Alcohol and oxidative liver injury


  • Aparajita Dey,

    1. Department of Pharmacology & Biological Chemistry, Mount Sinai School of Medicine, New York, NY
    Search for more papers by this author
  • Arthur I. Cederbaum

    Corresponding author
    1. Department of Pharmacology & Biological Chemistry, Mount Sinai School of Medicine, New York, NY
    • Department of Pharmacology and Biological Chemistry, Mount Sinai School of Medicine, Box 1603, One Gustave L Levy Place, New York, NY 10029
    Search for more papers by this author
    • fax: 212-996-7214

  • Potential conflict of interest: Nothing to report.


Acute and chronic ethanol treatment has been shown to increase the production of reactive oxygen species, lower cellular antioxidant levels, and enhance oxidative stress in many tissues, especially the liver. Ethanol-induced oxidative stress plays a major role in the mechanisms by which ethanol produces liver injury. Many pathways play a key role in how ethanol induces oxidative stress. This review summarizes some of the leading pathways and discusses the evidence for their contribution to alcohol-induced liver injury. Many of the seminal reports in this topic have been published in Hepatology , and it is fitting to review this research area for the 25th Anniversary Issue of the Journal. (Hepatology 2006;43:S63–S74.)

Alcohol, Oxidative Stress, and Cell Injury

Many pathways have been suggested as playing a key role in how ethanol induces “oxidative stress.”1–8 Some of these include redox state changes; production of the reactive product acetaldehyde; damage to mitochondria; direct or membrane effects caused by hydrophobic ethanol; ethanol-induced hypoxia; ethanol effects on the immune system and altered cytokine production; ethanol induction of CYP2E1; ethanol mobilization of iron; effects on antioxidant enzymes and chemicals, particularly mitochondrial and cytosolic glutathione; 1-electron oxidation of ethanol to the 1-hydroxyl ethyl radical. Many of these pathways are not exclusive of one another, and several, indeed many, systems likely contribute to the ability of ethanol to induce a state of oxidative stress. A review of these pathways and how they contribute to alcoholic liver disease (ALD) is summarized here.

The most convincing data that oxidative stress contributes to alcohol-induced liver injury come from the studies using the intragastric infusion model of alcohol administration. In these studies, alcohol-induced liver injury was associated with enhanced lipid peroxidation, protein carbonyl formation, formation of the 1-hydroxyl ethyl radical, formation of lipid radicals, and decreases in hepatic antioxidant defense, especially GSH.9–13 Replacement of polyunsaturated fat (required for lipid peroxidation to occur) with saturated fat or medium chain triglycerides in the diets fed to rats intragastrically lowered or prevented the lipid peroxidation and the alcohol-induced liver injury.14 Addition of iron, known to generate · OH and promote oxidative stress, to these diets exacerbated the liver injury.15 Importantly, addition of antioxidants such as vitamin E, ebselen, superoxide dismutase, GSH precursors prevented alcohol-induced liver injury.16–19

Because alcohol-induced liver disease has been linked to oxidative stress, we investigated the effect of a compromised antioxidant defense system, Cu, and Zn-superoxide dismutase (Sod 1) deficiency on alcohol-induced liver injury.20 C57BL/129SV wild-type (Sod 1+/+) and Sod1 knockout (Sod 1−/−) mice were fed dextrose or ethanol (10% of total calories) liquid diets for 3 weeks. Histological evaluation of liver specimens of Sod 1−/− mice fed ethanol showed the development of liver injury ranging from mild to extensive centrilobular necrosis and inflammation (Fig. 1). Sod 1+/+ mice fed ethanol showed mild steatosis; both Sod 1+/+ and Sod 1−/− mice fed the dextrose diet had normal histology. Alanine aminotransferase levels were significantly elevated only in Sod 1−/− mice fed ethanol. Ethanol consumption increased levels of protein carbonyls and lipid peroxidation aldehydic products in the liver of Sod 1−/− mice. Hepatic adenosine triphosphate (ATP) content was reduced dramatically in the Sod 1−/− mice fed ethanol in association with a decrease in the mitochondrial reduced GSH level and activity of MnSOD. Immunohistochemical determination of 3-nitrotyrosine (3NT) residues in liver sections of the Sod1 knockout mice treated with ethanol showed a significant increase of 3NT staining in the centrilobular areas (Fig. 2). A rather moderate ethanol consumption promoted oxidative stress in Sod 1−/− mice, with increased formation of peroxynitrate, protein carbonyls, and lipid peroxidation and subsequent liver injury. Although this review focuses on ethanol-induced oxidative stress in the liver, ethanol also produces oxidative stress in other tissues, including the brain, heart, pancreas, and testis.21–25

Figure 1.

