Cardiovascular Abnormalities in Folr1 KnockoutMice and Folate Rescue

Huiping Zhu,1* Bogdan J. Wlodarczyk,1 Melissa Scott,1 Wei Yu,2 Michelle Merriweather,1

Janee Gelineau-van Waes,3 Robert J. Schwartz,2 and Richard H. Finnell1,3,4

1Center for Environmental and Genetic Medicine, Institute of Biosciences and Technology,Texas A&M University System Health Science Center, Houston, Texas

2Center for Molecular Development and Disease, Institute of Biosciences and Technology,Texas A&M University System Health Science Center, Houston, Texas

3Department of Pediatrics, University of Nebraska Medical Center, Omaha, Nebraska4Center for Environmental and Rural Health, Texas A&M University, College Station, Texas

Received 26 September 2006; Revised 10 November 2006; Accepted 6 December 2006

BACKGROUND: Periconceptional folic acid supplementation is widely believed to aid in the prevention ofneural tube defects (NTDs), orofacial clefts, and congenital heart defects. Folate-binding proteins or receptorsserve to bind folic acid and 5-methyltetrahydrofolate, representing one of the two major mechanisms of cel-lular folate uptake. METHODS: We herein describe abnormal cardiovascular development in mouse fetuseslacking a functional folate-binding protein gene (Folr1). We also performed a dose-response study with fo-linic acid and determined the impact of maternal folate supplementation on Folr1 nullizygous cardiac devel-opment. RESULTS: Partially rescued preterm Folr1�/� (formerly referred to as Folbp1) fetuses were found tohave outflow tract defects, aortic arch artery abnormalities, and isolated dextracardia. Maternal supplemen-tation with folinic acid rescued the embryonic lethality and the observed cardiovascular phenotypes in adose-dependant manner. Maternal genotype exhibited significant impact on the rescue efficiency, suggestingan important role of in utero folate status in embryonic development. Abnormal heart looping was observedduring early development of Folr1�/� embryos partially rescued by maternal folinic acid supplementation.Migration pattern of cardiac neural crest cells, genetic signals in pharyngeal arches, and the secondary heartfield were also found to be affected in the mutant embryos. CONCLUSIONS: Our observations suggest thatthe beneficial effect of folic acid for congenital heart defects might be mediated via its impact on neural crestcells and by gene regulation of signaling pathways involved in the development of the pharyngeal archesand the secondary heart field. Birth Defects Research (Part A) 79:257–268, 2007. � 2007 Wiley-Liss, Inc.

Key words: folate; folate receptor alpha (Folr1); knockout; congenital heart defects; conotruncal defect; neuralcrest cell; secondary heart field

INTRODUCTION

Congenital heart defects (CHDs) are the most commonhuman congenital malformation, and the leading causeof infant death secondary to birth anomalies. Approxi-mately 35,000 babies are born with a CHD each year inthe United States. The crude infant (<1 year) death ratesare 44.0 per 100,000 for White babies, and 56.2 per100,000 Black babies (American Heart Association, 2001).At present, there are at least 35 well-described CHDs.The most common defects include those of the atrial andventricular septa, transposition of great arteries, coarcta-tion of the aorta, and Tetralogy of Fallot. For the mostpart, the causes of CHDs are largely unknown. Some ofthe known causes and risk factors include: gene muta-

tions (e.g., Nkx2-5 mutation; Reamon-Buettner andBorlak, 2004), chromosome abnormalities (e.g., 22q dele-tion syndrome; Baldini, 2005), infection (e.g., rubella;Rittler et al., 2004), and in utero alcohol or drug exposure

Presented as a poster at the CERH Scientific Symposium, Texas A&M University,December 16, 2005, Hilton Hotel and Conference Center, College Station, Texas.Grant sponsor: NHLBI Program Project; Grant number: 1PO1HL66398.;

The contents of this article are solely the responsibility of the authors and donot necessarily represent the official views of the NIH.*Correspondence to: Huiping Zhu, Institute of Biosciences and Technology,Texas A&M University System Health Science Center, 2121 W. HolcombeBoulevard, Houston, TX. E-mail: hzhu@ibt.tamhsc.eduPublished online 7 February 2007 in Wiley InterScience (www.interscience.wiley.com).DOI: 10.1002/bdra.20347

Birth Defects Research (Part A): Clinical and Molecular Teratology 79:257–268 (2007)

� 2007 Wiley-Liss, Inc. Birth Defects Research (Part A) 79:257–268 (2007)

(Williams et al., 2004). Maternal health status and mater-nal age also contribute to the incidence of CHDs.

Maternal folic acid intake has been shown to modifyrisks for selected birth defects, including NTDs, in anumber of well-designed epidemiology studies from thelast decade (Czeizel and Dudas, 1992; Czeizel et al., 1996,1998; MRC Vitamin Study Research Group, 1991; Berryet al., 1999). This evidence nourished the hope that pri-mary prevention might also be possible for other birthdefects, such as selected CHDs. The results obtainedfrom a number of clinical trials and case-control studiessuggested that periconceptional folic acid supplementa-tion might also decrease the risk for some heart defects,specifically conotruncal and ventricular septal defects(Czeizel, 1998; Shaw et al., 1995; Botto et al., 1996, 2000,2003; Scanlon et al., 1998; Werler et al., 1999; Bailey andBerry, 2005). Recent human epidemiological studies haveshown that elevated maternal homocysteine levels maycontribute to the risk for human conotruncal heartdefects (Hobbs et al., 2006), providing further indirectevidence that folate deficiency is a risk factor for thesedefects. Finally, genetic variations affecting cellular folatetransportation are also known to be a risk factor forthis group of heart defects (Shaw et al., 2003; Pei et al.,2005).

For decades, animal studies have provided evidencethat maternal folate deficiency can result in abnormal fe-tal heart development. Conotruncal defects were ob-served among offspring of female mice maintained on afolate-deficient diet (Warkany and Nelson, 1940; Nelsonet al., 1952; Nelson, 1960). These folate-deficient micehave decreased plasma folate levels and increased homo-cysteine concentrations (Burgoon et al., 2002). Abnormalcardiovascular development was also observed in geneti-cally modified mice with a mild methylenetetrahydrofo-late reductase deficiency, especially when the dam wasmaintained on a folate-deficient diet (Li et al., 2005).Studies using chicken embryos provided clear and con-vincing evidence supporting the cardiac toxicity of highhomocysteine concentrations. Rosenquist et al. (1996)determined that exogenous homocysteine can cause dys-morphogenesis of the heart and neural tube, as well as ofthe ventral wall in chicken embryos. Homocysteine levelshave been shown to decrease with increased consump-tion of folate, B6, and B12. Boot et al. (2003b) observedabnormal differentiation of neural crest cells in the pres-ence of high homocysteine concentrations using anin vitro system. In a recent study, Boot et al. (2004)injected homocysteine into either the neural tube lumenor the circulation system of chicken embryos andobserved 60% less apoptosis and 25% reduced myocardi-alization in the outflow tract of embryos from eggs, sug-gesting homocysteine disturbances cause apoptosis andmyocardialization of the outflow tract by affecting thecardiac neural crest cells, which leads to abnormal heartdevelopment in these embryos. The transport of folatesinto cells is mediated either by folate receptors or thereduced folate carrier system (Matherly and Goldman,2003). Folate receptors are anchored to the cell surface byglycosylphosphatidylinositol (GPI) linkages. Folates boundto these receptors are incorporated into cells through anendocytotic mechanism (Kamen et al., 1988). The folatereceptors are thought to have multiple important roles inembryogenesis. Conventional knockout of the Folr1 geneproduced mouse embryos with severe structural abnor-

