Bile Duct Obstruction Is not a Prerequisite for Type I Biliary Epithelial Cell Hyperplasia

TOXICOLOGY AND APPLIED PHARMACOLOGY ARTICLE NO.

152, 327–338 (1998)

TO988507

Bile Duct Obstruction Is not a Prerequisite for Type I Biliary Epithelial Cell Hyperplasia David C. Kossor,1 Paul C. Meunier,2 Deanne M. Dulik, Thomas B. Leonard, and Robin S. Goldstein3 Departments of Toxicology and Drug Metabolism and Pharmacokinetics, SmithKline Beecham Pharmaceuticals, King of Prussia, Pennsylvania 19406 Received December 23, 1997; accepted June 9, 1998

Bile Duct Obstruction Is not a Prerequisite for Type I Biliary Epithelial Cell Hyperplasia. Kossor, D. C., Meunier, P. C., Dulik, D. M., Leonard, T. B., and Goldstein, R. S. (1998). Toxicol. Appl. Pharmacol. 152, 327–338. Biliary obstruction, produced by common bile duct ligation or a-naphthylisothiocyanate (ANIT) treatment in rats, has been associated with the development of type I biliary epithelial cell (BEC) hyperplasia. However, the exact mechanism(s) by which bile duct obstruction lead(s) to this proliferative lesion are not clear. The present studies were designed to determine if cholestasis, in the absence of biliary obstruction, would result in type I BEC hyperplasia. Male Sprague–Dawley rats were given a single oral dose of 150 mg/kg ANIT or iv doses of estradiol glucuronide (E2–17G; 21 mmol/kg/h for 48 h) to produce obstructive and non-obstructive cholestasis, respectively. E2–17G treatment resulted in cholestasis that was comparable in extent and duration to that observed following ANIT treatment. E2–17G and ANIT treatments produced comparable increases in serum bile acids (55- to 60-fold) and activities of ALT (36- to 38-fold), ALP (4- to 5-fold), and 5*-nucleotidase (7- to 11-fold), respectively, compared to controls. Both ANIT and E2–17G also increased serum bilirubin concentrations. ANIT treatment resulted in significant increases in biliary glucose concentrations that were associated with BEC damage/necrosis and obstruction of the bile duct lumen. Conversely, no evidence of BEC damage was observed in E2–17Gtreated rats. Nonetheless, BEC hyperplasia was observed in the majority of rats following treatment with either ANIT or E2–17G, assessed by light microscopy and by BrdU immunohistochemistry. These data indicate that E2–17G treatment produces nonobstructive cholestasis and type I BEC hyperplasia, suggesting that biliary obstruction is not a prerequisite for type I BEC hyperplasia in rats. Differences in the time of onset of hyperplasia were observed: hyperplasia was noted immediately following 48 h of E2–17Ginduced cholestasis but occurred several days after ANIT-induced cholestasis had subsided. Since the magnitude/duration of cholestasis was similar in the two models but the temporal association between cholestasis and type I BEC hyperplasia were different, 1

Present address: Animal Nutrition & Health, Roche Vitamins Inc., Nutley, NJ 07110. 2 Present address: Department of Safety Assessment, The Du Pont Merck Pharmaceutical Co., Newark, DE 19714. 3 Present address: Global Project Management, Novartis Pharmaceuticals Inc., East Hanover, NJ 07936.

these data suggest that the proliferative stimulus may be different in the two models and that E2–17G-induced type I BEC hyperplasia may not be attributed solely to cholestasis. © 1998 Academic Press

Biliary epithelial cell (BEC) hyperplasia is a cellular reaction frequently observed in most forms of human liver disease and in a variety of experimental conditions associated with liver injury (Tavoloni, 1987; Slott et al., 1990). Generally, two forms of BEC hyperplasia have been described. Type I, or typical BEC hyperplasia, involves the preexisting biliary epithelium and has been shown to regress upon removal of the proliferative stimulus. In contrast, Type II or oval cell proliferation is associated with hepatocarcinogenesis, and the progenitor cell has not been well defined (Tavoloni, 1987). Bile duct obstruction, produced by either common bile duct ligation (Cameron and Oakley, 1932; Cameron and Hassan, 1958; MacDonald and Pechet, 1961; Steiner et al., 1962; Johnstone and Lee, 1975) or a-naphthylisothiocyanate (ANIT) treatment (Lopez and Mazzanti, 1955), is known to result in Type I BEC hyperplasia. At approximately 16 h after a single oral dose of ANIT to rats ($50 mg/kg), hepatocanalicular dysfunction (decreased biliary taurocholate transport and erythritol clearance) was observed, which was followed by extensive BEC danage/ necrosis at 24 to 48 h after treatment. Sloughing of damaged BECs into the bile duct lumen resulted in obstruction of intrahepatic bile ducts, which further impeded bile flow. At approximately 72 h after treatment, debris was cleared from the bile ducts, permitting bile flow to resume; choleresis was observed at 96 h after treatment with ANIT. By 168 h after treatment, BEC hyperplasia was observed in ANIT-treated rats, at a time when bile flow was not different than controls (Desmet et al., 1968; Kossor et al., 1993, 1995). These data indicate that bile duct obstruction precedes and plays a role in ANIT-induced BEC hyperplasia in rats. Indeed, Slott et al. (1990) have suggested a causal relationship between bile duct obstruction, increased intralumenal biliary pressure, and BEC proliferation. However, biliary obstruction also results in the cessation of bile flow and the contribution of cholestasis per se (or the biochemical sequelae to cholestasis) to the development of type I BEC hyperplasia is unknown.

