Anal Bioanal Chem (2016) 408:503–516 DOI 10.1007/s00216-015-9116-1
RESEARCH PAPER
Determination of urinary metabolites of XLR-11 by liquid chromatography–quadrupole time-of-flight mass spectrometry Moonhee Jang 1 & In Sook Kim 2 & Yu Na Park 2 & Jihyun Kim 1 & Inhoi Han 1 & Seungkyung Baeck 1 & Wonkyung Yang 1 & Hye Hyun Yoo 2
Received: 24 July 2015 / Revised: 6 October 2015 / Accepted: 13 October 2015 / Published online: 29 October 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Recently, use of novel synthetic cannabinoids has increased greatly despite worldwide efforts to regulate these drugs. XLR-11 ((1-[5'-fluoropentyl]indol-3-yl)-(2,2,3,3tetramethylcyclopropyl)methanone), a fluorinated synthetic cannabinoid with a tetramethylcyclopropyl moiety, has been frequently abused since 2012. XLR-11 produces a number of metabolites in common with its non-fluorinated parent analogue, UR-144 ((1-pentylindol-3-yl)-(2,2,3,3tetramethylcyclopropyl)methanone). Therefore, it is essential to develop effective urinary markers to distinguish between these drugs. In this study, we investigated the metabolic profile of authentic human urine specimens from suspected users of XLR-11 using liquid chromatography–quadrupole time-offlight mass spectrometry. Furthermore, we quantified four potential XLR-11 metabolites by using commercially available reference standards. In vitro metabolism of XLR-11 and UR144 using human liver microsomes was also investigated to compare patterns of production of hydroxypentyl metabolites. Urine samples were prepared with and without enzymatic hydrolysis, and subjected to solid-phase extraction. We identified 19 metabolites generated by oxidative defluorination, hydroxylation, carboxylation, dehydrogenation, glucuronidation, and combinations of these reactions. Among
* Wonkyung Yang
[email protected] * Hye Hyun Yoo
[email protected] 1
National Forensic Service, 139 Jiyang-ro, Yangcheon-gu, Seoul 158-707, Republic of Korea
2
Institute of Pharmaceutical Science and Technology and College of Pharmacy, Hanyang University, Ansan, Gyeonggi-do 426-791, Republic of Korea
the identified metabolites, 12 were generated from a cyclopropyl ring-opened XLR-11 degradation product formed during smoking. The XLR-11 metabolite with a hydroxylated 2,4-dimethylpent-1-ene moiety was detected in most specimens after hydrolysis and could be utilized as a specific marker for XLR-11 intake. Quantitative results showed that the concentration ratio of 5- and 4-hydroxypentyl metabolites should also be considered as a useful marker for differentiating between the abuse of XLR-11 and UR-144. Keywords Synthetic cannabinoid . XLR-11 . UR-144 . Urinary metabolites . Liquid chromatography–quadrupole time-of-flight mass spectrometry
Introduction A number of synthetic cannabinoids were developed for therapeutic applications following the discovery of the CB1 and CB2 cannabinoid receptors in the late 1980s. However, unexpectedly, they began to acquire increased popularity as legal substitutes for cannabis in Europe in 2008. Over the past several years, synthetic cannabinoids have been abused worldwide, and legal authorities of many countries are engaged in a multilateral effort to control the use of these drugs. Despite these efforts, the life cycle of synthetic cannabinoid derivatives from emergence to disappearance has become continually shorter. According to a report on changes in synthetic cannabinoid abuse in Japan, classic naphthoylindoles comprised up to 80 % of synthetic cannabinoids identified in herbal products until late 2011, but comprised less than 10 % in early 2012 [1]. They have been replaced by phenylacetylindole- and benzoylindole-type derivatives, and more recently by adamantylindoles and adamantylindazoles such as APICA and APINACA. In late 2011, a new non-
504
