Forensic Science International 200 (2010) 141–147

Contents lists available at ScienceDirect

Forensic Science International
journal homepage:

Detection of JWH-018 metabolites in smoking mixture post-administration urine
Tim Sobolevsky *, Ilya Prasolov, Grigory Rodchenkov
Moscow Antidoping Centre, Elizavetinsky per. 10, 105005 Moscow, Russia



Article history: Received 25 January 2010 Received in revised form 30 March 2010 Accepted 2 April 2010 Available online 28 April 2010 Keywords: JWH-018 Synthetic cannabinoid Metabolism GC–MS LC–MS

Smoking mixtures containing the cannabimimetic indoles may still be available over-the-counter in several countries. Due to the high affinity of these compounds to the cannabinoid receptors, their effective dose is lower than that of the marijuana products resulting in a low concentration of the excreted metabolites accompanied by a higher psychoactive potency. Up to now the in vivo metabolism of the cannabimimetic indoles seems to be insufficiently investigated and no data have been published on an assay of JWH-018 in urine. In this publication the urinary metabolites of JWH-018 are reported. Using gas and liquid chromatography combined with tandem mass spectrometry two main monohydroxylated metabolites were identified in the forensic urine samples. Based on the differences in their electron ionization MS/MS spectra it is supposed that one is formed by hydroxylation of the indole ring whilst the other by hydroxylation of the N-alkyl chain. The main metabolites are almost completely glucuroconjugated, whereas minor ones (N-despentyl hydroxy-, carboxy-, dihydroxy-, and reduced di- and trihydroxy metabolites) were also present in the free fraction. The parent compound was not detected in urine. ß 2010 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Recently, aminoalkylindoles have been found to act as the agonists of cannabinoid CB1 and CB2 receptors [1,2]. Due to their high receptor affinity and existing potential to separate therapeutic action from the psychotropic side effects, these compounds are screened as pharmaceutical candidates for the treatment of various diseases [3–5]. On the other hand, during the last two years the synthetic cannabinoids, mostly JWH-018, were identified as the components of smoking mixtures like ‘‘Spice’’ and its clones, openly sold as an incense [6]. They are unofficially advertised as a semi-legal and more powerful substitute for the marijuana products. One pack usually contains 3 g of a so-called herbal composition with approximately 0.5–1.5% of active component impregnated on it [7,8] that corresponds to ca. 30 mg of the active component per pack, although a large variation is expected. Following the prohibition of JWH-018 in several countries during 2009, the smoking mixtures containing this compound were legal in Russia until the middle of January, 2010. Presently, Russian drug enforcement bodies are trying to restrict their use because of increasing number of medical reports concerning severe complications related to the administration of the smoking

mixtures. This is not surprising as a typical marijuana abuser would anticipate a comparable activity of the synthetic cannabinoid, and this may lead to the overdose. Up to now only the in vitro data on the metabolic behavior of JWH-015 are available [9], whilst a number of publications deal with the detection and structural elucidation of the synthetic cannabinoids in herbal preparations [10–12]. To the best of our knowledge, no reports have been published on the human metabolism of JWH-018. Therefore, the aim of present study was to identify the urinary markers of JWH-018 administration and propose the analytical procedure for their assay.
2. Materials and methods 2.1. Reagents JWH-018, or naphthalene-1-yl-(1-pentylindole-3-yl)methanone, was purchased from Tocris Bioscience (Bristol, UK). HPLC solvents (acetonitrile, methanol, and water) were of gradient grade and were purchased from VWR (Leuven, Belgium). Diethyl ether was obtained from Medkhimprom (Moscow, Russia). b-Glucuronidase from Escherichia coli K12 (solution in 50% glycerol) was purchased from Roche Diagnostics (Mannheim, Germany) and used as supplied. N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) was obtained from Macherey-Nagel (Duren, Germany). All other chemicals were obtained from Sigma–Aldrich (St. Louis, MO). 2.2. Urine samples Urine samples were collected from 3 persons (2 males, 1 female, 22 Æ 1 year old) who were seized by police in a condition of drug intoxication. As reported, the behavioral effects were similar to those typical of the administration of the marijuana products, such as the reddening of eyes, tachycardia, anxiety, paranoia and hallucinations accompanied by a short-term memory defects and the impaired sense of time.

* Corresponding author. Tel.: +7 916 624 4389/495 258 3749; fax: +7 499 261 9943. E-mail address: (T. Sobolevsky). 0379-0738/$ – see front matter ß 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2010.04.003


T. Sobolevsky et al. / Forensic Science International 200 (2010) 141–147

Fig. 1. EI mass spectrum and structure of JWH-018.

