Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

A great deal of evidence suggests that reactive oxygen species (ROS) are implicated in the pathogenesis of a number of diseases, including atherosclerosis, cancer, and Alzheimer's disease (Halliwell and Gutteridge, 1990; Cerutti, 1991; Simonian and Coyle, 1996). Their toxic effects, which are amplified by pathological events including neutrophil activation, hyperoxia, metabolism of redox-active drugs, radiation exposure, and ischemia/reperfusion, include membrane damage resulting from lipid peroxidation, as well as attack by ROS on proteins and nucleic acids (Halliwell and Gutteridge, 1990). Therefore, antioxidant enzymes have been considered as therapeutic agents for ROS-mediated injuries and diseases. The use of superoxide dismutase (SOD) and catalase (CAT), representative antioxidant enzymes catabolizing superoxide and hydrogen peroxide, respectively, have been examined for various ROS-mediated injuries, especially those associated with ischemia/reperfusion (Atalla et al., 1985; Chiu and Toledo-Pereyra, 1987; Flye and Yu, 1987; Lambotte et al., 1988; Romani et al., 1988; Castillo et al., 1990; Cho et al., 1990; Nauta et al., 1990). However, both enzymes are known to be rapidly eliminated from the circulation after i.v. bolus injection (Pyatak et al., 1980; Turrens et al., 1984), which limits their therapeutic potential. To increase the plasma half-life of these enzymes, conjugation with macromolecules such as polyethylene glycol derivatives or poly(styrene comaleic acid butyl ester) has been carried out (Abuchowski et al., 1977; Pyatak et al., 1980; Beckman et al., 1988; Kawamoto et al., 1990). Through these modifications, the concentrations of these enzymes in plasma would be maintained for a longer period than those of unmodified enzymes and result in more efficient exposure to the target sites.

It has been generally accepted that ROS contribute to ischemia/reperfusion injury of the liver (McCord, 1985; Granger, 1988). The local hepatic injury is considered to involve two phases, with the initial injury being mediated by activated Kupffer cells (Jaeschke and Farhood, 1991; Jaeschke et al., 1991). Neutrophils are primed during this initial period and play a central role in the later phase of the hepatic injury (Jaeschke et al., 1990). In addition to the approaches to prolong the retention of antioxidant enzymes in plasma, targeted delivery of such enzymes to the sites where ROS are generated should also be a promising way of preventing ROS-mediated injuries. Based on this consideration, we developed by chemical modification SOD derivatives possessing high affinity for liver cells (Fujita et al., 1992a). Among them, mannosylated SOD (Man-SOD) and galactosylated SOD (Gal-SOD) showed improved effects on preventing ischemia/reperfusion injuries of the liver (Fujita et al., 1992b; Kondo et al., 1996; Mizoe et al., 1997), supporting the usefulness of the targeted delivery of antioxidant enzymes. These results led us to apply the same strategy to CAT.

As a first step toward this approach, we recently studied the pharmacokinetics and therapeutic effect of native bovine liver CAT in mice (Yabe et al., 1999). We found that native CAT itself rapidly accumulated selectively in the hepatocytes (liver parenchymal cells; PC) after i.v. injection. This rapid hepatic uptake was saturable and was not inhibited by a number of compounds, suggesting that CAT is taken up by the cells via a CAT-specific mechanism, which is unknown at present. We also observed that CAT could prevent an experimental hepatic ischemia/reperfusion injury in mice in a dose-dependent manner. However, there was no significant distribution of CAT to liver nonparenchymal cells (NPC) such as Kupffer cells and endothelial cells. As mentioned above, Kupffer cells produce ROS in the initial phase of hepatic ischemia/reperfusion. In addition, the liver endothelial cells are much more susceptible to ROS than hepatocytes, probably due to the lack of hydrogen peroxide detoxifying enzymes, including CAT (Hamer et al., 1995), which leads to vascular destruction and access of neutrophils to the hepatocytes. Therefore, targeting CAT to these liver NPC could be more beneficial for preventing the injury.

