Genetic and Phenotypic Response of Neural Tube Defect

GENETIC AND PHENOTYPIC RESPONSE OF NEURAL TUBE DEFECT

MOUSE MUTANTS TO FOLIC ACID

GHUNWA A. NAKOUZI

Submitted in partial fulfillment of the requirements

For the degree of Doctor of Philosophy

Dissertation Adviser: Dr. Joseph H. Nadeau

Department of Genetics

CASE WESTERN RESERVE UNIVERSITY

August, 2009

CASE WESTERN RESREVER UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

GHUNWA A. NAKOUZI ______

candidate for the PhD degree*.

Shawn McCandless (signed)______(chair of the committee)

Joseph Nadeau ______

Mark Adams ______

Donald Jacobsen ______

Radhika Atit ______

June 30, 2009 (date) ______

*We also certify that written approval has been obtained for any proprietary material contained therein.

TABLE OF CONTENTS

TABLE OF CONTENTS ...... 1 LIST OF TABLES ...... 3 LIST OF FIGURES ...... 5 ACKNOWLEDGEMENTS AND DEDICATIONS ...... 6 Abstract ...... 8 CHAPTER 1 ...... 10 NTDs in Humans ...... 11 Clinical significance...... 11 Genetic and environmental risk factors...... 13 Folate-Homocysteine association...... 14 Prevention measures...... 16 Adverse effects of high levels FA...... 18 NTDs in Mice ...... 20 Mice as models for human NTDs...... 20 Supplementation studies and metabolic pathways...... 24 Previous response predictive profiles...... 29 Summary of mouse mutants selected for modeling human NTDs...... 30 Research Objectives ...... 35 CHAPTER 2 ...... 40 Abstract ...... 41 Introduction ...... 42 Materials and Methods ...... 45 Results ...... 50 Discussion ...... 58 CHAPTER 3 ...... 62 Abstract ...... 64 Introduction ...... 65 Materials and Methods ...... 69 Results ...... 76 Discussion ...... 84 CHAPTER 4 ...... 88 Abstract ...... 90

Introduction ...... 92 Materials and Methods ...... 94 Results ...... 102 Discussion ...... 117 CHAPTER 5 ...... 121 Summary ...... 122 Future Directions ...... 126 APPENDIX ...... 136 APPENDIX 1 ...... 137 APPENDIX 2 ...... 138 APPENDIX 3 ...... 139 APPENDIX 4 ...... 139 APPENDIX 5 ...... 139 APPENDIX 6 ...... 139 APPENDIX 7 ...... 140 APPENDIX 8 ...... 143 BIBLIOGRAPHY ...... 144

LIST OF TABLES

CHAPTER 1 Table 1.1 Mouse NTD responses to different nutritional supplementations or treatments ...... 28 Table 1.2 NTD mutants chosen for study ...... 34

CHAPTER 2 Table 2.1 Apob: effect of maternal FA supplementation on occurrence of resorptions 51 Table 2.2 Effect of maternal FA supplementation on incidence of exencephaly in Apob-/- embryos ...... 51 Table 2.3 Lp: effect of maternal FA supplementation on occurrence of resorptions .... 53 Table 2.4 Effect of maternal FA supplementation on incidence of craniorachischisis in Lp-/- embryos ...... 53 Table 2.5 Effect of maternal FA supplementation on incidence of looped tail in Lp+/- embryos ...... 53 Table 2.6 Genotype distribution of observed Apob embryos ...... 55 Table 2.7 Genotype distribution of observed Lp embryos ...... 55 Table 2.8 Gender specific distribution of exencephaly frequency and embryonic loss in Apob...... 56 Table 2.9 Gender specific distribution of craniorachischisis and looped tail frequency

and embryonic loss in Lp...... 57

CHAPTER 3 Table 3.1 Test and control mice used for experiments ...... 71 Table 3.2 Comparison of metabolite levels in each mutant to its corresponding control

(unpaired t-tests with Welch’s correction) ...... 82

CHAPTER 4 Table 4.1 Early lethality of FA-supplemented Lrp6-/- embryos ...... 104 Table 4.2 Effect of maternal FA supplementation on incidence of NTDs in Lrp6-/- embryos ...... 105 Table 4.3 Genotype distribution of observed embryos from Lrp6+/- x Lrp6 +/- intercrosses ...... 105 Table 4.4 Effect of maternal FA supplementation on occurrence of resorptions ...... 105 Table 4.5 Frequency of resorptions in Lrp6+/+ x Lrp6+/+ intercrosses maintained on 2 ppm vs. 10 ppm FA diet ...... 106 Table 4.6 Five pathways most significantly associated with diet-genotype interaction dataset ...... 110

LIST OF FIGURES

CHAPTER 1 Figure 1.1 The three types of spina bifida ...... 12 Figure 1.2 Folate-Homocysteine metabolic pathway ...... 16 Figure 1.3 Sequential change of typical neural fold morphology during the elevation process...... 20 Figure 1.4 Three sites of closure of the rostral neural tube ...... 22

CHAPTER 2 Figure 2.1 Supplementation study outline ...... 45

CHAPTER 3 Figure 3.1 Weaning and breeding timeline ...... 70 Figure 3.2 The two signaling pathways commonly associated with all six NTD mutants ...... 78 Figure 3.3 Pathways defining the genes specific to each FA-response group ...... 80

CHAPTER 4 Figure 4.1 Incidence of NTD affected pregnancies in FA supplemented mice ...... 104 Figure 4.2 Diet-genotype interaction plots for genes significantly associated with the Wnt/β-catenin signaling pathway ...... 109 Figure 4.3 The two networks most significantly associated with the diet-genotype interaction dataset ...... 111 Figure 4.4 Maternal FA supplementation effects on proliferation in the neural tube of E9.5 Lrp6 wild-type and mutant embryos ...... 115 Figure 4.5 Impact of FA on the canonical Wnt signaling pathway in vitro ...... 116

ACKNOWLEDGEMENTS AND DEDICATIONS

Many special people have impacted my journey through graduate school. First, I would

like to thank my mentor Dr. Joseph Nadeau for providing me with the opportunity to

pursue my PhD in his lab. Dr. Nadeau has taught me to ask important scientific questions,

to think and work independently, and to write effectively. I would also like to thank my

thesis committee members for their continuous guidance and useful critique of my work.

Members of the Nadeau lab, both former and current, have been a source of great support

and help throughout this process. I would especially like to thank my bay mate and

lifelong friend, Soha Yazbek, with whom I have shared the stress of PhD life as well as

its excitement and joy.

A special thanks to Ricky Chan, a wonderful graduate school friend and a tremendously

helpful person. I would also like to thank Lorie Rice for brightening my days with her

joyful personality and beautiful smile. I must also acknowledge the administrative help of

Malana Bey.

I would like to thank my dear brothers, Samer and Fady, for keeping me close to their hearts as they have been close to mine despite the physical distance that has separated us.

My longing for their company has motivated me throughout this long process to work harder and faster to be with them again.

I also could not have done this without my soul mate and husband, Samir Shaia. He was always by my side through the tough and stressful times as through the good and happy times. He held my hand and encouraged me every step of the way. I love you forever.

Most importantly, I am forever grateful to my precious parents, Akram and Josiane

Nakouzi. It is because of their dedication to me and to my education, and because of their tremendous love and support that I have obtained my PhD. They have always taught me the value of hard work and determination, and their teachings have paid off.

Genetic and Phenotypic Response of Neural Tube Defect Mouse Mutants to Folic Acid

Abstract

GHUNWA A. NAKOUZI

Supplementation with folic acid (FA) is a public health policy to reduce the risk of neural tube defects (NTDs). However, compliance with recommended levels is low, and ≈50% of women and their babies are resistant to the beneficial effect of FA. If NTD response to

FA could be predicted, compliance might improve and treatment could be targeted directly to responsive women. To test for maternal markers of fetal NTD risk and FA response, mouse models were used. Females of six single gene mouse mutants that are prone to NTD affected pregnancies were studied, among which three had an untested response to FA. We found that the NTD phenotype of these three mutants was resistant to maternal FA supplementation. Unexpectedly, these three mutants showed a substantial

FA-induced loss of homozygous and heterozygous pre-implantation embryos. With two responsive and four resistant NTD mutants, whole genome expression and metabolite profiles were then analyzed to identify maternal markers for occurrence of fetal NTDs and FA response. Canonical pathways related to retinoid X receptor (RXR) function were associated with all NTD mutants regardless of FA response. By contrast, no specific pathways distinguished NTD-responsive and -resistant mutants, suggesting that multiple mechanisms control response to FA. Finally, gene expression analysis verified with in vitro assays showed that FA supplementation affected both Wnt/β-catenin signaling and

cell cycle regulation in the Lrp6 mutant. Overall, FA-induced loss of homozygous as well

as otherwise healthy and viable heterozygous embryos highlight the need to determine the cause and stage of this early embryonic lethality. Moreover, the association of RXR related pathways with all six NTD mutants suggests a shared mechanism for NTD

pathogenesis in which the role of RXR should be investigated. Our studies of the Lrp6

loss-of-function mutant, together with published work of the Lrp6 gain-of-function

mutant, suggest that FA supplementation attenuates canonical Wnt signaling to normal

levels in the latter mutant thereby restoring normal neural tube development, and to sub-

optimal levels in the former mutant thereby compromising embryonic viability. The

mechanisms by which FA affects Wnt signaling and the functional dependence of neural

tube development on the level of Wnt signaling remain to be determined.

CHAPTER 1

Introduction and Research Objectives

NTDs in Humans

Clinical significance. Neural tube defects (NTDs) are serious and common birth

defects with an estimated worldwide incidence of 1 per 1,000 live births (Northrup and

Volcik 2000). Neural tube closure is an early developmental process that gives rise to the

central nervous system, including the spinal cord and brain (Detrait et al. 2005).

Normally, closure of the neural tube is complete 28 days post-fertilization. Failure of

neural tube closure leads to different clinical types of NTDs depending on the site of

closure failure (Botto et al. 1999).

There are two main types of NTDs - anencephaly and spina bifida. Infants with

anencephaly, the most severe form of NTDs, are miscarried or stillborn due to failed

closure of the rostral (upper) portion of the neural tube. Failure is characterized by a total

or partial absence of the cranial vault and cerebral hemispheres (Detrait et al. 2005).

Sometimes an encephalocele containing tissue and cerebrospinal fluid (CSF) protrudes

through the skull (Botto et al. 1999). A more severe form of cranial NTD is

craniorachischisis, which is characterized by anencephaly accompanied by a bony defect

of the spine and exposure of neural tissue (Botto et al. 1999). In contrast, spina bifida

affects the lower portion of the neural tube. The three types of spina bifida ranging from

mild to severe are spina bifida occulta, meningocele, and myelomeningocele (Figure 1.1)

(Botto et al. 1999). Spina bifida occulta is a bony defect of the spine usually covered by

normal skin with no spinal cord and nerve abnormalities and few if any complications.

Meningocele is characterized by a sac containing the meninges and CSF, which protrudes

through the opening in the vertebrae leaving the spinal cord and nerves intact. Cysts can

be surgically removed after which normal development ensues. Myelomeningocele is the most severe form and is frequently incurable. The sac that forms in this case contains meninges, nerve roots and part of the spinal cord that is damaged or incompletely developed (Kaplan et al. 2005).

Spina bifida patients who survive often have lifelong disabilities such as paralysis, problems with bowel and bladder control, and hydrocephalus (Botto et al. 1999). In addition to the emotional cost of NTDs, the economic cost for managing spina bifida is high. The direct cost of care may reach up to $560,000 per live birth (Grosse et al. 2008).

Thus, effective preventive measures for this disease are urgently needed.

Figure 1.1 The three types of spina bifida

(Botto et al. 1999)

Genetic and environmental risk factors. Epidemiologic studies suggest that the

combined action of genetic and environmental factors cause NTDs, which therefore

constitute classic examples of multifactorial disorders (Botto et al. 1999; Detrait et al.

2005). Interactions may occur between alleles of the same gene, among genes, and

between genotypes and certain environmental factors giving rise to the multiple and

complex etiologies of NTDs (Botto et al. 1999).

The genetic contribution to NTDs is evident through the association of this phenotype

with certain genetic syndromes such as Meckel syndrome (Detrait et al. 2005; van der Put

et al. 2001). Various chromosomal disorders, such as trisomy 13 and trisomy 18, are

frequently associated with NTDs. The influence of genetics is also demonstrated by the

2-5% recurrence risk for siblings (Detrait et al. 2005), presence of a positive family history in affected individuals, and a higher frequency of the disease in certain ethnic

groups such as Hispanics in the United States (Botto et al. 1999).

Several environmental factors have been implicated in NTDs, but the most consistent

include maternal diabetes and obesity (Detrait et al. 2005) and the maternal use of anti-

convulsant medication, such as valproic acid, for the treatment of epilepsy (Kaneko et al.

1999; Mitchell 2005). Low socioeconomic status is also an important risk factor

suggesting that poor nutrition contributes to NTDs. In 1976, Smithells and colleagues

reported that NTDs in babies were related to low maternal serum levels of micronutrients,

including several vitamins (Smithells et al. 1976). In 1983, a nonrandomized intervention study was conducted by the same group in which women planning to become pregnant

after having a previous NTD affected fetus or infant were supplemented with a multivitamin containing 360 µg of folic acid (FA) (Smithells et al. 1983). This study reported an 86 per cent lower risk of NTD-affected pregnancies in women who took the multivitamin compared to women who did not (Smithells et al. 1983). This provided evidence for the association between NTDs and FA (Botto et al. 1999; Mitchell 2005).

Recent studies have also suggested potential associations between NTDs and maternal myo-inositol, zinc and glucose levels (Groenen et al. 2003a; Groenen et al. 2003b).

Folate-Homocysteine association. Folate is a water soluble B vitamin (B9), found naturally in leafy vegetables, citrus fruits, beans, and whole grains. FA is the synthetic form of the vitamin that is found in dietary and vitamin supplements. Folate acts as a substrate for several enzymes involved in purine and pyrimidine synthesis that are essential for DNA synthesis (van der Put and Blom 2000). Folate also has a role in protein and lipid metabolism. Although FA supplementation prevents many NTD cases, the mechanism by which it exerts its beneficial effect is unknown.

The focus on the association between FA and homocysteine (HCY) resulted from the observation that increased HCY is found in many NTD affected pregnancies and that supplementation with FA and vitamin B12 reduces maternal HCY levels (Hague 2003; van Guldener and Stehouwer 2001). It has been proposed that supplying a significantly high dose of FA corrects NTDs by correcting a metabolic defect in the HCY pathway

(Figure 1.2).

HCY is a sulfur-containing amino acid that is not supplied by any dietary source but is

the result of normal methionine metabolism (Bottiglieri 2005). HCY has two fates; it is

metabolized through either the methylation or the transsulfuration pathways (Figure 1.2).

It is remethylated into methionine through the methylation cycle, which is needed both for the removal and formation of HCY (Bottiglieri 2005). This pathway is catalyzed by methionine synthase and utilizes N5-methyltetrahydrofolate as a methyl group donor

substrate, and vitamin B12 as a cofactors (Bottiglieri 2005). Through this pathway, HCY

acts as a substrate for the recycling of intracellular folates and the catabolism of betaine

and choline (Finkelstein 1998). In addition, HCY is involved in the transsulfuration pathway that is catalyzed by cystathionine β-synthase and cystathionine γ-lyase, both of

which require vitamin B6 as a cofactor (van der Put and Blom 2000). The

transsulfuration pathway leads to the formation of cystathionine, cysteine, and

glutathione (Finkelstein 1998). S-adenosylmethionine (SAM), the activated form of

methionine that is synthesized in the methionine cycle, is the principal biological methyl

group donor leading to methylation of nucleic acids, proteins, phospholipids, myelin,

polysaccharides, choline, catecholamines and a large number of other small molecules

(Bottiglieri 2005). Studies suggest that the cause for NTDs is a metabolic block in the

folate-HCY pathway or an increased requirement for one of the many folate-dependent

enzymes rather than simple folate deficiency (Scott et al. 1994).

Figure 1.2 Folate-Homocysteine metabolic pathway

Purine 10-formyl THF biosynthesis

THF Methionine DMG SAM Folate Methylene Cycle THF Betaine SAH Methyl Homocysteine THF

Thymidine biosynthesis CystathionineCystathionine CGL B MTHFR=methylenetetrahydrofolate 6 reductase Cysteine MS=methionine synthase CBS=cystathionine beta synthase CGL=cystathione gamma lyase Glutathione

Adapted from: www.cerefolin.com/Homocysteine

Prevention measures. Routine prenatal screening for NTDs is possible through the maternal serum alpha-feto protein (AFP) screening test, which is part of the “triple screen” (Graves et al. 2002). The triple screen measures levels of AFP, human chorionic gonadotropin hormone (hCG), and estriol. The results of these measures are combined with the mother's age, weight, ethnicity and diabetic status to determine the extent of the risk of certain genetic disorders (Graves et al. 2002). At 16-18 weeks of gestation, if a high AFP level is detected indicating a high risk for NTDs, two additional tests are performed: ultrasound to examine the spine of the fetus and amniocentesis to detect amniotic AFP and acetylcholinesterase levels (Graves et al. 2002). Detecting a high risk

for NTDs subsequently results in the option of termination of the pregnancy. About 75% of open spina bifida cases produce a positive AFP test (Jallo 2005). Therefore, a negative

AFP test does not rule out NTDs. Although this test provides a basis for managing the current pregnancy, it does not prevent the disease. NTDs arise during the first 28 days of pregnancy usually before a woman even knows that she is pregnant (Sadler 1998).

Consequently, preventive measures are necessary before the screening stage.

Although evidence for the association between NTDs and the vitamin FA accumulated rapidly in the 1970s and 1980s (Mitchell 2005), it was not until a decade later that the findings of two FA supplementation clinical trials conducted by the British Medical

Research Council (MRC) and a Hungarian group fueled public health efforts to reduce the prevalence of NTDs using dietary supplementations of FA (Botto et al. 1999).

Although the mechanism by which FA prevents NTDs is unknown, several clinical studies showed that as many as 70% of NTDs can be prevented with FA supplementation before and during the early weeks of pregnancy.

The US Public Health Service (USPHS) recommends that women of child bearing age receive 400 µg (0.4 mg) of FA daily. Moreover, a daily intake of 4000 µg (4 mg) of FA is recommended by the Center for Disease Control and Prevention (CDC 1992) for women with a previous NTD affected pregnancy. The average diet in the United States contains

200 µg of naturally occurring folate that has less bioavailability than synthetic FA used in supplements (1999). Mandatory fortification of flour products with FA by the US federal government provides the average woman an additional 100 µg of FA daily (1999;

Brouwer et al. 2001). Thus, prevention of NTDs depends on the daily intake of a FA

containing supplement in addition to dietary folate.

Unfortunately, despite this widespread public health effort, about 30% of cases appear resistant to the beneficial effect of FA supplementation (Botto et al. 1999). Many pregnancies are unplanned, and even when they are planned many women do not comply with the recommended preventive supplementation. A poll in 1998 showed that 70% of women between the age of 18 and 45 years are not following the USPHS recommendation even six years after its publication (1999). As a result, noncompliant individuals who are potentially FA-responsive are subjected to an otherwise preventable higher risk for NTD pregnancies. Thus, by predicting responsiveness compliance could dramatically improve and the full individual benefit of FA supplementation would be realized.

Adverse effects of high levels FA. Identifying FA-resistant cases helps avoid unnecessary and possibly unfavorable intake of high levels of FA by the general population and allows the development of alternative approaches for prevention of NTDs.

The Institute of Medicine (IOM) has set the tolerable upper intake level of FA from fortified foods or supplements as 1000 µg/day (IOM 1998). Greater consumption of FA might delay diagnosis of vitamin B12 deficiency by masking the symptoms of the disease

(Lamers et al. 2004). In addition, FA supplementation in cases of vitamin B12 deficiency may lead to worsening of vitamin B12’s enzymatic functions and is associated with rapid cognitive decline (Selhub et al. 2007). Intakes of more than 500 µg/day of FA have no

additional functional bioefficacy, at least in healthy individuals (Brouwer et al. 2001).

When a single dose of more than 250 µg FA is fed, unmetabolized FA is present in the serum (Kelly et al. 1997). Also, the ingestion of 800 µg FA by pregnant women shows the unmetabolized form in their circulation and in the cord blood of their infants

(Johnston 2008). It is unclear however what health risks unmetabolised FA can have

(Brouwer et al. 2001). Although FA can prevent the development of colon cancer, the timing of this supplementation is crucial as it may have a negative effect by enhancing the progression of malignant or premalignant lesions (Cole et al. 2007; Kim 2003). High

FA levels could also affect the control of seizures by lowering circulating levels of some anti-seizure medications (Johnston 2008). High levels of FA hinder anti-malarial treatment in pregnant women (Ouma et al. 2006) as well as in children (Dzinjalamala et al. 2005), which is of high concern especially in some African countries where women are supplemented with the easily available 5 mg dose of FA tablets (Ouma et al. 2006).

Many questions remain unanswered in regard to the mechanism by which NTDs arise and by which FA exerts its beneficial effect. Thus, studies are necessary to identify pathways involved in the mechanism of NTD development and in the mechanism of

NTD response to FA supplementation. In addition, identifying maternal genetic factors and readily tested biomarkers that could predict the outcome of NTD risk in the fetus due to maternal FA supplementation is of importance. Such predictors would allow development of individual-oriented prevention plans eliminating the need for food fortification with FA. Thus, treatment could be targeted to maximize benefits and minimize risks.

NTDs in Mice

Mice as models for human NTDs. More than 190 mouse mutants and strains exhibit

NTDs, and still others are emerging from ongoing gene targeting studies and mutagenesis

screens (Greene and Copp 2005; Harris and Juriloff 2007). Mice have been extensively

used to study the developmental biology of NTDs because embryos can be directly

examined, and specialized strains and crosses can be used. Based on developmental

observations, the process of neural tube closure in mice and humans is fundamentally

similar (Juriloff and Harris 1998). Neurulation takes place as a coordinated set of morphogenetic events within the neural plate, a region of specialized dorsal ectoderm

(Copp et al. 2003). Neurulation events involve the elongation and shaping of the neural

plate to develop bilateral neural folds at its junction with surface (non-neural) ectoderm,

elevation of the neural folds from a semi-horizontal to a vertical position leading to

apposition and fusion of the folds at the midline (Figure 1.3) (Copp et al. 2003; Juriloff

and Harris 2000).

Figure 1.3 Sequential change of typical neural fold morphology during the elevation process

(Juriloff and Harris 2000)

The mouse neural tube consists of several cranio-caudal zones each having distinct initiation sites (Juriloff and Harris 2000). In these zones elevation takes place in a wavelike longitudinal manner leading to fusion of the folds within each zone upon apposition and contact (Juriloff and Harris 2000). A ziperlike extension of the contact and fusion leads to collision and fusion of the distinct initiation sites along the length of the folds (Juriloff and Harris 2000). In mouse embryos the neural tube forms during day 8-10 of gestation (Juriloff and Harris 2000). Fusion of the rostral neural tube is at three sites

(Figure 1.4) with closure progressing from each site to form two cranial neuropores (Van

Allen et al. 1993). The first contact between the neural folds occurs in the hindbrain/cervical region and fusion proceeds rostrally and caudally from this site on late day 8 and early day 9 of gestation and is termed closure 1 (Macdonald et al. 1989). The second closure event, closure 2, occurs at the forebrain/midbrain boundary (Copp et al.

2003). The caudal spread of closure from this site meets the rostral progression from closure 1 to complete closure at the hindbrain neuropore (Copp et al., 2003). The position of closure 2 is polymorphic being relatively caudal in the midbrain in some strains and rostral in the forebrain in others (Copp et al. 2003). Closure 3 initiates at the rostral extremity of the forebrain and progresses caudally from this site to meet the rostral spread of fusion from closure 2 completing closure at the anterior neuropore, the most anterior opening of the neural tube (Copp et al. 2003). Unlike closure 1-3, closure 4 does not occur by fusion of the neural folds but rather by elongation of a membrane that covers the rombencephalon (hindbrain) (Golden and Chernoff 1993).

Failure of closure 1 causes craniorachischisis (Copp et al. 2003). Failure of closure 2 on

day 9 leads to exencephaly –the equivalent of anencephaly in humans; failure to close at

the base of the tail on day 10 leads to spina bifida (Juriloff and Harris 1998; Kaufman

1992). Human cases of NTDs, anencephaly and spina bifida, are anatomically similar to

many mouse cases, suggesting that the failure of elevation occurs at similar elevation

zones (Juriloff and Harris 2000; Van Allen et al. 1993).

Figure 1.4 Three sites of closure of the rostral neural tube

(Copp et al. 2003)

More than 80% of NTD cases in humans are nonsyndromic, meaning that the only obvious birth defect present is the NTD (Harris and Juriloff 1997). Furthermore, the genetics of common human NTDs is characterized by low and incomplete penetrance, a nonsyndromic nature, survival into the fetal period, and in some cases dietary modification of penetrance (Juriloff and Harris 2000). Thus, a good genetic model of common human NTDs is most likely a non-syndromic NTD with a degree of genetic complexity, low penetrance, and a response to nutrient supplementation (Juriloff and

Harris 2000). However, few of the NTD mutations in mice fit these criteria fully (Juriloff

and Harris 2000). In mice NTDs from most genetic causes are syndromic (Harris and

Juriloff 1997). However, these too serve as models for human NTDs. Different types of

single gene and multigenic mouse models exist for NTDs. Most of these mutants

demonstrate variable low penetrance and some show complex inheritance patterns

(Juriloff and Harris 2000). In addition, some have specific responses to different

nutritional supplementations (Table 1.1) (Juriloff and Harris 1998).