Photomicrographs of liver samples stained with hematoxylin-eosin. (A) Sod 1−/− mice after 3 weeks of ethanol feeding (10% of total calories) with severe centrilobular hepatocellular necrosis and inflammation (arrows). (B) Sod 1−/− mice after 3 weeks of ethanol feeding (10% of total calories) with mild, focal areas of necrosis. (C) Sod 1−/− mice fed the dextrose diet. The histological picture shows no signs of pathological changes in the liver. (D) Sod 1+/+ mice fed ethanol (10% of total calories) for 3 weeks show moderate fatty liver only. (E) Sod 1+/+ mice fed the dextrose diet show no pathological changes.

Figure 2.

Photomicrographs of immunohistochemical staining of 3NT residues in proteins of liver specimens. (A) An increase in the level of 3NT staining in the centrilobular area in liver of Sod 1−/− mice (arrows) after 3 weeks of ethanol feeding (10% of total calories). (B) Sod 1−/− mice fed the dextrose diet. (C) Sod 1+/+mice fed ethanol (10% of total calories) for 3 weeks. (D) Sod 1+/+ mice fed the dextrose diet.


ALD, alcoholic liver disease; GSH, glutathione; CYP2E1, cytochrome P4502E1; Sod 1, superoxide dismutase; ATP, adenosine triphosphate; 3NT, 3-nitrotyrosine; TNFα, tumor necrosis factor alpha; ROS, reactive oxygen species; HER, 1-hydroxyethyl radical; NF-κB, nuclear factor-kappaB; iNOS, inducible nitric oxide synthase; MDA, malondialdehyde; HNE, 4-hydroxy-2-nonenal; MAA, malondialdehyde-acetaldehyde.

Kupffer Cells and Alcoholic Liver Disease

Kupffer cells are stimulated by chronic ethanol treatment to produce free radicals and cytokines, including tumor necrosis factor alpha (TNFα), which plays a role in ALD.26, 27 This stimulation is mediated by bacterial-derived endotoxin, and ALD is decreased when gram-negative bacteria are depleted from the gut by treatment with lactobacillus or antibiotics.28 Destruction of Kupffer cells with gadolinium chloride attenuated ALD.26 A major advance was the finding that anti-TNFα antibodies protect against ALD.27 NADPH oxidase was identified as a key enzyme for generating reactive oxygen species (ROS) in Kupffer cells after ethanol treatment, because diphenylene iodonium chloride, an NADPH oxidase inhibitor, lowered 1-hydroxyethyl radical (HER) production and ethanol-induced liver injury.29 Moreover, in mice deficient in a subunit of NADPH oxidase, p47phox, the ethanol-induced increase in ROS and TNFα and liver injury was decreased.30 The role of TNFα in ALD was further validated by the findings that the ethanol-induced pathology was nearly blocked in TNFα receptor1 knockout mice.31 Interestingly, although liver pathology was blunted in the TNFα receptor1 knockout mice fed ethanol, no decline was seen in the intensity of free radical signals, consistent with the hypothesis that free radicals act as redox signals for TNFα production but do not directly damage cells.32

The transcription factor nuclear factor-kappaB (NF-κB) regulates activation of many inflammatory genes, including TNFα. Endotoxin activates NF-κB, leading to the hypothesis that inhibition of NF-κB would prevent ALD.33 Administration of an adenovirus encoding the transgene for the IκB superrepressor to rats chronically infused with ethanol blunted the ethanol-induced activation of NF-κB, TNFα production and pathological changes. A general scheme to explain these results is that chronic ethanol treatment elevates endotoxin levels, endotoxin activates Kupffer cells to produce free radicals via NADPH oxidase, the free radicals activate NF-κB, leading to an increase in production of TNFα, followed eventually by tissue damage.34 Recent reviews on the roles of Kupffer cells in ALD can be found in Wheeler et al.35 and Takei et al.36

