| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,
* Department of Pharmacology, University of Toronto, Medical Sciences Building, Toronto Ontario, Canada M5S 1A8;
Department of Genetics, The Hospital for Sick Children, Toronto, Ontario, Canada M5G 1X8;
Department of Molecular and Medical Genetics, and Institute of Medical Sciences, University of Toronto, Toronto, Ontario, Canada M5S 1A8; and
§ Faculty of Pharmacy, University of Toronto, Toronto, Ontario, Canada M5S 2S2
1Correspondence: Faculty of Pharmacy, University of Toronto, 19 Russell Street, Toronto, Ontario, Canada M5S 2S2. E-mail: pg.wells{at}utoronto.ca
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Key Words: reactive oxygen species • development • birth defects • phenytoin • human risk
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
, 2)
.
During cellular oxidative stress, whether endogenous in origin or
initiated by drugs or environmental chemicals, collectively referred to
as xenobiotics, NADPH is critical for maintaining glutathione (GSH) in
its reduced form, which is essential for detoxification of reactive
free radicals and lipid hydroperoxides (3
4
5)
(Fig. 1
). Another important role for NADPH is the maintenance of the catalytic
activity of catalase (6
7
8)
; hence, NADPH also is
important for its role in the detoxification of hydrogen peroxide. The
production of ribose by the HMS is relevant to the synthesis of
nucleotides used in RNA and DNA replication and, hence, cell division
and possibly DNA repair (2
, 9)
. The gene for G6PD,
containing 13 exons, has been localized to the X chromosome (Xq28), and
the coding sequence has been reported for a number of species including
mice, rats, Drosophila, yeast, and humans (10
, 11)
. Hereditary deficiencies in G6PD, first identified in the
late 1950s based on studies of differential susceptibility to the
hemolytic effects of primaquine (12)
, are the most common
enzymopathy known, affecting well over 400 million people worldwide,
and particularly those from the Mediterranean region and selected
African and Asian countries, wherein the incidence of G6PD deficiency
may approach 60% of some populations (2)
. The degree of
G6PD deficiency varies from negligible to severe, according to both
whether one or two alleles are affected and the nature of the gene
mutation and protein/enzyme variant (1)
. To date, 100
human genetic mutations involving the 12 coding exons, and up to 400
enzyme variants, have been described (10
, 13
14
15)
.
Hematological problems arising in these G6PD-deficient populations from
exposure to oxidizing xenobiotics have been well characterized, ranging
from hemolysis of red blood cells and hereditary nonspherocytic
hemolytic anemia to sepsis and life-threatening kernicterus in the
newborn (9
, 16)
.
|
It currently is believed that G6PD deficiencies constitute a
problem only for mature red blood cells, which are non-nucleated and
cannot synthesize more protective enzyme under conditions of oxidative
stress (9
, 13)
. However, embryonic tissues up to and
including the critical period of organogenesis are remarkably deficient
in the synthesis of many enzymes, including most of those providing
cytoprotection against oxidative stress, such as GSH reductase, GSH
peroxidase, superoxide dismutase, and catalase (4
, 17)
. On
the other hand, elevated G6PD activity during embryonic development
corresponds to periods of both increased cellular proliferation and DNA
synthesis (18
, 19)
, suggesting that G6PD activity may be
important for normal development. Accordingly, we hypothesized that
G6PD-deficient embryos would be highly susceptible to normal
developmental oxidative stress, and even more so to that initiated by
oxidizing xenobiotics (Fig. 1)
. This hypothesis was tested in pregnant
mutant C3H mice with heterozygous (+/-) or homozygous (-/-)
deficiencies in G6PD activity, compared with congenic G6PD-normal
controls (+/+). Dams either were allowed to deliver untreated or were
treated during organogenesis with either the most commonly used
anticonvulsant drug in North America, phenytoin (Dilantin), a human
teratogen that is representative of xenobiotics known to initiate
embryonic oxidative stress (20)
or its vehicle. To
determine the cytoprotective role of G6PD with respect to
xenobiotic-initiated embryonic DNA damage, after maternal treatment
with phenytoin during organogenesis, individual embryos were analyzed
for both G6PD activity and DNA oxidation. The recently reported
functional mutation in the mouse G6PD gene (21)
was
confirmed by a combination of direct sequencing and the development of
a polymerase chain reaction (PCR)-based genotyping method, which was
used to determine the frequency of the G6PD-deficient genotype in the
remnants (resorptions) of embryos that died in utero. The
results provide the first direct evidence of a critical
embryoprotective role for G6PD in both endogenous and
xenobiotic-initiated oxidative stress and DNA damage.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Animals
Breeding pairs of G6PD-mutant C3H mice were purchased from the
Medical Research Council (MRC) of England (Genetics Division, MRC
Radiobiology Unit, Chilton, U.K.). Same-sex animals were housed not
more than three to one microisolator cage containing ground corncob
bedding (Beta Chip, Northeastern Products, Warrensburg, N.Y.) and were
maintained in a temperature-controlled animal facility with a 12 h
light/dark cycle. Food (Laboratory Rodent Chow 5001; PMI Feeds, St.
Louis, Mo.) and tap water were provided ad libitum. The
genotype of all animals was confirmed phenotypically by analysis of RBC
G6PD activity. To establish a breeding colony, three females were
housed overnight with one male breeder starting at 5:00
P.M. Females were checked by 9:00
A.M. the next morning, and the presence of a
vaginal plug was designated as gestational day (GD) 1. Pregnant females
were placed in their own microisolator cages and allowed to deliver
spontaneously, with an average gestation of 20.5 days. The number of
pups was recorded daily, and the pups were left with their mothers
until weaning, 21 days after birth. The number and sex of weaned pups
were recorded. Pups were ear-notched for identification and phenotyped
by G6PD activity using tail vein blood.
