A Validated Stable Isotope Dilution Liquid Chromatography Tandem

Anal. Chem. 1999, 71, 2021-2027

A Validated Stable Isotope Dilution Liquid Chromatography Tandem Mass Spectrometry Assay for the Trace Analysis of Cocaine and Its Major Metabolites in Plasma Gurkeerat Singh, Vinod Arora,† P. Thomas Fenn, Berend Mets,‡ and Ian A. Blair*

Center for Cancer Pharmacology, Department of Pharmacology, University of Pennsylvania, Civic Center Blvd. and Osler Circle, Philadelphia, Pennsylvania 19104-4318

A validated method has been developed for the simultaneous quantitation of cocaine and its major metabolites (ecgonine methyl ester, benzoylecgonine, and norcocaine) in rat plasma. The method is based upon the use of stable isotope dilution liquid chromatography/atmospheric pressure chemical ionization/tandem mass spectrometry. Previously reported methods do not have the sensitivity and specificity that can be attained with this method. Plasma samples required no cleanup apart from protein precipitation, and no derivatization was required. Selected reaction monitoring was performed on the transitions of m/z 200 to m/z 182 (ecgonine methyl ester), m/z 290 to m/z 168 (benzoylecgonine), m/z 304 to m/z 182 (cocaine), and m/z 290 to m/z 168 (norcocaine). The standard curves were linear over the range from 2 ng/ mL (benzoylecgonine, cocaine, and norcocaine) or 5 ng/ mL (ecgonine methyl ester) to 1000 ng/mL in rat plasma. The lower limit of quantitation (LLQ) for benzoylecgonine, cocaine, and norcocaine was 2 ng/mL, and for ecgonine methyl ester, the LLQ was 5 ng/mL for plasma. This simple, rapid, reliable, and sensitive method of quantitation had excellent accuracy and precision for the four analytes. The method was sensitive enough to permit a detailed study of the pharmacokinetics of cocaine and its metabolites after administration of a bolus intravenous dose to rats. Cocaine, a naturally occurring stimulant found in the leaves of coca plants is a drug, which is used by over 2 million people in the U. S.1 Cardiotoxicity is the most commonly observed toxic side effect resulting from cocaine abuse. However, central nervous system sensory disorders, pulmonary toxicity, and hepatotoxicity * Corresponding author: Center for Cancer Pharmacology, University of Pennsylvania, 302 Abramson Research Building, Civic Center Blvd. and Osler Circle, Philadelphia, PA 19104-4318, (tel.) (215) 573-9885, (fax) (215) 573-9889, (e-mail) [email protected]. † Current address: Pharmaceutical Research Institute, Bristol-Myers Squibb, 5 Research Parkway, Wallingford, CT 06492. ‡ Department of Anesthesiology, College of Physicians and Surgeons of Columbia University, New York, NY 10032. (1) Fox, B. S.; Kantak, K. M.; Edwards, M. A.; Black, K. M.; Bollinger, B. K.; Botka, A. J.; French, T. L.; Thompson, T. L.; Shad, V. C.; Greenstein, J. L.; Gefter, M. L.; Exley, M. A.; Swain, P. A.; Briner, T. J. Nat. Med. (NY) 1996, 2, 1129. 10.1021/ac981060e CCC: $18.00 Published on Web 04/16/1999

© 1999 American Chemical Society

Figure 1. Cocaine and its primary metabolites.

can also occur.2 Cocaine is similar to amphetamine in that it has very high volume of distribution. It is also cleared very rapidly, which implies that cocaine toxicity is mediated by one of its metabolites.3 The major routes of in vivo biotransformation of cocaine can be described by two distinct biochemical processes, namely, hydrolysis and oxidation (Figure 1).4 Cocaine is oxidized by specific cytochromes P450 (CYPs) to norcocaine and to ecgonine methyl ester and benzoylecgonine by serum cholinesterases and liver esterases.5 Further oxidative metabolism of norcocaine gives N-hydroxy cocaine, whereas further hydrolysis of benzoylecgonine and ecgonine methyl ester leads to the formation of ecgonine. Approximately 86-90% of the administered dose of cocaine is recovered in urine, but only 1-5% is eliminated as the unchanged parent drug.6 N-Demethylation of cocaine by CYPs 3A and 2B in hepatic microsomes produces norcocaine, the only pharmacologically active primary metabolite.7-10 Ecgonine (2) Das, G. J. Clin. Pharmacol. Toxicol. 1993, 31, 521. (3) Figliomeni, M. L.; Abdel-Rahman, M. S. J. Appl. Toxicol. 1997, 17, 105. (4) Kloss, M. W.; Rosen, G.; Rauckman, E. J. Biochem. Pharmacol. 1984, 33, 169. (5) Kloss, M. W.; Rosen, G.; Rauckman, E. J. Mol. Pharmacol. 1983, 23, 482. (6) Peterson, K. L.; Logan, B. K.; Christian, G. D. Forensic Sci. Int. 1995, 73, 183. (7) Benuck, M.; Reith, M. E. A.; Sershen, H.; Wiener, H. L.; Lajtha, A. Proc. Soc. Exp. Biol. Med. 1989, 190, 7. (8) Boelsterli, U.; Lanzotti, A.; Goldlin, C.; Oertle, M. Drug Metab. Dispos. 1992, 20, 96.