327

0041-008X/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.

328

KOSSOR ET AL.

The present studies were designed to determine if Type I BEC hyperplasia (hereafter referred to as “BEC hyperplasia”) would occur following nonobstructive cholestasis. Apart from mechanical obstruction of the biliary system, cholestasis can be produced by agents that impair the hepatocanalicular transport of biliary constituents. This so-called hepatocanalicular cholestasis occurs following estradiol–17b-glucuronide (E2–17G) treatment (Kan et al., 1989; Adinolfi et al., 1984; Utili et al., 1990; Abernathy et al., 1992). Electron microscopic evaluation of isolated, perfused livers treated with E2–17G for 90 min revealed fragmentation and loss of canalicular microvilli, dilatation of canaliculi, and thickening of pericanalicular ectoplasm, with no observed changes in the biliary epithelium (Abernathy et al., 1992). Dose-dependent decreases in bile flow, bile acid excretion, and erythritol clearance (a marker for canalicular flow) also were observed in rats following iv E2– 17G infusion. Single iv doses of 21 mmol/kg produced an immediate decrease in bile flow to less than 10% of control values, and bile flow returned to control values by 3 h after dosing (Meyers et al., 1980). Furthermore, selective binding sites for E2–17G, which appear to be associated with bile acid transport, have been identified in hepatocellular and canalicular membrane preparations (Changchit et al., 1990). These data suggest that E2–17G-induced cholestasis results from a decrease in canalicular bile production and not from bile duct obstruction. Thus, to test the hypothesis that cholestasis per se plays a role in BEC proliferation leading to BEC hyperplasia, we used E2–17G to produce nonobstructive cholestasis. We hypothesized that if BEC proliferation occurred with E2–17G at doses that mimicked the magnitude and duration of cholestasis associated with ANIT-induced BEC proliferation, then the relative importance of bile duct obstruction vs cholestasis to the pathogenesis of BEC hyperplasia could be assessed.

FIG. 2. Time course of changes in bile flow after ANIT treatment in preliminary experiment. Rats were treated with a single dose of either vehicle (corn oil) or 150 mg/kg of ANIT as described in Methods and bile flow was evaluated at 16 h and 1, 2, 3, 5, and 7 days later. Values are means 6 SEM of three to four rats per group. *Significant differences compared to controls ( p , 0.05).

METHODS Animals. Adult male Sprague–Dawley rats (CD-VAF)(Charles River Laboratories, Raleigh, NC) weighing between 350 and 500 g and approximately 10 to 12 weeks of age were used in these studies. Rats were housed individually in stainless steel cages in an environmentally controlled room (72 6 4°F; 50 6 10% relative humidity) with a 12-h light– dark cycle. Filtered tap water and food (Purina Rodeny Chow 5001, Purina Mills, Inc., St. Louis, MO) were provided ad libitum. All rats were handled in a humane manner in accordance with established guidelines for the care and use of laboratory animals (NIH publication no. 86-23, 1985). Materials. ANIT (95%), E2–17G, dimethyl sulfoxide (DMSO; 99.5%) and 5-bromo-29-deoxyuridine (BrdU) were purchased from Sigma Chem. Co. (St. Louis, MO). All chemicals were from commercial sources. Mouse anti-BrdU antibody was purchased from Becton Dickinson (San Jose, CA), and a peroxidase vectastain (avitin– biotin) detection system was purchased from Zymed Labs (San Francisco, CA). PE-10 and PE-50 tubing was purchased from Clay Adams (Parsippany, NJ).

FIG. 1. Treatment schedules for ANIT and E2–17G.

Comparison of ANIT and E2–17G effects on bile flow. Preliminary studies were conducted to assess and compare the time course and magnitude of cholestasis produced by ANIT and for E2–17G. The first series of preliminary studies was designed to assess the time course and magnitude of ANITinduced cholestasis associated with biliary hyperplasia. Rats were treated with a single oral dose of either vehicle (corn oil) or 150 mg/kg ANIT by gavage; this dose of ANIT has been reported previously to result in BEC hyperplasia (Kossor et al., 1993, 1995). At 16, 24, 48, 72, 96, and 168 h following treatment, rates of bile flow were measured (see below). A second series of preliminary studies were conducted to identify a dosing regimen for E2–17G that would mimic the extent and duration of cholestasis observed following ANIT treatment. In the preliminary studies, it was learned that a bolus iv dose of 21 mmol/kg E2–17G produced immediate but transient cholestasis, similar to data reported previously (Meyers et al., 1980). However, cholestasis could not be maintained by a continuous iv infusion of E2–17G (data not shown).

329

BILIARY HYPERPLASIA WITHOUT BILE DUCT OBSTRUCTION

allowed to recover for 6 h prior to the initiation of studies. On Day 0, rats with implanted jugular vein cannulae were given either vehicle or E2–17G (21 mmol z kg21 z h21; 5 ml/kg) for 48 h or a single oral dose of 150 mg/kg ANIT by gavage (5 ml/kg). This latter group also received hourly injections of vehicle (5 ml/kg) during the 48-h period of cholestasis (i.e., Days 1–3). In this manner, control, E2–17G-, and ANIT-treated rats were handled/treated similarly (Fig. 1). Hepatobiliary function and morphology were evaluated immediately following a 48-hour period of cholestasis induced by ANIT (study day 3) or E2–17G (study day 2). In addition, hepatobiliary function and morphology also were studied 4 days following cholestasis induced by ANIT (Study Day 7) or by E2–17G (Study Day 6), since significant BEC hyperplasia previously had been observed in ANIT-treated rats at this time (Kossor et al., 1993, 1995) (Fig. 1). To more definitively assess ANIT- or E2–17G-induced BEC proliferation, BrdU labeling of BEC nuclei was evaluated. Osmotic pumps (Model 2ML1; Alza Corp., Palo Alto, CA) containing BrdU (20 mg/mL in 0.9% NaCl) were implanted sc in the middorsal region of the back immediately following jugular vein cannulation. BrdU (200 mg/h) was delivered for 6 to 7 days and BrdU labeling of BEC nuclei was used to assess DNA synthesis.

