Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Alzheimer's disease affects about 4 million people in the United States and is characterized by a degradation of cholinergic nerves in the cerebral cortex and hippocampus leading to a decrease in cholinergic transmission (Owens, 1993). Drug therapy to increase cholinergic transmission has been one strategy used to combat the symptoms of Alzheimer's disease (Starr, 1992; Volger, 1991). Tetrahydroaminoacridine (also known as THA, tacrine, and Cognex) is the first agent approved by the Food and Drug Administration for treatment of Alzheimer's disease. Tacrine acts as an acetylcholinesterase inhibitor to block the degradation of acetylcholine in the neurons of cerebral cortex thereby increasing cholinergic transmission (Wu and Yang, 1989; Hunter et al., 1989). Preclinical trials showed that many Alzheimer's patients had a significant improvement in symptoms when tacrine (2-3 mg/kg/day) was administered on a daily basis (Farlow et al., 1992; Gamzu et al., 1990; Summers et al., 1981). However, prolonged use of tacrine has proven to be hepatotoxic in about 30% of the patients (Elinder et al., 1989; Ames et al., 1990; Forsyth et al., 1989b; Hammel et al., 1990; Marx, 1987; O'Brien et al., 1991; Summers et al., 1981; Summers et al., 1989) by unknown mechanisms.

Liver injury due to tacrine is characterized by an elevation of the liver-specific enzyme AST (Summers et al., 1989). Histopathology, documented by liver biopsy, revealed pericentral necrosis and fat accumulation in midzonal regions as well as a few cases of drug-induced granulomatous hepatitis (Waktkins et al., 1994; O'Brien et al., 1991; Ames et al., 1990; Forsyth et al., 1989a). Because liver biopsy is not a practical alternative to monitor hepatotoxicity, blood must be drawn weekly to determine if liver enzyme levels are within the normal range. If AST levels are greater than two times normal, the patient is removed from tacrine for 4 to 6 wk or until AST levels decline to values within normal limits (Cognex insert, Parke-Davis, Ann Arbor, MI, 1993).

The mechanism by which tacrine causes hepatotoxicity has not been elucidated, in part, because of the lack of an animal model to study the injury. In safety studies using rats, mice and dogs, chronic tacrine treatment failed to produce liver injury (Fitten et al., 1987; Woolf et al., 1993). From these studies, it was suggested that these species are less vulnerable to tacrine hepatotoxicity because they produce fewer toxic metabolites (Hsu et al., 1990; Madden et al., 1993; Woolf et al., 1993). However, this hypothesis is not likely because hepatotoxicity of tacrine in humans correlates directly with the concentration of tacrine in the blood and not with the concentration of tacrine metabolites (Ford et al., 1993).

Our first goal was to develop an acute animal model in the rat to characterize the hepatotoxicity of tacrine by completing dose and time course studies of tacrine toxicity. Because tacrine caused a midzonal and pericentral injury, we hypothesized that it produced hypoxia in the liver leading to pericentral cell death. Accordingly, the second goal of these experiments was to use this new model to study the mechanism of acute tacrine hepatotoxicity. Preliminary accounts of this work have appeared elsewhere (Stachlewitz and Thurman, 1996).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals. Tacrine (9-amino-1,2,3,4-tetra-hydroacridine hydrochloride, catalog #A3773), trypan blue (#T0776), and catechin (#C1251, 98% pure) were purchased from Sigma Chemical Company (St. Louis, MO). -(4-pyridyl-1-oxide)-N-tert-butylnitrone (4-POBN, #P-0020) was purchased from OMRF Spin Trap Source (Oklahoma City, OK). Pimonidazole hydrochloride was synthesized according to published procedures (Smithen and Hardy, 1982) and characterized using standard chromatographic, elemental analysis and spectrographic techniques. Chemicals for the preparation of formalin-fixed, paraffin-embedded tissue sections were reagent grade purity from local suppliers. Vector Laboratories Inc. (Burlingame, CA) supplied the ABC peroxidase Vectastain kit, avidin-biotin blocking kit, rat adsorbed horse-antimouse antibodies and the DAB peroxidase substrate. The antibody to pimonidazole (Raleigh and Koch, 1990) was isotyped using a Clonotyping System/AP kit purchased from Fisher Scientific Company (Pittsburgh, PA).

Experimental animals. Female Sprague-Dawley rats (weight range 175-225 g) were given food and water ad libitum and treated intragastrically with up to 50 mg/kg tacrine dissolved in saline. Animals were anesthetized with sodium pentobarbital (50 mg/kg) or methoxyflurane before all surgical procedures. Blood was drawn from the descending vena cava for enzyme analysis and the liver was perfused as described below. All animals received humane care in accordance with institutional guidelines.

