Content uploaded by Justin Thomas Fischedick
Author content
All content in this area was uploaded by Justin Thomas Fischedick on Jan 08, 2015
Content may be subject to copyright.
Cannabis - from cultivar to chemovar
A. Hazekamp
a
* and J. T. Fischedick
b
The medicinal use of Cannabis is increasing as countries worldwide are setting up official programs to provide patients with access to
safe sources of medicinal-grade Cannabis. An important question that remains to be answered is which of the many varieties of
Cannabis should be made available for medicinal use. Drug varieties of Cannabis are commonly distinguished through the use of
popular names, with a major distinction being made between Indica and Sativa types. Although more than 700 different cultivars
have already been described, it is unclear whether such classification reflects any relevant differences in chemical composition. Some
attempts have been made to classify Cannabis varieties based on chemical composition, but they have mainly been useful for
forensic applications, distinguishing drug varieties, with high THC content, from the non-drug hemp varieties. The biologically active
terpenoids have not been included in these approaches. For a clearer understanding of the medicinal properties of the Cannabis
plant, a better classification system, based on a range of potentially active constituents, is needed. The cannabinoids and terpenoids,
present in high concentrations in Cannabis flowers, are the main candidates. In this study, we compared cultivars obtained from
multiple sources. Based on the analysis of 28 major compounds present in these samples, followed by principal component
analysis (PCA) of the quantitative data, we were able to identify the Cannabis constituents that defined the samples into distinct
chemovar groups. The study indicates the usefulness of a PCA approach for chemotaxonomic classification of Cannabis varieties.
Copyright © 2012 John Wiley & Sons, Ltd.
Keywords: cannabis; cannabinoids; terpenoids; chemical profiling; cultivar; principle component analysis
Introduction
Cannabis as a medicine
Although there are an estimated 165 million frequent users of
Cannabis worldwide,
[1]
it is presently unclear how many of these
are medicinal users. Nevertheless, through persistent lobbying by
patients as well as through mounting scientific evidence, the
medicinal use of Cannabis is slowly gaining acceptance from
authorities. Over the last decade, both the Netherlands
[2]
and
Canada
[3]
have implemented state-run medicinal Cannabis
programmes, and other countries are considering a similar move
(Israel, Brazil) or are importing herbal material from the Dutch
programme (Italy, Finland, Germany; pers. comm. Office of
Medicinal Cannabis, the Netherlands).
[4]
In the USA, a growing
number of states have adopted medical marijuana laws to
provide safer access of Cannabis for medicinal use to patients,
despite the fact that this is vehemently opposed by the Federal
government.
[5]
Besides obvious legal implications,Cannabisasanherbalmedicine
poses serious challenges to modern medicine, which operates
according to the ‘single compound, single target’paradigm of
pharmacology. An obvious question therefore is how the chemical
constituents found in Cannabis reflect different medicinal properties,
and what types of Cannabis should consequently be made available
to patients. In fact, the Canadian programme is currently under
review, after increasing complaints from patients that the single
variety of Cannabis that is currently available is not effective for a
large proportion of patients.
[6]
It is now widely accepted that Cannabis is monotypic and consists
only of a single species Cannabis sativa, as described by Leonard
Fuchs in the sixteenth century.
[7,8]
But as a result of centuries of
breeding and selection, a large variation of cultivated varieties
(cultivars) have been developed. These are commonly distinguished,
by plant breeders, recreational users, and medical Cannabis patients
alike, through the use of popular names such as White Widow,
Northern Lights, Amnesia, or Haze. Already, over 700 different
varieties have been described
[9]
and many more are thought to exist.
However, it is unclear whether or not these names reflect any
relevant differences in chemical composition.
Classification systems
Some attempts have been made to classify Cannabis varieties based
on chemical composition. A first study was done by Grlic,
[10]
who
recognized different ripening stages. Later, Fettermann
[11]
described
different phenotypes based on quantitative differences in the
content of main cannabinoids and he was the first to distinguish
the drug- and fibre-type. Further extension of this approach was
done by Small and Beckstead,
[12]
Turner,
[13]
and Brenneisen.
[14]
However, it was found that a single plant could be classified into
different phenotypes, according to plant age. More recently, a
classification system was developed by de Meijer,
[15]
who recognized
five different Cannabis types based on the (relative) content of three
major cannabinoids.
For forensic and legislative purposes, the most important classifica-
tion of Cannabis types is that into the drug-type and the fibre-type
(hemp). The main difference between these two is found in the
content of the psychotropic component delta-9-tetrahydrocannabinol
(THC): a high content of THC classifies as drug-type Cannabis, while a
low content is found in fibre-type Cannabis (max. 0.2–0.3% THC on
basis of dry matter in the upper reproductive part of the plants), which
* Correspondence to: A. Hazekamp, R&D Department, Bedrocan BV, P.O. Box
2009 Veendam, the Netherlands. E-mail: ahazekamp@bedrocan.nl
aR&D Department, Bedrocan BV, Veendam, the Netherlands
bNatural Products Laboratory, Institute of Biology, Leiden University, the
Netherlands
Drug Test. Analysis (2012) Copyright © 2012 John Wiley & Sons, Ltd.
Research article
Drug Testin
g
and Anal
y
sis
Received: 18 November 2011 Revised: 29 November 2011 Accepted: 29 November 2011 Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI 10.1002/dta.407
may also be cultivated for its seeds for human or animal consumption.
The content of the closely related but psychotropically inactive canna-
bidiol (CBD) is not regulated by law, and its levels tend to be higher in
Cannabis cultivated for seed or fibre.
[16]
With almost no exception, Cannabis varieties presently used for
medicinal purposes in official programs belong to the drug-type,
because of their high content of the biologically active THC. But
although fibre-type Cannabis is not commonly used for either
medicinal or recreational purposes, it does contain components that
were found to be biologically active such as CBD,
[17]
indicating that
the distinction between the two types has limited relevance for
medicinal research into Cannabis. Moreover, it is becoming increas-
ingly clear that components in Cannabis beyond THC and CBD, such
as other minor plant-cannabinoids and volatile secondary metabo-
lites called terpenoids are involved in the drug’soveralleffect.
