- Email: [email protected]
Hepatotoxicity induced by iron overload and alcohol
Journal of Hepatology 1996; 25:538-546 Printed in Denmark •All rights reserved Munksgaard. Copenhagen
Copyright © EuropeanAssociation for the Study of the Liver 1996 Journal of Hepatology
ISSN 0168-8278
Hepatotoxicity induced by iron overload and alcohol Studies on the role o f chelatable iron, c y t o c h r o m e P450 2 E l a n d lipid p e r o x i d a t i o n Per St~l l, Inger Johansson2, Magnus Ingelman-Sundberg2, Karin Hagen3 and Rolf Hultcrantz3 1Department of Gastroenterology and Hepatology, Karolinska lnstitutet, Huddinge University Hospital, 2Department of Medical Biochemistry and Biophysics, Karolinska Institutet, and 3Department of Medicine, Karolinska Instituter, Karolinska Hospital, Stockholm, Sweden
Background~Aims: Clinical experience and studies with experimental animal models indicate a synergistic hepatotoxic effect of dietary iron overload and chronic alcohol ingestion. In order to elucidate the mechanism underlying this synergism, we examined the hepatic levels of ethanol-inducible cytochrome P450 2E1, glutathione and malondialdehyde, and the effect of iron chelation with desferrioxamine, in livers from rats treated with iron and/or ethanol. Methods: Animals received diets with or without 2.5-3% carbonyi iron for 6-9 weeks, followed by an ethanol-containing diet or a liquid control diet for 5-9 weeks. Desferrioxamine was administered subcutaneously with mini-osmotic pumps. Aianine aminotransferase activity in serum and hepatic contents of glutathione and malondialdehyde were determined. The hepatic level of cytochrome P450 2E1 was determined with Western Blotting using a specific polyclonal antibody. Results: The combination of iron and alcohol led to a marked increase in serum alanine aminotransferase activity as compared with all other treatment groups, and iron chelation with desferrioxamine reversed these increases. Treatment with al-
LINICALEXPERIENCEindicates a possible synergistic hepatotoxic effect when iron overload is combined with overconsumption of alcohol (1). We have previously demonstrated that rats fed carbonyl iron followed by a liquid ethanol-containing diet develop elevated levels of aminotransferases in their plasma
C
Received 8 June; revised 6 December; accepted 19 December 1995
Correspondence: Per StY, MD, Department of Gastroenterology and Hepatology, K 63, Karolinska Institutet, Huddinge Hospital, S-141 86 Huddinge, Sweden. Tel: +46-8746 22 63, Fax: +46-8-608 22 41. 538
cohol alone led to slightly increased aminotransferases compared with controls. The level of cytochrome P450 2E1 was significantly elevated in microsomes isolated from ethanol-treated rats, but neither additional iron supplementation nor desferrioxamine influenced this level significantly. Glutathione contents were increased in the livers of animals treated with iron and/or ethanol. Malondialdehyde values were increased in iron-treated animals, whereas neither ethanol nor desferrioxamine altered malondialdehyde levels significantly. Conclusions: The toxic effects exerted by the combination of iron overload and chronic ethanol feeding on rat liver are dependent on a pool of chelatable iron. The hepatic level of cytochrome P450 2El is markedly induced by ethanol but not further altered by iron overload. Neither increased lipid peroxidation nor depletion of hepatic glutathione levels can explain the synergistic hepatotoxic effects of iron and ethanol in this model.
Key words: Alcohol; Cytochrome P450 2E1; Desferrioxamine; Hemochromatosis; Hepatotoxicity; Iron.
and ultrastructural evidence of necrosis in the liver, alterations which are not seen in animals fed iron or alcohol alone (2). Lipid peroxidation of polyunsaturated fatty acids in membranes has been implicated as a mechanism by which iron, as well as ethanol, causes liver damage (3-7). Iron ions catalyze the formation of free species of reactive oxygen, which initiate the peroxidation process (8). Iron-induced peroxidation of intracellular membranes may lead to cellular dysfunction and eventually sideronecrosis (9-11). Alcohol ingestion by animals causes generation of
Hepatotoxicity by iron and alcohol
reactive oxyradicals and initiation of lipid peroxidation (12), probably via the induction of cytochrome P450 2El (CYP2E1) (5,13). This enzyme can generate free radicals, and initiate lipid peroxidation in microsomes (13). Livers from animals chronically exposed to ethanol, and thus having induced levels of CYP2E 1, display increased susceptibility to peroxidative damage caused by carbon tetrachloride, ferric citrate or ethanol itself (14,15). In addition, experimental evidence suggests that metabolism of acetaldehyde by aldehyde oxidase or xanthine oxidase may generate free radicals (16), or superoxide, which in turn mobilizes iron from ferritin (17). Thus, the increased hepatic iron stores seen in alcoholic liver disease may potentiate alcohol-induced hepatotoxicity (18). Hence, experimental evidence suggests the existence of a synergism between iron and alcohol with respect to increased free radical formation and lipid Experiment
A
t Weeks 7 18 19 I1° 111112113114115 I
11213141516
Control
I
Ethanol
I
I
I
I
I
[
I
I
ethanol,
I I ~h~ol
J
I
I
I
I
, ,
Experiment Control
I
I
I
I
I
I
B
t
Weeks I 11 z I 3 14 I s I 6 17 18 19 11°1111121131TM
I
8hanoi
IIII
Iron
Iron+
I
Eth..o,
E t h a n o l l l l l l l l l l l l l l l
ethanol
I
I I I I I I I I
Iron
Iron+
I
I
!
1
I
I
I
I
]
l
I
I
I
~ha.o.
peroxidation. The purpose of the present study was to investigate whether hepatotoxicity induced by a combination of iron and alcohol in vivo is mediated by a chelatable, low-molecular-weight iron pool, induction of CYP2E1, depletion of glutathione and/or increased hepatic levels of malondialdehyde (MDA). Materials
and
Methods
Animals
Male Sprague-Dawley rats were fed a rat chow diet (Altromin 1314, Altromin International, Lage, Germany) supplemented with 2.5 or 3% carbonyl iron (Sigma Chemicals, St Louis, MO, USA) (19) for 6-9 weeks after weaning, as illustrated in Fig. 1. Control groups were fed the same diet, but without iron supplementation. Following the period of iron supplementation, animals were pair-fed a liquid ethanolcontaining or non-alcoholic diet (Bioserv, New Jersey, USA), according to Lieber & DeCarli (20). The ethanol-containing and liquid control diets were identical, except that dextrin-maltose was replaced by isocaloric amounts of ethanol (65 g ethanol/l). The mean animal weights in the groups did not differ significantly when feeding with the liquid diet was started. Blood ethanol concentrations at sacrifice were similar, as was previously reported in ethanol-treated animals (2), without any significant difference between groups treated with or without iron and/or desferrioxamine. Animals who received the non-alcoholic diet did not have detectable amounts of ethanol in the blood. During the alcohol feeding period, animals were housed in separate cages and pair-fed individually. They were fed at noon every day and not given access to any food or drinking water other than that in the diet. The food was prepared fresh every other day. Simultaneous treatment with desferrioxamine did not alter the consumption of the ethanol-containing diet. Similarly, iron-treated and control animals consumed similar amounts of the liquid diet. Animals were maintained on a 12-h light-dark cycle.
I
Iron+ ! I ! I I I 1 I I I ~-~ ethanol + Etna.o, / DFO Fig. 1. Schematic presentation of the treated groups in Experiment A and B. Treatments as indicated. Five animals were included in each treated group. Experiment A: 3% carbonyl iron for 6 weeks + alcohol for 9 weeks. (Experiment A was also repeated with 3% carbonyl iron for 6 weeks + alcohol for 6 weeks). Experiment B: 2.5% carbonyl iron for 9 weeks + alcohol for 5 weeks; one of these groups also received DFO (desferrioxamine; 150 mg/kg/ day)for i week.
Experimental design
Two different experimental designs, designated A and B, were employed, as depicted in Fig. 1. In Experiment A, animals were divided into four different groups with five animals in each. The first group served as controls, receiving a normal chow diet for 6 weeks, followed by the liquid control diet for 9 weeks. Animals in the second group received the normal rat chow followed by an ethanol-containing liquid diet. In the third group, animals were fed a diet containing 3% carbonyl iron, followed by the liquid control diet without alcohol. In the fourth group ani539
P. St&l et al.
mals received the carbonyl iron-containing diet, followed by the ethanol liquid diet. Experiment A was repeated once, but with the modification that the alcohol-feeding period was shortened from 9 to 6 weeks. Animal weights at sacrifice did not differ significantly between treatment groups (control animals weighed 408+9 g, ethanol-treated 393+9 g, iron-treated 392+38 g, and iron + ethanol-treated 386+13 g). In Experiment B, 2.5% carbonyl iron was administered for 9 weeks, followed by an ethanol-containing diet for 5 weeks. In this experiment, the first four groups of treated animals were identical to those of Experiment A, but an additional group receiving iron, alcohol and subcutaneous desferrioxamine (Desferal ®, Ciba-Geigy, Switzerland) was also employed. The animals in group four, receiving iron + alcohol, were implanted with mini-osmotic pumps containing 0.9% saline. Animal weights did not differ between treatment groups at sacrifice (control animals weighed 391+28 g, ethanol-treated 375+39 g, irontreated 370+28 g, iron + ethanol-treated 366+26 g, and those treated with iron + ethanol + desferrioxamine 365+17 g).
