Journal of Pharmaceutical and Biomedical Analysis 143 (2017) 228–236

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

journal homepage: www.elsevier.com/locate/jpba

Development of a new extraction technique and HPLC method for the

analysis of non-psychoactive cannabinoids in fibre-type Cannabis
sativa L. (hemp)
Virginia Brighenti a , Federica Pellati a,∗ , Marleen Steinbach b , Davide Maran a ,
Stefania Benvenuti a
a
Department of Life Sciences, University of Modena and Reggio Emilia, Via G. Campi 103, Modena 41125, Italy
b
Faculty of Agricultural Sciences, Nutritional Sciences, and Environmental Management, Justus-Liebig University of Giessen, Goethestrasse 58, Giessen
35390, Germany

a r t i c l e i n f o a b s t r a c t

Article history: The present work was aimed at the development and validation of a new, efficient and reliable technique
Received 10 February 2017 for the analysis of the main non-psychoactive cannabinoids in fibre-type Cannabis sativa L. (hemp) inflo-
Received in revised form 29 May 2017 rescences belonging to different varieties. This study was designed to identify samples with a high content
Accepted 31 May 2017
of bioactive compounds, with a view to underscoring the importance of quality control in derived products
Available online 4 June 2017
as well. Different extraction methods, including dynamic maceration (DM), ultrasound-assisted extrac-
tion (UAE), microwave-assisted extraction (MAE) and supercritical-fluid extraction (SFE) were applied
Keywords:
and compared in order to obtain a high yield of the target analytes from hemp. Dynamic maceration
Cannabis sativa L.
Hemp
for 45 min with ethanol (EtOH) at room temperature proved to be the most suitable technique for the
Cannabinoids extraction of cannabinoids in hemp samples.
Extraction The analysis of the target analytes in hemp extracts was carried out by developing a new reversed-
HPLC phase high-performance liquid chromatography (HPLC) method coupled with diode array (UV/DAD) and
MS electrospray ionization-mass spectrometry (ESI–MS) detection, by using an ion trap mass analyser. An
Ascentis Express C18 column (150 mm × 3.0 mm I.D., 2.7 ␮m) was selected for the HPLC analysis, with a
mobile phase composed of 0.1% formic acid in both water and acetonitrile, under gradient elution. The
application of the fused-core technology allowed us to obtain a significant improvement of the HPLC
performance compared with that of conventional particulate stationary phases, with a shorter analysis
time and a remarkable reduction of solvent usage.
The analytical method optimized in this study was fully validated to show compliance with interna-
tional requirements. Furthermore, it was applied to the characterization of nine hemp samples and six
hemp-based pharmaceutical products. As such, it was demonstrated to be a very useful tool for the anal-
ysis of cannabinoids in both the plant material and its derivatives for pharmaceutical and nutraceutical
applications.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction

Cannabis sativa L. is well known, due to its history, pharma-

Abbreviations: CBD, cannabidiol; CBDA, cannabidiolic acid; CBG, cannabigerol; cology and social impact. It is a dioicous plant belonging to the
CBGA, cannabigerolic acid; CBC, cannabichromene; CBCA, cannabichromenic acid; Cannabaceae family, which is widely cultivated worldwide [1]; its
CBN, cannabinol; CBNA, cannabinolic acid; 9 -THC, 9 -tetrahydrocannabinol;
use as a psychoactive drug, as a folk medicine ingredient and a
THCA, 9 -tetrahydrocannabinolic acid; PFP, pentafluorophenyl; PEG, polyethy-
lene glycol; DM, dynamic maceration; UAE, ultrasound-assisted extraction; MAE, source of textile fibre has been widespread since ancient times [2].
microwave-assisted extraction; HCOOH, formic acid; EtOH, ethanol; MeOH, The genetic variability of Cannabis has turned the taxonomic clas-
methanol; ACN, acetonitrile; CHCl3 , chloroform; H2 O, water; CO2 , carbon dioxide; sification of this plant into a troublesome issue and the discussion
CID, collision induced dissociation.
of its proper classification is still ongoing [1,3]. Initially, the genus
∗ Corresponding author at: Department of Life Sciences, University of Modena and
Cannabis was divided into three main species [1,3,4]: a fibre-type
Reggio Emilia, Via G. Campi 103, 41125 Modena, Italy.
E-mail address: [email protected] (F. Pellati). one, named C. sativa L., a drug-type one, containing high levels of

http://dx.doi.org/10.1016/j.jpba.2017.05.049
0731-7085/© 2017 Elsevier B.V. All rights reserved.
 V. Brighenti et al. / Journal of Pharmaceutical and Biomedical Analysis 143 (2017) 228–236 229

(Z) (Z) (Z) OH O

OH O OH OH O
OH
OH (R) OH
(S) (S) HO
HO HO HO
CBDA CBGA
CBDA CBD

OH O OH Δ -CO
2
Δ -CO
2

OH (E)
(Z) OH
HO HO OH
(E)

CBGA CBG (R) HO

HO
Fig. 1. Chemical structures of cannabidiolic acid (CBDA), cannabidiol (CBD),
cannabigerolic acid (CBGA) and cannabigerol (CBG). CBD CBG

