Chemical Analysis of Cannabinoids
Analytes and Methods
- Fig.1: Examples of the 4 most common methods for the separation and analysis of cannabinoids; A) Gas chromatography GC-FID chromatogram of TMS derivatized (with BSTFA) cannabinoids using Rxi-35SIL MS column (Restek, 15 m x 0.25 mm x 0.25 mm), H2 at 2.5 mL/min, oven temperature: 180°C (0.1 m), 25°C/min to 320°C, (https://blog.restek.com/?p=14961); B) HPLC-UV chromatogram with Kintex 2.6 mm C18 (Phenomenex, 50 mm x 2.1 mm) with 0.1 % formic acid in water – 0.1 formic acid in methanol (40:60 →15:85 in 10 min), 240 bar at 50 oC and 230 nm for UV detection (Phenomenex Applications TN-1225); C) HPTLC chromatogram with HPTLC Si 60 F254 plate (Merck, 20 x 10 cm) and cyclohexane-diisopropyl ether-diethyl amine (52:40:8, v/v/v) as a mobile phase, Fast Blue salt B as detection reagent (CAMAG Application note A-98.1); D) 1H NMR spectrum of a drug type cannabis extract together with 1 mg of anthracene as internal standard (* the signal of anthracene used for quantitation).
- Tab.: Assessments of the various tools for the analysis of cannabis preparations in various disciplines and for different materials of interest.
- Fig.2: Various Cannabinoids
In the past years the legal acceptance of medicinal cannabis and the gradual international acceptance of recreational cannabis, and last but not least, because of the belief in CBD-oil as a wonder drug, there is a high demand for proper methods of analysis for cannabis preparations.
With the introduction of gas chromatography some 50 years ago various chromatography-based methods as well as other methods have been proposed for the analysis of cannabis. There are numerous papers dealing with this analysis: GC (364), LC (407), TLC (29), CE (7), Mass spectrometry (748), NMR (76), NIR (3) and DNA barcoding (2) (Number of publications Web of Science, March 2019). So, what method to choose for your analysis? It all depends on the goal. What requirements and limitations do the methods have? Scheme 1 gives a general assessment of the various methods. Important characteristics are the quality of the data (e.g. ease of identification and absolute quantitation), the sensitivity, the (long term) reproducibility, ease of sample preparation, and costs. The costs are determined by the throughput per time, running costs and the investment needed. Below the different methods will be briefly discussed.
The plant Cannabis sativa is native to central Asia. It is a dioecious species belonging to the Cannabaceae family . Since the first uses probably already some 29,000 years ago as fiber plant to make nets  has developed into a widespread multipurpose plant: fibers for ropes and textiles, seeds and seed oil for food and oil lamps, resin from flowers and leaves for medicinal and narcotic purposes [1, 3]. Extensive breeding for the different uses and spreading of the cultivation over Asia and Europe resulted in a various diversification. This is causing some taxonomical difficulties. Small and Cronquist  (1976) simplified the taxonomy by distinguishing Cannabis sativa into two subspecies: C. sativa L. and C. sativa subsp. indica with low and high narcotic potential respectively. The problem is that nature is a continuum and to try to classify all plants in a simple classification will always lead to discussion [1, 4, 5].
A plant’s phenotype is dependent on many variables, like diurnal variation, plant development stage, climatic changes, pests and diseases, and interactions with microorganisms (e.g. in rhizosphere and endophytes). So even genetically identical plants may differ in phenotype, both in morphology and chemistry. Highly standardized growth conditions and strict quality control is required for a herbal drug like cannabis that is prescribed to vulnerable patients. As more than one active principle might be present in a herbal drug, a metabolomics like approach is required, with both quantitative and qualitative information.
