Ó 2006 Wiley-Liss, Inc.
Birth Defects Research (Part A) 76:544–552 (2006)
Integrity of the Methylation Cycle Is Essential for
Mammalian Neural Tube Closure
Louisa P.E. Dunlevy,1 Katie A. Burren,1 Kevin Mills,2 Lyn S. Chitty,3
Andrew J. Copp,1 and Nicholas D.E. Greene1*
2
1
Neural Development Unit, Institute of Child Health, University College London, United Kingdom
Biochemistry, Endocrinology and Metabolism Unit, Institute of Child Health, University College London, United Kingdom
3
Clinical and Molecular Genetics Unit, Institute of Child Health, University College London, United Kingdom
Received 6 March 2006; Revised 30 May 2006; Accepted 22 June 2006
BACKGROUND: Closure of the cranial neural tube during embryogenesis is a crucial process in development
of the brain. Failure of this event results in the severe neural tube defect (NTD) exencephaly, the developmental forerunner of anencephaly. METHODS: The requirement for methylation cycle function in cranial neural
tube closure was tested by treatment of cultured mouse embryos with cycloleucine or ethionine, inhibitors of
methionine adenosyl transferase. Embryonic phenotypes were investigated by histological analysis, and immunostaining was performed for markers of proliferation and apoptosis. Methylation cycle intermediates s-adenosylmethionine and s-adenosylhomocysteine were also quantitated by tandem mass spectrometry. RESULTS:
Ethionine and cycloleucine treatments significantly reduced the ratio of abundance of s-adenosylmethionine to
s-adenosylhomocysteine and are, therefore, predicted to suppress the methylation cycle. Exposure to these
inhibitors during the period of cranial neurulation caused a high incidence of exencephaly, in the absence of
generalized toxicity, growth retardation, or developmental delay. Reduced neuroepithelial thickness and
reduced density of cranial mesenchyme were detected in ethionine-treated but not cycloleucine-treated
embryos that developed exencephaly. Reduced mesenchymal density is a potential cause of ethionine-induced
exencephaly, although we could not detect a causative alteration in proliferation or apoptosis prior to failure of
neural tube closure. CONCLUSIONS: Adequate functioning of the methylation cycle is essential for cranial neural tube closure in the mouse, suggesting that suppression of the methylation cycle could also increase the risk
of human NTDs. We hypothesize that inhibition of the methylation cycle causes NTDs due to disruption of
crucial reactions involving methylation of DNA, proteins or other biomolecules. Birth Defects Research (Part A)
76:544–552, 2006. Ó 2006 Wiley-Liss, Inc.
Key words: neural tube defects; exencephaly; mouse; embryo; ethionine
INTRODUCTION
The embryonic precursor of the brain and spinal cord,
the neural tube, is formed by the process of neurulation,
in which the lateral edges of the neural plate elevate to
form neural folds which fuse in the midline (Copp et al.,
2003). Incomplete closure of the neural tube results in
neural tube defects (NTDs), severe birth defects that
include anencephaly and spina bifida. Although NTDs
are among the most common human birth defects, the
causes are still poorly defined. Mouse models reveal a
functional requirement for more than 100 different genes
and indicate key developmental mechanisms (Juriloff
and Harris, 2000; Copp et al., 2003; Greene and Copp,
2005). However, the sporadic nature of human NTDs
supports the hypothesis that NTDs are multifactorial
with contributions from environmental and genetic risk
factors.
Among environmental factors, several lines of evidence suggest that 1-carbon metabolism, comprising the
interlinked folate and methylation cycles, plays a key
role in determining susceptibility to NTDs. The risk of
an affected pregnancy is reduced by maternal folic acid
Grant sponsor: Birth Defects Foundation (BDF) Newlife; Grant sponsor: Medical Research Council UK; Grant sponsor: Wellcome Trust; Grant sponsor:
Genzyme.
*Correspondence to: N.D.E. Greene, Neural Development Unit, Institute of
Child Health, Guilford Street, London, UK, WC1N 1EH.
E-mail: n.greene@ich.ucl.ac.uk
Published online 28 August 2006 in Wiley InterScience (www.interscience.
wiley.com).
DOI: 10.1002/bdra.20286
Birth Defects Research (Part A): Clinical and Molecular Teratology 76:544–552 (2006)
METHYLATION CYCLE AND NEURAL TUBE DEFECTS
supplementation and, conversely, increased by suboptimal folate status (Wald et al., 1991; Kirke et al., 1993;
Kalter, 2000). Additional risk factors associated with
abnormal 1-carbon metabolism are elevated levels of homocysteine in maternal blood or reduced levels of vitamin B12, the cofactor for methionine synthase (Kirke
et al., 1993, 1996; Steegers-Theunissen et al., 1994). These
studies focused attention on the potential teratogenic
role of homocysteine, supported by a study in chick
embryos, in which NTDs were induced by homocysteine
(Rosenquist et al., 1996). However, while homocysteine
is toxic to mouse and rat embryos it does not specifically
cause NTDs (Greene et al., 2003; VanAerts et al., 1994).
