Chip-Based P450 Drug Metabolism Coupled to Electrospray

Initial experiments described here employed a small commercial guard column in ..... The identity of the metabolites was confirmed by the mass spectra...
0 downloads 0 Views 111KB Size
Download PDF
Anal. Chem. 2003, 75, 6430-6436

Chip-Based P450 Drug Metabolism Coupled to Electrospray Ionization-Mass Spectrometry Detection Salete Benetton,† Jun Kameoka,‡ Aimin Tan,† Timothy Wachs,† Harold Craighead,‡ and Jack D. Henion*,†

Analytical Toxicology, College of Veterinary Medicine, Cornell University, 927 Warren Drive, Ithaca, New York 14850, and School of Applied and Engineering Physics, Cornell University, 212 Clark Hall, Ithaca, New York 14850

A chip-based P450 in vitro metabolism assay coupled with ESI-MS and ESI-MS/MS detection is described in this paper. The chips were made of a cyclic olefin polymer using a hot embossing process. The introduction of reagent solutions into the chip was carried out using fused-silica capillaries coupled to two syringes with the flow rate controlled by a syringe pump. Initial experiments described here employed a small commercial guard column in an off-chip format to desalt and concentrate the products of the enzymatic reaction prior to ESI-MS analysis. The system was used both to yield the Michaelis constant (Km) of the P450 biotransformation of imipramine into desipramine and to determine the IC50 value of a chemical inhibitor (tranylcypromine) for this CYP2C19-mediated reaction. The results demonstrated that the kinetics of the reaction inside the 4-µL volume within the channels of the cyclic olefin polymer chip provided results in agreement with those reported in the literature using conventional assays. The above reactions were carried out using human liver microsomes, and the metabolites were detected by ESI-MS showing the potential of the chip-based P450 reaction for metabolite screening studies as well as for P450 inhibition assays. A porous monolithic column was subsequently integrated into the chip to perform the reaction mixture cleanup process in an integrated fashion on the chip that is necessary for ESIMS detection. The miniature monolithic SPE column was prepared in situ inside the chip via UV-initiated polymerization. The results obtained using the integrated system demonstrated the possibility of performing P450 enzymatic reactions in a microvolume reaction chamber coupled directly to ESI-MS detection and required less than 4 µg of HLM protein. Research on microfabricated analytical devices has grown considerably in the last 10 years. Many publications have reported results and accomplishments demonstrating the advantages of these analytical microsystems, including speed, reduced sample/ * Corresponding author: (e-mail) [email protected]; (phone) 607-253-3971; (fax) 607-253-3973. † College of Veterinary Medicine. ‡ School of Applied and Engineering Physics.

