Journal Pre-proof

Analysis of endocannabinoids in plasma samples by biocompatible solid-phase

microextraction devices coupled to Mass Spectrometry

Vinicius R. Acquaro Junior, Germán Augusto Goméz-Ríos, Marcos Tascon, Maria

Eugênia Costa Queiroz, Janusz Pawliszyn

PII: S0003-2670(19)31045-1
DOI: https://doi.org/10.1016/j.aca.2019.09.002
Reference: ACA 237054

To appear in: Analytica Chimica Acta

Received Date: 22 May 2019

Revised Date: 30 August 2019
Accepted Date: 1 September 2019

Please cite this article as: V.R Acquaro Junior, G.A. Goméz-Ríos, M. Tascon, M.E. Costa Queiroz,
J. Pawliszyn, Analysis of endocannabinoids in plasma samples by biocompatible solid-phase
microextraction devices coupled to Mass Spectrometry, Analytica Chimica Acta, https://doi.org/10.1016/
j.aca.2019.09.002.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2019 Elsevier B.V. All rights reserved.

 Analysis of endocannabinoids in plasma samples by biocompatible
solid-phase microextraction devices coupled to Mass Spectrometry

Vinicius R Acquaro Junior1,2, Germán Augusto Goméz-Ríos2,3, Marcos Tascon2,4,

Maria Eugênia Costa Queiroz1, Janusz Pawliszyn2✉

1
Departamento de Química, Faculdade de Filosofia Ciências e Letras de Ribeirão Preto,
Universidade de São Paulo, São Paulo, Brazil.
2
Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada
3
Restek Corporation, Bellefonte, Pennsylvania, United States of America, current
affiliation
4
Instituto de Investigación e Ingenieria Ambiental (3iA), Universidad Nacional de San
Martín (UNSAM), San Martín, Buenos Aires, Argentina, current affiliation

Correspondence: Professor Janusz Pawliszyn, Department of Chemistry, University of

Waterloo, Waterloo, Ontario, N2L 3G1 Canada
E-mail: janusz@uwaterloo.ca

1
ABSTRACT

Anandamide (AEA) and 2-arachidonoyl glycerol (2-AG) represent two of the most
important endocannabinoids (ECs) investigated in neurobiology as therapeutic targets
for several mental disorders. However, the determination of these ECs in biological
matrices remains a challenging task because of the low concentrations, low stability and
high protein-bound (LogP ~ 6). This work describes innovative analytical methods
based on biocompatible SPME (Bio-SPME), SPME-UHPLC-MS/MS and Bio-SPME-
Nano-ESI-MS/MS, to determine AEA and 2-AG in human plasma samples. The direct
coupling of Bio-SPME with nano-ESI-MS/MS can be considered an alternative tool for
faster analysis.
Different Bio-SPME fibers based on silica and polymeric coating (i.e. C18, C30, and
HLB) were evaluated. Different desorption solvents based on combinations of
methanol, acetonitrile, and isopropanol were also evaluated for efficient elution with
minimum carry-over. Given the high protein binding analytes and the fact that SPME
extracts the free-concentration of the analytes, the plasma samples were modified with
additives such as guanidine hydrochloride (Gu-HCl), trifluoroacetic acid, and
acetonitrile. This study was carried out by experimental design to achieve complete
protein denaturation and the release of target analytes. The maximum extraction
efficiency was obtained under the following conditions: HLB coated fibers (10 mm
length, 20 µm coating thickness), matrix modified (300 µL of plasma) with 50 µL of
Gu-HCL 1 mol L-1, 75 µL of ACN and 75 µL of water, and desorption with
methanol/isopropanol solution (50:50, v/v). Both methods were validated based on
current international guidelines and can be applied for monitoring of concentrations of
endocannabinoids in plasma samples.
SPME-UHPLC-MS/MS method presented lower LOQ values than SPME-nanoESI-
MS/MS. The additional separation (chromatographic column) favored the detectability
of LC-MS/MS method. However, the SPME-nano-ESI-MS/MS decrease the total
analysis time, due to significant reductions in desorption and detection times.

