Formation of the Neural Tube

Among the earliest morphological specializations in the embryo is the formation of the neural placode, followed by neural plate and then neural tube [Sadler, 2005]. At the end of the second gestational week, the human embryo is a bilaminar disc of epiblast cells overlying hypoblast cells [Sadler, 2004]. At the beginning of the third week, the disc develops a midline groove, the primitive streak, in the caudal third (Figure 22-1A), which marks the initiation of gastrulation and the formation of three germ layers – ectoderm (giving rise to skin and the nervous system), mesoderm (providing inductive signals to ectoderm and contributing to morphogenesis), and endoderm (giving rise to viscera). The primitive node at the cranial end of the streak contains cells that act to organize the embryonic axes (see Figure 22-1). At the same stage, thickening of the rostral ectoderm by the apical–basal elongation of cells into a pseudostratified columnar shape produces the neural placode that marks the initiation of neurulation (see Figure 22-1B) [Schoenwolf and Powers, 1987; Schoenwolf, 1988; Colas and Schoenwolf, 2001]. Cells migrating through the primitive streak and node displace the hypoblast cells to form endoderm and subsequently middle layer mesoderm. Cells that migrate through the node in the midline form the prechordal plate and notochord (Figure 22-2), which are important for induction of the ventral CNS structures, starting with the neural plate. The remainder of ectoderm surrounding the neural plate becomes epidermis.

Fig. 22-1 Early stages of gastrulation and neurulation in human embryogenesis: views of the dorsal surface.

A, The primitive streak emerges as a groove at the caudal pole of the embryo, with the primitive node at the cranial pole. Epiblast cells moving through the primitive streak and node form the three germ layers – ectoderm, mesoderm, and endoderm. B, A thickening of the ectoderm forms the neural plate, initiated at the cranial pole, while the primitive streak at the caudal end initiates gastrulation. By 20 days’ gestation, the neural folds have elevated from the neural plate, and tips of the folds are closing. Somites, comprised of mesoderm cells, support the elevation of the neural folds and are the precursors of vertebrae.

(Adapted from Sadler TW. Embryology of neural tube development. Am J Med Genet C Semin Med Genet 2005;135C:2–8.)

Fig. 22-2 Cross-sections through the primitive streak.

A, A transaxial view shows movement of epiblast cells through the streak to form a mesodermal layer between the original epiblast (becoming the ectodermal layer) and hypoblast cells (becoming the endoderm). B, A midsagittal section shows the disc-shaped embryo suspended between yolk sac and amniotic cavities. The notochord and prechordal plate have formed by the displacement of cells through the primitive node.

(Adapted from Sadler TW. Embryology of neural tube development. Am J Med Genet C Semin Med Genet 2005;135C:2–8.)

Neural plate formation is actually a default state of the ectoderm, and formation of epidermis involves the inhibition of bone morphogenetic protein (BMP) and the wingless (Wnt) signaling pathway [Harland, 2000]. As the plate begins to emerge, morphogenesis, or shape changes, involving groups of cells in the neuroepithelium and surround, is essential to formation of brain and spinal cord (Figure 22-3). By 20 days of gestation, the neural plate appears indented in the midline, forming a groove or medial hinge region flanked by ridges – the neural folds. These folds elevate from the plane of the neural plate through the combined influences of proliferation of neural cells and underlying mesenchymal cells [Greene and Copp, 2009]. Bending inward of the neural folds at the dorsolateral hinge points occurs through morphological shape changes of the neural cells, which become radially elongated, while their apical (luminal) poles constrict, through a combination of cell-cycle regulation that moves nuclei to the basal end of cells in the hinge region, and actin-myosin contraction at the apical poles of cells in the hinge region, to bring the tips of the neural folds into apposition [Greene and Copp, 2009]. In the head region, proliferation and movement of epidermis and mesenchyme cells also help to push the neural folds into apposition, while at spinal levels, neighboring mesenchyme may be less critical to neural tube closure [Greene and Copp, 2009]. In addition to cell elongation and proliferation, the shape changes in the neural plate are affected by cell motility in the form of convergent extension, in which laterally placed cells move to the midline and migrate rostrocaudally in a process that is mediated by noncanonical Wnt signaling (Figure 22-4) [Wallingford and Harland, 2001; Wallingford et al., 2002]. Thus, once elevation and bending of the neural folds occur, the lateral margins or tips of the folds join and then fuse in the midline to become the neural tube.

