Regular Article Involvement of Human Blood Arylesterases and Liver Microsomal Carboxylesterases in Nafamostat Hydrolysis

Drug Metab. Pharmacokinet. 21 (2): 147–155 (2006).

Satoshi YAMAORI1, Nobuhiro FUJIYAMA1,MikaKUSHIHARA1, Tatsuya FUNAHASHI1, Toshiyuki KIMURA1,IkuoYAMAMOTO2, Tomomichi SONE3, Masakazu ISOBE3, Tohru OHSHIMA4,KenjiMATSUMURA5,MinoruODA5 and Kazuhito WATANABE1 1Department of Hygienic Chemistry, Faculty of Pharmaceutical Sciences, Hokuriku University, Kanazawa, Japan 2Department of Hygienic Chemistry, School of Pharmaceutical Sciences, Kyushu University of Health and Welfare, Nobeoka, Miyazaki, Japan 3Department of Toxicology, Faculty of Pharmaceutical Sciences, Setsunan University, Osaka, Japan 4Forensic and Social Environmental Medicine, Graduate School of Medical Science, Kanazawa University, Kanazawa, Japan 5Drug Information Department, Torii Pharmaceutical Co., Ltd., Tokyo, Japan

Full text of this paper is available at http://www.jstage.jst.go.jp/browse/dmpk

Summary: Metabolism of nafamostat, a clinically used serine protease inhibitor, was investigated with human blood and liver enzyme sources. All the enzyme sources examined (whole blood, erythrocytes,

plasma and liver microsomes) showed nafamostat hydrolytic activity. Vmax and Km values for the nafamostat hydrolysis in erythrocytes were 278 nmolWminWmL blood fraction and 628 mM; those in plasma were 160 nmolWminWmL blood fraction and 8890 mM, respectively. Human liver microsomes

exhibited a Vmax value of 26.9 nmolWminWmg protein and a Km value of 1790 mM. Hydrolytic activity of the erythrocytes and plasma was inhibited by 5, 5?-dithiobis(2-nitrobenzoic acid), an arylesterase inhibitor, in a concentration-dependent manner. In contrast, little or no suppression of these activities was seen with phenylmethylsulfonyl ‰uoride (PMSF), diisopropyl ‰uorophosphate (DFP), bis(p- nitrophenyl)phosphate (BNPP), BW284C51 and ethopropazine. The liver microsomal activity was markedly inhibited by PMSF, DFP and BNPP, indicating that carboxylesterase was involved in the nafamostat hydrolysis. Human carboxylesterase 2 expressed in COS-1 cells was capable of hydrolyzing nafamostat at 10 and 100 mM, whereas recombinant carboxylesterase 1 showed signiˆcant activity only at a higher substrate concentration (100 mM). The nafamostat hydrolysis in 18 human liver microsomes correlated with aspirin hydrolytic activity speciˆc for carboxylesterase 2 (r＝0.815, pº0.01) but not with imidapril hydrolysis catalyzed by carboxylesterase 1 (r＝0.156, p＝0.54). These results suggest that human arylesterases and carboxylesterase 2 may be predominantly responsible for the metabolism of nafamostat in the blood and liver, respectively.

Key words: nafamostat; human; blood; liver; arylesterase; carboxylesterase

tion. In addition, nafamostat is also prescribed as an Introduction anticoagulant in extracorporeal circulation. Nafamostat Nafamostat is a potent inhibitor of various serine is an ester conjugate of 6-amidino-2-naphthol (AN) and proteases including trypsin, thrombin and C1 esterase.1) p-guanidinobenzoic acid (pGBA), both of which them- This drug is widely used for the treatment of acute selves do not inhibit protease activities.2,3) Therefore, pancreatitis and disseminated intravascular coagula- the pharmacological activity of nafamostat may be

Received; August 8, 2005, Accepted; December 12, 2005 To whom correspondence should be addressed:KazuhitoWATANABE,Ph.D.,Department of Hygienic Chemistry, Faculty of Pharmaceutical Sciences, Hokuriku University, Ho-3, Kanagawa-machi, Kanazawa 920-1181, Japan. Tel. ＋81-76-229-6220, Fax. ＋81-76-229-6221, E-mail: k-watanabe＠hokuriku-u.ac.jp

147 148 Satoshi YAMAORI et al.

