Electrochemical Generation of Electrophilic Drug Metabolites

928

Chem. Res. Toxicol. 2008, 21, 928–935

Electrochemical Generation of Electrophilic Drug Metabolites: Characterization of Amodiaquine Quinoneimine and Cysteinyl Conjugates by MS, IR, and NMR Ulrik Jurva,* Anders Holmén, Gunnar Grönberg, Collen Masimirembwa, and Lars Weidolf AstraZeneca R&D Mölndal, S-431 83 Mölndal, Sweden ReceiVed NoVember 8, 2007

The chemical reactivity of electrophilic metabolites usually prevents their detection in vivo since, by definition, they are relatively short-lived and are likely to undergo one or more structural modifications to form more stable final products. Electrochemical oxidation provides a means to generate reactive metabolites in an environment without the presence of such nucleophiles. This paper describes the results of our MS, MS/MS, NMR, IR, and computational studies on oxidation products (and conjugates) that have been generated electrochemically from the antimalarial agent amodiaquine. The electrophilic quinoneimine metabolite of amodiaquine was the major oxidation product following electrochemical oxidation at +600 mV. The absence of biological nucleophiles in the electrochemical experiment facilitated (i) the acquisition of a clean IR spectrum of the amodiaquine quinoneimine and (ii) the addition of biologically relevant nucleophiles under controlled conditions. The addition of cysteine gave four cysteinyl conjugates, while the addition of glutathione gave four glutathionyl conjugates. The product ion spectra of the conjugates formed in the electrochemical experiment were used to identify suitable fragments for selected reaction monitoring (SRM) to selectively search for these conjugates in human liver microsomal (HLM) incubations. The four cysteinyl conjugates, as well as the four glutathionyl conjugates, were also detected as metabolites in HLM. The experiment with cysteine was repeated on a preparative scale that allowed characterization of the major conjugates by 1H NMR. Desethylamodiaquine, the major metabolite formed in human liver microsomes, was also generated electrochemically by oxidation of amodiaquine at +1200 mV followed by reduction at -800 mV. In conclusion, the EC-ESI/MS technique provides the unique opportunity to generate reactive metabolites in the absence of biological nucleophiles, which enables studies that can give insight into the nature of these reactive intermediates. Such knowledge is valuable for risk assessment of new compound classes and can be complementary to computer-based structure–activity relationships of carcinogenicity, mutagenicity, and teratogenicity. Introduction Drug-induced hypersensitivity reactions represent one of the most severe and unpredictable side effects associated with drug therapy and have become a matter of great concern for the pharmaceutical industry (1). Because relatively few individuals become hypersensitized to drugs, the problem is often not discovered until the drug has been released onto the market and exposed to large populations (2). The underlying mechanisms responsible for drug hypersensitization reactions are not fully understood, but reactive metabolites have been demonstrated to play an important role in many instances. Most drugs are too small to be effective immunogens per se, but proteins that have been covalently modified through reaction with reactive metabolites to form drug–protein conjugates may be recognized by the immune system as foreign to the body (3, 4). The formation of drug–protein conjugates most often is dependent on the metabolic bioactivation of the parent compounds to electrophilic metabolites, such as epoxides, quinones, and quinoneimines. In addition to drug hypersensitivity, electrophilic metabolites can be directly toxic (1). For example, binding to DNA may lead to carcinogenic outcomes, while binding to cellular proteins may cause apoptosis and/or necrosis (1). * To whom correspondence should be addressed. Tel: +46 31 7065356. Fax: +46 31 7763748. E-mail: [email protected].

Members of the cytochrome P450 class of enzymes are responsible for the majority of the biotransformations leading to reactive electrophilic metabolites (4). Other enzymes, such as NADPH oxidase and myeloperoxidase present in neutrophiles, also may catalyze the bioactivation of drugs and other xenobiotics to reactive metabolites (5). The chemical reactivity of electrophilic metabolites usually prevents their detection in vivo since, by definition, they are relatively short-lived and are likely to undergo one or more structural modifications to form more stable final products (6). Chemical systems capable of mimicking these oxidations in a more controlled manner therefore are being pursued in an effort to provide an approach that will help in the characterization of electrophilic metabolites. Electrochemistry online with electrospray mass spectrometry (7) (EC-ESI/MS)1 offers a clean and efficient technique to examine the oxidation of xenobiotics and to characterize the resulting oxidation products that often correspond to those 1 Abbreviations: EC-ESI/MS, electrochemistry online with electrospray mass spectrometry; MV, mass voltammogram; NL, neutral loss; SRM, selected reaction monitoring; J, J-coupling, through-bond coupling constant; COSY, correlation spectroscopy; NOE, nuclear Overhauser effect; br, broad; s, singlet; d, doublet; dd, doublet of doublets; t, triplet; GSH, glutathione; HLM, human liver microsomes; RLM, rat liver microsomes; PNM, polymorphonuclear leucocytes.