Fig. 22-3 A coronal cross-section through the cranial neural folds as they come together in the midline.

The pseudostratified appearance of the neuroepithelium is due to the translocation of cell nuclei (interkinetic nuclear migration) during cell-cycle progression, in which M-phase nuclei are at the apical (central) surface and S-phase nuclei are at the basal (outer) surface. Bending of the folds takes place at a midline, median hinge point (black dot) and at two dorsolateral hinge points (double black dots). Bending of the dorsolateral hinge points is facilitated by cell-shape changes, in which the apical poles are narrowed by actin-myosin-dependent constriction and basal expansion of cells, more of which contain translocated, S-phase nuclei. Closure is also facilitated by the movement toward the midline of surface ectoderm (ee) cells. n, notochord.

(Adapted from Sadler TW. Embryology of neural tube development. Am J Med Genet C Semin Med Genet 2005;135C:2–8.)

Fig. 22-4 Cell movements affecting planar cell polarity (PCP) signaling in the early developing nervous system.

A and B, Schematics showing cell movement toward the midline into the neural groove and longitudinally to narrow and elongate the embryo. C, This movement serves to narrow the ventral floorplate (red), facilitating median hinge-point bending and apposition of the neural folds. D, When PCP signaling is impaired, as in the looptail mouse-bearing mutation in Vangl2, the floorplate is wider (intervening cells remain in the floorplate region), which prevents the folds from meeting in the midline.

(A, From Ybot-Gonzalez P et al. Convergent extension, planar-cell-polarity signalling and initiation of mouse neural tube closure. Development 2007;134:789–799. B–D, From Copp AJ et al. Trends Neurosci: Dishevelled: Linking convergent extension with neural tube closure, 2003;26:453–455.)

In order to achieve this rising and bending inward of the neural folds, the cells of the neural plate must proliferate in an ordered manner, called interkinetic nuclear migration of progenitors, forming a pseudostratified epithelium in which S-phase occurs at the basal (outer) surface of the neural folds, mitosis (M-phase) occurs at the apical (central or luminal) surface, and G1 and G2 phase nuclei are positioned at intermediate locations. In addition, in the process of convergent extension, cells move medially and through the medial hinge region to migrate rostrocaudally and elongate the neuraxis, narrowing the ventral floorplate (see Figure 22-3). If the floorplate is too wide or the neural folds fail to elevate and bend, the folds will not appose and NTDs will ensue (see Figure 22-4). The broadening of the ventral floorplate and shortening of the rostrocaudal axis reflect impaired convergent extension; they are demonstrated in the looptail mouse mutant, which bears a mutation in the Van Gogh-like 2 (Vangl2) gene that functions in the planar cell polarity (PCP) pathway [Ybot-Gonzalez et al., 2007]. When apposition is successful, midline fusion of the neural folds, a process known as neurulation, occurs first at primary closure points, and extends by adding multiple closure points like the teeth of a zipper rostrally and caudally from each node to complete neural tube closure [Pyrgaki et al., 2010]. In the mouse, there are typically three closure points:

In humans, there are two, points 1 and 3, which are located at the hindbrain–spine junction and the frontal neuropore, respectively. The anterior neuropore, the region that will eventually give rise to the brain, closes approximately by day 26 of human gestation. The posterior neuropore, the region that will give rise to the caudal spinal column, closes approximately at day 29 of human embryogenesis. After the neural tube closes, it separates from the overlying ectoderm in a process termed dysjunction. Cells of the somitic mesoderm invade the space between the ectoderm and neural tube to form somites that eventually give rise to the posterior elements of the vertebral bodies and the paraspinal muscles. Specific neural cells at the tips of the folds are excluded from the neural tube; these cells form the neural crest, which is the anlage of the peripheral and autonomic nervous systems, and which also contributes the meninges, and portions of the skull and face.

