The Amniotic Fluid As a Source of Neural Stem Cells in the Setting of Experimental Neural Tube Defects

Abstract

We sought to determine whether neural stem cells (NSCs) can be isolated from the amniotic fluid in the setting of neural tube defects (NTDs), as a prerequisite for eventual autologous perinatal therapies. Pregnant Sprague-Dawley dams (n=62) were divided into experimental (n=42) and control (n=20) groups, depending on prenatal exposure to retinoic acid for the induction of fetal NTDs. Animals were killed before term for analysis (n=685 fetuses). Amniotic fluid samples from both groups underwent epigenetic selection for NSCs, followed by exposure to neural differentiation media. Representative cell samples underwent multiple morphological and phenotypical analyses at different time points. No control fetus (n=267) had any structural abnormality, whereas at least one type of NTD developed in 52% (217/418) of the experimental fetuses (namely, isolated spina bifida, n=144; isolated exencephaly, n=24; or a combination of the two, n=49). Only amniotic samples from fetuses with a NTD yielded cells with typical neural progenitor morphology and robust expression of both Nestin and Sox-2, primary markers of NSCs. These cells responded to differentiation media by displaying typical morphological changes, along with expression of beta-tubulin III, glial fibrillary acidic protein, and/or O4, markers for immature neurons, astrocytes, and oligodendrocytes, respectively. This was concurrent with downregulation of Nestin and Sox-2. We conclude that the amniotic fluid can harbor disease-specific stem cells, for example, NSCs in the setting of experimental NTDs. The amniotic fluid may be a practical source of autologous NSCs applicable to novel forms of therapies for spina bifida.

Introduction

N eural tube defects (NTDs) stem from the failure of the neural tube to close by the fourth week of embryonic development. These defects are commonly classified as open or closed based on the presence or absence of exposed neural tissue, and may involve any portion of the brain and/or spinal cord. Spina bifida is the most common NTD compatible with life. It is characterized by a midline vertebral defect, typically in the dorsal portion of the lumbosacral vertebrae. Depending on whether only the meningeal sac, or the meningeal sac and the spinal cord protrude through the defect, it is referred to as meningocele or myelomeningocele, respectively. Isolated NTDs are multifactorial in inheritance in the vast majority of cases [1]. At the same time, folic acid deficiency is known to be a causative factor [2]. Folic acid supplementation during the first trimester of gestation can reduce the risk of NTDs by as much as 50%–70% [3,4]. However, despite mandatory folate supplementation in all enriched cereal grain products, the incidence of spina bifida in the United States remains fairly high, at 3.4 per 10,000 live births [5].

Spina bifida leads to significant spinal cord damage at and below the lesion and is often accompanied by hydrocephalus. Common manifestations include paraplegia, sensory loss, bowel, bladder, and sexual dysfunction, and secondary musculoskeletal deformities [6,7]. It is among the most morbid survivable congenital anomalies; so much so, that it has prompted the first ever prospective randomized control trial of prenatal surgical repair of a nonlife threatening anomaly, widely known as the Management of Myelomeningocele Study (MOMS) [8]. This sentinel undertaking has shown that, compared with the conventional postnatal coverage of the lesion, prenatal repair does offer some benefits, but to a limited extent and only in a select subset of fetuses, not without notable maternal risks [8].

Using a large animal model, we have previously introduced the notion of therapeutic NSC delivery to the spinal cord in fetuses with spina bifida, as a potential means of promoting actual neural tissue repair, rather than solely preventing further damage by simple coverage, as originally aimed in the MOMS [9]. In that initial study, donor NSCs were necessarily of a xenologous origin. Certainly, for the eventual clinical translation of this therapeutic concept to be rendered viable, an autologous source of NSCs would be preferred. Neural cells have long been known to be present in the amniotic fluid in the setting of NTDs [10 –13]. In this study, we aimed at testing our hypothesis that a population of NSCs is also present therein in this setting. This would be a prerequisite for the development of minimally invasive autologous perinatal NSC-based strategies for the treatment of these anomalies.

Materials and Methods

This study was approved by the Children's Hospital Boston's Institutional Animal Care and Use Committee under protocol # 09-05-1353.