Analytical Chemistry, Vol. 71, No. 10, May 15, 1999 2021

methyl ester and benzoylecgonine are of interest in pharmacological/toxicological studies because of their longer half-lives compared with that of cocaine itself.11-14 We required sensitive and specific assay methodology in order to gain insight into the in vivo disposition of cocaine and its major metabolites. The ability to perform such determinations would make it possible to assess in vivo pharmacokinetic parameters predicted by in vitro studies.15 Currently available quantitative methodology includes high-performance liquid chromatography (HPLC) with ultraviolet14,16-19 or fluorescence detection,12 gas chromatography (GC)/mass spectrometry (MS),11,13,18,20-22 timeof-flight (TOF)/MS,23 and liquid chromatography (LC)/MS.24,25 A wide variety of other analytical methods are also used for semiquantitative screening procedures such as enzyme-multiplied immunoassay,26,27 radioimmunoassay,28 thin-layer chromatography,29 and fluorescence-polarization immunoassay.26,27 A recent report has described the use of liquid chromatography (LC) coupled with electrospray (ESI)/MS/MS for qualitative determinations of cocaine and several metabolites.12 A major disadvantage of quantitative GC/MS methodology is the requirement for extensive sample cleanup followed by derivatization. Quantitative methodology based on HPLC and LC/MS also requires a significant amount of sample cleanup in order to attain reasonable limits of detection. Methodology based on TOF/ MS is relatively insensitive and subject to interference from the biological matrix. It was reasoned that LC/atmospheric pressure chemical ionization (APCI)/tandem mass spectrometry (MS/MS) would provide the opportunity to conduct high throughput assays at high sensitivity.30 We report a method for the simultaneous analysis of cocaine, norcocaine, benzoylecgonine, and ecgonine (9) Bornheim, L. M.; Everhart, E. T.; Li, J.; Correia, M. A. Biochem. Pharmacol. 1994, 48, 161. (10) Pellinen, P.; Honkakoski, P. S. F.; Niemitz, M.; Alhava, E.; Pelkonen, O.; Lang, M.; Pasanen, M. Eur. J. Pharmacol. 1994, 270, 35. (11) De La Torre, R.; Ortun ˜o, J.; Gonza´lez, M. L.; Farre´, M.; Cami, J.; Segura, J. J. Pharm. Biomed. Anal. 1995, 13, 305. (12) Clauwaert, K. M.; Van Bocxlaer, J. F.; Lambert, W. E.; Van den Eeckhout, E. G.; Lemiere, F.; Esmans, E. L.; De Leenheer, A. P. Anal. Chem. 1998, 70, 2336. (13) Thompson, W. C.; Dasgupta, A. Clin. Chem. 1994, 187. (14) Clauwaert, K. M.; Van Bocxlaer, J. F.; Lambert, W. E.; De Leenheer, A. P. Anal. Chem. 1996, 68, 3021. (15) Bornheim, L. M. Toxicol. Appl. Pharmacol. 1998, 158. (16) Pan, W.; Hedaya, M. A. J. Chromatogr., B 1997, 703, 129. (17) Ma, F.; Zhang, J.; Lau, C. E. J. Chromatogr., B 1997, 693, 307. (18) Phillips, D. L.; Tebbett, I. R. J. Anal. Toxicol. 1996, 20, 305. (19) Barat, S. A.; Kardos, S. A.; Abdel-Rahman, M. S. J. Appl. Toxicol. 1996, 16, 215. (20) Jindal, S. P.; Lutz, T. J. Anal. Toxicol. 1986, 10, 150. (21) Morrison, J. F.; Chesler, S. N.; Yoo, W. J.; Selavka, C. M. Anal. Chem. 1998, 70, 163. (22) Gerlits, J. J. Forensic Sci. 1993, 38, 1210. (23) Muddiman, D. C.; Gussev, A. I.; Martin, L. B.; Hercules, D. M. Anal. Chem. 1996, 68, 103. (24) Nishikawa, M.; Nakajima, K.; Tatsuno, M.; Kasuya, F.; Igarashi, K.; Fukui, M.; Tsuchihashi, H. Forensic Sci. Int. 1994, 66, 149. (25) Tatsuno, M.; Nishikawa, M.; Katagi, M.; Tsuchihashi, H. J. Anal. Toxicol. 1996, 20, 281. (26) Bogusz, M.; Aderjan, R.; Schmitt, G.; Nadler, E.; Neureither, B. Forensic Sci. Int. 1990, 48, 27. (27) Armbruster, D. A.; Schwarzhoff, R. H.; Hubster, E. C.; Liserio, M. K. Clin. Chem. 1993, 39, 2137. (28) Hornbeck, C. L.; Barton, K. M.; Czarny, R. J. J. Anal. Toxicol. 1995, 19, 133. (29) Wolff, K.; Sanderson, M. J.; Hay, A. W. Clin. Biochem. 1990, 27, 482. (30) Huang, E. C.; Wachs, T.; Conboy, J. J.; Henion, J. D. Anal. Chem. 1990, 62, 713A.

