Regular Articles
Elucidation of Degradation Behavior of Tricyclic Antidepressant Amoxapine in Artificial Gastric Juice
2020 Volume 68 Issue 9 Pages 848-854
Details
2020 Volume 68 Issue 9 Pages 848-854
The degradation behavior of eight tricyclic antidepressants (TCAs; amitriptyline, amoxapine (AMX), imipramine, clomipramine, desipramine, doxepin, dothiepin, and nortriptyline) in artificial gastric juice was investigated to estimate their pharmacokinetics in the stomach. As a result, among the eight TCAs, only AMX was degraded in artificial gastric juice. The degradation was a pseudo first-order reaction; activation energy (Ea) was 88.70 kJ/mol and activation entropy (ΔS) was −80.73 J/K·mol. On the other hand, the recovery experiment revealed that the degradation product did not revert to AMX and accordingly, this reaction was considered to be irreversible. In the AMX degradation experiment, peaks considered to be degradation products A (I) and B (II) were detected at retention times of around 3 min and 30 min in LC/UV measurements, respectively. Structural analysis revealed that compound (I) was [2-(2-aminophenoxy)-5-chlorophenyl]-piperazin-1-yl-methanone, a new compound, and compound (II) was 2-chlorodibenzo[b,f][1,4]oxazepin-11(10H)-one. As for the degradation behavior, it was estimated that AMX was degraded into (II) via (I), i.e., (II) was the final product. The results are expected to be useful in clinical chemistry and forensic science, including the estimation of drugs to be used at the time of judicial dissection and suspected drug addiction.
Tricyclic antidepressants (TCAs) are psychotropic drugs that are used for the treatment of depression. However, some TCAs have many side effects and are likely to cause acute drug addiction through overdosing.1) The symptoms are often life-threatening and include arrhythmia, prolonged convulsion, seizure, coma, metabolic acidosis, and acute respiratory failure.2) It was reported that generalized seizures occurred repeatedly due to dothiepin overdose and the patient died despite lifesaving treatment including gastric lavage for 30 h.3) In addition, there have been cases of death due to cardiac paralysis as a result of chronic use of amoxapine (AMX).4)
For patients who present with consciousness disorder and seem to be suffering from drug addiction, drug identification is an urgent task. A pharmacokinetic study5) showed that AMX, after rapid absorption, is metabolized to 7-hydroxyamoxapine and 8-hydroxyamoxapine by the hepatic microsomal enzyme system. As a biological specimen, gastric lavage fluid may be useful to identify the drug ingested. That is, the detecting TCAs in gastric lavage fluid has demonstrated to be very important, as it can lead to evaluating therapeutic efficacy or determining the cause of death. Therefore, it is important to understand the pharmacokinetics, such as the degradation behavior of drugs in stomach. Although there are reports6–10) of the determination of TCAs and their in vivo metabolites in plasma or urine based on HPLC, there is little information on their behavior in the stomach, particularly their behavior under increased gastric acidity. On the other hand, a similar psychotropic drug, benzodiazepines (BZPs), has been used in criminal cases,11) such as murder, robbery, and rape after the victim fell into a coma. Thereof, the related researches for the degradation behavior of BZPs in acidic condition have already been in progress.12–14)
In this study, we investigated the degradation behavior of eight TCAs in artificial gastric juice to infer their pharmacokinetics in the stomach. For drugs that were degradable, their degradation rates were determined by conducting experiments to examine whether the degradation proceeds faster by increasing the storage temperature, or the degradation is reversible by varying the pH of the solution. The effect of storage conditions on drug degradability was examined by determining the physicochemical parameters. Furthermore, the degradation products were isolated and purified for LC/photodiode array detection (PDA), LC/time-of-flight mass spectrometer (TOF-MS), IR, NMR, and X-ray diffraction measurements, etc., and structural analysis was performed.
AMX (biochemistry grade) was purchased from LKT Laboratories, Inc. (St. Paul, MN, U.S.A.). Doxepin and dothiepin (both biochemistry grade) were purchased from Sigma-Aldrich Corp. (St. Louis, MO, U.S.A.). Amitriptyline, imipramine, clomipramine, desipramine, and nortriptyline (all biochemistry grade) were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). Each standard was dissolved in acetonitrile, methanol or water to make a 1 mg/mL standard stock solution. Working standard solutions were then prepared from the standard stock solutions by dilution with the same solvent, respectively.