Any plant species has a number of species-specific metabolites, the secondary metabolites. Cannabis is one of the best studied plants from the point of view of secondary metabolites. For example, through forensic studies [6, 7, 8]. The most important constituents are the cannabinoids, terpenoids, and flavonoids. The monoterpenoid pathways serves also as a source of the C10-unit that is coupled to a stilbene type of phenolic to yield cannabinoids [8, 9]. In the plant the cannabinoids are mostly present as acids, but these decompose by decarboxylation. For example, Δ9-THC acid during smoking gives Δ9-THC, that is responsible for the narcotic effects. In extraction one should thus avoid heating to be able to know the real content of plant. The Δ9-THC is also slowly converted to the more stable but less active Δ8-THC. Oxidation of Δ9-THC yields cannabinol. Various biological activities have reported for minor cannabinoids including the acids, but still little is known about their role in the medical uses of cannabis [8, 9]. In fact, CBD now is hot, as many beneficial effects are reported mostly based on casuistic observations, though for children’s epilepsy good evidence is available [10, 11]. Recently the FDA approved a CBD formulation . A formulation with THC and CBD (1:1) is on the market as medicine since 2010 . From the pharmacological point of view analysis of cannabinoids in body fluids is commonly applied. A number of metabolites of Δ9-THC have been identified over the years, with 11-hydroxy-Δ9-THC being the major and most psychoactive active one [13, 14].
Gas chromatography of cannabis was already described in 1963 . Since then the quality of the separation has increased every year, resulting in hundreds of papers [14, 15, 16]. The advantage of GC is that is a rather robust form of chromatography with good reproducibility. The GC-MS method  for plant metabolomics is a good example for that, and this method would certainly also be useful for the analysis of cannabis. However, the cannabinoids are rather unstable, and it isn’t possible to analyze the acid forms of the cannabinoids. For Δ9-THC GC is fine and could give a good insight in the expected activity. Also, the terpenes can be analyzed by GC, and many systems have been described for their GC-analysis [14, 15, 16]. Flavonoids, biosynthetically closely related to the cannabinoids, would not be detected by GC. The coupling with mass spectrometry offers a second dimension of separation and is thereby very useful for identification, particularly in case of having good fragmentation spectra. However, for absolute quantitation calibrations curves are needed for all compounds to be quantified. This requires high purity standard compounds. By using FID-detection the quantification of all compounds would be possible as similar compounds would give similar detector responses. The United Nations Office on Drugs and Crime  supports both types of detection. The most commonly used columns are of low polarity. Cannabinoids can be analyzed both, underivatized and derivatized. Underivatized cannabinoids showed slightly lower LODs compared to their silylated versions. The sample preparation for GC is in general more elaborate than for LC, and derivatization adds further challenges.
Considering the limitations of GC, the applications will particularly be in the field of the pharmacology where mostly the decarboxylated compounds are used. For the analysis of formulations with only decarboxylated cannabinoids and a fingerprint of decarboxylated plant material GC would also be suited.
The instability of the cannabinoids is the reason why LC is prevalent in the analysis of cannabis [7, 14, 15, 16]. The most common is the use of reverse phase columns (mainly C18 or biphenyl variants) with a solvent of methanol - water containing 0.1 % formic acid. LC-MS affords high sensitivity in combination with identification. But just like in GC-MS and the LC-UV the detection suffers from huge differences in detector response and consequently for quantitative analysis separate calibration curves have to be made for each compound that is to be quantified in absolute amounts. For relative quantitation the signal of each compound signal can be compared, and peak height or intensity can be used for relative quantitation. In LC-MS an ion trap is often used to obtain multistage fragmentation adding extra information for the identification. Compared with other MS detectors like quadruple (TQ) or TOF, there are obvious limitations in ion trap detectors such as low resolution and limited dynamic range. LC-MS is the most versatile of all the methods, both in terms of dynamic range and in terms of giving a real picture of the material analyzed. Though in all methods the preanalytical phase determines what you will see in the results of the analysis. The solvent used for the extraction is determining what compounds you will see. One should also be aware of the possibility that a major compound that is poorly soluble in the extraction solvent, will always show the same level, i.e. the level of its maximal solubility. In terms of analysis of the terpenes, LC-DAD isn’t able to detect most of the terpenes. Coupling to the MS is obligatory. The flavonoids can be analyzed very well in all detection methods coupled with the LC.
Thin layer chromatography (TLC) is the first method used for chemical analysis of cannabis. It has the advantage of parallel analyses, though with limitations in resolution and sensitivity. But low costs, simple sample preparation and non-destructive method are advantageous. Moreover, it allows the use of a wide array of chemical reagents for detection including in-situ bioassays. TLC has been developed into HPTLC, by improving the equipment allowing automatization and optimal control of chromatographic conditions and particularly for detection which include the automated desorption of spots from the plate and a direct interface from plate to MS. Currently, TLC is a low-cost method for cannabinoid analysis and approved by the United Nation Office on Drugs and Crimes  for routine control of cannabinoid content and of Cannabis origin. Obviously (HP)TLC is very useful in the laboratory for rapid screening for purity or cannabinoid contents in a sample [15, 16]. For applications that require high sensitivity, as e.g. in the pharmacology, it is not the method of choice. For fast fingerprinting of a series of samples it is quite useful.
