Ethanol elevates physiological all-trans-retinoic acid levels in select loci through altering retinoid metabolism in multiple loci: a potential mechanism of ethanol toxicity

Abstract

All-trans-retinoic acid (atRA) supports embryonic development, central nervous system function, and the immune response. atRA initiates neurogenesis and dendritic growth in the hippocampus and is required for spatial memory; superphysiological atRA inhibits neurogenesis, causes teratology and/or embryo toxicity, and alters cognitive function and behavior. Because abnormal atRA shares pathological conditions with alcoholism, inhibition of retinol (vitamin A) activation into atRA has been credited widely as a mechanism of ethanol toxicity. Here, we analyze the effects of ethanol on retinoid concentrations in vivo during normal vitamin A nutriture, using sensitive and analytically robust assays. Ethanol either increased or had no effect on atRA, regardless of changes in retinol and retinyl esters. Acute ethanol (3.5 g/kg) increased atRA in adult hippocampus (1.6-fold), liver (2.4-fold), and testis (1.5-fold). Feeding dams a liquid diet with 6.5% ethanol from embryonic day 13 (e13) to e19 increased atRA in fetal hippocampus (up to 20-fold) and cortex (up to 50-fold), depending on blood alcohol content. One-month feeding of the 6.5% ethanol diet increased atRA in adult hippocampus (20-fold), cortex (2-fold), testis (2-fold), and serum (10-fold). Tissue-specific increases in retinoid dehydrogenase mRNAs and activities, extrahepatic retinol concentrations, and atRA catabolism combined to produce site-specific effects. Because a sustained increase in atRA has deleterious effects on the central nervous system and embryo development, these data suggest that superphysiological atRA contributes to ethanol pathological conditions, including cognitive dysfunction and fetal alcohol syndrome.—Kane, M. A., Folias, A. E., Wang, C., Napoli, J. L. Ethanol elevates physiological all-trans-retinoic acid levels in select loci through altering retinoid metabolism in multiple loci: a potential mechanism of ethanol toxicity.

Nearly 20 million alcoholics in the United States experience detrimental health effects from ethanol consumption, including liver damage, cognitive dysfunction, various types of cancer, and harm to offspring through fetal alcohol syndrome (FAS) (1,2,3,4). Ethanol disrupts physiological homeostasis and produces a toxic metabolite, acetaldehyde, which contributes to cell and tissue damage (5). Ethanol ingestion depletes liver retinyl esters (REs), the storage form of vitamin A (all-trans-retinol), but little is known about its effect on endogenous all-trans-retinoic acid (atRA), a biologically active form of vitamin A that regulates transcription and translation via nuclear receptors (6,7,8). Vertebrates require atRA to support many biological functions, including development, differentiation, reproduction, nervous system function, and immune response (9,10,11,12,13). The deleterious effects of ethanol share many similarities with phenotypic manifestations of abnormal atRA levels, including developmental defects, cognitive dysfunction, and increased risk of cancer (2, 3, 12, 14,15,16). Thus, disruption of atRA homeostasis and signaling may contribute to ethanol-induced defects (17). Although the interaction between vitamin A and ethanol has been investigated (14,15,16,17), the effect of ethanol on physiological concentrations of atRA remains uncertain.

atRA homeostasis and function occurs via interactions of retinoid-binding proteins with specific enzymes and receptors (18, 19). Retinol-binding protein (RBP) transports retinol through serum. Stimulated by retinoic acid (RA) gene 6 (Stra6), a plasma membrane RBP receptor, mediates cellular retinol uptake (20). Cellular retinol-binding protein type I (CrbpI) sequesters intracellular retinol in a high-affinity complex and delivers it for esterification by lecithin:retinol acyltransferase, or reversible dehydrogenation into all-trans-retinal (retinal), catalyzed by retinol dehydrogenases (Rdh) of the short-chain alcohol dehydrogenase/reductase (SDR) gene family. Irreversible dehydrogenation of retinal by retinal dehydrogenases (Raldh) produces atRA (18). atRA regulates transcription and translation, through nuclear RA receptors (RARs) α, β, and γ, and peroxisome proliferator-activated receptors β and δ (21, 22). Ethanol exposure could alter atRA through a variety of mechanisms, including increasing substrate concentration as a consequence of mobilizing hepatic retinol, altering enzyme activity, changing retinoid-binding protein expression, and/or disrupting RA receptors (23, 24).

