2023 年 71 巻 4 号 p. 312-317
The degradation behavior of three benzodiazepines (BZPs)—lormetazepam (LMZ), lorazepam, and oxazepam—with hydroxy groups on the diazepine ring in artificial gastric juice and the effect of storage pH conditions on drug degradability were monitored using an LC/photodiode array detector (PDA) to estimate their pharmacokinetics in the stomach. Although the three BZPs degraded in artificial gastric juice, none could be restored, despite increasing the storage pH, implying that the degradation reaction was irreversible. As for LMZ, we discussed the physicochemical parameters, such as the activation energy and activation entropy involved in the degradation reaction as well as the reaction kinetics; one of the degradation products was isolated and purified for structural analysis. In the LMZ degradation experiment, peaks corresponding to degradation products, (A) and (B), were detected through the LC/PDA measurements. Regarding the degradation behavior, we hypothesized that LMZ was degraded into (B) via (A), where (A) was an intermediate and (B) was the final product. Although the isolation of degradation product (A) was challenging, degradation product (B) could be isolated and was confirmed to be “methanone, [5-chloro-2-(methylamino)phenyl](2-chlorophenyl)-” based on structure determination using various instrumental analyses. The compound exhibited axis asymmetry as determined using single-crystal X-ray structure analysis. Because the formation of degradation product (B) was irreversible, it would be prudent to target the final degradation product (B) and LMZ for identification when detecting LMZ in human stomach contents, such as during forensic dissection.
Benzodiazepines (BZPs) are frequently used as psychotropic drugs because of their anticonvulsant, anxiolytic, and sedative properties.1) In contrast, they are also used as “rape drugs” in crimes such as robbery, rape and murder by mixing them in beverages as they impart hypnotic and amnesic effects.2,3) BZPs have also been reported to increase suicidal tendencies.4,5) In cases where such drug use is suspected, it is necessary to demonstrate its usage. However, in most cases, drug appraisal is often performed sometime after the crime,6,7) thereby possibly making it difficult to detect drugs that do not have long-term stability in the body.8) In fact, in a murder case with suspected flunitrazepam use, detection of the drug in the biological sample of the victim was challenging,9) because flunitrazepam degrades after long-term exposure to acid in gastric juice. With regard to the physical properties of BZPs, it is known that the azomethine bond of the diazepine ring found in several BZPs is cleaved under acidic conditions, although ring closure occurs again under neutral conditions.10–12) These studies suggest that orally ingested BZPs are degraded in the stomach but are restored to their original state in the small intestine.
For many years, we have earnestly studied the degradability and restoration of psychotropic drugs gastrointestinal tract. Our efforts have resulted in the elucidation of the degradation pathway of amoxapine, a tricyclic antidepressant, and the determination of the structure of its new degradant (“Study on Degradability of Drugs in Stomach (I)”).13) In addition, we have also studied the degradation behavior of eight BZPs commonly used in medical practice in the stomach, and have revealed significant differences in degradability (stability) depending on the structural formula, namely, extremely high or low degradability and almost 100% restoration or no restoration at all (“Study on Degradability of Drugs in Stomach (II)”).14) In this study, we focused on the fact that lorazepam, which did not show any restoration in previous study (II), has a hydroxyl group on the diazepine ring, and decided to verify its degradability and restoration again together with lormetazepam (LMZ) and oxazepam, which have similar structures (Fig. 1).
Additionally, the physicochemical parameters involved in the degradation reaction of LMZ were determined, and the structure of the degradation product of LMZ was analyzed by instrumental analysis, such as LC/time-of-flight (TOF)-MS, NMR spectroscopy, and X-ray crystallography.
LMZ (for pharmacological research), lorazepam, and oxazepam (biochemical grade) were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). Each of the standards was dissolved in methanol to prepare a 1 mg/mL standard stock solution, which was stored in an amber glass bottle at 4 °C in a refrigerator. Working standard solutions were prepared from each standard stock solution by dilution with water. Acetonitrile, methanol (HPLC grade), acetone, formic acid, methanol-d4 (99.8%, NMR grade), and deuterium water (99.8%, NMR grade) were purchased from Kanto Chemical Co., Inc. (Tokyo, Japan). Ethanol, phosphoric acid, potassium dihydrogen phosphate, disodium hydrogen phosphate, sodium dihydrogen phosphate, boric acid, sodium tetraborate decahydrate (borax), acetic acid, sodium acetate, 28% ammonia, naphthalene, D-fructose, (all special grade), paraffin (Wako 1st grade), hydrochloric acid (HCl, precision analysis grade), and leucine-enkephalin (biochemical grade) were purchased from FUJIFILM Wako Pure Chemical Corporation. Water was purified using a Milli-Q Gradient A10 system equipped with an EDS-PAK® polisher (Merck KGaA, Darmstadt, Germany). All other chemicals were of special grade.