The many mouse models that display genetic complexity include the curly tail mutant

mouse (referred to hereafter as the ct mutant) in which the ct mutation arose

spontaneously and acts as a component of a multifactorial system with its modifier mct1

(Harris and Juriloff 1997). Another such model, SELH/Bc, is an inbred strain that has

relatively high incidence of spontaneous NTDs demonstrating a genetically multifactorial

liability (Harris and Juriloff 1997).

Many single gene mouse models also exist for NTDs. Looptail (Lp) is an example of a

single gene mutation that is inherited in a semi-dominant manner resulting in

craniorachischisis. The spontaneous mutation Axd (axial defects) displays semi-dominant inheritance and gives rise to a potentially nonsyndromic spina bifida model (Essien

1992). Mice with the fully penetrant and semidominant Splotch (Sp) mutation are a

Mendelian syndromic model of NTDs (Juriloff and Harris 1998). Crooked tail (Cd) is a

spontaneously arising mutation that produces exencephaly (Carter et al. 1999). A targeted

mutation of the Apob gene results in exencephaly alone or accompanied by

hydrocephalus (Juriloff and Harris 1998). Mouse knockouts for Cart1 as well as Cited2 also exhibit NTDs (Greene and Copp 2005). Additional details about the mouse mutants chosen to model human NTDs are discussed in the next two sections.

Supplementation studies and metabolic pathways. Among the more than 190 genetic mouse models that exhibit NTDs, only 10 have been tested for response of the fetal

NTDs to maternal FA supplementation (Table 1.1).

FA-resistant mutants. Although several NTD mutants support responsiveness to FA supplementation, many NTDs are resistant. Remarkably some are preventable with supplementation with other nutrients suggesting that other metabolic pathways contribute to protection against NTDs. ct is a FA-resistant mouse model that is however inositol responsive; inositol seems to exert this protective effect on NTD development by increasing flux through the inositol/lipid cycle and consequently downstream stimulation of PKC activity (Greene and Copp 1997). In another FA-resistant mutant, Axd, maternal supplementation with the amino acid methionine reduces the frequency of spina bifida among homozygous embryos by 41% (Essien 1992). The Bent tail mouse (Bn) is resistant to maternal supplementation with folinic acid, myo-inositol and zinc (Franke et al. 2003), leaving the metabolic pathway involved in its NTD phenotype to be identified. SELH/Bc mice are resistant to FA as well as methionine supplementation. However, the risk for

NTDs in SELH/Bc is reduced 3-fold by substituting one normal commercial mouse chow for another indicating that a dietary agent or a metabolic pathway influencing neural tube closure in these mice remains to be identified (Juriloff and Harris 1998). When Ski

mutant mice were treated with FA during the first half of gestation there was no effect on

the phenotype of Ski mutant homozygotes (Ernest 2004). No other nutrients were studied in the Ski mutant and no metabolic pathways were identified.

FA-responsive mutants. It is surprising that in the responsive mouse models, FA is not only beneficial in conditions of folate deficiency, such as in mice with a deletion of folate receptor 1 (Folr1- formerly Folbp1) where folate pathway is directly affected (Kappen

2005), but also in others such as Cart1 deficient mutants (Zhao et al. 1996) and Cited2

deficient mutants that do not have any known defects in folate metabolism (Barbera et al.

2002). A common finding in the two folate responsive models, Cart1 and Cited2 mutants,

is that the mutation affects a gene that encodes a protein functioning in transcriptional

regulation; the cartilage homeoprotein 1, a homeobox-containing transcription factor in

Cart1, and a member of the CITED family of transcriptional regulatory proteins in the

Cited2 mutant (Greene and Copp 2005). The Sp mutation in the paired box containing

transcription factor, Pax3, gives rise to another FA-responsive mutant with no known

defect in folate metabolism (Greene and Copp 2005). In mice with either of the three

mutations, Cart1, Cited2, and Sp, the pathways involved in folate responsiveness remain

unidentified, although a deoxyuridine suppression test suggested that a metabolic

deficiency in the supply of folate for pyrimidine biosynthesis exists in Sp mutant embryos

(Fleming and Copp 1998). These three mouse models might also indicate an interaction

of folate status with developmental pathways such as those controlled by Cart1 and

Cited2 (Kappen 2005). In mice with the Cd mutation, FA supplementation ameliorates

NTDs although the mutated gene, Lrp6, is not a direct participant in FA transport or metabolism (Carter et al. 1999).

Untested mutants. Many NTD mutants are still not tested for response of embryonic

NTDs to maternal FA supplementation. Some of these mutants have known metabolic or signaling pathway perturbations while others have unknown pathway defects. The untested Apob mutant has an established genetic perturbation in the cholesterol metabolism pathway (Homanics et al. 1993). Another FA-untested mutant, the Lrp6 knockout mutant (referred to as the Lrp6 mutant), has a perturbation in the Wnt/β-catenin signaling pathway (Khan et al. 2007; Pinson et al. 2000). The Lp mutant is a result of a loss-of-function (LOF) mutation in the Vangl2 gene, which prevents binding of Vangl2 to the cytoplasmic protein Dishevelled (Dvl), which in turn is involved in the Wnt/β-catenin signaling and planar cell polarity (PCP) pathways (Torban et al. 2004). It is important to add these mutants among others to the list of known FA-response mutants, facilitating the identification of pathways involved in the pathogenesis of NTDs and in the response to

FA.

Importance of the different FA-response mutants. The availability of NTD mouse models with specific responses to different nutritional supplements (Table 1.1) is important for modeling human responsiveness and resistance to FA and provides a model for alternative approaches of NTD prevention in cases of FA-resistance. The availability of these well characterized mouse models is also needed to better understand the mechanism by which NTDs arise and by which FA exerts its effect to decrease the risk of this defect.

In addition, these FA-tested mutants may provide a means for identification of maternal markers that distinguish individuals with alternative response to FA supplementation, thus avoiding the intake of high levels of FA in cases where it is not needed.

Table 1.1 Mouse NTD responses to different nutritional supplementations or treatments

NTD mutant Folate Methionine Thymidine Inositol Vit B12 Zinc

1Cart1  N/D N/D N/D N/D N/D 2Cd  N/D N/D N/D N/D N/D 3Cited2  N/D N/D N/D N/D N/D 4Folbp1  N/D N/D N/D N/D N/D 5Sp  X  N/D N/D N/D 6Axd X  N/D N/D X N/D 7Bn X N/D N/D X N/D N/D 8ct X X N/D  N/D N/D 9SELH/Bc X X N/D N/D N/D X 10Ski X N/D N/D N/D N/D N/D

 = Responsive; X=Resistant; N/D= Not done 1Cartilage homeoprotein 1 (Zhao et al. 1996) 2Crooked tail (Carter et al. 1999) 3CBP/p300 interacting transactivators with glutamic acid (E)/(D)-rich C-terminal domain (Barbera et al. 2002) 4Folate binding protein 1 (Piedrahita et al. 1999; Spiegelstein et al. 2004) 5PAX3 transcription factor (Fleming and Copp 1998) 6Axial defects (Essien 1992) 7Bent tail (Franke et al. 2003) 8Curly tail (Greene and Copp 1997; Tran et al. 2002; van Straaten et al. 1995) 9SELH/Bc (Juriloff and Harris 1998; Tom et al. 1991) 10Proto-oncogene (Colmenares et al. 2002)

Previous response predictive profiles. Work by a previous member of the Nadeau lab

suggests that gene expression and metabolite surveys can be used to develop profiles that are diagnostic for particular disease conditions or treatment responses (Ernest et al.

2006). Ernest et al. compared expression profiles of the Cd mouse, a NTD FA-responsive model, with those of previously studied mouse mutants of various phenotypic defects.

Mice with mutations in the Apob, Gli3, Pax3, Ptch and Ski genes were used as models for

NTDs, and mice with a mutation in the Apc gene as a model for colon cancer (Ernest et

al. 2002). These single gene mutation mice were chosen because they have phenotypes

similar to those associated with anomalies in folate-HCY metabolism in humans. These

mutants were primarily used to test whether they have altered HCY levels and expression

profiles of most of the genes involved in folate-HCY metabolism. Their expression

profiles later demonstrated that mice with phenotypic defects that are folate-responsive

clustered together: Cd, Apc, and Pax3 mutants; and those with unknown response

clustered together: Apob, Ptch, and Ski mutants. This predicted that mutants of the latter

group would be folate-resistant or would have a functionally different response from the

mutants in the former group. Interestingly, the Ski mutant was later treated with FA and

the NTD was found to be resistant (Ernest et al. 2006). These results suggest that

development of tests predictive of folate-response may be feasible. Thus, we aimed to

study gene expression and metabolite profiles of a number of NTD mutants with

confirmed response to FA supplementation to identify maternal markers (genes,

pathways, and metabolites) predictive of embryonic NTD response to FA. In addition to

identifying pathways involved in FA-responsiveness or -resistance.

Summary of mouse mutants selected for modeling human NTDs. The availability of various mouse models of NTDs involving a variety of known and unknown genes and consequently relating to different signaling pathways, and the specific cases in which these NTDs can be prevented by exogenous agents, specifically FA, provides the opportunity to search for genetic, molecular and pathway markers that are diagnostic of fetal NTD response to maternal dietary FA supplementation.

We have chosen to study a diverse set of mutants that have known and unknown metabolic and signaling pathway perturbations and that can be classified into three groups based on the NTD response to FA supplementation: responsive, resistant, and unknown (Table 1.2).

- Splotch (Sp). A semi-dominant point mutation lies within intron 3 of the paired

homeobox 3 (Pax3) gene on mouse Chromosome 1 (Epstein et al. 1993). Mice

with this Pax3Sp2H mutation (referred to as Sp mutants) are characterized by white

spotting on the abdomen, tail and feet in heterozygotes whereas homozygotes

display a pleiotropic phenotype whose features include exencephaly,

meningomyelocele, and spina bifida (Epstein et al. 1993). Pax3 codes for a

transcription factor that regulates morphogenetic cell behavior and regulates a

conserved planar cell polarity (PCP)/non-canonical Wnt-signaling cascade

entailing c-JUN-N-terminal Kinase (JNK) (Wiggan and Hamel 2002). The Sp

mutant is a model of FA-responsive NTDs; it also responds to thymidine but not

to methionine (Fleming and Copp 1998; Greene and Copp 2005; Juriloff and

Harris 2000).

- Crooked tail (Cd). A spontaneous missense gain-of-function (GOF) mutation

replaces a single highly conserved amino acid in the low-density lipoprotein

(LDL) receptor-related protein, Lrp6 (Carter et al. 2005). This Lrp6Cd mutation

produces a crooked tail in heterozygotes and an increased incidence of

exencephaly among homozygotes (Carter et al. 1999). LRP6 is an LDL co-

receptor for wingless (Wnt) and is required for Wnt signaling via the frizzled (Fz)

receptor-β-catenin pathway (Carter et al. 2005; Khan et al. 2007; Pinson et al.

2000). The Lrp6Cd mutant (referred to as the Cd mutant) is a model of a FA-

responsive NTD (Carter et al. 1999).

- Lrp6. The Lrp6 gene was mutated using a gene trap that causes LOF (Pinson et al.

2000). Embryos homozygous for this mutation show several birth defects

including caudal axis truncation and limb deformities as well as exencephaly and

spina bifida in some embryos (Pinson et al. 2000). This Lrp6 mutant however, has

not yet been studied for the effect of FA or other nutrients on its NTDs.

- Apob. The apolipoprotein B (Apob) targeted mutation (Apobtm1Unc) replaced the

low density lipoprotein binding domain of the endogenous exon 26 with

oligonucleotides encoding human β-globin (Homanics et al. 1993). ApoB

normally exists in two forms, the ApoB100 and the ApoB48 that are both the

product of the same gene (Homanics et al. 1993). However, in homozygotes and

heterozygotes, the ApoB48 isoform is unaffected by this mutation, whereas a

truncated mutant ApoB100 isoform, ApoB70, is produced from this allele

(Homanics et al. 1993). Homozygotes display perinatal mortality associated with

exencephaly (Homanics et al. 1993) in addition to displaying hallmarks of the

human disorder: low plasma concentration of ApoB, β-lipoproteins, cholesterol

and vitamin E (Homanics et al. 1995). Approximately 30% of the perinatal

homozygotes are exencephalic (Homanics et al. 1995). Furthermore, by 8 weeks

of age, those that have closed neural tubes show hydrocephalus in 32% of

homozygotes and in 1% of heterozygotes. However some unaffected

homozygotes usually survive to maturity, appear healthy, and reproduce but some

of their offspring can be exencephalic or hydrocephalic (Homanics et al. 1993).

ApoB is a major structural component of several lipoproteins that transport

cholesterol, lipids and vitamin E in the circulation (Homanics et al. 1995),

suggesting a role of the cholesterol metabolism pathway in neural tube closure.

The NTD in the Apobtm1Unc mutant (referred to as the Apob mutant) has not been

tested for response to nutrient supplementation.

- Looptail (Lp). The Vangl2 (Van Gogh-like, formerly “loop tail associated

protein,” Ltap) is the gene mutated in this mouse model of severe NTDs. Two

independent alleles have been described for Vangl2Lp mice (referred to as Lp

mutants), a naturally-occurring and a chemically-induced allele, both within the

Vangl2 cytoplasmic domain (Torban et al. 2004). Both alleles represent a LOF

mutation associated with the same severe NTD phenotype (Torban et al. 2004).

Mouse embryos homozygous for the Lp mutation fail to initiate neural tube

closure at E8.5 at the cervical/hindbrain boundary leading to craniorachischisis

(Greene et al. 1998). The mutation is inherited in a co-dominant fashion and the

heterozygous phenotype is characterized by a looped tail (Torban et al. 2004).

Vangl2 is a 521 amino acid transmembrane (TM) protein composed of four

putative TM domains in the N-terminal half; the C-terminal half seems to be

cytoplasmic and possibly involved in intracellular signaling (Torban et al. 2004).

Vangl2 interacts with mammalian Dishevelled (Dvl) proteins, which in turn is

involved in Wnt/β-catenin and PCP pathways (Torban et al. 2004). No nutrient

supplementation studies have been carried out to identify response of the NTD

phenotype in this mutant.

- Curly tail (ct). the ct mutation arose spontaneously in 1950 and displays variable

expression and an incomplete penetrance; the ct strain is maintained as a random

bred closed colony, in which all individuals are considered to be of genotype ct/ct

(van Straaten and Copp 2001). At least three modifier loci were identified for ct,

which points to a multifactorial inheritance (van Straaten and Copp 2001). 1-3%

of homozygous (ct/ct) mice have exencephaly, 15-20% show spina bifida aperta,

and 40-50% have tail flexion defects, a curled tail (Cockroft et al. 1992). The

nature of the ct gene product remains unidentified (van Straaten and Copp 2001).

NTDs in curly tail are resistant to FA and methionine but can be prevented with

myo-inositol (Greene and Copp 1997; van Straaten et al. 1995).

- CBA. This is an inbred strain of mice, to which the original curly tail mutant was

crossed, which is therefore thought to have a similar genetic background to ct and

can be used as its control (Cockroft et al. 1992). CBA embryos show a similar

high incidence of cranial NTDs after culture in inositol deficient medium, similar

to ct/ct embryos (Cockroft et al. 1992). The lack of response of ct/ct embryos to

supplementation with low doses of inositol distinguishes them from the

responsive CBA strain (Cockroft et al. 1992).

- C57BL/6J. This is an inbred strain of mice that is the genetic background for the

Pax3Sp2H and Apobtm1Unc mutations. Thus this strain is used as the control for these

two mutants.

Table 1.2 NTD mutants chosen for study

Allele Mutant Mutation type

Sp2H Responsive Pax3 Splotch (Sp) LOF Lrp6Cd Crooked tail (Cd) GOF Lrp6KO Lrp6 LOF Unknown Apobtm1Unc Apob LOF Vangl2Lp Loop tail (Lp) LOF Resistant unknown Curly tail (ct) ??

Research Objectives

Only 30% of women comply with recommendations for maternal dietary FA supplementation, a treatment that has been shown to decrease the prevalence of NTDs by

50-70%. However, the mechanism by which NTDs are corrected is unknown and there are no markers that distinguish individuals with alternative responses. If maternal genetic factors could predict fetal NTD response to maternal FA supplementation, compliance might dramatically improve. In addition, possibly adverse effects of high levels of FA could be avoided in the 30-50% of cases that show no decrease in NTD risk. The mouse is an important model for studying NTDs. In addition to having a mechanism of neural tube closure fundamentally similar to that in humans, NTD mouse mutants provide models for the variable responses to FA supplementation seen in humans.

A hypothesis-driven approach, also known as supervised classification, is typically used in studies of class-prediction (McShane et al. 2003; Pusztai and Hess 2004). With this method, an algorithm or classifier is developed to assign cases to a priori defined categories. Classifiers are typically developed based on a training data set with known class membership, in our case responsive & resistant mutant groups, and evaluated based on an independent test data set also with known classifications (Allison et al. 2006;

Brazma and Vilo 2001). The test data set is not initially included in the classification process, but is used later to evaluate the accuracy of the generated classifier (Allison et al.

2006; Brazma and Vilo 2001). With a smaller sample size in the training set, classifiers tend to predict response well on the training set but fail to predict accurately on a test set of independent data (Allison et al. 2006). Thus, a better and more accurate classifier is

possible with a larger training data set. However, among the more than 190 mouse NTD

models, only 10 have been tested for the effect of FA. As a result, additional mutants

need to be tested to facilitate the identification of maternal markers of fetal NTD response

using the supervised classification method and to aid in the identification of pathways and mechanisms involved in the determination of this response. In Chapter 2, a dietary

FA supplementation study was conducted in three mutants whose response had been previously unknown: the Apob and Lp mutants studied in our lab, in addition to the Lrp6 mutant that was studied as part of a collaboration with the laboratory of Dr. Margaret

Elizabeth Ross at Cornell University (referred to hereafter as the “Ross lab”). The embryos of all three mutants were examined visually for the effect of maternal FA supplementation on their NTD phenotype; the cartilaginous skeletons of the Apob and Lp mutant embryos were also examined to identify subtle effects FA might have on skeletal development. None of the three mutants showed a decrease in the frequency of embryonic NTDs with a supplemented diet (10 ppm FA) compared to a control diet (2 ppm FA). In addition, FA did not have any detectable effects on skeletal development of both the Apob and the Lp mutant embryos. Thus all three mutants fall within the FA- resistant group with respect to neural tube development. However, supplemental FA showed an unexpected and negative effect on viability of the mutant embryos, where we found an increased loss of homozygous and more importantly of otherwise unaffected heterozygous embryos with the higher FA concentration. These findings raise the question of whether the decrease in the risk of NTDs with FA supplementation in humans results from an increase in loss of affected embryos rather than a correction of the NTD phenotype and whether healthy embryos are also lost in the process.

In a previous study of mouse mutants whose phenotypic defects are associated with anomalies in HCY metabolism, gene expression and metabolite profiles related to the folate-HCY metabolic pathway distinguished mouse mutants based on the responsiveness

or resistance of their phenotype to FA supplementation and correctly predicted FA

response in a previously untested NTD mutant (Ernest et al. 2006). Our hypothesis is that

a broad survey of maternal gene expression and metabolite profiles would identify genes

and pathways that predict embryonic NTD response to maternal FA supplementation and

would identify pathways and mechanisms involved in the determination of this response.

Thus, in Chapter 3, whole-genome expression data and metabolite data for a panel of two

FA-responsive (the Sp and Cd mutants) and four FA-resistant (the Lrp6, Apob, Lp, and ct

mutants) female NTD mutants were analyzed. Liver was used as the tissue of choice for

gene expression as it is the major site for many folate-HCY metabolism studies

(Finkelstein 1998; Finkelstein et al. 1990) and because this tissue was previously used by

Ernest et al. in the study of FA-response predictive profiles. Folate and vitamin B12 were

measured in erythrocytes as a reflection of tissue stores; however these levels were not

indicative of FA-response in the different mutants. Genes significantly affected by each

of the NTD causing mutations were identified in each of the mutants. These genes were

then associated with metabolic and signaling canonical pathways using Ingenuity

Pathways Analysis (IPA) with the Sp and Cd mutants known to be responsive and the

Lrp6, Apob, Lp and ct mutants as resistant. Two pathways involved in inhibition of

retinoid X receptor (RXR) and activation of the heterodimer, pregnane X receptor

(PXR)/RXR were significantly associated with the genes in each of the six NTD mutants

independent of FA-response. The nuclear receptor RXR binds retinoic acid receptor

(RAR) forming a heterodimer to which retinoic acid (RA) then binds (Germain et al.

2006). RXR-RAR bound RA ultimately activates genes involved in cell differentiation and determines positioning along the anterior/posterior embryonic axis (Holland 2007).

On the other hand, accumulated toxic compounds in the body activate another nuclear receptor, PXR. Activated PXR forms a heterodimer with RXR to activate transcription of certain genes involved in the body’s defense mechanism (Kliewer et al. 2002). Thus, these two pathways can potentially be involved in a shared mechanism of NTDs. Using a different approach, we identified two independent gene lists, one unique to all responsive mutants and the other unique to all resistant mutants where all pathways associated with these gene lists were also unique to the specific response type. This finding suggests that multiple mechanisms exist for FA-response.

The study of individual mutants whose NTDs arise due to different pathway perturbations are essential for the determination of the mechanism by which FA affects these pathways.

In Chapter 4, the impact of FA supplementation on Lrp6 function and Wnt canonical

signaling was studied in the Lrp6 LOF mutant (referred to as the Lrp6 mutant). The Ross lab previously identified Cd as a FA-responsive NTD mouse mutant that shows increased viability with FA supplementation. This NTD mouse has a mutation in the Lrp6 gene, a co-receptor for canonical Wnt signaling. In collaboration with this group we studied whether the effect of FA is specific to the hyperactive Lrp6Cd mutation or whether FA has

a direct impact on Lrp6 function and Wnt signaling. In addition to the GOF Cd mutation, the LOF mutation of Lrp6 was also characterized by NTDs in the homozygous embryos.

Based on the Cd results we predicted a similar FA-response in the Lrp6 mutant. We conducted a liver gene expression study on Lrp6 heterozygous and wild-type female mice fed either the 2 ppm control diet or the 10 ppm supplemented diet to determine the effect of FA on Lrp6 genotype. Diet-genotype interaction analysis identified 232 genes that change significantly between the heterozygous and wild-type Lrp6 genotype depending on the level of FA in the diet2 ppm10 ppm. IPA associated those genes to cellular proliferation and Wnt canonical signaling pathways. In vitro assays validated the gene expression results showing an interaction between Lrp6 gene dosage and FA level that influence proliferation of neural tube cells. In addition, FA levels influenced cytosolic ß- catenin accumulation and transcriptional activation in Wnt stimulated cultured cells.

However, the NTD in the Lrp6 mutant was identified as resistant to FA supplementation with an increase in the loss of homozygous and heterozygous embryos indicating that the

Wnt signaling pathway in this mutant responds to FA but this response does not cause a correction of the NTD phenotype. Based on these findings it is interesting to understand why these two mutations of the same gene would cause different NTD responses to FA supplementation. Characterizing the Wnt signaling response to FA in the Cd mutant would be an important step towards answering that question.

CHAPTER 2

Folic acid effect on the outcome of neural tube defects (NTDs) and embryonic

viability in three NTD mouse models

All the work in this chapter was performed by the candidate.

Abstract

Neural tube defects (NTDs) are the second most common birth defect. The risk for this

defect decreases by 50-70% with maternal dietary folic acid (FA) supplementation.

However, the mechanisms of NTD development and the ways in which FA exerts its

beneficial effect remain unidentified. To better understand these aspects of NTDs,

characterized NTD mouse models with known etiology of the defect and response to FA

supplementation are needed. Among more than 190 mouse mutants and strains that

exhibit NTDs, only ten have been tested for the response to FA. Thus, we chose to examine the effect of FA supplementation on the NTD phenotype of three NTD mutants:

Apobtm1Unc, Vangl2Lp, and Lrp6KO (referred to as the Apob, Lp, and Lrp6 mutant,

respectively). All three mutants fell within the FA-resistant group as no decrease in the

frequency of embryonic NTDs was observed with a supplemented diet (10 ppm FA)

compared to a control diet (2 ppm FA). Interestingly, supplemental FA caused a negative

effect characterized by increased loss of homozygous and otherwise unaffected

heterozygous mutant embryos. These findings raise a concern of whether decrease in the

risk of NTDs with FA supplementation in humans is due to an increased loss of affected

embryos rather than a correction of the NTD phenotype and whether otherwise healthy

viable embryos are compromised. Based on reports in the human literature of female

prevalence of NTDs, this phenomenon was also examined in the Apob and Lp mutants.