Key words: Endocannabinoids, biofluid, SPME, nano-ESI, human plasma, MS/MS

2
1. INTRODUCTION

Endocannabinoids are defined as endogenous ligands of cannabinoid receptors

(CB1 and CB2). As the role of the endocannabinoid system in regulating several
physiological conditions and neurological diseases has become clearer in recent years as
a therapeutic target for a series of neurological disorders, including multiple sclerosis [1,
2]], Alzheimer’s disease [1, 3, 4], Parkinson’s disease [5-7], and Schizophrenia [8-10],
it has garnered increased interest within the scientific community, spurring the demand
for efficient analytical methods for monitoring of endocannabinoids in biomatrices.
In this respect, several studies have reported analysis of anandamide (AEA) and
2-arachidonyl glycerol (2-AG) in human biofluids. In general, AEA and 2-AG can be
analyzed via gas chromatography coupled to mass spectrometry (GC-MS) or liquid
chromatography coupled with tandem mass spectrometry (LC-MS/MS) [11, 12]. Due to
the high sensitivity and selectivity associated with LC-MS/MS liquid chromatography
coupled with tandem mass spectrometry, this technique have been frequently used for
quantitative determination of endocannabinoids [11, 12] . Regarding sample preparation
processes reported in the literature, the most frequent approaches involve solid phase
extraction (SPE) [13, 14], liquid-liquid extraction (LLE) [15, 16], protein precipitation
(PPT) [17, 18], salting out liquid-liquid extraction (SALLE) [19].
Most methods described in the literature for analysis of endocannabinoids are
based on more than one sample preparation step which, beyond the large volumes
required, are time consuming. In addition, as LLE and PPT are not selective techniques,
resulting extracts often contain high concentrations of matrix interferences such as
unprecipitated proteins, salts, and phospholipids that can hinder MS analysis (ion
suppression or enhancement). While SPE provides satisfactory selectivity, it requires
large volumes of biological sample and solvent. As well, SPE can be hindered by
clogging in cases where samples are not clean enough. In these cases, the sample
preparation techniques LLE, PPT, and SPE are often time-consuming and might present
relatively poor reproducibility [20, 21]. Also, some micro-extraction techniques such as
micro-solid phase extraction (µ-SPE) [22], column switching (CS) [23], and in-tube
SPME [24] have been reported. Although the time of analysis and the sample
consumption is reduced drastically, these methods deal with a packed bed of sorbent
which can cause clogging limiting the type of sample. Further, in some cases an extra
sample preparation step is required [22]. Additionally, determination of AEA and 2-AG

3
in biological matrices remains a challenging task given that (1) these compounds are
normally present in low concentrations; (2) some of these endogenous compounds
(ECs) possess low chemical stability; and (3) only low volumes of biological samples
can be collected for analysis [11, 25, 26]. In addition, 2-AG can rapidly isomerize both
in vitro and in vivo into 1-arachidonyl glycerol (1-AG). 1-AG presents the same
molecular weight as 2-AG, and is formed due to an acyl migration [11]. This
isomerization process has been observed in fresh collected plasma samples preserved
for 2 h at 4ºC or room temperature, and in samples stored for 20 days at -80ºC [27].
Several factors can accelerate the isomeric conversion of 2-AG into 1-AG in biological
samples, including elevated temperatures, high pH, polar protic solvents, and the
presence of serum albumin [11, 28, 29].
To overcome these issues, a sample preparation technology capable of
efficiently enriching target analytes while stabilizing ECs after collection is highly
desired. Efficient cleanup of the sample can drastically minimize matrix effects,
enhance the sensitivity of the method, and prevent damage to the chromatographic
system, particularly the stationary phase of the analytical column [30].
In this context, biocompatible solid phase microextraction (Bio-SPME) emerges
as a suitable technology for sampling of complex matrices [31, 32], since it prevents
biological macromolecules (e.g., proteins and phospholipids) sorption into coating
(sorbent), and selective extraction by the direct immersion of the fiber into the sample
[20, 33, 34].
Among the different strategies available for direct coupling of SPME with mass
spectrometry [32], nano-ESI was initially proposed by Walles et al. with the aim of
minimizing total analysis time while enhancing the method sensitivity. Due to the
potential of such approaches in bioanalysis, different innovations have since been
introduced [35-39]. In fact, direct insertion of the SPME fiber into the nanoelectrospray
(nano-ESI) emitter for desorption and ionization of target analytes represents an
innovative advance in bioanalysis [37], as it allows for desorption of analytes pre-
concentrated in the biocompatible phase using only a small volume of solvent inside the
emitter (Vdes <10 µL). Owing to these features, Bio-SPME-nano-ESI-MS/MS methods
have been shown to provide faster analysis times than LC-MS/MS methods while
delivering adequate limits of quantification for a range of molecules of interest [32, 37,
40].

4
 In this study, an innovative bio SPME-UHPLC-MS/MS method is introduced for
analysis of the endocannabinoids AEA and 2-AG in plasma samples. In addition, this
work also presents the evaluation of the direct coupling of Bio-SPME to nano-ESI-
MS/MS as an alternative tool aimed at faster analysis of AEA and 2-AG. A matrix
modifier solution was optimized to denature plasma sample with aims of increasing the
free concentrations of the targeted compounds, thus facilitating their collection by the
SPME fiber [41]. The methods were validated based on current international guidelines
issued by EMA (European Medicines Agency) and FDA (Food and Drug
Administration).