Chemicals and procedures Desferrioxamine was administered by mini-osmotic pumps (Alzet 2ML1) implanted subcutaneously on the back of the animal. Desferrioxamine was dissolved in sterile water at a concentration of 250 mg/ ml and the mini-osmotic pump (release of 10 ~tl/h) was filled with 2 ml of this solution. The substance was released at a rate of 150-180 mg/kg per day until sacrifice (animal weight 333-400 g). Animals were killed after the alcohol feeding period. They were anesthetized with ether, the abdomen cut open and blood drawn from the aorta for serum alanine aminotransferase (ALT) assays. Slices of liver tissue were fixed for light microscopic examination. Liver tissue was also immediately frozen in liquid nitrogen for determination of iron, protein, total hepatic malondialdehyde and glutathione levels, CYP2E1, MDA in liver microsomes and NADPHdependent lipid peroxidation. Light microscopy Liver tissue was fixed in formaldehyde and embedded in paraffin, after which sections were stained with hematoxylin-eosin, reticulin and Perls' blue, the latter to visualize hepatic iron deposits. Chemical and immunochemical determinations The iron contents of liver tissue samples were measured in an atomic absorption spectrophotometer (Vat540
ian SpectrAA 400) using background correction. The samples were hydrolyzed in 3.5% HNO 3 overnight before analysis (21). Protein was determined according to the method of Lowry et al. (22). For determination of CYP2E1 levels, liver tissue was homogenized in 10 mM sodium/potassium phosphate buffer (pH 7.4) containing 1.14% KC1 and the homogenate subfractionated by centrifugation and the microsomal fraction washed once before resuspension in 50 mM potassium phosphate buffer (pH 7.4) (5). Rat liver CYP2E1 was then quantitated by Westernblotting according to a procedure described previously (5), using specific rabbit polyclonal anti-CYP2E1 IgG characterized earlier (23). MDA levels and NADPHdependent lipid peroxidation in liver microsomes were determined as described previously (5,13). For determination of total hepatic malondialdehyde (MDA), slices of liver tissue were frozen in liquid nitrogen and stored at -70°C. The production of malondialdehyde (MDA) was assayed as thiobarbituric acid reactive products (TBAR), measured according to a method described by Ohkawa et al. (24). In short, the liver tissue sample weighing approximately 0.3 g was homogenized in 0.15 M potassium phosphate buffer (pH 7.2) to a concentration of 15%, and kept on ice or at 4°C. Twenty-five mM desferrioxamine was included in the buffer (Ciba-Geigy, Switzerland) to avoid TBAR formation in the sample. Two hundred and fifty microliters of the homogenates was pipetted into a glass tube, and 500 p.1 of 20% trichloroacetic acid and 40 I.tl of 2% 2,6-ditertbutyl-4-methylphenol was added. Two hundred milligrams of thiobarbituric acid (TBA) was dissolved in 30 ml distilled water (to a concentration of 6.7%) at 70°C. One milliliter of the 6.7% TBA solution was added and the homogenates were boiled at 100°C for 15 min and allowed to cool to room temperature. Following centrifugation at 3000 G (15°C) for 10 min, 1 ml of the supernatant was read spectrophotometrically (UV 160, Shimadzu, Kyoto, Japan) at 535 nm using appropriate blanks and standards. As MDA standard, 1,1,3,3-tetraethoxypropane (Sigma T9889) was mixed with buffer at different concentrations in the interval 0.1-20 I.tM. Glutathione was measured using the GSH-400 assay (Bioxytech S.A. Bonneuil/Marne, Cedex, France). In this procedure, the first step involves formation of thioethers between a chromogenic reagent and all mercaptans in the sample, whereas the second step transforms these substitution products into a chromophoric thione with maximal absorbance at a wavelength of 400 nm. Ten per cent (w/v) liver homogenates in distilled water were prepared, sup-
Hepatotoxicity by iron and alcohol
plemented with 5% (w/v) metaphosphoric acid (in order to precipitate high-molecular weight components) and centrifuged at 2500 g (10 min at 4°C). The assay was performed in duplicate on the supernatants obtained. A standard curve was prepared using 25200 ~tM glutathione (reduced form, 98-100%, Sigma Chemicals, St Louis, MO, USA) dissolved in 5% metaphosphoric acid and the glutathione contents of the samples were determined using linear regression. Statistical methods
The values for the different groups were compared using analysis of the variance (ANOVA). When two treatment groups were compared, Student's unpaired t-test (2-tail) was employed.
a 450 4O0 350
3O0 25O ,,¢
200 "k'k
150 100 50 0
_ w
•
a-w
Control
Ethanol
Iron
-= Iron+
ethanol
Group
b 140
100 80
60 40 20 0
This study was approved by the Stockholm North Local Committee for Ethical Review of Animal Experiments.
Results Serum alanine aminotransferase (ALT)
Fig. 2a shows the ALT activities in the serum of all animals in Experiment A, including the repeated experiment in which animals were treated with ethanol for 6 weeks. There was no statistically significant difference in ALT activities for animals fed ethanol for 6 or 9 weeks. Ethanol-feeding alone led to significantly increased ALT values as compared with controis (33.7+6.1 vs. 16.1+4.7 U/I). Also, iron feeding alone led to significantly increased ALT activities compared with controls (36.0-&-_19.2 U/I). ALT activities were markedly increased in animals receiving both iron and ethanol compared with all other groups (140.9+115.6 U/l). Figure 2b presents the ALT activities for the animals in experiment B. Here, treatment with ethanol alone led to significantly increased ALT activities as compared with controls (23.0+3.8 vs. 9.0-L-_3.0 U/I), whereas treatment with iron alone did not. As in experiment A, a marked increase in ALT was seen after treatment with iron plus ethanol (56.6+38.8 U/ 1), whereas treatment with desferrioxamine reversed this increase in serum ALT activity for animals exposed to both iron and ethanol (to ALT levels of 16.0+_9.7 U/I). Experiment A: Hepatic contents o f iron, MDA, glutathione and CYP2E1 (Table 1)
120
h-
Ethical approval
Control
Ethanol
Iron
Group
Iron+ ethanol
Iron+ ethanol+ DFO
Fig. 2. a) The activities of serum alanine aminotransferases (ALT) in the treated groups in Experiment A (since experiment A was repeated, each treatment group includes 5 + 5 animals), b) The activities of serum alanine aminotransferases (ALT) in the treated groups in Experiment B. Each dot represents one animal and a horizontal line depicts the mean of each group. * = p
The hepatic contents of iron, MDA, glutathione and CYP2E1 for the different groups in Experiment A are shown in Table 1. Values were similar in the repeated experiment A in which animals were treated with ethanol for 6 instead of 9 weeks (data not shown). The hepatic iron concentration was 0.93 ~tg/mg protein in the control animals fed a diet without iron supplementation,, while animals fed ethanol demonstrated only about 30% of this level. As expected, dietary iron supplementation led to a marked and significant increase in the content of hepatic iron for all groups. Animals fed the combination of iron and ethanol had hepatic levels of iron contents similar to those fed iron alone. The hepatic concentrations of MDA were significantly increased in iron-treated animals, whereas supplementation of ethanol did not further increase these levels. Treatment with iron and/or ethanol led to sig541
P. Stdl et al.
hepatic microsomes were significantly increased in iron-treated animals as compared with animals receiving ethanol only or a control diet. Microsomal NADPH-dependent lipid peroxidation did not differ between the treatment groups.
nificant increases in hepatic glutathione concentrations, compared with controls. The levels of CYP2E1 in liver microsomes are shown in Table 1. Ethanol treatment caused significant increases in the levels of CYP2E1 compared with controls or iron-treated animals. Iron overload did not result in any significant alteration in the amount of immunodetectable CYP2E1. In ethanoltreated animals, simultaneous iron overload did not alter the levels of CYP2E1. MDA levels of isolated
Experiment B: Hepatic contents of iron, MDA and CYP2E1 (Table 2) The hepatic contents of iron, MDA and CYP2E1 of the different groups in Experiment B are presented in
TABLE 1 Hepatic contents of iron, malondialdehyde (MDA) and glutathione, and levels of ethanol-inducible cytochrome P450 2El in the livers of the different groups (Experiment A) Control (n=5)
Alcohol (n=5)
Iron (n=5)
Iron+alcohol (n=5)
Liver iron (~mol/mg protein)
0.93_+0.10
0.33_+0.152
4.32_+0.921
4.06+1.001
MDA (nmol/g liver wet weight)
18.7+2.5
16.7+1.6
37.7+4.71
35.4_+1.31
Glutathione (~mol/g liver wet weight)
3.2_+0.6
4.6_+0.82
4.4_+0.62
4.3_+0.62
Cytochrome P450 2E 1 (Arbritrary units)
100-+39
512+893
87_+33
450-2_1463
Microsomal MDA (nmol/mg microsomal protein)
1.02_+0.03
0.88_+0.05
1.4-+0.321
1.26_+0.124
NADPH-dependent lipid peroxidation (nmol/mg microsomal protein)/min
0.14_+0.03
0.17_+0.06
0.14_+0.02
0.13_+0.04
Results expressed as means + S.D. MDA = malondialdehyde (measured as TBA-reactive substances). 1 = p < 0.05 compared with controls or animals treated with alcohol alone. 2 = p < 0.05 compared with controls. 3 = p < 0.05 compared with controls or animals treated with iron alone. 4 = p < 0.05 compared with animals treated with alcohol alone.
