NADH-dependent reductive stress and ferritin-bound iron in allyl alcohol-induced lipid peroxidation in vivo: the protective effect of vitamin E

The role of iron in allyl alcohol-induced lipid peroxidation and hepatic necrosis was investigated in male NMRI mice in vivo. Ferrous sulfate (0.36 mmol/kg) or a low dose of ally alcohol (0.6 mmol/kg) itself caused only minor lipid peroxidation and injury to the liver within 1 h. When FeSO4 was administered before allyl alcohol, lipid peroxidation and liver injury were potentiated 50-100-fold. Pretreatment with DL-tocopherol acetate 5 h before allyl alcohol protected dose-dependently against allyl alcohol-induced lipid peroxidation and liver injury in vivo. Products of allyl alcohol metabolism, i.e. NADH and acrolein, both mobilized trace amounts of iron from ferritin in vitro. Catalytic concentrations of FMN greatly facilitated the NADH-induced reductive release of ferritin-bound iron. NADH effectively reduced ferric iron in solution. Consequently, a mixture of NADH and Fe3+ or NADH and ferritin induced lipid peroxidation in mouse liver microsomes in vitro. Our results suggest that the reductive stress (excessive NADH formation) during allyl alcohol metabolism can release ferrous iron from ferritin and can reduce chelated ferric iron. These findings provide a rationale for the strict iron-dependency of allyl alcohol-induced lipid peroxidation and hepatotoxicity in mice in vivo and document iron mobilization and reduction as one of several essential steps in the pathogenesis.

Role of nitric oxide in the oxidant stress during ischemia/reperfusion injury of the liver

The potential role of nitric oxide (NO) and its reaction product with superoxide, peroxynitrite, was investigated in a model of hepatic ischemia-reperfusion injury in male Fischer rats in vivo. Pretreatment with the NO synthase inhibitor nitro-L-arginine (10 mg/kg) did neither affect the post-ischemic oxidant stress and liver injury during the initial reperfusion phase nor the subsequent infiltration of neutrophils into the liver and the later, neutrophil-induced injury phase. Furthermore, no evidence was found for a postischemic increase of the urinary excretion of nitrite, a stable oxidation metabolite of NO. In contrast, the administration of Salmonella enteritidis endotoxin (1 mg/kg) induced a significant diuresis in Fischer rats and an 800-fold enhancement of the urinary nitrite excretion. Nitro-L-arginine pretreatment inhibited the endotoxin-induced nitrite formation by 97%. Hepatic cGMP levels, as index of NO formation in the liver, were only increased significantly after endotoxin administration but not after ischemia and reperfusion. Our results provide no evidence for any enhanced generation of NO or peroxynitrite either systemically or locally during reperfusion and therefore it is unlikely that any of these metabolites are involved in the oxidant stress and liver injury during reperfusion after hepatic ischemia.

Neutrophil-induced liver cell injury in endotoxin shock is a CD11b/CD18-dependent mechanism

To investigate the role of neutrophils (PMNs) and PMN-dependent adhesion molecules in the pathogenesis of liver injury in a model of endotoxin shock, male ICR mice received a dose of 700 mg/kg galactosamine and 100 micrograms/kg Salmonella abortus equi endotoxin. PMNs accumulated continuously in the liver, reaching values of 446 +/- 71 PMNs/50 high-power fields at 9 h (basal value 18 +/- 7). Plasma alanine aminotransferase activities as index of parenchymal cell injury did not change up to 5 h posttreatment (basal value 35 +/- 5 U/l) but increased to 1,950 +/- 460 U/l at 9 h. The formation of glutathione disulfide (GSSG) in plasma as an index of an extracellular oxidant stress also increased only at 9 h. Pretreatment of animals with monoclonal antibodies against the CD11b and CD18 subunits of the CD11/CD18 integrin family on the surface of the PMN reduced the number of PMNs in the liver by 50% and significantly attenuated liver injury and GSSG formation. An anti-CD11a and a nonbinding control antibody were ineffective. It is concluded that PMNs are actively involved in the pathogenesis of galactosamine and endotoxin shock and that at least in part the accumulation of PMNs, the subsequent oxidant stress, and the tissue injury in this model of experimental hepatitis are CD11b/CD18 (Mac-1) dependent.

