Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Prostaglandin E₁ (PGE₁) is used clinically to treat peripheral vascular disorders. In addition, PGE₁ is known to be effective in fulminant or subfulminant viral hepatitis (Sinclair and Levy, 1991) due to its cytoprotective activity (Stachura et al., 1981; Ueda et al., 1987; Beck et al., 1993; Helling et al., 1995). However, repetitive or persistent administration of PGE₁ is required for the treatment of hepatitis due to its low hydrophilicity and poor physiological stability (Monkhouse et al., 1973; Younger and Szabo, 1986). Although an autoradiographic study demonstrated that [³H]PGE₁ is mainly distributed in the liver and kidneys (Hansson and Samuelsson, 1965), as much as 80% of PGE₁ is metabolized and inactivated by - or -oxidation during the first passage through the lung (Porst, 1996). Furthermore, a number of side effects (i.e., abdominal pain, diarrhea, hypotension, and peripheral edema) have occurred during hepatitis therapy with PGE₁. Therefore, development of a suitable delivery system that achieves hepatic targeting of pharmacologically active PGE₁ is of great interest.

Recently, delivery systems for PGE₁ have been investigated using cyclodextrins (see, for example, Uekama et al., 1992), lipid microspheres (see, for example, Mizushima and Hoshi, 1993), liposomes (see, for example, Rossetti et al., 1994), and a heparin conjugate (Jacobs and Kim, 1986). However, there have been few studies of the hepatic targeting of PGE₁ directed at the treatment of hepatitis. In the past few years, we achieved sugar receptor-mediated selective targeting of various pharmaceutical agents to the liver (Fujita et al., 1992a,b; Nishikawa et al., 1992, 1993, 1995a,b, 1998; Hirabayashi et al., 1996; Hashida et al., 1997; Mahato et al., 1997). Based on these investigations, PGE₁ attachment to galactosylated poly(L-glutamic acid) (Gal-PLGA) by the carbonyldiimidazole method was developed to deliver PGE₁ to liver parenchymal cells through recognition by the asialoglycoprotein receptor (Akamatsu et al., 1997). The prostaglandin E₁ prodrug using Gal-PLGA as a carrier synthesized by this method showed no pharmacological activity, although it could be successfully delivered to liver parenchymal cells.

To retain the pharmacological activity of PGE₁ after conjugation, the synthetic procedure must be altered. In this study, a hydrazide group was introduced into PLGA to couple PGE₁ through its ketone group in a weakly acidic buffer at room temperature, conditions under which PGE₁ is expected to be chemically (Ferruti et al., 1979) most stable (Monkhouse et al., 1973; Younger and Szabo, 1986). The usefulness of newly developed PLGA derivatives as hepatocyte-specific carriers was assessed by biodistribution experiments, and then the pharmacological activity of the PGE₁ conjugate was examined in mice with carbon tetrachloride-induced hepatitis.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals

Poly(-benzyl L-glutamate) derivatives with average molecular weights of approximately 17,300 [degree of polymerization (DP) 79] and 21,200 (DP 97) were purchased from Sigma Chemical Co. (St. Louis, MO). D-Galactose and hydrazine monohydrate were obtained from Wako Pure Chemicals (Osaka, Japan). Diethylenetriamine-N,N,N',N",N"-pentaacetic dianhydride (DTPA anhydride) was obtained from Dojindo Laboratory (Kumamoto, Japan). [¹¹¹In]InCl₃ was supplied by Nihon Medi-physics (Takarazuka, Japan). PGE₁ was obtained from Ono Pharmaceutical (Osaka, Japan). [5,6(n)-³H]PGE₁ was purchased from Amersham Japan (Tokyo, Japan). All other chemicals were reagent grade products obtained commercially.

Animals

Male ddY mice (27-31 g) were obtained from the Shizuoka Agricultural Cooperative Association for Laboratory Animals (Shizuoka, Japan).

