Comparative Metabolomic Profiling of the Metabolic Differences of Δ9-Tetrahydrocannabinol and Cannabidiol

2.2. THC and CBD Metabolites in HLM and MLM Profiles

A total of 52 phase I THC and CBD metabolites were identified. Twenty-eight metabolites in phase I were derived from THC, including 21 in HLM, and 24 in MLM (Table 1). Between the HLM and MLM profiles, in terms of THC metabolites, 17 metabolites (T1 to T12, T22 and T23, and T25 to T27) were found in two different species. The rest were found to be unique to three metabolites (T13, T14, and T17) in the HLM profile, and seven metabolites (T15, 16, T18 to T21, and T24) in the MLM profile.

Twenty-four CBD metabolites were identified in the HLM and MLM profiles, including 14 metabolites in the HLM profile and 21 in the MLM profile (Table S1, Supplementary Materials). C1 and C2, C8–C10, C12–C15, C18, C20, and C21 were consistently found between the two systems, whereas C3, C7, and C23 were only observed in the HLM system.

The details of the MS/MS of THC and CBD are shown in Figure S1. The formation of desaturated THC and CBD metabolites has not been described in previous studies. THC and CBD were desaturated and di-desaturated, which generated T1 and T2, and T3 to T7 metabolites, and C1 to C3, and C4 to C7 metabolites, respectively. The chemical formula of metabolites T1 and T2, C1 to C3 was C₂₁H₂₈O₂, in accordance with the observed m/z [M + H]⁺ in 313.2129–313.2174; these values which were 2 Da (2H) lower than that of the THC and CBD parent compounds, thus indicating that these metabolites were desaturation products. The retention time of C7 was 7.85 min, thus suggesting that it was derived from the cyclization of the methyl group (C10) and the nearby hydroxyl group at C5. Other details concerning the identification and structural elucidation of these THC and CBD metabolites are provided in Figures S2 and S3 (Supplementary Materials).

2.3. Identification of GSH and Cys Conjugates of THC and CBD

Four additional metabolites (T29–T32) that metabolize THC, and three additional metabolites (C25–C27) that metabolize CBD, were identified in the MLM and HLM profiles, and they were treated with GSH or Cys. GSH and Cys, conjugated with THC, were not dependent on NAPDH (Table 1), whereas Cys conjugated with CBD was provided by NAPDH (Table S1, Supplementary Materials). Five of these (T29, T31, and T32, and C26 and C27) were novel metabolites that have not been previously reported. The chromatogram, MS/MS spectrum, and fragmentation pathways of the representative THC and CBD metabolites are presented in Figure 3 and Figure 4.

The THC metabolites, T29 and T31, were assigned the molecular formulas C₃₁H₄₃N₃O₈S and C₃₁H₄₃N₃O₉S, based on the m/z ratio of the [M + H]⁺ ion at 618.2827⁺ and 634.2780⁺, respectively. The elution times of T29 and T30 were 4.53 and 4.48 min, respectively. T29 was identified as a novel GSH-conjugated metabolite, as confirmed by the presence of the MS/MS fragment and the [M + H]⁺ ion at 311.1997⁺; this was equivalent to the elimination of the GSH adduct from the parent ion. Metabolites of T31 and T32 were assigned the molecular formulas C₂₄H₃₃NO₄S and C₂₄H₃₃NO₅S, and they were eluted at 4.87 and 4.83 min, based on the m/z ratio of the protonated molecule at 432.2223⁺ and 448.2141⁺, respectively. The MS/MS spectrum of T31 also produced a major fragment with a m/z = 311.1994⁺, which corresponded with the loss of Cys adduct from the parent ion. The GSH/Cys adduct can be bound to the C2 or C4 position of the benzene ring of the THC metabolites. The chemical formula of the metabolite C25 was C₃₁H₄₃N₃O₉S, according to the [M + H]⁺ at m/z 634.2788⁺; this value was 319 Da (GSH+O-H₂) higher than that of the CBD parent compound, thus indicating that it was a GSH-conjugated CBD metabolite in a HLM/MLM incubation system. C26 and C27 were assigned the formula C₂₄H₃₃NO₄S, and they produced a major fragment with a m/z 311.2007⁺, which is representative of the loss of the Cys adduct from the parent ion. The positions of Cys were unable to be determined.

2.4. Molecular Docking Prediction of THC and CBD Metabolites

A software of Discovery Studio 2020 docking simulation was utilized to determine whether THC metabolites could activate a CB1 receptor. This tool allowed the tension between the binding energy and the active site to be resolved. The score of the -CDOCKER energy, -CDOCKER INTERACTION energy, and binding energy were used to display the molecular docking appearance. A higher -CDOCKER INTERACTION energy score indicated better binding. A negative binding energy indicated a more stable binding. When the -CDOCKER energy score was negative, it indicated a steric hindrance between receptor and ligand. THC (T0) and CBD (C0) generated a -CDOCKER INTERACTION energy score of 47.84 and 45.16, respectively (Table S5). The -CDOCKER energy was also found to be different between T0 (8.27) and C0 (−1.98). Twenty-three of the THC metabolites (T2, T7–21, and T23–32) had a higher affinity for the CB1 receptor than the THC parent compound, as shown the higher -CDOCKER INTERACTION energy score. These THC metabolites also displayed positive values for energy scores, thus suggesting that they were capable of docking into the CB1 binding pocket with minimal steric hindrance. Moreover, C2, C8, C15, C19, C23, and C24 had a higher -CDOCKER INTERACTION energy score than THC, but their CDOCKER energy scores were negative or close to zero, except for C23, thus suggesting the poor binding of these CBD metabolites due to steric hindrance.

