Báo cáo Y học: Excessive vitamin A toxicity in mice genetically deficient in either alcohol dehydrogenase Adh1 or Adh3 docx

Excessive vitamin A toxicity in mice genetically deficient in either
alcohol dehydrogenase
Adh1
or
Adh3
Andrei Molotkov, Xiaohong Fan and Gregg Duester
Gene Regulation Program, Burnham Institute, La Jolla, CA, USA
Alcohol dehydrogenase (ADH) deficiency results in
decreased retinol utilization, but it is unclear what physio-
logical roles the several known ADHs play in retinoid
signaling. Here, Adh1, Adh3,andAdh4 null mutant mice
have been examined following acute and chronic vitamin A
excess. Following an acute dose of retinol (50 mgÆkg
)1
),
metabolism of retinol to retinoic acid in liver was reduced
10-fold in Adh1 mutants and 3.8-fold in Adh3 mutants, but
was not significantly reduced in Adh4 mutants. Acute retinol
toxicity, assessed by determination of the LD
50
value, was
greatly increased in Adh1 mutants and moderately increased
in Adh3 mutants, but only a minor effect was observed in
Adh4 mutants. When mice were propagated for one gen-
eration on a retinol-supplemented diet containing 10-fold
higher vitamin A than normal, Adh3 and Adh4 mutants had
essentially the same postnatal survival to adulthood as wild-
type (92–95%), but only 36% of Adh1 mutants survived to
adulthood with the remainder dying by postnatal day 3.
Adh1 mutants surviving to adulthood on the retinol-
supplemented diet had elevated serum retinol signifying a

clearance defect and elevated aspartate aminotransferase
indicative of increased liver damage. These findings indicate
that ADH1 functions as the primary enzyme responsible for
efficient oxidative clearance of excess retinol, thus providing
protection and increased survival during vitamin A toxicity.
ADH3 plays a secondary role. Our results also show that
retinoic acid is not the toxic moiety during vitamin A excess,
as Adh1 mutants have less retinoic acid production while
experiencing increased toxicity.
Keywords: alcohol dehydrogenase; retinol; retinoic acid;
vitamin A; toxicity.
Mammalian alcohol dehydrogenase (ADH, EC 1.1.1.1) is
encoded by a family of genes closely linked on human
chromosome 4 and mouse chromosome 3, with all genes in
the same transcriptional orientation [1]. Five distinct classes
of mammalian ADH have been identified that differ
significantly in primary structure, catalytic activity with
various alcohols, and gene expression patterns [2]. The
physiological roles of these several classes of ADH are not
fully established, but studies along this line have led to the
hypothesis that retinoid metabolism is one of these roles [3].
The ability of horse liver ADH to oxidize retinol to retinal
in vitro was recognized early [4] and further studies
demonstrated retinol activity using purified class I
(ADH1), class II (ADH2), and class IV (ADH4) enzymes
[5–9]. Class III (ADH3) was not originally associated with
activity for retinol oxidation, but recent studies using a more
sensitive assay have demonstrated its ability to catalyze
retinol oxidation [10]. ADHs are cytosolic enzymes with
many forms having relatively high catalytic activity for

retinol oxidation compared with microsomal enzymes.
Several microsomal short-chain dehydrogenase/reductase
(SDR) enzymes have been reported to oxidize retinol to
retinal, but with activities that are 100-fold less than that of
ADH1 [3,11].
The physiological functions of ADHs in retinoid meta-
bolism are now being examined genetically in null mutant
mice. A role for ADH4 in protection against vitamin A
deficiency has been demonstrated in Adh4
–/–
mice that suffer
an increased rate of stillbirths relative to wild-type mice
when maintained on a vitamin A deficient diet during
gestation [12]. Adh1
–/–
mice have been shown to have
reduced metabolism of both ethanol and retinol, indicating
that ADH1 is likely to play a role in retinoid metabolism
in vivo [13]. In addition to functioning in the production of
retinoic acid (RA) for development, which is particularly
critical during vitamin A deficiency as pointed out by studies
on Adh4
–/–
mice [12], it is possible that ADHs also function
in the oxidative elimination of excess retinol to prevent
vitamin A toxicity.
The toxicity of excess vitamin A has been well established
[14–16] resulting in recommendations that consumption of
liver or vitamin A supplements be limited to avoid excess
exposure to retinol [17]. The pathways for clearance of