Teratogenesis
Homozygous (+/+) G6PD-normal and heterozygous (+/-) and
homozygous (-/-) G6PD-deficient females were mated with males that
were hemizygous (-/y) G6PD-deficient as described above. Dams were
either untreated, treated intraperitoneally at 9:00 A.M. on
GDs 12 and 13 with vehicle alone or treated with a subteratogenic (20
mg/kg) or teratogenic dose (65 mg/kg) of phenytoin in saline containing
0.002N NaOH (17)
and killed by cervical dislocation on GD
19. The uterus was exteriorized, implantations (fetuses and
resorptions/in utero deaths) were noted, and fetuses and
dissectable resorptions were removed. Fetuses were examined to
determine sex, weight, and external anomalies. Viable fetuses were kept
warm under a heat lamp (30°C) for 2 h to assess postpartum
lethality. Fetuses subsequently were bled by decapitation and
phenotyped for RBC G6PD activity. Resorptions and fetal tails were
stored at -80°C for future study. Pinpoint resorptions were noted
and left in the uterus, which was stored, similar to fetal heads and
bodies, in Carnoy’s solution for future analysis. Fetuses were later
examined for internal anomalies.
G6PD gene sequencing
RNA was isolated and purified from +/+ G6PD-normal and -/-
G6PD-deficient mouse spleen samples using a Qiagen RNeasy total RNA
purification kit (Qiagen, Chatsworth, Calif.) according to the
manufacturer’s instructions. Purified RNA was subsequently converted
to cDNA using a Gibco BRL Superscript Preamplification System for First
Strand cDNA Synthesis (Gibco BRL [Canada], Burlington, Ontario) with
oligo dT primers according to the manufacturer’s instructions using a
Perkin Elmer GeneAmp PCR System 2400 thermal cycler (Perkin-Elmer
[Canada], Mississauga, Ontario). Oligonucleotide primers for PCR
reactions were synthesized based on previously reported C57BL/6 mouse
G6PD cDNA sequences (11)
. cDNA coding for G6PD was
subsequently amplified using the PCR. Amplified cDNA templates were
subsequently used for direct cycle sequencing using a Thermo Sequenase
radiolabeled terminator cycle sequencing kit (Amersham [Canada],
Oakville, Ontario) according to the manufacturer’s instructions, where
sequencing primers for both sense and antisense strands, designed to
produce overlapping sequences, were based on the published normal G6PD
mouse cDNA sequence (11)
. Samples were run on standard 6%
polyacrylamide gels and exposed overnight prior to developing film.
Data for G6PD mutant mouse intronic regions were sequenced from genomic
mouse DNA samples derived from tail snips.
G6PD genotyping
DNA was isolated from late-stage (dissectable) fetal resorptions
by the method of Gupta (22)
. G6PD mouse PCR primers
(sense: GGAAACTGGCTGTGCGCTAC, antisense: TCAGCTCCGGCTCTCTTCTG)
were made between exon 1 and intron 1, around the reported mutation
site (21)
. PCR conditions on a Perkin Elmer 9600 thermal
cycler (Perkin-Elmer [Canada]) were 94°C for 2 min, 20 s at
94°C, 20 s at 58°C, and 30 s at 72°C for a total of 35
cycles, with a 5 min extension at 72°C and kept at 4°C until ready
for digestion. PCR products were digested using DdeI
restriction enzyme (Gibco BRL) at 37°C for 1 h and run on 3%
agarose gels to determine G6PD genotype.
G6PD phenotyping
G6PD activity was measured in RBCs, whole embryo homogenates,
and the 9,000 g supernatant from homogenized maternal organs
using a standard reagent kit purchased from Sigma. Activities were
measured over a 5 min interval at 37°C on a UV/vis spectrophotometer
(model Lambda 3, Perkin-Elmer [Canada]) using a computer-assisted
kinetic program. All results were standardized with respect to total
protein content and reported in International Units per gram of protein
(U/g). G6PD normal control standards (Sigma) were run concurrently with
samples.
DNA oxidation
Females were mated as in the teratological studies. Dams were
killed 6 h after maternal treatment with phenytoin (65 mg/kg i.p.)
on GD 13. The uterus was exteriorized, and embryos were removed and
homogenized separately. Once G6PD activity was measured, DNA was
isolated from the remainder of the individual whole embryo homogenates
by the method of Gupta (22)
, as modified in Winn and Wells
(20)
. Embryonic DNA oxidation was measured by the method
of Shigenaga and Ames (23)
, using high-performance liquid
chromatography with electrochemical detection of
8-hydroxy-2'-deoxyguanosine (8-OH-2'-dG).
Protein concentration assay
Protein content was analyzed using the standard Bio-Rad protocol
(Bio-Rad, Hercules, Calif.), as detected by spectrophotometric
absorbance at 595 nm, using bovine serum albumin concentrate as a
standard.
Statistical analysis
Binomial data were analyzed using
2 analysis or Fisher’s exact test where
appropriate. Continuous data were analyzed using a two-way analysis of
variance (ANOVA) and the Student-Newman-Keuls test. The level of
significance was P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
presents differences in red blood cell (RBC) G6PD activities among G6PD
mutant adult progenitor mice (3 months of age, from our colony), with
hemizygous (-/y), heterozygous (+/-), and homozygous (-/-) mutants
having 21, 53, and 17%, respectively, of normal RBC G6PD activity.
Phenotyping of congenic G6PD normal and mutant animals was based on our
progenitor RBC G6PD activities (see Fig. 2
) and corroborated by
comparisons with expected outcomes, from both parental matings and
neonatal sex, given that the G6PD mutation is inherited via the X
chromosome.
|
Untreated mice
In untreated dams allowed to deliver spontaneously, compared with
congenic +/+ G6PD-normal controls, litter sizes for -/-
G6PD-deficient animals were 50% smaller at birth (Fig. 3
, upper panel) (P<0.05). Subsequently, by the time of
weaning, litter sizes were 90% smaller, and the incidence of
preweaning offspring death was threefold higher for untreated
-/-G6PD-deficient dams compared with +/+ G6PD-normal controls
(P<0.0001) (Fig. 3
, lower panel).
|
Vehicle control mice
Homozygous G6PD-normal (+/+) and heterozygous (+/-) and
homozygous (-/-) G6PD-deficient mice were injected intraperitoneally
on GD 12 and 13 with the saline/NaOH vehicle for phenytoin and killed
on GD 19. Compared with +/+ G6PD-normal dams, +/- and -/-
G6PD-deficient dams had 6- and 7-fold increases, respectively, in fetal
resorptions (in utero deaths) (P<0.0001), 6- and
11-fold increases, respectively, in postpartum lethality
(P<0.0001), and 7 and 16% decreases, respectively, in
fetal body weight (P<0.05) (Fig. 4
).
