Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Long-term treatment of alcoholic liver disease continues to incorporate vitamins, nutrients, and trace elements (Fulton and McCullough, 1998; McCullough et al., 1998). In fact, the role of specific pharmacological agents remains unproven. Clearly, the development of more effective nutritional or pharmacological therapy will depend on further elucidating the mechanisms that contribute to liver injury.

Several lines of investigation indicate that dietary fat can modulate the severity of alcoholic liver injury (Mezey, 1998). In experimental animals, for example, diets enriched with saturated fatty acids protect against alcohol-induced liver injury, whereas diets containing polyunsaturated fatty acids promote liver injury (Nanji and French, 1989; Nanji et al., 1989, 1994a). Saturated fatty acids have also been reported to reverse established alcoholic liver injury (Nanji et al., 1995, 1996, 1997b). Importantly, in previous studies, use of alcohol was discontinued at the time that dietary treatment was initiated. This model represented the alcoholic patient who stopped drinking at the time of hospitalization (French, 1995).

Discontinuation of alcohol remains pivotal in the treatment of alcoholic liver disease. Although this goal can frequently be achieved in the short-term, the majority of patients resume alcohol consumption, often with sudden deterioration in liver disease (Pares et al., 1986). Hence, it is important to develop therapeutic strategies that simulate the clinical condition in which alcohol use is continued despite the presence of alcoholic liver disease.

Previously, we used the intragastric feeding rat model to study the pathogenesis of alcoholic liver disease (Nanji et al., 1999). In addition to being useful for elucidating mechanisms of injury, this model has been used to evaluate various strategies to prevent or reverse alcoholic liver disease (Nanji et al., 1995, 1997b). The results of previous studies suggest that elevated levels of endotoxin and lipid peroxides in alcohol-fed animals activate nuclear factor-B (NF-B), leading to enhanced expression of tumor necrosis factor- (TNF-), cyclooxygenase-2 (Cox-2), and proinflammatory cytokines (Nanji et al., 1997a, 1999). In the current study, we investigated whether treatment with dietary saturated fatty acids could reverse established alcoholic liver injury despite continued administration of ethanol. We show that diets enriched in saturated fatty acids improved both histological liver injury and biochemical parameters that have been etiologically linked to liver injury.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animal Model. Male Wistar rats (Harlan Bioproducts for Science, Indianapolis, IN) weighing between 225 and 250 g were fed a liquid diet by continuous infusion through permanently implanted gastric tubes as previously described (French et al., 1986; Tsukamoto et al., 1990). The rats received their total nutrient intake by intragastric infusion. This was achieved by joining two tubes, one carrying ethanol from one syringe pump and the other carrying diet from a second pump, so that ethanol and diet could be varied at will. Vitamins and minerals were given as described previously (French et al., 1993) The dose of ethanol was increased slowly, as tolerance developed, to maintain blood alcohol levels in the range of 150 to 300 mg/dl. The starting dose was 10 g/kg/day; the final dose was 16 g/kg/day. Each ethanol-fed rat underwent at least two measurements of blood alcohol level.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Schematic diagram of experimental design.

Histopathological Analysis Including Sirius Red Staining for Collagen. A small sample of liver was obtained by biopsy or at death and fixed in formalin. Hematoxylin and eosin stain was used for light microscopy. The severity of liver pathology was assessed as follows: steatosis (the percentage of liver cells containing fat), 1+, 25% of cells containing fat; 2+, 26 to 50%; 3+, 51 to 75%; and 4+, >75%. Necrosis was evaluated as the number of necrotic foci per square millimeter; inflammation was scored as the number of inflammatory cells per square millimeter. At least three different sections were examined per sample of liver. The pathologist evaluating these sections was unaware of the treatment the rats had received.

Measurement of Blood Alcohol Levels. Blood was collected from the tail vein, and ethanol concentration was measured using the alcohol dehydrogenase kit from Sigma Chemical (St. Louis, MO).

Measurement of Plasma Endotoxin Levels. Blood samples were collected in endotoxin-free vials (Sigma Chemical) and measured as previously described (Nanji et al., 1997a).

Determination of Thiobarbituric Acid-Reactive Substances (TBARS) and Conjugated Dienes. Levels of liver TBARS were measured according to the method of Ohkawa et al. (1979). Conjugated dienes in the total lipid extracted from liver homogenates were identified by their optical density of between 220 and 300 nm as described by Recknagel and Glende (1984).