excess retinol have been proposed to involve both oxidative
and nonoxidative mechanisms as reviewed [18]. The main
oxidative pathway for retinol turnover has been suggested
to involve the oxidation of retinol to retinal, then oxidation
of retinal to RA followed by glucuronide conjugation of the
acid and/or 4-hydroxylation of RA. The key enzyme
initiating this catabolic pathway, i.e. the one responsible
for oxidizing excess retinol to retinal, has not been identified.
In order to further investigate this metabolic pathway using
genetic means, we have compared several strains of ADH
Correspondence to G. Duester, Gene Regulation Program, Burnham
Institute, 10901 North Torrey Pines Road, La Jolla, CA 92037, USA.
Fax: + 1 858 646 3195, Tel.: + 1 858 646–3138,
E-mail:
Abbreviations: ADH, alcohol dehydrogenase; Adh1, mouse class I
ADH gene; Adh3, mouse class III ADH gene; Adh4, mouse class IV
ADH gene; AST, aspartate aminotransferase; RA, retinoic acid; SDR,
short-chain dehydrogenase/reductase.
Enzyme: alcohol dehydrogenase (EC 1.1.1.1).
(Received 9 January 2002, revised 15 April 2002,
accepted 16 April 2002)
Eur. J. Biochem. 269, 2607–2612 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02935.x
null mutant mice (Adh1
–/–
, Adh3
–/–
,andAdh4
–/–
) with wild-
type mice for the effects of acute and chronic retinol

treatment. The results demonstrate that ADH1 is the key
enzyme essential for efficient elimination of excess retinol,
thus indicating that it functions as the initiator of the
oxidative pathway. These findings also have implications for
the mechanism of vitamin A toxicity.
MATERIALS AND METHODS
Maintenance of mouse strains
Mice carrying targeted disruptions of Adh1, Adh3 [13] and
Adh4 [12] have been previously described. These null mutant
mice as well as wild-type litter-mates were propagated on
Purina 5015 Mouse Chow unless specified otherwise. This is
a standard mouse diet containing 30 IUÆg
)1
vitamin A.
Dietary retinol supplementation
Mice were propagated on Purina 5755 Basal Diet supple-
mented with additional retinyl acetate to bring the total
vitamin A concentration to 300 IUÆg
)1
, all in the form of
retinyl acetate, which is quickly hydrolyzed to retinol in the
digestive tract. Adult female mice were placed on the retinol-
supplemented diet for 2 weeks, then mated with a male
while still on this diet to generate offspring. Offspring were
maintained on the retinol-supplemented diet after weaning.
Lethal dosing of retinol
For lethal dose evaluation, mice were given oral doses of
retinol, as described previously [14]. Male 14-week-old mice
were used for all strains examined. All-trans-retinol (Sigma
Chemical Co., St Louis, MO, USA) was dissolved in corn

oil and administered by oral intubation at 0.2 mL per 10 g
of body weight. Doses ranged from 0.5 to 3.5 gÆkg
)1
.
Lethality was monitored daily over 14 days after retinol
administration. Doses resulting in the death of 16% (LD
16
),
50% (LD
50
), or 84% (LD
84
) of the mice by day 14, plus the
95% confidence limits for the LD
50
dose, were calculated
using the methods of Litchfield & Wilcoxon [19].
Acute retinol administration for tissue retinoic acid
determination
Retinol was administered essentially as described previously
[20]. All-trans-retinol (Sigma) was dissolved in acetone/
Tween 20/water (0.25 : 5 : 4.75, v/v) and a dose of
50 mgÆkg
)1
was injected orally to age- and weight-matched
female mice. After 2 h, liver was collected and stored at
)20 °C until HPLC analysis.
HPLC quantitation of retinoic acid and retinol
For tissue retinoic acid determination, liver (250 mg) was
homogenized on ice in 2 mL of methanol/acetone (50 : 50,

v/v). For serum retinol determination, blood was collected
andstoredat)20 °C until analysis. Serum (200 lL) was
extracted with 2 mL of methanol/acetone (50 : 50, v/v).
After centrifugation at 10 000 g for 10 min at 4 °C, the
organic phases from liver or serum extracts were evaporated
under vacuum. Residues were dissolved in 200 lLof
methanol/dimethylsulfoxide (50 : 50, v/v) and injected into
the HPLC system to quantitate retinoids using all-trans-
retinol and all-trans-retinoic acid (Sigma) as standards.
Reverse-phase HPLC analysis was performed using a
MICROSORB-MVTM 100 C18 column (4.5 · 250 mm;
Varian) at a flow rate of 1 mLÆmin
)1
. UV detection was
carried out at 340 nm. Mobile phase consisted of 0.5
M
ammonium acetate/methanol/acetonitrile (25 : 65 : 10, v/v/
v; solvent A) and acetonitrile (solvent B). The gradient
composition was (only solvent B is mentioned): 0% at the
time of injection; 30% at 1 min; 35% at 14 min; and 100%
at 16 min.
Measurement of aspartate aminotransferase levels
in serum
Aspartate aminotransferase/glutamic oxalacetic transami-
nase (AST/GOT) activity was measured in mouse serum
using the Sigma Diagnostics Transaminase kit following
manufacturer’s procedure. In brief, 200 lLofserumwas
mixed with substrate and incubated for 1 h at 37 °C. After
1 h, colour reagent was added and samples were left at
room temperature for 20 min. The reaction was stopped by