TABLE 2 Hepatic contents of iron and malondialdehyde (MDA), and levels of ethanol-inducible cytochrome P450 2El in the livers of the different groups (Experiment B). Control (n=5)
Alcohol (n=5)
Iron (n=5)
Iron+alcohol (n=5)
Iron+alcohol +DFO
Liver iron (~tmol/mg protein)
0.62_+0.17
0.38-+0.03
3.44--1.07 x
2.94-+0.181
2.43-+0.64 l
MDA (nmol/g liver wet weight)
12.9+2.3
11.4-+1.4
25.1+3.31
22.5+2.61
23.4-+5.71
Cytochrome P450 2E 1
100-2_21
1050-2-4753,4
524+-2773,4
344-+2632
(n=5)
(Arbritrary units) Results expressed as means + S.D. DFO = desferrioxamine. M D A = malondialdehyde. 1 = p < 0.05 compared with controls or animals treated with alcohol alone. 2 = p < 0.05 compared with controls. 3 = p < 0.01 compared with controls. 4 = p < 0.05 compared with animals treated with iron alone.
542
142-+302
Hepatotoxicity by iron and alcohol
Table 2. Treatment with desferrioxamine did not result in significantly decreased hepatic iron concentrations. Hepatic concentrations of MDA were significantly increased in iron-treated animals, but neither alcohol nor desferrioxamine further altered MDA levels. There was a slight increase in the hepatic levels of CYP2E1 in the iron-overloaded animals as compared with controls. As in Experiment A, adding iron to ethanol-treated animals did not alter the CYP2E1 levels. Desferrioxamine had no significant effect on CYP 2El levels in animals treated with iron plus ethanol. In animals treated with ethanol alone, there was a significant correlation between ALT activities and hepatic levels of CYP2E1 (r=0.92, p<0.05). No such correlation was found in animals treated with iron plus ethanol, or iron plus ethanol and desferrioxamine.
Light microscopy The findings from the light microscopic examinations were similar to those described previously (3). Liver tissue from animals fed ethanol displayed fat droplets in the central areas of the lobule. Perls' blue stain showed a positive granular reaction in periportal hepatocytes in animals fed carbonyl iron, but not in the other groups. In addition, occasional iron-containing Kupffer cells were present in the former livers. In animals receiving both iron and ethanol occasional inflammatory cells, as well as lipid droplets, granularity in hepatocytes and iron-containing Kupffer cells, were seen. The reticulin stain did not show any features of increased fibrosis in any of the livers. The cellular and lobular iron distributions were similar in animals exposed to desferrioxamine for 1 week as compared to the other groups treated with iron.
Discussion The aim of the present study was to elucidate the mechanisms underlying the synergistic hepatotoxicity exerted by a combination of iron overload and ethanol. We demonstrated that this synergistic hepatotoxicity is dependent on chelatable iron, since desferrioxamine added to the treatment with iron and ethanol decreases ALT activities in serum, although the total hepatic iron contents and MDA levels are unchanged. The exact mechanisms by which chelatable iron is involved in the hepatotoxic events in this model are unclear. In iron overload, non-transferrin-bound iron is markedly increased in serum and taken up by the hepatocytes, but less efficiently excreted into the bile
as compared with the iron release from control livers (25). It has been demonstrated that the pool of chelatable iron is enlarged in animals exposed to dietary carbonyl iron overload, although this enlargement is not as marked as the increase in total hepatic iron content (26,27). In addition, the pool of low-molecular-weight iron has been reported to increase during ethanol metabolism (28). Hypothetically, an increased pool of catalytically active iron ions could enhance alcohol-induced lipid peroxidation of cellular membranes and the leakage of aminotransferases into plasma. In the present study, iron alone increased lipid peroxidation significantly. We had no signs of increased lipid peroxidation in animals treated with ethanol alone, and adding ethanol to iron-treated animals did not further potentiate the formation of lipid peroxides (measured as TBAR formation) in the liver. Chelation of iron with desferrioxamine did not decrease the hepatic contents of malondialdehyde in animals exposed to the combination of iron and alcohol, although the aminotransferases were normalized. Hence, in this animal model we have no evidence that hepatotoxicity induced by iron and alcohol involves enhanced lipid peroxidation. In the present study, ethanol was given in the diet as described by Lieber & deCarli (20), in which ethanol-treated animals are fed ad libitum and the corresponding controls are given a similar amount of the iso-caloric, non-alcoholic diet. Using the LieberdeCarli ethanol-containing diet in combination with iron overload, only moderate liver injury such as steatosis and mild necrosis is produced (2), even if the duration of the ethanol-feeding period is prolonged, as in a study by Olynyk et al. (29). Butcher et al. used electron paramagnetic resonance spectroscopy on livers from rats subjected to a similar treatment, and failed to detect increased levels of stable free radicals in these livers (30). Others have demonstrated results similar to those of the present study, with unchanged levels of lipid peroxides of livers exposed to iron overload plus ethanol (31). In spite of this, lipid peroxidation has been implicated as a major mechanism in alcohol-related liver injury (5-7,12-17). In the present study, feeding with ethanol alone led to a slight but significant increase in plasma ALT activities, although MDA levels were unchanged. Sixteen weeks of intragastric feeding with ethanol plus a high-fat diet was shown to produce increased levels of malondialdehyde, 4-hydroxynonenal and signs of hepatic fibrosis, whereas 5 weeks with the same treatment led to elevated aminotransferases in plasma, but unaltered hepatic levels of malondialdehyde and 4-hydroxynonenal (12).
543
P. St&l et al.
Thus, ethanol infused intragastrically for prolonged time periods results in more severe signs of liver damage (including induced fibrogenesis) as compared with the Lieber-deCarli diet (2,12,29,32). Tsukamoto et al. further demonstrated that small amounts of iron (not resulting in iron overload) added to the intragastric ethanol plus high-fat diet enhanced lipid peroxidation and induced fibrogenesis (32). Similarly, iron potentiated fibrogenesis when the fibrotic stimulus was low-dose carbon tetrachloride plus alcohol (33). These results indicate that in the model used in the present study and the study by Olynyk et al. (29), iron overload combined with ethanol according to the Lieber-de Carli protocol causes increased aminotransferase levels and mild necrosis but neither enhanced lipid peroxidation nor fibrosis, as compared with iron overload alone. In the different model used by Tsukamoto et al., ethanol and a highfat diet given intragastrically have fibrogenic properties by itself, and iron in small amounts added to that protocol potentiates lipid peroxidation and fibrogenesis (32). Thus, a pool of chelatable low-molecularweight iron possibly interacts with other liver-cell damaging agents such as alcohol (32) or carbon tetrachloride (33), which is consistent with the decrease in aminotransferases seen if desferrioxamine is administered to animals treated with iron plus ethanol in the present study. Another possible interaction of iron in ethanolinduced liver damage could be iron-induced effects on the ethanol-inducible cytochrome P450 (CYP2E1). CYP2E1 may constitute the major source of superoxide anion production, leading to the formation of hydroxyl radicals via an iron-catalyzed HaberWeiss reaction (13,34), indicating a potential role for iron in CYP2El-induced free radical production (13, 16,34). In the present study, iron overload alone slightly increased the hepatic levels of CYP2E1 compared with controls (in experiment B). In contrast, in ethanol-treated animals CYP2E1 levels were neither significantly altered by iron nor by desferrioxamine. Hence, increased CYP2E1 levels were not the reason for the potentiating effects of iron on ethanol-induced liver damage in this model. However, we cannot rule out the importance of this enzyme in the pathogenesis of ethanol-induced hepatotoxicity, since hepatic levels of CYP2E1 correlated significantly to ALT values in animals treated with ethanol alone. Another mechanism by which iron and ethanol might exert hepatotoxic effects would be depletion of antioxidant defense systems. In the present study depletion of total hepatic glutathione stores could be excluded as a possible mechanism leading to hepato544
cellular injury. The reported effect of iron overload on hepatic glutathione levels has been variable, i.e. an increase (35), decrease (36) or unaltered levels (4) as compared with controls. Alcohol was found to depress hepatic glutathione in patients with chronic liver disease (7). Hepatic glutathione levels following chronic exposure to ethanol in animals are variable. Three weeks of ethanol-feeding increased hepatic glutathione stores (37), and glutathione turnover and hepatic glutathione contents were similarly increased following 8 weeks of ethanol treatment in rats (38). These results are thus in line with those of the present study. However, 16 weeks of intragastric infusion with ethanol plus a high-fat diet resulted in depletion of glutathione, whereas 5 weeks with the same treatment did not (12), indicating that a more aggressive ethanol exposure may lead to depletion of the total hepatic glutathione stores. Long-term feeding with ethanol selectively reduces the mitochondrial glutathione pool at first, and this reduction may escape detection when measuring the total glutathione content, since the cytosolic pool remains unchanged (39,40). Iron added to an ethanol-high-fat diet was not found to deplete hepatic glutathione (32). We conclude that hepatotoxicity induced by a combination of iron overload and chronic ethanol treatment in this animal model depends on a pool of chelatable iron, but does not involve any detectable increase in lipid peroxidation or glutathione depletion. Hepatic levels of CYP2E1 are markedly increased by ethanol, but not further altered by iron overload. Although previous evidence speaks in favor of a free-radical mediated process, the exact mechanisms by which iron overload work in concert with chronic ethanol exposition to induce hepatocellular damage remains to be elucidated.