[18]
All major terpenoids present in Cannabis (including e.g. myrcene,
alpha-pinene, beta-caryophyllene) can be found ubiquitously in
nature.
[19]
For this reason these components did not receive much
scientific attention, until it was suggested that the terpenoid profile
of Cannabis products may help in determining the origin of Cannabis
in custom seizures.
[20]
However, no further reports on this approach
were subsequently made in the scientific literature, most likely
because of limited accuracy for forensic use.
The most common way currently used to classify Cannabis
cultivars is through plant morphology (phenotype) with two types
typically recognized: Sativa and Indica. Cannabis cultivars of the
Indica type are smaller in height with broader leaves, while Sativa
types are taller with long, thin-fingered leaves.
[21,22]
Indica plants typ-
ically mature faster than Sativa types under similar conditions, and
the types tend to have a different smell, perhaps reflecting a different
profile of terpenoids.
[23,24]
Bridging the gap between culture and science
As a result of limited understanding and support from the medical
community, medicinal users of Cannabis generally adopt the termi-
nology derived from recreational users to describe the therapeutic
effects they experience. Although it is hard to study the popular
Cannabis literature and come to a single clear conclusion, the
following general picture emerges about the difference between
typical Sativa and Indica effects upon smoking.
The Sativa high is often characterized as uplifting and energetic.
The effects are mostly cerebral (head-high), also described as spacey
or hallucinogenic. This type gives a feeling of optimism and well-
being, as well as providing a good measure of pain relief for certain
symptoms. Although Indicas are generally said to contain more
THC, a few pure Sativa types are also very high in THC content.
Sativas are considered a good choice for daytime smoking. In con-
trast, the Indica high is most often described as a pleasant body buzz
(body-high). Indicas are primarily enjoyed for relaxation, stress relief,
and for an overall sense of calm and serenity. Indicas are supposedly
effective for overall body pain relief, and often used in the treatment
of insomnia; they are the late-evening choice of many smokers as an
aid for uninterrupted sleep. A few pure Indica strains are very potent
in THC, and may cause the ‘couchlock’effect, enabling the smoker to
simply sit still and enjoy the experience of the Cannabis.
[22]
It has not been properly confirmed whether subjective descriptions
such as these are correlated in any way to the morphological distinc-
tions between Indica and Sativa cultivars, or to any other chemical
classification described above. It is obvious that a better understand-
ing of chemical differences between Cannabis cultivars could help
to bridge the gap between the vast knowledge on Cannabis that
exists within the community of recreational users, and the information
needed by medicinal users and health professionals. However, the
high number of (potential) active components present in Cannabis
significantly complicates a conventional reductionist approach using
analytical chemistry, animal studies, and clinical trials, where an active
ingredient needs to be identified before further study is possible.
From cultivar to chemovar
An alternative approach to this multiple component problem may be
to simultaneously identify and quantify all major components pres-
ent in various Cannabis types, and use a multi-variant data analysis
tool such as principal component analysis (PCA) to classify cultivars
in a small number of chemically distinct groups. With well-designed
animal studies and/or clinical trials, and using a range of distinct
chemical varieties, correlations may then be observed between spe-
cific chemical characteristics, and potentially beneficial biological
effects. Of course, such an approach fits exactly within the paradigm
of Systems Biology, which is recognized as a way to better under-
stand the complex interactions that can be involved in the effects
of medicinal plants with multiple active ingredients.
[25]
With this ap-
proachitmaybepossibletomoveawayfromCannabiscultivars,with
often vague and unsubstantiated characteristics, towards a new clas-
sification using chemovars with a complex, but nevertheless well-
defined chemical composition (also known as a chemical ‘fingerprint’).
In this study we attempt for the first time to directly compare
cultivars obtained from official as well as illicit sources through a
comprehensive chemical profile, including cannabinoids such as
THC and CBD, but also minor cannabinoids and a range of mono-
and sesqui-terpenoids. Two popular and widely available Cannabis
cultivars were obtained from a number of Dutch coffee shops
(highly regulated outlets of small amounts of Cannabis for recrea-
tional use, tolerated under Dutch law).
[2]
All samples were chemi-
cally profiled using gas chromatography with flame ionization
detection (GC/FID). Because major components are more likely to
be involved in medicinal effects, only compounds present above
a 0.5 mg/gram threshold level were selected for further analysis,
in order to determine the chemical variation present in both
varieties. This approach resulted in the quantification of 28 different
sample components. The results were then compared to those
obtained by analyzing three medicinal-grade Cannabis varieties
currently available from Dutch pharmacies.
[26]
The study indicated
the usefulness of PCA analysis for chemotaxonomic classification of
Cannabis varieties, and may assist in a better identification of
Cannabis cultivars with a potential for medicinal use.
Materials and methods
Sample collection
In order to evaluate the chemical variation found in some major
Cannabis cultivars, ten coffee shops in four major, geographically
dispersed, cities in the Netherlands (Amsterdam, Utrecht, Groningen,
and Maastricht) were selected for sample collection. All locations
were visited a second time about two months later to obtain a larger
variation (different batches) of Cannabis samples. Two popular strains
were selected for analysis, because they were generally available at
all locations throughout the year: Amnesia (a Sativa-dominant type
Cannabis) and White Widow (an Indica-dominant type).
[27]
One gram
of each sample was purchased. All samples were delivered in small
zip-lock bags, as typically provided by Dutch coffee shops.
