration to the pharyngeal arches. After their passage, Sdf1 would be downregulated in the ectoderm. Conversely, once NCs reach the circumpharyngeal region and pause along the portion of the lateral ectoderm still expressing Sdf1, they confront a prolonged SDF1 exposure, causing progressively CXCR4 downregulation on their surface. This would in turn switch SDF1 expression in the ectoderm off, allowing cardiac NC cells to resume migration. A similar regulatory loop has been already proposed for FGF8 effect on cardiac NC.22,30 Such molecular cross-talks between cardiac NCs and their environment would then be crucial for their chemotactic guidance toward the heart.
SDF1 Role in the Segregation Between Cardiac and Enteric NCs
Although it has been known for long that the cardiac and enteric NC lineages derive from neural tube regions that overlap in the vagal region at the level of the first 3 somites,9,31 mechanisms involved in their segregation have remained elusive. Recent studies showed that cardiac and enteric NCs follow different migration pathways immediately after exiting the neural tube, with first cardiac NCs migrating under the ectoderm and enteric NCs undergoing ventral migration slightly later.10 Our results demonstrate for the first time that because of a selective SDF1 chemotactic effect on cardiac NCs, these progenitors are early segregated at onset of migration. In addition, they provide molecular explanation to previous embryology experiments showing that chimeric embryos in which the chick cardiac neural tube has been replaced by quail cranial or truncal neural tubes exhibit cardiac anomalies that mirror cardiac NC ablation.32 It is likely that cranial and truncal NCs cannot substitute for cardiac NCs because they do not express CXCR4 and, hence, cannot respond to SDF1 chemotaxis. Likewise, when quail neural tubes from somites 1 to 3 are transplanted into the same axial level in older embryos or to somites 5 to 7 level in embryos of the same age, quail NCs are never seen migrating under the ectoderm.10 We suggest that NCs coming from the somites 1 to 3 fail to migrate laterally in older embryos because Sdf1 has already been downregulated in the ectoderm, whereas NCs coming from the somites 5 to 7 do not express CXCR4 and thus cannot interpret the SDF1-guidance cue. However, this model fits with the recent observations that cardiac NCs can only regenerate and repopulate the heart region when ablated before HH9 and not later.33 Indeed, at this early stage, Sdf1 is still expressed in the ectoderm and is, therefore, able to attract the regenerating NCs expressing CXCR4.
Misregulation of SDF1 Signaling Leads to Cardiac Anomalies That Phenocopy Cardiac Congenital Pathologies
Cxcr4 loss-of-function in early migrating cardiac NCs or Sdf1 misexpression along their migration routes gave rise to a variety of conotruncal malformations. In nearly all cases, VSD, the most frequent heart defect arising after cardiac NC ablation,4 was invariably observed. In addition to VSD, DORV and GAT were seen, although less frequently, whereas PTA, considered as the most severe phenotype, was identified in a significant proportion of Sdf1-misexpressing embryos. The severity of phenotypes obtained may be related to the strategy used to target SDF1 signaling. Indeed, while Cxcr4 loss-of-function by miRNA decreased CXCR4 protein level from 24 hpe, leading to a relatively late NC defect between 24 and 48 hpe, only 15 hours were necessary to reduce NC migration in the DN-Cxcr4–treated embryos. Likewise, the strong effect of Sdf1 misexpression in the neural tube induced an even more rapid and massive accumulation of NCs 6 hpe, resulting in an almost-complete NC depletion from the lateral pathway. Thus, whereas in miRNA-Cxcr4 embryos, a number of NCs could reach the outflow tract and contribute later to septation of the aorticopulmonary trunk, but not in sufficient amount to form the ventricular septum, in the Sdf1-misexpressing specimens, the nearly complete depletion of cardiac NC impacts also on the formation of the outflow tract.
These heart phenotypes mirror those observed in mice defective for Sdf1, Cxcr4, and Cxcr7, which all exhibit VSD14–16 and, occasionally, misalignment of the aorta.34 Although expression of Cxcr4 and Cxcr7 in murine cardiac NCs has not been formally established,34,35 conditional deletions of Cxcr4 in cardiomyocytes36 and of Cxcr7 in endothelial cells34 have not been found to produce VSD, therefore, supporting the idea that heart defects in null mutant mice are related to NC defaults. It will be of interest to analyze the incidence on heart development of conditional Cxcr4 and Cxcr7 deletions in NCs. However, heart phenotypes observed in SDF1-signaling–deficient mice are less severe than in chick. This discrepancy could be attributed to species differences in heart colonization by NCs. Indeed, the use of NC reporters in mice to trace cardiac NCs revealed differences in NC-derived structures in the heart, notably in the aorta and pulmonary trunk, compared with quail-chick tracing.8,37
In humans, mutations in the CXCR4 gene have been found so far associated with warts, hypogammaglobulinemia, infections, and myelokathexis syndrome, a combined immunodeficiency disease.38 Although cardiac anomalies are not commonly detected in this pathology, a recent analysis revealed cardiac defects in some patients with WHIM syndrome.39 The DiGeorge syndrome is a well-described pathology characterized by severe cardiovascular anomalies related to cardiac NC defects.40 Interestingly, an important reduction of Cxcr4 in interneurons has been reported recently in the LgDel mouse model, which phenocopies the 22q11.2 deletion of the DiGeorge syndrome, possibly accounting for the cortical circuit disorders observed in this syndrome.41 This raises the intriguing possibility that cardiac anomalies of the DiGeorge syndrome might be related at least partly to diminished Cxcr4 expression.