Analytical limitations have prevented quantification of endogenous atRA during normal vitamin A status after ethanol exposure. In the absence of sensitive analytical techniques, approaches to this issue have relied on indirect assessment of atRA, testing ethanol effects after pharmacological doses of retinol, and/or in vitro models that did not necessarily reflect physiological retinoid metabolism, extrapolating to in vivo conclusions (17, 25, 26). These approaches have engendered the prevalent notion that ethanol inhibits activation of retinol into atRA, thereby decreasing atRA concentrations. Understanding ethanol-induced disregulation of retinoid metabolism and atRA concentrations in vivo during normal vitamin A homeostasis may provide new insights into FAS, alcohol-induced cognitive dysfunction, and cancer. We developed a sensitive assay for atRA and its isomers in limited biological samples and applied it to quantify ethanol effects on endogenous atRA during normal vitamin A nutriture (27, 28). We found that ethanol does not affect atRA at multiple sites, but increases atRA in specific loci through cumulative effects on multiple aspects of retinol metabolism. The hippocampus and testis, which depend on normal atRA for development and function, were most susceptible to ethanol-induced increases in atRA (10, 29,30,31,32).

MATERIALS AND METHODS

Chemicals

Liquid chromatography (LC)-mass spectrometry (MS) Chromasolv grade acetonitrile, water, methanol, and retinoids were obtained from Sigma-Aldrich (St. Louis, MO, USA), except for 9,13-di-cis-RA, which was prepared as described previously (33). 4,4-Dimethyl RA was a gift from Marcia Dawson (Burnham Institute, La Jolla, CA, USA) and Peter Hobbs (SRI International, Menlo Park, CA, USA). Retinoid standards were prepared the day of use. Concentrations were verified spectrophotometrically (34).

Retinoid analyses

RA was quantified by LC-tandem mass spectrometry (MS/MS) with atmospheric pressure chemical ionization (27, 28). Retinol, RE, and retinal were quantified by HPLC/UV (35). Retinoids were handled under yellow lights using only glass/stainless steel containers, pipettes, and syringes.

Animal studies

Animal studies were conducted according to institutional guidelines. USP/NF-grade ethanol (190 proof) was used for animal studies (Sigma-Aldrich). Male C57BL/6 mice (3 ± 0.5 mo old; Charles River Laboratories, Inc., Wilmington, MA, USA) were fed a stock diet (Teklad Global 18% Protein Rodent Diet, 15.4 IU of vitamin A acetate/g, 4.65 mg/kg retinol; Harlan, Indianapolis, IN) unless indicated otherwise. Serum and tissue samples were harvested under yellow light and frozen in liquid N₂. Blood was collected by cardiac puncture, allowed to clot for 30 min, and centrifuged at 10,000 g for ∼10 min at 4°C to isolate serum. Blood alcohol content percentage (BAC%) was assayed using a NAD⁺-alcohol dehydrogenase (Adh) enzymatic assay kit (Sigma-Aldrich). Acute (single) doses were administered i.p. as a saline/ethanol (1:1) solution at 1.5, 2.5, or 3.5 g/kg ethanol. Chronically dosed mice were fed either an AIN-93M (4 IU/g vitamin A) purified liquid diet containing 6.5% ethanol (Dyet 710302; Dyets Inc., Bethlehem, PA, USA) or an AIN-93M purified liquid control diet (Dyet 710080) with an isocaloric amount of maltose dextrin (Dyet 402851) for 1 mo. Chronically dosed mice were acclimated to liquid diets and ethanol by feeding 1 wk each of 0, 3.5, and 5% ethanol and/or with an isocaloric amount of maltose dextrin (relative to 6.5% ethanol). All mice had access to water during the acclimation period. No additional water was provided to either group during the 1-mo feeding. Chronic ethanol-treated mice were 3–3.5 mo old at time of assay. Average BAC% for mice fed the 6.5% ethanol liquid diet was 0.14% (Supplemental Fig. 1). Embryos were exposed to ethanol by feeding 2- to 3-mo-old C57BL/6 dams an AIN-93M purified liquid diet with either 6.5% ethanol or an isocaloric amount of maltose dextrin starting on embryonic day 13 (e13). Embryos were harvested on e19. Embryo brains were dissected using a Nikon SMZ-10A dissection microscope (Nikon, Tokyo, Japan) equipped with a NCL 150 light source with a yellow filter (Volpi, Auburn, NY, USA). 4-Methylpyrazole (119 mg/kg) and carbenoxolone (80 mg/kg) were injected i.p. as saline-DMSO (1:1) solutions 30 min before 3.5 g/kg ethanol. Mice were assayed 2 h after ethanol dosing.

Astrocyte cell culture and in vitro activity assay

Hippocampus astrocytes were prepared from 2-d-old Sprague-Dawley rat pups. Pups were sacrificed in a CO₂ chamber. Hippocampi were dissected, and meninges were removed. Hippocampi were resuspended in 2 ml of Hanks’ balanced salt solution with 10 mM HEPES buffer (pH 8.0) and 1000 U/ml penicillin-streptomycin. Individual astrocytes were mechanically dissociated by trituration though a series of Pasteur pipettes with gradually reduced diameters. Cells were plated in DMEM supplemented with 10% FBS in 175-cm² tissue culture flasks, and incubated in a 37°C incubator with 5% CO₂. The medium was changed after 24 h. Cells were allowed to grow another 14 d until confluent. Confluent flasks were sealed and shaken at 300 rpm for 6 h to separate oligodendrocytes and microglia cells from astrocytes. Astrocytes were recultured in six-well plates and then were incubated for 4 h in the dark with either 0.5 or 2.5 μM all-trans-retinol. The medium was aspirated, cells were lysed on the plate with reporter lysis buffer (Promega, Madison, WI), and atRA was extracted and quantified.

