Drug-induced intrahepatic cholestasis: characterization of different pathomechanisms

The pathogenesis of intrahepatic cholestasis in rats was studied using isolated perfused livers as an experimental model. Three basic mechanisms were differentiated: Permeabilization of the bilio-sinusoidal barrier associated with electron microscopic alterations of the tight junctional complexes was found in livers of rats treated with alpha-naphthylisothiocyanate (ANIT, 250 mg/kg body weight). Consequences of these alterations were: reflux of bile constituents such as taurocholate and sulfobromophthalein and increased access to the biliary space of paracellular markers such as inulin and sucrose. The clear-cut mechanism of ANIT cholestasis was used to distinguish other mechanisms of intrahepatic cholestasis. Inhibition of the basic process of fluid secretion was found to be the primary event in the development of cholestasis induced by estrogens. After 5 days of treating rats with ethinyl estradiol (5 mg/kg/day), bile flow was diminished in isolated livers while the permeability of the biliary tract to sucrose and inulin was not affected. Accordingly, the maximal concentration of taurocholate in bile was increased, indicating that its secretion was sustained. The same effect was observed after 1 week of treatment with the depot estrogen estradiol valerate (1 mg/kg/week). After 3 weeks of treatment, however, the taurocholate concentration in bile was lowered and the clearance of sucrose was increased. Bile flow remained at the same cholestatic level for 20 weeks. These results suggest that estrogens have the potency to increase tight junctional permeability only in a second step in the development of cholestasis, following the inhibition of bile flow. An additional mode of secretory inhibition was induced by lowering the concentration of Ca2+ in the perfusate of isolated liver.(ABSTRACT TRUNCATED AT 250 WORDS)

Quantitative estimation of transcellular and paracellular pathways of biliary sucrose in isolated perfused rat liver

A method was developed to estimate the relative contributions of paracellular and transcellular pathways to the total biliary clearance of sucrose in isolated perfused rat liver. When livers were perfused with a sucrose-containing medium (1 mM), biliary sucrose concentration reached an equilibrium of 165 +/- 27 microM within 10 min, without further significant change up to 40 min. After removal of sucrose from the perfusate, the decrease of the sucrose concentration in bile was found to obey biphasic first-order kinetics, showing a rapid initial decrease (half-life 3.3 +/- 0.5 min) and then a slower decrease (half-life 29.4 +/- 5.7 min). Both phases of decrease were further characterized. Pretreating rats with the cholestatic agents alpha-naphthyl isothiocyanate (ANIT), oestradiol valerate (OV) and colchicine increased the biliary equilibrium concentration and decreased the half-life of the fast phase of the biliary sucrose elimination. The slow phase was unaffected in the livers of ANIT- and OV-treated rats. The slow phase of biliary sucrose efflux was sensitive to colchicine treatment. A close correlation was observed between the slow-phase fraction of the biliary sucrose and the corresponding sucrose content of the liver. By quantitative analysis of the efflux kinetics the relative contribution of the paracellular pathway to the biliary clearance of sucrose was estimated to be 83 +/- 2% in control livers, which increased to about 90% in livers of pretreated animals. These results are important in view of the use of sucrose in evaluating the paracellular-pathway permeability in intra- and extra-hepatic cholestasis.

Disposition and hepatoprotection by phosphatidyl choline liposomes in mouse liver

