Rapid Screening of Glutathione-Trapped Reactive Metabolites by

Jun 13, 2008 - In this approach, an isotope pattern-dependent scanning method was applied to ... Recorded full-scan MS and MS/MS data sets were furthe...
0 downloads 0 Views 552KB Size
Download PDF
Chem. Res. Toxicol. 2008, 21, 1477–1483

1477

Rapid Screening of Glutathione-Trapped Reactive Metabolites by Linear Ion Trap Mass Spectrometry with Isotope Pattern-Dependent Scanning and Postacquisition Data Mining Li Ma, Bo Wen,† Qian Ruan, and Mingshe Zhu* Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Research and DeVelopment, Princeton, New Jersey 08543, ReceiVed March 19, 2008

The present study describes a novel integrated approach for rapid analysis of reactive metabolites with a linear ion trap mass spectrometer (LTQ). In this approach, an isotope pattern-dependent scanning method was applied to the data acquisition of glutathione (GSH)-trapped reactive metabolites. Recorded fullscan MS and MS/MS data sets were further processed with neutral loss filtering, product ion filtering, and extracted ion chromatographic analysis to search for protonated molecules and MS/MS spectra of GSH adducts. To evaluate the effectiveness and reliability of the approach, GSH adducts of carbamazepine, diclofenac, 4-ethylphenol, acetaminophen, p-cresol, and omeprazole were analyzed, which were formed in human liver microsome incubations fortified with a mixture of nonlabeled GSH and stable isotopelabeled GSH at a 1:0.8 ratio. Results demonstrate that the combination of the isotope pattern-dependent scanning with the postacquisition data mining was very effective in detecting low levels of GSH adducts, regardless of their fragmentation patterns. As compared to a neutral loss scanning method performed with a triple quadrupole mass spectrometer, the LTQ-based approach had several major advantages, including the superior selectivity and sensitivity in detecting different classes of GSH adducts and the higher throughput capability of the detection and MS/MS spectral acquisition of GSH adducts in a single LC/MS run. Overall, this analytical approach provides a simple and efficient means for screening for reactive metabolites using a linear ion trap LC/MS platform. Introduction Drug metabolite profiling using a variety of liquid chromatography-mass spectrometry (LC/MS) techniques has become an integral part of drug metabolism and pharmacokinetics research in the discovery process (1–4). In particular, the determination of metabolic soft spots, the sites of metabolic modifications that play a major role in drug clearance in vitro and in animals, can provide valuable information for designing more metabolically stable drug candidates. Traditionally, precursor ion (PI)1 and neutral loss (NL) scanning analyses with a triple quadrupole mass spectrometer are employed for the detection and structural characterization of drug metabolites in complex biological matrices (1). For analyzing oxidative metabolites, the MS/MS spectrum of a parent drug is predetermined, and selected product ions or NLs are incorporated into acquisition methods prior to NL or PI scanning (5). For detecting conjugated metabolites, NL or PI scanning is performed by following their common fragmentation pathways, such as NL of 176 Da for glucuronides. Once metabolites are detected by NL or PI scanning, an additional LC/MS run with product ion scanning is performed to acquire their MS/MS spectra for structural elucidation. NL and PI scanning methods are especially useful in searching for uncommon or unpredicted metabolites. Recently, an updated * To whom correspondence should be addressed. Tel: 609-252-3324. E-mail: [email protected]. † Present address: Drug Metabolism and Pharmacokinetics, M/S S3-2-E 218A, Roche Palo Alto, Palo Alto, CA 94304. 1 Abbreviations: CID, collision-induced disassociation; EIC, extracted ion chromatography; GSH, glutathione; HLM, human liver microsomes; + MH , protonated molecule; NL, neutral loss; NLF, neutral loss filtering; PI, percursor ion; PIF, product ion filtering; TIC, total ion chromatogram.

triple quadrupole instrument (TSQ Quantum Ultra) has implemented a new scanning function that uses a NL or PI scan as the survey scan to trigger MS/MS acquisition, which would significantly improve the throughput of the NL or PI scan analysis. Alternatively, full-scan MS-based data-dependent MS/MS acquisition with ion trap mass spectrometers, including threedimensional ion traps (6), linear ion traps, and hybrid linear ion trap quadrupole instruments (7), has been widely employed for the detection of common metabolites. Two types of datadependent scanning functions, list-dependent and intensitydependent scans, are often used in drug metabolite profiling and identification. In the list-dependent MS/MS experiment (8), fullscan MS analysis serves as a survey scan to search for metabolite ions listed in an acquisition method. Once a listed metabolite ion is found in the survey scan, MS/MS acquisition of the metabolite is automatically triggered. In the intensity-dependent MS/MS experiment (9), MS/MS acquisition is carried out for most intense ion(s) detected by the survey scanning. In addition, an isotope pattern data-dependent MS/MS scanning method was developed for selective data acquisition of analytes that display a distinct isotope pattern, such as bromine-containing compounds (10). However, this scanning method has not been widely employed in drug metabolism research due to a limited number of pharmaceutical compounds with one or more chlorine or bromine atom(s). Ion trap instruments feature a rapid scanning speed, which allows the sensitive detection of metabolites and the acquisition of their MSn spectra in a single LC/MS run. In addition to the determination of metabolic soft spots, the screening of glutathione (GSH)-trapped reactive metabolites formed in human liver microsome (HLM) incubations is often