adding 0.4 N NaOH and the A
505
was then measured. Data
are reported as Sigma–Frankel (SF) units per mL of serum.
Statistics
Statistical significance was determined for raw data using
the unpaired Student’s t-test (
STATISTICA
version 5.0).
RESULTS
Metabolism of acute dose of retinol
The main pathway for retinol turnover begins with the
oxidation of retinol to retinal followed very quickly by
further oxidation of retinal to RA, which then accumulates
before it is further metabolized [18]. In order to examine the
in vivo contribution of ADH1, ADH3, and ADH4 to retinol
metabolism, wild-type mice as well as null mutant mice
deficient in these ADHs were treated orally with a
50-mgÆkg
)1
dose of all-trans-retinol and 2 h later, all-trans-
RA was quantitated in liver. Under these conditions, wild-
type mice produced a large amount of RA (2.0 lgÆg
)1
)
whereas Adh1
–/–
mice produced 10-fold less RA
(0.21 lgÆg
)1

) (Fig. 1). Adh3
–/–
mice exhibited a 3.8-fold
reduction in RA production (0.53 lgÆg
)1
) compared to
wild-type, and Adh4
–/–
mice exhibited a small decrease in
RA (1.49 lgÆg
)1
) that was not statistically significant
(Fig. 1). These results indicate that ADH1 plays a dominant
role in clearance of an acute dose of retinol, and that ADH3
also contributes to a lesser extent, but that ADH4 plays little
or no role in liver retinol metabolism.
Retinol lethal dose
In order to determine if ADH reduces the toxicity of a
supraphysiological dose of retinol, we determined the LD
50
values for each mutant strain. Our wild-type mice exhibited
a retinol LD
50
value of 2.72 gÆkg
)1
, very close to the value of
2.52 gÆkg
)1
previously reported for mice [14]. The retinol
LD

50
value for Adh1
–/–
mice was reduced threefold to
2608 A. Molotkov et al. (Eur. J. Biochem. 269) Ó FEBS 2002
0.9 gÆkg
)1
,whereastheLD
50
value for Adh3
–/–
mice was
reduced 1.8-fold to 1.55 gÆkg
)1
(Table 1). For Adh4
–/–
mice
the LD
50
was reduced only 1.5-fold to 1.74 gÆkg
)1
, thus on
the border of being statistically signficant when considering
the confidence limits (Table 1). Thus, ADH1 plays a
dominant role in providing protection against large acute
doses of vitamin A that are life-threatening, with ADH3
playing a significant secondary role, and ADH4 playing a
very minor role.
Chronic retinol treatment during development
The effect of a chronic modest increase in vitamin A was

examined by propagating mice for one generation on a
retinol-supplemented diet containing 10-fold higher vitamin
A than normal mouse chow (300 IUÆg
)1
in the supplemen-
ted diet vs. 30 IUÆg
)1
in normal chow). The amount of
vitamin A present in the supplemented diet is not beyond
the range that could be ingested naturally if one considers
that it could also be obtained from a diet high in liver,
known to contain 660–1300 IUÆg
)1
vitamin A [17].
This level of retinol supplementation did not have a
negative effect on development of Adh3
–/–
and Adh4
–/–
mice, which behaved similarly to wild-type mice with respect
to survival to adulthood (92–95% survival for all three
strains), but Adh1
–/–
mice exhibited a large reduction in
survival to adulthood (36% survival) (Fig. 2A,B). Adh1
–/–
mice that did not survive were effected very early after birth
as they were found to have decreased maternal suckling
resulting in death by postnatal day 3. No gross malforma-
tions were observed in any of the mice that died (including

limbs and craniofacial region) indicating that overt retinoid
teratogenicity had not occurred with this moderate retinol
treatment. Instead, death was likely due to more subtle
effects of general vitamin A toxicity, perhaps leading to
reduced intake of maternal milk resulting in dehydration
and lack of nourishment. Vitamin A toxicity is known to
reduce food intake leading to weight loss and eventually
death [16].
Liver toxicity following chronic retinol treatment
Adh1
–/–
mice that survived development on the retinol-
supplemented diet had growth rates that were similar to
those of wild-type, Adh3
–/–
,andAdh4
–/–
mice through
7 weeks of age (Fig. 2C). Thus, we examined surviving mice
further to see if any differences could be found.
Compared to mice generated on normal mouse chow, all
strains of mice generated on the retinol-supplemented diet
exhibited higher serum retinol levels (Fig. 3). Compared to
wild-type mice generated on the retinol-supplemented diet,
Adh1
–/–
mice, and to a lesser extent Adh3
–/–
mice, exhibited
significantly higher serum retinol levels; Adh4