Sixty-two time-dated pregnant Sprague-Dawley dams (Charles River Laboratories, Inc., Wilmington, MA) were housed under a standard dark/light cycle and fed a normal diet ad libitum. Animals were divided into 2 groups. One group, control (n=20), underwent no manipulations. The other group, experimental (n=42), was exposed to retinoic acid at a specific time in gestation for the creation of fetal NTDs.

Fetal NTD model

Fetal NTDs were induced as previously described [14]. Briefly, after brief exposure to isoflurane (Abbot Laboratories, North Chicago, IL), chamber inhaled at 2%–4% in 100% oxygen, the dams received 60 mg/kg of all-trans retinoic acid (Sigma-Aldrich Chemical, St. Louis, MO) dissolved in olive oil at 10 mg/mL, through gavage, on embryonic day ten (E10), between 6:00 and 8:00 p.m.

Amniotic fluid procurement

In both groups, amniotic fluid samples from all fetuses (n=685) were obtained between E19 and E21 (term=E22). To that end, the dams were euthanized with chamber inhaled CO₂. A large midline laparotomy was made and the bicornuate uterus eviscerated. Both horns were evaluated to determine the total number of fetuses. A separate set of 30G needle (Becton Dickinson, Franklin Lakes, NJ) on a 1-mL syringe (Becton Dickinson) was introduced into each amniotic cavity upon the ventral aspect of the fetus, through which amniotic fluid was procured. After procurement, the uterus and gestational membranes were incised and the fetus examined for the presence or absence of a NTD, namely, an isolated spina bifida, an isolated exencephaly, or a combination of the two.

Amniotic fluid processing

Amniotic fluid samples were divided into 2 groups, depending on the presence or absence of a NTD. Samples from all fetuses were processed. Whenever necessary, samples from more than one fetus were combined to create an aliquot of at least 1 mL before processing, which consisted of epigenetic selection for the presence of NSCs, followed by exposure to neural differentiation media. To that end, aliquots were centrifuged at 3,000 rpm for 3 min, and the pellets removed and resuspended in a commercially available rat NSC expansion medium (Millipore, Billerica, MA), supplemented with 20 ng/mL fibroblast growth factor (FGF)-2 (Promega, Madison, WI); 5 μg/mL amphotericin (X-Gen Pharmaceuticals, Big Flats, NY); the Dulbecco's modified Eagle's medium/F12, 1× B27 chemically defined medium, 2 mM L-glutamine; 1× penicillin; and 1× streptomycin (all from Millipore). The cell suspensions were plated in poly-L-ornithine/laminin-coated 6-well plates (BD Biosciences, Bedford, MA) and placed in a 95% humidified, 5% CO₂ incubator at 37°C. The culture medium was changed every other day. Cells were maintained in culture for 6–8 weeks with 2–3 passages using an accutase solution (Millipore), after which time, a subset of cells derived from each aliquot underwent morphological and phenotypical analyses as specified below. The remaining cells were then exposed to the neural differentiation medium, which consisted of the rat NSC expansion medium described above (Millipore) without FGF-2 supplementation, also changed every other day. Cells were maintained in the differentiation medium for 5–7 days before morphological and phenotypical analyses.

Morphological and phenotypical analyses

Throughout processing, cells were inspected for morphological changes with an inverted microscope (Carl Zeiss, Jena, Germany). Upon completion of exposure to the NSC expansion and differentiation media, as described above, representative cell samples underwent immunostaining for the undifferentiated neural cell markers, Nestin and Sox-2, as well as markers for differentiation into the 3 major neural cell lineages, namely, the immature neuronal marker, β-III tubulin, the astroglial marker, glial fibrillary acidic protein (GFAP), and the oligodendrocyte maker, O4, as previously described [15]. Briefly, cells were fixed with 4% paraformaldehyde at room temperature for 10 min, and then washed 3 times with phosphate-buffered saline. After nonspecific antibody reactions were blocked with 10% goat serum (Sigma Chemical, St Louis, MO) and 0.1% triton X (Sigma) for 30 min, cells were incubated with primary antibodies directed against Nestin (5 μg/mL; Beckton Dickinson and Company); Sox-2 (1:50; Santa Cruz Biotechnology, Santa Cruz, CA); GFAP (1:80; Sigma); β-III tubulin (25 μg/mL; R&D Systems, Minneapolis, MN); and O4 (20 μg/mL; Millipore) overnight at 4°C. After 3 washes, cells were incubated with Dylight 488- or 549-conjugated goat secondary antibodies (Jackson Immunoresearch, West Grove, PA) for 1 h at room temperature. Adult rat hippocampal NSCs (Millipore) were used as positive controls. Amniotic mesenchymal stem cells obtained from the pregnant dams that were not exposed to retinoic acid were used as negative controls.

