The Role of Retinoid X Receptor α in Regulating Alcohol Metabolism

In a previous study, we reported that a deficiency in the expression of a nuclear receptor, retinoid X receptor (RXR) α caused a reduction of S-adenosylmethionine (SAMe) and glutathione (GSH) levels and resulted in a more serious alcohol-induced liver injury (Dai et al., 2003). It has been reported that the level of RXRα is constitutively lower in peroxisome proliferator-activated receptor (PPAR) α-null mice and that PPARα-null mice are more susceptible to hepatotoxicity induced by alcohol (Tanaka et al., 2003; Nakajima et al., 2004). Thus, RXRα might play a protective role against liver injury induced by ethanol. However, little is known about the regulation of alcohol-metabolizing enzymes by RXRα.

Ethanol is metabolized in the liver by enzymes located in different subcellular compartments of the hepatocyte. Alcohol dehydrogenase (ADH or Adh), an NAD⁺-dependent enzyme located in the hepatic cytosol, catalyzes the conversion of ethanol to acetaldehyde, a potent toxicant that accounts for most of the toxic effects of ethanol (Niemela et al., 1995). Acetaldehyde produced from ethanol is further converted to the nontoxic acetate by a mitochondrial aldehyde dehydrogenase (ALDH or Aldh) (Sladek et al., 1989). Both ADH and ALDH exist as a family of isozymes in mammals (Cheung et al., 2003). In the mouse, ADH exists as four major classes, with ADH1, ADH2, and ADH3 predominantly localized in liver (Cheung et al., 2003). Of these isozymes, only ADH1 has a low K_m for ethanol and is responsible for hepatic ethanol clearance (Julia et al., 1987). ADH4, absent in normal liver, has a high K_m value for ethanol, but it does play a major role in extrahepatic ethanol metabolism, especially in the stomach (Farres et al., 1994). The most important enzymes for acetaldehyde oxidation are mitochondrial ALDH2 and cytosolic ALDH (Sladek et al., 1989; Cheung et al., 2003). ADH1, ADH3, and ADH4 also exhibit retinol dehydrogenase activity and participate in retinol (vitamin A) metabolism to form retinaldehyde. Catalytic efficiency studies indicate that ADH1 prefers all-trans-retinol as a substrate, whereas ADH4 has a high affinity for 9-cis-retinol (Yang et al., 1994; Allali-Hassani et al., 1998). Oxidation of retinaldehyde to retinoic acid is performed by the cytosolic aldehyde dehydrogenases ALDH1A1, ALDH1A2, and ALDH1A3 (Duester et al., 2003).

RXRα is a nuclear receptor for 9-cis-retinoic acid (RA), the metabolite of vitamin A (Mangelsdorf et al., 1992). RXRα also dimerizes with retinoic acid receptors (RARα, -β, and -γ), which are the receptors for all-trans-RA, the active metabolite of vitamin A. RA signaling pathways are involved in many biological processes, including cell growth, differentiation, morphogenesis, homeostasis, and metabolism (Mangelsdorf et al., 1993). Excessive vitamin A consumption decreases 5-methyltetrahydrofolate and SAMe levels, potentiating alcohol-induced toxicity (Fell and Steele, 1986; Lee et al., 2004). Moreover, ethanol intoxication during embryogenesis leads to fetal alcohol syndrome, characterized by craniofacial defects of the eyes, upper lips, and jaw due to improper development of the cranial neural crest (Deltour et al., 1996). Likewise, severe vitamin A deficiency during pregnancy results in eye defects and cleft face in embryos (Morriss-Kay and Sokolova, 1996). Thus, alcohol toxicity and vitamin A deficiency cause similar congenital abnormalities.

Besides ADH, another enzyme that affects ethanol action is cytochrome P450 2E1 (CYP2E1), which oxidizes both ethanol and vitamin A (Liu et al., 2001). Oxidative stress resulting in the depletion of cellular biomolecules such as GSH, vitamin E, and SAMe is considered one of the mechanisms responsible for alcohol-induced liver injury (Rouach et al., 1997; Lee et al., 2004). Previously, we have shown that reduced SAMe and GSH levels in RXRα KO mice potentiate alcohol-induced liver injury (Dai et al., 2003). To understand the role of the receptor for retinoic acid in alcohol metabolism and alcoholic liver disease, it is important to examine the effect of RXRα deficiency on alcohol-metabolizing enzymes.