Fig. 2. Breakdown products of the main cannabinoic acids present in hemp (CBDA,
CBGA).
the psychoactive compound 9 -tetrahydrocannabinol (9 -THC),
named C. indica Lam., and another one with intermediate char-
acteristics, named C. ruderalis Janisch. However, the continuous been found to possess antimicrobial and anti-nausea properties
crossbreeding of these species to generate hybrids has led to the [1,11,14], while CBG has anti-inflammatory, antimicrobial and
use of a monotypic classification, in which all plants are classified analgesic activities [1,11,14,15]. In this respect, the selection of
as belonging to C. sativa species and subdivided into chemotypes botanical varieties of fibre-type C. sativa with a high content of
[1,3,4]. Fibre-type C. sativa is also commonly known as hemp. these non-psychoactive constituents by means of advanced ana-
In general, the remarkable pharmacological activity of psy- lytical techniques is as important as the assessment of a suitable
choactive cannabinoids makes drug-type C. sativa one of the most extraction procedure, with a view to the biological activity cited
thoroughly investigated medicinal plant [3]. Fibre-type plants above [1,3,11–14].
remain at the moment underused in the pharmaceutical ambit, The analysis of cannabinoids in the Cannabis species usually
where drug-type C. sativa is employed as medicinal Cannabis [1]. involves the determination of the main compounds by means of
Nevertheless, there has also been increasing interest in hemp gas chromatography (GC) combined with both a flame ioniza-
varieties containing non-psychoactive compounds [1], and those tion detector (FID) and mass spectrometry (MS) [5,16]. However,
approved for commercial use by the European Union amount to a a decarboxylation of the native components (the acids) to their
total of 57 [5]. Most of the European Union countries and Canada neutral forms occurs by using GC techniques, due to the high
have recognized the value of fibre-type hemp and they have defined temperature reached. Therefore, a reliable metabolite profiling of
a legal limit of 0.3% 9 -THC [1]. cannabinoids in the plant material is not possible [4,7,8,16,17].
Hemp is characterized by an extremely complex chemical In this context, high-performance liquid chromatography (HPLC)
composition, which encompasses terpenoids, sugars, alkaloids, stil- methods offer a good alternative to GC. In particular, HPLC coupled
benoids, quinones and the specific compounds of this plant, namely with diode-array detector (UV/DAD) and electrospray ionization-
cannabinoids [6]. Cannabinoids are a class of meroterpenoids mass spectrometry detector (ESI–MS) has been frequently applied
(specifically terpenophenolics), derived from the alkylation of an for the analysis of cannabinoids in hemp [4,7,8,16,18–23]. Peschel
alkyl resorcinol with a monoterpene unit [3]. They are mainly and Politi [8] also combined proton nuclear magnetic resonance (1 H
synthesised in glandular trichomes more abundant in female inflo- NMR) and HPLC-UV/DAD for the identification of chemical markers
rescences [7]. Hemp includes several chemotypes, changing in both in herbal drugs containing medicinal Cannabis extracts for chemo-
the qualitative and the quantitative profile of cannabinoids [4,8]. type distinction.
Until now, more than 100 of these compounds have been isolated, Despite the large number of studies on cannabinoid composition
identified and grouped in 11 chemical classes [4,9]. In general, the in drug-type Cannabis plants through HPLC [7,8,18,19,21], to the
main cannabinoids present in fibre-type plants are the acid ones, best of our knowledge there is no validated analytical method for
such as cannabidiolic acid (CBDA) and cannabigerolic acid (CBGA) the analysis of the main non-psychoactive cannabinoids in fibre-
(Fig. 1), followed by their decarboxylated forms, namely cannabid- type hemp plants and derived pharmaceutical products. In the
iol (CBD) and cannabigerol (CBG) (Fig. 1). light of all the above, the development and validation of a new,
Other minor cannabinoids comprise cannabichromenic acid reliable and highly efficient analytical method for the metabo-
(CBCA), cannabichromene (CBC), cannabinolic acid (CBNA) and lite profiling of non-psychoactive cannabinoids in fibre-type C.
cannabinol (CBN), with the last two being the oxidation products sativa by means of HPLC with UV/DAD, ESI–MS and MS2 detection
of 9 -tetrahydrocannabinolic acid (THCA) and THC, respectively was performed for the first time in this study. The chromato-
[1,3,4,7,10]. It is worth observing that cannabinoids are biosynthe- graphic performance of four columns, including a fully porous C18 ,
sized in the acid form in plant tissues, where they can undergo a pentafluorophenyl (PFP), a polyethylene glycol (PEG) and a fused-
a spontaneous decarboxylation process and generate the decar- core C18 stationary phase, was evaluated. Indeed, the most recently
boxylated counterparts under the action of heat and light (Fig. 2) reported strategy for improving chromatographic performance in
[1,3,4,7,10]. the analysis of complex matrices is based on fused-core technology
From a pharmaceutical point of view, CBD represents the most (also known as core-shell technology) [24]. Compared with fully
valuable compound, since it possesses a high anti-oxidant and porous particles, fused-core ones have a much shorter diffusion
anti-inflammatory activity, as well as antibiotic, neuroprotective, path because of the solid core. This tends to reduce the axial disper-
anxiolytic and anticonvulsant properties [1,3,11–14]. CBDA has sion of solutes and minimizes peak broadening. This type of support
230 V. Brighenti et al. / Journal of Pharmaceutical and Biomedical Analysis 143 (2017) 228–236

Table 1 Hemp-based pharmaceutical products, including two samples

Hemp varieties considered in this study
of oil (indicated in the text as oil 1 and 2, respectively), three
Sample Origin Code samples of balm (balm 1–3) and one extract equipped with a
Futura 75 France C1 syringe-based oral applicator, were kindly provided by the Uni-
Felina 32 France C2 versity Hospital of Modena and Reggio Emilia, Italy.
Fedora 17 France C3
Bernabeo France C4
Carmaleonte Italy C5
2.3. Extraction of non-psychoactive cannabinoids from hemp
Codimono Italy C6 plant material
Carmagnola Italy C7
2077/12 Italy C8 Four extraction techniques were applied and compared in order
Ermo Italy C9
to obtain a high yield of non-psychoactive cannabinoids from hemp
samples, including DM, UAE, MAE and SFE. As regards DM, UAE and
MAE the same sample-to-solvent ratio (w/v) and extraction solvent
can provide speed and efficiency similar to columns packed with
(EtOH) were applied. All the extraction procedures were carried out
sub–2 ␮m particles, but with reduced backpressure. This allows for
in duplicate.
the application of fused-core columns to conventional HPLC equip-
ment. As regards sample preparation, four extraction techniques
were compared, including dynamic maceration (DM), ultrasound- 2.3.1. Dynamic maceration (DM)
assisted extraction (UAE), microwave-assisted extraction (MAE) DM was performed on a weighed amount of hemp inflores-
and supercritical-fluid extraction (SFE), in order to obtain a high cences (0.25 g) with 10 mL of EtOH as the extraction solvent at room
recovery of the compounds of interest from the plant material. temperature for 15 min, under magnetic stirring. The solution was
The method developed in this study represents a powerful tool for then paper filtered and the residue was extracted with the same
the extraction and analysis of non-psychoactive cannabinoids from procedure twice more with 10 and 5 mL of solvent, respectively. The
hemp varieties to be used for the preparations of extracts with a filtrates of the three extractions were then combined and brought
high content of bioactive compounds for both pharmaceutical and to 25 mL with the solvent in a volumetric flask. Before the injection
nutraceutical applications. The acquisition of the chromatographic into the HPLC system, the extracts were filtered by using a 0.45 ␮m
profile of cannabinoids in hemp-based pharmaceutical products is PTFE filter in a HPLC vial.
also very useful for the assessment of their quality, efficacy and
safety. 2.3.2. Ultrasound-assisted extraction (UAE)
UAE was performed on 0.25 g of sample with 10 mL of EtOH at
2. Materials and methods 40 ◦ C for 15 min, by using an ultrasonic bath (Sonorex RK-100H,
Bandelin, Berlin, Germany). The solution was then filtered through
2.1. Chemicals and solvents a paper filter into a volumetric flask. The procedure was repeated
twice more, by adding 10 and 5 mL of solvent to the residue, respec-
Cannabidiolic acid (CBDA), cannabigerolic acid (CBGA), tively. The filtrates of the three extractions were then combined and
cannabigerol (CBG) and cannabidiol (CBD) standard solutions brought to the final volume of 25 mL with the extraction solvent.
(1 mg/mL in methanol or acetonitrile) were purchased from Ceril- Before the injection into the HPLC system, the extracts were filtered
liant (Round Rock, Texas, USA). Formic acid (HCOOH), HPLC-grade by using a 0.45 ␮m PTFE filter in a HPLC vial.
solvents, including ethanol (EtOH), methanol (MeOH), acetonitrile
(ACN), isopropanol, hexane, chloroform (CHCl3 ) and acetone were 2.3.3. Microwave-assisted extraction (MAE)
from Sigma-Aldrich-Fluka (Milan, Italy). Water (H2 O) was purified With regard to MAE, a weighed amount of sample (0.25 g) was
by using a Milli-Q Plus185 system from Millipore (Milford, MA, extracted with 10 mL of EtOH into a 20 mL glass vessel at 60 ◦ C
USA). Carbon dioxide (CO2 ) was 4.5 purity grade (99.995%) and it for 5 min under stirring, by using a monomode focused microwave
was supplied by SOL s.p.a. apparatus with a closed-vessel system (Biotage Initiator Robot
Sixty, Biotage, Uppsala, Sweden). At the end of the extraction pro-
2.2. Hemp plant material and hemp-based pharmaceutical cedure, the solution was paper filtered into a volumetric flask. The
products residue of the first extraction was extracted twice more, firstly with
10 mL more of EtOH and finally with a further 5 mL. The filtrates of
Nine samples of fibre-type C. sativa female inflorescences (indi- the three extractions were then combined and brought to the final
cated in the text as C1-C9, respectively) were analysed in this study volume of 25 mL in a volumetric flask with the solvent. Before the
(Table 1). These samples (about 100 g dry material each), belonging injection into the HPLC system, the extracts were filtered by using
to different breeding lines, were cultivated under the same agro- a 0.45 ␮m PTFE filter in a HPLC vial.
nomic conditions. The plant material was kindly provided by the
research center CREA-CIN (Rovigo, Italy) and by different Italian 2.3.4. Supercritical fluid extraction (SFE)
farms, and certified with a content of 9 -THC below 0.3% (w/w). All Supercritical fluid extraction (SFE) was carried out by using an
the samples considered in this study were hemp varieties approved Applied Separations extractor (Allentown, PA, USA) model Spe-
for commercial use by the European Union [5], with the exception edTM SFE Prime. The extraction was performed with supercritical
of samples “Bernabeo”, “2077/12” and “Ermo”, which were experi- CO2 and 20% of EtOH as a co-solvent [19,25]. A 50 mL extraction ves-
mental varieties. The “Futura 75” sample (C1) was selected for the sel was loaded with 5 g of powdered sample mixed with an equal
purpose of method development. At the same time, the “Bernabeo” amount of Spe-edTM Matrix. After the sample compression, a plug
sample (C4) was employed for method validation, since it displayed of wool (Spe-edTM Wool) was placed on the top and the empty
a broader cannabinoid profile. For each sample, hemp inflores- space was filled with the matrix. The CO2 flow rate was maintained
cences were manually separated from twigs and seeds. After this at an average level of 2.5 L/min, and the process was performed at
procedure, the samples were stored at +4.0 ◦ C until required for 100 bar at 35 ◦ C for 5 min of static extraction, followed by 15 min
chemical analysis. Hemp inflorescences were grounded in a mortar of dynamic extraction. EtOH was used to collect the extract. The
before the extraction. extract was then brought to dryness and dissolved in 10 mL EtOH.
 V. Brighenti et al. / Journal of Pharmaceutical and Biomedical Analysis 143 (2017) 228–236 231

