Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Evidence shows that dietary lipid can modulate the severity of alcohol-induced liver injury. None of the histologic features of alcoholic liver injury develop in rats fed ethanol and saturated lipid, whereas fatty liver, necrosis, inflammation and fibrosis develop in rats fed ethanol and lipids enriched in polyunsaturated fatty acids (Nanji et al., 1989; Nanji and French, 1989). One of the polyunsaturated fatty acids that promotes alcoholic liver injury is linoleic acid (Nanji and French, 1989). To gain a better understanding of the role of linoleic acid in ALD, we investigated the metabolism of linoleic acid to arachidonic acid and its subsequent conversion to eicosanoids in the intragastric feeding rat model for ALD (Nanji, 1993). Our studies showed that increased production of TXB₂ and decreased production of prostaglandin E₂ by liver nonparenchymal cells correlated with the presence of pathologic liver injury in ethanol-fed rats (Nanji et al., 1994b). Additionally, a significant correlation was observed between plasma levels of TXB₂ and severity of liver injury in rats fed corn oil and ethanol (Nanji et al., 1993b). A significant correlation was also seen between pathologic severity and levels of endotoxin in plasma (Nanji et al., 1993b). The Kupffer cell is believed to be the major site of thromboxane production in the liver in response to stimuli such endotoxin (Decker, 1990; Winwood and Arthur, 1993). A role for thromboxanes in promoting necroinflammatory changes in alcoholic liver injury is further supported by recent observations that show that inhibition of thromboxane synthesis or blocking the action of thromboxane at the receptor level leads to a significant decrease in the degree of necrosis and inflammation in corn oil-ethanol-fed rats (Nanji et al., 1997).

In the current study, we extended our observations relating thromboxanes to ALD. First, we investigated the cell type in the liver responsible for enhanced thromboxane synthesis in ALD. We used immunohistochemical analysis to identify thromboxane synthase in liver tissue, and the RT-PCR to assess the relationship between TX synthase mRNA and liver levels of TXB₂ and its metabolites. To further define the cell type and mechanisms involved in enhanced TX synthesis, we isolated the individual cell types from the livers of rats fed ethanol or dextrose with either saturated fat or corn oil. In each group, TX synthase mRNA expression in whole liver and the different cell types was evaluated and related to the immunohistochemical analyses for TX synthase and the presence of pathologic liver injury.

Some of the actions of TXA₂ which are likely to be important in the pathogenesis of ALD include a reduction in hepatic blood flow, platelet aggregation and formation of plasma membrane blebs on hepatocytes (Oates et al., 1988; Smith, 1992). There is overwhelming evidence that shows that many of these biological actions of TXA₂ are mediated by its specific receptor on the cell surface (Armstrong and Wilson, 1995; Morinelli and Halushka, 1991; Negishi et al., 1993). In view of the potential importance of thromboxanes in liver injury, it is of interest to investigate the role of the TXA₂-receptor system in ALD. TXA₂ receptors are found in a variety of tissues and cell types in mammals (Negishi et al., 1993). In the liver, recent investigations have identified TXA₂-receptors on the hepatic sinusoidal endothelial cells (Ishigoru et al., 1994). Activation of these receptors leads to alterations in the sinusoidal microcirculation which is important in the pathogenesis of ALD (Lieber, 1994; Tsukamoto et al., 1990). To further define the role of TXA₂-receptors in ALD, we used RT-PCR to investigate the changes in TXA₂-receptor mRNA concentrations in the livers and individual cell types in rats fed different dietary fats with either ethanol or dextrose. Alteration in TXA₂-receptor mRNA levels were related to pathologic liver injury.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animal model and treatment groups. Male Wistar rats weighing between 225 and 250 g were fed a liquid diet via permanently implanted gastric cannulas as described previously (Tsukamoto et al., 1990). The rats were administered their total nutrient intake by intragastric infusion. The rats were fed freshly prepared diets and either MCT or corn oil provided the fatty acids which contributed 35% of total calories. The fatty acid composition of the diet has been described previously (Nanji et al., 1994a). Vitamins and minerals were given as described previously (French et al., 1993). The liquid diet was infused at a rate of 180 ml/kg b.wt./day to achieve adequate weight gain (1 ml = 1 kcal). Ethanol was infused to maintain blood alcohol levels between 150 and 300 mg/dl (33-66 µmol/l). The amount was initially 10 g/kg/day and was increased up to 16 g/kg/day as tolerance developed. All animals received humane care in compliance with the National Institutes of Health criteria for care of laboratory animals.