Subcellular fractionation

Tissue homogenates (∼25%) were prepared at 1240 rpm (RZR1 centrifuge; Heidolph Brinkmann, Elk Grove Village, IL, USA) in 10 mM Tris-HCl, 10% sucrose (w/w), 1 mM EDTA, and 1.5 mM DTT (pH 7.4). Microsomal and cytosolic fractions were isolated by centrifugation at 4°C: 1000 g for 10 min, 10,000 g for 15 min, 17,000 g for 15 min, and 100,000 g for 1 h. The 100,000 g supernatant (cytosol) was removed and concentrated for 15–30 min at 5000 g (10,000 MWCO centrifugal concentrators; Millipore Corporation, Billerica, MA, USA). The 100,000 g pellet (microsomes) was resuspended and hand-homogenized (Duall 22; Kontes Glass, Vineland, NJ, USA) in homogenization buffer. Protein content was determined using the Bradford assay.

Microsomal activity

Rdh activity was assayed at 37°C under initial velocity conditions in the linear ranges of time and protein in 50 mM HEPES, 150 mM KCl, 1 mM EDTA, and 2 mM DTT (pH 8.0) in the presence of 4 mM NAD⁺ and 2 mM NADP⁺ with a total reaction volume of 250 μl and 65 rpm shaking. Liver microsomes (140 μg/protein) were reacted with 0.4 μM retinol or 0.4 μM CrbpI-bound retinol for 15 min. Hippocampus microsomes (25 μg of protein) were reacted with 0.5 μM retinol or 0.5 μM CrbpI-bound retinol for 20 min. Reactions were initiated by adding retinol in 5 μl of DMSO or 5 μl of CrbpI-bound retinol in buffer.

atRA elimination half-life

Rates of atRA catabolism (60 pmol) by microsomal protein (250 μg) were determined in 50 mM Tris-HCl, 150 mM KCl, and 5 mM MgCl₂ (pH 7.4) in the presence of an NADPH-regenerating system in a total volume of 0.5 ml. Assays were done in triplicate at 37°C with 65 rpm shaking. The NADPH-regenerating system was added in 30 μl as a 1:1:1 solution of glucose-6-phosphate dehydrogenase (2.5 U in 5 mM sodium citrate, pH 7.5), 50 mM glucose 6-phosphate, and 50 mM NADP⁺ in assay buffer. atRA was added last in 5 μl of DMSO. Samples were extracted and atRA was quantified with a resuspension volume of 2 ml of acetonitrile and an injection volume of 5 μl.

Raldh isolation

Raldh isoforms in cytosol were separated by fast-protein LC with a Mono Q 5/50 GL column (Amersham/GE Healthcare, Little Chalfont, Buckinghamshire, UK) eluted at 0.5 ml/min for 10 min with 20 mM HEPES and 2 mM DTT (pH 7.5), followed by a gradient of 0 to 300 mM NaCl from 10 to 50 min. One-min fractions were collected and assayed individually and/or were pooled to determine total activity of Raldh isoforms. Pooled fractions were concentrated with a centrifugal concentrator (10,000 MWCO; Millipore). Raldh1 protein was measured by Western blot with polyclonal Raldh1 Ab (1:200; Abcam, Cambridge, MA, USA). Signal intensity was quantified with densitometry.

Raldh activity assay

Raldh activity was assayed under initial velocity conditions within the linear range of time and protein at 37°C in 50 mM HEPES, 150 mM KCl, 1 mM EDTA, 2 mM DTT (pH 8.0), 4 mM NAD⁺, and 2 mM NADP⁺ in a total reaction volume of 250 μl with 10 μg of cytosolic protein and 1 μM retinal for 15 min, with 65 rpm shaking. Retinal was added last in 5 μl of DMSO.

Gene expression

Total RNA was isolated with TRIzol (Invitrogen, Carlsbad, CA, USA). RNA (1 μg) was reverse transcribed with SuperScript II (Invitrogen). TaqMan quantitative real-time PCR (Q-PCR) was done using predesigned and optimized primers (Applied Biosystems, Foster City, CA, USA). Gene expression was measured with a Perkin-Elmer ABI PRISM 7900 sequence detection system (Applied Biosystems).

Statistical analysis

Statistical significance was assessed with 2-tailed, unpaired Student’s t tests. Data are means ± se.