Therefore, the current study focuses on the role of RXRα in the alcohol detoxification pathway assessed by metabolic activities, immunoblot, and semiquantitative reverse-transcriptase-polymerase chain reaction (RT-PCR). We also compared ethanol and acetaldehyde metabolic and elimination rates in wild-type and RXRα KO mice. Our results provide evidence that RXRα deficiency leads to increased ADH activity and decreased ALDH2 and cytosolic ALDH activities. These changes resulted in a difference between wild-type and RXRα KO mice with regard to ethanol and acetaldehyde clearance from the blood and liver. The accumulation of acetaldehyde after ethanol ingestion in RXRα KO mice may in part explain why hepatocyte RXRα KO mice are more susceptible to ethanol-induced liver damage than wild-type mice.

Materials and Methods

Animals. Mice carrying the RXRα mutation in hepatocytes have been described previously (Wan et al., 2000). The animals used in all the experiments were age-matched male wild-type and hepatocyte RXRα KO mice. The mice were housed at 22°C with a 12/12-h light/dark cycle and provided food and water ad libitum. All procedures were conducted in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and were approved by the Kansas University Medical Center Institutional Animal Care and Use Committee.

Preparation of Mitochondria, Cytosol, and Microsomes. Mice were sacrificed by decapitation. Cytosolic, mitochondrial, and microsomal fractions were separated from fresh liver tissue. In brief, the liver was homogenized in 0.01 M potassium phosphate buffer, pH, 7.0, containing 0.25 M sucrose and 0.3 mM EDTA. The crude homogenate was centrifuged at 600g for 10 min at 4°C, and the supernatant was centrifuged further at 5000g for 10 min at 4°C to give a mitochondrial pellet and a supernatant containing the cytosolic and microsomal fractions. The mitochondrial pellet was washed twice, kept at -80°C, and used for ALDH enzyme activity as well as Western blot analysis. The supernatant was centrifuged at 105,000g for 60 min at 4°C to obtain the cytosolic and microsomal fractions. The pellet (microsomes) was further homogenized in 0.15 M Tris-HCl buffer, pH, 8.0, and centrifuged at 105,000g for 60 min at 4°C to give one-washed liver microsomes for CYP2E1 enzyme activity assay. The cytosol and microsomes were immediately frozen and kept at -80°C until use. The supernatant (cytosol) was used for ADH, ALDH, and glutathione S-transferase (GST) enzyme activity and immunoblot assay.

Measurements of ADH, ALDH, GST, and CYP2E1 Enzyme Activities. ADH isozyme activities in the liver cytosol were measured by detecting the reduction of NAD⁺ at 340 nm. ADH1 activity was measured with 0.3 mM NAD⁺ and 10 mM ethanol in 0.1 M glycine-NaOH buffer, pH 10.5, as described previously (Martras et al., 2004). The ADH2 activity was determined with 5 mM benzyl alcohol and 2.4 mM NAD⁺ in 0.1 M glycine-NaOH buffer, pH, 10, as reported previously (Stromberg et al., 2004). ADH3, identical to glutathione-dependent formaldehyde dehydrogenase, is the only ADH that exhibits high specificity for S-hydroxymethylglutathione (HMGSH) (Holmquist and Vallee, 1991). The oxidation of HMGSH, the actual substrate formed spontaneously from GSH and formaldehyde, was therefore measured as an index of ADH3 activity. The assay mixture contained 1 mM formaldehyde, 1 mM GSH, 2.4 mM NAD⁺, and cytosol in 0.1 M sodium phosphate buffer, pH 8.0, as described previously (Hedberg et al., 2000). During assay, GSH and formaldehyde were allowed to equilibrate for 2 min in 0.1 M sodium phosphate buffer, pH 8.0, at 25°C before the addition of NAD⁺ and cytosol. The formaldehyde oxidation rate of each sample was used as background and subtracted from that obtained for HMGSH. The mitochondrial and cytosolic ALDH activity was measured by detecting the reduction of NAD⁺ at 340 nm as described previously with a minor modification (Tottmar et al., 1973; Lindahl and Evces, 1984). The reaction mixture contained 50 mM sodium pyrophosphate buffer, pH 8.8, 0.5 mM NAD⁺, 0.1 mM pyrazole, and 100 μMor5mM acetaldehyde or propionaldehyde. When assaying mitochondrial fractions, the reaction mixture also contained 2 μM rotenone. Total GST activity in the cytosol and mitochondria was measured using 1-chloro-2,4-dinitrobenzene as described previously (Habig et al., 1974). CYP2E1 activity in the liver microsomes was estimated colorimetrically by measuring the hydroxylation of p-nitrophenol to 4-nitrocathecol (Reinke and Moyer, 1985).