FIG. 3. Time course of changes in bile flow after E2–17G treatment. Rats were given hourly iv injections of either vehicle or 21 mmol/kg of E2–17G and bile flow was evaluated in separate groups of rats after 0, 3, 6, or 12 h and 1 or 2 days of continuous treatment or at 4 days after the last dose (i.e., Study Day 6). Rats euthanized on Study Days 2 and 6 also were used for comparison of effects on the biliary tree (hepatic morphology and clinical chemistry). Values are means 6 SEM of three to five rats per group. *Significant differences compared to time-matched vehicle-treated control groups ( p , 0.05).

Ultimately, it was determined that cholestasis could be maintained by administering bolus injections of E2–17G (21 mmol z kg21 z h21) in a vehicle consisting of 10% DMSO, 5% EtOH, 0.9% NaCl, administered at hourly intervals. This was performed in anesthetized rats for 3 or 6 h and in conscious rats with implanted jugular vein cannulae for 12 and 24 h. Four hours before the end of the treatment period, conscious rats were anaesthetized with Na pentobarbital (50 mg/kg ip) and bile was collected (see below). Comparison of ANIT and E2–17G effects on the biliary tree. Under light isoflurane anesthesia, rats were equipped with PE-50 jugular vein cannulae and

Measurement of bile flow and serum enzymes. Four hours prior to the scheduled necropsy, each rat was anesthetized with sodium pentobarbital (50 mg/kg ip). The femoral vein was cannulated with PE-50 tubing and 0.9% NaCl containing 6 U/mL of heparin was infused (1 mL/h) to replenish fluids. Through a midline abdominal incision, the common bile duct was isolated and cannulated above the pancreas using PE-10 tubing. Body temperature was maintained at 37°C using a heat lamp. Bile was collected at 20-min intervals over 4 h and rates of bile flow were determined gravimetrically. Immediately after bile collection, a terminal blood sample was collected and rats were exsanguinated. Serum alanine aminotransferase (ALT), 59 nucleotidase (59N), alkaline phosphatase (ALP) activities, and concentrations of total bililrubin and total bile acids were determined spectrophotometrically using Sigma Diagnostic Kits 505, 265-UV, 245, 550, and 450, respectively. Biliary glucose concentrations also were determined spectrophotometrically using Sigma Diagnostic Kit 115-A. Hepatic morphology. Rats were killed by exsanguination and livers were excised and perfused with cold 0.9% NaCl and 10% buffered Formalin. Sections (4 –5 mm) from the left and right liver lobes were stained with hematoxylin and eosin for light microscopic evaluation. BrdU incorporation was assessed immunohistochemically in an adjacent serial section of liver using mouse anti-BrdU antibody (Becton Dickinson) and visualized using a peroxidase vectastain detection system (Zymed Labs) to stain nuclei red. Hematoxylin was used as a counterstain. Sections of the duodenum were prepared as a positive control for BrdU incorporation.

TABLE 1 Serum Activities of Hepatic Enzymes and Serum Concentrations of Bile Acids, Total Bilirubin, and Glucose Treatmenta Parameter 59 Nucleotidase (U/Liter) Alanine aminotransferase (U/Liter) Aspartate aminotransferase (U/Liter) Bile acids (mM) Total bilirubin (mg/dL) Glucose (mg/dL)

Vehicle

E2–17G

ANIT

14.0 6 1.80 54.0 6 11.0 463 6 27.0 14.7 6 1.60 0.158 6 0.099 79.9 6 5.4

100 6 11.7b 2034 6 415b 2145 6 249b 747 6 197b 3.00 6 0.900b 78.5 6 2.2b

163 6 19.8b 2158 6 299b 1847 6 157b 1030 6 147b 12.39 6 1.60b,c 124.7 6 8.3b

Note. Values are means 6 SEM for three or four rats per group. a Measurements were recorded after 48 h of treatment with either iv vehicle, iv E2–17G (21 mmol z kg21 z h21) or a single oral dose of ANIT (150 mg/kg) followed by iv vehicle, as described in Methods. b Significantly different than vehicle control ( p , 0.05). c Significantly different than E2–17G-treated group ( p , 0.05).

330

KOSSOR ET AL.

331

FIG. 4 —Continued

BILIARY HYPERPLASIA WITHOUT BILE DUCT OBSTRUCTION

332

KOSSOR ET AL.

FIG. 4. Photomicrograph of liver sections from rats treated with a single dose of ANIT (150 mg/kg) followed by iv vehicle or vehicle alone. (a) Section obtained on Day 7 after treatment with vehicle, illustrating the frequency of bile duct profiles (arrows) (stained with H & E. Magnification 2003). (b) Adjacent section from control liver on Day 7 after treatment with vehicle; nuclei incorporating 5-bromo-29-deoxyuridine (BrdU) are stained as described in Methods and principally involve nonparencymal cells. (c) Section of liver obtained on Day 3 after a single dose of ANIT (150 mg/kg); bile ductules are obstructed and the epithelium is damaged (arrow), periductular edema and inflammatory cells are also present (stained with H & E. Magnification 6003). (d) Adjacent section from liver obtained on Day 3 after ANIT treatment. Nuclei have been stained for BrdU incorporation. A significant increase in BrdU-labeled BECs was observed compared to vehicle-treated controls. (e) Section of liver on Day 7 after ANIT treatment; numerous bile duct profiles are present (arrows)(stained with H & E. Magnification 2003).

