Dauvé, Véronique; McLin, Valérie A.
Molecular embryology has been instrumental in advancing liver stem cell differentiation in vitro. In contrast, its contribution to gastrointestinal (GI) development has been to unravel the genetic underpinnings of GI malformations. Clinically, this is significant for the purposes of antenatal screening and genetic counselling. In addition, from an embryologist's perspective, the study of developmental anomalies offers insight into normal development. Thus, this review focuses on malformations of the GI tract along the craniocaudal axis from the oesophagus to the rectum. It aims to correlate molecular and genetic regulation of development with known clinical phenotypes of developmental defects.
BRIEF OVERVIEW OF GI DEVELOPMENT
Vertebrate development is characterized by the formation of the 3 primitive germ layers (endoderm, mesoderm, and ectoderm) during gastrulation. The multipotent endoderm gives rise to the liver, epithelial lining of the GI tract, lung, thyroid, and pancreas, whereas the mesoderm is the precursor of the intestinal mesenchyme and visceral musculature. Starting at gastrulation, an array of transcription factors and signalling pathways in the endoderm and the adjacent mesoderm contribute to cell-fate decisions, which direct the fate of endoderm and mesoderm progenitors (Table 1).
Development of the GI tract is governed by a vast molecular network comprising signalling pathways and transcription factors, which are still incompletely characterized. The basic paradigm of GI development is that cell-fate decisions occur along the following 4 axes: dorsoventral, anterioposterior, left–right, and radial (Fig. 1). Anterior–posterior patterning is necessary for the regional specification of the gut and a prerequisite for differential function from mouth to anus. Second, lumen formation is essential for epithelial morphogenesis. In other words, lumen formation is a prerequisite for the development of the large and specialized surface area required for sufficient nutrient absorption to sustain life. Many of these processes are directed by complex molecular and physical interactions between the developing endoderm and mesoderm. Furthermore, the vertebrate GI tract is characterized by its extensive coiling. These morphogenetic movements are “necessary” for spatial organization in the abdominal cavity, but their contribution to intestinal function per se is unclear.
The weakness of such a complex system is that a single, erroneous cell-fate decision can lead to severe developmental defects, often incompatible with extrauterine life. Mutations in ubiquitous signalling pathways are commonly assumed to underlie systemic syndromes. It follows that many other signalling cascades and transcription factors are probably involved in GI development as evidenced by the number of syndromes comprising GI malformations (1–3).
The purpose of this review is to illustrate this principle using examples from human, mouse, Xenopus laevis (African clawed frog), and zebrafish because it is accepted that developmental paradigms are conserved across vertebrate species. We will not expand on early specification and regional patterning for 2 reasons: they have been well summarized by experts in the field (4,5) and signalling defects at these early stages are likely associated with embryonic lethality. Rather, the focus is on the developing luminal digestive system along the craniocaudal axis from the oesophagus to the rectum after the initial patterning events. The reader is referred to recent reviews for molecular advances of liver and pancreas development (6–8). Conventional gene and protein nomenclature and symbols are used throughout. Figure 2 summarises genes and pathways that are implicated in malformations according to anatomic origin.
Oesophageal atresia is diagnosed either prenatally on ultrasound or promptly noticed in the first hours of life. It occurs in 1 in 3500 births, and several anatomic variants have been described (9).
Oesophageal and tracheal development are tightly linked. During human gestation, early embryonic events lead to the development of the respiratory primordium at approximately 4 weeks of development (9). Then, the early foregut endoderm and mesenchyme give rise to the oesophagus and trachea. Therefore, the advent of abnormal communications between the 2 neighbouring, tubular organs is a logical corollary. Typically, anomalous communications present as 1 of the following: oesophageal atresia with a distal tracheooesophageal fistula (TOF) (85% of cases), oesophageal atresia without fistula (10% of cases), oesophageal atresia with a proximal TOF (≤1% of cases), oesophageal atresia with both proximal and distal TOF (<1% of cases), and H-type fistula characterised by the absence of oesophageal atresia but the isolated presence of TOF (4% of cases).
It is widely accepted that although genetic defects may underlie familial cases of oesophageal atresia, few candidate genes have been identified. In a mouse model, the absence of Nkx2.1 expression in the ventral foregut and/or altered or absent Sox2 expression in the dorsal foregut (10) leads to incomplete separation of the oesophagus and trachea (11). Furthermore, the sonic hedgehog (Shh) signalling pathway likely exerts independent or additional effects, explaining some cases of oesophageal atresia/TOF (9,12). It is unclear whether these defects originate in the endoderm via SHH signalling or in the mesenchyme where SHH receptors GLI2, GLI3 (13), and SHH-target genes are expressed (12,14).
In humans, SOX2 mutations are associated with anophthalmia, oesophageal, and genital syndromes (15,16). Oesophageal atresia may be associated with Coloboma, Heart anomaly, choanal Atresia, Retardation, Genital and Ear anomalies (CHARGE) syndrome, possibly implicating the chromodomain helicase DNA-binding protein 7 gene (CHD7), a relatively ubiquitous nuclear protein (17–19). Patients with Feingold syndrome have also been described as presenting with oesophageal atresia. Genetically, Feingold syndrome is characterized by a microdeletion in the Shh target myelocytomatosis-viral-related-oncogene-neuroblastoma-derived (MYCN). The fact that this gene is known to be downstream of Shh suggests that akin to what is known in the mouse, SHH is likely involved in human oesophageal development and malformations (20).
Stomach and Pylorus
Gastric and pyloric malformations occur either as an isolated finding or as part of a syndrome. Obstructions or malformations of this anatomical region usually lead to unmistakable, obstructive clinical symptoms in the neonate.
The first such malformation is extremely rare: agastria. Only 1 case report was identified in the literature and the outcome was fatal (21). Congenital microgastria is a little more common and can be associated with other foregut malformations (22). The only genetic association identified in humans to date is with 22q11.2 deletion and ciliary dyskinesia in a newborn with DiGeorge syndrome (23).