Microsurgical sympathectomy in the liver. Rats were anesthetized and maintained with methoxyflurane. The peritoneum was opened and the portal vein, hepatic artery and bile duct were exposed. A 3-mm wide region of tissue around the vessels and bile duct was removed to cut the sympathetic nerve (Cucchiaro et al., 1990). Additionally, ligaments surrounding the liver were also severed to remove sympathetic nerves entering via this route. Sham operations were also performed by opening the abdomen and exposing the portal vein for 10 min. The incision was closed and the rat was allowed to recover at least 72 hr before tacrine treatment.

Liver perfusion. Livers were perfused with Krebs-Henseleit-bicarbonate buffer (pH 7.4, 37°C) saturated with an oxygen-carbon dioxide mixture (95:5) in a nonrecirculating system as described elsewhere (Scholz et al., 1973). Perfusate leaving the liver was allowed to flow past a Clark-type oxygen electrode shielded by Teflon (Instech Laboratories, Plymouth Meeting, PA) to measure oxygen content in the perfusate. Oxygen uptake was calculated from the difference between oxygen concentration in the perfusate entering and leaving the system, the flow rate and the liver weight. Oxygen uptake values were used to assess tissue viability during perfusion.

Trypan blue distribution. Trypan blue infusion has been used to assess changes in tissue perfusion (Beckh et al., 1985; Ji et al., 1984; Zhong et al., 1996). A 0.5 mM trypan blue solution (Sigma) in Krebs-Henseleit-bicarbonate buffer saturated with an oxygen-carbon dioxide mixture (95:5) was infused into the perfused liver for up to 15 min at the end of each perfusion experiment (Belinsky et al., 1984). The time for trypan blue to distribute evenly to all regions of the liver was recorded to assess changes in tissue perfusion.

Use of pimonidazole to detect hypoxia in the liver. One hour after treatment with tacrine (35 mg/kg) or saline, animals received pimonidazole (120 mg/kg i.p.) (Arteel et al., 1995, 1996). Pimonidazole detects hypoxia in vivo by binding in cells at oxygen concentrations of 14 µM or less (Raleigh and Koch, 1990). After 2 hr, animals were anesthetized with pentobarbital (50 mg/kg i.p.) and livers were isolated surgically. The liver was perfusion-fixed with 10% formalin, and tissue sections were collected from the right lobe for immunohistochemical analysis of pimonidazole adduct binding.

Analysis of tissue-bound pimonidazole by immunohistochemistry. Paraffin blocks of formalin-fixed liver tissue were sectioned at 6 µm and pimonidazole adducts were detected with a biotin-streptavidin-peroxidase indirect immunostaining method modified for rat livers (Arteel et al., 1995). Sections were hydrated and treated briefly with 0.01% protease (pronase E) and exposed to mouse antipimonidazole IgG₁ antibody in PBS-Tween for 2 hr at 37°C. Rat adsorbed horse anti-mouse antibody was then applied to the sections for 30 min. Once the antibody-biotin-peroxidase complex was formed, DAB chromogen was added as the peroxidase substrate. After the immunostaining procedure, a light counterstain of hematoxylin was applied followed by mounting with crystal mount solution.

Image analysis of immunohistochemistry. A Universal Imaging Corp. image acquisition and analysis system (Chester, PA) incorporating an Axioskop 50 microscope (Carl Zeiss, Inc., Thornwood, NY) was used to capture and analyze the immunostained tissue sections at 40 × magnification as described previously (Arteel et al., 1995). Five pericentral fields were chosen randomly from each tissue section and positioned so that vessel lumenae were in the center of each field. Color detection thresholds were set for the red-brown color the DAB chromogen based on an intensely labeled point; subsequently, a default color threshold range was assigned. The percentage of pericentral area positive for pimonidazole was determined from the percentage of the field labeled intensely with DAB chromogen minus the acellular space. For uniformity, comparison of labeling was restricted to lumenae whose diameters fell in the range from 5 to 8 µm. To avoid the effects of nonspecific staining, regions near the edge of tissue sections were not evaluated.

In vivo portal pressure. A piece of tygon tubing (i.d. 3/16 inch) was attached to a meter stick that was positioned vertically as a manometer and a 22-gauge needle was placed at the end. The tubing was filled with Kreb's-Heinseliet buffer and clamped to keep the fluid from escaping. The needle was inserted into the portal vein and the clamp was removed. After the buffer stopped moving in the tube, a measurement was read from the meter stick in centimeters. This measurement represents the portal pressure in vivo in centimeters of water. The apparatus was routinely calibrated with a manometer.

