A Study of the Identification, Fragmentation Mode and Metabolic Pathways of Imatinib in Rats Using UHPLC-Q-TOF-MS/MS

In this study, The metabolites, metabolic pathways, and metabolic fragmentation mode of a tyrosine kinase inhibitor- (TKI-) imatinib in rats were investigated. The samples for analysis were pretreated via solid-phase extraction, and the metabolism of imatinib in rats was studied using ultra-high-performance liquid chromatography-quadrupole-time-of-flight mass spectrometry (UHPLC-Q-TOF-MS/MS). Eighteen imatinib metabolites were identified in rat plasma, 21 in bile, 18 in urine, and 12 in feces. Twenty-seven of the above compounds were confirmed as metabolites of imatinib and 9 of them were newly discovered for the first time. Oxidation, hydroxylation, dealkylation, and catalytic dehydrogenation are the main metabolic pathways in phase I. For phase II, the main metabolic pathways were N-acetylation, methylation, cysteine, and glucuronidation binding. The fragment ions of imatinib and its metabolites were confirmed to be produced by the cleavage of the C-N bond at the amide bond. The newly discovered metabolite of imatinib was identified by UHPLC-Q-TOF-MS/MS. The metabolic pathway of imatinib and its fragmentation pattern were summarized. These results could be helpful to study the safety of imatinib for clinical use.


Introduction
Imatinib (IM) is the first generation of tyrosine kinase inhibitor [1,2] mainly used for the treatment of chronic, accelerated phase or blast crisis of adult chronic myeloid leukemia with positive Philadelphia chromosome and malignant gastrointestinal stromal tumors that cannot be resected and/or metastasized [3,4]. IM was first tested in clinical trials in the 1990s and was approved by the US Food and Drug Administration (FDA) at the beginning of this century. Its antitumor mechanism specifically blocks the binding site of adenosine triphosphate (ATP) on tyrosine kinase to inhibit the autophosphorylation and substrate phosphorylation of BCR-ABL target protein. IM could also stabilize the inactive conformation of tyrosine kinase to inhibit its activity, leading to reduced proliferation of chronic myelocytic leukemia (CML) cells and cell apoptosis [5][6][7]. IM has excellent pharmacokinetic characteristics and its oral bioavailability was reported to be ≥ 98%, with a halflife of 20 hours and peak time ranging from 2 to 4 hours. Its excretion pathway mainly depends on feces and urine [8].
Most drugs were mainly eliminated from the body after metabolism [9][10][11]. Metabolism can be understood as the process of converting drugs into their metabolites under the catalysis of enzymes [12]. Drug metabolism may produce one or more active metabolites with the same pharmacological effects and targets as the original drug [13,14]. Sometimes, it may also generate overactive or toxic metabolites, which may cause serious adverse side-effects [15]. For example, some CML patients developed hepatotoxicity symptoms after using IM, which may be related to one or more metabolites of IM [16]. In addition, studies showed that the main metabolites of IM may be involved in IM resistance [17]. erefore, to systematically understand the safety of IM for its clinical application, it is necessary to investigate its metabolites. Up to date, researchers have done a lot of work on IM metabolism and achieved reliable results. MarcMarull et al., with the aid of liquid chromatography combined with triple quadrupole mass spectrometer (TSQ-MS) and linear ion trap mass spectrometer (LTQ-MS), successfully isolated one demethylated metabolite, two hydroxylated metabolites, and three N-oxidation metabolites of IM from microsomes containing cytochrome P450 (CYP) isozymes [18]. Li et al. found 7 cyano metabolites of IM from human liver microsomes (in vitro) and identified their structures [19]. Friedecký et al. have identified 90 metabolites of IM from the plasma of patients with CML [20]. erefore, a thorough study of the metabolites and metabolic pathways of imatinib will help to use IM safely and rationally. In addition, the study of drug metabolites of IM will help with its clinical efficacy and provide more information for the scientist to develop safer and more effective formulations than the raw materials [21,22]. e purpose of this study was to investigate the metabolism of IM in rats. e metabolites of IM in plasma, bile, urine, and feces were identified using UHPLC-Q-TOF-MS/ MS. e metabolic pathways of phases I and II and the fragmentation pattern of metabolites were also summarized.