However, gender-specific prevalence of NTDs was not evident in either mutant. Thus,

Apob, Lp, and Lrp6 mutants are NTD models of FA-resistance, with a potential adverse

effect of treatment.

Introduction

In humans, maternal supplementation with FA, the synthetic form of folate - a water soluble B vitamin (B9), decreases the prevalence of NTDs by about 70% (1991; Antony and Hansen 2000). However, the mechanism by which FA exerts its beneficial effect is unknown and markers are not available to distinguish individuals with alternative responses to supplementation. NTD mouse models with known etiology of the defect and known response to FA supplementation are needed to elucidate these important questions.

More than 190 mouse mutants and strains model NTDs (Greene and Copp 2005; Harris and Juriloff 2007). However, only 10 of these mutants have been tested for the effect of maternal FA supplementation (Table 1.1).

To generate a good and accurate classifier of NTD response to FA following the supervised classification approach, typically used in studies of class-prediction analysis

(McShane et al. 2003; Pusztai and Hess 2004), a big number of known FA-response NTD mouse mutants is required. Thus, a big data set will train the classifier to predict response well not only on the training data but also on a variable set of independent data (Allison et al. 2006). To increase the list of known FA-response NTD models, we conducted a dietary FA supplementation study in three mutants whose FA-response had been previously unknown: the Apob and the Lp mutant, which we studied, in addition to the

Lrp6 mutant that was studied by the Ross lab at Cornell University as part of a collaboration (refer to Chapter 4).

Apob is a single gene mutant with a loss-of-function (LOF) mutation in the apolipoprotien (apo) B gene (Homanics et al. 1993). Homozygous mutants show a 30% penetrance of exencephaly, the equivalent of anencephaly in humans, alone or accompanied by hydrocephalus (Juriloff and Harris 1998) in addition to hallmarks of the human disorder, low plasma concentration of APOB, β-lipoproteins, cholesterol and vitamin E (Homanics et al. 1995). By 8 weeks of age, the mutants that have closed neural tubes show hydrocephalus in 32% of homozygotes and in 1% of heterozygotes. Several lipoproteins that transport cholesterol, lipids and vitamin E in the circulation have APOB as a major structural component (Homanics et al. 1995), suggesting a role of the cholesterol metabolism pathway in neural tube closure.

The Lp mutant is characterized by a LOF mutation in the Vangl2 gene (Torban et al.

2004). This prevents binding of Vangl2 to the cytoplasmic protein Dishevelled (Dvl), which in turn is involved in the Wnt/β-catenin signaling and planar cell polarity (PCP) pathways (Torban et al. 2004). Embryos homozygous for the Lp mutation have craniorachischisis due to the failure of initiating neural tube closure at E8.5 at the cervical/hindbrain boundary (Greene et al. 1998). This mutation is inherited in a co- dominant fashion and the heterozygous phenotype is characterized by a looped tail

(Torban et al. 2004).

In the Lrp6 mutant, a gene trap was used to cause LOF in the low-density lipoprotein

(LDL) receptor-related protein (Lrp6) gene (Pinson et al. 2000). In addition to showing a number of birth defects including caudal axis truncation and limb deformities, some

homozygous mutant embryos also show exencephaly and spina bifida (Pinson et al.

2000). Lrp6 is a co-receptor for wingless (Wnt) and might be required for Wnt signaling via the frizzled (Fz) receptor-β-catenin pathway (Carter et al. 2005; Khan et al. 2007;

Pinson et al. 2000).

We show that in all three mutants, maternal FA supplementation does not prevent NTDs.

However, in this chapter we only present the work done on the Apob and Lp mutants but discuss their results in accordance with similar results of the Lrp6 mutant whose experiments will be detailed in chapter 4. Interestingly, maternal FA supplementation caused a considerable decrease in the numbers of homozygous and heterozygous embryos. The missing embryos were not accounted for by the number of resorptions observed. Thus, the effect of FA could be explained as negative represented by homozygous and heterozygous pre-implantation losses; or a positive effect by increasing the number of wild-type implantations relative to other genotypes. Based on reports in the human literature of female prevalence of NTDs (Whiteman et al. 2000), this phenomenon was also examined in the mutants. However, gender did not influence the frequency of

NTDs in Apob nor in Lp mutants. Thus, we can conclude that these mutants are unique models of FA-resistance, displaying an adverse effect of FA supplementation.

Materials and Methods

Mouse colonies. Apob and Lp mutants were purchased from the Jackson Laboratory.

All mice were raised on the PMI Nutrition Laboratory Autoclavable Rodent Diet #5010

and maintained by trio matings. Heterozygous males and females from both mutants were

weaned at 3 weeks of age on either a control diet containing 2 ppm FA (D05072702,

Research Diets) or a supplemented diet containing 10 ppm FA (D05072701, Research

Diets) and maintained on that for at least 3 weeks before they are mated. After at least 3

weeks of being on the assigned diet, timed pregnancies were generated by mating a 6-10

week old female with a male overnight. Upon discovery of a plug the females were kept

another 13 days on that same diet. Between E12.5 and E14.5 the pregnant females were

dissected and the embryos examined. The supplementation plan is outlined in Figure 2.1.

All the mice shared the same animal room with controlled temperature, humidity, and 12

hour light-dark cycle. Mice were provided food and water ad libitum.

Figure 2.1 Supplementation study outline 3 week old female and male heterozygous mice (Apob or Lp mutants) were weaned on either the 2 ppm or 10 ppm FA diet. At 6 weeks of age they were mated and between 12.5-14.5 days post coitum (dpc) the females were dissected and their embryos examined.

Special diets. The only difference between the two diets used for the supplementation

study is the concentration of FA that is 5 times more in the supplemented diet than the

control diet (Appendix 1). We determined 2 ppm FA to be the control concentration required for proper breeding and fetal development based on many factors; these include the report by the National Research Council (NRC) in 1995 stating that the estimated minimal FA requirement in mice is 0.5ppm, however this concentration does not include a margin of safety (NRC 1995). Any concentration added to the diet should be higher than the minimal cutoff to account for any losses during mixing and storage of the diet. In addition, a study similar to ours showed that 0ppm FA caused embryonic lethality of the

Cd/Cd mutant embryos with a shift to display the expected exencephalic phenotype at

4ppm. The percentage of affected Cd/Cd embryos decreased with higher concentrations of FA (7ppm & 10 ppm) indicating that 4ppm is acting as the control diet that allows the observation of the expected penetrance of NTDs which is not possible with 0ppm (Carter et al. 1999).

Phenotype assessment. After weaning, heterozygous males and females were assigned to either the control or supplemented diet and maintained on it for at least 3 weeks before timed pregnancies were generated. Between E12.5 and E14.5 the pregnant females were dissected and the embryos phenotyped for resorption, NTD or normal phenotype. Tissues were taken from all embryos for DNA extraction, genotyping and sex typing. In addition, the cartilaginous skeleton of all embryos was stained and examined. Statistical comparisons using the chi-square and Fisher’s exact tests, as appropriate, were performed

using GraphPad QuickCalcs Web site: http://graphpad.com/quickcalcs/chisquared1.cfm and http://graphpad.com/quickcalcs/contingency1.cfm (accessed April 2009).

Genotyping and sex typing. Genotyping for Apob was done according to the protocol provided by the Jackson laboratory. Lp genotyping protocol was also previously described (Copp et al. 1994). Embryos were sex typed using primers for Sry, the sex determining gene on the Y chromosome (Heaney et al. 2009).

Fetal cartilaginous skeletal staining. Alcian blue was used to stain the cartilaginous skeleton of all Apob and Lp mutant embryos following the protocol described in (Nagy et al. 2003). These skeletons were examined under the microscope for any changes contributed by the different FA concentrations in the different diets.

Determination of embryo number required for study. The chi-square goodness of fit test (Pollard 1977; Rigby 2001) was used to determine the number of embryos needed to be examined to reach statistical significance when assessing whether FA supplementation affects the NTD phenotype in each mutant. For these calculations, we considered the smallest decrease in the incidence of the NTD phenotype ≈( 55%) observed in previous

FA supplementation studies (Appendix 2) to be the one expected for Apob and Lp homozygous mutant embryos after FA supplementation.

Assuming a p-value of 0.05 and a chi-square (χ 2) value of 3.841, the calculations for determining the number of embryos to be examined are as follows.

Calculations: Let the total # of animals =Y; with a Mendelian ratio of 1:2:1, let the total # of -/- mutants = X = 0.25Y

1. Apob mutant.

Phenotype of Apob-/- Control diet Supplemented diet Expected # of Apob-/- Observed # of Apob-/- NTDs (30% penetrence) 0.3X (0.3/2)X (50% decrease in incidence compared to control) No NTDs 0.7X X-((0.3/2)X) Total # of Apob-/- X X

χ 2 = ∑ (O-E)2 / E

χ 2 = {[X – (0.3/2)X] – 0.7X}2 / 0.7X + {(0.3/2)X – 0.3X}2 /0.3X

3.841= {[X – (0.3/2)X] – 0.7X}2 / 0.7X + {(0.3/2)X – 0.3X}2 /0.3X

X = 36

Therefore, we needed to examine at least 36 -/- embryos for each of the control and FA- supplemented diet to achieve a statistically significant difference for the effect of FA on

NTDs. Following a 1:2:1 ratio, the total number of embryos required was (36)(4)=144

2. Lp mutant.

Phenotype of Lp-/- Control diet Supplemented diet Expected # of Lp-/- Observed # of Lp-/- NTDs (100% penetrence) X X/2 (50% decrease in incidence compared to control) No NTDs 0 X-(X/2) Total # of Lp-/- X X

In the case where at least one of the cells of the table has a small number (0 in our case),

Fisher’s exact contingency test should be applied (Pollard 1977; Sokal and Rohlf 1995).

Fisher’s exact test required the use of a contingency table that looked like this:

Phenotype of Lp-/- # of Lp-/- on # of Lp-/- on Row total control diet supplemented diet NTDs (100% penetrance) a b (50% decrease in incidence a + b compared to control) No NTDs c d c + d Column total a + c b + d n= a + b + c + d

The formula for calculating the p-value using Fisher’s exact test (Sokal and Rohlf 1995)

is complicated when trying to solve for X. Thus, we used empirical numbers to calculate

Fisher’s p-value and chose the minimum number of -/- mutants that yielded a significant

p-value (0.05).

Therefore, we needed to examine at least 8 -/- embryos for each of the control and FA- supplemented diets to obtain a statistically significant difference for the effect on NTDs.

Because of a 1:2:1 ratio, the total number of embryos required was (8)(4)=32.

Results

Effect of maternal FA supplementation on embryonic NTDs. The availability of

mouse models for NTDs whose phenotype could be corrected by FA supplementation is

necessary for understanding the mechanism by which FA exerts this beneficial effect. To

determine the response of the two NTD models Apob and Lp, we generated timed

pregnancies supplemented with FA before and during conception. We dissected out the

embryos between E12.5-E14.5, the period after which the neural tube has completely

developed under normal conditions in mice. A chi-square statistic test was used to

calculate the number of homozygous embryos required to observe a significant difference

in NTDs between the two FA diets. This number was calculated based on a p-value of

less than 0.05 and on the penetrance of the phenotype in each of these mutants in the

literature (Greene et al. 1998; Homanics et al. 1995).

1. Apob mutant.

We began by testing whether the 10 ppm FA reduced the incidence of NTDs in Apob-/-

embryos. We also monitored the numbers of resorptions and the Mendelian ratio of

genotypes as a measure of embryonic loss. The penetrance of exencephaly in Apob-/-

embryos was determined to be 30% (Homanics et al. 1995). Based on that, we needed to

look at 144 embryos with each of the two FA concentrations to achieve significance.

However, our Apob mutant colony showed about 96% penetrance of the phenotype.

Thus, a total of 204 embryos collected from Apob+/- pregnant females maintained on either the 2 ppm (99 embryos) or 10 ppm (105 embryos) FA diet were examined. A slight but non-significant decrease in the frequency of resorptions out of total conceptions was

observed (Table 2.1). There was no decrease in the frequency of exencephaly seen in

Apob-/- embryos with the FA-supplemented diet compared to the control diet (Table 2.2).

Thus, the NTD in the Apob mutant is resistant to maternal FA supplementation.

Table 2.1 Apob: effect of maternal FA supplementation on occurrence of resorptions

# observed* Not Diet resorbed Resorbed P-value** 2ppm 99 14 NS 10ppm 105 9 *Total implantations visible at E13.5 **Chi-square test of independence shows that the association between FA concentration and resorption is not statistically significant

Table 2.2 Effect of maternal FA supplementation on incidence of exencephaly in Apob-/- embryos

# observed* Diet No exencephaly Exencephaly P-value** 2ppm 1 26 NS 10ppm 1 26 * Total Apob-/- embryos collected at E13.5 **Fisher’s exact test shows that the association between FA concentration and exencephaly phenotype is not statistically significant

2. Lp mutant.

We next tested whether FA supplementation reduced the NTD or looped tail incidence in the Lp-/- or Lp+/- embryos, respectively. The numbers of resorptions and the Mendelian

ratio of genotypes were also examined to determine embryonic loss. The penetrance of

both craniorachischisis and looped tail in the Lp-/-and Lp+/- embryos respectively is 100%.

A total of 135 embryos collected from Lp +/- pregnant females maintained on either the 2 ppm (69 embryos) or 10 ppm (66 embryos) FA diet were examined. There was no

significant decrease in the frequency of resorptions out of the total conceptions with FA

supplementation (Table 2.3). We also found no significant decrease in the frequency of

either craniorachischisis (Table 2.4) or looped tail (Table 2.5) in their respective genotype

groups on the supplemented diet compared to the control one. Thus like Apob, the NTD

in the Lp mutant is resistant to maternal FA supplementation.

Table 2.3 Lp: effect of maternal FA supplementation on occurrence of resorptions

# observed* Not Diet resorbed Resorbed P-value** 2ppm 69 7 NS 10ppm 66 7 *Total implantations visible at E13.5 **Chi-square test of independence shows that the association between FA concentration and resorption is not statistically significant

Table 2.4 Effect of maternal FA supplementation on incidence of craniorachischisis in Lp-/- embryos

# observed* No Diet Craniorachischisis Craniorachischisis P-value** 2ppm 0 15 NS 10ppm 0 11 *Total Lp-/- embryos collected at E13.5 **Fisher’s exact test shows that the association between FA concentration and craniorachischisis phenotype is not statistically significant

Table 2.5 Effect of maternal FA supplementation on incidence of looped tail in Lp+/- embryos

# observed* No Diet looped tail Looped tail P-value** 2ppm 1 40 NS 10ppm 1 31 *Total Lp+/- embryos collected at E13.5 **Fisher’s exact test shows that the association between FA concentration and looped tail phenotype is not statistically significant

Effect of maternal FA supplementation on viability of embryos. The distribution of

genotypes in both Apob and Lp mutants followed the expected Mendelian ratio of 1:2:1 with the control diet (Tables 2.6 & 2.7). Interestingly, with FA supplementation, we noticed a significant deviation from the 1:2:1 expected Mendelian ratio of genotypes for the Apob mutant with a decrease in the number of homozygous and more significantly the number of heterozygous embryos (Table 2.6). In the supplemented Lp mutant, although the chi-square statistic suggested no significant deviation from the expected 1:2:1 ratio of genotypes, it is obvious that the observed number of Lp-/- and Lp+/- embryos is reduced relative to expectation with the percent loss comparable to results of Apob (Table 2.7).

These results suggest that FA supplementation not only causes loss of NTD affected embryos (homozygous), but also of heterozygous embryos that would otherwise be unaffected and viable. Since the loss occurs in both homozygous and heterozygous embryos, it is not related to the NTD phenotype but suggests that the mutation in each mutant sensitizes to FA-induced embryonic lethality. This embryonic loss most likely occurs pre-implantation as the number of resorptions observed at E13.5 (Table 2.1 & 2.3) does not account for the number of lost embryos. Moreover, the average litter size in both supplemented mutants compared to the control diet does not decrease as would be expected due to the observed embryonic loss. Thus, preferential implantation of wild- type embryos compensating for the heterozygous and homozygous losses to achieve an optimal litter size is suggested.

Table 2.6 Genotype distribution of observed Apob embryos

# observed Average P-value* % Apob+/- % Apob-/- litter size Diet +/+ +/- -/- (χ2) lost** lost** X (n;L,H)***

2ppm 24 48 27 NS 0% 0% 6.18 (16;1,9)

10ppm 40 38 27 =0.003 52.5% 32.5% 5.83 (18;1,9) (11.23) *null hypothesis follows mendelian ratio 1:2:1; χ2: chi-square test statistic **+/+ is control number for 1:2:1 ratio; +/- and -/- expected numbers calculated based on that, ex. 1:2:1=40:80:40 (10ppm); % lost calculated based on difference between expected and observed numbers ***X :mean litter size; n:number of litters examined; L:smallest litter size observed; H:biggest litter size observed

Table 2.7 Genotype distribution of observed Lp embryos

# observed Average P-value* % Lp+/- % Lp-/- litter size Diet +/+ +/- -/- (χ2) lost** lost** X (n;L,H)***

2ppm 13 41 15 NS 0% 0% 4.31 (16;2,7)

10ppm 23 32 11 =0.1 30.4% 52.2% 4.42 (15;1,7) (4.4) *null hypothesis follows mendelian ratio 1:2:1; χ2: chi-square test statistic **+/+ is control number for 1:2:1 ratio; +/- and -/- expected numbers calculated based on that, ex. 1:2:1=40:80:40 (10ppm); % lost calculated based on difference between expected and observed numbers ***X :mean litter size; n:number of litters examined; L:smallest litter size observed; H:biggest litter size observed

Effect of maternal FA supplementation on gender specific prevalence of NTDs and on

preferential gender viability. Based on reports in the human literature of female

prevalence of NTDs (Whiteman et al. 2000), we tested for a significant difference in the

gender distribution of NTDs (in Apob-/- and Lp-/-) and looped tail defect (in Lp+/-) on the

control FA level. We also examined the gender distribution of Apob and Lp embryos on

the 10 ppm diet in both the homozygous and heterozygous genotype groups to determine

if the loss that was previously seen within these groups (Table 2.6 & Table 2.7) is due to

a preferential gender loss.

For Apob however, the chi-square p-values show no gender bias in the frequency of

exencephaly on either concentration of FA (Table 2.8). Supplementation also did not

result in a preferential gender loss of Apob+/- or Apob-/- embryos (Table 2.8). By contrast, we found that the Apob+/- embryos showed a deviation from the 1:1 female to male ratio

with significantly higher male prevalence on the 2 ppm FA diet (Table 2.8).

Table 2.8 Gender specific distribution of exencephaly frequency and embryonic loss in Apob

For the Lp mutant, the craniorachischisis and looped tail phenotypes that are associated with the Lp-/- and Lp+/- embryos, respectively, were equally represented in males and females on the 2 ppm FA diet (Table 2.9). Moreover, FA supplementation resulted in a significantly higher loss of heterozygous females rather than males, a trend that was also observed in the supplemented homozygous embryos despite its non statistical significance in this case (Table 2.9).

Table 2.9 Gender specific distribution of craniorachischisis and looped tail frequency and embryonic loss in Lp

Effect of maternal FA supplementation on embryonic skeletal development. We had suspected that in the case where the Apob and the Lp mutants would be responsive to FA supplementation, the effect might not be a complete correction of the phenotype but rather a transformation from a severe to a subtle defect that is less obvious to the naked eye. Examining the fetal skeleton in this case would allow a test for more subtle defects.

Thus, the E12.5-14.5 embryos whose phenotypes were examined were also stained with

Alcian Blue for cartilaginous examination. These mutants however, turned out to be resistant with no correction of the phenotype due to FA supplementation. We still

examined the skeletons to identify any changes that are not big enough to correct the

phenotype, but found none attributed to the higher level of FA.

Discussion

NTDs are common birth defect with an incidence of 1 in 1000 live births in humans

(Northrup and Volcik 2000). FA supplementation of maternal diet before and during early pregnancy decreases the risk of this defect significantly (1991; Antony and Hansen

2000). To better understand the mechanism by which NTDs arise and by which FA exerts its effect and to provide a means for identification of maternal markers to identify individuals with alternative response to FA supplementation, well characterized mouse models with known etiology of the defect and an identified response to FA supplementation need to be studied. Among more than 190 NTD mouse models and strains, only 10 have been tested for the effect of maternal FA supplementation on the outcome of embryonic NTDs, which has shown to be effective in some mutants but not others. Thus, we chose to conduct a supplementation study on an additional three mutants, the Apob, Lp, and Lrp6 mutants displaying exencephaly, craniorachischisis, and a combination of exencephaly and spina bifida ,respectively.

Maternal FA supplementation did not show a protective effect on the exencephaly observed in Apob-/- embryos, or on the craniorachischisis or looped tail phenotypes

observed in Lp-/- and Lp+/- embryos, respectively. The NTDs in Lrp6-/- embryos were also

not corrected by maternal FA supplementation as discussed in Chapter 4. Embryos at

12.5-14.5 days post coitum (dpc) subjected to the maternal control diet (2 ppm FA)

showed no significant difference in the frequency of the defects compared to the supplemented diet group with 5 times more FA. Thus, these three mutants represent NTD mouse models of FA-resistance with respect to neural tube development.

The higher FA concentration, interestingly, suggested a negative effect characterized by an increased pre-implantation loss of homozygous and heterozygous embryos in both mutants that we tested (the Apob and Lp mutants) as well as in the Lrp6 mutant (refer to

Chapter 4). There is no mention in the literature of this effect in any of the 10 NTD mutants that were previously tested with nutrient supplementations. A concern was raised of whether we were supplementing the mice with a concentration of FA that exceeded its equivalent in humans thus resulting in embryonic loss. However, based on some recommendations, supplemental FA concentrations in humans can reach up to 5 mg/day

(1999; Antony 2007; Ouma et al. 2006), which is about 12 times higher than the minimal requirement of 400 µg/day. Due to a difference in metabolic rates between mice and humans, the minimal requirement for dietary FA in mice is 2 ppm. Thus, the supplemental FA concentration provides 5 times more the minimal requirement, not exceeding the supplementation equivalent in humans. With this clarification, the findings suggest a negative selection for homozygous and heterozygous implantations which are compensated for by increased wild-type implantations with FA supplementation. Thus a fixed litter size is maintained on both FA levels despite the significant embryonic losses on the higher FA diet. In humans the decrease in the risk of NTDs is determined based on the observed total conceptions, since specific genotypes causing this defect are not well known and cannot be distinguished from other genotypes. Thus, a loss of pre/peri-

implantation embryos due to a specific genotype cannot be identified raising the concern

of whether the decrease in the risk of NTDs with FA supplementation in humans is

actually due to an increased loss of affected embryos rather than a correction of the NTD

phenotype. Due to the loss of heterozygous embryos, in certain cases with a higher extent

than homozygous losses, which we have found with our study, another concern is raised.

This concern that should be highly considered in humans is whether FA supplementation

that has been so far assumed beneficial may be causing the loss of embryos that are

otherwise healthy and viable.

We showed that although the loss of heterozygous and homozygous embryos that was

observed with FA supplementation was equally contributed by females and males in the

Apob mutant, the Lp mutant interestingly showed a preferential loss of females in both

genotype groups indicating that the negative effect of FA supplementation is more

pronounced in female embryos in certain cases.

Based on reports in the human literature for gender prevalence of NTDs (Whiteman et al.

2000), preferential occurrence of the defect in females was also examined in the Apob

and Lp mutants. However, neither mutant on the 2 ppm control diet was characterized by a higher frequency of NTDs in female embryos (exenephaly in Apob-/-; craniorachischisis

and looped tail in Lp-/- and Lp+/- respectively). Thus, FA supplementation does not affect

gender prevalence of NTDs in these two mutants.

Overall, the Apob, Lp, and Lrp6 mutants represent models of FA-resistance with respect

to correcting the neural tube defect. However, these mutants display an unexpected

response to maternal FA supplementation characterized by embryonic loss, raising a serious concern of similar results in humans.

CHAPTER 3

Maternal gene expression and metabolite profiling to identify biomarkers of fetal

response to folic acid supplementation

The work described in this chapter is a collaborative effort among many investigators.

The husbandry, genotyping, and RNA extraction for the Lrp6 mutant were performed by the laboratory of Dr. Margaret Elizabeth Ross at Cornell University (referred to hereafter as the “Ross lab”). RNA hybridizations to Affymetrix microarrays were performed by the

Gene Expression and Genotyping Core Facility at the Case Comprehensive Cancer

Center. Dr. Donald Jacobsen at the Cleveland Clinic Foundation performed the erythrocyte folate and vitamin B12 assays. All other work was performed by the candidate.