Hence, homocysteine may not be the primary factor
leading to NTDs in human pregnancy.
Risk factors such as elevated homocysteine and suboptimal folate or B12 status could be associated with
reduced methionine production. Moreover, exogenous
methionine is required to prevent NTDs in rat embryos
cultured in cow serum, suggesting that sufficient methionine may be critical for neurulation (Coelho et al., 1989).
However, the functional requirement for methionine in
neural tube closure has yet to be defined. One possibility
is that there is an essential requirement for methionine in
novel proteins, as the stage of neurulation encompasses a
period of rapid protein synthesis (Greene et al., 2002). As
methionine also plays other key metabolic roles, demand
may be higher than for other amino acids. For example,
methionine is a critical component of the methylation
cycle, being converted to S-adenosylmethionine (SAM),
the methyl donor for methylation of a range of biomolecules including proteins, DNA, and lipids (Scott, 1999).
Donation of a methyl group converts SAM to S-adenosylhomocysteine (SAH), which is subsequently hydrolyzed
to homocysteine (Finkelstein, 1998). The cycle is completed by the remethylation of homocysteine to methionine. Decreased production of methionine could lead to
suppression of the methylation cycle and consequent
reduction in methylation potential. In this study, we set
out to test the hypothesis that adequate flux through the
methylation cycle is essential for neural tube closure in
mammals.
MATERIALS AND METHODS
Mouse Strains and Whole Embryo Culture
Non-mutant random-bred CD1 mice were purchased
from Charles River Laboratories, United Kingdom. All
mouse procedures were in accordance with regulations
set out by the UK Government Home Office. Mice were
paired overnight and females were checked for copulation plugs the following morning, designated embryonic
day (ED) 0.5. Embryos were explanted at ED 8.5 and cultured for 24 hr in immediately centrifuged, heat-inactivated rat serum at 388C, as described previously (Cockroft, 1990; Greene et al., 2002). Ethionine, cycloleucine,
and 5-azacytidine (Sigma-Aldrich, Dorset, UK) were prepared as stock solutions in PBS. Ethionine (1003 stock)
was added to the culture medium as 1% additions (vol/
vol), and cycloleucine (503 stock) was added as a 2%
addition (vol/vol). The equivalent volume of PBS was
added to control groups. Embryos were randomly allocated to treatment groups to minimize the effect of litterto-litter variation. At the end of the culture period the
545
yolk sac circulation was observed as an indication of viability and quantified on a scale from 0 (no circulation) to
3 (vigorous circulation throughout the entire yolk sac).
Quantitative Analysis of Cell Density and
Neuroepithelial Thickness
After culture, embryos were fixed in 4% paraformaldehyde (PFA), dehydrated, embedded in paraffin wax, and
sectioned transversely at 7-lm thickness followed by hematoxylin-eosin staining. Quantitative analysis of mesenchymal cell density was carried out using the section at
the anterior limit of the optic vesicle and the second section posterior to this. Areas for cell counting were
defined by boxes of defined dimensions located in central
(8.5% 3 33.5% of section width) and lateral (8.5% 3
16.7% of section width) sites. On the section at the anterior limit of the optic vesicle, the neuroepithelial thickness in the forebrain and hindbrain was measured at the
midpoint of the dorsoventral axis. A section at 25% of
the distance between the rostral limit of the embryo and
the optic vesicle was used for measurement of the thickness of midbrain neuroepithelium. Midbrain measurements were made at points 25% and 75% of the distance
along the dorsoventral axis. Analysis was performed
‘‘blind’’ to embryo treatment.
Immunohistochemistry and Quantification
of Labeling
For detection of activated caspase-3 or phosphohistone
H3, alternate coronal sections were collected (as above)
and dewaxed in Declere (Cell Marque, Hot Springs, AR)
or HistoClear (National Diagnostics, Atlanta, GA), respectively. Sections were rehydrated, bleached in 3%
hydrogen peroxide, blocked in 5% goat serum, and then
exposed to primary antibodies diluted at 1:1000 (anti-activated caspase-3; Cell Signalling Technology, Danvers,
MA) or 1:500 (anti-phosphohistone H3; Upstate Biotechnology, Charlottesville, VA) in 1% fetal calf serum in
Tris-buffered saline. After washing in PBS, sections were
exposed to biotinylated anti-rabbit secondary antibody
(DAKO, Ely, UK) and signal was developed using ABC
reagent (Vectastain Elite; Vector Laboratories, Burlingame, CA) and 3,3-diaminobenzidine (DAB; Vector Laboratories). The number of positive cells in the mesenchyme
was counted in each of 5 sections evenly spaced in a
112-lm interval centered on the anterior limit of the optic
vesicles. The labeling index is the number of labeled cells
expressed as a percentage of the total number of mesenchymal cells in these sections. Analysis was performed
‘‘blinded’’ to embryo treatment.