6430 Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

reagent consumption, integration, and automation.1-8 Among the various areas of application of these chip-based analytical systems are enzymatic assays.9 Current on-chip enzymatic assays use soluble or immobilized enzymes and rely on the conversion of a nondetectable species to a detectable one.9-15 Fluorescence and amperometry have been the main types of detection used in these assays.9 In this work, we have studied the feasibility of performing enzymatic reactions in a chip-based microfluidic format coupled directly with electrospray ionization (ESI)-mass spectrometry (MS) detection. The enzymes investigated here belong to the cytochrome P450 system, which is a membrane-bound (waterinsoluble) enzymatic system responsible for the biotransformation of xenobiotics.16 Cytochrome P450 enzymes have long been recognized as the primary system responsible for human drug metabolism.16 The P450 enzymatic system is a superfamily of heme-containing monooxygenases located within the endoplasmic reticulum. The enzyme system consists of two protein components: a hemoprotein called cytochrome P450 and a flavoprotein called NADPHcytochrome P450 reductase. Cytochrome P450 is the substrateand oxygen-binding site of the enzyme system, while the reductase (1) Kricka, L. J. Clin. Chem. 1998, 44, 2008-2014. (2) Waters, L. C.; Jacobson, S. C.; Kroutchinina, N.; Khandurina, J.; Foote, R. S.; Ramsey, J. M. Anal. Chem. 1998, 70, 158-162. (3) Wang, J.; Chatrati, M. P.; Tian, B.; Polsky, R. Anal. Chem. 2000, 72, 25142518. (4) Deng, Y.; Henion, J.; Li, J.; Thibault, P.; Wang, C.; Harrison, D. J. Anal. Chem. 2001, 73, 639-646. (5) Deng, Y.; Zhang, H.; Henion, J. Anal. Chem. 2001, 73, 1432-1439. (6) Kameoka, J.; Craighead, H. G.; Zhang, H.; Henion, J. Anal. Chem. 2001, 73, 1935-1941. (7) Khandurina, J.; Guttman, A. J. Chromatogr., A 2002, 943, 159-183. (8) Kameoka, J.; Orth, R.; Ilic, B.; Czaplewski, D.; Wachs, T.; Craighead, H. G. Anal. Chem. 2002, 74, 5897-5901 (9) Wang, J. Electrophoresis 2002, 23, 713-718. (10) Hadd, A. G.; Raymond, D. E.; Halliwell, J. W.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1997, 69, 3407-3412. (11) Hadd, A. G.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1999, 71, 52065212. (12) Duffy, M.; Gillis, H. L.; Lin, J.; Sheppard, N. F.; Kellog, G. J. Anal. Chem. 1999, 71, 4669-4678. (13) Chen, C. B.; Chin-Dixon, E.; Jeong, S.; Nikiforov, T. Anal. Biochem. 1999, 273, 89-97. (14) Kerby, M.; Chien, R. L. Electrophoresis 2001, 22, 3916-3923. (15) Mao, H.; Yang, T.; Cremer, P. S. Anal. Chem. 2002, 74, 379-385. (16) Anzenbacher, P.; Anzenbacherova, E. Cell. Mol. Life Sci. 2001, 58, 737747. 10.1021/ac030249+ CCC: $25.00

© 2003 American Chemical Society Published on Web 10/30/2003

serves as an electron carrier, shuttling electrons from NADPH to the cytochrome P450 substrate complex.17 The hepatic microsomal cytochrome P450s catalyze the Phase I metabolism of xenobiotics, the initial step in the biotransformation and elimination of a wide variety of drugs and environmental pollutants.17,18 In the process of drug discovery, various in vitro methodologies are routinely used in pharmaceutical companies to screen a vast number of candidate drugs to determine rate of metabolism, to predict drug-drug interactions and metabolic profiles, and to determine the involvement of enzymes that present genetic polymorphism.19 Traditionally, these in vitro methods use either liver tissue slices, primary culture of hepatocytes, liver microsomes, or recombinant drug-metabolizing enzymes.19,20 Essential characteristics desired for an analytical drug discovery method include sensitivity, speed, automation, simultaneous testing of microvolumes of candidate drug compounds, and use of minimum quantities of costly material. In the case of P450 enzymatic reactions, besides quantification, identification of the metabolites generated by the candidate drug compounds is very important. Hence, the combination of a chip-based P450 reaction with a more universal detection system is highly desirable. ESI-MS is a sensitive and selective detection system that is compatible with chip flow rates and can also provide structural information that is essential for metabolite screening studies. The advantage of miniaturizing P450 enzymatic reactions includes the low sample volumes and hence the use of minimum quantities of costly materials and the potential for development of highthroughput assays. To monitor an on-chip reaction by ESI-MS, the reaction product mixture must be desalted in addition to removal of the enzyme molecules. Also, preconcentration of metabolites is desirable to bring them to a higher concentration and thus increase the sensitivity of the assay. Initial experiments described here employed a small commercial guard column in an off-chip format to desalt and concentrate the products of the enzymatic reaction prior to ESI-MS analysis. A porous monolithic column was later integrated into the chip to perform an integrated solid-phase extraction (SPE) cleanup process necessary for ESI-MS detection. The miniature monolithic SPE columns were prepared in situ inside the chip via UV-initiated polymerization. Using imipramine, doxepin, and amitriptyline (antidepressant drugs) as test small-molecule drug compounds, we investigated the biotransformation of these compounds in a chip-based format. The system was also used both to yield the Michaelis constant (Km) of the P450 biotransformation of imipramine into desipramine and to determine the IC50 value of a chemical inhibitor (tranylcypromine) for this CYP2C19-mediated reaction. The chips were made of Zeonor (a cyclic olefin) polymer that has already been shown to be suitable for making chips because it is readily embossed and bonded as previously reported.6 This report describes the successful on-chip P450 in vitro metabolism of three representative drugs followed by on-chip SPE sample cleanup and direct electrospray analysis of the resulting samples. (17) Meyer, U. A. J. Pharmacokinet. Biopharm. 1996, 24, 449-459. (18) Guengerich, F. P. J. Pharmacokinet. Biopharm. 1996, 24, 521-533. (19) Venkatakrishnan, K.; Von Moltke, L. L.; Greenblatt, D. J. J. Clin. Pharmacol. 2001, 41, 1149-1179. (20) Rodrigues, A. D. Biochem. Pharmacol. 1994, 48, 2147-2156.