Small unilamellar liposomes with an average diameter of 80 nm were prepared from phosphatidyl choline of various sources using the dialysis method with cholate as a detergent. When 14C-labeled soybean liposomes were intravenously injected into male NMRI mice, up to 10% of the total label was found in the liver lipid. The uptake was dose-dependent and reached an apparent saturation 4 h after injection. The liver maintained a constant radioactivity corresponding to 1.9 +/- 0.13 mg phospholipid/g liver until ten hours after injection of 850 mg labeled phosphatidyl choline/kg body wt. Little radioactivity was taken up by the spleen. Analogous doses of liposomes prepared from egg yolk phosphatidyl choline led to a radioactivity corresponding to 1.3 +/- 0.4 mg lipid/g liver 4 h after injection. Liposomes with a similar size were prepared from hydrated, i.e., saturated phosphatidyl choline. After intravenous administration of these liposomes, an amount of 5.3 +/- 0.5 mg labeled lipid was found per g liver after 4 h. In contrast to unsaturated liposomes, 5.8 +/- 0.8 mg lipid per gram spleen was trapped by the spleen. The pharmacodynamic effect of these different liposomes was studied in benzo[a]pyrene-pretreated mice intoxicated with 400 mg/kg paracetamol. Animals which received paracetamol exhibited serum alanine aminotransferase activities of 4220 +/- 1140 units/l after 4 h and exhaled 120 +/- 19 nmol ethane kg-1 h-1. When pretreated with 850 mg soybean phosphatidyl choline/kg body wt. (i.v.) 2 h prior to paracetamol, the increase in serum transaminase activity was reduced to 117 +/- 104 units/l and ethane exhalation amounted to 18 +/- 8 nmol kg-1 h-1. In contrast, similar pretreatment with egg yolk phosphatidyl choline or hydrated phosphatidyl choline failed to protect against paracetamol-induced hepatotoxicity. The different pharmacodynamic effects of the two phosphatidyl cholines of plant or animal origin cannot be explained on the basis of their different pharmacokinetics. In the case of soybean phosphatidyl choline liposomes, the amount of radioactive lipid found in the liver correlated with the hepatoprotective potency.

The role of acrolein in allyl alcohol-induced lipid peroxidation and liver cell damage in mice

Male NMRI mice were fed a sucrose diet for 48 hr in order to reduce the hepatic glutathione content and to level off its diurnal variation. After administration of allyl alcohol (AA: 1.1 mmol/kg), hepatic glutathione (24.3 +/- 7.0 nmol GSH/mg protein) was almost totally lost within the first 15 min (less than 0.5 nmol GSH/mg protein). Subsequently, a massive lipid peroxidation was observed, i.e. the animals exhaled 414 +/- 186 nmol ethane/kg/hr compared to 0.9 +/- 0.8 of controls, and the hepatic TBA-reactive compounds had increased from 55 +/- 16 pmol/mg protein in controls to 317 +/- 163 after 1 hr. Concomitantly, a 40-45% loss of the polyunsaturated fatty acids (arachidonic and docosahexaenoic acid) in the liver lipids was observed. About 80% of the cytosolic alcohol dehydrogenase activity and about 50% of the microsomal P450-content were destroyed. In vivo-inhibition of alcohol dehydrogenase by pyrazole or induction of aldehyde dehydrogenase by phenobarbital abolished AA-induced liver damage as well as glutathione depletion and lipid peroxidation, while inhibition of aldehyde dehydrogenase by cyanamide made a subtoxic dose of AA (0.60 mmol/kg) highly toxic. These results strongly favour the importance of acrylic acid formation as an additional detoxification pathway. Enhanced hepatic levels of glutathione protected in vivo against the damaging effects of AA. Depletion of the liver glutathione content by phorone or diethylmaleate alone caused marginally enhanced lipid peroxidation (phorone) but not liver cell damage. Monooxygenase inhibitors (metyrapone, diethyldithiocarbamate, alpha-naphthoflavone) or an inducer (benz(a)pyrene) did not affect AA-induced toxicity. The ferric iron chelator desferoxaminemethanesulfonate prevented AA-induced lipid peroxidation and liver cell damage in vivo. In vitro, acrolein alone failed to initiate lipid peroxidation in soy bean phospholipid liposomes or in mouse liver microsomes. Thus, acrolein not only impairs the glutathione defense system but also directly destroys cellular proteins and evokes lipid peroxidation by an indirect iron-depending mechanism.

Manipulation of mouse organ glutathione contents. II: Time and dose-dependent induction of the glutathione conjugation system by phenolic antioxidants

