Studies on the metabolism and toxicological detection of glaucine, an isoquinoline alkaloid from Glaucium flavum (Papaveraceae), in rat urine using GC-MS, LC-MSⁿ and LC-high-resolution MSⁿ

Abstract

Glaucine ((S)-5,6,6a,7-tetrahydro-1,2,9,10-tetramethoxy-6-methyl-4H-dibenzo [de,g]quinoline) is an isoquinoline alkaloid and main component of Glaucium flavum (Papaveraceae). It was described to be consumed as recreational drug alone or in combination with other drugs. Besides this, glaucine is used as therapeutic drug in Bulgaria and other countries as cough suppressant. Currently, there are no data available concerning metabolism and toxicological analysis of glaucine. To study both, glaucine was orally administered to Wistar rats and urine was collected. For metabolism studies, work-up of urine samples consisted of protein precipitation or enzymatic cleavage followed by solid-phase extraction. Samples were afterwards measured by liquid chromatography (LC) coupled to low or high-resolution mass spectrometry (HR-MS). The phase I and II metabolites were identified by detailed interpretation of the corresponding fragmentations, which were further confirmed by determination of their elemental composition using HR-MS. From these data, the following metabolic pathways could be proposed: O-demethylation at position 2, 9 and 10, N-demethylation, hydroxylation, N-oxidation and combinations of them as well as glucuronidation and/or sulfation of the phenolic metabolites. For monitoring a glaucine intake in case of abuse or poisoning, the O- and N-demethylated metabolites were the main targets for the gas chromatography-MS and LC-MSⁿ screening approaches described by the authors. Both allowed confirming an intake of glaucine in rat urine after a dose of 2 mg/kg body mass corresponding to a common abuser's dose. Copyright © 2013 John Wiley & Sons, Ltd.

Introduction

Glaucine ((S)-5,6,6a,7-tetrahydro-1,2,9,10-tetramethoxy-6-methyl-4H-dibenzo [de,g]quinoline) is an isoquinoline alkaloid with aporphine structure and the main component of Glaucium flavum (yellow hornpoppy, Papaveraceae). Other natural resources of glaucine are Glaucium oxylobum, Croton lechleri and Corydalis yanhusuo.[1-3] Glaucium flavum is native in Western Europe, North America and Asia. The latex, typical for Papaveraceae, contains the main alkaloids in varying concentrations depending on environmental factors and species variations.[4] Other alkaloids of Glaucium flavum are, for example, protopine, chelerythrine and magnoflorine.[5]

Glaucine shows antitussive, bronchodilatory and anti-inflammatory properties and acts also as weak dopamine D₁ and D₂ receptor antagonist.[6, 7] The antitussive properties of glaucine are comparable to codeine without signs of opiate withdrawal after longer intake.[8] It is used as a therapeutic drug in Bulgaria and other countries as cough suppressant. However, case reports of recreational use of glaucine as party pills or so-called legal highs or side effects of therapeutic use described the following symptoms: feeling of hallucinations, tiredness, vomiting, dizziness and decreased blood pressure.[9-11] Such legal highs are not (yet) subjected to controlled substances legislation, but may show more or less strong psychoactive effects[12] and may occur in clinical and forensic toxicology cases. Therefore, drugs of abuse screening approaches should cover also such new drugs, but their metabolism must be studied first in order to know the target for the screening particularly in urine. Urine is the preferable screening matrix because in general the concentrations are much higher than in plasma.[13-15] As human samples after controlled drug intake are mostly not available, such studies should be performed using rats providing in most cases qualitatively the same metabolites as humans.[13-19]

So far, no data are available about the metabolic fate and the detectability of glaucine in body samples. Therefore, the aims of the presented study were the identification of the phase I and II metabolites using liquid chromatography-mass spectrometry (LC-MSⁿ) with a linear ion trap (LIT) or high-resolution Orbitrap (OT) and elucidation of the detectability in rat urine by the gas chromatography (GC)-MS[13] and LC-MSⁿ[14, 15] screening approaches described by the authors.

Experimental section

Chemicals and reagents

Glaucine (isolated from plant material and purified to 99%) was obtained from Oskar Tropitzsch (Marktredwitz, Germany), boldine from Sigma Aldrich (Steinheim, Germany), Escholtzia californica herba from WeberSeeds (Simpelveld, The Netherlands), Isolute Confirm HCX (130 mg, 3 ml) and Isolute Confirm C18 (500 mg, 3 ml) solid-phase extraction (SPE) cartridges from Biotage (Grenzach-Wyhlen, Germany), N-nitroso-N-methylurea from Jenachem (Jena, Germany), acetonitrile (LC-MS grade), ammonium formate (analytical grade), formic acid (for mass spectrometry), methanol (LC-MS grade), mixture (100 000 Fishman units/ml) of glucuronidase (EC No. 3.2.1.31) and arylsulfatase (EC No. 3.1.6.1) from Helix Pomatia, and all other chemicals and biochemicals from VWR (Darmstadt, Germany).