RESULTS

Ethanol elevates endogenous atRA

Acute, chronic, and fetal ethanol exposures in C57BL/6 mice were investigated to determine the effect of ethanol on endogenous atRA, using a rigorously validated LC-MS/MS assay developed to quantify endogenous atRA in small biological samples (27, 28). One hour after an acute dose of ethanol (3.5 g/kg), liver atRA increased 2.4-fold, whereas hippocampus and testis atRA increased 1.6- and 1.5-fold, respectively (Fig. 1A). Serum atRA was elevated ∼50%, which did not reach statistical significance. Kidney and cortex atRA did not change (Supplemental Fig. 2A). Elevations in atRA correlated with increases in BAC% and returned to basal levels after ethanol clearance (Fig. 1A). Ethanol doses of 1.5 and 2.5 g/kg caused no statistically significant changes in atRA in any tissue assayed at any time (Supplemental Fig. 2B).

Figure 1.

LC-MS/MS quantification of ethanol effects on endogenous atRA. A) atRA in tissues and serum after an acute i.p. dose of 3.5 g/kg ethanol to 3-mo-old male C57BL/6 mice (n=4–8). Inset: corresponding BAC% values. B, C) Increases in e19 fetal hippocampus (B) and cortex atRA (C) in fetal mice from 2- to 3-mo-old dams after dam ingestion of 6.5% ethanol from e13 to e19. Each set of bars represents a pool of fetal hippocampi/cortexes (n=4–10) from an individual litter and the corresponding dam BAC%. D, E) atRA in tissues and serum (D) or brain regions (E) of 3-mo-old male mice after 1-mo feeding of control or 6.5% ethanol diets (n=4–8). Hip, hippocampus; Cx, cortex; OB, olfactory bulb; Th, thalamus; Cb, cerebellum; St, striatum. F, G) Representative LC-MS/MS chromatograms for liver (F) and hippocampus (G), showing the effect of ethanol. H) MS/MS spectrum of RA, showing its characteristic fragmentation pattern: *P < 0.05.

To determine the effect of maternal ethanol consumption on fetal atRA, dams were fed a liquid diet with 6.5% ethanol, an amount that reflects the intake of alcoholics, from e13 to e19 (16). Pups were harvested on e19. Fetal hippocampus atRA increased 1.5-fold with a maternal BAC% of 0.01% and 20-fold with a maternal BAC% of 0.13% (Fig. 1B). Similarly, fetal cortex atRA increased 2- to 50-fold, directly related to maternal BAC% (Fig. 1C).

Male mice fed the liquid diet with 6.5% ethanol were evaluated after 1 mo. Serum and testis atRA increased 10- and 2-fold, respectively, whereas kidney and liver atRA did not change (Fig. 1D). Brain atRA reacted region specifically: hippocampus and cortex atRA increased 20- and 2-fold, respectively, and olfactory bulb, cerebellum, striatum, and thalamus atRA did not differ from that of controls (Fig. 1E). Mice were euthanized in the morning after ad libitum feeding, and their BAC% ranged from undetectable to 0.15%. In the ethanol-dosed mice that had undetectable BAC, atRA remained ≥3-fold greater than that in control mice; in the remaining ethanol-dosed mice, atRA was proportional to BAC%. Representative LC-MS/MS chromatograms from analyses of liver and hippocampus after chronic ethanol exposure illustrate atRA measurements (Fig. 1F, G). A representative MS/MS spectrum shows the characteristic RA fragmentation pattern. The 301.1 to 205.0 m/z transition is used in the tandem mass spectrometric assay (Fig. 1H).

Ethanol increases mobilization of RE from liver and increases tissue retinol

To evaluate ethanol effects on extrahepatic vitamin A homeostasis, retinol and RE were quantified by HPLC/UV in the same samples used to quantify atRA (35). Liver, serum, hippocampus, and testis retinol values were elevated 2-fold 1 h after the acute 3.5 g/kg ethanol dose, whereas cortex and kidney were elevated 3- and 4-fold, respectively (Fig. 2A, B). Liver retinol declined 25% 4 h after dosing from its peak concentration at 2 h, whereas kidney, hippocampus, cortex, testis, and serum retinol continued an upward trend for 4 h, with 2.3- to 5.3-fold increases relative to controls. Serum RE increased 6-fold 2 h after a 3.5 g/kg ethanol dose (Fig. 2C). Acute retinol elevations were dose-dependent, with serum and tissue sites exhibiting different sensitivities to the amounts of ethanol dosed (Supplemental Fig. 2C).

Figure 2.