Abnormalities of the pylorus are much more common. The pyloric sphincter is an anatomic region composed of characteristic epithelial and mesodermal structures, which, under normal conditions, are slightly hypertrophied to limit luminal flow into the duodenum. When hypertrophy is more severe, it obstructs gastric emptying and the infant develops refractory vomiting. Obstructions can be more or less complete. A total of 1 to 3/1000 live births in Europe display a partial obstruction known to be infantile hypertrophic pyloric stenosis (IHPS) (24). New genetic approaches in humans have demonstrated that the underlying defect is a mutation of muscleblind-like (MBNL1) gene product (24). The muscleblind protein family of zinc finger proteins is known for its role in myotonic dystrophy. Its function is to regulate splicing in the early postnatal period, which is essential for extensive muscle remodelling during the first 3 weeks of extrauterine life. This timing fits with the clinical presentation of IHPS, which occurs almost exclusively between 2 and 8 weeks postnatally.
As may be expected, no single gene is implicated in the morphogenesis of such a functionally important structure. Rather, several other signalling pathways and transcription factors may contribute to the onset of IHPS. First, mutations in nitric oxide synthase (nNOS) have been identified in a small number of human cases IHPS (3,25), in keeping with observations in Nos-1–deficient mice (26). Nitric oxide mediates transient pyloric relaxation; thus, defects in NOS-1, and consequently in smooth muscle relaxation, could explain some cases of IHPS.
Another locus has recently been shown to be associated with cases of IHPS: the SHH target and homeobox-containing transcription factor NKX2.5, located on chromosome 5q35.2 (24). During chick embryonic development, NKX2.5 is expressed during a finite developmental window and is necessary for normal pyloric muscle (27). Its expression is limited to the pyloric sphincter (27). NKX2.5 may play a role in pyloric development in 1 of 2 ways: either directly through its effect on differentiating muscle precursors or because it is a target of SHH signalling, known to be implicated in the development of the foregut and pylorus.
Sox9 is another early SHH target located in the mouse pylorus. Mesodermal bone morphogenetic protein-4 (BMP4) signalling leads to activation of Sox9-Gremlin, shown to be necessary and sufficient in mice for normal development of the pylorus. To date, however, SOX9 mutations have not been identified in humans with IHPS (28).
At the morphological level, duodenal development is characterized by alternating windows of cell proliferation and cell death, in keeping with a fundamental theory of lumen formation (12,29). The proliferative phase is associated with a decrease in lumen diameter, ultimately leading to occlusion, whereas waves of programmed cell death (apoptosis) lead to vacuole formation, recanalization of the lumen, and villi formation (30).
Duodenal malformations diagnosed in neonates include duodenal atresia, duodenal web, annular pancreas, duodenal stenosis, and duodenal diverticulum. As many as half of these malformations are associated with other anomalies (31). The estimated incidence of congenital duodenal abnormalities varies between 1:4000 and 1:15,000 live births (32).
Based on studies in mice, it is probable that abnormal epithelial–mesenchymal interactions are at the root of many duodenal malformations. There is some consensus that the SHH pathway is an important player in this process by regulating gene expression in the developing duodenum (2,33). In Shh−/− mice, villi overgrowth leads to luminal occlusion similar to human duodenal stenosis (2,33,34).
In contrast, little data from humans are available. Duodenal atresia has been identified in Feingold syndrome, described in the setting of oesophageal atresia (20), thus implicating MYCN and SHH in human duodenal development. Duodenal atresia has also been identified in human infants with alveolar capillary dysplasia who display deletions in the FOXC1/FOXL1/FOXF1 cluster (35).
The fibroblast growth factor (FGF) family is known to participate in epithelial–mesenchymal interactions and as such is another likely signalling pathway in duodenal (mal−) development. Fgfr2b−/− or Fgf10−/− mutant mice display variable phenotypes of duodenal atresia, resembling what is observed in humans (36,37). No case of FGF mutation has yet been reported in human duodenal malformations.
Defects in genes regulating vacuolization and lumen formation may be the cause of duodenal diverticula. This rare anatomical anomaly mostly occurs along the pancreatic or mesenteric border in the second or third portions of the duodenum. Rarely, the diverticulum can be intraluminal, presenting as upper GI obstruction (38,39). Although little is known in humans and mice about lumen formation, it has been shown in a zebrafish model that genes regulating ion flux participate in normal lumen formation. Mutations in this cascade lead to multiple lumina (40). Looking for mutations in human homologues seems the obvious next step.
Jejunum and Ileum
The midgut also displays an array of anatomical anomalies accepted to occur secondary to developmental events. The 2 broad categories are atresias and duplications.
Jejunal and Ileal Atresias
Jejunal or ileal atresias may present as GI obstruction early in postnatal life (41). Their incidence varies from 1:300 to 1:3000 live births. Atresias are defined according to their morphology and blood supply. The morphological description refers to the nature of the obstruction between bowel segments (diaphragm vs fibrous cord) and to the blood supply via the mesentery (41). The importance of emphasizing blood supply is that a defect in blood supply may be the primum movens of several of these neonatal defects.
Although jejunal and ileal atresias are usually bundled together as one and the same disease, their clinical presentation and postoperative course often differ, suggesting 1 of 2 possibilities: the aetiologies are different or the 2 segments respond differently to obstruction and altered blood supply owing to regional, anatomical, and functional differences (42,43).
The aetiology of jejunal and ileal atresia is unclear. Historically, there are 2 theories. The first and older of the 2 proposes that imperfect recanalization during development is the root of the obstruction (44). Today, the obvious question is “what are the genes governing this recanalization defect?” The other theory proposed by Louw and Barnard suggests that vascular events underlie these malformations (45,46). Canine models favour the vascular theory, something that is also supported in a reported case of ileal atresia associated with vascular band anomaly (47).