|
With respect to embryonic phenotype, compared with +/y G6PD-normal
fetuses, independent of sex, +/- and -/- and -/y G6PD-deficient
fetuses had enhanced postpartum lethality, following a gene-dose
pattern with 0% in +/y G6PD-normal fetuses, 10% in +/-
G6PD-deficient fetuses, and 50% in combined -/- and -/y
G6PD-deficient fetuses (P<0.0001). A similar gene-dose
pattern was observed with fetal body weight, which, compared with +/y
G6PD-normal fetuses, was decreased by 6% in fetuses with a mutation in
one G6PD allele (+/-) and by 18% in fetuses with a mutation in all
alleles (-/- and -/y) (P<0.05) (Fig. 5
). Furthermore, the mean weight of fetuses with a mutation in all G6PD
alleles (-/- and -/y) was 13% lower than that of fetuses with a
mutation in one allele (+/-) (P<0.05).
|
Phenytoin-treated mice
With respect to maternal phenotype, compared with vehicle-treated
controls of the same phenotype, a standard teratogenic dose of
phenytoin (65 mg/kg) enhanced fetal resorptions twofold in both +/- and -/- G6PD-deficient dams (P<0.05), and in +/- G6PD-deficient dams, enhanced postpartum lethality threefold and
decreased fetal weight by 27% (P<0.05) (Fig. 4)
. Among
phenytoin-treated dams (65 mg/kg), compared with +/+ G6PD-normal
controls, in both +/- and -/- G6PD-deficient dams respectively,
the incidence of fetal resorptions was enhanced 4.8- and 5.5-fold
(P<0.0003), the incidence of postpartum lethality was
enhanced 16.6- and 15.4-fold (P<0.002), and fetal weight
was decreased by 18–29% (P<0.05) (Fig. 4)
.
With a lower dose of phenytoin (20 mg/kg) that is nonteratogenic
in other strains (4
, 5)
as well as the wild type of this
G6PD mutant, all embryopathies in phenytoin-treated +/- and -/-
G6PD-deficient dams were increased above the values observed in
phenytoin-treated +/+ G6PD-normal controls (P<0.05).
However, when analyzed by maternal phenotype, unlike embryonic
phenotype (see below), these values were only different from vehicle
controls of the same maternal phenotype among +/+ dams (Fig. 4)
.
Embryonic phenotype
With respect to embryonic phenotype, the higher dose of phenytoin
(65 mg/kg) caused a 2.4-fold increase in embryonic DNA oxidation in
G6PD-deficient fetuses compared to G6PD-normal littermates
(P<0.03) (Fig. 5
, upper panel, inset), and a decrease in
fetal body weight in all phenotypes, with the weight loss being
progressively worse with one or two mutated G6PD alleles (Fig. 5
, lower
panel). Thus, compared to vehicle controls of the same phenotype, the
phenytoin-initiated decrease in fetal weight was 5% in +/y G6PD-normal
fetuses, 9.5% in +/- G6PD-deficient fetuses, and 9% in -/-
and -/y G6PD-deficient fetuses (P<0.05). The high dose of
phenytoin also appeared to enhance postpartum lethality in all
G6PD-deficient fetuses, although these differences were not
statistically significant (Fig. 5
, upper panel).
Among only those fetuses exposed to phenytoin (65 mg/kg), compared with
+/y G6PD-normal fetuses, which had no postpartum lethality,
phenytoin-initiated postpartum lethality and decreased fetal body
weight were substantially worse in +/- G6PD-deficient fetuses,
and even more so in -/- and -/y G6PD-deficient fetuses
(P<0.05) (Fig. 5)
. Thus, postpartum lethality initiated by
phenytoin (65 mg/kg) was increased 2.9-fold in -/- and -/y fetuses
compared with +/- fetuses, which in turn were substantially more
affected than +/y G6PD-normal fetuses (25 vs. 0%)
(P<0.05). Similarly, with phenytoin-exposed fetuses, the
mean weight of combined -/- and -/y fetuses was decreased by 12.9%
compared with +/- fetuses, which in turn had weights 10.5% lower
than those in +/y G6PD-normal fetuses (P<0.05). There were
no apparent gender differences in teratological susceptibility
(data not shown).
Similar to the effects seen among +/+ dams for mean fetal body weight,
when the data were analyzed by embryonic phenotype, there was evidence
for an embryopathic effect of the lower dose of phenytoin (20 mg/kg).
This pattern was observed for both enhanced postpartum lethality and
decreased fetal body weight but was statistically significant only for
the latter in +/- G6PD-deficient embryos (P<0.05)
(Fig. 5)
.
Teratological syndrome
Viable fetuses from the teratological studies were examined in a
blinded fashion for both external and internal anomalies (Fig. 6
). These structural anomalies, collectively referred to here as a
syndrome, included cleft palate, club foot, dilated bladder, dilated
cerebral ventricles, ectopic kidney, hematoma, microcephaly,
micrognathia, omphalocele, open eye, red nevus, and underdeveloped
renal papilla. A gene-dose response was observed in both vehicle- and
phenytoin-treated groups, and a drug-dose response was observed in
phenytoin-treated animals. This pattern was observed by both embryonic
(Fig. 6)
and maternal (Fig. 6
, inset) phenotype, but was most clearly
defined by embryonic phenotype. Data for -/-(female) and -/y (male)
G6PD-deficient fetuses were analyzed independently and found to be
identical, hence these two phenotypes with all alleles mutated were
combined for all final analyses.
|
For fetuses exposed only to vehicle, anomalies in -/- and -/y
G6PD-deficient fetuses were over twofold higher than in either +/- G6PD-deficient or +/y G6PD-normal fetuses (P<0.05)
(Fig. 6)
. A similar twofold enhancement was observed by maternal
phenotype, but was not statistically significant (Fig. 6
, inset).
For fetuses exposed to the highest dose of phenytoin (65 mg/kg),
compared with +/y G6PD-normal fetuses, anomalies were increased
2.5-fold in +/- G6PD-deficient fetuses and 4.5-fold
(P<0.05) in -/- and -/y G6PD-deficient fetuses (Fig. 6)
.