Measurement of Nonheme Iron. Nonheme iron was determined in liver homogenate as described previously with ferene S, as an indicator with the molar absorptivity of 35,500 M¹ cm¹ at 594 nm (Artiss et al., 1982).

Aniline Hydroxylase Assay. Aniline hydroxylase assays were performed as described previously (Imai et al., 1966; Waxman et al., 1989). Liver microsomes were incubated for 60 min at 37°C in 0.45 ml of 0.1 mol/l potassium phosphate, pH 7.4, containing 8 mM aniline and 1 mM NADPH. Reactions were terminated with 90 µl of 40% trichloroacetic acid. Samples were then placed on ice for 10 min followed by 10 min of centrifugation. An aliquot of the supernatant (0.36 ml) was mixed with 10% Na₂CO₃ (0.24 ml) and 2% phenol (0.36 ml). A₆₃₀ values were determined after incubation for 45 min in the dark. Specific activities were calculated from a standard curve prepared with the reaction product 4-aminophenol (Aldrich Chemical, Milwaukee, WI).

Determination of NF-B Binding Activity and IB Protein Levels in Liver. Electrophoretic mobility shift assays (EMSAs) were used to determine the binding activity of NF-B and were performed essentially as described in previous studies (Lin et al., 1997; Nanji et al.; 1999). Equal amounts of protein were incubated with a 5'-³²P-labeled oligonucleotide containing a NF-B consensus site. The incubation mixtures were separated on a 7% nondenaturing polyacrylamide gel and bands were detected by autoradiography. The specificity of binding was determined by prior addition of 100-fold excess of unlabeled competitor consensus oligonucleotide.

Determination of mRNA Levels for TNF-, Cox-2, Cox-1, and -Actin. To examine the expression of Cox-1, Cox-2, TNF-, and -actin in liver tissue, total RNA was isolated according to the guanidinium isothiocyanate method (Chomczynski and Sacchi, 1987). The integrity of RNA was assessed by agarose gel electrophoresis and ethidium bromide staining. Reverse transcription-polymerase chain reaction was performed as previously described (Nanji et al., 1997a). The sequences of primer pairs, 5' and 3', and predicted sizes of the amplified PCR fragments of Cox-1, Cox-2, TNF-, and -actin have been reported previously (Nanji et al., 1994c, 1997a,b). PCR products and molecular size markers were subjected to electrophoresis on 1% agarose gels and visualized by means of ethidium bromide staining. The gels were analyzed by laser scanning densitometry with a Molecular Dynamics densitometer and Image Quant software (Molecular Dynamics, Sunnyvale, CA). Each experiment included a negative control (sample RNA that had not been subjected to reverse transcription). This sample did not yield a PCR product, confirming the absence of extraneous genomic DNA or PCR products contaminating the samples. Varying the number of PCR cycles did not change the relative differences between the samples, indicating that the PCR conditions were not within the plateau phase of amplification. All amplification reactions of one experiment were performed in parallel in the same heating block to ensure compatible conditions.

Statistical Analysis. Results are presented as means ± S.D. Comparison among groups was performed by one-way analysis of variance. The paired t test was used to evaluate differences in pathological changes in the same animals after switching to the saturated fatty acid-enriched diets.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

There was no significant difference in weight gain or levels of blood alcohol among the various groups. The levels of blood alcohol (mg/dl) in the different groups were fish oil-ethanol, 8 weeks, 226 ± 62; fish oil-ethanol, 6 weeks, 238 ± 53; fish oil-ethanol and palm oil-ethanol, 257 ± 62; and fish oil-ethanol and MCT-ethanol, 221 ± 64.