Abstract

Despite public health recommendations for folic acid (FA) supplementation, about 70% of women are non-compliant, subjecting their fetuses to an otherwise preventable risk of neural tube defects (NTDs). Even in compliant cases, 30-50% of women do not show a decreased risk of NTD affected pregnancies and could avoid the intake of unfavorably high levels of FA that contribute to some adverse health effects. Thus, if fetal NTD response to FA could be predicted based on maternal genetic factors and biomarkers, treatments could be targeted specifically to responsive women, perhaps improving their compliance and more efficiently decreasing the risk of NTDs. With six known responsive and resistant NTD mutants, we analyzed whole genome expression profiles and erythrocyte levels of folate and vitamin B12 to identify maternal markers of fetal FA response. Metabolite analysis did not identify markers that reliably distinguished NTD- responsive from -resistant mutants. However, we identified two distinctive gene expression profiles; each profile, along with the biological pathways it defined, uniquely characterized all mutants within a FA-response group. Interestingly, the six NTD responsive and resistant mutants collectively influenced pathways related to the function of the two nuclear receptors, retinoid X receptor (RXR) and pregnane X receptor (PXR), suggesting that these pathways may be related to the mechanism of NTD development.

These findings combined suggest that although FA-responsiveness and –resistance may be achieved through idiosyncratic mechanisms, the pathogenesis of NTDs may be mediated through a shared mechanism among all six mutants.

Introduction

Since the implementation of FA food fortification for the general population, a 19-49% reduction in the incidence of NTDs has been observed (Rader and Schneeman 2006; Ray et al. 2002; Stevenson et al. 2000). Moreover, public health organizations recommend maternal FA supplementation in the form of pills to decrease the risk of NTD affected pregnancies by up to 70% (1991; Antony and Hansen 2000). However, compliance with the recommended preventive supplementation is relatively poor with only about 33% of

U.S. women between the age of 18 and 45 years regularly using vitamins containing FA

(1999; CDC 2005). As a result, noncompliant individuals who are potentially FA- responsive subject their fetuses to an otherwise preventable risk of NTDs. Thus, groups concerned with prevention of NTDs argue for higher food fortification with FA (Johnston

2008).

Although FA supplementation is beneficial in many cases, high levels can lead to unfavorable health outcomes (Brouwer et al. 2001; Cole et al. 2007; Dzinjalamala et al.

2005; Kelly et al. 1997; Kim 2003; Lamers et al. 2004; Ouma et al. 2006). To prevent the intake of unnecessarily high levels of FA in the general population and in the approximate 30% of women who are resistant to FA, and to increase compliance in responsive cases, individualized treatment plans are required. Development of these plans could be achieved by conducting genetic and biochemical studies to identify maternal markers that could predict response of fetal NTDs to maternal FA supplementation. The focus on maternal and not fetal markers is due to the fact that several maternal factors such as diabetes, obesity, elevated serum HCY and low serum FA levels are associated

with an increased risk of fetal NTDs, and maternal supplementation with the vitamins FA and B12 tend to decrease the fetal risk for this defect (Detrait et al. 2005; Hague 2003; van Guldener and Stehouwer 2001), indicating a relationship between maternal metabolic imbalances and fetal defects. Moreover, the initiation of maternal FA supplementation is necessary before conception to ensure its effectiveness. In addition to identifying FA- response predictive profiles, these studies could aid in understanding the mechanisms of

NTD pathogenesis and FA-response.

NTD mouse models can be used to identify candidate maternal markers for fetal NTD response. These mouse mutants, in addition to having a mechanism of neural tube closure fundamentally similar to that in humans (Juriloff and Harris 1998), provide models for both responsiveness and resistance to FA supplementation. Previous studies showed that gene expression and metabolite surveys in mice can be used to develop profiles that are diagnostic for particular disease conditions or treatment responses (Ernest et al. 2006).

Ernest et al. compared expression profiles for several single gene mutants that have phenotypes similar to those associated with anomalies in folate-HCY metabolism in humans. Analysis of the expression profiles revealed two clusters, one for mutants whose phenotypic defects are FA-responsive and the other for mutants whose response to FA had not been tested. Mutants in the latter group were predicted to either be resistant to FA or respond in a functionally different manner than the mutants in the responsive group.

Interestingly, FA treatment of Ski deficient mice, one of the unknown response mutants, showed that their NTD was resistant as predicted by the gene expression profiles (Ernest et al. 2006). Another unknown response mutant that clustered with the Ski mutant, the

Apobtm1Unc mutant (referred to as the Apob mutant), was identified as part of the present work to be indeed resistant to FA (see Chapter 2). These results verify the predictions

made by Ernest et al.’s gene expression profiles for the mutants she studied and support

the feasibility of finding maternal markers of fetal NTD response to FA supplementation.

The two NTD mutants, Pax3Sp2H and Lrp6Cd (referred to as the Sp and Cd mutant,

respectively) are FA-responsive (Carter et al. 1999; Fleming and Copp 1998), whereas

another NTD mutant, curly tail (referred to as ct), is FA-resistant (Greene and Copp

1997; Tran et al. 2002; van Straaten et al. 1995). With these three mutants and an additional three (Lrp6KO, Apobtm1Unc, and Vangl2Lp- referred to as the Lrp6, Apob, and Lp

mutant, respectively) that were shown, in Chapter 2, to be resistant to FA

supplementation, whole genome expression and metabolite profiles were analyzed to

identify maternal markers (genes, pathways, and metabolites) for fetal NTD response.

Heterozygous females representing a group at risk for a NTD-affected pregnancy (when mated to a heterozygous male) were chosen for this study. Because, in humans, maternal markers for fetal NTD response should be identified before the time of conception and supplementation, the mouse experiments were conducted in females at the age of reproductive maturity but prior to pregnancy and after being maintained on a non-FA-

supplemented diet.

Our gene expression analysis identified two pathways that were significantly associated

with all six NTD mutants regardless of their response to FA supplementation. The first

pathway is involved in lipopolysaccharide/interleukin-1 (LPS/IL-1) inhibition of retinoid

X receptor (RXR) function. RXR is a nuclear receptor that heterodimerizes to bind retinoic acid (RA), which then binds to retinoic acid response elements (RAREs) in the regulatory regions of target genes such as Hox genes (Holland 2007). RA activates gene transcription ultimately controlling anterior/posterior patterning in the early stages of embryonic development (Allenby et al. 1993; Germain et al. 2006; Marletaz et al. 2006).

The second pathway involves the pregnane X receptor (PXR), another nuclear receptor that heterodimerizes with RXR upon activation with a ligand (Kliewer et al. 2002).

Activated PXR regulates certain genes involved in the clearance of harmful compounds from the body (Kliewer et al. 2002); these accumulated toxic compounds may be involved in the pathogenesis of NTDs. These two pathways may therefore represent a shared mechanism of NTD development at least among the six mutants tested.

Next we tested whether metabolite, gene expression, or pathway profiles can predict response to FA supplementation in the six studied mutants. Differences in the intracellular levels of folate and vitamin B12 were heterogeneous among the mutants and their controls, but with no obvious pattern to distinguish between the two FA-response types. However, we identified two distinctive gene expression profiles; one profile was unique to the FA-responsive mutants (Sp and Cd) whereas the other was unique to the

FA-resistant mutants (Lrp6, Apob, Lp, and ct). Moreover, the biological pathways that defined the genes in one response-specific expression profile were exclusive of the pathways defining the other gene expression profile. Together these results raise the possibility that the mechanism of FA-response is not shared between responsive and resistant mutants.

Materials and Methods

Mouse colonies and diets. The Sp and Apob mutants along with their control

C57BL/6J, in addition to the Lp and ct mutants with CBA (control for the ct mutant), were purchased from the Jackson Laboratory. The Cd mutant was imported from the

Ross lab as part of an ongoing collaboration. The Lrp6 mutant was part of the same

collaboration; however these Lrp6 mutant mice were bred and genotyped at Cornell

University and only RNA samples from this mutant were sent to us for microarray hybridization and analysis. All mice were raised on the PMI Nutrition Laboratory

Autoclavable Rodent Diet #5010 and maintained by trio matings. Wild-type females and heterozygous males for each of the Sp, Cd, Lrp6, Apob and Lp mutants, were then weaned at 3 weeks of age and maintained on a control diet containing 2 ppm FA

(D05072702, Research Diets- Appendix 1) for at least 3 weeks before they were mated.

For the ct strain that is maintained as a random bred closed colony in which all individuals are considered to be ct/ct homozygotes (van Straaten and Copp 2001), these males were mated with females from the CBA strain, the control for ct (Cockroft et al.

1992). Heterozygous (test group) and wild-type (control group) female pups from these trio matings were also maintained on the same diet (since conception) and used for this study at 6-8 weeks, the age of reproductive maturity (Figure 3.1). All mice shared the same animal room with controlled temperature, humidity, and 12 hour light-dark cycle.

Mice were provided food and water ad libitum except in the case of the study mice that were fasted 4-5 hours before dissection.

Figure 3.1 Weaning and breeding timeline Breeding pairs were weaned at 3 weeks of age on the 2 ppm FA diet. They were kept on the same diet until they were mated at 6 weeks of age. Female offspring from these matings were weaned in turn at 3weeks of age and maintained on the same diet until they were dissected at reproductive maturity. Liver samples were collected for gene expression analysis from four females of each specific genotype. Blood was collected from a separate set of 12 mice per genotype.

Genotyping. Genotyping for Apob was done according to the protocol provided by the

Jackson Laboratory. The Lp genotyping protocol was previously described (Copp et al.

1994). Cd was genotyped according to the protocol provided by the Ross Lab at Cornell

University, using the primers listed in Appendix 3. White spotting on the belly and occasionally on the back, feet, and tail were used to distinguish heterozygous Sp mice from their wild-type siblings (Epstein et al. 1993).

Sample collection. At 6-8 weeks of age, four females per genotype group for each of the mutants (Table 3.1) were fasted for 4-5 hours (to match metabolite assay condition) and were then anesthetized with isoflurane to avoid the use of injected anesthetic drugs that may affect liver function (Hedenqvist 2008). Deeply anesthetized mice were dissected and livers collected for microarray experiments, after which they were euthanized. Liver was used for this study because it is a major site of folate-HCY

metabolism studies (Finkelstein 1998; Finkelstein et al. 1990) and because our lab had

previously studied gene expression profiling in liver (Ernest et al. 2006). Because erythrocytes are a good indication of the tissue load of folate (Bailey 1990), blood collection for folate and vitamin B12 assays was done by heart puncture in a different set of 9-15 females per genotype group following the same conditions as liver dissection.

Table 3.1 Test and control mice used for experiments The test group includes heterozygous female mice for all the mutants except the ct mutant where the homozygotes are used. The control group includes the genetic background strain as a control for mutants with a pure genetic background (C57BL/6J for the Sp and Apob mutants). CBA/CaJ is used as the control for the ct mutant. However, in the case of mutants that have a mixed genetic background (Cd, Lrp6, and Lp mutants), the wild-type littermates are used as controls. For the gene expression experiment, 4 biological replicates were used for each of the test and control groups of each mutant. 9-15 biological replicates were used for the metabolite assays.

Mutant Test group Control group (genotype) (genotype/strain) Splotch (Sp) Sp+/- Sp+/+ Responsive Crooked tail (Cd) Cd+/- Cd+/+ Lrp6 Lrp6+/- Lrp6+/+ Apob Apob+/- Apob+/+ Resistant Loop tail (Lp) Lp+/- Lp+/+ Curly tail (ct) ct/ct CBA/CaJ

RNA isolation and microarray hybridization. Approximately 0.5 cm3 of liver was

excised from each of the experimental female mice (Table 3.1) and placed in RNA-Later

solution (Ambion) at 4ºC overnight for RNA stabilization. The Qiagen RNeasy Mini Kit

was used to isolate total RNA from liver. This method included a DNAse treatment to

digest genomic DNA. The RNA samples were hybridized to the Affymetrix mouse 430

2.0 expression array that includes over 39,000 transcripts. Procedures were carried out

according to Affymetrix protocols for single round amplifications.

Normalization of microarray data. The MAS 5.0 algorithm was used for background

adjustment of the microarray data. With this method, mismatch probes are utilized to

adjust the perfect match (PM) intensity. Linear scaling of the feature level intensity

values, using the trimmed mean, is the default to make the means equal for all arrays

being analyzed (Irizarry et al. 2003a).

Gene expression analysis. To identify the differentially expressed genes for each test

mutant and its wild-type control, we used Variance-Modeled Posterior Inference with

Regional Exponentials (VAMPIRE), which is a web-based microarray analysis tool that uses a robust Bayesian approach (Hsiao et al. 2005). Statistical significance in VAMPIRE for a certain gene depends on the variance of that gene in samples within a group, which should be lower than the variance between two groups. In addition, those genes that are more highly expressed than others are considered more significant. We chose a false discovery rate (FDR) threshold of 0.05 to determine significance. The genes that were identified as significant for each of the six mutants thus represented the effect of the NTD causing mutation independent of any genetic background effect.

Pathway analysis. Each of the gene lists identified with VAMPIRE for each of the six mutants, was interpreted biologically by associating them with canonical pathways using

Ingenuity Pathways Analysis (IPA) software (Ingenuity Systems®, www.ingenuity.com).

IPA associates genes within a dataset to a manually currated database of canonical

pathways and gene networks. The significance of the association between the dataset and

a canonical pathway is determined by calculating a ratio of the number of genes within

our dataset that map to a particular pathway divided by the total number of genes for that

specific pathway within IPA’s database. In addition, Fisher’s exact test is used to determine the probability that the association between the genes in the dataset and the canonical pathway is explained by chance alone. The p-value in this method is calculated by considering: 1) the number of genes within our dataset that participate in that pathway,

2) the total number of knowledge base genes known to be associated with that pathway,

3) the total number of genes within our dataset, and 4) the total number of genes in IPA’s database. In the right-tailed Fisher's exact test, only pathways that have significantly more differentially expressed genes than expected by chance are shown. In this discovery study to identify candidate pathways that warrant further investigation, pathways passing the p-value threshold of 0.05 were considered significantly associated with our dataset.

(We note that because genes can appear in several pathways, these tests are not independent and a simple correction for multiple hypothesis testing is overly conservative).

Erythrocyte folate and vitamin B12 assay. EDTA-anti-coagulated blood was collected and the hematocrit of each sample was determined. Then 100 μl of the anti-coagulated

blood was lysed with 2 ml of 0.2% ascorbic acid for 60-90 min at room temperature in

the dark. The hemolysates were then sent to the laboratory of Dr. Donald Jacobsen at the

Cleveland Clinic Foundation (CCF) where folate and vitamin B12 measures were

determined on 200 μl aliquot of the hemolysates using a competitive binding radioassay-

SimulTRAC Radioassay (MP Biomedicals in Orangeburg, NY).

Complete blood cell Counts (CBCs) and white blood cell (WBC) differentials. In humans, for a biomarker to be clinically useful, its detection in peripheral blood is favored. In anticipation of conducting gene expression studies on whole-blood for identifying FA-response predictive profiles in mice, we performed CBCs and WBC differentials on heterozygous mutant females and their wild-type controls.

The only cells in a whole-blood sample that would yield RNA are the leukocytes or

WBCs (Vu et al. 2004). Because different cell types have distinct gene expression profiles, a variation in the relative proportions of the different peripheral WBC types

(neutrophils, lymphocytes, monocytes, and eosinophils) between each mutant and its control might lead to a variable gene expression pattern that is not attributed to the NTD- causing mutation. To test for variability in the proportions of WBC types, the percentage of each of the four types present in the blood (percent differential) was compared between each of the six mutants and their controls using an unpaired t-test with Welch’s correction and a two-tailed p-value (using GraphPad Prism 3.03 for Windows, GraphPad

Software, San Diego California USA, www.graphpad.com).

The mice were maintained on the PMI Nutrition Laboratory Autoclavable Rodent Diet

#5010 until the blood was collected by retro-orbital sinus bleeding at 6-8 weeks of age

(matching age condition of gene expression study). Fifteen mice per genotype of each of

the six mouse strains were used for CBC (except in the case of the Lrp6 mutant where only eleven heterozygous and nine wild-type samples were available).

Results

Significant gene expression changes. To identify genes that differed significantly

between each of the six NTD mutants and their controls we used VAMPIRE and a FDR

significance threshold of 0.05. These six gene lists included 400, 218, 216, 2148, 390,

and 436 genes for the Sp, Cd, Lrp6, Apob, Lp, and ct mutants, respectively (Appendix 4).

The identified genes represent the effect of the NTD-causing mutation in each of the mutants on each of their particular inbred genetic background.

Pathways for response predictions. To biological interpret each of the six significant gene lists that were identified by VAMPIRE, we used IPA. IPA associates genes with

signaling as well as metabolic canonical pathways. There were 15, 11, 10, 21, 18, and 11

metabolic pathways, and 16, 13, 10, 63, 19, and 15 signaling pathways significantly

associated with the Sp, Cd, Lrp6, Apob, Lp, and ct mutants respectively (Appendix 5). To

determine whether these pathways can distinguish FA-responsive and –resistant mutants,

we divided the mutants into their respective response groups (responsive: Sp and Cd;

resistant: Lrp6, Apob, Lp, ct) and tested for pathways that were commonly associated

with all mutants within one FA-response group and not associated with any of the

mutants in the other group. None of the significant pathways showed this pattern.

Interestingly, two canonical signaling pathways showed a significant association with all

NTD mutants regardless of their response to dietary FA supplementation. The two

pathways were: lipopolysaccharide/Interleukin-1 (LPS/IL-1) mediated inhibition of

retinoid X receptor (RXR) function and pregnane X receptor (PXR)/RXR activation

(Figure 3.2). These findings suggest that the six tested mutants may have a common

mechanism of NTD pathogenesis rather than a common mechanism of response to FA.

In Chapter 2, we showed that substantial embryonic loss of Lrp6, Apob, and Lp mutants results from maternal FA supplementation. However, all other NTD mutants, including the three mutants in our study (Sp, Cd, and ct) that were previously tested by others for

FA-response did not show a similar embryonic loss. Thus, we analyzed the significant pathways for the six mutants (Appendix 5) to identify those that predict the risk of embryonic loss due to maternal FA supplementation regardless of NTD response. We grouped the mutants into an embryonic loss group and a non-embryonic loss group, and tested for metabolic and signaling pathways that were commonly associated with all mutants within one group while absent from all mutants within the other group. However, no such pattern was found, suggesting a variable mechanism for the effect of FA on embryonic viability.

Figure 3.2 The two signaling pathways commonly associated with all six NTD mutants The bars represent the –log(p-value) for the association of the specified pathway with the significant genes in each of the mutants. The horizontal red bar represents a p-value threshold of 0.05 (Fisher’s exact test p-value). All the bars that pass this threshold represent a significant association. FA-responsive mutants are labeled in red (Sp, Cd) and resistant mutants are labeled in black (Lrp6, Apob, Lp, and ct).

Gene lists for response predictions. To identify a FA response-predictive set of genes in the six mutants, we started with the lists of differentially expressed genes that were initially generated with VAMPIRE to characterize each of the mutants (Appendix 4). We divided the mutants into a responsive group (Sp and Cd) and a resistant group (Lrp6,

Apob, Lp, and ct). Only those genes that were differentially expressed among all mutants within a group (response-specific) were used for this analysis, resulting in a small number of 24 and 7 genes in the responsive and resistant groups, respectively (Appendix 6). We then aimed to identify those genes that were differentially expressed in both FA- responsive and –resistant groups but displayed an opposite direction of expression in one group compared to the other. However, no examples of this pattern were found and the response-specific lists were characterized by genes that were differentially expressed

among all mutants within a response group regardless of their direction of expression,

while not differentially expressed among any of the mutants within the other response

group (Appendix 6). The two unique response-specific gene expression profiles were then independently associated with biological pathways that were exclusive to each of the profiles (Figure 3.3). These findings suggest that different mechanisms of FA-response

exist between the responsive and resistant mutants and those mechanisms are affected by multiple pathways within each response group. Furthermore, the identification of two response-specific gene lists in the present analysis of NTD mutants and the findings by

Ernest et al. that a separate set of genes distinguishes FA-responsiveness and -resistance in mutants with phenotypes related to folate-HCY metabolism defects, emphasize the possibility that a broad survey of NTD mutants can lead to a reproducible set of maternal markers of fetal NTD response to FA.

Figure 3.3 Pathways defining the genes specific to each FA-response group Each bar represents the –log (p-value) for the association of a specific pathway with the genes within a response group. The red bar represents a p-value threshold of 0.05. All the bars that pass this threshold represent a significant association. The dark blue bars represent pathways associated with the genes specific to the FA-responsive group; light blue bars represent pathways associated with the genes specific to the resistant group.

Glutathione metabolism

Pentose phosphate pathway

Taurine &hypotaurine metabolism

Cyanoamino acid metabolism Responsive Responsive ( Sp

Celenoamino acid metabolism & Cd )

Retinol Metabolism

Purine metabolism

Urea cycle & metabolism of amino groups

Alanine & aspartate metabolism

Endrogen & estrogen metabolism ( Lrp6, Oxidative phosphorylation Resistant , Lp Apob Acute phase response signaling

Mitochondrial dysfunction , ct )

Sulfur metabolism

Erythrocyte folate and vitamin B12 measurements. Because folate and vitamin B12

are implicated in the response of elevated maternal plasma HCY levels to FA

supplementation (Hague 2003; van Guldener and Stehouwer 2001), and because

erythrocyte folate is a reflection of tissue stores (Bailey 1990), both erythrocyte folate

and vitamin B12 levels were analyzed in the six mutants to test for an association with

FA-response, where a defect in cellular uptake of folate could lead to resistance. To

reveal how each of the six mutations on the specific genetic background affected the

metabolite levels, each mutant was compared to its wild-type control using the unpaired

t-test with Welch’s correction and two-tailed p-value (using GraphPad Prism 3.03 for

Windows, GraphPad Software, San Diego California USA, www.graphpad.com) (Table

3.2). Within the responsive group, while the Sp mutation did not affect the levels of either folate or vitamin B12, the Cd mutation was associated with an increase in the levels of both metabolites. In the resistant group, the Apob mutation led to increased levels of both folate and vitamin B12. In the Lp mutant, no change from the wild-type state for either

metabolite was observed. The ct mutation although not affecting the level of folate,

caused an increase in vitamin B12 level. Because the intracellular levels of folate and

vitamin B12 did not show a consistent pattern among mutants of the same response

group, these metabolites do not serve as predictors of response. Finally, whenever a

difference between mutant and control was present, folate and vitamin B12 levels were

increased in the mutant. These findings suggest that the NTD is not a result of folate or

vitamin B12 depletion and the correction of this defect in certain cases is not directly

related to the levels of folate or vitamin B12.

Table 3.2 Comparison of metabolite levels in each mutant to its corresponding control (unpaired t-tests with Welch’s correction)

Erythrocyte folate (ng/ml) Comparison Mutant Control T-test (mutant vs control) mean SEM (N) mean SEM (N) (two-tailed p-value)* Sp+/- vs C57BL/6J 569.4 21.23 (15) 596.0 17.16 (15) NS** Responsive Cd+/- vs Cd+/+ 647.7 21.81 (15) 559.7 18.87 (15) 0.005 Lrp6+/- vs Lrp6+/+ 336.3 16.77 (11) 373.2 13.60 (9) NS Apob+/- vs C57BL/6J 737.2 28.14 (15) 596.0 17.16 (15) 0.0003 Resistant Lp+/- vs Lp+/+ 660.6 23.70 (15) 649.8 25.99 (15) NS ct vs CBA 637.0 25.27 (15) 584.7 25.41 (14) NS

*All significant levels indicate an increase in the mutant compared to the control **NS=Not Significant

Erythrocyte vitamin B12 (pg/ml) Comparison Mutant Control T-test (mutant vs control) mean SEM, N mean SEM, N (two-tailed p-value)* Sp+/- vs C57BL/6J 7396 588.7 (15) 8533 919.7 (15) NS** Responsive Cd+/- vs Cd+/+ 22070 1107 (15) 11390 721.2 (15) <0.0001 Lrp6+/- vs Lrp6+/+ 10690 717.5 (11) 12230 386.1 (9) NS Apob+/- vs C57BL/6J 29310 898.4 (15) 8533 919.7 (15) <0.0001 Resistant Lp+/- vs Lp+/+ 14120 868.0 (15) 13220 534.2 (15) NS ct vs CBA 18710 1572 (15) 9131 1327 (14) <0.0001

*All significant levels indicate an increase in the mutant compared to the control **NS=Not Significant

WBC differentials. With the goal of conducting gene expression studies on whole-blood where only the WBCs yield RNA (Vu et al. 2004), we needed to determine first whether any differential gene expression in a mutant is the result of the NTD-causing mutation or the result of variable proportions of the four different WBC types between the mutant and its control. We found no significant differences in the percent differentials for each of the four WBC types between any of the five mutants, Sp, Cd, Lrp6, Apob, Lp, and their corresponding controls (Appendix 7). For these five mutants, the mutation did not affect the proportions of WBC types. When whole-blood gene expression studies are carried out in these mutants as we had anticipated, any differential gene expression would be associated with the NTD-causing mutation. However, in the case of ct and its control

(CBA) a significant difference was observed for each of the neutrophils and the eosinophils. Thus any gene expression difference between these two strains should take into consideration the variable proportions of the two WBC types.