Acetonitrile and methanol (both HPLC grade); formic acid, phosphoric acid, potassium dihydrogen phosphate, dipotassium hydrogen phosphate, and sodium tetraborate,10-hydrate (borax; all special grade (> 97%)); 2-mercaptoethanol and hydrochloric acid (both precision analysis grade); and leucine–enkephalin (biochemical grade) were purchased from FUJIFILM Wako Pure Chemical Corporation. Ortho-phthalaldehyde (fluorescence analysis grade) was from Nacalai Tesque Inc. (Kyoto, Japan). Sodium chloride, formic acid, and methanol-d4 (all special grade) were from Kanto Chemical Co., Inc. (Tokyo, Japan). Dimethyl sulfoxide-d6 (DMSO-d6; special grade) was from Merck Ltd. (Tokyo, Japan) and water was purified with a Milli-Q Gradient A10 system equipped with an EDS-PAK® polisher (Merck Ltd.). All other chemicals were of special grade.
An Oasis® HLB cartridge (200 mg, particle size 30 µm; Waters Co., Milford, MA, U.S.A.) was used. The cartridge was conditioned with methanol and water prior to use.
Apparatus and Operating ConditionsLC/PDAAn LC-10AD pump system equipped with an SPD-10A (Shimadzu, Kyoto, Japan) and an L-6300 Intelligent pump (Hitachi, Tokyo, Japan) system equipped with an MD-910 photodiode array detector (PDA; JASCO Corporation, Tokyo, Japan) were used. LC separation was performed with a Poroshell 120 EC-C18 column (100 × 4.6 mm I.D., 2.7 µm; Agilent Technologies, Santa Clara, CA, U.S.A.). Column temperature was maintained at 50°C. The mobile phase was a mixture of 10 mmol/L potassium phosphate buffer solution (pH 3) and acetonitrile in the ratio of 70 : 30 (v/v), and was delivered in the isocratic elution mode at the flow rate of 0.25 or 0.5 mL/min. The UV detection or monitoring wavelength was set at 220 nm. A 10 µL aliquot of the sample was injected into the system.
LC/TOF-MSAn Alliance HT 2795 HPLC system equipped with an LCT Premier XE TOF-MS (Waters Corporation, Milford, MA, U.S.A.) was used. LC separation was performed with a Poroshell 120 EC-C18 column (100 × 4.6 mm I.D., 2.7 µm; Agilent Technologies). Column temperature was maintained at 50°C. The mobile phase was a mixture of 10 mmol/L formic acid aqueous solution and acetonitrile in the ratio of 70 : 30 (v/v), and was delivered in the isocratic elution mode at the flow rate of 0.5 mL/min. A 10 µL aliquot of the sample was injected into the system.
The optimum working parameters for TOF-MS were as follows: electrospray ionization (ESI): positive mode, capillary voltage: 2000 V, cone voltage: 50 V, aperture #1 voltage: 15 V, desolvation temperature: 350°C, source temperature: 120°C, desolvation gas flow (N2): 800 L/h, and cone gas flow (N2): 50 L/h. Leucine-enkephalin solution (0.25 µL/mL) was delivered at the flow rate of 5 µL/min. Mass accuracy was maintained using Lock–Spray with the leucine–enkephalin [M + H]+ ion, m/z = 556.2771, as the lock mass. The resolution was at least 10000 as calculated by using the full width at half-maximum method.
IRAn FT/IR-4100 Fourier transform IR spectrometer (JASCO) was used as the IR device. The transmission method (MagHoldIR, using 5 × 5 mm KBr plate) was used as the measurement method, and Deuterated L-Alanine Triglycine Sulphate (DLATGS) was used as the detector. As for the operating conditions, the resolution was 4 cm−1 and one window plate (5 × 5 mm) was set. Approximately 5 µL of sample solution was dropped on the plate and left to dry. The same operation was performed eight times to produce a thin film (approximately 40 µL of sample solution was used). The thin film was fixed with a cover plate, set to the holder for transmission, and measured.
