Synthesis of [13C4]-labeled ∆9-Tetrahydrocannabinol and 11-nor-9-Carboxy-∆9-tetrahydrocannabinol as Internal Standards for Reducing Ion Suppressing/Alteration Effects in LC/MS-MS Quantification

(−)-∆9-Tetrahydrocannabinol is the principal psychoactive component of the cannabis plant and also the active ingredient in some prescribed drugs. To detect and control misuse and monitor administration in clinical settings, reference samples of the native drugs and their metabolites are needed. The accuracy of liquid chromatography/mass spectrometric quantification of drugs in biological samples depends among others on ion suppressing/alteration effects. Especially, 13C-labeled drug analogues are useful for minimzing such interferences. Thus, to provide internal standards for more accurate quantification and for identification purpose, synthesis of [13C4]-∆9-tetrahydro-cannabinol and [13C4]-11-nor-9-carboxy-∆9-tetrahydrocannabinol was developed via [13C4]-olivetol. Starting from [13C4]-olivetol the synthesis of [13C4]-11-nor-9-carboxy-∆9-tetrahydrocannabinol was shortened from three to two steps by employing nitromethane as a co-solvent in condensation with (+)-apoverbenone.


Introduction
(−)-∆ 9 -Tetrahydrocannabinol (∆ 9 -THC, 1), is a product of the female flowering parts of Cannabis sativa (marijuana), and is the main psychoactive substance in the plant. As compound 1 and analogues act on the cannabinoid receptors, the cannabinoid group of compounds is also medicinally useful. Among others, development have led to the ∆ 9 -THC-containing drug Sativex ® , indicated for treatment of moderate to severe spasticity due to multiple sclerosis [1], and Marinol ® , used in the treatment of chemotherapy-induced nausea and vomiting [2]. The ∆ 9 -THC analogue Nabilone™, also an approved drug, is indicated efficient in treatment of the same conditions [2,3]. However, recreational use and continued illicit use of marijuana have increased the importance of having methods to determine usage by individuals.
In vivo ∆ 9 -THC (1) undergoes phase I metabolism to yield 11-hydroxy-∆ 9 -tetrahydrocannabinol (11-OH-THC, 2) peaking immediately after smoking, Scheme 1. Metabolite 2 is also psychoactive, but is rapidly oxidized to the inactive metabolite (-)-11-nor-9-carboxy-∆ 9 -tetrahydrocannabinol (THC-COOH, 3), which slowly increases and plateaus after 2-4 h [4]. Metabolism occurs mainly in the liver by cytochrome P450 enzymes 2C9, 2C19, and 3A4 [5]. The main urinary metabolites occur as phase II conjugates of glucuronic acid, and less commonly as sulphate, glutathione, amino acids, and fatty acid conjugates. The main site for glucuronidation is the carboxylate at C-11, but 11-OH-THC (2) may undergo phase II metabolism as well. Scheme 1. Major phase I metabolites derived from ∆ 9 -THC (1). Efficient analysis of both drugs and their metabolites are required in the fields of pharmacology, clinical toxicology and forensic toxicology, but also for workplace drug testing, testing of driving under the influence of drugs, doping analysis and rehabilitation programs. Since these analyses are often performed by mass spectroscopic techniques, reference samples of the native drug, its main metabolites and suitable standards are needed for identification and quantification purposes. Stable isotope labeled internal standards (SIL IS) are added to correct for error in the sample preparation and co-eluting substances that could alter or suppress the signal [6]. Deuterated internal standards with varying degree of labeling are currently used as IS in quantification of cannabinoids [7][8][9][10][11].