After 14 days of oral butylated hydroxyanisole (BHA) administration (1000 mg/kg/day) the tissue glutathione levels of male NMRI mice were increased by 74-141% in liver, lung, duodenum and intestine and after similar butylated hydroxytoluene (BHT) treatment by 18-85% in the liver, lung, spleen and the gastrointestinal tract. Doses of 100 mg/kg/day significantly elevated the glutathione content in the lung (BHA, BHT), duodenum (BHA) and intestine (BHA), while 10 mg/kg/day affected only lung glutathione content (BHA). BHA treatment (1000 mg/kg/day) induced GST activities significantly (138-1335%) in all organs investigated except the spleen, i.e. liver, lung, kidney and the entire gastrointestinal tract, while a similar dose of BHT increased GST activities in the liver, duodenum, intestine and colon by 26-339%. Daily doses of 100 mg/kg/day significantly induced GST activities only in the liver (BHA, BHT), lung (BHA) and kidney (BHA). Lower doses of BHA or BHT did not significantly affect GST activities in the organs investigated (except 10 mg BHA/kg/day in the lung). Comparison of the time course of induction of the glutathione conjugation system in various organs after different doses of antioxidants indicated no change between 5 and 14 days of treatment with all doses used (1-1000 mg/kg). Only the lung glutathione level showed a tendency to increase with low dose BHA by extending the time of treatment. The time course of the liver glutathione content between single doses of 100 mg/kg BHA or BHT revealed an initial decline followed by an increase above control values 2 days (BHA) or 5 days (BHT) after the first application. The glutathione levels of the lung and the duodenum increased without a preceding decline. Only the second dose of BHT caused a temporary decrease to control values of the elevated glutathione level in the duodenum. All animals (at any dose of BHA or BHT) showed control values of serum transaminase activities. These results suggest: The induction threshold of the glutathione conjugation system in various mouse organs is greater than or equal to 100 mg/kg for BHA and BHT. Chronic administration of these compounds did not change these results (except the lung glutathione level after low dose BHA). Elevated hepatic glutathione levels might be the result of an activated synthesis caused by a preceding loss of glutathione. Chronic BHA or BHT treatment did not cause hepatotoxic effects, as evaluated by serum transaminases, in male mice.

Regulation of canalicular bile formation by alpha-adrenergic action and by external ATP in the isolated perfused rat liver

In isolated perfused rat liver, addition of adrenaline induced a complex response of bile flow including rapid, reversible stimulation (1/2-2 min), reversible inhibition (2-10 min), and prolonged stimulation. Both the reversible stimulation and the inhibition were mimicked by the alpha-sympathomimetic agonist phenylephrine but not by the beta-agonist isoproterenol. The reversible stimulation was a very early effect being terminated prior to all other alpha-adrenergic responses of liver. External ATP considerably lowered bile flow while inducing release of glucose and lactate, inhibition of respiration, and a reversible efflux of Ca2+. Variations of mannitol clearance parallel to those of bile flow indicate a canalicular origin of all changes.

Manipulation of mouse organ glutathione contents I: Enhancement by oral administration of butylated hydroxyanisole and butylated hydroxytoluene

Administration of either butylated hydroxyanisole (BHA) or butylated hydroxytoluene (BHT) (1000 mg/kg/day for 5 days) to male mice increased the content of reduced glutathione by 50-100% in liver, lung, duodenum and intestine. In colon, glandular stomach, spleen and kidney no effect on glutathione level was observed. BHA and BHT led also to 100-1000% induction of glutathione transferases in liver, lung (only BHA), kidney and digestive tract (except the colon); the relative increase in transferase activity was greater with 1-chloro-2,4-dinitrobenzene (DCNB) as a substrate than with CDNB in all organs investigated. The effects of BHA, administered in olive oil by gavage, on different parts of the gastrointestinal tract revealed maximum increase of the glutathione content and transferase activities in the duodenum, smaller increase of these parameters in the upper intestine and no significant effects in the lower intestine and the colon. Starving mice for 1 day decreased the glutathione content of the liver by 50% to 21.3 +/- 4.5 nmol/mg protein in controls and to 39.4 +/- 3.3 in BHA-treated animals. Intravenous injection of 0.5 mmol GSH/kg restored the fed state (C: 37.4 +/- 2.8 nmol GSH/mg protein; BHA: 84.9 +/- 7.7) within 2 h. This indicates a much faster de novo synthesis of liver glutathione in BHA-pretreated animals. The mechanistic aspects of phenolic antioxidant effects on GSH metabolism are discussed.