10.1021/tx8001115 CCC: $40.75  2008 American Chemical Society Published on Web 06/13/2008

1478

Chem. Res. Toxicol., Vol. 21, No. 7, 2008

conducted in the selection and optimization of lead compounds in the drug discovery process (11–14). NL scanning in the positive ion mode with a triple quadrupole mass spectrometer is traditionally used for detecting GSH adducts (15–17). In the NL analysis, GSH adducts are monitored for a NL of 129 Da, which is a common primary fragmentation pathway of most GSH adducts. The NL experiment can be rapidly performed using a generic acquisition protocol without prior knowledge of the fragmentation of a test compound or the prediction of molecular weights of potential drug-GSH adducts. Alternatively, a PI scanning method that monitors an anion at m/z 272 in the negative electrospray ionization mode can be employed for screening for GSH adducts (18). The PI scanning has been shown to be more sensitive and selective in detecting different classes of GSH adducts than NL scanning analysis (19). One major disadvantage of using these NL and PI methods for detection and characterization of GSH adducts is a requirement for multiple-step processes. In the first LC/MS run, GSH adducts are monitored by NL or PI scanning. Then, each peak present in the corresponding ion chromatograms is manually inspected. Once protonated molecules of potential GSH adduct peaks are identified, their MS/MS spectra are acquired using specific product ion scanning methods in a second LC/MS run. Recently, hybrid quadrupole-linear ion trap mass spectrometry was employed to improve the throughput of PI scanning analysis of GSH adducts, in which the PI scanning of m/z 272 in the negative ion mode serves the survey scan that triggers dependent MS/MS acquisition in the positive ion mode with polarity switching (20, 21). There are relatively few literature reports of using ion trap instruments for the rapid screening of reactive metabolites. An intensity-dependent scanning method on an LTQ instrument was applied to the data acquisition of reactive metabolites trapped by stable isotope labeled GSH (22). Recorded MS/MS data were further processed with NL filtering of 129 Da. The study demonstrated the feasibility of NL filtering of MS/MS data in detecting GSH adducts and the utility of using stable isotopelabeled trapping agents in identifying false positive peaks. The main objective of the current study was to develop and evaluate a new linear ion trap mass spectrometry-based analytical strategy for rapid screening of reactive metabolites in which isotope pattern data-dependent acquisition and postacquisition data mining techniques were applied. The effectiveness of this approach was examined by analyzing GSH adducts of model compounds, carbamazepine, diclofenac, acetaminophen, 4-ethylphenol, p-cresol, and omeprazole, which are known to form reactive metabolites in HLMs. Results were also compared with those from NL scanning of 129 Da with a triple quadrupole mass spectrometer.

Experimental Procedures Materials. Pooled HLMs were purchased from BD Biosciences (Woburn, MA). GSH, NADPH, acetaminophen, carbamazepine, diclofenac, p-cresol, 4-ethylphenyl, and omeprazole were purchased from Sigma-Aldrich (St. Louis, MO). Stable isotope-labeled GSH (γ-glutamylcysteinylglycine-13C2-15N) was purchased from Cambridge Isotope Laboratories (Andover, MA). Microsomal Incubations. Test compounds (50 µM) were incubated separately with pooled HLM (1.0 mg/mL), GSH (1 mM), or a mixture of GSH and 13C215N-GSH (a 1:0.8 ratio, 1 mM total) and NADPH (1 mM) in potassium phosphate buffer (100 mM, pH 7.4) for 30 min. The total incubation volume was 1 mL. The biotransformation reactions were initiated by the addition of the NADPH solution after a 3 min preincubation and were stopped by the addition of 150 µL of trichloroacetic acid (10%). After