Animal Experiments.
Male Sprague-Dawley (SD) rats (n � 18, weighing 220-250 g, 12-14 weeks old) were provided by the Experimental Animal Center of Shenyang Pharmaceutical University (Shenyang, China). e SD rats were fed in metabolic cages for at least 3 days and fasted for 12 h before the experiments. All experiments in vivo were conducted under the guidelines of the administration of experimental animals in China.
IM (50 mg/mL) was dissolved in 0.5% carboxymethyl cellulose sodium (CMC-Na) aqueous solution. After administering IM solution at a dose of 50 mg/kg, the rats were randomly divided into three groups (six rats in each group).
Group I was for taking the plasma samples. e blood was taken from the ophthalmic veins at time points at 0.5, 1, 2, 4, 8, and 12 h after oral administration of IM, and then the blood collected in heparinized tubes was centrifuged immediately at 5000 rpm for 10 min to obtain the plasma. For Group II (bile samples), the rats were anesthetized with urethane solution (0.05 mL/10 g). Duct cannulation operation was performed to collect blank bile first, and the rats were allowed to recover from anesthesia before drug administration. e samples were continuously collected but samples were divided according to the period of 0-4, 4-8, 8-12, 12-24, and 24-48 h. Group III was used for obtaining the urine and feces samples. e rats were kept in metabolic cages individually and orally given IM, and then, the samples were collected at different time intervals of 0-4, 4-8, 8-12, 12-24, and 24-48 h. All the blank samples of each group were collected before oral administration of IM. All collected samples were stored at − 80°C until further treatment.

Sample Pretreatment.
Pretreatment of plasma: 200 μL of plasma sample was extracted using 600 uL of methanol and then vortexed for 1 min to precipitate the proteins. After centrifugation at 12000 rpm for 10 min at 4°C, the supernatant was transferred into a clean tube and dried under nitrogen gas at 40°C. Pretreatment of feces: feces sample (0.2 mg) was added to methanol (2 mL), then vortexed for 2 min, followed by sonication for 30 min, and vortexed again for another 2 min. After the mixture was centrifuged (13000 rpm, 10 min) at 4°C, the supernatant was collected and dried with nitrogen flow at 40°C.
Pretreatment of urine and bile: the solid-phase extraction cartridges (Angela Bond Elut C-18, 1 g) were activated via duplicated wash using water (1 mL), followed by methanol (1 mL). After centrifugation (13000 rpm) for 5 min, the supernatant was filtrated through a 0.22 μm membrane filter, followed by further purification (1 mL of the filtered sample) with a solid-phase extraction cartridge at a flow rate of 30 drops ·min − 1 . e cartridges were washed with water (1 mL) first and then washed for another 3 times using 1 mL of 0.1% formic acid in methanol. e eluent of 0.1% formic acid in methanol was collected and dried with nitrogen flow at 40°C. Before analysis, all the pretreatment samples have been dried with nitrogen, and the residues were reconstituted with 200 μL of a mixed solvent of acetonitrile and 0.1% formic acid in water in a ratio of 10 : 90 (v/v) and then sonicated for 10 min before centrifugation (12000 rpm) at 4°C. 3 μL of the supernatant was finally injected for UHPLC-Q-TOF-MS/MS analysis.
Mass spectrometry detection was performed using an AB SCIEX X500R UPLC-Q/TOF-MS (AB SCIEX, Framingham, MA, USA). Mass spectrum acquisition was in positive electrospray ionization mode and the conditions used were as follows: the pressure of the nebulizer gas (gas 1), the heater gas (gas 2), and the curtain gas was set to 55, 55, and 35 psi, respectively; ion source temperature was 600°C. TOF-MS/ MS parameter settings are as follows: scanning range, m/z 50∼1000; collision gas, 7 psi; ion spray voltage, 5.5 kV; declustering potential (DP), 80 V; the collision energy (CE), 10 V; TOF-MS/MS, m/z 50-1000; ion spray voltage, 5.5 kV; declustering potential (DP), 80 V; the collision energy (CE), from 20 to 50 V; accumulation time, 25 min. Apply AB SCIEX OS for processing. In the experiment, the CDS automatic calibration system was used to calibrate the experimental data collection.

