Investigation of phase II metabolism of 11-hydroxy-Δ-9-tetrahydrocannabinol and metabolite verification by chemical synthesis of 11-hydroxy-Δ-9-tetrahydrocannabinol-glucuronide

(−)-Δ-9-tetrahydrocannabinol ((−)-Δ-9-THC) is the main psychoactive constituent in cannabis. During phase I metabolism, it is metabolized to (−)-11-hydroxy-Δ-9-tetrahydrocannabinol ((−)-11-OH-Δ-9-THC), which is psychoactive, and to (−)-11-nor-9-carboxy-Δ-9-tetrahydrocannabinol ((−)-Δ-9-THC-COOH), which is psychoinactive. It is glucuronidated during phase II metabolism. The biotransformation of (−)-Δ-9-tetrahydrocannabinol-glucuronide ((−)-Δ-9-THC-Glc) and (−)-11-nor-9-carboxy-Δ-9-tetrahydrocannabinol-glucuronide ((−)-Δ-9-THC-COOH-Glc) is well understood, which is mainly due to the availability of commercial reference standards. Since such a standardized reference is not yet available for (−)-11-hydroxy-Δ-9-tetrahydrocannabinol-glucuronide ((−)-11-OH-Δ-9-THC-Glc), its biotransformation is harder to study and the nature of the glucuronide bonding—alcoholic and/or phenolic—remains unclear. Consequently, the aim of this study was to investigate the biotransformation of (−)-11-OH-Δ-9-THC-Glc in vitro as well as in vivo and to identify the glucuronide by chemically synthesis of a reference standard. For in vitro analysis, pooled human S9 liver fraction was incubated with (−)-Δ-9-THC. Resulting metabolites were detected by high-performance liquid chromatography system coupled to a high-resolution mass spectrometer (HPLC-HRMS) with heated electrospray ionization (HESI) in positive and negative full scan mode. Five different chromatographic peaks of OH-Δ-9-THC-Glc have been detected in HESI positive and negative mode, respectively. The experiment set up according to Wen et al. indicates the two main metabolites being an alcoholic and a phenolic glucuronide metabolite. In vivo analysis of urine (n = 10) and serum (n = 10) samples from cannabis users confirmed these two main metabolites. Thus, OH-Δ-9-THC is glucuronidated at either the phenolic or the alcoholic hydroxy group. A double glucuronidation was not observed. The alcoholic (−)-11-OH-Δ-9-THC-Glc was successfully chemically synthesized and identified the main alcoholic glucuronide in vitro and in vivo. (−)-11-OH-Δ-9-THC-Glc is the first reference standard for direct identification and quantification. This enables future research to answer the question whether phenolic or alcoholic glucuronidation forms the predominant way of metabolism. Electronic supplementary material The online version of this article (10.1007/s00414-020-02387-w) contains supplementary material, which is available to authorized users.

The aim of this study was to investigate the biotransformation of (−)-11-OH-Δ-9-THC-Glc in vitro as well as in vivo and to synthesize a reference standard for direct identification and quantification to fill the gap in main metabolite reference standards. For in vitro analysis, a human S9 liver fraction assay was selected, because it has proved to be a simple tool for metabolism studies for cannabinoids [9][10][11] and because it is used since a long time in our working group [16,17]. Metabolite detection was performed by high-performance liquid chromatography system coupled to high-resolution mass spectrometry (HPLC-HRMS) with heated electro spray ionization (HESI) in positive and negative full scan modes. To check the in vitro results, urine and serum samples of cannabis users, routinely analyzed at the Department of Forensic Toxicology Münster (Germany), are analyzed for (−)-11-OH-Δ-9-THC-glucuronide and verified with the reference standard synthesized.

Analysis of synthesis intermediates
Isolation of synthesis intermediates Flash chromatography was performed on Merck silica gel 60 (40-63 μm) with an excess argon pressure up to 1.0 bar. Merck silica gel 60 F254 plates were used for thin-layer chromatography (TLC) using UV light (254/366 nm), KMnO 4 (1.5 g in 200 mL H 2 O, 5 g NaHCO 3 ) for detection.
MS analysis of synthesis intermediates HRMS ESI measurements were performed using a Bruker MicroTof (Bremen, Germany).