Ma et al. centrifugation, supernatants were loaded onto solid-phase extraction cartridges (Oasis extraction cartridges, Waters Corp., Milford, MA). The cartridges were rinsed with 1 mL of water and then eluted with 2 mL of methanol. The methanol fractions were dried and reconstituted with 200 µL of 30% methanol. Aliquots (20 µL) of the reconstituted solutions were injected into the LC/MS/MS system. Linear Ion Trap LC/MS. An LTQ linear ion trap mass spectrometer (Thermo Electron, San Jose, CA) was coupled with a Agilent 1100 HPLC system (Agilent Technologies, Palo Alto, CA) and a Zorbax SB C-18 HPLC column (2.1 mm × 150 mm, 5 µm, Agilent Technologies). Mobile phase A was formic acid in water (0.1%), and mobile phase B was acetonitrile (100%). The gradient started with 2% B for 2 min, ramped linearly to 40% B over 18 min and then to 90% B in 2 min, and held at 90% B over 5 min, followed by return to the initial condition over 2 min. The HPLC flow rate was 0.3 mL/min. The LTQ instrument was operated in the positive ion electrospray mode at an ion spray voltage of 4.0 kV and a source temperature of 250 °C. An isotope datadependent acquisition was performed using Xcalibur softwere, in which full-scan MS analysis from 300 to 800 Da served as the survey scan. MS/MS acquisition was triggered when an isotope doublet with a mass difference of 3 Da and a ratio of 1:0.8 was detected in the survey scan. Match tolerance for the intensities of the doublet ions was 0.15. The intensity threshold for MS/MS acquisition was 500. The normalized collision energy was 35%, and the isolation width was 1.5 Da. Triple Quadrupole LC/MS. The HPLC system consisted of Shimadzu LC10ADvp bipumps (Columbia, MD) and an HTC PAL autosampler (Leap Technologies, Cary, NC). The HPLC column and mobile phases used were the same as those described above. The initial mobile phase composition was 2% B for 2 min, ramped linearly to 90% B over 27 min, and then held at 90% B over 5 min, followed by return to the initial condition over 2 min. The HPLC flow rate was 0.25 mL/min. The HPLC system was interfaced to a Finnigan TSQ quantum triple quadrupole mass spectrometer (Thermo Electron). The MS/MS analyses were operated in the positive ion electrospray mode. The spray voltage was set at 4.0 kV, and the capillary temperature was set at 350 °C. NL scans of 129 and 307 Da were performed at a scan range from 350 to 800 Da using argon as the collision gas and a collision energy set at 26 eV.

Results Analytical Strategy. The integrated approach applied to the detection and structural characterization of in vitro reactive metabolites using isotope pattern-dependent acquisition on a linear ion trap mass spectrometer and multiple postacquisition data mining techniques is illustrated in Figure 1. In this approach, a mixture of GSH and stable isotope-labeled GSH at a ratio of approximately 1:1 was employed for trapping reactive metabolites formed in the HLM incubations. Data acquisition was carried out using a generic, simple isotope pattern-dependent scanning method with full-scan MS scanning as a survey scan and MS/MS acquisition as a dependent scan. Once a GSH adduct with a predefined isotope pattern was detected by the survey scan, the acquisition of its product ion spectrum was triggered. As a result, a full-scan MS data set of all analytes and an MS/MS data set acquired by the data-dependent scanning were recorded in a single injection. GSH adduct ions were identified mainly via postacquisition processing of the MS/MS data set with neutral loss filtering (NLF) and PIF techniques and further confirmed by their MS isotope signatures displayed in full-scan MS spectra. Structures of the detected GSH adducts were characterized based on molecular weights of the adducts and the interpretation of their MS/MS spectra. Analysis of GSH Adducts of Carbamazepine by Linear Ion Trap LC/MS. The total ion chromatogram (TIC) of MS/ MS spectral data recorded from the isotope pattern-dependent

GSH Adduct Screening by LTQ LC/MS

Chem. Res. Toxicol., Vol. 21, No. 7, 2008 1479

Figure 1. Integrated analytical strategy for detection and identification of reactive metabolites trapped by stable isotope GSH using linear ion trap LC/MS. Full-scan MS and MS/MS data sets are acquired with a generic isotope pattern-dependent method. NLF, PIF, and EIC techniques are employed for postacquisition data mining.

Figure 3. Mass spectra of carbamazepine-GSH adducts acquired by the isotope pattern-dependent scanning using the LTQ instrument. (A) Full-scan MS spectrum of CM2. (B) Full-scan MS spectrum of CM8. (C) MS/MS spectrum and proposed structure of CM2.

Figure 2. Analysis of carbamazepine-GSH adducts formed in an HLM incubation with stable isotope GSH by the LTQ mass spectrometry. (A) TIC of MS/MS data set. (B) Ion chromatogram obtained from NLF (129 Da) of MS/MS data. (C) EIC for m/z 560. (D) EIC for m/z 558. (E) EIC for m/z 556. (F) EIC for m/z 576. Each EIC was obtained by processing full-scan MS data set.