malities involving the prosencephalon, mesencephalon,optic vesicle, otic placode, and branchial arches (Piedrahitaet al., 1999; Tang and Finnell, 2003; Speigelstein et al.,2004; Tang et al., 2004, 2005). Such embryos die in utero byembryonic day 10 (E10). Previous work with thesemutants revealed that impaired folic acid transport inFolr1�/� embryos results in extensive apoptosis-mediatedcell death, which concentrated in the interventricular sep-tum and truncus arteriosus (Tang et al., 2004). Daily oralsupplementation of dams with N5-formyl-tetrahydrofolate(folinic acid, 5F-THF), which is an active form of folatesfound naturally in foods, prior to and throughout gesta-tion reduces embryonic mortality and ameliorates a signif-icant portion of the adverse developmental effects in Folr1nullizygotes. External defects involving the neural tube,craniofacial structures, eyes, and abdominal wall wereobserved in Folr1 knockout mice exposed in uterus to fo-linic acid (Spiegelstein et al., 2004). However, this studyfailed to describe the cardiovascular phenotype of theFolr1 knockout mice. The plasma folate levels in Folr1�/�

mice are known to be �3-fold lower than in wild-typemice (Spiegelstein et al., 2003). Heterozygotes only exhib-ited a mild decrease in their plasma folate levels (Piedra-hita et al., 1999). Folr1 gene is highly expressed in the yolksac during early development, suggesting an importantrole of the Folr1 gene in embryonic folate supply (Saitsuet al., 2003). Therefore, the absence of the Folr1 gene in thepregnant dam may exacerbate the embryonic developmentof nullizygous embryos, and compromise the rescue effi-ciency of folate supplementation.The primary goal of this study was to characterize the

cardiovascular phenotypes and altered expression ofselected cardiac genes in Folr1�/� mice. We further reportthe results of our attempts to rescue these mice andrestore normal cardiac development by providing folatesupplementation to the pregnant dams. Our resultsdescribe a dose-related effect similar to that previouslyreported for NTDs (Spiegelstein et al., 2004). Finally, wereport the effect of the maternal genotype, that is, fornullizygotic dams, a higher dose of folate supplementa-tion was required to rescue their fetuses.

METHODS

Experimental Procedures

Animal husbandry. Animal experimentation wasapproved by the Institute of Biosciences and TechnologyAnimal Care Committee (IACUC). All mice were housedin clear polycarbonate micro-isolator cages, allowed freeaccess to water and food, and maintained on a 12-hlight/dark cycle in the Vivarium at the Institute of Bio-sciences and Technology in Houston, Texas. Folr1-defi-cient mice were generated by standard gene targetingmethodologies (Piedrahita et al., 1999). The Folr1 trans-gene was transferred by a speed congenics approach(Visscher, 1999) to the highly inbred LM/Bc genetic back-ground (L-Folr1), and was maintained by brother-sistermatings for at least 10 generations. Considering the em-bryonic lethality of the Folr1 knockout, mouse breedingstock were provided with a modified Clifford/Koury fo-late-deficient diet supplemented with 200 mg/kg folicacid and succinyl sulfathiazole (Dyets Inc., Bethlehem,PA) in order to produce viable nullizygous individuals.

258 ZHU ET AL.

Birth Defects Research (Part A) 79:257–268 (2007)

All live-born animals were provided access to regularchow (Dyets Inc.) after weaning. Both heterozygous andnullizygous embryos were generated by timed matingsbetween L-Folr1 heterozygous females and nullizygousmales. Folr1 nullizygous incrosses were also used inorder to evaluate the effect of maternal genotype. Mat-ings between wild type LM/Bc (Folr1þ/þ) mice were usedto generate control fetuses. The day on which the vaginalplug was detected was designated as E0.5 (Snell et al.,1948).

Genotyping. Genomic DNA was extracted from fetaltail tissue or embryonic yolk sacs using the PuregeneDNA extraction Kit (Gentra, Minneapolis, MN). Exon 2(Piedrahita et al., 1999) of the Folr1 gene was amplifiedusing the following primer pair: 50-AATGTCAAGGCTG-CATGTGG-30 and 50- CATTCCGATGTCATAGTTCCGC-30 in order to detect the wild-type Folr1. To detect themutant allele, the neo cassette was amplified using primerpair: 50-CTTGGGTGGAGAGGCTATTC-30 and 50-TGC-ATTCCGATGTCATAGTTCCG-30. The PCR conditionsincluded an initial denaturizing at 958C for 5 min, fol-lowed by 30 cycles of denaturing (958C for 1 min),annealing (608C for 1 min) and extension (728C for2 min), and a final extension at 728C for 10 min. PCRproduct was examined on 2% agarose gel under UVlight. The 179-bp and 1.2-kb products corresponded tothe wild-type and mutant alleles, respectively.

Folate rescue of Folr1 nullizygotes. In order to rescueFolr1�/� embryos from in utero death while retaining ab-normal cardiovascular developmental phenotypes, 6.25 mg/kg (6s) 5-formyl tetrahydrofolate (folinic acid), which isan active form of folate naturally found in food, wasadministered by gavage daily from E0.5 to E18.5 to preg-nant heterozygous mice. Similarly, in order to determinewhether maternal supplementation with folinic acid canactually rescue the cardiovascular phenotype, we per-formed a dose-response study with folinic acid (6.25,12.5, 25.0, 36, 44, and 66 mg/kg) and determined theimpact on Folr1 nullizygous cardiac development.

Morphological examination. The pregnant dams weresacrificed at E18.5 by CO2 inhalation, and the fetuseswere harvested for gross morphological evaluation. Theabdomens of the E18.5 dams were opened through rou-tine surgical procedures, and the gravid uteri wereremoved. The locations of all viable fetuses and resorp-tions were recorded, and the euthanized fetuses werecarefully examined for the presence of gross cardiacdefects, with particular attention to the potential presenceof aortic arch artery malformations. The tail of each fetuswas removed, and DNA was extracted to determine theFolr1 genotype by PCR analysis. Following the grossmorphological examination, E18.5 fetuses were fixed in10% formalin/PBS overnight. The heart was dissectedfree of the thorax and embedded in paraffin. Heart sec-tions (10 lM) were cut using a Leica 2030 microtome,and the sections were stained with Mayer’s hematoxylin-eosin. The histopathology sections were subsequentlyexamined under a Leica microscope (Bannockburn, IL).

Whole-mount in situ hybridization. Expression ofCrabp1, Nkx2-5, Tbx1, Fgf10, and Isl1 genes was examinedin E10.5 Folr1 nullizygous and wild-type embryos usingwhole-mount in situ hybridization. Embryos were col-lected in DEPC-treated PBS, fixed in 4% paraformalde-hyde (PFA) at 48C overnight, dehydrated through agraded methanol series (25, 50, 75, and 100%), and stored

at �208C until used. Prior to hybridization, the embryoswere bleached with 5% H2O2 and digested with protein-ase K. Hybridization was performed with digoxigenin(DIG)-labeled riboprobes at 638C overnight and detectedwith preabsorbed alkaline phosphatase–conjugated anti-DIG antibodies followed by reaction with nitrotetrazo-lium chloride and 5-bromo-4-chloro-3-indoly-phosphate.Stained embryos were subsequently postfixed with PFAand photographed digitally. Images were processedusing Adobe PhotoShop (Adobe Systems Inc., San Jose,CA).Reverse transcription PCR (RT-PCR). Total RNA was

extracted from E9.5, E10.5, and E12.5 embryonic heart tis-sue and whole embryos using a Picopure RNA IsolationKit (Acturus, CA). Five micrograms of total RNA wassubsequently reverse-transcribed into cDNA using HighCapacity cDNA Archive Kit (Applied Biosystems, FosterCity, CA). Expression of the Folr1 gene was detectedusing the following primers: 50-atg agt gtt ccc cga act tg-30 and 50-aca cag agc agc aga tgt cg-30, flanking exons 4,5, and 6 of the mouse Folr1 gene. The PCR conditionsincluded an initial denaturing step at 958C for 5 min, fol-lowed by 30 cycles of denaturing (958C for 30 s), anneal-ing (558C for 30 s), and extension (728C for 45 s), and afinal extension at 728C for 3 min. The mouse Gapdh genewas amplified as an internal control using the followingprimers: 50-cca tga caa ctt tgg cat tg-30 and 50-cct gct tcacca cct tct tg-30. PCR product was examined on 2% aga-rose gel and visualized under UV light. Agarose gel con-taining the PCR product of Folr1 was cut out using asterile razor blade. The PCR product was then purifiedfrom the agarose gel using the QIAEX II Gel ExtractionKit (Qiagen, Valencia, CA) and sequenced using the Big-Dye Terminator Kit on an ABI 3730 Genetic Analyzer(Applied Biosystems).Statistical analysis. Differences in frequencies of em-

bryonic abnormalities were assessed either by a v2 orFisher’s Exact test. The latter was also used to determinewhether or not different types of cardiovascular abnor-malities co-occur with neural tube closure failure. Teststatistics with p values < .05 were considered to be sig-nificant.