Isolation of Kupffer cells, endothelial cells and hepatocytes. Cells were isolated from anesthetized rats by previously described procedures (Petrick et al., 1994; Toth et al., 1985) and isolation buffers described by Seglen (1973). After intravenous administration of sodium heparin (100 U), the livers were exsanguinated in situ by portal vein perfusion with Ca⁺⁺-free buffer. Livers were excised, minced and subsequently incubated in 0.05% collagenase in buffer (0.1 M HEPES with 0.39% sodium chloride, 0.05% potassium chloride, 0.05 M calcium chloride, pH 7.6) at 37°C for 45 min. The resulting cell suspension was strained, pelleted and resuspended in fresh collagenase buffer with gentle shaking for 30 min. The cell suspension was centrifuged (×3) at 50 × g for 2 min. The pelleted fraction contained mainly hepatocytes. The cells remaining after centrifugation were washed several times with Hanks' balanced salt solution and centrifuged in a 17.5% metrizamide gradient in Hanks' balanced salt solution. This fraction contained approximately 65% Kupffer cells with the balance being liver endothelial and stellate cells. Further purification of Kupffer cells was done by incubation for 2 hr in 48-well tissue culture dishes at 37°C in a humid atmosphere of 5% carbon dioxide for 16 hr. The medium consisted of RPMI-1640, 10% fetal bovine serum with 2 mM L-glutamine and 50 µg/ml penicillin/streptomycin. Adherent cells formed a monolayer on the culture dish and >85% of these cells were macrophages. The different cell types were identified by morphology, and immunohistochemical markers which included peroxidase, acid phosphatase, ₁-antichymotrypsin, ED1 and cytokeratins (Alpini et al., 1994). Endothelial cells were the nonadherent cells from the Kupffer cell preparation and were plated onto type I collagen-coated dishes. The cells were stored frozen at 70°C.

Histopathology. A small sample of the liver was obtained when the rats were sacrificed and formalin-fixed. 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) was scored 1+ when <25% of the cells contained fat; 2+, with 26 to 50% fat; 3+, with 51 to 75% fat; and 4+, with >75% fat. Necrosis was evaluated as the number of necrotic foci per square millimeter, and 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 the sections was unaware of the treatment groups when assessing the histology.

Blood alcohol levels. Blood was collected from the tail vein, and ethanol concentration was measured with an alcohol dehydrogenase kit from Sigma Chemical Co. (St. Louis MO).

TXB₂ and 2,3-dinor-TXB₂ levels in liver. Because the measurement of TXB₂, the chemically stable hydrolysis product of TXA₂, is subject to artifactual increases (Lawson et al., 1986), it has been suggested that measurement of a longer-lived stable metabolite such as 2,3-dinor-TXB₂ represents a better measure of thromboxane production (Lawson et al., 1986). Approximately 1 g of liver from the rats in each of the different experimental groups treated for 4 weeks was rapidly homogenized in 10 ml of ice-cold methanol for 30 sec. After centrifugation, the supernatant was dried and resuspended in 0.1 mol/l of potassium phosphate buffer (pH 7.4) and purified by elutriation through an octadecyl silyl SEP-PAK C18 cartridge (Waters Associates, Milford, MA). The 80% methanol eluent was assayed for TXB₂ and 2,3-dinor-TXB₂ (Cayman Chemical, Ann Arbor, MI).