–/–
mice were
not significantly different to wild-type (Fig. 3). These
findings indicate that ADH1 provides the greatest protec-
tion against retinol accumulation in the serum when the diet
contains excess vitamin A.
We also examined liver toxicity in these mice by
examination of aspartate aminotransferase (AST) levels in
serum. Serum AST levels were elevated in all mice generated
on the retinol-supplemented diet relative to normal chow,
but the elevation was particularly high in Adh1
–/–
mice
(Fig. 4). Relative to serum AST levels in wild-type mice
generated on the retinol-supplemented diet, Adh1
–/–
mice
displayed a 92% increase, whereas Adh3
–/–
mice exhibited a
37% increase and Adh4
–/–
mice had no significant difference
(Fig. 4). These results essentially mirror the serum retinol
results discussed above. Thus, an increase in serum retinol
due to an absence of ADH1 (and to a lesser extent ADH3)
leads to an increase in liver toxicity.
DISCUSSION
The results reported here establish that ADH1 functions as
a major protective factor against vitamin A toxicity. The

ability of several classes of ADH to perform retinol
WT Adh1 -/- Adh3 -/- Adh4 -/-
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
RA (µg/g)
*
**
Fig. 1. Metabolism of retinol to RA in liver of Adh deficient mice. Liver
all-trans-RA levels were quantitated by HPLC 2 h after a 50 mgÆkg
)1
oral dose of all-trans-retinol for wild-type mice (WT) and each null
mutant strain indicated; all were age-matched adult female mice
(n ¼ 4). Values are mean ± SEM. *P <0.01; **P < 0.03 (null
mutant vs. wild-type).
Table 1. Effect of Adh genotype on retinol lethal dose. *P <0.05(Adh1
–/–
, Adh3
–/–
,andAdh4
–/–
vs. wild-type). The confidence limits for LD

50
are in
parentheses.
Genotype n LD
16
(gÆkg
)1
)LD
50
(gÆkg
)1
)LD
84
(gÆkg
)1
)
Wild-type 15 2.31 2.72 (2.32/3.18) 3.25
Adh1
–/–
12 0.73 0.90 (0.65/1.25)* 1.12
Adh3
–/–
24 1.22 1.55 (1.18/2.03)* 1.97
Adh4
–/–
12 1.37 1.74 (1.32/2.31)* 2.24
Ó FEBS 2002 Vitamin A toxicity and alcohol dehydrogenase (Eur. J. Biochem. 269) 2609
oxidation has been apparent for many years [4–9]. However,
it has previously been unclear what in vivo functions the
several classes of ADH might perform. Also, identification

of microsomal enzymes of the SDR family that can oxidize
retinol to retinal has produced conflicting results as to
whether ADHs or SDRs contribute significantly to retinol
metabolism [3,11]. Previous genetic studies have identified a
physiological role for ADH4 in maintaining sufficient
retinol oxidation during vitamin A deficiency to generate
RA for development [10,12]. It is now apparent from the
genetic studies presented here that ADH1 is responsible for
most of the in vivo oxidation of retinol to RA in liver during
vitamin A excess. We demonstrated a 10-fold reduction in
metabolism of retinol to RA in liver of Adh1
–/–
mice. Thus,
ADH1 is important as the main initiator of the oxidative
pathway for retinol turnover. That this is significant in vivo
is shown by results indicating that Adh1
–/–
mice are much
more sensitive to vitamin A toxicity as demonstrated by a
greatly reduced LD
50
value for retinol as well as greatly
reduced survival during development on a retinol-supple-
mented diet.
The LD
50
studies reported here indicate that ADH3 and
to a lesser extent ADH4 also provide some protection
against death induced by massive nonphysiological doses of
retinol, but that both provide much less protection than that

afforded by ADH1. Our results with Adh3
–/–
and Adh4
–/–
WT
Adh1
Adh3
Adh
4
WT
Adh1
Adh
3
Adh
4
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
normal mouse chow
serum retinol (µg/ml)
*
**
retinol-supplemented