is very low as compared with that of erythrocytes. Furthermore, our study suggests that arylesterases (EC 3.1.1.2) and carboxylesterase 2 (EC 3.1.1.1) are major enzymes responsible for nafamostat hydrolysis in human blood and liver, respectively. Materials and Methods Materials: Nafamostat mesilate and AN were synthesized in Torii Pharmaceutical Co., Ltd. (Chiba, Japan). Imidapril hydrochloride and imidaprilat were provided by Tanabe Seiyaku (Osaka, Japan). Methyl p-hydroxybenzoate, diisopropyl ‰uorophosphate (DFP), 5, 5?-dithiobis(2-nitrobenzoic acid) (DTNB) and aspirin were obtained from Wako Pure Chemical Ind. (Osaka, Japan). Phenylmethylsulfonyl ‰uoride (PMSF), BW284C51 and bis(p-nitrophenyl)phosphate Fig. 1. Proposed metabolic pathway of nafamostat (BNPP) were purchased from Sigma Chemical Co. (St. Louis, MO). Ethopropazine was obtained from The United States Pharmacopeial Convention, Inc. (Rock- aŠected by the rate of hydrolysis. ville, MD). Dulbecco's modiêd Eagle's medium, fetal A pharmacokinetic study in rats was reported by calf serum, penicillin, streptomycin and Lipofectamine Nanpo et al.4) After nafamostat was administered plus reagent were obtained from Life Technologies intravenously to rats, this drug was hydrolyzed and (Tokyo, Japan). Other chemicals and solvents used were eliminated from the blood at a half-life of approximate- of the highest quality commercially available. ly 2 min. The unchanged drug was transferred rapidly to Preparation of enzyme sources: Blood samples were various tissues including the liver, where approximately collected into tubes containing EDTA from healthy 5z of the administered drug was estimated to reach as volunteers (4 males and 4 females aged 22–54 years, an unchanged form at 2 min after the injection. The HB-1 to -8) from the scientiˆc staŠ. Informed consent elimination of nafamostat from rat liver had a half-life was obtained from all subjects. A part of whole blood of about 17 min and was faster than that from other was stored at －809C before analysis. Erythrocytes and tissues, such as the pancreas, lung and kidney, suggest- plasma were prepared as reported previously.7) Brie‰y, ing that the metabolism might contribute to nafamostat erythrocytes and plasma were separated by centrifuga- elimination from the liver. Major metabolites of tion at 880×g for 5 min. The plasma was removed and nafamostat in rats were AN, its glucuronide conjugate stored at －809C before analysis. The erythrocytes were and pGBA (Fig. 1) and the unchanged form of washed using an equal volume of saline. After centrifu- nafamostat was less detectable in urine and feces of rats. gation at 880×g for 10 min, the upper layer was re- These ˆndings suggest that nafamostat may be hydro- moved and discarded. This procedure was repeated. The lyzed mainly by rat blood esterases and, to a minor erythrocytes were then stored at －809C before analysis. extent, by liver esterases. In humans, a blood half-life of Thawing prior to assay lysed the whole blood and nafamostat elimination was approximately 23 min after erythrocytes. continuous intravenous infusion5) and longer than that Human liver samples were obtained from a 57-year- of rats.4) Such a species diŠerence in drug clearance old Japanese female (HL-1) and 16-year-old Japanese from the blood has been reported previously6) and could male (HL-2) who died in tra‹c accidents. The use of the be explained by a higher esterase activity of rat blood. human livers for these studies was approved by the Accordingly, the contribution of human liver to the Ethics Committee of Kanazawa University, Faculty of nafamostat hydrolysis would be larger than that of rat Medicine. Human liver microsomes were prepared as liver, suggesting that, in addition to the blood, human reported previously.8) Sixteen human liver microsomes liver may be an important metabolic site of nafamostat. (HL-3 to -18) for the Reaction Phenotyping Kit Ver. 6 However, enzymes responsible for the hydrolysis of were purchased from XenoTech (Kansas, KS). nafamostat in human blood and liver remain unclear. Expression of each human carboxylesterase isoform In the present study, we investigated the in vitro in COS-1 cells was carried out as described below. The metabolism of nafamostat using human blood and liver pBluescript vector containing each open reading frame enzymesources.Wereporthereinthatnafamostatis of human carboxylesterase 1 and 2 cDNAs9) was hydrolyzed by human erythrocytes, plasma and liver digested with EcoR I. The cDNA fragment encoding microsomes, although the hydrolytic activity of plasma carboxylesterase was ligated into a linearized mam- Nafamostat Hydrolysis by Arylesterases and Carboxylesterases 149 malian expression vector p91023(B),10) which was kindly To determine kinetic parameters for the nafamostat provided by Dr. R. J. Kaufman (Massachusetts Institute hydrolysis catalyzed by human blood fractions (HB-1) Technology) through Dr. K. Nagata (Tohoku Univer- and liver microsomes (HL-1), nafamostat was incubated sity, Sendai, Japan). COS-1 cells were cultured in with either of these enzyme sources under the same Dulbecco's modiêd Eagle's medium supplemented conditions as mentioned above. In preliminary experi- with 10z heat-inactivated fetal calf serum, 50 IUWmL ments, these reaction conditions were conˆrmed to penicillin and 50 mgWmL streptomycin in a humidiêd ensure initial rates for the formation of AN. The atmosphere containing 5z CO2 at 379C. The cells (2× nafamostat concentrations were 60–1250 mMfor 106) were transfected with 2 mg of the expression