Out of all the THC metabolites, T29–T32 were shown to be the top four metabolites with the highest -CDOCKER INTERACTION energy score (T29 = 81.27, T30 = 69.35, T31 = 55.09, and T32 = 57.20), and thus they represent the best binding modes that anchored at the active site and interacted with the key residues. On the other hand, the binding energy of T29 (THC+GSH-H₂) was −130.96, for T30 (THC+GSH+O-H₂) it was −17.62, for T31 (THC+Cys-H₂) it was −47.54, and for T32 (THC+Cys+O-H₂) it was −74.59, thus suggesting that T29 and T32 could be well docked into the binding pocket of CB1.

The binding pocket of CB1 attracts THC-bound amino acid residues comprising Ala380, Ile-105, Ser-383, Met-103, Cys-386, Trp-356, Val-196, Phe-268, and Phe-108, as shown in Figure S4A (Supplementary Materials). This binding pocket also interacted with CBD through the formation of one hydrogen bond with Ser-383, six π-alkyl bonds with Cys-386, Leu-387, Phe-108, Pro-269, Ala-380, and Ile-105, and one π–π bond with Phe-268, as shown in Figure S4B. As the novel THC metabolite T29 (GSH-conjugated metabolite) had the highest affinity with the receptor, and was most stably bound to the receptor compared with the other two novel metabolites, T30–T31 (Cys-conjugated metabolites which had a similar affinity for CB1), we examined the ligand–receptor interaction more closely using T29 and T31. THC conjugated with GSH (T29) interacted with the binding pocket through the formation of two hydrogen bonds with Ser-383 and Cys-386, one π–π bond with Phe-108, two π-sulfur bonds with Met-384 and Phe-170, and seven π-Alkyl bonds with Cys-107, Phe-268, Phe-102, Ala-380, Met-103, Phe-174, and His-178 (Figure 5A). Such interactions facilitated the proper positioning of the compound in the active site. Many residues also participated in the molecular docking of T31 (Figure 5B), including Met-384 (H-bond), Ile-105, Met-103, Phe-102, Phe-268, Phe-379, Trp-356, Cys-386, and Val-196 (π-alkyl bond).

4.2. Metabolism of THC and CBD in the MLM and HLM Systems

In vitro metabolism of THC and CBD was conducted in a 96 wells plate, and each incubation was conducted in triplicate [26]. After a series of pre-experimental tests, the final concentrations of GSH and CBD, at 0.2, 1, 5, and 25 μM, were tested with HLM co-incubation. The incubation system (180 μL) contained 25 μM THC or 25 μM CBD, and the final concentration of the HLM or MLM was placed in a 0.5 mg protein/mL phosphate-buffered saline solution (PBS, pH = 7.4). After pre-incubation for 5 min at 37 °C, with shaking at 400 rpm, this system was fortified with 20 μL of 10 mM NADPH, and it continued for 40 min at 37 °C, with shaking at 400 rpm in a MicBio Ⅱ incubator (Abson, USA). Reactions were terminated with the addition of 200 μL of ice-cold acetonitrile. The mixture was vortexed for 60 s, followed by centrifugation at 16,200× g for 20 min at 4 °C. Two control groups were pre-prepared using the same protocol, although the incubation system was not treated with NADPH or drugs, respectively. A 5 μL aliquot of the supernatant was collected for Thermo Scientific™ Vanquish™ flex ultra-high-performance liquid chromatography Q Exactive^TM MS analysis (UHPLC Q-Exactive).

4.3. Trapping Reactive Metabolites using GSH and Cys

To investigate the potential of THC and CBD when forming electrophilic metabolites in the presence of nucleophiles, either 400 μM GSH or 400 μM Cys was added to MLM/HLM treated with THC or CBD, as described in the previous section [5]. Five μL of supernatants were collected for UHPLC-Q-Exactive MS analysis.

4.4. Metabolism of THC and CBD in Human Recombinant CYP450s

In order to understand the contribution of different CYPs in the metabolism of THC and CBD, a 25 μM substrate (THC or CBD) was mixed with 2 pmol/mL of each of the cDNA-expressed P450 enzymes (control, CYP1A1, CYP1A2, CYP1B1, CYP2A6, CYP2B6, CYP2C19, CYP2C8, CYP2C9, CYP2D6, CYP2E1, CYP3A4, CYP3A5, CYP4A11) to comprise a final volume of 180 μL. The reaction was initiated with the addition of 20 μL NADPH (10 mM) and it stopped after 40 min with the addition of 200 μL of ice-cold acetonitrile. The solution was centrifuged at 16,200× g for 20 min at 4 °C, and a 5 μL aliquot of the supernatant was collected for UHPLC Q-Exactive MS analysis.

