- Email: [email protected]
In vitro metabolism studies of 18F-labeled 1-phenylpiperazine using mouse liver S9 fraction
Nuclear Medicine and Biology 33 (2006) 165 – 172 www.elsevier.com/locate/nucmedbio
In vitro metabolism studies of 18F-labeled 1-phenylpiperazine using mouse liver S9 fraction Eun Kyoung Ryu, Yearn Seong Choe4, Dong Hyun Kim, Bong-Ho Ko, Yong Choi, Kyung-Han Lee, Byung-Tae Kim Department of Nuclear Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul 135-710, South Korea Received 18 May 2005; received in revised form 26 November 2005; accepted 4 December 2005
Abstract The in vitro metabolism of 1-(4-[18F]fluoromethylbenzyl)-4-phenylpiperazine ([18F]1) and 1-(4-[18F]fluorobenzyl)-4-phenylpiperazine ([ F]2) was investigated using mouse liver S9 fraction. Results were compared to those of in vivo metabolism using mouse blood and bone and to in vitro metabolism using mouse liver microsomes. Defluorination was the main metabolic pathway for [18F]1 in vitro and in vivo. Based on TLC, HPLC and LC-MS data, [18F]fluoride ion and less polar radioactive metabolites derived from aromatic ring oxidation were detected in vitro, and the latter metabolites were rapidly converted into the former with time, whereas only the [18F]fluoride ion was detected in vivo. Similarly, the in vitro metabolism of [18F]2 using either S9 fraction or microsomes showed the same pattern as the in vivo method using blood; however, the radioactive metabolites derived from aromatic ring oxidation were not detected in vivo. These results demonstrate that liver S9 fraction can be widely used to investigate the intermediate radioactive metabolites and to predict the in vivo metabolism of radiotracers. D 2006 Elsevier Inc. All rights reserved. 18
Keywords: Metabolism; In vitro method; Liver S9 fraction; Radiotracers;
18
1. Introduction When a novel radiotracer is prepared for biological evaluation, the characterization of its metabolism is a prerequisite. Although drug metabolism can occur in lungs, skin, or kidneys, it is mainly performed in the liver. In the same vein, radiopharmaceuticals are metabolized mainly in the liver [1–3]. Most radiotracers are metabolized for detoxification purposes, but some radiotracers undergo metabolic activation, and thus act as prodrugs. Metabolism studies of radiotracers have been performed in vivo using animals, whereby radiotracers are injected into animals, and samples of blood and tissue are collected and analyzed for radioactive metabolites [4–6]. This requires relatively high doses of radiotracers and involves the costs of managing animal experiment systems. Our previous in vitro metabolism studies using mouse liver microsomes demonstrated 4 Corresponding author. Tel.: +82 2 3410 2623; fax: +82 2 3410 2639. E-mail address: [email protected] (Y.S. Choe). 0969-8051/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.nucmedbio.2005.12.002
F
that this method can be used to predict the in vivo metabolism of radiotracers [7]. Liver S9 fraction has been used to assess mutagenicity in the Ames test, in which cytochrome P-450 in the S9 fraction catalyzes the metabolic activation of nonmutagenic substances [8]. As a result, S9 fraction was used to investigate the metabolic conversion of various compounds [9–11], because liver S9 fraction is not only easily obtained during the early stage of liver microsomal preparation [10–12], but also contains both microsomal and cytosolic fractions, which can provide more metabolic information than microsomes alone due to the presence of cytosolic enzymes. Therefore, metabolism by S9 fraction may be preferable to that by microsomes if the former shows a similar metabolism pattern to the latter. In this study, we investigated a simple and efficient in vitro method for performing metabolism studies using model radiotracers, [18F]1 and [18F]2, and mouse liver S9 fraction. Results obtained were compared to in vivo metabolism findings using mouse bone and blood and to in vitro metabolism findings using mouse liver microsomes.
166
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172
2. Materials and methods Chemicals were purchased from Aldrich (Milwaukee, WI, USA), and NADPH and calcium phosphate tribasic were from Sigma (St. Louis, MO, USA). 1H NMR spectra were performed on a Varian 500NB spectrometer. Chemical shifts (d) were reported in parts per million downfield from tetramethylsilane as an internal reference. Electron impact (EI) and fast atom bombardment (FAB) mass spectra were obtained on a JMS-700 Mstation (JEOL Ltd) instrument. In vitro incubations were performed at 378C using a digital block heater (Digi-Block Laboratory Devices, Holliston, MA, USA). LC-MS was performed on an 1100LC/MSD trap (Agilent Technologies, Palo Alto, CA, USA) using an HPLC column (Phenomenex Gemini C18, 5 A, 4.6250 mm). For purification and analysis of radioligand, HPLC was carried out on a Thermo Separation Products System (Fremont, CA, USA) equipped with a semipreparative column (Alltech Econosil silica gel, 10 A, 10250 mm) or an analytical column (YMC-Pack C18, 5 A, 4.6250 mm). The eluant was simultaneously monitored by a UV detector (254 nm) and a NaI(T1) radioactivity detector. Thin-layer chromatography (TLC) was performed on Merck F254 silica plates and analyzed using a Bioscan scanner (Washington, DC, USA). The radioactivity was measured in a dose calibrator (Biodex Medical Systems, Shirley, NY, USA) and tissue radioactivity in a Wallac automatic gamma counter (Perkin Elmer, Wellesly, MA, USA). All animal experiments were performed in compliance with the rules of the Samsung Medical Center Laboratory Animal Care based on the NIH guidelines. 2.1. Synthesis of nonradioactive standards 2.1.1. Synthesis of 1-(4-fluoromethylbenzyl)-4phenylpiperazine (1) 4-Fluoromethylbenzyl methanesulfonate ester (40 mg, 0.18 mmol) and Et3N (28 Al, 0.20 mmol) were added to 1phenylpiperazine (35 Al, 0.23 mmol) in CH3CN (4 ml). The reaction mixture was refluxed for 1 h. Flash column chromatography (3:1 hexane–ethyl acetate) afforded 1 as a colorless oil (32 mg, 62.5%). 1H NMR (500 MHz, CDCl3, d): 2.61 (t, J =5 Hz, 4H), 3.20 (t, J = 5 Hz, 4H), 3.58 (d, J =1.5 Hz, 2H), 5.42 (t, J = 47.5 Hz, 2H), 6.84 (m, 1H), 6.92 (dd, J =8.75, 0.5, 2H), 7.25 (m, 2H), 7.36 (m, 4H); MS m/z (FAB): 285 (M + +H); HRMS calcd for C 18 H 22 FN 2 , 285.1767; found, 285.1754. 2.1.2. Synthesis of 1-(4-fluorobenzyl)-4phenylpiperazine (2) To 4-fluorobenzaldehyde (180 Al, 1.68 mmol) in methanol (5 ml) was added 1-phenylpiperazine (254 Al, 1.66 mmol). The reaction mixture was adjusted with acetic acid to pH 5 and then reacted with NaBH3CN (285 mg, 4.54 mmol) at 808C overnight. After removal of the solvent in vacuo and basification with 10% NaOH, the resulting solution was extracted with ethyl acetate, washed with water and dried over Na2SO4. Flash column chromatography
(4:1 hexane–ethyl acetate) gave 2 as a white solid (312 mg, 69%). 1H NMR (500 MHz, CDCl3) 1H NMR (500 MHz, CDCl3, d): 2.59 (t, J = 3.75 Hz, 4H), 3.19 (t, J = 5 Hz, 4H), 3.53 (s, 2H), 6.84 (td, J =7.25, 0.5, 1H), 6.91 (dd, J = 10.5, 0.5, 2H), 7.01 (m, 2H), 7.27 (m, 4H); MS m/z (FAB): 271 (M++H); HRMS calcd for C17H20FN2 271.1611; found, 271.1612. 2.1.3. Synthesis of 1-(4-fluoromethylbenzyl)-4(4-hydroxyphenyl)piperazine (3) 4-Fluoromethylbenzyl methanesulfonate ester (20 mg, 0.09 mmol) and Et3N (15.3 Al, 0.11 mmol) were added to 1-(4-hydroxyphenyl)piperazine (32.7 mg, 0.18 mmol) in CH3CN (3 ml). The reaction mixture was refluxed for 1 h. Flash column chromatography (3:1 hexane–ethyl acetate) gave 1 as a colorless oil (23 mg, 85.1%). 1H NMR (500 MHz, CDCl3, d): 2.65 (t, J = 4.5 Hz, 4H), 3.1 (t, J = 5 Hz, 4H), 3.58 (m, 2H), 5.4 (d, J = 47.5 Hz, 2H), 6.76 (m, 2H), 6.84 (dd, J = 9.75, 2, 2H), 7.37 (m, 4H); MS m/z (EI): 300 (M+); HRMS calcd for C18H21FN2O 300.1638; found, 300.1636. 2.1.4. Synthesis of 1-(4-fluorobenzyl)-4(4-hydroxyphenyl)piperazine (4) 1-(4-Hydroxyphenyl)piperazine (398.7 mg, 2.24 mmol) was added to 4-fluorobenzaldehyde (200 Al, 1.86 mmol) in methanol (5 ml). The reaction mixture was adjusted with acetic acid to pH 5 and then reacted with NaBH3CN (288 mg, 4.58 mmol) at 808C overnight. Flash column chromatography (3:1 hexane–ethyl acetate) afforded 2 as a white solid (398 mg, 74.6%). 1H NMR (500 MHz, CDCl3, d) 2.62 (t, J =7 Hz, 4H), 3.1 (t, J = 5 Hz, 4H), 3.55 (s, 2H), 6.76 (dd, J = 7.5, 3, 2H), 6.85 (dd, J =9.3, 2, 2H), 7.03 (m, 2H), 7.33 (m, 2H); MS m/z (EI): 286 (M+); HRMS calcd for C17H19FN2O 286.1481; found, 286.1471. 2.2. Preparation of radiotracers 2.2.1. Preparation of 1-(4-[18F]fluoromethylbenzyl)4-phenylpiperazine ([18F]1) 4-[18F]Fluoromethylbenzyl methanesulfonate ester was synthesized from 1,4-benzenedimethanol bismethanesulfonate ester (1 mg, 3.4 Amol) and nBu4N[18F]F in CH3CN (200 Al) at 908C for 10 min [7]. To the resulting mixture were added 1-phenylpiperazine (2.5 Al, 16.4 Amol), Et3N (9 Al, 64.6 Amol) and CH3CN (200 Al). The reaction mixture was then heated at 1308C for 15 min and then passed through a short plug filled with 1-cm silica gel and 1-cm Na2SO4 using a 9:1 mixture of CH2Cl2 and methanol. The eluate was concentrated under a stream of N2 and purified by HPLC using a semipreparative column eluted with an 85:15 mixture of hexane and 95:5:0.1 CH2Cl2–2-propanol– NH4OH at a flow rate of 4.0 ml/min. The desired product was eluted at 17–20 min. Effective specific activity was determined by comparing the UV peak area of the desired radioactive peak and those of different concentrations of