plas- erythrocytes, 25–2000 mM for plasma and 10–1560 mM mid carrying each human carboxylesterase isoform by for liver microsomes. Data points were ˆtted to the Lipofectamine plus reagent according to the manufac- Michaelis-Menten equation by nonlinear least-squares turer's instructions. After 48 hr, the transfected cells regression analysis with Origin 7.5J software (Origin- were rinsed and harvested by trypsinization. The Lab, Northampton, MA). All kinetic parameters are collected cells were suspended in phosphate-buŠered given as mean±S.E. saline and sonicated to prepare cell lysates. The lysates The activity of imidapril hydrolysis was determined as were centrifuged at 9000×g for 20 min. The resultant described previously11) with minor modiˆcations. supernatant (S9 fraction) was homogenized and kept at Brie‰y, imidapril (100 mM) was incubated with human －809C until use. The S9 fraction of COS-1 cells trans- liver microsomes (200 mg protein) and 100 mM Tris-HCl fected with a plasmid carrying the carboxylesterase 2 buŠer (pH 7.4) in a ˆnal volume of 500 mLat379Cfor cDNA ligated in the reverse orientation was used as a 50 min, and the reaction was terminated by adding 450 negative control. mL of acetonitrile containing p-hydroxybenzoic acid Enzyme assays: The in vitro half-lives of (0.5 mg) as an I.S. After centrifugation, the supernatant nafamostat in human blood samples (HB-1 to -3) were was subjected to a high-performance liquid chro- determined as described below. Each blood fraction matography (Hitachi L-7100 pump and L-7400 UV (450 mL of whole blood, erythrocytes and plasma) was detector, Hitachi) equipped with a Mightysil RP-18 GP preincubated at 379C for 5 min. Reactions were initiat- column (4.6×250 mm, 5 mm, Kanto Chemical). The ed by addition of 50 mL of nafamostat solution (ˆnal mobile phase was 82 mM sodium acetate buŠer (pH 3.0) concentration of 100 mM). Incubations were carried out containing 25 mM sodium 1-heptanesulfonate: acetoni- at 379C for 0, 5, 10, 20, 30 and 40 min, and terminated trile (8.5:1.5). Elution was performed at a ‰ow rate of by the addition of 500 mL of acetonitrile containing 1.0 mLWmin. The imidaprilat formation was monitored methyl p-hydroxybenzoate (5 mg) as an internal stan- at a wavelength of 237 nm. dard (I.S.). After centrifugation, the supernatant was The activity of aspirin hydrolysis was determined as subjected to a high-performance liquid chromatography described below. Aspirin (100 mM) was incubated with (Hitachi L-2130 pump, L-2200 autosampler and L-2400 human liver microsomes (120 mgprotein)and100mM UV detector, Hitachi, Tokyo, Japan) equipped with a Tris-HCl buŠer (pH 7.4) to make a ˆnal volume of 500 Mightysil RP-18 GP column (4.6×250 mm, 5 mm, mL. Incubations were carried out at 379C for 40 min. Kanto Chemical, Tokyo, Japan). The mobile phase Other methods were the same as described for the was 100 mM sodium acetate buŠer (pH 3.0) containing imidapril hydrolysis, except for the detection of salicylic 30 mM sodium 1-heptanesulfonate: acetonitrile (8:2). acid formation. Elution was performed at a ‰ow rate of 0.5 mLWmin. Inhibition studies: Human erythrocytes, plasma Nafamostat was monitored at a wavelength of 250 nm. and liver microsomes were preincubated with The hydrolytic activity for nafamostat was deter- BW284C51 or ethopropazine at 379C for 30 min. mined as described below. Nafamostat was incubated Human erythrocytes, plasma and the liver microsomes with each enzyme source (20 mL of whole blood, 13 mL were preincubated with PMSF, DFP, BNPP or DTNB of erythrocytes, 100 mL of plasma, 145 mgproteinof at 379C for 5 min. Subsequently, the mixture was liver microsomes and 50 mg protein of S9 fraction of incubated with nafamostat (500 mM) in the same COS-1 cells expressing each human carboxylesterase) manner as described in the enzyme assays. All inhibitors and 100 mM Tris-HCl buŠer (pH 7.4) to make a ˆnal except for DTNB in water were dissolved in dimethyl- volume of 500 mL. Incubations were carried out at 379C sulfoxide and added to the incubation mixture at a ˆnal for 15 min (whole blood), 10 min (erythrocytes), 20 min dimethylsulfoxide concentration of 1z.IC50 values (plasma and liver microsomes) or 40 min (S9 of COS-1 were calculated by nonlinear least-squares regression cells expressing each human carboxylesterase). Other analysis with Origin 7.5J software (OriginLab). methods were the same as described for the in vitro half- Statistical analysis: The statistical signiˆcance of lives of nafamostat in human blood samples, except for diŠerences between nafamostat hydrolytic activities of the detection of AN formation. COS-1 cells expressing human carboxylesterases and the 150 Satoshi YAMAORI et al. control cells was evaluated by means of the unpaired activity was examined with human liver microsomes t-test. The correlation between hydrolytic activities of (HL-1 to -18). The liver microsomal activity at a human liver microsomes was assessed by linear regres- substrate concentration of 10 mM was 0.0870±0.0368 sion analysis. All statistical analyses were carried out nmolWminWmg protein (n＝18), ranging from 0.0521 to with the program InStat (GraphPad Software, San 0.187 nmolWminWmg protein. Diego, CA). Kinetic analysis for the nafamostat hydrolysis was Other method: Protein concentration was estimated carried out with human blood fractions (HB-1) and the 12) by the method of Lowry et al., using bovine serum liver microsomes (HL-1) (Table 2). Vmax and Km values albumin as a standard. for the erythrocytes were 278 nmolWminWmL blood