Neutrophil and Kupffer cell-induced oxidant stress and ischemia-reperfusion injury in rat liver

The hypothesis that Kupffer cells and infiltrating neutrophils generate reactive oxygen in the hepatic sinusoids and may contribute to ischemia-reperfusion injury in the liver was investigated in a model of partial no-flow ischemia and reperfusion in male Fischer rats in vivo. During the reperfusion period of 60 min, plasma concentrations of glutathione disulfide (GSSG; index of oxidant stress) increased from 1.62 +/- 0.20 microM glutathione (GSH) equivalents to maximal values of 11.82 +/- 1.45 (45 min ischemia), 24.19 +/- 2.35 (60 min ischemia), and 70.20 +/- 7.8 (120 min ischemia). The basal tissue GSSG content in the postischemic lobes (0.19 +/- 0.02 nmol GSH eq/mg protein) increased by 50-100%. Although the number of neutrophils in liver and lung increased by 3- to 10-fold during reperfusion, there was no positive correlation between the number of neutrophils and the GSSG concentrations measured in plasma or tissue. However, activation of Kupffer cells with high doses of retinol or with Propionibacterium acnes significantly enhanced plasma GSSG levels, while inactivation of Kupffer cells with methyl palmitate or gadolinium chloride significantly attenuated the increase of plasma GSSG. Inactivation of Kupffer cells protected the liver significantly against ischemia-reperfusion injury. It is concluded that Kupffer cells are the predominant source of reactive oxygen formed during the initial reperfusion period and that Kupffer cell activity (including reactive oxygen formation) contributes to reperfusion injury in the liver in vivo.

Vascular oxidant stress and hepatic ischemia/reperfusion injury

The objective of this study was to test the hypothesis that the extracellular oxidation of glutathione (GSH) may represent an important mechanism to limit hepatic ischemia/reperfusion injury in male Fischer rats in vivo. Basal plasma levels of glutathione disulfide (GSSG: 1.5 +/- 0.2 microM GSH-equivalents), glutathione (GSH: 6.2 +/- 0.4 microM) and alanine aminotransferase activities (ALT: 12 +/- 2 U/l) were significantly increased during the 1 h reperfusion period following 1 h of partial hepatic no-flow ischemia (GSSG: 19.7 +/- 2.2 microM; GSH 36.9 +/- 7.4 microM; ALT: 2260 +/- 355 U/l). Pretreatment with 1,3-bis-(2-chloroethyl)-1-nitrosourea (40 mg BCNU/kg), which inhibited glutathione reductase activity in the liver by 60%, did not affect any of these parameters. Biliary GSSG and GSH efflux rates were reduced and the GSSG-to-GSH ratio was not altered in controls and BCNU-treated rats at any time during ischemia and reperfusion. A 90% depletion of the hepatic glutathione content by phorone treatment (300 mg/kg) reduced the increase of plasma GSSG levels by 54%, totally suppressed the rise of plasma GSH concentrations and increased plasma ALT to 4290 +/- 755 U/l during reperfusion. The data suggest that hepatic glutathione serves to limit ischemia/reperfusion injury as a source of extracellular glutathione, not as a cofactor for the intracellular enzymatic detoxification of reactive oxygen species.

Reactive oxygen and ischemia/reperfusion injury of the liver

Pharmacological experiments suggested that reactive oxygen species contribute to ischemia-reperfusion injury of the liver. Since there is no evidence that quantitatively sufficient amounts of reactive oxygen are generated intracellularly to overwhelm the strong antioxidant defense mechanisms in the liver and cause parenchymal cell injury, the role of reactive oxygen in the pathogenesis remains controversial. This paper reviews the data and conclusions obtained with pharmacological intervention studies in vivo, the sources of reactive oxygen in the liver as well as the growing evidence for the importance of liver macrophages (Kupffer cells) and infiltrating neutrophils in the pathogenesis. A comprehensive hypothesis is presented that focuses on the extracellular generation of reactive oxygen in the hepatic sinusoids, where Kupffer cell-derived reactive oxygen species seem to be involved in the initial vascular and parenchymal cell injury and indirectly also in the recruitment of neutrophils into the liver. Reactive oxygen species may also contribute to the subsequent neutrophil-dependent injury phase as one of the toxic mediators released by these inflammatory cells.