The purpose of this study was to control the biodistribution at a cellular level and subsequent biological activity of CAT by chemical modification of the enzyme. For targeting to the liver NPC, mannosylation was used because mannosylated macromolecules are selectively delivered to liver NPC after i.v. injection via mannose receptor-mediated uptake (Fujita et al., 1992a; Nishikawa et al., 1992). Because these NPC also possess scavenger receptors that recognize polyanions, including succinylated or maleylated proteins (Takakura et al., 1994; Furitsu et al., 1997), succinylation was also carried out. To deliver CAT to the liver PC via the asialoglycoprotein receptor, CAT was also galactosylated. For comparison, CAT was coupled with a polyethylene glycol derivative. The in vivo distribution characteristics of these CAT derivatives and their potential for preventing hepatic ischemia/reperfusion injury was investigated in mice. Moreover, we assessed the feasibility of the CAT derivatives in combination with SOD derivatives for further improving the treatment of ROS-mediated hepatic injuries.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. Male ddY mice (5 weeks old, 25-28 g and 10 weeks old, 40-50 g) were purchased from the Shizuoka Agricultural Cooperative Association for Laboratory Animals, Shizuoka, Japan. Animals were maintained under conventional housing conditions. The 10-week-old mice were used only in the hepatic ischemia/reperfusion experiment.

Chemicals. Bovine liver CAT (40,000 U/mg) and collagenase (type IA) was purchased from Sigma Chemical., St. Louis, MO. Recombinant human SOD (111-Ser) was supplied by Asahi Kasei, Tokyo, Japan. Methoxypolyethylene glycolyl N-succinimidyl succinate (activated PEG; average molecular weight 5200) was supplied by Nippon-Yushi, Tokyo, Japan. Diethylenetriaminepentaacetic acid (DTPA) anhydride was purchased from Dojindo Laboratory, Kumamoto, Japan. ¹¹¹Indium chloride ([¹¹¹In]InCl₃) was supplied by Nihon Medi-Physics, Takarazuka, Japan. Man-SOD (2750 U/mg) and Gal-SOD (2530 U/mg) synthesized as previously reported (Fujita et al., 1992a) were used. Succinylated BSA (Suc-BSA) and mannosylated BSA (Man-BSA) were also prepared as reported (Nishikawa et al., 1992; Takakura et al., 1994). All other chemicals were of the finest grade available.

Synthesis and Characterization of CAT Derivatives. Succinylated CAT (Suc-CAT) was synthesized as reported previously (Takakura et al., 1994). In brief, 100 mg CAT was dissolved in 10 ml 0.2 M Tris buffer (pH 8.65) and 46 mg succinic anhydride was added. The mixture was stirred for 18 h at room temperature. CAT-polyethylene glycol conjugate (PEG-CAT) was synthesized by reacting activated PEG with CAT in 50 mM borate buffer (pH 9.2) for 24 h at 4°C in the dark (Abuchowski et al., 1977). The reaction mixture was applied to a column (Toyopearl HW-55S, Tosoh Co., Tokyo, Japan) and fractions with larger molecular weights were collected. Galactosylated (Gal-CAT) and mannosylated CAT (Man-CAT) were synthesized by reacting CAT with 2-imino-2-methoxyethyl 1-thioglycoside as previously described (Fujita et al., 1992a). Each derivative was washed and concentrated by ultrafiltration (molecular weight cut-off, 200,000) against distilled water, and lyophilized.