fraction and 628 mM, respectively, indicating the Vmax W Results Km value of 443 mLWminWmL blood fraction. The Vmax

In vitro half-lives of nafamostat in human blood and Km values for the plasma fraction were 160 nmolW samples: To determine the in vitro half-lives of minWmL blood fraction and 8890 mM, respectively. nafamostat in human blood samples, nafamostat was incubated with human blood fractions (Fig. 2). Nafamostat was decreased with a half-life of 18.8 min in whole blood. A similar half-life of the drug was seen in erythrocytes (18.1 min). In contrast, the half-life in plasma was 41.9 min and approximately twice longer than those in whole blood and erythrocytes. Metabolism of nafamostat by human blood fractions and liver microsomes: Representative chromatograms are shown in Fig. 3. Nafamostat was hydrolyzed by human whole blood, erythrocytes (Fig. 3B), plasma and liver microsomes to produce AN. Hydrolytic activities of whole blood, erythrocytes and plasma at 100 mM nafamostat were 27.7±2.3, 68.5±6.7 and 6.48±0.40 nmolWminWmL blood fraction (mean±S.D., n＝8), respectively (Table 1). Since the hematocrit value in 8 subjects examined was 0.445±0.030, mean calculated activities of erythrocytes and plasma in whole blood were 30.5 and 3.60 nmolWminWmL whole blood, respec- Fig. 2. Plots of nafamostat concentration as a function of incuba- tively. The sum of both calculated activities was roughly tion time with human whole blood (closed circle), erythrocytes (open similar to the hydrolytic activity measured in the whole triangle) and plasma (open square). Individual points and bars blood. On the other hand, the nafamostat hydrolytic represent the mean±S.D. of three subjects (HB-1 to -3).

Fig. 3. Typical chromatograms of nafamostat and its metabolite. (A) Authentic compounds including nafamostat, AN and methyl p-hydroxybenzoate as an I.S. (B) Human erythrocytes were incubated with nafamostat (100 mM). Peak 1; AN (retention time; 8.9 min), peak 2; I.S. (11.6 min), peak 3; nafamostat (16.1 min). Nafamostat Hydrolysis by Arylesterases and Carboxylesterases 151

Table 1. Hydrolytic activities of nafamostat in human blood fractions

Nafamostat hydrolysis (nmolWminWmL blood fraction) Subject Sex Whole blood Erythrocytes Plasma

HB-1 Male 26.9 57.4 5.87 HB-2 Female 25.3 63.8 6.36 HB-3 Male 25.9 73.5 7.02 HB-4 Female 25.1 66.8 6.81 HB-5 Female 28.6 65.8 6.40 HB-6 Male 29.5 78.9 6.73 HB-7 Male 31.8 73.1 6.69 HB-8 Female 28.1 69.1 5.98

All determinations were performed in duplicate.

Table 2. Kinetic parameters of nafamostat hydrolysis catalyzed by hydrolytic activity of human liver microsomes was human erythrocytes, plasma and liver microsomes potently inhibited by PMSF, DFP and BNPP. IC50 values for PMSF (0.25–25 mM), DFP (0.005–0.5 mM) a Km Enzyme source Vmax (mM) and BNPP (0.05–5 mM) were 9.64, 0.266 and 0.262 mM, Erythrocytes 278±16 628±69 respectively. In contrast, the liver microsomal activity Plasma 160±28 8890±1790 was less decreased by the addition of BW284C51, Liver microsomes 26.9±4.2 1790±440 ethopropazine and DTNB. Metabolism of nafamostat by recombinant human Values are represented as mean±S.E. of kinetic parameters. a carboxylesterases: Metabolism of nafamostat was Units of Vmax values for the blood fractions and liver microsomes are showninnmolWminWmL blood fraction and nmolWminWmg protein, investigated with S9 fractions from COS-1 cells express- respectively. ing human carboxylesterases 1 and 2 (Fig. 4). The hydrolytic activities at 10 and 100 mM nafamostat in the cells expressing carboxylesterase 1 were 0.0671±0.0138

Thus, the Vmax WKm value was 18.0 mLWminWmL blood and 0.473±0.080 nmolWminWmg protein (mean±S.D., fraction and only 1W25 of that for the erythrocytes. n＝3), respectively. On the other hand, the corre-

