Experimental Procedures

Materials.

Reagents were obtained from Sigma Chemical Co. (St. Louis, MO), Gibco BRL (Gaithersburg, MD), and other commercial sources. Prestained SDS-polyacrylamide gel electrophoresis (SDS-PAGE) standards were purchased from Bio-Rad Laboratories (Hercules, CA), and recombinant human CYP2E1 expressed in Escherichia coli was from PanVera (Madison, WI). Nitrocellulose membrane (NitroBind, 0.45 μm pore size) was from Micron Separations, Inc. (Westboro, MA). Rabbit anti-human cytochrome P-4502E1 antibody was obtained from Research Diagnostics, Inc. (Flanders, NJ). Horseradish peroxidase-conjugated goat anti-rabbit IgG and SuperSignal ULTRA chemiluminescent substrate were purchased from Pierce (Rockford, IL). Radiolabeled compounds ([α-³²P]UTP, 800 Ci/mmol and [³H]CTP, 30 Ci/mmol) were from New England Nuclear (Boston, MA). Total RNA from adult human brain was obtained from Clontech (Palo Alto, CA).

Tissues.

Brain tissues from human embryos and fetuses were acquired through the Birth Defects Research Laboratory at the University of Washington (Department of Pediatrics) following dilatation and curettage procedures. Handling of these tissues was in accordance with the guidelines of the Human Subjects Review Committee at the same institution. The gestational ages ranged between 46 to 113 days, as estimated from measurements of foot length. The prenatal tissues were snap-frozen and stored under liquid nitrogen until analyzed. An adult female Sprague-Dawley rat, purchased from B&K Laboratories (Freemont, CA), was euthanized by exsanguination while under deep halothane anesthesia. The brain and liver tissues were rapidly removed and placed in homogenization buffer for subsequent microsomal preparations.

Protein and RNA Preparations.

For immunoblotting experiments, rodent tissues and prenatal human brain tissues (<2 g each) were homogenized in approximately 10 ml of 0.1 M Tris buffer containing 1.15% KCl, 0.1 mM ethylenediaminetetraacetic acid, 0.1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, and 10% (v/v) glycerol at pH 7.4. The homogenate was diluted with buffer to 50 ml and centrifuged at 12,000g for 30 min. The supernatant was transferred to a clean tube and centrifuged at 104,000g for 1 h. The resulting pellet was resuspended in the above buffer (without glycerol) and centrifuged at 104,000g for 1 h. The washed pellet was then resuspended in 0.5 ml buffer per gram starting tissue. For enzymatic assays, 10 mM phosphate buffer containing 0.25 M sucrose at pH 7.4 was used throughout the preparation of microsomes and the final wash step was not performed. Protein concentrations were determined by the Bio-Rad protein assay or Lowry method (Lowry et al., 1951) with BSA as a standard.

Total RNA was isolated from prenatal human brain tissues by a guanidinium thiocyanate/phenol-chloroform extraction procedure (Chomczynski and Sacchi, 1987). To eliminate potential guanidinium thiocyanate contamination, the RNA pellets obtained after the second precipitation were resuspended in 1 mM ethylenediaminetetraacetic acid and precipitated with 2 M sodium acetate, pH 4.0, and isopropanol. RNA purity was evaluated spectrophotometrically using the A₂₆₀/A₂₈₀ and A₂₆₀/A₂₃₀ ratios; concentrations were calculated from the A₂₆₀values. Integrity of the purified RNA was confirmed by comparison of the 28S and 18S ribosomal RNA band intensities visualized by agarose gel electrophoresis and ethidium bromide staining.

Enzymatic Assay.