For all G6PD-deficient fetuses, this enhancement also was observed
relative to vehicle controls of the same phenotype
(P<0.05). The higher dose of phenytoin produced more
anomalies than the lower dose (20 mg/kg) in -/- and-/y
G6PD-deficient fetuses (P<0.05). With the lower phenytoin
dose, compared with +/y G6PD-normal phenytoin-treated controls,
anomalies were not increased in +/- fetuses, but were increased
almost twofold in -/- and -/y fetuses (P<0.05). When
compared irrespective of embryonic phenotype, the lower dose of
phenytoin caused a 1.6-fold increase in anomalies compared with vehicle
controls (P<0.05).
By maternal phenotype, a similar pattern of phenytoin-enhanced
anomalies was observed, except that +/- G6PD-deficient dams were
as susceptible as -/- dams, and even the lower dose of phenytoin was
teratogenic. The lower phenytoin dose caused over a threefold, albeit
nonsignificant, increase in anomalies in +/- G6PD-deficient dams
compared with +/+ G6PD-normal phenytoin-treated dams and a significant
fourfold increase when compared with pooled +/+ and +/- vehicle
controls (P<0.05) (Fig. 6
, inset). With the higher dose of
phenytoin compared with respective vehicle controls, there was a
fivefold increase in anomalies in +/- G6PD-deficient dams
(P<0.05). A similar but lower twofold enhancement was
observed in -/- 6PD-deficient dams because of the 2.6-fold increase
in anomalies in the vehicle controls for this phenotype. When combined,
+/- and -/- G6PD-deficient dams treated with the higher dose of
phenytoin had over fourfold more anomalies than respective
phenytoin-treated +/+ G6PD-normal dams (P<0.05). Similarly,
combined +/- and -/- G6PD-deficient dams treated with either a
high or low dose of phenytoin had three- and twofold more anomalies,
respectively, compared with their combined vehicle controls
(P<0.05). The higher dose of phenytoin was more teratogenic
than the lower dose, but unlike by embryonic phenotype, the difference
by maternal phenotype was statistically significant only when the data
for +/- and -/- G6PD phenotypes were combined
(P<0.05).
In G6PD-normal animals, phenytoin in either dose did not increase fetal
anomalies compared with vehicle controls in either +/y fetuses (Fig. 6)
or +/+ dams (Fig. 6
, inset).
Mutational analysis and genotyping for late in utero
embryonic death
We were able to confirm the reported mutation (21)
causing the heritable decrease in G6PD activity in our strain of mutant
mice. This involved direct sequencing of not only the full-length
mutant mouse cDNA, which includes the 5' and 3' untranslated regions
and the entire coding region for the G6PD protein, but also 6 of the 12
intronic G6PD genomic DNA regions, including introns 4, 6, 7, and
10–12. Apart from the reported functional mutation in our C3H mutant
mouse strain, we found a single silent A to C mutation, located at base
718 of the mouse cDNA sequence. This silent mutation differs from the
reported C57 mouse strain cDNA (11)
and corresponds to the
human DNA sequence at base 15361 (24)
, maintaining the
amino acid sequence code of GGC for glycine.
We also noted that the single functional point mutation, an A to T
transversion in the 5' untranslated region of the gene at the
penultimate base of the 3' end of exon 1, results in the destruction of
a DdeI restriction enzyme site in the mutant allele. We used
this information to develop a genotyping assay to characterize the
functional G6PD mutation in the remnants of dissectable fetal
resorptions from our teratological studies (Fig. 7
). In short, we designed a pair of PCR primers around the mutation site,
based on the known sequence of the G6PD gene, expected to produce a PCR
fragment from mouse genomic DNA of 269 base pairs (bp) in length. For
normal mice, subsequent digestion of PCR products with the
DdeI restriction enzyme would produce two cleaved fragments
of 214 bp and 55 bp, whereas mice with mutations in all G6PD alleles
would not show a change in fragment size, and heterozygotes were
expected to show all three fragments. All three bands were initially
observed, as expected, by polyacrylamide gel electrophoresis (PAGE)
analysis, but subsequent analysis of samples was performed using
agarose gel electrophoresis, in which the 55 bp fragment did not
resolve from the leading xylene cyanole band of the loading buffer
(Fig. 7)
.
|
Based on an approach previously established in our laboratory
(25)
, these studies allowed us to determine the potential
protective role of G6PD with respect to in utero fetal death
(Fig. 8
). Among the dissectable fetal resorptions reflecting in
utero death late in gestation, compared with +/y G6PD-normal
fetuses, G6PD-deficient fetuses with a mutation in either one (+/-) or
all (-/- and -/y) G6PD alleles had over six- and fivefold increases,
respectively, in in utero deaths when compared independent
of treatment (P<0.0001) (Fig. 8)
. When analyzed by
treatment, compared with the +/y G6PD-normal embryonic genotype, the
incidence of late fetal resorptions was substantially increased to
approximately the same extent in all G6PD-deficient embryonic genotypes
by both vehicle and phenytoin treatments (P<0.03), although
the difference was statistically marginal in vehicle-exposed +/-
fetuses (P<0.06) (Fig. 8
, inset).
|
Early in utero embryonic death
In dams treated with phenytoin, there also were a
substantial number of pinpoint resorptions that were too small to
dissect, reflecting in utero embryonic death early in
gestation (Fig. 9
). This early embryonic death, which was both G6PD gene dose- and
phenytoin dose-related, occurred postimplantation, as the number of
implantations did not differ among the maternal phenotypes. The number
of implantations (mean ± SD), irrespective
of treatment, was 9.79 +/- 0.85 for +/+ G6PD-normal dams vs.