Figure 1. Design of the chip for the P450 reaction. The microchannels before the reaction region are 100 µm wide and 100 µm deep. In the reaction region, the channels are 200 µm wide and 100 µm deep with a total volume from the Y-junction to the outlet of 4 µL. (A) Enzyme inlet; (B) substrate + cofactor inlet; (C) Nanoport connection; (D) reaction region.

EXPERIMENTAL SECTION Chemicals. Imipramine, desipramine, doxepin, amitriptyline, tranylcypromine, β-nicotinamide adenine dinucleotide phosphate (NADPH), magnesium chloride, and Tris base were purchased from Sigma Chemical Co. (St. Louis, MO). Pooled human liver microsomes (HLM) containing 20 mg of microsomal protein/mL were purchased from Gentest Corp. (Woburn, MA). Ethylene dimethacrylate (EDMA), butyl methacrylate (BMA), 2, 2-dimethoxy2-phenylacetophenone (DAP), 2-acrylamido-2-methyl-1-propanesulfonic acid (AMPS), and hexane were all obtained from Aldrich Chemical Co. (Milwaukee, WI). EDMA and BMA were filtered through a 0.45-µm PTFE syringe filter (13-mm diameter, Nalge Nunc International, Rochester, NY) filled with fresh basic alumina powder (mesh 60-325, Fisher, Pittsburgh, PA). Other reagents were hydrochloric acid (EM Science, Gibbstown, NJ) and HPLC grade acetonitrile (J. T. Baker, Phillipsburg, NJ). Zeonor polymer was purchased from Zeon Chemicals (Louisville, KY). The 2.5 and 25 mM Tris buffer was prepared by dissolving Tris base in deionized water with the pH adjusted to 7.4 using hydrochloric acid. All solutions in this work were prepared using NANOpure deionized water (Barnstead, Boston, MA). Chip Fabrication. Figure 1 shows the layout and exterior dimensions of the polymer microfluidic chip used in this report. The dimensions of these chips were 4 cm × 6 cm (3.7 mm thick) with two inlet channels meeting at a Y-junction and forming the mixing region (reaction region). The dimensions of the inlet channels are 100 µm × 100 µm. The microchannels in the reaction region are 200 µm wide and 100 µm deep with a total volume of 4 µL from the Y-junction to the outlet. The fabrication process for the polymeric microfluidic chip, including the fabrication of etched silicon wafer masters, was as described previously.6 Chip Operation. The introduction of reagent solutions into the chip was carried out using fused-silica capillaries coupled to two syringes with the flow rate (1 µL/min) controlled by a twochannel infusion pump (Harvard Apparatus, South Natick, MA). The fused-silica capillaries are connected to the chip using commercial microconnectors (N-124S, Nanoports from UpChurch Scientific, Oak Harbor, WA) attached to the two inlets and to the outlet of the chip by an epoxy-based glue (DP420, 3M Engineered Adhesives Division, St. Paul, MN). In Figure 2 is shown a Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

6431

Figure 2. Schematic representation of on-chip P450 reaction using an off-chip guard column to clean up and concentrate the reaction mixture before detection by ESI-MS. (A) Zeonor chip; (B) fused-silica capillary connecting the chip to the guard column (flow rate, 2 µL/ min); (C) fused-silica capillary connecting the infusion pump delivering the elution solvent to the guard column (flow rate, 10 µL/min); (D) infusion pump for delivering the elution solvent; (E) C18 guard column; (F) ion spray unit; (G) PE Sciex API 3000 mass spectrometer.