Results and Discussion
e MS/MS spectrum of IM standard is shown in Figure 2; the ion at m/z 394 was formed after the cleavage of the 4methyl-1-piperazinyl group on benzamide on the phenyl group of 2-aniline pyrimidine, and the ion at m/z 394 generated m/z 378 by the loss of O of the benzoylamino group. e ion at m/z 217 was formed after the cleavage of the amide part of benzamide, and the ion at m/z 217 generated m/z 174 by RDA (Reverse-Diels-Alder reaction) cleavage reaction. As shown in Figure 3    bonds between C 1 and N breaks, m/z 247, and fragment ions are generated. e fragment ion at m/z 222 was generated by the fracture of C 1 and C 6 , C 4 , and C 5 in the fragment ion structure of m/z 247 and simultaneously intramolecular proton migration rearrangement. e fragment ion at m/z 231 was generated after the amide bond fracture of carbonylsubstituted IM, and the ionic fracture mode of other fragments was the same as that of M6. M13 and M21 are the IM metabolites of the carbonylation of 4-methylpiperazine rings.
Metabolites e fragment of m/z 203 further lost one CO and generated the fragment ion of m/z 175. e other fragment ions share the same metabolite pathway as the above. erefore, the metabolite is considered to be the IM metabolite of carbonylation of N-methyl. e m/z 408 fragment ion of M23 was the M0 methylation followed by the removal of the 4-methyl-1-piperazinyl group, and it is presumed to  15 16 17 18 19 20 21 22 23 24   1 2 3 4 5 6 7 8 9 10 11 12 13 14 Time (min)   15 16 17 18 19 20 21 22 23 24   1 2 3 4 5 6 7 8 9 10 11 12 13 14 Time (min)   15 16 17 18 19 20 21 22 23 24   1 2 3 4 5 6 7 8 9 10 11 12 13 14 Time (min)   15 16 17 18 19 20 21 22 23  be an N-methylation reaction on the amino group of the benzamide group or the amino group of 2-aniline on the pyrimidine ring. e fragment ion at m/z 264 was generated by the breakage of the amide bond after the breakage of the methyl group of C 4 on the benzene ring of 2-anilinopyrimidine and the formation of a five-membered ring at this site with C 1 -N of the pyrimidine ring. Other fragment ions were also broken in the same way as above. erefore, it was predicted that M23 is an amino group from the benzamide group of IM or an N-methylated metabolite from the amino group of 2-aniline on the pyrimidine ring.
Metabolite M15 (C 28 H 30 N 7 O), with a [M + H] + of m/z 480.2505, was eluted at 10.52 min, which was 14 Da less than that of IM. M15 had the same fragment ion of m/z 394 as the above metabolites and two new fragments. One (m/z 119) formed via the separation of the carbonyl group from the benzene ring. e breakage of the amide bond to the benzene ring after N-demethylation of IM yields fragment ion of m/z 203, and the m/z 222 fragment ion was the methyl group of C 4 from the benzene ring of 2-anilinopyrimidine after the breakage of this site and the C 1 -N of the pyrimidine ring to form a five-membered ring. C 1 , C 6 , C 4 , and C 5 are broken and the intramolecular proton migration rearrangement was generated. e fragment ion at m/z 394 was broken in the same way as M6, so it is presumed that M15 was the N-demethylated metabolite of IM.
In addition, M11 (C 18      e structure of the 1methylpiperazine group cracked the metabolite, resulting in the fragment ion m/z 99. As a result, M12 is molecular imatinib, a member of hydroxyl and carbonyl replacement metabolites (replacement occurred in 4-methyl-3-[4-(3pyridyl)-2-pyrimidine structures). e fragment of m/z 290 was created via the breakage of the methyl group of C 4 from the benzene ring of 2-anilinopyrimidine after the formation of a five-membered ring with C 1 -N of the pyrimidine ring and the break of the carbonyl group and N on the benzoylamino group, which also generates the m/z 217 fragment ion, and the m/z 99 fragment ion produced from the break of the 1-methyl piperazinyl group break generated presumably M12, that is, the 4-methyl-3-[4-(3-pyridyl)-2-pyrimidine of IM with one molecule of hydroxyl and one molecule of carbonyl-substituted metabolite. e m/z 478 fragment ion of M22 was generated by N-demethylation in the prototype of IM, and the m/z 247 fragment ion further deamidates the remaining carbonyl CO to produce the m/z 219 fragment ion of M22. erefore, it was presumed to be oxidized to carboxylic acid on N-methyl, and M22 was a metabolite of oxidation of methyl to carboxylic acid on N-methyl of IM.
e extracted ion chromatograms of all bile metabolites are shown in Figure 6. e chromatographic retention behavior of the metabolites, quasimolecular ions, and secondary fragment ions was comprehensively analyzed (Table 1). It was found that 12 of them were consistent with the metabolites in plasma, and the other 9 were different. Fragmental ions are shown in Figure 5. erefore, it was speculated that the methyl group of 4methylbenzamide was oxidized to a carboxylic acid. Based on the above results, it was inferred that M25 was a metabolite of IM depleted 4-methyl-1-piperazinyl carboxylic acid. e fragment ion at m/z 133 was generated by the CO-N cleavage of the amide bond of benzamide after M20 carbonyl substitution. e fragmentation of other ions was the same as the analysis. According to the metabolic law, it was speculated that the methyl group of 4-methylbenzamide was oxidized to aldehyde. us, it was inferred that M20 was formed by removal of N-      quasimolecular ions, and secondary fragments ions are shown in Table 1.