To investigate the contribution of CYPs, and to identify the CYPs responsible for the formation of the GSH- or Cys-conjugated products of THC, inhibitor assays were carried out in a THC-treated MLM system in the presence of 400 μM GSH/Cys. The final concentrations of each chemical inhibitor, and the corresponding CYP target, were shown as follows [5]: α-naphthoflavone (1.0 μM; CYP1A1/2 inhibitor), ketoconazole (100 μM; CYP3A4 inhibitor), ticlopidine (100 μM; CYP2C19 and CYP2B6 inhibitor), methoxsalen (20 μM; CYP2A6/13 inhibitor), trimethoprim (2.5 μM; CYP2C8 inhibitor), sulfaphenazole (100 μM; CYP2C9 inhibitor), quinidine (5.0 μM; CYP2D6 inhibitor), and resveratrol (10 μM; CYP1B1 inhibitor). The control group was carried out in the same vehicle, but in the absence of inhibitors. The reaction mixtures were collected for UHPLC Q-Exactive MS analysis to determine whether THC conjugated with GSH/Cys.

4.5. UHPLC Q-Exactive MS Analysis

The analysis of the microsomes was conducted using the UHPLC Q-Exactive (Thermo Fisher Scientific, San Jose, CA, USA) system, and analyte separations were achieved using a HSS T3–C18 column (2.1 mm × 100 mm, 1.8 μm) at 25 °C. The analytes were eluted with a gradient elution of 0.1% formic acid in water (A) and acetonitrile (B). The gradient procedure was performed as follows: 1% B for 0–1 min, 1–70% B for 1–3 min, 70–100% B for 3–15 min. 100–1% B for 15–18 min. Other conditions were set as follows: flow rate: 0.3 mL/min; capillary temperature: 350 °C; gas flow: 15 L/min; sheath gas flow rate: 35 L/min; capillary voltage: 3.7 kV; injection volume: 5 μL. The mass spectrometer was operated in a positive ionization mode. The MS/MS fragment information was detected using the following MS/MS conditions: the collision energy was set at 15, 25 and 35 eV, respectively; the orbitrap resolution was set to 70,000; Xcalibur software (Version 4.3, Thermo Fisher Scientific, San Jose, CA, USA) was used to control the LC-HRMS system and for data acquisition [26].

4.6. Data Processing, Feature Selection, and Metabolite Annotation

Thermo Xcalibur Qual Browser software (Thermo Fisher Scientific, San Jose, CA, USA) was used to detect THC and CBD metabolites. Chromatographic analysis and data processing was performed using MS-DIAL software. The parameters were set as follows: retention time range = 0–15 min, minimum peak height = 10,000, and the parent ions of [M + H]⁺ were selected for screening metabolites when the software was set to the positive mode. The processed data matrix was exported as an excel file containing the peak number, retention time, m/z, and integrated peak area. Unsupervised principal analysis (PCA) was applied to select discriminatory features (SIMCA v.14.1 software (Kinnelon, NJ, USA). The selected features were searched against the MS/MS analysis to identify the THC and CBD metabolites. Finally, the structures of these tentatively assigned features were analyzed and confirmed using MS/MS fragmentation patterns.

The structures of THC and CBD metabolites were drawn using ChemBioDraw Ultra v.13.0 software (Cambridge Soft Corporation, Cambridge, MA, USA), and the Clog P values of THC and CBD metabolites were calculated using the chemical properties window. Clog P is a potential indicator that reflects the polarity of a compound. A compound with a low Clog P value exhibits high polarity, which corresponds to an early elution time in reversed-phase chromatography [27]. The isomeric metabolites of THC and CBD can therefore be distinguished from one another through the use of the retention time and Clog P value.

4.7. Molecular Docking Prediction Concerning How the CB1 Receptor Interacts with Metabolites

To further understand the relationship between the metabolism of THC and drug addiction, a molecular docking simulation was utilized to understand the mechanism of interaction between the CB1 receptor and THC and CBD metabolites. The crystalline structure of the human CB1 receptor was confirmed in a previous study, which also depicted a model of the CB1 active site based on X-ray diffraction [28]. Molecular docking was simulated using the Discovery Studio 2020 in order to visually characterize the affinity between metabolites and the CB1 receptor. The standard three-dimensional structure of metabolites (*mol) as ligands was established by Chemdraw 3D v.19.0 software (Cambridge, MA, USA). The receptor structure of CB1 (PDB ID:5U09) was downloaded from the RCSB Protein Data Bank (https://www.rcsb.org/) (accessed on 14 October 2021) and imported into the Discovery Studio 2020 (Accelrys Software Inc., San Diego, CA, USA). The structure of the CB1 receptor was modified by removing water and adding hydrogen. The partially flexible CDOCKER program was used to determine special binding sites and the receptor radius. The molecular docking results were analyzed to determine the -CDOCKER energy score, -CDOCKER INTERATION energy, and binding energy.

We would like to thank the Xin Chuan Mass Spectrometry Center of West China Hospital for the MS tests.