8.13 ± 3.20 for ± and 8.82 ± 3.57 for -/-
G6PD-deficient dams. In +/- and -/- G6PD-deficient dams, both
the low dose (20 mg/kg) and the high dose (65 mg/kg) of phenytoin
substantially enhanced the incidence of pinpoint resorptions to a
similar extent compared with respective vehicle controls, which had no
such early resorptions in any treatment group (P<0.01)
(Fig. 9)
. With the high dose of phenytoin, the incidence of early
embryonic death was enhanced three- and fivefold (P<0.02),
respectively, in +/- and -/- G6PD-deficient dams compared with
+/+ G6PD-normal dams, and the incidence was 1.6-fold higher in -/-
dams compared with +/- dams (P<0.03). A remarkably
identical pattern of maternal phenotypic susceptibility to, and
magnitude of, early embryonic death was observed with the low dose of
phenytoin (P<0.01), although the enhancement in +/-
dams was statistically marginal (P<0.07).
|
Embryonic and maternal G6PD activities
On GD 13, within the period of organogenesis, whole embryo G6PD
activity was 43% lower in G6PD-deficient embryos compared with
G6PD-normal littermates (Fig. 10
). In this case, G6PD-deficient embryos included +/-, -/y, and -/-
genotypes, because these studies preceded the establishment of our
genotyping technique. At the end of gestation, on day 19, G6PD activity
in normal fetuses was lower than that during organogenesis and remained
at this same lower level at the time of weaning (Fig. 10)
and into
adulthood (see Fig. 2
). A similar pattern was observed in
G6PD-deficient fetuses, whereby compared with GD 13 whole embryo
activity, RBC activity on day 19 among combined fetuses with a mutation
in one or all G6PD alleles had declined by 64% (P<0.05)
and remained constant thereafter. On day 19, compared with +/+
G6PD-normal littermates, RBC G6PD activity in +/- G6PD-deficient
fetuses was 48% of normal (P<0.05) and in -/y and -/-
G6PD-deficient fetuses was 18% of normal (P<0.05). This
pattern and the respective RBC activities remained similar at weaning,
58 and 19%, respectively, of normal (P<0.05) (Fig. 10)
.
|
The relative distribution of G6PD activities in different organs was
similar for all phenotypes, with up to an 11-fold difference between
the highest (spleen) and the lowest (heart) activities (Fig. 11
). In spleen, compared with +/+ G6PD-normal dams, G6PD activities
in +/- and -/- G6PD-deficient dams were 79 and 27%,
respectively, of normal (P<0.05), and similar patterns were
evident for other organs and blood.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
, 16
, 26)
. However, most embryonic tissues in utero are
known to synthesize only low levels of most enzymes involved in the
detoxification of xenobiotic reactive intermediates and reactive oxygen
species (ROS), particularly up to the end of organogenesis, the
critical period of teratological susceptibility (4
, 17)
.
On the other hand, elevated G6PD activity during embryonic development
corresponds to periods of both increased cellular proliferation and DNA
synthesis (18
, 19)
, suggesting that G6PD activity may be
important for normal development. Indeed several groups have shown peak
G6PD activities in various nonmurine whole embryos and their organs
occur during the early stages of development and subsequently decline
to adult levels near the end of gestation (18
, 27
28
29)
.
These results correspond to our own and are suggestive of the critical
importance of G6PD during embryonic development. Moreover, similar to
avian species, which have nucleated RBCs (30
, 31)
, samples
from human maternal circulation found the amount of nucleated fetal
RBCs per 40 ml of blood increased as gestation progressed, from 0.1%
at 6 wk to 1% at term (32)
; the latter is comparable to
the proportion of reticulocytes among all cells in normal human adult
blood (33)
. If this increased percentage of nucleated RBCs
occurs similarly within the fetus, then this may assist in providing
protection for a normal developing fetus at a particular gestational
age because of the ability to synthesize more G6PD enzyme under
conditions of oxidative stress. However, although chickens have
nucleated RBCs, they are still susceptible to phenytoin teratogenesis
(34)
. We accordingly hypothesized that G6PD-deficient
embryos would be unable to mount an adequate antioxidative response to
normal developmental oxidative stress, as well as that initiated by
xenobiotics, and hence be highly susceptible to ROS-mediated damage
(Fig. 1)
.
In our mutant C3H mouse model, RBC G6PD activities at weaning of
heterozygous (+/-) and homozygous (-/-) G6PD-deficient females and
hemizygous (-/y) G6PD-deficient males were 56, 16, and 21% ,
respectively, of the activity in congenic wild-type G6PD-normal female
(+/+) and male (+/y) mice, similar to previously reported respective
activities of 60, 15, and 20% (35)
. More importantly,
during the period of organogenesis, while normal embryonic G6PD
activity was 1.6-fold higher than RBC activity at weaning,
G6PD-deficient embryos had <58% of the activity in G6PD-normal
littermates, potentially leaving the G6PD-deficient embryos more
susceptible to embryopathic oxidative stress. G6PD activities likely
were substantially lower in -/- and -/y than +/-
G6PD-deficient embryos, but these studies were performed before the
development of our genotyping assay, precluding subclassification by
embryonic genotype. However, relative differences can be inferred from
fetal G6PD activities on GD 19 when, compared with +/y G6PD-normal
littermates, +/- G6PD-deficient fetuses had only 48% of normal fetal
activity, and fetuses with all alleles mutated (-/- and -/y) had
only 18% of normal activity. The observed enhanced susceptibility of
G6PD-deficient mice to the embryopathic effects of both normal
developmental oxidative stress, and even more so to that initiated by
the ROS-initiating anticonvulsant drug phenytoin, indicates that G6PD
is a critical embryoprotective enzyme. Embryonic DNA oxidation and a
broad spectrum of apparent ROS-mediated embryopathies were enhanced in
G6PD-deficient animals, including fetal resorptions (in
utero death), postpartum and preweaning lethality, decreased fetal
body weight, and a syndrome of structural anomalies, including open
eye, facial, renal, bladder, and cerebral ventricular defects. This
C3H-derived mutant strain, similar to previous findings using inbred
C3H mice (36)
, is resistant to one of the hallmarks of
phenytoin teratogenicity, cleft palates. Although the number of fetuses
affected was too small for statistical analysis, it was interesting
that cleft palates were observed only in G6PD-deficient fetuses treated
either with vehicle (1 out of 50 +/- fetuses and 1 out of 15 -/y
fetuses) or phenytoin (65 mg/kg) (1 out of 28 +/- fetuses). The
enhanced susceptibility of G6PD-deficient animals also corroborates
other evidence indicating the importance of oxidative stress in the
molecular mechanism of phenytoin teratogenesis (37)
and
further suggests that, contrary to the apparent safety of oxidizing
drugs like phenytoin in adult G6PD-deficient patients (9)
,
embryos in general, and G6PD-deficient embryos in particular, may be at
serious risk.