2022 Analytical Chemistry, Vol. 71, No. 10, May 15, 1999

methyl ester with low ng/mL limits of detection. This method, which is based upon the use of stable isotope dilution LC/APCI/ MS/MS, makes it possible for the first time to simultaneously monitor the pharmacokinetic characteristics of cocaine together with those of its three major metabolites following the intravenous administration of cocaine to rats. EXPERIMENTAL SECTION Apparatus. LC was performed on a Waters 2690 Separation Module equipped with an autosampler, a vacuum degasser, and a column heater (Waters, Milford, MA). This was coupled to a Finnigan TSQ 7000, a triple-stage quadrupole mass spectrometer fitted with an APCI source (Finnigan Corp., San Jose, CA). Evaporations under nitrogen were conducted in an N-evap analytical evaporator (Organomation, Berlin, MA). Samples for LC/MS analysis were filtered through CoStar SPIN-X microcentrifuge filters (0.2 µm) (Costar, Cambridge, MA). The microcentrifuge was from VWR Scientific products (Bridgeport, NJ) and the cooled centrifuge was a Beckman GS-6KR (Beckman, Fullerton, CA). LC was performed on a YMCbasic column obtained from YMC Inc. (Wilmington, NC). Precolumn filters (2 µm) were from Alltech, Deerfield, IL. The Alzet infusion pump (2002) was obtained from Alta Corporation, Palo Alto, CA. All standards and quality control (QC) solutions were prepared using certified volumetric flasks. Reagents. HPLC grade water was obtained from Fisher Scientific Co. (Fair Lawn, NJ). HPLC grade acetonitrile was purchased from Burdick and Jackson (Muskegon, MI). Ammonium acetate was obtained from J. T. Baker (Phillipsburg, NJ). Blank rat plasma was purchased from Biological Specialty Corporation (Lansdale, PA). Study samples were collected at the Department of Anesthesiology, College of Physicians and Surgeons of Columbia University (New York, NY). Standards for ecgonine methyl ester, benzoylecgonine, cocaine, and norcocaine and deuterated internal standards N-C[2H3]ecgonine methyl ester, N-C[2H3]benzoylecgonine, [2H5]cocaine ([2H5]benzoylecgonine methyl ester), and [2H5]norcocaine ([2H5]benzoylnorecgonine methyl ester) were obtained from the National Institute on Drug Abuse, Basic Neurobiology and Biological Systems Research Branch (Bethesda, MD). Preparation of Standard Solutions. A stock solution of each compound was prepared by dissolving 10 mg of the pure standard in 2 mL of an acetonitrile/water mixture (50:50, v/v). A mixture of working solution was prepared by diluting a known aliquot of all four stock solutions in acetonitrile/water (10:90, v/v). Drug free rat plasma from Male Sprague Dawley rats was spiked with the working solutions to make plasma standards at concentrations ranging from 2 to 1,000 ng/mL. Immediately after spiking, the plasma standards (250 µL) were transferred in aliquots to 1.8-mL Eppendorf tubes and frozen instantly in an acetone-dry ice mixture. The standards were stored at -80 °C until ready for use. All concentrations for ecgonine methyl ester, benzoylecgonine, cocaine, and norcocaine were expressed as free base. A stock solution of each deuterated internal standard was prepared by dissolving 2 mg of the deuterated compound in 2 mL of an acetonitrile/water mixture (50:50, v/v). A mixture of the internal standards was prepared by diluting an aliquot of all four stock solutions in acetonitrile/water (10:90, v/v) to give a final concentration of 2.5 µg/mL for each standard. An aliquot of