Results

None of the fetuses in the control up (n=267) had a structural abnormality, whereas at least one NTD was present in 52% (217/418) of the fetuses in the group exposed to retinoic acid. Such NTDs were distributed as follows: isolated spina bifida, n=144; isolated exencephaly, n=24; and a combination of the two, n=49.

During exposure to the NSC medium, cells derived from amniotic samples procured from fetuses with a NTD underwent cell division and proliferation in culture, exhibiting clusters commonly referred to as neurospheres, the morphological hallmark of neural progenitors (Fig. 1) [16]. Such cells were immunoreactive for the neural progenitor markers, Nestin and Sox-2 (Fig. 2). Upon exposure to the neural differentiation media, these cells became increasingly heterogeneous morphologically, in arrangements indistinguishable from what would be expected of differentiated neural cells of diverse lineages, many displaying neurites (Fig. 3). For this proposed initial nature of qualitative assessment, further differentiation between dendrites and axons was deemed unnecessary. Assorted subpopulations of these cells stained positively for any of the 3 major neural cell lineages tested; namely, β-III tubulin, GFAP, and O4 (Fig. 4), concurrently with downregulation of Nestin and Sox-2. Although we did not measure the exact proportion of neurons, astrocytes, or oligodendrocytes, no particular cell type proved dominant, with all appearing fairly evenly distributed.

FIG. 1.

Phase-contrast microscopy of amniotic fluid-derived neural stem cells in culture forming neurospheres typical of undifferentiated neural progenitors. (Magnification 200×).

FIG. 2.

Undifferentiated amniotic fluid-derived neural stem cells (B, D) and positive controls isolated from rat hipoccampus (A, C) immunostained for Nestin (green), Sox-2 (pink), and DAPI (blue). (Magnification 200×). Color images available online at www.liebertpub.com/scd

FIG. 3.

Phase-contrast microscopy of amniotic fluid-derived neural stem cells in neural differentiation medium exhibiting heterogeneous morphological changes expected of neural cells of diverse lineages, including neurites. (Magnification 200×).

FIG. 4.

Amniotic fluid-derived neural stem cells (B, D, F) and positive controls isolated from rat hipoccampus (A, C, E) undergoing differentiation in culture. Cells were immunostained for beta-tubulin III ( green – A, B), glial fibrillary acidic protein (GFAP; green – C, D), O4 ( green – E, F), Nestin (pink), and DAPI (blue). (Magnification 200×). Color images available online at www.liebertpub.com/scd

Cells derived from amniotic samples procured in the control group did not proliferate, did not form neurospheres, nor exhibit any of the morphological changes observed during differentiation of the cells from the group with NTDs, as described above. These cells were negative for all of the neural markers tested.

Discussion

Spina bifida leads to spinal cord damage by both primary and secondary mechanisms. The former stems from the abnormal spinal cord development associated with the incomplete closure of the neural tube, while the latter is a result of spinal cord exposure to both the amniotic fluid (chemical insult) and local trauma (mechanical insult) within the uterus. Several studies point to the secondary component as being the most relevant clinically. Analyses performed in stillborns have shown that embryos, or very young fetuses, with spina bifida actually have little damage to the neural tube or spinal cord [17 –20]. Imaging studies have shown that fetal leg movement tends to be normal in fetuses with spina bifida early on in gestation, only to worsen as pregnancy progresses [21,22]. In addition, delivery of these fetuses by cesarean section is associated with improved neurological outcome, when compared with vaginal birth [23,24]. Animal models have further underlined the significance of the secondary insult to the spinal cord in this disease, also by demonstrating neurological improvement after experimental prenatal closure of the defect [25 –27]. All these data have justified fetal repair clinically, as performed at the MOMS.