Fig. 3. Serum retinol levels in mice generated on a retinol-supplemented
diet. For wild-type and each Adh null mutant strain indicated, HPLC
was used to quantitate serum retinol levels in mice generated on nor-
mal mouse chow (30 IUÆg
)1
vitamin A) or a retinol-supplemented diet
(300 IUÆg
)1
vitamin A). All mice were first generation 10-week-old
females (n ¼ 5). Values are mean ± SEM. *P < 0.02; **P <0.01
(null mutant vs. wild-type, retinol supplemented).
WT
Adh1
Adh3
Adh4
WT
Adh1
Adh3
Adh4
0
20
40
60
80
100
120
normal mouse chow
*
**
AST (SF units/ml)

retinol-supplemented
Fig. 4. AST levels in wild-type and Adh null mutant mice. Serum AST
levels are shown for mice generated on either normal chow or a retinol-
supplemented diet. All mice were first generation 10-week-old males
(n ¼ 4). Values are mean ± SEM. *P < 0.03; **P < 0.05 (null
mutant vs. wild-type, retinol-supplemented).
A
B
0 5 10 15 20 25 30 35 40 45
0
5
10
15
20
25
30
Adh1
WT
Adh3
Adh4
0 5 10 15 20 25 30 35 40 45
0
20
40
60
80
100
Adh1
WT
Adh3

Adh4
0 5 10 15 20 25 30 35 40 45 50
0
5
10
15
20
25
30
Adh1
WT
Adh
4
Adh
3
C
Age (days)
A
g
e (da
y
s)
Number of mice alive
% survival
Weight (g)
A
g
e (da
y
s)

Fig. 2. Postnatal lethality in Adh1
–/–
mice
generated on a retinol-supplemented diet. (A)
Shown is the number of first generation off-
spring born for wild-type and each Adh null
mutant strain maintained on a retinol-sup-
plemented diet, plus the numbers of offspring
surviving until postnatal day 40 (P40); exten-
sive postnatal death occurred in Adh1
–/–
mice
between birth and P3. (B) The percentage
survival for each mouse strain is shown; only
36% survival was observed for Adh1
–/–
mice
by P40. (C) Shown is the weight gain for the
above wild-type and Adh null mutant mice.
Data for some of the mice reported here is also
described elsewhere [21].
2610 A. Molotkov et al. (Eur. J. Biochem. 269) Ó FEBS 2002
mice also indicate that ADH3 plays a significant role in
metabolism of a dose of retinol to RA in the liver, albeit less
than that of ADH1, but that ADH4 does not effect retinol
turnover significantly in the liver perhaps due to its lack of
expression in liver as detailed further below. These findings
are therefore in agreement with the retinol LD
50
values

observed for these mice.
The data presented here on retinol toxicity provide
additional evidence that ADH3 functions in retinol meta-
bolism. Several studies had originally failed to recognize a
role for ADH3 in retinol metabolism [5,6,8], but recent
studies using a more sensitive assay have shown that ADH3
can oxidize retinol in vitro, but with much lower activity than
ADH1 or ADH4 [10]. Also, Adh3
–/–
mice have a growth
deficiency that can be rescued by dietary retinol supplemen-
tation, and they have greatly reduced survival during
vitamin A deficiency compared to wild-type mice [10].
Thus, there is now much evidence of a role for ADH3 in
metabolism of retinol to RA under physiological conditions
as well as during vitamin A toxicity and deficiency.
In contrast to the LD
50
studies, which examine vitamin A
toxicity under nonphysiological conditions, our results using
the retinol-supplemented diet are applicable to physiological
conditions as the amount of retinol used could be obtained
by natural diets rich in vitamin A (i.e. liver). Under these
dietary conditions we observed that ADH1 provided the
most protection (against both postnatal lethality and liver
toxicity assessed by serum AST), ADH3 provided less
protection (against liver toxicity only), and ADH4 provided
no protection. Thus, our results suggest that ADH1 has
evolved to be more efficient than ADH3 for retinol
turnover, and that ADH4 has not evolved for this function.

In addition, a recent description of mice deficient in both
Adh1 and Adh4 has demonstrated that the loss of both
activities does not result in increased vitamin A toxicity over
that seen for mice deficient in only Adh1 [21]. The role
observed for ADH1 in prevention of vitamin A toxicity also
suggests that the microsomal SDRs reported to metabolize
retinol probably do not play major roles in retinol turnover
or protection against vitamin A toxicity, as their activities
and expression are relatively low compared to ADH1.
The expression patterns of the ADH gene family provide
further understanding into the roles of these enzymes in
retinol turnover observed in the null mutant mice. ADH1
mRNA and protein is expressed at very high levels in mouse
liver, intestine, and kidney [22,23] and it accounts for 0.9%
of mouse liver protein [24]. ADH3 is expressed ubiquitously
[22,23] and accounts for 0.2% of mouse liver protein [24].
ADH4 is not expressed in liver, but is found at highest levels
in the stomach, esophagus, and skin [22,23] and accounts
for 0.07% of mouse stomach protein [24]. Thus, high
expression of ADH1 in liver makes it well-equipped to
handle turnover of large amounts of retinol as we observed.
The ubiquitous expression of ADH3, with a high concen-
tration in the liver, allows it to also contribute significantly
to retinol turnover as we observed, but the lack of ADH4
expression in liver and relatively low expression in other
organs precludes it from being a major player in systemic
retinol turnover consistent with the results provided here.
Mammalian ADH genes were derived from duplications
of an ancestral ADH3 gene conserved in lower vertebrates
(cartilaginous fishes) and invertebrates including Amphi-

