Soybean polyenylphosphatidylcholine (PPC)⁵ is a mixture of 94-96% of polyunsaturated phosphatidylcholines, about half of which is dilinoleoylphosphatidylcholine (Lieber et al. 1994). At variance with mammalian phospholipids, this plant phospholipid contains two unsaturated fatty acids in both the 1 and 2 positions of the glycerol backbone, which confers a high bioavailability, mainly because of the reacylation of the unsaturated 1-acyl-lysophosphatidylcholine with additional unsaturated fatty acids during intestinal absorption (Zierenberg and Grundy 1982). Chronic administration of PPC prevents the development of septal fibrosis in alcohol-fed baboons (Lieber et al. 1990 and 1994). Moreover, dilinoleoylphosphatidylcholine specifically stimulates collagenase activity in cultured stellate cells (Lieber et al. 1994), an effect that may contribute to prevention of the fibrosis in alcoholic as well as in nonalcoholic forms of liver injury (Ma et al. 1996).

In addition to its antifibrogenic effects, PPC could also affect earlier changes induced by alcohol consumption, such as those occurring at the fatty liver stage. The lack of significant differences in hepatic triglyceride concentration between baboons with fibrosis or cirrhosis and those without (because of the treatement with PPC) does not preclude earlier differences in the rate of hepatic lipid accumulation. Indeed, hepatic triglyceride concentration was shown to decrease with the development of fibrosis and cirrhosis (Savolainen et al. 1984). Moreover, beneficial effects of PPC have been reported in the recovery of alcoholics with fatty liver (Knüchel 1979, Schüller Pérez and González San Martin 1985). Therefore, this study was undertaken to determine whether dietary supplementation with PPC affects the development of alcoholic fatty liver and to investigate the mechanisms for such an effect.

MATERIALS AND METHODS

The diets were supplemented either with PPC (3 g/L) or with additional safflower oil and choline bitartrate (1 g/L) to provide the same amount of linoleic acid (2.1 g/L) and choline base (0.483 g/L). Thus, the concentration of choline base in both diets was 0.75 g/L, which is three times higher than that in the unsupplemented diets. The fatty acid composition (g/100 g fatty acids) of both diets consisted of 12.4 palmitic, 0.9 palmitoleic, 3.2 stearic, 56.6 oleic, 26.8 linoleic and 0.12 linolenic acids. This represents a 15.2% increase in linoleic acid, compared with the the nonsupplemented diets (Kim et al.1988, Lieber and DeCarli 1994). The diets were prepared twice a week, and the stability of linoleate and PPC in the diets was verified by gas liquid chromatography and HPLC, respectively. PPC (containing 40-52% dilinoleoylphosphatidylcholine) was kindly provided by Rhône-Poulenc Rorer GmbH (Köln, Germany).

All other reagents were purchased from Sigma Chemical (St. Louis, MO) and were at least 98% pure. All of the organic solvents were of a chromatographic purity grade.

After reaching a weight of 130-170 g, 28 rats were divided in groups of four (alcohol without PPC, control without PPC, alcohol with PPC, control with PPC). For each dietary treatment (with or without PPC), the rats fed control diets (without ethanol) were pair-fed daily (on an isoenergetic basis) with the corresponding littermates fed the ethanol-containing diet, for 21 d. The ethanol concentration in the diet was increased gradually, starting with 30 g/L for 2 d, 40 g/L for the next 2 d, and reaching the final concentration of 50 g/L on d 5. Half of the pairs received the diets with PPC and the other half were given instead the diets supplemented with safflower oil and choline. To assess for possible effects of these supplements, we compared a limited number of measurements (hepatic triglyceride accumulation and serum enzymes) with an additional group of 18 littermates of similar weight, pair-fed alcohol-containing and control diets without supplementation with safflower oil and choline and subjected to the same experimental procedures.

During the last week of pair-feeding, feces were collected on a wire mesh placed 5 cm below the cage for 4-6 d, weighed and stored at 70°C until analysis. On the day preceding the killing, differences in feeding pattern were minimized by giving one third of the usual daily ration of the corresponding diets in the morning and two thirds in the evening. On the day of the killing, 6 mL of the corresponding diets per 100 g body weight was given by gastric tube 90 min before blood collection. In the case of ethanol-containing diets, this intragastric feeding corresponded to a short-term ethanol administration of 3 g/kg, which is equivalent to one fifth of the daily alcohol intake.

The rats were killed while under pentobarbital anesthesia (40 mg/kg, intraperitoneally) by collecting the blood from the aorta into EDTA-containing tubes. Livers were excised, weighed and analyzed as described below.