Statistics. Data are expressed as means 6 SEM and homogeneity of variance was tested using Bartlett’s test (Sokal and Rohlf, 1981). Data were analyzed by analysis of variance followed by the Least Significant Differences procedure (Sokal and Rohlf, 1981). The criterion of significance was p , 0.05.

RESULTS

Bile flow at 16 and 24 h after treatment with ANIT was comparable to vehicle-treated controls (Fig. 2). At 48 and 72 h

after treatment, bile flow was significantly decreased to approximately 10 and 50% of control values, respectively. On Days 5 and 7 following ANIT treatment, bile flow was greater than and comparable to control values, respectively (Fig. 2). In contrast to the delayed onset for ANIT-induced cholestasis, a single dose of 21 mmol/kg E2–17G resulted in an immediate decrease in bile flow to approximately 10% of control values. Administration of E2–17G at 21 mmol z kg21 z h21 for either 3,

333

BILIARY HYPERPLASIA WITHOUT BILE DUCT OBSTRUCTION

6, 12, 24, or 48 h maintained bile flow at approximately 25% of control values (Fig. 3). Based on these results, it was concluded that treatment of rats with 21 mmol z kg21 z h21 of E2–17G resulted in cholestasis that was fairly comparable in magnitude to that produced by ANIT. Serum bile acid concentrations were significantly increased 55- fold in E2–17G-treated rats and 60-fold in ANIT-treated rats on Days 2/3 (Table 1), suggesting that treatment with either E2–17G or ANIT resulted in a similar degree of hepatobiliary dysfunction. In addition, serum 59-N, ALP, and ALT activities were increased 7- to 11-, 4- to 5-, and 36- to 38-fold in ANIT- and E2–17G-treated rats, respectively, compared to rats given vehicle alone (Table 1), further suggesting that E2–17G and ANIT produced a similar extent of hepatocellular damage. Serum bilirubin concentrations following treatment also were increased by 200- and 700-fold following treatment with either E2–17G or ANIT, respectively, compared to vehicle-treated controls (Table 1). Biliary glucose concentrations in ANIT-treated rats were approximately 50% greater compared to values in vehicletreated controls (Table 1). This finding is consistent with previous studies in ANIT-treated rats (Krstulovic et al., 1967, Goldfarb et al., 1962; Kossor et al., 1993) and suggests that ANIT may interfere with biliary glucose reabsorption by the BECs (Guzelian and Boyer, 1974). In contrast, biliary glucose concentrations in E2–17G-treated rats were not significantly different from those of vehicle-treated controls (Table 1). Compared to liver sections from control rats treated with iv vehicle only (Fig. 4A), ANIT treatment resulted in widespread necrosis of the biliary epithelium accompanied by multifocal biliary obstruction on Day 3 (Fig. 4C). The basement membrane of the majority of bile ducts was naked or covered with an attenuated epithelium, suggestive of epithelial cell damage. There was multifocal hypertrophy of the bile ductular epithelium. Adjacent liver sections from ANIT-treated rats were stained for BrdU incorporation to measure rates of proliferation (Fig. 4D). Increased BrdU incorporation by BEC nuclei can be visualized on Day 3 following treatment (Fig. 4D), which is greater than that observed in liver sections from control rats treated with vehicle only (Fig 4B), where BrdU labeling is present in nonparenchymal cells, but not BECs. Importantly, there was no evidence of biliary epithelial cell hyperplasia at this time. However, on Day 7 following ANIT treatment, biliary hyperplasia accompanied by periductular fibrosis was observed in the majority of ANIT-treated rats (Fig. 4E). These data are consistent with the results of previous studies of ANIT-treated rats (Kossor et al., 1993). In contrast to the findings in ANIT-treated rats, microscopic evaluation of livers in E2–17G-treated rats immediately following 48 h of cholestasis revealed mild BEC hyperplasia in three of five rats accompanied by marked hypertrophy of the epithelium of bile ducts and ductules (Table 2, Fig. 5A). Moreover, BrdU incorporation by BEC nuclei from E2–17Gtreated rats was significantly increased on Day 6 compared

TABLE 2 Summary and Incidences of Histologic Findings Treatment groups Vehicle

E2–17G

ANIT

0/5a 0/5 0/5 0/5 0/5 0/5 0/5 1/5 0/5

4/5 0/5 0/5 5/5 5/5 3/5 5/5 0/5 5/5

0/5 5/5 5/5 0/5 5/5 0/5 5/5 2/5 0/5

0/4 0/4

1/4 4/4

4/5 5/5

Days 2/3 Necrosis, biliary epithelium, minimal Necrosis, biliary epithelium, marked Obstruction, biliary Hypertrophy, bile duct epithelium Hypertrophy, bile ductular epithelium Hyperplasia, biliary epithelium Necrosis, hepatocellular, single cell Necrosis, hepatocellular, clusters Hypertrophy, hepatocellular, periportal Days 6/7 Hyperplasia, biliary epithelium Fibrosis, periductular a

Values represent number of rats affected/number of rats evaluated.