Normal developmental oxidative stress and teratogenesis
In untreated mice allowed to deliver spontaneously in the breeding
colony, the decreased litter size at birth, and increased preweaning
death in homozygous (-/-) G6PD-deficient dams (Fig. 3)
provides the
first direct evidence that a physiological level of endogenous
oxidative stress during development can be embryopathic and may
contribute substantially to both apparent infertility and postnatal
death. In addition, this is the first direct evidence that G6PD is a
critical embryoprotective enzyme for normal development.
In vehicle-treated mice killed just before the time of delivery and
examined more comprehensively, a striking, broader spectrum of
increased embryopathies, including fetal resorptions, postpartum
lethality, decreased fetal weight, and a syndrome of teratological
anomalies, was observed in G6PD-deficient animals (Figs. 4
5
6)
,
suggesting an extensive embryopathic potential for developmental levels
of ROS production and a broad embryoprotective role for G6PD. Given
that no pinpoint resorptions reflecting early in utero death
were observed in vehicle controls, and there were no differences in the
number of implantations among all groups, it appears that the apparent
infertility in G6PD-deficient animals is because of postimplantational,
relatively late in utero death. The obligatory role of ROS
in the mechanism of vehicle-initiated embryopathies is suggested by the
observation that these effects were found only in G6PD-deficient dams
and fetuses. By both maternal and embryonic genotype, all developmental
parameters for vehicle-treated G6PD-normal animals were comparable to
the normal range for extensive control groups from other murine strains
studied in our laboratory (17)
. Previous in
vivo studies in our laboratory also have shown that the vehicle
does not measurably enhance oxidative stress and hydroxyl radical
formation (38)
, indicating that the observed embryopathies
in vehicle-treated controls are the result of normal developmental
oxidative stress. For many embryopathies (fetal resorptions, postpartum
lethality, decreased fetal weight), increased susceptibility was as
great in +/- as in -/- G6PD-deficient dams, indicating a
potentially widespread developmental relevance of G6PD deficiencies. By
embryonic phenotype, the substantially increased severity of
vehicle-initiated embryopathies (Fig. 5)
, including teratologic
anomalies (Fig. 6)
, in G6PD-deficient fetuses provided direct and
proximal evidence that G6PD-deficient fetuses are exquisitely
susceptible to developmental oxidative stress. While -/- and -/y
G6PD-deficient fetuses were most adversely affected for several
embryopathies, there was evidence of a gene-dose effect, wherein +/- fetuses also had significantly decreased body weight and appeared
to be at increased risk of postpartum lethality compared with +/y
G6PD-normal fetuses. The significant intermediary embryopathic risk
for +/- dams and fetuses is consistent with their intermediate
deficiency in G6PD activity.
Using direct sequencing in the mutant C3H model, we did not find any
functional mutations in either the entire coding region or the 3'
untranslated region of the cDNA, nor in at least 6 of the 12 intronic
regions of genomic DNA. We did find a strain difference between our C3H
mutant mouse strain and the reported C57 mouse cDNA (11)
at base 718; however, this sequence alteration does not change the
amino acid code for glycine at this position and as such is not
expected to have any effect on G6PD activity. At that time, the
functional mutation was published, identifying a single point mutation,
an A to T transversion, in the 5' untranslated region of the gene at
the penultimate base of the 3' end of exon 1 (21)
. We
confirmed this mutation and then developed a genotyping assay that
allowed us to determine the G6PD genotype in the dissectable remnants
(resorptions) of embryos that died in utero (Fig. 7)
, as we
previously had successfully used in characterizing the role of p53 as a
teratological suppressor gene (25)
. This approach is
applicable to embryos that die later in gestation leaving a sufficient
amount of tissue to be reliably dissectable without maternal tissue
contamination. In vehicle-exposed embryos, the threefold greater
incidence of resorptions for -/- and -/y G6PD-deficient fetuses
compared with +/y G6PD-normal littermates (Fig. 8
, inset) shows
directly that normal developmental oxidative stress can play a major
role in in utero death, with G6PD serving as a critical
embryoprotective pathway. Remarkably, a similar threefold increase in
fetal resorptions was observed even in +/- G6PD-deficient
embryos, although this apparent enhancement was only marginally
significant (P<0.06). However, the incidence of these
resorptions appeared to be at a maximal level that was not further
enhanced by phenytoin, and when the data were analyzed independent of
treatment, heterozygous G6PD-deficient embryos were as susceptible as
embryos with mutations in all G6PD alleles, demonstrating a highly
significant sixfold increase over G6PD-normal littermates (Fig. 8)
and
indicating a potentially broad population at risk.
Embryos dying early in gestation were observed as nondissectable pinpoint resorptions, however this earlier embryopathy was not observed with vehicle-exposed embryos of any genotype, suggesting that the rate as well as the extent of embryolethality initiated by developmental oxidative stress are less than that initiated by xenobiotics. As well, the extent of this early embryopathy, especially in G6PD-deficient dams, likely contributed to the apparent lack of a xenobiotic-initiated effect among resorptions occurring later in gestation. In general, the developmental risk in G6PD-deficient heterozygotes, in some cases equivalent to that for dams and fetuses with mutations in all G6PD alleles, reveals a potentially substantial clinical risk from G6PD deficiencies under conditions of normal developmental oxidative stress.
Xenobiotic-initiated oxidative stress and teratogenesis
Phenytoin and related proteratogens, including the sedative drug
thalidomide, are bioactivated by embryonic prostaglandin H synthases,
lipoxygenases, and related enzymes to a free radical intermediate,
which causes embryonic oxidative stress, hydroxyl radical formation,
and oxidative damage to DNA and other cellular macromolecules in
embryonic tissues (4
, 17
, 37
, 39
, 40)
. Thus, the
dose-dependent increase in embryopathies caused by phenytoin by both
maternal and embryonic analyses, together with the enhanced
susceptibility of G6PD-deficient dams and embryos compared with vehicle
controls (Figs. 4
5
6
, 8)
and the enhanced phenytoin-initiated DNA
oxidation in G6PD-deficient embryos (Fig. 5
, inset), provide the first
direct evidence that G6PD is a major embryoprotective enzyme for
xenobiotic-initiated oxidative stress and embryonic macromolecular
target damage. The susceptibility of heterozygous (+/-) G6PD-deficient
animals to phenytoin embryopathies was particularly remarkable,
generally exhibiting a risk intermediate, if not equal, to that of
homozygous (-/-) and hemizygous (-/y) G6PD-deficient dams and
fetuses, depending on the parameter. This pattern differed somewhat
only with in utero death, wherein the incidence of early but
not late fetal resorptions was enhanced over vehicle controls (Figs. 8
, 9)
, although both types of in utero death were substantially
enhanced in G6PD-deficient embryos, and the risk in heterozygotes was
similar to that for homozygous and hemizygous G6PD-deficient embryos.