Multiple scan modes in the hybrid tandem mass spectrometric screening and characterization of the glutathione conjugate of 2-furamide

The glutathione conjugate of 2-furamide has been screened for and structurally characterized by tandem mass spectrometry (MS(MS) by using a hybrid instrument of BEqQ design. Mass spectrometry experiments employed fast atom bombardment (FAB) ionization of a crude bile extract from a rat dosed with a 1:1 mixture of unlabeled and [ (13)C12-furamide. Initial screening for glutathione conjugates employed constant neutral loss scanning to detect the loss of 129 u, corresponding to the loss of the γ-glutamyl moiety of the conjugates. By direct comparison with control bile, [M + H] (+) ions of m/z 417 and 418 were readily identified as candidate ions corresponding to the glutathione conjugates of unlabeled and (13)C-labeled 2-furamide. Complementary screening information was generated by using a methylated bile extract, with constant neutral loss scanning to detect the loss of the methylated γ-glutamyl moiety (143 u). An alternative screening procedure employing parent ion scanning to detect the sodium adducts of methylated glutathione conjugates was also developed. Structural information was generated by frrst-generation product ion scanning of the protonated and sodium cationized forms of the candidate species, both native and derivatized. This provided a body of internally consistent evidence that the conjugate retains the pseudoaromatic furan ring system without ring hydroxylation. The utility of sequential mass spectrometry (MS(MS(MS) capability of the hybrid instrument in the analysis of complex biological mixtures was also demonstrated. Using the bile extract, first-generation product ions that formed in either the first or second field-free region of the double-focusing portion of the instrument were subsequently collisionally activated in the rf-only quadrupole followed by mass analysis of the second-generation product ions. Structural information so provided for the glutathione conjugate of 2-furamide further substantiated its retention of the pseudoaromatic furan ring system and facilitated plausible assignment of structures to ionic species generated through multiple decomposition events.

Superoxide generation by Kupffer cells and priming of neutrophils during reperfusion after hepatic ischemia

The objective of this study was to identify the cellular source of the vascular oxidant stress in hepatic ischemia-reperfusion injury in male Fischer rats. Nonparenchymal cells (Kupffer cells, endothelial cells) and neutrophils were isolated from postischemic liver lobes by collagenase-pronase digestion followed by centrifugal elutriation. The spontaneous and stimulated generation of superoxide by these cells were subsequently quantified in vitro. Large Kupffer cells from the postischemic lobes spontaneously generated 300% more superoxide than similar cells from control animals. No difference in spontaneous superoxide formation was found when the small Kupffer cells were compared. No other cells isolated from the postischemic lobes or control liver including neutrophils released any detectable superoxide spontaneously. In contrast, small Kupffer cells and neutrophils from the postischemic liver generated significantly more superoxide after stimulation with phorbol ester or opsonized zymosan than the controls. The considerably higher response with zymosan stimulation compared to phorbol ester indicates a particular priming for a receptor-mediated signal transduction pathway during reperfusion. These studies demonstrate that Kupffer cells are the principal source of the oxidant stress during the initial reperfusion phase after hepatic ischemia. The priming of neutrophils during this time may be an important factor for the later neutrophil-induced injury phase.

Glutathione disulfide formation and oxidant stress during acetaminophen-induced hepatotoxicity in mice in vivo: the protective effect of allopurinol

Acetaminophen (500 mg/kg i.p.) induced hepatotoxicity in fasted ICR mice in vivo. Acetaminophen also caused a long-lasting 50% reduction of the hepatic ATP content, an irreversible loss of hepatic xanthine dehydrogenase activity and a transient increase of the xanthine oxidase activity. All effects occurred before parenchymal cell damage, i.e., the release of cellular enzymes. The hepatic content of GSH and GSSG was initially depleted by acetaminophen without affecting the GSSG:GSH ratio (1:200), however, during the recovery phase of the hepatic GSH levels the GSSG content increased faster than GSH, resulting in a GSSG:GSH ratio of 1:18 24 h after acetaminophen administration. The mitochondrial GSSG content increased from 2% in controls to greater than 20% in acetaminophen-treated mice. The extremely elevated tissue GSSG levels were accompanied by a 4-fold increase of the plasma GSSG concentrations but not by an enhanced biliary efflux, although hepatic GSSG formation and biliary excretion were not affected by acetaminophen. Allopurinol protected dose-dependently against acetaminophen-induced cell injury, the loss of ATP and the increase of the GSSG content in the total liver and in the mitochondrial compartment without inhibiting reactive metabolite formation. High, protective as well as low, nonprotective doses of allopurinol almost completely inhibited hepatic xanthine oxidase and dehydrogenase activity, but only high doses prevented the increase of the mitochondrial GSSG content. The data indicate a long-lasting, primarily intracellular oxidant stress during the progression phase of acetaminophen-induced cell necrosis. The protective effect of allopurinol is unlikely to involve the inhibition of reactive oxygen formation by xanthine oxidase but could be the result of its antioxidant property.