to time-matched controls, providing additional evidence that E2–17G treatment stimulated BEC proliferation (Fig. 5B). Although minimal BEC damage was produced by E2–17G treatment, there was no microscopic evidence of bile duct obstruction (Table 2). All rats treated with 21 mmol z kg21 z h21 of E2–17G for 48 h and evaluated on Day 6 displayed an increase in periductular fibrous connective tissue (periductular fibrosis), but BEC hyperplasia was observed in only one of four rats (Table 2). The presence of periductular fibrosis suggests that the biliary epithelial changes that were present after 48 h of treatment had regressed following discontinuation of drug treatment in the majority of rats. In addition to the changes in the biliary epithelium observed at the end of the 48-h cholestatic period, there was widespread single-cell hepatocellular necrosis in the liver of both ANIT (Fig. 4C) and E2–17G-treated rats (Fig. 5A). Marked hypertrophy of periportal hepatocytes in the liver of E2–17G-treated rats was also observed (Table 2). For E2–17G-treated rats, hepatocellular necrosis was associated with a regenerative hepatocellular response, evidenced by the substantial increase in BrdU labeling of hepatocytes that was observed on Day 6 (Fig. 5B). DISCUSSION

In the present studies, treatment of rats with either a single oral dose of ANIT or multiple iv bolus doses of E2–17G resulted in severe intrahepatic cholestasis. Following a single oral dose of ANIT, bile flow was significantly decreased at 48 and 72 h after treatment, while at 120 h, bile flow was restored. The time course of ANIT-induced cholestasis in the present studies were very similar to the results of previous studies using an identical ANIT treatment protocol, where the onset of

334

KOSSOR ET AL.

335

BILIARY HYPERPLASIA WITHOUT BILE DUCT OBSTRUCTION

TABLE 3 Temporal Relationships between Cholestasis and BEC Hyperplasia Treatment day Treatment ANIT Cholestasis BEC hyperplasia: E2–17G Cholestasis BEC hyperplasia:

0

1

2

3

4

5

6

l l l l l l l l l l l l l l l l l

7

l l

Following a single oral dose of ANIT, the onset of cholestasis occurs at approximately Day 1, and bile flow is restored at approximately Day 3; no BEC hyperplasia was observed immediately following the period of ANIT-induced cholestasis (Day 3), while substantial evidence for BEC hyperplasia was observed on Day 7 after treatment. Treatment with iv E2–17G immediately induced cholestasis, which persisted throughout the 48-h treatment period. Immediately following the period of E2–17G-induced cholestasis, substantial evidence for BEC hyperplasia was observed. At Day 6, BEC hyperplasia also was observed, but at a lower incidence/severity compared to that observed on Day 2.

cholestasis was observed at 24 h of treatment, and restoration of bile flow occurred between 72 and 96 h after treatment (Kossor et al., 1993, 1995). Thus, a single oral dose of 150 mg/kg of ANIT in rats results in severe cholestasis for a period of approximately 48 h. In the present studies, an iv dose of 21 mmol z kg21 z h21 E2–17G produced an immediate but transient cessation of bile flow; repeated doses at hourly intervals were needed to maintain bile flow at approximately 25% of control values for up to 48 h. Thus, in the present studies treatment with either ANIT or E2–17G resulted in cholestasis of comparable magnitude and duration. The observed decreases in bile flow rates following treatment with either ANIT or E2–17G also were accompanied by similar increases in serum bile acid concentrations and hepatic enzyme activities (Table 1) in E2–17G- or ANITtreated rats, further indicative of a comparable extent of hepatic damage and dysfunction produced by the two treatments. Increased serum bilirubin concentrations also were observed following treatment with either ANIT or E2–17G (Table 1). However, the ANIT-induced increase in serum bilirubin was substantially greater than that produced by treatment with E2–17G. This difference may be related to the combined effects of hepatobiliary damage and biliary obstruction. Hyperbilirubinemia following ANIT treatment in rats principally consists of conjugated bilirubin suggestive of hepatocellular damage (Becker and Plaa, 1965). However, increased serum concentrations of unconjugated bilirubin also have been observed and may be due to ANIT-

induced hemolysis (Goldfarb et al., 1962; Moran et al., 1961). Indeed, hemolysis is known to contribute to the increase in total serum bilirubin after short-term biliary obstruction in rats (Salomon et al., 1986). Thus, the more extensive hyperbilirubinemia following ANIT treatment may be due to ANIT-induced hepatocellular damage as well as hemolysis associated with ANIT-induced bile duct obstruction. In the normal, healthy liver, concentrations of glucose in bile are less than those in serum, due to extensive glucose reabsorption by the biliary epithelium (Guzelian and Boyer, 1974). Recently, Kanz et al. (1992) had demonstrated that selective BEC damage is associated with an increase in biliary glucose, indicative of a decreased reabsorptive capacity of the damaged biliary epithelium. Biliary glucose concentrations in rats treated with E2–17G were not significantly different from those of vehicle-treated controls (Table 1), suggesting that E2–17G treatment did not impair glucose reabsorption by the biliary epithelium. Conversely, biliary glucose concentrations in ANIT-treated rats were 50% greater than control values (Table 1), a finding that is consistent with previous studies reporting ANIT-induced BEC necrosis (Krstulovic et al., 1967; Goldfarb et al., 1962; Kossor et al., 1993). Thus, the reduced capacity for the biliary epithelium to reabsorb biliary glucose is associated with BEC damage/necrosis following treatment with ANIT, but not E2–17G. Since E2–17G was administered via iv injections, it also was

FIG. 5. Photomicrograph of liver sections from rats treated with E2–17G (21 mmol z kg21 z h21 for 48 h). (a) Section of liver obtained on Day 2 (immediately after treatment with E2–17G); numerous bile duct profiles are present (arrows). Hepatocellular hypertrophy is also present (magnified 2003). (b) Section of liver obtained on Day 6 (4 days after termination of treatment with E2–17G); nuclei have been stained for BrdU incorporation. Significant increases in BrdU incorporation in BEC (arrows) and hepatocyte (heavy arrows) nuclei were observed compared to control. However, no histologic evidence of BEC hyperplasia is present at this time (Magnification 2003).