schematic representation of a coupled on-chip P450 metabolism region, an off-chip guard column to desalt and concentrate the reaction mixture, followed by detection via ESI-MS. The guard column was connected directly to the turbo ion spray unit of the LC-MS/MS system (Figure 2). When the guard column was used for the sample desalting step, a fused-silica capillary was also connected to the outlet of the chip to transfer the product reaction mixture to the guard column. A 22.3 × 50.8 mm Thermal-Clear Heater (Minco Prod., Inc. Minneapolis, MN) is positioned under the chip to maintain the reaction region temperature at 37 °C. The temperature of the Zeonor polymer chip is monitored using a reversible surface thermometer (Omega Eng., Inc. Stanford, CT). MS/MS Conditions. The MS/MS system was a PE Sciex API 3000 tandem triple quadrupole mass spectrometer (Concord, ON, Canada) equipped with a turbo ion spray LC/MS interface operated in the ion spray mode at 3300 V in the positive ion mode. Operating conditions, optimized by infusion of a mixture of imipramine and desipramine (both at 1 µg/mL in acetonitrile that contained 0.1% formic acid) at a flow rate of 10 µL/min, were determined as follows: the nebulizer, curtain, and collision gases were set to 4, 12, and 4, respectively. The declustering potential was set to 26 V and the ring voltage at 220 V. Selected reaction monitoring (SRM) in the positive ionization mode was performed using a dwell time of 200 ms per transition to detect transitions at m/z 281.2 f 86.2 (imipramine), and m/z 267.2 f 72.0 (desipramine) using a collision energy of 28 eV and nitrogen as the collision gas. Enzymatic Reaction. The on-chip metabolic reaction was performed by pumping enzyme (3.3 mg of microsomal protein/ mL in 25 mM Tris buffer, pH 7.4) and substrate solution (20 µM in 25 mM Tris buffer, pH 7.4) through the reaction region (mixing section of the chip). After the channels of the reaction region were filled, the flow was stopped and the mixed solution was allowed to incubate for 30 min at 37 °C. Pumping enzyme and substrate solution at the same flow rates (1 µL/min) resulted in the 1:1 dilution of the sample in the reaction region. Other components of the reaction, including NADPH and MgCl2, were premixed offchip with the substrate. A mixture of the three antidepressant drugs, imipramine, doxepin, and amitriptyline, all at an on-chip concentration of 10 µM, was incubated with HLM (6.6 µg of microsomal protein) in the 4-µL channel (reaction region) on the chip. The incubation took place in 25 mM Tris buffer at a pH of 7.4, with 1 mM NADPH and 3 mM MgCl2. After 30 min of 6432

Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

incubation, the reaction mixture was pumped to and collected in a C18 guard column (15 mm long, 1 mm i.d., Opti-Guard column, Optimize Technologies, Inc. Oregon City, OR) at a flow rate of 2 µL/min to desalt and preconcentrate the sample before mass spectrometric analysis. After washing the column with Tris buffer 2.5 mM (ph 7.4) for 10 min, the column was connected to an PE Sciex API 3000 tandem triple quadrupole mass spectrometer, and the reaction mixture was eluted from the column with acetonitrile containing 0.1% formic acid at a flow rate of 10 µL/min using an infusion pump. The mass spectral data were acquired across the range m/z 250-350. A control experiment was performed following the same procedure described above but using inactivated human liver microsomes (microsomes previously kept at 100 °C for 5 min). Enzyme Assay. The biotransformation of imipramine into desipramine was used as a model to study a P450 enzymatic assay in the described chip-based format. The activity of an enzyme can be measured by determining the rate of product formation.21 The velocity of the substrate conversion to product versus substrate concentration can be described by the Michaelis-Menten kinetics that gives hyperbolic kinetics in which the Km and Vmax are determined by the following equation:

v ) VmaxS/(Km + S)