Detection of free radical adducts. To assess free radical formation after tacrine treatment, the spin trapping reagent 4-POBN (250 mg/kg, dissolved in saline) was injected into the tail vein 7 hr after tacrine treatment. The abdomen was opened while the rat was under ether anesthesia and bile was collected for 1 hr via a cannula (polyethylene tubing #50) placed in the common bile duct into 50 µl dipyridyl on ice to prevent ex vivo free radical formation and stored on dry ice until analysis (Knecht et al., 1990). Bile samples were then thawed, placed in a quartz ESR cell, and scanned. Free radical adducts were detected with a Bruker ESP 106 ESR spectrometer. Instrument conditions were as follows: 20-mW microwave power; 1.007-G modulation amplitude and 80-G scan range.

Histology and serum enzymes. One centimeter thick sections of liver were placed in 1% paraformaldahyde for at least 24 hr. The fixed tissue was embedded in paraffin, prepared for light microscopy and stained with hemtatoxylin and eosin. Blood was collected from the descending vena cava or tail vein, placed in 1.5-ml plastic tubes and allowed to clot. Serum was separated by centrifugation (500 × g and stored at -20°C until AST activity was measured by standard enzymatic methods (Bergmeyer, 1983).

Statistics. All data are presented as mean ± S.E.M. Comparisons between groups were performed by Student's t test, analysis of variance or repeated measures analysis of variance followed by Bonferroni's post hoc test for multiple comparisons. The criterion for significance of P < .05 was selected before the study.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Characterization of tacrine-induced liver injury. Animals were treated with 10 to 50 mg/kg tacrine or a comparable amount of saline intragastrically 24 hr before serum was collected for measurement of AST as described in "Materials and Methods." Tacrine caused a significant increase in serum AST activity at 35 mg/kg (table 1). Animals treated with a higher dose of tacrine (50 mg/kg) died within 4 hr of acute cholinergic toxicity as evidenced by salivation, tremor and difficult breathing (Taylor, 1990) without any liver pathology; therefore, 35 mg/kg was selected for all subsequent studies.

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Time course of AST release after tacrine administration. Blood was drawn from the tail vein every 4 hr for 32 hr after tacrine (35 mg/kg) treatment for determination of AST as described in "Materials and Methods." Values are the mean ± S.E.M. (n = 4). Representative error bars are shown.

View larger version (167K): [in this window] [in a new window]   Fig. 2.   Liver histology from rats 24 hr after treatment with tacrine. Animals were given saline (A and B) or 35 mg/kg tacrine intragastrically (C and D) Livers were removed 24 hr later and prepared for light microscopy as described in "Materials and Methods." Representative morphology from six rats per group is shown. A and C are at low (100 × power) and B and D are at high (400×) power.