Iron- and Alcohol-Induced Oxidative Stress

Iron promotes oxidative stress by catalyzing the conversion of less reactive oxidants such as superoxide or H2O2 to more powerful oxidants such as hydroxyl radical or perferryl-type oxidants. An increase in hepatic iron concentrations occurs in alcohol-dependent individuals and elevated hepatic iron uptake is seen in patients with alcohol-induced cirrhosis.37, 38 An increase in the cellular pool of low-molecular-weight iron occurs during ethanol metabolism in rat hepatocyte cultures.39 Acute ethanol administration to rats elevated the iron content in liver and cerebellum.40 Chronic ethanol treatment increases iron uptake by hepatocytes.41 In rats, chronic ethanol feeding for 8 weeks elevated iron content in the hepatocytes and Kupffer cells.42 Treatment of rats with ethanol plus carbonyl iron strikingly elevated liver iron levels and produced significant liver injury.42, 43 In the intragastric infusion model, addition of a small amount of iron, which only elevated hepatic iron levels 2- to 3-fold, enhanced lipid peroxidation, serum transaminase levels, and induced fibrosis.15 Ethanol administration elevated the iron content of Kupffer cells, and this was suggested to prime Kupffer cells for NF-κB activation and ultimately for TNFα production and ALD.44 Addition of Fe2+ but not Fe3+ increased TNFα release by rat Kupffer cells in an NF-κB–dependent manner.45 Further studies supported a role for iron as a direct agonist to induce intracellular signaling for NF-κB activation in Kupffer cells by an oxidant-dependent process.45, 46 Oral iron chelators such as deferiprone attenuated these effects, reducing the elevations in non-heme iron, lipid peroxidation, and liver fat accumulation and injury.44, 47

ROS production, lipid peroxidation, and interaction with iron chelates were found to be enhanced with microsomes from ethanol-treated rats. This was associated with elevated levels of CYP2E1 and could be blocked by inhibitors of CYP2E1 or by anti-CYP2E1 immunoglobulin. In HepG2 cells expressing CYP2E1, addition of an iron chelator, ferric-nitrilotri-acetate, produced greater toxicity than that found with control HepG2 cells. This enhanced synergistic toxicity was associated with increased lipid peroxidation and oxidant stress and could be blocked by anti-oxidants. Damage to the mitochondria played a critical role in the CYP2E1 plus iron–dependent toxicity. In the CYP2E1-expressing HepG2 cells, synergistic interactions between iron and polyunsaturated fatty acids were observed.48 The proceedings of a recent symposium on the role of iron in alcoholic liver disease can be found in Alcohol 30, No. 2, 2003.

1-Hydroxyethyl Radical (HER)

Ethanol is a hydroxyl radical scavenger; the product of the interaction of ethanol with hydroxyl radical is HER. Liver microsomes can oxidize ethanol to HER in an NADPH-dependent manner.49 The mechanism involves production of superoxide and H2O2 by cytochrome P450, followed by an iron-catalyzed generation of hydroxyl radical like–oxidants, which interact with ethanol to yield HER.50, 51 Human liver microsomes in the presence of either NADPH or NADH could produce HER from ethanol by an iron-catalyzed process.52 Microsomes isolated from rats treated chronically with ethanol were more reactive in producing HER from ethanol than control microsomes.53 This was due to induction of CYP2E1. HER production from ethanol has been demonstrated in vivo, as a spin-trapped HER adduct was detected in bile from mice or rats treated with ethanol.54, 55 The role of HER adducts in ALD is not known. HER binds readily to proteins to produce ethanol-derived protein adducts, which are immunogenic, and production of antibodies that specifically recognize HER protein adducts was found after chronic ethanol consumption,56 as well as in patients with alcohol-induced cirrhosis.57 HER was shown to interact with chemical antioxidants such as GSH, ascorbate and α-tocopherol and to inactivate antioxidant enzymes.58 Interaction of HER with cellular antioxidants could contribute to mechanisms by which ethanol produces a state of oxidative stress.58 A review on HER production in vitro and in vivo can be found in Reinke, 2002.59