Terminal Deoxynucleotidyl Transferase-Mediated
Deoxyuridine Triphosphate-Biotin Nick-End
Labeling of Whole Embryos
Embryos were fixed in 4% PFA and whole-mount terminal deoxynucleotidyl transferase deoxyuridine triphosphate (dUTP) nick-end labeling (TUNEL) was performed,
as described previously (Martinez-Barbera et al., 2002).
Birth Defects Research (Part A) 76:544–552 (2006)
546
DUNLEVY ET AL.
Figure 1. Inhibition of the methylation cycle causes
cranial NTDs in mouse embryos. Embryos were cultured from ED 8.5 to 9.5 in the presence of ethionine
or cycloleucine and scored for the presence of exencephaly, defined as persistently open cranial neural
folds in an embryo with 16 or more somites (numbers of embryos are indicated in Table 1). Asterisks
indicate significantly increased incidence compared
to PBS controls (*P < .001; **P < .05).
Assay of s-Adenosylmethionine and
s-Adenosylhomocysteine
Samples, consisting of a pool of 3–4 embryos from a
specific treatment group, were sonicated in 120 ll of
aqueous mobile phase (4 mM ammonium acetate, 0.1%
formic acid, 0.1% heptafluorobutyric acid, pH 2.5), heattreated, and analyzed in duplicate by liquid chromatography coupled to tandem mass spectrometry, as
described previously (Burren et al., 2006).
Statistical Analysis
Comparison of proportions of affected and unaffected
embryos were made by v2 analysis with pairwise comparison by Fisher’s exact test. Quantitative measurements
of growth parameters were compared by 1-way analysis
of variance (ANOVA) on ranks and, where significant
variation was detected, pairwise comparison was made
by use of the Mann-Whitney rank-sum test. Statistical
tests were performed using SigmaStat (version 2.03; SPSS,
Chicago, IL).
RESULTS
Mouse embryos were exposed to the methylation cycle
inhibitors, ethionine and cycloleucine, throughout the period of cranial neural tube closure (ED 8.5–9.5), in order
to directly test the functional requirement of the methylation cycle in neurulation. Ethionine, a methionine analog,
is converted to s-adenosyl ethionine, which is not utilized
by methyltransferases and in turn acts as a competitive
inhibitor of methionine adenosyltransferase (E.C.2.5.1.6)
(Miller et al., 1994), whereas cycloleucine directly inhibits
this enzyme. During the 24-hr culture period, from
ED 8.5, cranial neural tube closure was completed in the
majority of PBS-exposed control embryos (Figs. 1–2;
Table 1). In contrast, embryos exposed to ethionine or
cycloleucine exhibited a high frequency of exencephaly,
in which the cranial neural folds failed to fuse in the
midline (Figs. 1–2). This defect represented a true failure
of closure, as opposed to a temporary delay in closure,
since by the end of the culture period embryos had
developed to a stage which was 4 somites (equivalent
to 8 hr of development) beyond the 16-somite stage, at
Figure 2. Exencephaly induced by methylation cycle inhibitors occurs in the absence of additional morphological abnormalities. Representative embryos showing that cranial neural tube closure is complete in the great majority of control embryos (a), whereas the neural
folds remain open in >50% of embryos exposed to 5 mM ethionine (b) or 15 mM cycloleucine (c). The extent of the region of open neural folds is indicated by arrowheads and illustrates open FB and MB (b) and open FB, MB, and HB (c), as described in Table 2. FB, forebrain; HB, hindbrain; MB, midbrain. Scale bars represent 200 lm.
Birth Defects Research (Part A) 76:544–552 (2006)
METHYLATION CYCLE AND NEURAL TUBE DEFECTS
547
Table 1
Growth and Development of Mouse Embryos Cultured in the Presence
of Methylation Cycle Inhibitors*
Reagent/
concentration
(mM)
Ethionine
0.00
0.50
1.00
2.00
5.00
10.00
Cycloleucine
0.0
15.0
Number of
embryos
cultured
Number of
live embryos
60
15
9
15
46
16
55
15
9
15
45
13
2.7
2.6
2.2
2.9
2.8
2.3
25
28
24
24
2.3 (0.2)
2.4 (0.2)
Yolk sac
circulationa
(0.1)
(0.2)
(0.3)
(0.1)
(0.1)
(0.3)
Somites
18.9
19.0
18.9
19.4
19.0
16.5
(0.3)
(0.5)
(0.8)
(0.6)
(0.4)
(0.5)b
18.6 (0.6)
19.3 (0.5)
Crown-rump
length
2.38
2.37
2.33
2.55
2.30
2.15
(0.03)
(0.06)
(0.07)
(0.08)
(0.04)
(0.09)b
2.53 (0.05)
2.50 (0.05)
Number
of cranial
NTDs (%)
1
2
1
3
26
13
(1.8)
(14.3)
(12.5)
(21.4b)
(61.9b)
(100b)
3 (14.3)
14 (58.3b)
*Embryos cultured for 24 hr from ED 8.5 to 9.5 were assessed for yolk sac circulation and developmental
parameters. Values are given as mean 6 SEM. Embryos with no yolk sac circulation were excluded from further analysis. Cranial NTDs are defined as failure to complete cranial neural tube closure in live embryos that
had 16 or more somites.