Before the injection into the HPLC system, the extract was diluted 2.7. HPLC-UV/DAD method validation
1:50 (v/v) and filtered by using a 0.45 ␮m PTFE filter in a HPLC vial.
The validation of the HPLC-UV/DAD method was performed in
agreement with the international guidelines on analytical tech-
2.4. Sample preparation from hemp-based pharmaceutical
niques for the quality control of pharmaceuticals (ICH guidelines)
products
[26]. UV/DAD detectors are indeed widely used for quantitative
analyses of cannabinoids in plant extracts [7,8,22].
As regards sample preparation from pharmaceutical products,
Regarding linearity, the stock standard solution of each com-
50 ␮L of hemp oil were diluted to 10 mL with isopropanol in a vol-
pound (CBDA, CBGA, CBG and CBD) was prepared as follows: an
umetric flask. In the case of the balm, a 0.25 g sample was dissolved
accurate volume of each standard solution was placed into a volu-
in isopropanol and diluted to 25 mL volume in a volumetric flask. As
metric flask (200 ␮L of the standard solution in 1 mL). The standards
for the hemp-based extract with the oral applicator, finally, 0.020 g
where then brought to volume with EtOH and the external stan-
of product was dissolved in 10 mL of isopropanol and the resulting
dard calibration curve was generated by using six data points. Three
solution was further diluted to 1:5 (v/v) with isopropanol. Before
␮L of each sample were used for the HPLC analysis and the injec-
the injection into the HPLC system, the extracts were filtered by
tions were performed in triplicate for each concentration level. The
using a 0.45 ␮m PTFE filter in a HPLC vial.
calibration curve was generated by plotting the peak area of the
The sample preparation procedure was carried out in duplicate
compound at each level versus its concentration in the solution.
for each product.
For reference compounds, the limit of detection (LOD) and
the limit of quantification (LOQ) were experimentally determined
2.5. HPLC-UV/DAD analysis through HPLC analysis of serial dilutions of a standard solution to
reach a signal-to-noise (S/N) ratio of 3 and 10, respectively.
The HPLC-UV/DAD analyses were performed on an Agilent The accuracy of the analytical method was evaluated by means
Technologies (Waldbronn, Germany) modular model 1100 sys- of the recovery test. This involved the addition of a known quan-
tem, consisting of a vacuum degasser, a quaternary pump, an tity of standard compound to half the sample weight of grounded
autosampler, a thermostated column compartment and a diode sample (sample C1) to reach 100% test concentration. The forti-
array detector (UV/DAD). Chromatograms were recorded by using fied samples were then extracted and analysed with the proposed
an Agilent Chemstation for LC and LC–MS systems (Rev. B.01.03). method.
An Ascentis Express C18 column (150 mm × 3.0 mm I.D., 2.7 ␮m, The precision of the extraction technique was validated by
Supelco, Bellefonte, PA, USA) was used, with a mobile phase com- repeating the extraction procedure on the inflorescences of the
posed of 0.1% HCOOH in both (A) H2 O and (B) ACN. The gradient same hemp sample (sample C1) six times. An aliquot of each extract
elution was modified as follows: 0–13 min 60% B, 13–17 min from was then injected and quantified. The precision of the chromato-
60% to 80% B, 17–22 min from 80% to 90% B, which was kept graphic system was tested by performing intra- and inter-day
for 8 min. The post-running time was 15 min. The flow-rate was multiple injections of one extract (sample C4) and then by check-
0.4 mL/min. The column temperature was set at 30 ◦ C. The sample ing the%RSD of retention times and peak areas. Six injections were
injection volume was 3 ␮L. The UV/DAD acquisitions were carried performed each day for three consecutive days.
out in the range 190–600 nm, while chromatograms were acquired
at 210 nm (for decarboxylated cannabinoids) and at 220 nm (for 2.8. Statistical analysis
cannabinoic acids). Three injections were performed for each sam-
ple. All statistical analyses were performed by using Excel software
Office Home and Business 2016 and R-Studio version 3.3.1 for Mac.
2.6. HPLC-ESI–MS and MS2 analysis Concentrations, recovery, means and standard deviation were cal-
culated with Excel. R-Studio was used for the analysis of variances
The HPLC-ESI–MS and MS2 analyses were performed on an Agi- (ANOVA) and Tukey’s adjustment for multiple comparisons was
lent Technologies modular 1200 system, equipped with a vacuum implemented to test for significant differences between the means.
degasser, a binary pump, a thermostated autosampler, a ther- P-values under the significance level (␣) of 0.05 were considered
mostated column compartment and a 6310A ion trap mass analyser statistically significant.
with an ESI ion source. The HPLC column and the applied chro-
matographic conditions were the same as those reported for the 3. Results and discussion
HPLC-UV/DAD system.
The HPLC-ESI–MS system was operated both in the positive and 3.1. HPLC method development
in the negative ion mode. For the positive ion mode, the experimen-
tal parameters were set as follows: the capillary voltage was 3.5 kV, The aim of this work was to develop a new analytical method
the nebulizer (N2 ) pressure was 32 psi, the drying gas temperature for the extraction and determination of non-psychoactive cannabi-
was 350 ◦ C, the drying gas flow was 10 L/min and the skimmer volt- noids in fibre-type C. sativa, commonly known as hemp, and derived
age was 40 V. For the negative ion mode, the MS conditions were products. Given the complex composition of natural products, a
the same as described above. crucial aspect for the accurate identification and reliable quantifica-
Data were acquired by Agilent 6300 Series Ion Trap LC/MS sys- tion of the analytes is the optimization of the separation conditions.
tem software (version 6.2). The mass spectrometer was operated From this perspective, RP-HPLC is the most frequently applied tech-
in the full-scan mode in the m/z range 200–1200. MS2 spectra were nique for the analysis of phenolic compounds in plant extracts
automatically performed with helium as the collision gas in the m/z [27,28]. In the case of hemp, RP-HPLC on C18 columns, with acidic
range 50–1500, by using the SmartFrag function. The SmartFrag or buffered MeOH or ACN used as the strong solvent eluent and
function automatically selects the precursor ion to be fragmented acidic H2 O as the weak one, is the technique of choice for the analy-
in order to provide richer MS2 spectra, and it eliminates the need sis of cannabinoids [4,8,16,18,19,22]. For acidification, 0.1% HCOOH
for a time-consuming manual collision-induced dissociation (CID) or trifluoroacetic acid is usually employed [7,18,19,23], while dif-
voltage optimization. ferent buffers in the organic phase (e.g. ammonium formate and
232 V. Brighenti et al. / Journal of Pharmaceutical and Biomedical Analysis 143 (2017) 228–236