A. Hazekamp and J. T. Fischedick
Drug Testin
g
and Anal
y
sis
wileyonlinelibrary.com/journal/dta Copyright © 2012 John Wiley & Sons, Ltd. Drug Test. Analysis (2012)
Three varieties of pharmaceutical-grade Cannabis (official product
name: Bedrocan
W
,Bedrobinol
W
,andBedica
W
)wereobtained
directly from the official Dutch cultivator, Bedrocan BV (Veendam,
the Netherlands). Plants were grown from clones under standardized
indoor conditions, and quality-controlled for chemical composition,
water content, and the absence of adulterants.
[26]
Flowertops were
harvested and air-dried for one week at controlled temperature
and humidity. Samples were delivered in airtight, triple-laminate
bags. All samples were kept in a freezer until analyzed.
Solvents and chemicals
Authentic standards for alpha-bisabolol, myrcene, alpha-pinene,
beta-pinene, gamma-terpineol and beta-caryophyllene were pur-
chased from Sigma-Aldrich (Steinheim, Germany). alpha-Humulene
was from Fluka (Steinheim, Germany). Terpinolene, gamma-
cadinene, cis-ocimene and beta-phellandrene were from a chemical
database of the authors.
Calibrated standards for THC, CBD, cannabigerol (CBG), cannabi-
chromene (CBC), trans-()-delta-9-tetrahydrocannabivarin (THCV),
and cannabinol (CBN) were purified and quantified as previously
described.
[28,29]
All cannabinoid references were >98% pure in
ethanol. For the structures of the cannabinoids(Figure1).Astandard
for cannabigerol-monomethylether (CBGM) was not available.
All organic solvents used were of analytical reagent (AR) grade
(Biosolve BV, Valkenswaard, the Netherlands).
Extraction
Extracts were prepared as described recently by Fischedick et al.
[30]
In short, 500 mg (+/2 mg) of each sample was extracted with
40 ml of absolute ethanol in plastic serum tubes (total volume
50 ml) while mechanically shaking for 10 min. Tubes were
centrifuged and clear supernatant was transferred to a 100-ml
volumetric flask. For exhaustive extraction, the procedure was
repeated twice more with 25 ml of ethanol, and supernatants were
combined. Volumes were adjusted to 100 ml and mixed well. Finally,
the solution was filtered through a 0.45 mm PTFE syringe filter.
Filtrates were stored at 20 C until analysis.
GC-analysis
An Agilent GC 6890 series (Agilent Technologies Inc., Santa Clara, CA,
USA) equipped with a 7683 autosampler and a flame ionization
detector (FID) was used for the simultaneous analysis of monoterpe-
noids, sesquiterpenoids, and cannabinoids, as previously described
by Fischedick.
[30]
The instrument was equipped with a DB5 capillary
column (30 m length, 0.25 mm internal diameter, film thickness
0.25 mm; J&W Scientific Inc., Folsom, CA, USA). The injector
temperature was 230 C, with an injection volume of 4 ml, a split ratio
of 1:20 and a carrier gas (N
2
)flow rate of 1.2 ml/min. The temperature
gradient started at 60 C and increased at a rate of 3 C/min until
240 C which was held for 5 min making a total run time of
65 min/sample. The FID detector temperature was set to 250C.
The GC-FID was controlled by Agilent GC Chemstation software
version B.04.01.
All terpenoids were quantified from a 4-point standard curve of
gamma-terpinene while all cannabinoids were quantified from a
4-point standard curve of THC, which was previously validated to give
adequate quantitative results.
[30]
The final result, expressing the
concentration of each component in mg/gram of plant material,
was used for PCA data analysis. Values were not corrected for water
content of the samples.
In order to confirm peak identification, samples and available
standards were further analyzed by gas chromatography–mass
spectrometry (GC-MS) as recently described
[30]
using a single
quadrupole MS in Total Ion Count (TIC) mode. Compounds were
identified by comparing their mass spectra and retention times with
authentic references as well as literature reports.
[23,31–34]
The NIST
library was also used to assist in compound identification (version
2.0f; Standard Reference Data Program of the National Institute of
Standards and Technology, as distributed by Agilent Technologies).
PCA data analysis
All compounds present at a mean concentration of ≥0.5 mg/gr in at
least one of the Cannabis varieties were identified and quantified
as described above. Resulting data was subjected to principal
component analysis (PCA ) with the aid of SIMCAP + version 12.0
software (Umetrics, Umeå, Sweden). Unit variance scaling was used.
The first two principal components (PC1 and PC2) were visualized
in a scatter-plot, in combination with its accompanying loading-plot.
For more information on interpreting PCA data, you are referred to
the tutorials section of the Umetrics website.
[35]
O
OH
HO
OH
O
OH OH
HO
OH
O O
OH
OCH3
HO
THC
tetrahydrocannabinol
CBD
Cannabidiol
CBG
Cannabigerol
CBN
Cannabinol
CBC
Cannabichromene
THCV
Delta-9-tetrahydrocannabivarin
CBGM
Cannabigerol-monomethyl-ether
Figure 1. Structures of main cannabinoids.
Towards a better definition of Cannabis potency
Drug Testin
g
and Anal
y
sis
Drug Test. Analysis (2012) Copyright © 2012 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/dta
Results and discussion
Sampling and analysis
In total, we intended to obtain 40 Cannabis samples for our study
(10coffeeshopsx2visitsx2varieties).However,infive cases the
desired product was not available from the coffee shop visited (2x
Amnesia, 3x White Widow), which resulted in a total of 18 Amnesia
samples and 17 White Widow samples available for analysis. In total,
28 different components were found to be present in one or more
Cannabis samples at a mean level of ≥0.5 mg/gr. These are indicated
in Table 1 (coffee shop samples) and Table 2 (pharmacy samples).
Values were not corrected for water content of the samples as
this may require more sample (cost) and it may induce the loss
of volatile constituents. All components could be positively iden-
tified with the help of retention time, mass-data and authentic
standards, except one single compound (m/z: 356 [M+], 313,
297 [base], 243, 231). Because we expected this component to
be a cannabinoid, it was used in the final data analysis (com-
pound 24 in Tables 1 and 2).