Recently, it has been suggested that the gene regulatory factor X6 (RFX6) may be involved in intestinal atresias. The rationale is that several patients with Mitchell-Riley syndrome, characterised by jejunal or ileal atresia and others malformations (pancreatic hypoplasia, gallbladder aplasia, or hypoplasia, extrahepatic biliary atresia with or without TOF and neonatal diabetes), have mutations of the RFX6 gene (48–50). (References 51–158 are available online only at http://links.lww.com/MPG/A216.) Because RFX6 is expressed in the early pancreatic endoderm it is unclear how it may be linked to intestinal atresias (50–52). An obvious candidate partner is the pancreatic duodenal homeobox 1 (PDX1), which is an important transcription factor in early foregut specification (53,54). Gene profiling technologies should assist in unravelling the molecular network governing duodenal development (54).
Ileal and Jejunal Duplications
Atresias and duplications also occur in the midgut: jejunum and ileum, both presenting with obstructive symptoms. Usually, GI duplications presents before age 2 years. They can occur anywhere along the digestive tract and typically present in 1 of 2 forms: cystic (≥80%) or tubular. The frequency of their distribution is as follows: ileum 30%, ileocecal valve 30%, duodenum 10%, stomach 8%, jejunum 8%, colon 7%, and rectum 5% (42).
Killpack described a cystic, ileal duplication as a double organ “with distinct mucosa composed of stratified polygonal almost squamous, cells, the most superficial of which formed a continuous layer of ciliated epithelium” (55). Thus, although little is known in mammals about the molecular mechanisms underlying duplications, it appears that they likely recapitulate most steps of GI development. Again, one may postulate that genes involved in lumen formation via vacuolization would be likely candidates to explore.
During human embryonic development, the yolk sac and the developing foetus are joined by the vitelline duct. This duct generally narrows and disappears by the ninth week of gestation. If this fails to occur, the proximal part of the vitelline duct persists as Meckel diverticulum (56). It is the most frequent congenital anomaly of the GI tract, affecting as much as 2% of the general population (57). Although this malformation does not compromise life expectancy, as many as 25% of individuals will present with a complication. Complications include inflammation (diverticulitis), haemorrhage, intussusception, small bowel obstruction, stone formation, and malignant transformation (57).
The molecular mechanisms governing the involution of the vitelline duct are incompletely understood; however; evidence from human samples suggests that CDX2 may be one of the actors (58). In human samples of Meckel diverticulum associated with gastric heterotopias, CDX2 expression was absent from the epithelium (59). Likewise, diverticula and gastric heterotopia were observed in Cdx2+/− mice, suggesting that this association arises in the absence of 1 wild-type allele (60). Finally, SHH overexpression in human surgical specimens is in support of a role for CDX2 because CDX2 is known to repress SHH signalling (61,62).
Colonic atresia is a rarer event than congenital luminal obstruction of the small bowel. The incidence is estimated to be approximately 1 in 20,000 live births (63). Colonic atresia is often associated to other developmental anomalies such as jejunoileal atresias, Hirschsprung disease, and genitourinary malformations.
The most common hypothesis for colonic atresia is that a vascular accident leads to decreased intestinal perfusion and ischaemia (46), resulting in tissue necrosis and, in severe cases, complete luminal obstruction. This hypothesis was verified for the small bowel, but no study has demonstrated a role for vascular problems in the advent of colonic atresia (45,46).
Recent animal studies challenge this theory by suggesting that developmental, genetic mechanisms may play a role. Both Fgf10−/− and Fgfr2b−/− animals display colonic atresia in addition to duodenal atresia. The proposed mechanism is that altered mesoderm–endoderm signalling leads to decreased proliferation and increased apoptosis of epithelial cells (36,64).
Another mouse model offering insight into the heritable origins of colonic atresia is the Cdx2 null mouse. CDX2 is a homeobox-containing transcription factor, which is expressed in the developing foregut and hindgut, where it regulates proliferation and differentiation along the crypt–villus axis (58,65,66). Again, this model suggests that misregulated epithelial proliferation contributes to maldevelopment of the gut lumen, which is in keeping with the recanalization theory (12); however, to date, CDX2 has been shown to participate in human GI neoplasias, but has not been identified in samples from infants with colonic atresia (67).
Based on this and other studies, it is safe to conclude that molecular regulation of epithelial–mesenchymal interactions in the gut is essential in controlling epithelial proliferation. Intestinal atresias may be the results of localized misregulation of epithelial proliferation in utero. What is unclear is why most atresias are focal or localized rather extending over long segments. One postulate is that multiple, redundant signalling pathways overlap to ensure lumen patency and that maldevelopment ensues in areas where they fail to overlap.
Colonic and rectal duplications are among the rarest forms of intestinal duplication. Most are silent and detected during the investigation of genitourinary malformation. Because colonic duplications have the potential for malignant transformation, detection is important (68).
The cystic form (80%) is more common than the tubular form (20%), and both types may communicate with the lumen. They are located on the mesenteric border of the bowel and may affect any segment of the colon, including the rectum (69). By histology, mucosa, submucosa, and muscularis propria are readily identified (42). From a developmental point of view, it is remarkable that all intestinal segments are prone to atresias and duplications, although at varying degrees, because molecular regulation and plasticity of the intestine differs from one segment to another (70).
Little is known about the genetic causes of colonic duplication. Defects in vacuolization and recanalization are the overriding theory, much like what is assumed for the jejunum and ileum (38,39). Others have suggested that caudal syndromes may affect the distal bowel together with the notochord and genitourinary organs (split notochord, embryonic diverticula) (68).
Anorectal malformations (ARMs) are a frequent clinical problem for paediatric surgeons and gastroenterologists alike. The spectrum of malformations extends from anal stenosis to imperforate anus with or without fistula (rectovesical, prostatic, urethral, or vaginal) (71). The incidence is 1:2500 to 1:5000 (72). More than 50% of ARMs are associated with other malformations affecting the trachea, oesophagus, duodenum, colon, spinal cord, kidneys, urinary tract, and heart (71). That ARMs are associated with multiple other anomalies and syndromes speaks to the molecular complexity regulating the development of the distal GI tract in concert with neighbouring structures.