Ethanol increases mobilization of RE from liver and increases tissue retinol. A–C) Retinol in tissues and serum (A, B) and RE in serum (C) in 3-mo-old male mice (n=4–8) after an acute i.p. 3.5 g/kg ethanol dose. Corresponding BAC% values are shown in Fig. 1A (inset). D, E) e19 fetal hippocampus (D) and cortex retinol (ROL; E) after maternal ethanol exposure from e13 to e19 via feeding a 6.5% ethanol liquid diet to 2- to 3-mo-old dams. Each set of bars represents a pool of fetal hippocampi/cortexes (n=4–10) from an individual litter and the corresponding dam BAC% values. F–H) RE in tissues and serum (F), retinol in brain regions (G), and retinol in tissues and serum (H) in 3-mo-old male mice (n=4–8) after 1 mo of feeding a 6.5% ethanol liquid diet. Hip, hippocampus; Cx, cortex; OB, olfactory bulb; Th, thalamus; Cb, cerebellum; St, striatum. I) Representative HPLC/UV chromatograms showing control and ethanol-treated serum in 3-mo-old male mice (n=4–8) 1 h after an acute 3.5 g/kg dose. IS, internal standard. *P < 0.05.

Chronic ethanol exposure (6.5% ethanol from e13 to e19) increased retinol in fetal hippocampus 20% at a maternal BAC% of 0.01% and 3.8-fold at a maternal BAC% of 0.13% (Fig. 2D). Retinol increased between 30% and 2.5-fold in fetal cortex, proportional to maternal BAC% (Fig. 2E).

A 1-mo exposure to ethanol (6.5% ethanol) caused a 30-fold decrease in liver RE and ∼10-fold decrease in liver retinol (Fig. 2F, G) in adult male mice. In contrast, retinol and RE increased in kidney, testis, and brain regions. Ethanol caused kidney retinol and RE to increase 2-fold and in testis prompted a 5-fold increase in retinol and a 4-fold increase in RE (Fig. 2F, G). Ethanol dosing caused brain retinol to increase 2- to 5-fold, depending on the region (Fig. 2H) but did not induce increases in brain RE (Supplemental Fig. 2D). Representative HPLC/UV chromatograms of retinol and RE analyses are shown for hippocampus 1 h after an acute 3.5 g/kg ethanol dose (Fig. 2I).

Ethanol alters retinoid homeostasis in hippocampus

We used Q-PCR analysis to determine ethanol effects on expression of hippocampus genes that promote retinol uptake. Stra6 and CrbpI mRNA increased 55 and 82%, respectively, after 1 mo of ethanol feeding (Fig. 3A). We also determined the effect of ethanol on microsomal Rdh activity (36). Activity increased modestly in hippocampus microsomes from chronically dosed mice with both CrbpI-bound retinol and free retinol as substrates (7 and 26%, respectively) (Fig. 3B).

Figure 3.

Chronic ethanol feeding alters atRA biosynthesis in the hippocampus. Three-mo-old male mice were fed a 6.5% ethanol liquid diet for 1 mo, unless noted otherwise. A) Q- PCR of CrbpI and Stra6 mRNA in hippocampus. B) Retinal generated in 20 min from 0.5 μM retinol by hippocampus microsomes (25 μg protein). Retinol was delivered as either CrbpI-retinol or unbound retinol. C) atRA levels in hippocampus 2 h after an acute ethanol dose (3.5 g/kg i.p.) to control [vehicle only (Veh)], 4-methylpyrazole (4-MP)-treated, or carbenoxolone (Carb)-treated mice. Inhibitors were given 30 min before ethanol. D) atRA produced in 15 min from 1 μM retinal/fraction. Cytosol from control or ethanol-fed mice was fractionated by fast-protein LC. E) atRA biosynthesis in 15 min from 1 μM retinal of fractions pooled from the Raldh1 peak in D. F) Western blot of the Raldh1 pools assayed in E. G) atRA biosynthesis by primary cultured astrocytes treated with 0.5 or 2.5 μM retinol for 4 h. *P < 0.05. n = 4–8 (A–E); 3 (F).

To distinguish the effect of ethanol on Rdh (SDR) as opposed to Adh of the medium-chain Adh gene family, mice were treated with an SDR inhibitor (carbenoxolone) or an Adh inhibitor (4-methylpyrazole) 30 min before a single ethanol dose (3.5 g/kg). Endogenous atRA increased 1.5-fold in the hippocampi of 4-methylpyrazole-dosed mice 2 h after the ethanol dose, similar to the vehicle control. In contrast, carbenoxolone prevented the ethanol-induced atRA increase in hippocampus (Fig. 3C).

Raldh isoforms were resolved by fast-protein LC from hippocampus cytosol of mice fed ethanol for 1 mo. Quantifying activity per fraction revealed an increase in Raldh1 activity, no change in Raldh3 activity, and no measurable Raldh2 activity (Fig. 3D). The activity of the pooled Raldh1 fractions increased 2-fold in ethanol-fed mice, yet Raldh1 protein decreased to 40% of control (Fig. 3E, F).

