Anticoagulants Prevent Monocrotaline-Induced Hepatic Parenchymal Cell Injury but Not Endothelial Cell Injury in the Rat

Toxicology and Applied Pharmacology 180, 186 –196 (2002) doi:10.1006/taap.2002.9394, available online at http://www.idealibrary.com on

Anticoagulants Prevent Monocrotaline-Induced Hepatic Parenchymal Cell Injury but Not Endothelial Cell Injury in the Rat Bryan L. Copple, Brook Woolley, Amy Banes, Patricia E. Ganey, and Robert A. Roth Department of Pharmacology and Toxicology Michigan State University, East Lansing, Michigan 48824 Received November 9, 2001; accepted February 27, 2002

Anticoagulants Prevent Monocrotaline-Induced Hepatic Parenchymal Cell Injury but Not Endothelial Cell Injury in the Rat. Copple, B. L., Woolley, B., Banes, A., Ganey, P. E., and Roth, R. A. (2002). Toxicol. Appl. Pharmacol. 180, 186 –196. Monocrotaline (MCT) is a pyrrolizidine alkaloid plant toxin that produces hepatotoxicity in humans and animals. Human exposure to MCT occurs through consumption of contaminated grains and herbal medicines. Administration of MCT to rats stimulates activation of the coagulation system and fibrin deposition in the liver. Fibrin deposition occurs simultaneously with endothelial cell damage and prior to hepatic parenchymal cell injury. Accordingly, the hypothesis that activation of the coagulation system is required for MCT-induced liver injury was tested. Treatment of rats with either heparin or warfarin significantly reduced MCTinduced activation of the coagulation system and the increase in alanine aminotransferase activity in the plasma, a biomarker of hepatic parenchymal cell injury. Histopathological examination of liver sections revealed that heparin decreased parenchymal cell necrosis but did not affect central venular endothelial cell damage, congestion and dilation of the sinusoids, or hemorrhage in the liver. Morphometric analysis revealed that 28% of the area of livers from MCT-treated rats contained regions of coagulative necrosis, whereas less than 5% of the area of livers from rats treated with MCT and heparin contained these regions. By contrast, neither heparin nor warfarin prevented MCT-induced damage to endothelial cells in the liver as estimated by increased plasma hyaluronic acid concentration. These results suggest that activation of the coagulation system is required for MCT-induced parenchymal cell injury but not endothelial cell injury in the liver. © 2002 Elsevier Science (USA) Key Words: monocrotaline; liver; rat; fibrin deposition; heparin; warfarin; parenchymal cell injury; coagulation; thrombin; endothelial cell injury; pyrrolizidine alkaloid; liver injury; hepatotoxicity.

Monocrotaline (MCT) is a pyrrolizidine alkaloid (PA) plant toxin that produces hepatotoxicity which can progress to hepatic veno-occlusive disease (HVOD) in animals and humans (Brooks et al., 1970; Allen et al., 1969; Schoental and Head, 1955). People become exposed to PAs through ingestion of herbal medicines containing components of PA-containing 0041-008X/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.

plants (Schultze and Roth, 1998; Huxtable, 1989). Although cases of HVOD still result from PA poisonings, most cases today result from high-dose cytoreductive therapy for bone marrow and stem cell transplantation (DeLeve, 2001; Carreras, 2000; Bearman, 2000). This disease is one of the leading causes of mortality in bone marrow transplant patients. In addition, risk of developing this disease precludes the use of some treatment protocols for patients with preexisting conditions that confer a high risk of developing HVOD (DeLeve, 2001). Accordingly, it is important to understand what factors contribute to the initiation and progression of this disease. Because of the similarity of HVOD produced after bone marrow transplantation and PA exposure, treatment of rats with MCT has been used as a model of HVOD (DeLeve et al., 1999). Endothelial cell damage in the liver may be the initial insult that promotes development of HVOD (DeLeve, et al., 1999). Treatment of rats with MCT produces endothelial cell damage, congestion and dilation of the sinusoids, and hemorrhage in the liver (Schoental and Head, 1955; DeLeve, et al., 1999; Wang et al., 2000). Histopathological examination of liver sections by light microscopy shows evidence of central venular endothelial cell (CVEC) damage (Schoental and Head, 1955; DeLeve, et al., 1999; Wang et al., 2000). In addition, damage to sinusoidal endothelial cells (SECs) in the centrilobular region occurs after MCT exposure (DeLeve, et al., 1999; Copple et al., 2002). These early changes are followed by coagulative, centrilobular hepatic parenchymal cell necrosis, and, in later stages, fibrosis, which leads to partial or complete occlusion of central and sublobular veins (DeLeve, et al., 1999). In vitro, SECs are more sensitive than hepatic parenchymal cells to the toxic effects of compounds that produce HVOD, such as MCT (DeLeve et al., 1996; DeLeve, 1994, 1996). Therefore, it has been suggested that MCT-induced damage to SECs in vivo results in microcirculatory disturbances that lead to hypoperfusion of the liver and ischemic parenchymal cell injury (DeLeve, et al., 1996). Another contributing factor may be the coagulation system. The coagulation system is activated in rats after exposure to MCT and in human patients who develop HVOD after bone marrow transplantation (Schoental and Head, 1955; Butler et al., 1970; Korte, 1997). We have

186

HEPARIN PREVENTS MONOCROTALINE-INDUCED LIVER INJURY

shown recently that treatment of rats with MCT activates the coagulation system and causes fibrin deposition in the liver (Copple et al., 2002). Fibrin deposition occurs simultaneously with endothelial cell damage and before hepatic parenchymal cell injury. In addition, fibrin deposition occurs only in centrilobular regions where endothelial cell and parenchymal cell injury are most severe. This suggests the possibility of a causal link between fibrin deposition and hepatocellular injury, which is supported by observations with other models of liver injury. Indeed, several hepatotoxicants such as bacterial lipopolysaccharide, dimethylnitrosamine and 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea (CCNU) promote activation of the coagulation system and fibrin deposition in the liver (Pearson et al., 1996; Arai et al., 1996; Fujiwara et al., 1988; Ahmed et al., 1987). Moreover, inhibition of the coagulation system completely prevents hepatic parenchymal cell injury from these agents. Thus, the coagulation system may be a critical mediator of liver injury after exposure to certain hepatotoxicants. Considering the extent of fibrin deposition in the liver and the nature and location of the hepatic lesions, it is reasonable to hypothesize that coagulation system activation and fibrin deposition are required for MCT-induced liver injury. To test this, rats were treated with MCT and the anticoagulants heparin and warfarin. The results show that anticoagulant treatment decreased MCT-induced activation of the coagulation system and hepatic parenchymal cell injury. By contrast, neither anticoagulant had an effect on damage to endothelial cells in the liver after MCT treatment. These results suggest that the coagulation system may be a critical component of MCT-induced hepatic parenchymal cell injury.