Understanding of the molecular network governing the coordinated organogenesis of the caudal part of the embryo is still in its infancy. Nonetheless, numerous chromosomal abnormalities have been identified, the most significant of which is the 7q36 deletion, for its effect on the function of the homeobox gene HLXB9 (73). HLXB9 is now recognized as causally linked to the Currarino triad, which includes a presacral mass, an anterior meningocele, and rectal duplications (71,74,75). Previously, HLXB9 was known for its role in pancreatic development (74). By virtue of the fact that HLXB9 is a homeobox gene, its role in anterior–posterior patterning is intuitive; however, it is only 1 actor in an intricate ontogenic process regulating the complex, caudal development of the embryo. Indeed, it is not causally linked to other ARM such as caudal regression syndrome (75). Furthermore, the role of HLXB9 in the Currarino triad remains unclear, and presently there is no available mouse model to explore its function because the Hlxb9−/− mouse does not exhibit vertebral or ARMs (73,76). It is possible that the molecular regulation of a sphincter region differs between lower mammals and humans.
SHH and HLXB9 are in proximity at the 7q36 locus (73); however, to date, no SHH mutation has been identified in human ARM. Animal models are plentiful to demonstrate the critical role of SHH signalling in hindgut development. Shh−/− mice show ARM resembling human disease. Gli2−/−or Gli3−/− knockouts phenocopy the Shh−/− mouse (77); thus, SHH signalling, discussed for its role in proximal GI development, is also a likely determinant of spatial and temporal morphogenesis of the anorectal region.
Studies in zebrafish and chick have, respectively, identified proprotein convertase subtilisin kexin 5 (pcsk5) (78) and its target, the growth differentiation factor (GDF11) (79), a member of the transforming growth factor-β (TGF-β) signalling cascade, as potential actors in the intricate morphogenetic events occurring in the genitourinary and anorectal regions of the developing embryo. Rodent models for pcsk5 or Gdf11 loss-of-function display ARM and altered Hlxb9 expression, resembling what is observed in humans (80,81). Although to date no mutation in these genes has been identified in humans, it is likely that a PCSK5–GDF11 cascade participates in distal Hox gene expression, which in turn, coordinates early distal patterning of urological, skeletal, genitourinary, and anorectal development primordia (81).
Persistent cloaca is a rare malformation affecting 1:250,000 live-born girls, in which the urinary tract, vagina, and rectum are fused, creating a single channel, the cloaca. Although this malformation is at the crossroads of GI, genital, and urinary development, and thus slightly beyond the scope of this review, it is worth expanding on briefly because it is a poignant illustration of an intricate developmental process gone awry.
In brief, development of the distal part of the embryo requires a coordinated orchestration for normal morphogenesis of 3 tubular systems: urinary, GI, and genital. It is accepted that the caudal lateral mesenchyme is central to this process. Based on mouse and zebrafish studies, the SHH signalling cascade is essential (77,82–84), although no mutation has been identified in human cases. How SHH interacts with the PCSK5–GDF11–HLXB9 sequence is unclear. Finally, SHH is central to so many embryonic processes, which may explain why no mutations have been identified to date in humans; it is likely that in humans any such mutation leads to embryonic or perinatal mortality.
Finally, from an ontogenic perspective, it is tempting to suggest that distal anorectal and genitourinary development is a sort of “developmental Achilles heel” of the vertebrate embryo. Indeed, developmental expression patterns of Hox genes and other transcription factors, as well as signalling pathways, are enriched in this region (Fig. 3), as they are in the foregut. Yet, the molecular and cellular processes involved are so complex that minor variations of the normal developmental course lead to significant morphological (and clinical) events, suggesting that any redundant signalling mechanism is at least in part insufficient to prevent against ARM. In other words, one may hypothesize that areas of little or no overlap between signalling cascades or morphogens may be particularly vulnerable to developmental errors (Fig. 3).
Hirschprung disease (HD) affects approximately 1 in 5000 infants. It is a congenital disease characterized by colonic dysmotility owing to the absence of myenteric innervation necessary for normal motility. The enteric nervous system (ENS) is derived from neural crest cells (NCCs), which migrate to constitute the vagal and sacral plexuses during development. Improper migration through the sacral ganglia leads to impaired distal motility, characterized on imaging by a thin distal segment and a dilated proximal colon. Furthermore, NCC interact with neighbouring cells: both interactions with the epithelium and the mesenchyme are essential for NCC maintenance and function. Naturally, such a complex system is regulated by numerous signalling cascades and transcription factors, making for a vast array of candidate genes involved in defective NCC migration, differentiation, and homeostasis (85).
Briefly, NCC receptors recognize signalling molecules synthesized by intestinal mesenchymal cells to which they respond by activating cell-fate and cell-survival transcription factors (Fig. 4). The 2 important pairs are the glial-cell-line–derived neurotrophic factor (GDNF)-ret receptor tyrosine kinase (86) involved in proliferation of progenitors located at the distal extremity of the migrating ENS, and the endothelin-3 (ET-3)–endothelin-B receptor pathway regulating the timing of NCC differentiation during colonization of the gut (87). Mutations in both of these ligand-receptor pairs have been identified in rodent models and human cases of HD (88–90). Recently, mutations of RET ligands have also been identified, probably associated with lack of receptor binding (91).