10.1021/tx700400c CCC: $40.75  2008 American Chemical Society Published on Web 03/25/2008

EC-Generated Electrophilic Drug Metabolites

generated metabolically. The potential of EC-ESI/MS to mimic cytochrome P450-catalyzed oxidations has been investigated (8, 9). Many enzymatic oxidations leading to electrophilic metabolites derived from, for example, phenols, catechols, and aminophenols, can be mimicked by this technique. Already in 1989, Getek and co-workers demonstrated that paracetamol could be oxidized electrochemically to a quinoneimine intermediate that could be trapped by glutathione (10). Despite the great potential of EC-ESI/MS as a tool in the study of reactive metabolites, there have been few publications in this field until recent years when the technique has gained some new attention (11–14). The unique power of EC-ESI/MS is the capability of characterizing by MS/MS oxidation products that are too reactive to survive in vivo. Although MS/MS data can provide useful structural information on the molecular composition and, to some extent, on the regiochemical details of these oxidation products, often additional spectroscopic data are needed to characterize unambiguously the structure of the product. It is possible to carry out EC reactions on a semipreparative scale to provide amounts of compound sufficient for characterization by IR and NMR. With the recent advances in NMR instrumentation, it is now possible to acquire a proton NMR spectrum from a few nanomoles of a pure sample (15). As a complement to MS and NMR techniques, modern IR spectroscopy also may be exploited to aid in structure elucidation studies. In general, IR has not been very useful in working with biological samples because of the presence of interfering components derived from the biological matrix, a problem that may be overcome with EC-generated compounds. As will be described, the EC system used in this work can yield microgram quantities of the targeted products in relatively pure form. These preparations can be made suitable for IR and NMR analysis following either conventional chromatographic fractionation or automated fractionation into 96-well plates. A second approach to aid in structural studies that is explored in this study involves quantum chemical calculations of vibrational spectra. It is now possible to predict the IR spectrum of small molecules with high accuracy (16). Thus, we have a tool where candidate structures of metabolites can be investigated computationally and the results matched against the experimentally determined spectra. This paper describes the results of our MS, MS/MS, NMR, IR, and computational studies on oxidation products (and conjugates) that have been generated electrochemically from the antimalarial agent amodiaquine whose therapeutic utility has been restricted because of its toxicity. Amodiaquine 1 is a 4-aminoquinoline antimalarial agent that is effective against chloroquine resistant strains of Plasmodium falciparum, the cause of cerebral malaria (17). The drug has been widely used for the treatment of malaria over the past 50 years but was withdrawn from prophylactic use because of its hepatotoxicity that may lead to agranulocytosis (18, 19). When conjugated with a high molecular weight protein, amodiaquine is capable of initiating an immune response; the majority of patients suffering from adverse reactions to the drug have detectable IgG antiamodiaquine antibodies (20–22). Several studies suggest that amodiaquine is biotransformed into the electrophilic quinoneimine metabolite 2. Conjugation of 2 with protein thiol groups is thought to be responsible for the observed agranulocytosis (6, 23–27). The bioactivation reaction is catalyzed by liver microsomal enzymes but has also been observed in human polymorphonuclear leucocytes (PNM) (5). Amodiaquine was selected as a candidate to explore the potential of the EC-ESI/

Chem. Res. Toxicol., Vol. 21, No. 4, 2008 929

Figure 1. Molecular structures of amodiaquine 1, amodiaquine quinoneimine 2, desethylamodiaquine 3, and desethylamodiaquine quinoneimine 4.

MS system to model this biotransformation. As part of these studies, we have also examined the electrochemical properties of the amodiaquine metabolite desethylamodiaquine (28) 3 that is converted to the corresponding desethylamodiaquine quinoneimine 4. The main aims of these studies include (i) evaluation of the EC-ESI/MS system as an efficient and clean method for the generation and characterization of electrophilic metabolites and (ii) exploration of extending this technique to include IR and NMR characterization of the EC-generated products. As part of these studies, we have also repeated parts of previously reported in vitro microsomal metabolism work on amodiaquine (27, 28) for comparison of spectral data between metabolites and electrochemically generated products.

Experimental Procedures Chemicals. Amodiaquine (1) dihydrochloride and desethylamodiaquine (3) hydrochloride were obtained from Karolinska Institute (Stockholm, Sweden). The following chemicals were obtained commercially: L-cysteine, 97%, Sigma; glutathione, 98%, Sigma; β-NADPH, reduced form tetrasodium salt, 98%, Sigma; and deuterated methanol, 99.95% deuteration degree, Dr. Glaser AG. All solvents were of analytical grade, and the water used in the experiments was obtained from a water purification system (Elgastat Maxima, ELGA, Lane End, United Kingdom). Microsome Incubations. Human liver microsomes (HLM) were obtained from an in-house bank of liver microsomes maintained at AstraZeneca R&D Mölndal. The incubation mixtures, containing 0.5 mg/mL of human liver microsomal protein and 10 µM amodiaquine 1 in 50 mM phosphate buffer, were preincubated for 2 min at 37 °C. Cysteine or glutathione was added to a final concentration of 10 mM to the incubations that were used to investigate the reactivity of amodiaquine quinoneimine 2 toward nucleophiles. The reaction was initiated by adding NADPH to a final concentration of 1 mM, and the reaction mixtures (final volume 600 µL) were incubated for 60 min at 37 °C in a shaking water bath. Control incubations were conducted in the absence of NADPH, and blank samples without test compound were prepared. The reactions were terminated by adding 600 µL of acetonitrile on ice. The samples were vortexed for 10 s (Technokartell TK3S) and were then centrifuged for 10 min at +4 °C, 20000g, in a Microcentrifuge 157 MP (OLE DICH Instrument makers APS, Denmark). The supernatants were stored at –18 °C until analysis by LC/MS/MS. General Electrochemical Reaction Conditions. Samples were dissolved in 50% methanol/50% 10 mM formic acid for all electrochemical experiments. The substrate concentration was 10 µM for electrochemical oxidation followed by LC/MS/MS analysis and for acquisition of potential scans and 50–100 µM for electrochemical synthesis for characterization by IR and NMR. EC-ESI/MS System. The EC-ESI/MS system was set up as previously reported with the following modifications (8, 9). Samples were infused through an ESA Coulochem 5011 analytical cell (ESA Biosciences Inc., Chelmsford, MA) by a syringe pump at a flow of 50 µL/min. The electrochemical cell was controlled by an ESA Coulochem II potentiostat (ESA Inc., Bedford, MA). The ESA