RNA extraction from liver tissue and cells and analysis of mRNA for TX synthase, TXA₂-receptors and -actin by RT-PCR. To examine the expression of TX synthase, TXA₂-receptors and -actin in both liver tissue and cells, total RNA was isolated according to the guanidium isothiocyanate method (Chomczynski and Sacchi, 1987). The total RNA concentration of each sample was determined from absorbance at 260 nm, and the quality of each RNA preparation was determined by agarose-formaldehyde gel electrophoresis and ethidium bromide staining. We reverse-transcribed 0.5 to1 µg of total RNA by adding 30 µl of a master mix with reverse transcriptase buffer (0.6 mmol/l MgCl₂, 15 mmol/l KCl, 10 mmol/l Tris HCl [pH 8.3]), 40 pmol of downstream primer, 0.5 mmol/L dNTP mixture, 1 U/µl RNase inhibitor and 13.3 U/µl Moloney murine leukemia virus reverse transcriptase (GIBCO-BRL, Grand Island, NY) (final concentrations indicated). Samples were incubated, first for 60 min at 42°C and then at 75°C for 10 min, then chilled on ice. Then 2 µl of each sample was added to 20 µl of 1.5 mmol/l MgCl₂, 50 mmol/l KCl, 10 mmol/l Tris HCl (pH 8.3), 0.2 mmol/l of each dNTP and 0.01% gelatin, 5 U/100 µl Taq DNA polymerase (Perkin Elmer Cetus) and 50 pmol of sense primer and 10 pmol of antisense primer. The sequences of primer pairs, 5 and 3, and predicted sizes of the amplified PCR fragments are shown in table 1. Amplification was performed in an automated thermal cycler at 94°C for 60 sec, 50°C for 90 sec and 72°C for 2 min for 35 cycles, followed by 72°C for 10 min. To measure the efficiency of the extraction of RNA and of reverse transcription, we amplified 2 µl of the same reverse transcriptase reaction with -actin-specific primers as an internal control. PCR products and molecular weight markers were subjected to electrophoresis on 1% agarose gels and visualized by means of ethidium bromide staining. Location of the predicted PCR products was confirmed by using a 100-base pair ladder (GIBCO-BRL) as a standard size marker. For quantitation, the expression of the products was quantitated using densitometric scan analysis. The index of the various mRNA signals was standardized against that of the -actin signal from the same RNA.

TXA₂ synthase in liver by immunohistochemistry. For identification of cells staining for TX synthase, 6-µm-thick sections were prepared from paraffin blocks, after deparaffinization through graded ethanol. The sections were then washed in phosphate-buffered saline. Immunocytochemical staining for TX synthase was performed by an antibody against thromboxane synthase (Cayman Chemical, Ann Arbor, MI) and by the avidin-biotin complex method (Vector Laboratories, Burlingame, CA). The number of positively stained cells were counted and the numbers expressed as cells/mm². The nature of positively stained sinusoidal cell was further defined by characteristic morphology and staining with vimentin which in these animals stained Kupffer cells (Marugg et al., 1990; Nanji et al., 1996). These cells failed to stain positively for desmin.

Statistical analysis. Analysis of variance and multiple comparison with the Student-Neuman-Keuls method were used for determination of statistical significance. Pearson's correlation coefficient "r" was used for evaluation of associations.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

There was no significant difference in the amount of weight gained in the different experimental groups. There was also no significant difference in blood alcohol levels (mg/dl) (mean ± S.E.) in the ethanol-fed groups at 4 and 8 weeks. At 4 weeks, the blood alcohol levels were: MCTE, 216 ± 39 (47 ± 8.5 µmol/l); CE, 231 ± 41 (50 ± 9 µmol/l); at 8 weeks they were: MCTE, 241 ± 32 (52 ± 7 µmol/l); CE, 251 ± 30 (54 ± 6.5 µmol/l).