NMRAn ECA-600II NMR spectrometer (JEOL, Tokyo, Japan) was used as the NMR apparatus. The frequency for 1H measurement was set at 600.1723 MHz, and that for 13C measurement was set at 150.9134 MHz. As the reference peak, tetramethylsilane (0 ppm) or solvent peaks of methanol-d4 (1H: 3.31 ppm, 13C: 49.0 ppm) and DMSO-d6 (1H: 2.49 ppm, 13C: 39.5 ppm) were used.
X-Ray Crystallography (XRC)A Super Nova (Rigaku, Tokyo, Japan) was used as the X-ray source and the operating conditions were as follows: X-ray was generated by Kα radiation (λ = 1.54187 Å) of Cu. X-ray crystallographic data were collected at 100 K. As the other apparatuses, an SCI-165 CO2 incubator (ASTEC, Tokyo, Japan) and an MOV-112S dry heat sterilizer (SANYO, Tokyo, Japan) were used as the storage device at 38 to 70°C.
Degradation Behavior of TCAs in Artificial Gastric JuiceDegradation ExperimentTo 900 µL of artificial gastric juice (2.0 g of sodium chloride dissolved in 7.0 mL of hydrochloric acid and purified water added to make 1000 mL; pH 1.2) was added 100 µL of each TCA standard solution (100 µg/mL each of amitriptyline, AMX, imipramine, clomipramine, desipramine, doxepin, dothiepin, and nortriptyline). Then, the solutions were stored in an incubator set to 38°C and time-dependent degradation over an approximately four-week period was determined by LC/UV measurements. In the accelerated degradation test, AMX, which was confirmed to be degraded in the degradation experiment, was dissolved in artificial gastric juice similar to the above and stored in an incubator set at 50, 60, and 70°C. The time course of AMX degradation was determined by LC/UV measurements. As control experiment, 900 µL of AMX standard solution (100 µg/mL, acetonitrile solution) was added to 100 µL of purified water and the solution was stored at 70°C and then measured.
Recovery ExperimentIn experiments to verify the recovery behavior of AMX degradation products by pH readjustment, 1 M phosphate buffer adjusted to pH 5 and 7, and 900 µL of 0.1 M borate buffer adjusted to pH 9 and 11 were used. To 900 µL of each buffered solution, 100 µL of AMX degradation solution, which was preliminarily prepared by adding 100 µL of AMX standard solution (100 µg/mL, acetonitrile solution) to artificial gastric juice (pH 1.2) and storing at 70°C for 4 d, was added. Then, the solutions were stored in an incubator set at 38°C and time-dependent changes over an approximately one-week period were determined by LC/UV measurements.
Isolation and Purification of AMX Degradation ProductsDegradation Product B (II)Twenty milligrams of AMX was dissolved in 30 mL of artificial gastric juice and the solution was stored at 70°C for four days to precipitate crystals. The solution containing the crystals was centrifuged (2580 × g, 10 min) and the supernatant was removed. The remaining crystals were washed with 1 mL of purified water and centrifuged again (2580 × g, 10 min). After repeating this operation twice, the precipitated crystals were redissolved in methanol, transferred into a microtube, and allowed to stand at room temperature to isolate (II) as colorless needle crystals.
Degradation Product A (I)The supernatant (approx. 30 mL) obtained when (II) was isolated and purified, as described above, was purified using the solid-phase extraction cartridge Oasis® HLB cartridge (150 mg × 6). Five milliliters of supernatant was processed in one Oasis® HLB cartridge. After loading the supernatant and washing with 2 mL of purified water, 2 mL of 70% methanol was used for elution from the solid phase. All the eluted solutions were combined, 30 mL of 100% ethanol was added, and azeotropic distillation was carried out with an evaporator. The residue was redissolved in 100% methanol, transferred into a microtube, and dried by nitrogen purge to isolate (I) as a colorless liquid.