classic compound, UR-144 ((1-pentylindol-3-yl)-(2,2,3,3tetramethylcyclopropyl)methanone), and its fluorinated analogue, XLR-11 ((1-[5'-fluoropentyl]indol-3-yl)-(2,2,3,3tetramethylcyclopropyl)methanone), were introduced to the drug market [2]. UR-144 and XLR-11 were developed to explore the activity of selective CB2 receptor agonists [3, 4]; however, they also bind to the CB1 receptor with nanomolar Ki values ( **) Urine_MSMS_30155.d 3.75 125.0958 m/z 230.1170 3.5 O 3.25 3 2.75 2.5 2.25 N 2 1.75 1.5 m/z 328.2248 230.1170 1.25 Glu 1 328.2248 O 0.75 -176 Da 0.5 69.0705 174.9708 290.1251 504.2590 393.0258 0.25 0 50 100 150 200 250 300 350 400 450 500 550 600 Counts vs. Mass-to-Charge (m/z)
m/z 328.2253 Glu
O 85.0273
50
M7-2
m/z 230.1170
100
186.1268
150
200
272.1295
-176 Da
250 300 350 400 450 Counts vs. Mass-to-Charge (m/z)
504.2555
500
550
600
M8 m/z 244.0953
Fig. 2 (continued)
previous metabolism studies of synthetic cannabinoids containing a 5-fluoropentyl side chain [12–15]. Carboxylated metabolites were found at m/z 342.2058 (M3-1) and m/z 342.2053 (M3-2) and showed a characteristic fragment at m/z 244.0979 indicating oxidation of the terminal carbon to a carboxylic acid. The chemical structures of M3-1 and M3-2 were confirmed by comparison with the chromatographic and mass spectrometric data of the reference standards, resulting in their identification as UR-144 degradation product NCOOH M and UR-144N-COOH M, respectively. The elemental composition of M2-1 and M2-2 was C21H25NO3. M2-1 was observed as a major metabolite of XLR-11 in urine samples. When the product ion mass spectra of M2-1 was compared with those of M3-1 and M3-2, the presence of fragment ions at m/z 144.0454, 144.0475, and 244.0979 revealed the loss of two hydrogen atoms from the 2,4-dimethylpent-1ene or cyclopropyl ring moieties. Therefore, it can be deduced that M2 metabolites were generated by dehydration following hydroxylation of M3 metabolites. The M2 metabolites were postulated to be derived from XLR-11 degradation and intact XLR-11 on the basis of the retention times of M3-1 and M3-2, respectively. A metabolite with a molecular ion at m/z 340 was reported previously [11], but its structure was not assigned. M4 was found at m/z 346.2163 as a protonated ion and
identified as a monohydroxylated metabolite of the XLR-11 degradation product. The characteristic fragment m/z 99.0803 suggests hydroxylation of the 2,4-dimethylpent-1-ene moiety, and a fragment at m/z 232.1140 is associated with an unchanged fluoropentylindole moiety. However, the location of the hydroxyl group in M4 could not be assigned in this study. The M5 metabolites exhibited an elemental composition of C21H27NO4 and were formed by further oxidation of M3-1 or M3-2. The presence of fragment ions at m/z 144, 200, and 244 suggested hydroxylation (or oxygenation) of the 2,4dimethylpent-1-ene (M5-1 and M5-2) or cyclopropyl ring moieties (M5-3 and M5-4). M6-1 and M6-2 (C21H27NO5) may have resulted from further oxidation of the hydroxyl group of the M5 metabolites to a carboxylic acid, as evidenced by fragment ions at m/z 244.0956 and 244.0925. The M6 metabolites were also observed as isobaric forms. The early eluting metabolite (M6-1) was proposed to be the metabolite of the XLR-11 degradation product on the basis of the fragment ion at m/z 99.0806. M7–M10 were postulated to be glucuronide conjugated metabolites, which commonly show fragment ions resulting from the loss of a glucuronic acid (−176 Da), with the exception of M8. M7-1 and M7-2 were determined to be glucuronides of 5-hydroxypentyl metabolites. The presence of two isobaric forms suggests that they
Determination of urinary metabolites of XLR-11
511
M9-2
M9-1 m/z 230.1165
m/z 230.1165
OH
O
m/z 99.0800 m/z 99.0810 246.1478
Glu N
-176 Da -176 Da
OH
M10-2
M10-1
x10 4 +ESI Product Ion (5.2 min) Frag=90.0V
[email protected] (538.2447[z=1] -> **) Urine_MSMS_30155.d 1.1 OH m/z 248.1071 248.1071 1