However, under certain conditions the MS ionization of the analytes is influenced differently from that of its deuterium labeled analogues, due to slight differences in retention. This potentially leads to inaccuracies in the quantification [12,13]. Urine analysis of ∆ 9 -THC (1) and its metabolites by LC-MS/MS is especially challenging, due to the presence of unstable compounds and strong adsorption to hydrophobic surfaces and the need for enzymatic pre-processing of the sample prior to analysis. Further, the importance of an effective sample clean-up prior to cannabinoid analysis to remove matrix interferences and maintaining a high extraction efficiency has been highlighted [14]. Also, Scheidweiler et al. [15] concluded that urine samples from different individuals gave matrix effects not observed during method validation, contributing to inaccurate cannabinoid quantification. It is also likely that the challenge with ion suppressing/alteration or matrix effects will be even more pronounced by the use of high resolution techniques such as UHPLC-MS/MS. This indicates that there is a need for a better SIL IS for cannabinoid analysis. 13 C-Labeled IS are particularly suitable for minimizing ion suppression/alteration effects in LC-MS/MS analysis, therefore the quantitative analysis in various biological samples are particularly accurate and reproducible. However, to avoid "overlap" with the natural 13 C in the native compound in MS detection, the number of labeled atoms must preferably be at least three. Recently, the success of substituting deuterated IS with 13 C IS has been shown in the case of amphetamine and methamphetamine quantification [16]. The use of 13 C as labeled compounds as IS is also widely known and appreciated in other fields of analytical chemistry [17][18][19][20][21]. Based on this background we disclose herein a synthesis of [ 13 C 4 ]-labeled Δ 9 -THC (1) and Δ 9 -THC-COOH (3) made via [ 13 C 4 ]-olivetol.
A key intermediate was assumed to be protected 5-(bromomethyl)benzene-1,3-diol derivatives, II, which can be converted to olivetol analogues (I) by metal catalysed coupling [29] or olefination chemistry [30] using IV or V. Another precursor is [ 13 C 6 ]-3,5-dihydroxybenzoic acid (VII), available by synthesis from [ 13 C 6 ]-benzoic acid [31]. Depending on the following strategy (olefination or C-C coupling) additional steps are needed. As the benzoic acid routes appeared long with expected lower 13 C atom efficiency, we decided to prepare an olivetol derivative with four carbons in the alkyl chain by using commercially available [ 13 C 4 ]-n-bromobutane. The chemistry performed is shown in Scheme 3.
A Wurtz type reaction between 1-(bromomethyl)-3,5-dimethoxy-benzene (4) and the Grignard reagent made from [ 13 C 4 ]-n-bromobutane in the presence of dilithium tetrachlorocuprate gave 78% isolated yield of [ 13 C 4 ]-1-(3,5-dimethoxyphenyl)pentane (5). The major by-product observed was the 3,5-dimethoxybenzyl dimer. By-product formation was somewhat increased by running the reaction in tetrahydrofuran instead of diethyl ether. To obtain [ 13 C 4 ]-olivetol (6) the dimethoxy ether groups were removed by heating compound 5 in pyridine hydrochloride at 200 °C. Trimethylsilyl iodide could also be used, although purification was more tedious in this case. The Wittig strategy using butyltriphenylphosphonium salt in combination with the non-labeled benzaldehyde (V) was also briefly tested. The first step involving formation of the Wittig salt was slow in toluene and needed several days at reflux to reach full conversion. On the other hand both the alkene forming step, and the hydrogenation of the olefin with palladium on carbon as catalyst proceeded smoothly. Thus, by tuning of the first step, this route might be highly useful.

[ 13 C 4 ]-Labeled Δ 9 -THC and Δ 9 -THC-COOH
In the synthesis of [ 13 C 4 ]-Δ 9 -THC (1), [ 13 C 4 ]-olivetol (6) was simply condensed with the terpene 7 in the presence of boron trifluoride as Lewis acid following the procedure of Silverberg et al. [25], Scheme 4. Alternative acid catalysts for such transformations has been published by Rosati et al. [32]. After purification by preparative HPLC the product was isolated in 61% yield as a colourless oil (purity >99%). The product gradually darkened upon storage at −20 °C, however, no change in chromatographic purity was noticed.