scanning of an HLM incubation of carbamazepine with stable isotope-labeled GSH revealed nine GSH adducts, CM1-CM9, along with a few minor peaks associated with endogenous components or background noise (Figure 2A). NLF (loss of 129 Da) of the MS/MS data file significantly reduced the false positive peaks, resulting in a clearer ion chromatogram (Figure 2B), from which the protonated molecules and MS/MS spectra of adduct peaks were directly retrieved (Figure 3C). Conse-

quently, extracted ion chromatograms (EIC) of these carbamazepine adducts (Figure 2B-F) were generated via processing the full-scan MS data file, which separated overlapped adduct peaks, such as CM2 (15.83 min, Figure 2C) and CM3 (15.86 min, Figure 2F). Full-scan MS spectra of carbamazepine-GSH adducts were recovered directly from the EIC, which exhibited a unique isotope doublet with a mass difference of 3 Da (Figure 3A,B). Carbamazepine adducts, CM1, CM2, and CM4, had the same protonated molecule at m/z 560 (protonated carbamazepine + GSH + O) (Figures 2C and 3A) and similar MS/MS spectra (Figure 3C), indicating that these adducts were formed after an epoxidation and ring opening by a nucleophilic attack of GSH. CM5, CM6, and CM7 had the same protonated molecule at m/z 558 (protonated carbamazepine + GSH + O - 2H) (Figure 2D) and similar MS/MS spectra (data not shown), suggesting that an oxygen atom was incorporated into a reactive intermediate prior to GSH trapping. CM8 and CM3 had the same protonated molecule at m/z 576 (protonated carbamazepine + GSH + 2O) (Figure 2F), suggesting that both adducts may be

1480

Chem. Res. Toxicol., Vol. 21, No. 7, 2008

Ma et al.

Figure 4. Analysis of diclofenac-GSH adducts formed in an HLM incubation with stable isotope GSH by the LTQ instrument. (A) TIC of full-scan MS data. (B) TIC of MS/MS data. (C) Ion chromatogram obtained from NLF (129 Da) of MS/MS data.

formed via an expoxide derived from monohydroxylated carbamazepine. The carbamazepine adducts, CM1-CM7, were also detected and structurally characterized in HLM in previous studies, in which the structures and formation pathways of these adducts were discussed (19, 23). Analysis of GSH Adducts of Diclofenac by Linear Ion Trap LC/MS. The TIC of the MS/MS data acquired for an HLM incubation of diclofenac with stable isotope-labeled GSH displayed two GSH adducts (DM1 and DM2) and some interference peaks (Figure 4B). The processing of the MS/MS TIC with NLF of 129 Da resulted in a chromatogram with less false positive peaks (Figure 4C), from which the MS/MS spectra of DM1 (m/z 583) and DM2 (m/z 617) were retrieved and are shown in Figure 5B and D, respectively. Furthermore, extracted ion chromatograms of the diclofenac adducts were generated from the full-scan MS data file, which clearly displayed DM1 and DM2 peaks (data not shown). The full-scan MS spectra of DM1 (Figure 5A) and DM2 (Figure 5C) displayed the characteristic isotope doublets with a mass difference of 3 Da. Both DM1 and DM2 were also identified in HLM incubations previously (23–25). Analysis of GSH Adducts of 4-Ethylphenol by Linear Ion Trap LC/MS. The TIC of data-dependent MS/MS analysis of an HLM incubation of 4-ethylphenol with stable isotopelabeled GSH revealed two GSH-trapped reactive metabolites, EM1 and EM2 (Figure 6A). The data mining of the MS/MS data set with NLF for 129 Da clearly revealed EM2 but not EM1 (Figure 6B). In contrast, the processing with PIF for m/z 308 was able to detect EM1 but not EM2 (Figure 6C). On the basis of their MS/MS spectra, EM1 was tentatively identified as a benzylic thioether GSH adduct (Figure 6E), and EM2 was tentatively determined as an aromatic GSH adduct (Figure 6D).

Figure 5. Mass spectra of diclofenac-GSH adducts acquired by the isotope pattern-dependent scanning using the LTQ instrument. (A) Fullscan MS spectrum of DM1. (B) MS/MS spectrum and proposed structure of DM1. (C) Full-scan MS spectrum of DM2. (D) MS/MS spectrum and proposed structure of DM2.

The same GSH adducts were previously identified in HLM incubations (23). Analysis of GSH Adducts of Acetaminophen, p-Cresol, and Omeprazole by Linear Ion Trap LC/MS. NLF (loss of 129 Da) of the MS/MS data set acquired for the acetaminophen incubation sample revealed a GSH adduct, AM2 (Figure 7B). On the basis of its protonated molecule (m/z 457, drug + GSH - 2H) and MS/MS spectrum (data not shown), AM2 was tentatively identified as 3′-glutathion-S-yl-acetaminophen, which was previously observed as the single major GSH adduct of acetaminophen in HLM incubations (16, 26, 27). MS/MS data processing of the p-cresol sample with NLF for 129 Da revealed two GSH adducts, PM1 (m/z 414, drug + GSH - 2H) and PM2 (m/z 430, drug + GSH + O - 2H) (Figure 7C). The structures and formation pathways of these adducts were previously discussed in the literature (23). The NLF-processed profile of the omeprazole incubation sample displayed two major GSH adducts, RM1 (m/z 635, drug + GSH - O - 2H) and RM2 (m/z 621, drug + GSH - O - CH2- 2H), along with multiple minor GSH adducts (Figure 7D). All of these adducts displayed the unique isotope douplet in their full-scan MS spectra (data not shown). MS/MS spectra and proposed structures of RM1 and RM2 are presented in Figure 8. As previously reported, omeprazole undergoes an acid-catalyzed chemical rearrangement reaction to form a reactive sulfenic acid that covalently binds to thiol groups in proteins by losing an oxygen atom (28). Therefore, it is tentatively proposed that omeprazole and its