Neutrophils contribute to ischemia/reperfusion injury in rat liver in vivo

To determine the role of neutrophils in the pathogenesis of hepatic ischemia/reperfusion injury, livers from male Fischer rats were subjected to 45 min of no-flow ischemia followed by reperfusion for up to 24 h. Two phases of liver injury were identified, an initial phase during the first hour of reperfusion and a later progression phase with 80 +/- 3% hepatocyte necrosis and an 80-fold increase of neutrophil infiltration in the liver after 24 h. Pretreatment with a monoclonal antibody against neutrophils, which caused consistent neutropenia, protected the liver from reperfusion injury as indicated by 28 +/- 10% necrosis, and 84% reduction of hepatic neutrophil accumulation and a complete recovery of the hepatic ATP content. Our data suggest that the later progression phase of reperfusion injury after hepatic no-flow ischemia is mediated mainly by neutrophils.

Glutathione disulfide as index of oxidant stress in rat liver during hypoxia

Formation of glutathione disulfide (GSSG) was used as an index of reactive oxygen generation in the isolated perfused liver of male Fischer rats during normoxia and hypoxia. Low oxygen tension may affect GSSG formation, rereduction, and transport mechanisms. The effect of short-term hypoxia (15 min) on the biliary and sinusoidal transport of GSSG was tested with the glutathione S-conjugates of sulfobromophthalein and 1-chloro-2,4-dinitrobenzene. Hypoxia inhibited S-conjugate excretion through both pathways by 15-20%. tert-Butyl hydroperoxide (75 microM tBHP) or diquat (200 microM) in the perfusate increased hepatic GSSG release by 430 and 1,550%, respectively, and increased the tissue GSSG content by 47 and 124%, respectively, under normoxia. Hypoxia reduced the stimulated GSSG export by 38 (tBHP) and 83% (diquat) and also caused an additional increase of the tissue GSSG content by 112% during tBHP infusion but caused a reduction by 32% during diquat infusion. Inhibition of the biliary export of GSSG and S-conjugates is mainly compensated by the sinusoidal efflux. Therefore, it is concluded that hypoxia reduces GSSG formation predominantly through suppression of reactive oxygen formation with only marginal effects on the biliary and sinusoidal excretion mechanism. Thus hepatic GSSG formation is a sensitive indicator of oxidant stress during normoxia and hypoxia. Because single parameters may vary considerably, simultaneous monitoring of GSSG in bile, perfusate, and tissue is essential for qualitative and quantitative estimation of reactive oxygen formation.

Diquat hepatotoxicity in the Fischer-344 rat: the role of covalent binding to tissue proteins and lipids

Diquat produces hepatic necrosis in the Fischer-344 rat, and although reactive oxygen species generated by redox cycling are thought to mediate the damage, the possibility that covalent binding of diquat or diquat metabolites to tissue macromolecules contributes to the observed hepatotoxicity of diquat needed to be examined experimentally. Intraperitoneal administration of [ethylene-14C]diquat (0.1 mmol/kg) results in distribution of radioactivity to all organs examined. Measurable radioactivity remains associated with hepatic and renal protein even after extensive solvent extraction, but the amount (12-16 pmol/mg protein) is 100-fold less than the extent of covalent binding observed with comparably hepatotoxic doses of other hepatotoxins such as acetaminophen and bromobenzene. Similarly, although small amounts of radioactivity remain in Folch lipid extracts of liver and kidney (56-179 pmol/mg lipid), this is virtually completely removed by transesterification of the lipid (less than 5 pmol/mg lipid), indicating that the radioactivity does not represent an alkylation of electroneutral alkyl residues of the lipid. The diquat-induced increase in biliary excretion of glutathione disulfide temporally parallels the biliary excretion of radioactivity. Although the present results do not prove the absence of a contribution by alkylation mechanisms to diquat hepatotoxicity, the extremely low upper limits placed on possible covalent interactions reinforce the confidence with which the diquat-treated Fischer-344 rat can be employed as an animal model for mechanisms of cell death mediated by reactive oxygen species.