The 2 ligand-receptor pairs mentioned above are regulated by 2 transcription factors: SRY-related HMG-box transcription factor 10 (SOX10) and paired-like homeobox factor 2B (PHOX2B). SOX10 is a key transcription factor expressed in NCC during early human development, which contributes to peripheral nervous system development including enteric neurons (92). The transcription factor PHOX2B is involved in the development of several noradrenergic neuron populations and autonomic nervous system regulation (93). Mice heterozygous for the Sox10 locus exhibit a phenotype similar to the human disease characterized by colonic aganglionosis and hypopigmentation (92), and enteric ganglia are absent in mice with a homozygous disruption of the PHOX2B gene (71). Mutations in both SOX10 and PHOX2B have been identified in human subjects with aganglionic megacolon (89,93). The molecular network involved in NCC migration to the colonic ganglia, and thus in colonic aganglionosis, is summarized in simple terms in Figure 3.
Regulation and misregulation of NCC migration, proliferation, and differentiation illustrate the limits of the 1 gene–1 phenotype concept of heritable disease. Clearly, mutations at any level in the cascade can lead to similar phenotypes. This point cannot be overemphasized at the bedside: a disease cannot be ruled out on the sole basis of an absent mutation in 1 of the candidate genes. As the unravelling of the molecular basis of disease evolves, clinicians will increasingly think in terms of molecular networks and cascades when approaching congenital malformations and heritable diseases with variable phenotypes.
Abdominal Wall Defects
Gastroschisis is a defect of the anterior abdominal wall affecting approximately 3/10,000 live births, in which the intestines (and sometimes the stomach and other organs) of the newborn protrude into the amniotic cavity. The abdominal wall defect occurs usually to the right of the umbilicus (94).
Gastroschisis is a typical clinical example of an early developmental accident (between 5 and 10 weeks’ gestation) (94). It is a common birth defect and its incidence is on the rise (95). Nonetheless, its aetiology is unknown and has generated numerous hypotheses. The first is that the somatic layer of the mesoderm does not form normally owing to teratogen exposure (96). Second, in 1975, Shaw (97) proposed that gastroschisis may be caused by a rupture of the amniotic membrane at the base of the umbilical cord secondary to ring closure. This theory did not hold because the umbilical ring is intact in nearly all patients with gastroschisis. Third, deVries (98) suggested a vascular theory whereby premature atrophy or abnormal persistence of the right umbilical vein leads to impaired growth and viability of the surrounding mesenchyme; however, this interesting hypothesis was incorrect because the umbilical vein does not supply the intestinal mesenchyme. Another theory proposed that a vitelline artery abnormality (contributing to ischaemia) leads to abdominal wall weakness and subsequent rupture of the intestine through this defective region, leading to gastroschisis (99); however, this hypothesis is improbable because vitelline arteries supply the yolk sac rather than the body wall. Furthermore, considering that the vitelline artery supplies the yolk sack as a network, it is unclear why one area of the abdominal wall would be more prone to ischaemia than another (100). The most recent and plausible hypothesis is that an abnormal body fold closure may be the cause (101). Improper closure of a body fold could impede merging of the yolk stalk with the connecting stalk. The result is then that the intestine rather than the umbilical cord herniates into the amniotic cavity with the vitelline duct (101).
The genetic basis of gastroschisis is under investigation. As in other developmental malformations, many of the studies have been performed in lower mammals, limiting the generalizability of the findings to humans. Today, polymorphisms in 4 genes have been identified as possibly linked to human cases of gastroschisis: intracellular adhesion molecule 1 (ICAM1), endothelial nitric oxide synthase 3 (NOS3), atrial natriuretic peptide (NPPA), and alpha-adducin (ADD1) (102).
ICAM1 induces and controls endothelial cell activation (103), ADD1 is important in cell proliferation and epidermal differentiation (104), NOS3 plays a critical role in neovascularization (105), and NPPA is implicated in the control of extracellular fluid volume and electrolyte homeostasis. Although the interplay among these genes is unclear, it has been shown that the risk of gastroschisis is higher in children with ICAM1 and NOS3 single-nucleotide polymorphisms (SNPs) whose mothers smoked during pregnancy (102). Lammer et al (106) hypothesized that tobacco impairs VEGF–NOS3 and ICAM1 signalling, which in turn leads to defective angiogenesis and vascular remodelling, in keeping with earlier suggestions that vascular defects are at the root of gastroschisis. This theory is an elegant illustration of the potential role of environment on gene expression during morphogenesis.
It is noteworthy that to date, no association has been observed between early somatic (IROQUOIS homeobox factor) and splanchnic mesoderm (FOXF1) genes and abdominal wall defects reviewed by McLin et al (70). Although there is no established animal model of gastroschisis, Shroom3−/− mice display massive gastroschisis (107). Shroom3 is a PDZ (postsynaptic density protein [PSD95], Drosophila disc large tumour suppressor [Dlg1], and zonula occludens-1 protein [zo-1]) protein that regulates cell shape.
Omphalocele represents nearly 50% of congenital abdominal wall defects (95). Unlike gastroschisis, which occurs independently of the umbilical cord, omphalocele is a hernia around the base of the umbilical cord. Severity is proportional to the size of the intestinal or liver protrusion through the opening (108).
The causes of omphalocele are mostly unknown. Omphalocele is often associated with other birth defects, including single gene disorders, neural tube defects, diaphragmatic defects, and >20 syndromes of unknown aetiology (108). Among these syndromes, 2 are noteworthy for their association with specific gene defects. The first is the OtoPalatoDigital (OPD) syndrome-spectrum disorder, which has been linked to filamin A (FLNA) mutations (109); however, no FLNA mutation has ever been identified in a patient presenting with omphalocele only (110). Next, Rieger syndrome, which is characterized by eye and tooth abnormalities together with abdominal wall defects (111), has been linked to both haploinsufficiency and gain-of-function mutations of PITX2 (112). PITX2 is a bicoid-related Hox gene expressed in the somites and lateral plate mesoderm, where it plays an important role in primary myogenesis (113). It has recently been incriminated in loss of abdominal musculature in a mouse model (114). Interestingly, a human study suggests that mutations in PITX2 may also rarely be linked to isolated omphalocele (110). Unlike experimental models of gastroschisis in which extrapolation to humans is difficult, Pitx2−/− mice demonstrate abnormal cellularity of the body wall and failure of body wall closure similar to what is observed in human omphalocele (115). Finally, in keeping with the environmental exposure theory put forward for gastroschisis, chicks exposed to cadmium show an increased incidence of omphalocele and a decrease in PITX2 mRNA expression (116).