To an extent greater than that in the vehicle controls, the
embryopathic susceptibility of even heterozygous G6PD-deficient dams
and embryos to phenytoin suggests that G6PD deficiencies may have broad
developmental relevance for exposures to drugs and environmental
chemicals like phenytoin that initiate oxidative stress.
With the lower, 20 mg/kg dose of phenytoin, the enhanced embryopathies
(decreased fetal body weight, early resorptions, teratological
anomalies) compared with vehicle controls was remarkable, as this is
<40% of a threshold teratogenic dose (55 mg/kg) in normal mice
(17)
. For some embryopathies (e.g., early resorptions,
teratological anomalies), a maximal response was achieved with the
lower 20 mg/kg dose in both +/- and -/- and -/y G6PD-deficient
littermates. These results demonstrate that G6PD-deficient animals are
exceptionally susceptible to the embryopathic effects of
xenobiotic-initiated oxidative stress.
By both maternal and embryonic analysis, the dose-dependent increase in
most embryopathies, including teratological anomalies, produced in
G6PD-deficient animals by phenytoin compared with vehicle controls,
corroborates the role of oxidative stress in the molecular mechanism of
phenytoin teratogenesis. The enhanced embryonic DNA oxidation observed
in phenytoin-exposed G6PD-deficient embryos further suggests that
oxidative damage to embryonic cellular macromolecules may play a
proximate role in embryopathic initiation. These results implicating
reactive oxygen species and oxidative damage to embryonic cellular
macromolecules in the mechanism of phenytoin teratogenicity are
consistent with other murine studies in vivo, in embryo
culture, and in vitro demonstrating phenytoin-initiated
formation of reactive oxygen species, oxidative damage to embryonic
cellular macromolecules, and a protective role for antioxidants
(vitamin E, caffeic acid, glutathione) and other antioxidative enzymes
such as glutathione reductase, glutathione peroxidase, superoxide
dismutase, and catalase (37
, 41)
. The embryopathic
importance of enhanced DNA oxidation in G6PD-deficient embryos is
consistent with previous studies of phenytoin and benzo[a]pyrene,
another teratogen known to initiate embryonic oxidative stress, wherein
p53-deficient mice with reduced DNA repair were shown to be more
susceptible to the embryopathic effects of these xenobiotics (25
, 42
, 43)
.
It has been reported that under conditions of oxidative stress, the
increased expression of G6PD activity is primarily a result of an
increased rate of transcription, with a minor contribution from
posttranscriptional modifications (44)
, possibly due to
posttranslational regulation of G6PD by small heat shock proteins
(45)
. Given the above evidence for the involvement of
reactive oxygen species in in utero death and teratogenesis,
it is likely that the biochemical mechanism underlying the
embyroprotective role of G6PD is its production of NADPH essential for
both glutathione reductase-dependent, and possibly catalase-dependent,
detoxification of lipid hydroperoxides and hydrogen peroxide, rather
than G6PD-dependent pentose production. This is consistent with a
recent study of diamide-initiated oxidative stress, which identified
the essential protective role for G6PD as NADPH production, as distinct
from its dispensable role in pentose synthesis, determined by cloning
efficiency of mouse embryonic stem cells wherein the G6PD gene was
selectively knocked out (46)
. A reduction in catalase
function may contribute to the enhanced teratologic susceptibility
observed with G6PD deficiencies, because NADPH is known to maintain
catalase activity not only by preventing the formation of one of the
inactive states of the enzyme (compound II), possibly through electron
tunneling between surface-bound NADPH and the internal heme group, but
also by reducing oxidized states and internal groups of catalase
distinct from compound II, possibly including one of the active states
of the enzyme (compound I) (6
7
8)
. This mechanism also is
consistent with other studies demonstrating a protective effect of
catalase therapy against phenytoin teratogenicity in embryo culture
(20)
and in vivo (41)
. Our results
also show that G6PD, as distinct from isocitrate dehydrogenase and
malic enzyme (47)
, provides the major supply of embryonic
NADPH during organogenesis, and alternative embryonic sources are
inadequate in the face of hereditary G6PD deficiencies.
Epidemiological considerations
The observed decreased litter size at birth in untreated
-/- G6PD-deficient mice and, in vehicle-treated mice, the increased
fetal resorptions in both +/- and -/- G6PD-deficient dams,
would present in humans as an apparent decrease in fertility. Given
that implantations were not decreased in G6PD-deficient dams, the
decrease in viable offspring was because of postimplantation
embryolethality rather than increased oocyte loss and/or increased
preimplantation embryonic death. The gestational timing of
susceptibility to oxidative stress may be because of, at least in part,
the change from anaerobic to aerobic metabolism that begins after
implantation. This change occurs as the mitochondrial electron
transport system and associated enzymes start to become functional and
in rodents is complete with the establishment of the allantoic
circulation. For rat embryos, this process starts between GD 10 and 11
and ends by GD 12.5, with blood circulation through the yolk sac and
embryo by GD 11 (48
49
50)
. For mouse embryos, this process
may run from GD 6 to 9.5, with allantoic circulation through the
placenta by GD 9 (51
52
53)
. Although humans do not have a
yolk sac placenta, comparisons based on stage of development and somite
count between rat and human indicate that GD 11 and 12.5 in the rat are
equivalent to GD 28 and 37 in the human, after which time the switch
from anaerobic to aerobic metabolism should be complete (48
, 54)
.