Molecular Patterning of the Neural Tube

Work on frog, mouse, and zebrafish species has identified a host of factors that are necessary for proper neurulation and neural tube closure. These data have direct implications for human disease because vertebrates display a conserved body plan. A key observation was that a small fragment of non-neural tissue (mesoderm), when transplanted, was capable of duplicating the neural tube, indicating that secreted factors from mesoderm are sufficient to induce neurulation. Thus, the factors necessary for neurulation are intrinsic to and secreted from this mesodermal region. These factors include chordin, noggin, sonic hedgehog (SHH), and several others, which are both necessary and sufficient for proper neurulation and control of the amount and fate of the neuroectoderm (reviewed in De Robertis and Kuroda [2004]). In mammals, two distinct groups of non-neural cells appear to provide these early patterning signals: axial mesodermal cells of the notochord, which underlie the midline of the neural plate, and the cells of the epidermal ectoderm, which flank its lateral edges. The notochord is the source of ventralizing inductive factor, and the epidermal ectoderm is the source of dorsalizing factors. Opposing actions of these two signals appear to be critical in establishing the identity and pattern of cell types generated along the dorsal–ventral axis of the neural tube.

The nervous system is patterned along the anteroposterior, dorsal–ventral, and right–left axes in response to the expression of intrinsic patterning genes and critical embryonic signaling pathways involving secreted factors and cell–cell interactions [Altmann and Brivanlou, 2001]. The homeotic, Hox, transcription factor genes provide a positional “address” or identity of spinal cord and hindbrain neurons [Carpenter, 2002]. Fate determination along the dorsal–ventral axis also requires the action of three opposing signaling pathways:

In this model, SHH released by the notochord diffuses toward the ventral neural tube and induces the differentiation of the floor plate. The floor plate then becomes capable of producing additional SHH, which diffuses and establishes a gradient along the dorsal–ventral axis of the neural tube. Differentiation of cells in the dorsal half of the neural tube depends on signals such as BMPs provided by the lateral epidermal ectoderm. BMP4 and BMP7, released by the epidermal ectoderm, diffuse toward the dorsal neural tube and induce the differentiation of the roof plate. The roof plate then becomes capable of producing additional BMPs, which diffuse and establish a gradient along the ventral–dorsal axis of the neural tube. Regions of the spinal cord that are exposed to the highest concentrations of SHH and lowest concentrations of BMPs give rise to cells of ventral fate, including motor neurons, whereas cells exposed to lowest SHH and highest BMP concentrations give rise to cells of dorsal fate, such as commissural projection neurons [Cowan et al., 1997]. Fibroblast growth factor (FGF), NOTCH, and retinoic acid (RA) signaling is also known to act in dorsal–ventral patterning and neural fate determination, though their roles are less well understood. Rostrocaudally, the neural tube is regionalized into four major divisions: forebrain, midbrain, hindbrain, and spinal cord. Actions of FGF, Wnt and RA pathways confer a caudal identity [Diez del Corral and Storey, 2004; Aboitiz and Montiel, 2007] and also impact telencephalic patterning [Takahashi and Liu, 2006]. Homeotic Hox and Engrailed family transcription factors are particularly important for establishing segmental identities and morphogenetic features of hindbrain [Barrow et al., 2000; Cheng et al., 2010] (see also Chapter 24).

Incidence

Among live births, females are more affected by NTDs than males by 2:1 [Elwood and Little, 1992; Little and Elwood, 1992; Shaw et al., 1994]. Additionally, the prevalence of NTDs varies across time, by region, and by ethnicity, and approximates 0.5 cases per 1000 in the United States but 10 per 1000 in parts of China and India [Northrup and Volcik, 2000; Gelineau-van Waes and Finnell, 2001]. These numbers do not reflect the decrease in incidence that has followed FA supplementation – with current estimates of 0.37 cases per 1000 reported in the US [Mills and Signore, 2004] – or the availability of prenatal diagnosis and the elective termination of some pregnancies with affected fetuses. The geographic and population variabilities in NTD rates are in keeping with the complexities of the etiology of neural tube closure failure, which involve both genetic and environmental factors [Janerich and Piper, 1978; Ross, 2010].