Synthesis of PGE₁ Conjugate Using PLGA Hydrazide (PLGA-HZ)

Synthesis of PLGA-Hydrazide. PLGA-HZ was synthesized according to the method of Hurwitz et al. (1980) with slight modification (Fig. 1). In brief, poly(-benzyl L-glutamate) (460 mg) was dissolved in 3 ml of dimethyl formamide and 15 ml of hydrazine monohydrate was added to the solution. After 3 h with stirring at room temperature, the reaction mixture was transferred to dialysis tubing (3500 molecular weight cut-off) and dialyzed thoroughly against distilled water. The dialysate (if a gel formed, 3 M HCl solution was added until a homogeneous solution was obtained) was concentrated by ultrafiltration (10,000 cut-off) and lyophilized to obtain PLGA-HZ. The number of hydrazide groups in PLGA-HZ was measured photometrically by the -naphthoquinone-4-sulfonate method (Pratt, 1953).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Synthetic process of PGE1 conjugate.

Synthesis of 2-Imino-2-Methoxyethyl-1 (IME)-Thiogalactoside. IME-thiogalactoside was prepared as described previously (Nishikawa et al., 1995a). Briefly, cyanomethyl-1-thiogalactoside was treated with 0.01 M sodium methoxide methanolic solution at room temperature for 24 h. The solvent was evaporated in vacuo to obtain a syrupy product (IME-thiogalactoside).

Galactosylation of PLGA-HZ Using IME-Thiogalactoside. PLGA-HZ (140 mg) was dissolved in 1 ml of 2 M HCl and neutralized by the addition of 2 M NaOH. Then, the solution was slowly added to IME-thiogalactoside dissolved in 3 ml of 50 mM borate buffer (pH 9.4). After stirring for 5 h at room temperature, the reaction mixture was transferred to dialysis tubing (3500 cut-off) and dialyzed thoroughly against distilled water. The dialysate was concentrated by ultrafiltration (10,000 cut-off) and lyophilized to obtain Gal-HZ-PLGA. The number of galactose residues in synthetic Gal-HZ-PLGA was determined by the anthrone-sulfuric acid method.

Conjugation of PGE₁ to Gal-HZ-PLGA. An ethanolic solution of PGE₁ (4.0 mg/0.5 ml), with or without 0.02 ml of [³H]PGE₁ solution (18.5 MBq/ml), was slowly added to Gal-HZ-PLGA (20 mg) dissolved in 5 ml of 0.01 M citrate buffer (pH 5), and the mixture was stirred at room temperature overnight. Then, it was dialyzed against 0.9% aqueous NaCl, and the dialysate that was obtained was maintained at 4°C without lyophilizing. The PGE₁ content of the conjugate was determined by counting ³H radioactivity. The numbers of carboxyl, hydrazide, galactose, and PGE₁ residues and the molecular weights of the PLGA derivatives were theoretically calculated by using the results of the quantitative determinations of hydrazide, galactose, and PGE₁ (Table 1). The number of carboxyl residues was calculated by subtracting that of hydrazide, galactose, and PGE₁ from the DP value of poly(-benzyl-L-glutamate) used. Because there were no substantial differences in results obtained with two PLGA derivatives with different DP values, only the results with derivatives synthesized from a PLGA with a DP of 97 are reported here.

¹¹¹In-Labeling of PLGA-HZ and Gal-HZ-PLGA

PLGA-HZ and Gal-HZ-PLGA were radiolabeled with ¹¹¹In using DTPA anhydride as described previously (Nishikawa et al., 1995a). Each radiolabeled derivative was purified by gel-filtration chromatography using a Sephadex G-25 column (1.5 × 5.0 cm, 0.1 M acetate buffer, pH 6.0), and the solution was replaced with 0.9% NaCl by ultrafiltration. The specific activity of each derivative was approximately 37 MBq/mg.

Biodistribution Experiment

Mice received a 1 mg/kg dose of ¹¹¹In-PLGA-HZ, ¹¹¹In-Gal-HZ-PLGA, or [³H]PGE₁ conjugate in saline by tail vein injection and were housed in metabolic cages for urine collection. At given time points, blood was collected from the vena cava with the animal under ether anesthesia and plasma was obtained by centrifugation. The heart, lung, liver, spleen, and kidney were excised; rinsed with saline; weighed; and examined for radioactivity. The amount of radioactivity in urine was determined by collecting urine both excreted and remaining in the bladder. ¹¹¹In radioactivity was counted in a well-type NaI scintillation counter (ARC-500; Aloka, Tokyo, Japan). ³H radioactivity was counted with a liquid scintillation counter (LSC-5000; Beckman, Tokyo, Japan) after dissolution with Soluene-350 (Packard, Groningen, the Netherlands) and the addition of scintillation medium, Clear-sol I (Nakalai Tesque, Tokyo, Japan). Radioactivity originating from the plasma in each tissue sample was corrected using the distribution data for ¹¹¹In-labeled BSA at 10 min after i.v. injection (Nishikawa et al., 1995b), assuming that ¹¹¹In-labeled BSA was not taken up by tissue during this 10-min period.