Discussion

Evidence exists that high levels of FA contribute to some adverse health effects in humans. In addition, the interesting finding in Chapter 2 that maternal FA supplementation leads to embryonic loss in certain NTD mutants raises another concern of this supplementation in humans. Identifying maternal markers of embryonic NTD response would eliminate the risk of high FA intake in the general population by limiting the treatments to responsive individuals. Thus, we analyzed whole-genome expression data, and erythrocyte folate and vitamin B12 data for a panel of two responsive (Sp and

Cd) and four resistant (Lrp6, Apob, Lp and ct) NTD mouse models. Differentially expressed genes as a result of the NTD-causing mutation in each of the six mutants were identified using VAMPIRE (Appendix 4). These genes that separately characterize each mutant were then associated with metabolic and signaling canonical pathways using IPA

(Appendix 5). Among the many IPA significant pathways, none could distinguish FA- responsiveness and -resistance, where pathways would be commonly associated with all mutants of one response type and not associated with any of the mutants of the other response type.

Findings of the FA supplementation study in Chapter 2 showed that although dietary FA did not correct the NTD phenotype of Lrp6, Apob, and Lp, supplementation still had an unusual and unexpected effect characterized by loss of many homozygous and heterozygous embryos of all three mutants. This effect was not reported in the literature for any other previously tested NTD mutant, including the 3 mutants used in this study

(Sp, Cd, and ct). However, we were unable to identify any pathways that distinguish the mutants affected by embryonic loss from those that were unaffected.

Interestingly, we found that two signaling pathways related to LPS/IL-1 mediated inhibition of RXR function and PXR/RXR activation were commonly associated with the differentially expressed genes in each of the six NTD mutants regardless of FA-response.

These pathways could thus be involved in NTD pathogenesis. Two nuclear receptors

RXR and RAR (retinoic acid receptor) bind together forming a heterodimer (Germain et al. 2006). RA, the oxidized form of vitamin A, functions by binding RAR/RXR dimers, in turn activating transcription of target genes that function in cell differentiation. Hox genes that control positioning along the anterior/posterior axis during early embryonic development are among the targets of RA (Holland 2007). RARα, RARβ, and RARγ are the three types of RAR (Germain et al. 2006). Spina bifida may be mediated through liganded RARα-RXR heterodimerization (Elmazar et al. 1997). Moreover, a double knockout for RARα and RARγ in addition to a mutant with reduced RAR activity shows exencephaly (Lohnes et al. 1994; Yao et al. 1998). These findings suggest that an inhibition of RXR function might lead to similar defects by hindering RAR/RXR heterodimerization and preventing the binding of RA. PXR, on the other hand, is another nuclear receptor that is activated by toxic substances including foreign chemicals or xenobiotics, steroids, and bile acids as part of the body’s defense mechanism (Kliewer et al. 2002). Once activated PXR heterodimerizes with RXR to bind hormone response elements eliciting the expression of certain genes involved in the clearance of the toxic substances from the body (Kliewer et al. 2002). Cytochrome P-450 monooxygenase 3A4

(CYP3A4), one of the primary targets of PXR activation, is responsible for the metabolism of xenobiotics including many clinical drugs (Lehmann et al. 1998). PXR also up-regulates glutathione-S-transferase (GST), which conjugates reduced glutathione, an end-product of the HCY transsulfuration pathway (Falkner et al. 2001; Finkelstein

1998). The significant association of PXR/RXR activation with all the NTD mutants might suggest its involvement in the clearance of an accumulated compound that results from an altered metabolic process related to NTD pathogenesis.

Although the pathways associated with each mutant’s differentially expressed genes could not distinguish between FA-responsiveness and –resistance, we identified two response-distinctive gene expression profiles. Each profile included genes that were differentially expressed among all mutants of the same response type, while not differentially expressed among any mutant of the other response type (Appendix 6).

Interestingly too, the biological pathways that associated with each response-specific gene profile were unique to that profile and not associated with the other response type

(Figure 3.3). These results suggest that the mechanism of FA-responsiveness and - resistance does not involve a shared biological pathway.

Because FA and vitamin B12 are beneficial in cases of elevated maternal plasma HCY levels, which are in turn correlated with an increased risk of NTDs (Hague 2003; van

Guldener and Stehouwer 2001), and because the metabolic pathway of HCY requires folate and vitamin B12 as cofactors (Bottiglieri 2005), we thought that distinguishing mutants that benefited from FA supplementation from those that did not would be

feasible through the measurements of erythrocyte levels of folate and vitamin B12 as an

established measure of tissue status (Bailey 1990). However, the change in the level of

each of the two metabolites between each mutant and its control was not consistent

within a response group and none of the six mutants showed depletion in the levels of

either metabolite indicating that the defect and its correction are not directly related to

folate or vitamin B12 levels in these six NTD mutants (Table 3.1).

We conclude that although erythrocyte levels of either folate or vitamin B12 were not

indicative of the FA-response type in the six studied NTD mutants, response-specific

gene lists were identified using these six mutants. In addition, separate previously

studied gene expression profiles could predict FA-resistance of NTDs in the ski mutant

(Ernest et al. 2006) and the Apob mutant (present study). Thus, development of a classifier that accurately predicts FA-response in the training set of NTD mutants as well as in independent sets of variable NTD mutants is possible using a supervised classification method which requires a large sample size to initially train the classifier

(Allison et al. 2006; McShane et al. 2003; Pusztai and Hess 2004).

CHAPTER 4

Functional interactions between the LRP6 Wnt co-receptor and folate metabolism

The work presented in this chapter is the result of collaboration among many investigators. The laboratory of Dr. Margaret Elizabeth Ross (referred to as the “Ross lab” throughout the rest of this chapter) at Cornell University-New York performed all embryo viability and Wnt signaling cell assays. RNA hybridizations to Affymetrix microarrays were performed by the Gene Expression and Genotyping Core Facility at the

Case Comprehensive Cancer Center. Dr. J. Sunil Rao and Dr. Jean-Eudes Dazard from the department of epidemiology and biostatistics at CWRU performed the Bayesian

ANOVA analysis of the gene expression data. All other work was performed by the candidate.

Authors: Jason Gray*, Ghunwa A. Nakouzi*, Bozena Slowinska-Castaldo, Jean- Eudes Dazard, J. Sunil Rao, Joseph H. Nadeau, M. Elizabeth Ross (*Co-first authors)

Reference: Manuscript submitted

Abstract

A hypermorphic mutation of the low density lipoprotein receptor-related protein 6 (Lrp6),

one of the Wnt signaling requisite co-receptors, causes NTDs in the Crooked tail (Cd)

mouse. Dietary folic acid (FA) supplementation ameliorates this significant birth defect,

although Lrp6 is not a direct participant in folate transport or metabolism. Whether this connection to FA is idiosyncratic for the Lrp6Cd mutation or whether FA has a direct and

general influence on LRP6 function in Wnt signaling was sought using a mouse model

with a knockout mutation of the Lrp6 gene (referred to hereafter as the Lrp6 mutant). A

FA supplementation study showed that a FA-enriched (10 ppm FA) diet did not reduce

the occurrence of NTDs in Lrp6-/- embryos compared to a control diet (2 ppm FA).

However, a shift to earlier lethality was found in the Lrp6 homozygous embryos as well as an increase in pre-implantation loss of homozygous and, more interestingly, of otherwise viable and healthy heterozygous embryos. This unexpected detrimental effect of FA translates into a decreased occurrence of NTDs when total implantations, regardless of genotype, are examined. Moreover, gene expression analysis revealed an association between the Lrp6 knockout allele and FA-sensitive expression changes in several gene clusters relevant to cell cycle regulation and Wnt signaling. An in vivo assay revealed that the cell proliferation in the midbrain/hindbrain region was decreased in

Lrp6-/- embryos compared to wild-type siblings. However, FA supplementation increased

proliferation in the Lrp6+/+ neuroepithelium at E9.5, but was insufficient to improve cell

proliferation and embryonic viability in the Lrp6-/- embryos. FA levels also modulated the canonical Wnt response in NIH 3T3 cells grown in defined media. These data indicate that although FA is required for optimal Wnt signaling, FA over-supplementation by as little as two-fold attenuates LRP5/6-dependent canonical Wnt responses. Thus, it is likely that FA supplementation rescues defects in Lrp6Cd/Cd fetuses by normalizing hyperactive

Wnt activity, whereas in the Lrp6-/- mutant FA further attenuates residual Wnt signaling

leading to earlier lethality of the otherwise viable and non-viable homozygous embryos

as well as some otherwise viable heterozygous embryos. This study provides the first

evidence that FA supplementation directly impacts LRP6 function and Wnt signaling.

Introduction

NTDs have been associated with more than 190 mouse mutants and strains (Greene and

Copp 2005; Harris and Juriloff 2007). FA supplementation during pregnancy prevents

neurulation failure in several cases (Gray and Ross 2009; Harris and Juriloff 2007). Mice

with the spontaneous Crooked tail (Cd) mutation are a folate-responsive model of rostral

NTDs (exencephaly) (Carter et al. 1999). Cd results from a single-nucleotide point mutation in the low density lipoprotein receptor-related protein 6 (Lrp6) gene (Carter et

al. 2005). Lrp6 (Arrow in Drosophila) encodes a single pass transmembrane receptor

that, like its paralog Lrp5, is a co-receptor with Frizzled (Fz) in the canonical wingless

(Wnt) signaling pathway to activate ß-catenin-dependent TCF/LEF transcription. The

Lrp6Cd mutation replaces a highly conserved glycine with aspartate in the extracellular domain of Lrp6 and interferes with the ability of dickopf1 (Dkk1) to inhibit canonical

Wnt signaling, resulting in sustained elevation of cytosolic ß-catenin levels in the presence of Wnt. On the other hand, the loss of LRP6 function produces several birth defects including caudal axis truncation and limb deformities as well as exencephaly and spina bifida (Pinson et al. 2000). Moreover, a hypomorphic allele of Lrp6, caused by the point mutation ringelschwanz, is associated with spina bifida in otherwise viable homozygous mutant pups (Kokubu et al. 2004). Thus, both gain- and loss-of-function

mutations in Lrp6 can result in NTDs.

Prenatal dietary FA supplementation shifts the phenotype of Lrp6Cd/Cd embryos to increased embryonic viability and decreased NTD incidence, showing a normal

Mendelian distribution of genotypes at all FA levels (Carter et al. 1999). Moreover,

analyses of gene expression arrays and biochemical markers of folate-homocysteine

(folate-HCY) metabolism indicate a defect in intracellular FA utilization associated with

homozygosity for the Lrp6Cd mutation (Ernest et al. 2006). However, it is not yet known

whether this effect is specific to the Lrp6Cd allele, dependent on the Cd mouse strain

background, or whether FA supplementation influences the action of the Lrp6 gene itself.

In this study, the response of LRP6 deficient mice to FA supplementation was examined

to determine the impact of FA on LRP6 function, NTD incidence, and Wnt canonical

signaling. We show that although FA supplementation stimulates canonical Wnt signaling, this effect is inadequate to correct the neurulation defects in the Lrp6 mutant.

Materials and Methods

Animals. All procedures were carried out in accordance with the National Institutes of

Health Guide for the Care and Use of Laboratory Animals and were approved by the

Institutional Animal Care and Use Committee at Weill Medical College of Cornell

University. Gene-trap mice in which the Lrp6 locus was inactivated (Pinson et al. 2000),

have been backcrossed more than 12 generations to the C3H/HEJ background. Mice were

housed in climate-controlled Thoren units with a 12-hour light-dark cycle.

NTD Phenotyping. Mating pairs of Lrp6+/- mice were maintained on a defined diet containing 10 ppm or 2 ppm FA (Appendix 1) for two generations prior to tissue or embryo collection. Timed pregnancies were generated from Lrp6+/- x Lrp6+/-

intercrosses. Upon discovery of a plug, which is designated embryonic day 0.5 (E0.5) the

mating pairs were separated. Embryos from timed-pregnant females were harvested at or

near E13 and fixed in 4% paraformaldehyde to preserve them for characterization of

NTDs and histology. Implantations were scored by visual inspection as undergoing

resorption, early lethality (dead at harvest with development halted at least 24-48 hours

earlier), or as live embryos that displayed developmental staging comparable to wild-type

littermates and signs of active cardiovascular circulation at the time of collection.

Embryos were further scored as having one or more defects including spina bifida (open

caudal neural tube), exencephaly (open cranial folds), caudal axis truncation, or limb

deformities. Genotyping of embryos and postnatal pups was performed by PCR as

previously described using tissue from the embryonic yolk sac or tail snips, respectively

(Pinson et al. 2000). Sex determination was not performed on these embryos. Statistical

comparisons using the chi-square and Fisher’s exact tests, as appropriate, were performed using Microsoft Excel Software and GraphPad QuickCalcs Web site:

http://graphpad.com/quickcalcs/chisquared1.cfm and

http://graphpad.com/quickcalcs/contingency1.cfm (accessed April 2009).

Tissue collection and microarray. Approximately 0.5 cm3 of liver was excised from each of the experimental female mice and placed in RNA-Later solution (Ambion) at 4ºC overnight for RNA stabilization. The RNeasy Mini Kit (Qiagen), was used for isolation of total RNA from liver. This method included a DNAse treatment to digest genomic

DNA. The RNA samples were used to make cRNA probes that were hybridized to the

Affymetrix mouse 430 2.0 expression array that includes over 39,000 transcripts.

Procedures were conducted according to Affymetrix protocols for single round amplifications.

Normalization of data. Differential expression analysis consisted of two groups; the first included four of each of the Lrp6 wild-type and heterozygous animals treated with 2 ppm FA, whereas the second group was treated with a 10 ppm FA diet. The “Robust

Multichip Analysis” method (RMA) was used for background adjustment on the input data (Irizarry et al. 2003b). With this method, perfect match probe intensities were corrected by using a global model for the distribution of probe intensities (Irizarry et al.

2003b).

Diet-genotype interaction effects. The interaction effect of FA supplementation with

the Lrp6 variants was evaluated using a Bayesian ANOVA in BAMarray™ 2.0 software available at http://ora.ra.cwru.edu/bamarray/ (Ishwaran et al. 2006). BAM identifies

differentially expressed genes from multigroup high throughput microarray experiments.

BAM is robust to non-normality of gene expression measurements and to correlations between expression measurements on a given chip (Ishwaran and Rao 2003). BAM relies on a special type of inferential regularization (i.e. borrowing strength across the data) allowing it to balance the number of false detections against false non-detections, hence

detecting more true positive differentially expressed genes (Ishwaran and Rao 2003;

Ishwaran and Rao 2005). This is an ideal property guaranteeing lower total gene

misclassification. It reliably differs from current statistical methods that protect false

detection rates (Ishwaran and Rao 2005). Controlling false detection rates tends to

identify obviously varying genes but misses more subtle changes. For multigroup

designs, BAM adaptively reduces correlations between test statistics on a given gene,

enabling signals to be extracted from noise more efficiently, thus allowing true

differential gene expression patterns to be readily identified and reducing the number of

implausible patterns (Ishwaran and Rao 2005). An interaction effect was deemed

important not only if it met the BAM-derived level of significance, but also if at least one

of its corresponding main effects were also found to be significant. Based on this

analysis, diet-genotype interactions identified genes that were differentially expressed

between the wild-type and heterozygous states of Lrp6 and that differed between the 2

ppm and the 10 ppm FA diets.

Pathway and network analysis. Ingenuity Pathways Analysis (IPA) software

(Ingenuity Systems®, www.ingenuity.com), was used to identify biological pathways and networks defined by the 232 interaction significant genes. Affymetrix identifiers for these genes were uploaded into the software. Each gene identifier was mapped to its

corresponding gene object in the IPA knowledge base. Differentially expressed genes

were then mapped to the global molecular network that is developed from information in

the Ingenuity Knowledge Base.

Many of the genes that mapped to the global molecular network were associated with a

canonical pathway in the Ingenuity Knowledge Base and were thus eligible for canonical

pathway analysis. Significance of the association between the dataset and the canonical

pathway was measured in two ways: 1) a ratio of the number of differentially expressed

genes from the dataset that map to the pathway divided by the total number of all the

genes that exist in the canonical pathway, 2) Fisher’s exact test was used to calculate a p-

value for the probability that the association between the genes in the dataset and the

canonical pathway is explained by chance alone. The p-value in this method is calculated by considering: 1) the number of genes within our dataset that participate in that pathway,

2) the total number of knowledge base genes known to be associated with that pathway,

3) the total number of genes within our dataset, and 4) the total number of genes in IPA’s database. Only pathways that have significantly more genes than expected by chance passed the p-value threshold of 0.05 (Fisher’s exact test p-value) and were considered significantly associated with our dataset.

Because many of the pathways in IPA are not exclusively independent but show a certain

level of overlap in the involvement of genes, applying any multiple testing correction

would be overly conservative.

IPA also associated many differentially expressed genes with networks in the Ingenuity

Knowledge Base. Networks for these genes were then algorithmically generated based on

their connectivity. A score that was the negative log of the p-value of the right-tailed

Fisher’s exact test was assigned for each network. This score takes into account the number of eligible genes in our dataset and the size of the network to calculate the fit between each network and the genes in our dataset.

Immunohistochemistry. Embryos from Lrp6+/- x Lrp6+/- intercrosses were harvested at

E9.5 and the yolk sacs were collected for genotyping. Embryos were fixed in 4% paraformaldehyde overnight at 4ºC, and then transferred to 0.1M PBS prior to paraffin processing (Tissue Tek 2000, Miles Laboratories). Embedded tissues were sectioned coronally at 6 µm, mounted on adhesive-coated slides (Fisher Scientific), deparaffinized, and rehydrated. Antigen retrieval was performed with Reveal (BioCare Medical), and quenched in 3% hydrogen peroxide. Tissues were blocked in SNIPER (BioCare Medical,

BS966-BM) for 30 minutes at room temperature prior to incubation in anti- phosphohistone H3 (Upstate Biotechnology, 16-189, 1:1,000) overnight at 4ºC. Specific immunolabeling was visualized by incubation in secondary antibody conjugated to a fluorophore (Alexa Fluor 488, Molecular Probes, A-11070 1:500). Slides were cover-

slipped with DAPI mounting media (Vectastatin) and photographed on a Nikon Optiphot-

2 compound microscope fitted with a Spot Insight CCD camera (Diagnostic Instruments).

β-catenin stabilization assay. The levels of cytosolic β-catenin in Wnt-sitmulated cultured cells were measured as previously detailed (Carter et al. 2005; Giarre et al.

1998), with the following modifications. Briefly, FA powder (Sigma) was dissolved in

DMEM without FA (Specialty Media, Phillipsburg, NJ) and culture media was prepared with 10% calf serum+DMEM with specific FA concentrations (0, 1, 4, 10, 50, or 100 µg

FA/ml DMEM). Two hours after NIH 3T3 cells were seeded in a 6-well plate, growth media containing 10% calf serum+DMEM was replaced with culture media of specific

FA concentrations. Cells were cultured for 72 hours in FA specific media prior to assays, which were run as previously described (Giarre et al. 1998). Cultures were incubated in vehicle with or without 40 ng recombinant Wnt3a (R&D Systems, 1324-WN) for 2 hours. Following collection of the cytosolic fraction, cell lysates were normalized for protein concentration using the BCA Assay (Pierce, 23250) and Western-transferred to nitrocellulose membranes. Blots were blocked at room temperature with 1% non-fat dairy milk for 1 hour before incubation in primary antibody in block at 4ºC overnight with anti-ß-catenin (Santa Cruz, sc-7963 at 1:1,000), and anti-Erk2 (Santa Cruz, sc-154 at

1:5,000). Blots were washed 2x15 minutes in TBS w/ 0.05% Tween-20 before incubation with horseradish peroxidase-conjugated secondary antibodies for 1 hour at room temperature, subsequent washing in TBS, and detection (chemiluminescence reagents,

Pierce, pico- 34078 or femto- 34096). X-ray film exposures were scanned with an optical densitometer and analyzed using Quantity One software (Bio-Rad). Band densities of ß-

catenin were normalized for protein loading to the concentration of Erk2 present in the

lane.

Transfection and TopFlash reporter assay. A TCF/LEF in vitro reporter assay was

used to examine the effect of FA on Wnt-induced transcriptional activity. NIH3T3 cells

(100,000 cells per well of a 24-well plate) were plated in DMEM containing 0, 1, 4, 10,

50 or 100 μg/ml FA. After 3 days, cells were transfected with 500 ng of TopFlash or

FopFlash reporter plasmids using Lipofectamine per manufacturer’s protocol (BioRad,

#170-3350). Cells were allowed to recover for 36 hours in the FA-defined medium after

which, either 0 ng or 200 ng of recombinant Wnt3a (R&D Systems, 1324-WN) was

added per well and cultures were incubated another 8 hours prior to collection and lysis

of cells for luciferase reporter assay using a Microlumat Plus luminometer with

automated sample injection (EG&G Berthold, LB96V). Wnt activity (TopFlash) was

normalized to a Wnt-insensitive luciferase control (FopFlash) plasmid.

In situ assessment of canonical Wnt activity. A TCF/LEF-LacZ reporter mouse line

(Liu et al. 2003; Mohamed et al. 2004) was crossed with Lrp6 mutants and the progeny of

Lpr6+/-::TCF-LacZ double heterozygous parents were evaluated for canonical Wnt activity in the cranial folds at E9.5. Embryos were fixed with 4% paraformaldehyde for

30 minutes at 4ºC, then washed three times for 20 minutes in 0.1M PBS at 4ºC before

incubation in X-gal Wash Buffer (0.02% Igepal, 0.01% Na-Deoxycholate, 2mM MgCl2,

100 mM Na-Phosphate) for 5 minutes at room temperature, transfer to X-gal staining

solution (1 mg/ml X-Gal, 5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6 in X-Gal Wash Buffer)

100

and incubation at 37ºC. After 30 minutes, embryos were rinsed in 0.1M PBS, 3x5 minutes then 2x1 hour and post-fixed with 4% paraformaldehyde overnight at 4ºC.

Images were collected with Leica Applications Suite software running a Leica DFC310

FX camera fitted to a Lecia M165 FC stereomicroscope.

101

Results

FA supplementation and birth defects in the Lrp6 mutant. To determine the effect of

FA supplementation on the occurrence of birth defects in the Lrp6 mutant, two hundred and eighty seven embryos collected from dams fed diets containing 2 ppm FA (130 embryos) or 10 ppm FA (157 embryos) were examined and genotyped. Most deformities were observed in Lrp6-/- embryos, with only two instances of spina bifida occurring in

Lrp6+/- pups, one from each diet group. No instances of exencephaly, limb defects, or

axis truncation were found in any wild-type or Lrp6+/- embryos. Lrp6-/- embryos

exhibited at least two defects, and several were recovered that displayed three or more

defects. Two Lrp6-/- embryos from the 2 ppm diet group at E12.5-13.5 are shown in

Figure 4.1B, in which both display axis truncation and limb defects; the embryo on the

left shows spina bifida and the one on the right shows exencephaly. We tested the effect

of FA supplementation on the Lrp6 mutant embryos. Interestingly, a higher proportion of

Lrp6-/- mice succumbed in early embryogenesis (prior to E12) with 10 ppm versus 2 ppm

FA supplementation (Table 4.1). Among the viable Lrp6-/- embryos, FA supplementation

did not reduce the frequency of NTDs (Table 4.2). These findings raised the possibility

that FA supplementation increased the severity of developmental defects occurring before

E12 in Lrp6-/- embryos.

The distribution of embryonic genotypes as percent of conceptions (Table 4.3) shows that

while genotypes on the 2 ppm diet approach Mendelian expectations for Lrp6+/- x Lrp6+/-

intercrosses, there is a clear deficit of Lrp6-/-and interestingly of otherwise viable Lrp6+/-

genotypes conceived and maintained on a 10 ppm FA diet. Occurrence of resorptions was

102

similar for the 2 ppm and 10 ppm diets (Table 4.4). Nevertheless, the 10 ppm group from

Lrp6+/- x Lrp6+/- intercrosses displayed a higher resorption frequency than observed for

Lrp6+/+ x Lrp6+/+ intercrosses (approximately 10% at 10 ppm, Table 4.5), indicating that the fetal loss above 10% on the high FA diet was genotype-dependent. Although some resorptions could not be genotyped, those from which DNA could be recovered genotyped as homozygotes. Even if resorptions above the baseline10% are assumed to represent Lrp6-/- lost embryos and are included in the totals, the 10 ppm group still falls

short of Mendelian expectations with a high statistical significance (Table 4.3),

suggesting that some resorptions in the 10 ppm group occurred pre-implantation or too

early in embryogenesis to be detected at E12.5.

Given the striking deficiency of otherwise affected Lrp6-/- embryos due to supplementation, total conceptions maintained on a 10 ppm FA diet exhibited fewer

NTDs (exencephaly and spina bifida) and fewer morphogenic defects associated with

Lrp6 deficiency (limb malformation and axis truncation) when compared to total conceptions maintained on 2 ppm FA diet (Figure 4.1A). A similar effect of FA supplementation in humans may influence the assumption that the NTD is being corrected when FA is actually detrimental to the affected genotype group.

103

Figure 4.1 Incidence of NTD affected pregnancies in FA supplemented mice Observed implantations were scored as resorptions, embryos dead at harvest (early lethality), or live embryos with open cranial folds (exencephalic), open caudal neuropore (spina bifida), axis truncation, or limb defects, and graphed as a percentage of total observed implantations. (A) The incidence of each phenotype is shown as a percentage of all implantations, with individual embryos often displaying more than one defect. The incidence of NTDs was lower among embryos maintained on the high FA (10 ppm) diet. (B) Lrp6-/- embryos at E13.5 that were maintained on the 2 ppm FA diet demonstrate each of the phenotypes scored. The embryo at left shows tail truncation and limb defects (arrowheads) and spina bifida (arrows), while the one at right exhibits limb, trunk and exencephaly defects (arrows).