GSH Adduct Screening by LTQ LC/MS

Chem. Res. Toxicol., Vol. 21, No. 7, 2008 1481

Figure 7. Analysis of GSH adducts formed in HLM incubations with stable isotope GSH by the LTQ instrument. Full-scan MS and MS/MS data files were acquired by the isotope pattern-dependent scanning. (A) TIC of MS/MS data set for an acetaminophen incubation sample. (B) Ion chromatogram obtained from NLF (129 Da) of the acetaminophen MS/MS data set. (C) Ion chromatogram obtained from NLF (129 Da) of the p-cresol MS/MS data file. (D) Ion chromatogram obtained from NLF (129 Da) of the omeprazole MS/MS data set. Figure 6. Analysis of 4-ethylphenyl-GSH adducts in an HLM incubation with stable isotope GSH by the LTQ instrument. (A) TIC of MS/MS data set. (B) Ion chromatogram obtained from NLF (129 Da) of MS/MS data set. (C) Ion chromatogram obtained from PIF (m/z 308) of MS/MS data set. (D) MS/MS spectrum and proposed structure of EM2. (E) MS/MS spectrum and proposed structure of EM1.

5-demethylated metabolite, a major metabolite of omeprazole in HLM (29), were converted to sulfenic acid intermediates that led to RM1 and RM2, respectively. The formation of RM1 and RM2 was significantly reduced in an HLM incubation sample that was not treated with trichloroacetic acid for stopping the incubation reaction, suggesting that the formation of RM1 and RM2 was mediated by the acid-catalyzed rearrangement reaction. Analysis of GSH Adducts of Diclofenac, Carbamazepine, Acetaminophen, and p-Cresol by Triple Quadrupole LC/ MS. For comparison purposes, diclofenac, carbamazepine, acetaminophen, and p-cresol were incubated with HLM in the presence of nonlabeled GSH. GSH-trapped reactive metabolites were analyzed by NL scanning of 129 Da with a triple quadrupole instrument. As a result, carbamazepine adducts (CM2-CM4, CM6, and CM7) (Figure 9A), diclofenac adducts (DM1 and DM2) (Figure 9B), an acetaminophen adduct (AM2) (Figure 9C), and p-cresol adducts (PM1 and PM2) (Figure 9D) were detected in these HLM incubations. In addition to the GSH adducts, several significant false positive peaks were also present in the NL ion chromatograms (Figure 9).

Discussion Linear ion trap LC/MS is one of the most often used LC/MS platforms for drug metabolite profiling (8, 30). However,

because of the lack of the triple quadrupole-based scanning functions, such as NL and PI scans, applications of ion trap instruments for rapid screening of GSH adducts are not common in the literature (14, 17). The ultimate goal of this study was to develop and validate an integrated analytical approach that enables the LTQ linear ion trap instruments to be utilized routinely for screening for GSH-trapped reactive metabolites in the drug discovery process. The analytical approach illustrated in Figure 1 takes advantages of the high sensitivity of full-scan MS analysis and capability of data-dependent MS/MS acquisition of linear ion trap instruments. Most importantly, the isotope pattern-dependent scanning analysis, which is a standard scanning function for most ion trap instruments, was successfully applied to selective and sensitive collection of MS/MS data of GSH adducts. As demonstrated in the analyses of GSH adducts of carbamazepine (Figures 2A), diclofenac (Figure 4B), 4-ethylphenol (Figure 6A), and acetaminophen (Figure 7A), the MS/ MS TICs of these incubation samples displayed GSH adducts as major components with little or no interference peaks. Interestingly, the isotope pattern-dependent acquisition was also effective in detecting diclofenac adducts, DM1 and DM2, which contained one and two chlorine atom(s), respectively (Figure 5A,C). In contrast, the TIC of MS/MS data of an acetaminophen incubation sample acquired with intensity-dependent scanning generated many false positive peaks (22). The isotope patterndependent scanning analysis provides a selective method for detection of GSH adducts using a linear ion trap instrument. The MS/MS data sets acquired with the isotope patterndependent scanning method can be further processed with NLF and PIF data-mining techniques. As demonstrated in the analyses of carbamazepine (Figure 2B vs A) and diclofenac adducts

1482

Chem. Res. Toxicol., Vol. 21, No. 7, 2008

Ma et al.