(1)

where S is the substrate concentration, v is the velocity of the reaction, Vmax is the velocity of the enzyme reactions when all the substrates are at saturating concentrations, Km is the concentration of substrate required to achieve half the maximum velocity of the enzyme. Km, and Vmax are the parameters that define a rectangular hyperbola relating v to s. The derivation of the equation relies on three assumptions: (1) the rate of the reverse reaction is negligible during the measurement period, (2) the measurement is made in the steady state when the concentration of enzyme-substrate complex is unchanging, and (3) the formation of enzyme-substrate complex does not significantly deplete the concentration of free substrate.21 The measured velocity should be linearly proportional to enzyme concentration and to the time of reaction, both requirements to ensure the assumptions underlying the Michaelis-Menten relationship. Hence, following the same procedure described above, the standard incubation conditions were chosen on the basis of the results of preliminary experiments that showed that the reaction for the biotransformation of imipramine to desipramine was linear with respect to the incubation time (20 min) and the amount of enzyme protein (6.6 µg), and it gave sufficient response by ESI-MS/MS detection to proceed with the planned kinetic and inhibition experiments. To determine the value of Km for the enzymatic formation of desipramine, enzyme kinetic experiments were performed (in duplicate on-chip incubations) with four concentrations of imipramine ranging from 5 to 100 µM. The metabolite formation (area of the metabolite’s SRM ion current)-substrate concentration data were fitted to the Michaelis-Menten equation, and the Km value was determined by nonlinear regression analysis (SigmaPlot 2000, Version 6.0). (21) Eisenthal, R., Danson, M. J., Eds. Enzyme Assays. A Practical Approach; Oxford University Press: New York, 1992.

Figure 3. (A) Schematic diagram of the Zeonor polymeric chip (Z) containing an incorporated SPE monolithic column. C1, C2, inlet connecting capillaries; C3, outlet connecting capillary; RC, reaction channels; M, monolithic SPE column. (B) The substrate plate with embossed reaction channels (RC). (C) The cover plate embossed against three fused-silica capillaries (EC1, EC2, EC3) to make channels for connecting capillaries and monolithic SPE column (see text for further details).

Imipramine is converted to its N-demethylated metabolite primarily by CYP2C19,22 which may be inhibited by the action of tranylcypromine.23,24 Tranylcypromine was therefore used as a chemical inhibitor to determine its IC50 value for the CYP2C19 imipramine N-demethylation-mediated reaction. The concentrations of tranylcypromine (0, 5, 10, and 15 µM) used for the calculation of the IC50 were chosen according to those previously shown for this compound to inhibit CYP2C19-mediated reactions.25,26 Tranylcypromine was dissolved in 50% acetonitrile (final concentration of organic solvent inside the chip was e0.5% v/v) and was premixed off-chip with the substrate. For the determination of the IC50, the following equation was used to determine the formation of metabolite: Am/Am+Ad (where Am and Ad are the area of the metabolite’s and the area of the parent drug’s SRM ion current, respectively). Comparison was made to experimental control incubations (0 µM inhibitor) and the enzyme activity expressed as the percentage of control remaining. The IC50 value was determined by graphical linear interpolation. Integration of a Porous Monolithic SPE Column on the Chip. Figure 3A shows the layout of the polymer microfluidic chip containing an incorporated monolithic SPE column. Zeonor polymer plate was cut into 4 cm by 6 cm pieces. The substrate plate was embossed against an etched silicon wafer master to obtain reaction channels (Figure 3B). The cover plate was then embossed against three conventional fused-silica capillaries (360µm o.d., 50-µm i.d.) as shown in Figure 3C. After completing this embossing process, the three capillaries were peeled off and three corresponding embossed channels (18 mm long for the two inlet channels, 35 mm long for the outlet channel) were left in this (22) Madsen, H.; Rasmussen, B. B.; Brosen, K. Clin. Pharmacol. Ther. 1997, 61, 319-324. (23) Bu, H. Z.; Knuth, K.; Magis, L.; Teitelbaum, P. J. Pharm. Biomed. Anal. 2001, 25, 437-442. (24) Hartter, S.; Tybring, G.; Friedberg, T.; Weigmann, H.; Hiemke, C. Pharm. Res. 2002, 19, 1034-1037. (25) Yu, C.; Shin, Y. G.; Kosmeder, J. W.; Pezzuto, J. M.; van Breemen, R. B. Rapid Commun. Mass Spectrom. 2003, 17, 307-313. (26) Dierks, E. A.; Stams, K. S.; Lim, H.; Cornelius, G.; Zhang, H.; Ball, S. E. Drug Metab. Dispos. 2001, 29, 23-29.