In all studies, rats were killed 14 h after MCT treatment to assess hepatic injury and activation of the coagulation system. The injection volume for all chemicals was 2 ␮l/g of rat. Assesment of hepatic injury and plasma fibrinogen. Fourteen hours after treatment with MCT or its vehicle, rats were anesthetized with sodium pentobarbital (50 mg/kg ip). A midline abdominal incision was made, and blood was collected from the descending aorta into a syringe containing sodium citrate (final concentration, 0.38%). Hepatic parenchymal cell injury was evaluated by measuring the activity of alanine aminotransferase (ALT) in the plasma using Sigma kit 59-UV (Sigma Chemical Co.). Fibrinogen concentration was determined from the thrombin clotting time of diluted samples by using a fibrometer and a commercially available kit (Sigma Kit 886-A; Sigma Chemical Co.). Plasma hyaluronic acid was measured using an enzyme-linked immunosorbent assay (Chugai Diagnostic Science Co. Tokyo, Japan). One transverse section from the middle of the left lateral liver lobe and one from the right lateral lobe were fixed in 10% buffered formalin and processed for light microscopy. Paraffin-embedded sections were cut at 5 ␮m, stained with H & E, and evaluated for lesion size and severity as described previously with slight modification (Yee et al., 2000). A computerized image analysis system was employed to quantify treatment-induced changes in liver morphology. This system consisted of a light microscope (Olympus AX-80T, Olympus, Lake Success, NY) interfaced with a SPOT II camera (Diagnostic Instruments, Sterling Heights, MI) and a computer with SPOT Advanced Software. Digitized color images of H & E-stained liver sections were evaluated with a 20-point lattice grid (Adobe Photoshop, Adobe Systems, Inc, San Jose, CA) to determine (1) the total area of the liver analyzed, (2) the area of the lesion, (3) the area of single-cell necrosis, (4) the area of coagulative necrosis, (5) the area of normal parenchyma, and (6) the area of nonparenchymal space. Necrotic hepatic parenchymal cells were defined as parenchymal cells with either swollen, eosinophilic cytoplasm and karyolytic or pyknotic nuclei (i.e., oncosis) or cells with shrunken cytoplasm and karyorrhexic nuclei or apoptotic bodies (i.e., apoptosis) (Yee et al., 2000; Levin et al., 1999). Coagulative necrosis was defined as a region of parenchymal cell necrosis containing four or more juxtaposed necrotic parenchymal cells. Nonparenchymal space was defined as nonparenchymal tissue, vessel lumen, and regions outside the perimeter of the liver section. The area of each object of interest was calculated from the following expression (Yee et al., 2000):

MATERIALS AND METHODS AreaInterest ⫽ Animals. Male Sprague–Dawley rats (Crl:CD BR(SD) VAF/Plus, Charles River, Portage, MI) weighing 100 –130 g were used in all studies. Animals were allowed food (Rodent Chow/Tek 8640, Harlan Teklad, Madison, WI) and water ad libitum. They were housed no more than three to a cage on Aspen chip bedding (Northeastern Products Company, Warrenburg, NY) and were maintained on a 12-h light/dark cycle in a controlled temperature (18 –21°C) and humidity (55 ⫾ 5%) environment for a period of 1 week before use. All procedures on animals followed the guidelines for humane treatment set by the American Association of Laboratory Animal Sciences and the University Laboratory Animal Research Unit at Michigan State University. Treatment protocol. Rats were fasted for 24 h before treatment with MCT. They received MCT (300 mg/kg; Trans World Chemicals, Inc., Rockville, MD) or an equivalent volume of sterile saline vehicle by ip injection. Food was returned to the rats after MCT treatment. They then received heparin (3000 U/kg; Sigma Chemical Co., St. Louis, MO) or an equivalent volume of sterile saline vehicle by sc injection 6 h later. MCT was dissolved in sterile saline, minimally acidified with 0.2 M HCl. The pH was brought to 7 by addition of 2 M NaOH, and the volume was adjusted with sterile saline to the appropriate final concentration. To investigate the effect of warfarin on MCT-induced liver injury, rats were treated with 5 or 10 mg warfarin/kg or an equivalent volume of dimethylsulfoxide (DMSO) vehicle by ip injection. They then received a second injection of warfarin (5 or 10 mg/kg) or DMSO vehicle 24 h later. Finally, the rats were treated with MCT (300 mg/kg ip) or saline vehicle 10 h after the second warfarin administration.

187

冘

pointsInterest ⫻ area/point,

where Area/point ⫽ (distance between points) 2. Distance between points was 70 ␮m. Accordingly, the area represented by each point was 4900 ␮m 2. Two sections from different lobes of the liver were analyzed for each animal in a treatment group. Samples were coded such that the evaluator was not aware of the treatment. They were systematically scanned using adjacent, nonoverlapping microscopic fields. The first image field analyzed in each section was chosen using a random number table (i.e., any image field between 1 and 20). Thereafter, every 10th field containing hepatic parenchymal cells was evaluated (minimum of 20 fields measured per section). A minimum of five animals per group was analyzed. The percentage lesion area (i.e., total, single-cell, or coagulative necrosis) was estimated based on the following formula: [Area Lesion/(Area Lesion ⫹ Area Normal Parenchyma)] ⫻ 100. The Ehrlich assay was used to measure the concentration of pyrroles in liver as a marker of bioactivation of MCT to toxic pyrrolic metabolites (Mattocks and White, 1970). This was done to determine whether warfarin prevented bioactivation of MCT to MCT pyrrole. The Ehrlich assay was performed as described previously (Yee et al., 2000). Liver hemoglobin. An additional piece of liver was snap frozen in liquid nitrogen and stored at ⫺70°C for evaluation of liver hemoglobin content. The concentration of liver hemoglobin was used as a biomarker of hemorrhage and was measured using a commercially available kit (Total Hemoglobin Kit, Sigma Chemical) as described previously (Jaeschke et al., 2000; Copple et al., 2002). A 20% homogenate was made from samples of frozen liver in 50 mM