Pharmacokinetic Analysis

Tissue distribution patterns of ¹¹¹In-PLGA derivatives were evaluated using organ uptake clearance according to the method reported previously (Takakura et al., 1987). In the early period after injection, the efflux of ¹¹¹In radioactivity from organs is assumed to be negligible because the degradation products of ¹¹¹In-labeled ligands using DTPA anhydride cannot easily pass through biological membranes (Duncan and Welch, 1993; Arano et al., 1994). With this assumption, organ uptake clearance was calculated by dividing the amount of radioactivity in an organ at 10 min by the area under the plasma concentration-time curve (AUC) up to the same time point. AUC and total body clearance (CL_total) were calculated by fitting an equation to the plasma concentration data of the derivatives using the nonlinear least-squares program MULTI (Yamaoka et al., 1981).

Hepatic Cellular Localization of PLGA Derivatives

The intrahepatic distribution of ¹¹¹In-PLGA-HZ and ¹¹¹In-Gal-HZ-PLGA between parenchymal and nonparenchymal cells was determined by separating these cells using collagenase after i.v. injection, as reported previously (Blomhoff et al., 1985). Mice were anesthetized with pentobarbital sodium and injected i.v. with ¹¹¹In-PLGA-HZ or ¹¹¹In-Gal-HZ-PLGA. At 15 min after administration, the liver was perfused first with preperfusion buffer (Ca²⁺, Mg²⁺-free HEPES solution, pH 7.2) for 10 min and then with HEPES solution containing 5 mM CaCl₂ and 0.05% (w/v) collagenase (type I) (pH 7.5) for approximately 10 min. As soon as perfusion was started, the vena cava and aorta were cut. After the discontinuation of perfusion, the liver was excised and deprived of the capsule membranes. The cells were dispersed by gentle stirring in ice-cold Hanks-HEPES buffer containing 0.1% BSA. The dispersed cells were filtered through the cotton-mesh sieves, followed by centrifugation at 50g for 1 min. The pellets containing parenchymal cells were washed twice with Hanks-HEPES buffer by centrifugation at 50g for 1 min. The supernatant containing nonparenchymal cells was similarly centrifuged two additional times. The resulting supernatant was then centrifuged twice at 200g for 2 min. Parenchymal and nonparenchymal cells were resuspended separately in ice-cold Hanks-HEPES buffer. The cell number was determined by the trypan blue exclusion method. The cells (0.5 ml) were digested with Soluene-350 (1 ml) through incubation overnight at 45°C. After digestion, 0.3 ml of 2 N HCl and 5 ml of Clear-sol I were added, the mixture was stored overnight, and radioactivity was measured using a scintillation counter. The amount of radioactivity on each cell fraction was calculated as the percentage of dose/10⁸ cells.

Degradation of Gal-HZ-PLGA in Liver

Mice were injected with ¹¹¹In-Gal-HZ-PLGA in saline at a dose of 1 mg/kg. At 10, 30, and 120 min after injection, they were sacrificed, and the liver was excised. The liver was homogenized with 4 ml of distilled water, and 1 ml of concentrated KCl solution was added to the homogenate, followed by vortexing. After cooling overnight at 4°C, the homogenate was centrifuged at 26,000g for 30 min at 4°C. The supernatant of the liver homogenate was applied to a Sephadex G-25 column (1 × 40 cm) and eluted with 0.1 M acetate buffer (pH 6.0), and the radioactivity of each fraction was counted. The percentage of degradation was calculated from the elution profiles.

Release of PGE₁ Derivative from PGE₁ Polymeric Conjugate in Liver Homogenate

Degradation in liver homogenate was examined according to the method of Gonsho et al. (1994). Mouse liver was homogenized in 3 ml of PBS (pH 7.4), and the supernatant (liver homogenate) was obtained by centrifugation at 4000g for 5 min. Five hundred microliters of the [³H]PGE₁ conjugate (8 mg/ml) was added to 0.5 ml of liver homogenate and incubated at 37°C. At various intervals, 50 µl was withdrawn from the mixture and placed in 0.4 ml of ethanol together with 0.1 ml of 5% BSA. After centrifugation at 12,000g for 2 min, the radioactivity in the supernatant (50 µl) was measured.