mAU
A MeOH EtOH Acetone MeOH/CHCl3 9:1 Hexane
700 1 30

600
25 c
b b b
500
20 a
400
4

mg/g
300 15

200
10

100
2 3 5
c d b b,c a
0
0 5 10 15 20 25 min b,c c b b,c a
mAU 0
B CBDA CBGA CBD
350 3
Fig. 4. Comparison of the extraction yield (mg/g) of hemp components (sample
300
C1), by using different extraction solvents. Data shown are means ± SD. Mean values
250 marked with the same letter are not significantly different from each other (p > 0.05).

2
200
nature of cannabinoids, alcohols such as MeOH and EtOH were con-
150 sidered as the more suitable extraction solvents, together with less
4
polar solvents such as acetone, MeOH/CHCl3 9:1 (v/v) and hexane,
100
in agreement with the literature [7,18,23]. As regards the optimiza-
50 tion of sample-to-solvent ratio and time, different conditions were
1
0
tested and the optimal ones were found to be 1:100 (w/v) for an
0 5 10 15 20 25 min overall extraction time of 45 min by DM.
As shown in Fig. 4, there was no marked difference between the
Fig. 3. Chromatograms obtained through HPLC-UV/DAD analysis of hemp ethanolic solvents tested in this study in terms of cannabinoid yield. In partic-
extracts from sample C1 (A) and C4 (B) at 210 nm. Experimental conditions as in
ular, only a slight or no significant difference was observed among
Section 2.5. For peak identification, see Table 2.
MeOH, MeOH/CHCl3 9:1 (v/v) and acetone, in terms of extraction of
the main compounds. In general, more polar solvents are preferable
ammonium acetate) have also been described [7,21]. The use of an to non-polar solvents such as hexane, especially for the extraction
acidic mobile phase improves separation performance, providing of cannabinoic acids (Fig. 4). As a matter of fact, hexane was found
better peak shape and enhanced resolution especially in the case of to be the worst performer in terms of recovery of all cannabinoids
cannabinoic acids, along with the enhancement of their ionization considered in this study, while EtOH was shown to possess the most
in HPLC-ESI–MS and MS2 experiments [18,19]. appropriate polarity for these compounds (Fig. 4). For this reason,
In this study, the chromatographic performance of four columns, it was eventually selected.
including a fully porous C18 (Zorbax C18 SB, 150 × 4.6 mm I.D., As regards extraction method optimization, several techniques
5 ␮m, Supelco), a PFP (Discovery HS F5, 150 × 4.6 mm I.D., 5 ␮m, were considered in this study, including DM, UAE, MAE and SFE,
Sigma-Aldrich), a PEG (Discovery HS PEG, 150 × 4.6 mm I.D., which were evaluated in order to obtain a high yield of non-
5 ␮m, Sigma-Aldrich) and a fused-core C18 (Ascentis Express C18 , psychoactive cannabinoids. DM is an extraction process of plant
150 × 3.0 mm I.D., 2.7 ␮m, Supelco), was evaluated. For our pur- material based on the use of a solvent with several daily shak-
pose, a mobile phase composed of H2 O and ACN, both containing ings or stirring at room temperature, while UAE and MAE are
0.1% HCOOH, was selected. Despite the acceptable separation procedures that use sound-waves or microwaves to accelerate the
achieved by using an isocratic elution, the use of a gradient allowed extraction and increase the yield of the secondary metabolites [29].
for a better separation of the compounds of interest in a shorter SFE with CO2 is a promising alternative technique with the advan-
time. Considering the selectivity, symmetry, resolution and number tage of no flammability or toxicity issues, given the simple solvent
of theoretical plates of the main chromatographic peaks, the best removal [18,30]. Each extraction technique was optimized through
results were achieved by using the Ascentis Express C18 column. the appropriate adjustment of extraction time and temperature.
The fused-core stationary phase led to a considerable improvement Finally, the extraction techniques were compared under their opti-
of chromatographic performance (in terms of both resolution and mal conditions. For a reliable comparison, a quantitative analysis
sensitivity) compared to the others, and to a marked reduction of of the main cannabinoids in hemp inflorescences was carried out
solvent consumption, working at a flow-rate of 0.4 mL/min instead through HPLC. As shown in Fig. 5, there was no significant difference
of 1 mL/min. Therefore, it was selected for our study. Representa- between UAE and SFE, which provided the lowest amount of CBDA,
tive chromatograms of two hemp samples (C1 and C4) are shown CBD and CBGA. DM gave a significantly high amount of CBDA. The
in Fig. 3. highest content of CBD was found with MAE. As regards CBGA, no
significant difference was observed between DM and MAE. Over-
3.2. Optimization of the extraction conditions all, DM was the best technique for the acidic cannabinoid CBDA,
whereas MAE was for its neutral counterpart CBD. It should be
The optimization of an effective extraction technique of pointed out that MAE requires a very low extraction time with
cannabinoids from hemp plays a key role in the outcome of the respect to DM. However, in the case of MAE, the increase in CBD
overall analytical method. Therefore, several extraction solvents was accompanied by a decrease in CBDA, which suggests a partial
and techniques were evaluated in this work. conversion of the cannabinoic acid into its decarboxylated counter-
Solvent selection, which is typically influenced by the polarity part. This is due to the high extraction temperature reached with
of the secondary metabolites in the plant material, is crucial to the MAE, thus confirming the critical role played by temperature in the
development of the extraction method [29]. Owing to the polar extraction of these compounds. To optimise extraction recovery by
 V. Brighenti et al. / Journal of Pharmaceutical and Biomedical Analysis 143 (2017) 228–236 233