Acknowledgements The authors are grateful to Mrs. Kristina Eckes for skillful technical assistance. This study was supported by grants from the Swedish Medical Research Council, the Nanna Swartz Foundation, the Swedish Society of Medicine (Bengt Ihres Fund), and Rut and Richard Juhlins Foundation.
References 1. LeSage GD, Baldus WP, Fairbanks VF, Baggenstoss AH, McCall JT, Moore SB, Taswell HF, Gordon H. Hemochromatosis: genetic or alcohol-induced? Gastroenterology 1983; 84: 1471-7. 2. Sthl P, Hultcrantz R. Iron increases ethanol toxicity in rat liver. J Hepatol 1993; 17: 108-15. 3. Dresow B, Albert C, Zimmerman I, Nielsen P. Ethane exhalation and vitamin E/ubiquinol status as markers of lipid per-
Hepatotoxicity by iron and alcohol
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20. 21.
oxidation in ferrocene iron-loaded rats. Hepatology 1995; 21: 1099-105. Fletcher LM, Roberts FD, Irving MG, Powell LW, Halliday JW. Effects of iron loading on free radical scavenging enzymes and lipid peroxidation in rat liver. Gastroenterology 1989; 97: 1011-8. Ekstr/Sm G, Ingelman-Sundberg M. Rat liver microsomal NADPH-supported oxidase activity and lipid peroxidation dependent on ethanol-inducible cytochrome P-450IIE1. Biochem Pharmacol 1989; 38: 1313-9. Situnayake RD, Crump B J, Thurnham DI, Davies JA, Gearty J, Davis M. Lipid peroxidation and hepatic antioxidants in alcoholic liver disease. Gut 1990; 31:1311-7. Shaw S, Rubin KP, Lieber CS. Depressed hepatic glutathione and increased diene conjugates in alcoholic liver disease. Evidence of lipid peroxidation. Dig Dis Sci 1983; 28: 585-9. Bacon BR, Britton RS. The pathology of hepatic iron overload: a free radical mediated process? Hepatology 1990; 11: 127-37. Myers BM, Prendergast FG, Holman R, Kuntz SM, LaRusso NE Alterations in the structure, physicochemical properties, and pH of hepatocyte lysosomes in experimental iron overload. J Clin Invest 1991; 88: 1207-15. Bacon BR, Healey JF, Brittenham GM, Park CH, Nunnari J, Tavill AS, Bonkovsky HL. Hepatic microsomal function in rats with chronic dietary overload. Gastroenterology 1986; 90: 1844-53. St~ P, Hultcrantz R, Glaumann H. Liver cell damage and lysosomal iron storage in patients with idiopathic hemochromatosis. A light and electron microscopic study. J Hepatol 1990; 11: 172-80. Kamimura S, Gaal K, Britton RS, Bacon BR, Triadafilopoulos G, Tsukamoto H. Increased 4-hydroxynonenal levels in experimental alcoholic liver disease: association of lipid peroxidation with liver fibrogenesis. Hepatology 1992; 16: 448-53. Ingelman-Sundberg M, Johansson I. Mechanisms of hydroxyl radical formation and ethanol oxidation by ethanol-inducible and other forms of rabbit liver microsomal cytochromes P-450. J Biol Chem 1984, 259; 6447-58. Castillo T, Koop DR, Kamimura S, Triadafilopoulos G, Tsukamoto H. Role of cytochrome P-450 2El in ethanol-, carbon tetrachloride- and iron-dependent microsomal lipid peroxidation. Hepatology 1992; 16: 992--6. Shaw S, Jayatilleke E, Ross WA, Gordon E, Lieber CS. Ethanol-induced lipid peroxidation: potentiation by long-term alcohol feeding and attenuation by methionine. J Lab Clin Med 1981; 98: 417-25. Shaw S, Jayatilleke E. The role of aldehyde oxidase in ethanol-induced hepatic lipid peroxidation in the rat. Biochem J 1990; 268: 579-83. Shaw S. Lipid peroxidation, iron mobilization and radical generation induced by alcohol. Free Rad Biol Med 1989; 7: 541---47. Irving MG, Halliday JW, Powell LW. Association between alcoholism and increased hepatic iron stores. Alcohol Clin Exp Res 1988; 12: 7-13. Park CH, Bacon BR, Brittenham GM, Tavill AS. Pathology of dietary carbonyl iron overload in rats. Lab Invest 1987; 57: 555-63. Lieber CS, DeCarli LM. The feeding of ethanol in liquid diets. Alcohol Clin Exp Res 1986; 10:550-3 Hultcrantz R, Glaumann H. Studies on the rat liver following
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
iron overload. Biochemical studies after iron removal. Lab Invest 1982; 46: 383-93. Lowry OH, Rosebrough N J, Fan" AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Biol Chem 1951; 193: 265-75. Johansson I, Ingelman-Sundberg M. Benzene metabolism by ethanol-, acetone-, and benzene-inducible cytochrome P-450 (IIE1) in rat and rabbit liver microsomes. Cancer Res 1988; 48: 5387-90. Ohkawa H, Ohishi N, Yagi K. Assay for lipid peroxides in animal tissues by thiobarbituric acid reaction. Anal Biochem 1979; 95: 351. Brissot P, Zanninelli G, Guyader D, Zeind J, Gollan J. Biliary excretion of plasma non-transferrin-bound iron in rats: pathogenetic importance in iron-overload disorders. Am J Physiol 1994; 267: G135--42. Britton RS, Ferrali M, Magiera CJ, Recknagel RO, Bacon BR. Increased prooxidant action of hepatic cytosolic lowmolecular-weight iron in experimental iron overload. Hepatology 1990; 11: 1038--43. Nielsen P, Diillmann J, Wulfhekel U, Heinrich HC. Nontransferrin-bound-iron in serum and low-molecular-weightiron in the liver of dietary iron-loaded rats. Int J Biochem 1993; 25: 223-32. Sergent O, Morel I, Cogrel P, Lescoat G, Pasdeloup N, Brissot P, Cillard P, Cillard J. Cellular pool of low molecular weight iron during metabolism of ethanol in rat hepatocytes. (Abstract) Hepatology 1993; 18: 276A. Olynyk J, Hall P, Reed W, Williams P, Kerr R, Mackinnon M. A long-term study of the interaction between iron and alcohol in an animal model of iron overload. J Hepatol 1995; 22: 671--6. Butcher GP, Deighton N, Batt RM, Ding B, Haywood S, Hoffman J, Jackson MJ, Symons MCR, Rhodes JM. Electron paramagnetic resonance spectroscopy of stable free radicals in the liver compared with ultrastructural and functional damage in a rat model of alcohol- and iron-overload. Clin Sci 1993; 84: 339--48. Tector AJ, Britton RS, O'Neill R, Bacon BR. Additive effect of iron overload and ethanol on hepatic damage. (Abstract). Gastroentrology 1993; 102: A944. Tsukamoto H, Home W, Kamimura S, Niemel~i O, Parkkila S, Yl~i-Herttuala S, Brittenham GM. Experimental liver cirrhosis induced by alcohol and iron. J Clin Invest 1995; 96: 620-30. Mackinnon M, Clayton C, Plummer J, Ahem M, Cmielewski P, Ilsley A, Hall P. Iron overload facilitates hepatic fibrosis in the rat alcohol/low-dose carbon tetrachloride model. Hepatology 1995; 21: 1083-8. Nordmann R, Ribi~re C, Rouach H. Involvement of iron and iron-catalysed free radical production in ethanol metabolism and toxicity. Enzyme 1987; 37: 57--69. Kawabata T, Ogino T, Awai M. Protective effects of glutathione against lipid peroxidation in chronically iron-loaded mice. Biochim Biophys Acta 1989; 1004: 89-94. Bonkovsky HL, Healey JF, Lincoln B, Bacon BR, Bishop DF, Elder GH. Hepatic heme synthesis in a new model of experimental hemochromatosis: studies in rats fed finely divided elemental iron. Hepatology 1987; 7:1195-203. Kawase T, Kato S, Lieber CS. Lipid peroxidation and antioxidant defense systems in rat liver after chronic ethanol feeding. Hepatology 1989; 10: 815-21.
545
P. St&l et al.
38. Morton S, Mitchell MC. Effects of chronic ethanol feeding on glutathione turnover in the rat. Biochem Pharmacol 1985; 34: 1559--63. 39. Hirano T, Kaplowitz N, Tsukamoto H, Kamimura S, Femandez-Checa JC. Hepatic mitochondrial glutathione depletion and progression of experimental alcoholic liver disease in rats. Hepatology 1992; 16: 1423-7.
546
40. Fernandez-Checa JC, Ookhtens M, Kaplowitz N. Effects of chronic ethanol feeding on rat hepatocytic glutathione. Relationship of cytosolic glutathione to efflux and mitochondrial sequestration. J Clin Invest 1989; 83: 1247-52.