Radiolabeling of CAT. For the disposition experiments, each derivative was radiolabeled with ¹¹¹In using the bifunctional chelating agent, DTPA anhydride, according to the method of Hnatowich et al. (1982). This radiolabeling method is suitable for examining the distribution phase of macromolecules from plasma to various tissues, because radioactive metabolites, if produced after cellular uptake, are retained within cells where the uptake takes place (Duncan and Welch, 1993; Arano et al., 1994). In brief, a CAT derivative (2 mg) was dissolved in 1 ml HEPES buffer (0.1 M, pH 7.0) and a 2-fold molar excess of DTPA anhydride in 10 µl dimethyl sulfoxide was added. After stirring 30 min at room temperature, the mixture was purified by gel filtration chromatography using a Sephadex G-25 column (1 × 40 cm) and eluted with acetate buffer (0.1 M, pH 6.0) to separate the unreacted DTPA. Fractions containing DTPA-coupled CAT derivative were selected using spectrophotometry and concentrated by ultrafiltration. To 30 µl of sodium acetate buffer, an equivalent volume of ¹¹¹InCl₃ solution was added, then 60 µl of DTPA-coupled CAT derivative (about 1 mg/ml) solution was added to the mixture. After 30 min at room temperature, the mixture was purified by gel filtration chromatography using a PD-10 column (Pharmacia, Uppsala, Sweden) and eluted with acetate buffer (0.1 M, pH 6.0). The appropriate fractions were selected based on their radioactivity and concentrated by ultrafiltration.

In Vivo Disposition Experiment. ¹¹¹In-CAT derivative was dissolved in saline, and the protein concentration was adjusted by the addition of nonradiolabeled CAT derivative to the solution. The solution of ¹¹¹In-CAT derivative was injected into the tail vein of 5-week-old mice at a dose of 0.1 mg/kg. At appropriate intervals after injection, blood was collected from the vena cava under ether anesthesia and mice were sacrificed. Then, plasma was obtained by centrifugation of the blood collected. The kidney, liver, spleen, heart, and lung were removed, rinsed with saline, and weighed. Urine in the bladder and excreted were also collected. The radioactivity in each sample was counted in a well-type NaI scintillation counter (ARC-500, Aloka, Tokyo).

Hepatic Cellular Localization of ¹¹¹In-CAT Derivatives. Groups of three mice each were injected with ¹¹¹In-CAT derivative at a dose of 0.1 mg/kg and anesthetized by peritoneal administration of pentobarbital sodium. At 10 min after injection, the liver was perfused from the portal vein first with preperfusion buffer (Ca²⁺-, Mg²⁺-free HEPES buffer, pH 7.2) for 10 min and then with HEPES buffer (pH 7.5) containing 5 mM CaCl₂ and 0.05% (w/v) collagenase for about 20 min. Then, the liver was excised and the cells were dispersed by gentle stirring in ice-cold Hanks-HEPES buffer containing 0.1% BSA. The dispersed cells were filtered through cotton mesh sieves and centrifuged for 1 min at 50g to sediment the liver PC. The supernatant was removed and kept as the source of nonparenchymal cells (NPC). The pellet of PC was resuspended in buffer and centrifuged again to remove other cells. The NPC suspension was centrifuged twice at 50g for 1 min to remove the PC, then subjected to centrifugation at 200g for 2 min to obtain the pellet of NPC. Both PC and NPC fractions were resuspended in buffer and the number of cells was determined by the trypan blue exclusion method. The radioactivity of both samples was also determined as in the in vivo disposition experiment.

Calculation of Area Under the Plasma Concentration-Time Curve (AUC) and Clearances. The ¹¹¹In-radioactivity concentration in plasma was normalized to percentage of dose per milliliter and analyzed using the nonlinear least-squares program MULTI (Yamaoka et al., 1981), and the AUC and total body clearance (CL_total) were calculated. The tissue distribution pattern of ¹¹¹In-CAT derivatives was evaluated by the organ uptake clearance (CL_org) as previously reported (Takakura et al., 1987). In brief, with the assumption of negligible efflux of radioactivity from tissues, CL_org was calculated by dividing the amount of radioactivity in an organ at an appropriate interval of time by the AUC up to the same time point.

Hepatic Ischemia/Reperfusion Experiment. Male ddY mice were anesthetized with a peritoneal injection of pentobarbital sodium (50 mg/kg). An incision was made in the abdomen, and the portal vein and the hepatic artery were occluded with a vascular clamp for 30 min to induce hepatic ischemia. Then, blood was allowed to reflow through the liver (reperfusion). Saline, BSA derivatives (control), CAT derivatives (10,000 or 20,000 U/kg), SOD derivatives (10,000 U/kg), or both CAT and SOD derivatives (10,000 U/kg for each) were administered through the tail vein 5 min before blood reflow. After 60 min of reperfusion, blood was collected from the vena cava and plasma was obtained as above. The activities of glutamic pyruvic transaminase (GPT) and glutamic oxaloacetic transaminase (GOT), as indicators of hepatocyte injury during reperfusion, were assayed using commercial test reagents.