a
See methods for definition of yolk sac circulation.
b
Significant difference from control group (P < .05).
which cranial neural tube closure is normally complete in
the CD1 mouse strain. In the majority of affected embryos, the failure of closure phenotype was severe, with
the neural folds remaining open throughout the entire
forebrain and midbrain, with or without an effect on the
hindbrain, in 100% of ethionine-treated and 70% of cycloleucine-treated embryos, in which exencephaly occurred
(Table 2). In those embryos in which the forebrain remained open, the initial closure event at the rostral limit
of the neuroepithelium had occurred, but closure had not
progressed caudally from that site.
The mean yolk sac circulation score did not differ
between PBS-treated controls and the inhibitor treatment
groups indicating that neither ethionine nor cycloleucine
causes generalized toxicity (Table 1). The crown-rump
length and number of somites were assessed after culture
as measures of embryonic growth and developmental
progression, respectively. Embryos exposed to 10 mM
ethionine exhibited reduced crown-rump length and
mean number of somites compared to PBS-treated controls, suggesting that this dose caused growth retardation
and developmental delay (Table 1). However, there was
no apparent effect of cycloleucine or lower doses of ethionine on crown-rump length or somite number. There-
fore, in these treatment groups, induction of exencephaly
by methylation cycle inhibitors appears to occur in the
absence of growth retardation or developmental delay
(Table 1). Moreover, other than NTDs, no other obvious
morphological defects (Fig. 2) were detected despite the
severity of the exencephaly phenotype.
To investigate the possible modes of pathogenesis of
NTDs, we examined tissue sections from the cranial
region of cultured embryos (Fig. 3). No obvious histological abnormalities were associated with cranial NTDs in
cycloleucine-treated embryos, whereas ethionine-treated
embryos appeared to show reduced density of the cranial
mesenchyme and reduced thickness of the neuroepithelium compared to controls. These changes were present
at doses (5 mM) that did not have an effect on overall
growth or developmental progression (Table 1). Measurements of neuroepithelial thickness at matched levels of
the cranial region (Fig. 3a–c) confirmed that there was a
significant reduction in thickness in the midbrain of ethionine-treated embryos (both with and without NTDs)
compared with PBS-treated controls (Fig. 3h). There was
also a slight, but nonsignificant, reduction of the neuroepithelium thickness in the hindbrain of these embryos,
but no changes in the forebrain. This observation of a
Table 2
Extent of Cranial Neural Tube Defect in Embryos Exposed to Methylation Cycle Inhibitors*
Treatment
PBS
Ethionine
Cycloleucine
Region of open neural folds (% of exencephalic embryos)
Number of
embryos
FB
FB þ MB
MB
MB þ HB
HB
FB þ MB þ HB
8
27
23
0
0
4.3
33.3
14.3
52.2
33.3
0
4.3
11.1
0
17.4
11.1
0
4.3
11.1
85.7
17.4
*The region of open neural folds in exencephalic embryos was recorded at the end of the culture period in
a subset of the cultures presented in Table 1. The proportion of embryos with a particular phenotype is
expressed as the percentage of the embryos with exencephaly. Owing to the low rate of exencephaly in PBStreated embryos, data are included from concurrent cultures (cultured simultaneously but not necessarily littermates of treated embryos). Concentrations of ethionine have been pooled (excluding 10 mM dose).
FB, forebrain; MB, midbrain; HB, hindbrain.
Birth Defects Research (Part A) 76:544–552 (2006)
548
DUNLEVY ET AL.
Figure 3. Reduced thickness of neuroepithelium and
density of cranial mesenchyme in ethionine-treated
embryos. In the intact embryo (a), dashed lines indicate
the position of sections (b,c; rostral to top) used for
measurement of neuroepithelial thickness. Black lines in
(b) and (c) indicate the position in the neuroepithelium
at which thickness was measured. Transverse sections
(at the same level as b) of control (d), ethionine-treated
(e) and cycloleucine-treated (f) embryos reveal the open
neural folds (exencephaly) in examples of treated
embryos, compared to the control in which closure is
complete. Boxes in (d) indicate the areas from which
mesenchyme density was calculated. Cranial mesenchyme density (g) and neuroepithelial thickness (h) was
quantified (mean 6 SEM) in control embryos (n ¼ 6), or
embryos exposed to 5 mM ethionine or 15 mM cycloleucine. Inhibitor-treated embryos are shown separately in
cases where neural tube closure was complete (closed;
n ¼ 3 for ethionine; n ¼ 4 for cycloleucine) or had failed
(NTDs; n ¼ 5 for ethionine; n ¼ 5 for cycloleucine).
Asterisks indicate significant difference from values at
the equivalent region of control embryos (P < .05). EX,
exencephaly; FB, forebrain; HB, hindbrain; MB, midbrain; OV, optic vesicle. Scale bars represent 200 lm.