In Vivo Pharmacological Activity of PGE₁ Polymeric Conjugate

The pharmacological activity of PGE₁ conjugate was evaluated by measuring glutamic pyruvic transaminase (GPT) activity in the plasma of mice with carbon tetrachloride (CCl₄)-induced hepatitis. Mice received an i.p. injection of CCl₄ (2% solution in sesame oil) at a dose of 10 ml/kg. Immediately after CCl₄ administration, the drug solution was injected i.v. into the tail vein. The mice were then starved for 18 h after CCl₄ administration, and blood was collected from the vena cava under ether anesthesia and plasma was obtained by centrifugation. GPT activity in the plasma samples was determined by the UV-rate method (Wroblewski and La Due, 1956).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Plasma Clearance and Tissue Distribution of ¹¹¹In-PLGA Derivatives and [³H]PGE₁ Conjugate

Figure 2 shows the concentrations in the plasma (Fig. 2a) and amounts in tissues (Fig. 2, b-e) of ¹¹¹In-PLGA-HZ, ¹¹¹In-Gal-HZ-PLGA, and [³H]PGE₁ conjugate after i.v. injection in mice. ¹¹¹In-PLGA-HZ was largely recovered in the kidney and urine, and hardly any accumulated in other tissues (Fig. 2b). ¹¹¹In-Gal-HZ-PLGA was rapidly eliminated from plasma (Fig. 2a) and was largely recovered in the liver at about 75% of the dose within 10 min after a dose of 1 mg/kg (Fig. 2c), whereas at a dose of 10 mg/kg, it accumulated in the liver up to only 40% of dose even within 60 min after injection (Fig. 2d) by means of saturation of asialoglycoprotein receptor, which has been demonstrated in an inhibition experiment by coinjection with galactosylated BSA (data not shown). The disposition characteristics of [³H]PGE₁ conjugate were comparable to those of ¹¹¹In-Gal-HZ-PLGA, indicating that little PGE₁ was released from the carrier before delivery to the liver. The ³H radioactivity in the liver after injection of the [³H]PGE₁ conjugate remained at a high level throughout the experimental period (72 h).

View larger version (25K): [in this window] [in a new window]   Fig. 2.   Plasma concentration, tissue accumulation-time courses of 111In-PLGA-HZ, 111In-Gal-HZ-PLGA, and 3H-labeled PGE1 prodrug using Gal-HZ-PLGA as a carrier ([3H]PGE1 conjugate) (DP of these polymers is 97) after i.v. injection in mice. Results are expressed as the mean ± S.D. of three mice. a, plasma concentration-time courses of 111In-PLGA-HZ at a dose of 1 mg/kg (), 111In-Gal-HZ-PLGA at a dose of 1 mg/kg (), 111In-Gal-HZ-PLGA at a dose of 10 mg/kg (), and 3H-PGE1 conjugate at a dose of 1 mg/kg (0.074 mg of equivalent PGE1/kg) (). b-e, tissue accumulation-time curves of 111In-PLGA-HZ at a dose of 1 mg/kg, 111In-Gal-HZ-PLGA at a dose of 1 mg/kg, 111In-Gal-HZ-PLGA at a dose of 10 mg/kg, and 3H-PGE1 conjugate at a dose of 1 mg/kg, respectively (kidney, ; spleen, ; liver, ; lung, ; urine, ).

Pharmacokinetic Analysis of ¹¹¹In-PLGA Derivatives and [³H]PGE₁ Conjugate

Table 2 summarizes the clearances for liver (CL_liver), kidney, urine, spleen, and lung; CL_total; and AUC for each PLGA derivative. ¹¹¹In-Gal-HZ-PLGA had a larger CL_liver than ¹¹¹In-PLGA-HZ. Increasing the dose reduced the CL_liver of ¹¹¹In-Gal-HZ-PLGA. This tendency was similar to the disposition characteristics of other galactosylated macromolecules whose uptake by the liver is considered to be mediated by asialoglycoprotein receptors (Nishikawa et al., 1993; Akamatsu et al., 1997). [³H]PGE₁ conjugate had a much greater CL_liver than ¹¹¹In-Gal-HZ-PLGA. Such an excess CL_liver value was considered to be led by rapid elimination of PGE₁ conjugate from plasma (Fig. 2a), which is responsible for a small AUC value. The galactose density of Gal-HZ-PLGA was thought to become higher by its conformational change by PGE₁ bound, to be easily recognized by asialoglycoprotein receptors on hepatic parenchymal cells (Nishikawa et al., 1995a).