In summary, the role of PITX2 in GI development is manifold. It is known for its role in regulating mesenchymal genes that contribute to the first gut tilt to the left (117). Therefore, it is tempting to speculate that PITX2 may act on other mesenchymal components besides the dorsal mesentery. Finally, it has recently been shown in X laevis to be upstream of shroom3 together with other Pitx proteins whereby it regulates epithelial morphogenesis and cell shape (118). Although Pitx proteins are likely actors in the advent of omphalocele, they play several other roles in gut development.
Recent studies show that SNPs in genes related to homocysteine metabolism via vitamin B12 and folate have been associated with omphalocele. These include transcobalamin receptor (TCblR), transcobalamin 2 (TCN2), and methylenetetrahydrofolate reductase (MTHFR). On the one hand, it is reasonable to assume that akin to what has been described in neural tube defects, the administration of folate and vitamin B12 during pregnancy may be justified to prevent omphalocele (119). On the other hand, folate is of increasing interest for its epigenetic footprint, raising the possibility of the likely contribution of epigenetics.
Owing to the complexity of intestinal development and gut coiling, it is reasonable to assume that no single gene is responsible for any of these abdominal wall defects. Rather than a 1 gene–1 malformation situation, it is more likely that omphalocele, or gastroschisis, is the product of any number of signalling defects or cell-fate decisions, controlled by genes, the environment, and the epigenome.
Malrotation and Other Syndromes
The midgut includes the duodenum, the jejunum, the ileum, the cecum, the ascending colon and approximately two-thirds of the transverse colon. Its development is characterized by 2 major morphogenetic events: rotation and fixation of the colon and dorsal mesentery (120). These events require complex and tightly orchestrated molecular and cellular events and as such, are vulnerable to frequent developmental “glitches.” Malrotation occurs in as many as 1:500 live births and is the presumed consequence of abnormal GI rotation or reinsertion into the body cavity (121,122). Its morphological characteristics include right-sided small bowel, a cecum in the epigastrium, narrow mesenteric base, and rightward shift of the ligament of Treitz (123,124). The clinical consequences include obstructive symptoms owing to acute or chronic volvulus and acute bowel necrosis (121). On occasion, malrotation is diagnosed in isolation. More typically, however, it is associated with heterotaxy and other congenital malformations of the GI tract.
Recent studies show that intestinal rotation is initiated by key ultrastructural changes in the dorsal mesentery. Pitx2 regulates cell-shape changes, which in turn induce left–right asymmetries via islet-1 (Isl-1), thereby determining the chirality of midgut looping (117). This important study is the first to identify early molecular regulators of this critical morphogenetic event in GI development. To date, however, these findings have not been confirmed in human subjects. Pitx proteins were discussed earlier with respect to gastroschisis and their effect on Shroom 3, another regulator of cell shape, raising the obvious question of the link between chirality and the advent of gastroschisis.
In contrast, little data from humans are available. Duodenal atresia has been identified in Feingold syndrome, described in the setting of oesophageal atresia (35). In the first instance, the associated gene defects were identified by screening cases of malrotation associated with other laterality defects. For example, mutations of cryptic family 1 (CFC1) were identified in patients presenting with heart and stomach laterality or chirality defects (125). This gene is known to be involved in the NODAL signalling pathway, critical in early left–right patterning (125,126). The other mutations that have been identified in syndromic malrotation (as opposed to isolated) are the following: zinc finger protein ZIC3 (ZIC3), previously shown to regulate left–right asymmetry in X laevis (127–130), the growth and differentiation factors activin receptor type 2B (ACVR2B) (131), and the highly homologous factor LEFTYA (human homologue of the mouse LEFTY1), known for its role in left–right patterning (131,132). Mutations in the gene encoding the early mesoderm transcription factor NKX2.5 deserve special mention because the NKX2.5 promoter has PITX2 binding sites, further highlighting the many functions of PITX2 during early vertebrate gut development (133); however, at present, none of these are specific enough to be of any use for antenatal screening purposes.
Of note, Stankiewicz et al (35) showed that human neonates experiencing the rare and fatal alveolar capillary dysplasia with misalignment of pulmonary veins (ACDMPV), attributed to FOXF1 heterozygosity, often display associated malrotation. These findings are compelling because they implicate the key transcription factor of the visceral mesoderm in a syndrome, which includes malrotation. On the contrary, it does not prove causality; rather, it merely adds a candidate to the long list of partners involved in this intricate process. Furthermore, Foxf1 loss-of-function is well characterized in both mouse and X laevis, and malrotation is not a predominant feature of the phenotype in either species (14,134,135).
In conclusion, although a few genetic associations have been established in human syndromes presenting with malrotation, it is obvious that gut coiling is governed by a regulatory network rather than a single gene. It follows that any number of clinical syndromes can present with associated malrotation. In fact, it is generally accepted in the field of GI embryology and patterning that any number of genetic or environmental modifications can disrupt the delicate sequence of events leading to normal gut looping, something readily observed in the amphibian X laevis.
Heterotaxy and Situs Inversus
Malrotation presents as part of heterotaxy syndromes, which stem from abnormal left–right cell-fate decisions. Heterotaxy comprises situs inversus and situs ambiguous and may present as any of a number of visceral malpositions above and below the diaphragm (136). As detailed, the NODAL signalling pathway is most often involved because of its role in the lateral plate mesoderm in left–right decisions. Importantly, for the gastroenterologist, these syndromes are often associated with malrotation (137), biliary atresia (138), or heart defects (137). As may be predicted, mutations in the following genes discussed in the section on malrotation were identified in human subjects: NODAL (139), ZIC3 (140), ACVR2B (141), LEFTYA (142), CFC1 (143), and NKX2.5 (133). Mutations of these genes involved in the propagation of left–right asymmetry to the left lateral plate mesoderm are causes of partial situs inversus, whereas mutations of gene that generate left–right information in the node are responsible of complete situs inversus.