Taken together, our results demonstrate that disrupting SDF1 signaling in cardiac NCs impairs their migration massively, thus leading to heart anomalies, and opens perspectives for identifying new genes in cardiac congenital pathologies in humans.
We thank S. Gournet for illustration, our colleagues for constructs, probes, and antibodies, and the GFP+transgenic-egg resource at the Roslin Institute.
Sources of Funding
This work was supported by the CNRS, the Université Pierre et Marie Curie, and the Association Française contre les Myopathies (Grant No. 11405).
Sophie Escot was a recipient of a doctoral fellowship from the Ministère de l’Enseignement Supérieur et de la Recherche.
In June 2013, the average time from submission to first decision for all original research papers submitted to Circulation Research was 13.67 days.
The online-only Data Supplement is available with this article at http://circres.ahajournals.org/lookup/suppl/doi:10.1161/CIRCRESAHA.113.301333/-/DC1.
- Nonstandard Abbreviations and Acronyms
- double-outlet right ventricle
- embryonic day
- transposition of great arteries
- green-fluorescent protein
- Hamburger and Hamilton stage
- hours postelectroporation
- neural crest cells
- persistent truncus arteriosus
- stromal-derived factor-1
- ventricular septal defects
- Received March 8, 2013.
- Revision received July 8, 2013.
- Accepted July 9, 2013.
- © 2013 American Heart Association, Inc.
- Evans SM,
- Yelon D,
- Conlon FL,
- Kirby ML
- Kasemeier-Kulesa JC,
- McLennan R,
- Romine MH,
- Kulesa PM,
- Lefcort F
- Ma Q,
- Jones D,
- Borghesani PR,
- Segal RA,
- Nagasawa T,
- Kishimoto T,
- Bronson RT,
- Springer TA
- Sierro F,
- Biben C,
- Martínez-Muñoz L,
- Mellado M,
- Ransohoff RM,
- Li M,
- Woehl B,
- Leung H,
- Groom J,
- Batten M,
- Harvey RP,
- Martínez-A C,
- Mackay CR,
- Mackay F
- Waldo KL,
- Kumiski DH,
- Wallis KT,
- Stadt HA,
- Hutson MR,
- Platt DH,
- Kirby ML
- Vasyutina E,
- Stebler J,
- Brand-Saberi B,
- Schulz S,
- Raz E,
- Birchmeier C
- Ahr B,
- Denizot M,
- Robert-Hebmann V,
- Brelot A,
- Biard-Piechaczyk M
- Mazzinghi B,
- Ronconi E,
- Lazzeri E,
- et al
- Delgado-Martín C,
- Escribano C,
- Pablos JL,
- Riol-Blanco L,
- Rodríguez-Fernández JL
- Odemis V,
- Boosmann K,
- Heinen A,
- Küry P,
- Engele J
- Kirby ML,
- Gale TF,
- Stewart DE
- Agarwal U,
- Ghalayini W,
- Dong F,
- Weber K,
- Zou YR,
- Rabbany SY,
- Rafii S,
- Penn MS
- Jiang X,
- Rowitch DH,
- Soriano P,
- McMahon AP,
- Sucov HM
- Meechan DW,
- Tucker ES,
- Maynard TM,
- LaMantia AS
Novelty and Significance
What Is Known?
During embryogenesis, cardiac neural crest cells (NC) migrate from the future spinal cord to give rise to the smooth muscle of the cardiac outflow tract and the great vessels, as well as the septum that separate ventricles.
Some of the signaling pathways that regulate cardiac NC migration have been identified.
Abnormalities in cardiac NC migration have been implicated in congenital heart diseases in humans.
What New Information Does This Article Contribute?
In the chick embryo, expression of the chemokine, stromal-derived factor-1 (SDF1), is tightly coordinated with the progression of cardiac NCs expressing its CXCR4 receptor during their early step of migration toward the heart.
SDF1 signaling selectively drives cardiac NC migration toward the heart via a chemotactic mechanism.
Misregulation of SDF1 signaling results in severe cardiovascular defects, identifying SDF1 and its receptor CXCR4 as candidate genes responsible for cardiac congenital pathologies in human.
Cardiac NCs contribute to heart morphogenesis by giving rise to smooth muscle of the outflow tract and great vessels, as well as part of the ventricular septum. Failure in cardiac NC development results in outflow tract and ventricular septation defects commonly observed in congenital heart diseases. However, the molecular mechanisms driving cardiac NC migration toward the heart have not been completely elucidated. We show that in the chick embryo, the SDF1 signaling pathway plays an essential role in cardiac NC migration toward the heart. Loss-of-function of CXCR4 (the SDF1 receptor) causes delayed migration and enhanced cell death of cardiac NCs, whereas misexpression of SDF1 provokes cardiac NC diversion from their normal migration pathway, indicating that SDF1 acts as a chemoattractant for cardiac NCs. These findings suggest that SDF1 and its receptor CXCR4 may be candidate genes responsible for cardiac congenital pathologies in humans.