336

KOSSOR ET AL.

necessary to administer sham vehicle injections to the ANITtreated and control groups. No adverse findings were oberved in the livers from vehicle-treated controls (Fig. 4A). In contrast, ANIT treatment resulted in widespread necrosis of the biliary epithelium accompanied by multifocal biliary obstruction on Day 3 (Fig. 4C). At this time, increased BrdU incorporation by BEC nuclei also was observed (Fig. 4D). Importantly, there was no evidence of biliary epithelial cell hyperplasia at this time. However, on Day 7 following ANIT treatment, biliary hyperplasia accompanied by periductular fibrosis was observed in the majority of ANIT-treated rats (Fig. 4E). These data are consistent with the results of previous studies of ANIT-treated rats (Kossor et al., 1993). Thus, ANIT treatment resulted in a period of severe cholestasis that persisted for approximately 48 h, and BEC hyperplasia was observed several days after bile flow was restored. A comparison of the temporal association between cholestasis and BEC hyperplasia produced by either ANIT or E2–17G is presented in Table 3. In contrast to the findings in ANIT-treated rats, bile flow was restored immediately upon cessation of E2–17G treatment (data not shown), and microscopic evaluation of livers from E2–17G-treated rats immediately following 48 h of cholestasis revealed mild BEC hyperplasia in three of five rats accompanied by marked hypertrophy of the epithelium of bile ducts and ductules (Table 2, Fig. 5A). BrdU incorporation by BEC nuclei from E2–17G-treated rats was significantly increased compared to time-matched controls, providing additional evidence that E2–17G treatment stimulated BEC proliferation. Although minimal BEC damage was produced by E2–17G treatment, there was no microscopic evidence of bile duct obstruction (Table 2). These observations demonstrate that E2–17G treatment results in BEC proliferation in the absence of bile duct obstruction, suggesting that bile duct obstruction is not essential for the pathogenesis of BEC hyperplasia in rats. All rats treated with 21 mmol z kg21 z h21 or E2–17G for 48 h and evaluated on Day 6 displayed an increase in periductular fibrous connective tissue (periductular fibrosis). However, BEC hyperplasia was observed in only one of four rats at this time. The presence of periductular fibrosis suggests that the biliary epithelial changes that were present after 48 h of treatment had regressed following discontinuation of drug treatment in the majority of rats (Fig. 5B). A similar pattern of effects have been observed in bile ductligated rats when removal of the proliferative stimulus (bile duct ligation) was associated with a rapid regression of the epithelial component of biliary epithelial hyperplasia by the process of apoptosis, which preceded regression of the supportive stroma (Bathal and Gall, 1985). Thus, BEC hyperplasia was observed after 48 h of E2–17G-induced cholestasis and regressed following discontinuation of E2–17G treatment. Taken collectively, these data indicate that ANIT or E2–

17G, at dosages that result in a comparable magnitude and duration of cholestasis, are associated with BEC hyperplasia in rats. E2–17G-induced BEC hyperplasia was not associated with biliary obstruction, suggesting that bile duct obstruction is not a prerequisite for BEC hyperplasia and that cholestasis per se (or the sequelae to cholestasis), regardless of its origin, may be an important pathogenetic factor for BEC hyperplasia. The compounds used in these studies (ANIT and E2–17G) resulted in BEC hyperplasia that is not associated with hepatocarcinogenesis. This is an important distinction, since many previous studies that report chemically induced BEC hyperplasia without biliary obstruction have utilized compounds that are associated with hepatocarcinogenesis and/or oval cell proliferation (Svoboda et al., 1966; Nolan et al., 1966). Therefore, data from these earlier studies do not pertain to the type of BEC hyperplasia observed in this present study. The exact mechanisms by which cholestasis may contribute to BEC proliferation are not known. Conceivably, mitogens or comitogens, which are normally excreted in bile, may be increased in the vascular or tissue compartments under conditions of cholestasis. For example, hepatocyte growth factor and epithelial growth factor (EGF) are peptide growth factors that are excreted in bile (Liu et al., 1992; Appasamy et al., 1993; Burwen et al., 1984; Kong et al., 1992) and have been shown to stimulate BEC growth in culture (Mathis and Sirica, 1990; Joplin et al., 1991). Receptor-mediated endocytosis of EGF also has been demonstrated in isolated BECs (Ishii et al., 1990). In addition, the thyroid hormone triiodothyronine, which is excreted in bile, recently has been shown to augment BEC proliferation in bile duct-ligated rats (Apostoli et al., 1994). Thus, cholestasis may result in increased BEC exposure to growth factors that are normally excreted in bile, resulting in BEC proliferation and hyperplasia. Although the magnitude and duration of ANIT- and E2–17Ginduced cholestasis were comparable, treatment-related differences in the time of onset of BEC hyperplasia were noted. Specifically, E2–17G-induced BEC hyperplasia was noted immediately after the 48-h period of cholestasis whereas ANITinduced BEC hyperplasia was noted several days after the 48-h cholestatic period (Table 2). These differences in the onset of BEC hyperplasia relative to the period of cholestasis suggest that the onset of BEC hyperplasia cannot be attributed solely to cholestasis (or the biochemical sequelae to cholestasis) and further suggest differences in the proliferative stimuli. Thus, factors other than, or in addition to, cholestasis likely contribute to the development of this E2–17G-induced hyperplastic lesion. Several studies have shown that estradiol and other synthetic estrogens are mitogenic for hepatocytes (Francavilla et al., 1984, 1989; Mayol et al., 1992) and, conceivably, E2–17G or a metabolite may be mitogenic to the biliary epithelium. In the present study, a substantial increase in BrdU