The effects of ethanol on total hepatic GSH levels are variable, with reports of decreases, no effects, or even an increase.60–62 Lowering of mitochondrial GSH by chronic ethanol treatment has been a more consistent observation and appears to be a key lesion contributing to ALD.63, 64 Because liver mitochondria lack catalase, mitochondrial GSH in association with glutathione peroxidase is the major mechanism by which H2O2 is detoxified by mitochondria. Fernandez-Checa and collaborators63–65 have extensively documented that chronic ethanol intake either in the Lieber–DeCarli model or the intragastric infusion model selectively lowers levels of mitochondrial GSH in hepatocytes. A progressive depletion of mitochondrial GSH occurred with ethanol intake that preceded signs of liver injury in ethanol-fed mice.64 Depletion of mitochondrial GSH by chronic ethanol feeding occurs preferentially in pericentral hepatocytes, where most of the liver injury originates.66 This depletion by ethanol is attributable to defective transport of GSH from the cytosol into the mitochondria and can be prevented by fluidization of the mitochondrial membrane by S-adenosylmethionine.66, 67 Lowering of mitochondrial GSH by ethanol has been suggested to sensitize hepatocytes to TNFα-induced cell death, and replenishment of mitochondrial GSH with S-adenosylmethionine protects hepatocytes from alcohol-treated rats to TNF toxicity.67 Bailey et al.,68 however, found that mitochondrial GSH levels were increased after chronic ethanol feeding in the Lieber–De Carli model by approximately 25%. This finding was suggested to reflect an adaptive response to counteract ethanol-related increases in mitochondrial production of ROS. Deaciuc et al.69 reported no change in mitochondrial GSH levels after 7 weeks of ethanol intake. Thus, the effects of ethanol on mitochondrial GSH, as with total GSH, remain controversial. Recent reviews on mitochondrial GSH and the effects of ethanol have been published by Reed70 and by Fernandez-Checa and Kaplowitz.71

Mitochondria, Oxidative Stress, and ALD

Chronic ethanol treatment has long been known to depress mitochondrial function, including oxygen uptake and at high blood alcohol levels, liver ATP concentration is reduced.72, 73 Recent studies have suggested that the ethanol impairment of mitochondrial structure and function may produce an increase in production of ROS and cause cell toxicity. An increase occurs in lipid peroxidation–derived products such as thiobarbituric acid-reactive substances in mitochondria isolated from chronic ethanol-fed rats.74, 75 An increase also occurs in the content of protein carbonyl groups in mitochondrial proteins as compared with cytosolic proteins after ethanol treatment.68 Increased superoxide, H2O2, and hydroxyl radical production were observed in mitochondria from ethanol-fed rats.76, 77 Bailey and Cunningham78 and Bailey et al.79 have proposed that the excess of reducing equivalents generated when ethanol is oxidized by liver alcohol dehydrogenase produce a more reduced electron transfer chain, which will facilitate transfer of an electron to molecular oxygen to produce superoxide. ROS generation will be further elevated after chronic ethanol consumption because of the decreased activity of the respiratory chain, resulting in accumulation of reduced respiratory carriers in complexes I and III. Similar experiments with hepatocytes from chronic ethanol-fed rats showed an enhanced production of ROS by ethanol.80 Increases in ROS production by ethanol metabolism were associated with small decreases in hepatocyte viability and increases in mitochondrial protein carbonyl levels reflecting oxidized protein accumulation.68 Thus, the mitochondria contribute to the increase in oxidant levels in hepatocytes exposed to ethanol acutely or chronically. An exciting advance in this area is the use of mitochondrial proteomics to identify alterations to the mitochondrial proteome in the development of ALD.81

A single oral dose of ethanol had no effect on nuclear DNA integrity of mouse liver, whereas hepatic mitochondrial (mt) DNA was extensively damaged and depleted/degraded.82 This mtDNA depletion was prevented by 4-MP, indicating a role for ethanol metabolism. Ethanol-induced mtDNA depletion was also found in brain, heart, and skeletal muscle.83 Daily ethanol administration for 4 days caused a longer-lasting depletion of mtDNA in mouse liver, perhaps because of an impaired ability to synthesize mtDNA.84 This was associated with an increase in lipid peroxidation and CYP2E1 protein levels. A so-called common 4,977-base pair mtDNA deletion was also found to be more prevalent in patients with alcohol-induced liver disease than in controls.85 Feeding rats the Lieber–Decarli diet caused a significant decrease in mtDNA levels after 4 to 6 months, in association with an increase in 8OHdG levels, indicative of an increase in oxidative modification of mtDNA.86 Despite dramatic changes in mtDNA elicited by ethanol, significant ALD is not observed in these models; therefore, the role of mtDNA deletion and oxidation in the pathophysiology of ALD remains to be determined.