To overcome the mentioned problems in absolute quantitation, nuclear magnetic resonance (NMR) spectroscopy offers the solution [20, 21]. In 1H NMR spectroscopy the signal intensity of a proton is only dependent on it molar concentration, that means with one internal standard every compound visible in the NMR can be quantified in absolute amounts. Hazekamp and colleagues [20, 21] reported on the application of 1HNMR to the quantitation of the cannabinoids THCA, THC, CBDA, CBD and CBN. The relative signal intensities of cannabinoids (H-4, H-10, or H-9) to that of internal standard (anthracene) were used to calculate the mole concentrations of the cannabinoids in cannabis.
Only a few papers report capillary electrophoresis and supercritical chromatography . For identification of plant material NIR-spectroscopy  and DNA barcoding  could be used, but these methods will not give much information on the levels of cannabinoids. The 12C:13C ratio can tell something about the latitude (North/South) a plant has been grown which could be of interest for quality and forensic aspects. For the isolation of single pure cannabinoids counter current chromatography has been reported . Additionally, supercritical extraction is an elegant way to get extracts with high levels of the cannabinoids.
Conclusion and Perspectives
There is an enormous amount of literature available on the analysis of cannabis. What is missing are standard protocols, for example in metabolomics. By using a metabolomics approach, all the results can be put together and by using biostatistical methods known compounds can readily be identified and new ones can be easily recognized. With this approach a (public) database can be build up for long term use.
For each application there have been good methods reported. However, from the field we learn that in different labs different results are obtained for the same sample. Obvious causes are the preanalytical processing and the (non) availability of high-quality standards of the major compounds to be detected. Only NMR can do quantitation without the need of a 100 % pure standard, and in fact could be used as the golden standard for a mixture of cannabinoids that then is used for making calibration curves for chromatography-based methods.
Young Hae Choi1, Robert Verpoorte1
1Natural Products Laboratory, Institute of Biology, Leiden University, Leiden, The Netherlands
Prof. Dr. Rob Verpoorte
Natural Products Laboratory, IBL, Leiden University
Leiden, The Netherlands
1 . E. Small, Evolution and Classification of Cannabis sativa (Marijuana, Hemp) in Relation to Human Utilization Bot. Rev.: 81(2015) 189-294; doi:10.1007/s12229-015-9157-3
2 . H. Pringle, H., Ice age communities may be earliest known net hunters. Science 1997, 277, 1203–1204. DOI: 10.1126/science.277.5330.1203
3 . S.A. Bonini, M. Premoli, S. Tambaro. A. Kumar, G. Maccarinelli, M.Memo, and A. Mastinu. Cannabis sativa: A comprehensive ethnopharmacological review of a medicinal plant with a long history J. Ethnopharmacol. 227(2018) 300-315; doi.org/ doi:10.1016/j.jep.2018.09.004
4 . Small, E., Cronquist, A. 1976. A practical and natural taxonomy for Cannabis. Taxon, 25, 405-435. DOI:10.2307/1220524
5 . J.M. McPartland, and G.W. Guy. Models of Cannabis Taxonomy, Cultural Bias, and Conflicts between Scientific and Vernacular Names. Bot. Rev. 83 (2013) 327-381. doi:10.1007/s12229-017-9187-0
6 . A.Hazekamp, J.T. Fischedick, M.L. Diez, A. Lubbe, and R.L. Ruhaak. Chemistry of Cannabis. Comprehensive Natural Products II Chemistry and Biology; L. Mander and H.-W Lui (Editors). Elsevier, Oxford, 2010. Volume 3, pp 1033-1084; ISBN: 978-0-08-045382-8