Cellular Distribution of ¹¹¹In-PLGA Derivatives in Liver

Figure 3 shows the intrahepatic distribution of ¹¹¹In-PLGA-HZ and ¹¹¹In-Gal-HZ-PLGA between parenchymal and nonparenchymal cells in the liver at 30 min after i.v. injection. ¹¹¹In-PLGA-HZ was hardly recovered in both parenchymal and nonparenchymal cells, reflecting its poor uptake by the liver (Fig. 2). On the other hand, ¹¹¹In-Gal-HZ-PLGA was selectively taken up by parenchymal cells, and the amount recovered in parenchymal cells was more than twice that in nonparenchymal cells on a cell-number basis. From the liver cell numbers [parenchymal and nonparenchymal cells are 1.25 × 10⁸ and 0.65 × 10⁸ cells/g liver in mouse, respectively (Blomhoff et al., 1985)] and the average liver weight (about 1.1 g) of mice used in our study, the recovery in the whole liver was calculated to be approximately 84% of the dose (68 and 16% in parenchymal and nonparenchymal cells, respectively). The calculated value (84%) corresponds well to the biodistribution data in Fig. 2.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Hepatic cellular localization of radioactivity after i.v. injection of [111In]PLGA-HZ and [111In]Gal-HZ-PLGA (DP 97) in mice at a dose of 1 mg/kg. At 15 min after administration, the liver cells were separated into parenchymal cells () and nonparenchymal cells () by collagenase perfusion and subjected to radioactivity counting. The value was calculated as the percentage of dose/108 cells to compare their contribution to hepatic uptake of the carrier on the cell-number basis. Statistically significant difference based on Student's t test (*P < .05, **P < .01). Results are expressed as the mean ± S.D. of three mice.

Degradation of ¹¹¹In-Gal-HZ-PLGA in Liver

The biodegradability of ¹¹¹In-Gal-HZ-PLGA in the liver was evaluated for the elution profiles of ¹¹¹In-radioactivity recovered in the liver homogenate of mice injected with the derivative (Fig. 4). The amounts of radioactivity in the liver homogenate eluted in low-molecular-weight fractions increased with time, and about 60% of the radioactivity was recovered in those fractions 2 h after injection. These results suggest that the polymer chain of Gal-HZ-PLGA, which is linked via the amide bond, is rapidly digested enzymatically in the liver. This characteristic may be advantageous as a polymeric prodrug for the release of active component conjugated with the polymer backbone.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Gel-filtration (Sephadex G-50) patterns of 111In-labeled Gal-HZ-PLGA (DP 97) in liver for 10 (), 30 (), and 120 () min after i.v. injection in mice at a dose of 1 mg/kg. The first and second peaks correspond to high-molecular-weight (polymer) and low-molecular-weight (degradation products) fractions, respectively.

Release of PGE₁ Derivatives from PGE₁ Conjugate in Liver Homogenate

To exhibit therapeutic effect, PGE₁ should be released from the polymeric prodrug in the hepatocytes after endocytosis. To address this issue, we carried out stability experiments using [³H]PGE₁ conjugate. Figure 5 shows the release profile of ethanol-soluble ³H radioactivity generated from the [³H]PGE₁ conjugate during incubation in the liver homogenate. Ethanol-soluble ³H radioactivity, presumably [³H]PGE₁ and/or its derivatives, gradually increased with time, reaching about 30% after a 24-h incubation. On the other hand, [³H]PGE₁ conjugate remained stable during incubation in buffered solution (pH 5.5 and 7.4) or in mouse plasma at 37°C for 24 h (data not shown). These results imply that the PGE₁ conjugate could release active drug selectively in the hepatocytes after endocytosis.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   In vitro release profile of ethanol-soluble 3H-compounds derived from [3H]PGE1 conjugate during incubation in the mouse liver homogenate. [3H]PGE1 conjugate (DP 79) was incubated in the homogenate at 37°C, and aliquots were withdrawn at appropriate intervals. 3H-radioactivity soluble in ethanol was measured as compounds ([3H]PGE1 or its derivatives) released from the conjugate, assuming that intact PGE1 conjugate is insoluble in the solvent. The values are expressed as the percentage of applied radioactivity ([3H]PGE1 conjugate).