Copyright © EuropeanAssociation for the Study of the Liver 1996 Journal of Hepatology
ISSN 0168-8278
Hepatotoxicity induced by iron overload and alcohol Studies on the role o f chelatable iron, c y t o c h r o m e P450 2 E l a n d lipid p e r o x i d a t i o n Per St~l l, Inger Johansson2, Magnus Ingelman-Sundberg2, Karin Hagen3 and Rolf Hultcrantz3 1Department of Gastroenterology and Hepatology, Karolinska lnstitutet, Huddinge University Hospital, 2Department of Medical Biochemistry and Biophysics, Karolinska Institutet, and 3Department of Medicine, Karolinska Instituter, Karolinska Hospital, Stockholm, Sweden
Background~Aims: Clinical experience and studies with experimental animal models indicate a synergistic hepatotoxic effect of dietary iron overload and chronic alcohol ingestion. In order to elucidate the mechanism underlying this synergism, we examined the hepatic levels of ethanol-inducible cytochrome P450 2E1, glutathione and malondialdehyde, and the effect of iron chelation with desferrioxamine, in livers from rats treated with iron and/or ethanol. Methods: Animals received diets with or without 2.5-3% carbonyi iron for 6-9 weeks, followed by an ethanol-containing diet or a liquid control diet for 5-9 weeks. Desferrioxamine was administered subcutaneously with mini-osmotic pumps. Aianine aminotransferase activity in serum and hepatic contents of glutathione and malondialdehyde were determined. The hepatic level of cytochrome P450 2E1 was determined with Western Blotting using a specific polyclonal antibody. Results: The combination of iron and alcohol led to a marked increase in serum alanine aminotransferase activity as compared with all other treatment groups, and iron chelation with desferrioxamine reversed these increases. Treatment with al-
LINICALEXPERIENCEindicates a possible synergistic hepatotoxic effect when iron overload is combined with overconsumption of alcohol (1). We have previously demonstrated that rats fed carbonyl iron followed by a liquid ethanol-containing diet develop elevated levels of aminotransferases in their plasma
C
Received 8 June; revised 6 December; accepted 19 December 1995
Correspondence: Per StY, MD, Department of Gastroenterology and Hepatology, K 63, Karolinska Institutet, Huddinge Hospital, S-141 86 Huddinge, Sweden. Tel: +46-8746 22 63, Fax: +46-8-608 22 41. 538
cohol alone led to slightly increased aminotransferases compared with controls. The level of cytochrome P450 2E1 was significantly elevated in microsomes isolated from ethanol-treated rats, but neither additional iron supplementation nor desferrioxamine influenced this level significantly. Glutathione contents were increased in the livers of animals treated with iron and/or ethanol. Malondialdehyde values were increased in iron-treated animals, whereas neither ethanol nor desferrioxamine altered malondialdehyde levels significantly. Conclusions: The toxic effects exerted by the combination of iron overload and chronic ethanol feeding on rat liver are dependent on a pool of chelatable iron. The hepatic level of cytochrome P450 2El is markedly induced by ethanol but not further altered by iron overload. Neither increased lipid peroxidation nor depletion of hepatic glutathione levels can explain the synergistic hepatotoxic effects of iron and ethanol in this model.
Key words: Alcohol; Cytochrome P450 2E1; Desferrioxamine; Hemochromatosis; Hepatotoxicity; Iron.
and ultrastructural evidence of necrosis in the liver, alterations which are not seen in animals fed iron or alcohol alone (2). Lipid peroxidation of polyunsaturated fatty acids in membranes has been implicated as a mechanism by which iron, as well as ethanol, causes liver damage (3-7). Iron ions catalyze the formation of free species of reactive oxygen, which initiate the peroxidation process (8). Iron-induced peroxidation of intracellular membranes may lead to cellular dysfunction and eventually sideronecrosis (9-11). Alcohol ingestion by animals causes generation of
Hepatotoxicity by iron and alcohol
reactive oxyradicals and initiation of lipid peroxidation (12), probably via the induction of cytochrome P450 2El (CYP2E1) (5,13). This enzyme can generate free radicals, and initiate lipid peroxidation in microsomes (13). Livers from animals chronically exposed to ethanol, and thus having induced levels of CYP2E 1, display increased susceptibility to peroxidative damage caused by carbon tetrachloride, ferric citrate or ethanol itself (14,15). In addition, experimental evidence suggests that metabolism of acetaldehyde by aldehyde oxidase or xanthine oxidase may generate free radicals (16), or superoxide, which in turn mobilizes iron from ferritin (17). Thus, the increased hepatic iron stores seen in alcoholic liver disease may potentiate alcohol-induced hepatotoxicity (18). Hence, experimental evidence suggests the existence of a synergism between iron and alcohol with respect to increased free radical formation and lipid Experiment
A
t Weeks 7 18 19 I1° 111112113114115 I
11213141516
Control
I
Ethanol
I
I
I
I
I
[
I
I
ethanol,
I I ~h~ol
J
I
I
I
I
, ,
Experiment Control
I
I
I
I
I
I
B
t
Weeks I 11 z I 3 14 I s I 6 17 18 19 11°1111121131TM
I
8hanoi
IIII
Iron
Iron+
I
Eth..o,
E t h a n o l l l l l l l l l l l l l l l
ethanol
I
I I I I I I I I
Iron
Iron+
I
I
!
1
I
I
I
I
]
l
I
I
I
~ha.o.
peroxidation. The purpose of the present study was to investigate whether hepatotoxicity induced by a combination of iron and alcohol in vivo is mediated by a chelatable, low-molecular-weight iron pool, induction of CYP2E1, depletion of glutathione and/or increased hepatic levels of malondialdehyde (MDA). Materials
and
Methods
Animals
Male Sprague-Dawley rats were fed a rat chow diet (Altromin 1314, Altromin International, Lage, Germany) supplemented with 2.5 or 3% carbonyl iron (Sigma Chemicals, St Louis, MO, USA) (19) for 6-9 weeks after weaning, as illustrated in Fig. 1. Control groups were fed the same diet, but without iron supplementation. Following the period of iron supplementation, animals were pair-fed a liquid ethanolcontaining or non-alcoholic diet (Bioserv, New Jersey, USA), according to Lieber & DeCarli (20). The ethanol-containing and liquid control diets were identical, except that dextrin-maltose was replaced by isocaloric amounts of ethanol (65 g ethanol/l). The mean animal weights in the groups did not differ significantly when feeding with the liquid diet was started. Blood ethanol concentrations at sacrifice were similar, as was previously reported in ethanol-treated animals (2), without any significant difference between groups treated with or without iron and/or desferrioxamine. Animals who received the non-alcoholic diet did not have detectable amounts of ethanol in the blood. During the alcohol feeding period, animals were housed in separate cages and pair-fed individually. They were fed at noon every day and not given access to any food or drinking water other than that in the diet. The food was prepared fresh every other day. Simultaneous treatment with desferrioxamine did not alter the consumption of the ethanol-containing diet. Similarly, iron-treated and control animals consumed similar amounts of the liquid diet. Animals were maintained on a 12-h light-dark cycle.
I
Iron+ ! I ! I I I 1 I I I ~-~ ethanol + Etna.o, / DFO Fig. 1. Schematic presentation of the treated groups in Experiment A and B. Treatments as indicated. Five animals were included in each treated group. Experiment A: 3% carbonyl iron for 6 weeks + alcohol for 9 weeks. (Experiment A was also repeated with 3% carbonyl iron for 6 weeks + alcohol for 6 weeks). Experiment B: 2.5% carbonyl iron for 9 weeks + alcohol for 5 weeks; one of these groups also received DFO (desferrioxamine; 150 mg/kg/ day)for i week.
Experimental design
Two different experimental designs, designated A and B, were employed, as depicted in Fig. 1. In Experiment A, animals were divided into four different groups with five animals in each. The first group served as controls, receiving a normal chow diet for 6 weeks, followed by the liquid control diet for 9 weeks. Animals in the second group received the normal rat chow followed by an ethanol-containing liquid diet. In the third group, animals were fed a diet containing 3% carbonyl iron, followed by the liquid control diet without alcohol. In the fourth group ani539
P. St&l et al.
mals received the carbonyl iron-containing diet, followed by the ethanol liquid diet. Experiment A was repeated once, but with the modification that the alcohol-feeding period was shortened from 9 to 6 weeks. Animal weights at sacrifice did not differ significantly between treatment groups (control animals weighed 408+9 g, ethanol-treated 393+9 g, iron-treated 392+38 g, and iron + ethanol-treated 386+13 g). In Experiment B, 2.5% carbonyl iron was administered for 9 weeks, followed by an ethanol-containing diet for 5 weeks. In this experiment, the first four groups of treated animals were identical to those of Experiment A, but an additional group receiving iron, alcohol and subcutaneous desferrioxamine (Desferal ®, Ciba-Geigy, Switzerland) was also employed. The animals in group four, receiving iron + alcohol, were implanted with mini-osmotic pumps containing 0.9% saline. Animal weights did not differ between treatment groups at sacrifice (control animals weighed 391+28 g, ethanol-treated 375+39 g, irontreated 370+28 g, iron + ethanol-treated 366+26 g, and those treated with iron + ethanol + desferrioxamine 365+17 g).