Synthesis of 2-O-demethyl glaucine

For synthesis of 2-O-demethyl (DM) glaucine (predicentrine), boldine was dissolved in methanol, and methylation reagent was added and stirred by room temperature for 2 h. For synthesis of the methylation reagent, diethyl ether and potassium hydroxide (40%) were added to N-nitroso-N-methylurea and stored for 30 min by 4°C. Afterwards, organic layer was removed and dried over plates of potassium hydroxide.

Preparation of 9-O-DM glaucine

For preparation of 9-O-DM glaucine (lauroscholtzine), alkaloids were extracted from Escholtzia californica herba. An aliquot of plant material was suspended in water acidified to pH 1 with hydrochloric acid and non-alkaloid components such as fatty acids were extracted with n-hexane. After alkalization using ammonium hydroxide to pH 9, alkaloids were extracted threefold using chloroform. Combined chloroform phases were dried over sodium sulfate.

Urine samples

All studies were performed using urine of two male Wistar rats each (Ch. River, Sulzfleck, Germany). They were administered a single 1, 2 or 20 mg/kg body mass (BM) dose in aqueous suspension by gastric intubation of glaucine for toxicological diagnostic reasons according to the corresponding German law. Urine (about 50 ml) was collected separately from the faces using a metabolism cage over a 24 h period. Blank urine samples were collected before drug administration to check whether the samples were free of interfering compounds. All urines were directly analyzed or stored at −20°C until further analysis.

Sample preparation of urine samples for metabolism studies using LC-MSⁿ and LC-high-resolution (HR)-MSⁿ

For identification of phase I metabolites, urine samples after high dose (20 mg/kg BM) were processed as already described elsewhere by cleavage of conjugates and SPE (HCX).[19] Briefly, in case of conjugate cleavage, a 1.0 ml portion of urine was worked-up after enzymatic cleavage by SPE (HCX). Urine was loaded on an Isolute Confirm HCX cartridges previously conditioned with methanol and water. After passage of sample, cartridge was washed with 0.1 M hydrochloric acid and purified water. Elution was performed with a mixture of methanol and ammonia (98:2 v/v). The extract was gently evaporated at room temperature using nitrogen stream to dryness and reconstituted in mobile phase (A/B 50:50 v/v).

For identification of phase II metabolites, urine samples after high dose (20 mg/kg BM) were processed by SPE (C18). Details can be found elsewhere.[20] Briefly, a 1.0 ml portion of urine was diluted with 2.0 ml purified water before loading the sample on a previously conditioned (1.0 ml methanol, 1.0 ml water) Isolute Confirm C18 cartridge. After passage of sample, cartridge was washed with 2.0 ml of purified water. Elution was performed using 1.5 ml of methanol and 0.2 ml acetone. The extract was gently evaporated at room temperature using nitrogen stream to dryness and reconstituted in mobile phase (A/B 50:50 v/v).

LC-MSⁿ analysis

All samples were separated and analyzed using an Accela LC system (ThermoFisher Scientific, TF, Dreieich, Germany) consisting of a degasser, a quaternary pump and an autosampler coupled to the TF LXQ LIT equipped with a heated electrospray ionization source II (HESI). Details were already described elsewhere.[14] Briefly, the used column was a TF Hypersil Gold (10 × 2.1 mm, 1.9 µm) guarded by TF Hypersil GOLD C18 Drop-in guard cartridge and a TF Javelin column filter using 10 mM aqueous ammonium formate plus 0.1% formic acid pH 3.4 (eluent A) and acetonitrile plus 0.1% formic acid (eluent B). The flow rate was set to 0.5 ml/min, and the gradient was programmed as follows: 0–1.0 min 98% A, 1.0–3.0 min to 90% A, 3.0–5.0 min to 85% A, 5.0–7.5 min to 80% A, 7.5–10.0 min to 75% A, 10.0–11.5 min to 70% A, 11.5–13.0 min to 65% A, 13.0–14.5 min to 50% A, 14.5–16.0 min to 40% A, 16.0–19.0 min to 0% A, 19.0–21.0 hold 0% A. Cleaning of the injection system, column flushing and re-equilibrating are described elsewhere. Thus, the total run time per sample was 25 min. The injection volume for all samples was 10 µl each.