188

COPPLE ET AL.

sodium phosphate buffer (120 mM NaCl, 10 mM EDTA). The samples were centrifuged at 16,000g for 10 minutes at 4°C, and 200 ␮l of the supernatant was diluted in Drabkin’s solution. After a 15-min incubation, the absorbance was measured at 540 nm, and the concentration of hemoglobin was calculated from a standard curve. Immunohistochemistry. A 1-cm 3 block cut from the middle of the left lateral liver lobe was frozen for 8 min in isopentane immersed in liquid nitrogen. For liver endothelial cell immunostaining, 8-␮m-thick sections of frozen liver were fixed for 10 min in acetone (⫺20°C). Sections were blocked for 30 min with phosphate-buffered saline (PBS) containing 10% goat serum (i.e., blocking solution; Vector Laboratories, Burlingame, CA) and then incubated with mouse anti-rat RECA-1 (rat endothelial cell antigen-1, Serotec, Inc., Raleigh, NC) diluted (1:20) in blocking solution overnight at 4°C. The RECA-1 antibody binds to rat endothelium but not other cell types (Duijvestijn et al., 1992). In the liver, this antibody stains both SECs and endothelial cells of larger venules. After incubation with the RECA-1 antibody, sections were incubated for 3 h with goat anti-mouse secondary antibody conjugated to Alexa 594 (1:1000; Molecular Probes, Eugene OR) in blocking solution containing 2% rat serum. Sections were washed three times, for 5 min each, with PBS and visualized using a fluorescent microscope. For fibrin immunostaining, 8-␮m-thick sections of frozen liver were fixed in 10% buffered formalin containing 2% acetic acid for 30 min at room temperature. This fixation protocol solubilizes all fibrinogen and fibrin species except for cross-linked fibrin; therefore, only cross-linked fibrin stains in sections of liver (Schnitt et al., 1993). Sections were blocked for 30 min with PBS containing 10% horse serum (i.e., blocking solution; Vector Laboratories) and then incubated with goat anti-rat fibrinogen diluted (1:2000, ICN Pharmaceuticals, Aurora, OH) in blocking solution overnight at 4°C. Next, sections were incubated for 3 h with donkey anti-goat secondary antibody conjugated to Alexa 594 (1:1000 dilution, Molecular Probes) in blocking solution. Sections were washed three times, for 5 min each, with PBS and visualized using a fluorescent microscope. For both protocols, no staining was observed in controls in which the primary or secondary antibody was omitted from the staining protocol. All treatment groups that were compared morphometrically were immunohistochemically stained at the same time. Morphometric quantitation of endothelial cells and fibrin in the liver. Endothelial cells and fibrin deposition in the liver were quantified morphometrically by analyzing the area of immunohistochemical staining for each in sections of liver (Copple et al., 2002). A decrease in the area of immunohistochemical staining for endothelial cells suggests a loss of these cells in the liver. An increase in the area of immunohistochemical staining for fibrin indicates fibrin deposition. For morphometric analysis of the area of endothelial cell or fibrin staining in a liver section, digital images of 10 randomly chosen, low-power (100⫻ magnification) fields per tissue section were captured using a SPOT II camera and SPOT Advanced Software. Samples were coded so that the evaluator was not aware of the treatment. The same exposure time was used for all captured images, and each digital image encompassed a total area of 1.4 mm 2. The area of immunohistochemical staining (number of pixels) was quantified using Scion Image Software (Scion Corporation, Frederick, MD). For endothelial cell quantification, a density slice was taken from an inverted, gray-scale digital image of a liver section from a saline-treated rat. A density slice allows analysis of pixels in a defined range of gray values (i.e., densities). Next, the threshold was adjusted such that background staining was minimized and only endothelial cell staining was counted. For quantification of fibrin deposition, the threshold was selected so that little positive staining was present in saline-treated controls. The same threshold value was used to analyze digital images from all treatment groups. For quantification of endothelial cells or fibrin in a low-power field of liver, the area of positive staining was measured and divided by the total area of the image. The staining is expressed as a fraction of the total area. The 10 random fields analyzed for each rat liver were averaged and counted as one replicate, i.e., each replicate represents a different rat.

FIG. 1. MCT-induced changes in plasma fibrinogen in rats treated with and without heparin or warfarin. (A) Rats were treated with 300 mg MCT/kg ip or saline vehicle and then with 3000 U heparin/kg sc or saline vehicle 6 h later. (B) Rats were treated with 5 or 10 mg warfarin/kg ip or DMSO vehicle. They then received a second injection of warfarin (5 or 10 mg/kg) or DMSO vehicle 24 h later followed by treatment with MCT (300 mg/kg) or saline vehicle 10 h after the second warfarin injection. Plasma fibrinogen was evaluated 14 h after MCT treatment. Data are expressed as means ⫾ SEM. (A) n ⫽ 13 rats per group; (B) n ⫽ 6 rats per group. *Significantly different ( p ⬍ 0.05) from saline-treated control. **Significantly different ( p ⬍ 0.05) from MCT/saline-treated rats.

Statistical analysis. Results are presented as the mean ⫾ SEM. Data from studies of the effect of heparin or warfarin on MCT-induced changes in plasma fibrinogen, plasma hyaluronic acid, fibrin deposition, plasma ALT, RECA-1 staining, and liver hemoglobin were analyzed by using a 2 ⫻ 2 multifactorial, completely random ANOVA. ANOVAs were performed on log X-transformed data in instances in which variances were not homogeneous. Data expressed as a fraction of total or percentage were transformed by arc sine square root prior to analysis. Comparisons among group means were made using the Student– Newman–Keuls test. The criterion for significance was p ⬍ 0.05 for all studies.