Role of hypoxia and the sympathetic nervous system in tacrine-mediated hepatotoxicity. To determine if tacrine caused cell death, livers were perfused and then infused with trypan blue at the end of the perfusion. About 10% of the cells in the pericentral region stained positive for trypan blue in histology. Yet, during perfusion, it was observed that the distribution of trypan blue in the circulation of the tacrine-treated liver was significantly impaired. Not only can trypan blue be used in histologically to stain dead cells in the liver (Belinsky et al., 1984), but time for vital dye to distribute evenly can be used to detect changes in intrahepatic perfusion (Beckh et al., 1985; Ji et al., 1984; Zhong et al., 1996). Therefore, the time for trypan blue to distribute evenly in the perfused liver was used as an index of changes in intrahepatic distribution of flow (fig. 3). Overall, there was a significant, nearly 7-fold increase in the time for trypan blue to distribute when tacrine-treated animals (6.9 ± 1.7 min) were compared to controls (1.0 ± 0.3 min), showing decreased tissue perfusion. However, there was no significant difference in the oxygen uptake of perfused livers ex vivo from tacrine-treated animals (112 ± 10 µmol/g/hr) compared to controls (105 ± 7 µmol/g/hr). When the hepatic nerve, which is known to control liver microcirculation, was severed 72 hr before tacrine treatment, the increase in trypan blue distribution time by tacrine was prevented.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Time for trypan blue to fully distribute in the perfused liver. Rats received tacrine (35 mg/kg) or a comparable volume of saline. Some animals had hepatic sympathectomy or an appropriate sham operation 72 hr before treatment with tacrine (35 mg/kg). Livers were perfused as in figure 3 and the time for trypan blue to fully distribute in the liver was recorded. Values are mean ± S.E.M. (n = 4). Statistical comparisons were made using analysis of variance and Bonferroni's post hoc test. a, different from saline-treated control. b, different from sham operated animal receiving tacrine.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Pimonidazole binding in pericentral regions of the liver lobule. Animals received 35 mg/kg tacrine or a comparable volume of saline intragastrically. Some animals had hepatic sympathectomy or a sham operation 72 hr before treatment with tacrine. One hour after tacrine, the animals received the hypoxia marker pimonidazole (120 mg/kg, i.p.) and the liver was perfusion-fixed 2 hr later with formalin. Pimonidazole binding was detected with immunohistochemistry and the average region of binding around the pericentral vein was determined by image analysis as described in "Materials and Methods." Values are mean ± S.E.M. (n = 4). Statistical comparisons were made using analysis of variance and Bonferroni's post hoc test. a, different from saline-treated control. b, different from sham operated animal receiving tacrine.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Effect of tacrine on portal pressure in vivo. Animals were treated with 35 mg/kg tacrine or a comparable volume of saline intragastrically. Some animals had hepatic sympathectomy or a sham operation 72 hr before treatment with tacrine. Portal pressure was measured 3 hr after tacrine treatment as described in "Materials and Methods." Values are the mean ± S.E.M. (n = 4). Statistical comparisons were made using analysis of variance and Bonferroni's post hoc test. a, different from saline-treated control. b, different from sham operated animal receiving tacrine.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Prevention of acute, tacrine-induced liver injury by severing the hepatic nerve. Animals were treated with 35 mg/kg tacrine or a comparable volume of saline intragastrically. Some animals had hepatic sympathectomy or a sham operation 72 hr before tacrine. AST levels in serum were measured 24 hr later as described in "Materials and Methods." Values are the mean ± SEM (n = 6, 35 mg/kg tacrine treatment, n = 4 all other groups). Statistical comparisons were made using analysis of variance and Bonferroni's post hoc test. a, different from saline treated control. b, different from sham operated animal receiving tacrine.

Prevention of tacrine hepatotoxicity by hepatic denervation is not due to inability to regulate body temperature. Because tacrine is extensively metabolized by the liver and hepatotoxicity has been proposed to be caused by production of toxic metabolites, one possible explanation for the observations in these studies is that severing the hepatic nerve may decrease body temperature and inhibit metabolism of tacrine. Accordingly, the time course of changes in body temperature was measured after tacrine treatment (fig. 7). As with other acetylcholinesterase inhibitors (Kooka et al., 1987), tacrine treatment caused a significant decrease in body temperature of about 3°C at 4 and 8 hr that was unaffected by prior surgery. Body temperature returned to near pretreatment values at 24 hr in all groups.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Time course of changes in body temperature after tacrine treatment. Basal body temperature was determined at 0 hr and animals were given 35 mg/kg tacrine. Some animals had hepatic sympathectomy or comparable sham operation 72 hr before tacrine. Body temperature was measured every four hr. Values are mean ± S.E.M. (n = 4). Statistical comparisons were made using repeated measures analysis of variance and Bonferroni's post hoc test. a, all points at this time are different from respective controls at 0 hr. b, denotes point is different from control at 0 hr.

Evidence for free radical production and reperfusion injury after tacrine treatment. It is possible that the injury due to tacrine treatment is not the result of prolonged hypoxia but instead occurs upon reperfusion when the vascular space increases (i.e., after tacrine is metabolized) and oxygen reenters previously hypoxic cells leading to the production of free radicals that cause cell injury. Using the spin trapping technique, a free radical adduct was detected in the bile of tacrine-treated animals 8 hr after treatment (fig. 8). The radical spectrum simulated as a mixture of two radical species (species I: a^N = 15.8 and a^H = 2.1, species II: a^N = 15.6 and a^H = 3.4). The coupling constants of species I is consistent with a carbon-centered free radical while species II is consistent with an oxygen-centered free radical.

View larger version (14K): [in this window] [in a new window]   Fig. 8.   Detection of free radical adducts in the bile after tacrine treatment. Tacrine (35 mg/kg) was given intragastrically. After 7 hr, rats were anesthetized and 4-POBN (250 mg/kg) was injected i.v. Bile was collected for the next hour into dipyridyl to prevent ex vivo radical formation, and free radical adducts were detected by ESR as described in "Materials and Methods." Data shown are of a typical spectrum.