These persons confessed that within the last 12 h they have administered one pack of the smoking mixture ‘‘Tropical Synergy’’. The maximal administered dose was ca. 1 g per person. They claimed this was their first experience with smoking mixtures. These persons were requested to provide the urine samples and upon signing the informed consent the respective samples were collected and delivered to the laboratory. The remainder of confiscated smoking mixture was also sent to the laboratory. As a blank urine a certified laboratory negative control urine was used. 2.3. Testing the smoking mixture for identity To check the composition of the smoking mixture, 5 mg of the raw material were extracted with 1 ml of methanol in an ultrasonic bath. Following centrifugation, 5 ml of the solution was taken to dryness and treated with MSTFA/NH4I/ dithiotreitol (1000/2/1.5; v/w/w) at 70 8C for 30 min. The reaction mixture was transferred into a 100-ml vial and 1 ml was injected into GC–MS.

2.4. Preparation of urine samples To 3 ml of urine was added 1 ml of phosphate buffer (0.8 M, pH 6.3) and 30 ml of b-glucuronidase. Following brief vortexing the samples were placed in an incubator where enzymolysis was allowed to proceed at 55 8C for 60 min. After that 1 ml of carbonate buffer (3 M, pH 10.1) was added and the samples were extracted with 5 ml of diethyl ether by rigorous vortexing in the presence of Na2SO4 as a salting out agent. After centrifugation at 3000 rpm for 4 min the aqueous layer was frozen in a low-temperature bath (À30 8C) and the ethereal extract was poured out into another test tube followed by evaporation at 40 8C under nitrogen flow. The dry residue was either dissolved in a methanol/water mixture (60/40; v/v) for the LC–MS analysis or treated with the same MSTFA/NH4I/ dithiotreitol reagent at 70 8C for 30 min. When the urinary free fraction was analyzed, adding b-glucuronidase and incubation at 55 8C were omitted with all other steps being the same.

Fig. 2. RIC plotted from GC–MS fullscan data against m/z 636 (A) and m/z 270 for post-administration (B) and laboratory negative control (C) urine (14.31 min: tetrahydrocortisol, 14.46 min: allo-tetrahydrocortisol, both as TMS artifacts).

T. Sobolevsky et al. / Forensic Science International 200 (2010) 141–147 2.5. GC–MS/MS The gas chromatographic analyses were carried out on the system comprising a Trace GC Ultra gas chromatograph (Thermo Scientific, Rodano, Italy) coupled to a TSQ Quantum GC triple quadrupole mass spectrometer (ThermoFisher Scientific, San Jose, CA, USA). The separation was achieved on a Thermo TR-5MS column (15 m  0.25 mm  0.1 mm) applying temperature programming as follows: 150 8C (1 min), heating at 10 8C/min to 300 8C (5 min). One ml injections were done at 250 8C in the split mode (1:30) with a carrier gas flow rate set to 1.2 ml/min (helium 99.9999%). Transfer line temperature was 300 8C, the ion source was held at


250 8C. The mass spectrometer was operated in the fullscan, MS/MS and SRM modes. In the latter two cases, argon (99.998%) was admitted into a Q2 at the pressure of 1.0  10À3 Torr and the collision energy (CE) of 10 or 20 eV. 2.6. LC–MS/MS The HPLC analyses were performed on a TSQ Vantage AM triple quadrupole mass spectrometer connected to an Accela liquid chromatograph (ThermoFisher Scientific, San Jose, CA, USA). Thermo Hypersil Gold column (50 mm  2.1 mm, particle size 1.9 mm) maintained at 40 8C was used for the separation. Injection

Fig. 3. EI mass spectra of the M1 (A), M2 (C) and MS/MS spectra of the M1 (B), M2 (D).


T. Sobolevsky et al. / Forensic Science International 200 (2010) 141–147

volume was 5 ml. The mobile phase flow rate was set to 250 ml/min with the gradient elution starting at 70% of 0.1% formic acid in water and 30% of acetonitrile followed by a linear increase to 90% of acetonitrile in 7 min, holding 90% of acetonitrile for 1 min and then re-equilibration until the end of analysis (10 min). The heated ESI II ion source was used for ionization. Positive and negative ions were detected in the fullscan, MS/MS (product ion scan, parent ion scan) and SRM modes. Argon was admitted into a Q2 at the pressure of 1.5 Â 10À3 Torr resulting in the Pirani gauge high vacuum reading of 2.0 Â 10À5 Torr. Sheath gas pressure (nitrogen from nitrogen generator, 99.9% purity) was set at 35 arbitrary units corresponding to nitrogen consumption of ca. 10 liters per min. The vaporizer and capillary temperatures were set at 350 8C with the spray voltage set at +3 kV (or À2.5 kV in negative ionization mode).