An Oasis® HLB 1 cc Vac Cartridge (30 mg, particle size 30 µm) and an Oasis® MCX 1 cc Vac Cartridge (30 mg, particle size 30 µm; Waters Corporation, Milford, MA, U.S.A.) were used. The HLB cartridge was conditioned with methanol and water; whereas the MCX cartridge was conditioned with methanol and 1% formic acid before use.
The artificial gastric juice and each pH buffer solution were prepared as follows:
Artificial gastric juice (pH 1.2): 2.0 g of sodium chloride was dissolved in 7.0 mL of hydrochloric acid, and purified water (1000 mL) was added.
A buffer solution (pH 5) was prepared by adding 1 mol/L acetic acid aqueous solution to a 1 mol/L sodium acetate aqueous solution until the pH reached 5.0. Another buffer solution (pH 7) was prepared by adding 1 mol/L disodium hydrogen phosphate aqueous solution to a 1 mol/L sodium dihydrogen phosphate aqueous solution until the pH reached 7.0. A third buffer solution (pH 9) was prepared by adding 1 mol/L boric acid aqueous solution to a 0.1 mol/L borax aqueous solution until the pH reached 9.0.
Apparatus and Operating ConditionsLC/Photodiode Array Detector (PDA)An L-6300 Intelligent pump (Hitachi, Ltd., Tokyo, Japan) system equipped with an MD-910 PDA (JASCO Corporation, Tokyo, Japan) was used. LC separation was performed using an Atlantis dC18 column (150 × 4.6 mm I.D., 5 µm; Waters Corporation). The column temperature was maintained at 40 °C. The mobile phase was a mixture of 50 mmol/L potassium phosphate buffer solution (pH 6.8) and acetonitrile at a ratio of 60 : 40 (v/v) and was delivered in the isocratic elution mode at a flow rate of 0.5 mL/min. The UV monitoring wavelength was 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) was used. LC separation was performed using an L-column2 ODS (150 × 2.1 mm I.D, 3.0 µm; CERI, Tokyo, Japan). The column temperature was maintained at 40 °C. The mobile phase was a mixture of 10 mmol/L formic acid aqueous solution and acetonitrile at a ratio of 40 : 60 (v/v), and was delivered in the isocratic elution mode at a flow rate of 0.2 mL/min. A 5 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; cone gas flow (N2): 50 L/h. Leucine-enkephalin solution (0.25 µL/mL) was delivered at a 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 using the full-width-at-half-maximum method.
NMRAn ECA-600II NMR spectrometer (JEOL, Tokyo, Japan) was used for the NMR analysis. The frequency for 1H-NMR measurements was set at 600.1723 MHz, and that for 13C-NMR measurements was set at 150.9134 MHz. As the reference peak, solvent peaks of methanol-d4 (1H: 3.31 ppm, 13C: 49.0 ppm) were used.
X-Ray Crystallography (XRC)An R-AXIS RAPID II (Rigaku, Tokyo, Japan) was used as the X-ray source and the operating conditions were as follows: X-rays were generated using Kα radiation (λ = 1.54187 Å) of Cu. The X-ray crystallographic data were collected at 93 K.
Degradation Behavior of Three BZPs with Hydroxy Group in Artificial Gastric JuiceDegradation ExperimentTo 900 µL of artificial gastric juice, 100 µL of each BZP standard solution (1000 µg/mL, LMZ, lorazepam, and oxazepam) was added. The solutions were then stored in an incubator set to 38 °C. Time-dependent degradation over a nine days period was determined by LC/PDA measurements. In the accelerated degradation test for LMZ, (confirmed to be degraded in the degradation experiment), LMZ was dissolved in the artificial gastric juice, similar to the above, and stored in different temperature conditions maintained at 50 and 70 °C in addition to 38 °C. The time course of LMZ degradation was determined by LC/PDA measurements.
Restoration Experiment of Three BZPs with Hydroxy GroupTo verify the restoration behavior of the three BZPs with hydroxy groups (LMZ, lorazepam, and oxazepam) from their degradation products by pH readjustment, three types of buffer solutions adjusted to pH 5, 7, and 9 were used. To 900 µL of each buffer solution, 100 µL of each BZP degradation solution, preliminarily prepared by adding 100 µL of BZP standard stock solution (1 mg/mL) to artificial gastric juice (pH 1.2) and storing it at 38 °C for 24 h, was added. Then, the solutions were stored again at room temperature (25 °C), and time-dependent changes over an approximately 24 h duration (0, 1, 2, and 24 h) were determined by LC/PDA measurements.