Full situs inversus is associated with ciliary defects, which are known to control left–right decisions in the early embryo. Primary ciliary dyskinesia or infantile nephronophthisis (NPHP2) is associated with situs inversus in 50% of patients. It has been shown to be associated with mutations in the following dynein-related genes (dynein is an important ciliary protein): the dynein intermediate chain (DNAI1) (144) and the dynein heavy chains DNAH5 (145) or DNAH11 (146).
Gap junctions have also been incriminated in heterotaxy because mutations in the connexin subunit CX43 were identified in patients with laterality defects (147), something shown in X laevis (148,149).
Congenital Short Bowel Syndrome
Short bowel syndrome is a frequent cause of intestinal failure in infancy and the most common indication for intestinal transplantation in children. It is usually an acquired condition following massive intestinal resection. A few cases of congenital short bowel have been described (150–153). Recently, van der Werf et al (154) linked mutations in coxsackie and adenovirus receptor–like membrane protein (CLMP) to familial short bowel. They showed that CLMP is a membrane protein expressed in the intestinal crypts of human foetal samples. Using a zebrafish model, they showed that clmp loss-of-function is associated with multiple abnormalities, including short bowel. Although it is unclear whether the shortened intestinal length in these morphant embryos is primary or secondary to global developmental abnormalities, it nonetheless is an interesting finding for 2 reasons: first, it is the first such description in an amphibian model, and second, it implicates an epithelial protein in gut elongation, something classically attributed to the mesoderm (Fig. 2).
Much of the molecular underpinnings of GI malformations and syndromes are being understood through the use of mammalian and nonmammalian developmental biology models. The mechanisms governed by the genes identified to date are still incompletely understood, although genetic mutations are increasingly recognized in humans. Comparative embryology should pave the way to deciphering the molecular network governing the complexity of GI development and how environmental influences affect phenotypic expression. The hope is also that these advances will ultimately facilitate in vitro cellular differentiation for the purposes of cell-based therapies and tissue engineering, in the footsteps of recent accomplishments in the field of hepatic embryology.
1. Spitz L. Esophageal atresia. Lessons I have learned in a 40-year experience. J Pediatr Surg 2006; 41:1635–1640.
2. Ramalho-Santos M, Melton DA, McMahon AP. Hedgehog signals regulate multiple aspects of gastrointestinal development. Development 2000; 127:2763–2772.
3. Chung E. Infantile hypertrophic pyloric stenosis: genes and environment. Arch Dis Child 2008; 93:1003–1004.
4. Noah TK, Donahue B, Shroyer NF. Intestinal development and differentiation. Exp Cell Res 2011; 317:2702–2710.
5. Zorn AM, Wells JM. Vertebrate endoderm development and organ formation. Annu Rev Cell Dev Biol 2009; 25:221–251.
6. Zorn AM. Girard L. Liver development. Harvard Stem Cell Institute, StemBook. Cambridge, MA:2008.
7. Zhao R, Duncan SA. Embryonic development of the liver. Hepatology 2005; 41:956–967.
8. Bonal C, Herrera PL. Genes controlling pancreas ontogeny. Int J Dev Biol 2008; 52:823–835.
9. Shaw-Smith C. Oesophageal atresia, tracheo-oesophageal fistula, and the VACTERL association: review of genetics and epidemiology. J Med Genet 2006; 43:545–554.
10. Hajduk P, Sato H, Puri P, et al. Abnormal notochord branching is associated with foregut malformations in the adriamycin treated mouse model. PLoS One 2011; 6:e27635.
11. Minoo P, Su G, Drum H, et al. Defects in tracheoesophageal and lung morphogenesis in Nkx2.1(−/−) mouse embryos. Dev Biol 1999; 209:60–71.
12. Ioannides AS, Massa V, Ferraro E, et al. Foregut separation and tracheo-oesophageal malformations: the role of tracheal outgrowth, dorso-ventral patterning and programmed cell death. Dev Biol 2010; 337:351–362.
13. Arsic D, Cameron V, Ellmers L, et al. Adriamycin disruption of the Shh-Gli pathway is associated with abnormalities of foregut development. J Pediatr Surg 2004; 39:1747–1753.
14. Mahlapuu M, Enerback S, Carlsson P. Haploinsufficiency of the forkhead gene Foxf1, a target for sonic hedgehog signaling, causes lung and foregut malformations. Development 2001; 128:2397–2406.
15. van Noort M, Clevers H. TCF transcription factors, mediators of Wnt-signaling in development and cancer. Dev Biol 2002; 244:1–8.
16. Petrackova I, Pozler O, Kokstein Z, et al. Association of oesophageal atresia, anophthalmia and renal duplex. Eur J Pediatr 2004; 163:333–334.
17. Delahaye A, Sznajer Y, Lyonnet S, et al. Familial CHARGE syndrome because of CHD7 mutation: clinical intra- and interfamilial variability. Clin Genet 2007; 72:112–121.
18. Sanlaville D, Etchevers HC, Gonzales M, et al. Phenotypic spectrum of CHARGE syndrome in fetuses with CHD7 truncating mutations correlates with expression during human development. J Med Genet 2006; 43:211–217.
19. Jongmans MC, van Ravenswaaij-Arts CM, Pitteloud N, et al. CHD7 mutations in patients initially diagnosed with Kallmann syndrome: the clinical overlap with CHARGE syndrome. Clin Genet 2009; 75:65–71.
20. van Bokhoven H, Celli J, van Reeuwijk J, et al. MYCN haploinsufficiency is associated with reduced brain size and intestinal atresias in Feingold syndrome. Nat Genet 2005; 37:465–467.