Statistical Analysis. Differences were statistically evaluated by one-way ANOVA followed by the Student-Newmann-Keuls multiple comparison test, at a significance level of P < .05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Physicochemical Characteristics of CAT Derivatives. Table 1 summarizes the physicochemical characteristics of the synthesized CAT derivatives. About 70% of the total amino groups of CAT (112 residues) was used for chemical modification for Suc-CAT, Gal-CAT, and Man-CAT. On the other hand, only about 26 residues of the amino groups were modified in PEG-CAT. However, PEG conjugation resulted in an increase in the apparent molecular weight to 840,000. The apparent molecular weight of Suc-CAT was slightly larger than that of CAT. Glycosylation hardly affected the molecular weight of both derivatives. Each chemical modification slightly reduced the enzymatic activity of CAT, but more than 86% of the activity was retained in all the CAT derivatives.

In Vivo Disposition of ¹¹¹In-CAT Derivatives. Fig. 1 shows the plasma concentration of ¹¹¹In-CAT derivatives following i.v. injection into mice at a dose of 0.1 mg/kg. ¹¹¹In-Suc-CAT, ¹¹¹In-Gal-CAT, and ¹¹¹In-Man-CAT rapidly disappeared from plasma. These disappearance rates were faster than that of ¹¹¹In-CAT. On the other hand, ¹¹¹In-PEG-CAT showed a prolonged retention in plasma.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Plasma concentration of 111In-CAT derivatives after i.v. injection in mice at doses of 0.1 mg/kg. , 111In-CAT; , 111In-Suc-CAT; , 111In-PEG-CAT; , 111In-Gal-CAT; , 111In-Man-CAT. Results are expressed as mean ± S.D. of three mice.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Tissue distribution of 111In-CAT derivatives after i.v. injection in mice at a dose of 0.1 mg/kg. A, 111In-CAT; B, 111In-Suc-CAT; C, 111In-Gal-CAT; D, 111In-Man-CAT; E, 111In-PEG-CAT. Results are expressed as mean ± S.D. of three mice. , Kidney; , liver; , spleen; , heart; , lung; , urine.

Cellular Distribution of ¹¹¹In-CAT Derivatives in the Liver. Fig. 3 shows the cellular distribution profiles of ¹¹¹In-CAT derivatives in the liver after i.v. injection into mice. ¹¹¹In-CAT was largely taken up by the PC of the liver via a CAT-specific and concentration-dependent process (Yabe et al., 1999). ¹¹¹In-Gal-CAT was also recovered mainly in the PC fraction after collagenase perfusion, and the amount recovered in PC accounted for more than 98% of the hepatic uptake, which was calculated using the number of liver cells reported (PC, 1.25 × 10⁸ cells/g liver; NPC, 0.65 × 10⁸ cells/g liver) (Blomhoff et al., 1985). On the other hand, a significant amount was recovered in the NPC fraction in the case of ¹¹¹In-Suc-CAT or ¹¹¹In-Man-CAT. The NPC uptake was estimated to account for about 51% and 43% of the hepatic uptake for ¹¹¹In-Suc-CAT and ¹¹¹In-Man-CAT, respectively.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Cellular localization of 111In-CAT derivatives in the liver after i.v. injection in mice at a dose of 0.1 mg/kg. (open column), PC; (closed column), NPC. Results are expressed as mean ± S.D. of at least three experiments.

Pharmacokinetic Analysis of ¹¹¹In-CAT Derivatives. The AUC, CL_total, hepatic uptake clearance (CL_liver), renal uptake clearance (CL_kidney), and urinary excretion clearance (CL_urine) of ¹¹¹In-CAT derivatives are summarized in Table 2. The CL_liver of any ¹¹¹In-CAT derivative accounted for a large portion of its CL_total. The CL_liver of ¹¹¹In-Suc-CAT, ¹¹¹In-Gal-CAT and ¹¹¹In-Man-CAT was larger than that of ¹¹¹In-CAT, and was close to the plasma flow rate to the liver (85,000 µl/h).