Birth Defects Research (Part A) 76:544–552 (2006)
METHYLATION CYCLE AND NEURAL TUBE DEFECTS
549
Figure 4. TUNEL staining of the cranial region of cultured embryos. Apoptotic cells (labeled in blue) are scattered in the cranial mesenchyme and neuroepithelium after culture of embryos for 24 hr in the presence of PBS (controls; a,b) or 5 mM ethionine (c,d). a,c: Left lateral view. b,d: posterior view. Particularly intense staining is present in the proximal region of the first branchial arch (arrows in a,c).
No obvious differences in the extent of staining were observed between control and treated embryos. White arrowheads (c) indicate the
region of open neural folds in the ethionine-treated embryo. Scale bars represent 100 lm.
tissue change solely in the midbrain contrasts with the
finding that forebrain and hindbrain also frequently
remained open in ethionine-treated embryos.
Reduced density of the cranial mesenchyme has previously been described in association with development of
exencephaly in Twist and Cart1 null embryos (Zhao et al.,
1996; Soo et al., 2002). In order to quantify the apparent
effect of ethionine on the cranial mesenchyme, we determined mesenchymal density at matched levels within the
cranial region of inhibitor-treated and PBS control
embryos (Fig. 3d–f). Since the density varied within sections in different regions of the mesenchyme, both central
and lateral areas were analyzed (Fig. 3d). In the ethionine-treated embryos in which neural tube closure had
failed there was a significant reduction in mesenchyme
density in both central and lateral areas (Fig. 3g),
whereas no such change was observed in cycloleucine-
treated embryos irrespective of whether or not they
exhibited exencephaly. Reduced density of the cranial
mesenchyme could potentially contribute to failure of
neural tube closure, but could also be secondary to failure of closure. The latter possibility seems likely, since
the mesenchyme density in ethionine-treated embryos
that successfully completed closure was indistinguishable
from that in controls. On the other hand, the absence of
an effect of cycloleucine on mesenchyme density, both in
embryos that displayed NTDs and in embryos in which
closure was complete (Fig. 3g), argues against a secondary effect.
A possible cause of the observed depletion of cranial
mesenchyme could be excessive cell death or a reduced
rate of proliferation. Whole-mount TUNEL staining of
embryos after the 24-hr culture period did not reveal any
obvious differences in the extent of apoptosis between
Birth Defects Research (Part A) 76:544–552 (2006)
550
DUNLEVY ET AL.
Table 3
Proliferation and Apoptosis in the Cranial Mesenchyme of Embryos Exposed
to Methylation Cycle Inhibitors*
Treatment group
Number of
embryos
Mean number
of somites
Number of
sections analyzed
Labeling index
(mean 6 SEM)
17
9
9
13.8
13.9
14.0
83
45
44
7.85 6 0.60
7.58 6 0.36
7.86 6 0.54
14
8
6
13.5
14.4
14.0
68
39
28
4.27 6 1.28
3.15 6 1.05
7.11 6 1.75
a. Phosphohistone H3
PBS
Ethionine
Cycloleucine
b. Cleaved caspase-3
PBS
Ethionine
Cycloleucine
*Labeling indices for phosphohistone H3 and cleaved caspase-3 were calculated from cell counting of immunostained sections, following culture of embryos for 12 hr in the presence of PBS alone (control), 5 mM
ethionine, or 15 mM cycloleucine. No significant differences were detected between treatment groups in either
mean number of somites or labeling indices.
treatment groups (Fig. 4). However, if the induction of
NTDs by ethionine or cycloleucine is mediated via an
effect on apoptosis or cellular proliferation, this would be
expected to occur prior to the failure of neural tube closure. Therefore, we cultured an additional series of
embryos and analyzed rates of apoptosis and proliferation
after only 12 hr of culture. At this time, embryos had
developed to the 12–15 somite stage, immediately prior to
the stage (16 somites) at which closure is completed in
control and unaffected treated embryos. Immunohistochemical staining was used to calculate labeling indices
for phosphorylated histone 3 (Table 3a) and cleaved caspase 3 (Table 3b) in the cranial mesenchyme, to give an indication of the extent of proliferation and apoptosis,
respectively. However, we did not detect significant differences in labeling indices for either marker, suggesting that
neither ethionine nor cycloleucine acts through a primary
effect on rates of apoptosis or proliferation.
In order to determine whether ethionine and cycloleucine act to suppress the methylation cycle, embryos were
cultured for 17 hr, to a stage immediately preceding closure. Embryos were exposed to 5 mM ethionine, 15 mM
cycloleucine, or PBS only, and the levels of SAM and
SAH were quantified. Both ethionine and cycloleucine
caused a significant increase in the concentration of SAH,
and a significant decrease in the ratio of abundance of
SAM to SAH in comparison to PBS-treated control
embryos (Table 4). In addition, ethionine-treated embryos
had significantly lower levels of SAM (Table 4), which
explains the greater effect on SAM/SAH ratio than for
cycloleucine. A key function of the methylation cycle is
the provision of methyl groups for methyltransferasemediated methylation of a range of biomolecules including genomic DNA. The elevation of SAH levels and
reduction of the SAM/SAH ratio that we detected in cultured embryos (Table 4) is expected to result in suppression of methyltransferase activity.