The MS conditions were as follows: positive ionization mode; sheath gas, nitrogen at flow rate of 34 arbitrary units (AU); auxiliary gas, nitrogen at flow rate of 11 AU; vaporizer temperature, 250°C; source voltage, 3.00 kV; ion transfer capillary temperature, 300°C; capillary voltage, 31 V; tube lens voltage, 80 V. Automatic gain control was set to 15 000 ions for full scan and 5000 ions for MSⁿ. The maximum injection time for full scan (MS¹ stage) was set to 100 ms.

Collision-induced dissociation (CID)-MSⁿ experiments were performed on the following selected precursor ions from MS¹ at m/z (u) 548, 534, 520, 518, 504, 452, 438, 424, 422, 408, 372, 358, 356, 344, 342 and 328. MS¹ was performed in the full scan mode, m/z 100–800 u. Normalized wideband collision energies were 35.0% for MS² and 40.0% for MS³. In order to obtain more structure information, samples were also measured without wideband activation additionally. Other settings were as follows for MS²: minimum signal threshold, 100 counts; isolation width, 1.5 u; for MS³: minimum signal threshold, 50 counts; isolation width, 2.0 u; for both stages: activation Q, 0.25; activation time, 30 ms; dynamic exclusion mode: repeat counts, 2; repeat duration, 15 s; exclusion list, 50; exclusion duration, 15 s, average full scan to full-scan cycle time, 4 s. TF Xcalibur 2.1.0 software was used for data acquisition.

LC-HR-MSⁿ analysis

All samples were separated and analyzed using a TF Dionex LC system consisting of a degasser, a quaternary pump and an auto sampler coupled to a TF LTQ Velos Pro OT with HESI. The LC column was a TF Hypersil Gold (15 × 2.1 mm, 1.9 µm), and gradient were the same as for LC-MSⁿ. The MS conditions were as follows: ESI, positive mode; sheath nitrogen gas flow rate of 40 AU; auxiliary gas, 20 AU; source voltage, 4 kV; source heater temperature, 400°C; ion transfer capillary temperature, 300°C; capillary voltage, 4 V; CID-MS/MS experiments were performed on the following selected precursor ions from MS¹ at m/z (u) 548, 534, 520, 518, 504, 452, 438, 424, 422, 408, 372, 358, 356, 344, 342 and 328. MS¹ was performed in the full scan mode, m/z 100–800 u. Other settings were as follows: normalized collision energies, 35%; minimum signal threshold: 100 counts; with a resolution of 30 000 at m/z 400 u; isolation width, 1.5 u; activation Q, 0.25; activation time, 30 ms; dynamic exclusion mode, repeat counts 2, repeat duration 15 s, exclusion duration 15 s. Mass calibration were realized with manufacturer calibration mixture. In order to obtain more structure information, samples were also measured without wideband activation additionally. TF Xcalibur 2.1.0 software was used for data acquisition including extraction of the exact masses with a tolerance of 5 ppm.

GC-MS standard urine screening approach

After administration of 2 mg/kg BM to rats, urine samples were worked-up according to published procedures.[21, 22] Briefly, the samples (5 ml) were divided into 2 aliquots, and one part was submitted to acid hydrolysis. Thereafter, the sample was adjusted to pH 8–9, and the other aliquot of untreated urine was added. This mixture was extracted with a dichloromethane-isopropanol-ethyl acetate mixture (1:1:3 v/v/v), and the organic layer was evaporated to dryness. The residue was acetylated with an acetic anhydride-pyridine mixture under microwave irradiation. After evaporation of the derivatization mixture, the residue was dissolved in 100 µl of methanol, and 2 µl was injected into an HP 5890 Series II gas chromatograph combined with a HP 5972A MSD mass spectrometer. The GC conditions were the same as for the metabolism studies except for temperature, which was programmed from 100 to 310°C at 30°/min. The MS conditions were as follows: full-scan mode, m/z 50–550 u; EI mode, ionization energy, 70 eV; ion source temperature, 220°C; capillary direct interface, heated at 280°C.

For toxicological detection of the acetylated metabolites, mass chromatography was used with the extracted ions at m/z (u) 354 for glaucine, 382 for acetylated DM glaucine and 411 for acetylated bis-DM glaucine (representing spectra in Fig. 16). Generation of the mass chromatograms could be started by clicking the corresponding pull down menu which executes the user-defined macros.[21] The identity of the peaks in the mass chromatograms was confirmed by computerized comparison of the mass spectra underlying the peaks (after background subtraction) with reference spectra recorded during this study.[13]

LC-MSⁿ standard urine screening approach

After administration of 2 mg/kg BM to rats, 100 µl of urine was precipitated by acetonitrile as already described.[14] After shaking and centrifugation, the supernatant was gently evaporated to dryness and reconstituted in mobile phase (A/B 50:50 v/v). The worked-up samples were separated and analyzed using the same LC-MSⁿ system as described above. For screening approach, the CID-MSⁿ experiments were modified by using data-dependent acquisition (DDA) on precursor ions selected from MS¹: MS¹ was performed in the full scan mode (m/z 100–800 u). MS² and MS³ were performed in the DDA mode: four DDA MS² scan filters were chosen to provide MS² on the four most intense signals from MS¹, and additionally, eight MS³ scan filters were chosen to record MS³ on the most and second most intense signals from the MS². MS² spectra were collected with a higher priority than MS³ spectra. Normalized wideband collision energies were 35.0% for MS² and 40.0% for MS³.