Over 30 years of clinical and basic science investigation are behind the insight that NTDs in humans are not caused by a single-gene defect but instead arise from the interplay of multiple genes, as well as gene–environment interactions [Holmes et al., 1976]. The twin concordance rate among same-sex twins (presumed monozygotic) is significantly increased (to 6.8 percent), supporting a genetic contribution [Janerich and Piper, 1978]. Furthermore, compared to an incidence in the general population of 1 per 1000, the risk of NTD recurrence in a family with one affected child increases to 1.8 per 100, but does not approach the 1 in 4 recurrence risk of an autosomal-recessive mutation with complete penetrance. Studies have documented the higher occurrence rate of NTDs in females compared to males, and the significantly higher incidence of consanguinity among parents of infants with NTDs [Elwood and Little, 1992; Shaw et al., 1994]. These and many other studies indicate prominent genetic, heritable contributions to failed neural tube closure that may well require the concerted action of multiple gene polymorphisms to manifest the disorder.

Complex Genetic Contributions

The complexity of the genetic underpinnings of neurulation is reflected in the NTD-prone mouse lines, for which over 200 mutations have been associated with failure of neural tube closure [Harris and Juriloff, 2007; Harris, 2009]. Despite this wealth of information, there is no single-gene polymorphism in the human homologs of these mouse genes that confers a robust, broadly reproducible enhanced risk of developing an NTD [Greene et al., 2009; Ross, 2010]. This has led to the supposition that either neurulation in the mouse is significantly different from that in humans, despite the similarities in morphogenesis and cell behaviors, or the genesis of NTDs requires the compounding of multiple gene polymorphisms. Supporting this latter premise, there are a number of examples in which modifier loci have been detected that increase the penetrance of NTD in mutant mouse models. A classic example is found the curly tail (ct) mouse, in which spina bifida is associated with a primary risk gene and at least three distinct modifier loci [Neumann et al., 1994]. The hypomorphic, partial loss-of-function mutation most responsible for NTDs in ct/ct mouse embryos occurs in a transcription factor called Grainy-head-like 3 (Grhl3). Another mutant with demonstrated multiple gene interactions is the SELHBc mouse, for which four loci have been mapped that conspire to result in exencephaly [Harris and Juriloff, 2007]. There are also examples of digenic mutations, or mutations in two genes, that must occur together to produce NTD in the mouse [Harris and Juriloff, 2007]. These digenic mutations in compound heterozygous mice affect processes from cell polarity (Cobl/Vangl2, Dvl1/Dvl2, Fzd3/Fzd6, Vangl2/Celsr1, Vangl2/Scrb, Vangl2/Ptk7), to actin cytoskeletal regulation (Enah/Vasp, Enah/Pfn1), to cell–cell adhesion (Gja1/Gja5, Itga1/Itga6), intracellular protein transport (Snx1/Snx2), intracellular signaling (Jnk1/Jnk2, Prkaca/Prkacb), or transcription regulation (Msx1/Msx2, Rara/Rarg). Genetic interactions that increase NTD risk certainly do occur in mice and can be expected to occur in humans as well [Ross, 2010].

The several hundred genes that have been associated with NTDs in mouse models are beginning to provide insights into molecular networks that are critically important for neurulation and may become clinically useful [Ross, in press]. For example, just as mutation in Vangl2 renders the looptail mouse prone to NTD, polymorphisms in human Vangl1 [Kibar et al., 2007] and Vangl2 [Lei et al., 2010] have recently been implicated in NTD patients. However, evidence that these single nucleotide polymorphisms (SNPs) can occur in unaffected individuals as well suggests that the picture is too incomplete at present to permit using these SNPs for risk assessment. The context of the SNP or genetic background and environmental factors must be taken into consideration. Toward establishing this capability, pathway relationships associated with NTDs are emerging that include PCP signaling pathways encompassing Wnt and SHH pathways, chromatin remodeling and DNA methylation pathways, cell-cycle regulation, apoptosis, and cytoskeletal regulation, as well as transcription factors [Harris and Juriloff, 2007; Harris, 2009; Ross, 2010]. The complexity of the genetic underpinnings of NTD indicates that meeting the challenge for determining individual risk – and the optimal preventative therapy – will require evaluation of multiple genes (in signaling, metabolic, and transcriptional pathways) in a single person to detect compounding effects of gene polymorphisms that alone might not be significant [Greene et al., 2009; Ross, 2010]. Advanced technologies for high-throughput genomic DNA sequencing and analysis, and detection of the epigenome and perhaps the microbiome, as well as untargeted metabolomic screening, will all play a role in the clinical evaluation of individual patient assessments of NTD risk and prevention.