Chemicals and procedures Desferrioxamine was administered by mini-osmotic pumps (Alzet 2ML1) implanted subcutaneously on the back of the animal. Desferrioxamine was dissolved in sterile water at a concentration of 250 mg/ ml and the mini-osmotic pump (release of 10 ~tl/h) was filled with 2 ml of this solution. The substance was released at a rate of 150-180 mg/kg per day until sacrifice (animal weight 333-400 g). Animals were killed after the alcohol feeding period. They were anesthetized with ether, the abdomen cut open and blood drawn from the aorta for serum alanine aminotransferase (ALT) assays. Slices of liver tissue were fixed for light microscopic examination. Liver tissue was also immediately frozen in liquid nitrogen for determination of iron, protein, total hepatic malondialdehyde and glutathione levels, CYP2E1, MDA in liver microsomes and NADPHdependent lipid peroxidation. Light microscopy Liver tissue was fixed in formaldehyde and embedded in paraffin, after which sections were stained with hematoxylin-eosin, reticulin and Perls' blue, the latter to visualize hepatic iron deposits. Chemical and immunochemical determinations The iron contents of liver tissue samples were measured in an atomic absorption spectrophotometer (Vat540
ian SpectrAA 400) using background correction. The samples were hydrolyzed in 3.5% HNO 3 overnight before analysis (21). Protein was determined according to the method of Lowry et al. (22). For determination of CYP2E1 levels, liver tissue was homogenized in 10 mM sodium/potassium phosphate buffer (pH 7.4) containing 1.14% KC1 and the homogenate subfractionated by centrifugation and the microsomal fraction washed once before resuspension in 50 mM potassium phosphate buffer (pH 7.4) (5). Rat liver CYP2E1 was then quantitated by Westernblotting according to a procedure described previously (5), using specific rabbit polyclonal anti-CYP2E1 IgG characterized earlier (23). MDA levels and NADPHdependent lipid peroxidation in liver microsomes were determined as described previously (5,13). For determination of total hepatic malondialdehyde (MDA), slices of liver tissue were frozen in liquid nitrogen and stored at -70°C. The production of malondialdehyde (MDA) was assayed as thiobarbituric acid reactive products (TBAR), measured according to a method described by Ohkawa et al. (24). In short, the liver tissue sample weighing approximately 0.3 g was homogenized in 0.15 M potassium phosphate buffer (pH 7.2) to a concentration of 15%, and kept on ice or at 4°C. Twenty-five mM desferrioxamine was included in the buffer (Ciba-Geigy, Switzerland) to avoid TBAR formation in the sample. Two hundred and fifty microliters of the homogenates was pipetted into a glass tube, and 500 p.1 of 20% trichloroacetic acid and 40 I.tl of 2% 2,6-ditertbutyl-4-methylphenol was added. Two hundred milligrams of thiobarbituric acid (TBA) was dissolved in 30 ml distilled water (to a concentration of 6.7%) at 70°C. One milliliter of the 6.7% TBA solution was added and the homogenates were boiled at 100°C for 15 min and allowed to cool to room temperature. Following centrifugation at 3000 G (15°C) for 10 min, 1 ml of the supernatant was read spectrophotometrically (UV 160, Shimadzu, Kyoto, Japan) at 535 nm using appropriate blanks and standards. As MDA standard, 1,1,3,3-tetraethoxypropane (Sigma T9889) was mixed with buffer at different concentrations in the interval 0.1-20 I.tM. Glutathione was measured using the GSH-400 assay (Bioxytech S.A. Bonneuil/Marne, Cedex, France). In this procedure, the first step involves formation of thioethers between a chromogenic reagent and all mercaptans in the sample, whereas the second step transforms these substitution products into a chromophoric thione with maximal absorbance at a wavelength of 400 nm. Ten per cent (w/v) liver homogenates in distilled water were prepared, sup-
Hepatotoxicity by iron and alcohol
plemented with 5% (w/v) metaphosphoric acid (in order to precipitate high-molecular weight components) and centrifuged at 2500 g (10 min at 4°C). The assay was performed in duplicate on the supernatants obtained. A standard curve was prepared using 25200 ~tM glutathione (reduced form, 98-100%, Sigma Chemicals, St Louis, MO, USA) dissolved in 5% metaphosphoric acid and the glutathione contents of the samples were determined using linear regression. Statistical methods
The values for the different groups were compared using analysis of the variance (ANOVA). When two treatment groups were compared, Student's unpaired t-test (2-tail) was employed.
a 450 4O0 350
3O0 25O ,,¢
200 "k'k
150 100 50 0
_ w
•
a-w
Control
Ethanol
Iron
-= Iron+
ethanol
Group
b 140
100 80
60 40 20 0
This study was approved by the Stockholm North Local Committee for Ethical Review of Animal Experiments.
Results Serum alanine aminotransferase (ALT)
Fig. 2a shows the ALT activities in the serum of all animals in Experiment A, including the repeated experiment in which animals were treated with ethanol for 6 weeks. There was no statistically significant difference in ALT activities for animals fed ethanol for 6 or 9 weeks. Ethanol-feeding alone led to significantly increased ALT values as compared with controis (33.7+6.1 vs. 16.1+4.7 U/I). Also, iron feeding alone led to significantly increased ALT activities compared with controls (36.0-&-_19.2 U/I). ALT activities were markedly increased in animals receiving both iron and ethanol compared with all other groups (140.9+115.6 U/l). Figure 2b presents the ALT activities for the animals in experiment B. Here, treatment with ethanol alone led to significantly increased ALT activities as compared with controls (23.0+3.8 vs. 9.0-L-_3.0 U/I), whereas treatment with iron alone did not. As in experiment A, a marked increase in ALT was seen after treatment with iron plus ethanol (56.6+38.8 U/ 1), whereas treatment with desferrioxamine reversed this increase in serum ALT activity for animals exposed to both iron and ethanol (to ALT levels of 16.0+_9.7 U/I). Experiment A: Hepatic contents o f iron, MDA, glutathione and CYP2E1 (Table 1)
120
h-
Ethical approval
Control
Ethanol
Iron
Group
Iron+ ethanol
Iron+ ethanol+ DFO
Fig. 2. a) The activities of serum alanine aminotransferases (ALT) in the treated groups in Experiment A (since experiment A was repeated, each treatment group includes 5 + 5 animals), b) The activities of serum alanine aminotransferases (ALT) in the treated groups in Experiment B. Each dot represents one animal and a horizontal line depicts the mean of each group. * = p
The hepatic contents of iron, MDA, glutathione and CYP2E1 for the different groups in Experiment A are shown in Table 1. Values were similar in the repeated experiment A in which animals were treated with ethanol for 6 instead of 9 weeks (data not shown). The hepatic iron concentration was 0.93 ~tg/mg protein in the control animals fed a diet without iron supplementation,, while animals fed ethanol demonstrated only about 30% of this level. As expected, dietary iron supplementation led to a marked and significant increase in the content of hepatic iron for all groups. Animals fed the combination of iron and ethanol had hepatic levels of iron contents similar to those fed iron alone. The hepatic concentrations of MDA were significantly increased in iron-treated animals, whereas supplementation of ethanol did not further increase these levels. Treatment with iron and/or ethanol led to sig541
P. Stdl et al.
hepatic microsomes were significantly increased in iron-treated animals as compared with animals receiving ethanol only or a control diet. Microsomal NADPH-dependent lipid peroxidation did not differ between the treatment groups.
nificant increases in hepatic glutathione concentrations, compared with controls. The levels of CYP2E1 in liver microsomes are shown in Table 1. Ethanol treatment caused significant increases in the levels of CYP2E1 compared with controls or iron-treated animals. Iron overload did not result in any significant alteration in the amount of immunodetectable CYP2E1. In ethanoltreated animals, simultaneous iron overload did not alter the levels of CYP2E1. MDA levels of isolated
Experiment B: Hepatic contents of iron, MDA and CYP2E1 (Table 2) The hepatic contents of iron, MDA and CYP2E1 of the different groups in Experiment B are presented in
TABLE 1 Hepatic contents of iron, malondialdehyde (MDA) and glutathione, and levels of ethanol-inducible cytochrome P450 2El in the livers of the different groups (Experiment A) Control (n=5)
Alcohol (n=5)
Iron (n=5)
Iron+alcohol (n=5)
Liver iron (~mol/mg protein)
0.93_+0.10
0.33_+0.152
4.32_+0.921
4.06+1.001
MDA (nmol/g liver wet weight)
18.7+2.5
16.7+1.6
37.7+4.71
35.4_+1.31
Glutathione (~mol/g liver wet weight)
3.2_+0.6
4.6_+0.82
4.4_+0.62
4.3_+0.62
Cytochrome P450 2E 1 (Arbritrary units)
100-+39
512+893
87_+33
450-2_1463
Microsomal MDA (nmol/mg microsomal protein)
1.02_+0.03
0.88_+0.05
1.4-+0.321
1.26_+0.124
NADPH-dependent lipid peroxidation (nmol/mg microsomal protein)/min
0.14_+0.03
0.17_+0.06
0.14_+0.02
0.13_+0.04
Results expressed as means + S.D. MDA = malondialdehyde (measured as TBA-reactive substances). 1 = p < 0.05 compared with controls or animals treated with alcohol alone. 2 = p < 0.05 compared with controls. 3 = p < 0.05 compared with controls or animals treated with iron alone. 4 = p < 0.05 compared with animals treated with alcohol alone.