the internal standard solution (20 µL), corresponding to 50 ng of each deuterated compound, was used for each sample. Collection of Samples from Pharmacokinetic Study. Male Sprague-Dawley rats (400-450 g) were weighed and anesthetized with intraperitoneal ketamine (70 mg/kg) and xylazine (10 mg/kg), and then through a ventral neck incision the right internal jugular vein was cannulated using PE 60 tubing. This was connected to an Alzet infusion pump and the time was noted. The pump was then tunneled subcutaneously to lie at the nape and the incision was closed with sutures. No antibiotics were administered. The Alzet osmotic pump was primed 12 h earlier to administer saline continuously, over a 13-day period. Animals were then allowed to recover on a warming blanket, fed rat chow and water ad libitum, and kept on a 12-h light-dark cycle for 13 days. On day 14, the femoral artery and vein were catheterized using PE 50 tubing, and this was tunneled through to the nape and exteriorized under a plastic hood. The incisional wounds were sutured and animals recovered as before. They were then fed with water and rat chow ad libitum. The following day (day 15) the rats were weighed, and then the femoral artery catheter was connected to an oscilloscope for measurement of blood pressure after the arterial sample was taken (0.5 mL) and replaced with an equivalent volume of saline. Cocaine hydrochloride (2.5 mg/kg) diluted into 1 mL of saline was injected over 90 s. Arterial blood samples (0.5 mL) were taken at 0.5, 1, 2, 5, 10, 15, 30, 45, 60, 90, 120, and 150 min, respectively. The rats were euthanized under carbon dioxide. All blood samples were injected into heparinized Eppendorf tubes pretreated with saturated NaF solution (20 µL) and stored on ice. They were centrifuged within 2 h of sampling; the plasma was separated and stored at -80 °C, until LC/MS analysis. Sample Preparation for Analysis by Liquid Chromatography/Mass Spectrometry. The prespiked rat plasma standards and study samples were placed in a water/ice mixture and partially thawed with the “easy defrost” cycle for 2.5 min in a Whirlpool microwave oven. Each sample was vortex mixed and microcentrifuged at 14 000 rpm for 2 min. An aliquot (200 µL) of each sample was then transferred to a 15-mL screw-cap conical-bottom centrifuge tube. An aliquot of the internal-standard solution (20 µL) was added to each sample using a 25-µL Hamilton syringe. The samples were then vortex mixed for 10 s, and acetonitrile (700 µL) was added. Samples were again vortex-mixed for 10 s and centrifuged at 4 000 rpm for 10 min at 10 °C. The supernatant from each sample was transferred into a clean 15-mL screw-cap conical-bottom centrifuge tube, and formic acid (10 µL) was added. The solution was vortex mixed for 10 s and evaporated to dryness under nitrogen at 37 °C. The dried samples were reconstituted in 250 µL of water/acetonitrile (96:4, v/v), vortex mixed for 10 s, and centrifuged at 4 000 rpm for 10 min at 10 °C. The reconstituted solutions were then filtered through CoStar SPIN-X microcentrifuge filters in a microcentrifuge at 14 000 rpm for 4 min. The filtrates were transferred into 0.2-mL glass inserts contained within 2-mL auto-sampler vials. Fifty samples (standard-curve samples and QC samples) were analyzed for each day of validation. This number of samples was maintained in all subsequent analytical runs. Liquid Chromatography. Chromatographic separation of ecgonine methyl ester, benzoylecgonine, cocaine, and norcocaine