Gene–Environment Interactions Influencing Neural Tube Defects

Environmental influences contributing to NTDs have long been recognized [Janerich and Piper, 1978; Zhu et al., 2009]. The variables that have been implicated as risk factors for nonsyndromic forms of spina bifida are listed in Table 22-1. Risk factors include maternal diabetes, in which both hyperglycemia and associated hyperinsulinemia increase NTD risk [Eriksson et al., 2003]. Maternal obesity and pre-pregnancy weight gain are also associated with increased NTD risk [Werler et al., 1996; Hendricks et al., 2001]. Moreover, maternal periconceptional elevations in simple sugars that raise the glycemic index have been associated with increased NTD risk, even among nondiabetic women [Shaw et al., 2003]. In animal models, exposing rat embryos to a hyperglycemic environment induces dysmorphisms, accompanied by increases in biomarkers of oxidative stress and inositol depletion [Wentzel et al., 2001].

Table 22-1 Risk Factors for Spina Bifida

Gene–Diet Interactions in Neural Tube Defects: Role of Metabolism of Folic Acid and Other Nutrients

The important role of the folate pathway in the pathogenesis of NTDs has steadily gained acceptance since the landmark clinical investigations of Smithells and colleagues in the early 1980s prompted a large randomized trial of prenatal FA supplementation for women with previous NTD-affected pregnancies [Smithells et al., 1976; Smithells et al., 1983; MRC, 1991; Czeizel and Dudas, 1992]. Studies in the United Kingdom, later corroborated in Eastern Europe and elsewhere, indicated that the prevalence of NTDs could be reduced by 70 percent or more by prenatal supplementation with FA, even in the absence of maternal folate deficiency [MRC, 1991; Czeizel and Dudas, 1992]. However, some populations demonstrated only a small or no significant reduction in NTD rates with prenatal FA supplementation [Shaw et al., 1995], suggesting that differences in genetic background, diet, or other environmental exposures could impact the efficacy of folate supplementation. Inadequate intake of natural folate before and during early pregnancy is associated with a 2- to 8-fold increased risk of MMC and anencephaly, as indicated by several series of case-controlled, randomized clinical trials and community-based interventions [Wald, 1993]. Moreover, the risk of having a child affected by an NTD is indirectly related to both maternal folate intake and maternal folate intracellular level [Daly et al., 1995; Wald et al., 2001; Moore et al., 2003].

The mechanism underlying the association between NTDs and folate has not been established [Blom et al., 2006]. In view of the extraordinary variety of genes and molecular pathways contributing to NTD formation, perhaps the success of FA supplementation is due to the many biological functions to which folate metabolism contributes [Beaudin and Stover, 2009; Ross, 2010]. The exchange of single carbon units takes place in part in mitochondria, where serine cleavage generates glycine and formate, and glycine cleavage enzymes also generate formate that is transported to the cytoplasm. In the cytoplasm, formate enters the folate pathway for one-carbon metabolism (OCM) as formyl and then methenyl and methylene groups, added to or donated by dihydrofolates and tetrahydrofolates [Beaudin and Stover, 2007, 2009]. In the cytoplasm, the FA metabolic pathway is needed for physiological processes ranging from nucleotide biosynthesis, crucial for proliferation, to generation of pterin coactors impacting biochemical reactions, and generation of the body’s principal methyl donor, S-adenosyl methionine (SAM or Ado-met), which participates in methylation of DNA and histones modulating gene expression, and methylation of proteins and lipids [Blom et al., 2006; Beaudin and Stover, 2007; Miller, 2008].

Recent studies have demonstrated that NTDs can occur in the offspring of pregnant women who had maternal autoantibodies against folate receptors [Rothenberg et al., 2004]. These autoantibodies blocked the binding of ^3HFA to folate receptors on placental membranes, presumably resulting in reduced levels of CNS folate that contributed to development of the NTD. This result was not replicated in a population study of a large Irish cohort [Molloy et al., 2009]. However, that research was weakened by the fact that a number of serum samples were taken from women up to 10 years after pregnancy. Another study, which was confined to measuring folate receptor antibodies in samples drawn during midgestation, found a significant association between antibody levels, the folate transport-blocking capability of the antibody, and NTD occurrence [Cabrera et al., 2008]. This may lead in future to clinical testing during first pregnancies that will be used to adjust the recommended level of FA supplementation in subsequent pregnancies.

