Strategies to distinguish new synthetic cannabinoid FUBIMINA (BIM-2201) intake from its isomer THJ-2201: metabolism of FUBIMINA in human hepatocytes

Since 2013, a new drugs-of-abuse trend attempts to bypass drug legislation by marketing isomers of scheduled synthetic cannabinoids (SCs), e.g., FUBIMINA (BIM-2201) and THJ-2201. It is much more challenging to confirm a specific isomer’s intake and distinguish it from its structural analog because the isomers and their major metabolites usually have identical molecular weights and display the same product ions. Here, we investigated isomers FUBIMINA and THJ-2201 and propose strategies to distinguish their consumption. THJ-2201 was scheduled in the US, Japan, and Europe; however, FUBIMINA is easily available on the Internet. We previously investigated THJ-2201 metabolism in human hepatocytes, but human FUBIMINA metabolism is unknown. We aim to characterize FUBIMINA metabolism in human hepatocytes, recommend optimal metabolites to confirm its consumption, and propose strategies to distinguish between intakes of FUBIMINA and THJ-2201. FUBIMINA (10 μM) was incubated in human hepatocytes for 3 h, and metabolites were characterized with high-resolution mass spectrometry (HR-MS). We identified 35 metabolites generated by oxidative defluorination, further carboxylation, hydroxylation, dihydrodiol formation, glucuronidation, and their combinations. We recommend 5′-OH-BIM-018 (M34), BIM-018 pentanoic acid (M33), and BIM-018 pentanoic acid dihydrodiol (M7) as FUBIMINA specific metabolites. THJ-2201 produced specific metabolite markers 5′-OH-THJ-018 (F26), THJ-018 pentanoic acid (F25), and hydroxylated THJ-2201 (F13). Optimized chromatographic conditions to achieve different retention times and careful selection of specific product ion spectra enabled differentiation of isomeric metabolites, in this case FUBIMINA from THJ-2201. Our HR-MS approach should be applicable for differentiating future isomeric SCs, which is especially important when different isomers have different legal status.


Introduction
Synthetic cannabinoids (SCs) were originally developed as tools to study the endogenous endocannabinoid system [1,2]. Many SCs are CB 1 and CB 2 receptors agonists, eliciting greater cannabimimetic effects than D 9 -tetrahydrocannabinol [3,4]. However, SCs marketed as drugs-ofabuse since the 2000s are an important global public health and safety issue [5]. SC drug abuse can produce significant human toxicity including agitation, seizures, hypertension, emesis, myocardial infarction, and even death [6][7][8][9]. For these reasons, many SCs were scheduled across the globe, and many more structural analogs emerged. However, little is known about metabolism of these new SCs, which makes it challenging to document their intake in clinical screenings, complicated by the fact that the parent SCs rarely are found in urine, the most common analytical matrix.
Compared to structural analogs, it is much more challenging to confirm the intake of isomers of scheduled SCs by mass spectrometry (MS) because the SCs and their major metabolites display identical molecular ions and often identical product ions on a mass spectrometer. FUBIMINA (BIM-2201), BIM-018, AB-CHMINACA, and 1-n-pentyl-3-benzoylindole are isomers of THJ-2201, THJ-018, AB-CHMINACA 2H-indazole analog, and 1-benzoyl-3-n-pentylindole, respectively ( Fig. 1) [10][11][12]. In many cases, the isomers endow different scheduling status. For instance, THJ-2201 is scheduled in the US, Japan, and some European countries; however, FUBI-MINA, while being scheduled in Japan since August 2014, is currently available over the Internet in the US. Therefore, it is essential to confirm which SC is consumed for the purpose of legal prosecution. Previously, DeRuiter et al. [12] described the differentiation of 1-npentyl-3-benzoylindole and 1-benzoyl-3-n-pentylindole. However, the methods are limited to the parent drugs without considering metabolites. Because all previously studied SCs are extensively metabolized, major metabolites information is more important than parents for confirmation of intake.
All SCs investigated so far were highly metabolized, with rare detection of the parent SCs in urine, making identification of urinary metabolites essential to document SC intake in this matrix [13,14]. However, for newly emerging SCs, metabolism data are generally unavailable. We previously determined THJ-2201 metabolism with human hepatocytes and high resolution mass spectrometry (HR-MS) [15]. Currently, the only available FUBIMINA metabolism data were reported by Wiley et al. [3], who focused on FUBIMINA affinity and potency, only mentioning detection of some metabolites. In rat urine, hydroxylation, oxidative defluorination, and carboxylation metabolites were observed. However, these helpful preliminary data also have limitations. Metabolism was in rats, rather than humans; metabolites were mentioned, but no structures were characterized; the relative abundance of metabolites was not provided; and no marker metabolites were recommended for future identification of FUBIMINA intake.
Since 2005, HR-MS has become more widely available for characterizing new drug metabolism due to its many advantages over traditional unit-resolution MS [16], including metabolite identification and bioactivation mechanism elucidation [17][18][19]. Accurate mass measurement efficiently confirms metabolite parent ion and fragment ion formulas and achieves metabolite identification with improved productivity and quality. Traditional unitresolution MS requires multiple injections with different acquisition methods by experienced staff [16]; data analysis is time-consuming and less reliable due to the uncertainty of fragment elemental composition. Our aims were to characterize human FUBIMINA metabolism with human hepatocytes, recommend appropriate metabolites for clinical screening, and propose strategies to distinguish FUBIMINA from THJ-2201 intake. Human hepatocytes contain all required cofactors, enzymes, and an intact membrane, producing ratios of metabolites more closely matched to those in authentic human urine than those produced from human liver microsomes (HLM) [20][21][22]. Our approach using HLM to determine parent compound half-life and properly design human hepatocyte experiments, human hepatocyte incubation to produce phase I and phase II metabolites followed by HR-MS analysis, and sophisticated data analysis have proved successful in predicting major human urine metabolites for other SCs including AB-PINACA [23], AB-FUBINACA [24], FDU-PB-22, and FUB-PB-22 [14]. The spectral data for major metabolites will be highly useful for clinical scientists to incorporate into their urine screening methods to identify FUBIMINA intake.