The use of only one terpenoid (gamma-terpinene) and one
cannabinoid (THC) for quantification of all sample components
may demand some further explanation. In a previously published
report, we showed that the differences in FID-detector response
between different types of cannabinoids, or between a range of
mono- and sesqui-terpenoids, were relatively small.
[30]
Obviously,
quantifying all 28 sample components with their own standard
curves would be preferred, but this is a tremendous and costly task.
Our current approach may give small deviations from actual
concentrations, but the method is easy to perform and it does not
invalidate our findings. In our case, the method is used for selecting
the main sample components (above ca. 0.5mg/gr) and to subse-
quently compare the relative differencesbetween samples. Because
our intention is to keep extending our database of cultivars on a
regular basis, and therefore would need a rapid and simple method-
ology, we believe the limitations of our method are acceptable.
The data presented in Table 1 prompt some intriguing ques-
tions about the popular name used for these cultivars. For exam-
ple, Amnesia has a mean titer of alpha-pinene that is less than
half that of White Widow. If Amnesia actually has more amnesic
effects, this distinction is noteworthy, since that component is
an anticholinesterase inhibitor, and may reduce associated
short-term memory deficits engendered by THC.
[18]
Similarly,
the myrcene titre of White Widow is more than four times that
in Amnesia. If White Widow, being of the Indica chemotype, is
indeed noted for its sedating properties (see Introduction), this
could be correlated, since myrcene has been associated with
‘couch-lock’.
[18]
Table 1. GC quantitative data of coffee shop samples.
white widow amnesia
ID# Compound RRT
MEAN STDEV RSD% MEAN STDEV RSD %
1alpha-pinene 0.257 1.1 0.24 21.1 0.5 0.04 6.9
2beta-pinene 0.306 0.5 0.08 16.3 0.7 0.16 22.7
3myrcene 0.319 2.9 1.16 39.5 0.7 0.16 23.4
4beta-phellandrene 0.371 0.7 0.13 17.9 0.7 0.13 19.9
5cis-ocimene 0.397 ———0.7 0.14 19.3
6terpinolene 0.458 ———1.9 0.77 39.6
7terpineol 0.622 0.6 * 0.0 —— —
8beta-caryophyllene 1.000 1.2 0.39 31.1 2.1 0.59 27.6
9alpha-guaiene 1.030 ———0.5 * 0.0
10 humulene 1.054 0.6 0.11 19.2 0.8 0.21 25.2
11 beta-farnesene 1.057 0.5 0.06 11.2 0.5 0.00 0.7
12 gamma-selinene 1.136 —0.7 0.25 —34.6 —
13 delta-guaiene 1.136 ——————
14 gamma-cadinene 1.180 0.6 0.22 34.3 0.6 0.08 14.2
15 eudesma-3,7(11)-diene 1.190 0.8 0.27 34.9 0.6 0.09 15.2
16 elemene 1.214 0.7 0.18 27.2 0.5 0.07 13.8
17 guaiol 1.274 0.6 0.12 20.0 —— —
18 gamma-eudesmol 1.307 0.6 0.21 35.7 —— —
19 beta-eudesmol 1.355 0.5 0.06 11.2 —— —
20 alpha-bisabolol 1.398 0.5 * 0.0 —— —
21 THCV 2.179 1.0 0.22 23.1 1.2 0.21 17.2
22 CBD 2.292 0.7 0.10 14.3 0.6 0.05 9.1
23 CBC 2.303 1.8 0.33 18.4 2.8 0.73 26.3
24 Unknown compound 2.348 0.7 0.26 36.2 0.7 0.12 17.4
25 CBGM 2.358 1.9 * 0.0 —— —
26 THC 2.399 159.5 26.76 16.8 167.5 21.88 13.1
27 CBG 2.456 5.4 2.34 43.2 12.5 2.95 23.6
28 CBN 2.461 0.9 0.19 21.0 1.4 0.41 29.6
All levels are expressed in mg of analyte per gram of cannabis sample.
* Analyte detected in only one sample, so standard deviation could not be calculated. RRT: relative retention time compared to beta-caryophyllene.
A. Hazekamp and J. T. Fischedick
Drug Testin
g
and Anal
y
sis
wileyonlinelibrary.com/journal/dta Copyright © 2012 John Wiley & Sons, Ltd. Drug Test. Analysis (2012)
PCA and quantitative analysis of coffee shop samples
PCA is a mathematical way of identifying patterns in data, and
expressing the data in such a way as to highlight their similarities
and differences. Since patterns can be hard to find in data of higher
dimension, where the luxury of graphical representation (i.e. a 2- or
3-dimensional diagram) is not available, PCA is a powerful tool for
analyzing data.
[36]
In this study, PCA was used to analyze our data set,
which consisted of 44 samples with up to 28 compounds each, equal-
ing 1232 data points. Each compound represented a variable while its
quantity represented the observation. Figure 2 shows an evaluation
of the first two principal components of all samples in the form of
a scatter plot (Figure 2a) and its related loading plot (Figure 2b).
The data as represented in the PCA scatter plot clearly shows that
Cannabis cultivars Amnesia and White Widow cluster in two chem-
ically distinct groups. Principal component 1 (PC1; 27.2%) and PC2
(17.0%) together explained 44.2% of the total variance found in the
sample set, indicating a relatively high degree of confidence. Anal-
ysis of PC3 added only another 11.4%. Interestingly, the mean THC
content found in Amnesia (167.5mg/gr, or 16.75 %) and White
Widow (159.5 mg/gr, or 15.95 %) samples showed no significant
difference, providing no basis for distinguishing both popular
varieties. However, the loading plot (Figure 2b) helps us under-
stand how exactly the two cultivars differ from one another. Based
on our PCA analysis, it becomes clear that monoterpenoids,
sesquiterpenoids, and minor cannabinoids are important in distin-
guishing the varieties.