was achieved on a reversed-phase YMCbasic column (150 mm × 4.6 mm, i.d.: 5 µm) at a flow rate of 0.8 mL/min with the column oven set at 25 °C. A 2-µm precolumn filter was used to prevent column clogging. Solvent A consisted of 5 mM ammonium acetate/0.1% formic acid in water/acetonitrile (96:4, v/v), and solvent B consisted of 5-mM ammonium acetate/0.1% formic acid in water/acetonitrile (4:96, v/v). The LC conditions were as follows: 100% A for 2 min then a linear gradient to 10% A over 9.0 min, followed by a linear gradient to 100% A over 3 min and equilibration for 5 min. A set of 50 samples (50 µL of each) was run at one time with the autosampler maintained at 20 °C. Mass Spectrometry. LC/APCI/MS and MS/MS data were collected in the positive-ion mode with the vaporizer temperature at 550 °C and the heated capillary at 230 °C. Gas-phase chemical ionization was induced by a corona discharge needle (5 µA). The sheath gas (nitrogen) and auxiliary gas (nitrogen) pressures were set at 80 psi and 20 units, respectively. The protonated molecular ions (MH+) of cocaine and its metabolites were filtered through the first quadrupole, and collision induced dissociation (CID) was performed using argon as the collision gas at 2.5 mTorr in the second (Rf only) quadrupole. Specific product ions for cocaine and its metabolites were detected in the third quadrupole. Collision energies were optimized for each analyte, and ranged from -20 to -23 eV. Selected reaction monitoring (SRM) was performed on the transitions of m/z 200 f m/z 182 at -22 eV, m/z 290 f m/z 168 at -23 eV, m/z 304 f m/z 182 at -22 eV, and m/z 290 f m/z 168 at -20 eV for ecgonine methyl ester, benzoylecgonine, cocaine, and norcocaine, respectively. Deuterated analogues of cocaine and its metabolites were used as internal standards. SRM of the following transitions were performed for the internal standards: m/z 203 f m/z 185 at -22 eV, m/z 293 f m/z 171 at -23 eV, m/z 309 f m/z 182 at -22 eV, and m/z 295 f m/z 168 at -20 eV, for N-C[2H3]ecgonine methyl ester, N-C[2H3]benzoylecgonine, [2H5]cocaine, and [2H5]norcocaine, respectively. Unit resolution was maintained for the SRM analyses. The collision-energy offset for the deuterium labeled internal standard was the same as that used for the unlabeled analyte. SRM transitions were monitored in three distinct stages during the course of the LC run. Transitions for ecgonine methyl ester and its internal standard were monitored from 0.00 to 5.00 min, transitions for benzoylecgonine and its internal standard were monitored from 5.00 to 8.75 min, and transitions for cocaine and norcocaine and their internal standards were monitored from 8.75 to 19.00 min. Each of the three time periods employed tuning parameters, which were optimized for the ions being monitored at that particular time. Data Analysis. Calibration curves ranged from 2 to 1000 ng/ mL for benzoylecgonine, cocaine and norcocaine, and 5 to 1 000 ng/mL for ecgonine methyl ester. For each curve, 10 different concentrations distributed over the entire concentration range were used. Peak-height ratios between ecgonine methyl ester and its internal standard and peak-area ratios between benzoylecgonine, cocaine, and norcocaine and their respective internal standards were calculated for each concentration using Finnigan LCquan version 1.2 software. The data were fit to a linear leastsquares regression curve with a weighting index of 1/x. A water blank, plasma blank, and plasma blank spiked with internal standard were also processed with each calibration curve. PharAnalytical Chemistry, Vol. 71, No. 10, May 15, 1999

2023

Figure 2. Product ion spectra from CID of MH+ (A) ecgonine methyl ester (MH+ ) m/z 200) at a collision energy of -22 eV; (B) benzoylecgonine (MH+ ) m/z 290) at collision energy of -23 eV; (C) cocaine (MH+ ) m/z 304) at collision energy of -22 eV; and (D) norcocaine (MH+ ) m/z 290) at collision energy of -20 eV.

macokinetic parameters were determined using WinNonlin (SIS, Scientific Software, CA). Accuracy and Precision. Validation of the assay was performed with QC samples at four different concentrations for each analyte. Lower limit of quantitation (LLQ) QCs were 2 ng/mL for benzoylecgonine, cocaine, and norcocaine and 5 ng/mL for ecgonine methyl ester. The lower quality control (LQC) was 10 ng/mL, the middle quality control (MQC) was 100 ng/mL, and the high quality control (HQC) was 500 ng/mL for each analyte. Five replicates of each QC sample at each concentration level were processed and analyzed together with the 10 standard-curve samples. The LQC, MQC, and HQC samples were examined on three separate days, whereas LLQ QC samples were analyzed on 1 day only. Assay accuracy was assessed by comparing means of the measured ecgonine methyl ester, benzoylecgonine, cocaine, and norcocaine concentrations with the theoretical concentrations in the QC samples expressed as percentages. Within-day precision expressed percent relative standard deviation (% RSD) and was obtained by calculating the percent ratio between the relative standard deviation of five replicates (n ) 5) and their mean at each concentration within the same validation run. Interday precision, defined as percent relative standard deviation (% RSD) of three different validation runs (n ) 3) was also assessed. RESULTS AND DISCUSSION Liquid Chromatography/Atmospheric Pressure Chemical Ionization/Tandem Mass Spectrometry. The ability to monitor the plasma pharmacokinetics of cocaine and its metabolites has previously been limited by inadequate sensitivity and specificity. The two major problems encountered are, first, ecgonine methyl 2024 Analytical Chemistry, Vol. 71, No. 10, May 15, 1999