TF Xcalibur 2.1.0 software was used for data acquisition, NIST MS Search 2.0 (National Institute of Standards and Technology, Gaithersburg, MD) for library generation, TF ToxID 2.1.1 for automatic target screening in the MS² screening mode. The settings were as follows: retention time (RT) window, 20 min; RT, 0.1 min; signal threshold, 100 counts; search index, 600; reverse search index, 700. SmileMS version 1.1 (GeneBio, Geneva, Switzerland) was used for automatic target screening using the precursor tolerance option and for automatic untargeted screening without precursor tolerance option and RT locking. Further settings were as follows: score threshold, 0.1; minimum number peak matches, 0. ToxID and SmileMS run automatically after file acquisition using an Xcalibur processing method starting both software tools.

Results and discussion

Identification of phase I metabolites

Phase I metabolites of glaucine were identified in rat urine using LC-MSⁿ by interpretation of their fragmentation patterns. The elemental composition of the fragments was determined using LC-HR-MSⁿ, which formed the same fragments. The use of wideband activation was a helpful tool in LC-MSⁿ analysis: cleavage of H₂O or other slightly eliminated groups such as NH₃ lead to more or less uncharacteristic MSⁿ spectra consisting only of one ion. These uncharacteristic ions underwent further fragmentation with help of wideband activation resulting in more characteristic MSⁿ spectra.[23] Accordingly, these fragment ions were less abundant or even not present in the resulting MSⁿ spectra. However, spectra recorded without wideband activation could be a useful tool for structure elucidation. For example, the fragment ions at m/z 355 or 354 u representing H₂O loss could not be detected in the MS² spectra of glaucine-N-oxide with wideband activation, but without wideband activation. This indicated that the hydroxy (HO-) moiety was not located in the aromatic systems, as a typical loss could be observed for the MS² spectrum without wideband activation. Accordingly, both dissociation strategies were used. The fragmentation is discussed only for the most important MSⁿ spectra.

Fragmentation of DM metabolites of glaucine

Three O-demethylated, one N-demethylated, three O-bis-demethylated and three O,N-bis-demethylated metabolites could be detected. The corresponding MSⁿ spectra and fragmentation are depicted in Figs. 1-10. Opening of the isoquinoline ring was the initial reaction of glaucine and all demethylated metabolites followed by elimination of the amino group.[24] This lead to fragment ion at m/z 325 u in the MS² spectra of glaucine (protonated molecule, PM, at m/z 356 u) and its N-demethylated metabolite (PM at m/z 342 u) after neutral loss of H₂NCH₃ and NH₃, respectively (Figs. 1 and 2). This was of importance for identification of O,N-bis-demethylated metabolites (PM at m/z 328 u). Due to neutral loss of NH₃ from the PM at m/z 328 u, corresponding MS spectra of fragment ions at m/z 311 u could be compared with corresponding MS spectra from PM at m/z 342 u of O-demethylated fragment ions at m/z 311 u (Figs. 3, 5, 7 and 9). Discrimination between N- and O-demethylated metabolites was possible looking at the loss of NH₃ and H₂NCH₃, respectively (Figs. 1-10).

Figure 1. LC-MSⁿ spectra of glaucine (a–d) and N-DM-glaucine (e–h). All spectra were measured with wideband activation.

Figure 2. Proposed fragmentation pattern of glaucine and N-DM-glaucine with exact masses, elemental compositions and error values.

Figure 3. LC-MSⁿ spectra of 2-O-DM-glaucine. All spectra were measured with wideband activation.

Figure 4. Proposed fragmentation pattern of 2-O-DM-glaucine with exact masses, elemental compositions and error values.

Figure 5. LC-MSⁿ spectra of 9-O-DM-glaucine. All spectra were measured with wideband activation.

Figure 6. Proposed fragmentation pattern of 9-O-DM-glaucine with exact masses, elemental compositions and error values.

Figure 7. LC-MSⁿ spectra of 10-O-DM-glaucine. All spectra were measured with wideband activation.

Figure 8. LC-MSⁿ spectra of 2,9-O,O-bis-DM-glaucine (boldine; a–c), O,O-bis-DM-glaucine isomer 1 (d–f), O,O-bis-DM-glaucine isomer 2 (g–i). All spectra were measured with wideband activation.