Table 4.1 Early lethality of FA-supplemented Lrp6-/- embryos

# observed* Viable Non-viable Diet embryos Embryos† P-value** 2ppm 18 2 0.003 10ppm 6 10 *Total Lrp6-/- embryos harvested at E13 **Fisher’s exact test shows that the association between FA concentration and early lethality is statistically significant †Dead at harvest with development halted at least 24-48 hours earlier

104

Table 4.2 Effect of maternal FA supplementation on incidence of NTDs in Lrp6-/- embryos

# observed* Diet No NTDs NTDs P-value** 2ppm 6 12 NS 10ppm 0 6 *Total Lrp6-/- viable embryos harvested at E13 **Fisher’s exact test shows that the association between FA concentration and NTDs is not statistically significant

Table 4.3 Genotype distribution of observed embryos from Lrp6+/- x Lrp6 +/- intercrosses

Average # observed % of Lrp6+/- % of Lrp6-/- litter size Diet +/+ +/- -/- P-value* lost** lost** X(n;L,H)*** 2ppm 37 73 20 0.04 0% 45.9% 7.94 (20;4,12) 10ppm 55 86 16 0.0001 21.8% 70.9% 8.45 (22;2,12) *null hypothesis follows Mendelian ratio 1:2:1 **+/+ is control number for 1:2:1 ratio; +/- and -/- expected numbers calculated based on that, ex. 1:2:1=55:110:55 (10ppm); % lost calculated based on difference between expected and observed numbers *** X :mean litter size; n:number of litters examined; L:smallest litter size observed; H:biggest litter size observed

Table 4.4 Effect of maternal FA supplementation on occurrence of resorptions

# observed* Not Diet resorbed Resorbed P-value** 2ppm 130 29 NS 10ppm 157 29 *Total implantations; non-resorptions include visible, viable and non-viable embryos **Chi-square test of independence shows that the association between FA concentration and resorption is not statistically significant

105

Table 4.5 Frequency of resorptions in Lrp6+/+ x Lrp6+/+ intercrosses maintained on 2 ppm vs. 10 ppm FA diet

Diet Litters Implantations Resorption collected observed rate 2ppm 9 71 18.3% 10ppm 8 71 9.9%

Gene and diet effects on mRNA profiles. Gene expression arrays were examined to

elucidate the impact of dietary FA on the Lrp6 genotype. Liver samples from adult wild-

type and Lrp6+/- female mice were used in these experiments for several reasons. First,

liver is the tissue where folate-HCY metabolism has been most comprehensively studied

(Finkelstein et al. 1990). Second, gene expression in an adult maternal tissue is relevant

to maternal-fetal interactions, because certain maternal metabolic features are established risk factors in the pathogenesis of NTDs (Detrait et al. 2005; Hague 2003; van Guldener and Stehouwer 2001). Third, gene expression patterns identified in adult heterozygous

Lrp6 deficient tissues may be relevant for screening assays designed to evaluate NTD risk and response to FA supplementation.

Nearly 40,000 sequences covering the mouse transcriptome were interrogated.

Significant diet-genotype interactions were identified for 232 genes. Interaction plots show how the significant genes with this interaction test co-vary in the wild-type and

heterozygous states of Lrp6 at the 2 and 10 ppm levels of FA (Appendix 8). For example,

Figure 4.2 shows the interaction plots for four genes belonging to the Wnt/β-catenin

signaling pathway where the expression of gap junction protein-alpha 1 (GJA1) on the 2

106

ppm diet is not different between the wild-type and the heterozygous states, but is more highly expressed in the wild-type state as compared to the heterozygous state on the 10 ppm diet. Retinoic acid receptor-gamma (RARG) by contrast, shows higher expression in wild-types relative to heterozygotes on the 2 ppm FA diet, but this relationship is reversed on the 10 ppm FA diet. On the 10 ppm diet expression of Smoothened homolog

(SMO) shows higher expression in heterozygotes compared to the wild-types, whereas that expression shows an opposite but slightly smaller change on the 2 ppm FA diet.

Wingless-type MMTV integration site family, member 7A (WNT7A) expression shows an opposite relationship to the one seen with SMO, here the expression is lower in the heterozygotes as compared to the wild-types on the 10 ppm diet and there is only a slight increase in expression in the heterozygotes on the 2 ppm diet relative to wild-types.

The 232 genes that showed diet-genotype interactions were analyzed with IPA. This analysis showed that the canonical pathways most significantly associated with our interaction dataset were the cell cycle: G1/S checkpoint regulation (p=0.0137) followed by the Wnt/β-catenin signaling (p=0.0162), indicating that the observed phenotype of

NTDs and response to FA supplementation is more likely explained by these pathways than others among the list, which includes aryl hydrocarbon receptor signaling, and nitrogen metabolism (Table 4.6).

IPA analysis also identified network interactions of genes whose expression was affected by Lrp6 gene dosage and FA concentration. The two most significant networks are displayed in Figure 4.3 with those genes that displayed co-variance with gene and diet

107

conditions indicated in red. Genes in the first network are involved in cellular proliferation and growth, and encode proteins included in the cell cycle-G1/S checkpoint

regulation pathway. The second network encompasses genes such as Disheveled 1

(Dvl1), Wnt, Wnt7a, which also overlap with the Wnt signaling pathway. The genes in

this network are involved in cellular differentiation, nervous system development and

function, and lipid metabolism. Wnt7a involvement in this context is particularly

interesting, as hypomorphic mutations in human WNT7A have been associated with

significant limb and caudal axis truncation defects that are highly reminiscent of defects

seen in the Lrp6-/- mutant mice (Woods et al. 2006). These findings suggest that the

effect of FA supplementation in the Lrp6 loss-of-function (LOF) mutant is functionally

connected to Wnt signaling and cell cycle regulation.

108

Figure 4.2 Diet-genotype interaction plots for genes significantly associated with the Wnt/β-catenin signaling pathway The X-axis represents the gene factor (GF) with the gene expression change from wild- type to heterozygous state shown from left to right. The intensity of the gene expression is represented by the scale on the Y-axis. The red dotted line represents the change in gene expression on the 2 ppm FA diet. The blue solid line represents the change in gene expression on the 10 ppm FA diet. The number in parentheses represents the significance rank of each gene among the 232 interaction significant genes.

109

Table 4.6 Five pathways most significantly associated with diet-genotype interaction dataset

Pathway P-value* Ratio** Genes***

Cell Cycle: G1/S Checkpoint Regulation 1.37E-02 5.17E-02 E2F6, SUV39H1, CDKN1B

Wnt/ β-catenin Signaling 1.62E-02 3.03E-02 GJA1, WNT7A, SMO, RARG

Aryl Hydrocarbon Receptor Signaling 3.94E-02 2.63E-02 TGM2, NRIP1, CDKN1B, RARG

Nitrogen Metabolism 5.21E-02 1.50E-02 ASRGL1, PTPRZ1

* Significance of association between the genes in the dataset and the canonical pathway (Fisher’s exact test p-value) ** Number of differentially expressed genes from the dataset that map to the pathway divided by the total number of all the genes that exist in the canonical pathway within IPA’s database ***Genes within our interaction dataset that significantly associated with the canonical pathways

110

Figure 4.3 The two networks most significantly associated with the diet-genotype interaction dataset Genes in red represent those within the diet-genotype interaction dataset that are associated with the network.

Network 1

111

Network 2

112

Effect of FA on cell proliferation in the Lrp6 deficient neural tube. Gene expression

profiling results indicate an important interaction between Lrp6 genotype and dietary FA that impacts cell cycle regulators and cell proliferation. A hypothesized influence of FA supplementation on NTD mechanisms is that added FA enhances cell division in the neural folds and perhaps underlying mesenchyme to promote neural tube closure. This could potentially correlate with altered Wnt signaling because this developmentally crucial pathway is known to promote cell proliferation. We therefore examined cell proliferation in the cranial neural tube of Lrp6+/+, Lrp6+/- and Lrp6-/- embryos by

quantifying the number of cells in M-phase in the cranial neural folds of E9.5 embryos

(Figure 4.4). Interestingly, FA supplementation significantly improved cell proliferation

in Lrp6+/+ embryos, but this effect was abolished in heterozygous and homozygous

mutant embryos on either the 2 ppm or 10 ppm FA diet relative to wild-type embryos on

the 2 ppm control diet, suggesting an interaction between gene dosage and FA. Because

the Lrp6 mutants showed increased embryonic lethality with higher FA supplementation,

this FA-induced improvement in cell proliferation seen in the wild-type embryos was

apparently insufficient to compensate for reduced LRP6 function in Lrp6 mutants.

Canonical Wnt signaling in the presence of defined FA levels. Gene expression

profiling indicated significant gene-diet interactions with Wnt signaling pathways. To

validate this observation, we next explored whether FA supplementation modulates

canonical responses to recombinant Wnt3a stimulation of NIH-3T3 cells in vitro. The

accumulation of cytosolic β-catenin was examined in cells treated with FA over

concentrations ranging from 0-10 μgFA/ml DMEM (Figure 4.5B). Inclusion of FA in the

113

defined medium did not alter the basal levels of β-catenin in the cytosol. Significantly,

FA depletion suppressed responses to Wnt stimulation, with optimal response observed at

4 μg/ml FA. In contrast, FA added to the medium in excess of 10 ug/ml FA attenuated

Wnt responses. Stabilization of β-catenin could have a number of effects including cytosolic and nuclear functions. Wnt stimulated transcriptional activity was therefore further examined in this paradigm using NIH3T3 cells transfected with the TopFlash reporter plasmid and showed a similar influence of FA levels on Wnt activity (Figure

4.5C). These data indicate that FA levels can modulate the cellular responses to Wnt stimulation, influencing cytosolic accumulation of β-catenin and TCF/LEF transcriptional activity. The importance of Lrp6 gene dosage to canonical Wnt activity in the rostral neural tube was confirmed using the TCF/LEF-LacZ reporter mouse on the Lrp6-/- background, which showed a substantial decrease in the immunohistochemical fluorescence of β-galactosidase expression in the midbrain-hindbrain compared to the wild-type embryos at E9.5 (Figure 4.5D). Thus, continued Lrp5 gene expression does not fully compensate for LRP6 loss in the region of cranial neural tube closure point 2, which is a common location for neurulation failure in these mutants.

114

Figure 4.4 Maternal FA supplementation effects on proliferation in the neural tube of E9.5 Lrp6 wild-type and mutant embryos Panels compare the dorsal midbrain of Lrp6+/+, Lrp6+/-, and Lrp6-/- E9.5 embryos from each diet group. A. Cells labeled in M-phase with PH3 (green) and counterstained with DAPI (blue). B. Stereological assessment of PH3 labeled/total cells are compared by genotype. The labeling index of mitotic cells in Lrp6+/- embryos is the same as wild-type but is significantly reduced in the Lrp6-/- embryos compared among siblings maintained on the 2 ppm FA diet. On the 10 ppm FA diet, proliferation in the cranial neural folds was nearly doubled in wild-type, but was unchanged in Lrp6+/- embryos and was further suppressed in the Lrp6-/- embryos. * p<0.05 with respect to wild-type on 2 ppm FA.

115

Figure 4.5 Impact of FA on the canonical Wnt signaling pathway in vitro and demonstration of reduced Wnt activity in situ of E9.5 Lrp6-/- embryos A. Western blot of cytosolic ß-catenin levels in NIH3T3 cells that were cultured for 3 days in FA-supplemented defined media prior to 2 hr stimulation with recombinant Wnt3a. B. Quantification of cytosolic β-catenin in lysates from cells, using Erk2 as protein loading control. Normalized optical density values are expressed as percentage of β-catenin level in unstimulated cells after 3 days in 0 ug/ml FA C. Wnt-regulated transcriptional activity measured by TOP/FOP reporter assay. NIH3T3 cells were cultured for 3 days in folate-supplemented defined media prior to transfection with the TOP and FOP reporter plasmids, followed by 36 hr recovery before stimulation with recombinant Wnt3a for 8 hr. Normalized luminescence values are expressed as fold increase above unstimulated (Basal) cells maintained in 0 FA media over the course of the experiment. n=6 separate experiments, each run in duplicate, in both B and C. D. Lrp6-/-::TCF-LacZ embryos (middle) showed almost no Wnt activity in the midline cranial folds and only residual activity in the mesencephalon, likely due to remaining LRP5 function. Lrp6-/- embryos (right), in which a LacZ sequence interrupts the Lrp6 locus, demonstrated no visible β-galactosidase activity at 30 minutes.

116

Discussion

In contrast to our studies in the Cd mouse line, the present study indicates that NTDs

associated with a LOF mutation in Lrp6 are FA-resistant, with no reduction in the

occurrence of these defects among embryos exposed to prenatal diets with moderately

elevated FA content. Moreover, unlike the Cd mouse strain, FA supplementation did not

rescue Lrp6-/- offspring, but instead shifted an already severe phenotype to earlier

embryonic lethality. Another detrimental effect of FA supplementation was evident

through the deficit of not only affected Lrp6-/- embryos but also, more strikingly, of

otherwise viable and unaffected Lrp6+/- embryos. That FA levels influence Wnt signaling

was evident in the effect of FA supplementation on gene expression profiles in Lrp6+/-

tissue. In fact, of the hundreds of pathways that are traceable with the IPA algorithms, it

is highly significant that cell cycle regulation and Wnt signaling pathways and networks

displayed the greatest gene-diet interaction. The relevance of Wnt signaling was

validated using two different in vitro assays of Wnt canonical pathways in cultured cells.

In these assays, FA deficiency blunted pathway responses to Wnt signaling, reflecting the importance of FA metabolism for early development (Gray and Ross 2009). However, these assays also indicate that FA supplementation even to a modestly elevated level attenuates Wnt signaling, so that the Lrp6Cd defect could be ameliorated by countering

this net hyperactive Cd allele, while the further suppression of Wnt signaling- reducing

residual LRP6 co-receptor action-exacerbates the LOF in the Lrp6 knockout line. This indicates a more direct relationship between the FA supplementation and the LRP6/Wnt signaling pathway than was previously appreciated. Such a direct relationship would be

117

consistent with our studies indicating that the Lrp6Cd allele also impacts intracellular FA

utilization (Ernest et al. 2006).

Gene expression profiling suggested that interaction between Lrp6 LOF and FA supplementation impact cell proliferation as a mechanism influencing neurulation. PH3 labeling in neuroepithelial cells around the time of cranial neural tube closure (E9.5) showed that raising dietary levels of FA promotes proliferation in wild-type mice.

However, Lrp6+/- embryos on both the 2 ppm and 10 ppm diets showed levels of

proliferation equivalent to the wild-type 2 ppm baseline level, suggesting an interaction

between gene dosage and FA level. Lrp6-/- mutant embryos displayed an impaired proliferative response to FA supplementation where the 10 ppm FA diet was associated with further reduced proliferation in the neural folds. While these results could explain the failure of added FA to rescue NTDs in Lrp6-/- embryos, they could not explain the

increased loss of heterozygous animals, since there was no difference in mitoses between

diets in the Lrp6+/- cranial neural folds. Moreover, Lrp6-/- embryos of equivalent stage did

not display a statistically significant difference in mitoses on the two diets, undermining

the possibility that a proliferation defect was the sole reason for earlier embryonic loss of

Lrp6-/- embryos receiving the 10 ppm FA diet.

Involvement of LRP6 in canonical Wnt signaling suggested that FA effects on neurulation could involve transcriptional influences on embryogenesis. Indeed, canonical

Wnt signaling in the cranial neural folds and midbrain of E9.5 embryos was diminished as assessed by the comparison of in situ β-galactosidase activity in the Lrp6+/+::TCF-

118

LacZ vs. Lrp6-/-::TCF-LacZ double mutants, when sibling embryos were processed in parallel. Therefore, the loss of LRP6 in the cranial folds was not fully compensated for by

continued Lrp5 expression.

In vitro assessment of Wnt canonical pathway stimulation in the presence of varied FA

levels revealed a bimodal relationship between Wnt signaling and FA concentration.

There was no effect of FA alone on the basal levels of cytosolic β-catenin or TopFlash

reporter activity in NIH3T3 fibroblast cultures. Interestingly, the level of FA that

provided optimal Wnt activation in vitro, 4 μg/ml, is the same concentration used in the

complete Dulbecco’s Modified Eagle’s Medium that was previously optimized for

growth and maintenance of cell lines (Dulbecco and Freeman 1959). FA levels around 10

μg/ml or higher blunted cellular response to recombinant Wnt3a in both β-catenin

stabilization and TopFlash reporter assays. Therefore, FA supplementation has an impact

on transcriptional activation by the canonical Wnt signaling pathway.

A relationship between FA deficiency and Wnt signaling has been reported in the cancer

literature, where 90% of colorectal cancers are thought to begin with aberrant Wnt

pathway gene expression (Crott et al. 2008; Liu et al. 2007; Song et al. 2000). Folate

depletion alters the canonical Wnt pathway in human colorectal adenomas (Jaszewski et al. 2004), in human colon cell lines (Crott et al. 2008), human fibroblast cell lines (Katula et al. 2007), and in mouse colonic mucosa (Liu et al. 2007). The mechanism of FA deficiency effects on Wnt signaling has been proposed to require uracil misincorporation into DNA, encouraging dsDNA breaks and subsequent loss of heterozygosity (LOH) in

119

key genes such as APC (Crott et al. 2008; Dianov et al. 1991; Fodde et al. 2001). This is

consistent with the observation from animal studies that, for example, APCmin mutant

mice had to be maintained on FA-deficient diets for several months before altered gene expression was detected in tissues (Liu et al. 2007; Song et al. 2000). In vitro studies suggest more rapid effects may also occur, since gene expression changes were detected in FA-deprived NIH3T3 cells that were apparent only after 4 days in FA-deficient

medium, with little or no effect before then (Katula et al. 2007). In the present study, in

vitro effects of FA levels on Wnt stimulation were measured after as little as 3 days in

defined media, before any change in genome stability would seem a plausible explanation

of the altered activity. This assertion is supported in our study by the fact that basal levels

of β-catenin or luciferase reporter activity in unstimulated cells were not altered by

increasing FA in the absence of added Wnt. The results reported here indicate, for the

first time, a direct influence of FA on canonical Wnt signaling, in which both FA

depletion and excess impair cytosolic β-catenin accumulation and transcription in

response to Wnt stimulation.

These data also demonstrate that the mechanism of birth defect prevention by FA

supplementation (reduction of NTD incidence by amelioration vs. increased embryonic

loss) can vary not only with the genes associated with increased risk, but also with the

type of mutation within those genes (GOF vs. LOF). Clearly, a detailed understanding of

genetic risk and interactions between folate metabolism and other developmentally

important pathways will be required to optimize birth defect prevention strategies for

individual families.

120

CHAPTER 5

Summary and Future Directions

121

Summary

The mechanism of neural tube defect (NTD) pathogenesis and the process by which the

vitamin folic acid (FA) corrects this defect is unknown. However, public health

organizations have established recommendations for maternal supplementation with FA

to decrease the risk of the defect and have mandated the fortification of enriched grain

products with FA. However, many women are noncompliant with the recommended

levels and some women are biologically resistant to the beneficial effects of this

supplementation. Moreover, high levels of FA have been associated with adverse health

effects in certain populations. Thus, it is important to identify maternal markers of fetal

NTD risk and fetal FA-response to target treatments to responsive women. These studies of maternal markers are facilitated with the presence of NTD mouse models that have a mechanism of neural tube closure similar to that in humans and whose fetal NTDs are either responsive or resistant to FA supplementation. Only a few of the many NTD mouse models have been tested for FA-response. Thus, FA supplementation studies

should be carried out on more mutants to aid in the identification of maternal FA-

response markers and of pathways associated with the mechanism of NTD formation and

of FA response.

The Apobtm1Unc, Vangl2Lp, and Lrp6KO mutants (referred to as the Apob, Lp, and Lrp6

mutant, respectively) used in our studies, are among many NTD mouse models that had

not been tested for their response to FA supplementation. As described in Chapter 2 and

in collaboration with the Laboratory of Dr. Margaret Elizabeth Ross at Cornell University

(referred to as the “Ross lab”), we found that maternal dietary FA supplementation did

122

not significantly alter the frequency of fetal NTDs in any of the three studied mutants.

We also found that FA did not affect the cartilaginous skeletons of the E13.5 Apob and

Lp mutant embryos. Based on these findings we classified the Apob, Lp, and Lrp6

mutants as NTD models of FA-resistance. Unexpectedly, FA supplementation caused a

significant loss of homozygous and heterozygous embryos in all three FA-resistant

mutants. This unpredicted FA-mediated loss of mutant embryos raised two concerns

critically important to the understanding of human NTD prevention. The first was the loss

of affected embryos rather than the widely accepted expectation of defect correction as a

means of decreasing the incidence of NTDs. More alarming was the loss of

heterozygotes, which are otherwise unaffected, healthy, viable fetuses.

Identifying the Apob, Lp, and Lrp6 mutants as FA-resistant, in addition to having others

known to be FA-responsive or -resistant, such as Pax3Sp2H, Lrp6Cd, and ct (referred to as

the Sp, Cd, and ct mutant, respectively), allowed us, in Chapter 3, to test for genes,

pathways and, metabolites that predict the effect of maternal FA supplementation on fetal

NTDs, and provide insight into the mechanisms of NTD pathogenesis and FA-response.

Using the Sp and the Cd mutants as the FA-responsive mutants and the Apob, Lp, Lrp6, and ct mutants as the resistant mutants, we identified the differentially expressed genes specific to each NTD-causing mutation. We found that the biological pathways defining each of the six mutants did not establish patterns that could distinguish FA- responsiveness and -resistance. We also showed that these pathways did not associate specifically with the mutants in which FA unexpectedly led to embryonic loss. However, two pathways, lipopolysaccharide/interleukin-1 (LPS/IL-1) mediated inhibition of

123

retinoid X receptor (RXR) function and pregnane X receptor (PXR)/RXR activation, significantly associated with each of the six NTD mutants regardless of FA-response,

suggesting a shared mechanism of NTD pathogenesis in which these pathways are

involved. Because a shared predictive pathway of FA-response was not identified, we

alternatively characterized genes differentially expressed in all mutants of the same

response type and not the other type. These genes constituted two unique response-

specific expression profiles. Each of the two distinctive gene lists was then associated

with biological pathways. None of the pathways commonly associated with the two

response types, suggesting a variable mechanism of FA-responsiveness and -resistance.

Finally, maternal erythrocyte levels of FA and vitamin B12, which are two important

cofactors of folate-homocysteine metabolism, were not predictive of fetal FA-response.

Interestingly, none of the six NTD causing mutations depleted these metabolites,

suggesting that the pathogenesis of NTDs and the correction of this defect are not directly

related to FA or vitamin B12 levels.

Chapter 4 was a collaborative work with the Ross lab. We investigated the effect of FA

on canonical Wnt signaling. These studies included a functional analysis of the Wnt co-

receptor LRP6, using the Lrp6 mutant that is characterized by NTDs in the homozygous

embryos. We conducted a whole-genome expression study on Lrp6 heterozygous mutants

and wild-type females fed either a FA control or supplemented diet. Analysis of this diet-

genotype interaction revealed a set of 232 genes that were affected by heterozygosity of

Lrp6 mutation. Interestingly, different effects on gene expression were observed between

the control and supplemented states. Many of the interaction significant genes, which

124

included RARG, SMO, Wnt and Wnt7A, are significantly associated with cellular proliferation and Wnt canonical signaling pathways, suggesting a modification of these pathways by FA levels. These findings were confirmed by an in vivo assay, where FA levels interacted with Lrp6 gene dosage to influence neural tube cell proliferation, and an

in vitro assay where FA influenced transcriptional activation in Wnt stimulated cultured

cells. Although these findings reveal a response of Wnt signaling to FA in the Lrp6

mutant, a FA supplementation study showed that this response does not correct the NTD

phenotype, but instead causes an increased loss of homozygous and heterozygous

embryos.

125

Future Directions

Effect of maternal FA supplementation on embryonic loss. In Chapter 2, we reported the results of a maternal FA supplementation study, which demonstrated that the NTDs in the Apob, Lp, and Lrp6 mutant embryos cannot be corrected with FA supplementation.

However, a significant embryonic loss of homozygotes and more importantly, of otherwise unaffected and perfectly viable heterozygotes was evident in significant deviations from the 1:2:1 Mendelian ratio of genotypes expected for heterozygous intercrosses of these mutants. Interestingly, the number of observed resorptions did not differ between 2 ppm and 10 ppm FA supplementation suggesting that the embryonic loss occurred prior to implantation on the 10 ppm diet. Moreover, these losses were also not accompanied by a decrease in the mean litter size compared to that seen with the control FA diet. Thus, we hypothesize that embryonic loss in the FA supplemented mutants occurs before implantation and an optimal litter size is maintained by increased implantation of wild-type embryos.