RESULTS

Folr1 Expression in Developing Heart

Expression of Folr1 in wild-type developing heart tissue(E9.5, E10.5, and E12.5) was detected using RT-PCR (Fig. 1).The 300-bp bands were visible in embryonic hearts in allthree stages during early development, corresponding tothe size of the PCR product given by the primer pair used.The bands in the lower part corresponded to the PCR prod-uct of Gapdh in each sample, which served as an internalcontrol. Sequencing results of RT-PCR products showeda complete match with the Folr1 cDNA (NM_008034)sequence, thereby confirming the presence of Folr1 mRNAin these tissue samples, leaving little doubt that the Folr1gene is expressed in developing heart tissue.

Pregnancy Outcome

When heterozygous dams were mated to nullizygousmales and supplemented with low dose s-folinic acid(6.25 mg/kg po/day) during pregnancy, there was a

259CARDIOVASCULAR ABNORMALITIES IN FOLR1 KNOCKOUT MICE

Birth Defects Research (Part A) 79:257–268 (2007)

definite departure from Mendelian expectations. Weobserved that 42.7% (90/211) of the total implantationswere resorbed, 43.1% (91/211) were live-born heterozy-gotes, and 13.3% (28/211) were live-born Folr1 nullizy-gotes that survived to term. The resorption rate declinedto 18.3% (28/153) when the dams were supplementedwith 36.0 mg/kg po/day folinic acid. At a higher level ofsupplementation (66 mg/kg po/day), the resorption ratewas reduced to 8.5% (5/59). No fetuses survived to E18.5when nullizygous females were mated to nullizygousmales without folate supplementation, or when folatesupplementation was limited to <36.0 mg/kg po/day. Atthis folinic acid supplementation level, the nullizygousdams still resorbed two-thirds of their implantations. Thesurviving nullizygous fetuses were developmentallydelayed, and 87.1% exhibited NTDs (Table 1).

Abnormal Heart Looping in the Progeny ofHeterozygous 3 Nullizygous Crosses

Supplemented with 6.25 mg/kg S-Folinic Acid

Under normal developmental conditions, the hearttube first loops between the conus cordis (future proxi-mal outflow tract) and bulbus cordis (future right ventri-cle), and bends ventrally and towards the right side togenerate a D-shaped heart. The second loop occursbetween the primitive atrium and the left ventricle andbends dorsally towards the left. When the Folr1 nullizy-gous embryos were evaluated at a midgestational timepoint (E10.5), 35.4% (17/48) of the nullizygous embryoshad significant cardiac looping defects. These included:

1. Reversed looping, in which the cono truncus islocated on the left side and the primitive atriumand sinus venosus are located on the right. Thismay later cause transposition of great arteries and/or isolated dextracardium.

2. Midline looping, in which the cono truncus and thebulbus cordis bend ventrally, this time towards themidline, therefore located cranially to the primitiveleft ventricle and atrioventricular canal.

3. Tighter looping, in which the conotruncal regionappears to be shorter than normal and the anglebetween conus cordis and ventricle chamber issharper than that observed in their heterozygouslittermates (Fig. 2).

Abnormal Expression of Neural Crest Cell (NCC)Marker Crabp1

In wild-type embryos, there were two separate streamsof cardiac NCCs entering branchial arches 3 and 4, whereasin the Folr1 nullizygous embryos, these streams fused intoa single stream of migrating cells (Fig. 3a,b). The signal in-tensity of the Crabp1 marker was also attenuated in thehomozygous mutant embryos compared to wild-type.

Down-Regulation of Isl1, Nkx2-5, Tbx1, and Fgf10in the Pharynx and Branchial Arches of Folr1

Nullizygous Embryos

The Isl1 expression was blocked in the branchial archesand dorsal root ganglia in Folr1 nullizygous embryos

Figure 1. Folr1 expression inmouse embryonic heart tissue(LM/Bc, E9.5, E10.5, and E12.5),RT-PCR: lane 1, E9.5 heart; lane 2,E9.5 embryo; lane 3, E10.5 heart;lane 4, E10.5 embryo; lane 5, E12.5heart; lane 6, E12.5 embryo; lane 7,negative control; and lane 8, 100-bp DNA ladder.

260 ZHU ET AL.

Birth Defects Research (Part A) 79:257–268 (2007)

(Fig. 3c,d). Nkx2-5 was expressed in the heart, pharynx,and branchial arches of wild-type embryos. The Folr1mutant embryos exhibited a significant reduction ofNkx2-5 expression in the pharyngeal arches and second-ary heart field, but no substantial reduction was observedin the abnormal looping heart (Fig. 3e,f). Tbx1 regulatescell contribution to the outflow tract by supporting cellproliferation in the secondary heart field. Tbx1 expressionis also required in cells expressing Nkx2-5 for the forma-tion of the aorto-pulmonary septum (Xu et al., 2004).Fgf10 is a direct target of Tbx1. We observed that Tbx1gene activity was significantly reduced in the pharyngealarches of Folr1 nullizygous embryos compared to thewild-type (Fig. 3g,h). Fgf10 also exhibited a vast reduc-tion of expression in pharyngeal arches of the Folr1 mu-tant embryos (Fig. 3i,j).

Spectrum of Cardiovascular Abnormalities WasObserved in Preterm Fetuses

Most of the cardiovascular phenotypes were detectableat E18.5 and included profound abnormalities affectingthe outflow tract and the aortic arch arteries.

Outflow tract defects in Folr1�/� fetuses (Fig. 4). Inthe experimental group supplemented with 6.25 mg/kgpo/day folinic acid, all 28 (100.0%) of the nullizygousfetuses had ventricular septal defects (VSDs). In addition,60.7% (17/28) of Folr1�/� fetuses had outflow tractdefects; 28.6% (8/28) of Folr1�/� fetuses had double-out-let right ventricle (DORV), which is characterized by thepresence of both the pulmonary trunk and the aorta aris-ing from the right ventricle. Three of 28 (10.7%) Folr1�/�

fetuses exhibited transposition of the great arteries(TGA), with the pulmonary trunk arising from the leftventricle and the aorta coming off the right ventricle. Wealso observed rightward displacement of persistent trun-cus arteriosus (PTA) in 10.7% (3/28) of the Folr1 nullizy-gotes, in which the aorta and pulmonary trunk did notseparate properly, forming a common trunk vessel origi-nating from the right ventricle. None of the aforemen-tioned malformations were observed in Folr1þ/� or wild-type fetuses (Table 2).