Histopathology. Only rats fed corn oil and ethanol for 4 and 8 weeks developed pathologic liver injury (table 2, fig. 1). None of dextrose-fed rats or rats in the MCTE group developed pathologic changes (fig. 2).

View larger version (168K): [in this window] [in a new window]   Fig. 1.   Liver from a rat fed corn oil and ethanol for 8 weeks showing the presence of fat, necrosis and inflammation (arrow). The necrosis and inflammation in these animals is focal in nature. Hematoxylin and eosin, ×155.

View larger version (155K): [in this window] [in a new window]   Fig. 2.   Liver from a rat fed MCT and ethanol for 8 weeks. There is no evidence of fatty liver, necrosis or inflammation. Hematoxylin and eosin, ×155.

Liver thromboxane synthase (mRNA and immunohistochemistry) and liver TXB₂ and 2,3-dinor-TXB₂ levels. The number of sinusoidal lining cells, identified as Kupffer cells based on morphologic characteristics, that stained positive with the antibody against thromboxane synthase, were increased in rats fed corn oil and ethanol (table 3, fig. 3). The number of positively stained cells increased significantly between 2 and 4 weeks in the CE group (P < .01, table 3). An increase in TX synthase positive cells was also seen in the MCTE group compared with the dextrose-fed control group; however, the degree of increase was not as great as was seen in the CE group in the same time periods (fig. 4). Occasional positively staining cells (<0.1/mm²) were seen in rat livers before feeding.

View larger version (117K): [in this window] [in a new window]   Fig. 3.   Liver from a rat fed corn oil and ethanol for 8 weeks stained for TX synthase. Note the positive staining in Kupffer cells lining the hepatic sinusoid (arrow).

View larger version (108K): [in this window] [in a new window]   Fig. 4.   Liver stained for TX synthase from a rat fed MCT and ethanol for 8 weeks. There are occasional Kupffer cells that stain positively for TX synthase.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Numbers of TX synthase positive cells/mm2 (A), TX synthase mRNA (B) and liver levels of TXB2 and 2,3-dinor-TXB2 (C) in rats sacrificed at 4 weeks from the different experimental groups. The number of TX synthase positive cells was significantly higher (P < .01) in the corn oil-ethanol group (9.8 ± 5.0/mm2) versus the dextrose-fed control group (CD, 0.3 ± 0.2) and rats fed MCTE (2.1 ± 1.0). The number of cells in MCTE were higher (P < .01) than MCTD (0.2 ± 0.2). TX synthase mRNA was highest in CE (6.0 ± 0.9 arbitrary units) compared with CD (1.0), MCTE (2.8 ± 1.1) and MCTD (0.2 ± 0.05) (fig. 6B). The increase in liver levels of TXB2 and 2,3-dinor-TXB2 in the CE group (fig. 6C) reflected the increased expression of TX synthase. The level of TXB2 (pg/mg protein) (212 ± 36) and 2,3-dinor-TXB2 (239 ± 44) was higher (P < .01) than in the CD group (64 ± 21 and 64 ± 19), MCTE group (64 ± 21 and 73 ± 26) and MCTD group (17 ± 9 and 16 ± 6). The levels in MCTE were significantly greater than in the MCTD group (P < .01).

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Correlation between TX synthase mRNA and the number of TX synthase positive cells/mm2 (r = 0.94, P < .01). TX synthase mRNA was measured in whole liver by RT-PCR and assessed densitometrically (see "Materials and Methods"). When the saturated fat (MCT) and corn oil-treated groups were assessed separately, the correlation between the two parameters was significant in each group. r = 0.90, P < .01 in the MCT group; r = 0.95 in the corn oil group.