Three specific aims are therefore proposed to test the relationship between increased FA and embryonic loss with a fixed litter size. The first aim tests whether production of sperm and in particular ovulated oocytes by heterozygous males and females, respectively, is greater on 10 ppm compared to 2 ppm FA, and whether this production is biased towards wild-type versus mutant genotype. If sperm and oocytes have normal genotype distributions for heterozygous animals, the second aim will test whether fertilization with the wild-type genotype is favored over mutant with 10 ppm compared to the control 2 ppm FA diet. Fertilized oocytes can be flushed from the oviduct and

126

genotyped. In addition the fertilizations can take place in vitro under either concentration of FA, after which they will be genotyped. The third aim will test whether increased FA favors the implantation of a specific genotype over the other. We will examine the genotypic proportions at implantation relative to the proportions at fertilization. Fertilized

oocytes could be cultured in vitro with test 10 ppm or control 2 ppm FA concentrations.

After genotyping the embryos will be transferred to a pseudo-pregnant female fed the

corresponding level of FA. The genotypic percentage of implanted relative to transferred

embryos will determine if the wild-type genotype is favored for implantation.

Alternative nutrients for protection against NTDs. Nutrients other than FA have successfully decreased the incidence of NTDs in some mouse models of this defect.

These nutrients, including methionine, thymidine and inositol, may serve as alternative means of protection against NTDs in cases that are resistant to FA. The possibility of alternative protection is evident through studies in the ct mutant, which is resistant to FA and methionine but responsive to myo-inositol (Greene and Copp 1997; van Straaten et al. 1995). In another FA-resistant mutant, Axd (axial defects), the frequency of spina bifida was reduced due to maternal supplementation with the amino acid methionine

(Essien 1992). The Sp mutant, on the other hand, is a FA-responsive NTD model that also responds to thymidine but not to methionine (Fleming and Copp 1998; Greene and Copp

2005; Juriloff and Harris 2000). Other potential alternatives to FA include choline and its oxidized form, betaine, which are both major methyl donors in the HCY methylation pathway (Ueland et al. 2005). Moreover, choline deficiency is associated with an increased risk of NTDs in humans (Shaw et al. 2004). To determine whether nutrients

127

other than FA would serve as alternative treatments for the correction of otherwise FA-

resistant NTDs or cause detrimental effects like FA in some cases, it is important to test

their effect on the outcome of NTDs in previously identified FA-resistant mutants (i.e.

Apob, Lrp6, and Lp). Once favorable effects of these nutrients are identified in mice,

human studies can be conducted for alternative approaches to NTD correction. In

addition, the different outcomes these nutrients have on NTDs may provide clues about

the mechanisms of NTD pathogenesis by identifying different pathways involved in the

correction of this defect.

Effect of FA supplementation on skeletal defects. For FA-responsive mouse models, it

is widely accepted that the decrease in incidence of NTDs results from correction of the phenotype. However, the extent of this correction at the skeletal level was not previously examined. In humans, supplementation might leave unidentified residual subtle defects.

Thus, in Chapter 2, in addition to response determination, we tested the effect of maternal

FA supplementation on embryonic skeletal changes. However, both mutants studied,

Apob and Lp, turned out to be resistant to FA supplementation and their cartilaginous skeletons were not changed. Therefore, to properly understand the extent of NTD correction, it is important to examine the skeletal changes of supplemented FA- responsive mutants and compare them to their non-supplemented controls within a NTD

affected genotype.

Validation of a shared mechanism in NTD pathogenesis. In Chapter 3, we used six

NTD mouse models to identify two signaling pathways potentially involved in a shared

128

mechanism of NTD pathogenesis. The first pathway is LPS/IL-1 mediated inhibition of

RXR function. Retinoic acid (RA) binds to heterodimers of retinoic acid receptor (RAR) and RXR to target and activate Hox genes, which function during early embryonic development to specify the anterior-posterior axis (Germain et al. 2006; Holland 2007).

Thus, inhibition of RXR function might ultimately interfere with this important developmental process. The second pathway, PXR/RXR activation, is involved in the clearance of toxic chemicals from the body (Kliewer et al. 2002). Thus, it is possible that

NTD pathogenesis is due to an altered metabolic process leading to the accumulation of a certain compound, which is perceived as harmful and for which PXR/RXR is activated to relieve the body of its damaging effect. These two pathways should be studied as potential participants in a common mechanism of NTD pathogenesis. As a first step, gene expression and pathway analysis of additional mutants, regardless of FA-response, should show a significant association with the same two RXR related pathways, validating their involvement in the defect’s development. Secondly, perturbations to these pathways could then be designed to determine what downstream aspects are involved in the mechanism of NTDs. These perturbations may include inactivation of genes such as

PXR, RXR, or RAR, or the use of specific antagonists where available. In addition, these nuclear receptors’ downstream targets could be inactivated, such as the cytochrome P450 gene in the case of PXR. Moreover, inactivation of other nuclear receptors, which function through heterodimerizing with PXR or RXR, may be a clue to whether the mechanism of NTD pathogenesis is directly affected by PXR or RXR, or indirectly by their partners.

129

Metabolite profiles for FA-response predictions. Supplementation with folate and

vitamin B12 lowers maternal plasma homocysteine (HCY) levels, which are associated

with the risk of NTDs in many pregnancies (Hague 2003; van Guldener and Stehouwer

2001). In Chapter 3 we showed that levels of folate as well as vitamin B12 were not

depleted in any of the six NTD mutants suggesting that neither the defect nor its

correction was directly related to the levels of these metabolites. However, other studies

suggest that the cause for NTDs is a metabolic block in the folate-HCY pathway, in

which folate is an essential methyl donor (Bottiglieri 2005), or an increased requirement

for one of the many folate dependent enzymes rather than simple folate deficiency (Scott

et al. 1994). Thus, we propose to measure levels of metabolic compounds in the folate-

HCY pathway or related to folate processing. These compounds may include 1) HCY whose increased level is associated with NTDs in some cases (Hague 2003); 2) betaine

and choline, which act as alternative methyl donors in absence of folate; 3) methionine

and S-adenosylmethionine (SAM), which are by-products of the methylation cycle of

HCY metabolism; 4) SAM/S-adenosylhomocysteine (SAM/SAH) ratio, an indicator of

the cellular methylation potential (Melnyk et al. 2000); 5) vitamin B6, which acts as a

cofactor in the transsulfuration pathway of HCY metabolism; and 6) cysteine and

glutathione, which are end products of the transsulfuration pathway (Bottiglieri 2005;

Finkelstein 1998; van der Put and Blom 2000). In addition, cholesterol, which is

important for formation of folate receptor clusters in cell membranes, is also an important

compound although it is not directly related to the folate-HCY pathway (Wolf 1998).

Comparing levels of these compounds between folate-responsive and -resistant NTD

mice, and depending on how these metabolites are affected by dietary FA

130

supplementation, might provide important clues to the prediction of FA-responsiveness and -resistance and insight into the effect of FA levels on the regulation of one-carbon

metabolism in NTDs.

Response predictive profiles in blood. Peripheral blood is mostly used in a clinical

setting for non-invasive detection of biomarkers. Thus, we propose to conduct whole-

blood gene expression studies for the identification of FA-response predictive profiles in

mice. Because the only source of RNA in whole-blood are the white blood cells (WBCs)

(Vu et al. 2004), a difference in the relative proportion of the four WBC types between

each mutant and its control may lead to gene expression differences. In Chapter 3 we

determined that in the five mutants (Sp, Cd, Lrp6, Apob, and Lp), the differential gene

expression in whole blood would largely be the consequence of the NTD-causing

mutation rather than the variable proportion of the four different peripheral WBC types

between each of the mutants and its control. However, in the case of ct and its control

(CBA), any gene expression difference between these two strains should take into

consideration the variable proportions of the two WBC types, neutrophils and

eosinophils. In light of these findings, gene expression analysis on whole-blood in these

six mutants can be interpreted properly. Genes and pathways that consistently display

FA-response specific patterns could serve as potential markers in humans.

Recently, attention has been directed towards microRNAs (miRNAs) as novel blood-

based biomarkers for diagnosis of cancers and other diseases (Chen et al. 2008; Mitchell

et al. 2008). miRNAs are small non-coding RNA molecules that control gene expression

131

and translation, and have a crucial role in many physiologic and pathologic processes

(Kloosterman and Plasterk 2006; Stefani and Slack 2008). Being highly stable in both serum and plasma, where their levels reflect altered physiologic conditions, miRNAs represent an easily accessible biomarker for clinical use (Chen et al. 2008; Mitchell et al.

2008). Moreover, miRNAs exist not only in human blood, but also in many animal species including mice (Chen et al. 2008). Thus miRNA expression profiles can be used to potentially identify biomarkers of FA-response in NTD mouse models.

Pathways affected by FA supplementation. Since the mechanism by which FA exerts its effect on NTDs is currently unknown, we can approach this issue by performing additional experiments on the six NTD mutants. Gene expression analysis in these mutants was performed in the heterozygous females under conditions of a low (control)

FA level. The same experiments could be done on the same mutants, however, in wild- type, heterozygous, and homozygous embryos and under both conditions of low and high

FA levels. Comparing gene expression and pathway associations between the control and supplemented FA levels for each of these mutants could provide insight into the pathways that are affected with supplementation. In addition, comparisons among all three genotypes could indicate how these supplementation-affected pathways are altered by mutation dosage. Because two of the mutants are FA-responsive (Sp & Cd) whereas the remaining four are FA-resistant (Lrp6, Apob, Lp, and ct), these genes and pathways that are significantly affected by the increased FA level could be compared between the two response groups to identify differences specific to each response type. These

132

experiments could provide clues to the mechanism of FA action and how that mechanism

differs in responsive and resistant mutants.

Involvement of the Folate-HCY pathway in the response to FA supplementation. FA

plays a role in both the folate cycle for the production of thymidylate and purines

mediating cell division, and in the methylation cycle of homocysteine metabolism

resulting in methylation of DNA and regulation of many genes. Our Lrp6 study suggested

that FA supplementation affects cell proliferation. However, the proliferation in the

neural tube of Lrp6 mutants was not sufficient to correct the defect, but cell death or

proliferation at other sites might have been sufficient to promote early embryonic loss.

Thus, our question is whether the effect of FA on NTD mutants (correcting the phenotype

or causing early embryonic lethality) is mediated through the folate cycle or the

methylation cycle.

We suggest four specific aims to test the involvement of the folate-HCY pathway in the

response to FA supplementation. The first aim will determine whether the effect of FA is

independent of its involvement in the folate cycle. NTD mutants that responded to FA

supplementation either by inducing embryonic loss or by correcting the NTD will be

supplemented with choline or its oxidized form, betaine, which is an alternative methyl

donor surpassing the folate cycle but still affecting the methylation cycle in the same way

that FA does. If betaine supplementation affects the NTD mutants in the same way that

FA did, this will prove the role of the methylation cycle independent of the folate cycle in

the FA effect. A second aim will test whether the effect of FA in these mutants is

133

independent of an increased methylation potential. The mutants will be supplemented with thymidine and purine, the end products of the folate cycle, with no direct influence on the methylation cycle. If this supplementation corrects the NTD phenotype or leads to embryonic loss as FA did in the respective mutants, then the folate cycle through the thymidine and purine pools is sufficient to mediate the FA effect. In the third aim we will use a BrdU assay to determine the rate of cell division in the neural tube of E9.5 mutant embryos (around the time of neural tube closure) supplemented with betaine or FA and whose NTDs in both cases are corrected. The embryos could be supplemented in vivo through the maternal diet, or in vitro through culture with the specific nutrient. The presence of increased proliferation with FA but not with betaine supplementation will indicate that the correction of the phenotype is influenced by the methylation potential rather than the purine and thymidine pool which stimulates cell division. The fourth aim, will examine cell proliferation and death either with FA or betaine supplementation in the mutants affected by increased embryonic loss. A caspase assay will show what types of cells are dying and a BrdU assay will indicate cell proliferation in pre-implantation embryos under either condition of supplementation. The fertilized oocytes will be cultured in a medium with either nutrient and assayed at consecutive stages until E5, right before the implantation stage. Thus, the exact pre-implantation stage at which lethality happens can be identified. Moreover, if the FA supplementation results in cell division changes that are not seen with betaine, then the common embryonic loss phenotype is mediated through the methylation cycle and is independent of the folate cycle and its effect on cell division. Seeing the same cell proliferation changes with both nutrients

134

would suggest that the embryonic loss phenotype is mediated by the methylation cycle through DNA methylation which regulates certain cell cycle genes.

Characterization of Wnt signaling in NTD models of mutated Lrp6. The Lrp6Cd

mutant (referred to as the Cd mutant) with a gain-of-function (GOF) mutation in the Lrp6

gene is a FA-responsive model of NTDs (Carter et al. 2005; Carter et al. 1999). However, we found that the loss-of-function (LOF) mutation in the same gene gives rise to a FA- resistant NTD model (referred to as the Lrp6 mutant). Could it be that FA attenuates the

Lrp6 gene function causing correction of the phenotype in the Cd mutant, while further attenuation of the effect of the LOF mutation leads to FA-resistance? It is interesting to know what is affected downstream of LRP6 in the Wnt/β-catenin signaling and how FA

alters that pathway to influence the NTD outcome. Furthermore, we showed in Chapter 4

that cell proliferation and Wnt/β-catenin signaling pathways were stimulated by FA

levels in the Lrp6 mutant. However, the effect was insufficient to rescue neurulation

failure. It is important to similarly characterize the Wnt signaling response to FA in the

Cd mutant, to identify differences that might be involved in the two types of FA-response

in these two mutants.

135

APPENDIX

136

APPENDIX 1

Chemical composition of FA control diet (D05072702) and supplemented diet (D05072701). Both diets have the same composition except for folic acid concentration.

D05072702: 2 mg FA contributed by V10001 and casein contributes a trace level of around 0.08 mg - this is in 1,055.05 gm of diet, hence 2 ppm (mg/Kg) FA.

D05072701: 11 mg FA (9 mg added and 2 mg contributed by V10001) and casein contributes a trace level of around 0.08 mg - this is in 1,055.059 gm of diet, hence10 ppm (mg/kg) FA.

Product # D05072701 D05072702 % gm kcal gm kcal Protein 19.2 20 19.2 20 Carbohydrate 67.3 70 67.3 70 Fat 4.3 10 4.3 10 Total 90.8 100 90.8 100 kcal/gm 3.85 3.85

Ingredient gm kcal gm kcal Casein, 80 Mesh 200 800 200 800 L-Cystine 3 12 3 12 Corn Starch 315 1260 315 1260 Maltodextrin 10 35 140 35 140 Sucrose 350 1400 350 1400 Cellulose, BW200 50 0 50 0 Soybean Oil 45 405 45 405 Lard 0 0 0 0 Mineral Mix S10026 10 0 10 0 DiCalcium Phosphate 13 0 13 0 Calcium Carbonate 5.5 0 5.5 0 Potassium Citrate, 1 H2O 16.5 0 16.5 0 Vitamin Mix V10001 10 40 10 40 Choline Bitartrate 2 0 2 0 Folic Acid 0.009 0 0 0 FD&C Yellow Dye #5 0 0 0.05 0 FD&C Red Dye #40 0.05 0 0 0 Total 1055.059 4057 1055.05 4057

137

APPENDIX 2

Penetrance of NTD phenotypes in different FA-responsive mutants.

NTD NTD Penetrance of FA Supplementation mutant phenotype phenotype response Effect (% decrease in NTDs) 1Cart1 Acrania/meroanencephaly 100%  87% 2Cd Exencephaly 20-30%  55% 3Cited2 Exencephaly 80%  84% 4Folbp1 Exencephaly 100%  92% 5Sp Exencephaly & Spina Bifida 100%  65%

 = Responsive 1Cartilage homeoprotein 1 (Zhao et al. 1996) 2Crooked tail (Carter et al. 1999) 3CBP/p300 interacting transactivators with glutamic acid (E)/(D)-rich C-terminal domain (Barbera et al. 2002) 4Folate binding protein 1 (Piedrahita et al. 1999; Spiegelstein et al. 2004) 5PAX3 transcription factor (Fleming and Copp 1998)

138

APPENDIX 3

Primer sequences used for Cd genotyping; provided by the Ross Lab at Cornell University.

D6Mit61-FW CACTTGTTGCTCCTGCTGAG D6Mit61-RV TACAGAGGCTAGAACACTCCTGG D6Mit135-FW CCAGCCCCCAATTTGATATA D6Mit135-RV CCTAACAGTTCAATTTGTCAGCC D6Mit374-FW TACATATGCCAATGATATTCTCCC D6Mit374-RV TTCTGGCTCTTAACAGTCTGTCC

APPENDIX 4

Differentially expressed gene lists for each of the six NTD mutants.

(Refer to: http://genetics.case.edu/?page_id=5&LN=Nadeau&FN=Joseph)

APPENDIX 5

Signaling as well as metabolic pathways associated with each of the six NTD mutants’ significant gene lists.

(Refer to: http://genetics.case.edu/?page_id=5&LN=Nadeau&FN=Joseph)

APPENDIX 6

FA-responsive mutant group and FA-resistant mutant group unique gene lists.

(Refer to: http://genetics.case.edu/?page_id=5&LN=Nadeau&FN=Joseph)

139

APPENDIX 7

Graph representation of the unpaired T-test with Welch’s correction for each of the six mutants and their corresponding controls for each of the four different white blood cell (WBC) types. Each graph represents four comparisons, each comparing the percentage of one of the four different WBC types present in the blood (%-Differential; Y-axis) of the heterozygous mutant to its wild-type control. The blue color represents the heterozygous mutants in each comparison while the pink represents the controls. The horizontal bars represent the means. An asterisk indicates a significant difference with the reported two- tailed p-value.

Sp+/- vs B6

100

%-Differential 40

0 Sp+/- B6 Sp+/- B6 Sp+/- B6 Sp+/- B6 neutrophils lymphocytes monocytes eosinophils

140

Cd+/- vs Cd+/+ 100

%-Differential 40

0 Cd+/- Cd+/+ Cd+/- Cd+/+ Cd +/- Cd+/+ Cd+/- Cd+/+ neutrophils lymphocytes monocytes eosinophils

Lrp6+/- vs Lrp6+/+ 90

40 % Differential 30

0 Lrp6+/- Lrp6+/+ Lrp6+/- Lrp6+/+ Lrp6+/- Lrp6+/+ Lrp6+/- Lrp6+/+ neutrophils lymphocytes monocytes eosinophils

141

Apob+/- vs B6 100

%-Differential 40

0 Apob+/- B6 Apob+/- B6 Apob+/- B6 Apob+/- B6 neutrophils lymphocytes monocytes eosinophils

Lp+/- vs Lp+/+ 100

%-Differential 40

0 Lp+/- Lp+/+ Lp+/- Lp+/+ Lp+/- Lp+/+ Lp+/- Lp+/+ neutrophils lymphocytes monocytes eosinophils

142

Ct vs CBA 100

%-Differential 40 * P=0.025 30 * 20 P=0.004 10

0 Ct CBA Ct CBA Ct CBA Ct CBA neutrophils lymphocytes monocytes eosinophils

APPENDIX 8

Plots representing the significant diet-genotype interactions identified for 232 genes. The X-axis represents the gene factor (GF) with the gene expression change from wild- type to heterozygous state shown from left to right. The intensity of the gene expression is represented by the scale on the Y-axis. The red dotted line represents the change in gene expression on the 2 ppm FA diet. The blue solid line represents the change in gene expression on the 10 ppm FA diet. The number in parentheses represents the significance rank of each gene among the 232 interaction significant genes.

(Refer to: http://genetics.case.edu/?page_id=5&LN=Nadeau&FN=Joseph)

143

BIBLIOGRAPHY

1991. Prevention of neural tube defects: results of the Medical Research Council Vitamin Study. MRC Vitamin Study Research Group. Lancet 338: 131-137.

1999. Folic acid for the prevention of neural tube defects. American Academy of Pediatrics. Committee on Genetics. Pediatrics 104: 325-327.

Allenby, G., M.T. Bocquel, M. Saunders, S. Kazmer, J. Speck, M. Rosenberger, A. Lovey, P. Kastner, J.F. Grippo, P. Chambon, and et al. 1993. Retinoic acid receptors and retinoid X receptors: interactions with endogenous retinoic acids. Proc Natl Acad Sci U S A 90: 30-34.

Allison, D.B., X. Cui, G.P. Page, and M. Sabripour. 2006. Microarray data analysis: from disarray to consolidation and consensus. Nat Rev Genet 7: 55-65.

Antony, A.C. 2007. In utero physiology: role of folic acid in nutrient delivery and fetal development. Am J Clin Nutr 85: 598S-603S.

Antony, A.C. and D.K. Hansen. 2000. Hypothesis: folate-responsive neural tube defects and neurocristopathies. Teratology 62: 42-50.

Bailey, L.B. 1990. Folate status assessment. J Nutr 120 Suppl 11: 1508-1511.

Barbera, J.P., T.A. Rodriguez, N.D. Greene, W.J. Weninger, A. Simeone, A.J. Copp, R.S. Beddington, and S. Dunwoodie. 2002. Folic acid prevents exencephaly in Cited2 deficient mice. Hum Mol Genet 11: 283-293.

Bottiglieri, T. 2005. Homocysteine and folate metabolism in depression. Prog Neuropsychopharmacol Biol Psychiatry 29: 1103-1112.

Botto, L.D., C.A. Moore, M.J. Khoury, and J.D. Erickson. 1999. Neural-tube defects. N Engl J Med 341: 1509-1519.

Brazma, A. and J. Vilo. 2001. Gene expression data analysis. Microbes Infect 3: 823-829.

Brouwer, I.A., M. van Dusseldorp, C.E. West, and R.P. Steegers-Theunissen. 2001. Bioavailability and bioefficacy of folate and folic acid in man. Nutr Res Rev 14: 267-294.

Carter, M., X. Chen, B. Slowinska, S. Minnerath, S. Glickstein, L. Shi, F. Campagne, H. Weinstein, and M.E. Ross. 2005. Crooked tail (Cd) model of human folate-

144

responsive neural tube defects is mutated in Wnt coreceptor lipoprotein receptor- related protein 6. Proc Natl Acad Sci U S A 102: 12843-12848.

Carter, M., S. Ulrich, Y. Oofuji, D.A. Williams, and M.E. Ross. 1999. Crooked tail (Cd) models human folate-responsive neural tube defects. Hum Mol Genet 8: 2199- 2204.

CDC. 1992. Recommendations for the Use of Folic Acid to Reduce the Number of Cases of Spina Bifida and Other Neural Tube Defects. MMWR 41(RR-14): 001.

CDC. 2005. Use of Dietary Supplements Containing Folic Acid Among Women of Childbearing Age - United States. MMWR 54: 955-958.

Chen, X., Y. Ba, L. Ma, X. Cai, Y. Yin, K. Wang, J. Guo, Y. Zhang, J. Chen, X. Guo, Q. Li, X. Li, W. Wang, J. Wang, X. Jiang, Y. Xiang, C. Xu, P. Zheng, J. Zhang, R. Li, H. Zhang, X. Shang, T. Gong, G. Ning, K. Zen, and C.Y. Zhang. 2008. Characterization of microRNAs in serum: a novel class of biomarkers for diagnosis of cancer and other diseases. Cell Res 18: 997-1006.

Cockroft, D.L., F.A. Brook, and A.J. Copp. 1992. Inositol deficiency increases the susceptibility to neural tube defects of genetically predisposed (curly tail) mouse embryos in vitro. Teratology 45: 223-232.

Cole, B.F., J.A. Baron, R.S. Sandler, R.W. Haile, D.J. Ahnen, R.S. Bresalier, G. McKeown-Eyssen, R.W. Summers, R.I. Rothstein, C.A. Burke, D.C. Snover, T.R. Church, J.I. Allen, D.J. Robertson, G.J. Beck, J.H. Bond, T. Byers, J.S. Mandel, L.A. Mott, L.H. Pearson, E.L. Barry, J.R. Rees, N. Marcon, F. Saibil, P.M. Ueland, and E.R. Greenberg. 2007. Folic acid for the prevention of colorectal adenomas: a randomized clinical trial. Jama 297: 2351-2359.

Colmenares, C., H.A. Heilstedt, L.G. Shaffer, S. Schwartz, M. Berk, J.C. Murray, and E. Stavnezer. 2002. Loss of the SKI proto-oncogene in individuals affected with 1p36 deletion syndrome is predicted by strain-dependent defects in Ski-/- mice. Nat Genet 30: 106-109.

Copp, A.J., I. Checiu, and J.N. Henson. 1994. Developmental basis of severe neural tube defects in the loop-tail (Lp) mutant mouse: use of microsatellite DNA markers to identify embryonic genotype. Dev Biol 165: 20-29.

Copp, A.J., N.D. Greene, and J.N. Murdoch. 2003. The genetic basis of mammalian neurulation. Nat Rev Genet 4: 784-793.

145

Crott, J.W., Z. Liu, M.K. Keyes, S.W. Choi, H. Jang, M.P. Moyer, and J.B. Mason. 2008. Moderate folate depletion modulates the expression of selected genes involved in cell cycle, intracellular signaling and folate uptake in human colonic epithelial cell lines. J Nutr Biochem 19: 328-335.