Amnesia samples were characterized mainly by the presence of
the terpenoids terpinolene, alpha-guaiene and gamma-selinene,
which were not detected (i.e. below threshold levels) in White
Widow. Levels of beta-caryophyllene, as well as cannabinoids THCV,
CBC and, particularly, CBG were higher in Amnesia than in White
Widow. In contrast, White Widow samples showed significantly
higher content of myrcene and alpha-pinene. Samples of this variety
were further characterized by the presence of guaiol, beta-eudesmol,
gamma-eudesmol, and alpha-bisabolol, which were not detected in
Amnesia samples. Such sesquiterpene alcohols have previously been
reported to be important for the chemotaxonomic discrimination of
Indica Cannabis varieties originating from Afghanistan.
[24]
In three cases, chemical analysis showed that the obtained sample
was likely not of the desired cultivar type. More specifically, two
samples that were purchased as White Widow were found to
resemble the typical chemical fingerprint of Amnesia, while for one
Amnesia sample the opposite was true. Although these samples
are shown in Figure 2, we assumedthatamix-uphadoccurredat
the coffee shop, and these outliers were not included in our final
evaluation of chemical diversity. However, it is also possible that
these samples represented actual specimens of the desired cultivars,
which would even further increase the chemical diversity (i.e. STDEV)
found in both cultivar groups (Table 1). Naturally, a mix-up of samples
Table 2. GC quantitative data of pharmacy samples.
BEDROCAN BEDROBINOL BEDICA
ID# Compound
MEAN STDEV RSD % MEAN STDEV RSD % MEAN STDEV RSD %
1alpha-pinene 0.5 0.03 6.3 1.9 0.03 1.8 1.1 0.02 2.0
2beta-pinene 0.8 0.03 3.8 0.5 0.02 3.3 0.3 0.28 86.7
3myrcene 2.9 0.11 3.6 5.9 0.24 4.1 3.7 0.18 4.8
4beta-phellandrene 1.2 0.05 4.0 —— ——— —
5cis-ocimene 2.0 0.07 3.8 —— ——— —
6terpinolene 5.8 0.27 4.6 ——— —— —
7terpineol —— ——— ——— —
8beta-caryophyllene 1.5 0.05 3.3 0.9 0.04 4.2 1.3 0.01 0.6
9alpha-guaiene ——— —— ——— —
10 humulene 0.5 0.03 5.8 ——— —— —
11 beta-farnesene —— ——— ——— —
12 gamma-selinene ——— —— ——— —
13 delta-guaiene 0.7 0.02 3.5 ——— —— —
14 gamma-cadinene 0.6 0.03 5.2 ——— —— —
15 eudesma-3,7(11)-diene 0.8 0.00 0.6 —— —0.6 0.03 5.5
16 elemene 0.9 0.03 3.0 ——— —— —
17 guaiol —— ——— —0.6 0.03 5.3
18 gamma-eudesmol ——— —— —0.6 0.04 6.8
19 beta-eudesmol —— ——— ——— —
20 alpha-bisabolol —— ——— —0.7 0.03 3.6
21 THCV 1.3 0.04 3.2 0.9 0.08 8.4 1.0 0.09 8.5
22 CBD 0.8 0.03 4.3 —0.6 0.01 ——1.7
23 CBC 2.8 0.02 0.8 2.5 0.02 0.6 2.6 0.11 4.2
24 Unknown compound ——— —— ——— —
25 CBGM —— ——— ——— —
26 THC 211.9 6.48 3.1 135.0 5.02 3.7 174.5 6.14 3.5
27 CBG 12.0 0.64 5.3 2.4 0.11 4.4 12.4 0.47 3.8
28 CBN 0.8 0.16 19.0 ———0.7 0.08 11.1
All levels are expressed in mg of analyte per gram of cannabis sample.
Towards a better definition of Cannabis potency
Drug Testin
g
and Anal
y
sis
Drug Test. Analysis (2012) Copyright © 2012 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/dta
during extraction or processing may also be considered as a possible
reason for the observed effect. However, this risk was minimized by
clearly marking all samples directly after purchase, while keeping
them in their original packaging. Samples were then extracted and
processed in groups of eight sequential samples by two people
checking each other.
It may be concluded that, based on the overall chemical profile
consisting of 28 major components, it is indeed possible to
distinguish between two cultivars that are popularly recognized by
recreational Cannabis users: Amnesia and White Widow. However,
an important issue when using Cannabis as a medicine is the repro-
ducibility of the chemical composition of the product that the patient
is trying to obtain. For both Cannabis varieties analyzed in this study,
chemical variability was considerable, as shown in Table 1 and visual-
ized in the PCA score plot. Deviations (relative standard deviation) of
more than 25% were found for 14 analyzed compounds. To put
this variability in perspective, it makes sense to compare these results
to the chemical variation commonly observed in Cannabis obtained
from standardized sources (see next section).
Comparing coffee shop to pharmacy samples
Three different types of medicinal-grade Cannabis, currently avail-
able on prescription in the Netherlands, were analyzed according
to the procedures mentioned. In order to obtain better statistical
results, these analyses were done on triplicate samples for each
variety (compared to obtaining samples in 20-fold for both coffee
shop varieties). Chemical profiles were then compared to
determine similarities and differences between coffee shop and
pharmacy samples.
Although varieties Bedrobinol and Bedica have quite different
genetic backgrounds, PCA analysis indicated that both closely
resemble the typical fingerprint of the White Widow variety.
Indeed, variety Bedica is genetically derived from an Indica type,
so this would explain their similarity in chemical profile, such as
the presence of sesquiterpene alcohols. The fact that variety
Bedrobinol resembles White Widow is more surprising, as it was
genetically derived from a Haze variety, which has a large propor-
tion of Sativa genes in its background.
[27]
A potential explanation
is the fact that variety Bedrobinol has very low content of
terpenes in general, with only a few components detectable
above our 0.5 mg/gram threshold level. As a result, there may
not have been enough distinguishing features to clearly show
differences with the other analyzed samples. Finally, the differ-
ence between Bedrobinol and Bedica is mainly found in the
presence of sesquiterpenes such as guaiol, gamma-eudesmol,
and alpha-bisabolol in Bedica, while Bedrobinol was largely
devoid of these compounds.