TABLE 2 Hepatic contents of iron and malondialdehyde (MDA), and levels of ethanol-inducible cytochrome P450 2El in the livers of the different groups (Experiment B). Control (n=5)
Alcohol (n=5)
Iron (n=5)
Iron+alcohol (n=5)
Iron+alcohol +DFO
Liver iron (~tmol/mg protein)
0.62_+0.17
0.38-+0.03
3.44--1.07 x
2.94-+0.181
2.43-+0.64 l
MDA (nmol/g liver wet weight)
12.9+2.3
11.4-+1.4
25.1+3.31
22.5+2.61
23.4-+5.71
Cytochrome P450 2E 1
100-2_21
1050-2-4753,4
524+-2773,4
344-+2632
(n=5)
(Arbritrary units) Results expressed as means + S.D. DFO = desferrioxamine. M D A = malondialdehyde. 1 = p < 0.05 compared with controls or animals treated with alcohol alone. 2 = p < 0.05 compared with controls. 3 = p < 0.01 compared with controls. 4 = p < 0.05 compared with animals treated with iron alone.
542
142-+302
Hepatotoxicity by iron and alcohol
Table 2. Treatment with desferrioxamine did not result in significantly decreased hepatic iron concentrations. Hepatic concentrations of MDA were significantly increased in iron-treated animals, but neither alcohol nor desferrioxamine further altered MDA levels. There was a slight increase in the hepatic levels of CYP2E1 in the iron-overloaded animals as compared with controls. As in Experiment A, adding iron to ethanol-treated animals did not alter the CYP2E1 levels. Desferrioxamine had no significant effect on CYP 2El levels in animals treated with iron plus ethanol. In animals treated with ethanol alone, there was a significant correlation between ALT activities and hepatic levels of CYP2E1 (r=0.92, p<0.05). No such correlation was found in animals treated with iron plus ethanol, or iron plus ethanol and desferrioxamine.
Light microscopy The findings from the light microscopic examinations were similar to those described previously (3). Liver tissue from animals fed ethanol displayed fat droplets in the central areas of the lobule. Perls' blue stain showed a positive granular reaction in periportal hepatocytes in animals fed carbonyl iron, but not in the other groups. In addition, occasional iron-containing Kupffer cells were present in the former livers. In animals receiving both iron and ethanol occasional inflammatory cells, as well as lipid droplets, granularity in hepatocytes and iron-containing Kupffer cells, were seen. The reticulin stain did not show any features of increased fibrosis in any of the livers. The cellular and lobular iron distributions were similar in animals exposed to desferrioxamine for 1 week as compared to the other groups treated with iron.
Discussion The aim of the present study was to elucidate the mechanisms underlying the synergistic hepatotoxicity exerted by a combination of iron overload and ethanol. We demonstrated that this synergistic hepatotoxicity is dependent on chelatable iron, since desferrioxamine added to the treatment with iron and ethanol decreases ALT activities in serum, although the total hepatic iron contents and MDA levels are unchanged. The exact mechanisms by which chelatable iron is involved in the hepatotoxic events in this model are unclear. In iron overload, non-transferrin-bound iron is markedly increased in serum and taken up by the hepatocytes, but less efficiently excreted into the bile
as compared with the iron release from control livers (25). It has been demonstrated that the pool of chelatable iron is enlarged in animals exposed to dietary carbonyl iron overload, although this enlargement is not as marked as the increase in total hepatic iron content (26,27). In addition, the pool of low-molecular-weight iron has been reported to increase during ethanol metabolism (28). Hypothetically, an increased pool of catalytically active iron ions could enhance alcohol-induced lipid peroxidation of cellular membranes and the leakage of aminotransferases into plasma. In the present study, iron alone increased lipid peroxidation significantly. We had no signs of increased lipid peroxidation in animals treated with ethanol alone, and adding ethanol to iron-treated animals did not further potentiate the formation of lipid peroxides (measured as TBAR formation) in the liver. Chelation of iron with desferrioxamine did not decrease the hepatic contents of malondialdehyde in animals exposed to the combination of iron and alcohol, although the aminotransferases were normalized. Hence, in this animal model we have no evidence that hepatotoxicity induced by iron and alcohol involves enhanced lipid peroxidation. In the present study, ethanol was given in the diet as described by Lieber & deCarli (20), in which ethanol-treated animals are fed ad libitum and the corresponding controls are given a similar amount of the iso-caloric, non-alcoholic diet. Using the LieberdeCarli ethanol-containing diet in combination with iron overload, only moderate liver injury such as steatosis and mild necrosis is produced (2), even if the duration of the ethanol-feeding period is prolonged, as in a study by Olynyk et al. (29). Butcher et al. used electron paramagnetic resonance spectroscopy on livers from rats subjected to a similar treatment, and failed to detect increased levels of stable free radicals in these livers (30). Others have demonstrated results similar to those of the present study, with unchanged levels of lipid peroxides of livers exposed to iron overload plus ethanol (31). In spite of this, lipid peroxidation has been implicated as a major mechanism in alcohol-related liver injury (5-7,12-17). In the present study, feeding with ethanol alone led to a slight but significant increase in plasma ALT activities, although MDA levels were unchanged. Sixteen weeks of intragastric feeding with ethanol plus a high-fat diet was shown to produce increased levels of malondialdehyde, 4-hydroxynonenal and signs of hepatic fibrosis, whereas 5 weeks with the same treatment led to elevated aminotransferases in plasma, but unaltered hepatic levels of malondialdehyde and 4-hydroxynonenal (12).
543
P. St&l et al.
Thus, ethanol infused intragastrically for prolonged time periods results in more severe signs of liver damage (including induced fibrogenesis) as compared with the Lieber-deCarli diet (2,12,29,32). Tsukamoto et al. further demonstrated that small amounts of iron (not resulting in iron overload) added to the intragastric ethanol plus high-fat diet enhanced lipid peroxidation and induced fibrogenesis (32). Similarly, iron potentiated fibrogenesis when the fibrotic stimulus was low-dose carbon tetrachloride plus alcohol (33). These results indicate that in the model used in the present study and the study by Olynyk et al. (29), iron overload combined with ethanol according to the Lieber-de Carli protocol causes increased aminotransferase levels and mild necrosis but neither enhanced lipid peroxidation nor fibrosis, as compared with iron overload alone. In the different model used by Tsukamoto et al., ethanol and a highfat diet given intragastrically have fibrogenic properties by itself, and iron in small amounts added to that protocol potentiates lipid peroxidation and fibrogenesis (32). Thus, a pool of chelatable low-molecularweight iron possibly interacts with other liver-cell damaging agents such as alcohol (32) or carbon tetrachloride (33), which is consistent with the decrease in aminotransferases seen if desferrioxamine is administered to animals treated with iron plus ethanol in the present study. Another possible interaction of iron in ethanolinduced liver damage could be iron-induced effects on the ethanol-inducible cytochrome P450 (CYP2E1). CYP2E1 may constitute the major source of superoxide anion production, leading to the formation of hydroxyl radicals via an iron-catalyzed HaberWeiss reaction (13,34), indicating a potential role for iron in CYP2El-induced free radical production (13, 16,34). In the present study, iron overload alone slightly increased the hepatic levels of CYP2E1 compared with controls (in experiment B). In contrast, in ethanol-treated animals CYP2E1 levels were neither significantly altered by iron nor by desferrioxamine. Hence, increased CYP2E1 levels were not the reason for the potentiating effects of iron on ethanol-induced liver damage in this model. However, we cannot rule out the importance of this enzyme in the pathogenesis of ethanol-induced hepatotoxicity, since hepatic levels of CYP2E1 correlated significantly to ALT values in animals treated with ethanol alone. Another mechanism by which iron and ethanol might exert hepatotoxic effects would be depletion of antioxidant defense systems. In the present study depletion of total hepatic glutathione stores could be excluded as a possible mechanism leading to hepato544
cellular injury. The reported effect of iron overload on hepatic glutathione levels has been variable, i.e. an increase (35), decrease (36) or unaltered levels (4) as compared with controls. Alcohol was found to depress hepatic glutathione in patients with chronic liver disease (7). Hepatic glutathione levels following chronic exposure to ethanol in animals are variable. Three weeks of ethanol-feeding increased hepatic glutathione stores (37), and glutathione turnover and hepatic glutathione contents were similarly increased following 8 weeks of ethanol treatment in rats (38). These results are thus in line with those of the present study. However, 16 weeks of intragastric infusion with ethanol plus a high-fat diet resulted in depletion of glutathione, whereas 5 weeks with the same treatment did not (12), indicating that a more aggressive ethanol exposure may lead to depletion of the total hepatic glutathione stores. Long-term feeding with ethanol selectively reduces the mitochondrial glutathione pool at first, and this reduction may escape detection when measuring the total glutathione content, since the cytosolic pool remains unchanged (39,40). Iron added to an ethanol-high-fat diet was not found to deplete hepatic glutathione (32). We conclude that hepatotoxicity induced by a combination of iron overload and chronic ethanol treatment in this animal model depends on a pool of chelatable iron, but does not involve any detectable increase in lipid peroxidation or glutathione depletion. Hepatic levels of CYP2E1 are markedly increased by ethanol, but not further altered by iron overload. Although previous evidence speaks in favor of a free-radical mediated process, the exact mechanisms by which iron overload work in concert with chronic ethanol exposition to induce hepatocellular damage remains to be elucidated.