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172 18
nonradioactive standard 1 on HPLC. An aliquot of [ F]1 was coinjected with 1 into an HPLC to confirm its identity. For specific activity determination and identification of [18F]1, the sample was analyzed by HPLC using an analytical column at a flow rate of 1.0 ml/min under the same conditions as described above. 2.2.2. Preparation of 1-(4-[18F]fluorobenzyl)-4phenylpiperazine ([18F]2) 4-[18F]Fluorobenzaldehyde was synthesized from 4trimethylammoniumbenzaldehyde trifluoromethanesulfonate (1 mg, 3.2 Amol) and nBu4N[18F]F in DMSO (300 Al) at 908C for 5 min [7,13]. 1-Phenylpiperazine (2.5 Al, 16.4 Amol), NaBH3CN (4 mg, 63.6 Amol), methanol (200 Al) and acetic acid (8 Al) were then added to the reaction mixture, which was then heated at 1308C for 15 min. At the end of the reaction, the mixture was cooled, diluted with water (2 ml) and extracted with ethyl acetate (2 ml). The organic layer was washed with water and passed through a 5-cm Na2SO4 plug, and solvents were removed under a stream of N2 at 508C (water bath). The residue was purified by HPLC using the conditions described above. The desired fraction, which eluted at 16–20 min, was collected. Effective specific activity measurement and [18F]2 identification were carried out as described above. 2.2.3. Synthesis of 4-[18F]fluorobenzoic acid 4-[18F]Fluorobenzoic acid was prepared as previously described [7]. Briefly, KMnO4 (4 mg, 25 Amol) in 0.5 ml of H2O was added to 4-[18F]fluorobenzaldehyde and heated at 908C for 10 min. The mixture was then acidified with 200 Al HCl (0.1 N) and extracted with diethyl ether (2 ml). The organic layer was passed through a small plug of Na2SO4 and used as a standard for metabolite analysis. 2.3. Preparation of mouse liver S9 fraction and microsomes Mouse liver S9 fraction and microsomes were prepared from mice (ICR, male, 25–30 g) as described in the literature [10–12]. Briefly, mouse liver tissue was cut into small pieces and washed with ice-cold 1.15% KCl. After adding three volumes of Tris acetate buffer (0.1 M, pH 7.4) containing 0.1 M KCl and 1 mM EDTA, the liver pieces were minced and homogenized on ice with a Polytron homogenizer (Kinematica, Littau-Lucerne, Switzerland). Centrifugation of the homogenate at 10,000g for 20 min at 48C gave the supernatant (S9 fraction). The supernatant was then further centrifuged at 100,000g for 60 min at 48C, and the phosphate buffer (0.1 M, pH 7.4) equal in volume to the supernatant removed was added to the precipitate. The mixture was homogenized using a glass homogenizer and centrifuged at 100,000g for 60 min at 48C. Precipitates were further homogenized using a glass homogenizer in a minimal volume of Tris acetate buffer (0.01 M, pH 7.4) containing 1 mM EDTA and 20% glycerol to yield the microsomes. S9 fraction and microsomes were stored in aliquots at 708C until use,
167
and their protein contents were determined by the method of Bradford [14]. 2.4. In vivo metabolism method Mice (ICR, male 25–30 g) were injected with radiotracers (8.89–10.4 MBq/mouse) via a tail vein and sacrificed at 1, 5, 15, 30 and 60 min postinjection. At each time point, bone sample was removed, weighed and counted. The data were expressed as the percent injected dose per gram of tissue (%ID/g). Samples of blood and liver were also collected, homogenized in 1 ml of chilled absolute ethanol and centrifuged. Metabolites in the supernatant were analyzed by TLC using a 1:1 mixture of hexane–ethyl acetate (A) and a 1:1:0.01 mixture of CH2Cl2–CH3OH– NEt3 (B). Ratios of metabolites were determined from % peak areas on TLC. After analyzing metabolites using developing solvent mixture B, the proportion of the metabolite comigrated with the parent radiotracer was calculated by subtracting the % peak area of the parent radiotracer from the total peak area. 2.5. In vitro metabolism method Radiotracers (0.95–1.6 MBq/time point) dissolved in a minimal volume of ethanol were preincubated with either mouse liver S9 fraction (0.5 mg/ml) or microsomes (0.5 mg/ml), which was rapidly thawed at room temperature before use, in phosphate buffer (0.1 M, pH 7.4) at 378C for 3 min. The reaction was initiated by adding NADPH (0.25 mM) and incubating the mixture at 378C. The total volume of the incubation mixture was adjusted with the phosphate buffer to 500 Al per time point. At the designated time points (1, 5, 15, 30 and 60 min), an aliquot (500 Al) was removed and passed through a 2-cm celite plug using 2 ml of absolute ethanol and analyzed by TLC using developing solvent mixtures A and B. A control experiment was simultaneously carried out in the absence of NADPH. Ratios of metabolites were determined as described above. 2.6. Incubation with calcium phosphate An aliquot (500 Al) was removed from the mouse liver S9 or microsomal reaction mixture after 60 min of incubation and passed through a 2-cm celite plug. The eluate was added to calcium phosphate (10 mg) in an Eppendorf tube and incubated at 378C for 15 min, centrifuged at 3000 rpm for 5 min, and the supernatant was removed. The pellet obtained was treated with water (500 Al) and centrifuged again, and the water layer combined with the supernatant and the pellet were counted. 2.7. HPLC analysis of metabolites Radiotracer was incubated with S9 fraction in the presence of NADPH at 378C for 30 min as described above. The incubation mixture was injected into an analytical HPLC column (YMC-Pack C18), and the analyses were performed using a combination of solution A, which contained 5 mM ammonium acetate in water, and
168
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172
solution B, which contained 100% acetonitrile, and flow rate was 0.5 ml/min. HPLC elution program for [18F]1 metabolites was 100% to 99% solution A and 0% to 1% solution B in a linear gradient over 0 to 10 min, 99% to 0% solution A and 1% to 100% solution B in a linear gradient over 10 to 30 min and 100% solution B over 30 to 60 min. HPLC elution program for [18F]2 metabolites was 100% to 50% solution A and 0% to 50% solution B in a linear gradient over 0 to 20 min, 50% to 10% solution A and 50% to 90% solution B in a linear gradient over 20 to 30 min and 10% solution A and 90% solution B over 30 to 60 min. The incubation mixture was coinjected with the corresponding nonradioactive standard (3 or 4), parent radiotracer ([18F]1 or [18F]2) or [18F]F into HPLC to identify the radioactive metabolites. 2.8. LC-MS analysis of metabolites 1 (54.1 Ag, 0.20 mmol) or 2 (56.9 Ag, 0.20 mmol) was dissolved in a minimal volume of ethanol, and then preincubated with S9 fraction (2 mg/ml) in phosphate buffer (0.1 M, pH 7.4) at 378C for 3 min. Incubation was carried out at 378C for 10 min, after the addition of NADPH (4 mM). The total volume of the incubation mixture adjusted with buffer was 500 Al. At the end of the incubation, the mixture was passed through a membrane filter (0.22 Am) and analyzed by LC-MS using the HPLC conditions as described above. 3. Results and discussion Radiotracers [18F]1 and [18F]2 were prepared as described in the literature (Scheme 1) [15]. [18F]1 was synthesized from the nucleophilic substitution of 1-phenylpiperazine with 4-[18F]fluoromethylbenzyl methanesulfonate ester (radiochemical yield, 25–28%; effective specific activity, 79.2 GBq/Amol). [18F]2 was prepared from the reductive alkylation of 1-phenylpiperazine with 4-[18F]fluorobenzaldehyde (radiochemical yield, 32–35%; effective specific activity, 57.9–61.1 GBq/Amol). The authenticity of the radiotracers was confirmed by coelution with the nonradioactive ligand 1 or 2 on HPLC. The metabolite
Fig. 1. TLC analyses of the metabolites of [18F]1 obtained in vivo (A) and in vitro using either S9 fraction (B) or microsomes (C). The chromatograms are presented in order of incubation time (1, 5, 15, 30 and 60 min). TLC was performed using developing solvent mixture A; Rf=0.65 ([18F]1).