Aortic arch arteries anomalies in Folr1�/� mice (Fig. 5).In the group supplemented with 6.25 mg/kg po/day fo-linic acid, 50% (14/28) of Folr1�/� fetuses had malforma-tions in the aortic arch arteries. Five of 28 (17.9%) ofFolr1�/� fetuses had a right-sided aortic arch (RAA),right-sided ductus arteriosus (DA), and right-sided de-scending aorta, with or without a retro-esophageal rightsub-clavian artery (RSA). Three of 28 (10.7%) of L-Folr1�/�

fetuses had coarctation of the aortic arch or interruptedaortic arch (IAA). Complete interruption of the aorta, orIAA, is defined as a complete absence of a segment ofthe aortic arch. In Folr1�/� fetuses, interruption occurredmostly proximal to the origin of the left common carotidartery (LCC), with the descending aorta connected to thepulmonary artery by a large patent ductus arteriosus.Coarctation of aortic arch was defined as a narrowing ofthe lumen of the aorta that produces an obstruction toflow. Vascular ring and double aortic arch were observedin 14.3% (4/28) of Folr1�/� mice. None of these malfor-mations was seen in Folr1þ/� littermates or wild-typefetuses. In 10.7% (3/28) of nullizygotes, we observed lackof brachiacephalic trunk; however, this phenotype was

Tab

le1

Pregnan

cyOutcomes

ofFolr1

Miceunder

DifferentS-Folinic

Acid(SFA)Treatmen

ts

No.of

implants

No.of

resorptions

(%)*

No.offetuses

exam

ined

Noofcard

iac

malform

ations(%

)yNo.ofneu

ral

tubedefects

(%)

No.oforal

facial

clefts

(%)

No.ofab

dominal

walldefects

(%)

No.with

dextracardia

(%)

Heterozy

goticdam

Folr1

�/�(SFA

6.25

)21

190

(42.7)

2828

(100

.00)

22(78.6)

21(75.0)

11(39.3)

2(7.1)

Folr1

�/�(SFA

12.5)

119

24(20.2)

2413

(54.2)

20(83.3)

7(29.2)

3(12.5)

1(4.2)

Folr1

�/�(SFA

36.0)

153

28(18.3)

6816

(23.5)

13(19.1)

5(7.4)

02(2.9)

Folr1

�/�(SFA

66.0)

595(8.5)

301(3.3)

00

00

Nulldam

Folr1

�/�(SFA

36.0)

9362

(66.7)

3116

(51.6)

27(87.1)

13(41.9)

14(45.2)

0Folr1

�/�(SFA

66.0)

9125

(27.5)

6621

(31.8)

28(42.4)

2(3.0)

5(7.6)

0{F

olr1

þ/�

172

N/A

172

4(2.3)

00

00

§Folr1

þ/þ

362(5.6)

361(2.8)

00

00

SFA

inmg/kgbodyweight.

*v2test:heterozy

goticdam

:p<

.001;nulldam

:p<

.001.

y v2test:heterozy

goticdam

:p<

.001

;nulldam

:p<

.001.

{Folr1

þ/�fetuseswerefrom

theHet

Xnullmating(90from

SFA

6.25,57

from

theSFA

36.0

group,an

d25

from

SFA

60.0

group).

§Folr1

þ/þfetuseswerefrom

wild-typeto

wild-typematingwithnosu

pplemen

tation.

261CARDIOVASCULAR ABNORMALITIES IN FOLR1 KNOCKOUT MICE

Birth Defects Research (Part A) 79:257–268 (2007)

also detected in 2.3% of heterozygous littermates and2.8% of wild-type mice (Table 3).

Dextrocardia in Folr1–/– mice. Two of 28 (7.1%) Folr1�/�

fetuses exposed to low folinic acid treatment in uteropresented with dextrocardia (situs solitus reversus),which was not observed in heterozygous littermates orwild-type fetuses (Fig. 4). No other situs defects wereobserved.

Rescue of Cardiovascular Abnormalities byMaternal Supplementation of Folinic Acid

When heterozygous dams were mated to nullizygousmales, abnormal cardiovascular phenotypes could be res-cued by maternal folinic acid supplementation (Table 1).S-folinic acid at 66 mg/kg po daily can completely rescueexternal defects (NTDs, or facial clefts and abdominalwall defects) and outflow tract abnormalities, whereas 1/30 (3.3%) still presented with a right aortic arch malfor-mation. We also examined nullizygous fetuses from nulli-zygous to nullizygous matings to evaluate the effect ofthe maternal genotype on embryonic heart development.At low-dose folinic acid supplementation, all embryosdied in utero (100% resorption, data not shown). At 36mg/kg, 12/31 (38.7%) nullizygous fetuses from Folr1knockout dams had outflow tract defects, and 13/31(41.9%) had aortic arch artery defects. Frequencies ofboth defects were higher than those seen in fetuses fromthe Folr1 heterozygous dams (p < .05). At 66 mg/kg, 8/66 (12.1%) of the nullizygous embryos had outflow tractdefects and 20/66 (30.3%) had aortic arch artery defects(Tables 1–3).

Failure of Anterior Neural Tube Closure andCardiovascular Phenotypes

Table 4 depicts the statistical association between thefailure of anterior neural tube closure and the risk fordeveloping cardiovascular abnormalities. In order toincrease the numbers in each group, 6.25 mg/kg and12.5 mg/kg were combined together as a ‘‘low-dose’’group; 36 mg/kg was considered a ‘‘medium’’ dose,and 66 mg/kg was considered a ‘‘high’’ dose. Whenpartially rescued (medium dose/heterozygous dam andhigh dose/nullizygous dam), outflow tract defects weremore likely to occur when an anterior NTD (exence-phaly and encephalocele) was also present (Fisher’sExact test: p ¼ .001 and .008, respectively). Aortic archartery defects showed significant association with neuraltube closure failure when the Folr1 nullizygous damswere given a high dose of folinic acid (p ¼ .043). VSDsdid not show statistical association with anterior NTDs(p > .05).

DISCUSSION

The purpose of our study was to determine if inacti-vating the Folr1 gene by homologous recombinationcould compromise cardiovascular development, as previ-ous descriptions of this mouse model failed to addressheart development. As we observed in this study, all ofthe Folr1 nullizygous embryos supplemented in uterowith low levels of folinic acid presented at term withCHDs, resembling the most prevalent category of majorbirth defects in humans. Understanding the embryonic

Figure 2. Abnormal heart looping in L-Folr1 nullizygotes.E10.5: (a) normal looping; (b) midline looping: cono truncusand bulbus cordis bend ventrally but towards the midline,therefore located cranially to the primitive left ventricle andAV canal; (c) normal looped heart (left) and reverselylooped heart (right); (d) normal heart (left) and heart withshorter outflow tract and tighter loop (right). E11.5: (e) nor-mal heart; (f) reversed heart.

262 ZHU ET AL.

Birth Defects Research (Part A) 79:257–268 (2007)

Figure 4. Outflow track defects in L-Folr1 nullizygotes. (a,b) Normal heart; (c) persistent truncus arteriosus, dextracardia; (d) overridingaorta; (e,f) DORV; (g,h) transposition of great arteries, interrupted aorta arch.

Figure 3. Whole-mount in situ hybridization. (a,b) Expression of NCC marker Crabp1 demonstrates induction and migration but reduced in-tensity and an abnormal pattern in the Folr1 mutant. The distinct streams of NCCs in arches 3–4 were fused into one (white arrow) in nullembryos. Yellow arrow indicates open anterior neural tube in the Folr1 mutant. (c,d) Lateral view of E10.5 whole embryos. In the wild-typeembryo, Isl1 is expressed strongly in the future spinal cord and dorsal root ganglia (yellow arrows). Weaker expression was observed in pha-ryngeal arches (white arrows). In the null embryo, Isl1 expression in these regions is down-regulated. (e,f) Homogenous expression of Nkx2-5in the heart with no significant difference in signal level between wild-type and null embryos. Nkx2-5 is also present in pharyngeal arches(white arrows). A reduction of expression in the pharyngeal arches was observed in the null embryo. (g,h) Tbx1 expression in pharyngealarches. White arrow, 1st arch; yellow arrow, 2nd arch; green arrow, presumptive 4th/6th arches; red arrow, otic vesicle. Compared to thewild-type embryo, Tbx1 expression is down-regulated and has a different expression pattern in null embryos. In the 1st and 2nd pharyngealarches, the stream of Tbx1 expression in null embryo is shorter. In the presumptive 4th/6th arches, there is no stream of Tbx1 expression. (i,j)Fgf10 expression in the pharyngeal arches: white arrows show the reduction of Fgf10 expression in null embryos. The majority of the reduc-tion occurred in the 2nd pharyngeal arch (white arrows). Reduction of Fgf10 expression is also seen in the 3rd arch (yellow arrows). Theexpression pattern appeared to be less condensed in the null embryo than in the wild-type embryo. Red arrow: otic vesicle.