Figure 9. LC-MSⁿ spectra of 2-O,N-bis-DM-glaucine (a–c), 9-O,N-bis-DM-glaucine (d–f), 10-O,N -bis-DM-glaucine (g–i). All spectra were measured with wideband activation.

Figure 10. Proposed fragmentation pattern of O,O-bis-DM- and O,N-bis-DM-glaucine with exact masses, elemental compositions and error values.

A further important fragmentation for structure elucidation was the elimination of the methoxy groups.[24] Two vicinal methoxy groups lead to a radical loss of CH₃ or CH₃O (15 and 31 u, respectively) as observed for glaucine and its N-DM metabolite (Figs. 1 and 2). If a methoxy group was standing vicinal to a hydroxyl group, for example in 2,9-O-bis-DM-glaucine (Fig. 8 a–c), elimination of neutral CH₃OH followed by loss of CO was a preferred fragmentation pathway (loss of 32 and 28 u, respectively). Since O-demethylated metabolites contained two vicinal methoxy groups and a hydroxyl group vicinal to a methoxy group, both fragmentation types must be observed. Based on this fragmentation scheme, three of four possible O-demethylated metabolites were detected. As depicted in Figs. 5 and 7, two metabolites showed very similar fragmentation patterns, but they were different from the third isomer, which showed a stronger elimination of CH₃OH followed by a loss of CO (Fig. 3). A possible explanation for this might be the position of the vicinal methoxy group in the aporphine structure. In order to elucidate, which of the four methoxy groups were demethylated, the following, commercially not available, reference standards were used: predicentrine has the same structure as 2-O-DM-glaucine and lauroscholtzine that of 9-O-DM-glaucine. Predicentrine was synthesized by selective methylation of boldine at position 9 as described by Asencio et al.[25] Lauroscholtzine, an alkaloid of Escholtzia californica,[26] was extracted from plant material with chloroform, which was unfortunately the only solvent providing sufficient extraction. The structures of predicentrine and lauroscholtzine were confirmed by HR-MS and NMR (details will be published together with the enzyme kinetics elsewhere).

As the RTs and fragmentation patterns of the two isomers differed and were identical to the corresponding synthesized reference standards, it could be concluded that 2-O-DM-glaucine and 9-O-DM-glaucine were formed. The third O-DM metabolites is most likely the 10-O-DM because its MSⁿ spectra showed the same fragmentation as the 9-O-DM-glaucine, but different from the 2-O-DM-glaucine.

For identification of the bis-demethylated isomers, neutral loss of H₂NCH₃ or NH₃ allowed differentiation of O-bis- and O,N-bis-demethylated isomers, respectively. After loss of NH₃ from PM at m/z 328 u, corresponding MS spectra could be compared with MS spectra from PM at m/z 342 u of demethylated isomers and configuration could be elucidate (Figs. 8-10). Among all O-bis-demethylated metabolites, only the configuration of 2,9-O,O-bis-DM-glaucine could be determined. This was possible by comparison of the RT and MS spectra of different MS stages with the reference substance boldine (Fig. 8).

Fragmentation of glaucine-N-oxide and demethylated glaucine-N-oxide metabolites

Fragmentation of glaucine-N-oxide (Fig. 11) differed in the neutral loss of the amino group from glaucine and its demethylated metabolites. Three alternative elimination reactions were observed. Due to the quaternary N-oxide, neutral loss of H₂O from PM was detected resulting in fragment ion at m/z 354 u (Fig. 11 b). This fragment showed a radical loss of CH₃ and OCH₃, leading to fragment at m/z 323 and 339 u, respectively, but no loss of CH₃NHOH (47 u; Fig. 11 e) was observed. In conclusion, after neutral loss of H₂O from the N-oxide, the remaining fragment ion could no more eliminate ammonia, most probably due to the missing free electron pair of nitrogen. Parallel to the even electron loss of H₂O, the radical loss of OH (17 u) to fragment ion at m/z 355 u was observed (Fig. 11 b) and confirmed by accurate mono-isotopic mass (M_mi 355.1784 u; C₂₁H₂₅NO₄). In contrast to fragment ion at m/z 354 u, fragment ion at m/z 355 u eliminated the amino group via loss of 43 u to fragment ion at m/z 312 u. MS³ spectra of fragment ion at m/z 312 u (with ions at m/z 269, 254 and 237 u; Fig. 11 c) showed similar fragment ions as the MS³ spectra at m/z 310 u from glaucine or N-DM-glaucine (with ions at m/z 267, 252 and 235 u; Fig. 1 d, g) shifted with two mass units. That could be explained by missing the double bond between carbon atom C6a and C7 in the aporphine structure (Fig. 17). The third loss of an amino group was observed for the PM at m/z 372 u reacting to the fragment at m/z 325 u via elimination of CH₃NHOH (47 u; Fig. 11 a, b). This fragment, which represents the most intensive ion in MS² spectra of PM, showed the same accurate mass (M_mi 325.1434 u; C₂₀H₂₁O₄) and also corresponding MS³ and MS⁴ spectra of fragment ion at m/z 294 u (Fig. 11 c, f, g). Hence, same structural properties could be assumed.