Total liver lipids were extracted according to Folch et al. (1957). After separation by thin layer chromatography (Amenta 1964), triglycerides were measured colorimetrically by the method of Snyder and Stephens (1959). The cholesterol esters were calculated from the differences between total and free cholesterol, measured by the colorimetric method of Searcy and Berquist (1960). Total phospholipids were measured by the colorimetric method of Bartlett (1959). Protein concentrations were determined by the colorimetric method of Lowry et al. (1951), with bovine serum albumin as the standard.

Plasma was obtained from blood collected in tubes containing EDTA (1 g/L) by centrifugation at 1200 × g for 20 min. Ethanol was measured in the plasma by head-space gas chromatography (Korsten et al. 1975). Triglycerides were assayed enzymatically with glycerol kinase after lipase hydrolysis (Wahlefeld 1974), with the use of a commercially available kit (Triglyceride kit no. 336, Sigma Chemical). Plasma free fatty acids were determined by the colorimetric method of Novák (1965). Serum alanine-aminotransferase (ALT, EC 2.6.1.2) and aspartate aminotransferase (AST, EC 2.6.1.1) activities were measured by the enzymatic method of Bergmeyer et al. (1978), with the use of commercially available kits (ALT kit no. 59-UV, and AST kit no. 58-UV, Sigma Chemical). Serum glutamate dehydrogenase (GDH, EC 1.4.1.3) activity was determined enzymatically by the method of Ellis and Goldberg (1972).

Total fecal fatty acids were measured by titration after saponification, according to the procedure of van de Kamer et al. (1949).

RESULTS

Compared with their pair-fed controls, both groups of alcohol-fed rats developed hepatomegaly characterized by significantly higher liver lipid and protein contents (Table 1). The supplementation with PPC did not alter these values in the pair-fed controls; however, it significantly (P < 0.01) lowered (by approximately half ) the ethanol-induced increases in liver weight (from 1.29 ± 0.13 to 0.56 ± 0.15 g/100 g body weight with PPC), lipids (from 324 ± 46 to 175 ± 27 mg/100 g body weight) and proteins (from 404 ± 59 to 222 ± 66 mg/100 g body weight).

Table 1. Effects of polyenylphosphatidylcholine (PPC) on the ethanol-induced increases in liver weight, lipid and protein contents in rats pair-fed control or ethanol-containing diets1 [View Table]

The hepatic concentration of triglycerides (expressed per gram of liver), and to a smaller extent, that of cholesterol esters, were greater in rats fed the ethanol-containing diets than in controls, whereas there were no significant differences in the concentration of total phospholipids (Table 2). Supplementation with PPC did not affect the concentration of total phospholipids, but significantly attenuated the ethanolinduced rise in triglyceride and cholesterol ester concentrations. PPC did not affect significantly the lipid concentrations in the livers of the controls. The alcohol-induced rises in concentration reflected much larger elevations in total liver contents when the alcoholic hepatomegaly was taken into account.

Table 2. Effects of polyenylphopshatidylcholine (PPC) on the ethanol-induced changes in liver lipid components in rats pair-fed control or ethanol-containing diets1 [View Table]

The ethanol-induced accumulation of triglycerides in the liver of rats supplemented with safflower oil and choline did not differ significantly (P = 0.143) from that in a separate group of rats without safflower oil and choline supplementation (361 ± 41 in the ethanol-fed vs. 46 ± 8 mg/100 g body weight in the pair-fed controls; P < 0.01), despite a slightly, but not significantly (P = 0.1462) lower ethanol consumption [14.3 g/(kg·d)] in the latter group.

The ethanol-induced accumulation of lipids in the liver was associated with postprandial hypertriglyceridemia, which was significantly lower when the alcohol was given with PPC than without (Fig. 1). PPC did not affect significantly the serum triglycerides of the pair-fed controls. When compared with the respective controls, plasma free fatty acids were higher in rats fed ethanol either with PPC (1.15 ± 0.23 vs. 0.70 ± 0.08 mmol/L; P < 0.05) or without PPC (1.21 ± 0.19 vs. 0.63 ± 0.09 mmol/L; P < 0.05). PPC did not affect the ethanol-induced rise in free fatty acids.

Ethanol feeding did not affect either the concentration of fat in the feces or the daily fecal fat excretion, whereas supplementation with PPC lowered the fecal fat excretion in both the alcohol-fed and the pair-fed control rats (Table 3).