Ishii and colleagues87, 88 showed that incubation of rat liver hepatocytes with 50 mmol/L ethanol increased dichlorofluorescein fluorescence, a measure of ROS, largely H2O2 production.87, 88 This increased fluorescence occurred within only 20 minutes of ethanol incubation. Depletion of GSH led to enhanced fluorescence and loss of hepatocyte viability induced by ethanol. Ethanol caused a decline in the mitochondrial membrane potential (ΔΨm) and triggered opening of the mitochondrial permeability pore. These effects were blocked by an inhibitor of ethanol metabolism (4-MP) and an antioxidant.89 The decline in ΔΨm and the increase in membrane permeability subsequently led to cytochrome c release, activation of caspase 9, and then caspase 3, followed by apoptosis. Interestingly, mitochondria from chronic ethanol-fed rats are more sensitive to the mitochondrial permeability pore induced by calcium and other apoptotic stimuli than control mitochondria.90 Ethanol-induced oxidative stress causes mitochondrial dysfunction, and mitochondrial depolarization and permeability changes are early events in ethanol-induced hepatocyte injury. Recent reviews on ethanol, mitochondria, and ROS production can be found in references91 through94.

Nitrosative Stress

Depending on conditions, nitric oxide either can be hepatoprotective or might potentiate liver injury.95–98 Chronic ethanol administration increases NO production in rat liver.99 Peroxynitrite, derived from the interaction of NO with superoxide, has been suggested to play a role in ethanol-induced hypoxic liver injury.100 Reports also show that the ethanol-induced increase in NO lowers superoxide levels and is therefore protective.101 The role of NO in ALD is not clear. Nanji et al. showed that an inhibitor of nitric oxide synthase increased the severity of ALD in the intra-gastric infusion model, whereas L -arginine supplementation completely prevented the liver injury.102, 103 The authors concluded that NO was protective against ALD, although which isoform of nitric oxide synthase was responsible for the protection was not identified. In contrast to this protective role of NO, inducible nitric oxide synthase was shown to be required for ALD, because ethanol toxicity was significantly blunted in inducible nitric oxide synthase (iNOS) knockout mice.104 A similar protection was found if wild-type mice were treated with a relatively specific iNOS inhibitor, 1400W. However, N(G)-nitro-L-arginine-methyl ester, a more effective inhibitor of endothelial nitric oxide synthase than iNOS, enhanced liver damage.104 The concept that NO produced from endothelial nitric oxide synthase may be protective, whereas NO derived from Kupffer cell iNOS is critical for ALD, was advanced to explain the divergent inhibitor results.104 What regulates this balance or distribution of NO production in ALD is not known. A recent study found that sildenafil was protective against the ethanol-induced hypoxia via delivery of NO to increase blood flow; it also increased the pathology score, thus acting as a two-edged sword.105

Chronic ethanol consumption increased the sensitivity of rat liver mitochondrial oxygen consumption to inhibition by NO.106 State 3 respiration, as well as respiratory control ratios, were depressed by 30% to 40% in mitochondria from the ethanol-treated rats, as was cytochrome oxidase activity. Generation of NO from an NO donor progressively inhibited mitochondrial respiration, and there was a greater inhibition of respiration in mitochondria from the ethanol-fed rats. This greater sensitivity of mitochondria from ethanol-treated rats to NO was suggested to contribute to ethanol-induced hypoxia and to depression of the energy state of the liver.106 In a follow-up study, mitochondria isolated from iNOS knockout mice fed ethanol chronically did not exhibit this increased sensitivity to NO as did mitochondria from wild-type controls.107 Somewhat surprisingly, ethanol feeding did not decrease state 3 mitochondrial respiration in mice as it does in rats; this needs to be further studied because impairment of mitochondrial function is believed to be an important component in alcohol-induced liver injury.

Cellular Repair

The 26S proteosome is important for the catabolism of damaged proteins produced by oxidative stress for which improper removal can result in cell toxicity.108 Chronic ethanol consumption causes protein accumulation in the liver because of a decreased rate of protein catabolism.109, 110 Chronic ethanol consumption in the intragastric infusion model, but not in the oral Lieber–DeCarli model, caused a decrease in chymotrypsin and trypsin-like activities of the proteosome.111 An inverse correlation was found between proteosomal chymotrypsin activity and hepatic lipid peroxidation, suggesting that ethanol-induced oxidative stress can inactivate the proteosome.112 This may contribute to the accumulation of proteins, especially oxidized proteins, in the liver. Intragastric ethanol administration had no effect on proteosome activity in CYP2E1 knockout mice,113 and chlormethiazole, a CYP2E1 inhibitor, prevented the ethanol-mediated decrease in proteosome activity.114 These results suggest that CYP2E1-derived oxidant stress plays a role in the inactivation of the proteosome by ethanol. Another consequence of the lowering of proteosome activity by ethanol would be an elevation of CYP2E1 levels, because CYP2E1 is degraded by the proteosome.115, 116 The ubiquitin-proteosome system and its role in ethanol toxicity has recently been reviewed.117