Metabolic stability of FUBIMINA in human liver microsomes
The HLM metabolic stability assays were performed in the same manner as in our previous manuscripts [25,26]. Final concentration of FUBIMINA in HLM incubation system was 1 lM. Samples were centrifuged at 15,000g (4°C, 5 min); supernatant was stored at -80°C until analysis. When analyzing, the samples were thawed, vortexed, and centrifuged again. Then, 10 lL supernatant was diluted in 990 lL mobile phase A/B (90: 10, v/v). Ten microliters was injected to determine remaining FUBIMINA.
The high-performance liquid chromatography (HPLC) system consisted of two LC-20ADxr pumps, a DGU-20A3R degasser, a SIL-20ACxr autosampler, and a CTO-20A column oven (Shimadzu, Columbia, MD, USA). The Kinetex C 18 column (100 mm 9 2.1 mm, 2.6 lm; Phenomenex, Torrance, CA, USA) was fitted with a Krud-Katcher Ultra HPLC in-line filter (0.5 lm 9 0.1 mm, Phenomenex). Mobile phases were 0.1 % formic acid in water (A) and 0.1 % formic acid in acetonitrile (B). The gradient elution started at 10 % B for 0.5 min, ramped to 95 % B at 10 min, then held until 12.5 min before reequilibrating at 10 % B for 2.5 min. Total run time was 15 min with a flow rate of 0.3 mL/min. Column and autosampler temperatures were 40 and 4°C, respectively.
FUBIMINA peak areas were plotted against time, and in vitro microsomal half-life (T 1/2 ) and intrinsic clearance (CL int, micr ) were calculated. Microsomal intrinsic clearance was then scaled to whole-liver dimensions, yielding intrinsic clearance (CL int ). Human hepatic clearance (CL H ) and extraction ratio (ER) were predicated based on CL int without considering plasma protein binding.

Metabolite identification in human hepatocytes
Hepatocyte incubation was carried out as previously described [15,25,26]. Cryopreserved human hepatocytes were washed twice with GRO CP medium and GRO KHB buffer. Cell viability was evaluated with Trypan blue (0.4 %, v/v) exclusion method, assuring greater than 80 % viability. FUBIMINA was incubated at a final concentration of 10 lM with human hepatocytes (1 9 10 6 cells/mL) in a 24-well plate (BD Biosciences, San Jose, CA, USA). Final incubation volume was 500 lL and samples were incubated at 37°C. Chemical stability of FUBIMINA in the KHB buffer, and hepatocyte incubation without FUBIMINA also were performed (37°C, 3 h) to determine whether metabolites are dependent on hepatocytes enzymes or are impurities from hepatocyte itself. Diclofenac was incubated as a positive control to ensure metabolic activity of the hepatocytes by monitoring formation of 4 0 -hydroxydiclofenac and acyl b-D-glucuronide diclofenac. The reactions were quenched by adding 500 lL ice-cold acetonitrile at 0 and 3 h incubation. Samples were stored at -80°C until analysis. Before analysis, samples were thawed, vortexed thoroughly, and 100 lL acetonitrile added to 100 lL samples. After vortexing and centrifugation at 15,000 g (4°C, 5 min), supernatant was transferred to a new 10-mL plastic tube, evaporated to dryness under nitrogen at 40°C, and reconstituted in 150 lL mobile phase A/B (80:20, v/v). Fifteen microliters of the reconstituted solution was injected for metabolite identification.
The HPLC system consisted of two LC-20ADxr pumps, a DGU-20A5R degasser, a SIL-20ACxr autosampler, and a CTO-20 AC column oven (Shimadzu). Chromatographic separation was achieved on an Ultra Biphenyl column (100 9 2.1 mm, 3 lm; Restek, Bellefonte, PA, USA) with a guard column containing the same packing material. Gradient elution was performed with 0.1 % formic acid in water (A) and 0.1 % formic acid in acetonitrile (B) at a flow rate of 0.5 mL/min. Initial gradient conditions were 20 % B, held for 0.5 min; then increased to 95 % B over 10.5 min, held until 13.0 min; and returned to 20 % B at 13.1 min and held until 15.0 min. The column and autosampler were maintained at 30 and 4°C, respectively.
Data were acquired on a 5600 ? TripleTOF MS (Sciex) in ?ESI mode. MS data were acquired by informationdependent acquisition (IDA) in combination with multiple mass defect filters (MDF) and dynamic background subtraction (DBS). Ion source parameters were as follows: curtain gas, 45 psi; gas 1, 60 psi; gas 2, 75 psi; source temperature 650°C; ion spray voltage, 4000 V; declustering potential, 80 V; entrance potential, 10 V. For IDA, spectra exceeding 100 cps were selected for the metabolism-dependent tandem mass spectrometry MS/MS scan, isotopes within 1.5 Da were excluded, mass tolerance was 50 mDa, and collision energy was set to 35 ± 15 eV. The MS was calibrated automatically every three injections.
MetabolitePilot (version 1.5, Sciex) characterized metabolites with different peak-finding algorithms including MDF, predicted biotransformation, generic liquid chromatography (LC) peak finding, common product ion and neutral loss; LC peak intensity threshold was 500 cps, MS 200 cps, and MS/MS 50 cps. Special attention was given to phase II metabolites that were susceptible to insource fragmentation.
To distinguish FUBIMINA from THJ-2201, THJ-2201 hepatocyte samples incubated in the same procedure were analyzed with the same acquisition and processing methods.