Prevention of Hepatic Ischemia/Reperfusion Injury by CAT Derivatives. The effect of the ischemia/reperfusion on plasma GOT and GPT activities was examined first. Within the 30-min ischemia, plasma GOT and GPT levels did not significantly increase. After the reperfusion of the hepatic flows, however, both activities significantly augmented with time up to 1 h (data not shown), indicating that hepatic injuries are induced by this treatment. At 60 min after reperfusion, plasma GOT and GPT levels significantly increased from 32 to 190 and from 8.7 to 110 U/liter, respectively. Figure 4 shows the GPT and GOT levels in plasma of hepatic ischemia/reperfusion-injured mice. At a dose of 10,000 U/kg, CAT and Gal-CAT, which accumulated in the liver parenchymal cells, showed a moderate ability to suppress these increases in plasma enzymes; significant differences were detected in GOT activity for CAT-treated mice (P < .05) and in GPT activity for CAT- (P < .01) and Gal-CAT-treated mice (p < .001), compared with corresponding activities for saline-treated mice. PEG-CAT was not effective in the present experiment. On the other hand, Suc-CAT and Man-CAT significantly suppressed the increase in both GOT (P < .001 for Suc-CAT, P < .01 for Man-CAT, compared with saline) and GPT (P < .001), and their efficacies were more potent than the other CAT derivatives examined. Neither Suc-BSA nor Man-BSA showed any significant inhibitory effects on the increases in plasma GOT and GPT activities (data not shown).

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Effect of CAT and SOD derivatives on plasma GOT (A) and GPT (B) levels in mice following hepatic ischemia/reperfusion injury at a dose of (hatched column) 10,000 U/kg or (closed column) 20,000 U/kg. Results are expressed as mean ± S.E.M. for at least three mice. *, **, ***, Significantly different from that of saline-treated mice (*p < .05; **p < .01; ***p < .001). , , , Significantly different from that of control mice (p < .05; p < .01; p < .001).

View larger version (136K): [in this window] [in a new window]   Fig. 5.   Histological observation of frozen sections of livers at 60 min after reperfusion. a, liver structure was disrupted and severe damages of hepatocytes were seen in livers of mice treated with saline. b, structures of hepatocytes tended to remain unchanged and integrity of sinusoids was almost maintained in livers of mice treated with Suc-CAT. Original magnification, 200×

In Vivo Disposition of ¹¹¹In-Labeled CAT and SOD Derivatives after Coadministration with Other Derivatives. Figure 6 illustrates the effect of coadministration of cold Man-SOD, Man-CAT, and Suc-CAT on the plasma concentration and liver accumulation of ¹¹¹In-labeled Man-CAT, Suc-CAT, and Man-SOD at a therapeutic dose of 10,000U/kg. No significant change was observed in any of these, although a slightly reduced hepatic uptake was observed only when ¹¹¹In-Man-CAT was injected with Man-SOD (P < .05 at time points after 3 min), suggesting that effective hepatic delivery of these enzyme derivatives could be expected in combination therapy.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Effect of coadministration of CAT and SOD derivatives (10,000 U/kg) on plasma concentration and liver accumulation of 111In-labeled CAT and SOD derivatives (10,000 U/kg). Results are expressed as mean ± S.D. of at least three experiments. Dose (10,000 U/kg) corresponds to 0.25 mg/kg and 4 mg/kg for CAT and SOD derivatives, respectively. A, , 111In-Suc-CAT alone; , 111In-Suc-CAT + Man-SOD; B, , 111In-Man-CAT alone; , 111In-Man-CAT + Man-SOD; C, , 111In-Man-SOD alone; , 111In-Man-SOD + Man-CAT; , 111In-Man-SOD + Suc-CAT.