Oxidative DNA adducts, mutagenic apurinic/apyrimidinic sites, and expression of DNA repair enzymes such as 8-oxoguanine DNA glycosylase/lyase1, endonuclease1, polymeraseβ, and poly (ADP-ribose) polymerase were elevated by ethanol.118 The latter repair enzymes are a sensitive marker for oxidative stress–induced DNA damage.119 Interestingly, no increase in any of these endpoints was observed in ethanol-treated CYP2E1 knockout mice, but it was observed in NADPH oxidase knockout mice.118 The increase in expression of DNA repair enzymes was abolished by treatment with a broad-spectrum P450 inhibitor. The authors indicated that CYP2E1 is required for the induction of oxidative stress to DNA and may play a role in ethanol-associated hepatocarcinogenesis. These studies emphasize the importance of cellular repair enzymes such as the proteosome and DNA repair enzymes in preventing ethanol toxicity and indicate a key role for CYP2E1-derived oxidants in modulating the activity or up-regulation of these cellular repair enzymes.

Ethanol-Induced Protein Adducts

Reactive aldehydes produced from ethanol metabolism and ethanol-induced oxidant stress such as acetaldehyde, malondialdehyde (MDA), and 4-hydroxy-2-nonenal (HNE) can bind to proteins to produce protein adducts. The various types of protein-adducts that can be generated as a result of ethanol consumption have been summarized and characterized by Niemela.120 Acetaldehyde, MDA, and HNE adducts have been found in rats chronically consuming ethanol, in human alcoholics, and in a micropig model of ALD.121–123 Such adducts may produce toxicity because the adducted protein may lose function. Alternatively, the protein adducts may be immunogenic and provoke an immune response.124–127 Tuma128 and others129 have characterized the malondialdehyde-acetaldehyde (MAA) adduct, which results from the ability of acetaldehyde and MDA to increase each other's binding to proteins to produce hybrid adducts. These adducts elicit an immune response and can be found in the circulation of ethanol-fed rats125 and in alcoholics.127 Targeting of MAA-adducted proteins to scavenger receptors on professional antigen-presenting cells appears to be a mechanism by which antibody and T cells invoke an immune response.130In vitro experiments showed that the viability of antigen-presenting cells, lymphocytes, and hepatocytes was decreased on incubation with an MAA hen egg lysosome adduct by necrotic and apoptotic modes of cell death.131 Circulating antibodies against MAA protein adducts were increased in patients with alcohol-induced cirrhosis and hepatitis and correlated with the severity of liver injury.132 Specifically identifying protein adducts in ethanol-fed rat livers is difficult, because the amount of adduct formed involves a low percentage of the total protein. Recent reviews on how immune reactions toward the liver elicited by acetaldehyde, MDA, HNE, and HER protein adducts contribute to the pathogenesis of ALD can be found in Albano133 and Duryee et al.134


Interest in CYP2E1 revolves around the ability of this P450 to metabolize and activate many toxicologically important compounds such as ethanol, carbon tetrachloride, acetaminophen, benzene, halothane, and many other halogenated substrates.135–138 Procarcinogens including nitrosamines and azo compounds are effective substrates for CYP2E1.139 Toxicity by these compounds is enhanced after induction of CYP2E1, for example, by ethanol treatment, and toxicity is reduced by inhibitors of CYP2E1 or in CYP2E1 knockout mice.140 CYP2E1 displays high NADPH oxidase activity because it appears to be poorly coupled with NADPH-cytochrome P450 reductase.141 The increase in ROS production and lipid peroxidation found with microsomes from chronic ethanol-treated rats was blocked by chemical inhibitors of CYP2E1 and anti-CYP2E1 immunoglobulin G.48, 141 CYP2E1 is induced by ethanol, several low-molecular-weight compounds and under a variety of metabolic, nutritional, and pathophysiological conditions.135–138, 142