cover plate. For the two embossed inlet channels, two connecting capillaries (360-µm o.d., 50-µm i.d., 280 mm long) were inserted into each channel until 10 mm of each of the two capillaries was inside the embossed channels. Another connecting capillary (360µm o.d., 50-µm i.d., 180 mm long) was inserted into the embossed outlet channel until 10 mm of the capillary was inside the embossed channel. Then, the embossed substrate plate (Figure 3B) and the embossed cover plate (Figure 3C) were aligned and thermally bonded for 10 min. After this thermal bonding process, epoxy glue was applied to the edges of the bonded chip around the connecting capillaries to reinforce the bonding in that area. Preparation of the Monolithic SPE Column. The monomer and porogen mixture was similar to those reported by others.27-29 A typical formulation of monomer and porogen mixture was as follows. Five milligrams of DAP and 3 mg of AMPS were dissolved in 500 µL of methanol and 250 µL of hexane, to which 150 µL of BMA and 100 µL of EDMA were added. After being purged with nitrogen for 5 min, the mixture was aspirated into the chip using a syringe. A mask containing one layer of black tape and one layer of aluminum foil was placed over the chip with the 25-mm-long embossed channel in the right side of the chip uncovered. The polymerization was initiated by UV irradiation for 20 min under UV lamp (UVL-28, 365 nm, VWR, West Chester, PA). When the polymerization process was complete, the remaining monomer mixture inside the connecting capillaries and nonpolymerized monomer in the chip were removed with methanol. In Figure 3A is shown a schematic representation of the completed chip with the integrated monolithic column. Using the chip with the integrated monolithic column, the P450 enzymatic reaction was performed as described above using 10 µM imipramine as the parent drug. A control experiment was also performed following the same procedure described above but using inactivated human liver microsomes. RESULTS AND DISCUSSION This work focused on demostrating the feasibility of performing P450 enzymatic reactions using human liver microsomes in a chipbased format coupled with ESI-MS detection. The chip was designed to permit the mixing of the reagents, notably the enzyme and substrate, by a molecular diffusion process in a pressuredriven laminar flow. Experiments using a fluorescent substrate were performed to confirm the mixing process under a fluorescent microscope (data not shown). A pressure-driven flow inside the chip was selected because the reactions of P450 enzymes take place in a complex environment (two membrane-bound protein components, cofactor, MgCl2). Although electrokinetic driven flow is easy to implement and is used in most of the current on-chip enzymatic reactions, it is also sensitive to the physicochemical properties of the fluid being pumped (pH, ionic strength, viscosity) and to bubbles in the channel, and it is surface sensitive. To ensure adequate mixing and preclude these problems, a syringe pump was used to provide pressure-driven flow. Enzymatic Reaction. In vitro metabolism using the described system was demonstrated by the transformation of a mixture of (27) Yu, C.; Davey, M. H.; Svec, F.; Frechet, J. M. J. Anal. Chem. 2001, 73, 5088-5096. (28) Rohr, T.; Yu, C.; Davey, M. H.; Svec, F.; Frechet, J. M. J. Electrophoresis 2001, 22, 3959-3967. (29) Tan, A.; Benetton, S.; Henion, J., Anal. Chem., submitted.

Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

6433

Figure 5. Michaelis-Menten curve showing the results of the kinetics of on-chip desipramine formation. The results are the average of duplicate measurements. The Km value was found to be 46 µM.

Figure 6. Inhibition of desipramine formation by tranylcypromine. The results are the average of triplicate measurements (error bar is the standard error of the mean). The IC50 value, calculated by graphical linear interpolation, is 8 µM.