RESULTS

Effect of heparin or warfarin on MCT-induced activation of the coagulation system. To investigate the role of the coagulation system in MCT-induced liver injury, two anticoagulants were used that inhibit thrombin-catalyzed conversion of circulating fibrinogen to fibrin by different mechanisms. Heparin inhibits thrombin by increasing its affinity for antithrombin III, thereby inhibiting its proteolytic activity. In contrast,

HEPARIN PREVENTS MONOCROTALINE-INDUCED LIVER INJURY

189

FIG. 2. Fibrin deposition in the livers of MCT-treated rats given heparin. Rats were treated with either 300 mg MCT/kg ip or saline vehicle and with either 3000 U heparin/kg sc or saline vehicle 6 h later. Livers were removed 14 h after MCT treatment and processed for immunohistochemical staining of fibrin, which appears black in the photomicrographs. Liver sections from (A) a saline/saline-treated rat, (B) a saline/heparin-treated rat, (C) a MCT/saline-treated rat, and (D) a MCT/heparin-treated rat. CV, centrilobular; PP, periportal. Bar represents 100 ␮m.

the anticoagulant effects of warfarin are due to an alteration in the vitamin K-dependent synthesis of prothrombin in the liver. Activation of the coagulation system leads to the conversion of plasma fibrinogen into fibrin, resulting in a decrease in circulating fibrinogen. This change in plasma fibrinogen was used as an indicator of activation of the coagulation system. Rats were treated with either 300 mg/kg MCT or saline vehicle followed 6 h later by either 3000 U/kg heparin or saline vehicle. Heparin was given at this time because it is before the significant decrease in plasma fibrinogen that occurs between 8 and 12 h after MCT treatment (Copple et al., 2002). In addition, by this time the bioactivation of MCT to MCT pyrrole, its toxic metabolite, is nearly complete (Allen et al., 1972). Treatment of rats with MCT/saline caused a significant decrease in

plasma fibrinogen 14 h after treatment (Fig. 1A). Heparin significantly modulated this decrease. To investigate the effect of warfarin on MCT-induced activation of the coagulation system, rats were treated with 5 or 10 mg warfarin/kg or an equivalent volume of DMSO vehicle. They then received a second injection of warfarin (5 or 10 mg/kg) or DMSO vehicle 24 h later. This was followed by treatment with MCT (300 mg/kg) or vehicle 10 h after the second warfarin administration. Both doses of warfarin completely prevented MCT-induced activation of the coagulation system (Fig. 1B). Liver tissue pyrrole concentration was measured to determine if warfarin exposure altered the bioactivation of MCT to MCT pyrrole. Rats were treated with warfarin and MCT as described above, and pyrrole activity was deter-

190

COPPLE ET AL.

FIG. 3. Quantitative morphometry of fibrin deposition in the liver. Rats were treated with either 300 mg MCT/kg ip or saline vehicle and then with either 3000 U heparin/kg sc or saline vehicle 6 h later. Livers were removed 14 h after MCT treatment and processed for immunohistochemical staining of fibrin. The total area of fibrin staining in 10 randomly chosen, low-power (100⫻) fields per tissue section; 2 sections per liver, and 5 animals per group was analyzed morphometrically. Data are expressed as means ⫾ SEM. *Significantly different ( p ⬍ 0.05) from respective saline-treated control. **Significantly different ( p ⬍ 0.05) from MCT/saline-treated rats.

mined in the liver 1.5 h after MCT administration. At this time, the bioactivation of MCT to MCT pyrrole is nearly complete (Allen et al., 1972). The concentration of pyrrole in the liver after treatment with MCT and DMSO vehicle was 39.59 ⫹/⫺ 5.00 ␮g/mg of liver tissue. The concentration of pyrrole in the liver after treatment with MCT and 5 or 10 mg warfarin/kg was 38.00 ⫹/⫺ 2.01 and 39.18 ⫹/⫺ 2.83 ␮g/mg liver tissue, respectively. Therefore, warfarin had no significant effect on the bioactivation of MCT. In liver sections from saline/saline- and saline/heparintreated rats, no immunohistochemical staining for fibrin was observed in the sinusoids (Figs. 2A and 2B, respectively). In both groups, staining was observed along the intima of large vessels in the centrilobular and periportal regions. This staining could not be eliminated with increased blocking or washing of the tissue and may have resulted from fibrin deposition that occurred after euthanasia of the rat. Treatment of rats with MCT resulted in extensive fibrin staining in the sinusoids of the liver 14 h after MCT treatment (Fig. 2C). This increase in fibrin staining was dramatically reduced by heparin treatment (Fig. 2D). Morphometric analysis revealed that MCT caused a significant increase in the area of fibrin staining in the liver 14 h after MCT treatment (Fig. 3). Heparin reduced this increase by approximately 70%. Effect of heparin or warfarin on MCT-induced liver injury. MCT/vehicle produced significant hepatic parenchymal cell injury 14 h after injection of MCT, as measured by release of ALT into the plasma (Fig. 4). This was significantly reduced by heparin (Fig. 4A) or warfarin (Fig. 4B) administration. No lesions were observed in livers from saline/saline- or saline/heparin-treated rats (data not shown). Centrilobular, hepatic parenchymal cell necrosis was typical in livers isolated from rats treated 14 h earlier with MCT/saline (Fig. 5A). Lesions consisted of areas of coagulative and single-cell ne-

crosis. Hemorrhage and widening of the sinusoids was evident and occurred predominantly in centrilobular areas. Damage to CVECs also occurred and frequently progressed to complete loss of vascular intima. In livers from MCT/heparin-treated rats, hepatic parenchymal cell necrosis was dramatically reduced 14 h after MCT treatment (Fig. 5B). Areas of coagulative necrosis in MCT-treated rats given heparin were atypical, and the remaining lesions consisted of single, necrotic parenchymal cells. Heparin treatment did not prevent MCT-induced hemorrhage, widening of the sinusoids, or damage to CVECs. Morphometric analysis of livers from rats treated with MCT revealed that approximately 35% of the liver contained lesions consisting of necrotic parenchymal cells (Fig. 6). By contrast, only 16% of the liver area contained lesions after treatment with MCT and heparin (Fig. 6). Next, morphometric analysis was performed to assess the areas of coagulative vs single-cell necrosis. Approximately 23% of the liver contained regions of coagulative necrosis after treatment with MCT (Fig. 6). Hep-

FIG. 4. Effect of heparin or warfarin on MCT-induced hepatic parenchymal cell injury. (A) Rats were treated with either 300 mg MCT/kg ip or saline vehicle and then with either 3000 U heparin/kg sc or saline vehicle 6 h later. (B) Rats were treated with 5 or 10 mg warfarin/kg ip or DMSO vehicle. They then received a second injection of warfarin (5 or 10 mg/kg) or DMSO vehicle 24 h later followed by treatment with MCT (300 mg/kg) or saline vehicle 10 h after the second warfarin injection. Plasma ALT activity was assessed 14 h after MCT treatment as a marker of hepatic parenchymal cell injury. Data are expressed as means ⫾ SEM. (A) n ⫽ 13 rats per group, (B) n ⫽ 6 rats per group. *Significantly different ( p ⬍ 0.05) from saline/saline-treated rats. **Significantly different ( p ⬍ 0.05) from MCT/saline-treated rats.