JWH analog [9] it was natural to anticipate that the in vivo metabolism of JWH-018 should occur similarly. 3.3. GC–MS/MS It is worthy to note that the parent compound was not found in the post-administration urines at the detection limit of ca. 50 pg/ml (in the SRM mode). The search for the metabolites was first attempted using GC–MS/MS following trimethylsilylation, and in this case hydroxylation should add 88 amu (MÀH+OSiMe3) to the molecular weight of parent compound. We detected two low abundance peaks (metabolites M1 and M2, see Fig. 6 for the proposed structures) with [M]+ = 429 amu eluting at the very end of the chromatograms after the distinctive doublet corresponding to the tetrahydrocortisol and allo-tetrahydrocortisol silylation artifacts. The reconstructed ion chromatograms (RIC) plotted for the tetrahydrocortisol doublet (m/z 636) and one of characteristic ions of the metabolites M1 and M2 (m/z 270) in the post-administration and laboratory negative control urine are presented in Fig. 2. We found that these two metabolites are almost completely glucuroconjugated, so enzymolysis is a must. Other metabolites listed in Fig. 6 were not detected by GC–MS/MS. Mass spectra of the metabolites M1 and M2 are given in Fig. 3A and C, respectively. As is seen from these data, both metabolites have their naphthalenyl moiety untouched, as the characteristic ions with m/z 127 and 155 are present in the mass spectra. The loss of trimethylsilanol (90 amu) resulting in the ion with m/z 339 is typical when a trimethylsilylated hydroxy group is available to cleave. The presence of ion with m/z 106 may be indicative of hydroxylation at the indole phenyl ring. The MS/MS spectra of the metabolites M1 and M2, both produced at the collision energy of 10 eV (Fig. 3B and D, respectively), clearly indicate that the metabolites have different positions of hydroxyls as fragmentation patterns are different. The ion with m/z 324 in the spectrum of metabolite M1 seems to be formed upon the loss of trimethylsilanol followed by the elimination of methyl group from the alkyl chain. This is favorable when hydroxyl is located at the indole phenyl ring and not at the Nalkyl side chain, therefore it is reasonable to assume that metabolite M2 is N-alkyl hydroxylated. The loss of 17 amu from

3. Results and discussion 3.1. Composition of smoking mixture Based on the GC–MS analysis of plant extract, the presence of JWH-018 in the mixture was confirmed by comparing the retention time and electron ionization (EI) mass spectrum with that of the reference compound. The extract was also found to contain a C8 homolog of CP 49 497, another cannabinoid receptor agonist occurring in herbal mixes [10], and a-tocopherol (data not shown). As is seen from Fig. 1, JWH-018 produces the informative mass spectrum. Interestingly, the molecule eliminates a methyl group and two protons resulting in the abundant ion at m/z 324 detected as an atypical loss of 17 amu. The ions at m/z 284 and 270 are produced by loosing the butyl and pentyl side chains, respectively. Of structural importance are the ions with m/z 127 and 155 which correspond to the naphthalenyl and carbonylnaphthalenyl fragments and, as discussed below, are diagnostic for the metabolites where the naphthalenyl moiety is not modified. The ion with m/z 214 is formed from the molecular ion upon the loss of the naphthalenyl group. 3.2. Urinary metabolites of JWH-018 It is well-known that hydroxylation is a common way of the metabolic transformation of xenobiotics as it facilitates their elimination from an organism by glucuronidation (or sulfation) of the hydroxylated metabolites. Based on the data reported for a

Fig. 4. SRM chromatograms for urine ca. 12 h after administration of JWH-018.