Detrait, E.R., T.M. George, H.C. Etchevers, J.R. Gilbert, M. Vekemans, and M.C. Speer. 2005. Human neural tube defects: developmental biology, epidemiology, and genetics. Neurotoxicol Teratol 27: 515-524.

Dianov, G.L., T.V. Timchenko, O.I. Sinitsina, A.V. Kuzminov, O.A. Medvedev, and R.I. Salganik. 1991. Repair of uracil residues closely spaced on the opposite strands of plasmid DNA results in double-strand break and deletion formation. Mol Gen Genet 225: 448-452.

Dulbecco, R. and G. Freeman. 1959. Plaque production by the polyoma virus. Virology 8: 396-397.

Dzinjalamala, F.K., A. Macheso, J.G. Kublin, T.E. Taylor, K.I. Barnes, M.E. Molyneux, C.V. Plowe, and P.J. Smith. 2005. Blood folate concentrations and in vivo sulfadoxine-pyrimethamine failure in Malawian children with uncomplicated Plasmodium falciparum malaria. Am J Trop Med Hyg 72: 267-272.

Elmazar, M.M., R. Ruhl, U. Reichert, B. Shroot, and H. Nau. 1997. RARalpha-mediated teratogenicity in mice is potentiated by an RXR agonist and reduced by an RAR antagonist: dissection of retinoid receptor-induced pathways. Toxicol Appl Pharmacol 146: 21-28.

Epstein, D.J., K.J. Vogan, D.G. Trasler, and P. Gros. 1993. A mutation within intron 3 of the Pax-3 gene produces aberrantly spliced mRNA transcripts in the splotch (Sp) mouse mutant. Proc Natl Acad Sci U S A 90: 532-536.

Ernest, S. 2004. Genetic and Molecular Disection of Homocysteinemia in Mice. In Genetics, pp. 237. Case Western Reserve University, Cleveland.

Ernest, S., M. Carter, H. Shao, A. Hosack, N. Lerner, C. Colmenares, D.S. Rosenblatt, Y.H. Pao, M.E. Ross, and J.H. Nadeau. 2006. Parallel changes in metabolite and expression profiles in crooked-tail mutant and folate-reduced wild-type mice. Hum Mol Genet 15: 3387-3393.

Ernest, S., B. Christensen, B.M. Gilfix, O.A. Mamer, A. Hosack, M. Rodier, C. Colmenares, J. McGrath, A. Bale, R. Balling, D. Sankoff, D.S. Rosenblatt, and J.H. Nadeau. 2002. Genetic and molecular control of folate-homocysteine metabolism in mutant mice. Mamm Genome 13: 259-267.

146

Essien, F.B. 1992. Maternal methionine supplementation promotes the remediation of axial defects in Axd mouse neural tube mutants. Teratology 45: 205-212.

Falkner, K.C., J.A. Pinaire, G.H. Xiao, T.E. Geoghegan, and R.A. Prough. 2001. Regulation of the rat glutathione S-transferase A2 gene by glucocorticoids: involvement of both the glucocorticoid and pregnane X receptors. Mol Pharmacol 60: 611-619.

Finkelstein, J.D. 1998. The metabolism of homocysteine: pathways and regulation. Eur J Pediatr 157 Suppl 2: S40-44.

Finkelstein, J.E., E.R. Hauser, C.O. Leonard, and S.W. Brusilow. 1990. Late-onset ornithine transcarbamylase deficiency in male patients. J Pediatr 117: 897-902.

Fleming, A. and A.J. Copp. 1998. Embryonic folate metabolism and mouse neural tube defects. Science 280: 2107-2109.

Fodde, R., R. Smits, and H. Clevers. 2001. APC, signal transduction and genetic instability in colorectal cancer. Nat Rev Cancer 1: 55-67.

Franke, B., R. Klootwijk, B. Lemmers, C.G. de Kovel, R.P. Steegers-Theunissen, and E.C. Mariman. 2003. Phenotype of the neural tube defect mouse model bent tail is not sensitive to maternal folinic acid, myo-inositol, or zinc supplementation. Birth Defects Res A Clin Mol Teratol 67: 979-984.

Germain, P., P. Chambon, G. Eichele, R.M. Evans, M.A. Lazar, M. Leid, A.R. De Lera, R. Lotan, D.J. Mangelsdorf, and H. Gronemeyer. 2006. International Union of Pharmacology. LXIII. Retinoid X receptors. Pharmacol Rev 58: 760-772.

Giarre, M., M.V. Semenov, and A.M. Brown. 1998. Wnt signaling stabilizes the dual- function protein beta-catenin in diverse cell types. Ann N Y Acad Sci 857: 43-55.

Golden, J.A. and G.F. Chernoff. 1993. Intermittent pattern of neural tube closure in two strains of mice. Teratology 47: 73-80.

Graves, J.C., K.E. Miller, and A.D. Sellers. 2002. Maternal serum triple analyte screening in pregnancy. Am Fam Physician 65: 915-920.

Gray, J.D. and M.E. Ross. 2009. Mechanistic insights into folate supplementation from Crooked tail and other NTD-prone mutant mice. Birth Defects Res A Clin Mol Teratol 85: 314-321.

147

Greene, N.D. and A.J. Copp. 1997. Inositol prevents folate-resistant neural tube defects in the mouse. Nat Med 3: 60-66.

Greene, N.D. and A.J. Copp. 2005. Mouse models of neural tube defects: investigating preventive mechanisms. Am J Med Genet C Semin Med Genet 135: 31-41.

Greene, N.D., D. Gerrelli, H.W. Van Straaten, and A.J. Copp. 1998. Abnormalities of floor plate, notochord and somite differentiation in the loop-tail (Lp) mouse: a model of severe neural tube defects. Mech Dev 73: 59-72.

Groenen, P.M., P.G. Peer, R.A. Wevers, D.W. Swinkels, B. Franke, E.C. Mariman, and R.P. Steegers-Theunissen. 2003a. Maternal myo-inositol, glucose, and zinc status is associated with the risk of offspring with spina bifida. Am J Obstet Gynecol 189: 1713-1719.

Groenen, P.M., R.A. Wevers, F.S. Janssen, J.H. Tuerlings, J.M. Merkus, and R.P. Steegers-Theunissen. 2003b. Are myo-inositol, glucose and zinc concentrations in amniotic fluid of fetuses with spina bifida different from controls? Early Hum Dev 71: 1-8.

Grosse, S.D., L. Ouyang, J.S. Collins, D. Green, J.H. Dean, and R.E. Stevenson. 2008. Economic evaluation of a neural tube defect recurrence-prevention program. Am J Prev Med 35: 572-577.

Hague, W.M. 2003. Homocysteine and pregnancy. Best Pract Res Clin Obstet Gynaecol 17: 459-469.

Harris, M.J. and D.M. Juriloff. 1997. Genetic landmarks for defects in mouse neural tube closure. Teratology 56: 177-187.

Harris, M.J. and D.M. Juriloff. 2007. Mouse mutants with neural tube closure defects and their role in understanding human neural tube defects. Birth Defects Res A Clin Mol Teratol 79: 187-210.

Heaney, J.D., M.V. Michelson, K.K. Youngren, M.Y. Lam, and J.H. Nadeau. 2009. Deletion of eIF2beta suppresses testicular cancer incidence and causes recessive lethality in agouti-yellow mice. Hum Mol Genet.

Hedenqvist, P. 2008. Anesthesia and analgesia for surgery in rabbits and rats: A comparison of the effects of different compounds. In Department of physiology and pharmacology, pp. 57. Karolinska Institutet, Stockholm.

148

Holland, L.Z. 2007. Developmental biology: a chordate with a difference. Nature 447: 153-155.

Homanics, G.E., N. Maeda, M.G. Traber, H.J. Kayden, D.B. Dehart, and K.K. Sulik. 1995. Exencephaly and hydrocephaly in mice with targeted modification of the apolipoprotein B (Apob) gene. Teratology 51: 1-10.

Homanics, G.E., T.J. Smith, S.H. Zhang, D. Lee, S.G. Young, and N. Maeda. 1993. Targeted modification of the apolipoprotein B gene results in hypobetalipoproteinemia and developmental abnormalities in mice. Proc Natl Acad Sci U S A 90: 2389-2393.

Hsiao, A., T. Ideker, J.M. Olefsky, and S. Subramaniam. 2005. VAMPIRE microarray suite: a web-based platform for the interpretation of gene expression data. Nucleic Acids Res 33: W627-632.

IOM. 1998. Institute of Medicine. Dietary reference intakes for thiamine, riboflavin, niacin, vitamin B6, folate, vitamin B12, pantothenic acid, biotin, and choline. Washington DC: National Academy Press.

Irizarry, R.A., B.M. Bolstad, F. Collin, L.M. Cope, B. Hobbs, and T.P. Speed. 2003a. Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res 31: e15- e15.

Irizarry, R.A., B. Hobbs, F. Collin, Y.D. Beazer-Barclay, K.J. Antonellis, U. Scherf, and T.P. Speed. 2003b. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4: 249-264.

Ishwaran, H. and J.S. Rao. 2003. Detecting Differentially Expressed Genes in Microarray Using Bayesian Model Selection. J Am Stat Assoc 98: 438-455.

Ishwaran, H. and J.S. Rao. 2005. Spike and slab gene selection for multigroup microarray data. J Am Stat Assoc 100: 764-780.

Ishwaran, H., J.S. Rao, and U.B. Kogalur. 2006. BAMarraytrade mark: Java software for Bayesian analysis of variance for microarray data. BMC Bioinformatics 7: 59.

Jallo, G. 2005. Neural Tube Defects. eMedicine eds R.S. Rust F. Talavera K.J. Mack M.J. Baker, and N. Lorenzo).

Jaszewski, R., B. Millar, J.S. Hatfield, K. Nogothu, R. Finkenauer, A.K. Rishi, J.A. Naumoff, O. Kucuk, B.N. Axelrod, and A.P. Majumdar. 2004. Folic acid reduces

149

nuclear translocation of beta-catenin in rectal mucosal crypts of patients with colorectal adenomas. Cancer Lett 206: 27-33.

Johnston, R.B., Jr. 2008. Will increasing folic acid in fortified grain products further reduce neural tube defects without causing harm?: consideration of the evidence. Pediatr Res 63: 2-8.

Juriloff, D.M. and M.J. Harris. 1998. Animal models of neural tube defects. Ment Retard Dev Disabil Res Rev 4: 254-263.

Juriloff, D.M. and M.J. Harris. 2000. Mouse models for neural tube closure defects. Hum Mol Genet 9: 993-1000.

Kaneko, S., D. Battino, E. Andermann, K. Wada, R. Kan, A. Takeda, Y. Nakane, Y. Ogawa, G. Avanzini, C. Fumarola, T. Granata, F. Molteni, G. Pardi, L. Minotti, R. Canger, L. Dansky, M. Oguni, I. Lopes-Cendas, A. Sherwin, F. Andermann, M.H. Seni, M. Okada, and T. Teranishi. 1999. Congenital malformations due to antiepileptic drugs. Epilepsy Res 33: 145-158.

Kaplan, K.M., J.M. Spivak, and J.A. Bendo. 2005. Embryology of the spine and associated congenital abnormalities. Spine J 5: 564-576.

Kappen, C. 2005. Folate supplementation in three genetic models: implications for understanding folate-dependent developmental pathways. Am J Med Genet C Semin Med Genet 135: 24-30.

Katula, K.S., A.N. Heinloth, and R.S. Paules. 2007. Folate deficiency in normal human fibroblasts leads to altered expression of genes primarily linked to cell signaling, the cytoskeleton and extracellular matrix. J Nutr Biochem 18: 541-552.

Kaufman, M.H. 1992. The Atlas of Mouse Development. Academic Press Limited.

Kelly, P., J. McPartlin, M. Goggins, D.G. Weir, and J.M. Scott. 1997. Unmetabolized folic acid in serum: acute studies in subjects consuming fortified food and supplements. Am J Clin Nutr 65: 1790-1795.

Khan, Z., S. Vijayakumar, T.V. de la Torre, S. Rotolo, and A. Bafico. 2007. Analysis of endogenous LRP6 function reveals a novel feedback mechanism by which Wnt negatively regulates its receptor. Mol Cell Biol 27: 7291-7301.

Kim, Y.I. 2003. Role of folate in colon cancer development and progression. J Nutr 133: 3731S-3739S.

150

Kliewer, S.A., B. Goodwin, and T.M. Willson. 2002. The nuclear pregnane X receptor: a key regulator of xenobiotic metabolism. Endocr Rev 23: 687-702.

Kloosterman, W.P. and R.H. Plasterk. 2006. The diverse functions of microRNAs in animal development and disease. Dev Cell 11: 441-450.

Kokubu, C., U. Heinzmann, T. Kokubu, N. Sakai, T. Kubota, M. Kawai, M.B. Wahl, J. Galceran, R. Grosschedl, K. Ozono, and K. Imai. 2004. Skeletal defects in ringelschwanz mutant mice reveal that Lrp6 is required for proper somitogenesis and osteogenesis. Development 131: 5469-5480.

Lamers, Y., R. Prinz-Langenohl, R. Moser, and K. Pietrzik. 2004. Supplementation with [6S]-5-methyltetrahydrofolate or folic acid equally reduces plasma total homocysteine concentrations in healthy women. Am J Clin Nutr 79: 473-478.

Lehmann, J.M., D.D. McKee, M.A. Watson, T.M. Willson, J.T. Moore, and S.A. Kliewer. 1998. The human orphan nuclear receptor PXR is activated by compounds that regulate CYP3A4 gene expression and cause drug interactions. J Clin Invest 102: 1016-1023.

Liu, H., O. Mohamed, D. Dufort, and V.A. Wallace. 2003. Characterization of Wnt signaling components and activation of the Wnt canonical pathway in the murine retina. Dev Dyn 227: 323-334.

Liu, Z., S.W. Choi, J.W. Crott, M.K. Keyes, H. Jang, D.E. Smith, M. Kim, P.W. Laird, R. Bronson, and J.B. Mason. 2007. Mild depletion of dietary folate combined with other B vitamins alters multiple components of the Wnt pathway in mouse colon. J Nutr 137: 2701-2708.

Lohnes, D., M. Mark, C. Mendelsohn, P. Dolle, A. Dierich, P. Gorry, A. Gansmuller, and P. Chambon. 1994. Function of the retinoic acid receptors (RARs) during development (I). Craniofacial and skeletal abnormalities in RAR double mutants. Development 120: 2723-2748.

Macdonald, K.B., D.M. Juriloff, and M.J. Harris. 1989. Developmental study of neural tube closure in a mouse stock with a high incidence of exencephaly. Teratology 39: 195-213.

Marletaz, F., L.Z. Holland, V. Laudet, and M. Schubert. 2006. Retinoic acid signaling and the evolution of chordates. Int J Biol Sci 2: 38-47.

151

McShane, L.M., J.H. Shih, and A.M. Michalowska. 2003. Statistical issues in the design and analysis of gene expression microarray studies of animal models. J Mammary Gland Biol Neoplasia 8: 359-374.

Melnyk, S., M. Pogribna, I.P. Pogribny, P. Yi, and S.J. James. 2000. Measurement of plasma and intracellular S-adenosylmethionine and S-adenosylhomocysteine utilizing coulometric electrochemical detection: alterations with plasma homocysteine and pyridoxal 5'-phosphate concentrations. Clin Chem 46: 265-272.

Mitchell, L.E. 2005. Epidemiology of neural tube defects. Am J Med Genet C Semin Med Genet 135: 88-94.

Mitchell, P.S., R.K. Parkin, E.M. Kroh, B.R. Fritz, S.K. Wyman, E.L. Pogosova- Agadjanyan, A. Peterson, J. Noteboom, K.C. O'Briant, A. Allen, D.W. Lin, N. Urban, C.W. Drescher, B.S. Knudsen, D.L. Stirewalt, R. Gentleman, R.L. Vessella, P.S. Nelson, D.B. Martin, and M. Tewari. 2008. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci U S A 105: 10513-10518.

Mohamed, O.A., H.J. Clarke, and D. Dufort. 2004. Beta-catenin signaling marks the prospective site of primitive streak formation in the mouse embryo. Dev Dyn 231: 416-424.

Nagy, A., M. Gertsenstein, K. Vintersten, and R. Behringer. 2003. Manipulating the Mouse Embryo: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor.

Northrup, H. and K.A. Volcik. 2000. Spina bifida and other neural tube defects. Curr Probl Pediatr 30: 313-332.

NRC. 1995. Nutrient requirements of the mouse. In Nutrient Requirements of Laboratory Animals, Fourth Revised Edition, pp. 80-102. National Academy Press, Washington, DC.

Ouma, P., M.E. Parise, M.J. Hamel, F.O. Ter Kuile, K. Otieno, J.G. Ayisi, P.A. Kager, R.W. Steketee, L. Slutsker, and A.M. van Eijk. 2006. A randomized controlled trial of folate supplementation when treating malaria in pregnancy with sulfadoxine-pyrimethamine. PLoS Clin Trials 1: e28.

Piedrahita, J.A., B. Oetama, G.D. Bennett, J. van Waes, B.A. Kamen, J. Richardson, S.W. Lacey, R.G. Anderson, and R.H. Finnell. 1999. Mice lacking the folic acid- binding protein Folbp1 are defective in early embryonic development. Nat Genet 23: 228-232.

152

Pinson, K.I., J. Brennan, S. Monkley, B.J. Avery, and W.C. Skarnes. 2000. An LDL- receptor-related protein mediates Wnt signalling in mice. Nature 407: 535-538.

Pollard, J.H. 1977. A handbook of numerical and statistical techniques with examples mainly from the life sciences. Cambridge University Press.

Pusztai, L. and K.R. Hess. 2004. Clinical trial design for microarray predictive marker discovery and assessment. Ann Oncol 15: 1731-1737.

Rader, J.I. and B.O. Schneeman. 2006. Prevalence of neural tube defects, folate status, and folate fortification of enriched cereal-grain products in the United States. Pediatrics 117: 1394-1399.

Ray, J.G., C. Meier, M.J. Vermeulen, S. Boss, P.R. Wyatt, and D.E. Cole. 2002. Association of neural tube defects and folic acid food fortification in Canada. Lancet 360: 2047-2048.

Rigby, A.S. 2001. Statistical methods in epidemiology. VII. An overview of the chi2 test for 2 x 2 contingency table analysis. Disabil Rehabil 23: 693-697.

Sadler, T.W. 1998. Mechanisms of neural tube closure and defects. Ment Retard Dev Disabil Res Rev 4: 247-253.

Scott, J.M., D.G. Weir, A. Molloy, J. McPartlin, L. Daly, and P. kirke. 1994. Folic acid metabolism and mechanisms of neural tube defects. In Ciba Foundation Symposium 181. Neural Tube Defects, pp. 180-191. John Wiley & Sons Ltd, Chichester.

Selhub, J., M.S. Morris, and P.F. Jacques. 2007. In vitamin B12 deficiency, higher serum folate is associated with increased total homocysteine and methylmalonic acid concentrations. Proc Natl Acad Sci U S A 104: 19995-20000.

Shaw, G.M., S.L. Carmichael, W. Yang, S. Selvin, and D.M. Schaffer. 2004. Periconceptional dietary intake of choline and betaine and neural tube defects in offspring. Am J Epidemiol 160: 102-109.

Smithells, R.W., N.C. Nevin, M.J. Seller, S. Sheppard, R. Harris, A.P. Read, D.W. Fielding, S. Walker, C.J. Schorah, and J. Wild. 1983. Further experience of vitamin supplementation for prevention of neural tube defect recurrences. Lancet 1: 1027-1031.

Smithells, R.W., S. Sheppard, and C.J. Schorah. 1976. Vitamin dificiencies and neural tube defects. Arch Dis Child 51: 944-950.

153

Sokal, R.R. and F.J. Rohlf. 1995. Biometry: the principles and practice of statistics in biological research. W. H. Freeman and Company, New York.

Song, J., A. Medline, J.B. Mason, S. Gallinger, and Y.I. Kim. 2000. Effects of dietary folate on intestinal tumorigenesis in the apcMin mouse. Cancer Res 60: 5434- 5440.

Spiegelstein, O., L.E. Mitchell, M.Y. Merriweather, N.J. Wicker, Q. Zhang, E.J. Lammer, and R.H. Finnell. 2004. Embryonic development of folate binding protein-1 (Folbp1) knockout mice: Effects of the chemical form, dose, and timing of maternal folate supplementation. Dev Dyn 231: 221-231.

Stefani, G. and F.J. Slack. 2008. Small non-coding RNAs in animal development. Nat Rev Mol Cell Biol 9: 219-230.

Stevenson, R.E., W.P. Allen, G.S. Pai, R. Best, L.H. Seaver, J. Dean, and S. Thompson. 2000. Decline in prevalence of neural tube defects in a high-risk region of the United States. Pediatrics 106: 677-683.

Tom, C., D.M. Juriloff, and M.J. Harris. 1991. Studies of the effect of retinoic acid on anterior neural tube closure in mice genetically liable to exencephaly. Teratology 43: 27-40.

Torban, E., H.J. Wang, N. Groulx, and P. Gros. 2004. Independent mutations in mouse Vangl2 that cause neural tube defects in looptail mice impair interaction with members of the Dishevelled family. J Biol Chem 279: 52703-52713.

Tran, P., F. Hiou-Tim, P. Frosst, S. Lussier-Cacan, P. Bagley, J. Selhub, T. Bottiglieri, and R. Rozen. 2002. The curly-tail (ct) mouse, an animal model of neural tube defects, displays altered homocysteine metabolism without folate responsiveness or a defect in Mthfr. Mol Genet Metab 76: 297-304.

Ueland, P.M., P.I. Holm, and S. Hustad. 2005. Betaine: a key modulator of one-carbon metabolism and homocysteine status. Clin Chem Lab Med 43: 1069-1075.

Van Allen, M.I., D.K. Kalousek, G.F. Chernoff, D. Juriloff, M. Harris, B.C. McGillivray, S.L. Yong, S. Langlois, P.M. MacLeod, D. Chitayat, and et al. 1993. Evidence for multi-site closure of the neural tube in humans. Am J Med Genet 47: 723-743.

van der Put, N.M. and H.J. Blom. 2000. Neural tube defects and a disturbed folate dependent homocysteine metabolism. Eur J Obstet Gynecol Reprod Biol 92: 57- 61.

154

van der Put, N.M., H.W. van Straaten, F.J. Trijbels, and H.J. Blom. 2001. Folate, homocysteine and neural tube defects: an overview. Exp Biol Med (Maywood) 226: 243-270. van Guldener, C. and C.D. Stehouwer. 2001. Homocysteine-lowering treatment: an overview. Expert Opin Pharmacother 2: 1449-1460. van Straaten, H.W., H. Blom, M.C. Peeters, A.M. Rousseau, K.J. Cole, and M.J. Seller. 1995. Dietary methionine does not reduce penetrance in curly tail mice but causes a phenotype-specific decrease in embryonic growth. J Nutr 125: 2733-2740. van Straaten, H.W. and A.J. Copp. 2001. Curly tail: a 50-year history of the mouse spina bifida model. Anat Embryol (Berl) 203: 225-237.

Vu, N.T., H. Zhu, E.D. Owuor, M.E. Huggins, L.V. White, A.K. Chaturvedi, D.V. Canfield, and J.E. Whinnery. 2004. Isolation of RNA from Peripheral Blood Cells: A Validation Study for Molecular Diagnostics by Microarray and Kinetic RT-PCR Assays. Application in Aerospace Medicine, pp. 1-8. Civil Aerospace Medical Institute. Federal Aviation Administration, Oklahoma City.

Whiteman, D., M. Murphy, K. Hey, M. O'Donnell, and M. Goldacre. 2000. Reproductive factors, subfertility, and risk of neural tube defects: a case-control study based on the Oxford Record Linkage Study Register. Am J Epidemiol 152: 823-828.

Wiggan, O. and P.A. Hamel. 2002. Pax3 regulates morphogenetic cell behavior in vitro coincident with activation of a PCP/non-canonical Wnt-signaling cascade. J Cell Sci 115: 531-541.

Wolf, G. 1998. Inhibition of cellular uptake of folate by blocking synthesis of the membrane folate receptor. Nutr Rev 56: 86-87.

Woods, C.G., S. Stricker, P. Seemann, R. Stern, J. Cox, E. Sherridan, E. Roberts, K. Springell, S. Scott, G. Karbani, S.M. Sharif, C. Toomes, J. Bond, D. Kumar, L. Al-Gazali, and S. Mundlos. 2006. Mutations in WNT7A cause a range of limb malformations, including Fuhrmann syndrome and Al-Awadi/Raas- Rothschild/Schinzel phocomelia syndrome. Am J Hum Genet 79: 402-408.

Yao, T.P., S.P. Oh, M. Fuchs, N.D. Zhou, L.E. Ch'ng, D. Newsome, R.T. Bronson, E. Li, D.M. Livingston, and R. Eckner. 1998. Gene dosage-dependent embryonic development and proliferation defects in mice lacking the transcriptional integrator p300. Cell 93: 361-372.

155

Zhao, Q., R.R. Behringer, and B. de Crombrugghe. 1996. Prenatal folic acid treatment suppresses acrania and meroanencephaly in mice mutant for the Cart1 homeobox gene. Nat Genet 13: 275-283.

156