With respect to human liver microsomes, the Vmax and sponding activities of the cells expressing carbox- Km values were 26.9 nmolWminWmg protein and 1790 ylesterase 2 were 0.0864±0.0163 and 0.753±0.163 mM, respectively. The nafamostat hydrolytic activities nmolWminWmg protein, respectively. The S9 fraction of at higher substrate concentrations were not determined control cells without human carboxylesterases showed with the plasma and liver microsomes because of the nafamostat hydrolytic activities with 0.0437±0.0158 low solubility of nafamostat in reaction mixtures and 0.286±0.083 nmolWminWmg protein at 10 and 100 containing these enzyme sources. mM nafamostat, respectively. The nafamostat hydrolyt- EŠects of inhibitors on nafamostat hydrolysis in ic activity of the cells expressing human carbox- human blood fractions and liver microsomes: To ylesterase 1 was signiˆcantly higher than that of the clarify human blood and liver esterases responsible for control cells only at a substrate concentration of 100 mM the nafamostat hydrolysis, the eŠects of various esterase (pº0.05). The S9 fraction of the cells expressing inhibitors on the nafamostat hydrolytic activity were carboxylesterase 2 showed signiˆcantly greater activities examined with human blood fractions (HB-1) and the at 10 and 100 mMnafamostatincomparisontothatof liver microsomes (HL-1) (Table 3). The hydrolytic the control cells (pº0.05). activity for nafamostat in human erythrocytes was Correlation between hydrolyses of nafamostat and inhibited by DTNB (arylesterase inhibitor)13) in a carboxylesterase isoform-speciˆc substrates in human concentration-dependent manner: the hydrolysis was liver microsomes: To clarify the involvement of car- completely suppressed by 1000 mM DTNB. The plasma boxylesterases 1 and 2 in the nafamostat hydrolysis in activity was also decreased by the addition of DTNB. In human liver microsomes, correlation analysis between contrast, little or no inhibition of hydrolytic activities in nafamostat hydrolytic activities and carboxylesterase the erythrocytes and plasma was seen with PMSF (serine isoform-speciˆc activities was carried out with 18 esterase inhibitor), DFP (inhibitor for cholinesterase human liver microsomes (HL-1 to -18) (Fig. 5). The and carboxylesterase), BNPP (carboxylesterase inhibi- nafamostat hydrolysis correlated highly with the tor),14) BW284C51 (acetylcholinesterase inhibitor) and hydrolytic activity for aspirin, which was a substrate of ethopropazine (butyrylcholinesterase inhibitor). The carboxylesterase 215) (r＝0.815, pº0.01). In contrast, 152 Satoshi YAMAORI et al.

Table 3. EŠects of various esterase inhibitors on nafamostat hydrolysis catalyzed by human erythrocytes, plasma and liver microsomes

Residual activity (z of control)a Inhibitor Concentration (mM) Erythrocytes Plasma Liver microsomes

PMSF 1 —b —88.5 10 — — 36.4 100 96.0 94.4 — DFP 0.25 — — 51.3 0.50 — — 33.3 100 94.4 94.7 — BW284C51 10 96.5 — 96.4 Ethopropazine 10 — 96.5 94.3 BNPP 0.25 — — 55.7 1——18.2 100 94.2 99.7 — DTNB 200 22.4 78.0 85.2 1000 0.0 16.4 100

All determinations were performed in duplicate. aControl activities (500 mM nafamostat) of human erythrocytes, plasma and liver microsomes without inhibitors were 130 and 8.45 nmolWminWmL blood fraction, and 3.87 nmolWminWmg protein, respectively. bNot analyzed.

pharmacological eŠects. However, the metabolism of nafamostat is not fully understood. At ˆrst, we focused on the involvement of blood esterases in the nafamostat hydrolysis because nafamostat was hydrolyzed rapidly after intravenous administration to humans5) and rats.4) In the present study, nafamostat was metabolized by whole blood, erythrocytes and plasma. The in vitro half-life of nafamostat disappearance in human whole blood and erythrocytes, in contrast to the plasma, was similar to the in vivo half-life in blood reported previously.5) Interestingly, kinetic analysis suggests that the hydrolytic activity for nafamostat is predominantly localized in the erythrocytes. A few examples of drugs that are mainly hydrolyzed by the erythrocytes in human blood have been previously reported.6,16–18) Fig. 4. Nafamostat hydrolysis catalyzed by human carboxylesterases 1 and 2 expressed in S9 fraction from COS-1 cells. Nafamostat (10 and Inhibitory eŠect of DTNB on the nafamostat hydrolysis 100 mM) was incubated with S9 fractions of COS-1 cells expressing in human erythrocytes suggests the involvement of human carboxylesterase 1 (closed column) and 2 (hatched column) arylesterase. This enzyme has been shown to be present isoforms, or without human carboxylesterases as a negative control in the cytosol of erythrocytes19) and hydrolyzes a small ± (open column). Individual columns and bars are the mean S.D. of number of drugs such as aspirin13) and TEI-9090 (isocar- three determinations. *Signiˆcantly diŠerent from the negative 6) control with pº0.05. bacyclin methyl ester). Other esterases localized in the erythrocytes include acetylcholinesterase (EC 3.1.1.7)20) and carboxylesterase.21) These esterases appear not to be there was no signiˆcant correlation between the involved in the nafamostat hydrolysis because PMSF, nafamostat hydrolysis and the imidapril hydrolysis DFP, BW284C51 and BNPP showed no signiˆcant inhi- catalyzed by carboxylesterase 115) (r＝0.156, p＝0.54). bition. When acetylthiocholine (1000 mM)wasusedasa substrate, the hydrolytic activity of the erythrocytes Discussion used in this study was similar to that obtained by Reiner A previous clinical study has demonstrated that et al.22) and was completely inhibited by 10 mM nafamostat is hydrolyzed to produce AN and pGBA in BW284C51 (data not shown). Thus, the blood sample humans.5) Since these metabolites have no ability to used contains a functional acetylcholinesterase. inhibit protease activities,2,3) the hydrolysis of Inhibitory eŠects of various esterase inhibitors on the nafamostat is an important determinant in‰uencing its plasma activity for nafamostat suggest that plasma Nafamostat Hydrolysis by Arylesterases and Carboxylesterases 153