Mono-oxygenase activity was measured by incubating approximately 1 to 3 mg of microsomal protein with 0.2 mMp-nitrophenol and 2 mM NADPH in 40 mM KH₂PO₄, and 5 mM MgCl₂, pH 7.4 (1 ml total volume) at 37^oC. Glucose 6-phosphate (1 mM) and glucose 6-phosphate dehydrogenase (5 U) were also included to regenerate NADPH. The buffer-substrate-cofactor mixture was equilibrated at 37^o for 3 min before the addition of protein. After 2 h, the reaction was stopped with 50 μl of cold trifluoroacetic acid. Samples were placed on ice for 10 min and then centrifuged to pellet the precipitated protein. The supernatant was filtered through an Ultrafree-MC centrifugal filter (Millipore, Bedford, MA) and analyzed by reversed-phase HPLC. A volume of 50 μl was injected onto a Waters C-18 Radial Pac cartridge (10 cm × 8 mm × 4 μm). Compounds were eluted isocratically with a mobile phase containing 25% acetonitrile and 0.1% trifluoroacetic acid, at a flow rate of 1.5 ml/min. The column eluate was monitored by electrochemical detection (Mishin et al., 1996) with an ESA Coulochem II detector set at 700 mV.

Western Immunoblotting.

Samples of microsomal protein and prestained standards were separated by SDS-PAGE (Laemmli, 1970) using the Mini-PROTEAN II electrophoresis system (Bio-Rad). The running and stacking gels were composed of 10% and 5% polyacrylamide, respectively, and both gels contained 0.1% SDS. The protein gel was electroblotted for 2 h onto a nitrocellulose membrane using the Bio-Rad Trans-Blot electrophoretic transfer system. The membrane was blocked for 1 h in 2% nonfat powdered milk and rinsed briefly. TBST buffer containing 0.1 M Tris, 0.9% NaCl, and 0.1% Tween 20 was used for all washes and dilutions. The membrane was incubated overnight with rabbit anti-human CYP2E1 antibody (1:10,000 dilution) then washed three times. After incubating 1 h with peroxidase-conjugated secondary antibody (1:250,000 dilution), the membrane was washed six times and immersed in SuperSignal ULTRA chemiluminescent substrate (Pierce) for 15 min. The immunoblot was exposed to Hyperfilm enhanced chemiluminescence (Amersham, Arlington Heights, IL) for several minutes, and the film was developed for visualization of signal. Bands of immunoreactive CYP2E1 protein were quantitated by scanning densitometry.

Ribonuclease Protection Assay.

The CYP2E1 antisense riboprobe, used in the RNase protection assays, was generated as described by Carpenter et al. (1996). Reverse transcription/PCR was used to synthesize and amplify a 475-bp fragment of CYP2E1 cDNA. The PCR product was then ligated into a cloning vector called pCR2.1 (Invitrogen, San Diego, CA). This construct was used to transform Epicurian Coli XLI-Blue competent cells (Stratagene, La Jolla, CA) for purification of plasmid DNA. A portion of the purified DNA was digested with the restriction endonuclease EcoRI then re-ligated to reverse the insert orientation, enabling synthesis of sense RNA transcript. The pCR2.1-CYP2E1 plasmids were linearized withHindIII restriction enzyme and treated with Proteinase K to inactivate ribonucleases before the transcription reaction, according to a protocol from Ambion (Austin, TX). The MAXIscript T7 In Vitro Transcription Kit (Ambion) was used in the production of radiolabeled riboprobes; reaction mixtures contained 1 μg linearized CYP2E1 or β-actin DNA template, T7 RNA Polymerase (10 U), transcription buffer, 0.5 mM each unlabeled NTP, and 3.1 μM [α-³²P]UTP (20 μl total volume). The human β-actin riboprobe served as an internal control. Sense-CYP2E1 RNA transcript was synthesized in the presence of 6.7 μM [³H]CTP; the yield and specific activity were calculated from the results of liquid scintillation counting following TCA precipitation of the reaction products. The antisense,³²P-labeled probes were gel-purified by denaturing PAGE, as described in the Ambion instruction manual.