930

Chem. Res. Toxicol., Vol. 21, No. 4, 2008

working electrode was porous graphite, and all reported cell potentials were measured vs a palladium reference electrode. The potentiostat was programmed to perform a scan from 0 to +1500 mV at a scan rate of 2 mV/s. The outlet from the ESA cell was connected to a Finnigan TSQ7000 triple stage quadrupole mass spectrometer (Finnigan MAT, San Jose, CA) equipped with an electrospray interface. Full scan spectra were acquired continuously. The delay between the electrochemical cell and the mass spectrometer was determined as follows. At a continuous flow of analyte, a potential step from 0 to +600 mV was performed. The time between the potential step and the mass spectral detection of the main oxidation product was measured. With the delay between the electrochemical cell and the mass spectrometer determined to be 45 s, spectra could be assigned to any given potential, and the signals from the different oxidation products could be extracted from the full scan data file and plotted against the potential. Electrochemical Synthesis for Characterization by LC/MS/ MS. The amodiaquine quinoneimine 2 was generated by infusion of amodiaquine (10 µM) through the electrochemical cell at a flow of 50 µL/min for 2 min with the potential maintained at +600 mV. The sample was collected and supplemented with 1 mL of 10 mM formic acid, and aliquots of 20 µL were injected onto the LC/MS/ MS system. For electrochemical generation of desethylamodiaquine 3 and amodiaquine aldehyde 5, amodiaquine (10 µM) was first infused through the electrochemical cell at a flow of 50 µL/min for 5 min with the potential maintained at +1200 mV. The oxidized sample was collected and infused through the electrochemical cell at a reductive potential (-800 mV) to reduce the quinoneimine moieties of 4 and 6 to the corresponding aminophenols 3 and 5, respectively. The sample was collected and supplemented with 1 mL of 10 mM formic acid, and aliquots of 20 µL were injected onto the LC/MS/ MS system. Cysteinyl and glutathionyl conjugates were generated by infusion of amodiaquine (10 µM) through the electrochemical cell at a flow of 50 µL/min for 2 min with the potential maintained at +600 mV. The oxidized sample was collected in a vial containing cysteine (10 µL, 0.1 M) or glutathione (10 µL, 0.1 M), supplemented with 1 mL of 10 mM formic acid, and injected onto the LC/MS/MS system. Electrochemical Synthesis and Purification for Characterization by IR and NMR. For electrochemical synthesis of amodiaquine quinoneimine 2, amodiaquine 1 (100 µM) was infused through the electrochemical cell at a flow of 50 µL/min for 1 h with the potential maintained at +600 mV. The oxidized sample was collected in a glass tube that was protected from light by aluminum foil, and the solvent was removed at room temperature (RT) under a stream of nitrogen gas. The residue was dissolved in 45% methanol/55% water and injected onto the LC-UV system. The fraction containing amodiaquine quinoneimine 2 was manually collected in an ice-cooled glass tube that was protected from light by aluminum foil. The sample was immediately evaporated under a stream of nitrogen gas at RT. The dry residue was kept at –18 °C until analysis by IR the following day. The cysteinyl conjugates 7, 8, and 9 were prepared and purified as follows. Amodiaquine (50 µM) was infused through the electrochemical cell at a flow of 50 µL/min for 3 h with the potential maintained at +600 mV. The oxidized sample was collected in a glass tube that was protected from light by aluminum foil. Cysteine (75 µL, 10 mM in water) was added slowly while stirring. The solution was stirred at RT for 5 min and was then evaporated at RT under a stream of nitrogen gas. The residue was dissolved in mobile phase (5% methanol/95% 10 mM formic acid), and aliquots of 100 µL were injected onto the HPLC system (mass spectrometer not connected). Fractions were collected in polystyrene tubes by a Gilson 202 fraction collector (Gilson, Villiers, France). The content of each tube was analyzed by flow injection analysis on the ESIMS system. Conjugates 7 and 8 coeluted in the used HPLC system, and the tubes containing these compounds were analyzed by LC/ MS. The tubes containing pure fractions were combined, and the solvent was removed under a stream of nitrogen gas to give formate