Figure 9. NL scanning (129 Da) of GSH adducts formed in HLM incubations with GSH by a triple quadrupole instrument. (A) TIC of the NL scanning of a carbamazepine incubation sample. (B) TIC of the NL scanning of a diclofenac incubation sample. (C) TIC of the NL scanning of an acetaminophen incubation sample. (D) TIC of the NL scanning of a p-cresol incubation sample.

Figure 8. Mass spectra of omeprazole-GSH adducts acquired by the isotope pattern-dependent scanning using the LTQ instrument. (A) Fullscan MS spectrum of RM1. (B) MS/MS spectrum and proposed structure of RM1. (C) Full-scan MS spectrum of RM2. (D) MS/MS spectrum and proposed structure of RM2.

(Figure 4C vs B), the NLF (loss of 129 Da) of the MS/MS data sets almost completely eliminated minor false positive peaks present in the TIC of the MS/MS data. In contrast, ion chromatograms from the NL scanning of GSH adducts with a triple quadrupole exhibited many more false positive peaks (Figure 9). Although the use of stable isotope-labeled GSH can help one to identify these false positive components, a timeconsuming process with visual inspection of each of peaks present in NL chromatograms is required (16). TICs of fullscan MS data sets acquired for individual incubations usually did not display GSH adducts as shown in the diclofenac example (Figure 4A). However, it can be utilized to generate extracted ion chromatograms (Figure 2C-F) and full-scan MS spectra (Figures 3A,B, 5A,C, and 8A,C) based on detected or predicted GSH adducts, which are very valuable for the confirmation of GSH adducts and elimination of false positive peaks. For example, a minor diclofenac adduct, DM2, was revealed from the data mining with NLF of 129 Da (Figure 4C). The fullscan MS spectrum of DM2 displayed a 3 Da isotope doublet, unambiguously confirming the protonated molecule (m/z 617) of DM2 (Figure 5C). Overall, the superior selectivity of the integrated approach was contributed by multiple factors, including the isotope pattern-dependent MS/MS acquisition, data processing with NLF and PIF, and the unique isotope pattern of GSH adducts shown in full-scan MS spectra. In addition to the selectivity, the sensitivity in detecting GSH adducts using the linear ion trap-based integrated approach was slightly better than NL scanning with a triple quadrupole instrument.

For example, the NLF analysis of the diclofenac sample with an LTQ instrument clearly revealed diclofenac adducts including the minor GSH adduct, DM2 (Figure 4C), while these adduct peaks were barely above the background noise levels with NL scanning on a triple quadrupole instrument (Figure 9B). The detection of the major GSH adducts of diclofenac in HLM under the current experiment conditions suggests that the LTQ-based method had relatively good sensitivity since diclofenac formed very low levels of protein covalent binding in HLM incubations (11). An additional example was the analysis of carbamazepine adducts. As shown in Figure 2B, nine GSH adducts of carbamazepine were discovered by NLF processing of MS/MS data acquired with an LTQ instrument. In contrast, the NL scanning analysis with a triple quadrupole instrument only detected five carbamazepine-GSH adducts (Figure 9A). The superior sensitivity of the LTQ-based approach may be contributed by the high sensitivity of full-scan MS analysis of the linear ion trap mass spectrometry and the high selectivity of the isotope pattern-dependent scanning and NLF. This study also demonstrates that the integrated approach is suitable for the detection of various classes of GSH adducts as illustrated in the analysis of the 4-ethylphenol sample. In the HLM incubation, 4-ethylphenol was converted to a quinone methide intermediate leading to the formation of a benzylic thioether GSH adduct, EM1. Like most benzylic thioether GSH adducts (18), EM1 did not afford the NL of 129 Da as a primary fragmentation pathway upon collision-induced disassociation (CID) (Figure 6E). Therefore, NL scanning of 129 Da was not effective in detecting this GSH adduct as reported in a previous study (23). In the current study, the isotope pattern-dependent scanning method was able to trigger the MS/MS acquisition of GSH adducts, including EM1, regardless of their fragmentation patterns. Furthermore, the postacquisition data mining enabled the use of multiple NLs and product ions to search for GSH adducts, some of which were more effective in detecting certain classes of GSH adducts (17, 18). For example, the PIF of m/z 308 or the NLF of 307 Da can be applied to the specific