T. Sobolevsky et al. / Forensic Science International 200 (2010) 141–147


Fig. 5. RIC plotted from LC–MS fullscan data for the metabolites identified in the post-administration urine.

the molecular ion, as seen in the mass spectra of both metabolites, is again atypical and might arise from the elimination of methyl fragment of the trimethylsilyl group with subsequent loss of two hydrogen atoms. The SRM chromatogram obtained for the post-administration urine sample is given in Fig. 4. This data clearly demonstrate that metabolite M2 does not produce the ion with m/z 324. At this point it may be concluded that GC–MS/MS is suitable for detection of the monohydroxylated metabolites of JWH-018, preferably for the M1. Nevertheless we found that the gas chromatographic behavior of these metabolites is dependent on

the condition of column’s stationary phase, as after ca. 1000 injections the metabolite M1 and M2 peaks in the same urine samples exhibited severe tailing, and their retention times were shifted right by 0.3 min. 3.4. LC–MS/MS Using liquid chromatography combined with a tandem mass spectrometry one major peak with [M+H]+ corresponding to the molecular weight of the monohydroxylated metabolite (m/z 358) was observed on the chromatograms, with other related peaks

Fig. 6. Tentative structure of the identified urinary metabolites of JWH-018.


T. Sobolevsky et al. / Forensic Science International 200 (2010) 141–147

(RT = 3.91 min and 4.72 min, see below) being of a low abundance. At this point it is unclear which one is the metabolite M2 detected using GC–MS/MS. To find more metabolites, the parent ion scan MS/MS experiments were performed to detect any ions fragmenting to m/z 127 and 155, as well as to m/z 171 and 189 that should reflect hydroxylation at the naphthalenyl moiety [9]. It allowed detecting the minor metabolites, in addition to the data obtained by GC–MS/MS, with the abundance being at least one order of magnitude lower than the M1. These metabolites include: the carboxy metabolite M4 (M5 could be either carboxy or hydroxy-methoxy metabolite), dihydroxy metabolites M6, M7, reduced trihydroxy metabolites M8, M9, reduced dihydroxy metabolites M10, M11, and the N-despentyl

hydroxy metabolites M12, M13. Fig. 5 shows the mass chromatograms plotted from the fullscan data against the m/z ratio of the [M+H]+ ions of the detected metabolites, except M12 and M13. The ion trace for metabolites M12, M13 is not presented, as they coelute with some endogenous compound(s) of the same molecular weight (m/z 288) and are masked in the fullscan data. The proposed structures of the identified metabolites are summarized in Fig. 6. The MS/MS product ion scans were performed for each metabolite at CE = 20 eV. It was found that ion transitions [M+H]+ ! 127, 155 or [M+H]+ ! 171, 189 provide a high selectivity of analysis. The ESI(+) MS/MS spectrum for the metabolite M1 is given in Fig. 7A. One may see that the same ions with m/z 127 and 155 as

Fig. 7. ESI(+) MS/MS spectra of the metabolites M1 (A), M4 (B), M6 (C), M7 (D), M8 (E) and M10 (F).

T. Sobolevsky et al. / Forensic Science International 200 (2010) 141–147


observed in the electron ionization are present confirming that naphthalenyl moiety of this metabolite is untouched. The ESI(+) MS/MS spectrum for the metabolite M4 is shown in Fig. 7B. It should be noted that in the ESI(+) mode the detection of this metabolite may be difficult if the main monohydroxylated metabolite M1 is simultaneously monitored, as the M1 and M4 elute closely and fragment to the same product ions. To additionally support the assumption that the metabolite M4 is carboxylated, the urine samples were analyzed with the mass spectrometer operated in the negative ESI mode, as it is wellknown that carboxylic acids easily loose proton resulting in the [MÀH]À ions. Having performed the product ion scan from the ion with m/z 370, one peak with an exceptional selectivity was detected at the same retention time as in the positive ESI mode. With the CE set to 20 eV, the MS/MS spectrum consisted of the only ion with m/z 270. Structurally, this fragment may correspond to the loss of the carboxylated N-pentyl side chain. As our data indicate there are two closely eluting dihydroxy metabolites M6 and M7, the first one with the monohydroxylated naphthalenyl moiety (RT = 3.12 min, 374 ! 171, Fig. 7C) and the second one where it seems to be untouched (RT = 3.20 min, 374 ! 155, Fig. 7D). In addition, the trihydroxy metabolites with the reduced naphthalenyl moiety (M8, M9) were identified. The MS/MS spectrum for the metabolite M8 is shown in Fig. 7E. The reduced dihydroxy metabolites (M10, M11) were also found (Fig. 7F), with metabolite M11 being present in a negligible amount. As is seen, the MS/MS spectrum of metabolite M10 has the same characteristic ion transitions as the metabolite M8, fragmenting to the same product ions m/z 171 and 189. Since the N-despentyl hydroxy metabolites M12, M13 coelute with some of endogenous compounds, their MS/MS spectra were not possible to subtract and therefore are not presented. Nevertheless, the metabolite M12 may be detected by its characteristic ion transition 288 ! 155 and 288 ! 127. The metabolites are mostly glucuroconjugated. About 10% of the metabolite M12 and 1% of the metabolite M10 were detected in the free fraction. The carboxylated metabolite M4 was found only partly conjugated. Overall, JWH-018 metabolism proceeds in a good agreement with the in vitro data for another aminoalkylindole [9], JWH-015, except that the carboxylated metabolite was not reported before. Due to their characteristic mass spectra and distinctive ion transitions, the metabolites may be easily detected by LC–MS/ MS. At present it is not clear which of the metabolites are the most long-term, but it may be anticipated that the monohydroxylated and carboxylated metabolites M1 and M4 should be valuable for