Following transplantation from an exogenous source, NSCs have been shown to mediate repair in a multitude of central nervous system injury or disease models, including in the spinal cord, often by invoking developmental mechanisms [15,28 –35]. In essentially all of these conditions, however, NSCs have been administered after birth. In our previous study, we introduced the notion of prenatal delivery of NSCs as a potential means of enhancing fetal surgical coverage of spina bifida, which is fundamentally a developmental problem [9]. Our fetal approach was predicated on the hypothesis that the regenerative/reparative impact of NSCs could be maximized if applied to the developing, rather than the more mature, spinal cord. Necessarily, however, our early work involved a well-characterized NSC clone xenologous to the host. At this time, any realistic hopes of clinical translation of this therapeutic concept would necessarily hinge on the consistent availability of a heterologous, or, preferably, an autologous prenatal source of NSCs.

While neural cells have long been known to exist in amniotic fluid in the presence of NTDs, and, in fact, often aid in the diagnosis of these anomalies, the existence of an undifferentiated, more primitive population of neural cells within the amniotic cavity had yet to be demonstrated [10 –13,36]. In this study, we have isolated a population of NSCs in the setting of experimental NTDs. Due to the fact that, in this rodent model, samples from more than one fetus with the same type of NTD, at times, had to be combined to generate an aliquot sizeable enough for processing, we could not quantify the likelihood of such an isolation being viable in each fetus. On the other hand, all of the samples from fetuses with NTDs with at least 1 mL yielded cells. Also, such quantitative data in rodents would arguably be of limited translational significance. Previous data on the abundant presence of neural cells in human amniotic fluid in the setting of fetal NTDs would suggest that a comparable isolation of NSCs may well be practical clinically [36]. Such a demonstration is one of our current endeavors.

We have previously established experimentally that the mesenchymal stem cells long known to be naturally occurring in the amniotic fluid can be used in tissue engineering strategies for the repair of a variety of congenital anomalies in the perinatal period, and that these cells can be manufactured within regulatory guidelines for clinical use [37 –45]. Translated clinically, from a simple amniocentesis, different grafts could be engineered, in parallel, to the remainder of gestation, so that an infant could benefit from having autologous, expanded tissue promptly available for surgical reconstruction at or before birth. In a similar fashion, the present study suggests that autologous NSCs could be obtained by amniocentesis in a fetus with spina bifida and expanded ex vivo for therapeutic applications, either prenatally or postnatally. This is also, to our knowledge, the first demonstration of disease-specific stem cells in the amniotic fluid.

The inherent limitations of an introductory study such as this notwithstanding, our findings, taken together with the large body of data on the use of NSCs in other forms of spinal cord injury, support further development of this multifaceted, perinatal therapeutic approach, combining local NSC delivery with mechanical/surgical strategies aimed at inducing protective and/or regenerative processes within the spinal cord in spina bifida.

Footnotes

Acknowledgment

D.O.F. was supported by a Translational Research Program Award from Children's Hospital Boston. Y.D.T. and his group were supported by DoD-CIMIT, VARR&D, and HeadNorth Project funding.

Authors Disclosure Statement

No competing financial interests exist.

References

Main

, Mennuti

. 1986. Neural tube defects: issues in prenatal diagnosis and counselling. Obstet Gynecol, 67:1–16.

Botto

, Moore

, Khoury

, Erickson

. 1999. Neural-tube defects. N Engl J Med, 341:1509–1519.

1991. Prevention of neural tube defects: results of the Medical Research Council Vitamin Study. Lancet, 338:131–137.

Czeizel

, Dudas

. 1992. Prevention of the first occurrence of neural-tube defects by periconceptional vitamin supplementation. N Engl J Med, 327:1832–1835.

Boulet

, Yang

, Mai

, Kirby

, Collins

, Robbins

, Meyer

, Canfield

, Mulinare

. 2008. Trends in the postfortification prevalence of spina bifida and anencephaly in the United States. Birth Defects Res A Clin Mol Teratol, 82:527–532.

Wong

, Paulozzi

. 2001. Survival of infants with spina bifida: a population study, 1979–94. Paediatr Perinat Epidemiol, 15:374–378.

Mitchell

, Adzick

, Melchionne

, Pasquariello

, Sutton

, Whitehead

. 2004. Spina bifida. Lancet, 364:1885–1895.

Adzick

, Thom

, Spong

, Brock

3rd , Burrows

, Johnson

, Howell

, Farrell

, Dabrowiak

et al. 2011. A randomized trial of prenatal versus postnatal repair of myelomeningocele. N Engl J Med, 364:993–1004.