Isolation and Purification of LMZ Degradation ProductsThirty milligrams of LMZ was dissolved in 3 mL of methanol, and then artificial gastric juice was added to obtain 30 mL solution and was stored at 70 °C for one week. The resulting solution was transferred to a 200 mL eggplant flask, ethanol was added to bring the total volume to approximately 100 mL, and the solvent was azeotropically distilled off in an evaporator (40 °C). Purified water of 1 mL was added to the residue contained in the eggplant flask to dissolve the highly polar residue (dissolves easily in water), and the resulting solution was removed using a Pasteur pipette. This operation was repeated eight times, and all the resulting solutions were combined to form the degradation product (A) fraction. Then, 2 mL of methanol was added to the residue in the flask to redissolve all the residue, and the solution was transferred into a 5 mL microtube and dried by purging with nitrogen to isolate the yellow powder as the degradation product (B).
Structural Analysis of LMZ Degradation ProductsThe molecular weight and molecular formula of the LMZ degradation products (A & B) were deduced by LC/TOF-MS using a solution of LMZ stored in artificial gastric juice at 70 °C for 96 h.
Degradation product (B) (28 mg) was dissolved in 1 mL of methanol-d4, of which 600 µL was subjected to NMR measurements (1H-NMR, 13C-NMR, heteronuclear multiple-bond connectivity (HMBC), and 1H–1H correlation spectroscopy (COSY)) for structural analysis.
The methanol solution of degradation product (B) used in the NMR measurement was recrystallized by storing it in a desiccator at room temperature, and the crystal structure was confirmed by single crystal X-ray structure analysis.
The melting point of degradation product (B) (crystal) was measured using a homemade micro-melting point apparatus. Paraffin (melting point; 54–56 °C) and D-fructose (melting point; 102–104 °C) were used as calibration samples.
In our previous work14) where we investigated the behavior of eight BZPs in artificial gastric juice, we found that the degradability of the eight BZPs was markedly dependent on their structures. Among them, lorazepam, which has a hydroxyl group on the diazepine ring, showed only approximately 40% degradation even after 24 h at 38 °C. Therefore, in this study, to verify the degradability of the three BZPs possessing hydroxyl groups (LMZ, lorazepam, and oxazepam), their degradation behavior was observed in artificial gastric juice at 38 °C until almost complete degradation was achieved.
As expected, the BZPs immediately degraded to approximately 36, 20, and 13%, respectively, after two days, and all three BZPs degraded and nearly disappeared after one week (Fig. 2).
A restoration experiment was conducted to confirm whether the degraded three BZPs with hydroxy groups could be restored to their original state by changing the pH from slightly acidic (pH 5) to neutral (pH 7) or basic (pH 9) by monitoring the change in each time-course. No increase in the peak area of the parent compound was observed in either case (data not shown). In other words, no restoration of the parent compound was detected, suggesting that no intermediates were reversible to the parent compound, as in other BZPs.14)
Verification of LMZ Degradation Products in Artificial Gastric JuiceTo examine the behavior of LMZ degradation products in artificial gastric juice, we conducted an accelerated test by storing LMZ under high-temperature conditions (70 °C) for a prolonged time period (1.5 months). As shown in Fig. 3, two main degradation products were detected after two days: one with a faster retention time (greater polarity) than LMZ (degradation product (A)) and the other with a slower retention time (less polarity) than LMZ (degradation product (B)). The behavior of LMZ, degradation product (A), and degradation product (B) was analyzed, and the results showed that degradation product (A) gradually increased at first as LMZ degraded, and then the degradation product (B) gradually increased with a decrease in the degradation product (A). Based on these results, we inferred that degradation product (A) was an irreversible degradation intermediate, and degradation product (B) was the final degraded product (Fig. 4).
(A) Lormetazepam degradation product (A) (tR: 6.3 min) lormetazepam (tR: 7.4 min). (B) Lormetazepam degradation product (B) (tR: 25.6 min) (Mobile phase: 50 mmol/L phosphate buffer (pH 3)/acetonitrile = 60 : 40).