BILIARY HYPERPLASIA WITHOUT BILE DUCT OBSTRUCTION

labeling also was observed in hepatocytes at the end of the 48-h period of cholestasis in E2–17G-treated rats, accompanied by marked hypertrophy of periportal hepatocytes. Although the increased BrdU labeling was associated with hepatocellular necrosis and is suggestive of a regenerative hepatocellular response, a mitogenic effect of E2–17G for hepatocytes must be considered. Regardless of the exact mechanism(s) mediating E2– 17G-induced BEC hyperplasia, however, the occurrence of this lesion in the absence of biliary obstruction suggests that bile duct obstruction is not a prerequisite for Type I BEC hyperplasia. To our knowledge, this is the first demonstration that Type I BEC hyperplasia is unrelated to biliary obstruction. ACKNOWLEDGMENTS The authors thank Dr. Lester Schwartz for his evaluation of liver histology and Dr. Jeffrey A. Handler, Ms. April R. Apostoli, and Ms. Caroline A. Genell for their assistance in the design and execution of these studies.

REFERENCES Abernathy, C. O., Zimmerman, H. J., Ishak, K. G., Utili, R., and Gillespie, J. (1992). Drug-induced cholestasis in the perfused rat liver and its reversal by tauroursodeoxycholate: An ultrastructural study (43328). Proc. Soc. Exp. Biol. Med. 199, 54 –58. Adinolfi, L. E., Utili, R., Gaeta, G. B., Abernathy, C. O., and Zimmerman, H. J. (1984). Cholestasis induced by estradiol-17b-D- glucuronide: Mechanisms and prevention by sodium taurocholate. Hepatology 4(1), 30 –37. Apostoli, A. R., Kossor, D. C., Meunier, P. C., Short, B. G., and Goldstein, R. S. (1994). Role of thyroid status in biliary epithelial cell proliferation following bile duct ligation. Toxicologist 14(1), 285. Appasamy, R., Tanabe, M., Murase, N., Zarnegar, R., Venkataramanan, R., Van Thiel, D. H., and Michalopoulos, G. K. (1993). Hepatocyte growth factor, blood clearance, organ uptake, and biliary excretion in normal and partially hepatectomized rats. Lab. Invest. 68(3), 270 –276. Bathal, P. S., and Gall, J. A. M. (1985). Deletion of hyperplastic biliary epithelial cells by apoptosis following removal of the proliferative stimulus. Liver 5, 311–325. Becker, B. A., and Plaa, G. L. (1965). Quantitative and temporal delineation of various parameters of liver dysfunction due to a-naphthylisothiocyanate. Toxicol. Appl. Pharmacol. 7, 708 –718. Burwen, S. J., Barker, M. E., Goldman, I. S., Hradek, G. T., Raper, S. E., and Jones, A. J. (1984). Transport of epidermal growth factor by rat liver: Evidence for a nonlysosomal pathway. J. Cell Biol. 99, 1259 –1265. Cameron, G. R., and Hassan, S. M. (1958). Disturbances of sturcture and function in the liver as the result of biliary obstruction. J. Pathol. Bacteriol. 75, 333–349. Cameron, G. R., and Oakley, C. (1932). Ligation of the common bile duct. J. Pathol. Bacteriol. 35, 769 –798. Changchit, A., Durham, S., and Vore, M. (1990). Characterization of [3H]estradiol-17b-(b-D-glucuronide) binding sites in basolateral and canalicular liver plasma membranes. Biochem. Pharmacol. 40(6), 1219 –1225. Desmet, V. J., Krstulovic, B., and Van Damme, B. (1968). Histochemical study of rat liver in alpha-naphthylisothiocyanate (ANIT) induced cholestasis. Am. J. Pathol. 52, 401– 421. Francavilla, A., DiLeo, A., Eagon, P. K., Wu, S., Ove, P., Van Thiel, D. H., and Starzl, T. E. (1984). Regenerating rat liver: Correlations between