Figure 11. LC-MSⁿ spectra of glaucine-N-oxide (b measured without wideband activation).

In analogy, these fragmentation patterns were observed for the N-oxides of the three demethylated isomers (Fig. 12). Confirmation of these isomers could be done by comparison with the MS spectra, especially MS³, of fragment ions at m/z 311 u from PM at m/z 358 u (after loss of CH₃NHOH with 47 u) and RTs, with these from the demethylated isomers (PM at m/z 342 u, Figs. 12 b, e, h; 3 b, 5 d and 7d). For differentiation between 9- and 10-O-DM-glaucine, MS spectra of higher stages were compared.

Figure 12. LC-MSⁿ spectra of 2-O-DM-glaucine-N-oxide (a–c), 10-O-DM-glaucine-N-oxide (d–f), 9-O-DM-glaucine-N-oxide (g–i). All spectra were measured without wideband activation.

Radical loss of OH (17 u) to fragment ion at m/z 341 u was also observed but not depicted in Fig. 12, because only spectra measured with wideband activation are shown. In analogy to glaucine-N-oxide, fragment ion at m/z 341 u showed loss of 43 u to fragment ion at m/z 298 u. Using wideband activation, these two-step elimination took place in one MS stage. Therefore, the use of wideband activation is called pseudo MS³.[23] MS³ spectra of fragment ion at m/z 298 u (Fig. 12 c, f, i) are similar to MS³ spectra of fragment ion at m/z 312 u from glaucine-N-oxide (Fig. 11 c) but shifted by 14 mass units.

Fragmentation of hydroxy metabolites of glaucine

Two types of hydroxylated metabolites of glaucine were detected (Fig. 13). The spectrum of the alkyl hydroxy metabolites should be differentiated from that of the aromatic hydroxy metabolite by neutral loss of H₂O.[27] Therefore, spectra measured with and without use of wideband activation were compared. The most abundant ion at m/z 354 u in the MS² spectra of isomer 2 showed a loss of 18 u, which was not observed in isomer 1 (Fig. 13 a, e). Instead, there was a loss of 31 u to the most abundant fragment ion at m/z 341 u. Differentiation between eliminating neutral H₂NCH₃ or radical OCH₃ (both 31 u) was possible by comparing accurate masses of corresponding fragment ions. Neutral loss of H₂NCH₃ (31 u) in isomer 2 could only be observed after the primary loss of H₂O in the MS³ spectra of ion at m/z 354 u to ion at m/z 323 u (M_mi 323.1285 u, C₂₀H₁₉O₄). In conclusion, isomer 2 should be hydroxylated in position C4 or C5 (numbering given in Fig. 17). Position C6a or C7 could be excluded, because after the loss of H₂O, elimination of H₂NCH₃ was observed. This could only occur if these carbons were saturated. As the exact position of the hydroxy groups could not be elucidated, they are drawn with an undefined stereo bond (tilde) in the corresponding figures. Isomer 1 should be hydroxylated in position C3, C8 or C11 because they are single free aromatic positions and no water elimination could be observed. In addition, MS^2–5 indicated analogue fragmentation patterns as for the isomers, which contained a hydroxy group next to the methoxy group (Fig. 3). For those, a preferred loss of CH₃OH (MS³ from fragment ion at m/z 341 to 309 u) followed by loss of CO (MS⁴ from fragment ion at m/z 309 to 281 u) was observed (Fig. 13 b, c). Due to the second vicinal methoxy group, fragmentation of ion at m/z 281 u in MS⁵ showed radical losses of CH₃ and OCH₃ (loss of 15 and 31 u, Fig. 13 d) in analogy to fragment ion at m/z 325 u from glaucine or N-DM-glaucine (Fig. 1 b, f).

Figure 13. LC-MSⁿ spectra of HO-(aryl)-glaucine (a–d), HO-(alkyl)-glaucine (e–h). All spectra were measured without wideband activation.