Accordingly, it is not surprising that this postimplantation
gestational period constitutes a critical window of susceptibility to
ROS-mediated lethality for G6PD-deficient embryos. Indeed, +/-
G6PD-deficient embryos were as susceptible as their -/- and -/y
G6PD-deficient littermates. In humans, expected incidences for
G6PD-deficient -/y males and -/- females born among all
G6PD-deficient groups in Europe were 0.7 and 5.1%, respectively,
whereas the observed incidences of these groups were only 0.3 and 2%,
respectively (2)
. In other words, there were 57% fewer
-/y G6PD-deficient males and 61% fewer -/- G6PD-deficient females
born than expected. These lower birth incidences in human
G6PD-deficient populations imply a lower survival rate than expected
for embryos deficient in G6PD, as was observed in our mouse model.
To date, no major deletions or frameshifts within G6PD have been
identified. Most variants result from point mutations or small
intragenic deletions leading to a decrease of only one or two residues
(14)
. There are few published cases of heritable mild G6PD
deficiencies in erythrocytes from other species, specifically in a
colony of rats (55)
and in one dog out of 3,300 screened
(56)
. In both cases, the mutation was not characterized,
but difficulties in maintaining the rat colony because of increased
mortality and sterility among the affected animals suggests that a
homozygous knockout of G6PD would likely prove embryolethal. Similarly,
there has never been a reported case of a complete human G6PD
deficiency (10)
.
A limited number of recent human studies provide indirect evidence of
other areas where G6PD deficiencies may be pathologically relevant.
G6PD is important for maintaining adequate concentrations of reduced
glutathione, which is necessary for a number of cytoprotective
antioxidative activities, including that of glutatione peroxidase (Fig. 1)
. Graf et al. (57)
found significantly lower glutathione
peroxidase activity among 37 children with the neural tube defect
myelomeningocele, compared with age-matched controls. Weber and
colleagues (58)
reported intractable seizures, repeated
infections, and intolerance to anticonvulsants in four children with
glutathione peroxidase deficiencies, one of whom also was G6PD
deficient. These adverse outcomes may have been because of
granulocytopenia and enhanced ROS-dependent signaling pathways,
respectively, both of which could result from oxidative stress and
inadequate protection from antioxidative enzymes. This is consistent
with results from a rat model of human chronic posttraumatic epileptic
seizures, in which pretreatment with antioxidants (alpha-tocopherol and
selenium) protected against a spectrum of iron-induced peroxidative
injuries, including cavitation, neuronal loss, astrogliosis, and
epileptiform discharges in rat isocortical regions, suggesting that
deficiencies in cytoprotection against peroxidative injury may increase
the risk of recurrent epileptic seizures (59)
. Finally, a
recent study found increased levels of G6PD activity in several
relevant regions of Alzheimer’s brains compared with controls,
possibly reflective of increased levels of cerebral oxidative challenge
(60
, 61)
. In our mouse model, normal adult brain G6PD
activity was one of the lowest of all tissues examined and was further
reduced by >21 and 72%, respectively, in +/- and -/-
G6PD-deficient animals (Fig. 11)
. If oxidative stress is involved in
the mechanism of neurodegenerative diseases (62
, 63)
, then
G6PD deficiencies may contribute to enhanced susceptibility. These
studies are consistent with our observation of enhanced embryopathies
in G6PD-deficient mice and suggest the potential for a broader range of
pathological susceptibilities with G6PD deficiencies that may be
further enhanced by exposure to xenobiotics like phenytoin that
initiate oxidative stress.
The commonly postulated evolutionary pressure for the widespread
prevalence of G6PD deficiencies is their advantage in providing
resistance to malaria. This view is supported by the almost complete
overlap of regions of increased incidence of G6PD deficiency with those
of elevated incidence of malarial infections (2
, 10)
.
Recently, both heterozygous and hemizygous G6PD-deficient people were
confirmed to be 46 and 58% more resistant, respectively, to severe
malarial infection (64)
. Malaria-infected normal RBCs
demonstrate increased lipid peroxidation (65)
presumably
because of parasite-generated
H2O2 (66)
and
decreased antioxidant levels, with no change in G6PD activity. The
authors concluded this to be a possible defense mechanism by which
infected host cells seek to produce and maintain unchecked oxidative
stress to their advantage in a self-sacrificing attempt to limit the
spread of malarial infection. As such, G6PD-deficient cells are more
sensitive to this increased oxidative damage (66)
, which
may help mediate their increased resistance to such infections, either
by predisposing the cell to premature destruction by the
reticuloendothelial system (67)
or
macrophage-initiated lysis (68)
and/or impaired growth of
the parasite in G6PD-deficient erythrocytes as suggested by in
vitro studies (69
, 70)
, thus abolishing a suitable
incubation environment for parasite development. However, this apparent
selective advantage must be balanced by equally disadvantageous
selective pressures, not the least of which include the developmental
risks observed in our mouse study and suggested in human
epidemiological studies discussed above. Developmental risks may well
be increasing as a result of more frequent exposure to drugs and
environmental chemicals such as phenytoin that initiate oxidative
stress exquisitely toxic to G6PD-deficient embryos. Such factors may be
contributing to the apparently restricted increase in the prevalence of
G6PD deficiencies in malaria-infected regions (64)
.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
, 10
, 26)
. Our results provide the
first evidence that G6PD is a developmentally critical enzyme,
protecting the embryo from both endogenous and xenobiotic-initiated
oxidative stress and DNA damage that produce a broad range of
embryopathies, including teratologic anomalies. The developmental risk
is potentially substantial, given both the common incidence of
hereditary G6PD deficiencies and the susceptibility of even
heterozygous G6PD-deficient animals.
The enhanced susceptibility of G6PD-deficient embryos to phenytoin
corroborates other studies implicating embryonic oxidative stress and
DNA damage in the molecular mechanism of embryopathies caused by
phenytoin and related teratogens (37)
. G6PD deficiencies
completely redefine thresholds for teratologic susceptibility to
xenobiotic exposure, as evidenced by the enhanced susceptibility of
even heterozygous G6PD-deficient embryos to phenytoin embryopathies at
a dose that is well below the minimal teratogenic dose in G6PD-normal
mice. The developmental risk from xenobiotics that initiate oxidative
stress appears to be considerably greater than that observed in adults,
because adult patients with G6PD deficiencies are reported to be
nonsusceptible to RBC hemolysis initiated by drugs like
phenytoin.
25 U/g protein.
Lower panel, mean fetal body weights. Asterisks indicate
a difference from respective G6PD-normal groups
(P<0.05), § indicates a difference from vehicle
controls from the same phenotype (P<0.05), and 