Fluorescence Derivatization of Degradation Product A (I)As the fluorescence derivatization reagent, 50 mM borax solution containing 5 mM OPA and 5 mM 2-mercaptoethanol was used. The preparation procedure is as follows: Borax (0.95 g) was added to a solution obtained by adding 3.0 mL of methanol to 33.5 mg of OPA in advance, and purified water was added thereto to make 50 mL. Then, 20 µL of 2-mercaptoethanol was further added. Degradation product A ((I): 1 mg) that was isolated and purified as described in “Degradation Product A (I)” was redissolved in 100% methanol (2 mL) and approximately 3 µL was spotted on a TLC plate (adsorbent: silica gel). The spot was allowed to dry naturally. Thereafter, 20 µL of a fluorescence derivatization reagent was dropped with a micropipette while irradiating the TLC plate with UV light (365 nm) in a dark room, and the presence or absence of fluorescence was observed.
LC/UV measurement conditions were examined with reference to published papers.6–10) When analyzing the eight TCAs simultaneously, the flow rate was set at 0.25 mL/min in consideration of their mutual separation, and when examining the degradation behavior of AMX alone, the flow rate was set at 0.5 mL/min. Figure 1 shows the LC/UV chromatograms of the standard solutions of the eight TCAs. Only AMX was degraded among the eight TCAs (Fig. 2). The chemical structure of AMX differs considerably from those of the other tricyclic antidepressants (Fig. 1) in that AMX has a piperazine ring bound to the tricyclic skeleton via the single bond of azomethine. Therefore, it was thought that this azomethine bond would likely rotate and be cleaved off, that is, AMX is likely to degrade more easily than the other tricyclic antidepressants.
Sampling times (h) were as follows: AMX: 0, 0.5, 96, 144, 456, 600; Clomipramine: 0, 0.5, 192, 360, 600; Desipramine: 0, 0.5, 24, 192, 312, 456, 600; Imipramine: 0, 96, 312, 456, 600; Amitriptyline: 0, 96, 144, 288, 600; Nortriptyline: 0, 0.5, 24, 192, 360, 600; Doxepin: 0, 0.5, 1, 24, 158, 246, 432, 600; Dothiepin: 0, 0.5, 24, 168, 312, 456, 600.
Focusing on the AMX degradation products, the peaks that were thought to be degradation products were detected at retention times of around 3 and 30 min (Fig. 3). Hereinafter, they are referred to as degradation product A (I) and degradation product B (II), respectively.
AMX stored at 38°C for 25 d was not completely degraded. Thus, an accelerated degradation test was carried out in which the storage temperature was set at 50, 60, and 70°C to reduce the reaction time. As a result, the degradation rate of AMX increased with increasing storage temperature, and AMX was almost completely degraded in four days at 70°C (Fig. 4). In addition, because both (I) and (II) were observed at all temperatures from the initial stage of the AMX degradation experiment, two pathways, namely, AMX →(I) and AMX →(II), were speculated as the AMX degradation pathway. However, AMX →(I) →(II) or AMX →(II) →(I) was also possible, i.e., one could be an intermediate degradation product and the other, the final degradation product. This was difficult to verify from the above experiment alone. Therefore, in order to verify this hypothesis, we decided to isolate and purify (I) and (II) and confirm their degradation behavior in artificial gastric juice. In addition, physicochemical parameters, such as the degradation rate of AMX and the production rates of (I) and (II), were also calculated. On the other hand, in a control experiment where AMX was dissolved in purified water without using artificial gastric juice, AMX was hardly degraded even when stored at 70°C, and neither (I) nor (II) was detected. These results indicate that both (I) and (II) degraded under acidic conditions.
In order to confirm the degradation pathway of AMX, physicochemical parameters, such as reaction rate constant, activation energy (Ea), and activation entropy (ΔS), in the degradation reaction were calculated using the following formulas.
 |
As a result, it was found that the degradation of AMX was a pseudo first-order reaction. Furthermore, when an Arrhenius plot of the degradation reaction of AMX was created and its physicochemical parameters were determined, Ea was 88.70 kJ/mol and ΔS was −80.73 J/K·mol. From these results, the degradation reaction of AMX had a high energy barrier at the start of the reaction, and because it was an endothermic reaction, enthalpy change ΔH was positive (> 0) and ΔS was negative (< 0). Gibbs free energy change (ΔG) was positive (> 0), suggesting that the AMX degradation reaction usually proceeded with difficulty. However, as the degradation reaction progressed even when stored at 38°C, which is nearly the same as human body temperature, it seemed necessary to pay close attention to the behavior of the degradation products in the body.