Bedrocan is currently the most popular medicinal Cannabis
variety in the Netherlands. It is interesting to conclude that it does
not resemble either coffee shop type, or the two other pharmacy
varieties. In Figure 2, the data for Bedrocan (left-lower corner) even
falls outside the 95% confidence interval region, underscoring the
fact that its composition is significantly different from any of the
other four varieties analyzed. Bedrocan was originally developed
from a Jack Herer genetic background, which supposedly is a
slightly Sativa dominant hybrid.
[27]
Indeed, Bedrocan is closer to
Amnesia, which is also considered Sativa dominant, than to White
Widow. The main distinguishing feature of Bedrocan compared
to Amnesia is the markedly higher content of myrcene, cis-
ocimene and terpinolene by at least a factor of 2. In contrast, a
few components found in Amnesia were not detected (below
threshold levels) in Bedrocan (i.e. beta-farnesene, gamma-selinene,
gamma-cadinene and eudesma-3,7(11)-diene).
It is clear that plant-based products can never be perfectly
standardized for content of active components, as they are
dependent on too many environmental factors to completely predict
the chemical composition of the final product. For that reason, a
variability of up to 15% is allowed for the THC and CBD levels in
the Dutch medicinal Cannabis. Although currently no regulations
exist for the levels of minor cannabinoids or terpenoids, any producer
of a medicinally used product should strive for a minimum of
variation in chemical composition. In our previous study on
medicinal-grade Cannabis
[30]
using the same analytical methodol-
ogy, an average variation (RSD) of about 7.6% was found for
cannabinoids, and 11.0% for terpenoids in a standard batch of
variety Bedrocan; for variety Bedica these numbers were 4.3%
(terpenoids) and 5.5% (cannabinoids). Incontrasttothecoffee
shop samples discussed above, none of the components analyzed
in the pharmacy varieties showed a variability of over 25%.
Obviously, this result may be expected based on the fact that the
Cannabis samples of pharmaceutical grade were cultivated under
-6
-4
-2
0
2
4
6
-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8
PC2 [17.0%]
PC1 [27.2%]
White widow
Amnesia
Bedrocan Bedrobinol
Bedica
-0.35
-0.30
-0.25
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
-0.30 -0.25 -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20 0.25
PC2
PC1
beta-caryophyllene
CBN
humulene
gamma-selinene
CBG
CBC
alpha-
guaiene
THCV
terpinolene
beta-phellandrene
cis-ocimene
THC
beta-pinene
CBD
elemene
unknown
cannabinoid
gamma-cadinene
eudesma-3,7(11)-diene
beta-farnesene
CBGM
beta-eudesmol
terpineol
alpha-bisabolol
guaiol
gamma-
eudesmol
alpha-pinene
myrcene
delta-guaiene
b
a
Figure 2. PCA scatter plot. Solid ellipse indicates 95% confidence inter-
val. Dotted lines were drawn by the authors to facilitate visual interpreta-
tion of this figure only; it does not signify significance or confidence
intervals. Note that 2 samples purchased as Amnesia (closed triangles)
were chemically similar to White Widow (closed circles), while for one
sample the opposite was observed. It was assumed these samples were
mixed up by the coffee shop, and they were not used for final statistical
evaluations. Amnesia and White Widow data all represents individually
purchased samples from different batches. Data for Bedrocan, Bedrobinol
and Bedica represents triplicate analyses of the same batch.
A. Hazekamp and J. T. Fischedick
Drug Testin
g
and Anal
y
sis
wileyonlinelibrary.com/journal/dta Copyright © 2012 John Wiley & Sons, Ltd. Drug Test. Analysis (2012)
strictly controlled indoor conditions, and each batch was ana-
lyzed for chemical content by validated laboratory methods.
Conclusion
The psychotropic compound THC is often considered to be the main
component responsible for (medicinal) effects of Cannabis products,
through activation of specific cannabinoid-binding receptors.
[37]
But
increasingly, other, non-psychotropic cannabinoids are showing their
medicinal potential,
[38,39]
sometimes even through mechanisms that
do not involve binding to the known cannabinoid receptors.
[17]
With
a growing number of medicinal Cannabis users worldwide, it there-
foremaybetimetodefine the ‘potency’of Cannabis products in a
more comprehensive manner than by THC content alone. With mod-
ern analytical techniques, a full- range analysis of cannabinoids and
terpenoids has become possible.
In previous studies, our group showed that a range of Cannabis
cultivars grown by clonal propagation under controlled environmen-
tal conditions could be accurately distinguished based on their
cannabinoid and terpenoid content. We also observed that the
chemical profile of cannabinoids and terpenoids in different batches
of the same cultivar (grown sequentially over the period of several
months) was highlyreproducible.
[30]
These data suggest it is possible
to fully standardize the chemical content of cannabinoids as well as
terpenoids in Cannabis, when cultivated in a controlled setting.
In this study we applied the same analytical methodology to eval-
uate the chemical variation of two popular Cannabis cultivars
obtained in Dutch coffee shops. With this study we intended to (1)
determine whether there is a chemical basis to consider the obtained
varietiesastrulydifferentproducts, and (2) evaluate the chemical
variability that exists within each cultivar type. In order to bridge
the gap between recreational and medicinal use of Cannabis, we
then compared the overall chemical composition of the obtained
coffee shop samples with three types of medicinal Cannabis from a
government-licensed facility in the Netherlands. The fourth variety
of Cannabis currently available on prescription in the Netherlands,
called Bediol
W
, was not included in this study because its high
content of CBD (7.5%) completely changes its pharmacological spec-
trum, compared to the varieties analyzed in this study (CBD <0.1%).