GSH Adduct Screening by LTQ LC/MS

detection of benzylic thioether GSH adducts, such as EM1 (Figures 6C). As compared to NL or PI scanning methods in the literature, the most attractive advantage of the isotope pattern-dependent analysis with an LTQ instrument is its higher throughput analysis capability. As illustrated in Figure 1, the integrated approach uses a generic data-dependent acquisition method regardless of fragmentation patterns and molecular weights of GSH adducts. Full-scan MS and MS/MS spectral data sets of GSH adducts present in an incubation sample are recorded without multiple LC/MS injections. Therefore, a relatively complete mass spectral data set of a large number of samples can be acquired by the LTQ instrument in a higher throughput fashion. At the same time, recorded full-scan MS and MS/MS data can be processed with NLF, PIF, and EIC techniques for discovery and structural elucidation of GSH adducts. In summary, the present study describes a novel integrated approach for rapid detection and characterization of GSHtrapped reactive metabolites using an LTQ linear ion trap mass spectrometer (Figure 1). The results from this study demonstrate that this linear ion trap-based integrated approach has several significant advantages over the traditional triple quadrupolebased NL scanning method: (i) much higher selectivity for detecting GSH adducts, which is accomplished by employing a combination of isotope pattern-dependent data acquisition and NL and PI filtering of MS/MS data; (ii) more sensitive in the detection of GSH adducts than NL scanning of 129 Da; (iii) highly capable of detecting a variety of classes of GSH adducts regardless of their fragmentation pathways; and (iv) suitable for the higher throughput screening of a large number of samples. Overall, the integrated analytical approach enables linear ion trap mass spectrometry to be used routinely for screening for reactive metabolites in the drug discovery process.

Chem. Res. Toxicol., Vol. 21, No. 7, 2008 1483

(10)

(11)

(12)

(13) (14) (15) (16) (17) (18)

(19)

(20)

(21)

Acknowledgment. We thank Dr. W. Griffith Humphreys for critical reading of this manuscript.

(22)

References (1) Clarke, N. J., Rindgen, D., Korfmacher, W. A., and Cox, K. A. (2001) Systematic LC/MS metabolite identification in drug discovery. Anal. Chem. 73, 430A–439A. (2) Liu, D. Q., and Hop, C. E. (2005) Strategies for characterization of drug metabolites using liquid chromatography-tandem mass spectrometry in conjunction with chemical derivatization and on-line H/D exchange approaches. J. Pharm. Biomed. Anal. 37, 1–18. (3) Ma, S., Chowdhury, S. K., and Alton, K. B. (2006) Application of mass spectrometry for metabolite identification. Curr. Drug Metab. 7, 503–523. (4) Prakash, C., Shaffer, C. L., and Nedderman, A. (2007) Analytical strategies for identifying drug metabolites. Mass Spectrom. ReV. 26, 340–369. (5) Zhu, M., Ma, L., Zhang, D., Ray, K., Zhao, W., Humphreys, W. G., Skiles, G., Sanders, M., and Zhang, H. (2006) Detection and characterization of metabolites in biological matrices using mass defect filtering of liquid chromatography/high resolution mass spectrometry data. Drug Metab. Dispos. 34, 1722–1733. (6) Gangl, E., Utkin, I., Gerber, N., and Vouros, P. (2002) Structural elucidation of metabolites of ritonavir and indinavir by liquid chromatography-mass spectrometry. J. Chromatogr. 974, 91–101. (7) Hopfgartner, G., Varesio, E., Tschappat, V., Grivet, C., Bourgogne, E., and Leuthold, L. A. (2004) Triple quadrupole linear ion trap mass spectrometer for the analysis of small molecules and macromolecules. J. Mass Spectrom. 39, 845–855. (8) Anari, M. R., Sanchez, R. I., Bakhtiar, R., Franklin, R. B., and Baillie, T. A. (2004) Integration of knowledge-based metabolic predictions with liquid chromatography data-dependent tandem mass spectrometry for drug metabolism studies: Application to studies on the biotransformation of indinavir. Anal. Chem. 76, 823–832. (9) Xia, Y. Q., Miller, J. D., Bakhtiar, R., Franklin, R. B., and Liu, D. Q. (2003) Use of a quadrupole linear ion trap mass spectrometer in

(23)

(24)

(25)

(26)

(27)

(28) (29) (30)