Results and discussion
Metabolic stability evaluation in human liver microsomes After determining remaining FUBIMINA at different time points, the in vitro T 1/2 was calculated to be 4.9 ± 0.03 min, and in vitro CL int, micr was 0.142 mL/min/mg, corresponding to intrinsic clearance (CL int ) of 134.1 mL/min/kg after scaling to whole-liver dimensions [27]. Without considering plasma protein binding and with a simplified Rowland's equation [21,28], we calculated the predicted human CL H as 17.4 mL/min/kg and ER as 0.87.
Metabolic stability, expressed as in vitro T 1/2 and CL int defines a drug's susceptibility to metabolism. These values facilitate the prediction of in vivo hepatic clearance, in vivo T 1/2 , and bioavailability [21,29]. According to criteria proposed by Lave et al. [30], short T 1/2 , large CL int , and high ER demonstrate that FUBIMINA is a high clearance drug. THJ-2201 eluted later at 9.21 min (Fig. 2b), and produced similar fragments (Fig. 2d), but with quite different relative ion abundances, as for m/z 177.0467 and 155.0499. Notably, the prominent fragment m/z 273.1041 in FUBI-MINA was not observed in THJ-2201.

Oxidative defluorination and further oxidation plus glucuronidation
M34 also underwent benzimidazole hydroxylation and glucuronidation to M14/M16. Their major fragment, m/z 375.1714 was generated via loss of glucuronide (Fig. 5c). Fragments m/z 289.0979 and 161.0343 were 15.9947 Da (?O) larger than corresponding M34 ions, suggesting benzimidazole hydroxylation. Glucuronidation may also occur on the benzimidazole hydroxyl group.
Urine is the most common routine matrix for drug screening because of easy accessibility and high drug concentration compared to blood or oral fluid. There are limitations to extrapolating human hepatocyte data to human urine. Extrahepatic metabolism [38], transporter involvement [39], and metabolite enrichment in urine [40] might change relative abundance of urinary metabolites, and thus selection of marker metabolites. Therefore, it is preferable to obtain FUBIMINA and THJ-2201 positive urine case samples, and confirm whether recommended hepatocyte marker metabolites match those present in authentic urine after SC intake. Unfortunately, such samples are currently unavailable despite our efforts to obtain them.

Conclusions
In summary, we characterized FUBIMINA metabolism in human hepatocytes with HR-MS. We recommend 5 0 -OH-BIM-018 (M34), BIM-018 pentanoic acid (M33), and BIM-018 pentanoic acid dihydrodiol (M7) as the most promising urinary marker metabolites for documenting FUBIMINA intake. Their retention times and unique product ion spectra facilitate differentiation of FUBI-MINA from THJ-2201 for clinical and forensic scientists. The data will enable linkage of potential adverse toxicological events to FUBIMINA or THJ-2201, in order that the public can be informed of their toxicity. These data also empower manufacturers to focus their synthesis efforts on optimal metabolites. We suggest workflow to differentiate isomeric SCs as follows: incubate SC isomers in human hepatocytes, analyze incubation samples by HPLC-HR-MS, characterize unique major metabolites' retention times and fragmentation profile, and confirm with case urine samples. Our strategies are applicable for distinguishing other SC isomers with different legal status.