HEPARIN PREVENTS MONOCROTALINE-INDUCED LIVER INJURY

191

FIG. 5. Photomicrographs of liver sections from rats treated with MCT and heparin. Rats were treated with either 300 mg MCT/kg ip or saline vehicle and then with either 3000 U heparin/kg sc or saline vehicle 6 h later, and livers were removed 14 h after MCT treatment and processed for H & E staining. (A) Liver of a rat treated with MCT/saline. Coagulative necrosis (arrow) was present in the centrilobular region. (B) Liver of a rat treated with MCT/heparin. Bar represents 50 ␮m.

arin treatment significantly reduced this to less than 5%. By contrast, there was no difference in the area of single-cell necrosis between the two MCT-treated groups.

Effect of heparin or warfarin on MCT-induced damage to liver endothelial cells. To determine whether anticoagulants prevented MCT-induced damage to SECs in the liver,

192

COPPLE ET AL.

Effect of heparin on MCT-induced hemorrhage in the liver. To quantify the degree of hemorrhage in the liver after MCT treatment, the concentration of hemoglobin in the liver was determined. Treatment of rats with MCT caused a significant increase in the concentration of hemoglobin in the liver 14 h after treatment (Fig. 10). This increase was not prevented by heparin. DISCUSSION

FIG. 6. The area of hepatic lesions in rats treated with MCT and heparin. Rats were treated with either 300 mg MCT/kg ip or saline vehicle and then with either 3000 U heparin/kg sc or saline vehicle 6 h later. Livers were removed 14 h after MCT treatment, processed, and the areas of total necrosis, single-cell necrosis, and coagulative necrosis were quantitated as described under Materials and Methods. Data are expressed as means ⫾ SEM. n ⫽ 8 rats. *Significantly different ( p ⬍ 0.05) from MCT/saline-treated rats.

plasma hyaluronic acid was measured as a marker of injury to these cells. Under normal conditions, approximately 90% of the hyaluronic acid circulating in the blood is taken up and degraded by SECs in the liver (Kobayashi et al., 1999). Damage to SECs can impair the ability of these cells to clear hyaluronic acid from the circulation, leading to an increase in its concentration in the blood. This method has been used to monitor SEC damage in vivo after exposure to toxicants (Deaciuc et al., 1993, 1994), including MCT (Copple et al., 2002). Treatment of rats with MCT produced a significant increase in the plasma concentration of hyaluronic acid (Fig. 7). This increase was not affected by heparin (Fig. 7A) or warfarin (Fig. 7B) treatment. To confirm the lack of effect of heparin on SEC injury, endothelial cells in sections of liver were stained immunohistochemically with the RECA-1 antibody. This antibody stains rat endothelial cells selectively (Duijvestijn et al., 1992). In addition, we have shown that this technique effectively measures the loss of endothelial cells in the liver after MCT treatment (Copple et al., 2002). In liver sections from rats treated with only saline (Fig. 8A) or with saline and heparin (Fig. 8B), RECA-1 staining was present along the sinusoids and lined the major vessels of the liver. In livers from rats treated with MCT, RECA-1 staining within the sinusoids was discontinuous or completely absent in centrilobular regions of the liver 14 h after MCT treatment (Fig. 8C). RECA-1 staining in livers from MCT-treated rats given heparin was not different from MCT-treated rats (Fig. 8D). Morphometric analysis revealed that MCT caused a significant decrease in the area of RECA-1 staining in the liver 14 hours after MCT treatment (Fig. 9). This decrease was not prevented by heparin.

In vitro, hepatic parenchymal cells are relatively resistant to MCT, which suggests that MCT may damage these cells in vivo by indirect mechanisms (DeLeve et al., 1996). It has been hypothesized that MCT-induced damage to SECs in vivo results in microcirculatory disturbances that lead to hypoperfusion of the liver and ischemic parenchymal cell injury (DeLeve et al., 1996). One factor that may contribute to this is an activated coagulation system. The coagulation system is activated, and fibrin deposits in

FIG. 7. Effect of heparin or warfarin on MCT-induced increase in plasma hyaluronic acid. (A) Rats were treated with 300 mg MCT/kg (ip) or saline vehicle and then with 3000 U heparin/kg (sc) or saline vehicle 6 h later. (B) Rats were treated with either 5 mg warfarin/kg or 10 mg warfarin/kg (ip) or DMSO vehicle. They then received a second injection of warfarin (5mg/kg or 10 mg/kg) or DMSO vehicle 24 h later followed by treatment with MCT (300 mg/kg) or saline vehicle 10 h after the second warfarin injection. Plasma hyaluronic acid was evaluated 14 h after MCT treatment as a marker of SEC injury. Data are expressed as means ⫾ SEM. n ⫽ 5 rats. *Significantly different ( p ⬍ 0.05) from saline-treated controls.

HEPARIN PREVENTS MONOCROTALINE-INDUCED LIVER INJURY

193

FIG. 8. MCT-induced changes in liver RECA-1 immunostaining in rats given heparin. Rats were treated with either 300 mg MCT/kg ip or saline vehicle and then with either 3000 U heparin/kg sc or saline vehicle 6 h later. Livers were removed 14 h after MCT treatment and processed for immunohistochemical staining of RECA-1, which appears black in the photomicrograph. Liver sections are from rats treated with (A) saline/saline, (B) saline/heparin, (C) MCT/saline, and (D) MCT/heparin. CV, centrilobular; PP, periportal. Bar represents 100 ␮m.