standards, 3 and 4, were prepared as in the synthesis of 1 and 2. Mouse liver S9 fraction was obtained during the early stage of mouse liver microsomal preparation. [18F]1 underwent severe metabolic defluorination both in vivo and in vitro. Radioactivity uptake by bone increased as a function of time (1.4 %ID/g at 1 min, 6.8 %ID/g at 30 min and 14.0 %ID/g at 60 min), indicative of [18F]fluoride ion as a major metabolite. When blood samples were analyzed by TLC using developing solvent mixture A, a polar radioactive metabolite appeared at 1 min after injection and increased with time, as follows: 19.5% at 1 min, 32.7% at 5 min, 46.2% at 15 min, 60.7% at 30 min and 82.2% at 60 min (Fig. 1). The radioactive metabolite comigrated with authentic [18F]fluoride ion on TLC. When in vivo incubation mixture was analyzed by TLC using developing solvent
Scheme 1. Radiochemical synthesis of 1-(4-[18F]fluoromethylbenzyl)-4-phenylpiperazine ([18F]1) and 1-(4-[18F]fluorobenzyl)-4-phenylpiperazine ([18F]2).
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172
169
Scheme 2. Proposed metabolic pathways of [18F]1.
mixture B, the [18F]fluoride ion was the only radioactive metabolite found (Fig. 3). The [18F]fluoride ion was also detected, but its production (b 10% of the total peak at 60 min) was extremely slow in liver homogenates, probably due to the low concentration of microsomal enzymes. In vitro methods using either S9 fraction or microsomes showed that [18F]1 is metabolized to a polar radioactive metabolite, like in vivo method. The percentages of radioactive metabolites formed from [18F]1 by S9 fraction and by microsomes based on TLC analysis using developing solvent mixture A were 78.1% and 79.6% at 1 min, 93.4% and 97.2% at 5 min, 97.2% and 100% at 15 min, 98.1% and 100% at 30 min and 100% and 100% at 60 min (Fig. 1). The radioactive metabolite was confirmed as the [18F]fluoride ion based on significant uptake of radioactivity by calcium phosphate (1:7 supernatant–calcium phosphate pellet). Adsorption of the [18F]fluoride ion by calcium phosphate as a bone substitute was found to be a useful method to detect the defluorination of 18F-labeled compounds [7]. When the incubation mixture using S9 fraction was analyzed by TLC using the developing solvent mixture B, two radioactive metabolites, [18F]fluoride ion (R f = 0) and a less polar compound (R f = 0.70), appeared with % peak area ratios of 57:34 at 5 min, 63:34 at 30 min and 71:29 at 60 min. In the case of the microsomal incubation mixture, the same radioactive metabolites appeared at ratios of 72:22 at 5 min, 85:15 at 30 min and 89:11 at 60 min (Fig. 3). This TLC result indicates that the less polar peak was rapidly converted to the [18F]fluoride ion with time. HPLC analysis of the incubation mixture of [18F]1 with S9 fraction showed three polar metabolite peaks and the parent peak, the most polar peak at the retention time of 7.9 min being coeluted with [18F]F , the peak at 26.1 min being unidentified and the peak at 34.3 min being coeluted with 3 (Fig. 4A and B). The latter two peaks gave a major fragment at m/z 301.3
(M++H) corresponding to the products of aromatic ring oxidation and were comigrated on TLC (R f = 0.70) (Fig. 3). Aromatic ring oxidation was reported as the major pathway of 5-HT1A antagonists, N-{2-[4-(2-methoxyphenyl)piperazino]ethyl}-N-(2-pyridyl) trans- and cis-4-fluorocyclohexanecarboxamide fluorocyclohexanecarboxamide in rat [16]. These results suggest that the in vitro metabolism of [18F]1 proceeded via both pathways a and b or pathway b (Scheme 2). On the other hand, the blood sample produced
Fig. 2. TLC analyses of the metabolites of [18F]2 obtained in vivo (A) and in vitro using either S9 fraction (B) or microsomes (C). The chromatograms are presented in incubation time order (1, 5, 15, 30 and 60 min). TLC was performed using developing solvent mixture A; Rf=0.67 ([18F]2).
170
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172 18
the [ F]fluoride ion as the only radioactive metabolite under the same condition. This may be because in the in vivo system, pathway b proceeds too rapidly for the intermediate radioactive metabolites to be detected or that pathway a is preferred (Scheme 2). The formation of [18F]fluoride ion was not surprising because of the presence of a reactive benzylic carbon in [18F]1, as shown in Scheme 2 [7,17], though a metabolically stable benzyl fluoride substituted with a chlorine atom meta to the benzylic position has been reported [18]. [18F]2 underwent severe metabolic transformation both in vivo and in vitro. A constant level of radioactivity was taken up by bone as a function of time (2.6 %ID/g at 1 min, 2.5 %ID/g at 30 min and 2.0 %ID/g at 60 min), indicative of little defluorination. TLC analysis of the blood sample using developing solvent mixture A showed a polar radioactive metabolite that appeared from 1 min postinjection and then increased with time, 27.2% at 1 min, 63.8% at 5 min, 91.8% at 15 min, 95.1% at 30 min and 100% at 60 min (Fig. 2). The same metabolite was detected in liver homogenates, but its formation was slower with less than 40% of the total peak at 60 min. The 30-min blood sample was reanalyzed using TLC developing solvent mixture B, and a radioactive metabolite peak was separated into two peaks with R f ’s of 0 and 0.2 (18:82) (Fig. 3). The generation of polar radioactive metabolites by either S9 fraction or by microsomes increased with time; 2.8% and 19.9% at 1 min, 22.2% and 62.1% at 5 min, 57.6% and 90.6% at 15 min, 79.1% and 100% at 30 min and 96.3% and 100% at 60 min (Fig. 2). Although the radioactive metabolite was not identified, it was separated into three peaks (R f = 0, 0.23 and 0.97) by TLC using developing solvent mixture B. The peak at R f of 0.97 was comigrated with the parent peak. In S9 fraction, the ratios of these three radioactive metabolites were 4:8:12 at 5 min, 14:29:37 at 30 min and 24:60:8 at 60 min (Fig. 3). In microsomal incubation mixtures, these ratios were 5:33:29 at 5 min, 28:49:23 at 30 min and 19:70:11 at 60 min (Fig. 3). The polar two peaks were not [18F]fluoride ion or 4-[18F]fluorobenzoic acid; the
Fig. 4. HPLC analyses of [18F]1 and [18F]2 metabolites obtained in vitro using S9 fraction at 30 min after incubation: (A) Radioactive metabolites of [18F]1; t R=7.9 min ([18F]F-), t R=26.1 min (unidentified), t R=34.3 min (coeluted with 3) and t R=40.3 min ([18F]1). (B) Nonradioactive standard 3. (C) Radioactive metabolites of [18F]2; t R=11.2, 24.3 and 30.7 min (unidentified), t R=35.1 min (coeluted with 4) and t R=42.5 min ([18F]2). (D) Nonradioactive standard 4.
latter was proposed to be derived from 4-[18F]fluorobenzaldehyde formed via N-dealkylation of [18F]2 [4,5]. [18F]Fluoride ion was confirmed to be absent by low-level calcium
Fig. 3. TLC analyses of [18F]1 and [18F]2 metabolites produced in vivo (A) or in vitro using either S9 fraction (B) or microsomes (C) at 30 min after injection or incubation. TLC was performed using developing solvent mixture B; Rf=0.93 ([18F]1) and Rf=0.97 ([18F]2).