Effect of Experimental Diets on Liver Pathology. Feeding a fish oil-ethanol diet for 6 (group 2) or 8 weeks (group 1) caused fatty liver, necrosis, and inflammation (Table 2). Control animals fed the fish oil-dextrose diet (group 5) showed no pathological changes. Significant improvement in hepatic pathology occurred when the fish oil-ethanol diet was discontinued and replaced with either palm oil-ethanol (group 3) or MCT-ethanol (group 4). Necrosis and inflammation were virtually absent in these two groups (3 and 4). In contrast, the degree of fatty liver did not decrease in these two groups.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Changes in amounts of collagen and pericellular fibrosis in rats fed fish oil and ethanol for 6 weeks followed by palm oil and ethanol (group 3) or MCT and ethanol (group 4) for 2 weeks. Comparison groups were animals fed fish oil and ethanol for 8 weeks (group 1) and fish oil and ethanol for 6 weeks (group 2). All animals receiving the palm oil-ethanol and MCT-ethanol diets underwent liver biopsy after 6 weeks of fish oil-ethanol treatment. The results of these initial biopsies served as a baseline for comparison with the result after treatment with the experimental diets. In the fish oil-ethanol groups treated with palm oil-ethanol (group 3), a significant decrease (p < 0.01) in the amount of central vein collagen (% area) (1.3 ± 0.2-0.4 ± 0.1) and pericellular fibrosis (19 ± 5-6 ± 1) was seen after 2 weeks of treatment. A significant decrease (p < 0.01) in central vein collagen (% area) (1.2 ± 0.2-0.4 ± 0.1) and pericellular fibrosis (20 ± 4-7 ± 1) was seen also in the group treated with MCT-ethanol. The amount of central vein collagen (%) and pericellular fibrosis were not significantly different in the rats treated with fish oil and ethanol for 8 weeks (group 1) (collagen, 1.4 ± 0.2, pericellular fibrosis 21 ± 5) versus rats treated with fish oil-ethanol for 6 weeks (group 2) (collagen 1.2 ± 0.5, pericellular fibrosis 16 ± 9). The amount of collagen and pericellular fibrosis in the rats treated with fish oil-ethanol for 6 weeks (group 2) were not different from the amount of collagen and pericellular fibrosis before dietary intervention at 6 weeks in groups 3 and 4. Each point represents an individual rat.

Dietary Modulation of Endotoxemia and Lipid Peroxidation. We also determined the effects of different diets on known mediators of hepatic injury (Table 3). Concentrations of endotoxin in plasma and hepatic TBARS and conjugated dienes decreased significantly after institution of the palm oil- and MCT-containing diets. Activity of CYP 2E1 and liver nonheme iron levels were also significantly decreased after administration of palm oil and MCT diets (Table 3).

Effects of Experimental Diets on Activation of NF-B. To evaluate the activation of NF-B, EMSA of nuclear extracts from whole liver were carried out. NF-B binding activity was increased in the fish oil-ethanol fed groups (Fig. 3). To prove specificity, a 100-fold excess of unlabeled NF-B or STAT (signal transducer and activator of transcription) oligonucleotide was added to an EMSA-binding reaction. Although addition of the NF-B oligonucleotide completely abrogated complex formation, addition of STAT 3 oligonucleotide had no effect (data not shown). Activation of NF-B was absent or barely detectable in the FE-PE and FE-MCTE groups. NF-B activation was absent in rats fed fish oil and dextrose.

View larger version (43K): [in this window] [in a new window]   Fig. 3.   Treatment with saturated fatty acids reverses ethanol-induced activation of NF-B. Representative electrophoretic mobility shift assay for NF-B and Western blot analysis for IkB. Increased NF-B binding activity was seen in the rats fed fish oil and ethanol (FE) (lanes 1-3). Binding activity was absent in the groups treated with palm oil-ethanol (FE-PE) (lanes 4-6), MCT-ethanol (FE-MCTE) (lanes 7-9), or fish oil-dextrose (FD) (lanes 10-12). IkB (molecular mass 37 kDa) in cytosol was decreased in the fish oil-ethanol group compared with other groups.

View this table: [in this window] [in a new window]   TABLE 4 Analysis of IB/, TNF-, and Cox-2 in different experimental groups

Effect of Experimental Diets on Cox-2 and TNF-. As mentioned above, we have proposed that lipid peroxidation and endotoxin activate NF-B which lead, in turn, to induction of TNF- and Cox-2 in alcoholic liver injury (Fig. 4). Thus, we also measured the effects of different diets on levels of mRNA for TNF-, Cox-2, Cox-1, and -actin. We have previously demonstrated that levels of TNF- and Cox-2 mRNAs are too low to be detected by Northern blot or ribonuclease protection assays (Nanji et al., 1997a). Reverse transcription-polymerase chain reaction was used for these measurements. Administration of fish oil-ethanol led to an increase in amounts of Cox-2 and TNF- mRNAs. In contrast, replacing the fish oil with palm oil (FE-PE) or MCT (FE-MCTE) led to marked down-regulation of both TNF- and Cox-2 (Table 4). Levels of mRNA for Cox-1, which is the constitutive isoform of the enzyme, were similar in all groups (Table 4).