ester is very polar and difficult to retain on the reversed-phase column, and second, the concentration of norcocaine in control nontreated animals is very low.6 We have used the power of LC coupled with APCI and MS/MS to overcome these problems. MH+ for ecgonine methyl ester, benzoylecgonine, cocaine, and norcocaine was observed at m/z 200, m/z 290, m/z 304, and m/z 290, respectively. CID of these parent ions coupled with analysis by MS/MS, provided the product-ion spectra shown in Figure 2. Major product ions arose from loss of benzoic acid (cocaine, benzoylecgonine, and norcocaine) or water (ecgonine methyl ester). To provide maximal sensitivity, these intense product ions were selected for SRM analyses. A further improvement to sensitivity was made by monitoring the SRM transitions in three distinct stages during the course of the LC run. Each of the three time periods employed tuning parameters, which were optimized for those ions being monitored. Thus, there was a maximum of four transitions being monitored during any one time period during the LC run. The retention times for ecgonine methyl ester, benzoylecgonine, cocaine, and norcocaine were 2.66, 8.30, 9.21, and 9.27 min, respectively (Figure 3). The deuterated internal standards eluted from 0.005 to 0.010 min ahead (2H3 analogues) or 0.02 to 0.03 min ahead ([2H5] analogues) of the corresponding protium analogues. A typical LC/MS/MS chromatogram of the plasma sample obtained at the 10-min time point in one of the pharmacokinetic studies is shown in Figure 3. Ionization under APCI conditions results from the effect of a corona discharge.31,32 This provides several advantages over the (31) Brewer, E.; Henion, J. J. Pharm. Sci. 1998, 87, 395. (32) Matuszewski, B. K.; Constanzer, M. L.; Chavez-Eng, C. M. Anal. Chem. 1998, 70, 882.

Figure 3. LC/APCI/MS/MS chromatograms of cocaine and its metabolites, and corresponding deuterated internal standards in a rat plasma sample obtained at 10-min postinjection during a pharmacokinetic study. (A) ecgonine methyl ester, retention time 2.66 min (Upper), ([2H3]ecgonine methyl ester, retention time 2.66 min (Lower); (B) benzoylecgonine, retention time 8.30 min (Upper), ([2H3]-benzoylecgonine, retention time 8.30 min (Lower); (C) cocaine, retention time 9.22 min (Upper), ([2H5]-cocaine, retention time 9.19 min (Lower); and (D) norcocaine, retention time 9.27 min (Upper), ([2H5]-norcocaine, retention time 9.26 min (Lower). The chromatographic separation was obtained with a YMCbasic column (150 × 4.6 mm, 5 µm).

other API techniques of ionspray, turboionspray, and ESI for accurate and precise analyses of drugs and their metabolites in biological fluids. First, APCI can be employed using LC flow rates in the 0.5 to 1.0 mL/min range making it possible to use conventional (4.6-mm i.d.) LC columns. These columns are more rugged and permit faster reequilibration after gradient elution than the microbore (2.0-mm i.d.) columns normally used for other API methods. Second, greater volumes can be injected on column, which makes it easier to filter and transfer aqueous solutions containing the analytes. Third, in contrast to other API methodology, suppression of ionization by the constituents of the biological matrix is minimal.32 Therefore, for assays that avoid extraction of the biological matrix, APCI is clearly the method of choice. Mobile-phase additives are often required either to improve chromatography or to improve the sensitivity of APCI.33 For cocaine and its metabolites, we found that optimal chromatographic separations coupled with high MS sensitivity could be obtained by adding 5-mM ammonium acetate and 0.1% formic acid to the mobile phase. Assays based on LC/MS tend to employ short elution times and rely on the power of the mass spectrometer to provide specificity so that large numbers of samples can be analyzed. However, great care has to be taken to ensure that none of the metabolites or constituents from the biological matrix interferes in the analysis.31,32 For example, norcocaine and benzoylecgonine have identical MH+ and product ions (Figure 2). If poor chromatography was performed, these two metabolites would (33) Shaefer, W. H.; Dixon, F., Jr. J. Am. Soc. Mass Spectrom. 1996, 7, 1059.