7 . B.F. Thomas and M.A. ElSohly, The Analytical Chemistry of Cannabis. Elsevier, Amsterdam 2016, ISBN: 9780128046463
8 . C.M. Andre, J.-F. Hausman, G. Guerriero. Cannabis sativa: the plant of the thousand and one molecules. Front. Plant Sci., 7 (2016) 1-17. doi:10.3389/fpls.2016.00019
9 . I.J. Flores-Sanchez, and R. Verpoorte. Secondary metabolism in Cannabis. Phytochem. Rev. 7(2008), 615–639. doi.10.1007/s11101-008-9094-4
10 . J. Elliott, D. DeJean, T. Clifford, D. Coyle, B.K. Potter, B. Skidmore, C. Alexander, A.E. Repetski, V. Shukla, B. McCoy, and G.A. Wells GA. Cannabis-based products for pediatric epilepsy: A systematic review. Epilepsia 60 (2019) 6-19;DOI:10.1111/epi.14608
11 . S. Lattanzi, F. Brigo, E. Trinka, G. Zaccara, C. Cagnetti, C. Del Giovane, M. Silvestrini. Efficacy and safety of cannabidiol in epilepsy: a systematic review and meta-analysis Drugs 78 (2018) 1791-1804;DOI:10.1007/s40265-018-0992-5
12 . Website: https://www.gwpharm.com/ assessed March 10, 2019
13 . M.E. Wall, and M. Perez-Reyes, The metabolism of delta 9-tetrahydrocannabinol and related cannabinoids in man. J. Clin. Pharmacol. 1981, 21, 178s–189s. doi.org/10.1002/j.1552-4604.1981.tb02594.x
14 . A.Leghissa, Z.L. Hildenbrand,and K.A. Schug. A review of methods for the chemical characterization of cannabis natural products. J. Sep. Sci.: 41 (2018) 398-415;DOI:10.1002/jssc.201701003
15 . T.J. Raharjo and R. Verpoorte. Methods for cannabinoids analysis in biological materials. Phytochem. Anal. 15(2004)79-94;DOI:10.1002/pca.753
16 . C. Citti, D. Braghiroli, M.A. Vandelli, and G Cannazza. Pharmaceutical and biomedical analysis of cannabinoids: A critical review. J Pharm Biomed Anal. 147(2018)565-579; doi:10.1016/j.jpba.2017.06.003
17 . T.W.M. Davis, C.G. Farmilo, and M. Osadchuk. Identification and origin eterminations of cannabis by gas and paper chromatography. Anal Chem: 35(1963) 751-755; DOI:10.1021/ac60199a020
18 . O. Fiehn, J. Kopka, P. Dormann, T. Altmann, R.N. Trethewey, and L. Willmitzer, Metabolite profiling for plant functional genomics. Nat Biotechnol 18 (2000)1157-1161; doi:10.1038/81137
19 . Laboratory and Scientific Section UNODC, Recommended Methods for the Identification and Analysis of Cannabis and Cannabis Products, United Nations, publications 2009, p.36-39.
20 . Hazekamp, A., Choi, Y.H., Verpoorte, R., 2004. Quantitative analysis of cannabinoids from Cannabis sativa using 1H-NMR. Chem. Pharm. Bull. 52, 718-721.
21 . Y.H. Choi, H.K. Kim, A. Hazekamp, C. Erkelens, A.W.M. Lefeber and R. Verpoorte. Metabolomic Differentiation of Cannabis sativa Cultivars Using 1H-NMR and Principal Component Analysis. J. Nat Prod. 67(2004)953-957;DOI:10.1021/np049919c
22 . C. Sánchez-Carnerero, N. Núñez-Sánchez, S. Casanoa, and C. Ferreiro-Vera. The potential of near infrared spectroscopy to estimate the content of cannabinoids in Cannabis sativa L.: A comparative study. Talanta 190(2018)147-157; doi:10.1016/j.talanta.2018.07.085
23 . D. Vergara, H. Baker, K. Clancy, K.G. Keepers, J.P.; Mendieta, C.S. Pauli, S.B. Tittes K.H. White, N.C. Kane. Genetic and Genomic Tools for Cannabis sativa. Crit. Rev. Plant Sci.: 35 (2016) 364-377; doi:10.1080/07352689.2016.1267496
24 . A. Hazekamp, R. Simons, A.M.G. Peltenburg-Looman, M. Sengers, R. van Zweden and R. Verpoorte. Preparative isolation of cannabinoids from Cannabis sativa by centrifugal partitioning chromatography. J. Liq. Chrom. & Rel. Technol. 27(2004)2421-2440; doi:10.1081/JLC-200028170