Teratogens

RA is a well-known teratogen when administered to nonhuman embryos, in which one of its many effects is to induce NTDs, including spina bifida, exencephaly, and anencephaly in several different species [Shenefelt, 1972; Tibbles and Wiley, 1988; Yasuda et al., 1989]. Exposure to excess RA has also been implicated in human embryopathy [Lammer et al., 1985; Rosa et al., 1986]. In contrast, inactivation of RA-synthesizing enzyme genes or receptor genes in mice leads to significantly increased rates of NTDs. High doses of RA (vitamin A) or retinoids in medications such as isotretinoin (Accutane) are to be avoided during pregnancy, although vitamin A deficiency can also increase NTD risk [Azais-Braesco and Pascal, 2000].

Most, if not all, antiepileptic drugs (AEDs) are known teratogens [Ornoy, 2006]. Different AEDs, however, are associated with different constellations of malformations. An increased risk of MMC is associated with in utero exposure to valproic acid or carbamazepine alone, or in combination with other AEDs [Lammer et al., 1987; Dansky and Finnell, 1991; Matalon et al., 2002]. In infants exposed to valproic acid or carbamazepine, the risk of MMC can be as high as 1–2 percent [Koren and Kennedy, 1999; Zhu et al., 2009]. Women who use these drugs for indications other than epilepsy (e.g., bipolar disease, migraine, chronic pain) also are at increased risk of having a child with MMC if they become pregnant while taking these drugs. The mechanisms by which valproic acid and carbamazepine increase the risk of NTD have not been established, but there is general consensus that genetic predisposition to its teratogenic effect is required for valproate to promote NTDs. Among the pathogenic mechanisms associated with valproic acid are increased homologous recombination that induces mutations in the genome, generation of reactive oxygen species, and epigenetic reprogramming through direct inhibition of histone deacetylases by valproate (reviewed in Ross, 2010). Folate administration does not appear to protect against the effects of valproic acid or carbamazepine on neural tube closure.

Nomenclature

The historical nomenclature of NTDs is imprecise. Many terms are based upon older descriptive terms that were derived with limited knowledge of the underlying embryologic defect. Broadly speaking, it is useful to separate those anomalies that arise from an early failure of neural tube formation, and those that arise from defects in subsequent developmental steps. MMC refers to the commonest form of spina bifida, in which the specific developmental defect is the presence of a flat neural placode, the unfolded derivative of the neural plate, elevated above a sac containing cerebrospinal fluid and continuous with the skin. Rather than representing a simple failure of fusion of the edges of the neural tube, the typical gross appearance of the neural placode of an MMC is most consistent with failure of folding of the neural plate at the median and dorsolateral hinge points.

Neural tube closure is usually required for subsequent developmental steps, such as the formation of mesodermal structures (e.g., dura, posterior spinal elements, and muscle). If the neural tube does not fold or close, the subsequent steps of mesodermal formation and ectodermal fusion at that segmental level do not occur. Milder NTDs will often result in near-normal formation of mesodermal structures and closure of the overlying skin. This basic difference, the presence or absence of skin, has led to the designation of spina bifida into open forms, spina bifida aperta, or closed forms, spina bifida occulta. Spina bifida occulta is a confusing term, since it can refer to either a broader group of anomalies that have normal skin overlying the spinal defect, or a specific anomaly that indicates a lack of fusion of the spinous processes in the lumbar area and has limited clinical significance (see below). Fortunately, both of these terms are not commonly used in the current era.

Embryologic Classification of Neural Tube Defects

One of the first visible steps in neurulation is the thickening of the ectoderm into the neural plate. In mammalian embryos, this is rapidly followed by a change in shape of the columnar epithelium in specific areas of the neural plate, resulting in formation of the neural tube. Once the edges of the neural groove fuse, the neural tube separates from the overlying ectoderm in a process known as dysjunction. This allows para-axial mesoderm to fill in the space between the neural tube and the ectoderm. This simplified series of events can be used to divide the commonly encountered NTDs on the basis of the embryologic defect:

anomalies due to defects of neural folding and formation – myelomeningocele and anencephaly

anomalies due to disordered postneurulation development – encephaloceles

anomalies due to incomplete dysjunction – dermal sinus and associated dermoid and epidermoid tumors

anomalies due to premature dysjunction – spinal cord lipomas

anomalies due to disorders of gastrulation – split cord malformation, neurenteric cysts

anomalies due to disordered secondary neurulation – thickened filum terminale, myelocystocele

anomalies due to failure of caudal neuraxial development – sacral agenesis.