An alternative procedure towards 8 that was previously used in the synthesis of the structurally related compound Nabilone™ [34] was also investigated. By this one-step reaction between (1R,5R)-6,6-dimethylbicyclo[3.1.1]hept-3-en-2-one ((+)-apoverbenone, 11) and 5-(1,1-dimethylheptyl)resorcinol in the presence of aluminum chloride, a low 16% yield was reported [34]. However, we were able to increase the yield in this transformation up to 67% by the use of nitromethane as co-solvent in the condensation between [ 13 C 4 ]-olivetol (6) and 11. Although the exact role of nitromethane has not been investigated it is assumed that the improvement in part is due to increased solubility and mixing of aluminum chloride. Nitromethane has also been reported to modify/reduce the activity of the aluminum chloride, which might also be an important aspect [35]. It should be noted that both of the terpenes 9 and 11 in the end yielded the natural enantiomeric form of (-)-Δ 9 -THC-COOH (3) in contrast to other possible strategies [36].
The precursor 8 was then transformed to [ 13 C 4 ]-Δ 9 -THC-COOH (3) in three operations. Treatment with triisopropylbenzenehydrazine yielded the corresponding hydrazone, which was carefully dried prior to treatment with butyl lithium in n-hexane in the presence of tetramethylethylenediamine (TMEDA) and then gaseous CO 2 . Scrupulously dry and pure reagents were needed in this transformation. Kachensky et al. [37] reported on a 9/1 ratio of Δ 9 /Δ 8 THC-COOH by using a 10% solution of TMEDA. The dilution factor was however not mentioned. Following these conditions, only the Δ 8 -isomer was obtained in our earlier test reactions. This is in line with that observed by Nikas et al. [33]. Decreasing the concentration by fifty percent by addition of more of the TMEDA/n-hexane solvent mixture had a positive effect. The crude product contained a 6/4 ratio of [ 13 C 4 ]-3 and the  8 -regioisomers, which corresponds with that reported by Nikas et al. [32]. Purification was done first by silica-gel column chromatography to remove structurally unrelated impurities, followed by two crystallisations to arrive at a final purity of 97.5%. A chromatogram of the prepared material as compared to the native substance is shown in Figure 2. Preparative HPLC was used to recover additional [ 13 C 4 ]-Δ 9 -THC-COOH (3) from the mother liquor giving a total yield of 18%. The enantiomeric excess of Δ 9 -THC-COOH (3) was determined by self-induced non-equivalence by 1 H-NMR spectroscopy of ketone intermediate [ 13 C 4 ]-8 [38] and was found to be 96% ee.

Chemicals and Analysis
Bulk solvents were purchased either from LabScan (Gliwice, Poland) or Merck (Darmstadt, Germany). Deuterated solvents were purchased from CDN Isotopes Inc (Pointe-Claire, QC, Canada). All chemicals or reagents used were of highest purity available and purchased from Sigma-Aldrich (Oslo, Norway) or Acros (Geel, Belgium). All solvents and chemicals were used as is without further purification unless otherwise stated. Anhydrous solvents were used as is and stored over activated molecular sieves. The silica-gel used for flash chromatography was Merck silica gel 60 (230-400 mesh). For chromatography, thin layer chromatography (TLC) silica gel 60F 254 Merck plates/sheets were employed with visualization under UV light at 254 nm. 1 H and 13 C-NMR spectra were recorded from Bruker Advance DPX instruments (400/100 MHz). Chemical shifts (δ) are reported in ppm rel. to tetramethylsilane. Due to the high intensity of the 13 C-labeled carbons as compared to those unlabeled, and multiple coupling, some NMR resonances were not detected. In this study the non-labeled benzoic carbonyl signal at around 166 ppm was difficult to detect, experiencing extensive coupling by the neighbouring 13 C-isotopes. Isotopic purity and accurate mass determination in positive and negative mode on the final product was performed on a "Synapt G2-S" Q-TOF instrument from Waters with a resolution of 5 ppm. Samples were ionized by the use of an atmospheric pressure solids analysis probe (ASAP). No chromatography separation was used prior to the mass analysis. HPLC analysis was performed on an Agilent 1200 with atmospheric pressure chemical ionization (APCI)/ electrospray ionization (ESI) multimode ionization. Acetonitrile was used as the mobile phase in combination with water buffered at pH 3 with formic acid. Column used was XBridge™ C18, 5 µm, 4.6 × 150 mm from Waters or Hilic Plus, 3.