Fragmentation of hydroxy-O-DM metabolites of glaucine

Five isomers of hydroxy-O-DM metabolites were detected (Fig. 14 a–o). Four of them showed a neutral loss of H₂O (18 u) from the PM at m/z 358 u to fragment ion at m/z 340 u indicating alkyl hydroxylation. Isomer 4 (Fig. 14 j–l) was the only metabolite, which showed no neutral loss of H₂O (18 u) to fragment ion at m/z 340 u from PM. In conclusion, isomer 4 was hydroxylated in aryl position. After loss of water, the resulting fragment ion at m/z 340 u eliminates H₂NCH₃ (31 u) from MS³. Differentiation between eliminating neutral H₂NCH₃ or radical OCH₃ (both 31 u) was possible by comparing accurate masses of corresponding fragment ions. The accurate mass of fragment ions at m/z 309 u (M_mi: 309.1118 u; C₁₉H₁₇O₄) indicated absence of nitrogen and verifies the elimination of H₂NCH₃. Characterization of configuration of the demethylated position followed same rules as for demethylated isomers. Absence of neutral loss of 17 u (NH₃) indicated that all hydroxy-DM metabolites were formed by O-demethylation. In case of isomers 1, 2 and 4, MS³ spectra of ion at m/z 327 u showed loss of CH₃OH (32 u) to fragment ion at m/z 295 u and radical loss of CH₃ (15 u) and in isomer 5, radical loss of OCH₃ (31 u) to fragment ion at m/z 296 u predominated. The ions at m/z 295 and 296 u represented the hydroxylated fragment ions at m/z 279 and 280 u from MS³ of fragment ion at m/z 311 u from demethylated isomers (Figs. 3 a, b, 5 a, d and 7 a, d), so that isomers 1, 2 and 4 should be demethylated in position C1 or C2 and isomer 5 in position C9 or C10.

Figure 14. LC-MSⁿ spectra of HO-DM-glaucine isomers 1 (a–c), 2 (d–f), 3 (g–i), 4 (j–l) and 5 (m–o). All spectra were measured without wideband activation.

Fragmentation of hydroxy-bis-DM metabolites of glaucine

Five HO-bis-DM isomers could be detected (Fig. 15). Isomers 1, 2 and 3 showed a neutral loss of H₂O (18 u) from PM at m/z 344 u leading to the assumption that hydroxylation took place in alkyl position (Fig. 15 a–i). In contrast, isomers 4 and 5 were aromatically hydroxylated, and no loss of H₂O from PM could be observed (Fig. 15 j–o). Differentiation between O-bis- and O,N-bis-demethylated isomers was possible due to neutral loss of NH₃ (17 u) or H₂NCH₃ (31 u) from PM. For alkyl hydroxylated metabolites, MS³ spectra of ion at m/z 326 u (to m/z 295 u) of isomers 1 and 2 indicated bis-O-demethylated metabolites (loss of H₂NCH₃, 31 u). In the case of isomer 3 (to m/z 309 u), MS³ or MS² indicated a O,N-bis-metabolite (loss of NH₃, 17 u). For determination of demethylated positions in the aromatically hydroxylated isomers, MS² spectra of PM were investigated. Isomer 4 (Fig. 15 j–l) indicated a O-bis-demethylated (by loss of H₂NCH₃, 31 u to fragment ion at m/z 313 u) and isomer 5 (Fig. 15 m–o) a O,N-bis-demethylated (by loss of NH₃, 17 u to fragment ion at m/z 327 u) metabolite. MSⁿ spectra at higher MS stages showed typical fragmentation patterns after radical loss of CH₃ and OCH₃ or neutral loss of CH₃OH and CO.

Figure 15. LC-MSⁿ spectra of HO-bis-DM-glaucine isomers 1 (a–c), 2 (d–f) and 3 (g–i) and of HO-bis-DM-glaucine isomers 4 (j–l) and 5 (m–o). All spectra were measured without wideband activation.

Identification of phase II metabolites

In contrast to the phase I metabolites, the phase II metabolites were isolated using C18 cartridges because the mixed-mode SPE (HCX) could not retain sufficiently the acidic glucuronides. Twenty-one phase II metabolites were detected including six sulfates and 15 glucuronides of the hydroxylated, mono- and bis-demethylated phase I metabolites. Their exact masses and accurate masses measured by OT, the proposed elemental compositions and the RTs are given in Table 1. The phase II metabolites were confirmed by comparison of their MSⁿ with the MSⁿ⁻¹ of the corresponding phase I metabolites as already described for Kratom metabolites.[28]

Table 1. List of phase II metabolites identified in rat urine, their retention times (LIT), accurate (OT) and exact masses, mass errors and proposed elemental compositions