Antenatal Diagnosis

Maternal serum α-fetoprotein (AFP) determination and ultrasound examination are used to identify fetuses that have or are likely to have spina bifida or anencephaly [Drugan et al., 2001]. Positive findings from either of these two screening tests can be monitored with amniocentesis and detailed sonography [Kooper et al., 2007]. AFP is a component of fetal cerebrospinal fluid, and it may leak into the amniotic fluid from the open neural tube. Elevated amniotic AFP concentrations correlate with open NTDs, while closed lesions usually do not lead to increased AFP concentration. Detection of NTDs correlates with the magnitude of increase in the amniotic fluid AFP level; NTDs are associated in a minority of pregnancies with mildly elevated AFP levels, in a majority of those with moderately elevated levels, and overwhelmingly in those with very elevated AFP levels [Canick et al., 2003]. An AFP level of two times the normal median or higher is found in approximately 2 percent of unaffected pregnancies, 80 percent of open spina bifida pregnancies, and 95 percent of anencephalic pregnancies.

Elevations in amniotic fluid acetylcholinesterase, produced by the fetal nervous system, correlate with maternal serum or amniotic fluid AFP, and can provide additional specificity and sensitivity in this screening approach. When acetylcholinesterase is elevated, more than 95 percent of patients with moderately or very elevated AFP levels had evidence of an NTD. Thus, it can differentiate between open ventral wall defects (i.e., gastroschisis and omphalocele), which will have normal acetylcholinesterase, and open NTDs, which will have elevated acetylcholinesterase.

Sonography also can be used to differentiate between ventral wall defects and NTDs [Lennon and Gray, 1999], and to identify additional structural malformations that are characteristic of fetuses with chromosomal abnormalities [Sepulveda et al., 2004]. When a diagnosis of spina bifida is confirmed, ultrasound examination is used to assess spontaneous leg and foot motion and to screen for leg and spine deformities, a Chiari II malformation, and other physical defects. Ultrasonography can detect or confirm the extent of the NTDs [Watson et al., 1991] and has had an enormous impact on the number of liveborn infants among populations in which termination of pregnancy is accepted [Zlotogora et al., 2002]. It is 60 percent accurate in low-risk pregnancies, which is equivalent to the accuracy of serum AFP screening (64 percent), 89 percent accurate in high-risk pregnancies, and 100 percent accurate for women referred for confirmation of a suspected spina bifida by another ultrasonographer [Chan et al., 1993, 1995]. The data indicate that neither sonography nor AFP screening alone provides sufficient sensitivity or specificity, but that, when these studies are used together, the predictive value is much higher [Chan et al., 1995].

Prenatal magnetic resonance imaging (MRI), with ultra-fast T2-weighted sequences, also can be used to characterize the Chiari II and other malformations (Figure 22-6). Such prenatal imaging studies might help to predict neurologic deficit [Cochrane et al., 1996] and ambulatory potential [Biggio et al., 2001]. Most fetuses with spina bifida that are not electively terminated receive no treatment until after birth. Several studies have investigated whether method of delivery influences the outcome for infants with the disorder. A study based on a review of this work concluded that, in general, conclusive evidence is lacking that, relative to vaginal delivery, cesarean section improves the outcome in children with spina bifida [Anteby and Yagel, 2003]. Cesarean section, however, might be justified for large lesions, to reduce the risk of trauma, and is done after in utero treatment of spina bifida because the forces of labor are likely to produce a wound dehiscence.

Fig. 22-6 A T2-weighted magnetic resonance image of a 22-week-old fetus with an open neural tube defect and cerebellar tonsils descended into the upper cervical canal.

These features are consistent with a Chiari II malformation.

Additionally, the fetal karyotype can be examined to rule out chromosomal anomalies. Cytogenetic analysis is justified in the setting of prenatally detected spina bifida based on the prevalence of chromosome abnormalities in 17 percent of fetuses (trisomy 18, trisomy 13, triploidy, and translocation).