Table 3. Effects of polyenylphosphatidylcholine (PPC) and ethanol feeding on fecal fat excretion in rats pair-fed control or ethanol-containing diets1 [View Table]

Without PPC, the capacity of the hepatic mitochondria to oxidize palmitoyl-1-carnitine was significantly lower in the ethanol-fed rats than in the controls. The changes were particularly manifest when judged by the respiratory control ratio between the ADP-stimulated and nonstimulated O₂ consumption (Fig. 2). This was associated with a lesser oxidation of substrates entering the site 1 (glutamate) and site 2 (succinate) of the electron transport chain. PPC supplementation did not affect the mitochondrial respiration in the control rats, whereas it prevented the alterations produced by chronic ethanol administration. When palmitoyl-1-carnitine was used as substrate, ethanol feeding lowered the ADP:O ratio (a measure of efficiency of coupling between oxidation and ATP production), but the lowering was prevented by PPC (Table 4).

Table 4. Effects of polyenylphosphatidylcholine (PPC) and ethanol feeding on the efficiency of coupling between mitochondrial oxidation and ATP production in rats pair-fed control or ethanol-containing diets1 [View Table]

Without PPC, the activity of liver mitochondrial cytochrome oxidase (a membrane-bound enzyme) was lower in ethanol-fed rats than in controls, but such a difference was abolished by the supplementation with PPC (Table 5). The activity of soluble mitochondrial enzymes, such as glutamate dehydrogenase, as well as that of succinate dehydrogenase, was not significantly affected by either the feeding of ethanol or the administration of PPC. The recovery of mitochondria (judged from the ratio between the specific activities of the enzymes in the homogenate and the mitochondria) was not affected by the treatments.

Table 5. Effects of polyenylphosphatidylcholine (PPC) and ethanol feeding on the recovery and specific activities of liver mitochondrial enzymes in rats pair-fed control or ethanol-containing diets1 [View Table]

The activities of serum alanine and aspartate aminotransferases, as well as that of serum glutamate dehydrogenase, were significantly higher in the ethanol-fed rats than in the pair-fed controls (Table 6). Supplementation with PPC did not affect the activities of these enzymes in the control rats, but prevented the alterations produced by ethanol administration.

Table 6. Effects of polyenylphosphatidylcholine (PPC) and ethanol feeding on serum aminotransferase and glutamate dehydrogenase activities in rats pair-fed control or ethanol-containing diets1 [View Table]

The changes in serum enzymes produced by ethanol in the rats fed the diets supplemented with safflower oil and choline did not differ significantly from those measured in a separate group of rats without these supplements. In the nonsupplemented rats, serum AST in the ethanol-fed animals was 1.01 ± 0.15 vs. 0.60 ± 0.03 µkat/L in the controls; serum ALT in the ethanol-fed rats was 0.56 ± 0.21 vs. 0.34 ± 0.03 µkat/L in the controls; and serum GDH in ethanol-fed rats was 15.82 ± 3.22 vs. 4.28 ± 0.54 U/L in the pair-fed controls.

DISCUSSION

The administration of PPC provides a supplement of choline and of the essential fatty acid, linoleate. To control for this, the diets without PPC were supplemented with equivalent amounts of linoleate and choline. In the rats fed the latter diet, the hepatic concentrations of triglycerides and (to a lesser extent) cholesterol esters were markedly higher in the ethanol-fed than in the control rats. The magnitude of the changes produced by the diets supplemented with linoleate and choline was comparable to those obtained in 18 rats of similar body weight that were pair-fed the same diet without the addition of linoleate and choline and subjected to the same experimental and analytical procedures. This indicates that the supplementation of these nutrients does not affect the development of alcoholic fatty liver. In baboons fed diets supplemented with a fivefold increment in choline for several years, there were signs of toxicity in the control animals, with electron microscopic changes and elevation of aminotransferase and glutamate dehydrogenase activities in the serum (Lieber et al.1985). This was not the case in our rats fed a diet with a threefold increment in choline for 3 wk; the serum activities of aminotransferases and glutamate dehydrogenase (Table 6) were not different than those in a separate group of rats fed the diet without additional choline (data not shown in the table).

By contrast, the supplementation of the diet with PPC significantly attenuated the ethanol-induced increases in triglycerides and cholesterol esters, whereas it had no significant effect on the controls. This indicates that PPC attenuates the ethanol-induced impairments of lipid metabolism. Moreover, these effects of PPC cannot be attributed simply to a larger intake of essential fatty acids or choline, but must be attributed to a more specific effect of the phospholipid administered, which is characterized by high bioavailability and is largely incorporated into cell membranes (Lieber et al. 1994).