DISCUSSION
Cranial neural tube closure is a complex morphological
process characterized by the initial formation of biconvex
neural folds whose tips subsequently ‘‘flip around’’ to
form concave neural folds with paired dorsolateral hinge
points that allow the tips of neural folds to appose and
fuse in the midline (Morriss-Kay, 1981; Copp, 2005). The
exquisite sensitivity of this process to perturbation is
revealed by the fact that the majority of NTDs arising in
genetic mutant mouse strains comprise exencephaly, with
or without a spinal defect (Juriloff and Harris, 2000;
Copp et al., 2003). Similarly, many teratogens induce cranial defects whereas relatively few cause isolated spina
bifida (Copp et al., 1990). In this study, induction of
exencephaly by exposure of embryos to inhibitors of the
methylation cycle indicates an essential requirement for
the methylation cycle in cranial neurulation. The extent
of the failure of closure, which includes the forebrain in
the majority of affected, inhibitor-treated embryos suggests that the phenotype at birth would be severe and
include a ‘‘split face’’ defect. The idea that methylation
reactions are essential for processes underlying neural
tube closure is also suggested by a recent study in chicks
Table 4
Quantification of SAM and SAH in Cultured Mouse Embryos*
Treatment
group
PBS
Ethionine
Cycloleucine
Number of
samples
SAM
(nmol/mg protein)
SAH
(nmol/mg protein)
Ratio
SAM/SAH
7
5
6
3.91 6 0.38
2.67 6 0.25a
4.16 6 0.60
0.023 6 0.003
0.034 6 0.004a
0.032 6 0.003a
178.0 6 11.9
79.2 6 3.2a
130.0 6 12.1a
*Embryos were cultured from ED 8.5 for 17 hr in the presence of PBS (control group), 5 mM ethionine, or
15 mM cycloleucine. SAM and SAH were quantified by liquid chromatography coupled to tandem mass spectrometry. Values were normalized to protein content and are expressed as mean 6 SEM.
a
Significant difference from values for PBS-treated controls, P < .05.
Birth Defects Research (Part A) 76:544–552 (2006)
METHYLATION CYCLE AND NEURAL TUBE DEFECTS
in which exposure to methylation cycle inhibitors caused
a delay in closure of the anterior neural folds (Afman
et al., 2005). Unlike the mouse, neural tube closure was
eventually completed in treated chick embryos, although
it is possible that this represents a difference of experimental approach rather than a major difference between
birds and mammals.
In several cases, exencephaly induced experimentally by
exogenous reagents appears to be associated with rather
non-specific generalized retardation of embryonic growth
or developmental progression (Copp et al., 1990). Moreover, the neural tube may remain persistently open following exposure to insults that are lethal at developmental
stages prior to completion of closure (Copp, 1995; Copp
et al., 2003). However, neither ethionine nor cycloleucine
suppressed growth or development at doses which induced a significant incidence of exencephaly, suggesting a
specific effect of these inhibitors on a process that is essential for closure of the cranial neural folds.
Ethionine and cycloleucine are known to inhibit methionine adenosyl transferase (Miller et al., 1994), and we
observed a significant decrease in abundance of SAM,
the product of this enzyme, in ethionine-treated embryos,
although not in cycloleucine-treated embryos. Both inhibitors caused an increase in the level of SAH, presumably
owing to feedback regulation around the methylation
cycle. The net effect is that both agents act to reduce the
SAM/SAH ratio, which is predicted to result in suppression of methyltransferase-mediated methylation reactions
(Caudill et al., 2001). In mice, investigation of the relationship between levels of SAM and SAH and the SAM/
SAH ratio in terms of their effect on DNA methylation,
suggests that elevation of intracellular SAH concentration
and, to a lesser extent, decreased SAM/SAH ratio are
most closely associated with reduced methylation status
(Caudill et al., 2001). Thus, although cycloleucine did not
apparently affect SAM levels in cultured mouse embryos,
the significant effects on SAH concentration and SAM/
SAH ratio strongly suggest that transmethylation reactions will be suppressed.
The subtle differences in the cellular effects of cycloleucine and ethionine in the cranial neuroepithelium and
mesenchyme may be due to differing extent of activity in
suppressing flux through the cycle. For example, ethionine appeared to have an effect on embryonic levels of
both SAH and SAM, such that the magnitude of the
effect on the SAM/SAH ratio was greater. In turn, this
could affect the subset of methyltransferases whose activity is compromised, due to differences in enzyme
kinetics. In addition to inhibition of methionine adenosyl
transferase, S-adenosyl ethionine is also known to activate cystathionine b-synthase, which acts to catabolize
homocysteine through the transsulfuration pathway
(Miller et al., 1994). Thus, a potential effect of ethionine
treatment would be enhanced clearance of homocysteine.