337

estrogen receptor localization and deoxyribonucleic acid synthesis. Gastroenterology 86, 552–557. Francavilla, A., Polimeno, L., DiLeo, A., Barone, M., Ove, P., Coetzee, M., Eagon, P., Makowka, L., Ambrosino, G., Mazzaferro, V., and Starzl, T. E. (1989). The effect of estrogen and tamoxifen on hepatocyte proliferation in vivo and in vitro. Hepatology 9(4), 614 – 620. Goldfarb, S., Singer, E. J., and Popper, H. (1962). Experimental cholangitis due to alpha-naphthylisothiocyanate (ANIT). Am. J. Pathol. 40(6), 685– 698. Guzelian, P., and Boyer, J. L. (1974). Glucose reabsorption from bile. J. Clin. Invest. 53, 526 –535. Ishii, M., Vroman, B., and LaRusso, N. F. (1990). Morphologic demonstration of receptor-mediated endocytosis of epidermal growth factor by isolated bile duct epithelial cells. Gastroenterology 98, 1284 –1291. Johnstone, J. M. S., and Lee, E. G. (1975). A quantitative assessment of the structural changes in the rat’s liver following obstruction of the common bile duct. Br. J. Exp. Pathol. 57, 85–94. Joplin, R., Ayres, R. C. S., Hishida, T., Daikuhara, Y., Tsubouchi, H., Neuberger, J. M., and Strain, A. J. (1991). Human hepatocyte growth factor induces proliferation of human intrahepatic biliary epithelial cells in vitro. Hepatology 14(4), 103A. Kan, K. S., Monte, M. J., Parslow, R. A., and Coleman, R. (1989). Oestradiol 17b-glucuronide increases tight-junctional permeability in rat liver. Biochem. J. 61, 297–300. Kanz, M. F., Kaphalia, L., Kaphalia, B. S., Romagnoli, E., and Ansari, G. A. S. (1992). Methylene dianiline: Acute toxicity and effects on biliary function. Toxicol. Appl. Pharmacol. 117, 88 –97. Kong, W., Koldovsky, O., and Rao, R. K. (1992). Appearance of exogenous epidermal growth factor in liver, bile and intestinal lumen of suckling rats. Gastroenterology 102, 661– 667. Kossor, D. C., Goldstein, R. S., Ngo, W., DeNicola, D. B., Leonard, T. B., Dulik, D. M., and Meunier, P. C. (1995). Biliary epithelial cell proliferation following alpha-naphthylisothiocyanate (ANIT) treatment: Relationship to bile duct obstruction. Fundam. Appl. Toxicol. 26, 51– 62. Kossor, D. C., Meunier, P. C., Handler, J. A., Sozio, R. A., and Goldstein, R. S. (1993). Temporal relationship of changes in hepatobiliary function and morphology in rats following a-naphthylisothiocyanate (ANIT) administration. Toxicol. Appl. Pharmacol. 119, 108 –114. Krstulovic, B., vanDamme, B., and Desmet, V. J. (1967). Comparative histochemical study of rat liver in bile-duct ligation and in alpha-naphthylisothiocyanate (ANIT) intoxication. Am. J. Pathol. 52(2), 423– 436. Liu, K., Kato, Y., Narukawa, M., Kim, D. C., Hanano, M., Higuchi, O., Nakamura, N., and Sugiyama, Y. (1992). Importance of the liver in plasma clearance of hepatocyte growth factor. Am. J. Physiol. 263, G642–G649. Lopez, M., and Mazzanti, L. (1955). Experimental investigation on alphanaphthylisothiocyanate as a hyperplastic agent of the biliary ducts in the rat. J. Pathol. Bacteriol. 69, 243–250. MacDonald, R. A., and Pechet, G. (1961). Liver cell regeneration due to biliary obstruction. Arch. Pathol. 72, 133–141. Mathis, G. A., and Sirica, A. (1990). Effects of medium and substratum conditions on the rates of DNA synthesis in primary cultures of bile ductular epithelial cells. In Vitro Cell. Dev. Biol. 26, 113–118. Mayol, X., Neal, G. E., Davies, R., Romero, A., and Domingo, J. (1992). Ethinyl estradiol-induced cell proliferation in rat liver. Involvement of specific populations of hepatocytes. Carcinogenesis 13(12), 2381–2388. Meyers, M., Slikker, W., Pascoe, G., and Vore, M. (1980). Characterization of cholestasis induced by estradiol-17b-D-glucuronide in the rat. J. Pharmacol. Exp. Ther. 214, 87–93.

338

KOSSOR ET AL.

Moran, E., Eliakim, M., Sucholowski, A., and Ungar, H. (1961). Serum vitamin B12 and glutamic-oxalacetic transaminase in experimental intrahepatic obstructive jaundice. Gastroenterology 40, 408 – 415. NIH Publication no. 86-23. (1985). Guide for the care and use of laboratory animals, National Research Council, Washington, DC. Nolan, J. P., Scheig, R. L., and Klatskin, G. (1966). Delayed hepatitis and cirrhosis in weanling rats following a single small dose of the senecio alkaloid, Lasiocarpine. Am. J. Pathol. 49(1), 129 –151. Salomon, W. L., Vreman, H. J., Kwong, L. K., and Stevenson, D. K. (1986). Red cell destruction and bilirubin production in adult rats with short-term biliary obstruction. J. Pediatr. Gastroenterol. Nutr. 5, 806 – 810. Slott, P. A., Liu, M. H., and Tavoloni, N. (1990). Origin, pattern, and mechanism of bile duct proliferation following biliary obstruction in the rat. Gastroenterology 99, 466 – 477.

Sokal, R. R., and Rohlf, F. J. (1981). Assumptions of Analysis of Variance. In Biometry, 2nd ed. (J. Wilson and S. Cotter, Eds.), Ch. 13, pp. 402– 412. W. H. Freeman, New York. Steiner, J. W., Carruthers, J. S., and Kalifat, S. R. (1962). The ductular cell reaction of rat liver in extrahepatic cholestasis. I. Proliferated biliary epithelial cells. Exp. Mol. Pathol. 1, 162–185. Svoboda, D., Grady, H. J., and Higginson, J. (1966). Aflatoxin B1 injury in rat and monkey liver. Am. J. Pathol. 49(6), 1023–1051. Tavoloni, N. (1987). The intrahepatic biliary epithelium: An area of growing interest in hepatology. Semin. Liver Dis. 7(4), 280 –292. Utili, R., Tripodi, M. F., Adinolfi, L. E., Gaeta, G. B., Abernathy, C. O., and Zimmerman, H. J. (1990). Estradiol-17b-D-glucuronide (E–17G) cholestasis in perfused rat liver: Fate of E–17G and choleretic responses to bile salts. Hepatology 11, 735–742.