Scheme 3. A major metabolic pathway of [18F]2.
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172
phosphate uptake (7:1 supernatant–calcium phosphate pellet), and 4-[18F]fluorobenzoic acid was confirmed absent by the different migration rates of the radioactive metabolites (R f = 0 and 0.1) and authentic 4-[18F]fluorobenzoic acid (R f = 0.3) on TLC using a 14:4:0.01 mixture of CH2Cl2CH3OH-NH4OH as a developing solvent mixture. The polar metabolites in either S9 or microsomes are not the product of glucuronide conjugation because the transformation requires a cofactor such as uridine 5V-diphosphoglucuronic acid [10,16]. HPLC analysis of the incubation mixture of [18F]2 with S9 fraction showed four polar metabolite peaks and the parent peak, the least polar peak at the retention time of 35.1 min being coeluted with 4, the other peaks at 11.2, 24.3 and 30.7 min being unidentified (Fig. 4C and D). All three peaks except the most polar peak gave a major fragment at m/z 287.3 (M++H) corresponding to the products of aromatic ring oxidation and were comigrated on TLC (R f = 0.97) when using developing solvent mixture B (Fig. 3) (Scheme 3). The rate of radioactive metabolite formation from [18F]1 was highest for microsomes and lowest for blood samples, whereas that of [18F]2 was highest in blood samples and lowest in S9 fraction. In the absence of NADPH, no radioactive metabolites were detected from either [18F]1 or [18F]2, indicating that cytochrome P-450 could be involved in the biotransformation. The in vivo method required high doses (N 8.89 MBq/ mouse) of radiotracer, whereas low doses ( N0.95 MBq/time point) were sufficient for the in vitro method. 4. Conclusion This study shows that the in vitro metabolism of [18F]1 using either mouse liver S9 fraction or microsomes produces the same results as the in vivo method using blood or bone. However, the intermediate radioactive metabolite derived from aromatic ring oxidation, which is believed to be converted into [18F]fluoride ion, was not observed in vivo (blood) by TLC analysis using developing solvent mixture B. By the same token, the in vitro metabolism of [18F]2 using S9 fraction or microsomes showed the same pattern of metabolite production as the in vivo blood or bone-based methods, although the radioactive metabolite derived from aromatic ring oxidation was not detected in vivo. Although the metabolism rates differed because of different concentrations of metabolizing enzymes, the results demonstrate that liver S9 fraction can be used instead of microsomes for the metabolism studies of radiotracers. Moreover, the liver S9 fraction is easily obtained during the early stage of liver microsomal preparation and contains both microsomal and cytosolic fractions. In addition, the in vitro metabolism method using either S9 fraction or microsomes produced intermediate radioactive metabolites not detected in vivo, which would provide valuable information on the in vivo metabolism pathways of radiotracers. Further studies using
171
various radiotracers are warranted to demonstrate the applicability of liver S9 fraction for the metabolism studies of radiotracers.
Acknowledgment This work was supported in part by the Samsung Biomedical Research Institute grant, #SBRI C-A5-124-3, and by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD) (R04-2002-000-20067-0).
References [1] Delaforge M. Importance of metabolism in pharmacological studies: possible in vitro predictability. Nucl Med Biol 1998;25:705 – 9. [2] Mazel P. General principles and procedures for drug metabolism in vitro. In: La Du BN, Mandel HG, Way EL, editors. Fundamentals of drug metabolism and drug deposition. Baltimore7 The Williams & Wilkins Company; 1971. p. 527 – 45. [3] Mazie`re B, Cantineau R, Coenen HH, Guillaume M, Halldin C, Luxen A, et al. PET radiopharmaceutical metabolism — plasma metabolite analysis. In: Stfcklin G, Pike VW, editors. Radiopharmaceuticals for positron emission tomography. Methodological aspects. Dordrecht7 Kluwer Academic Publishers; 1993. p. 151 – 78. [4] Mach RH, Nowak PA, Moroney DA, Morton TE, Jackson JR, Ehrenkaufer RLE. Synthesis and in vivo evaluation of [18F]-N-( pnitrobenzyl)spiperone ([18F]PNBS) in rats. Nucl Med Biol 1993; 20:269 – 78. [5] Welch MJ, Katzenellenbogen JA, Mathias CJ, Brodack JW, Carlson KE, Chi DY, et al. N-(3-[18F]Fluoropropyl)-spiperone: the preferred 18 F labeled spiperone analog for positron emission tomographic studies of the dopamine receptor. Nucl Med Biol 1988;15:83 – 97. [6] Tan P-Z, Baldwin RM, Van Dyck CH, Al-Tikriti M, Roth B, Khan N, et al. Characterization of radioactive metabolites of 5-HT2A receptor PET ligand [18F]altanserin in human and rodent. Nucl Med Biol 1999;26:601 – 8. [7] Lee S-Y, Choe YS, Kim DH, Park B-N, Kim SE, Choi Y, et al. A simple and efficient in vitro method for metabolism studies of radiotracers. Nucl Med Biol 2001;28:391 – 5. [8] Ames BN, McCann J, Yamasaki E. Methods for detecting carcinogens and mutagens with the Salmonella/mammalian-microsome mutagenicity test. Mutat Res 1975;31:347 – 64. [9] Liu K, Guengerich P, Yang SK. Enantioselective hydrolysis of lorazepam 3-acetate by esterases in human and rat liver microsome and rat brain S9 fraction. Drug Metab Dispos 1991;19:609 – 13. [10] Li AP, Hartman NR, Lu C, Collins JM, Strong JM. Effects of cytochrome P450 inducers on 17a-ethynyloestradiol (EE2) conjugation by primary human hepatocytes. Br J Clin Pharmacol 1999;48: 733 – 42. [11] Yoshihara S, Makishima M, Suzuki N, Ohta S. Metabolic activation of bisphenol A by rat liver S9 fraction. Toxicol Sci 2001;62:221 – 7. [12] Guengerich FP, Martin MV, Beaune PH, Kremers P, Wolff T, Waxman DJ. Characterization of rat and human liver microsomal cytochrome P-450 forms involved in nifedipine oxidation, a prototype for genetic polymorphism in oxidative drug metabolism. J Biol Chem 1986;261: 5051 – 60. [13] Haka MS, Kilbourn MR, Watkins GL, Toorongian SA. Aryltrimethylammonium trifluoromethanesulfonates as precursors to aryl [18F]fluoride: improved synthesis of [18F]GBR-13119. J Labelled Compd Radiopharm 1989;27:823 – 33. [14] Bradford M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of dye-binding. Anal Biochem 1976;72:248 – 54.
172
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172
[15] Choe YS, Oh SJ, Shim I, Naruto S, Chi DY, Kim SE, et al. Syntheses and biological evaluation of 18F-labeled 3-(1-benzyl-piperidin-4-yl)-1(1-methyl-1H-indol-3-yl)propan-1-ones for in vivo mapping of acetylcholinesterase. Nucl Med Biol 2000;27:263 – 7. [16] Ma Y, Lang L, Kiesewetter DO, Jagoda E, Sassaman MB, Der M, et al. Liquid chromatography-tandem mass spectrometry identification of metabolites of two 5-HT1A antagonists, N-{2-[4-(2-methoxyphenyl)piperazino]ethyl}-N-(2-pyridyl) trans- and cis-4-fluorocyclohex-
anecarboxamide, produced by human and rat hepatocytes. J Chromatogr B 2001;755:47 – 56. [17] Lee KC, Lee S-Y, Choe YS, Chi DY. Metabolic stability of [18F]fluoroalkylbiphenyls. Bull Korean Chem Soc 2004;25:1225 – 30. [18] Magata Y, Lang L, Kiesewetter DO, Jagoda EM, Channing MA, Eckelman WC. Biologically stable [18F]-labeled benzylfluoride derivatives. Nucl Med Biol 2000;27:163 – 8.