Alterations in mRNA levels for TXA₂-receptor. The highest concentrations of TXA₂-receptor levels (normalized for -actin) evaluated by RT-PCR were seen in the livers of animals in the CE group. The levels were about five times higher in the CE group than in the corn oil-dextrose-fed controls (P < .01) (fig. 7). In the CE group, the highest expression of TXA₂-receptor was seen in the Kupffer cells with lesser levels of induction in the endothelial cells and hepatocytes. Because the isolation procedure yielded cells with 80 to 85% purity, we cannot exclude the possibility that the low levels of TXA₂-receptor mRNA in endothelial cells and hepatocytes reflected contamination by Kupffer cells. A 2-fold increase in endothelial cell TXA₂-receptor mRNA levels and a 4-fold increase in Kupffer cell TXA₂-receptor mRNA levels was seen in the MCTE group compared with the MCT-dextrose-fed controls (P < .05). Thus ethanol, independent of the source of dietary fatty acids increased the level of TXA₂-receptor mRNA expression in liver.

View larger version (33K): [in this window] [in a new window]   Fig. 7.   Densitometric analysis of TXA2-receptor mRNA (normalized for -actin) in the liver and individual cell types in the different experimental groups. The livers were obtained from rats sacrificed at 4 weeks. The TXA2-receptor levels are significantly higher (P < .01) in the liver and various cell types in the CE-fed rats than in the other treatment groups. The fold-increase in the different groups was related to the corn oil-dextrose group as control. The TXA2-receptor mRNA levels in the liver were significantly higher in the CE group (4.6 ± 1.7) than in the MCTD (0.3 ± 0.03), MCTE (0.9 ± 0.2) and CD (1.0) groups. TXA2-receptor mRNA levels were the highest in the Kupffer cells in the CE group (13.0 ± 4.0). Ethanol administration in the MCT group also significantly (P < .01) increased TXA2-receptor mRNA levels in Kupffer cells.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We have previously shown that rats fed polyunsaturated fatty acids with ethanol develop pathologic liver injury, whereas rats fed saturated fatty acids with ethanol are relatively protected (Nanji et al., 1989; Nanji and French, 1989). In corn oil-ethanol-fed rats, the polyunsaturated fatty acid responsible for inducing liver injury is linoleic acid (Nanji and French, 1989). In subsequent studies, we determined that conversion of linoleic acid to arachidonic acid led to the enhanced synthesis of thromboxane by nonparenchymal liver cells (Nanji et al., 1993b). Furthermore, plasma levels of thromboxane B₂ correlated with the severity of liver injury. These observations led to the hypothesis that thromboxane(s) were important in the pathogenesis of alcoholic liver injury. In this respect, it was of interest to identify the cell type(s) responsible for increased synthesis of thromboxanes. The present study identified the Kupffer cell as the most likely source of the enhanced synthesis of thromboxane A₂ in ethanol-fed rats. The stimulus for enhanced synthesis of TXA₂ by Kupffer cells is probably endotoxin which originates from the cell wall of gram-negative bacteria and is increased in the plasma of rats fed corn oil and ethanol (Nanji et al., 1993b; Adachi et al., 1995). Kupffer cells are the primary effector cells in the liver which respond to endotoxin by producing many inflammatory mediators (Decker, 1990). Arachidonic acid metabolites, such as thromboxanes, are prominent mediators produced by macrophages in response to endotoxin (Decker, 1990). The importance of Kupffer cell-derived mediators in the pathogenesis of alcoholic liver injury in the intragastric feeding rat model is further demonstrated by the study of Adachi et al. (1994) who showed that inactivation of Kupffer cells by gadolinium chloride prevented alcoholic liver injury.