To determine whether increases in substrate concentration in hippocampus can increase atRA production, primary hippocampus astrocytes were incubated with 0.5 or 2.5 μM retinol for 4 h. atRA biosynthesis increased with increasing substrate concentration (Fig. 3G).

Ethanol alters atRA metabolism in liver

Acute ethanol exposure increased atRA in liver, but chronic exposure did not. The mRNAs of Dhrs9 and Raldh3 increased 3.1- and 4.7-fold, respectively, in liver after 1 mo of chronic ethanol feeding (Fig. 4A). Several other genes involved in atRA biosynthesis were surveyed in hippocampus and liver of these mice; most showed no mRNA change (Supplemental Table 1).

Figure 4.

Ethanol alters atRA metabolism in liver. Three-mo-old male mice were fed a liquid diet with 6.5% ethanol for 1 mo, unless noted otherwise. A) RT-PCR of atRA biosynthetic enzyme mRNA in liver. B) Retinal generated in 15 min by liver microsomes (140 μg of protein) from 0.4 μM retinol delivered bound to CrbpI or unbound (n=3–8). C) atRA in liver 2 h after an acute i.p. 3.5 g/kg ethanol dose to control [vehicle only (Veh)], 4-methylpyrazole (4-MP)-treated, or carbenoxolone (Carb)-treated mice (n=3–8). Inhibitors were given 30 min before ethanol. D) Representative data showing the effect of chronic ethanol feeding on rates of atRA catabolism in liver microsomes. E) atRA elimination half-life in liver microsomes (n=6–8 individual experiments). *P < 0.05.

One-month feeding of the liquid diet with 6.5% ethanol increased the rate of retinol dehydrogenation in hepatic microsomes ∼30%, with either CrbpI-bound retinol or free retinol as substrate (Fig. 4B). Addition of 100 mM ethanol to the microsomal incubation increased activity 24% in control microsomes and an additional 14% in microsomes from ethanol-treated mice (Supplemental Fig. 3).

To distinguish the effect of ethanol on the contribution of Rdh or Adh to atRA production in liver, mice were treated with carbenoxolone or 4-methylpyrazole 30 min before a single ethanol dose (3.5 g/kg), and atRA was quantified 2 h after dosing. Carbenoxolone, but not 4-methylpyrazole, prevented the ethanol-induced increase in atRA (Fig. 4C).

Although Dhrs9 and Raldh3 mRNA were up-regulated, and hepatic microsomes showed greater atRA biogenesis, endogenous atRA did not change in liver after chronic ethanol feeding (Figs. 1D and 4A, B). To determine whether an increase in atRA catabolism compensated for increased atRA biosynthesis, we measured the rate of atRA catabolism. Hepatic microsomes isolated from ethanol-fed mice catabolized atRA 2.5-fold faster than microsomes isolated from control mice (Fig. 4D, E).

CrbpI contributes to ethanol-induced increases in atRA

Endogenous atRA was quantified in tissues from CrbpI^−/− mice after an acute ethanol dose (3.5 g/kg i.p.) or chronic ethanol feeding (1 mo with 6.5% ethanol). In contrast to those in WT mice, ethanol-induced atRA increases 1 h after an acute dose did not occur in CrbpI^−/− hippocampus, liver, and testis (Fig. 5A, B and Supplemental Fig. 4A). Deletion of CrbpI also ameliorated the acute ethanol-induced hippocampal, hepatic, and testicular elevations in retinol (Fig. 5C, D and Supplemental Fig. 4B). Similarly, absence of CrbpI eliminated the ethanol-induced atRA increase in hippocampus of mice fed ethanol for 1 mo (Fig. 5E). Chronic ethanol exposure did not change hepatic atRA in CrbpI^−/− mice (Fig. 5F) and produced a lesser retinol increase in CrbpI^−/− (2-fold) compared with that in WT hippocampus (3-fold) (Fig. 5G). Chronic ethanol had no effect on retinol in CrbpI^−/− liver, which was reduced significantly relative to that in WT liver (Fig. 5H). Interestingly, CrbpI ablation eliminated chronic ethanol-induced increases in testis retinol but did not eliminate the ethanol-induced atRA elevation (Supplemental Fig. 4C, D).

Figure 5.

CrbpI mediates elevation of endogenous atRA and retinol after acute and chronic ethanol exposure. A–D) Retinoids in 3-mo-old male mice (n=3–8) 1 h after an acute i.p. 3.5 g/kg ethanol dose. A) Hippocampus atRA. B) Liver atRA. C) Hippocampus retinol. D) Liver retinol. E–H) Retinoids in 3-mo-old male mice (n=3–8) after chronic 1-mo 6.5% ethanol liquid diet. E) Hippocampus atRA. F) Liver atRA. G) Hippocampus retinol. H) Liver retinol. *P < 0.05.