In a total of 40 visits to the coffee shop, the desired product was
not available on five occasions, while in three cases a product was
obtained that significantly differed from the average composition
of the requested variety. As we assumed these three samples had
been mixed up in the coffee shop (either intentionally or
accidentally), we chose not to include them in our final analysis. If
our experiences can be considered as representative, it means that
in 20% of cases (8/40) a consumer is not able to receive a reliable
service when visiting an average coffee shop. The observed variabil-
ity in chemical composition may also be a cause of concern for
medicinalusers.Asitisoftenclaimed that the exact combination
of active constituents (synergy) gives each Cannabis variety its
unique (medicinal) properties, medicinal Cannabis users may
inadvertently purchase a product that has unexpected effects on
their health and/or psyche. It should be noted that changes in
chemical composition may not only be derived from genetic differ-
ences between batches of the same Cannabis variety, but could also
be caused by differences between coffee shops (or their suppliers) in,
for example, cultivation conditions, drying, processing, and storage.
However, no matter what the cause,Cannabisproductsfroma
standardized and quality-controlled source may be the safer choice
for medicinal users.
An important reason for patients to keep purchasing their
materials from illicit markets is the fact that, by trial and error,
they have found a strain that works optimally for treatment of
their specific symptoms. With the limited choice of Cannabis
varieties currently available from official sources, it is hard to
deny the value of such choice. By making use of the comprehen-
sive chemical fingerprint of herbal Cannabis, our results may help
medicinal users of Cannabis and their doctors to switch from a
beneficial variety obtained through the street market, to a strain
that is available through official state-run programmes. It may
also help these official programmes to narrow down the search
for future Cannabis varieties to be standardized and introduced
as an official medicine.
As this study intended to compare chemical data about
Cannabis samples with knowledge available from Cannabis
popular culture, we decided to use information from the Sensi
Seed Bank (the Netherlands) website as a reference for the Sativa
or Indica background of cultivars. Sensi Seed Bank is perhaps the
largest supplier of Cannabis seeds in the world, and an authorita-
tive voice among producers and users of recreational Cannabis.
In a next phase of our ongoing studies, we would like to obtain
typical ‘pure’Sativa and Indica strains to determine the two ex-
treme ends of the chemical diversity scale. Although this was
our intention for the present study, we were not able to identify
such strains that were available from all ten coffee shops we
visited for sample collection. According to Sensi Seed Bank, a
typical Sativa type Cannabis would be Durban (90% Sativa), while
some well-known examples of Indica types would be Hindu kush
(100% Indica), Afghani #1 (95%), or Northern lights (90%). As an
example of pure Sativa we could also select a non-drug variety
(i.e. hemp) available in the Netherlands, and grow it under
standardized conditions.
We consider this study a work in progress, and wish to add a
larger number of popular strains from the recreational scene,
as well as Cannabis varieties from official programmes to our
database. Over time, we hope this will lead to a better under-
standing of the overlap between medicinal use of Cannabis,
and the street culture around the same plant. It is our personal
belief that, with such an approach, the endless number of popu-
lar cultivars could be reduced to perhaps two dozen chemically
distinct chemical varieties, or chemovars. Further study into the
chemical similarities between popular coffee shop (or other illicit
sources) Cannabis types, and available medicinal-grade strains
may help medical users to accurately and efficiently find a more
reliable source of medicine within the various national medicinal
Cannabis programmes. Exchange of cultivars and information
between the various national programmes, in countries including
the Netherlands, Canada, and the USA, may greatly facilitate such
a transition.
Acknowledgement
The authors would like to pay their final respect to their former
colleague Gian van der Water, whose enthusiasm was a driving
force behind this study.
Disclaimer
AH has been employed by Bedrocan BV as the head of Research
and Development since February 2011.
Towards a better definition of Cannabis potency
Drug Testin
g
and Anal
y
sis
Drug Test. Analysis (2012) Copyright © 2012 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/dta
References
[1] United Nations Office on Drugs and Crime. World Drug Report.
2011.
[2] A. Hazekamp. An evaluation of medicinal grade cannabis in the
Netherlands. Cannabinoids 2006,1(1), 1.
[3] P.G. Lucas. Regulating compassion: An overview of Canada’s federal
medical cannabis policy and practice. Harm Reduct. J. 2008,5,1.
[4] OMC (Office of Medicinal Cannabis). Available at: www.cannabisbureau.
nl [November 2011].
[5] D.E. Hoffmann, E. Weber. Medical marijuana and the law. New Engl. J.
Med. 2010,362, 1453.
[6] Health Canada. Available at: http://www.hc-sc.gc.ca/dhp-mps/
marihuana/index-eng.php [November 2011].
[7] R. Brenneisen. Psychotrope drogen I. Cannabis sativa L. Pharm. Acta
Helv. 1983,58, 314.
[8] A. Hazekamp. Chemistry of Cannabis, in Comprehensive Natural Pro-
ducts II Chemistry and Biology, Vol. No. 3, (Eds: L. Mander, H.-W. Lui).
Elsevier, Oxford, 2010, pp. 1033–1084.
[9] W. Snoeijer. A Checklist of Some Cannabaceae Cultivars. Part 1:
Cannabis. Div. Pharmacognosy, Leiden/Amsterdam Centre for Drug
Research, Leiden, the Netherlands, 2001.
[10] L. Grlic. A combined spectrophotometric differentiation of samples
of cannabis. Bull. Narcot. 1968,20, 25.
[11] P.S. Fetterman, E.S. Keith, C.W. Waller, O. Guerrero, N.J. Doorenbos,
M.W. Quimby. Mississippi grown Cannabis sativa L: Preliminary
observation on chemical definition of phenotype and variations in
tetrahydrocannabinol versus age, sex, and plan part. J. Pharm. Sci.
1971,60, 1246.
[12] E. Small, H.D. Beckstead. Common cannabinoid phenotypes in 350
stocks of cannabis. Lloydia 1973,36, 144.