m/z 248.1071
O
0.9
m/z 99.0804
0.8
Glu
0.7
N
0.6
248.1071
0.5 0.4
-176 Da
OH F
0.3 344.2018 362.2148
0.2
538.2464
206.1332
99.0791
0.1
-176 Da 463.2712
0 50
100
150
200
250 300 350 400 450 Counts vs. Mass-to-Charge (m/z)
500
550
600
650
Fig. 2 (continued)
are produced from M1 and 5-hydroxypentyl UR-144, respectively. M8 was identified as a glucuronide of M3-1 on the basis of the characteristic fragment ions of carboxylic acid metabolites (M3 or M5) at m/z 200.1054 (a loss of CO2 from m/z 244.0953) and m/z 244.0953. M9-1 and M9-2 yielded major fragment ions at m/z 344.2194 and 344.2198 which may have resulted from hydroxylation of M1. The fragment ions at m/z 99 and 230 indicated additional hydroxylation of the 2,4-dimethylpent-1-ene moiety. Metabolites M10-1 and M10-2 were postulated as glucuronides of the dihydroxylated metabolites from XLR-11 degradation and intact XLR-11, respectively; however, they showed different fragmentation patterns. M10-1 fragment ions at m/z 99.0804 and 248.1071 suggested that glucuronidation occurred following further hydroxylation of M4 (a degradation metabolite), whereas M10-2 was proposed to be conjugated from the cyclopropyl ringhydroxylated metabolite on the basis of the absence of the fragment ion at m/z 99.0791. The metabolites identified in urine specimens are presented in Table 3. The tested specimens showed different metabolite profiles. Parent drug, XLR-11, was not detected in any specimen, whereas 1–15 metabolites were detected in each specimen. As shown in Table 3, UR-144 degradation product NCOOH M (M3-1) was detected in all analyzed samples. All
samples except sample no. 6 contained M5-1 and M5-3. M2-1 was also frequently detected and was one of the most abundant metabolites in specimen no. 1, along with M3-1, M5-1, and M5-3, as determined by comparing their peak areas. Metabolite M4, a specific metabolite of XLR-11, was identified in all hydrolyzed specimens except for specimen no. 6. The glucuronide conjugated metabolites (M7–M10) were detected in only three specimens (specimen nos. 1, 2, and 5). Most specimens showed increased levels of metabolites after hydrolysis by β-glucuronidase, indicating that XLR-11 generally undergoes phase II metabolism following phase I metabolism. In vitro metabolism study In previous studies, we showed that relative concentrations of some metabolites were important factors that could be used to determine synthetic cannabinoid abuse [12, 13]. Therefore, we performed an in vitro metabolism study of XLR-11 and UR-144 to compare the relative abundance of metabolites, with a focus on commercially available metabolites. XLR-11 readily underwent oxidative defluorination to generate common metabolites (UR-144 N-5-OH M and a trace amount of UR-144 N-COOH M) with UR-144. XLR-11 N-4-OH M, the only commercially available specific metabolite of XLR-11,
512 Table 3
M. Jang et al. Identification of XLR-11 metabolites in authentic specimens
Metabolites
XLR-11 M1
Specimen no. 1
1H
2
2H
3
3H
4
4H
5
5H
6
6H
– –
– ◯
– ◯
– ◯
– –
– –
– –
– –
– ◯
– ◯
– –
– –
M2-1
◯
◯
◯
◯
◯
◯
◯
–
–
M2-2 M3-1
◯ ◯
◯ ◯
– ◯
– ◯
– ◯
– ◯
◯ – ◯
– ◯
– ◯
– ◯
– ◯
– ◯
M3-2 M4
◯ –
◯ ◯
– –
– ◯
– –
– ◯
– –
– ◯
– ◯
– ◯
– –
– –
M5-1
◯
◯
◯
◯
◯
◯
◯
◯
◯
◯
–
–
M5-2 M5-3
– ◯
– ◯
– ◯
– ◯
◯ ◯
– ◯
– ◯
– ◯
– ◯
– ◯
– –
– –
M5-4 M6-1
◯ ◯
◯ ◯
◯ –
◯ –
– –
– –
– –
– –
◯ –
◯ –
– –
– –
M6-2 M7-1
◯ ◯
◯ –
◯ ◯
◯ –
– –
– –
– –
– –
◯ ◯
◯ –
– –
– –
M7-2 M8
◯ –
– –
– ◯
– –
– –
– –
– –
– –
– –
– –
– –
– –
M9-1 M9-2
◯ ◯
– –
– –
– –
– –
– –
– –
– –
– –
– –
– –
– –
M10-1 M10-2
◯ ◯
– –
– –
– –
– –
– –
– –
– –
– –
– –
– –
– –
H hydrolyzed sample, ◯ detected
was not produced during XLR-11 metabolism. As depicted in Fig. 3, UR-144 N-5-OH M was predominant in the XLR-11 incubation mixture, while both UR-144 N-5-OH M and UR-144 N-4-OH M were identified following UR144 metabolism.