In vitro metabolism studies of 18F-labeled 1-phenylpiperazine using mouse liver S9 fraction Eun Kyoung Ryu, Yearn Seong Choe4, Dong Hyun Kim, Bong-Ho Ko, Yong Choi, Kyung-Han Lee, Byung-Tae Kim Department of Nuclear Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul 135-710, South Korea Received 18 May 2005; received in revised form 26 November 2005; accepted 4 December 2005
Abstract The in vitro metabolism of 1-(4-[18F]fluoromethylbenzyl)-4-phenylpiperazine ([18F]1) and 1-(4-[18F]fluorobenzyl)-4-phenylpiperazine ([ F]2) was investigated using mouse liver S9 fraction. Results were compared to those of in vivo metabolism using mouse blood and bone and to in vitro metabolism using mouse liver microsomes. Defluorination was the main metabolic pathway for [18F]1 in vitro and in vivo. Based on TLC, HPLC and LC-MS data, [18F]fluoride ion and less polar radioactive metabolites derived from aromatic ring oxidation were detected in vitro, and the latter metabolites were rapidly converted into the former with time, whereas only the [18F]fluoride ion was detected in vivo. Similarly, the in vitro metabolism of [18F]2 using either S9 fraction or microsomes showed the same pattern as the in vivo method using blood; however, the radioactive metabolites derived from aromatic ring oxidation were not detected in vivo. These results demonstrate that liver S9 fraction can be widely used to investigate the intermediate radioactive metabolites and to predict the in vivo metabolism of radiotracers. D 2006 Elsevier Inc. All rights reserved. 18
Keywords: Metabolism; In vitro method; Liver S9 fraction; Radiotracers;
18
1. Introduction When a novel radiotracer is prepared for biological evaluation, the characterization of its metabolism is a prerequisite. Although drug metabolism can occur in lungs, skin, or kidneys, it is mainly performed in the liver. In the same vein, radiopharmaceuticals are metabolized mainly in the liver [1–3]. Most radiotracers are metabolized for detoxification purposes, but some radiotracers undergo metabolic activation, and thus act as prodrugs. Metabolism studies of radiotracers have been performed in vivo using animals, whereby radiotracers are injected into animals, and samples of blood and tissue are collected and analyzed for radioactive metabolites [4–6]. This requires relatively high doses of radiotracers and involves the costs of managing animal experiment systems. Our previous in vitro metabolism studies using mouse liver microsomes demonstrated 4 Corresponding author. Tel.: +82 2 3410 2623; fax: +82 2 3410 2639. E-mail address: [email protected] (Y.S. Choe). 0969-8051/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.nucmedbio.2005.12.002
F
that this method can be used to predict the in vivo metabolism of radiotracers [7]. Liver S9 fraction has been used to assess mutagenicity in the Ames test, in which cytochrome P-450 in the S9 fraction catalyzes the metabolic activation of nonmutagenic substances [8]. As a result, S9 fraction was used to investigate the metabolic conversion of various compounds [9–11], because liver S9 fraction is not only easily obtained during the early stage of liver microsomal preparation [10–12], but also contains both microsomal and cytosolic fractions, which can provide more metabolic information than microsomes alone due to the presence of cytosolic enzymes. Therefore, metabolism by S9 fraction may be preferable to that by microsomes if the former shows a similar metabolism pattern to the latter. In this study, we investigated a simple and efficient in vitro method for performing metabolism studies using model radiotracers, [18F]1 and [18F]2, and mouse liver S9 fraction. Results obtained were compared to in vivo metabolism findings using mouse bone and blood and to in vitro metabolism findings using mouse liver microsomes.
166
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172
2. Materials and methods Chemicals were purchased from Aldrich (Milwaukee, WI, USA), and NADPH and calcium phosphate tribasic were from Sigma (St. Louis, MO, USA). 1H NMR spectra were performed on a Varian 500NB spectrometer. Chemical shifts (d) were reported in parts per million downfield from tetramethylsilane as an internal reference. Electron impact (EI) and fast atom bombardment (FAB) mass spectra were obtained on a JMS-700 Mstation (JEOL Ltd) instrument. In vitro incubations were performed at 378C using a digital block heater (Digi-Block Laboratory Devices, Holliston, MA, USA). LC-MS was performed on an 1100LC/MSD trap (Agilent Technologies, Palo Alto, CA, USA) using an HPLC column (Phenomenex Gemini C18, 5 A, 4.6250 mm). For purification and analysis of radioligand, HPLC was carried out on a Thermo Separation Products System (Fremont, CA, USA) equipped with a semipreparative column (Alltech Econosil silica gel, 10 A, 10250 mm) or an analytical column (YMC-Pack C18, 5 A, 4.6250 mm). The eluant was simultaneously monitored by a UV detector (254 nm) and a NaI(T1) radioactivity detector. Thin-layer chromatography (TLC) was performed on Merck F254 silica plates and analyzed using a Bioscan scanner (Washington, DC, USA). The radioactivity was measured in a dose calibrator (Biodex Medical Systems, Shirley, NY, USA) and tissue radioactivity in a Wallac automatic gamma counter (Perkin Elmer, Wellesly, MA, USA). All animal experiments were performed in compliance with the rules of the Samsung Medical Center Laboratory Animal Care based on the NIH guidelines. 2.1. Synthesis of nonradioactive standards 2.1.1. Synthesis of 1-(4-fluoromethylbenzyl)-4phenylpiperazine (1) 4-Fluoromethylbenzyl methanesulfonate ester (40 mg, 0.18 mmol) and Et3N (28 Al, 0.20 mmol) were added to 1phenylpiperazine (35 Al, 0.23 mmol) in CH3CN (4 ml). The reaction mixture was refluxed for 1 h. Flash column chromatography (3:1 hexane–ethyl acetate) afforded 1 as a colorless oil (32 mg, 62.5%). 1H NMR (500 MHz, CDCl3, d): 2.61 (t, J =5 Hz, 4H), 3.20 (t, J = 5 Hz, 4H), 3.58 (d, J =1.5 Hz, 2H), 5.42 (t, J = 47.5 Hz, 2H), 6.84 (m, 1H), 6.92 (dd, J =8.75, 0.5, 2H), 7.25 (m, 2H), 7.36 (m, 4H); MS m/z (FAB): 285 (M + +H); HRMS calcd for C 18 H 22 FN 2 , 285.1767; found, 285.1754. 2.1.2. Synthesis of 1-(4-fluorobenzyl)-4phenylpiperazine (2) To 4-fluorobenzaldehyde (180 Al, 1.68 mmol) in methanol (5 ml) was added 1-phenylpiperazine (254 Al, 1.66 mmol). The reaction mixture was adjusted with acetic acid to pH 5 and then reacted with NaBH3CN (285 mg, 4.54 mmol) at 808C overnight. After removal of the solvent in vacuo and basification with 10% NaOH, the resulting solution was extracted with ethyl acetate, washed with water and dried over Na2SO4. Flash column chromatography
(4:1 hexane–ethyl acetate) gave 2 as a white solid (312 mg, 69%). 1H NMR (500 MHz, CDCl3) 1H NMR (500 MHz, CDCl3, d): 2.59 (t, J = 3.75 Hz, 4H), 3.19 (t, J = 5 Hz, 4H), 3.53 (s, 2H), 6.84 (td, J =7.25, 0.5, 1H), 6.91 (dd, J = 10.5, 0.5, 2H), 7.01 (m, 2H), 7.27 (m, 4H); MS m/z (FAB): 271 (M++H); HRMS calcd for C17H20FN2 271.1611; found, 271.1612. 2.1.3. Synthesis of 1-(4-fluoromethylbenzyl)-4(4-hydroxyphenyl)piperazine (3) 4-Fluoromethylbenzyl methanesulfonate ester (20 mg, 0.09 mmol) and Et3N (15.3 Al, 0.11 mmol) were added to 1-(4-hydroxyphenyl)piperazine (32.7 mg, 0.18 mmol) in CH3CN (3 ml). The reaction mixture was refluxed for 1 h. Flash column chromatography (3:1 hexane–ethyl acetate) gave 1 as a colorless oil (23 mg, 85.1%). 1H NMR (500 MHz, CDCl3, d): 2.65 (t, J = 4.5 Hz, 4H), 3.1 (t, J = 5 Hz, 4H), 3.58 (m, 2H), 5.4 (d, J = 47.5 Hz, 2H), 6.76 (m, 2H), 6.84 (dd, J = 9.75, 2, 2H), 7.37 (m, 4H); MS m/z (EI): 300 (M+); HRMS calcd for C18H21FN2O 300.1638; found, 300.1636. 2.1.4. Synthesis of 1-(4-fluorobenzyl)-4(4-hydroxyphenyl)piperazine (4) 1-(4-Hydroxyphenyl)piperazine (398.7 mg, 2.24 mmol) was added to 4-fluorobenzaldehyde (200 Al, 1.86 mmol) in methanol (5 ml). The reaction mixture was adjusted with acetic acid to pH 5 and then reacted with NaBH3CN (288 mg, 4.58 mmol) at 808C overnight. Flash column chromatography (3:1 hexane–ethyl acetate) afforded 2 as a white solid (398 mg, 74.6%). 1H NMR (500 MHz, CDCl3, d) 2.62 (t, J =7 Hz, 4H), 3.1 (t, J = 5 Hz, 4H), 3.55 (s, 2H), 6.76 (dd, J = 7.5, 3, 2H), 6.85 (dd, J =9.3, 2, 2H), 7.03 (m, 2H), 7.33 (m, 2H); MS m/z (EI): 286 (M+); HRMS calcd for C17H19FN2O 286.1481; found, 286.1471. 2.2. Preparation of radiotracers 2.2.1. Preparation of 1-(4-[18F]fluoromethylbenzyl)4-phenylpiperazine ([18F]1) 4-[18F]Fluoromethylbenzyl methanesulfonate ester was synthesized from 1,4-benzenedimethanol bismethanesulfonate ester (1 mg, 3.4 Amol) and nBu4N[18F]F in CH3CN (200 Al) at 908C for 10 min [7]. To the resulting mixture were added 1-phenylpiperazine (2.5 Al, 16.4 Amol), Et3N (9 Al, 64.6 Amol) and CH3CN (200 Al). The reaction mixture was then heated at 1308C for 15 min and then passed through a short plug filled with 1-cm silica gel and 1-cm Na2SO4 using a 9:1 mixture of CH2Cl2 and methanol. The eluate was concentrated under a stream of N2 and purified by HPLC using a semipreparative column eluted with an 85:15 mixture of hexane and 95:5:0.1 CH2Cl2–2-propanol– NH4OH at a flow rate of 4.0 ml/min. The desired product was eluted at 17–20 min. Effective specific activity was determined by comparing the UV peak area of the desired radioactive peak and those of different concentrations of