For the RNase protection assay, both CYP2E1 (250,000 cpm) and β-actin (25,000 cpm) antisense riboprobes were combined with the following samples: 50 μg of prenatal human brain RNA, 20 μg of adult human brain RNA, 20 μg of yeast RNA (Ambion), and 1 to 50 pg of tritiated CYP2E1 sense RNA. Reagents from the RPA II Ribonuclease Protection Assay Kit (Ambion) were used throughout the assay according to guidelines provided in the accompanying instruction manual. An aliquot of the RNA Century Marker Plus Template Set (Ambion) was transcribed in the presence of [α-³²P]UTP and ethanol-precipitated to generate RNA size standards. After separating radiolabeled standards and hybridized RNA fragments by gel electrophoresis, the 5% polyacrylamide/8 M urea gel was transferred to filter paper and dried. The dried gel was exposed to a phosphor screen for 2 days, and the screen was scanned with the Molecular Dynamics PhosphorImager (Sunnyvale, CA). Data were analyzed with ImageQuant software, Version 3.3.

Results

The developmental expression of CYP2E1 protein in human brain was quantitatively evaluated by immunoblot analysis. Microsomes were prepared from two separate pools of embryonic and two separate pools of fetal brain tissues with gestational ages ranging from 53 to 59 days and 67 to 98 days, respectively. The primary antibody used for these experiments was reported to be highly specific, detecting the CYP2E1 isoform exclusively in microsomes prepared from human, rat, rabbit, and hamster liver (Research Diagnostics, Inc.). An extremely sensitive chemiluminescent system was used to detect immunoreactive protein in the samples prepared from prenatal human brain and adult rat liver and brain (Fig. 1). Adult rat brain and liver tissues were included in the experiment for the following reasons: samples of adult human brain were unavailable, the highly homologous rat 2E1 isoform was expected to have the same mobility as the human isoform, and the numerous studies conducted with adult rat liver made this tissue a logical choice for purposes of comparison. The predominant bands comigrated with recombinant human CYP2E1 standards at a molecular weight consistent with the native monomeric protein, based on prestained molecular weight markers. The less intense band visible in the lane containing human embryonic brain was attributed to the affinity of the secondary antibody for an unidentified protein, whereas the doublet that appeared in the lane containing adult rat brain is unexplained at present. Scanning densitometry was performed on the immunoblots to quantitate positive signals and generate standard curves (r² = 0.95). From the results we determined that the CYP2E1 content differed by <23% in the embryonic and fetal samples analyzed. The enzyme was present at approximately 1.6 μg per mg microsomal protein, similar to the amount in adult rat brain and about 150-fold less than the amount measured in adult rat liver. In terms of tissue, this is equivalent to 0.9 μg or 17 pmol CYP2E1 protein per gram of prenatal brain (wet weight).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Immunoblot analysis of CYP2E1 content in microsomes from prenatal human brain and adult rat tissues. Microsomal protein and pure CYP2E1 standards were subjected to SDS-PAGE, then transferred to nitrocellulose membrane. An immunochemical staining procedure was used to detect CYP2E1 on the membrane, as described in Experimental Procedures. Aliquots are shown from adult rat liver (10 ng), adult rat brain (2.0 μg), human fetal brain (1.2 μg), and human embryonic brain (1.2 μg) where amounts are based on total protein concentration. Recombinant human CYP2E1 protein serves as the standard. A predominant band of immunoreactive protein is visible in each lane of sample and standard corresponding to CYP2E1 (53 kD).

The enzymatic activity of CYP2E1 was measured in prenatal human brain using an HPLC assay and p-nitrophenol as the substrate. Microsomal protein that served as the enzyme source was prepared from pools of tissues with gestational ages varying by <6 days in each pool. Single assays were performed with each sample due to the difficulty in obtaining sufficient amounts of brain tissues for the experiment and the large amount of protein needed to produce detectable levels of product during incubation. Unfortunately, antibody inhibition data could not be collected with these samples. The formation of 4-nitrocatechol was monitored electrochemically, a method that is 30-fold more sensitive than UV detection (Mishin et al., 1996). Peak height data from one 320-pmol standard was used to calculate quantities of product formed in the reaction mixtures. The results are summarized in Fig. 2 where specific activities are plotted as a function of gestational age. 4-Nitrocatechol generated in all samples (except days 45–49) was within the linear range of the calibration curve reported by Mishin et al. (1996). From the bar graph representing data from a single experiment, it seems CYP2E1-mediated hydroxylation increases during organogenesis and throughout the early fetal period. The specific activity at 81 gestational days was approximately twice that of the days 53 to 59 tissue pool. Increases in levels of P-450 reductase and/or cytochromeb₅ also could feasibly contribute to this rise in CYP2E1 activity. The specific activity displayed by prenatal human brain was about half that measured in adult rat brain, whereas the content of immunodetectable CYP2E1 protein was similar in both tissues. The specific activity was approximately four orders of magnitude lower in prenatal human brain compared to adult rat liver (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