Experiments on Recovery Behavior of AMX Degradation Products by pH ReadjustmentIn order to confirm whether AMX degradation products formed under acidic conditions showed reversibility, i.e., whether the products could revert to the original AMX by changing the pH from acid to neutral or basic, a recovery experiment was conducted. Both (I) and (II) did not revert to AMX even when stored at any of pH 5, 7, 9, and 11. Therefore, it was confirmed that this degradation reaction is an irreversible reaction.
Structural Analysis of Degradation ProductsIn “Isolation and Purification of AMX Degradation Products,” (I) was purified as a liquid and (II) was purified as crystals. The former was used in LC/PDA, LC/TOF-MS, IR, and NMR measurements and fluorescence derivatization, whereas the latter was subjected to LC/PDA, LC/TOF-MS, NMR, melting point measurements, and single-crystal X-ray structure analysis.
Degradation Product A (I)Compound (I) (5.09 mg) was obtained as a colorless liquid. UV λmax peaks appeared at 287, 227, and 203 nm, as shown in Fig. S1. From the result of high resolution (HR)-ESI-MS ([M + H]+ m/z 332.100, Calcd for C17H19ClN3O2: 332.1166) by LC/TOF-MS, the molecular formula of (I) was determined to be C17H18ClN3O2, i.e., it has one extra H2O molecule relative to AMX. The absorption peak at 1630 cm−1 in the IR spectrum of (I) indicated the presence of an amide group (Fig. S2). Compound (I) emitted blue fluorescence in the presence of OPA derivatization reagent. Therefore, it was suggested that (I) had a primary amino group. The NMR spectra of (I) measured in DMSO-d6 are shown in Supplementary Materials (Figs. S3–7). The 1H-NMR spectrum of (I) indicated the presence of a 1,2-disubstituted benzene ring [δ 6.95 (1H, t, J = 7.6 Hz), δ 6.85 (1H, d, J = 7.6 Hz), δ 6.80 (1H, d, J = 7.6 Hz), and δ 6.56 (1H, t, J = 7.6 Hz)] and a 1,2,4-trisubstituted benzene ring [δ 7.49 (1H, d, J = 2.8 Hz), δ 7.39 (1H, dd, J = 2.8, 8.9 Hz), and δ 6.69 (1H, d, J = 8.9 Hz)]. In addition, proton peaks indicating the presence of a piperazine ring were observed at δ 2.87–3.85 ppm (Table S1).
From the 13C-NMR spectrum of (I), 13 sp2 carbons including a carbonyl carbon (δ 164.8 ppm) derived from an amide group and four sp3 carbons were observed. These results suggested the presence of two benzene rings, one amide group, and four methylene carbons derived from a piperazine ring (Table S1). The correlations from H-6 to H-9 in the 1H–1H correlation spectroscopy (COSY) spectrum of (I) and the heteronuclear multiple-bond connectivity (HMBC) correlations from 10-NH2 to the sp2 carbon of C-5a, C-9, and C-9a of (I) indicated the presence of a 1,2-disubstituted benzene ring with a primary amino group.
HMBC correlations from H-1, H-3, and H-4 to C-4a (152.6 ppm), which was adjacent to an oxygen atom, and from H-1 to C-11 (164.8 ppm), the carbonyl carbon in the amide group, indicated the presence of a 1,2,4-trisubstituted benzene ring with an amide group and a neighboring oxygen atom.
1H–1H COSY correlations between H-13 and H-14, and H-16 and H-17 indicated the presence of the piperazine ring. HMBC correlation from H-13 to C-11 revealed that the carbonyl carbon and the piperazine ring were linked.