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172 18
nonradioactive standard 1 on HPLC. An aliquot of [ F]1 was coinjected with 1 into an HPLC to confirm its identity. For specific activity determination and identification of [18F]1, the sample was analyzed by HPLC using an analytical column at a flow rate of 1.0 ml/min under the same conditions as described above. 2.2.2. Preparation of 1-(4-[18F]fluorobenzyl)-4phenylpiperazine ([18F]2) 4-[18F]Fluorobenzaldehyde was synthesized from 4trimethylammoniumbenzaldehyde trifluoromethanesulfonate (1 mg, 3.2 Amol) and nBu4N[18F]F in DMSO (300 Al) at 908C for 5 min [7,13]. 1-Phenylpiperazine (2.5 Al, 16.4 Amol), NaBH3CN (4 mg, 63.6 Amol), methanol (200 Al) and acetic acid (8 Al) were then added to the reaction mixture, which was then heated at 1308C for 15 min. At the end of the reaction, the mixture was cooled, diluted with water (2 ml) and extracted with ethyl acetate (2 ml). The organic layer was washed with water and passed through a 5-cm Na2SO4 plug, and solvents were removed under a stream of N2 at 508C (water bath). The residue was purified by HPLC using the conditions described above. The desired fraction, which eluted at 16–20 min, was collected. Effective specific activity measurement and [18F]2 identification were carried out as described above. 2.2.3. Synthesis of 4-[18F]fluorobenzoic acid 4-[18F]Fluorobenzoic acid was prepared as previously described [7]. Briefly, KMnO4 (4 mg, 25 Amol) in 0.5 ml of H2O was added to 4-[18F]fluorobenzaldehyde and heated at 908C for 10 min. The mixture was then acidified with 200 Al HCl (0.1 N) and extracted with diethyl ether (2 ml). The organic layer was passed through a small plug of Na2SO4 and used as a standard for metabolite analysis. 2.3. Preparation of mouse liver S9 fraction and microsomes Mouse liver S9 fraction and microsomes were prepared from mice (ICR, male, 25–30 g) as described in the literature [10–12]. Briefly, mouse liver tissue was cut into small pieces and washed with ice-cold 1.15% KCl. After adding three volumes of Tris acetate buffer (0.1 M, pH 7.4) containing 0.1 M KCl and 1 mM EDTA, the liver pieces were minced and homogenized on ice with a Polytron homogenizer (Kinematica, Littau-Lucerne, Switzerland). Centrifugation of the homogenate at 10,000g for 20 min at 48C gave the supernatant (S9 fraction). The supernatant was then further centrifuged at 100,000g for 60 min at 48C, and the phosphate buffer (0.1 M, pH 7.4) equal in volume to the supernatant removed was added to the precipitate. The mixture was homogenized using a glass homogenizer and centrifuged at 100,000g for 60 min at 48C. Precipitates were further homogenized using a glass homogenizer in a minimal volume of Tris acetate buffer (0.01 M, pH 7.4) containing 1 mM EDTA and 20% glycerol to yield the microsomes. S9 fraction and microsomes were stored in aliquots at 708C until use,
167
and their protein contents were determined by the method of Bradford [14]. 2.4. In vivo metabolism method Mice (ICR, male 25–30 g) were injected with radiotracers (8.89–10.4 MBq/mouse) via a tail vein and sacrificed at 1, 5, 15, 30 and 60 min postinjection. At each time point, bone sample was removed, weighed and counted. The data were expressed as the percent injected dose per gram of tissue (%ID/g). Samples of blood and liver were also collected, homogenized in 1 ml of chilled absolute ethanol and centrifuged. Metabolites in the supernatant were analyzed by TLC using a 1:1 mixture of hexane–ethyl acetate (A) and a 1:1:0.01 mixture of CH2Cl2–CH3OH– NEt3 (B). Ratios of metabolites were determined from % peak areas on TLC. After analyzing metabolites using developing solvent mixture B, the proportion of the metabolite comigrated with the parent radiotracer was calculated by subtracting the % peak area of the parent radiotracer from the total peak area. 2.5. In vitro metabolism method Radiotracers (0.95–1.6 MBq/time point) dissolved in a minimal volume of ethanol were preincubated with either mouse liver S9 fraction (0.5 mg/ml) or microsomes (0.5 mg/ml), which was rapidly thawed at room temperature before use, in phosphate buffer (0.1 M, pH 7.4) at 378C for 3 min. The reaction was initiated by adding NADPH (0.25 mM) and incubating the mixture at 378C. The total volume of the incubation mixture was adjusted with the phosphate buffer to 500 Al per time point. At the designated time points (1, 5, 15, 30 and 60 min), an aliquot (500 Al) was removed and passed through a 2-cm celite plug using 2 ml of absolute ethanol and analyzed by TLC using developing solvent mixtures A and B. A control experiment was simultaneously carried out in the absence of NADPH. Ratios of metabolites were determined as described above. 2.6. Incubation with calcium phosphate An aliquot (500 Al) was removed from the mouse liver S9 or microsomal reaction mixture after 60 min of incubation and passed through a 2-cm celite plug. The eluate was added to calcium phosphate (10 mg) in an Eppendorf tube and incubated at 378C for 15 min, centrifuged at 3000 rpm for 5 min, and the supernatant was removed. The pellet obtained was treated with water (500 Al) and centrifuged again, and the water layer combined with the supernatant and the pellet were counted. 2.7. HPLC analysis of metabolites Radiotracer was incubated with S9 fraction in the presence of NADPH at 378C for 30 min as described above. The incubation mixture was injected into an analytical HPLC column (YMC-Pack C18), and the analyses were performed using a combination of solution A, which contained 5 mM ammonium acetate in water, and
168
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172
solution B, which contained 100% acetonitrile, and flow rate was 0.5 ml/min. HPLC elution program for [18F]1 metabolites was 100% to 99% solution A and 0% to 1% solution B in a linear gradient over 0 to 10 min, 99% to 0% solution A and 1% to 100% solution B in a linear gradient over 10 to 30 min and 100% solution B over 30 to 60 min. HPLC elution program for [18F]2 metabolites was 100% to 50% solution A and 0% to 50% solution B in a linear gradient over 0 to 20 min, 50% to 10% solution A and 50% to 90% solution B in a linear gradient over 20 to 30 min and 10% solution A and 90% solution B over 30 to 60 min. The incubation mixture was coinjected with the corresponding nonradioactive standard (3 or 4), parent radiotracer ([18F]1 or [18F]2) or [18F]F into HPLC to identify the radioactive metabolites. 2.8. LC-MS analysis of metabolites 1 (54.1 Ag, 0.20 mmol) or 2 (56.9 Ag, 0.20 mmol) was dissolved in a minimal volume of ethanol, and then preincubated with S9 fraction (2 mg/ml) in phosphate buffer (0.1 M, pH 7.4) at 378C for 3 min. Incubation was carried out at 378C for 10 min, after the addition of NADPH (4 mM). The total volume of the incubation mixture adjusted with buffer was 500 Al. At the end of the incubation, the mixture was passed through a membrane filter (0.22 Am) and analyzed by LC-MS using the HPLC conditions as described above. 3. Results and discussion Radiotracers [18F]1 and [18F]2 were prepared as described in the literature (Scheme 1) [15]. [18F]1 was synthesized from the nucleophilic substitution of 1-phenylpiperazine with 4-[18F]fluoromethylbenzyl methanesulfonate ester (radiochemical yield, 25–28%; effective specific activity, 79.2 GBq/Amol). [18F]2 was prepared from the reductive alkylation of 1-phenylpiperazine with 4-[18F]fluorobenzaldehyde (radiochemical yield, 32–35%; effective specific activity, 57.9–61.1 GBq/Amol). The authenticity of the radiotracers was confirmed by coelution with the nonradioactive ligand 1 or 2 on HPLC. The metabolite
Fig. 1. TLC analyses of the metabolites of [18F]1 obtained in vivo (A) and in vitro using either S9 fraction (B) or microsomes (C). The chromatograms are presented in order of incubation time (1, 5, 15, 30 and 60 min). TLC was performed using developing solvent mixture A; Rf=0.65 ([18F]1).