JurVa et al. salts of 7, 8, and 9. Each fraction was dissolved in water and purified on 200 mg Chromabond C18 ec solid phase columns (MachereyNagel, Düren, Germany). Prior to addition of the samples, the solid phase columns were washed according to the following procedure: (i) 3 mL of methanol, (ii) 3 mL of water, (iii) 3 mL of 100 mM formic acid, (iv) 3 mL of water, (v) 3 mL of 100 mM ammonia, (vi) 3 mL of water, (vii) 3 mL of acetonitrile, (viii) 9 mL of methanol, and (x) 9 mL of water. After addition of the sample, the column was washed with 9 mL of water. The compounds were eluted with 9 mL of 90% methanol/10% water into glass tubes, and the resulting solutions were evaporated at RT under a stream of nitrogen gas. The residues were kept at –18 °C until analysis by NMR. LC/MS/MS. A Hewlett-Packard HP1050 HPLC system (Palo Alto, CA) was used for injection of samples onto a reversed-phase HPLC column (Zorbax SB, C18, 4.6 mm × 150 mm, 5 µm, Agilent Technologies, Palo Alto, CA). The HP1050 pump was programmed to deliver a gradient from methanol and 10 mM formic acid at a total flow of 1 mL/min. At the start of the gradient, the methanol concentration was set to 5% and was then linearly increased to 25% over a period of 15 min. An additional linear increase to 50% methanol was performed over a period of 5 min, and finally, the mobile phase was linearly brought back to 5% methanol in 1 min. The system was allowed to equilibrate for at least 10 min between the injections. A splitter, introduced after the HPLC column, gave an approximate flow of 200 µL/min into a Finnigan TSQ7000 triple stage quadrupole mass spectrometer (Finnigan MAT, San Jose, CA) equipped with an electrospray interface. The heated capillary was set at 200 °C, and the spray voltage was kept at +4.5 kV. The sheath gas flow was set to 60 arbitrary pressure units. Capillary and lens voltages were optimized for amodiaquine 1 to maximize the signal from the [M + H]+ ion. Collision-induced dissociation (CID) was carried out in the collision cell on the [M + H]+ ion of each compound using argon as collision gas at a pressure of 1 mTorr and a collision offset of –35 V. LC/UV. The oxidized sample was injected onto a reversed-phase HPLC column (ACE 3, C8, 2.1 mm × 100 mm, ACT, Aberdeen, United Kingdom) that was connected to a Hewlett-Packard HP1050 diode array detector. The pumps were programmed to deliver 45% methanol/55% water at a flow of 200 µL/min. IR Spectroscopy and Quantum Chemical Calculations. IR spectra were collected on a Bio-Rad FTS 6000 FT-IR instrument equipped with a nitrogen-cooled MCT detector (Bio-Rad, Cambridge, MA). The ATR system consisted of a DuraSamplIR base plate with a 20 bounce SiComp/ZnSe (EL) Duradisk (SensIR Technologies, Danbury, CT). Samples were added to the ATR crystal as methanol solutions of the analytes. After a few minutes, the methanol had evaporated while forming a thin film of the analyte on the crystal, and a spectrum was recorded. The spectral resolution was 4 cm-1, and 128 scans were coadded for each spectrum. After the IR experiment was performed, the 2 analyte film on the ATR crystal was dissolved in methanol and analyzed by flow injection analysis on the ESI-MS system and the content of amodiaquine 1 in the sample was estimated from the relative intensity of m/z 356 to be less than 3%. Density functional theory calculations of isotropic IR absorption spectra were performed using the Gaussian 98 quantum chemistry package with the B3LYP hybrid functional and 6-311G** basis set (29). The absorption profiles were simulated, using an in-house Matlab code (30), from the calculated scaled frequencies (scale factor ) 0.985) and the dipole (Di) strengths using the following equation (31):

(ν˜ ) )

8π3NAν˜ 3000hc2.303

∑ Di fi(ν˜ ,ν˜ i)

(1)

i

The line shape, fi (ν˜ , ν˜ i), is assumed to be Lorentzian (31):

f(ν˜ ,ν˜ i) )

γ˜ π[(ν˜ - ν˜ i)2 - γ˜ 2]

with a half-width at half-height, γ˜ , equal to 5 cm-1.

(2)



Figure 2. Mass voltammograms with m/z values of protonated amodiaquine 1 and its oxidation products.

Scheme 1. Proposed Pathway for the Electrochemical Oxidation of Amodiaquine in 50% Methanol and 50% 10 mM Formic Acida

Figure 3. Extracted ion chromatograms (XICs) of electrochemicalgenerated products and metabolites of amodiaquine 1. (a) Electrochemical oxidation of 1 at +600 mV in 50% 10 mM formic acid and 50% methanol; XIC of m/z 356 and m/z 354. (b) Sixty minutes incubation of 1 with HLM; XIC of m/z 356 and m/z 328. (c) Electrochemical oxidation of 1 at +1200 mV followed by reduction at –800 mV; XIC of m/z 328 and m/z 299. The retention times in minutes are given within parentheses above each peak.

a The oxidation potentials given in the scheme correspond to the potentials at which the MS signal intensity of the oxidized compound reaches 50% of its maximal value. The reduction was carried out at a fixed potential of -800 mV.

NMR Spectroscopy. NMR spectra were recorded on a Varian Inova Spectrometer operating at 600 MHz 1H frequency, using an indirect detection 3 mm triple resonance Nalorac probe with z-gradients. Two-dimensional, gradient COSY and NOESY 1H experiments were acquired with standard Varian sequences supplied with VNMR 6.1C. The samples were dissolved in 180 µL of deuterated methanol and transferred to 3 mm NMR tubes. All experiments were run at +25 °C; chemical shifts are reported relative to the residual solvent signal of methanol set to 3.30 ppm.

Results and Discussion Electrochemical Oxidation Products and Metabolites of Amodiaquine. Mass voltammograms (MVs) with detection of m/z values of protonated amodiaquine 1 and its oxidation products in 10 mM formic acid and 50% methanol are presented in Figure 2 with the proposed electrochemical oxidation pathways given in Scheme 1. At an applied potential of approximately +400 mV, the [M + H]+ ion of amodiaquine 1 at m/z 356 starts to decrease and as the potential is increased it