Prevention of Hepatic Ischemia/Reperfusion Injury by CAT Derivatives in Combination with SOD Derivatives. Fig. 7 shows the plasma GPT and GOT of the hepatic injured mice treated with both CAT and SOD derivatives. Coadministration of Gal-SOD with Suc-CAT, Man-CAT, or Gal-CAT failed to increase the inhibitory effect of each CAT derivative on the injury. In contrast, simultaneous injection of Man-SOD with each CAT derivative examined showed excellent effects in suppressing the hepatic injury than its administration alone. Among them, the combination of Suc-CAT and Man-SOD showed the greatest inhibitory effect; the GPT and GOT in plasma were 53 and 23 U/liter, respectively.

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Effect of coadministration of SOD (10,000 U/kg) and CAT derivatives (10,000 U/kg) on plasma GOT (A) and GPT (B) levels in mice following hepatic ischemia/reperfusion injury. Results are expressed as mean ± S.E.M. for at least three mice. *, **, ***, significantly different from that of saline-treated mice (*p < .05; **p < .01; ***p < .001). , , , significantly different from that of control mice (p < .05; p < .01; p < .001).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Various endogenous macromolecular substances, especially enzymes and cytokines, now represent a new class of therapeutic agents because of recent progress in biotechnology. CAT is an enzyme that detoxifies hydrogen peroxide, a compound involved in various ROS-mediated injuries. Therefore, treatment with CAT alone or its combination with SOD has been used for hepatic ischemia/reperfusion injuries, but any pharmacological effects reported are so far controversial (Atalla et al., 1985; Flye and Yu, 1987; Chiu and Toledo-Pereyra, 1987; Nauta et al., 1990; Castillo et al., 1990). Experimental designs including the duration of ischemia and reperfusion, dosing schedules of the enzymes, and animals used could make a significant difference. The biodistribution of CAT and/or SOD is also one important factor determining its pharmacological activity, because it can detoxify only at the sites to which it distributes. However, this issue has not been extensively taken into considerations in previous studies.

Our recent work has clarified the relationship between the pharmacokinetic characteristics of native CAT and its therapeutic effect on a hepatic ischemia/reperfusion injury model in mice (Yabe et al., 1999). A dose-dependent efficacy of the enzyme against the injury seems to be ascribed to its rapid and preferential uptake by the liver PC (hepatocytes), the target cell of ROS generated in this injury, after i.v. injection. However, there was no significant uptake by the NPC, involving Kupffer cells and endothelial cells, other important cell populations in the liver as a source of ROS. These results led us to develop CAT derivatives that can alter the intrahepatic distribution of the native enzyme, particularly the CAT derivatives targetable to these cells.

Receptor-mediated endocytosis is an attractive mechanism for cell-specific targeting of biologically active agents to restricted cells that express representative receptors. Chemical modification was used to render CAT capable of being selectively recognized by receptors on liver cells. For hepatocyte-specific delivery via asialoglycoprotein receptors, CAT was modified with galactose moieties. This modification of CAT hardly altered its apparent intrahepatic distribution pattern (Fig. 3), because CAT itself is also rapidly taken up by the liver PC via an unknown mechanism specific to CAT (Yabe et al., 1999). To understand the uptake mechanism of Gal-CAT by the liver PC, ¹¹¹In-Gal-CAT (0.1 mg/kg) was injected along with an excess of cold Gal-BSA (20 mg/kg) or CAT (20 mg/kg). Neither Gal-BSA nor CAT alone significantly altered the hepatic uptake of ¹¹¹In-Gal-CAT, when the uptake was examined at 5 min postinjection (data not shown). However, significant inhibition of the hepatic uptake of ¹¹¹In-Gal-CAT was observed when both cold Gal-BSA and CAT were injected simultaneously. These results indicate that Gal-CAT can be recognized by both of the two mechanisms, the CAT-specific mechanism and the asialoglycoprotein receptor.