263CARDIOVASCULAR ABNORMALITIES IN FOLR1 KNOCKOUT MICE

Birth Defects Research (Part A) 79:257–268 (2007)

mechanisms underlying these abnormal heart defects willfacilitate the design of preventive measures and novelintervention strategies. The Folr1 knockout mouse canserve as a unique tool for studying folate-responsiveCHDs and the molecular mechanisms through whichadequate folate provides protection against heart malfor-mations. From the data presented herein, it is clear thatcardiovascular phenotypes in Folr1 knockout mice have aspecific embryonic pathogenesis, likely involving the car-diac NCCs, as well as myocardium cells from the second-ary heart field. The phenotypes mimic cardiac NCC abla-tion studies in chicken embryos. Tang and coworkers(2004) reported apoptosis and cell death at later-stage(E12.5) nullizygous hearts, possibly reflecting pathologi-cal consequences of abnormal folate homeostasis in thesecell populations.

Heart morphogenesis is a complex developmental pro-cess beginning with the formation of the tubular heartfrom cells of the cardiac crescent located in anterior lat-eral mesoderm (Perea-Gomez et al., 2001). Looping of theheart transforms the single heart tube to a more complexstructure with two atria and two ventricles and normallystarts at about E9.0 in mouse embryos (17 somite stage).During normal cardiac looping, the heart tube undergoesa rapid elongation in which cells are added to the grow-ing poles (Kelly and Buckingham, 2002; Kelly, 2005). Thestraight heart tube bends into a C-shape with the convexside pointed to the right. The proximal part of the inflowtract subsequently moves cranially towards the distalpart of the outflow tract and repositions from an initialcaudal position to a cranial position and nestles behindthe ventrally displaced outflow tract. This process is

Table 2S-Folinic Acid (SFA) Rescues the Outflow Tract Phenotypes in Folr1 Nullizygotic Fetuses (E18.5)

No. offetuses

examined

Outflow tract defects

No. offetuses with

OFT defects (%)*No. of

VSDs (%)yNo. of

PTAs (%)No. of

DORVs (%)No. of

TGAs (%)

No. ofoverridingaorta (%)

Heterozygotic damFolr1�/� (SFA 6.25) 28 17 (60.7) 28 (100.0) 3 (10.7) 8 (28.6) 3 (10.7) 3 (10.7)Folr1�/� (SFA 12.5) 24 5 (20.8) 4 (16.7) 1 (4.2) 2 (8.3) 0 4 (16.7)Folr1�/� (SFA 36.0){ 68 8 (11.8) 2 (2.9) 0 2 (2.9) 4 (5.9) 1 (1.5)Folr1�/� (SFA 66.0)§ 30 0 0 0 0 0 0

Null damFolr1�/� (SFA 36.0){ 31 12 (38.7) 12 (38.7) 2 (6.5) 8 (25.8) 2 (6.5) 1 (3.2)Folr1�/� (SFA 66.0)§ 66 8 (12.1) 4 (6.1) 2 (3.0) 3 (4.5) 2 (3.0) 1 (1.5)Folr1þ/� 21 0 0 0 0 0 0

Folr1þ/� fetuses were from the Het 3 null mating (90 from SFA 6.25, 57 from the SFA 36.0 group, and 25 from SFA 60.0 group).SFA in mg/kg body weight.*v2 test: heterozygotic dam: p < .001; null dam: p ¼ .003.yv2 test: heterozygotic dam: p < .001; null dam: p < .001.{Fisher’s Exact test: OFT defects: p ¼ .001; VSD: p ¼ .000.§Fisher’s Exact test: OFT defects: p ¼ .054; VSD: p ¼ .03.

Figure 5. Aortic arch artery abnormalities. (a) Normal alignment; (b) right aortic arch; (c) interrupted aortic arch (IAA, complete); (d)vascular ring (dorsal view); (e) right subclavian artery arises from the pulmonary trunk.

264 ZHU ET AL.

Birth Defects Research (Part A) 79:257–268 (2007)

referred to as convergence, as described in developingchicken embryos (Flynn et al., 1999). Abnormal loopingand convergence may be the result of an absence of thecervical/cranial flexure; cardiac neural crest ablation alsocan cause an incorrect convergence process (Manneret al., 1993). Looping defects were observed in neuralcrest–ablated chicken embryos, and the characteristics ofabnormal looping could be explained by decreased incor-poration of myocardial cells from secondary heart field,resulting in a shorter outflow tract (Yelbuz et al., 2002).We have observed shorter and straighter outflow limb inFolr1 knockout embryos, similar to those seen in chickenembryos after neural crest ablation; therefore, it may becausally related to the mal-alignment of the inflow andoutflow segments, and outflow tract defects of the defini-tive hearts.

One-third of human CHDs affect the outflow tractregion (Srivastava and Olson et al., 2000). We observed aseries of outflow tract defects in folinic acid–rescuedFolr1�/� fetuses, including: TGA, rightward PTA, DORV,and overriding aorta. Two populations of cells are knownto contribute to development of the outflow tract. CardiacNCCs are derived from the neural folds prior to fusionin the region of the developing hindbrain, migratingalong the caudal aortic arch arteries, to reach the cardiacoutflow tract and the developing aortic arch arteries(Kirby et al., 1993; Boot et al., 2003b). Cardiac NCCs thendifferentiate into smooth muscle cells, forming the tunica

media of the great arteries and the ascending aorta. Stud-ies in chicken embryos suggested that early-migratingcardiac NCCs largely invest the aorticopulmonary sep-tum and outflow tract, while the late-migrating cardiacNCCs are more restricted to the proximal part of the aor-tic arch arteries (Boot et al., 2003a). A series of concurrentstudies reported that myocardial wall of the outflow tractis formed from progenitor cells situated in the pharyn-geal mesoderm, referred to as anterior heart field (Mjaat-vedt et al., 2001; Kelly et al., 2001) or secondary heartfield (SHF) (Waldo et al., 2001) in both chicken andmouse embryos.Cardiovascular phenotypes involving outflow tract,

aorticopulmonary septum, and aortic arch arteriesobserved in Folr1 nullizygous fetuses were consistentwith the consequence of the ablation or reduction inNCC precursors. We observed reduced expression leveland an abnormal pattern of the NCC marker, Crabp1, inthe 3rd, 4th, and 6th pharyngeal arches in E10 Folr1 nulli-zygous embryos. In a previous study, we found a sub-stantial reduction in the expression of the Pax-3 gene,which had been viewed by some as a reasonable surro-gate marker for premigratory NCCs, in the same regionin Folr1 knockout embryos on the SWV/FNN back-ground. Taken together, the NCC population appeared tobe severely affected by inactivating the Folr1 gene (Tanget al., 2004). More detailed studies investigating the spa-tial and temporal expression pattern of NCC markers are

Table 3S-Folinic Acid (SFA) Rescues the Aortic Arch Artery Phenotypes in Folr1 Nullizygotic Fetuses (E18.5)

No. offetuses

examined

Aortic arch artery defects

No. of fetuseswith aortic arch

artery defects (%)*

No. of rightaortic

arches (%)

No. of IAAs/coarctation ofaorta (%)

No. of aortarings or double

aorta (%)

No. of missingbrachiacephalic

trunks (%)

Heterozygotic damFolr1�/� (SFA 6.25) 28 14 (50.0) 5 (17.9) 3 (10.7) 4 (14.3) 3 (10.7)Folr1�/� (SFA 12.5) 24 11 (45.8) 5 (20.8) 0 3 (12.5) 2 (8.3)Folr1�/� (SFA 36.0)y 68 11 (16.2) 5 (7.4) 0 4 (5.9) 2 (2.9)Folr1�/� (SFA 66.0){ 30 1 (3.3) 1 (3.3) 0 1 (3.3) 0

Null damFolr1�/� (SFA 36.0)y 31 13 (41.9) 10 (32.3) 0 2 (6.5) 3 (9.7)Folr1�/� (SFA 66.0){ 66 20 (30.3) 4 (6.1) 2 (3.0) 6 (9.1) 11 (16.7)Folr1þ/� 172 0 0 0 0 4 (2.3)Folr1þ/þ 36 0 0 0 0 1 (2.8)

Folr1þ/� fetuses were from the Het 3 null mating (90 from SFA 6.25, 57 from the SFA 36.0 group, and 25 from SFA 60.0 group);Folr1þ/þ fetuses were from wild-type to wild-type mating with no supplementation.SFA in mg/kg body weight.