In the intragastric model of ethanol feeding, large increases in microsomal lipid peroxidation have been observed, and the ethanol-induced liver pathology has been shown to correlate with CYP2E1 levels and elevated lipid peroxidation.10, 13, 15, 113, 143 Experimentally, a decrease in CYP2E1 induction was found to be associated with a reduction in alcohol-induced liver injury. CYP2E1 inhibitors blocked the lipid peroxidation and ameliorated the pathological changes in ethanol-fed rats.114, 144, 145 A CYP2E1 transgenic mouse model was developed that over-expressed CYP2E1. When treated with ethanol, the CYP2E1 over-expressing mice displayed higher transaminase levels and histological features of liver injury compared with control mice.146

Conversely, studies by Thurman and colleagues147, 148 suggested that CYP2E1 may not play a role in alcohol-induced liver injury based on studies with gadolinium chloride or CYP2E1 knockout mice. These issues have been discussed elsewhere.149, 150 Clearly further studies are necessary to resolve the above discrepancies. As summarized in this review, several mechanisms contribute to alcohol-induced liver injury, and ethanol-induced oxidant stress is likely to arise from several sources, including CYP2E1, mitochondria, and activated Kupffer cells. Bradford et al.118 recently reported that, whereas NADPH oxidase but not CYP2E1 was important for ethanol-induced liver injury in their model, CYP2E1 but not NADPH oxidase was critical for ethanol-induced oxidative DNA damage and ethanol-associated hepatocarcinogenesis.

An approach that our laboratory has used to understand basic effects and actions of CYP2E1 is to establish HepG2 cell lines that constitutively express human CYP2E1. We have characterized the toxicity of ethanol, polyunsaturated fatty acids, iron, the effect of GSH depletion, and the production of ROS and development of a state of oxidative stress in these cell lines. Results are summarized in recent reviews.149–151 Damage to mitochondria by CYP2E1-derived oxidants is an early event in the overall pathway of cellular toxicity. Similar findings were obtained with primary hepatocyte cultures after induction of CYP2E1 by treatment with pyrazole. Adaptation to oxidant stimuli is critical for short- and long-term survival of cells exposed to oxidative stress. We found that the levels of GSH and several antioxidant enzymes such as glutathione transferase, catalase, and heme oxygenase-1 were up-regulated in the CYP2E1- expressing cells. This upregulation was prevented by antioxidants, suggesting that ROS generated by CYP2E1 were responsible for the transcriptional activation of these antioxidant genes. Because of this activation of antioxidant genes, the CYP2E1-expressing cells were less sensitive to toxicity by H2O2, menadione, or HNE than control cells. We believe that the upregulation of these antioxidant genes reflect an adaptive mechanism to remove CYP2E1-derived oxidants.

Hepatic stellate cells are central to the fibrotic response of the liver to injury. A co-culture system was developed to evaluate whether mediators (ROS, cytokines) whose production may be elevated by CYP2E1 in liver cells can diffuse to hepatic stellate cells and activate smooth muscle actin and collagen production. Indeed, levels of smooth muscle actin and collagen were increased when hepatic stellate cells were co-cultured with CYP2E1-expressing HepG2 cells compared with control HepG2 cells or HepG2 cells expressing a different P450, CYP3A4. These increases were blocked by catalase or vitamin E, suggesting that the CYP2E1-expressing cells release ROS, such as H2O2 and lipid peroxidation products that stimulate type 1 collagen synthesis by stellate cells.149–151

A working model of CYP2E1-dependent oxidative stress and toxicity is shown in Fig. 3. Ethanol increases levels of CYP2E1, largely by a posttranscriptional mechanism involving enzyme stabilization against degradation. CYP2E1 generates ROS such as O2· − and H2O2 during its catalytic cycle. In the presence of iron, which is increased after ethanol treatment, more powerful oxidants including · OH, ferryl species, and 1-hydroxyethyl radical, are produced. Initially, the liver cells respond to the CYP2E1-related oxidative stress by transcriptionally inducing antioxidant enzymes via their antioxidant response elements. Ultimately, these protective mechanisms are overwhelmed, and the cells become sensitive to the CYP2E1-generated oxidants. These various oxidants can promote toxicity by protein oxidation and enzyme inactivation, oxidative damage to the DNA, and disturbing cell membranes via lipid peroxidation and production of reactive lipid aldehydes, such as MDA and HNE. Mitochondria appear to be among the critical cellular organelles damaged by CYP2E1-derived oxidants. A decrease of mitochondrial membrane potential and perhaps the mitochondrial membrane permeability transition causes release of proapoptotic factors, resulting in apoptosis. A decrease in ATP levels will cause necrosis. Some CYP2E1-derived reactive oxygen species, such as H2O2, LOOH, and HNE, are diffusible and may exit hepatocytes and enter other liver cell types such as stellate cells and stimulate these cells to produce collagen and elicit a fibrotic response. We believe that the linkage between CYP2E1-dependent oxidative stress, stellate cell activation, and mitochondrial injury and GSH homeostasis contribute to the toxic action of ethanol on the liver.