standards, 3 and 4, were prepared as in the synthesis of 1 and 2. Mouse liver S9 fraction was obtained during the early stage of mouse liver microsomal preparation. [18F]1 underwent severe metabolic defluorination both in vivo and in vitro. Radioactivity uptake by bone increased as a function of time (1.4 %ID/g at 1 min, 6.8 %ID/g at 30 min and 14.0 %ID/g at 60 min), indicative of [18F]fluoride ion as a major metabolite. When blood samples were analyzed by TLC using developing solvent mixture A, a polar radioactive metabolite appeared at 1 min after injection and increased with time, as follows: 19.5% at 1 min, 32.7% at 5 min, 46.2% at 15 min, 60.7% at 30 min and 82.2% at 60 min (Fig. 1). The radioactive metabolite comigrated with authentic [18F]fluoride ion on TLC. When in vivo incubation mixture was analyzed by TLC using developing solvent
Scheme 1. Radiochemical synthesis of 1-(4-[18F]fluoromethylbenzyl)-4-phenylpiperazine ([18F]1) and 1-(4-[18F]fluorobenzyl)-4-phenylpiperazine ([18F]2).
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172
169
Scheme 2. Proposed metabolic pathways of [18F]1.
mixture B, the [18F]fluoride ion was the only radioactive metabolite found (Fig. 3). The [18F]fluoride ion was also detected, but its production (b 10% of the total peak at 60 min) was extremely slow in liver homogenates, probably due to the low concentration of microsomal enzymes. In vitro methods using either S9 fraction or microsomes showed that [18F]1 is metabolized to a polar radioactive metabolite, like in vivo method. The percentages of radioactive metabolites formed from [18F]1 by S9 fraction and by microsomes based on TLC analysis using developing solvent mixture A were 78.1% and 79.6% at 1 min, 93.4% and 97.2% at 5 min, 97.2% and 100% at 15 min, 98.1% and 100% at 30 min and 100% and 100% at 60 min (Fig. 1). The radioactive metabolite was confirmed as the [18F]fluoride ion based on significant uptake of radioactivity by calcium phosphate (1:7 supernatant–calcium phosphate pellet). Adsorption of the [18F]fluoride ion by calcium phosphate as a bone substitute was found to be a useful method to detect the defluorination of 18F-labeled compounds [7]. When the incubation mixture using S9 fraction was analyzed by TLC using the developing solvent mixture B, two radioactive metabolites, [18F]fluoride ion (R f = 0) and a less polar compound (R f = 0.70), appeared with % peak area ratios of 57:34 at 5 min, 63:34 at 30 min and 71:29 at 60 min. In the case of the microsomal incubation mixture, the same radioactive metabolites appeared at ratios of 72:22 at 5 min, 85:15 at 30 min and 89:11 at 60 min (Fig. 3). This TLC result indicates that the less polar peak was rapidly converted to the [18F]fluoride ion with time. HPLC analysis of the incubation mixture of [18F]1 with S9 fraction showed three polar metabolite peaks and the parent peak, the most polar peak at the retention time of 7.9 min being coeluted with [18F]F , the peak at 26.1 min being unidentified and the peak at 34.3 min being coeluted with 3 (Fig. 4A and B). The latter two peaks gave a major fragment at m/z 301.3
(M++H) corresponding to the products of aromatic ring oxidation and were comigrated on TLC (R f = 0.70) (Fig. 3). Aromatic ring oxidation was reported as the major pathway of 5-HT1A antagonists, N-{2-[4-(2-methoxyphenyl)piperazino]ethyl}-N-(2-pyridyl) trans- and cis-4-fluorocyclohexanecarboxamide fluorocyclohexanecarboxamide in rat [16]. These results suggest that the in vitro metabolism of [18F]1 proceeded via both pathways a and b or pathway b (Scheme 2). On the other hand, the blood sample produced
Fig. 2. TLC analyses of the metabolites of [18F]2 obtained in vivo (A) and in vitro using either S9 fraction (B) or microsomes (C). The chromatograms are presented in incubation time order (1, 5, 15, 30 and 60 min). TLC was performed using developing solvent mixture A; Rf=0.67 ([18F]2).
170
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172 18
the [ F]fluoride ion as the only radioactive metabolite under the same condition. This may be because in the in vivo system, pathway b proceeds too rapidly for the intermediate radioactive metabolites to be detected or that pathway a is preferred (Scheme 2). The formation of [18F]fluoride ion was not surprising because of the presence of a reactive benzylic carbon in [18F]1, as shown in Scheme 2 [7,17], though a metabolically stable benzyl fluoride substituted with a chlorine atom meta to the benzylic position has been reported [18]. [18F]2 underwent severe metabolic transformation both in vivo and in vitro. A constant level of radioactivity was taken up by bone as a function of time (2.6 %ID/g at 1 min, 2.5 %ID/g at 30 min and 2.0 %ID/g at 60 min), indicative of little defluorination. TLC analysis of the blood sample using developing solvent mixture A showed a polar radioactive metabolite that appeared from 1 min postinjection and then increased with time, 27.2% at 1 min, 63.8% at 5 min, 91.8% at 15 min, 95.1% at 30 min and 100% at 60 min (Fig. 2). The same metabolite was detected in liver homogenates, but its formation was slower with less than 40% of the total peak at 60 min. The 30-min blood sample was reanalyzed using TLC developing solvent mixture B, and a radioactive metabolite peak was separated into two peaks with R f ’s of 0 and 0.2 (18:82) (Fig. 3). The generation of polar radioactive metabolites by either S9 fraction or by microsomes increased with time; 2.8% and 19.9% at 1 min, 22.2% and 62.1% at 5 min, 57.6% and 90.6% at 15 min, 79.1% and 100% at 30 min and 96.3% and 100% at 60 min (Fig. 2). Although the radioactive metabolite was not identified, it was separated into three peaks (R f = 0, 0.23 and 0.97) by TLC using developing solvent mixture B. The peak at R f of 0.97 was comigrated with the parent peak. In S9 fraction, the ratios of these three radioactive metabolites were 4:8:12 at 5 min, 14:29:37 at 30 min and 24:60:8 at 60 min (Fig. 3). In microsomal incubation mixtures, these ratios were 5:33:29 at 5 min, 28:49:23 at 30 min and 19:70:11 at 60 min (Fig. 3). The polar two peaks were not [18F]fluoride ion or 4-[18F]fluorobenzoic acid; the
Fig. 4. HPLC analyses of [18F]1 and [18F]2 metabolites obtained in vitro using S9 fraction at 30 min after incubation: (A) Radioactive metabolites of [18F]1; t R=7.9 min ([18F]F-), t R=26.1 min (unidentified), t R=34.3 min (coeluted with 3) and t R=40.3 min ([18F]1). (B) Nonradioactive standard 3. (C) Radioactive metabolites of [18F]2; t R=11.2, 24.3 and 30.7 min (unidentified), t R=35.1 min (coeluted with 4) and t R=42.5 min ([18F]2). (D) Nonradioactive standard 4.
latter was proposed to be derived from 4-[18F]fluorobenzaldehyde formed via N-dealkylation of [18F]2 [4,5]. [18F]Fluoride ion was confirmed to be absent by low-level calcium
Fig. 3. TLC analyses of [18F]1 and [18F]2 metabolites produced in vivo (A) or in vitro using either S9 fraction (B) or microsomes (C) at 30 min after injection or incubation. TLC was performed using developing solvent mixture B; Rf=0.93 ([18F]1) and Rf=0.97 ([18F]2).