Western Blot. Cytosolic and mitochondrial proteins (10 μg) were separated by 10% SDS-PAGE gels, electroblotted onto a polyvinylidene difluoride membrane, and immunoblotted with anti-mouse ADH1 antibody (kindly supplied by Dr. Michael Felder, Department of Biological Sciences, University of South Carolina, Columbia, SC), ALDH2 antibody (kindly supplied by Dr. Henry Weiner, Department of Biochemistry, Purdue University, West Lafayette, IN), ALDH1A1 antibody (EMD Biosciences, San Diego, CA), and ALDH1A2 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Blots were blocked at 4°C with Tris-buffered saline with 0.1% Tween 20 (TBST) plus 5% dry nonfat milk, and antibodies were diluted in TBST plus 1% dry nonfat milk. Blots were incubated in primary antibodies overnight at 4°C and washed three times in TBST. Blots were then incubated with the appropriate peroxidase-conjugated anti-rabbit (Bio-Rad, Hercules, CA) or donkey anti-goat IgG (Santa Cruz Biotechnology, Inc.) secondary antibodies diluted in TBST plus 1% milk for 1 h at room temperature. Following probing, blots were stripped and reprobed for β-actin with monoclonal anti-β-actin (Sigma-Aldrich, St. Louis, MO). Proteins were viewed using enhanced chemiluminescence (Pierce Chemical, Rockford, IL). Protein contents in the mitochondria, cytosol, and microsomes were determined by the Bradford method (Bradford, 1976).

RNA Isolation and Semiquantitative RT-PCR. Male wild-type and hepatocyte RXRα KO mice (10 weeks old; n = 5) were used. Total RNA was isolated from frozen liver tissues by using TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol. RNA concentration and quality were determined spectrophotometrically at 260 nm and by the A₂₆₀/A₂₈₀ ratio, respectively. RNA integrity was assessed by comparing the relative intensities of the 28S and 18S rRNA bands by electrophoresis in an agarose 1.2% (w/v) gel containing 2.2 M formaldehyde. Quantification of cDNA-specific mRNA for individual wild-type and RXRα KO mouse livers was accomplished by semiquantitative RT-PCR. In brief, total RNA (5 μg) was reverse transcribed with SuperScript II first-strand synthesis (Invitrogen). cDNA (2 μl) was amplified in a total reaction volume of 50 μl in a PTC-200 Peltier thermal cycler (MJ Research, Watertown, MA) in the presence of specific primers to Adh1, Adh2, Adh3, Adh4, Aldh1a1, and Aldh2. Primer pairs (Table 1) were designed using Primer 3 software (http://frodo.wi.mit.edu). In addition, primers specific to β-actin were included in each reaction to serve as an internal control. The ratio of specific primers to β-actin primer was 5:1. The PCR procedure consisted of an initial denaturation at 95°C for 5 min, followed by 15 to 35 cycles of denaturation at 95°C for 1 min, annealing at 55°C for 2 min, extension at 72°C for 2 min, and a final extension at 72°C for 10 min. To ensure that the amplification remained within the linear range, PCR with each set of primers was performed at three to five different cycles. The number of cycles found to be optimal for each gene was as follows: Adh1, 25; Adh2, 30; Adh3, 30; Adh4, 30; Aldh1a1, 25; and Aldh2, 28.