Figure 3.

Working model of CYP2E1-dependent oxidative stress and toxicity.

Future Directions for Research

More detailed information is needed on the mechanisms involved in some of the major proposed pathways by which alcohol produces oxidative stress. Other mechanisms remain highly controversial, such as the role of CYP2E1 or of various cytokines in alcohol-induced oxidative stress. Additional analyses need to determine the role of alcohol metabolism and its byproducts (e.g., acetaldehyde) in the production of ROS. How alcohol-induced oxidative stress is produced in tissues where only limited alcohol metabolism occurs is still unclear.

Many of these issues can be studied using animal models; however, extrapolation of findings from animals to humans will be a difficult task because ROS production and antioxidant status in humans are affected by numerous nutritional, environmental, and drug influences that are difficult to reproduce in animals. Scattered data suggest that the blood of human alcoholics contain lipids modified by radicals and other reactive molecules as well as immune molecules targeted at such modified lipids and proteins.57, 85, 126, 132 Many reports exist that various parameters of oxidative stress, such as ethane exhalation, lipid hydroperoxide levels in serum, 4-HNE protein adducts, salicylate hydroxylation products, and decreases in serum selenium and vitamin E levels, are elevated by alcohol in humans.152–157 Meagher et al.158 showed that alcohol increased urinary isoprostane levels in a time- and dose-dependent manner in volunteers given alcohol acutely.158 These data indicate that ROS and other reactive molecules are indeed formed in human alcoholics.

Other questions that should be addressed in future research include the following:

  • How do reactive nitrogen species (e.g., nitric oxide) play a role in alcohol-induced oxidative stress in addition to ROS? What is the link/balance between ROS and reactive nitrogen species (RNS)?

  • What is the impact of possible interactions between alcohol and environmental influences such as smoking, use of other drugs or medications, and viral infections (e.g., hepatitis C) on ROS production, oxidative stress, and tissue injury? These interactions must be better defined because most alcoholics are exposed to 1 or more of these influences in addition to alcohol.

  • How is oxidative stress affected by interactions between alcohol and nutritional factors, such as the levels and specific types of fats ingested? And how much iron is “safe” in a heavy-drinker?

  • Why does CYP2E1 have a short half-life and what determines its rapid turnover? More mechanistic details on the regulation of CYP2E1 and how ethanol modulates this regulation are needed. A valuable addition would be the development of specific, nontoxic inhibitors of CYP2E1.

  • The effects of ROS produced in the hepatocyte and ROS produced by non-parenchymal cells in the actions of ethanol need to be determined in view of recent results described in Bradford et al.118

  • What are the priming or sensitizing factors for ethanol-induced oxidant stress and cell injury? What are the interactions between TNF, CYP2E1, mitochondrial GSH, iron, electron transfer chain, NO, and acetaldehyde that are responsible for ethanol-induced tissue injury?

  • Can markers predictive of individuals particularly sensitive to ethanol-induced oxidative stress and tissue injury be developed, such as polymorphisms in alcohol dehydrogenase isoforms, CYP2E1, antioxidant enzymes, cytokines?

  • What are the effects, if any, of the sex, age, and race in the ability of ethanol to generate oxidative stress?

  • What are the effects of antioxidants (e.g., vitamin E, vitamin C, or carotenoids) in heavy drinkers? This question is important because some antioxidants can be toxic under certain conditions.

Whether oxidative stress in ALD is clinically relevant based on several human clinical trials remains unclear, and more data in this area are needed. The ability of alcohol to promote oxidative stress and the role of free radicals in alcohol-induced tissue injury clearly are important areas of research in the alcohol field, particularly because such information may be of major therapeutic significance in attempts to prevent or ameliorate alcohol's toxic effects, for example, by antioxidants, iron chelators, inhibitors of CYP2E1 or of cytokine production/actions or GSH replenishment. As basic information continues to emerge regarding the role of oxidative stress in disease development and the mechanisms underlying ROS-related cellular toxicity, these findings will lead to more rational antioxidant therapeutic approaches. Moreover, these findings could result in the development of more effective and selective new medications capable of blocking the actions of ROS and consequently the toxic effects of alcohol.