Fauza

, Jennings

, Teng

, Snyder

. 2008. Neural stem cell delivery to the spinal cord in an ovine model of fetal surgery for spina bifida. Surgery, 144:367–373.

10.

Gosden

, Brock

, Eason

. 1977. The origin of the rapidly adhering cells found in amniotic fluids from foetuses with neural tube defects. Clin Genet, 12:193–2U1.

11.

Aula

, von Koskull

, Teramo

, Karjalainen

, Virtanen

, Lehto

, Dahl

. 1980. Glial origin of rapidly adhering amniotic fluid cells. Br Med J, 281:1456–1457.

12.

Greenebaum

, Mansukhani

, Heller

, Timor-Tristsch

. 1997. Open neural tube defects: immunocytochemical demonstration of neuroepithelial cells in amniotic fluid. Diagn Cytopathol, 16:143–144.

13.

Mendonca

, Gutierrez

, Peres

. 2005. Brain tissue fragments in the amniotic fluid of rats with neural tube defect. Pathology, 37:152–156.

14.

Danzer

, Schwarz

, Wehrli

, Radu

, Adzick

, Flake

. 2005. Retinoic acid induced myelomeningocele in fetal rats: characterization by histopathological analysis and magnetic resonance imaging. Exp Neurol, 194:467–475.

15.

, Jones

, Snyder

, Tuszynski

. 2003. Neural stem cells constitutively secrete neurotrophic factors and promote extensive host axonal growth after spinal cord injury. Exp Neurol, 181:115–129.

16.

Reynolds

, Weiss

. 1992. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science, 255:1707–1710.

17.

Patten

. 1953. Embryological stages in the establishing of myeloschisis with spina bifida. Am J Anat, 93:365–395.

18.

Osaka

, Tanimura

, Hirayama

, Matsumoto

. 1978. Myelomeningocele before birth. J Neurosurg, 49:711–724.

19.

Hutchins

, Meuli

, Meuli-Simmen

, Jordan

, Heffez

, Blakemore

. 1996. Acquired spinal cord injury in human fetuses with myelomeningocele. Pediatr Pathol Lab Med, 16:701–712.

20.

Meuli

, Meuli-Simmen

, Hutchins

, Seller

, Harrison

, Adzick

. 1997. The spinal cord lesion in human fetuses with myelomeningocele: implications for fetal surgery. J Pediatr Surg, 32:448–452.

21.

Korenromp

, van Gool

, Bruinese

, Kriek

. 1986. Early fetal leg movements in myelomeningocele. Lancet, 1:917–918.

22.

Sival

, Begeer

, Staal-Schreinemachers

, Vos-Niel

, Beekhuis

, Prechtl

. 1997. Perinatal motor behaviour and neurological outcome in spina bifida aperta. Early Hum Dev, 50:27–37.

23.

Luthy

, Wardinsky

, Shurtleff

, Hollenbach

, Hickok

, Nyberg

, Benedetti

. 1991. Cesarean section before the onset of labor and subsequent motor function in infants with meningomyelocele diagnosed antenatally. N Engl J Med, 324:662–666.

24.

Shurtleff

, Luthy

, Nyberg

, Benedetti

, Mack

. 1994. Meningomyelocele: management in utero and post natum. Ciba Found Symp, 181:270–280discussion 280–276.

25.

Meuli

, Meuli-Simmen

, Yingling

, Hutchins

, Hoffman

, Harrison

, Adzick

. 1995. Creation of myelomeningocele in utero: a model of functional damage from spinal cord exposure in fetal sheep. J Pediatr Surg, 30:1028–1032discussion 1032–1023.

26.

Meuli

, Meuli-Simmen

, Hutchins

, Yingling

, Hoffman

, Harrison

, Adzick

. 1995. In utero surgery rescues neurological function at birth in sheep with spina bifida. Nat Med, 1:342–347.

27.

Meuli

, Meuli-Simmen

, Yingling

, Hutchins

, Timmel

, Harrison

, Adzick

. 1996. In utero repair of experimental myelomeningocele saves neurological function at birth. J Pediatr Surg, 31:397–402.

28.