Isolation of Polysomes and Northern Analysis. Polysomal distribution of Adh1 mRNA in wild-type and RXRα KO mouse livers was performed as described previously with some modification (Taylor and Schimke, 1973). After overnight fasting, mice were sacrificed and the liver was immediately excised. The liver (1 g) was homogenized in 5 volumes of polysome buffer (25 mM Tris-HCl buffer, pH 7.4, 10 mM MgCl₂, 25 mM NaCl, and 0.14 M sucrose) containing 100 μg/ml heparin. The homogenate was centrifuged at 14,600g for 10 min at 4°C, and the pellet was discarded. After addition of 1 volume of a solution of 10% Triton X-100 and 10% sodium deoxycholate to 9 volumes of the supernatant, the resultant solution was rehomogenized. A 1.0 ml of this mixture was layered over a 10 to 60% sucrose gradient made in 25 mM Tris-HCl, pH 7.4, 10 mM MgCl₂, and 25 mM NaCl. Ultracentrifugation was carried out at 4°C and 288,000g for 105 min in a Beckman SW41 Ti rotor (Beckman Coulter, Fullerton, CA). After centrifugation, the gradients were fractionated into 10 fractions from the bottom (1 ml/fraction). The absorbance of the gradients at 254 nm was determined for each fraction using a Cary 50 Bio spectrophotometer (Varian Analytical Instruments, Walnut Creek, CA). The RNA in each fraction was extracted using the TRIzol reagent. Because heparin, which potently inhibits PCR, was present in the polysomal gradients, Northern analysis, but not RT-PCR, was used to quantify the Adh1 mRNA in each fraction. An equal volume of total RNA extracted from polysomal fractions was separated by electrophoresis in an agarose 1.2% (w/v) gel containing 2.2 M formaldehyde. RNA integrity was assessed by comparing the relative intensities of the 28S and 18S rRNA bands. RNA was transferred to a nylon membrane by capillary blotting in 10× standard saline citrate and cross-linked by UV irradiation. The cDNA fragment specific to mouse Adh1 was synthesized by RT-PCR and used as a probe. Probe labeling and hybridization were performed as described previously (Wan et al., 2003).

Determination of Alcohol and Acetaldehyde Levels in Blood and Liver. Male wild-type and hepatocyte RXRα KO mice (10 weeks old) were used. Ethanol was administered as a single dose (3 g/kg) at a concentration of 35% (v/v) by gastric intubation after an overnight starvation. Mice (n = 5) were sacrificed 1, 2, 3, 4, and 6 h after ethanol administration. Blood (collected into heparinized tubes) was either used for blood alcohol level determination by the Analox GL5 alcohol analyzer (Analox Instrument USA Inc., Luneburg, MA) (Ronis et al., 2004) or blood acetaldehyde level determination. For blood alcohol determination, heparinized blood was deproteinized with 6% perchloric acid. The sample was then thoroughly mixed with 1.5 M K₂CO₃ and centrifuged at 2700g for 3 min. The supernatant was collected and used for the determination of ethanol concentration. Blood and liver tissue were placed in sealed vials and stored at -80°C for gas chromatography analysis of acetaldehyde levels. Immediately before analysis, the samples, 100 μl of blood or 100 mg of liver tissue were warmed at 50°C for 40 min. Headspace gas (100 μl) from each vial was removed through a septum on the cap with a tight syringe and transferred to the injection system on a Varian 3400 gas chromatography equipped with flame ionization detector. Acetaldehyde was separated on a 30-m capillary column (SP 2330; Supelco, Bellfonte, PA) with an inner diameter of 0.25 mm. The temperature was held isothermally at 50°C, and helium was used as the carrier gas at a flow rate of 6 ml/min. Under the specified conditions, separation of acetaldehyde from other compounds was complete within 2.5 min with a retention time of 1.9 min. Individual peaks were integrated with a programmable Star Chromatography Work Station, version 6.0 (Varian Analytical Instruments). Quantification of acetaldehyde was achieved by calibrating the gas chromatography peak area against those from headspace samples of authentic acetaldehyde standards.