CYP2E1 activity in human brain as a function of gestational age. Enzymatic activity was determined with an HPLC assay, as described in Experimental Procedures. The conversion of p-nitrophenol to 4-nitrocatechol was measured with an electrochemical detector following a 2-h incubation. Specific activity is expressed as pmole product formed per milligram microsomal protein per hour of reaction. Microsomes were prepared from pools of at least three tissues, except for the 81- day sample in which two tissues were pooled, and the 113-day sample in which one brain tissue was used. Gestational age was estimated from measurements of foot length.

The level of CYP2E1 mRNA expression was evaluated in prenatal human brain using a quantitative RNase protection assay. Samples contained 50 μg of total RNA isolated from tissues ranging from 46 to 108 gestational days. Radiolabeled, antisense riboprobe and sense transcript were synthesized from a 475-bp fragment of CYP2E1 cDNA that served as the template. Various amounts of tritiated sense RNA were included in each experiment to generate standard curves (r² ≥ 0.80) for quantitation of positive signals. Phosphorimaging was used for the detection of radioactive hybrids due to the increased sensitivity and shorter exposure time required. As illustrated in Fig.3, CYP2E1 transcripts were found in all samples analyzed. β-actin transcripts were also found that confirmed that the RNA was intact and enabled normalization of signal in each sample. The assay results presented in Table1 suggest that the CYP2E1 mRNA content increases in prenatal human brain from day 46 through day 53, then levels off with small variations (≤16% deviations in mean values) observed between days 53 and 84. The level of CYP2E1 message in one sample analyzed from the second trimester (108 days) was about 30% higher than levels detected at 58 and 74 gestational days (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Ribonuclease protection assay of human brain RNA using CYP2E1 riboprobe. CYP2E1 and β-actin antisense probes, prepared as described in Experimental Procedures, were hybridized to 50 μg of prenatal and 20 μg of adult human brain RNA, and to various amounts of CYP2E1 sense transcripts. Protected fragments were separated by denaturing gel electrophoresis and exposed to a phosphorimaging screen for 2 days. Based on positioning of radiolabeled RNA markers, the upper bands migrated at 500 bases and represent CYP2E1 mRNA sequences, whereas the lower set of bands migrated at 250 bases and represent β-actin mRNA sequences.

Discussion

This investigation provides evidence supporting the CYP2E1-mediated bioactivation of xenobiotics in prenatal human brain tissue. Our demonstration of CYP2E1 mRNA expression in brain during embryonic and fetal stages of development confirms earlier findings byBoutelet-Bochan et al. (1997) and, with the use of highly sensitive methods of detection, it was also possible to detect immunoreactive and functional CYP2E1 protein. Analysis of the quantitative results collected in this study supported a pattern of developmental expression. Significant increases in activity and message levels were apparent between gestational days 45 and 53 of embryogenesis. This time interval overlaps with the start of organogenesis (days 50–60), often described as the period when the conceptus is most vulnerable to chemical-induced injury. Small variations in message were observed in the individual RNA samples analyzed from day 53 through day 84 of gestation, whereas CYP2E1-specific activity appeared to be leveling off toward the end of the first trimester. The results tend to suggest that transcriptional activation causes a dramatic increase in human brain CYP2E1 content around day 50 and that a fairly constant level is maintained throughout the early fetal period, at least until gestational day 113.