the liver after MCT treatment (Schoental and Head, 1955; Butler et al., 1970; Allen et al., 1969; Copple et al., 2002). Considering that activation of the coagulation system is required for hepatic parenchymal cell injury after exposure to some hepatotoxicants (Pearson et al., 1996; Arai et al., 1996; Fujiwara et al., 1988), the studies presented herein tested the hypothesis that activation of the coagulation system is required for MCT-induced hepatic parenchymal cell injury. Treatment of rats with heparin or warfarin significantly reduced MCTinduced activation of the coagulation system as measured by a decrease in plasma fibrinogen (Fig. 1). In addition, heparin reduced fibrin deposition (Fig. 3) in the liver after MCT treatment. Both anticoagulants significantly lessened MCT-induced parenchymal cell injury (Fig. 4), which suggests that activation of the coagulation system is required for MCT-induced paren-

chymal cell injury. Although heparin did not completely prevent activation of the coagulation system, it afforded approximately 80% protection from hepatic parenchymal cell injury (Fig. 4). This suggests that activation of the coagulation system is required for MCT-induced parenchymal cell injury and that complete inhibition of this activation is not required for injury prevention. This is similar to what has been observed in other models of liver injury that are dependent upon coagulation system activation. For example, activation of the coagulation system is required for bacterial lipopolysaccharide-induced liver injury (Hewett and Roth, 1995). In this model, the anticoagulant warfarin decreased hepatic parenchymal cell injury by approximately 90%, while only reducing activation of the coagulation system (i.e., decrease in plasma fibrinogen) by approximately 40% (Hewett and Roth, 1995). This suggests

194

COPPLE ET AL.

FIG. 9. Quantitative morphometry of RECA-1 staining in the liver. Rats were treated with either 300 mg MCT/kg ip or saline vehicle and then with either 3000 U heparin/kg sc or saline vehicle 6 h later. Livers were removed 14 h after MCT treatment and processed for immunohistochemical staining of RECA-1. The area of RECA-1 staining in 10 randomly chosen, low-power fields per tissue section, 2 sections per liver, and 5 animals per group was analyzed morphometrically as described under Materials and Methods. Data are expressed as means ⫾ SEM. *Significantly different ( p ⬍ 0.05) from saline-treated controls.

that, in certain models of liver injury, complete inhibition of the coagulation system is not required to reduce tissue fibrin deposition markedly or to prevent hepatic parenchymal cell injury. In addition, although both heparin and warfarin reduced MCT-induced parenchymal cell injury 14 h after MCT treatment, a more extensive time-course study will be required in order to determine whether anticoagulant treatment completely prevented or delayed MCT-induced parenchymal cell injury. As discussed above, it has been suggested that MCT-induced hepatic parenchymal cell necrosis results from ischemia due to poor perfusion in centrilobular regions (DeLeve et al., 1996). This is supported by histopathological findings of coagulative, hepatic parenchymal cell necrosis, which is consistent with ischemic injury. Moreover, parenchymal cell injury is preceded by endothelial cell damage and evidence of microcirculatory disturbances such as hemorrhage, which also occur in centrilobular regions (DeLeve et al., 1999; Copple et al., 2002). Taken together, these results suggest that activation of the coagulation system may contribute to hypoperfusion of the liver and subsequent ischemic injury. After treatment with MCT, approximately 23% of the liver had regions of coagulative necrosis, whereas less than 5% had regions of coagulative necrosis in rats cotreated with heparin. This suggests that occlusion of the microcirculation by insoluble fibrin clots and consequent ischemia might be responsible for parenchymal cell necrosis. However, there is currently no evidence that ischemia contributes to parenchymal cell injury after MCT treatment, and it is possible that the coagulation system promotes injury by other mechanisms. We have shown that the coagulation system promotes lipopolysaccharide-induced liver injury independent of insoluble fibrin clot formation by a mechanism involving activation of a cellular receptor for thrombin (Copple et al., 2000). Accordingly, although activation of the coagulation system is required for MCT-induced coagulative necrosis,

whether it promotes injury through fibrin deposition and consequent ischemia or by some other mechanism remains to be elucidated. Interestingly, heparin did not prevent the occurrence of single, necrotic parenchymal cells in the liver after MCT treatment. Although these cells are relatively resistant to MCT, large concentrations can damage these cells in vitro (DeLeve et al., 1996). It seems possible that the occasional single, necrotic cells that remained in the livers of rats treated with MCT and heparin resulted from direct toxicity of MCT. In any case, the pronounced effect of heparin on coagulative necrosis in the face of no effect on single-cell necrosis suggests the occurrence of more than one contributor to cell death. Heparin has effects on various processes unrelated to its anticoagulant effect. Therefore, it is possible that heparin lessened MCT-induced parenchymal cell injury independent of effects on the coagulation system. It has been shown to inhibit neutrophil (PMN) functions such as degranulation, adherence to endothelial cells, and chemotaxis (Bazzoni et al., 1993; Nelson et al., 1993; Matzner et al., 1984). Although PMNs accumulate in the liver after MCT treatment, prior depletion of PMNs from rats does not prevent MCT-induced liver injury (Copple et al., 2001), suggesting that heparin does not act by altering neutrophil function. Heparin can also increase the extracellular activity of superoxide dismutase (Oyanagui and Sato, 1990). It is unlikely that heparin affords protection through this mechanism, however, because antioxidants do not prevent MCT-induced liver injury (Wang et al., 2000). Evidence in support of a role for the coagulation system in MCT-induced liver comes from the observation that warfarin, which has a different chemical structure and acts on the coagulation system by a different mechanism than heparin, also afforded protection against MCT-induced liver injury (Fig. 4). This suggests that the protective effects of heparin were due to its actions on the coagulation system. Inhibition of coagulation system activation did not prevent

FIG. 10. Effect of heparin on MCT-induced increase in hemoglobin in the liver. Rats were treated with either 300 mg MCT/kg ip or saline vehicle and then with either 3000 U heparin/kg sc or saline vehicle 6 h later. Livers were removed 14 h after MCT treatment and analyzed for hemoglobin. Data are expressed as means ⫾ SEM. n ⫽ 5. *Significantly different ( p ⬍ 0.05) from saline-treated controls.