is gradually replaced with the [M + H]+ ion of the quinoneimine oxidation product 2 at m/z 354. The increase in intensity of m/z 356 at +600 to +800 mV is due to an increase in the intensity of the [M + H]+ ion of the 37Cl isotopolog of 2. At approximately +700 mV, the tertiary amine function is also oxidized and an ion at m/z 326 appears, corresponding to the [M + H]+ ion 4 of desethylamodiaquine quinoneimine. At the same potential, two additional ions appear at m/z 297 and m/z 329, presumably the [M + H]+ ion and the [M + MeOH + H]+ ion of 6. Interpretation of the MVs shown in Figure 2 suggests that oxidation at potentials between +600 and +700 mV will give good yields of 2 and that oxidation at potentials above +1200 mV will generate mainly 4 and 6. Oxidation of desethylamodiaquine 3 in the EC-ESI/MS system gave mainly desethylamodiaquine quinoneimine 4. Cooxidation of amodiaquine 1 and desethylamodiaquine 3 revealed that the two compounds were oxidized at the same potential (50% consumed at a potential of +510 mV). Extracted ion chromatograms (XICs) for [M + H]+ ions of amodiaquine 1 electrochemical products and metabolites are given in Figure 3. Electrochemical oxidation at +600 mV gave amodiaquine quinoneimine 2 as the major oxidation product. XICs of metabolites formed following incubation of 1 in human liver microsomes (HLM) are presented Figure 3b. In agreement with previous studies (27, 28), desethylamodiaquine 3 was the major metabolite formed. To generate 3 electrochemically, amodiaquine 1 was first oxidized at a potential of +1200 mV. The collected oxidized sample was then injected with the EC set at –800 mV to reduce the quinoneimine moiety back to an aminophenol (Figure 3c and Scheme 1). The quinoneimine metabolite 2 was not observed in the HLM experiment, although its presence is evident from the experiments with cysteine and glutathione (discussed below). Reactions of Amodiaquine Quinoneimine (2) with Nucleophiles. Amodiaquine quinoneimine 2 is electrophilic, and adduct formation with cysteinyl groups of proteins has been proposed as a possible explanation for the toxicity of amodiaquine

932


JurVa et al.

Figure 4. Amodiaquine 1 oxidized at +600 mV followed by addition of 5 equiv cysteine. (a) Total ion current chromatogram. (b-e) Product ion spectra and proposed structures of the cysteinyl conjugates 7, 8, 9, and 10 in chromatogram a.

(6, 23–27). With the most suitable conditions for electrochemical generation of 2 from 1 now available from the EC-ESI/MS experiments, the reactions of 2 with added nucleophile could be investigated. In this section, we report the results from our studies of 2 with two biologically relevant nucleophiles, cysteine and glutathione. The experiments involved collection of the electrochemically generated quinoneimine 2 directly after the electrochemical cell in a vial containing the desired nucleophile. The total ion current chromatogram (TIC) of amodiaquine 1 oxidized at +600 mV followed by treatment with 5 equiv of cysteine is given in Figure 4a with the product ion spectra of the adducts presented in Figure 4e. The quinoneimine 2 was completely consumed, and four products appeared (7, 8, 9, and 10). Compounds 7, 8, and 9 all gave singly charged ions at m/z 475 corresponding to [M + H]+ ions of three isomeric compounds with a molecular mass of 474 Da. This is consistent with the addition of cysteine to the quinoneimine moiety of 2. The product ion spectra of compounds 7, 8, and 9 (Figure 4b-d) contain a characteristic neutral loss (NL) of 73 Da (diethylamine) to give a fragment at m/z 402. Subsequent NL of 87 Da (2aminoacrylic acid from the cysteinyl group) gave a fragment at m/z 315. The proposed fragmentation pathway for the cysteinyl conjugate 9 is presented in Scheme 2.

Scheme 2. Proposed Fragmentation Pathway of Amodiaquine Cysteinyl Conjugate 9

To further characterize these three isomers, the electrochemical oxidation of 1 was carried out on a preparative scale, and the products were purified by HPLC. Fractions containing individual isomers were collected, evaporated, and further purified by solid phase extraction to allow characterization of compounds 7 and 9 by 1H NMR. Analysis of the spectra allowed assignment of the positions of the cysteinyl function for the two regioisomers 7 and 9 as shown in Figure 4b,d. Compound 8 was tentatively assigned as the third possible regioisomeric cysteinyl conjugate at the aminophenol moiety of 1 as shown



Scheme 3. Proposed Pathways for the Formation of Amodiaquine Cysteinyl Conjugates 9 and 10

in Figure 4c (NMR interpretations are available in the Supporting Information). Compound 10, eluting after 1 at 19.7 min (Figure 4a), gave a singly charged ion at m/z 390 corresponding to the [M + H]+ ion of a compound with a molecular mass of 389 Da. The product ion spectrum of 10 (Figure 4e) does not contain the characteristic NL of 73 Da (diethylamine), indicating that 10 does not contain the tertiary aminyl function. The NL of 87 Da (2-aminoacrylic acid from the cysteinyl group) to give a fragment at m/z 303 indicates that cysteine is incorporated in the structure. A molecular mass of 389 Da is consistent with compound 10 (Figure 4e) where cysteine has replaced the tertiary aminyl function on amodiaquine. Further support for this structure was provided by electrochemical oxidation of desethylamodiaquine 3 followed by addition of cysteine. Compound 10 also was obtained in this experiment, supporting the assumption that the aminyl substituent is not present in compound 10. Proposed pathways for the formation of compounds 9 and 10 are given in Scheme 3. With the structures of potential metabolites now available from the electrochemical experiments, it was possible to search selectively for compounds 7, 8, 9, and 10 in human liver microsomal incubations of amodiaquine 1. The product ion spectra (Figure 4) were used to identify suitable fragments for selected reaction monitoring (SRM). The SRM chromatograms presented in Figure 5a,b show that the cysteinyl conjugates formed in the electrochemical experiments are also formed in HLM, although the amounts of 7 and 8 are relatively low. It should be noted that any attempt to predict relative amounts from these data is speculative since the MS response might differ for different compounds. The electrochemical experiment with amodiaquine 1 and cysteine was also repeated with glutathione as the nucleophile. Analogous to the experiments with cysteine, the quinoneimine 2 was completely consumed and four products (11, 12, 13, and 14) were detected (Figure 5c). Compounds 11, 12, and 13 all gave singly charged ions at m/z 661 corresponding to [M + H]+ ions of three isomeric compounds with a molecular mass of 660 Da, consistent with addition of glutathione to the quinoneimine moiety of 2. Compound 14, eluting after 1 at 20.7 min (chromatogram 5c), gave a singly charged ion at m/z 576, corresponding to a [M + H]+ ion of a compound with a