A possible role of plasma glutathione in glucose-mediated insulin secretion: in vitro and in vivo studies in rats

In isolated rat pancreatic islets exogenous glutathione which is not able to penetrate into cells, augmented glucose (11.1 mmol/l)-mediated insulin release. In the presence of a non-stimulatory glucose concentration (2.8 mmol/l) glutathione had no effect. The half-maximal synergistic action of glutathione on insulin secretion was observed at approximately 8.0 mumol/l. This concentration of glutathione is similar to that found in the plasma of non-fasted anaesthetised rats (6.5 mumol/l). Oral ingestion of glucose increased the arterial plasma glutathione in rats. Our data provide for the first time indirect evidence for a modulating effect of plasma glutathione in postprandial glucose-mediated insulin secretion which appears to be located at the extracellular site of islet cells.

Mitochondria and xanthine oxidase both generate reactive oxygen species in isolated perfused rat liver after hypoxic injury

Hypoxia caused severe damage in isolated perfused livers from fasted male Fischer rats without evidence of the formation of reactive oxygen species during hypoxia. Reoxygenation caused a significant increase in intracellular oxygen species in the injured liver, as indicated by increases in sinusoidal GSSG efflux and tissue GSSG levels. Both parameters were elevated further by addition of KCN (100 microM) or antimycin A (8 microM). Sinusoidal GSSG efflux was suppressed in part by addition of allopurinol (500 microM) and enhanced by hypoxanthine (250 microM). Xanthine oxidase appears to be a partial source, and damaged mitochondria a continuous and quantitatively greater source, of reactive oxygen as a result of liver injury following hypoxia.

Effect of acetaminophen on hepatic content and biliary efflux of glutathione disulfide in mice

The increased expiration of ethane and pentane by mice treated with hepatotoxic doses of acetaminophen suggests the possibility of oxidant mechanisms associated with the necrosis. However, studies in rats are not consistent with oxidant stress mechanisms causing the damage, because acetaminophen given to rats does not increase GSSG efflux, a sensitive index of intrahepatic oxidant stress. To compare the extent of oxidant stress generated by acetaminophen in mice versus rats, hepatic content and biliary efflux of GSSG and GSH in mice have been examined. Bile was collected from anesthetized male ICR mice before and after intraperitoneal administration of acetaminophen (325 mg/kg, 2.15 mmol/kg), t-butyl hydroperoxide (TBHP) (1.5 mmol/kg), diethyl maleate (400 mg/kg, 2.33 mmol/kg, in corn oil) or saline (control) and GSH and GSSG were measured by the enzymatic recycling method of Tietze. An increase in biliary GSSG efflux was produced by t-butyl hydroperoxide, but not by the other agents. Biliary GSH/GSSG ratios decreased in acetaminophen-treated animals, presumably reflecting the marked depletion of hepatic GSH, since a similar decrease was observed with non-hepatotoxic doses of diethyl maleate. The failure of acetaminophen to increase the hepatic content or biliary efflux of GSSG in ICR mice is not consistent with the view that oxidant stress mechanisms cause the damage, despite the increases in alkanes expired after acetaminophen administration in this specific animal model.

Choleresis and increased biliary efflux of glutathione induced by phenolic antioxidants in rats

The mechanism by which high doses of the synthetic antioxidants butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT) raise hepatic glutathione levels above physiological values was investigated in rats. A single dose of an antioxidant (200 mg/kg; p.o.) reduced the hepatic glutathione content by 17% (BHA) or 36% (BHT) after 5 h, but in contrast levels of 55% (BHA) or 34% (BHT) above controls (7.1 +/- 0.5 mumol GSH-equivalents/g liver wt) were measured 48 h after dosing. Both antioxidants increased basal bile flow (1.37 +/- 0.11 microliter/min per g liver wt) and biliary efflux of total glutathione, i.e. GSH and GSSG, (4.18 +/- 0.97 nmol GSH-eq./min per g) severalfold (up to 250%) over controls 24 h after in vivo antioxidant treatment. The sinusoidal efflux of reduced glutathione (14.9 +/- 2.2 nmol GSH-eq./min per g) was significantly reduced (BHA: 23%; BHT: 41%). The increased glutathione excretion into bile is likely to be independent of the induction of the choleresis. The secretion of bile salts was unaffected by BHA treatment and only temporarily reduced by BHT.

CONCLUSION

phenolic antioxidants increase the hepatic turnover of glutathione by stimulating the biliary efflux of GSH. The resulting shift from a predominantly sinusoidal efflux of GSH in controls (hepato-renal circulation) to a predominantly biliary efflux of GSH in antioxidant-treated animals (entero-hepatic circulation) may lead to increased concentrations of cysteine, glycine and glutamic acid in the portal vein and consequently may stimulate the biosynthesis of GSH by enhanced substrate availability in the liver.