interfere with each other. Similarly, there may be other compounds in the biological matrix which give rise to ions identical to those of the analytes. Such interfering substances may only be present in the clinical samples, so they would not be observed in control plasma. To eliminate such potential problems, we developed a gradient HPLC system that completely separates cocaine and its three major metabolites. The two potentially interfering compounds, norcocaine and benzoylecgonine, were separated by almost 1.0 min (Figure 3). Therefore, the assay had very high specificity coupled with high sensitivity. As noted above, ecgonine methyl ester is very polar and eluted with a relatively short retention time using a variety of different stationary phases. This raised the possibility that interference from the plasma matrix may result in poor reproducibility and accuracy. Indeed, two additional peaks were observed in the SRM chromatogram of the plasma control and that of the dosed rats. These peaks were, in fact, also observed in control water samples. No extra peaks have been observed in plasma samples from 10 different animals. In addition, no peaks have ever been detected with the same retention time as ecgonine methyl ester in untreated rat plasma samples. Therefore, we determined that the assay had adequate specificity for our pharmacokinetic determinations. Stable-isotope internal standards were employed in order to obviate any problems that could have arisen through suppression of ionization by endogenous contaminants or cocaine metabolites that may have been present in different plasma samples. Suppression of ionization would have resulted in a decrease of both Analytical Chemistry, Vol. 71, No. 10, May 15, 1999

2025

Table 1: Accuracy and Precision of Assay for LLQ Samples (n ) 5)

mean (ng/mL) accuracy (%) precision (%)

ecgonine methyl ester 5 ng/mL

benzoylecgonine 2 ng/mL

cocaine 2 ng/mL

norcocaine 2 ng/mL

4.3 90 5

2.0 100 5

1.9 95 5

1.9 95 4

Table 2: Accuracy and Precision of Assay for QC Samples (n ) 15) ecgonine methyl benzoylnorester ecgonine cocaine cocaine LQC 10 ng/mL

mean (ng/mL) accuracy (%) precision (%)

10.8 108 3.5

10.9 109 1.9

11.1 111.4 2.5

11.1 110.5 2.1

MQC 100 ng/mL mean (ng/mL) accuracy (%) precision (%)

100.5 100.5 1.8

92.6 92.6 1.4

95.3 95.3 0.7

93.3 93.3 1.0

HQC 500 ng/mL mean (ng/mL) accuracy (%) precision (%)

480.3 96.1 1.5

494.2 98.8 0.9

498.4 99.7 1.6

489.8 98.0 1.4

internal-standard and analyte signals. However, the ratio between the two signals would have remained the same. Standard Curves. The standard curves were plots of the ratios of analyte/internal standard responses (peak-height ratio for ecgonine methyl ester and peak-area ratios for benzoylecgonine, cocaine, and norcocaine) as a function of the analyte concentration. The concentration of the standards ranged from 5 ng/mL to 1 000 ng/mL for ecgonine methyl ester and 2 ng/mL to 1 000 ng/mL for benzoylecgonine, cocaine, and norcocaine. The means, standard deviations, and % RSDs were reproducible and demonstrated linearity for all of the 3 days of the assay validation. The data were fitted to a linear least-squares regression curve using a weighting index of 1/x. Equations of regression lines and correlation coefficients (r2) were as follows: y ) 0.0221 + 0.0037x (r2 ) 0.9991) for ecgonine methyl ester, y ) 0.0020 + 0.0041x (r2 ) 0.9992) for benzoylecgonine, y ) 0.0035 + 0.0038x (r2 ) 0.9994) for cocaine, and y ) 0.0040 + 0.0039x (r2 ) 0.9991) for norcocaine. The percentage error of back-calculated values for the 10 calibration standards from their actual values ranged from 88 to 112%. Specificity, Precision, and Accuracy. Water and plasma blanks devoid of analytes and internal standards showed no interfering signals at the appropriate retention times for cocaine, its metabolites, and its internal standards. Acceptable accuracy and precision were set at (20% for the LLQ QC and (15%, each, for the LQC, MQC, and HQC. The LLQ for this assay was determined to be 5 ng/mL for ecgonine methyl ester and 2 ng/ mL for benzoylecgonine, cocaine, and norcocaine (Table 1). The intraassay and interassay accuracy and precision (% RSD) for all four analytes of interest ranged from 92 to 109% and 0.8-4.1%, respectively (Table 2). For quantitative analyses, two LQCs, two MQCs, and two HQCs were included with the samples. The analyses were repeated if these samples were not within (15% of their theoretical values. The YMCbasic stationary phase used in this assay consists of several different alkyl silanes, with chain lengths of C8 and smaller, 2026 Analytical Chemistry, Vol. 71, No. 10, May 15, 1999

Figure 4. Pharmacokinetic profiles for cocaine and its metabolites. The concentrations of ecgonine methyl ester, benzoylecgonine, cocaine, and norcocaine were determined in rat plasma samples from cocaine treated rats following a chronic administration of saline over a 13-day period. Pharmacokinetic profiles are presented for (A) ecgonine methyl ester, (B) benzoylecgonine, (C) cocaine, and (D) norcocaine.