Fig. 5. Relationship of nafamostat hydrolysis with imidapril and aspirin hydrolyses in human liver microsomes. Eighteen human liver microsomes were incubated with nafamostat (10 mM), imidapril (100 mM) and aspirin (100 mM). Each point indicates an average value of duplicate determinations. A correlation coe‹cient represents statistically signiˆcant linear regression of the hydrolytic activities with pº0.01.

arylesterase may be responsible for nafamostat nafamostat hydrolysis in human liver microsomes hydrolysis (Table 3). A well-characterized plasma (Table 3), although this result is based on the data arylesterase is paraoxonase.23) However, the plasma obtained from only one liver sample due to the use of esterase hydrolyzing nafamostat found in this study is limited samples. Human carboxylesterase consists of at not paraoxonase because paraoxonase requires calcium least four isoforms: carboxylesterases 1,27,28) 2,9,29,30) for catalytic activity. Smith and Cole24) previously 331,32) and 4,33,34) although Satoh and Hosokawa35) identiêd a plasma arylesterase as the enzyme hydrolyz- proposed that these isoforms were classiêd into four ing heroin. This enzyme exhibits a hydrolytic activity families. Among them, carboxylesterases 1, 2 and 4 are for heroin without calcium, in contrast to paraoxonase. expressed in the liver.34,36) However, carboxylesterase 4 This ˆnding and our result suggest that calcium- is most likely not involved in the nafamostat hydrolysis independent arylesterase(s) may be present in human in human liver microsomes, since it has been previously plasma. Another well-known plasma esterase is butyryl- reported that this isoform is less sensitive to inhibition cholinesterase (EC 3.1.1.8), which plays an important by PMSF.33,34) The enzyme assay using recombinant role in drug metabolism in the blood.25) In this study, human carboxylesterases indicated that carbox- 10 mM ethopropazine failed to inhibit nafamostat ylesterase 2 alone was capable of hydrolyzing hydrolysis catalyzed by human plasma (Table 3). In nafamostat at a lower substrate concentration (10 mM) contrast, this inhibitor at a concentration of 10 mM (Fig. 4). Furthermore, correlation analysis with human showed a complete suppression of hydrolytic activity liver microsomes showed that the nafamostat hydrolytic for butyrylthiocholine (1000 mM) in the same enzyme activity at a substrate concentration of 10 mMcorrelated source (data not shown). These results exclude the only with hydrolytic activity of aspirin, a speciˆc sub- possibility that butyrylcholinesterase contributes to the strate for carboxylesterase 2 (Fig. 5). These results nafamostat hydrolysis in human plasma. suggest that carboxylesterase 2 is mainly responsible for Next, we investigated the involvement of liver es- the nafamostat hydrolysis in human liver microsomes. terases in nafamostat hydrolysis because the elimination However, we cannot exclude the possibility that carbox- of nafamostat from the liver was faster than those from ylesterase 1 might be involved in the nafamostat other tissues in rat.4) The present study demonstrated hydrolysis catalyzed by human liver microsomes be- that human liver microsomes were capable of hydrolyz- cause the recombinant carboxylesterase 1 revealed ing nafamostat. The hydrolytic activities of 18 human signiˆcant activity at a higher substrate concentration liver microsomes varied by 3.6-fold. A previous study (100 mM) (Fig. 4). Both relative activities and expression determined that microsomal carboxylesterase activities levels of carboxylesterases 1 and 2 would determine the in 12 human livers exhibited large interindividual contribution of these isoforms to the nafamostat variations, ranging from 4.8- (acetanilide) to 44.7- hydrolysis in human liver microsomes. (p-nitrophenyl propionate) fold.26) The variability in A recent report has proposed that carboxylesterase nafamostat hydrolytic activities of human liver micro- 2 recognizes a bulky alcohol group and a relatively somes was smaller than those in the carboxylesterase smaller acyl moiety, which are held as acyl-enzyme activities reported by Hosokawa et al.26) intermediates and released as the alcohol product, Inhibition study using PMSF, DFP and BNPP sug- respectively.37) Irinotecan is an ester prodrug of 7-ethyl- gests that carboxylesterase is responsible for the 10-hydroxycamptothecin (SN-38) and 4-piperidino- 154 Satoshi YAMAORI et al. piperidine-carboxylic acid, and is e‹ciently hydrolyzed 4) Nanpo, T., Ohtsuki, T., Jin, Y., Matsunaga, K., by this isoform.15) For aspirin, hydrolysis of the acetyl Takahashi, M., Shibuya, M., Sasaki, H. and Kurumi, ester bond is catalyzed by carboxylesterase 2 to form M.: Pharmacokinetic study of FUT-175 (nafamostat salicylic acid and acetic acid.15) Furthermore, it has been mesilate) (1)-Blood level proˆles, tissue distribution, reported that the p-guanidinobenzoyl ester linkage of metabolism and excretion in rats after intravenous camostat is hydrolyzed exclusively by this isoform.15) administration. The Clinical Report, 18: 467–488 (1984). 