Table 2
Retention times, UV, MS and MS2 data of non-psychoactive cannabinoids detected in hemp extracts

Peak number Compound tR (min) UV ␭max (nm) [M + H]+ MS2 (m/z) [M − H]− MS2 (m/z)
(m/z) (m/z)

Base peak Secondary peaks Base peak Secondary peaks

1 CBDA 13.8 225, 269, 307 359 341 285 (33), 261 (55), 357 – 339 (100), 313 (9)
233 (22), 219 (100)
2 CBGA 15.7 223, 268, 305 361 343 261 (49), 219 (100) 359 – 341 (100), 315 (6)
3 CBG 17.1 207, 231sh, 273 317 – 261 (39), 233 (40), 315 – 297 (49), 271 (100),
207 (100) 247 (51), 204 (80)
4 CBD 17.9 209, 232sh, 276 315 – 259 (100), 233 (26), 313 – 245 (100), 210 (6)
221 (24), 207 (24)

Experimental conditions as described in Section 2.5 and 2.6.

DM UAE MAE SFE As regards the cannabinoic acids CBDA and CBGA, they tended
30 to generate sodium adduct [M + Na]+ ions in the MS spectra [19],
together with their pseudo-molecular [M + H]+ ions. CBGA was
25 c easily distinguished from CBDA, thanks to its different molecular
b weight. In the MS2 spectra carried out in the positive ion mode,
20 a a both CBDA and CBGA provided base peaks at m/z 341 and 343,
respectively, due to a loss of water ( 18 u). In addition, these com-
mg/g

15
pounds generated product ions at m/z 219 and 261, with similar
relative intensity. CBDA showed an additional product ion at m/z
10
285, attributable to the loss of a C4 H8 group [21,31] and one at
m/z 233, which were both absent in the MS2 spectrum of CBGA. As
5 c
b a a regards the experiments carried out in the negative ion mode, the
b a b a acids provided [2 M H + Na]– precursor ions (at m/z 737 for CBDA
0
CBDA CBGA CBD and 741 for CBGA, respectively), even though their intensity was
much lower than those of the pseudo-molecular [M H]–– ions.
Fig. 5. Comparison of the extraction yield (mg/g) of hemp components (sample C1), Both CBDA and CBGA generated product ions at m/z 339 and 341,
by using different extraction techniques. Data shown are means ± SD. Mean values
respectively, corresponding to the loss of water ( 18 u), and at m/z
marked with the same letter are not significantly different from each other (p > 0.05).
313 and 315, respectively, attributable to the loss of a CO2 ( 44 u).
Both the decarboxylated cannabinoids CBG and CBD gave prod-
uct ions at m/z 233 and 207 in the MS2 experiments performed in
means of DM, three repeated extractions were preferred to one with
the positive ion mode, although their relative intensity was differ-
all the volume of the solvent. By taking into account all the above
ent (Table 2). CBD also showed a product ion at m/z 259, attributable
information, DM with EtOH as the extraction solvent at room tem-
to the loss of a C4 H8 group [21,31], which was absent in the MS2
perature for an overall time of 45 min was finally selected as the
spectrum of CBG. As regards the MS2 spectra acquired in the neg-
best extraction technique for non-psychoactive cannabinoids from
ative ion-mode, CBG generated a product ion at m/z 247, due to
hemp.
the loss of a C5 H8 group. Additional fragments were observed at
m/z 297, 271 and 204. The fragmentation of CBD in the negative
3.3. Identification of non-psychoactive cannabinoids in hemp ion-mode provided almost exclusively one product ion at m/z 245,
attributable to the loss of a C5 H8 group, following a retro Diels-
The identification of non-psychoactive cannabinoids in hemp Alder reaction, which is a common fragmentation pathway in the
extracts was carried out by combining the information obtained field of natural substances when an ESI ion source is employed [31].
from UV/DAD and MS detectors. UV/Vis, MS and MS2 data of all
compounds identified are shown in Table 2. For a reliable character-
isation, the experimental data were compared with the literature 3.4. Method validation
[9,18,19,22,23] and with commercially available reference stan-
dards. The HPLC method developed in this work, together with the
The chromatographic peaks were preliminarily assigned to a extraction procedure, were validated by following the ICH guide-
chemical class according to their UV spectra. In fact, cannabinoids lines [26]. These guidelines are internationally accepted as needed
have different UV behaviour on the basis of their chemical struc- for method validation for the quality control of pharmaceutical
ture: cannabinoic acids (CBDA and CBGA) are characterized by three products. The guideline used was the ICH Q2 (R1), which includes
absorption maxima (␭max ), one stronger at 220–223 nm, a second linearity, sensitivity, precision, accuracy and specificity.
one at 266–270 nm and a third one around 305 nm. Decarboxylated Linearity over the concentration range tested was good for the
cannabinoids (CBD and CBG) show a first ␭max at 210–215 nm and compounds used as cannabinoid standards, having a r2 > 0.998, as
an additional one at 270 nm [8]. shown in Table 3.
As regards MS and MS2 , the experiments were carried out both The LOD values ranged from 0.5 to 0.8 ␮g/mL, while the LOQ
in the positive and in the negative ion mode. Even if the positive range was 1.8-2.5 ␮g/mL (Table 3), which indicates that the method
ion mode has been previously considered as the most suitable one is sensitive. The LOD and LOQ values obtained for the cannabinoids
for the target compounds [18,19,21], these ionised well both in the were comparable with those of other methods previously described
positive and in the negative ion mode under the conditions applied in the literature [23].
in the present study. In general, cannabinoic acids gave better sig- The accuracy of the analytical procedure was evaluated by using
nals in the negative ion-mode, while decarboxylated compounds the recovery test. The percentage recovery values, obtained by com-
ionised better in the positive ion mode. paring the results from samples and fortified samples, were found
234 V. Brighenti et al. / Journal of Pharmaceutical and Biomedical Analysis 143 (2017) 228–236