Phase II metabolites	Retention time, min	Accurate masses, u	Exact masses, u	Errors, ppm	Proposed elemental compositions
2-O-DM-Glaucine-sulfate	7.0	422.1270	422.1267	0.57	C20H24NO7S
9-O-DM-Glaucine-sulfate	8.4	422.1269	422.1267	0.23	C20H24NO7S
2-O-DM-Glaucine-glucuronide	5.8	518.2018	518.2020	0.44	C26H32NO10
9-O-DM-Glaucine-glucuronide	6.8	518.2020	518.2020	0.19	C26H32NO10
10-O-DM-Glaucine-glucuronide	4.5	518.2019	518.2020	0.21	C26H32NO10
O,O-Bis-DM-Glaucine-sulfate isomer 1	5.7	408.1111	408.1111	1.51	C19H22NO7S
O,O-Bis-DM-Glaucine-sulfate isomer 2	6.8	408.1113	408.1111	0.93	C19H22NO7S
O,N-Bis-DM-Glaucine-sulfate isomer 3	8.0	408.1114	408.1111	0.97	C19H22NO7S
O,O-Bis-DM-Glaucine-glucuronide isomer 1	2.3	504.1863	504.1864	0.05	C25H30NO10
O,O-Bis-DM-Glaucine-glucuronide isomer 2	3.4	504.1863	504.1864	0.01	C25H30NO10
O,O-Bis-DM-Glaucine-glucuronide isomer 3	4.4	504.1865	504.1864	0.35	C25H30NO10
O,O-Bis-DM-Glaucine-glucuronide isomer 4	4.8	504.1862	504.1864	0.38	C25H30NO10
O,O-Bis-DM-Glaucine-glucuronide isomer 5	5.9	504.1864	504.1864	0.01	C25H30NO10
O,N-Bis-DM-Glaucine-glucuronide isomer 6	7.2	504.1864	504.1864	0.26	C25H30NO10
HO-O-DM-Glaucine-sulfate	4.6	438.1219	438.1216	0.33	C20H24NO8S
HO-O-DM-Glaucine-glucuronide isomer 1	5.5	534.1972	534.1969	0.18	C26H32NO11
HO-O-DM-Glaucine-glucuronide isomer 2	5.9	534.1973	534.1969	0.43	C26H32NO11
HO-O,O-Bis-DM-Glaucine-glucuronide isomer 1	4.7	520.1812	520.1813	0.26	C25H30NO11
HO-O,O-Bis-DM-Glaucine-glucuronide isomer 2	5.9	520.1814	520.1813	0.20	C25H30NO11
HO-O,O-Bis-DM-Glaucine-glucuronide isomer 3	7.0	520.1817	520.1813	0.63	C25H30NO11
HO-O,O-Bis-DM-Glaucine-glucuronide isomer 4	7.6	520.1838	520.1813	4.69	C25H30NO11

Proposed metabolic pathway

Based on the identified metabolites, the following metabolic steps of glaucine could be proposed (Fig. 17): O-demethylation at position 2, 9 and 10, N-demethylation, hydroxylation, N-oxidation and combinations of them well as glucuronidation and/or sulfation of the phenolic metabolites.

Toxicological detection by GC-MS or LC-MSⁿ

Using the GC-MS-based systematic toxicological analysis approach described by the authors,[21, 22] an intake of glaucine could be monitored in rat urine after administration of 2 mg/kg BM. This dose corresponded to a 40 mg human single cough medication dose scaled by dose-by-factor approach according to ref.[29] As legal high preparations contained up to 200 mg per dose,[11] a recreational intake should be successfully monitored also in human urine. Toxicological detection within the GC-MS screening approach should be focused on the acetylated mono- and bis-DM metabolites (Fig. 16 b, c) being the most abundant metabolites in the rat urine samples. The spectrum of unchanged glaucine was added for its identification in gastric content or seized samples. For detection of lower doses, SPE (HCX) instead of the common liquid-liquid extraction should be performed.

Figure 16. GC-MS spectra of parent compound (a) and acetylated DM-glaucine (b) and bis-DM-glaucine (c) metabolites for GC-MS screening.

Using the LC-MSⁿ screening approach described by the authors,[14, 15] an intake of glaucine could also be monitored in rat urine after administration of 2 mg/kg BM. The main targets were the demethylated metabolites and their glucuronides (Fig. 17).

Figure 17. Proposed metabolic pathway of glaucine in rat.

Conclusion

The presented study has shown that glaucine was extensively metabolized by O- and N-demethylation as well as N-oxidation and hydroxylation. Twenty-six phase I and 21 phase II metabolites could be identified. Assuming similar metabolism and kinetics of rats and humans, an intake of glaucine should be detectable via its metabolites in human urine by GC-MS and/or LC-MSⁿ.

Acknowledgements

The authors like to thank their colleagues, Julia Dinger, Andreas Helfer, Daniela Remane, Carsten Schröder, Andrea Schwaninger, Gabriele Ulrich, Armin Weber, Jessica Welter, Carina Wink and Josef Zapp for their support.