Therapeutic Activity of PGE₁ Polymeric Conjugate in Acute Hepatitis

The activity of the PGE₁ polymeric conjugate in hepatitis was evaluated by measuring GPT activity in the plasma of mice with experimentally induced hepatitis (Fig. 6). Intraperitoneal administration of CCl₄ resulted in marked increase in plasma GPT. Bolus i.v. injection of PGE₁ (2 mg/kg) showed no significant effect on the elevation of the GPT level. On the other hand, i.v. injection of PGE₁ conjugate, even at a lower dose of free PGE₁ (0.074 mg equivalent PGE₁/kg), significantly suppressed the GPT increase (P < .001, P < .01 compared with saline- and PGE₁-treated mice, respectively). A 10-fold increase in the dose injected once or twice was also effective, but the pharmacological activity was comparable with that obtained with the low dose of conjugate. These results suggest that the PGE₁ conjugate can exhibit sufficient suppression of hepatitis after bolus administration at a relatively low dose, whereas free PGE₁ should be infused continuously.

View larger version (33K): [in this window] [in a new window]   Fig. 6.   GPT levels in the plasma of hepatitis mice treated by PGE1 or PGE1 conjugate (DP 97, 5 mol PGE1/mol conjugate). Statistically significant difference based on ANOVA (Student-Newman-Keuls multiple comparisons test). *P < .001, compared with saline-treated mice; #P < .01, ##P < .05, compared with PGE1-treated mice. Results are expressed as the mean ± S.D. of at least five mice except for CCl4-untreated group (n = 2). Hepatitis mice were prepared by receiving an intraperitoneal injection of CCl4 (2% solution in sesame oil) at a dose of 10 ml/kg. In this experiment, four kinds of drug solutions were used: saline (0.9% NaCl), PGE1 (0.2 mg/ml polyethyleneglycol solution), and PGE1 conjugate (0.1, 0.5, or 1 mg/ml aqueous solution). Each solution except for 0.5 mg/ml PGE1 conjugate was injected i.v. into the tail vein at a dose of 2, 1, or 10 mg/kg, respectively. A group received 0.5 mg/kg PGE1 conjugate injected twice at an interval of 2 h (total dose, 10 mg/kg). GPT activity in the plasma samples excised from vena cava after keeping for 18 h was determined by the UV-rate method.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

To achieve cell-specific, targeted drug delivery, an appropriate drug delivery system, which can deliver and release pharmacologically active drugs at target sites, must be developed. To develop PGE₁ polymeric prodrug as a therapeutic agent for fulminant hepatitis, we previously reported a synthetic method and measured the biodistribution of PGE₁ polymeric conjugate using PLGA as a carrier (Akamatsu et al., 1997). In the conjugate, carbonyldiimidazole was used as a condensation agent to bind the carboxyl group of PGE₁ to PLGA-ethylenediamine. This type of PGE₁ conjugate had excellent liver specificity but no significant pharmacological activity in hepatitis, suggesting that pharmacologically active PGE₁ might not be released after systemic administration, probably due to inactivation after the conjugation reaction with carbonyldiimidazole.

PGE₁ is chemically unstable; the -hydroxyketone moiety is rapidly dehydrated under both acidic and basic conditions (Monkhouse et al., 1973). There are three possible methods to avoid the -elimination reaction of PGE₁: 1) optimizing the pH, 2) decreasing the temperature, and 3) shortening the time of the conjugation reaction. Monkhouse et al. (1973) demonstrated that the dehydration rate of PGE₁ by -elimination reached a minimum value around pH 4. Therefore, we first attempted to conjugate PGE₁ to PLGA through a Schiff's base between primary amines and aldehydes or ketones, which can be catalyzed by dilute acid under weakly acidic conditions. However, the reaction rate of amines with ketones is generally slower than that with aldehydes. In particular, the cyclopentanone-type carbonyl group, which is a component of PGE₁, is expected to be difficult to react with nucleophiles due to steric hindrance of the transition state. In fact, poly-D-lysine, possessing many primary amino groups, hardly reacted with PGE₁ (data not shown). On the other hand, hydrazine and hydroxylamine are thought to be more nucleophilic than general amines (Buncel et al., 1980). Therefore, polymers containing hydrazide groups, like poly(L-glutamic acid) hydrazide, are expected to be better candidates for preparing PGE₁ polymeric conjugate. In this study, we succeeded in synthesizing a PGE₁ polymeric conjugate where PGE₁ is bound to PLGA-HZ through a hydrazone bond.