Table 3
Linearity and sensitivity data for cannabinoids used as hemp standards.a

Compound Linearity range (␮g/mL) Slope (a) Intercept (b) r2 LOD (␮g/mL) LOQ (␮g/mL)

CBDA 2.5–200.0 29.7 (±0.1) −11.7 (±12.3) 0.9996 0.8 2.5

CBGA 2.5–200.0 29.9 (±0.2) −21.5 (±17.9) 0.9991 0.8 2.5
CBG 2.5–200.0 58.1 (±0.4) 107.4 (±37.5) 0.9995 0.5 1.8
CBD 2.5–200.0 46.9 (±0.6) −11.5 (±55.2) 0.9986 0.7 2.3

Experimental conditions as in Section 2.5.

a
For each curve the equation is y = ax + b, where y is the peak area, x the concentration of the analyte (␮g/mL), a is the slope, b is the intercept and r2 the correlation
coefficient. Standard error (S.E.) values are given in parenthesis. The p value was < 0.0001 for all calibration curves.

Table 4
Percentage recovery data of cannabinoids used as hemp standards (sample C1).

Compound Original amount (mg) Spiked amount (mg) Mean determined amount (mg) Mean recovery (%) ± RSD (%) (n = 3)

CBDA. 2.92 0.70 3.44 74.17 ± 1.00

CBGA 0.10 0.11 0.20 90.57 ± 4.46
CBG <LOD 0.20 0.16 81.05 ± 2.73
CBD 0.33 0.43 0.71 90.84 ± 0.60

Experimental conditions as described in Section 2.5.

Table 5
Intra- and inter-day precision data for retention time (tR ) and peak area of cannabinoids from hemp inflorescences (sample C4).

Intra-day precision (n = 6, mean) Inter-day precision (n = 18, mean)

Day 1 Day 2 Day 3

tR (min) ± RSD Area tR (min) ± RSD Area tR (min) ± RSD Area tR (min) ± RSD Area
(%) (mAU × s) ± RSD (%) (mAU × s) ± RSD (%) (mAU × s) ± RSD (%) (mAU × s) ± RSD
(%) (%) (%) (%)

CBDA 13.6 ± 0.2 386.6 ± 1.3 13.7 ± 0.7 385.0 ± 1.0 14.0 ± 0.4 387.0 ± 1.1 13.8 ± 1.3 386.2 ± 1.1
CBGA 15.5 ± 0.2 3441.5 ± 1.5 15.6 ± 0.8 3415.9 ± 0.6 15.9 ± 0.4 3402.6 ± 0.9 15.7 ± 1.4 3420.0 ± 1.1
CBG 17.0 ± 0.2 4359.7 ± 1.4 17.0 ± 0.7 4350.5 ± 0.5 17.4 ± 0.4 4343.3 ± 0.8 17.1 ± 1.2 4351.1 ± 0.9
CBD 17.5 ± 0.1 1726.0 ± 1.1 17.6 ± 0.5 1732.4 ± 1.3 17.9 ± 0.9 1713.4 ± 1.3 17.9 ± 1.0 1724.1 ± 1.3

Experimental conditions as described in Section 2.5.

Table 6 By taking into account all the information described above, it can
Intra- and inter-day precision data for the extraction of cannabinoids from hemp
be concluded that this method is a reliable tool for the identification
inflorescences (sample C1).
and quantification of non-psychoactive cannabinoids in fibre-type
Intra-day precision (n = 6, mean) Inter-day precision hemp, conforming to the ICH guidelines.
(n = 18, mean)

Day 1 Day 2 Day 3

3.5. Quantitative analysis of fibre-type hemp inflorescences and
mg/g ± SD mg/g ± SD mg/g ± SD mg/g ± SD hemp-based pharmaceutical products
CBDA 22.5 ± 0.3 23.3 ± 1.4 23.8 ± 1.4 23.2 ± 1.2
CBGA 0.7 ± 0.1 0.9 ± 0.1 0.8a 0.8 ± 0.1 The method developed in this study was applied to the quali-
CBG <LOQ <LOQ <LOQ <LOQ
tative and quantitative analysis of non-psychoactive cannabinoids
CBD 3.0a 3.2 ± 0.2 3.4 ± 0.2 3.2 ± 0.2
in nine varieties of hemp inflorescences. Quantitative data were
Experimental conditions as described in Section 2.5.
a
expressed as mg/g (dry weight) (Table 7). Variability in the content
SD < 0.05.
of the compounds of interest was observed among the samples ana-
lysed in this study. Indeed, it has to be taken into account that there
to be higher than 74% and they can be considered satisfactory is great genetic variability in fibre-type C. sativa plants, with a con-
(Table 4). sequently different composition of bioactive compounds [1,4]. In
The low intra- and inter-day %RSD for retention times (≤ 1.4%) addition, the content of cannabinoids may vary considerably with
and peak area (≤ 1.5%) relative to the target cannabinoids (Table 5) soil and climatic conditions at the place of cultivation [1,3,4].
and their low intra- and inter-day SD values for content (≤ 1.4 mg/g) The majority of the samples analysed showed CBDA as the most
(Table 6) indicate the high precision of both the chromatographic abundant compound, in agreement with the literature [3], with a
system and the extraction procedure. content variable from 0.1 to 46.8 mg/g. In particular, samples C1,

Table 7
Content (mg/g) of cannabinoids in hemp inflorescences (C1-C9) by HPLC-UV/DAD.a

C1 C2 C3 C4 C5 C6 C7 C8 C9

CBDA 23.2 ± 1.2 14.6 ± 1.5 16.2 ± 1.8 0.9 b

2.9 ± 0.4 5.9 ± 0.1 24.8 ± 0.5 46.8 ± 0.6 0.1b
CBGA 0.8 ± 0.1 0.4 ± 0.1 0.4b 9.8 ± 0.2 0.4b 0.1b 1.5 ± 0.2 1.6b 0.1b
CBG <LOQ <LOQ <LOQ 6.5 ± 0.1 0.2b <LOQ 1.3 ± 0.1 <LOQ <LOQ
CBD 3.2 ± 0.2 4.4 ± 0.4 3.4 ± 0.3 2.9 ± 0.1 2.2 ± 0.3 4.6 ± 0.1 23.9 ± 0.4 9.7 ± 1.4 0.1b

Experimental conditions as described in Section 2.5.

a
Data are expressed as mean (n = 6) ± SD. For sample C1, n = 18.
b
SD < 0.05.
 V. Brighenti et al. / Journal of Pharmaceutical and Biomedical Analysis 143 (2017) 228–236 235

Table 8 4. Conclusions
Content (mg/mL) of cannabinoids in hemp oil by HPLC-UV/DAD.a