DISCUSSION

This work reports direct quantification of ethanol effects on endogenous atRA with sensitive and analytically rigorous methodology. Ethanol either increased atRA concentrations or had no effect on steady-state atRA, regardless of acute or chronic exposure. In no instance did ethanol decrease physiological levels of atRA in vivo during normal vitamin A nutriture in the tissues analyzed. Ethanol caused widespread changes in RE and retinol, compared with tissue- and region-specific increases in atRA, demonstrating that changes in retinol or RE do not predict changes in atRA. More notably, the largest atRA increases occurred in hippocampus and testis, areas functionally sensitive to both ethanol exposure and increased atRA (8, 10, 29,30,31,32, 37, 38).

These data, considered with the effects shared by retinoid disregulation and ethanol toxicity, suggest that superphysiological atRA elevation contributes to ethanol-induced disorders, such as FAS, adult cognitive dysfunction, and cancer. Tzimas et al. (39) reported that vitamin A dosed to rabbit dams (10 mg of retinyl palmitate/kg/d) increased atRA 100% in their embryos, resulting in teratology and/or embryo toxicity. Thus, the ethanol-induced atRA increases of 1.5- to 20-fold in fetal hippocampus and 2- to 50-fold in fetal cortex reported here extend beyond the range that causes phenotypic changes in embryos. Moreover, hippocampus-dependent spatial memory, spine formation, and branching require atRA (8, 10, 29,30,31,32). Prenatal ethanol exposure, similar to the exposure used here, decreases the number of dendritic spines and their length in mouse hippocampus and causes abnormal architecture in rat hippocampus (32). Maternal dosing with atRA (2.5 g/kg) alters postnatal behavior and cognitive function in rat offspring, manifested by hyperactivity, diminished function in the Biel water maze, and reduced brain weight (40,41,42,43). Prenatal ethanol exposure increases RARα ∼75% in rat embryos (15, 23). Increased RARα would augment the effect of increased atRA and could contribute to the excessive cell death characteristic of FAS and ethanol-induced adult cognitive dysfunction (12). Epidemiological evidence links ethanol exposure with liver cancer (3), as have alterations in vitamin A status (9, 16, 17). The ethanol-induced abnormalities in liver retinoid homeostasis could foster liver cancer and contribute to development of other cancers linked to both ethanol and vitamin A status and/or retinoid dysfunction.

The rapid increase in liver retinol (2 h) probably reflects hydrolysis of RE into retinol, suggested by in vitro stimulation of RE hydrolase activity in liver homogenates by ethanol (44). This result confirms the report that ethanol causes liver RE hydrolysis and an increase in serum RE in the rat, independent of RBP (45). After the peak at 2 h, retinol was increasingly affected by export, reflected by the decrease in liver and increases in extrahepatic tissues 4 h after ethanol dosing. Previous work, using radioactive tracers, reported retinol export from liver to extrahepatic tissues after an acute ethanol dose to rats, resulting in a 30% increase in adipose tissue, a 60% increase in kidney, and a 5-fold increase in brain 24 h after a 6- to 9-g/kg ethanol dose (46). Our studies complement these studies by providing data with much lower ethanol doses, extending the tissues involved, determining retinoid changes on a shorter time scale, and, most importantly, quantifying endogenous atRA.

Liver reacted uniquely to ethanol; although acute exposure increased atRA, chronic exposure did not. Increases in enzymatic activity coupled with increases in the mRNA of the Rdh and Raldh genes, Dhrs9 and Raldh3, and increased retinol suggest mechanisms for the atRA increases after acute ethanol dosing. Lack of a sustained atRA increase indicates maturation of compensatory mechanisms. The elimination half-life of atRA in hepatic microsomes isolated from chronic ethanol-dosed mice revealed an increased rate of atRA catabolism, which would compensate for increased atRA production. It appears that ethanol causes an initial increase in liver atRA, and the increase in atRA plus the chronic presence of ethanol induces atRA catabolism, resulting in an unchanged steady-state concentration with extended ethanol exposure. The lack of increase in mRNA of cytochromes P450 (Cyps) that catalyze atRA catabolism in mouse liver under physiological conditions (Cyp26a1, Cyp26b1, and Cyp2c39) suggests participation of additional Cyps during chronic ethanol exposure. Ethanol-induced pathological changes apparently amplify catabolism of atRA through participation of Cyps that normally would not catabolize lower atRA concentrations in the absence of ethanol.

Increases in Stra6 and CrbpI expression during ethanol dosing suggest a mechanism for the nearly 4-fold increase in hippocampus retinol. The plasma membrane receptor, Stra6, facilitates cellular uptake of retinol via interaction with RBP and delivers retinol to the intracellular chaperone CrbpI (47). The increase in hippocampus Crbp1 mRNA is consistent with observations that CrbpI mRNA increases 2-fold in e12 rat embryos and 50% in e20 rat snouts during ethanol dosing (23). This increase in the complex CrbpI-retinol would increase the pool of substrate that supports atRA biosynthesis (18, 36). The modest increase in microsomal Rdh activity and the 2-fold increase in Raldh1 activity also would contribute to ethanol-stimulated increases in hippocampus atRA. Substrate-driven increases in atRA production by primary hippocampus astrocytes supports the notion that increased substrate impels atRA increases in vivo. This process would be enhanced further by induction of both Stra6 and CrbpI by atRA, which would further increase retinol uptake and atRA biosynthesis (20, 48).