In vitro analysis
The assay used for in vitro analysis is a combination of two previously published and in our working group established S9 fraction assays (Schwarzkopf et al. [16] and Holtfrerich et al. [17]). The combined assay was checked with respect to detergent use, incubation time and amount of protein used (data not shown). To verify the qualitative assay, paracetamol, as a standard test substance used in the working group, was first incubated with pooled human liver S9 fraction. Finally, (−)-Δ-9-THC reference solution and a negative control (phosphate buffer instead of NADPH/UDPGA solution) were incubated (n = 1, respectively) with the following assay details: All solutions, unless otherwise specified, were prepared in 0.1 M phosphate buffer. Final concentrations are given in brackets. To a solution of 5 μL alamethicin (100 μg/mL in EtOH/H 2 O, 1:1, v:v), 125 μL of pooled human liver S9 fraction (2 mg protein/mL) was added and then incubated for 15 min on ice. After that, 72 μL of 0.1 M phosphate buffer, 100 μL saccharolactone solution (4.9 mM), 100 μL of MgCl 2 solution (1.9 mM) and 3.14 μL of (−)-Δ-9-THC solution (1 mg/mL in EtOH/H 2 O, 1:1, v:v, 19.7 μM) are added sequentially. After incubating for 3 min at 37°C in a water bath, 100 μL of NADPH/UDPGA solution (0.9 mM NADPH/ 4.9 mM UDPGA) was added, leading to 505.14 μL total volume. The final EtOH concentration was 0.8%vol. Then, the mixture was incubated for 60 min at 37°C. The reaction was stopped with 200 μL ice-cold acetonitrile and the mixture was cooled at 0°C for 15 min. After 1 min vortexing and 5 min centrifugation at 12,000×g, 100 μL supernatant was removed, diluted with 50 μL methanol, and applied to HPLC-MS/MS analysis.

In vivo analysis
To verify the in vitro results, 10 urine samples and 10 serum samples from cannabis users, routinely analyzed at the Department of Forensic Toxicology Münster (Germany), were investigated. The urine sample (50 μL) was diluted 1:10 (v:v) by adding 200 μL 2 mM ammonium acetate buffer containing 0.1% formic acid and 250 μL acetonitrile with 0.1% formic acid. After centrifugation for 10 min at 12,000×g, 2 μL was injected into the HPLC-HRMS in full scan MS-1 mode. The serum samples were analyzed by solid phase extraction (SPE) followed by HPLC-HESI(+)-HRMS in parallel reaction monitoring (PRM) mode, because of its higher sensitivity compared with full scan mode. Five hundred microliters of serum was diluted with phosphate buffer (pH 6) and internal standard was added. Before loading, the SPE-C18 cartridge was activated in a common way. Successively, the SPE was washed with H 2 O, dried under vacuum and the analytes were eluted in two steps with acetone and methanol containing 0.1% formic acid. Then, the extracts were evaporated and resolved in 100 μL volume to achieve a higher sensitivity. The injection volume was 10 μL. The chromatographic separation was achieved with a gradient of 2 mM ammonium acetate buffer containing 0.1% formic acid and methanol with 0.1% formic acid (manuscript in preparation).

HPLC-HRMS and HPLC-HRMS/MS method
Metabolite screening was achieved by analyzing the abovementioned in vitro and in vivo samples with a Thermo Fisher Scientific UltiMate 3000 HPLC system coupled with a Q Exactive Focus™ Quadrupole-Orbitrap Mass Spectrometer (Thermo Scientific, Bremen, Germany). For chromatographic separation, a Thermo Acclaim™ (120 C18 3 μm 120 Å 2.1 × 100 mm) phase was used at 40°C with a multistep gradient (eluent A was 2 mM ammonium acetate buffer containing 0.1% formic acid and eluent B acetonitrile with 0.1% formic acid). Initially, 50% B was kept for 0.5 min, then raised to 95% until 9 min, kept for 2 min at 95% B, decreased to 50% B within 0.1 min and kept until 14 min for equilibration. A flow rate of 300 μL/min and an injection volume of 2 μL were used. Ionization was achieved with a heated electro spray (HESI) in switching mode. The source parameters were as follows: auxiliary gas flow rate, sheath gas flow rate, sweep gas flow rate: 11, 2 and 48 arbitrary units, respectively; auxiliary gas heater 413°C; capillary temperature 256°C; and spray voltage ± 3.5 kV. The mass spectrometer was operated in full MS-1 mode with a mass range from m/z 200 to 1000 and resolution of 70,000 (full width at half maximum (FWHM) at m/z 200). The automatic gain control (AGC) target was set to 1 × 10 6 and the maximum injection time to "auto". Data were recorded in profile data format.
For dd-MS-2 mode, the resolution was set to 35,000 (full width at half maximum (FWHM) at m/z 200), AGC target to 2 × 10 5 and max. injection time to "auto." Data were recorded in profile data format. The isolation window was set to 1.5 m/ z. For fragmentation, the normalized fragmentation energy was set to 40 eV.

In vitro analysis
The human liver S9 fraction assay generated phase I and phase II metabolites, and the positive paracetamol control was successfully glucuronidated. The negative control did not produce any metabolites as expected. The metabolites, detected with HPLC-HESI-HRMS in full scan MS-1 mode with a mass range from m/z 200 to 1000, are listed in Table 1. It shows HESI(+) data-sorted by calculated mass (m/z)-retention time (RT), chemical formula, adduct, calculated and found mass, and mass error in parts per million (ppm), and the three most abundant fragments of the detected metabolites. If available, the metabolites were verified with authentic standards. As identification criteria, we committed retention time errors below 0.1 min, mass errors less than 5 ppm, and consistent MS-2 spectra.    processed spectrum for database was used) and S9-assay with (−)-Δ-9-THC as substrate; weighted dot-product score: 0.918 (MassBank m = 2, n = 0.5 [27]) mass, and the fragment spectra) of the main peak at 3.4 min and the synthesized reference standard identified this as (−)-11-OH-THC, glucuronidated at the primary hydroxy group (hereafter referred to as alcoholic (−)-11-OH-Δ-9-THC-Glc). Figure 5 shows the MS-2 spectrum pairing, which yields a match of 91% [27]. Thus, the main peak at 3.4 min represents alcoholic (−)-11-OH-Δ-9-THC-Glc (1), synthesized in this study.