*v2 test: heterozygotic dam: p < .001; null dam: p ¼ .08.yFisher’s Exact test: aortic arch defects: p ¼ .007.{Fisher’s Exact test: aortic arch defects: p ¼ .016.

Table 4Association between Cardiovascular Abnormalities and Failure of Anterior Neural Tube Closure

S-folinic aciddosage

Maternalgenotype

No. offetuses

No. of VSDs Outflow tract defects Aortic arch artery defects

OR (95% CI) p OR (95% CI) p OR (95% CI) p

Low þ/� 52 0.70 (0.16, 3.09) .729 1.93 (0.44, 8.49) .488 1.02 (0.25, 4.17) 1.000Medium þ/� 68 NA NA 32.5 (3.35, 315.34) .001 0.80 (0.09, 7.50) 1.000High �/� 30 NA NA NA NA NA NAMedium �/� 31 1.79 (0.29, 11.13) .676 1.79 (0.29, 11.13) .676 3.60 (0.37, 34.94) .379High �/� 66 4.44 (0.44, 45.12) .304 12.33 (1.42, 107.23) .008 4.67 (1.11, 19.62) .043

Bold numbers represent significance at the p value indicated.

265CARDIOVASCULAR ABNORMALITIES IN FOLR1 KNOCKOUT MICE

Birth Defects Research (Part A) 79:257–268 (2007)

warranted in order to understand how NCCs are affectedin these genetically manipulated animals and, moreimportantly, are rescued by folate supplementation. Inaddition to outflow tract defects, we also observed vari-ous aortic arch artery patterning defects, including theright aorta arch, vascular ring, interrupted aorta arch,double aortic arch, and others. Neural crest-derived cellsare known to be also involved in the patterning of thesevessels (Yutzey and Kirby, 2002). Evidence suggests thatthe remodeling process is affected by accumulatingNCCs that migrate through the pharyngeal arches(Bockman et al., 1999; Kirby, 1990; Kirby and Waldo,1990, 1995). However, the mechanisms underlying thisprocess are still unclear, possibly involving NCC migra-tion, survival, proliferation, or interaction with local envi-ronment. Studies in avian embryos have suggested thatNCCs are essential for the persistence of aortic archarteries but not for their formation (Waldo et al., 1996).

Defective SHF in Folr1�/� mutants may be secondaryto cardiac NCC disruption: the SHF is known as a keycontributor of cardiogenic precursor cells to the outflowtract and right ventricle (Kelly and Buckingham, 2002;Waldo et al., 2001), providing both myocardium andsmooth muscle to the developing heart (Waldo et al.,2005a). Isl1 is required for development of the medialsplanchnic mesoderm from which FGF-expressing SHFcells arise and plays an early role in specifying the SHF(Cai et al., 2003). Isl1-expressing cells have been found tocontribute to both posterior and anterior segments of thedeveloping heart. Isl1 and other signals induce splanch-nic mesodermal cells to an outflow tract (OFT) myocar-dial fate. Cells in the SHF also express Nkx2-5 and Gata4,as the cardiac progenitor cells from the primary heartfield. Isl1 expression is substantially reduced in Folr1knockout embryos. The homeobox gene Nkx2-5 is the ear-liest known marker of vertebrate heart development(Komura and Izumo, 1993; Lints et al., 1993). In additionto the heart, Nkx2-5 is also expressed in developing pha-ryngeal arches, spleen, thyroid, stomach, and tongue.When we examined Nkx2-5 expression patterns in Folr1nullizygous fetuses, it was apparent that it is significantlydecreased in the pharyngeal arches and SHF (Fig. 3c,d).Xu et al. (2004) recently reported that Tbx1, a T-box tran-scription factor, is expressed in mouse SHF. Tbx1 regu-lates morphogenesis of aorticopulmonary septum as wellas cell proliferation in the SHF region. Tbx1 maintainscell contribution to the OFT at a sufficiently high level tosupport growth and remodeling of the OFT. In mouse,loss of Tbx1 function causes hypoproliferation in the SHFand hypoplasia of outflow tract myocardium. Fgf10marks a distinct population of splanchnic mesodermalcells in the SHF (von Both et al., 2004). Fgf10 is a directtarget of Tbx1. In Folr1 null embryos, Tbx1 and Fgf10expression was reduced in the pharynx and pharyngealarches. Therefore, in Folr1-null embryos, changes ingenetic pathways regulating SHF function could contrib-ute to the cardiovascular abnormalities observed in thedefinitive heart.

The Folr1 gene is expressed at low levels in premigra-tory NCCs as well as heart tissue during early and midg-estation periods, suggesting its direct role in heart devel-opment. The Folr1 knockout mice have low folate andhigh homocysteine levels and presented with a spectrumof external defects, including NTDs, orofacial clefts, andocular and abdominal wall defects, attributed to affected

NCC proliferation or migration abnormalities (Piedrahitaet al., 1999; Spiegelstein et al., 2004). Congenital heartdefects and NTDs coexist in a number of animal modelssuch as Splotch, Looptail, Cited2, and others. Homocysteineis involved in the etiology of both CHDs and NTDs(Rosenquist et al., 1996). Dietary folate deficiency affectsboth heart development and neurulation. Both externaldefects and cardiovascular phenotypes in Folr1 knockoutmice support the hypothesis that migrating NCCs haveelevated demand for folate, and these cells are specifi-cally vulnerable to folate deficiency and/or increased ho-mocysteine concentrations. In the present study, weobserved abnormal NCC migration patterns and possiblesubsequent impairment of the SHF myocardium.A number of transcription factors and signaling mole-

cules expressed in the neural crest and endocardiumhave been implicated in regulating endocardial cushionand outflow tract development (Epstein and Stoller,2005), including Wnt, Notch, and Semaphorin signaling,as well as specific adhesion molecules. The Folr1 geneexhibited an overlapping expression pattern with theWnt1 gene, suggesting a direct role of Folr1 in NCC func-tion (Parr et al., 1993; Saitsu et al., 2003). NCC defectscontribute to the NTDs in Folr1 knockout mice throughmechanisms involving dysregulation of Shh, Pax3, andother critical genes (Tang and Finnell, 2003). Recently,Shh was found to contribute to arch artery and outflowtract patterning defects (Washington et al., 2005). There-fore, it is possible that Folr1 gene ablation affects cardiacNCCs and heart development through Shh and other sig-naling molecules. Our data demonstrated that outflowtract defects in Folr1 nullizygotes were statistically associ-ated with a failure of anterior neural tube closure(Fisher’s exact test: p < .05), raising the question whetheror not these defects are secondary to the failure of neuraltube closure. The association between NTDs and VSDsappeared to be fairly weak (p > .05). Even though theventricular septum development coincides with earlymigrating NCCs, the defective development of the ven-tricular septum may be due to disruption of other eventsduring embryogenesis, such as failure of myocardial cellmigration from SHF or abnormal chamber alignment,and may be attributed less to the completion of anteriorneural tube closure.One unanswered question of interest is whether there

is severe cardiac dysfunction in Folr1 knockout animals,because the observed structural defects of the heart alonemay not be sufficient to produce the high level of prena-tal mortality observed in the nullizygous mice. In chickembryos, ablation of the cardiac neural crest also causessigns of severe cardiac dysfunction that appear to befrom impaired function at the myocyte level (Waldoet al., 1999, 2005b). NCC malformations in Splotch mice(Sp2H/Sp2H) led to depressed myocardial function anddeath in utero from cardiac failure (Conway et al., 1997;Creazzo et al., 1998). Therefore, if NCCs were affected inL-Folr1�/� mice, impaired myocardial function, myocar-dium maturation, or endocardial development may haveexisted and possibly caused early embryonic death inthese mice. Further experimentation to elucidate anypotential defective cardiac function in Folr1 knockoutmice is warranted.Experiments using the Folr1 knockout mouse demon-

strated that both outflow tract defects and aortic arch ar-tery malformations can be rescued by maternal supple-