Statistical Analysis. Data are presented as means ± S.D. Statistical analysis was performed by the Student's t test. P < 0.05 was considered statistically significant.

Results

ADH1 enzyme activity was significantly increased in the livers of RXRα KO mice compared with wild-type mice (Table 2). Among the ADH isozymes, ADH1 has been suggested to play a major role in hepatic ethanol clearance due to its low K_m for ethanol (Julia et al., 1987). Therefore, we examined the expression of the ADH1 protein in liver cytosol of wild-type and RXRα KO mice. In agreement with the increased ADH1 activity (1.2-fold), ADH1 protein level was higher (2.5-fold) in RXRα KO mice than wild-type mice (Fig. 1a). ADH2 and ADH3 enzyme activities were not different between RXRα KO mice and wild-type mice (Table 2). Using semiquantitative RT-PCR, the levels of Adh1, Adh2, Adh3, and Adh4 mRNAs in the livers were not different between the mutant and wild-type mice (Fig. 1b). Table 1 shows the size of the products obtained after RT-PCR. All the amplifications were within the linear range, and the data were reproduced more than three times. The low expression of the Adh4 mRNA observed in the liver (Fig. 1b) agrees with previous reports showing the expression of ADH4 is extrahepatic and is very low in normal liver (Farres et al., 1994).

Since Adh1 mRNA level remained unchanged, we hypothesized that increased ADH1 protein synthesis might account for the elevated ADH1 protein level in the mutant mice. To investigate this, we isolated polysomes from wild-type and RXRα KO mice. The absorbance profile of fractions collected after sucrose gradient was used to designate the fractions into polysomes and monosomes (Fig. 2a). Northern analysis revealed that fractions 2 to 5, obtained from the RXRα KO mice, which represented the actively translated polysomal fractions (polysomes), were found to contain higher levels of Adh1 mRNA (2.8-fold) compared with the same fractions isolated from wild-type mice (Fig. 2, b and c). This finding indicates increased translation of Adh1 protein in the mutant mice. We observed that although the -fold changes at the mRNA level (2.8-fold) (Fig. 2, b and c) and protein level (2.5-fold) (Fig. 1a) were similar, the -fold change in ADH1 activity was only 1.2 (Table 2).

Two different enzymes may participate in the oxidation of acetaldehyde generated from ethanol: mitochondrial ALDH2 and cytosolic ALDH1 (Klyosov et al., 1996). To provide specific information about the low and high K_m isoforms of these enzymes, ALDH activities were measured at two different concentrations (100 μM and 5 mM) of acetaldehyde and propionaldehyde. ALDH activity in the liver cytosol of wild-type and RXRα KO mice is shown in Table 3. The cytosolic ALDH activity using 100 μM or 5 mM acetaldehyde as substrate was significantly decreased in the RXRα KO mice to 50 and 67.4%, respectively, of the levels in wild-type mice (Table 3). However, although there was no difference in cytosolic ALDH activity when 100 μM propionaldehyde was used as substrate between the mutant and the wild-type mice, ALDH activity in the RXRα KO mice decreased to 68.1% of wild-type mice when 5 mM propionaldehyde was used (Table 3). Mitochondrial ALDH2 activity was also significantly lower in RXRα KO mice than wild-type mice (Table 3). Our results revealed that differences between the mutant and wild-type mice were more apparent in the cytosolic fraction when 100 μM acetaldehyde was the substrate. This suggests that the cytosolic low K_m ALDH1 with high acetaldehyde-oxidizing capacity is more affected by RXRα deficiency than other ALDH isoforms in the mutant mice. By Western analysis, ALDH1A1 and ALDH1A2 proteins were decreased by 70 and 60%, respectively, in RXRα KO mice compared with the wild-type mice (Fig. 3, a and b). However, ALDH2 protein level was not changed (Fig. 3c). At the mRNA level, the level of Aldh1a1 mRNA was significantly less in RXRα KO than wild-type mice (Fig. 4a). However, Aldh2 mRNA level was not changed (Fig. 4b). In the RXRα KO mice, cytosol and mitochondrial GST activities were decreased to 47.9 and 66.7%, respectively, of that of the wild-type mice (Table 4).