bonded to silica. Peak shapes and chromatographic reproducibility for cocaine and its metabolites were excellent if the column was maintained at 25 °C in a column heater. This was accomplished by adding 5-mM ammonium acetate and 0.1% formic acid in the mobile phase. When plasma samples were analyzed, the column effluent was diverted to waste for the first 1.85 min to prevent extraneous endogenous materials from entering the mass spectrometer. This maintained cleanliness of the source for extended periods. The addition of formic acid to the protein-free plasma proved to be absolutely critical as it prevented hydrolysis of cocaine and its metabolites. Ecgonine methyl ester proved to be particularly sensitive to hydrolysis in protein-free plasma that had not been acidified. Samples, which were reanalyzed after being stored at -80 °C for 2-3 days, gave identical analytical results when compared with the original analyses. Reported methods for the analysis of cocaine and its metabolites have demonstrated LLQs of 25 ng/mL.11,13,18,23,25,28 In contrast, the LC/APCI/MS/MS method, which we have developed, permits precise and accurate quantitation with LLQs that are at least an order of magnitude better. Analysis of Rat Plasma Samples from the Pharmacokinetic Study. The validated assay was used to examine the plasma pharmacokinetics of cocaine and its metabolites following an intravenous bolus dose of cocaine hydrochloride (2.5 mg/kg). Arterial-plasma samples were obtained at thirteen time points (predose and 0.5-150 min). The control rats had been pretreated with a saline infusion (vehicle) over the previous 13 days. These experiments were designed to lay the foundations for future studies in which various potential modulators of cocaine metabolism could be examined. The pharmacokinetic profile for cocaine (Figure 4) showed a maximum plasma concentration (Cmax) of 975.6 ng/mL at 1 min postinjection. There was a time-dependent

decrease in the cocaine concentrations over the next 150 min. In contrast, the plasma concentration of benzoylecgonine increased until Cmax of 183.9 ng/mL was observed 30 min after the cocaine injection. Ecgonine methyl ester reached a Cmax of 27.9 ng/mL at 10 min and norcocaine a Cmax of 3.4 ng/mL after 10 min (Figure 4). The cocaine data was fitted to a two-compartment model, because on a semilogarithmic plot the plasma concentrations arranged themselves as a biexponential function. There were two distinct half-lives for cocaine, t1/2R ) 4.38 min and t1/2β ) 31.14 min. The area under the plasma concentration-time curve (AUC) for cocaine was 13 924 ng/mL‚h and its clearance was 0.16 L/min. Cocaine exhibited a very high volume of distribution of 2.212 L/kg. AUCs for ecgonine methyl ester, benzoylecgonine, and norcocaine were, 2 029 ng/mL‚h, 18 171 ng/mL‚h, and 101 ng/mL‚h, respectively. The pharmacokinetic profiles of cocaine and its metabolites obtained with this LC/MS/MS methodology will make it possible to examine how pharmacologic manipulations affect cocaine metabolism. CONCLUSIONS Numerous studies have been conducted in the past in order to establish the relationship between cocaine metabolism and toxicity. However, it is only recently that significant progress has been made in understanding some of the factors involved in inducing cocaine metabolism by the use of in vitro techniques.15 The ability to translate these in vitro studies to the in vivo situation is seriously hampered by the lack of suitable methodology for

quantitation of cocaine and its metabolites in plasma. Using a combination of APCI and LC/MS, we have overcome this problem. A chromatographic technique has been developed which permits the separation of ecgonine methyl ester, benzoylecgonine, cocaine, and norcocaine. It was possible, for the first time, to analyze with high precision and accuracy ecgonine methyl ester (a very polar metabolite) in the presence of norcocaine (a relatively nonpolar metabolite). Highly reproducible retention times were possible under the assay conditions that were employed. As shown in Figure 3, an excellent separation of ecgonine methyl ester, benzoylecgonine, cocaine, and norcocaine was obtained with a total run time of 19 min (allowing for column reequilibration). Therefore, it was possible to analyze 50 plasma samples and standards for four analytes in a single overnight run. Minimal sample workup was necessary, and no derivatization was required so that large numbers of samples could be readily processed. The ruggedness, sensitivity, and specificity of the assay highlight the utility of APCI for routine determinations of drugs and their metabolites. The assay will provide the foundation for detailed in vivo studies in which pharmacological manipulation of cocaine metabolism will be assessed. It will also be useful for performing conventional toxicology studies in human subjects.

Received for review September 23, 1998. Accepted February 8, 1999. AC981060E

Analytical Chemistry, Vol. 71, No. 10, May 15, 1999

2027