266 ZHU ET AL.

Birth Defects Research (Part A) 79:257–268 (2007)

mentation with s-folinic acid in a dose-dependent man-ner. The Folr1 knockout mouse was generated as a folatetransport deficiency model, and early work documentedonly the external phenotypes, including exencephaly,orofacial clefts, and abdominal wall defects (Piedrahitaet al., 1999; Spiegelstein et al., 2004). The present studyshowed that cardiovascular abnormalities occur at a highfrequency in these animals, and that they can be rescuedby maternal folinic acid supplementation. The maternalgenotype has significant impact on the pregnancy out-come as well as the rescue efficiency. If the pregnantdam is nullizygous for the ablated Folr1 allele, higherdosages of s-folinic acid are required in order to rescuethe fetuses. This finding is consistent with previousknowledge that in utero nutrition plays an important rolein embryonic development. Folr1 is highly expressed inthe placenta and yolk sac (Saitsu et al., 2003). Pregnantmice lacking the Folr1 gene did not transport adequatequantities of folate through the yolk sac and the placenta,thereby exposing the embryos to a folate-deficient in uteroenvironment.

Other than transporting folate into the cell, the folate re-ceptor alpha may also act as a signaling molecule throughinteraction with other factors and plays an important rolein embryonic development through as yet poorly under-stood mechanisms. The folate receptor gene was alsoreported to interact with other signaling molecules func-tionally in the cell proliferation of ovarian carcinoma(Miotti et al., 2000), suggesting that the folate receptor mayhave a cell signaling role during embryogenesis. Whensupplemental folic acid is provided to the cell, critical pro-teins associated with the folate receptor are transportedinto the cell’s cytoplasm, where they provide substrates forcritical signal transduction reactions. Thus, folic acid maybenefit developing embryos not only by maintainingproper methylation status to ensure normal patterns ofgene expression, but it may also contribute to the mainte-nance of critical signal transduction reactions that wouldotherwise be compromised by folate deficiency. Futurestudy will focus on cellular and molecular mechanismsunderlying the cardiovascular defects in the Folr1 knockoutmice. Differentially expressed gene analysis of developingheart and pharyngeal arch tissues will provide importantclues pointing to the biological and molecular changes thatlead to the observed cardiovascular phenotypes in theFolr1 knockout mice. Furthermore, conditional knockoutstargeting cardiac progenitor cells and cardiac NCCs will beexcellent tools by which to determine functional roles ofFolr1 in cardiovascular system development.

ACKNOWLEDGMENTSWe would like to thank Dr. James Martin at the Insti-

tute of Biosciences and Technology, Texas A&M Univer-sity Health Science Center, who generously provided themouse Isl1 probe. We would also like to thank Dr. Huan-sheng Xu and Dr. Antonio Baldini, at the Institute of Bio-sciences and Technology, Texas A&M University HealthScience Center, for their kind help with mouse heartmorphological examinations.

REFERENCES

Bailey LB, Berry RJ. 2005. Folic acid supplementation and the occurrenceof congenital heart defects, orofacial clefts, multiple births, and mis-carriage. Am J Clin Nutr 81:1213S–1217S.

Baldini A. 2005. Dissecting contiguous gene defects: TBX1. Curr OpinGenet Dev 15:279–284.

Berry RJ, Li Z, Erickson JD, et al. 1999. Prevention of neural tube defectswith folic acid in China. N Engl J Med 341:1485–1490.

Bockman DE, Redmond ME, Kirby ML. 1999. Altered development ofpharyngeal arch vessels after neural crest ablation. Ann NY Acad Sci588:296–304.

Boot MJ, Gittenberger-De Groot AC, Van Iperen L, et al. 2003a. Spatio-temporally separated cardiac neural crest subpopulations that targetthe outflow tract septum and pharyngeal arch arteries. Anat Rec275A:1009–1018.

Boot MJ, Steegers-Theunissen RP, Poelmann RE, et al. 2004. Cardiac out-flow tract malformations in chick embryos exposed to homocysteine.Cardiovasc Res 64:365–373.

Boot MJ, Steegers-Theunissen RP, Poelmann RE, et al. 2003b. Folic acidand homocysteine affect neural crest and neuroepithelial cell out-growth and differentiation in vitro. Dev Dyn 227:301–308.

Botto LD, Khoury MJ, Mulinare J, et al. 1996. Periconceptional multivita-min use and the occurrence of conotruncal heart defects: results froma population-based, case-control study. Pediatrics 98:911–917.

Botto LD, Mulinare J, Erickson JD. 2000. Occurrence of congenital heartdefects in relation to maternal multivitamin use. Am J Epidemiol151:878–884.

Botto LD, Mulinare J, Erickson JD. 2003. Do multivitamin or folic acidsupplements reduce the risk for congenital heart defects? Evidenceand gaps. Am J Med Genet 121A:95–101.

Burgoon JM, Selhub J, Nadeau M, et al. 2002. Investigation of the Effectsof Folate Deficiency on Embryonic Development Through the Estab-lishment of a Folate Deficient Mouse Model. Teratology 65:219–227.

Cai CL, Liang X, Shi Y, et al. 2003. Isl1 Identifies a Cardiac ProgenitorPopulation that Proliferates Prior to Differentiation and Contributes aMajority of Cells to the Heart. Dev Cell 5:877–889.

Conway SJ, Godt RE, Hatcher CJ, et al. 1997. Neural crest is involved indevelopment of abnormal myocardial function. Mol Cell Cardiol29:2675–2685.

Creazzo TL, Godt RE, Leatherbury L, et al. 1998. Role of cardiac neural crestcells in cardiovascular development. Annu Rev Physiol 60:267–286.

Czeizel AE. 1998. Periconceptional folic acid containing multivitamin sup-plementation. Eur J Obstet Gynecol Reprod Biol 78:151–161.

Czeizel AE, Dudas I. 1992. Prevention of the first occurrence of neural-tube defects by periconceptional vitamin supplementation. N Engl JMed 327:1832–1835.

Czeizel AE, Toth M, Rockenbauer M. 1996. Population-based case controlstudy of folic acid supplementation during pregnancy. Teratology53:345–351.

Epstein JA, Stoller JZ. 2005. Cardiac neural crest. Semi Cell Dev Biol16:704–715.

Flynn ME, Pikalow AS, Kimmelman RS, et al. 1999. The mechanism ofcervical flexure formation in the chick. Anat Embryol 184:411–420.

Hobbs CA, James SJ, Parsian A, et al. 2006. Congenital heart defects andgenetic variants in the methylenetetrahydrofolate reductase gene. JMed Genet 43:162–166.

Kamen BA, Wang MT, Streckfuss AJ, et al. 1988. Delivery of folates to thecytoplasm of MA104 cells is mediated by a surface membrane recep-tor that recycles. J Biol Chem 263:13602–13609.

Kelly RG. 2005. Molecular inroads into the anterior heart field. TrendsCardio Med 15:51–56.

Kelly RG, Brown NA, Buckingham ME. 2001. The arterial pole of themouse heart forms from Fgf10-expressing cells in pharyngeal meso-derm. Dev Cell 1:435–440.

Kelly RG, Buckingham ME. 2002. The anterior heart-forming field: voyageto the arterial pole of the heart. Trends Genet 18:210–216.

Kirby ML. 1990. Alteration of cardiogenesis after neural crest ablation.Ann NY Acad Sci 588:289–295.

Kirby ML, Kumiski DH, Myers T, et al. 1993. Back transplantation ofchick cardiac neural crest cells cultured in LIF rescues heart develop-ment. Dev Dyn 198:296–311.

Kirby ML, Waldo KL. 1990. Role of neural crest in congenital heart dis-ease. Circulation 82:332–340.

Kirby ML, Waldo KL. 1995. Neural crest and cardiovascular patterning.Circ Res 77:211–215.