However, it appears that cystathionine b-synthase may
not be expressed during rodent neurulation (VanAerts
et al., 1995), suggesting that the observed effect of ethionine is most likely mediated through suppression of
methylation cycle flux.
Cellular changes in the cranial region of ethioninetreated embryos were manifested as reduced thickness of
the neuroepithelium that constitutes the midbrain neural
folds. This reduced thickness is unlikely to be the cause
of NTDs since both embryos with open and closed neural
551
folds were affected. It could be argued that the presence
of an open midbrain, with thin neuroepithelium, caused
the failure of closure in the hindbrain, which was
observed in most exencephalic ethionine-treated embryos.
However, this mechanism appears unlikely, as normal
closure progresses into the midbrain, caudally from the
forebrain/midbrain boundary and rostrally from the
hindbrain, and not vice versa (Copp et al., 2003).
In contrast to reduced neuroepithelial thickness, a
reduction in density of the cranial mesenchyme was
observed only in association with failure of neural tube
closure in ethionine-treated embryos. Expansion of the
cranial mesenchyme subjacent to the cranial neural folds
has previously been suggested to assist the elevation of
the cranial neural folds during the biconvex phase (Morriss and Solursh, 1978; Morris-Wiman and Brinkley,
1990). These observations suggest a possible causative
link between reduced cell density and failure of closure
following ethionine treatment. However, prior to failure
of closure there was no apparent abnormality in the rate
of cell proliferation or apoptosis that could explain the
reduced cell density, so we cannot currently rule out a
secondary effect.
Disruption of the methionine cycle could influence
multiple biochemical processes including DNA and protein methylation, polyamine synthesis, and regulation of
the folate cycle (Scott, 1999; Van der Put et al., 2001).
Methylation of DNA provides an important mechanism
for the epigenetic control of gene expression, whereby
methylation of cytosine residues within CpG islands
inhibits transcription (Dean et al., 2005). Suppression of
DNA methylation could thus lead to aberrant expression
of one or more genes, resulting in development of NTDs.
Consistent with this idea, embryos that are null for DNA
methyltransferase 3b develop cranial NTDs (Okano et al.,
1999). Alternatively, aberrant protein methylation could
also play a role since regulated methylation can influence
protein function. For example, cytoskeletal proteins such
as b-actin and tubulin are known to become methylated
at the stage of neural tube closure (Moephuli et al., 1997).
Integrity of the cytoskeleton is essential for cranial neurulation (Smedley and Stanisstreet, 1986; Matsuda and
Keino, 1994; Ybot-Gonzalez and Copp, 1999), suggesting
that altered function of cytoskeletal proteins through lack
of methylation could influence closure.
ACKNOWLEDGMENTS
We thank Dawn Savery for technical assistance and
members of the Neural Development Unit for helpful
discussion.
REFERENCES
Afman LA, Blom HJ, Drittij M-J, et al. 2005. Inhibition of transmethylation
disturbs neurulation in chick embryos. Brain Res Dev Brain Res
158:59–65.
Burren KA, Mills K, Copp AJ, Greene NDE. 2006. Quantitative analysis of
s-adenosylmethionine and s-adeno-sylhomocysteine in neurulationstage mouse embryos by liquid chromatography tandem mass spectrometry. J Chromatogr B.
Caudill MA, Wang JC, Melnyk S, et al. 2001. Intracellular s-adenosylhomocysteine concentrations predict global DNA hypomethylation in
tissues of methyl-deficient cystathionine b-synthase heterozygous
mice. J Nutr 131:2811–2818.
Cockroft DL. 1990. Dissection and culture of postimplantation embryos.
In: Copp AJ, Cockroft DL, editors. Postimplantation mammalian
embryos: a practical approach. Oxford: IRL Press. pp. 15–40.
Birth Defects Research (Part A) 76:544–552 (2006)
552
DUNLEVY ET AL.
Coelho CND, Weber JA, Klein NW, et al. 1989. Whole rat embryos
require methionine for neural tube closure when cultured on cow serum. J Nutr 119:1716–1725.
Copp AJ, Brook FA, Estibeiro JP, et al. 1990. The embryonic development
of mammalian neural tube defects. Prog Neurobiol 35:363–403.
Copp AJ. 1995. Death before birth: clues from gene knockouts and mutations in the mouse. Trends Genet 11:87–93.
Copp AJ, Greene NDE, Murdoch JN. 2003. The genetic basis of mammalian neurulation. Nat Rev Genet 4:784–793.
Copp AJ. 2005. Neurulation in the cranial region – normal and abnormal.
J Anat 207:623–635.
Dean W, Lucifero D, Santos F. 2005. DNA methylation in mammalian
development and disease. Birth Defects Res C Embryo Today 75:98–
111.
Finkelstein JD. 1998. The metabolism of homocysteine: pathways and regulation. Eur J Pediatr 157(Suppl 2):S40–S44.