Figure 5. Chromatograms from oxidation of amodiaquine 1 together with nucleophiles. (a) TIC chromatogram from electrochemical oxidation of 1 at +600 mV followed by addition of 5 equiv of cysteine. (b) Incubation of 1 for 60 min in HLM with 10 mM cysteine. SRM transitions: 475f402, 475f315, and 390f303. (c) Electrochemical oxidation of 1 at +600 mV followed by addition of 5 equiv of glutathione. SRM transitions: 661f315, 661f459, 661f558, 576f303, 576f447, and 576f493. (d) Incubation of 1 for 60 min in HLM with 10 mM glutathione. The same SRM transitions as in chromatogram c were used.

molecular mass of 575 Da. This is consistent with the addition of glutathione to the quinoneimine moiety in combination with loss of the tertiary aminyl group (see analogy with cysteinyl conjugate 10). No further characterization of the isomers was made, but interpretation of the product ion spectra from the electrochemically generated glutathionyl conjugates helped in the detection of compounds 11, 12, 13, and 14 by SRM in liver microsomal incubations of amodiaquine 1. The SRM chromatogram (Figure 5d) shows that compounds 13 and 14 are generated in the HLM and that there are trace amounts of compounds 11 and 12. The cysteinyl conjugates 7 and 8, as well as the glutathionyl conjugates 11 and 12, were formed in very low amounts in the liver microsomal incubations. These metabolites could not be separated from the background noise when the mass spectrometer was operated in the full-scan MS mode. Hence, it would have been a very challenging task to find these metabolites without the electrochemically generated standard compounds. Characterization of 2 by IR and Quantum Chemical Computational Methods. Reactive metabolites are, by definition, short-lived in vivo because of the presence of nucleophiles,

934


Figure 6. Mid-IR spectra of amodiaquine 1 (upper panel) and amodiaquine quinoneimine 2 (lower panel).

such as glutathione, thiol moieties of cysteine in proteins, DNA bases, etc (4). Electrochemical oxidation provides a means to generate reactive metabolites in an environment without the presence of such nucleophiles. This facilitates isolation and characterization by methods that would not be applicable to reactive metabolites formed in a biological environment. To test this concept, we decided to isolate the electrochemically generated amodiaquine quinoneimine (2) for further characterization with IR and NMR. Unfortunately, 2 was reduced to amodiaquine in the NMR probes by an unknown mechanism, and thus, NMR spectra could not be acquired. In contrast, we were able to generate a clean IR spectrum of solid 2 isolated on the surface of the ATR crystal. This was accomplished by dissolving purified 2 in a small amount of methanol, a quick transfer to the ATR crystal, and subsequent evaporation of solvent. The IR spectrum of solid 2 as well as the reference spectrum of solid 1 are presented in Figure 6. The structural change from the phenolic ring in 1 to the quinoneimine structure in 2 gave rise to some prominent differences in the IR spectra of 1 and 2. Notably, 2 has two strong bands at 1649 and 1631 cm-1 that are not present in the spectrum of 1. These two vibrations were assigned to the conjugated carbonyl stretch and the conjugated iminyl stretch, respectively. The positions of these two bands are in excellent agreement with literature data for these functional groups (32). To further strengthen the assignments of the bands in the IR spectrum of 2, we have undertaken a quantum chemical computational investigation of the spectrum using density functional theory (DFT). It should be noted that the calculated spectrum is obtained for an isolated molecule with no solvent interactions whereas the sample used to obtain the experimental spectrum is a solid sample of 2. The observed differences in band widths and intensity distributions may be due to such factors. The experimental and computed (DFT/B3LYP/6311G**) mid-IR spectra are compared in Figure 7. The carbonyl stretch (band 1) and the iminyl stretch (band 2) are computed to be of quite different intensity while experimentally this difference is not as pronounced. However, band 2 is overlapping at least with bands 1 and 3 that makes it difficult to accurately estimate band intensities. Also, for bands 4 and 5, there is overlap that also precludes an accurate estimate of their intensities. Despite these complications, correlations between the observed and the computed spectral features (bands 1–28) can be made that strongly support that the structure of the isolated compound is actually 2.

JurVa et al.

Figure 7. Experimental (upper panel) and calculated (lower panel) midIR spectra of 2. The displayed computed spectrum is derived from the calculated (B3LYP/6-311G**) vibrational frequencies and corresponding dipole strengths. The frequencies have been scaled by a single scale factor (0.985) to aid the comparison between experimental and calculated data. Lorentzian bandshapes (γ˜ ) 5 cm-1) have been used in the simulation of the computed spectrum. Calculated fundamental vibrations and their corresponding observed bands have been arbitrarily numbered 1–28.