the detection of JWH-018 in urine. Authors would also like to stress that until confirmed by synthesis, the proposed structures of the metabolites should be considered as tentative. In addition, since all the data were obtained as a result of the uncontrolled administration of smoking mixture by only 3 persons, the metabolites identified and their ratios may vary. 4. Conclusion The metabolites of a synthetic cannabimimetic indole JWH-018 were identified in the post-administration urines. The main monohydroxylated metabolites are excreted as conjugates with glucuronic acid and can be reliably detected either by GC–MS/MS or LC–MS/MS. However, LC–MS/MS seems to be a preferable method of detection as the minor metabolites could also support the analytical finding. References
[1] M.M. Aung, G. Griffin, J.W. Huffman, M.-J. Wu, C. Keel, B. Yang, V.M. Showalter, M.E. Abood, B.R. Martin, Influence of the N-1 alkyl chain length of cannabimimetic indoles upon CB1 and CB2 receptor binding, Drug Alcohol Depend. 60 (2000) 133– 140. [2] J.W. Huffman, Cannabimimetic indoles, pyrroles, and indenes: structure–activity relationships and receptor interactions, in: P.H. Reggio (Ed.), The Cannabinoid Receptors, Humana Press, Totowa, NJ, 2009, pp. 49–94. [3] M.E. Abood, Molecular biology of cannabinoid receptors, in: R. Pertwee (Ed.), Handbook of Experimental Pharmacology, vol. 168, Springer-Verlag, New York, 2005, pp. 81–115. [4] A.R. Stoit, J.H.M. Lange, A.P. den Hartog, E. Ronken, K. Tipker, H.H. van Stuivenberg, J.A.R. Dijksman, H.C. Wals, C.G. Kruse, Design, synthesis and biological activity of rigid cannabinoid CB1 receptor antagonists, Chem. Pharm. Bull. 50 (8) (2002) 1109–1113. [5] A. Makriyannis, Cannabimimetic indole derivatives. United States Patent No. 7,241,799 B2. (Jul. 10, 2007). [6] Understanding the ‘‘Spice’’ phenomenon. European Monitoring Centre for Drugs and Drug Addiction, Lisbon, 2009. Website: html.cfm/index90917EN.html (accessed 11.01.2010). [7] C. Steup, Untersuchung des Handelsproduktes ‘‘Spice’’. Note from THC Pharm GmbH. Website: (accessed 11.01.2010). [8] Synthetic cannabinoids and ‘‘Spice’’. European Monitoring Centre for Drugs and Drug Addiction, Lisbon, 2009. Website: index90414EN.html (accessed 11.01.2010). [9] Q. Zhang, P. Ma, R.B. Cole, Identification of in vitro metabolites of JWH-015, an aminoalkylindole agonist for the peripheral cannabinoid receptor (CB2) by HPLCMS/MS, Anal. Bioanal. Chem. 386 (2006) 1345–1355. [10] N. Uchiyama, R. Kikura-Hanajiri, N. Kawahara, Y. Goda, Identification of a cannabimimetic indole as a designer drug in a herbal product, Forensic Toxicol. 27 (2) (2009) 61–66. [11] N. Uchiyama, R. Kikura-Hanajiri, N. Kawahara, Y. Haishima, Y. Goda, Identification of a cannabinoid analog as a new type of designer drug in a herbal product, Chem. Pharm. Bull. 57 (4) (2009) 439–441. [12] R. Lindigkeit, A. Boehme, I. Eiserloh, M. Luebbecke, M. Wiggermann, L. Ernst, T. Beuerle, Spice: a never ending story? Forensic Sci. Int. 191 (1–3) (2009) 58–63.