The mechanism(s) by which thromboxane contributes to liver injury is unknown. TXA₂ is a potent vasoconstrictor and platelet aggregatory agent (Hamberg et al., 1975; Morinelli and Halushka, 1991). Vasoconstriction of the hepatic sinusoids can aggravate the hypoxia already caused by enhanced oxygen consumption and impaired oxygen utilization seen in ethanol-fed animals (Israel and Orrego, 1987; Lieber et al., 1989; Tsukamoto and Xi, 1989). Platelet aggregation leads to release of secretory products which cause cell injury (Schror and Braun, 1990). TXB₂ causes bleb formation in isolated hepatocytes (Horton and Wood, 1990); the formation of plasma membrane blebs is a consequence of toxic or ischemic cell injury (Gores et al., 1990). The actions of TXA₂ are generally believed to be mediated by specific TXA₂ receptors (Negishi et al., 1993). The TXA₂-TXA₂ receptor system is one of the factors involved in the regulation of the microcirculation in the hepatic sinusoid (Ishigoru et al., 1994). We found increased levels of TXA₂-receptors in all cell types studied, i.e., Kupffer cells, endothelial cells and hepatocytes, in rats fed corn oil and ethanol. The largest increase was seen in the Kupffer cells, with lesser degrees of increase seen in endothelial cells and hepatocytes. TXA₂ receptors have been detected on human (Halushka et al., 1989) and equine (Simmons et al., 1993) monocytes and it is believed that these receptors may function as autocrine regulators of cytokine synthesis. In line with this suggestion is the finding of increased levels of TNF in livers of rats fed corn oil and ethanol (Nanji et al., 1994d). Also, inhibition of thromboxane synthesis and action at the receptor levels leads to down-regulation of TNF- mRNA (Nanji et al., 1997). Ishigoru et al. (1994) showed that sinusoidal endothelial cells in rats have a single class of TXA₂ binding sites and that the number of binding sites is similar to that seen in endothelial cells of rat aorta. In endotoxin-treated rats, these investigators showed a reduction in TXA₂-receptors on sinusoidal endothelial cells and suggested that when TXA₂ binds to the receptor, the ligand-receptor complex is internalized , which results in a reduction in the number of cell surface receptors. Our study in which we assessed TXA₂ receptor-mRNA levels rather than surface binding of TXA₂, we observed an increase in TXA₂-receptor mRNA in the presence of increased thromboxane levels. It is quite conceivable, that under conditions of thromboxane overproduction, the synthesis of receptor mRNA may be increased in response to loss of surface receptors secondary to internalization. Down-regulation of TXA₂-receptor binding sites in response to increased TXA₂ production has been described. For example, in diabetic rats, down-regulation of TXA₂-receptor sites occurs in the glomeruli and mesangial cells (Wilkes et al., 1992). Thus down-regulation of TXA₂-receptor binding may be a generalized feature of TXA₂ overproduction (Spurney et al., 1994) necessitating increased synthesis of the receptor. TXA₂ stimulation of endothelial cells, in addition to causing microcirculatory disturbances, also leads to increased expression of genes necessary for endothelial cell proliferation (Kent et al., 1993). In this respect, it is of interest to note that the number of proliferating hepatic sinusoidal endothelial cells are increased in rats fed corn oil and ethanol (Nanji et al., 1994c). The relevance of this observation to ALD remains to be established. We also identified TXA₂ receptor mRNA in hepatocytes, which suggests that TXA₂ binding sites are also present on the hepatocyte surface. This observation may account for the ability of TXB₂ to cause hepatocyte blebbing (Horton and Wood, 1990). Whether up-regulation of receptor synthesis contributes to hepatocyte necrosis, as in corn oil-ethanol-fed rats, remains to be studied.

In summary, our results show that the Kupffer cell is the most probable source of the enhanced production of TXA₂ in experimental ALD. Increased levels of thromboxane metabolites have also been seen in the portal vein in human liver disease including ALD (Garcia-Valdecasas et al., 1995). However, other measurements such as endotoxin levels were not carried out and the cell type responsible for the production of thromboxane was not identified. Our findings in an experimental model of alcoholic liver injury lend credence to the hypothesis that the combination of enhanced production of TXA₂ and enhanced synthesis of TXA₂-receptors in the various cell types in the liver may contribute to enhanced cytokine production by Kupffer cells, microcirculatory disturbances and hepatocyte necrosis. This view is further supported by previous studies that show that inhibition of thromboxane synthesis or thromboxane action at the receptor level reduces the expression of TNF- in the liver and the severity of necroinflammatory changes in experimental ALD.