To study the impact of changes in ADH and ALDH activities in RXRα KO mice, ethanol and acetaldehyde elimination rates were determined after a single dose of ethanol by intragastric administration. The concentrations of ethanol and acetaldehyde in the blood and liver of mice were measured at different times. RXRα KO mice exhibited enhanced blood ethanol clearance compared with wild-type mice (Fig. 5). The blood ethanol levels in the RXRα KO mice were significantly lower at 1, 2, and 4 h after ethanol administration compared with wild-type mice. However, a comparison of blood acetaldehyde clearance between mutant and wild-type mice demonstrated that RXRα KO mice had a defect in the clearing of acetaldehyde (Fig. 6a). Likewise, liver acetaldehyde clearance was slower in RXRα KO mice than wild-type mice (Fig. 6b).

Discussion

In this study, we assessed the levels of expression of ADH, ALDH, CYP2E1, and GST enzymes in wild-type and RXRα KO mice. The data demonstrated that RXRα deficiency results in an induction of ADH activity through translational regulation of the ADH protein. In addition, ALDH, CYP2E1, and GST activities are reduced in the mutant mice. Furthermore, we demonstrated that RXRα KO mice have elevated blood acetaldehyde levels, which remained high for 3 h after ethanol administration. In contrast, acetaldehyde levels peaked at 1 h in the wild-type mice and were low 3 h after ethanol administration.

The toxicity of ethanol is mediated by acetaldehyde, a highly reactive molecule produced from ethanol by the cytosolic enzyme ADH (Niemela et al., 1995). Under normal circumstances, acetaldehyde is rapidly formed in the liver, but it is also immediately oxidized by ALDH to acetate. Hence, little or no acetaldehyde enters the blood. Thus, the high acetaldehyde levels we observed in RXRα KO mice, most probably resulted from the increased rate of ethanol metabolism in the liver of these mutant mice by ADH. Indeed, our results revealed that both hepatic ADH enzyme activity and ADH1 protein were higher in RXRα KO mice compared with wild-type mice. Our findings provide in vivo evidence that the level of ADH1 expression regulated by RXRα is an important factor in determining the rate of ethanol elimination and influences the risk for alcohol-induced liver injury (Bosron et al., 1993). Furthermore, the cyclic pattern of blood and urine ethanol levels has been suggested to involve the induction of ADH1 by ethanol to facilitate ethanol elimination (Badger et al., 2000).

In addition to ADH, ethanol is known to be metabolized by CYP2E1 to produce acetaldehyde (Guengerich et al., 1991). CYP2E1 activity could affect the rate of acetaldehyde formation. Our previous study revealed that the basal level of CYP2E1 mRNA was not different between RXRα KO mice and wild-type mice (Cai et al., 2003). However, our current study shows that CYP2E1 activity was significantly decreased in RXRα KO mice. Hence, the observed increase in ethanol clearance in RXRα KO mice does not seem to be related to CYP2E1 activity.

It was also observed that hepatic RXRα deficiency did not change Adh1, Adh2, Adh3, and Adh4 mRNA levels. This finding suggests that, the increase in ADH activity in the RXRα KO mice was due to an increase in ADH1 protein without affecting Adh1 gene transcription. In a recent report, leptin was found to increase both ADH activity and ADH protein without affecting Adh mRNA (Mezey et al., 2005). Although leptin did not affect ADH protein degradation, it enhanced ADH protein synthesis through translational initiation factors such as eukaryotic initiation factor (eIF) 2α, eIF2B activity, and the eIF4E-eIF4G complex formation (Mezey et al., 2005). A post-transcriptional mechanism is likely to be involved in the increase in ADH activity found in the C57BL/6J mice (Tussey and Felder, 1989). Leptin level is also increased in RXRα KO mice (Wan et al., 2003). Therefore, it is reasonable to hypothesize that the observed increase in ADH activity and ADH1 protein in RXRα KO mice might involve increased protein synthesis due to serum leptin induction in these mutant mice. Indeed, using polysomal distribution analysis, our results indicate that RXRα deficiency increased Adh1 mRNA association with polysomes, suggesting increased ADH1 protein synthesis in the mutant relative to wild-type mice.