Conclusions These studies have shown that electrochemical oxidation of amodiaquine 1 provides a clean and efficient means to generate the electrophilic amodiaquine quinoneimine 2. Because reactive metabolites are difficult to detect in vivo, their presence is usually hypothesized from the detection of, for example, glutathionyl conjugates. Electrochemical synthesis allows for characterization of the actual reactive metabolite by MS/MS and, for reasonably stable products like the amodiaquine quinoneimine 2, by additional techniques such as IR, and in the particular case with 2, IR spectroscopy in combination with quantum chemical calculations offered a complementary route to MS for structural elucidation. The absence of biological nucleophiles in the electrochemical experiment facilitates the addition of selected nucleophiles under controlled conditions. The product ion spectra of the conjugates formed in the electrochemical experiment can then be used to identify suitable fragments for SRM to selectively search for these conjugates in biological samples. Thus, the EC-ESI/MS technique provides the unique opportunity to study reactive metabolites without the presence of biological nucleophiles, which enables studies that can give insight into the nature of these reactive intermediates. Such knowledge is valuable for risk assessment of new compound classes and can be complementary to computer-based structure–activity relationships of carcinogenicity, mutagenicity, and teratogenicity. These experiments have also demonstrated that the ESA coulochem cells are well-suited for clean, small-scale synthesis of material that allows for characterization by IR and NMR. We believe that, in many cases, this is a more rapid and straightforward way to identify a metabolite than to separate the metabolite from a complicated biological matrix or to synthesize a reference compound by traditional means. Because both IR and NMR are nondestructive techniques, the compound can be recovered and used as reference material for future studies. Acknowledgment. Prof. Neal Castagnoli Jr. is gratefully acknowledged for valuable discussions and for proofreading of this manuscript.


Supporting Information Available: Characterization of the amodiaquine cysteinyl conjugates by NMR. This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Naisbitt, D. J., Williams, D. P., Pirmohamed, M., Kitteringham, N. R., and Park, B. K. (2001) Reactive metabolites and their role in drug reactions. Curr. Opin. Allergy Clin. Immunol. 1, 317–325. (2) Ju, C., and Uetrecht, J. P. (2002) Mechanism of idiosyncratic drug reactions: Reactive metabolite formation protein binding and the regulation of the immune system. Curr. Drug Metab. 3, 367–377. (3) Williams, D. P., Kitteringham, N. R., Naisbitt, D. J., Pirmohamed, M., Smith, D. A., and Park, B. K. (2002) Are chemically reactive metabolites responsible for adverse reactions to drugs? Curr. Drug Metab. 3, 351–366. (4) Kalgutkar, A. S., Gardner, I., Obach, R. S., Shaffer, C. L., Callegari, E., Henne, K. R., Mutlib, A. E., Dalvie, D. K., Lee, J. S., Nakai, Y., O’Donnell, J. P., Boer, J., and Harriman, S. P. (2005) A comprehensive listing of bioactivation pathways of organic functional groups. Curr. Drug Metab. 6, 161–225. (5) Park, B. K., Pirmohamed, M., and Kitteringham, N. R. (1998) Role of drug disposition in drug hypersensitivity: A chemical molecular, and clinical perspective. Chem. Res. Toxicol. 11, 969–988. (6) Harrison, A. C., Kitteringham, N. R., Clarke, J. B., and Park, B. K. (1992) The mechanism of bioactivation and antigen formation of amodiaquine in the rat. Biochem. Pharmacol. 43, 1421–1430. (7) Zhou, F., and Van Berkel, G. J. (1995) Electrochemistry combined online with electrospray mass spectrometry. Anal. Chem. 67, 3643– 3649. (8) Jurva, U., Wikström, H. V., Weidolf, L., and Bruins, A. P. (2003) Comparison between electrochemistry/mass spectrometry and cytochrome P450 catalyzed oxidation reactions. Rapid Commun. Mass Spectrom. 17, 800–810. (9) Johansson, T., Weidolf, L., and Jurva, U. (2007) Mimicry of phase I drug metabolismsNovel methods for metabolite characterization and synthesis. Rapid Commun. Mass Spectrom. 21, 2323–2331. (10) Getek, T. A., Korfmacher, W. A., McRae, T. A., and Hinson, J. A. (1989) Utility of solution electrochemistry mass spectrometry for investigation the formation and detection of biologically important conjugates of acetaminophen. J. Chromatogr. A 474, 256. (11) Gamache, P., Smith, R., McCarthy, R., Waraska, J., and Acworth, I. (2003) ADME/Tox profilingsUsing electrochemistry and electrospray ionisation mass spectrometry. Spectroscopy 18, 14–21. (12) van Leeuwen, S. M., Blankert, B., Kauffmann, J.-M., and Karst, U. (2005) Prediction of clozapine metabolism by on-line electrochemistry/ liquid chromatography/mass spectrometry. Anal. Bioanal. Chem. 382, 742–750. (13) Lohmann, W., and Karst, U. (2006) Simulation of the detoxification of paracetamol using on-line electrochemistry/liquid chromatografy/ mass spectrometry. Anal. Bioanal. Chem. 386, 1701–1708. (14) Madsen, K. G., Olsen, J., Skonberg, C., Hansen, S. H., and Jurva, U. (2007) Development and evaluation of an electrochemical method for studying reactive phase-I metabolites: Correlation to in vitro drug metabolism. Chem. Res. Toxicol. 20, 821–831. (15) Lewis, R. J., Bernstein, M. A., Duncan, S. J., and Sleigh, C. J. (2005) A comparison of capillary-scale LC-NMR with alternative techniques: Spectroscopic and practical considerations. Magn. Reson. Chem. 43, 783–789. (16) Yoshida, H., Takeda, K., Okamura, J., Ehara, A., and Matsuura, H. (2002) A new approach to vibrational analysis of large molecules by density functional theory: Wavenumber-linear scaling method. J. Phys. Chem. A 106, 3580–3586. (17) Watkins, W. M., Sixsmith, D. G., Spencer, H. C., Boriga, D. A., Kariuki, D. M., Kipingor, T., and Koech, D. K. (1984) Effectiveness of amodiaquine as treatment for chloroquine-resistant Plasmodium falciparum infections in Kenya. Lancet 1, 357–359.