View larger version (12K): [in this window] [in a new window]   Fig. 9.   Effects of catechin treatment on tacrine-induced liver injury. Catechin (100 mg/kg, i.p.) or a comparable amount of saline was administered 1 hr before treatment with tacrine (35 mg/kg, intragastrically). Sixteen hours after drug treatment, blood was collected and AST determined as described in "Materials and Methods." Values are mean ± S.E.M. (n = 4). Statistical comparisons were made using analysis of variance and Bonferroni's post hoc test. a, different from saline-treated control. b, different from tacrine-treated animals.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Development and characterization of a new model of tacrine-induced liver injury. In this study, a new model has been developed in the rat to study the hepatotoxicity of tacrine. Tacrine caused a significant increase in serum AST, a specific marker for liver toxicity (table 1), and injury in midzonal and pericentral regions of the liver lobule as well as fatty changes (fig. 2) at 35 mg/kg. Lower doses, as reported by others (Hunter et al., 1989), do not cause significant liver injury. This pattern of injury 24 hr after 35 mg/kg tacrine is similar to that observed with hypoxia (Lemasters et al., 1981). Additionally, it resembles the pathology observed in the clinic when patients are given tacrine chronically (Waktkins et al., 1994; O'Brien et al., 1991; Ames et al., 1990; Forsyth et al., 1989a; Hammel et al., 1990). Furthermore, tacrine decreased tissue perfusion (fig. 3), increased portal pressure in vivo (fig. 5) and increased pimonidazole adduct binding in the pericentral region of the lobule (fig. 4) providing evidence that tacrine causes hypoxia in the liver by altering intrahepatic blood flow and decreasing tissue perfusion. Collectively, these data support the hypothesis that acute tacrine treatment causes hypoxia.

Role of the sympathetic nervous system in the mechanism of tacrine-induced hepatotoxicity. It is well known that the sympathetic nervous system plays an important role in control of the microcirculation in the liver (Lautt, 1980; Gardemann et al., 1992). The pathway by which the liver is innervated is shown schematically in figure 10. The sympathetic nerve controlling the liver vasculature originates from the cholinergic celiac ganglion. Acetylcholine is released in the celiac ganglion and causes an action potential that propagates through the hepatic nerve, which is the afferent, sympathetic pathway to the liver. It is postulated that norepinephrine is released directly onto endothelial cells and stellate cells in periportal regions of the liver, decreasing sinusoidal vascular space and tissue perfusion (Gardemann et al., 1992). Infusion of noradrenaline or electrical stimulation of the sympathetic pathway in the perfused liver alters the distribution of trypan blue in the liver and decreases surface oxygen tension (Beckh et al., 1985; Ji et al., 1984). Prolonged hypoxia is known to cause cell death; therefore, it was hypothesized that tacrine, by increasing acetylcholine at the celiac ganglion, would increase sympathetic activity via the hepatic nerve and release norepinephrine in the liver causing injury. To test this hypothesis, the sympathetic afferent to the liver was severed. Hepatic sympathectomy prevented the decrease in tissue perfusion (fig. 3), the increase in portal pressure (fig. 5), and the increase in binding of the hypoxia marker pimonidazole in pericentral regions of the liver (fig. 4). The increase in AST (fig. 6) caused by tacrine was also reduced significantly by cutting the sympathetic hepatic nerve. These data support the hypothesis that the sympathetic hepatic nerve plays a central role in the mechanism of tacrine-induced liver injury.

View larger version (27K): [in this window] [in a new window]   Fig. 10.   Proposed mechanism of tacrine-induced hepatotoxicity. The sympathetic, afferent nerve bundle of the liver is illustrated. Based on data from this new model of tacrine-induced liver injury, it is hypothesized that tacrine acts at the celiac ganglion to increase sympathetic activity to the liver leading to diminished sinusoidal space, decreased blood flow and hypoxia. As tacrine is metabolized, free radicals are formed leading to reperfusion injury as oxygen delivery increases in previously hypoxic cells. AR, Acetylcholine release; NE, norepinephrine release.

Reperfusion injury occurs after tacrine. As tacrine is metabolized, it is proposed that resistance in the sinusoids decreases because of diminishing stimulation of the sympathetic hepatic nerve. Blood flow to the liver increases, restoring oxygen to previously hypoxic cells (fig. 10). Reperfusion of the liver after hypoxia could result in production of free radicals by activated Kupffer cells or xanthine oxidase leading to parenchymal cell injury (Lemasters and Thurman, 1993). Indeed, we have detected two free radical species 8 hr after tacrine treatment by spin trapping and ESR spectroscopy (fig. 8). Furthermore, the free radical scavenger catechin was shown to decrease hepatotoxicity resulting from tacrine treatment (fig. 9). Collectively, these data are consistent with the hypothesis that a significant component of tacrine-mediated hepatotoxicity may not be prolonged hypoxia per se, but rather the result of increased free radical formation upon reperfusion of previously hypoxic tissue.