A number of factors could potentially influence the amounts of CYP2E1 we report in the developing brain. Variable levels of induction could occur during gestation by post-transcriptional (Badger et al., 1993) and post-translational mechanisms (Koop and Tierney, 1990) depending on the extent of intrauterine exposure to environmental chemicals and drugs. The nutritional status of the embryo or fetus could affect CYP2E1 content because inadequate carbohydrate, fat, vitamin, and mineral intake reportedly alters its regulation in liver microsomes prepared from rats and mice (Yang et al., 1992). Interindividual differences could be the result of a genetic polymorphism such as the mutant c2 allele (Hayashi et al., 1991), which is associated with higher transcriptional activity in human liver tissue. Measurements of CYP2E1-specific activity could be influenced by changes in P-450 reductase and cytochrome b₅ levels in prenatal human brain. Finally, regional differences in CYP2E1 tissue content could also be a factor.

Several studies report findings that, collectively, may support the participation of CYP2E1 enzyme in the elicitation of alcohol-related birth defects. The significant and progressive reductions of CYP2E1 message and protein observed in pregnant rats (Casazza et al., 1994) may also occur in humans, thus, exposing the fetus to higher concentrations of ethanol from the maternal circulation and increasing the likelihood of localized damage to developing conceptal tissues. Induction of CYP2E1 was reported in fetal rat liver following a maternal ethanol diet (Carpenter et al., 1997) and in cultured human fetal hepatocytes after ethanol treatment (Carpenter et al., 1996). The combination of lower enzyme levels in adult tissues during pregnancy plus the presence of readily inducible CYP2E1 enzyme in the conceptus could lead to substantial ethanol metabolism in the fetus and increased risk of alcohol teratogenesis. Although earlier investigations reported that only a portion of immunochemically detected 2E1 isoform is catalytically active (Eliasson et al., 1988; Ronis and Ingelman-Sundberg, 1989), extremely low levels of P-450s were shown to be sufficient to catalyze bioactivation resulting in profound abnormalities in rodent embryos (Juchau et al., 1992). Additional studies are needed to determine the extent to which CYP2E1 regulation in prenatal human brain is similar to that observed in liver and among species.

Current speculation regarding the underlying mechanisms through which ethanol exerts its harmful effects on prenatal development suggests the involvement of human ethanol-oxidizing enzymes. These enzymes catalyze the initial step of ethanol metabolism producing acetaldehyde, a metabolite capable of inducing tissue damage by directly impairing nucleic acid and protein synthesis. The degree of neurological damage attributable to acetaldehyde levels attained in vivo is unknown, as well as which enzyme system(s) is responsible for metabolite generation in the CNS. Catalase-mediated oxidation of ethanol was demonstrated in vitro using adult (Aragon et al., 1992; Gill et al., 1992) and fetal rat brain homogenates (Hamby-Mason et al., 1997). The presence of catalase-specific inhibitors reduced the formation of acetaldehyde in a dose-dependent manner, whereas ADH and P-450-specific inhibitors had no apparent effect. Because an exogenous supply of the cosubstrate hydrogen peroxide was added to the system, it is questionable whether sufficient amounts are present in brain tissue/cells to support the peroxidation reaction (Hunt, 1996). We are unaware of any reports concerning active catalase enzyme in human embryonic or fetal tissues. The class I ADH isozymes that efficiently metabolize ethanol were not detected in prenatal human brain tissues from the second trimester by Northern blot analysis (Estonius et al., 1996), although a more sensitive mRNA detection method may yield different results. The role of class I, IV, or other ADH in eliciting neuroembryotoxic effects cannot be discounted until further studies are completed supporting their presence or absence in fetal brain. Interestingly, human class V ADH has been suggested to be a predominantly fetal ADH (Estonius et al., 1996).