[13] C.E. Turner, M.A. ElSohly, P.C. Cheng, G. Lewis. Constituents of
Cannabis sativa L, XIV: intrinsic problem in classifying Cannabis
based on a single cannabinoid analysis. J. Nat. Prod. 1979,42, 317.
[14] R. Brenneisen, T. Kessler. Die variabilität der Cannabinoidführung
von Cannabispflanzen aus Schweizer Kulturen in Abhängigkeit von
genetischen und ökologischen Faktoren. Pharm. Acta Helv. 1987,
62, 134.
[15] E.P.M. Meijer, K.M. de Hammond, A. Sutton. The inheritance of
chemical phenotype in Cannabis sativa L. (IV): Cannabinoid-free
plants. Euphytica 2009,168, 95.
[16] C. Rustichelli, V. Ferioli, M. Baraldi, P. Zanoli, G. Gamberini. Analysis of
cannabinoids in fiber hemp plant varieties (Cannabis sativa L.) by
high-performance liquid chromatography. Chromatographia 1998,
48, 215.
[17] A.A. Izzo, F. Borrelli, R. Capasso, V. Di Marzo, R. Mechoulam. Non-
psychotropic plant cannabinoids: new therapeutic opportunities
from an ancient herb. Trends Pharmacol. Sci. 2009,30, 515.
[18] E.B. Russo. Taming THC: Potential cannabis synergy and
phytocannabinoid-terpenoid entourage effects. Brit. J. Pharmacol.
2011,163, 1344.
[19] C.E. Turner, M.A. Elsohly, E.G. Boeren. Constituents of cannabis sativa L.
XVII. A review of the natural constituents. J. Nat. Prod. 1980,43, 169.
[20] R. Brenneisen, M.A. ElSohly. Chromatographic and spectroscopic
profiles of cannabis of different origins: Part I. J. Forens. Sci. 1988,
33, 1385.
[21] A. Cronquist. An integrated system of classification of flowering plants.
Cambridge University Press, New York, USA, 1981, p. 193.
[22] J. Holland. The Pot Book - A Complete Guide to Cannabis: Its Role in
Medicine, Polititcs, Science, and Culture. Park Street Press, Rochester,
NY, USA, 2010, p. 45.
[23] K.W. Hillig, P.G. Mahlberg. A chemotaxonomic analysis of cannabinoid
variation in Cannabis (cannabaceae). Am.J.Bot.2004,91, 966.
[24] K.W. Hillig. A chemotaxonomic analysis of terpenoid variation in
Cannabis.Biochem. Syst. Ecol. 2004,32, 875.
[25] N.D. Yuliana, A. Khatib, Y.H. Choi, R. Verpoorte. Metabolomics for
bioactivity assessment of natural prducts. Phytother. Res. 2011,
25,157.
[26] B.V. Bedrocan. Cultivator of medicinal grade cannabis in the Netherlands.
Available at: www.bedrocan.nl [November 2011].
[27] Sensi Seed Bank. Available at: www.sensiseeds.com [November 2011].
[28] A. Hazekamp, R. Simons, A. Peltenburg-Looman, M. Sengers, R. van
Zweden, R. Verpoorte. Preparative isolation of cannabinoids from
Cannabis sativa by centrifugal partition chromatography. J. Liq.
Chrom. Rel. Technol. 2004,27, 2421.
[29] A. Hazekamp, Y.H. Choi, R. Verpoorte. Quantitative analysis of
cannabinoids from Cannabis sativa using 1H-NMR. Chem. Pharm.
Bull. 2004,52, 718.
[30] J.T. Fischedick, A. Hazekamp, T. Erkelens, Y.H. Choi, R. Verpoorte.
Metabolic fingerprinting of Cannabis sativa L., cannabinoids and
terpenoids for chemotaxonomic and drug standardization purposes.
Phytochemistry 2010,71, 2058.
[31] T. Komori, T. Nohara, I. Hosokawa, T. Kawasaki. Cannabigerol
monomethyl ether, a new component of hemp. Chem. Pharm. Bull.
1968,16, 1164.
[32] R.P. Adams. Identification of Essential Oils by Ion Trap Mass
Spectrometry. Academic Press Inc., New York, USA, 1989.
[33] S.A. Ross, M.A. ElSohly. The volatile oil composition of fresh and air-dried
buds of Cannabis sativa. J. Nat. Prod. 1996,59,49.
[34] M. Rothschild, G. Bergström, S. Wångberg. Cannabis sativa: volatile
compounds from pollen and entire male and female plants of two
variants, Northern lights and Hawaiian indica. Bot. J. Linn. Soc.
2005,147, 387.
[35] Umetrics, developer of PCA analysis software. Available at: www.
umetrics.com [November 2011].
[36] L. Eriksson, E. Johansson, N. Kettaneh-Wold, J. Tyrgg, C. Wikström, S.
Wold. Multi- and Megavariate Data Analysis Part 1: Basic Principles
and Applications, 2nd Edn. Umetrics, Umeå, Sweden, 2006.
[37] R.G. Pertwee, A.C. Howlett, M.E. Abood, S.P. Alexander, V. Di Marzo,
M.R. Elphick, P.J. Greasley, H.S. Hansen, G. Kunos, K. Mackie, R.
Mechoulam, R.A. Ross. International Union of Basic and Clinical Phar-
macology. LXXIX. Cannabinoid receptors and their ligands: Beyond
CB1 and CB2. Pharmacol. Rev. 2010,62, 588.
[38] M. Ben Amar. Cannabinoids in medicine: A review of their therapeutic
potential. J. Ethnopharmacol. 2006,105,1.
[39] A. Hazekamp, F. Grotenhermen. Review on clinical studies with
cannabis and cannabinoids 2005–2009. Cannabinoids 2010,5,1.
A. Hazekamp and J. T. Fischedick
Drug Testin
g
and Anal
y
sis
wileyonlinelibrary.com/journal/dta Copyright © 2012 John Wiley & Sons, Ltd. Drug Test. Analysis (2012)