HEPARIN PREVENTS MONOCROTALINE-INDUCED LIVER INJURY

MCT-induced damage to endothelial cells in the liver. Histopathologic examination of liver sections from rats treated with MCT and saline showed evidence for damage to CVECs that frequently progressed to complete loss of the vascular intima. In addition, there was a significant increase in the plasma concentration of hyaluronic acid in rats treated with MCT and saline, which is indicative of SEC damage (Fig. 7), and a 70% reduction in sinusoidal RECA-1 staining, suggesting that these cells were lost from the liver (Fig. 9). Interestingly, anticoagulants had no effect on these measures of endothelial cell toxicity in the liver after MCT treatment (Figs. 7 and 9). Therefore, activation of the coagulation system is not required for MCT-induced damage to endothelial cells in the liver. Surprisingly, heparin had no effect on MCT-induced accumulation of red blood cells in the liver tissue (Fig. 10). This suggests that red blood cell accumulation results from hemorrhage after endothelial cell damage and not from congestion due to hepatic fibrin deposition or from parenchymal cell injury. Taken together, our findings suggest a scenario in which MCT produces damage to endothelial cells in the liver that leads to disruption of vascular architecture and hemorrhage. This is accompanied by activation of the coagulation system, which promotes hepatic parenchymal cell injury through deposition of insoluble fibrin clots in the vasculature and ischemia and/or through other mechanisms, such as activation of a cellular receptor for thrombin. Whether the coagulation system contributes to HVOD in humans remains a matter of debate. There is evidence for activation of the coagulation system in bone marrow transplant patients who developed HVOD, although this has not been a finding in all clinical studies (Korte, 1997). Several studies do, however, suggest a role for the coagulation system. Prophylactic treatment of bone marrow transplant patients with heparin decreased the incidence of HVOD (Or et al., 1996; Attal et al., 1992). In addition, preliminary studies have shown that stimulation of fibrinolysis with recombinant tissue plasminogen activator was effective in treating bone marrow transplant patients with established HVOD (Terra et al., 1997; Yu et al., 1994). These results suggest that the coagulation system may be a critical mediator in the development of HVOD after bone marrow transplantation. In summary, treatment of rats with anticoagulants prevented MCT-induced activation of the coagulation system and fibrin deposition in the liver. Anticoagulants also decreased hepatic parenchymal cell injury but had no effect on endothelial cell injury. These results raise the possibility that MCT damages hepatic endothelial cells, thereby causing activation of the coagulation system. The coagulation system then promotes parenchymal cell injury though fibrin deposition and consequent ischemia or through other mechanisms.

195

ACKNOWLEDGMENTS The authors thank Catherine Book and Catherine Rondelli for technical assistance. This study was supported by NIH Grants ES04139 and DK50728. B.L.C. was supported by NRSA ES05866 from NIH.

REFERENCES Ahmed, A. E., Grissom, M., el-Azhary, R., Haque, A., Boor, P. J., and Costanzi, J. (1987). Studies on the mechanism of 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea (CCNU)-induced hepatotoxicity. II. Biochemical and morphological characterization of the injury and its prevention by phenobarbital. Cancer Chemother. Pharmacol. 19, 103–108. Allen, J. R., Carstens, L. A., and Katagiri, G. J. (1969). Hepatic veins of monkeys with veno-occlusive disease: Sequential ultrastructural changes. Arch. Pathol. 87, 279 –289. Allen, J. R., Chesney, C. F., and Frazee, W. J. (1972). Modifications of pyrrolizidine alkaloid intoxication resulting from altered hepatic microsomal enzymes. Toxicol. Appl. Pharmacol. 23, 470 – 479. Arai, M., Mochida, S., Ohno, A., and Fujiwara, K. (1996). Blood coagulation in the hepatic sinusoids as a contributing factor in liver injury following orthotopic liver transplantation in the rat. Transplantation 62, 1398 –1401. Attal, M., Huguet, F., Rubie, H., Huynh, A., Charlet, J. P., Payen, J. L., Voigt, J. J., Brousset, P., Selves, J., and Muller, C. (1992). Prevention of hepatic veno-occlusive disease after bone marrow transplantation by continuous infusion of low-dose heparin: A prospective, randomized trial. Blood 79, 2834 –2840. Bazzoni, G., Beltran, N. A., Mascellani, G., Bianchini, P., DeJana, E., and del Maschio, A. (1993). Effect of heparin, dermatan sulfate, and related oligoderivatives on human polymorphonuclear leukocyte functions. J. Lab. Clin. Med. 121, 268 –275. Bearman, S. I. (2000). Veno-occlusive disease of the liver. Curr. Opin. Oncol. 12, 103–109. Brooks, S. E., Miller, C. G., McKenzie, K., Audretsch, J. J., and Bras, G. (1970). Acute veno-occlusive disease of the liver: Fine structure in Jamaican children. Arch. Pathol. 89, 507–520. Butler, W. H., Mattocks, A. R., and Barnes, J. M. (1970). Lesions in the liver and lungs of rats given pyrrole derivatives of pyrrolizidine alkaloids. J. Pathol. 100, 169 –175. Carreras, E. (2000). Veno-occlusive disease of the liver after hemopoietic cell transplantation. Eur. J. Haematol. 64, 281–291. Copple, B. L., Banes, A, Ganey, P. E., and Roth, R. A. (2002). Endothelial cell injury and fibrin deposition in rat liver after monocrotaline exposure. Toxicol. Sci. 65, 309 –318. Copple, B. L., Moulin, F., Ganey, P. E., and Roth, R. A. (2000). Endotoxininduced liver injury requires activation of protease-activated receptor-1. FASEB J. 14, A290. Copple, B. L., Perez Rivera, A. A., Ganey, P. E., and Roth, R. A. (2001). Neutrophils and monocrotaline-induced liver injury. Toxicologist 60, 351. Deaciuc, I. V., Bagby, G. J., Neisman, M. R., Skrepnik, N., and Spitzer, J. J. (1994). Modulation of hepatic sinusiodal endothelial cell function by Kupffer cells: An example of intercellular communication in the liver. Hepatology 19, 464 – 470. Deaciuc, I. V., McDonough, K. H., Bagby, G. J., and Spitzer, J. J. (1993). Alcohol consumption in rats potentiates the deleterious effect of gramnegative sepsis on hepatic hyaluronan uptake. Alcohol Clin. Exp. Res. 17, 1002–1008. DeLeve, L. D. (1994). Dacarbazine toxicity in murine liver cells: A model of hepatic endothelial injury and glutathione defense. J. Pharmacol. Exp. Ther. 268, 1261–1270. DeLeve, L. D. (1996). Cellular target of cyclophosphamide toxicity in the

196

COPPLE ET AL.