metabolite identification and bioanalysis. Rapid Commun. Mass Spectrom. 17, 1137–1145. Drexler, D. M., Tiller, P. R., Wilbert, S. M., Bramble, F. Q., and Schwartz, J. C. (1998) Automated identification of isotopically labeled pesticides and metabolites by intelligent “real time” liquid chromatography tandem mass spectrometry using a bench-top ion trap mass spectrometer. Rapid Commun. Mass Spectrom. 12, 1501–1507. Evans, D. C., Watt, A. P., Nicoll-Griffith, D. A., and Baillie, T. A. (2004) Drug-protein adducts: an industry perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chem. Res. Toxicol. 17, 3–16. Nelson, S. D. (2001) Structure toxicity relationshipssHow useful are they in predicting toxicities of new drugs. In Biological ReactiVe Intermediates (Dansette, P. M., Snyder, R. R., Delaforge, M., Gibson, G. G., Greim, H., Jollow, D. J., Monks, T. J., and Sipes, I. G., Eds.) Kluwer Academic/Plenum Publishers, New York Uetrecht, J. (2003) Screening for the potential of a drug candidate to cause idiosyncratic drug reactions. Drug DiscoVery Today 8, 832– 837. Guengerich, F. P., and MacDonald, J. S. (2007) Applying mechanisms of chemical toxicity to predict drug safety. Chem. Res. Toxicol. 20, 344–369. Baillie, T. A., and Davis, M. R. (1993) Mass spectrometry in the analysis of glutathione conjugates. Biol. Mass Spectrom. 22, 319– 325. Yan, Z., and Caldwell, G. W. (2004) Stable-isotope trapping and highthroughput screenings of reactive metabolites using the isotope MS signature. Anal. Chem. 76, 6835–6847. Ma, S., and Subramanian, R. (2006) Detecting and characterizing reactive metabolites by liquid chromatography/tandem mass spectrometry. J. Mass Spectrom. 41, 1121–1139. Dieckhaus, C. M., Fernandez-Metzler, C. L., King, R., Krolikowski, P. H., and Baillie, T. A. (2005) Negative ion tandem mass spectrometry for the detection of glutathione conjugates. Chem. Res. Toxicol. 18, 630–638. Zheng, J., Ma, L., Xin, B., Olah, T., Humphreys, W. G., and Zhu, M. (2007) Screening and identification of GSH-trapped reactive metabolites using hybrid triple quadruple linear ion trap mass spectrometry. Chem. Res. Toxicol. 20, 757–766. Wen, B., Ma, L., Nelson, S. D., and Zhu, M. (2008) High-throughput screening and characterization of reactive metabolites using polarity switching of hybrid triple quadrupole linear ion trap mass spectrometry. Anal. Chem. 80, 1788–1799. Wen, B., Ma, L., Rodrigues, D., and Zhu, M. (2008) Detection of novel reactive metabolites of trazodone: Evidence for CYP2D6-mediated bioactivation of m-chlorophenylpiperazine. Drug Metab. Dispos. Mutlib, A., Lam, W., Atherton, J., Chen, H., Galatsis, P., and Stolle, W. (2005) Application of stable isotope labeled glutathione and rapid scanning mass spectrometers in detecting and characterizing reactive metabolites. Rapid Commun. Mass Spectrom. 19, 3482–3492. Zhu, M., Ma, L., Zhang, H., and Humphreys, W. G. (2007) Detection and structural characterization of glutathione-trapped reactive metabolites using liquid chromatography-high-resolution mass spectrometry and mass defect filtering. Anal. Chem. 79, 8333–8341. Tang, W., and Abbott, F. S. (1996) Characterization of thiol-conjugated metabolites of 2-propylpent-4-enoic acid (4-ene VPA), a toxic metabolite of valproic acid, by electrospray tandem mass spectrometry. J. Mass Spectrom. 31, 926–936. Yu, L. J., Chen, Y., Deninno, M. P., O’Connell, T. N., and Hop, C. E. (2005) Identification of a novel glutathione adduct of diclofenac, 4′hydroxy-2′-glutathion-deschloro-diclofenac, upon incubation with human liver microsomes. Drug Metab. Dispos. 33, 484–488. Castro-Perez, J., Plumb, R., Liang, L., and Yang, E. (2005) A highthroughput liquid chromatography/tandem mass spectrometry method for screening glutathione conjugates using exact mass neutral loss acquisition. Rapid Commun. Mass Spectrom. 19, 798–804. Chen, W. G., Zhang, C., Avery, M. J., and Fouda, H. G. (2001) Reactive metabolite screen for reduci candidate attriton in drug discovery. In Biological Reactive Intermediates (Dansette, P. M., Snyder, R. R., Delaforge, M., Gibson, G. G., Greim, H., Jollow, D. J., Monks, T. J., and Sipes, I. G., Eds.) pp 521-524, Kluwer Academic/ Plenum Publishers, New York. Wallmark, B. (1989) Omeprazole: Mode of action and effect on acid secretion in animals. Scand. J. Gastroenterol. 166, 12–18. Abelo, A., Andersson, T. B., Antonsson, M., Naudot, A. K., Skanberg, I., and Weidolf, L. (2000) Stereoselective metabolism of omeprazole by human cytochrome P450 enzymes. Drug Metab. Dispos. 28, 966–972. Josephs, J. L., and Sanders, M. (2004) Creation and comparison of MS/MS spectral libraries using quadrupole ion trap and triplequadrupole mass spectrometers. Rapid Commun. Mass Spectrom. 18, 743–759.

TX8001115