oxus [25,26]. As ADH1 did not appear until bony fishes [27]
and ADH4 until amphibians [28], early vertebrates could
have relied upon ADH3 to oxidize retinol to retinal. As it
has now been demonstrated that retinol metabolism by
mouse ADH3 functions to catalyze turnover of excess
retinol, ADH3 can be thought of as a prototype retinol
dehydrogenase that was expanded upon later in evolution.
In particular, evolution of ADH1 may have allowed more
efficient turnover of excess retinol providing a selective
advantage against vitamin A toxicity observed in mice. In
the case of ADH4, evolution of this enzyme may have
provided added protection against vitamin A deficiency
encountered by land animals, as it did not appear until
amphibians and does give protection against vitamin A
deficiency in mice [10,12].
It has been proposed that both the toxic and teratogenic
properties of retinol may be due to its conversion to RA,
which is known to control a signaling pathway involving
RA receptors [16]. Studies in mice have shown that retinol
teratogenicity is associated with increased production of
RA, and that the teratogenicity of retinol can be reduced by
treatment with the ADH inhibitor 4-methylpyrazole, which
also reduces RA production [20]. Interestingly, among the
several classes of ADH, ADH1 is by far the most sensitive to
4-methylpyrazole [24,29]. Combining these inhibitor find-
ings with those here using Adh1
–/–
mice, it is now clear that
when retinol is given at very high doses, it is metabolized
primarily by ADH1 to produce high levels of RA, which are

teratogenic for embryonic development. However, our data
show that when metabolism of retinol to RA is greatly
reduced in Adh1
–/–
mice, there is an increase in retinol
toxicity (rather than teratogenicity) as demonstrated by a
decrease in the lethal dose for retinol in adult mice as well as
reduced survival of newborn mice generated on a retinol-
supplemented diet. In our developmental studies, we
provided a very modest increase in dietary retinol, much
less than that needed to produce retinoid teratogenicity, but
enough to produce toxicity when ADH1 is missing, as
shown by decreased survival of newborn mice and increased
serum AST in those that did survive to adulthood. Thus,
retinol toxicity, as opposed to teratogenicity, occurs when
there is a defect in the ability to turnover retinol oxidatively.
Our findings demonstrate that in order to avoid retinol
toxicity, it is more beneficial to metabolize retinol oxida-
tively to RA than to allow it to be disposed of in any other
way. When oxidation of retinol to RA is severely impaired
as it is in Adh1
–/–
mice, retinol may instead become a
substrate for P450s known to metabolize retinol to
4-hydroxyretinol [30]. This may lead to toxicity as P450-
mediated metabolism requires molecular oxygen and pro-
duces oxygen free radicals that can cause liver damage
[31,32]. In contrast, ADH-mediated metabolism occurs via
dehydrogenation with the cofactor NAD, thus does not
produce oxygen free radicals. Also, retinol toxicity has been

shown to be associated with formation of large amounts of
glucuronidated retinol leading to reduced levels of uridine
diphosphoglucuronic acid suggesting that retinol toxicity
may be due in part to reduction of the substrate needed to
perform other essential glucuronidations [33]. Our findings
point out a fundamental difference in retinol toxicity and
retinol teratogenicity, as the former is mediated by retinol
(or metabolites other than RA) whereas the latter is
mediated by excessive production of RA that leads to
aberrant retinoid signaling during embryogenesis.
Ó FEBS 2002 Vitamin A toxicity and alcohol dehydrogenase (Eur. J. Biochem. 269) 2611
ACKNOWLEDGEMENTS
We thank H. Freeze for providing access to an HPLC system and
F. Mic for useful discussions. This work was supported by National
Institutes of Health Grant AA09731 (G.D.).
REFERENCES
1. Szalai, G., Duester, G., Friedman, R., Jia, H., Lin, S., Roe, B.A. &
Felder, M.R. (2002) Organization of six functional mouse alcohol
dehydrogenase genes on two overlapping bacterial artificial
chromosomes. Eur. J. Biochem. 269, 224–232.
2. Duester, G., Farre
´
s, J., Felder, M.R., Holmes, R.S., Ho
¨
o
¨
g, J O.,
Pare
´
s, X., Plapp, B.V., Yin, S J. & Jo