Greene NDE, Leung KY, Wait R, et al. 2002. Differential protein expression at the stage of neural tube closure in the mouse embryo. J Biol
Chem 277:41645–41651.
Greene NDE, Dunlevy LE, Copp AJ. 2003. Homocysteine is embryotoxic
but does not cause neural tube defects in mouse embryos. Anat
Embryol 206:185–191.
Greene NDE, Copp AJ. 2005. Mouse models of neural tube defects: investigating preventive mechanisms. Am J Med Genet C Semin Med
Genet 135:31–41.
Juriloff DM, Harris MJ. 2000. Mouse models for neural tube closure
defects. Hum Mol Genet 9:993–1000.
Kalter H. 2000. Folic acid and human malformations: a summary and
evaluation. Reprod Toxicol 14:463–476.
Kirke PN, Molloy AM, Daly LE, et al. 1993. Maternal plasma folate and
vitamin B12 are independent risk factors for neural tube defects. Q J
Med 86:703–708.
Kirke PN, Daly LE, Molloy A, et al. 1996. Maternal folate status and risk
of neural tube defects. Lancet 348:67–68.
Martinez-Barbera JP, Rodriguez TA, Greene NDE, et al. 2002. Folic acid
prevents exencephaly in Cited2 deficient mice. Hum Mol Genet 11:283–
293.
Matsuda M, Keino H. 1994. An open cephalic neural tube reproducibly
induced by cytochalasin D in rat embryos in vitro. Zool Sci 11:547–
553.
Miller JW, Nadeau MR, Smith J, et al. 1994. Folate-deficiency-induced
homocysteinaemia in rats: disruption of S-adenosylmethionine’s coordinate regulation of homocysteine metabolism. Biochem J 298:415–
419.
Moephuli SR, Klein NW, Baldwin MT, Krider HM. 1997. Effects of methionine on the cytoplasmic distribution of actin and tubulin during
Birth Defects Research (Part A) 76:544–552 (2006)
neural tube closure in rat embryos. Proc Natl Acad Sci USA 94:543–
548.
Morris-Wiman J, Brinkley LL. 1990. Changes in mesenchymal cell and hyaluronate distribution correlate with in vivo elevation of the mouse
mesencephalic neural folds. Anat Rec 226:383–395.
Morriss GM, Solursh M. 1978. Regional differences in mesenchymal cell
morphology and glycosaminoglycans in early neural-fold stage rat
embryos. J Embryol Exp Morphol 46:37–52.
Morriss-Kay GM. 1981. Growth and development of pattern in the cranial
neural epithelium of rat embryos during neurulation. J Embryol Exp
Morphol 65(Suppl):225–241.
Okano M, Bell DW, Haber DA, Li E. 1999. DNA methyltransferases
Dnmt3a and Dnmt3b are essential for de novo methylation and
mammalian development. Cell 99:247–257.
Rosenquist TH, Ratashak SA, Selhub J. 1996. Homocysteine induces congenital defects of the heart and neural tube: effect of folic acid. Proc
Natl Acad Sci USA 93:15227–15232.
Scott JM. 1999. Folate and vitamin B12. Proc Nutr Soc 58:441–448.
Smedley MJ, Stanisstreet M. 1986. Calcium and neurulation in mammalian embryos. II. Effects of cytoskeletal inhibitors and calcium antagonists on the neural folds of rat embryos. J Embryol Exp Morphol
93:167–178.
Soo K, O’Rourke MP, Khoo PL, et al. 2002. Twist function is required for the
morphogenesis of the cephalic neural tube and the differentiation of the
cranial neural crest cells in the mouse embryo. Dev Biol 247:251–270.
Steegers-Theunissen RPM, Boers GHJ, Trijbels FJM, et al. 1994. Maternal
hyperhomocysteinemia: a risk factor for neural tube defects. Metabolism 43:1475–1480.
Van der Put NMJ, Van Straaten HWM, Trijbels FJM, Blom HJ. 2001. Folate, homocysteine and neural tube defects: an overview. Proc Soc
Exp Biol Med 226:243–270.
VanAerts LAGJM, Blom HJ, Deabreu RA, et al. 1994. Prevention of neural
tube defects by and toxicity of L-homocysteine in cultured postimplantation rat embryos. Teratology 50:348–360.
VanAerts LAGJM, Poirot CM, Herberts CA, et al. 1995. Development of methionine synthase, cystathionine-b-synthase and S-adenosyl-homocysteine hydrolase during gestation in rats. J Reprod Fertil 103:227–232.
Wald N, Sneddon J, Densem J, et al. 1991. Prevention of neural tube
defects: results of the Medical Research Council Vitamin Study. Lancet 338:131–137.
Ybot-Gonzalez P, Copp AJ. 1999. Bending of the neural plate during
mouse spinal neurulation is independent of actin microfilaments.
Dev Dyn 215:273–283.
Zhao Q, Behringer RR, De Crombrugghe B. 1996. Prenatal folic acid treatment suppresses acrania and meroanencephaly in mice mutant for
the Cart1 homeobox gene. Nat Genet 13:275–283.