Prevention

NTDs occur during early pregnancy, often before a woman knows she is pregnant; 50–70 percent of these defects can be prevented if a woman consumes sufficient FA daily before conception and throughout the first trimester of her pregnancy. Two landmark studies in the early 1990s found that administration of FA had a profound effect on the risk of NTDs [MRC, 1991; Czeizel and Dudas, 1992]. In 1992, to reduce the number of cases of MMC, the U.S. Public Health Service recommended that all women capable of becoming pregnant consume 400 μg of FA daily. Because approximately 50 percent of pregnancies in the United States are unplanned and the neural tube develops before most women know they are pregnant, it was also recommended that women consume this amount of FA routinely. Three approaches to increase FA consumption were cited:

Mandatory fortification of cereal grain products was legislated in the U.S. in 1996 and fully implemented as of January 1998. The estimated number of NTD-affected pregnancies in the U.S. declined from 4000 in the period 1995–1996 to 3000 in 1999–2000 [CDC, 2004; Mills and Signore, 2004]. This analysis controlled for declines in live births with NTDs as a result of recently implemented screening programs. Since the overall reduction of 25–30 percent in NTD rates in the U.S. since fortification was put into effect is below the expected 70 percent reduction, the debate continues regarding whether the fortification standard should be raised [Pitkin, 2007; Smith et al., 2008; Oakley, 2010].

Regardless of whether the optimal regimen for prenatal FA supplementation has been determined, there is still a significant proportion of pregnancies at risk for NTD for which FA will not effectively prevent NTD. Evidence in mouse models indicates that certain genetic backgrounds may include partial block in folate metabolism that can be circumvented by using alternative supplements. For example, the NTDs in the mouse mutant, Axial defects (Axd), are resistant to FA supplementation, but methionine – which will contribute to the methylation arm of the folate metabolic pathway – can reduce the occurrence of neurulation failure [Essien and Wannberg, 1993]. Another mouse mutant, curly tail (ct), which is prone to spina bifida that is not rescued by FA, can be protected against NTD by supplementation with myo-inositol [Greene and Copp, 1997]. Currently, clinical trials of prenatal inositol supplementation for the prevention of NTD are under way in the U.K. It appears likely that additional options for prevention of NTD will be available in the foreseeable future.

Clinical Features

The mortality rate for MMC is approximately 50 percent in the absence of therapy. Early surgery for closure of the lumbosacral defect is required to prevent meningitis. Later causes of death include hydrocephalus and renal failure. The renal complications are induced by chronic urinary tract infections, abnormal urodynamic function, and genitourinary tract abnormalities, such as progressive hydronephrosis. Patients at particular risk are the subgroup with high pressure within the bladder [Snodgrass and Adams, 2004].

MMCs may be situated at any longitudinal level of the neuroaxis. The location and extent of the defect determine the nature and degree of neurologic impairment; rating scales have been developed in an attempt to standardize the evaluation of affected children [Oi and Matsumoto, 1992]. Lumbosacral involvement is most common. Thoracic defects are the most complex and frequently are associated with serious complications. Cervical cord involvement is different from MMC of the lower spine and can be differentiated into two types:

Cervical lesions are clearly protuberant, covered at the base by full-thickness skin, and covered on the dome by a thick epithelium. Neural tissue is not superficially exposed but usually is tethered to adjacent dural or intrasaccular tissues. These differences likely are responsible for the more favorable outcome in cervical lesions [Meyer-Heim et al., 2003]. Varying degrees of paresis of the legs, usually profound, and sphincter dysfunction are the major clinical manifestations. Congenital dislocation of the hips or deformities of the feet, such as clubbing, may also occur. Severe sensory loss and accompanying trophic ulcers may complicate the condition. Occasionally, only sphincter disturbances are present. Radiographs reveal the primary defect of the vertebral arch.

Hydrocephalus is present in about 70–85 percent of patients with MMC, and occurs most frequently when the lesion is situated in the thoracolumbar area [Rintoul et al., 2002], which is the case in 90 percent of patients. Most persons with spina bifida have normal intelligence, but specific cognitive disabilities and language difficulties are common and can adversely affect educational and occupational achievements and the ability to live independently [Northrup and Volcik, 2000; Vachha and Adams, 2003].