21. Dorney SF, Middleton AW, Kozlowski K, et al. Congenital agastria. J Pediatr Gastroenterol Nutr 1987; 6:307–310.
22. Nagendran S, Johal N, Set P, et al. Bilateral communicating intralobar sequestration and microgastria. Ann Thorac Surg 2009; 88:2040.
23. Filippi L, Serafini L, Fiorini P, et al. Congenital microgastria and primary ciliary dyskinesia in a newborn with DiGeorge syndrome and 22q11.2 deletion. Eur J Pediatr Surg 2008; 18:195–197.
24. Feenstra B, Geller F, Krogh C, et al. Common variants near MBNL1 and NKX2-5 are associated with infantile hypertrophic pyloric stenosis. Nat Genet 2012; 44:334–337.
25. Chung E, Curtis D, Chen G, et al. Genetic evidence for the neuronal nitric oxide synthase gene (NOS1) as a susceptibility locus for infantile pyloric stenosis. Am J Hum Genet 1996; 58:363–370.
26. Huang PL, Dawson TM, Bredt DS, et al. Targeted disruption of the neuronal nitric oxide synthase gene. Cell 1993; 75:1273–1286.
27. Smith DM, Nielsen C, Tabin CJ, et al. Roles of BMP signaling and Nkx2.5 in patterning at the chick midgut-foregut boundary. Development 2000; 127:3671–3681.
28. Moniot B, Biau S, Faure S, et al. SOX9 specifies the pyloric sphincter epithelium through mesenchymal-epithelial signals. Development 2004; 131:3795–3804.
29. Qi BQ, Beasley SW. Stages of normal tracheo-bronchial development in rat embryos: resolution of a controversy. Dev Growth Differ 2000; 42:145–153.
30. Matsumoto A, Hashimoto K, Yoshioka T, et al. Occlusion and subsequent re-canalization in early duodenal development of human embryos: integrated organogenesis and histogenesis through a possible epithelial-mesenchymal interaction. Anat Embryol (Berl) 2002; 205:53–65.
31. Bianchi DW, Crombleholme TM, D’Alton ME. Duodenal atresia. In: Fetology: Diagnosis and Management of the Fetal Patient. 2000. New York: McGraw-Hill;
32. Nawaz A, Matta H, Hamchou M, et al. Situs inversus abdominus in association with congenital duodenal obstruction: a report of two cases and review of the literature. Pediatr Surg Int 2005; 21:589–592.
33. Hebrok M, Kim SK, St Jacques B, et al. Regulation of pancreas development by hedgehog signaling. Development 2000; 127:4905–4913.
34. Sukegawa A, Narita T, Kameda T, et al. The concentric structure of the developing gut is regulated by Sonic hedgehog derived from endodermal epithelium. Development 2000; 127:1971–1980.
35. Stankiewicz P, Sen P, Bhatt SS, et al. Genomic and genic deletions of the FOX gene cluster on 16q24.1 and inactivating mutations of FOXF1 cause alveolar capillary dysplasia and other malformations. Am J Hum Genet 2009; 84:780–791.
36. Fairbanks TJ, Kanard R, Del Moral PM, et al. Fibroblast growth factor receptor 2 IIIb invalidation: a potential cause of familial duodenal atresia. J Pediatr Surg 2004; 39:872–874.
37. Kanard RC, Fairbanks TJ, De Langhe SP, et al. Fibroblast growth factor-10 serves a regulatory role in duodenal development. J Pediatr Surg 2005; 40:313–316.
38. Pumberger W, Maier-Hiebl B, Kargl S. Recurrent pancreatitis due to an intraluminal duodenal diverticulum: report of a case. Surg Today 2012; 42:589–592.
39. Van Beers B, Trigaux JP, De Ronde T, et al. CT findings of perforated duodenal diverticulitis. J Comput Assist Tomogr 1989; 13:528–530.
40. Bagnat M, Cheung ID, Mostov KE, et al. Genetic control of single lumen formation in the zebrafish gut. Nat Cell Biol 2007; 9:954–960.
41. Tongsin A, Anuntkosol M, Niramis R. Atresia of the jejunum and ileum: what is the difference? J Med Assoc Thai 2008; 91 (Suppl 3):S85–S89.
42. Stern LE, Warner BW. Gastrointestinal duplications. Semin Pediatr Surg 2000; 9:135–140.
43. Nichol PF, Reeder A, Botham R. Humans, mice, and mechanisms of intestinal atresias: a window into understanding early intestinal development. J Gastrointest Surg 2011; 15:694–700.
44. Tandler J. Zur Entwicklungsgeschichte des Menschlichen Duodenum in Fruhen Embryonalstadien. Morphol Jahrb 1900; 29:187–216.
45. Louw JH. Congenital atresia and stenosis of the small intestine. The case for resection and primary end-to-end anastomosis. S Afr J Surg 1966; 4:57–64.
46. Louw JH, Barnard CN. Congenital intestinal atresia; observations on its origin. Lancet 1955; 269:1065–1067.
47. Nayci A, Avlan D, Polat A, et al. Ileal atresia associated with a congenital vascular band anomaly: observations on pathogenesis. Pediatr Surg Int 2003; 19:742–743.
48. Mitchell J, Punthakee Z, Lo B, et al. Neonatal diabetes, with hypoplastic pancreas, intestinal atresia and gall bladder hypoplasia: search for the aetiology of a new autosomal recessive syndrome. Diabetologia 2004; 47:2160–2167.
49. Pearl EJ, Jarikji Z, Horb ME. Functional analysis of Rfx6 and mutant variants associated with neonatal diabetes. Dev Biol 2011; 351:135–145.
50. Spiegel R, Dobbie A, Hartman C, et al. Clinical characterization of a newly described neonatal diabetes syndrome caused by RFX6 mutations. Am J Med Genet A 2011; 155A:2821–2825.
© 2013 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,