Diurnal fluctuation and pharmacological alteration of mouse organ glutathione content

Mouse liver glutathione content showed a diurnal variation with a maximum GSH + 2 GSSG content at 6 to 10 a.m. of 62 +/- 8 nmole per mg protein and a minimum of 42 +/- 7 at 6 p.m. Starvation for more than 24 hr decreased the hepatic glutathione content to 22 +/- 3 nmole/mg protein and abolished the diurnal rhythm. Artificial reversal of the feeding habit of the animals reversed the diurnal rhythm. Kidney, spleen and lung glutathione contents showed no such rhythm. The organ glutathione content decreased by 50% or more upon starvation. The increase of the liver glutathione content by injection of either free or liposomally entrapped GSH to starved animals was not dependent on the time of administration. The physiological maximum level could not be exceeded by this treatment. It was not possible to influence the glutathione content of kidney, lung or intestine by glutathione injections in either form. Intravenous injections of equimolar doses of 2,3-dimercaptopropanol, 2-mercaptoethanesulfonic acid, N-2-mercaptopropionylglycine, D-penicillamine, or cysteamine did not lead to any significant change in liver, kidney, spleen or lung glutathione contents 2 hr after administration. Intravenously given N-acetylcysteine, methionine, GSH or GSSG restored liver glutathione levels of starved animals to the contents observed in the fed state. The diurnal hepatic variation of GSH caused by the food intake habit of the animals may limit the capacity of the intracellular detoxication system.

No increase of biliary permeability in ethinylestradiol-treated rats

Ethinylestradiol, administered to male Wistar rats for 5 days (5 mg/kg X day), decreased bile flow in anesthetized animals and in isolated perfused livers. The bile salt secretion rate was diminished. The bile-to-perfusate ratios of [14C]sucrose and [14C]inulin increased significantly, but this could be attributed to the decline of bile flow as indicated by almost identical clearance rates. Theorectical analysis according to Forker and Wheeler yielded diffusion permeability coefficients (K) for sucrose and inulin of 0.243 and 0.037 in controls and 0.273 and 0.038 in ethinylestradiol-treated rats. In contrast, 9 h of alpha-naphthylisothiocyanate treatment (250 mg/kg) caused cholestasis with heavily decreased bile salt secretion rates. Here, K-values calculated for sucrose and inulin were 0.807 and 0.175. These findings suggest that altered permeability of the paracellular pathway is the cause for alpha-naphthylisothiocyanate-induced cholestasis, but not the primary event in ethinylestradiol-induced cholestasis.

Drug-induced lipid peroxidation in mice--III. Glutathione content of liver, kidney and spleen after intravenous administration of free and liposomally entrapped glutathione

The half-life of extracellular glutathione was found to be 1.9 min in fed mice with a hepatic glutathione content of 44 +/- 10 nmol glutathione per mg protein. It was 4.9 min in animals that had been fed for 48 hr a liquid sucrose diet resulting in a decreased hepatic glutathione of 25 +/- 7 nmol/mg. A single intravenous injection of 16.2 mumol liposomally entrapped glutathione led to an increase in hepatic glutathione to 45 nmol/mg in the sucrose-fed mice after 2 hr and had no effect in the fed group. The spleen glutathione content reached a maximum at 30 min after injection in both groups. The maximum uptake into liver was 21% of the applied dose, into the spleen 7% and into the kidneys 2.4%. Injection of glutathione in solution led to a similar increase of hepatic glutathione as observed with GSH-containing liposomes, while liposomes filled with the constituent amino acids had only a marginal effect. The spleen took up only liposomal GSH. In contrast, the kidney glutathione content increased within 10 min up to 150% upon injection of free glutathione. The findings are consistent with a rapid hydrolysis of extracellular free glutathione followed by an interorgan turnover utilizing the constituent amino acids for resynthesis in the liver. Pretreatment of the animals with the glutathione synthesis inhibitor buthionine sulfoximine essentially abolished the hepatic glutathione increase upon treatment with GSH-liposomes or with the free compound. The finding that only liposomally entrapped glutathione protects mice against liver necrosis induced by highly dosed paracetamol is discussed with respect to differential uptake and distribution of GSH-liposomes in the liver.

Drug-induced lipid peroxidation in mice--II. Protection against paracetamol-induced liver necrosis by intravenous liposomally entrapped glutathione

If injected intravenously 2 hr before the drug, a dose of more than 175 mg/kg body weight glutathione (0.57 mmol/kg) protected male mice from acute liver necrosis induced by intraperitoneal administration of 400 mg/kg (2.65 mmol/kg) paracetamol. Soluble glutathione yielded a limited, and liposomally entrapped glutathione an optimal dose-dependent protective effect against drug-induced lipid peroxidation (as measured by in vivo ethane exhalation) liver necrosis (assessed by serum transaminases) and hepatic glutathione depletion (determined post mortem). N-Acetylcysteine solution had no effect in this model.