Compound Oil 1 Oil 2 A new, reliable and efficient RP-HPLC method with UV/DAD,
CBDA 3.5 ± 0.5 <LOD ESI–MS and MS2 detection was developed for the qualitative and
CBGA <LOD <LOD quantitative analysis of non-psychoactive cannabinoids in fibre-
CBG 0.7 ± 0.2 <LOQ type C. sativa plant material and pharmaceutical products. The
CBD 78.6 ± 3.7 3.5 ± 0.1 fused-core technology of the stationary phase was applied for the
Experimental conditions as in Section 2.5. first time in this study to the cannabinoid profiling of this Cannabis
a
Data are expressed as mean (n = 6) ± SD. species. Its application allowed us to achieve improvements in
chromatographic performance in comparison with conventional
Table 9 particulate stationary phases, thus securing a good separation of
Content (mg/g) of cannabinoids in hemp balm and extract by HPLC-UV/DAD.a all constituents in a shorter time and with lower solvent usage.
A simple extraction procedure for non-psychoactive cannabi-
Compound Balm 1 Balm 2 Balm 3 Extract
noids from hemp was also optimized in this work, by using dynamic
CBDA 80.4 ± 0.4 44.7 ± 0.2 78.0 ± 3.2 <LOQ maceration with EtOH at room temperature.
CBGA 3.9 ± 0.1 2.0 ± 0.1 3.8 ± 0.3 <LOD
CBG 0.4 ± 0.1 <LOQ 0.4b <LOQ
The overall analytical method was fully validated in agreement
CBD 13.8 ± 0.1 7.6 ± 0.2 13.6 ± 0.5 193.7 ± 3.5 with international guidelines and then successfully applied to the
characterisation of nine hemp samples and six hemp-based prod-
Experimental conditions as described in Section 2.5.
a
Data are expressed as mean (n = 6) ± SD. For sample C1, n = 18. ucts, therefore proving a powerful and reliable tool for both the
b
SD < 0.05. selection of hemp varieties with a high content of bioactive com-
pounds and the quality control of its derivatives. This research also
shows the interesting potential that fibre-type C. sativa inflores-
C7 and C8 showed the highest amount of this compound (23.2, cences may have in agriculture, in relation to the preparation of
24.8 and 46.8 mg/g, respectively), while sample C9 had the lowest certified extracts with a high content of non-psychoactive cannabi-
content (0.1 mg/g). noids to be used in the pharmaceutical and nutraceutical ambit.
CBD was found to be the second main cannabinoid in almost all
the samples analysed (0.1-23.9 mg/g), with a content 4–10 times
lower than its acid form CBDA. These data were consistent with Acknowledgments
those described in the literature [3,7,9,19,23].
Samples C5, C6 and C7 showed a nearly 1:1 ratio between CBDA The work on the optimization of the extraction conditions of
and its neutral counterpart CBD. This fact is noteworthy, since cannabinoids from the plant material was carried out thanks to a
cannabinoids are biosynthesised in the plant material as acids. research grant awarded to Marleen Steinbach for her master the-
Thus, the decarboxylated compounds should be present in lower sis at the Department of Life Sciences of the University of Modena
amounts than their acidic precursors [1]. and Reggio Emilia (Italy) within the Erasmus+ program, under the
CBGA is the first biogenic cannabinoid produced in the plant. scientific supervision of Dr. Federica Pellati.
It represents the parent compound from which all other cannabi- The authors are grateful to Dr. Gianpaolo Grassi from CREA-
noids are bio-synthesised [1,3,9,15]. For this reason, it was detected CIN center (Rovigo, Italy), Farms Sandri (Faedo, TN, Italy), Marò
in low amounts in almost all the samples considered in this study (Suvereto, LI, Italy) and Formaggio (Rovigo, Italy) for providing the
(≤ 1.6 mg/g). Consequently, CBG was below the LOQ value in the fibre-type hemp samples.
majority of samples. Sample C4 was the only one with a differ- The authors are also grateful to Dr. Massimo Tacchini and Prof.
ent trend. As a matter of fact, the major cannabinoid present in Gianni Sacchetti (Department of Life Sciences and Biotechnology,
this sample was CBGA (9.8 mg/g), followed by its neutral counter- University of Ferrara, Italy), for allowing the use of the supercritical
part CBG (6.5 mg/g). In this specific sample, the content of CBD was fluid extractor in their lab.
higher than CBDA, in contrast to what was observed for all the other
samples analysed.
In the light of the promising biological activities described in the References
Introduction for CBD, samples C1, C7 and C8 represent a potentially
good source of CBD-type compounds to be used for the prepara- [1] J.A. Hartsel, J. Eades, B. Hickory, A. Makriyannis, Cannabis sativa and hemp, in:
R.C. Gupta (Ed.), Nutraceuticals: Efficacy, Safety and Toxicity, Elsevier Inc,
tion of bioactive extracts. On the other hand, sample C4 showed an London, 2016, pp. 735–754.
interesting panel of CBG-type cannabinoids, which make it a poten- [2] C.M. Andre, J.F. Hausman, G. Guerriero, Cannabis sativa: the plant of the
tial source of other biologically active compounds to be further thousand and one molecule, Front. Plant. Sci. 7 (2016) 1–17.
[3] G. Appendino, G. Chianese, O. Taglialatela-Scafati, Cannabinoids: occurrence
investigated for pharmaceutical applications. These results also and medicinal chemistry, Curr. Med. Chem. 18 (2011) 1085–1099.
demonstrate that inflorescences, which are usually a by-product [4] B.F. Thomas, M.A. ElSohly, The Analytical Chemistry of Cannabis: Quality
of fibre-type hemp, can be considered as a new product for the Assessment, Assurance and Regulation of Medicinal Marijuana and
Cannabinoid Preparations, 1st ed., Elsevier Amsterdam, Oxford, Waltham,
agricultural sector involved in hemp cultivation. 2015.
As regards the analysis of hemp pharmaceutical products, the [5] European Commission, Food safety, plant variety catalogues, databases and
results are shown in Tables 8 and 9. In the case of hemp oil information systems, https://ec.europa.eu/food/plant/
plant propagation material/plant variety catalogues databases en, 2017
(Table 8), great variability was observed for the two samples ana-
(accessed 27.01.17).
lysed in this study. The main component detected was CBD in both [6] R. Mechoulam, L.O. Hanuš, R. Pertwee, A.C. Howlett, Early phytocannabinoid
cases (78.6 mg/mL for oil 1 and 3.5 mg/mL for oil 2, respectively). chemistry to endo-cannabinoids and beyond, Nat. Rev. Neurosci. 15 (2014)
Hemp balm samples were found to be highly rich in CBDA (80.4- 757–764.
[7] B. De Backer, B. Debrus, P. Lebrun, L. Theunis, N. Dubois, L. Decock, A.
44.7 mg/g), followed by CBD (13.8-7.6 mg/g) (Table 9). The hemp Verstraete, P. Hubert, C. Charlier, Innovative development and validation of
extract with the syringe-based oral applicator was characterized an HPLC/DAD method for the qualitative and quantitative determination of
by a high content of CBD (193.7 mg/g). A decarboxylation process major cannabinoids in Cannabis plant material, J. Chromatogr. B 877 (2009)
4115–4124.
of CBDA into CBD might have occurred for both hemp oil and extract [8] W. Peschel, M. Politi, 1 H NMR and HPLC/DAD for Cannabis sativa L. chemotype
during the extraction or during storage over time. distinction, extract profiling and specification, Talanta 140 (2015) 150–165.
236 V. Brighenti et al. / Journal of Pharmaceutical and Biomedical Analysis 143 (2017) 228–236