5)Abe,T.,Kinoshita,T.,Matsuda,J.,Ogushi,T., These characteristics of carboxylesterase 2 appear to be Kawasugi, K., Yoshimura, Y. and Otomo, M.: Phase I true for nafamostat hydrolysis. study of FUT-175-Single and multiple dose study (in In this study, we failed to demonstrate the involve- Japanese). Jpn. Pharmacol. Ther., 12: 65–88 (1984). ment of erythrocyte carboxylesterase in nafamostat 6) Minagawa, T., Kohno, Y., Suwa, T. and Tsuji, A.: hydrolysis, in contrast to that of the liver microsomal Species diŠerences in hydrolysis of isocarbacyclin methyl carboxylesterases. It may be assumed that erythrocyte ester (TEI-9090) by blood esterases. Biochem. Phar- carboxylesterase has diŠerent substrate speciˆcity from macol., 49: 1361–1365 (1995). carboxylesterases expressed in human liver microsomes. 7) McCracken, N. W., Blain, P. G. and Williams, F. M.: The question of to what extent human blood and liver Human xenobiotic metabolizing esterases in liver and esterases contribute to the nafamostat hydrolysis in vivo blood. Biochem. Pharmacol., 46: 1125–1129 (1993). 8) Watanabe, K., Matsunaga, T., Yamamoto, I., Funae, is raised from our results. To our knowledge, however, Y. and Yoshimura, H.: Involvement of CYP2C in the detailed pharmacokinetic data of nafamostat in hu- metabolism of cannabinoids by human hepatic micro- mans, containing plasma protein binding and transport somes from an old woman. Biol. Pharm. Bull., 18: across erythrocyte membrane, have not been reported 1138–1141 (1995). 5) so far. Although it should be noted that nafamostat is 9) Sone, T. and Wang, C. Y.: Microsomal amidases and exposed to the blood at the ˆrst step, further studies are carboxylesterases. In Guengerich, F. P. (ed.): Compre- needed to quantify the compartment where this drug is hensive Toxicology: Biotransformation,NewYork, metabolized. Elsevier Science Inc, 1997, pp. 265–281. The present study conclusively determined that 10) Kaufman, R. J.: Identiˆcation of the components nafamostat was hydrolyzed by human blood fractions necessary for adenovirus translational control and their and liver microsomes. Our results suggest that utilization in cDNA expression vectors. Proc. Natl. Acad. Sci. USA, 82: 689–693 (1985). arylesterases and carboxylesterase 2 may be mainly 11) Luan, L., Sugiyama, T., Takai, S., Usami, Y., Adachi, responsible for nafamostat hydrolysis in human blood T., Katagiri, Y. and Hirano, K.: Puriˆcation and charac- and liver, respectively. terization of pranlukast hydrolase from rat liver microsomes: the hydrolase is identical to carboxylesterase pI Acknowledgements: We thank Tanabe Seiyaku for 6.2. Biol. Pharm. Bull., 20: 71–75 (1997). providing imidapril hydrochloride and imidaprilat. 12) Lowry, O. H., Rosebrough, N. J., Farr, A. L. and We thank Mrs. Fumiko Higashida (Dispensary, Faculty Randall, R. J.: Protein measurement with Folin-phenol of Pharmaceutical Sciences, Hokuriku University, reagent. J. Biol. Chem., 193: 265–275 (1951). Kanazawa, Japan) for blood sample collection from 13)Costello,P.B.andGreen,F.A.:Identiˆcationand subjects. partial puriˆcation of the major aspirin hydrolyzing This work was supported by `Àcademic Frontier'' enzyme in human blood. Arthritis Rheum., 26: 541–547 Project for Private Universities (2005–2009), the Special (1983). 14) Heymann, E. and Krisch, K.: Phosphoric acid-bis- Research Fund of Hokuriku University and the Torii (p-nitro-phenylester), a new inhibitor of microsomal Pharmaceutical Co., Ltd. carboxylesterases. Hoppe-Seyler's Z. Physiol. Chem., References 348: 609–619 (1967). 15)Takai,S.,Matsuda,A.,Usami,Y.,Adachi,T., 1) Fujii, S. and Hitomi, Y.: New synthetic inhibitors of Sugiyama, T., Katagiri, Y., Tatematsu, M. and Hirano, C1r, C1 esterase, thrombin, plasmin, kallikrein and K.: Hydrolytic proˆle for ester- or amide-linkage by trypsin. Biochim. Biophys. Acta, 661: 342–345 (1981). carboxylesterases pI 5.3 and 4.5 from human liver. Biol. 2) Aoyama, T., Okutome, T., Nakayama, T., Yaegashi, Pharm. Bull., 20: 869–873 (1997). T., Matsui, R., Nunomura, S., Kurumi, M., Sakurai, Y. 16) Costello,P.B.andGreen,F.A.:Aspirinsurvivalin and Fujii, S.: Synthesis and structure-activity study of human blood modulated by the concentration of protease inhibitors. IV. Amidinonaphthols and related erythrocytes. Arthritis Rheum., 25: 550–555 (1982). acyl derivatives. Chem.Pharm.Bull., 33: 1458–1471 17) Owen, J. A. and Nakatsu, K.: Diacetylmorphine (1985). (heroin) hydrolases in human blood. Can. J. Physiol. 3) Yang, H., Henkin, J., Kim, K. H. and Greer, J.: Pharmacol., 61: 870–875 (1983). Selective inhibition of urokinase by substituted phenyl- 18) Quon, C. Y. and Stamp‰i, H. F.: Biochemical properties guanidines: quantitative structure-activity relationship of blood esmolol esterase. Drug Metab. Dispos., 13: analyses. J. Med. Chem., 33: 2956–2961 (1990). 420–424 (1985). Nafamostat Hydrolysis by Arylesterases and Carboxylesterases 155