Acknowledgements The authors are grateful to Mrs. Kristina Eckes for skillful technical assistance. This study was supported by grants from the Swedish Medical Research Council, the Nanna Swartz Foundation, the Swedish Society of Medicine (Bengt Ihres Fund), and Rut and Richard Juhlins Foundation.
References 1. LeSage GD, Baldus WP, Fairbanks VF, Baggenstoss AH, McCall JT, Moore SB, Taswell HF, Gordon H. Hemochromatosis: genetic or alcohol-induced? Gastroenterology 1983; 84: 1471-7. 2. Sthl P, Hultcrantz R. Iron increases ethanol toxicity in rat liver. J Hepatol 1993; 17: 108-15. 3. Dresow B, Albert C, Zimmerman I, Nielsen P. Ethane exhalation and vitamin E/ubiquinol status as markers of lipid per-
Hepatotoxicity by iron and alcohol
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20. 21.
oxidation in ferrocene iron-loaded rats. Hepatology 1995; 21: 1099-105. Fletcher LM, Roberts FD, Irving MG, Powell LW, Halliday JW. Effects of iron loading on free radical scavenging enzymes and lipid peroxidation in rat liver. Gastroenterology 1989; 97: 1011-8. Ekstr/Sm G, Ingelman-Sundberg M. Rat liver microsomal NADPH-supported oxidase activity and lipid peroxidation dependent on ethanol-inducible cytochrome P-450IIE1. Biochem Pharmacol 1989; 38: 1313-9. Situnayake RD, Crump B J, Thurnham DI, Davies JA, Gearty J, Davis M. Lipid peroxidation and hepatic antioxidants in alcoholic liver disease. Gut 1990; 31:1311-7. Shaw S, Rubin KP, Lieber CS. Depressed hepatic glutathione and increased diene conjugates in alcoholic liver disease. Evidence of lipid peroxidation. Dig Dis Sci 1983; 28: 585-9. Bacon BR, Britton RS. The pathology of hepatic iron overload: a free radical mediated process? Hepatology 1990; 11: 127-37. Myers BM, Prendergast FG, Holman R, Kuntz SM, LaRusso NE Alterations in the structure, physicochemical properties, and pH of hepatocyte lysosomes in experimental iron overload. J Clin Invest 1991; 88: 1207-15. Bacon BR, Healey JF, Brittenham GM, Park CH, Nunnari J, Tavill AS, Bonkovsky HL. Hepatic microsomal function in rats with chronic dietary overload. Gastroenterology 1986; 90: 1844-53. St~ P, Hultcrantz R, Glaumann H. Liver cell damage and lysosomal iron storage in patients with idiopathic hemochromatosis. A light and electron microscopic study. J Hepatol 1990; 11: 172-80. Kamimura S, Gaal K, Britton RS, Bacon BR, Triadafilopoulos G, Tsukamoto H. Increased 4-hydroxynonenal levels in experimental alcoholic liver disease: association of lipid peroxidation with liver fibrogenesis. Hepatology 1992; 16: 448-53. Ingelman-Sundberg M, Johansson I. Mechanisms of hydroxyl radical formation and ethanol oxidation by ethanol-inducible and other forms of rabbit liver microsomal cytochromes P-450. J Biol Chem 1984, 259; 6447-58. Castillo T, Koop DR, Kamimura S, Triadafilopoulos G, Tsukamoto H. Role of cytochrome P-450 2El in ethanol-, carbon tetrachloride- and iron-dependent microsomal lipid peroxidation. Hepatology 1992; 16: 992--6. Shaw S, Jayatilleke E, Ross WA, Gordon E, Lieber CS. Ethanol-induced lipid peroxidation: potentiation by long-term alcohol feeding and attenuation by methionine. J Lab Clin Med 1981; 98: 417-25. Shaw S, Jayatilleke E. The role of aldehyde oxidase in ethanol-induced hepatic lipid peroxidation in the rat. Biochem J 1990; 268: 579-83. Shaw S. Lipid peroxidation, iron mobilization and radical generation induced by alcohol. Free Rad Biol Med 1989; 7: 541---47. Irving MG, Halliday JW, Powell LW. Association between alcoholism and increased hepatic iron stores. Alcohol Clin Exp Res 1988; 12: 7-13. Park CH, Bacon BR, Brittenham GM, Tavill AS. Pathology of dietary carbonyl iron overload in rats. Lab Invest 1987; 57: 555-63. Lieber CS, DeCarli LM. The feeding of ethanol in liquid diets. Alcohol Clin Exp Res 1986; 10:550-3 Hultcrantz R, Glaumann H. Studies on the rat liver following
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
iron overload. Biochemical studies after iron removal. Lab Invest 1982; 46: 383-93. Lowry OH, Rosebrough N J, Fan" AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Biol Chem 1951; 193: 265-75. Johansson I, Ingelman-Sundberg M. Benzene metabolism by ethanol-, acetone-, and benzene-inducible cytochrome P-450 (IIE1) in rat and rabbit liver microsomes. Cancer Res 1988; 48: 5387-90. Ohkawa H, Ohishi N, Yagi K. Assay for lipid peroxides in animal tissues by thiobarbituric acid reaction. Anal Biochem 1979; 95: 351. Brissot P, Zanninelli G, Guyader D, Zeind J, Gollan J. Biliary excretion of plasma non-transferrin-bound iron in rats: pathogenetic importance in iron-overload disorders. Am J Physiol 1994; 267: G135--42. Britton RS, Ferrali M, Magiera CJ, Recknagel RO, Bacon BR. Increased prooxidant action of hepatic cytosolic lowmolecular-weight iron in experimental iron overload. Hepatology 1990; 11: 1038--43. Nielsen P, Diillmann J, Wulfhekel U, Heinrich HC. Nontransferrin-bound-iron in serum and low-molecular-weightiron in the liver of dietary iron-loaded rats. Int J Biochem 1993; 25: 223-32. Sergent O, Morel I, Cogrel P, Lescoat G, Pasdeloup N, Brissot P, Cillard P, Cillard J. Cellular pool of low molecular weight iron during metabolism of ethanol in rat hepatocytes. (Abstract) Hepatology 1993; 18: 276A. Olynyk J, Hall P, Reed W, Williams P, Kerr R, Mackinnon M. A long-term study of the interaction between iron and alcohol in an animal model of iron overload. J Hepatol 1995; 22: 671--6. Butcher GP, Deighton N, Batt RM, Ding B, Haywood S, Hoffman J, Jackson MJ, Symons MCR, Rhodes JM. Electron paramagnetic resonance spectroscopy of stable free radicals in the liver compared with ultrastructural and functional damage in a rat model of alcohol- and iron-overload. Clin Sci 1993; 84: 339--48. Tector AJ, Britton RS, O'Neill R, Bacon BR. Additive effect of iron overload and ethanol on hepatic damage. (Abstract). Gastroentrology 1993; 102: A944. Tsukamoto H, Home W, Kamimura S, Niemel~i O, Parkkila S, Yl~i-Herttuala S, Brittenham GM. Experimental liver cirrhosis induced by alcohol and iron. J Clin Invest 1995; 96: 620-30. Mackinnon M, Clayton C, Plummer J, Ahem M, Cmielewski P, Ilsley A, Hall P. Iron overload facilitates hepatic fibrosis in the rat alcohol/low-dose carbon tetrachloride model. Hepatology 1995; 21: 1083-8. Nordmann R, Ribi~re C, Rouach H. Involvement of iron and iron-catalysed free radical production in ethanol metabolism and toxicity. Enzyme 1987; 37: 57--69. Kawabata T, Ogino T, Awai M. Protective effects of glutathione against lipid peroxidation in chronically iron-loaded mice. Biochim Biophys Acta 1989; 1004: 89-94. Bonkovsky HL, Healey JF, Lincoln B, Bacon BR, Bishop DF, Elder GH. Hepatic heme synthesis in a new model of experimental hemochromatosis: studies in rats fed finely divided elemental iron. Hepatology 1987; 7:1195-203. Kawase T, Kato S, Lieber CS. Lipid peroxidation and antioxidant defense systems in rat liver after chronic ethanol feeding. Hepatology 1989; 10: 815-21.
545
P. St&l et al.
38. Morton S, Mitchell MC. Effects of chronic ethanol feeding on glutathione turnover in the rat. Biochem Pharmacol 1985; 34: 1559--63. 39. Hirano T, Kaplowitz N, Tsukamoto H, Kamimura S, Femandez-Checa JC. Hepatic mitochondrial glutathione depletion and progression of experimental alcoholic liver disease in rats. Hepatology 1992; 16: 1423-7.
546
40. Fernandez-Checa JC, Ookhtens M, Kaplowitz N. Effects of chronic ethanol feeding on rat hepatocytic glutathione. Relationship of cytosolic glutathione to efflux and mitochondrial sequestration. J Clin Invest 1989; 83: 1247-52.