From the above results, the structure of (I) was determined to be [2-(2-aminophenoxy)-5-chlorophenyl]-piperazin-1-yl-methanone, as shown in Fig. 5. Compound (I) was thought to be formed by hydrolysis of the imine moiety of AMX. It had also been reported that the exogenous AMX derived from Cunninghamella elegans was metabolized to three metabolites, which were identified by NMR and MS as 7-hydroxyamoxapine, N-formyl-7-hydroxyamoxapine, and N-formylamoxapine.15) However, to date, there has been no mention of compound (I) in reports of the degradation products and metabolites of AMX. Therefore, (I) is a new compound.
Compound (II) (16.55 mg) was obtained as colorless needle crystals. Its melting point was at 242.3°C and UV λmax peaks were found at 275, 227, and 207 nm, as shown in Fig. S2. From the results of HR-ESI-MS ([M + H]+ m/z 246.026, calcd for C13H9ClNO2: 246.0322), the molecular formula of (II) was determined to be C13H8ClNO2. Compound (II) was confirmed to be 2-chlorodibenzo[b,f][1,4]oxazepin-11(10H)-one from the detailed analyses of its 1H- and 13C-NMR spectra and two-dimensional NMR spectra (Figs. S8–12, Table S2), which were measured in methanol-d4, and the structure is shown in Fig. 6. Compound (II) is a known substance16) (CAS Number 3158-91-6). When the melting point of compound (II) was measured (n = 3), the average value was 242.3°C, which coincided with the literature value (242–244°C) of 2-chlorodibenzo[b,f][1,4]oxazepin-11(10H)-one. Furthermore, the result of single-crystal X-ray structure analysis also supported the structure of (II) (Fig. 7). The crystallographic data have been deposited with the Cambridge Crystallographic Data Center as supplementary publication No. CCDC 1959210.
LC/UV measurement of a sample solution of (I), which was stored at 70°C for four days, revealed the formation of (II). Furthermore, (I) was further degraded by continuous heating for 12 d (16 d in total; Fig. 8).
Degradation product A (I) was added to the solution. The amount of degradation product A (I) is the residual amount, and that of degradation product B (II) is the amount produced in the solution.
On the other hand, storage of a sample solution of (II) at 70°C for four days revealed that (II) hardly degraded and no new peaks of degradation products were observed. From these results, it was inferred that (I) was an intermediate degradation product and (II) was the final degradation product in the AMX degradation reaction (Fig. 9).
Of the eight tricyclic antidepressants targeted for measurement, only AMX was degraded in artificial gastric juice, producing degradation products (I) and (II). The AMX degradation reaction was a pseudo first-order reaction; Ea was 88.70 kJ/mol and ΔS was −80.73 J/K·mol. As this degradation reaction was an endothermic reaction, ΔH was positive (> 0) and ΔS was negative (< 0), so ΔG was positive (> 0), suggesting that the AMX degradation reaction usually proceeded with difficulty. However, because the degradation reaction proceeded even when stored at 38°C, which is nearly the same as human body temperature, it seemed necessary to pay close attention to the behavior of the degradation products in the body. This degradation reaction was irreversible because AMX was not recovered even when the pH of the solution after the reaction was changed from neutral to basic.
Structural analysis revealed that compound (I) had a similar structure to AMX and its chemical formula was C17H18ClN3O2, [2-(2-aminophenoxy)-5-chlorophenyl]-piperazin-1-yl-methanone, making it a new compound. On the other hand, compound (II) had the same structure as AMX minus a piperazine ring, and its chemical formula was C13H8ClNO2, 2-chlorodibenzo[b,f][1,4]oxazepin-11(10H)-one.
From the confirmation experiment of the degradation pathway, it was deduced that compound (I) was an intermediate degradation product and compound (II) was the final degradation product in the AMX degradation reaction.
Based on the above results, in addition to AMX, compounds (I) and (II) would be a useful target substance for detecting AMX in stomach contents, such as during gastric lavage and judicial dissection in acute drug poisoning. For instance, the detection of AMX and compounds (I) and (II) in stomach may be useful for estimating gastric residence time to some extent. It may also enable verification of the estimated time of death in overdose or murder cases.
This study was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (Grant No. 18K06609). We would like to express our gratitude to JASCO Engineering Co., Ltd. (Tokyo, Japan) for cooperation in IR measurement.
The authors declare no conflict of interest.
The online version of this article contains supplementary materials.