Investigation of OH-THC metabolites
The other metabolites G1-G4, shown in Fig. 4, are assumed to also represent OH-THC-Glc metabolites, as the exact mass fits to OH-THC-Gluc, and their HESI(+)-MS-2 spectra all show the initial neutral loss of the glucuronic acid which leads to a fragment with m/z 331.2276. A comparison with the MS-2 spectrum of 11-OH-Δ-9-THC-Glc encourages this through the presence of most of the other fragments, but with different relative intensities.

Investigation of glucuronide bounding
To determine whether the metabolites G1-G4 are alcoholic glucuronides (glucuronidation of the primary hydroxy group) or phenolic glucuronides (glucuronidation of the phenolic hydroxy group), we followed the approach described by Wen et al. [30]. They characterized phenolic and alcoholic structures in general and stated that in negative ionization mode, the presence of the fragments m/z 113, m/z 175, and m/z 193 is typical for alcoholic glucuronides, whereas the absence of fragment m/z 193 is typical for phenolic glucuronides (Fig. 7).

Analysis of authentic specimen
To confirm the in vitro results, first ten urine samples of cannabis users were analyzed by dilute and shoot preparation and HPLC-HRMS in full scan mode. Figure 10 shows the overlaid EICs of OH-THC-Glc (HESI positive m/z 507.2588) of the ten urine samples. Only in two of the ten Δ-9-THC-COOH positive tested urines that OH-THC-glucuronides were detectable. Comparing in vitro data with in vivo urine data, it is noticeable that the urine samples contain only the two main metabolites of a total of five previously detected OH-THC-Glc of the in vitro assay. Diglucuronide of OH-THC was again not observed. In contrast to the in vitro assay, G2 is formed in urine in a greater extent than (−)-11-OH-Δ-9-THC-Glc. As described before, G2 metabolite seems to be a phenolic glucuronidated OH-THC, but a clear identification of this signal is still outstanding due to a lack of reference standards. The second main metabolite was successfully identified as the alcoholic (−)-11-OH-Δ-9-THC-Glc, confirmed by our novel reference standard.
The first positive urine sample belongs to a daily consuming person of medicinal cannabis. The second positive urine is from a person for checking fitness to drive with a concentration of 5.5 ng/mL (−)-Δ-9-THC and (−)-Δ-9-THC-COOH of 54 ng/mL in serum. The other eight negative samples for OH-THC-glucuronide have (−)-Δ-9-THC-COOH concentrations in serum between not detectable (only detectable in urine up to 20 ng/mL (−)-Δ-9-THC-COOH) and 26 ng/mL (−)-Δ-9-THC-COOH. These data indicate that OH-THCglucuronides in urine only occur at higher concentrations of the phase I metabolites. This may be explained by the fact that 11-OH-THC is excreted mainly via the feces and less via the kidneys [23,31].

Conclusion
The investigation of phase II metabolism of 11-hydroxy-Δ-9tetrahydrocannabinol reveals two main OH-Δ-9-THC-glucuronides in vitro and in vivo-an alcoholic and a presumably Fig. 11 Extracted ion chromatograms of m/z 313.2162 (PRM mode, precursor ion m/z 507) in HESI(+) for OH-Δ-9-THC-Glc of two authentic serum samples extracted with a newly developed solid phase extraction method phenolic glucuronide. A double glucuronidation was not observed. The alcoholic (−)-11-OH-Δ-9-THC-Glc was successfully chemically synthesized and can now be used as reference standard. HPLC-HRMS data of this novel reference standard were successfully matched with the data of the in vitro and in vivo samples (urine/serum) and have thus confirmed the biotransformation of alcoholic (−)-11-OH-Δ-9-THC-Glc in vivo. The other main metabolite is assumed to be a phenolic glucuronide, due to detailed analysis of MS-2 spectra. Confirmation by synthesis of a reference standard is still pending.
The newly developed synthesis strategy of alcoholic (−)-11-OH-Δ-9-THC-Glc provides a simple and straightforward way for the synthesis as reference standard. Furthermore, the availability of a reference standard for alcoholic (−)-11-OH-Δ-9-THC-Glc offers the possibility for direct identification and quantification. After availability of the phenolic glucuronide besides the alcoholic glucuronide, it can be investigated, if there is a toxicogenetic influence, e.g., of polymorphic UGT 1A9 [32,33], on the site and rate of glucuronidation (alcoholic/phenolic) of the 11-OH-Δ-9-THC.