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Schematic diagram of proposed mechanism of interaction between endotoxemia and lipid peroxidation in alcohol-induced liver injury. Endotoxin and lipid peroxides activate NF-B, which leads to up-regulation of TNF- and Cox-2. Synergism between TNF- and Cox-2 promotes liver injury.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Relationship of Rat Model of Alcohol-Induced Liver Injury to Human Alcoholic Liver Disease. The main problem in the management of alcoholic liver disease remains the inability of patients to abstain from alcohol. Many studies have been carried out, therefore, to determine the effectiveness of therapy in reducing the progression of liver injury in alcoholics (Fulton and McCullough, 1998). The current data are important in this regard because the experimental model used has striking similarities to the clinical setting in which alcoholic liver disease occurs.

Mechanisms by Which Saturated Fatty Acids Down-Regulate Pathological Events Triggered by Ethanol. It is well accepted that endotoxin and lipid peroxides are hepatotoxic and levels of both increase in alcohol-induced liver injury (Nanji et al., 1993; Adachi et al., 1995; Kaplowitz and Tsukamoto, 1996; Polavarapu et al., 1998). The observed differences in liver pathology in rats fed ethanol with fish oil and then switched to either palm oil or MCT can be explained, in part, by differences in levels of endotoxin and lipid peroxides. For example, rats fed fish oil and ethanol for either 6 or 8 weeks had the highest levels of endotoxin and lipid peroxidation. Animals treated with palm oil (FE-PE) or MCT (FE-MCTE) even with continued administration of ethanol had an approximately 60 to 80% reduction in endotoxin and lipid peroxide levels. In fact, in the FE-PE and FE-MCTE groups, the lipid peroxide levels were similar to those seen in the fish oil-dextrose-fed rats. The increase in endotoxin levels in ethanol-fed rats reflects a combination of increased gut permeability and a decreased ability of Kupffer cells to detoxify endotoxin (Nanji et al., 1993). The reduction in endotoxin levels after treatment with saturated fatty acids is likely a result of improvement in liver function although alterations in gut permeability cannot be excluded.

Role of NF-B, TNF-, and Cox-2. One pathway by which endotoxemia and lipid peroxidation act in concert to promote alcoholic liver injury is via NF-B. NF-B is a ubiquitous transcription factor that is implicated in the activation of many genes, including those involved in alcoholic liver injury (May and Ghosh, 1997; Nanji et al., 1999). The results of this study confirm our previous finding that activation of NF-B occurs in association with development of necroinflammatory changes in the liver (Nanji et al., 1999). In addition, we show that the reduction in necrosis and inflammation induced by treatment with saturated fatty acid is accompanied by a marked reduction of NF-B activation. The saturated fatty acid-induced inhibition of NF-B activation was accompanied by increased amounts of IB. The mechanism by which saturated fatty acids stabilize IB and suppress NF-B activation remains to be elucidated, but a role for decreased levels of endotoxin and lipid peroxidation is likely. Although we did not, in the present study, identify the specific cell types expressing NF-B, based on our previous study (Nanji et al., 1999) and those of others (Lin et al., 1997), we expect that the Kupffer cell is the major cell type showing activation of NF-B. We cannot, however, exclude the contribution of hepatocytes, and endothelial and stellate cells to activation of NF-B in liver. In contrast to the proinflammatory cascade of genes expressed in response to activation of NF-B in Kupffer cells, activation of NF-B in hepatocytes may serve a hepatoprotective role, by stimulating cellular regeneration and inhibition of apoptosis (Schmid and Adler, 2000). Thus, the cellular site of activation of NF-B may determine the balance between necroinflammatory changes and hepatoprotection.

Relevance of Observed Changes to Human Alcoholic Liver Disease. The development of an appropriate animal model for alcoholic liver disease has been difficult because multiple factors interact, leading to pathological changes characteristic of alcoholic liver disease. The liver pathology in rats fed ethanol intragastrically has been compared with histological findings in human alcoholic liver disease (Hall et al., 2001). Hepatomegaly, increased levels of transaminases in serum, balloon degeneration, apoptosis, megamitochondria, necrosis, inflammation, and fibrosis are seen in both clinical and experimental alcoholic liver injury. Also of note is that cirrhosis is not seen in rats administered ethanol unless carbonyl iron is supplemented in the diet (Tsukamoto et al., 1995).