Reactive oxygen species during ischemia-reflow injury in isolated perfused rat liver

The hypothesis that intracellular generation of reactive oxygen species in hepatocytes or reticuloendothelial cells may cause ischemia-reperfusion injury was tested in isolated perfused livers of male Fischer rats. GSSG was measured in perfusate, bile, and tissue as a sensitive index of oxidative stress. After a preperfusion phase of 30 min, the perfusion was stopped (global ischemia) for various times (30, 120 min) and the liver was reperfused for another 60 min. The bile flow (1.48 +/- 0.17 microliters/min X gram liver weight), the biliary efflux of total glutathione (6.54 +/- 0.94 nmol GSH eq/min X g), and GSSG (1.59 +/- 0.23 nmol GSH eq/min X g) recovered to 69-86% after short-term ischemia and to 36-72% after 2 h of ischemia when compared with values obtained from control livers perfused for the same period of time. During reperfusion, the sinusoidal efflux of total glutathione (16.4 +/- 2.1 nmol GSH eq/min X g) and GSSG (0.13 +/- 0.05 nmol GSH eq/min X g) did not change except for an initial 10-30-s increase during reperfusion washout. No increased GSSG secretion into bile was detectable at any time during reperfusion. The liver content of total glutathione (32.5 +/- 3.5 nmol GSH eq/mg protein) and GSSG (0.27 +/- 0.09 nmol GSH eq/mg protein) did not change significantly during any period of ischemia or reperfusion. We conclude, therefore, that at most only a minor amount of reactive oxygen species were generated during reperfusion. Thus, reactive oxygen species are unlikely to cause ischemia/reperfusion injury in rat liver by lipid peroxidation or tissue thiol oxidation.

Hypoxic damage generates reactive oxygen species in isolated perfused rat liver

The aim of the present study was to investigate the possible role of reactive oxygen species in the pathogenesis of hypoxic damage in isolated perfused rat liver. One hour of hypoxia caused severe cell damage (lactate dehydrogenase release of greater than 12,000 mU/min/g liver wt) and total irreversible cholestasis which was accompanied by a loss of cellular ATP and a marked decrease in lactate efflux. Tissue glutathione disulfide (GSSG) content and GSSG efflux as a measure of hepatic reactive oxygen formation was less than 1% of total glutathione before and during hypoxia. Upon reoxygenation, however, hepatic GSSG content increased sharply to about twice the control values and GSSG efflux increased several-fold to around 3-4 nmol GSH-equivalents/min/g. The release of lactate dehydrogenase decreased upon reoxygenation and tissue ATP content recovered partially. When livers were reoxygenated at an earlier time interval than 1 hr of hypoxia, i.e., before the onset of damage, no enhanced GSSG formation was observed. The results demonstrate that hypoxic damage is a prerequisite to reactive oxygen formation during the subsequent reoxygenation period. Thus, reactive oxygen species appear unlikely to play a crucial role in the pathogenesis of hypoxic liver damage in the hemoglobin-free, isolated perfused liver model.

Increase in biliary permeability subsequent to intrahepatic cholestasis by estradiol valerate in rats

To clarify the role of biliary permeability in estrogen-induced intrahepatic cholestasis, long-term experiments were performed in rats using estradiol 17 beta-valerate. Bile flow, secretion of taurocholate, biliary clearance of [14C]sucrose and [14C]inulin, and phosphate concentration in bile were determined in hemoglobin-free perfused livers excised from male rats pretreated for different time periods. Basal and taurocholate-stimulated bile flow were already reduced during the first 7-10 days of treatment and remained at the lower level for more than 12 wk. In contrast, the concentration of taurocholate in bile was elevated at 7 and 10 days but was lower than in controls after 3 wk. An increase in [14C]sucrose clearance after 3 wk indicated a moderate increase in the permeability of a paracellular pathway. The concentration of phosphate in bile was also increased after 3 wk. Detailed analysis of biliary sucrose and inulin clearances revealed that the diffusion permeability coefficient (k = 0.14) did not increase during the first 10 days of treatment but did increase to greater than 0.4 after 3 wk of treating rats with estradiol 17 beta-valerate. From the temporal sequence of the cholestatic responses it is concluded that the altered permeability of the biliary tree is not the primary event but occurs subsequent to the cholestasis induced by estrogens.