Metabolite Analysis in Feces.
Compared with blank stool samples, in addition to the IM prototype drug, 12 compounds were also detected. e extracted ion chromatograms of all feces metabolites are shown in Figure 8. Comprehensive analysis of chromatographic retention behavior, quasimolecular ions, and secondary fragments ions is presented in Table 1.

Metabolic Pathways of IM in Rats.
In summary, the UHPLC-Q-TOF-MS/MS method was used to study the metabolism of IM in rats. A total of 27 metabolites were found, including 22 phase I metabolites and 5 phase II metabolites. e phase I metabolic pathways contained hydroxylation, oxidation, catalytic dehydrogenation, and dealkylation; the phase II metabolic pathways were N-acetylation, methylation, cysteine binding reactions, and glucuronidation. ese metabolic pathways may also cross-link each other to generate secondary metabolites. Based on the above results, the main metabolic pathways of IM in rats are summarized in Figure 9.

Discussion
In the results of this experiment, M1-M2, M4-M8, M10, M12-M16, M18-M19, M21, and M24-M25 were previously reported metabolites of IM. Among them, the characteristic  fragment ions and structures of M6, M7, M10, M14, M15, and  M18 were consistent with those of M3, M8, M5, M6, N-demethyl metabolite, and M7 in the actual test results of Marull and Rochat [18]. In this experiment, the structures of M13, M21, and M25 were consistent with the previously reported structures of M29.6 and APG050 and M42.2 [23]. e oxides and dioxides in the results of Friedecký et al. (metabolites 59-61 and 65-74) may be consistent with M16 and M1, M4, M8, M12, and M19 in this study because they shared the same molecular formula and weight [20]. M5 was consistent with the previously reported dissociation products of M519 B because they had the same characteristic fragment ions and chemical structures [19]. M24 was also similar to the reported M609, and they all have fragment ions at m/z 215, m/ z 394, m/z 478, and m/z 565 [24]. In addition, the experimental results of M2 may be the same as metabolite No. 25 of Rochat et al. [25], but they did not provide complete characteristic fragments ion and structural formula. erefore, new content was added to this study. With regard to the effect of IM metabolites, N-demethylated metabolites of IM (M15 in this experiment) were the main metabolites of IM, although they had similar effects. However, some studies have shown that the pharmacological activity of N-demethyl IM was three times lower than that of IM [26]. Mlejnek et al. also found that N-demethyl IM had almost no therapeutic effect on CML through K562 cell experiments [17]. In addition, some studies have shown that some metabolic compounds, especially highly demethylated and sulfur-containing compounds (M15 and M24 in this experiment) may be related to IM adverse reactions [24,26].
With regard to the newly discovered IM metabolites in this experiment, M22 is an IM metabolite with a carboxyl group, which was widely used in drug design and synthesis and was often used to form pharmacophore of various drugs. In addition, this group was reported to play a key role in the interaction between drugs and targets [27]. M3 and M11 were newly discovered N-acetylated metabolites in this experiment. Studies have shown that acetylated compounds can induce apoptosis in CML cells (k562) and IM-resistant