19) La Du, B. N. and Snady, H.: Esterases of human tissues. characterization of a novel putative carboxylesterase, InBrodie,B.B.andGillette,J.R.(eds.):Handbook of present in human intestine and liver. Biochem. Biophys. Experimental Pharmacology, New York, Springer- Res. Commun., 233: 117–120 (1997). Verlag, 1971, pp. 477–499. 30) Pindel, E. V., Kedishvili, N. Y., Abraham, T. L., 20) Ott, P., Jenny, B. and Brodbeck, U.: Multiple molecular Brzezinski, M. R., Zhang, J., Dean, R. A. and Bosron, forms of puriˆed human erythrocyte acetyl- W. F.: Puriˆcation and cloning of a broad substrate cholinesterase. Eur. J. Biochem., 57: 469–480 (1975). speciˆcity human liver carboxylesterase that catalyzes 21) Oertel, J., Wirthmuller, R. and Kastner, M.: Alpha the hydrolysis of cocaine and heroin. J. Biol. Chem., naphthyl acetate esterase in human blood cells with 272: 14769–14775 (1997). diŠerent molecular weights. Blut, 46: 101–106 (1983). 31) Højring, N. and Svensmark, O.: Carboxylesterases of 22) Reiner, E., Bosak, A. and Simeon-Rudolf, V.: Activity human brain extract. Puriˆcation and properties of a of cholinesterases in human whole blood measured with butyrylesterase. Biochim. Biophys. Acta, 481: 500–514 acetylthiocholine as substrate and ethopropazine as (1977). selective inhibitor of plasma butyrylcholinesterase. Arh. 32) Mori, M., Hosokawa, M., Ogasawara, Y., Tsukada, E. Hig. Rada Toksikol., 55: 1–4 (2004). and Chiba, K.: cDNA cloning, characterization and 23) Mackness, B., Durrington, P. N. and Mackness, M. I.: stable expression of novel human brain carboxylesterase. Human serum paraoxonase. Gen. Pharmacol., 31: FEBS Lett., 458: 17–22 (1999). 329–336 (1998). 33) Probst, M. R., Jeno, P. and Meyer, U. A.: Puriˆcation 24) Smith, D. A. and Cole, W. J.: Identiˆcation of an and characterization of a human liver arylacetamide arylesterase as the enzyme hydrolysing diacetylmorphine deacetylase. Biochem. Biophys. Res. Commun., 177: (heroin) in human plasma. Biochem. Pharmacol., 25: 453–459 (1991). 367–370 (1976). 34) Probst,M.R.,Beer,M.,Beer,D.,Jeno,P.,Meyer,U. 25) Williams, F. M.: Clinical signiˆcance of esterases in A. and Gasser, R.: Human liver arylacetamide deacety- man. Clin. Pharmacokinet., 10: 392–403 (1985). lase. Molecular cloning of a novel esterase involved in 26) Hosokawa, M., Endo, T., Fujisawa, M., Hara, S., the metabolic activation of arylamine carcinogens with Iwata, N., Sato, Y. and Satoh, T.: Interindividual high sequence similarity to hormone-sensitive lipase. J. variation in carboxylesterase levels in human liver Biol. Chem., 269: 21650–21656 (1994). microsomes. Drug Metab. Dispos., 23: 1022–1027 35) Satoh, T. and Hosokawa, M.: The mammalian carbox- (1995). ylesterases: from molecules to functions. Annu. Rev. 27) Long, R. M., Calabrese, M. R., Martin, B. M. and Pohl, Pharmacol. Toxicol., 38: 257–288 (1998). L. R.: Cloning and sequencing of a human liver carbox- 36)Xie,M.,Yang,D.,Liu,L.,Xue,B.andYan,B.: ylesterase isoenzyme. Life Sci., 48: PL43–49 (1991). Human and rodent carboxylesterases: immunorelated- 28) Munger, J. S., Shi, G. P., Mark, E. A., Chin, D. T., ness, overlapping substrate speciˆcity, diŠerential Gerard, C. and Chapman, H. A.: A serine esterase sensitivity to serine enzyme inhibitors, and tumor-related released by human alveolar macrophages is closely expression. Drug Metab. Dispos., 30: 541–547 (2002). related to liver microsomal carboxylesterases. J. Biol. 37) Satoh,T.,Taylor,P.,Bosron,W.F.,Sanghani,S.P., Chem., 266: 18832–18838 (1991). Hosokawa, M. and La Du, B. N.: Current progress on 29) Schwer, H., Langmann, T., Daig, R., Becker, A., esterases: from molecular structure to function. Drug Aslanidis, C. and Schmitz, G.: Molecular cloning and Metab. Dispos., 30: 488–493 (2002).