Figure 4. On-chip incubation of doxepin, amitriptyline, and imipramine with HLM. Preconcentrantion of sample using a 1-mm C18 guard (Optiguard) and detection by Q1 full-scan (m/z 250-350) ESIMS (PE Sciex API 3000). Solid line, results from incubation; dotted line, results from blank (incubation with inactive HLM). (A) Consumption of amitriptyline (m/z 278.1), doxepin (m/z 280.1), and imipramine (m/z 281.1); (B) formation of N-demethylated metabolites of amitriptyline (m/z 264.2), doxepin (m/z 266.2), and imipramine (m/z 267.2), respectively; (C) formation of OH-amitriptyline (m/z 294.2), OHdoxepin (m/z 296.2), and OH-imipramine (m/z 297.2).

imipramine, doxazepin, and amitriptyline into their respective metabolites inside a 4-µL volume within the channel of a polymer chip by the action of human liver microsomes (Figure 4). The identities of the metabolites were confirmed by the mass spectra obtained under the abbreviated full-scan MS analysis. Figure 4A shows the consumption (solid traces) of the parent drugs (ami6434 Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

triptyline with a [M + H]+ ) 278.2, doxepin with a [M + H]+ ) 280.2, and imipramine with a [M + H]+ ) 281.2) compared to the control baseline ion current, which is from an incubation with inactivated enzyme (dotted traces). In Figure 4B is shown the formation of the N-demethylated metabolites of amitriptyline with a [M + H]+ ) 264.2, doxepin with a [M + H]+ ) 266.2, and imipramine with a [M + H]+ ) 267.2 (solid traces) compared to control baseline ion current (dotted line). In Figure 4C is shown the formation of the monohydroxylated metabolites of amitriptyline with a [M + H]+ ) 294.2, doxepin with a [M + H]+ ) 296.2, and imipramine with a [M + H]+ ) 297.2 (solid traces) compared to control baseline ion current (dotted line). Imipramine is reported to be hydroxylated in more than one position,30 and these isomers are not resolved under the conditions described here. Although, under the described conditions, both the Ndemethylated and hydroxylated metabolites of imipramine have the interference of an isotope of amitriptyline, it was still possible to observe the formation of these metabolites (Figure 4B,C). Enzyme Assay. On-chip experiments were performed to characterize the enzyme kinetics of the N-demethylation of imipramine with pooled human liver microsomes, and the results were plotted in a Michaelis-Menten curve (Figure 5). Imipramine was chosen as a model compound for the enzymatic assay because the kinetics of its transformation into desipramine is described in the literature along with the enzyme responsible for this biotransformation.31,32 The Km value was found to be 46 µM, which (30) Chen, A. G.; Wing, Y. K.; Chiu, H.; Lee, S.; Chen, C. N.; Chan, K. J. Chromatogr., B: Biomed. Sci. Appl. 1997, 693, 153-158. (31) Obach, R. S.; Reed-Hagen, A. E. Drug Metab. Dispos. 2002, 30, 831-837.

Figure 7. ESI-MS results from the incubation of imipramine with HLM in a chip with on-chip sample cleanup employing a monolithic SPE column. The data were acquired with Q1 scan across the range m/z 250-350. The spectrum shows the formation of the N-demethylated metabolite, desipramine, with a [M + H]+ ) 267.2, the formation of the monohydroxylated isomeric metabolites of imipramine with [M + H]+ ) 297.2, and the remaining parent drug imipramine with [M + H]+ ) 281.2.

is an acceptable result33 compared to those reported in the literature, 10-20 µM.31 On-chip incubation of imipramine with different concentrations of tranylcypromine allowed the determination of the IC50 value for the inhibition of the CYP2C19mediated formation of desipramine. The value of the IC50 for tranylcypromine was found to be 8 µM, which is in agreement with values (8.9, 9 µM) found for the inhibition of other CYP2C19mediated reactions that show correlation with the N-demethylation of imipramine.25,26,32 The experiments were carried out in triplicate on three different days and the precision of the assay, expressed as coefficient of variation (CV), varied from 7 to 23% at all concentrations used, which demonstrated the acceptable reproducibility of the on-chip experiments under the conditions described in this work (Figure 6). Integrated Porous Monolithic Column. The mass spectral results obtained using the monolithic column integrated on the chip are shown in Figure 7. In this experiment, imipramine was incubated on-chip in the described P450 in vitro mixture followed by on-chip SPE desalting employing the monolithic column. The identity of the metabolites was confirmed by the mass spectra obtained under the mass range m/z 250-350. Figure 7 shows the formation of the N-demethylated metabolite, desipramine, with [M + H]+ ) 267.2, and the formation of the monohydroxylated isomeric metabolites of imipramine with [M + H]+ ) 297.2 with a good response by ESI-MS. These results demonstrate the possibility of performing P450 enzymatic reactions in a microvolume and detecting the metabolic products of the reaction using ESI-MS with a fully integrated on-chip sample preparation procedure. (32) Chiba, K.; Saitoh, A.; Royana, E.; Tani, M.; Hayashi, M.; Ishizaki, T. Br. J. Clin. Pharmacol. 1994, 37, 237-242. (33) Tucker, G. T.; Houston, J. B.; Huang, S. M. Br. J. Clin. Pharmacol. 2001, 52, 107-117.