To verify the reactivity of LMZ degradation reaction in artificial gastric juice, three reaction temperatures (38, 50, and 70 °C) were set, and their respective time-course changes were observed (Supplementary Fig. S1). An Arrhenius plot (Supplementary Fig. S2) was drawn; the activation energy (Ea) and activation entropy (ΔS) were calculated with reference to a previous report, and the Gibbs free energy (ΔG) was calculated from these values. As a result, the Ea, ΔS, and ΔG of LMZ were 58.1 (kJ/mol), −160.7 (J/K-mol), and 103.8 (kJ/mol), respectively.
We previously calculated the physicochemical parameters of the degradation reactions in artificial gastric juice for other BZPs. Accordingly, LMZ had a higher ΔG value and was less reactive than nitrazepam (recalculated from the data13); ΔG = 98.98 kJ/mol), alprazolam (unpublished data; ΔG = 86.30 kJ/mol), triazolam (unpublished data; ΔG = 90.67 kJ/mol), and etizolam (unpublished data; ΔG = 97.43 kJ/mol). The measured degradability data13) for each BZP reported by the group were consistent with these ΔG values.
Therefore, the reason for LMZ to be less susceptible to degradation than other BZPs is discussed based on the structural formula as follows: LMZ has an electron-donating hydroxy group at the C3 position of the diazepine ring, which is believed to increase the electron density of the nitrogen atom at the N4 position and weaken the ability to draw an electron from the C5 position, thus exhibiting δ+ properties.
Isolation and Purification of LMZ Degradation ProductsWhen artificial gastric juice was added to LMZ to accelerate the degradation reaction at high temperatures, precipitates were formed in the solution. The precipitates were then separated, redissolved in methanol, and analyzed using LC/PDA. Because the main component was estimated to be the degradation product (B), the precipitates were used to isolate and purify the degradation product (B). The procedure was as follows: After the degradation reaction, the solvent, including the precipitate, was removed from the sample solution, and purified water was added to the residue and dissolved adequately. The aqueous solution was removed, and methanol was added to the residue to redissolve it. The resulting aqueous and methanol solutions were then analyzed using LC/PDA. As shown in Fig. 5, the results showed that degradation product (A) (Fig. 5-(I)) was mainly dissolved in the aqueous solution, while degradation product (B) (Fig. 5-(II)) was primarily dissolved in the methanol solution, suggesting that mutual separation of degradation products (A) and (B) is possible depending on the solvent used for redissolving. The degradation product (A) fraction dissolved in purified water was a clear solution, and a white powdery material was obtained when the solvent was removed. In contrast, the solution of the degradation product (B) fraction dissolved in methanol was yellow in color, and a yellow powdery substance was obtained upon removal of the solvent, with a yield of 15.5%. The isolation operation was repeated as required.
Although the degradation product (A) could be measured using LC/PDA or LC/TOF-MS, allowing immediate analysis of the solution, we found it difficult to isolate and purify because of its low stability in the solution. Therefore, the structural analysis described below focuses on degradation product (B), which could be isolated, and only speculation based on UV and MS spectral data is presented for degradation product (A).
Structural Analysis of LMZ Degradation ProductsWe performed UV spectroscopy, MS, and NMR spectroscopy (1H-NMR, 13C-NMR, COSY, and HMBC correlations) analysis.
The UV spectra of LMZ, degradation product (A), and degradation product (B) were measured using LC/PDA. LMZ exhibited UV λmax peaks at 211, 231, and 315 nm (Supplementary Fig. S3-(I)), whereas the degradation product (A) showed UV λmax peaks at 211 and 300 nm (Supplementary Fig. S3-(II)). A slight difference in the wavelength (λmax) of degradation product (A) was observed compared to that of LMZ. However, the spectral patterns of both LMZ and degradation product (A) were nearly the same. The basic skeleton of degradation product (A), that is, the resonance structure derived from the double bond, was similar to that of LMZ. On the other hand, the degradation product (B) showed UV λmax peaks at 235, 271, and 415 nm (Supplementary Fig. S3-(III)), and its spectral pattern was completely different from those of LMZ and degradation product (A).
From the high-resolution MS of LMZ (Supplementary Fig. S4-(I)), degradation product (A) (Supplementary Fig. S4-(II)), and degradation product (B) (Supplementary Fig. S4-(III)), measured using LC/TOF-MS, the protonated molecules ([M + H]+) m/z 335.0068, m/z 336.0011, and m/z 280.0054, respectively, were detected. The molecular formulae of degradation products (A) and (B) were hypothesized to be C16H13Cl2N2O2, and C14H11Cl2NO, respectively.