[9] L.O. Hanuš, S.M. Meyer, E. Muñoz, O. Taglialatela-Scafati, G. Appendino, [21] A.A.M. Stolker, J. van Schoonhoven, A.J. de Vries, I. Bolbedijk-Pastorova, W.H.J.
Phytocannabinoids: a unified critical inventory, Nat. Prod. Rep. 33 (2016) Vaes, R. van den Berg, Determination of cannabinoids in Cannabis products
1357–1392. using liquid chromatography-ion trap mass spectrometry, J. Chromatograph.
[10] M.A. ElSohly, D. Slade, Chemical constituents of marijuana: the complex A 1058 (2004) 143–151.
mixture of natural cannabinoids, Life Sci. 78 (2005) 539–548. [22] A. Hazekamp, A. Peltenburg, R. Verpoorte, C. Giroud, Chromatographic and
[11] S.P.H. Alexander, Therapeutic potential of cannabis-related drugs, Prog. spectroscopic data of cannabinoids from Cannabis sativa L, J. Liq.
Neuropsychopharmacol. Biol. Psychiatry 64 (2016) 157–166. Chromatograph. Relat. Technol. 28 (2005) 2361–2382.
[12] A.C. Campos, M.V. Fogaça, A.B. Sonego, F.S. Guimarães, Cannabidiol, [23] W. Gul, S.W. Gul, M.M. Radwan, A.S. Wanas, Z. Mehmedic, I.I. Khan, M.H.M.
neuroprotection and neuropsychiatric disorders, Pharmacol. Res. 112 (2016) Sharaf, M.A. ElSohly, Determination of 11 cannabinoids in biomass and
119–127. extracts of different varieties of Cannabis using high-performance liquid
[13] J. Fernàndez-Ruiz, M. Moreno-Martet, C. Rodrìguez-Cueto, C. Palomo-Garo, M. chromatography, J. AOAC Int. 98 (2015) 1523–1528.
Gòmez-Caña, S. Valdeolivas, C. Guaza, J. Romero, M. Guzmàn, R. Mechoulam, [24] J.J. Kirkland, S.A. Schuster, W.L. Johnson, B.E. Boyes, Fused-core particle
J.A. Ramos, Prospects for cannabinoid therapies in basal ganglia disorders, Br. technology in high-performance liquid chromatography: an overview, J.
J. Pharmacol. 163 (2011) 1365–1378. Pharm. Anal. 5 (2013) 303–312.
[14] A.A. Izzo, F. Borrelli, R. Capasso, V. Di Marzo, R. Mechoulam, Non-psychotropic [25] J. Omar, M. Olivares, M. Alzaga, N. Extebarria, Optimisation and
plant cannabinoids: new therapeutic opportunities from an ancient herb, characterization of marijuana extracts obtained by supercritical fluid
Trends Pharmacol. Sci. 30 (2009) 515–527. extraction and focused ultrasound extraction and retention time locking
[15] R. Brenneisen, Chemistry and analyss of phytocannabinoids andother GC–MS, J. Sep. Sci. 36 (2013) 1397–1404.
Cannabis constituents, in: M.A. ElSohly (Ed.), Marijuana and the Cannabinoids, [26] International Conference on Harmonization of Technical Requirements for
Humana Press Inc., Totowa, 2007, pp. 17–49. the Registration of Pharmaceuticals for Human Use (ICH), Guideline
[16] UNODC, Recommended methods for the identification and the analysis of Q2(R1)-Validation of Analytical Procedures: Text and Methodology, ICH
Cannabis and Cannabis products. Manual for use by national drug analysis Secretariat,c/o IFPMA, Geneva, 2005, pp. 1.
laboratories. United Nations, 2009, pp. 25–47. [27] X. Qiao, W. Yang, D. Guo, M. Ye, Extraction, separation, detection, and
[17] J.T. Fishedick, A. Hazekamp, T. Herkelens, Y.H. Choi, R. Verpoorte, Metabolic structural analysis of flavonoids, Curr. Org. Chem. 15 (2011) 2541–2566.
fingerprinting of Cannabis sativa L., cannabinoids and terpenoids for [28] A. Khoddami, M.A. Wilkes, T.H. Roberts, Techniques for analysis of plant
chemotaxonomic and drug standardization purposes, Phytochemistry 71 phenolic compounds, Molecules 18 (2013) 2328–2375.
(2010) 2058–2073. [29] J. Azmir, I.S.M. Zaidul, M.M. Rahman, K.M. Sharif, A. Mohamed, F. Sahena,
[18] C. Cinti, G. Ciccarella, D. Braghiroli, C. Parenti, M.A. Vandelli, G. Cannazza, M.H.A. Jahurul, K. Ghafoor, N.A.N. Norulaini, A.K.M. Omar, Techniques for
Medicinal Cannabis: principal cannabinoids concentration and their stability extraction of bioactive compounds from plant materials: a review, J.Food Eng.
evaluated by a high-performance liquid chromatography coupled to diode 117 (2013) 426–436.
array and quadrupole time of flight mass spectrometry method, J. Pharm. [30] H. Perrotin-Brunel, M.C. Kroon, M.J.E. van Roosmalen, J. van Spronsen, C.J.
Biomed. Anal. 128 (2016) 201–209. Peters, G.-J. Witkamp, Solubility of non-psychoactive cannabinoids in
[19] O. Aizpurua-Olaizola, J. Omar, P. Navarro, M. Olivares, N. Etxebarria, A. supercritical carbon dioxide and comparison with psychoactive cannabinoids,
Usobiaga, Identification and quantification of cannabinoids in Cannabis sativa J. Supercrit. Fluids 55 (2010) 603–608.
L. plants by high-performance liquid chromatography-mass spectrometry, [31] D.P. Demarque, A.E.M. Crotti, R. Vessecchi, J.L.C. Lopes, N.P. Lopes,
Anal. Bioanal. Chem. 406 (2014) 7549–7560. Fragmentation reactions using electrospray ionization mass spectrometry: an
[20] V. Gambaro, L. Dell’Acqua, F. Farè, R. Froldi, E. Saligari, G. Tassoni, important tool for the structural elucidation and characterization of synthetic
Determination of primary active constituents in Cannabis preparations by and natural products, Nat. Prod. Rep. 33 (2016) 432–455.
high-resolution gas chromatography/flame ionization detection and
high-performance liquid chromatography/UV detection, Anal. Chim. Acta 468
(2002) 245–254.