Chem. Res. Toxicol., Vol. 21, No. 4, 2008 935 (18) Hatton, C. S., Peto, T. E., Bunch, C., Pasvol, G., Russell, S. J., Singer, C. R., Edwards, G., and Winstanley, P. (1986) Frequency of severe neutropenia associated with amodiaquine prophylaxis against malaria. Lancet 1, 411–414. (19) Neftel, K. A., Woodtly, W., Schmid, M., Frick, P. G., and Fehr, J. (1986) Amodiaquine induced agranulocytosis and liver damage. Br. Med. J. (Clin. Res. Ed.) 292, 721–723. (20) Christie, G., Breckenridge, A. M., and Park, B. K. (1989) Drug-protein conjugatessXVIII. Detection of antibodies towards the antimalarial amodiaquine and its quinone imine metabolite in man and the rat. Biochem. Pharmacol. 38, 1451–1458. (21) Clarke, J. B., Maggs, J. L., Kitteringham, N. R., and Park, B. K. (1990) Immunogenicity of amodiaquine in the rat. Int. Arch. Allergy Immunol. 91, 335–342. (22) Clarke, J. B., Neftel, K., Kitteringham, N. R., and Park, B. K. (1991) Detection of antidrug IgG antibodies in patients with adverse drug reactions to amodiaquine. Int. Arch. Allergy Immunol. 95, 369–375. (23) Maggs, J. L., Tingle, M. D., Kitteringham, N. R., and Park, B. K. (1988) Drug-protein conjugatessXIV. Mechanisms of formation of protein-arylating intermediates from amodiaquine, a myelotoxin and hepatotoxin in man. Biochem. Pharmacol. 37, 303–311. (24) Naisbitt, D. J., Ruscoe, J. E., Williams, D., O’Neill, P. M., Pirmohamed, M., and Park, B. K. (1997) Disposition of amodiaquine and related antimalarial agents in human neutrophils: implications for drug design. J. Pharmacol. Exp. Ther. 280, 884–893. (25) Naisbitt, D. J., Williams, D. P., O’Neill, P. M., Maggs, J. L., Willock, D. J., Pirmohamed, M., and Park, B. K. (1998) Metabolism-dependent neutrophil cytotoxicity of amodiaquine: A comparison with pyronaridine and related antimalarial drugs. Chem. Res. Toxicol. 11, 1586– 1595. (26) Tingle, M. D., Jewell, H., Maggs, J. L., O’Neill, P. M., and Park, B. K. (1995) The bioactivation of amodiaquine by human polymorphonuclear leucocytes in vitro: Chemical mechanisms and the effects of fluorine substitution [erratum appears in Biochem. Pharmacol. (1995) 50 (12), 2119]. Biochem. Pharmacol. 50, 1113–1119. (27) Jewell, H., Maggs, J. L., Harrison, A. C., O’Neill, P. M., Ruscoe, J. E., and Park, B. K. (1995) Role of hepatic metabolism in the bioactivation and detoxication of amodiaquine. Xenobiotica 25, 199– 217. (28) Li, X.-Q., Björkman, A., Andersson, T. B., Ridderstrom, M., and Masimirembwa, C. M. (2002) Amodiaquine clearance and its metabolism to N-desethylamodiaquine is mediated by CYP2C8: A new high affinity and turnover enzyme-specific probe substrate. J. Pharmacol. Exp. Ther. 300, 399–407. (29) Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R., Zakrzewski, V. G., Montgomery, J. A., Stratmann, R. E., Burant, J. C., Dapprich, S., Millam, J. M., Daniels, A. D., Kudin, K. N., Strain, M. C., Farkas, O., Tomasi, J., Barone, V., Cossi, M., Cammi, R., Menucci, B., Pomelli, C., Adamo, C., Clifford, S., Ochterski, J., Petersson, G. A., Ayala, P. Y., Cui, Q., Morokuma, K., Malick, D. K., Rabuck, A. D., Raghavachari, K., Foresman, J. B., Cioslowski, J., Ortiz, J. V., Baboul, A. G., Stefanov, B. B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Gomperts, R., Martin, R. L., Fox, D. J., Keith, T., Al-Laham, M. A., Peng, C. Y., Nanayakkara, A., Gonzales, C., Challacombe, M., Gill, P. M. W., Johnson, B., Chen, W., Wong, M. W., Andres, J. L., Gonzalez, C., Head-Gordon, M., Replogle, E. S., and Pople, J. A. (1998) Gaussian 98, Gaussian, Inc., Pittsburg, PA. (30) MatLab 6.0, The Mathworks, Natick, MA. (31) Michl, J., and Thulstrup, E. W. (1995) Spectroscopy with Polarized Light, VCH Publishers, New York. (32) Daimay, L.-V., Colthup, N. B., Fately, W. G., and Grasselli, J. G., Eds. (1991) The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules, Academic Press Inc., San Diego, CA.

TX700400C

Electrochemical Generation of Electrophilic Drug Metabolites

Recommend Documents