The NMR spectrum of the isolated degradation product (B) dissolving in 600 µL deuterated methanol (CD3OD) indicated the presence of a 1,2-disubstituted benzene ring [δ 7.50 (1H, dd, J = 7.5, 1.8 Hz), δ 7.47 (1H, td, J = 7.5, 1.8 Hz), δ 7.42 (1H, td, J = 7.5, 1.8 Hz), and δ 7.30 (1H, dd, J = 7.5, 1.8 Hz)], a 1,2,4-trisubstituted benzene ring [δ 7.36 (1H, dd, J = 9.0, 2.4 Hz), δ 6.99 (1H, d, J = 2.4 Hz)), and δ 6.83 (1H, d, J = 9.0 Hz)], and three protons [δ 2.98 (3H, s)] derived from SP3 carbon (Supplementary Table S1). The number of protons was predicted to be 11 based on the high-resolution MS. Therefore, we inferred that the one-deficient hydrogen was derived from an amino group.
From the 13C-NMR spectrum of the degradation product (B), 13 SP2 carbons, including a carbonyl carbon (δ 197.6 ppm) and an SP3 carbon (δ 29.6 ppm), were assigned as shown in Supplementary Table S1.
The COSY and HMBC correlation results are presented in Supplementary Table S1 and Fig. 6, respectively. Structural analyses confirmed that the irreversible degradation product (B) was methanone, [5-chloro-2-(methylamino)phenyl](2-chlorophenyl)-. The raw NMR spectra of the degradation product (B) are shown in the Supplementary Figs. S5–8.
Only the main correlations are indicated by arrows.
Degradation product (B) is a known substance (CAS Number 5621-86-3). The melting point of the obtained compound was measured to be 77–80 °C, which is consistent with the melting point of the substance in question (78–80 °C) described in the literature.15)
Single-Crystal X-Ray Structure AnalysisThe isolated degradation product (B) was redissolved in a small amount of warm methanol and stored in a desiccator at room temperature to obtain yellow tetragonal crystals. The structure obtained using single-crystal X-ray structural analysis was consistent with that inferred by NMR. Single-crystal X-ray structural analysis revealed that the degradation product (B) was axially malformed because the rotation was controlled by steric hindrance between the Cl group of the benzene ring (A) and the NHCH3 group of the benzene ring (B) (Fig. 7). The crystallographic data were deposited with the Cambridge Crystallographic Data Center as supplementary publication No. CCDC 2215360.
The pathway of the degradation reaction from LMZ to degradation product (B) is discussed bellow. We speculated that the oxonium ion in the artificial gastric juice (acidic solution) attacks the carbon atom of the azomethine bond (nucleophilic addition reaction), thereby cleaving the diazepine ring, followed by nucleophilic addition of the oxonium ion to the carbonyl carbon atom at position 2. Then, the α-hydroxyglycine consisting of the hydroxy group at position 3 and the amino group at position 4 is removed to form degradation product (B) (Supplementary Fig. S9).
Because the chemical formula of degradation product (A) was estimated to be C16H13Cl2N2O2 based on accurate mass measurement results, we checked for the existence of the compound corresponding to the estimated chemical formula against the degradation path diagram shown in Supplementary Fig. S9. The nitrogen atom at position 4 of LMZ was protonated to form a quaternary ammonium ion (Supplementary Fig. S9, upper middle) corresponding to this compound. However, because the degradation product (A) was difficult to isolate and purify, the structural formula could not be determined and was left as speculated.
We showed that benzodiazepines with a hydroxy group on the diazepine ring (LMZ, lorazepam, and oxazepam) were degraded in artificial gastric juice, similar to other BZPs. Moreover, once the degradation products were formed, there was no restoration to its corresponding parent compound, even after readjustment of the pH of the stored solution from slightly acidic to basic. Therefore, this study demonstrated that the reaction was irreversible. Furthermore, the LMZ degradation process produced degradation product (B) via an intermediate degradation product (A). We found that the degradation product (B) was the final degradation product of LMZ, and its isolation, and purification, and structural analysis was confirmed it to be methanone, [5-chloro-2-(methylamino)phenyl]-(2-chlorophenyl)-.
As the formation of these degradation products was irreversible, we considered it effective to target not only LMZ but also degradation product (B), the final degradation product, when analyzing LMZ for forensic drug identification in gastric juice.
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 Professor Tomoo Hosoe of Hoshi University for giving us advice on inference of the degradation reaction pathway of LMZ, and to Dr. Hiroyasu Sato of Rigaku Corporation for giving us advice on the single-crystal X-ray structure analysis.
The authors declare no conflict of interest.
This article contains supplementary materials.