Quantification of XLR-11 metabolites in authentic specimens The method validation data are summarized in Tables 4 and 5. No endogenous interference with the signals of the analytes or internal standards was observed in blank urine samples (n= 10). Linear ranges were 0.25–100 ng/mL for UR-144 NCOOH M, UR-144 N-5-OH M, and UR-144 N-4-OH M and 0.25–50 ng/mL for UR-144 degradation product N-COOH M and XLR-11 N-4-OH (1/x2 weighted linear regression). Average correlation coefficients (r2) were greater than 0.99. The LOD and LOQ were 0.25 ng/mL for all analytes. Figure 4a shows the combined extracted ion chromatogram of blank urine sample fortified with analytes at low QC concentrations. Intra- and interday precision and accuracy were satisfactory at the three evaluated QC concentrations. Intra-and interday precision were 1.8–12.0 and 4.5–8.0 %, respectively. Intra- and interday accuracy ranged from –5.6 to 4.1 %. Mean values of matrix effect were between 78 and 96 %. For all analytes, the variation of the matrix effect from different sources was less than 8.4 %, which is lower than the limit of 15 % proposed by Viswanathan et al. [21]. Mean extraction recoveries ranged from 74 to 91 %. Dilution integrity was determined by calculating the accuracies of blank samples fortified at 500 ng/mL after a 1:10 dilution, showing that all analytes were within 15 % of the nominal concentration. No carry-over was observed after injection of the sample at 500 ng/mL. The validated method was applied to the 18 urine specimens collected from suspected XLR-11 users (Table 6). In addition, the representative extracted ion chromatograms of the analyzed urine specimens are shown in Fig. 4b–g. In the LC–MS/MS analysis, UR-144 N-5-OH M and UR-144 N-4OH M were chromatographically separated. As noted in
a XLR-11
UR-144 N-5-OH M 341.9 327.9
125.1 125.1
UR-144 N COOH M
b UR-144
UR-144 N-4-OH M UR-144 N-5-OH M
UR-144 N COOH M
Fig. 3 Representative combined extracted ion chromatograms of in vitro metabolites of XLR-11 (a) and UR-144 (b) after incubation with human liver microsomes. The chromatographic peaks of analyte are indicated with respective quantifier transition
Determination of urinary metabolites of XLR-11 Table 4
513
Linearity, matrix effect and recoveries for analysis of XLR-11 metabolites in human urine by LC–MS/MS (n=5)
Analyte
Calibration curve Slope
Intercept
Matrix effect (%) r2
Low
Extraction recovery (%) High
Low
High
Mean
CV
Mean
CV
Mean
CV
Mean
CV
UR-144 N-COOH M UR-144 N-5-OH M
0.0598 0.0590
0.0139 0.0051
0.997 0.996
88 90
3.6 2.9
94 91
3.6 4.3
76 86
3.6 2.9
74 90
5.3 5.9
UR-144 N-4-OH M UR-144 degradation product N-COOH M
0.0747 0.0680
0.0069 0.0028
0.995 0.995
91 78
3.6 8.4
96 79
2.5 7.2
88 89
6.5 7.6
90 82
3.7 3.5
XLR-11 N-4-OH M
0.0676
0.0028
0.995
83
4.2
81
7.4
91
6.4
86
5.9
Low quality control concentrations were 0.5 ng/mL, and high quality control concentrations were 75 ng/mL for UR-144 N-COOH M, UR-144 N-5-OH, and UR-144 N-4-OH, and 40 ng/mL for UR-144 degradation product N-COOH M and XLR-11 N-4-OH M CV coefficient of variation
were generated by oxidative defluorination of the pentyl side chain, hydroxylation of the tetramethylcyclopropane ring or 2, 4-dimethylpent-1-ene moiety, carboxylation, dehydrogenation, glucuronidation, and a combination of these reactions. The oxidative defluorination was the most predominant pathway among them as detailed in Table 2. A previous in vitro study using human hepatocytes reported 30 metabolites of XLR-11 [11]. Comparing the metabolic profile in human hepatocytes with that we obtained in this study, seven metabolites (M3-2, M5-3, M5-4, M6-2, M7-2, M8, and M10-2) were common to both studies including major metabolites M3-2 and M6-2. The previous study using human hepatocytes reported UR-144 N-5-OH M as a major metabolite of XLR-11, while its corresponding metabolite (M1), derived from XLR-11 degradation, was more prevalent in authentic urine samples in the present study. A small amount of UR-144 N-5-OH M was detected by LC–MS/MS analysis in all tested samples. Similarly, UR-144 N-COOH M (M3-2) was also detected in all the analyzed samples using LC–MS/MS; however, it was identified in only specimen no. 1 by QTOF/MS. In a human hepatocyte-based metabolism study [11], a metabolite with a carboxylated cyclopropane ring
previous studies of the metabolism of fluorinated synthetic cannabinoids [12–15, 18], UR-144 N-5-OH M but not UR144 N-4-OH M was detected in all urine specimens (