In the human liver, mitochondrial (low K_m) ALDH2 oxidizes most of the ethanol-derived acetaldehyde; however, in rodents both mitochondrial and cytosolic ALDH isozymes are important in acetaldehyde oxidation (Klyosov et al., 1996). The hepatic ALDH activity in alcoholics and rodents has been suggested as rate-limiting for hepatic oxidation of acetaldehyde (Enomoto et al., 1991). Our results showed elevated blood and hepatic acetaldehyde levels in RXRα KO compared with wild-type mice. Furthermore, both mitochondrial and cytosolic ALDH activities are significantly lower in RXRα KO compared with wild-type mice, which may result in accumulation of acetaldehyde following ethanol ingestion by the RXRα KO mice. Our results revealed a marked decrease in cytosolic ALDH activity compared with mitochondrial ALDH2 in the mutant mice, confirming an earlier observation that mouse cytosolic ALDH1 has a high affinity for acetaldehyde and is an important mediator of alcohol and acetaldehyde metabolism (Smolen et al., 1981; Algar and Holmes, 1986). Consistent with the decrease in the cytosolic ALDH activity and protein levels, Aldh1a1 mRNA was decreased in RXRα KO mice more than wild-type mice. Hence, the reduction of ALDH expression in RXRα KO mice might be regulated at the level of transcription. However, ALDH2 expression was not changed in the mutant mice, suggesting that the decrease in ALDH2 activity without a change in protein and mRNA level might be due to modification of the ALDH2 molecule, which leads to a decrease in substrate binding or catalysis. Therefore, it is persuasive to suggest that lower ALDH activity together with higher ADH1 activity in livers of RXRα KO mice may at least in part be responsible for the higher blood and liver acetaldehyde levels observed in these mutant mice after ethanol administration. There is substantial overlap in retinol and alcohol metabolism; therefore, these changes in ADH1 and ALDH enzyme activities will not only affect retinoid metabolism but also contribute to alcohol toxicity in the mutant mice. Since ALDH1A1 and ALDH1A2 have retinaldehyde dehydrogenase activity and are involved in the synthesis of retinoic acid, it is of interest to examine the contribution of accumulated levels of retinaldehyde on alcohol toxicity in RXRα KO mice.

The human ALDH2 promoter contains a retinoid response element with the potential to bind retinoid acid receptors (Pinaire et al., 2003). It has therefore been suggested that retinoic acid, a metabolite of vitamin A, is important for ALDH2 expression (Crabb et al., 2001). We observed that cytosolic ALDH activity, ALDH1A1, and ALDH1A2 proteins, and Aldh1a1 mRNA were decreased in RXRα KO mice. The effect of RXRα deficiency on ALDH1A proteins as well as Aldh1a1 mRNA suggests a diverse but a selective effect of RXRα on the ALDH gene. Although the interaction of the ALDH2 promoter with PPAR has also been suggested, the exact regulation of the ALDH2 gene by nuclear receptors has not been fully elucidated (Crabb et al., 2001). The molecular regulation of the Aldh1a1 gene expression by RXRα is important and needs further investigation.

Both mitochondrial and cytosolic GSTs were decreased in RXRα KO mice, suggesting that conjugation of toxic metabolites with GSH via GSTs is compromised. In addition, we have previously shown that GSH levels are reduced in RXRα KO mice (Dai et al., 2003). Hence, RXRα deficiency leads to impairment in the detoxification of ethanol through alteration of phase I (ADH and ALDH) as well as the toxic products of ethanol metabolism by phase II (GST) enzyme activities (Fig. 7). Among the nuclear receptors, RXRα, which is highly expressed in the liver, is unique in its ability to heterodimerize with other nuclear receptors, thereby controlling many metabolic processes in the liver (Wan et al., 2000). We believe the diverse effect we observed with the RXRα deletion in the liver might be due to the promiscuous nature of RXRα and that some of these effects might be direct or indirect.

Together, the results presented reveal differences in wild-type and RXRα KO mice with regard to ethanol and acetaldehyde clearance in the blood and liver. Our results strongly suggest a central role of RXRα in alcohol metabolism.