Komuro I, Izumo S. 1993. Csx: a murine homeobox-containing gene spe-cifically expressed in the developing heart. Proc Nat Acad Sci USA90:8145–8149.

Li D, Pickell L, Liu Y, et al. 2005. Maternal methylenetetrahydrofolate re-ductase deficiency and low dietary folate lead to adverse reproduc-tive outcomes and congenital heart defects in mice. Am J Clin Nutr82:188–195.

Lints TJ, Parsons LM, Hartley L, et al. 1993. Nkx-2.5: a novel murinehomeobox gene expressed in early heart progenitor cells and theirmyogenic descendants. Development 119:419–431.

267CARDIOVASCULAR ABNORMALITIES IN FOLR1 KNOCKOUT MICE

Birth Defects Research (Part A) 79:257–268 (2007)

Manner J, Siedl W, Steding G. 1993. Correlation between the embryonichead flexures and cardiac development. An experimental study inchick embryos. Anat Embryol 188:269–285.

Matherly LH, Goldman DI. 2003. Membrane transport of folates. VitamHorm 66:403–456.

Miotti S, Bagnoli M, Tomasseti A, et al. 2000. Interaction of folate receptorwith signaling molecules lyn and G(alpha)(I-3) in detergent-resistantcomplexes from the ovary carcinoma cell line IGROV1. J Cell Sci113:349–357.

Mjaatvedt CH, Nakaoka T, Moreno-Rodriguez R, et al. 2001. The OutflowTract of the Heart Is Recruited from a Novel Heart-Forming Field.Dev Biol 238:97–109.

MRC Vitamin Study Research Group. 1991. Prevention of neural tubedefects: results of Medical Research Council Vitamin Study. Lancet338:131–137.

Nelson MM. 1960. Teratogenic effects of pteroylglutamic acid deficiencyin the rat. CIBA Foundation Symp Cong Malfs 1960:1324–1357.

Nelson MM, Asling CW, Evans HM. 1952. Production of multiple congen-ital abnormalities in young by maternal pteroylglutamic acid defi-ciency during gestation. J Nutr 48:61–80.

Parr BA, Shea MJ, Vassileva G, et al. 1993. Mouse Wnt genes exhibit dis-crete domains of expression in the early embryonic CNS and limbbuds. Development 119:247–261.

Pei L, Zhu H, Zhu J, et al. 2005. Genetic Variation of Infant Reduced Fo-late Carrier (A80G) and Risk of Orofacial Defects and CongenitalHeart Defects in China. Ann Epidemiol 16:352–356.

Perea-Gomez A, Rhinn M, Ang SL. 2001. Role of the anterior visceralendoderm in restricting posterior signals in the mouse embryo. Int JDev Biol 45:311–320.

Piedrahita JA, Oetama B, Bennett GD, et al. 1999. Mice lacking the folicacid-binding protein Folr1 are defective in early embryonic develop-ment. Nat Genet 23:228–232.

Reamon-Buettner SM, Borlak J. 2004. Somatic NKX2–5 mutations as anovel mechanism of disease in complex congenital heart disease. JMed Genet 41:684–690.

Rittler M, Lopez-Camelo J, Castilla EE. 2004. Monitoring congenital rubellaembryopathy. Birth Defects Res A Clin Mol Teratol 70:939–943.

Rosenquist TH, Ratashak SA, Selhub J. 1996. Homocysteine induces con-genital defects of the heart and neural tube: effect of folic acid. ProcNatl Acad Sci USA 93:15227–15232.

Saitsu H, Ishibashi M, Nakano H, et al. 2003. Spatial and temporal expres-sion of folate-binding protein 1 (Fbp1) is closely associated with ante-rior neural tube closure in mice. Dev Dyn 226:112–117.

Scanlon KS, Ferencz C, Loffredo CA, et al. 1998. Preconceptional folateintake and malformations of the cardiac outflow tract. Baltimore-Washington Infant Study Group. Epidemiology 9:95–98.

Shaw GM, O’Malley CD, Wasserman CR, et al. 1995. Maternal pericon-ceptional use of multivitamins and reduced risk for conotruncal heartdefects and limb deficiencies among offspring. Am J Med Genet 59:536–545.

Shaw GM, Zhu H, Lammer EJ, et al. 2003. Genetic variation of infantreduced folate carrier (A80G) and risk of orofacial and conotruncalheart defects. Am J Epidemiol 158:747–752.

Snell GD, Fekete E, Hummel KP. 1948. The relation of mating, ovulation, andestrus smear in the house mouse to the time of day. Anat Rec 76:30–54.

Spiegelstine O, Lu X, Le CX, et al. 2003. Effects of dietary folate intakeand folate binding protein-1 (Folbp1) on urinary speciation of sodiumarsenate in mice. Toxic Lett 145:167–174.

Spiegelstein O, Mitchell LE, Merriweather MY, et al. 2004. Embryonic de-velopment of folate binding protein-1 (Folbp1) knockout mice: Effectsof the chemical form, dose, and timing of maternal folate supplemen-tation. Dev Dyn 231:221–231.

Srivastaba D, Olson EN. 2000. A genetic blueprint for cardiac develop-ment. Nature 407:221–226.

Tang LS, Finnell RH. 2003. Neural and orofacial defects in Folbp1 knock-out mice. Birth Defects Res A Clin Mol Teratol 67:209–218.

Tang LS, Santillano DR, Wlodarczyk BJ, et al. 2005. Role of Folbp1 in theregional regulation of apoptosis and cell proliferation in the develop-ing neural tube and craniofacies. Am J Med Genet C Semin MedGenet 135:48–58.

Tang LS, Wlodarczyk BJ, Santillano DR, et al. 2004. Developmental conse-quences of abnormal folate transport during murine heart morpho-genesis. Birth Defects Res A Clin Mol Teratol 70:449–458.

Visscher PM. 1999. Speed congenics: accelerated genome recovery usinggenetic markers. Genet Res 74:81–85.

von Both I, Silvestri C, Erdemir T, et al. 2004. Foxh1 is essential for devel-opment of the anterior heart field. Dev Cell 7:331–345.

Waldo KL, Hutson MR, Stadt HA, et al. 2005a. Cardiac neural crest isnecessary for normal addition of the myocardium to the arterial polefrom the secondary heart field. Dev Biol 281:66–77.

Waldo KL, Hutson MR, Ward CC, et al. 2005b. Secondary heart field con-tributes myocardium and smooth muscle to the arterial pole of thedeveloping heart. Dev Biol 281:78–90.

Waldo KL, Kumiski D, Kirby ML. 1996. Cardiac neural crest is essentialfor the persistence rather than the formation of an arch artery. DevDyn 205:281–292.

Waldo KL, Kumiski DH, Wallis KT, et al. 2001. Conotruncal myocardiumarises from a secondary heart field. Development 128:3179–3188.

Warkany J, Nelson RC. 1940. Appearance of skeletal abnormalities in theoffspring of rats reared on a deficient diet. Science 31:188–205.

Washington SI, Byrd NA, Abu-Issa R, et al. 2005. Sonic hedgehog isrequired for cardiac outflow tract and neural crest cell development.Dev Biol 283:357–372.

Werler MM, Hayes C, Louik C, et al. 1999. Multivitamin supplementationand risk of birth defects. Am J Epidemiol 150:675–682.

Williams LJ, Correa A, Rasmussen S. 2004. Maternal lifestyle factors andrisk for ventricular septal defects. Birth Defects Res A Clin Mol Tera-tol 70:59–64.

Xu H, Morishima M, Wylie JN, et al. 2004. Tbx1 has a dual role in themorphogenesis of the cardiac outflow tract. Development 131:3217–3227.

Yelbuz TM, Waldo KL, Kumiski DH, et al. 2002. Shortened outflow tractleads to altered cardiac looping after neural crest ablation. Circula-tion 106:504–510.

Yutzey KE, Kirby ML. 2002. Wherefore heart thou? Embryonic origins ofcardiogenic mesoderm. Dev Dyn 223:307–320.

268 ZHU ET AL.

Birth Defects Research (Part A) 79:257–268 (2007)