murine liver: Role of glutathione and site of metabolic activation. Hepatology 24, 830 – 837. DeLeve, L. D. (2001). Hepatic venoocclusive disease: A major complication of hematopoietic stem cell transplantation in cancer patients. Tumori 87, S27– S29. DeLeve, L. D., McCuskey, R. S., Wang, X., Hu, L., McCuskey, M. K., Epstein, R. B., and Kanel, G. C. (1999). Characterization of a reproducible rat model of hepatic veno-occlusive disease. Hepatology 29, 1779 –1791. DeLeve, L. D., Wang, X., Kuhlenkamp, J. F., and Kaplowitz, N. (1996). Toxicity of azathioprine and monocrotaline in murine sinusoidal endothelial cells and hepatocytes: The role of glutathione and relevance to hepatic venoocclusive disease. Hepatology 23, 589 –599. Duijvestijn, A. M., van Goor, H., Klatter, F., Majoor, G. D., van Bussel, E., and van, B. V. (1992). Antibodies defining rat endothelial cells: RECA-1, a panendothelial cell-specific monoclonal antibody. Lab. Invest. 66, 459 – 466. Fujiwara, K., Ogata, I., Ohta, Y., Hirata, K., Oka, Y., Yamada, S., Sato, Y., Masaki, N., and Oka, H. (1988). Intravascular coagulation in acute liver failure in rats and its treatment with antithrombin III. Gut 29, 1103–1108. Hewett, J. A., and Roth, R. A. (1995). The coagulation system, but not circulating fibrinogen, contributes to liver injury in rats exposed to lipopolysaccharide from gram-negative bacteria. J. Pharmacol. Exp. Ther. 272, 53– 62. Huxtable, R. J. (1989). Human health implications of pyrrolizidine alkaloids and herbs containing them. In Toxicants of Plant Origin (P. R. Cheeke, Ed.), pp. 41– 86. CRC Press, Boca Raton, FL. Jaeschke, H., Farhood, A., Cai, S. X., Tseng, B. Y., and Bajt, M. L. (2000). Protection against TNF-induced liver parenchymal cell apoptosis during endotoxemia by a novel caspase inhibitor in mice. Toxicol. Appl. Pharmacol. 169, 77– 83. Kobayashi, H., Horikoshi, K., Yamataka, A., Yamataka, T., Okazaki, T., Lane, G. J., and Miyano, T. (1999). Hyaluronic acid: A specific prognostic indicator of hepatic damage in biliary atresia. J. Pediatr. Surg. 34, 1791– 1794. Korte, W. (1997). Veno-occlusive disease of the liver after bone marrow transplantation: Is hypercoagulability really part of the problem? Blood Coagul. Fibrinolysis 8, 367–381. Levin, S., Bucci, T. J., Cohen, S. M., Fix, A. S., Hardisty, J. F., LeGrand, E. K., Maronpot, R. R., and Trump, B. F. (1999). The nomenclature of cell death: Recommendations of an ad hoc Committee of the Society of Toxicologic Pathologists. Toxicol. Pathol. 27, 484 – 490. Mattocks, A. R., and White, I. N. H. (1970). Estimation of metabolites of pyrrolizidine alkaloids in animal tissues. Anal. Biochem. 38, 529 –535.

Matzner, Y., Marx, G., Drexler, R., and Eldor, A. (1984). The inhibitory effect of heparin and related glycosaminoglycans on neutrophil chemotaxis. Thromb. Haemostasis 52, 134 –137. Nelson, R. M., Cecconi, O., Roberts, W. G., Aruffo, A., Linhardt, R. J., and Bevilacqua, M. P. (1993). Heparin oligosaccharides bind L- and P-selectin and inhibit acute inflammation. Blood 82, 3253–3258. Or, R., Nagler, A., Shpilberg, O., Elad, S., Naparstek, E., Kapelushnik, J., Cass, Y., Gillis, S., Chetrit, A., Slavin, S., and Eldor, A. (1996). Low molecular weight heparin for the prevention of veno-occlusive disease of the liver in bone marrow transplantation patients. Transplantation 61, 1067– 1071. Oyanagui, Y., and Sato, S. (1990). Heparin, a potent releasing agent of extracellular superoxide dismutase (EC-SOD C), suppresses ischaemic paw oedema in mice. Free Radical Res. Commun. 9, 87–99. Pearson, J. M., Schultze, A. E., Schwartz, K. A., Scott, M. A., Davis, J. M., and Roth, R. A. (1996). The thrombin inhibitor, hirudin, attenuates lipopolysaccharide-induced liver injury in the rat. J. Pharmacol. Exp. Ther. 278, 378 –383. Schnitt, S. J., Stillman, I. E., Owings, D. V., Kishimoto, C., Dvorak, H. F., and Abelmann, W. H. (1993). Myocardial fibrin deposition in experimental viral myocarditis that progresses to dilated cardiomyopathy. Circ. Res. 72, 914 – 920. Schoental, R., and Head, M. (1955). Pathological changes in rats as a result of treatment with monocrotaline. Br. J. Cancer 9, 229 –237. Schultze, A. E., and Roth, R. A. (1998). Chronic pulmonary hypertension: The monocrotaline model and involvement of the hemostatic system. J. Toxicol. Environ. Health 1, 271–346. Terra, S. G., Spitzer, T. R., and Tsunoda, S. M. (1997). A review of tissue plasminogen activator in the treatment of veno-occlusive liver disease after bone marrow transplantation. Pharmacotherapy 17, 929 –937. Wang, X., Kanel, G. C., and DeLeve, L. D. (2000). Support of sinusoidal endothelial cell glutathione prevents hepatic veno-occlusive disease in the rat. Hepatology 31, 428 – 434. Yee, S. B., Kinser, S., Hill, D. A., Barton, C. C., Hotchkiss, J. A., Harkema, J. R., Ganey, P. E., and Roth, R. A. (2000). Synergistic hepatotoxicity from coexposure to bacterial endotoxin and the pyrrolizidine alkaloid monocrotaline. Toxicol. Appl. Pharmacol. 166, 173–185. Yu, L. C., Malkani, I., Regueira, O., Ode, D. L., and Warrier, R. P. (1994). Recombinant tissue plasminogen activator (rt-PA) for veno-occlusive liver disease in pediatric autologous bone marrow transplant patients. Am. J. Hematol. 46, 194 –198.