¨
rnvall, H. (1999)
Recommended nomenclature for the vertebrate alcohol dehy-
drogenase gene family. Biochem. Pharmacol. 58, 389–395.
3. Duester, G. (2000) Families of retinoid dehydrogenases regulating
vitamin A function: production of visual pigment and retinoic
acid. Eur. J. Biochem. 267, 4315–4324.
4. Bliss, A.F. (1951) The equilibrium between vitamin A alcohol and
aldehyde in the presence of alcohol dehydrogenase. Arch. Bio-
chem. 31, 197–204.
5. Boleda, M.D., Saubi, N., Farre
´
s, J. & Pare
´
s, X. (1993) Physiolo-
gical substrates for rat alcohol dehydrogenase classes: aldehydes of
lipid peroxidation, omega-hydroxyfatty acids, and retinoids. Arch.
Biochem. Biophys. 307, 85–90.
6. Yang, Z N., Davis, G.J., Hurley, T.D., Stone, C.L., Li, T K. &
Bosron, W.F. (1994) Catalytic efficiency of human alcohol dehy-
drogenases for retinol oxidation and retinal reduction. Alcohol.
Clin.Exp.Res.18, 587–591.
7. Kedishvili, N.Y., Bosron, W.F., Stone, C.L., Hurley, T.D., Peggs,
C.F., Thomasson, H.R., Popov, K.M., Carr, L.G., Edenberg, H.J.
& Li, T K. (1995) Expression and kinetic characterization of
recombinant human stomach alcohol dehydrogenase. Active-site
amino acid sequence explains substrate specificity compared with
liver isozymes. J. Biol. Chem. 270, 3625–3630.
8. Han, C.L., Liao, C.S., Wu, C.W., Hwong, C.L., Lee, A.R. & Yin,
S.J. (1998) Contribution to first-pass metabolism of ethanol and

inhibition by ethanol for retinol oxidation in human alcohol
dehydrogenase family – implications for etiology of fetal alcohol
syndrome and alcohol-related diseases. Eur. J. Biochem. 254,
25–31.
9. Allali-Hassani, A., Peralba, J.M., Martras, S., Farre
´
s, J. & Pare
´
s,
X. (1998) Retinoids, omega-hydroxyfatty acids and cytotoxic
aldehydes as physiological substrates, and H
2
-receptor antagonists
as pharmacological inhibitors, of human class IV alcohol dehy-
drogenase. FEBS Lett. 426, 362–366.
10. Molotkov, A., Fan, X., Deltour, L., Foglio, M.H., Martras, S.,
Farre
´
s, J., Pare
´
s, X. & Duester, G. (2002) Stimulation of retinoic
acid production and growth by ubiquitously-expressed alcohol
dehydrogenase Adh3. Proc. Natl Acad. Sci. USA 99, 5337–5342.
11. Napoli, J.L. (1999) Interactions of retinoid binding proteins and
enzymes in retinoid metabolism. Biochim. Biophys. Acta Mol. Cell
Biol. Lipids 1440, 139–162.
12. Deltour, L., Foglio, M.H. & Duester, G. (1999) Impaired retinol
utilization in Adh4 alcohol dehydrogenase mutant mice. Dev.
Genet. 25, 1–10.
13. Deltour, L., Foglio, M.H. & Duester, G. (1999) Metabolic defi-