The third enzyme system catalyzing the conversion of ethanol to acetaldehyde consists of CYP enzymes and is commonly referred to as MEOS. We have provided evidence that the major enzymatic component of MEOS, CYP2E1, is present and functionally active in human prenatal brain during the first trimester. The elevated expression of CYP2E1 around gestational day 50 coincides with increasing oxygen requirements of the conceptus, triggered by the onset of organogenesis and cellular differentiation (Fantel, 1996). This suggests increased availability of molecular oxygen cosubstrate necessary for alcohol oxidation. CYP2E1 is induced by ethanol at relatively low concentrations; enzyme levels in brain were reported to increase 2- to 6-fold in response to a single dose (Warner and Gustafsson, 1994; Tindberg and Ingelman-Sundberg, 1996). With an apparent K_mapproximately 10-fold higher compared with those of catalase and ADH, CYP2E1 may play a more important role during periods of high ethanol concentrations or following repeated alcohol exposures, common in alcoholics. A study conducted by Alderman and coworkers (1987) revealed that MEOS contributes significantly to ethanol oxidation in deer mice, metabolizing 36% of ethanol at low blood concentrations and 73% at high blood levels based on alteration of kinetic parameters by deuterium isotope effects. Additional studies are needed to assess the contribution of each enzyme system to acetaldehyde production in prenatal human brain and to more fully determine the teratogenic potential of acetaldehyde in the CNS.

Alcohol-induced neurotoxicity in the human conceptus may be attributed, at least partially, to conditions of oxidative stress contributed by CYP2E1-mediated reactions. CYP2E1 is considered a leaky enzyme due to the uncoupling of the mixed function oxidase reaction and is capable of reducing molecular oxygen to superoxide anion and hydrogen peroxide, particularly in the presence of certain substrates (Lieber, 1997). These active oxygen intermediates may react sequentially with nonheme iron to form the highly potent and toxic hydroxyl radical (Fantel, 1996). Additionally, the production of hydroxyethyl radicals was demonstrated in the CYP2E1 active site using rat liver microsomes and a spin-trapping technique (Albano et al., 1991). The oxygen and ethanol-derived free radicals generated can elicit cellular injury by breaking single strands of DNA, inactivating metabolic enzymes, and initiating lipid peroxidation reactions. Embryonic and fetal brain tissues are especially susceptible to injury by the peroxidative process because these membranes are rich in easily oxidizable, polyunsaturated fatty acid side chains (Fantel, 1996). There is evidence linking CYP2E1 to the enhanced production of reactive oxygen species and lipid peroxidation in rat brain homogenates (Montoliu et al., 1994), rat liver microsomes (Ekström and Ingelman-Sundberg, 1989), and a transfected human hepatoma-derived cell line (Dai et al., 1993). However, there are currently few data concerning free radical formation in fetal tissues or the impact of membrane lipid peroxidation on prenatal development.

The human brain is particularly vulnerable to the damaging effects of reactive oxygen intermediates due to its complexity and long period of development relative to other organs. Several studies indicate that antioxidant enzymes and small molecule antioxidants that serve to protect cells from oxidative stress exhibit extremely low activities and message levels in prenatal tissues, especially the brain (De Haan et al., 1994; Fantel et al., 1995). Although this rudimentary antioxidant defense system is adequate for development under normal conditions, the system might be easily overcome by ethanol, resulting in neurological and morphological abnormalities characteristic of the FAS.

The findings in this study confirm the presence of CYP2E1 message, immunoreactive protein, and functionally active enzyme at relatively low levels in prenatal human brain tissue (gestational weeks 7–16). Based on these results, we suggest that the P-4502E1 isoform may play an important role in alcohol teratogenesis, more specifically, in eliciting neurotoxic effects. It is possible that, during ethanol metabolism in conceptal brain, CYP2E1 generates reactive chemical species including oxygen-derived free radicals, hydroxyethyl radical, acetaldehyde, and other aldehydes derived from lipid peroxides. Each of these chemical species is capable of contributing to alcohol-induced cellular injury that is often manifested as CNS dysfunction. Current understanding would be advanced in this area by investigating CYP2E1 distribution in human fetal brain and comparing with brain regions especially sensitive to ethanol-induced toxicity. These studies are underway.