The present findings suggest that PPC does not attenuate alcoholic fatty liver by reducing the supply of fat to the liver or by increasing the disposition of liver triglycerides as serum lipoproteins. With regard to the supply of dietary fat to the liver, not only was the consumption of dietary fatty acids the same in all rats, but PPC tended to enhance rather than lower fat absorption, as indicated by a lesser excretion of fatty acids in the feces (Table 3). The latter finding is consistent with observations indicating that phosphatidylcholine facilitates the formation of lymph chylomicrons (O'Doherty et al. 1973). Furthermore, the possibility that PPC could have decreased the mobilization of fatty acids from adipose tissue seems unlikely because the plasma free fatty acid concentrations in alcohol-fed rats did not differ in those fed diets with and without PPC. Finally, the possibility that PPC could have reduced the uptake of chylomicron remnants is also unlikely because this would have resulted in exacerbation rather than attenuation of the postprandial hyperlipemia. Thus, none of the three main sources of fatty acid supply to the liver were decreased by the administration of PPC.

The possibility that PPC might attenuate the alcoholic fatty liver by enhancing disposition of fat as serum lipoproteins was also inconsistent with the finding of a concomitant decrease in alcoholic hyperlipemia (Fig. 1). Previous studies in rats (Baraona and Lieber 1970, Baraona et al. 1973) and baboons (Savolainen et al. 1986) indicate that the development of alcoholic fatty liver is associated with increased output of triglycerides in the form of large chylomicron-like VLDL-particles from the liver. PPC had no significant effects on serum lipids in the control rats. Thus, the attenuation of the alcoholic hyperlipemia in the present study is more likely a consequence of a lesser fat accumulation in the liver than of a direct effect of PPC on serum lipoproteins.

A likely mechanism for the preventive effects of PPC on both the fatty liver and the hyperlipemia is the observed correction of the ethanol-induced impairments in the capacity of the hepatic mitochondria to oxidize fatty acids. Indeed, a decrease in mitochondrial fatty acid oxidation has been repeatedly demonstrated in ethanol-fed animals. This is a consequence not only of the redox shift engendered by the oxidation of ethanol, but also of structural and functional alterations of this organelle, including a decreased capacity to oxidize fatty acids (Cederbaum et al. 1975, Matsuzaki and Lieber 1977). Similar changes, including a significant decrease in cytochrome oxidase activity, were observed in the ethanol-fed rats of previous studies (Cederbaum et al. 1973), as well as in ethanol-fed baboons (Arai et al. 1984). These alterations were prevented when ethanol feeding was supplemented with PPC. In alcohol-fed baboons, these mitochondrial alterations have been linked to changes in the phospholipid composition of its membranes (Arai et al. 1984); the decrease in cytochrome oxidase activity was associated with decreased phosphatidylcholine and cardiolipin contents of the mitochondrial membranes. Furthermore, the activity of cytochrome oxidase was restored by the addition of these phospholipids. Among the phospholipids, the most efficient class for reactivating the cytochrome oxidase activity were the phosphaditylcholines. It is therefore possible that the effect of PPC could be exerted at this level. A PPC-induced improvement of mitochondrial functions could also account for the smaller increase in liver proteins, because hepatic protein accumulation has been shown to be due to retention of secretory proteins secondary to acetaldehyde- mediated alteration of microtubules (Baraona et al. 1977) and to accumulation of fatty acid-binding protein (Pignon et al. 1987), expected consequences of the impairments in mitochondrial acetaldehyde (Hasumura et al. 1976) and fatty acid (Cederbaum et et. 1975, Matsuzaki and& Lieber, 1977) oxidation, respectively. The preventive effects of PPC also included an attenuation of the elevations of serum ALT, AST and GDH produced by ethanol feeding, with a greater effect on those enzymes that reside either partially (AST) or totally (GDH) in mitochondria.

Thus, the preventive effect of PPC on the development of liver fibrosis after prolonged administration of ethanol to baboons (Lieber et al. 1990 and 1994) may have been preceded by an attenuation of earlier alterations which could have determined the progression of the disease. Indeed, in a large prospective study, Sørensen et al. (1984) found a stepwise increase in the prevalence of cirrhosis in alcoholics that correlated with the degree of steatosis in the initial biopsies. This does not preclude the possibility that the main component of PPC could also have direct antifibrotic effects, as suggested by enhanced collagenase activity in cultured stellate cells (Lieber et al. 1994) and decreased activation or transformation into myofibroblast-like cells in vivo (Lieber et al.1994) and in vitro (Poniachick et al. 1996).

ACKNOWLEDGMENT