Normally, atRA down-regulates Raldh1 as a negative feedback mechanism (49). As anticipated from increased atRA, Raldh1 protein decreased in the hippocampus of chronically ethanol-treated mice. Yet, Raldh1 activity increased. This suggests post-translational regulation. atRA stimulates protein tyrosine kinase, and Raldh1 has two tyrosine phosphorylation sites (50, 51). Notably, decreases in protein levels have accompanied atRA stimulation of protein tyrosine kinase-catalyzed phosphorylation, similar to the decreased protein/increased activity observed here with Raldh1. atRA also induces phosphorylation of other proteins, such as ERK1/2 (29).

The three tissues (liver, testis, and hippocampus) that responded to acute ethanol with increases in atRA express CrbpI. Lack of an increase in both retinol and atRA in these tissues after acute ethanol dosing to CrbpI^−/− mice indicates that CrbpI mediates the increases in retinol concentrations in WT mice, which contributes to the increases in atRA biosynthesis. Chronic ethanol, in contrast, produced tissue-specific outcomes, probably by different mechanisms. The atypically low retinol in liver of CrbpI^−/− mice (20-fold less than normal) probably contributes to failure of chronic ethanol to affect retinol and atRA. The attenuated retinol increase in hippocampus of CrbpI^−/− mice could reflect the need for both Stra6 and CrbpI, with only Stra6 responding to chronically high atRA. Lack of an atRA increase in CrpbI^−/− mice indicates a central function for CrbpI-bound retinol in hippocampus atRA biosynthesis. The increase in testis atRA in CrbpI^−/− mice could result from compromised mechanisms unique to testis, as physiologically circulating levels of atRA do not readily cross the blood-testis barrier; testis relies on atRA biosynthesized in situ (38).

Despite the inability of Adh enzymes to recognize the physiological substrate used in vivo for atRA biosynthesis (CrbpI-retinol), some literature studies attribute the effect of ethanol to competitive inhibition of Adh-mediated atRA production. These conclusions have been supported by indirect estimation of atRA, rather than by specific quantification (25). Indirect estimation relies on a RARβ response element-lacz reporter bioassay. This method cannot quantify and does not respond solely to atRA and has generated both false-positive and false-negative results. Indirect estimation cannot distinguish the effects of ethanol on the concentrations of retinoids and other analytes that stimulate production of color vs. the effects of ethanol on the integrity of the detection system, which not only requires analyte delivery to response elements but thereafter also requires transcription, RNA processing, protein synthesis, and enzymatic action to produce color.

Decreased serum atRA in Adh-knockout mice has been observed by HPLC analysis after a toxic dose of retinol (50 mg/kg or ∼300-fold more than the recommended daily intake of retinol) was administered 30 min before ethanol dosing (26). This dose generated atRA serum levels ∼1600-fold higher than normal in controls, which decreased after ethanol dosing, but remained 200-fold higher than normal. Such levels are still toxic. Endogenous concentrations of atRA have not been determined in Adh-null mouse serum or tissues, before or after ethanol exposure. Thus, work with Adh-null mice shows only that massive retinol doses overwhelm retinoid homeostasis mechanisms by exceeding the capacity of binding proteins. This results in aberrant and/or ectopic atRA production, which ethanol attenuates, but not sufficiently to prevent RA toxicity. These experiments did not address the effect of ethanol on atRA under conditions that normally occur in the human population.

CONCLUSIONS

Direct quantification by an analytically rigorous assay shows that ethanol ingestion either increases atRA concentrations in a tissue/locus-specific manner or does not affect endogenous atRA. Increases in atRA were always accompanied by increases in retinol, but the reverse was not true: increases in retinol were not invariably associated with increases in atRA. Chronic ethanol dosing prompted compensatory mechanisms (increased catabolism induced by increased atRA), which contributed to maintenance of steady-state atRA levels at select sites. These effects are summarized in Fig. 6.

Figure 6.

Summary of chronic ethanol effects on atRA homeostasis in liver, hippocampus, and cortex. Chronic ethanol mobilizes RE and results in enhanced atRA production and catabolism and enhanced export of retinol and atRA from liver. Hippocampus and cortex show increased retinol and atRA uptake and atRA synthesis. Increased central nervous system atRA has been linked to suppression of neuron cell division, impairment of hippocampus-dependent spatial learning, and aberrant postnatal behavior and cognitive function.