On the other hand, significant CAT delivery to the liver NPC was achieved by mannosylation or succinylation (Fig. 3). However, the specificity to NPC of ¹¹¹In-Man-CAT (43%) or ¹¹¹In-Suc-CAT (51%) was not so prominent compared with that of other proteins, such as Man-BSA (70%) (Nishikawa et al., 1992), Man-SOD (72%) (Fujita et al., 1992a), and Suc-BSA (76%) (Takakura et al., 1994). These results suggest that ¹¹¹In-Man-CAT and ¹¹¹In-Suc-CAT are recognized by not only the mannose and scavenger receptors on the NPC but also by the mechanism specific for CAT on hepatocytes, as shown in ¹¹¹In-Gal-CAT.

In the initial phase of hepatic ischemia/reperfusion injury within the first 1 to 2 h of reperfusion, ROS are mainly released from Kupffer cells (Jaeschke, 1991; Jaeschke et al., 1991), and these ROS recruit and activate neutrophils (Jaeschke et al., 1991; Hamer et al., 1995; Johnston et al., 1996). Once activated and attached to endothelial cells, neutrophils may exacerbate tissue injury by generating ROS and secreting several proteases such as myeloperoxidase, elastase, and collagenase (Jaeschke, 1991). If the sinusoidal lining cells (i.e., the endothelial cells) are damaged or even removed, for example during severe ischemia-reperfusion injury, neutrophils can directly attack hepatocytes without transmigration through the endothelial cell layer (McKeown et al., 1988; Caldwell-Kenkel et al., 1989; Fisher et al., 1997). Under these conditions, Kupffer cells can also directly access to parenchymal cells. Therefore, in addition to direct protection of the parenchymal cells against ROS, the protection of the endothelial cells can suppress the ROS injuries in the parenchymal cells. In the present study, to assess the preventive effect of the CAT and SOD derivatives on the initial hepatic injury mediated by ROS, we measured the plasma GPT and GOT, markers for parenchymal cell injury, at 1 h after reperfusion.

The ischemia followed by reperfusion resulted in a striking increase in plasma GPT and GOT (Fig. 4), indicating that liver PC is severely damaged by the treatment. This hepatic damage can be also confirmed by the histological observation (Fig. 5a). We already reported that native CAT attenuates the increase of the plasma levels of the indicators for hepatocyte injury in a dose-dependent manner in the same protocol as this study (Yabe et al., 1999). Suc-CAT and Man-CAT showed greater inhibitory effects on the ischemia/reperfusion injury than unmodified CAT at the same doses. The disposition of these CAT derivatives can be characterized by greater CL_liver and significant distribution to the liver NPC. However, Suc-BSA and Man-BSA, both of which are also delivered to the liver NPC, showed no inhibitory effects, indicating that the observed effects of Suc-CAT and Man-CAT result from the catalase activities of these derivatives. In addition, Gal-CAT, which has the greatest CL_liver among CAT derivatives examined but hardly distributes to the NPC, was not so effective as Suc-CAT and Man-CAT. Therefore, these results strongly suggest that targeted delivery of CAT to the liver NPC improves its pharmacological activity against the injury. In the present study, plasma GPT and GOT, markers for parenchymal cell injury, were evaluated to examine the hepatic injury, and Suc-CAT and Man-CAT successfully reduced these activities in plasma. Although their precise mechanism of the PC protection remains to be clarified, the protection of the endothelial cells which function as the `barrier' of parenchymal cells against ROS, as discussed above, can suppress the parenchymal cell injury.

The activity of endogenous antioxidant enzymes, especially CAT, in NPC is much less than in PC (Hamer et al., 1995; Spolarics and Wu, 1997), which suggests that the endothelial cells are much more susceptible to ROS than hepatocytes. Based on the CAT activities per cell protein (Spolarics and Wu, 1997), and protein contents of cells (Knook and Brouwer, 1980), the CAT activities can be calculated to be 280 to 330 U/10⁸ endothelial cells and 1100 to 1700 U/10⁸ Kupffer cells. From the biodistribution results, we can estimate that about 2000 to 3000 U/10⁸ cells of catalase activity can be delivered to the liver NPC when Suc-CAT or Man-CAT was injected at a dose of 10000 U/kg.