To exhibit activity, pharmacologically active drugs must be released from their carriers (Sezaki and Hashida, 1984). The hydrazone bond between an aldehyde and an acid hydrazide can be hydrolyzed. For example, Coessens et al. (1996) demonstrated that the hydrolysis half-lives of a hydrazone bond between streptomycin, an aldehyde compound, and polymeric hydrazide, in buffer solution (pH 5.2 and 7.4) at 37°C, were about 8 and over 12 h, respectively. However, in our experiment, we hypothesized that the hydrazone bond between the ketone and acid hydrazide would be difficult to hydrolyze (data not shown), probably because of the steric hindrance in the transition state. Although PGE₁ was not released in buffers and mouse plasma, it seemed to be released in liver homogenates (Fig. 5), probably due to enzymatic hydrolysis by lysosomal enzymes. Although we could not identify the compound, our preliminary analysis by thin-layer chromatography revealed that [³H]PGE₁ and/or its derivatives could be generated from the [³H]PGE₁ conjugate in the homogenate (data not shown). These results suggest that the PGE₁ conjugate could act as a prodrug; the conjugate might release its active component after being taken up by the hepatocytes without releasing it in the blood circulation.

In this study, CCl₄ dissolved in oil was injected intraperitoneally into mice to induce hepatitis. Although the mechanism of liver injury by CCl₄ is not fully understood, Kupffer cells may contribute to the injury (Edwards et al., 1993). These authors demonstrated that depletion of Kupffer cells dramatically suppressed the necrosis of hepatocytes induced by CCl₄. Therefore, suppression of Kupffer cell activation by PGE₁ could be responsible, at least in part, for its prevention of hepatic injuries. Kayano et al. (1995) demonstrated that production of cytotoxic cytokines and tumor necrosis factor- from activated Kupffer cells was dramatically inhibited by PGE₁ in rats. In our mouse model, PGE₁ generated in hepatocytes also may diffuse into Kupffer cells because dihomo--linoleic or arachidonic acid derivatives, including prostaglandins, in general, are known to be transported easily across cell membranes. On the other hand, PGE₁ is reported to directly protect hepatocytes from injuries induced by tert-butyl hydroperoxide in culture systems (Masaki et al., 1992). Although the protective mechanism of the PGE₁ conjugate remains to be clarified, direct protection of hepatocytes could be a major mechanism because PGE₁ is targeted to liver parenchymal cells.

The PGE₁ conjugate synthesized significantly and effectively suppressed the increase in plasma GPT in mice with hepatitis (Fig. 6). Therefore, pharmacologically active PGE₁ derivatives would be released in the liver parenchymal cells after receptor-mediated endocytosis and prevent the hepatic injury induced by CCl₄ administration. On the other hand, a bolus injection of free PGE₁ was ineffective even at a higher dose probably due to low availability in the liver. This is in good agreement in other studies demonstrating that continuous infusion of this drug for a long period with a large dose in total is required to obtain preventive effects in acute liver injuries (Masaki et al., 1992; Quiroga and Prieto, 1993). However, increasing the dose of the conjugate resulted in little enhancement of its pharmacological activity. The pharmacokinetic characteristics of the PGE₁ conjugate could explain its dose-independent efficacy to some extent. As shown in Fig. 2, the hepatic uptake of ¹¹¹In-Gal-HZ-PLGA decreased with an increase in the dose due to saturation of asialoglycoprotein receptor-mediated uptake.

In conclusion, the PGE₁ conjugate with Gal-HZ-PLGA is an effective prodrug of PGE₁ that can be targeted to liver parenchymal cells, and it exhibits enhanced suppression in the treatment of fulminant hepatitis.