CML cells (IR-K562). erefore, these metabolites may have the potential of value for the drug development industry [28]. M23 was a newly discovered N-methylated metabolite. It has been reported that methylation biomarkers were closely related to drug resistance, and it may help with the study on drug resistance of CML [29]. M9 and M27 were the metabolites of dealkylation. Studies showed that the dealkylation metabolites of TKI had the same effect as the original drug. erefore, M9 and M27, as newly discovered IM dealkylation metabolites, may also have the potential drug activity [21]. M26 and M20 were metabolites with aldehyde groups, which can lead to inactivation of cytochrome P450, drug interaction (DDI), and hepatotoxicity. ey may provide a new way to investigate the toxicity of TKI drugs [16]. M17 was the first-phase metabolite of N-demethylation and catalytic dehydrogenation, and it was a new metabolite discovered for the first time.
e metabolites of IM and the fragment ions of IM were mainly produced by the cleavage of the C-N bond at the amide site, and the metabolic position of IM metabolites was determined by the increase or decrease of the molecular weight of the corresponding fragment ions. For example, N-oxidized metabolites tend to lose O on N during mass spectrometry cleavage, thus forming fragment ions lacking 16 Da. erefore, the type and location of metabolic reaction can be determined by comparing the changes of the ion-to-nucleus ratio of fragments. In addition, in the analysis of IM and its phase II metabolites by two-phase full scan mass spectrometry, it was found that the glucuronic acid conjugate and cysteine conjugate first removed the binding group and obtained the corresponding mother nucleus.

Conclusions
In this study, the metabolic spectrum of IM in rats was studied using UHPLC-Q-TOF-MS/MS technique. 27 IM metabolites (18 known and 9 unknown) were found in plasma, bile, urine, and feces, including 18 in plasma, 21 in bile, 18 in urine, and 12 in feces. ere were 22 metabolites in phase I. M6, M7, M10, M14, and M18 were formed by hydroxylation. M13, M16, M21, and M26 were formed by oxidation, and M5 was formed by catalytic dehydrogenation. M9, M15, and M27 were generated by dealkylation metabolism. M1, M4, M8, M12, M17, M19, M20, M22, and M25 were produced by cross-reaction. ere were five phase II metabolites, M2 was formed by glucuronic acid-binding reaction, and M23 is through methylation. M11 and M3 were formed by N-acetylation and M24 was due to the cysteine binding reaction. IM and its phase I metabolites can be combined with endogenous substances to greatly improve the water solubility of metabolites and the level of drug excretion. However, the methylation metabolite M23 (II phase) of IM was on the contrary. Under the action of methylation, the methylated compound S-adenosyl methylthiamine (SAM) was used as a cofactor. After being catalyzed by methyltransferase, methyl was transferred to the O, S, N atoms of the substrate to form a less polar product, resulting in a decrease in the water solubility of the metabolite, making it difficult to be excreted. Since more metabolites were found in bile and urine, it was suggested that IM and its metabolites rely on bile and urine for metabolism and excretion.

Data Availability
All data generated or analyzed during this study are included within this article.