Apart from native CAT, whose handling by hepatocyte is unknown, Suc-CAT, and Man-CAT, and Gal-CAT should be internalized by the cells after binding to their respective surface receptors (Fujita et al., 1992b; Furitsu et al., 1997). Therefore, both the surface-bound and internalized CAT derivatives in endocytotic vesicles could be responsible for the pharmacological activity in hepatic injury. It has been reported that extracellular hydrogen peroxide can be degraded by intracelluar CAT because hydrogen peroxide is amphipathic and easily crosses cellular membranes (Inoue, 1994). Although further studies are required to clarify the intracellular localization of CAT derivatives, these results suggest that CAT delivery via receptor-mediated endocytosis is useful. In contrast, ¹¹¹In-PEG-CAT gradually accumulated in the liver with a plasma half-life of about 79 min, but the inhibitory effect of PEG-CAT was not pronounced. These results indicate that this long-circulating CAT derivative is less effective than CAT derivatives targeted to the liver NPC, at least in the initial phase of the ischemia/reperfusion injury.

SOD is an enzyme dismutating superoxide into hydrogen peroxide and oxygen. A beneficial effect has been reported in studies of hepatic ischemia/reperfusion in vivo for the pretreatment of SOD (Atalla et al., 1985; Flye and Yu, 1987; Romani et al., 1988; Cho et al., 1990; Nauta et al., 1990). We also showed that targeted delivery of SOD to the liver by galactosylation or mannosylation resulted in improved pharmacological activity against hepatic ischemia/reperfusion injuries in rat models (Fujita et al., 1992b; Kondo et al., 1996; Mizoe et al., 1997). In this study, the effectiveness of both Gal-SOD and Man-SOD was also demonstrated in the hepatic ischemia/reperfusion model in mice. These results prompted us to use both CAT and SOD derivatives, especially NPC-specific derivatives, in combination in this mouse model.

In receptor-mediated hepatic targeting, the dose of ligand is an important factor for efficient delivery (Nishikawa et al., 1992; Takakura et al., 1994). Therefore, before the combined use of CAT and SOD derivatives targetable to the liver NPC in the therapeutic experiments, we examined the biodistribution of these derivatives after simultaneous administration at a therapeutic dose of 10,000 U/kg (Fig. 6). In spite of being targeted to the same cells (NPC) or sharing the same receptors on the cells, the two types of antioxidant enzymes did not significantly affect the apparent hepatic uptake of their radiolabeled counterpart after coadministration. A slight reduction was observed only in the combination of ¹¹¹In-Man-CAT and Man-SOD. However, these results indicate that combination of two different ligands of the receptor(s) on the liver NPC is not critical for efficient delivery of the ligands within the dose range used in this study.

In combination therapy, coinjection of Gal-SOD with CAT derivatives failed to improve the pharmacological activity of the enzyme (Fig. 7). On the contrary, Man-SOD improved the inhibitory effect of native CAT and CAT derivatives (Suc-CAT, Man-CAT or Gal-CAT) against the injury when injected simultaneously, suggesting the usefulness of targeting these enzymes to the NPC, even in combination therapy. Of these, coadministration of Suc-CAT (10,000 U/kg) and Man-SOD (10,000 U/kg) exhibited the greatest inhibitory effect. However, this effect was comparable with that obtained by Suc-CAT or Man-CAT alone at a higher dose (20,000 U/kg). Although further studies employing other evaluation methods will be required to compare the efficacies of the different therapeutic regimens, these results suggest the combination of CAT and SOD derivatives could be a useful way of preventing this injury. To optimize the therapeutic use of CAT and/or SOD derivatives in single or combined use, the molecular mechanism(s) underlying the preventive effects of the enzyme derivatives need to be elucidated.

In conclusion, we successfully synthesized four CAT derivatives showing different pharmacokinetic properties, especially in terms of liver-cell specificity. Assessment of their pharmacological activity against a hepatic ischemia/reperfusion injury in mice indicates that targeted delivery of CAT to the NPC, the source of ROS in the liver, is a promising strategy for ROS-mediated injuries in the liver. Furthermore, the potential usefulness of the CAT derivatives in combination with SOD derivatives is also demonstrated.