The rate of the P450 enzymatic reaction, as with any enzymatic reaction, depends on the temperature and the pH. P450 enzymatic reactions were performed at a controlled temperature of 37 °C, and heating of the chip was achieved with the use of a 22.3 × 50.8 mm Thermal-Clear heater positioned under the chip. The actual temperature inside the chip was not determined, but it was inferred to be in the vicinity of 37 °C by monitoring the temperature of the upper part of the Zeonor polymer chip using a reversible surface thermometer. The optimal pH value of 7.4 used in P450 assays is not favorable for the retention of basic drugs and their respective metabolites in a C18 SPE column. The pKa of imipramine, for example, is 9.4, and making a direct connection between the reaction region and the SPE column, as shown in Figure 3, does not permit adjustment of the pH for an efficient extraction of these compounds, which would lead to the prospect of using less enzyme or a smaller volume. Despite working with a nonfavorable pH for extraction, the retention of metabolites was enough for determining the feasibility of the process. In future work, the integration of pH adjustment before introducing the enzymatic mixture into the column could potentially lead to a more efficient protocol. Another characteristic of a microscale device is the increase in the surface-to-volume ratio, sometimes by orders of magnitude.34 A well-known problem due to this phenomenon is that surface adsorption of biological macromolecules and small molecules can increase. Since the on-chip enzymatic reaction described in this work takes place in the whole volume of the chip, enzyme adsorption to the walls of the channels did not appear to impede the formation of the products. Additionally, P450 reactions are traditionally carried out in an open environment so that the (34) Locascio, L. E.; Perso, C. E.; Lee, C. S. J. Chromatogr., A 1999, 857, 275284.

Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

6435

reaction is exposed to atmospheric oxygen. Although this work presents only preliminary results on the feasibility of performing P450 enzymatic reactions on a chip format, it demonstrated that P450 biotransformation reactions can occur in a closed compartment. Generally, the amount of enzyme and the volume used for an assay will depend on several factors. These include the following: (1) the sensitivity of the analytical assay for the metabolites; (2) the kinetic parameters of enzyme catalysis for a particular substrate and P450; (3) the incubation period of the assay; and (4) the solubility of the test compound or substrate. Using three antidepressant drugs, we have demonstrated that, even using these microvolumes, the P450 enzymatic reaction can be performed for these compounds and form sufficient quantities of metabolites to permit their detection by ESI-MS. CONCLUSIONS In this work, we have shown the feasibility of performing P450 enzymatic reactions inside a 4-µL volume housed in a cyclic olefin polymer chip by the action of human liver microsomes. A commercial guard column was initially used to desalt the products of the enzymatic reaction and preconcentrate the analytes. A porous monolithic column was also integrated into the chip in

6436

Analytical Chemistry, Vol. 75, No. 23, December 1, 2003

order to perform, directly on the chip, the cleanup and concentration process necessary for ESI-MS detection. This combined system allowed the P450 enzymatic reactions to occur in a microvolume and the detection of expected metabolic products of the reaction using ESI-MS in chip-based integrated process. A key advantage of miniaturizing P450 enzymatic reactions is the use of low sample volumes, hence the use of minimum quantities of costly material. Furthermore, the use of ESI-MS provides a sensitive, selective detection strategy, which also can provide structural information. We suggest this strategy may present an important advance for both P450 inhibition assays and for metabolite screening studies in the future. ACKNOWLEDGMENT The authors thank Merck Research Laboratories (MRL) for financial support of the research project and Dr. Richard King from MRL for the helpful discussions. We also thank PE Sciex for the generous loan of the API 3000 mass spectrometer system.

Received for review July 2, 2003. Accepted August 28, 2003. AC030249+