Scheme 3. A major metabolic pathway of [18F]2.
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172
phosphate uptake (7:1 supernatant–calcium phosphate pellet), and 4-[18F]fluorobenzoic acid was confirmed absent by the different migration rates of the radioactive metabolites (R f = 0 and 0.1) and authentic 4-[18F]fluorobenzoic acid (R f = 0.3) on TLC using a 14:4:0.01 mixture of CH2Cl2CH3OH-NH4OH as a developing solvent mixture. The polar metabolites in either S9 or microsomes are not the product of glucuronide conjugation because the transformation requires a cofactor such as uridine 5V-diphosphoglucuronic acid [10,16]. HPLC analysis of the incubation mixture of [18F]2 with S9 fraction showed four polar metabolite peaks and the parent peak, the least polar peak at the retention time of 35.1 min being coeluted with 4, the other peaks at 11.2, 24.3 and 30.7 min being unidentified (Fig. 4C and D). All three peaks except the most polar peak gave a major fragment at m/z 287.3 (M++H) corresponding to the products of aromatic ring oxidation and were comigrated on TLC (R f = 0.97) when using developing solvent mixture B (Fig. 3) (Scheme 3). The rate of radioactive metabolite formation from [18F]1 was highest for microsomes and lowest for blood samples, whereas that of [18F]2 was highest in blood samples and lowest in S9 fraction. In the absence of NADPH, no radioactive metabolites were detected from either [18F]1 or [18F]2, indicating that cytochrome P-450 could be involved in the biotransformation. The in vivo method required high doses (N 8.89 MBq/ mouse) of radiotracer, whereas low doses ( N0.95 MBq/time point) were sufficient for the in vitro method. 4. Conclusion This study shows that the in vitro metabolism of [18F]1 using either mouse liver S9 fraction or microsomes produces the same results as the in vivo method using blood or bone. However, the intermediate radioactive metabolite derived from aromatic ring oxidation, which is believed to be converted into [18F]fluoride ion, was not observed in vivo (blood) by TLC analysis using developing solvent mixture B. By the same token, the in vitro metabolism of [18F]2 using S9 fraction or microsomes showed the same pattern of metabolite production as the in vivo blood or bone-based methods, although the radioactive metabolite derived from aromatic ring oxidation was not detected in vivo. Although the metabolism rates differed because of different concentrations of metabolizing enzymes, the results demonstrate that liver S9 fraction can be used instead of microsomes for the metabolism studies of radiotracers. Moreover, the liver S9 fraction is easily obtained during the early stage of liver microsomal preparation and contains both microsomal and cytosolic fractions. In addition, the in vitro metabolism method using either S9 fraction or microsomes produced intermediate radioactive metabolites not detected in vivo, which would provide valuable information on the in vivo metabolism pathways of radiotracers. Further studies using
171
various radiotracers are warranted to demonstrate the applicability of liver S9 fraction for the metabolism studies of radiotracers.
Acknowledgment This work was supported in part by the Samsung Biomedical Research Institute grant, #SBRI C-A5-124-3, and by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD) (R04-2002-000-20067-0).
References [1] Delaforge M. Importance of metabolism in pharmacological studies: possible in vitro predictability. Nucl Med Biol 1998;25:705 – 9. [2] Mazel P. General principles and procedures for drug metabolism in vitro. In: La Du BN, Mandel HG, Way EL, editors. Fundamentals of drug metabolism and drug deposition. Baltimore7 The Williams & Wilkins Company; 1971. p. 527 – 45. [3] Mazie`re B, Cantineau R, Coenen HH, Guillaume M, Halldin C, Luxen A, et al. PET radiopharmaceutical metabolism — plasma metabolite analysis. In: Stfcklin G, Pike VW, editors. Radiopharmaceuticals for positron emission tomography. Methodological aspects. Dordrecht7 Kluwer Academic Publishers; 1993. p. 151 – 78. [4] Mach RH, Nowak PA, Moroney DA, Morton TE, Jackson JR, Ehrenkaufer RLE. Synthesis and in vivo evaluation of [18F]-N-( pnitrobenzyl)spiperone ([18F]PNBS) in rats. Nucl Med Biol 1993; 20:269 – 78. [5] Welch MJ, Katzenellenbogen JA, Mathias CJ, Brodack JW, Carlson KE, Chi DY, et al. N-(3-[18F]Fluoropropyl)-spiperone: the preferred 18 F labeled spiperone analog for positron emission tomographic studies of the dopamine receptor. Nucl Med Biol 1988;15:83 – 97. [6] Tan P-Z, Baldwin RM, Van Dyck CH, Al-Tikriti M, Roth B, Khan N, et al. Characterization of radioactive metabolites of 5-HT2A receptor PET ligand [18F]altanserin in human and rodent. Nucl Med Biol 1999;26:601 – 8. [7] Lee S-Y, Choe YS, Kim DH, Park B-N, Kim SE, Choi Y, et al. A simple and efficient in vitro method for metabolism studies of radiotracers. Nucl Med Biol 2001;28:391 – 5. [8] Ames BN, McCann J, Yamasaki E. Methods for detecting carcinogens and mutagens with the Salmonella/mammalian-microsome mutagenicity test. Mutat Res 1975;31:347 – 64. [9] Liu K, Guengerich P, Yang SK. Enantioselective hydrolysis of lorazepam 3-acetate by esterases in human and rat liver microsome and rat brain S9 fraction. Drug Metab Dispos 1991;19:609 – 13. [10] Li AP, Hartman NR, Lu C, Collins JM, Strong JM. Effects of cytochrome P450 inducers on 17a-ethynyloestradiol (EE2) conjugation by primary human hepatocytes. Br J Clin Pharmacol 1999;48: 733 – 42. [11] Yoshihara S, Makishima M, Suzuki N, Ohta S. Metabolic activation of bisphenol A by rat liver S9 fraction. Toxicol Sci 2001;62:221 – 7. [12] Guengerich FP, Martin MV, Beaune PH, Kremers P, Wolff T, Waxman DJ. Characterization of rat and human liver microsomal cytochrome P-450 forms involved in nifedipine oxidation, a prototype for genetic polymorphism in oxidative drug metabolism. J Biol Chem 1986;261: 5051 – 60. [13] Haka MS, Kilbourn MR, Watkins GL, Toorongian SA. Aryltrimethylammonium trifluoromethanesulfonates as precursors to aryl [18F]fluoride: improved synthesis of [18F]GBR-13119. J Labelled Compd Radiopharm 1989;27:823 – 33. [14] Bradford M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of dye-binding. Anal Biochem 1976;72:248 – 54.
172
E.K. Ryu et al. / Nuclear Medicine and Biology 33 (2006) 165–172
[15] Choe YS, Oh SJ, Shim I, Naruto S, Chi DY, Kim SE, et al. Syntheses and biological evaluation of 18F-labeled 3-(1-benzyl-piperidin-4-yl)-1(1-methyl-1H-indol-3-yl)propan-1-ones for in vivo mapping of acetylcholinesterase. Nucl Med Biol 2000;27:263 – 7. [16] Ma Y, Lang L, Kiesewetter DO, Jagoda E, Sassaman MB, Der M, et al. Liquid chromatography-tandem mass spectrometry identification of metabolites of two 5-HT1A antagonists, N-{2-[4-(2-methoxyphenyl)piperazino]ethyl}-N-(2-pyridyl) trans- and cis-4-fluorocyclohex-
anecarboxamide, produced by human and rat hepatocytes. J Chromatogr B 2001;755:47 – 56. [17] Lee KC, Lee S-Y, Choe YS, Chi DY. Metabolic stability of [18F]fluoroalkylbiphenyls. Bull Korean Chem Soc 2004;25:1225 – 30. [18] Magata Y, Lang L, Kiesewetter DO, Jagoda EM, Channing MA, Eckelman WC. Biologically stable [18F]-labeled benzylfluoride derivatives. Nucl Med Biol 2000;27:163 – 8.