ciencies in alcohol dehydrogenase Adh1, Adh3,andAdh4 null
mutant mice: overlapping roles of Adh1 and Adh4.ethanolclear-
ance and metabolism of retinol to retinoic acid. J. Biol. Chem. 274,
16796–16801.
14. Kamm, J.J. (1982) Toxicology, carcinogenicity, and teratogenicity
of some orally administered retinoids. J. Am. Acad. Dermatol. 6,
652–659.
15. Biesalski, H.K. (1989) Comparative assessment of the toxicology
of vitamin A and retinoids in man. Toxicology 57, 117–161.
16. Armstrong, R.B., Ashenfelter, K.O., Eckhoff, C., Levin, A.A. &
Shapiro, S.S. (1994) General and reproductive toxicology of
retinoids. In The Retinoids: Biology, Chemistry, and Medicine,2nd
edn. (Sporn, M.B., Roberts, A.B. & Goodman, D.S., eds), pp.
545–572. Raven Press, Ltd., New York.
17. Nelson, M. (1990) Vitamin A, liver consumption, and risk of birth
defects. Liver is a cheap source of many nutrients. Br.Med.J.301,
1176.
18. Frolik, C.A. (1984) Metabolism of retinoids. In The Retinoids,
Vol. 2 (Sporn, M.B., Roberts, A.B. & Goodman, D.S., eds), pp.
177–208. Academic Press, Orlando.
19. Litchfield, J.T. Jr, & Wilcoxon, F. (1949) A simplified method of
evaluating dose–effect experiments. J. Pharmacol. Exp. Ther. 96,
99–117.
20. Collins, M.D., Eckhoff, C., Chahoud, I., Bochert, G. & Nau, H.
(1992) 4-methylpyrazole partially ameliorated the teratogenicity of
retinol and reduced the metabolic formation of all-trans-retinoic
acid in the mouse. Arch. Toxicol. 66, 652–659.
21. Molotkov, A., Deltour, L., Foglio, M.H., Cuenca, A.E. &
Duester, G. (2002) Distinct retinoid metabolic functions for
alcohol dehydrogenase genes Adh1 and Adh4 in protection against

vitamin A toxicity or deficiency revealed in double null mutant
mice. J. Biol. Chem. 277, 13804–13811.
22. Zgombic-Knight, M., Ang, H.L., Foglio, M.H. & Duester, G.
(1995) Cloning of the mouse class IV alcohol dehydrogenase
(retinol dehydrogenase) cDNA and tissue-specific expression
patterns of the murine ADH gene family. J. Biol. Chem. 270,
10868–10877.
23. Haselbeck, R.J. & Duester, G. (1997) Regional restriction of
alcohol/retinol dehydrogenases along the mouse gastrointestinal
epithelium. Alcohol. Clin. Exp. Res. 21, 1484–1490.
24. Algar, E.M., Seeley, T L. & Holmes, R.S. (1983) Purification and
molecular properties of mouse alcohol dehydrogenase isozymes.
Eur. J. Biochem. 137, 139–147.
25. Danielsson, O. & Jo
¨
rnvall, H. (1992) ÔEnzymogenesisÕ: classical
liver alcohol dehydrogenase origin from the glutathione-depend-
ent formaldehyde dehydrogenase line. Proc.NatlAcad.Sci.USA
89, 9247–9251.
26. Can
˜
estro, C., Hjelmqvist, L., Albalat, R., Garcia-Ferna
`
ndez, J.,
Gonza
`
lez-Duarte,R.&Jo
¨
rnvall, H. (2000) Amphioxus alcohol
dehydrogenase is a class 3 form of single type and of structural

conservation but with unique developmental expression. Eur.
J. Biochem. 267, 6511–6518.
27. Danielsson, O., Eklund, H. & Jo
¨
rnvall, H. (1992) The major pis-
cine liver alcohol dehydrogenase has class-mixed properties in
relation to mammalian alcohol dehydrogenases of classes I and
III. Biochemistry 31, 3751–3759.
28. Hoffmann, I., Ang, H.L. & Duester, G. (1998) Alcohol dehy-
drogenases in Xenopus development: conserved expression of
ADH1 and ADH4 in epithelial retinoid target tissues. Dev. Dyn.
213, 261–270.
29. Boleda, M.D., Julia
`
,P.,Moreno,A.&Pare
´
s, X. (1989) Role of
extrahepatic alcohol dehydrogenase in rat ethanol metabolism.
Arch. Biochem. Biophys. 274, 74–81.
30. Roberts, E.S., Vaz, A.D.N. & Coon, M.J. (1992) Role of isozymes
of rabbit microsomal cytochrome P-450 in the metabolism of
retinoic acid, retinol, and retinal. Mol. Pharmacol. 41, 427–433.
31. Lieber, C.S. (1994) Mechanisms of ethanol-drug–nutrition inter-
actions. J. Toxicol. Clin. Toxicol. 32, 631–681.
32. Kono, H., Bradford, B.U., Yin, M., Sulik, K.K., Koop, D.R.,
Peters, J.M., Gonzalez, F.J., McDonald, T., Dikalova, A.,
Kadiiska, M.B., Mason, R.P. & Thurman, R.G. (1999) CYP2E1
is not involved in early alcohol-induced liver injury. Am.J.Physiol.
Gastrointest. Liver Physiol. 277, G1259–G1267.
33. Bray, B.J. & Rosengren, R.J. (2001) Retinol potentiates acet-

aminophen-induced hepatotoxicity in the mouse: Mechanistic
studies. Toxicol. Appl. Pharmacol. 173, 129–136.
2612 A. Molotkov et al. (Eur. J. Biochem. 269) Ó FEBS 2002