Snyder

, Deitcher

, Walsh

, Arnold-Aldea

, Hartwieg

, Cepko

. 1992. Multipotent neural cell lines can engraft and participate in development of mouse cerebellum. Cell, 68:33–51.

29.

Snyder

, Wolfe

. 1996. Central nervous system cell transplantation: a novel therapy for storage diseases? Curr Opin Neurol, 9:126–136.

30.

Whittemore

, Snyder

. 1996. Physiological relevance and functional potential of central nervous system-derived cell lines. Mol Neurobiol, 12:13–38.

31.

Snyder

. 1998. Neural stem-like cells: developmental lessons with therapeutic potential. Neuroscientist, 4:408–425.

32.

Teng

, Lavik

, Qu

, Park

, Ourednik

, Zurakowski

, Langer

, Snyder

. 2002. Functional recovery following traumatic spinal cord injury mediated by a unique polymer scaffold seeded with neural stem cells. Proc Natl Acad Sci U S A, 99:3024–3029.

33.

Yip

, Aboody

, Burns

, Imitola

, Boockvar

, Allport

, Park

, Teng

, Lachyankar

et al. 2003. Neural stem cell biology may be well suited for improving brain tumor therapies. Cancer J, 9:189–204.

34.

Teng

, Choi

, Onario

, Zhu

, Desilets

, Lan

, Woodard

, Snyder

, Eichler

, Friedlander

. 2004. Minocycline inhibits contusion-triggered mitochondrial cytochrome c release and mitigates functional deficits after spinal cord injury. Proc Natl Acad Sci U S A, 101:3071–3076.

35.

Imitola

, Raddassi

, Park

, Mueller

, Nieto

, Teng

, Frenkel

, Li

, Sidman

et al. 2004. Directed migration of neural stem cells to sites of CNS injury by the stromal cell-derived factor 1alpha/CXC chemokine receptor 4 pathway. Proc Natl Acad Sci U S A, 101:18117–18122.

36.

Gosden

, Brock

. 1977. Morphology of rapidly adhering amniotic-fluid cells as an aid to the diagnosis of neural-tube defects. Lancet, 1:919–922.

37.

Fuchs

, Kaviani

, Oh

, LaVan

, Udagawa

, Jennings

, Wilson

, Fauza

. 2004. Diaphragmatic reconstruction with autologous tendon engineered from mesenchymal amniocytes. J Pediatr Surg, 39:834–838discussion 834–838.

38.

Kaviani

, Perry

, Dzakovic

, Jennings

, Ziegler

, Fauza

. 2001. The amniotic fluid as a source of cells for fetal tissue engineering. J Pediatr Surg, 36:1662–1665.

39.

Klein

, Turner

, Ahmed

, Steigman

, Zurakowski

, Fauza

. 2010. Chest wall repair with engineered fetal bone grafts: an efficacy analysis in an autologous leporine model. J Pediatr Surg, 45:1354–1360.

40.

Kunisaki

, Armant

, Kao

, Stevenson

, Kim

, Fauza

. 2007. Tissue engineering from human mesenchymal amniocytes: a prelude to clinical trials. J Pediatr Surg, 42:974–979discussion 979–980.

41.

Kunisaki

, Freedman

, Fauza

. 2006. Fetal tracheal reconstruction with cartilaginous grafts engineered from mesenchymal amniocytes. J Pediatr Surg, 41:675–682.

42.

Kunisaki

, Fuchs

, Kaviani

, Oh

, LaVan

, Vacanti

, Wilson

, Fauza

. 2006. Diaphragmatic repair through fetal tissue engineering: a comparison between mesenchymal amniocyte- and myoblast-based constructs. J Pediatr Surg, 41:34–39discussion 34–39.

43.

Steigman

, Ahmed

, Shanti

, Tuan

, Valim

, Fauza

. 2009. Sternal repair with bone grafts engineered from amniotic mesenchymal stem cells. J Pediatr Surg, 44:1120–1126discussion 1126.

44.

Steigman

, Armant

, Bayer-Zwirello

, Kao

, Silberstein

, Ritz

, Fauza

. 2008. Preclinical regulatory validation of a 3-stage amniotic mesenchymal stem cell manufacturing protocol. J Pediatr Surg, 43:1164–1169.

45.

Turner

, Fauza

. 2009. Fetal tissue engineering. Clin Perinatol, 36:473–488xii.