2. MATERIALS AND METHODS

2.1 Chemicals and solutions

Acetic acid (AA), trifluoroacetic acid (TFA), guanidine hydrochloride (Gu-

HCL), albumin, sodium chloride, potassium chloride, sodium phosphate, and potassium
phosphate were purchased from Sigma-Aldrich (Saint-Louis, MO, USA). LC-MS-grade
acetonitrile (ACN), ethanol, isopropanol (IPA), and water were purchased from Fisher
Scientific (Hampton, NH, USA). The endocannabinoids standards 2-Arachidonoyl
Glycerol (2-AG), 1- Arachidonoyl Glycerol (1-AG), and Arachidonoyl Ethanolamide
(Anandamide – AEA) were purchased from Cayman Chemical (Ann Arbor, MI, USA).
The internal standards (ISs), 2-Arachidonoyl Glycerol-d5 (2-AG-d5), 1- Arachidonoyl
Glycerol-d5 (1-AG-d5), and Arachidonoyl Ethanolamide-d4 (Anandamide-d4 – AEA-d4)
were also acquired from Cayman Chemical (Ann Arbor, MI, USA). Individual stock
solutions were prepared for each endocannabinoid. Both anandamide and anadamide-d4
were prepared in ethanol at concentrations of 1000 µg·mL-1, while 2-AG and 2-AG-d5
were prepared in ACN at 50 µg mL-1 and 500 µg·mL-1, respectively. All stock solutions
were stored in a -80 ºC until use. A phosphate-buffered saline (PBS) with albumin was
prepared in a concentration of 36 mg·mL-1 (experimental procedure detailed in
supplementary material) [42].
Nitinol wire (200 µm diameter) was cut in pieces of 10 cm and then coated using
three different slurries of particles (~ 5 µm): C18, C30, and Hydrophilic-lipophilic
balance (HLB). The particles were mixed with polyacrylonitrile (PAN) obtained from
Sigma-Aldrich (Oakville, ON, Canada) according to the protocol described by Gomez-
Rios et al. [41, 43]. C30 and C18 particles were kindly provided by Millipore-Sigma

5
(Bellefonte, PA, USA), whereas HLB particles were kindly provided by Waters
Corporation (Wimslow, UK). SPME fibers with a coating thickness of 20 µm and a
coating length of 10 mm were used in all the experiments herein described.

2.2 Plasma samples

Human plasma (K2EDTA) was purchased from BioIVT (WestBurry, New York,
USA). Standard solutions were mixed with plasma and agitated for 3 hours (1200 rpm),
whereupon spiked plasma samples were stored overnight in a 4ºC fridge to ensure
endocannabinoid-protein biding equilibrium prior analysis.

2.3 Matrix Modifier optimization

Aiming to increase the endocannabinoids free concentration of

endocannabinoids in plasma, an experimental design central composite (CCD) 23 was
utilized to investigate an optimal matrix modifier [44, 45]. Investigated matrix modifier
components (factors) included TFA (10%), Gu-HCL (1 mol·L-1), and ACN in 5 levels.
Studied concentration ranges (levels) included 0.0 - 50 µL TFA, 0.0 - 50 µL Gu-HCL,
and 0.0 - 150 µL ACN. Water (LC-MS grade) was used in order to correct the final
volume of matrix modified (200 µL) to maintain the volume constant among all
extractions. The evaluated dependent variable (response) was the recovery of the target
endocannabinoids. In total, 17 entirely randomized experiments were performed,
including 8 in the factorial points (-1, +1); 6 in the axial points (-1.68, +1.68) and 3 in
the central point (0, 0).

2.4 SPME-UHPLC-MS/MS workflow

The entire analytical procedure was carried out in three simple steps. Extraction
of endocannabinoids was performed by DI-SPME in a 750-µL glass vial containing 300
µL of plasma sample and 200 µL of the optimized matrix modifier solution (50 µL of
Gu-HCL 1 mol·L-1, 75 µL of ACN and 75 µL of LC-MS grade water) for 60 min under
1500 rpm orbital agitation. Next, fibers were submitted to two sequential rinsing cycles
of 5s in 300-µL glass vials containing 300 µL water (LC-MS grade) under 1500 rpm
orbital agitation. Fibers were then desorbed for 60 min into a 1.5 mL glass vial with a

6
50 µL glass insert prefilled with 17 µL of methanol/isopropanol (50:50, v/v) under 1500
rpm orbital agitation.
The SPME-UHPLC-MS/MS workflow was carried out on a Thermo Dionex
UltiMate 3000 autosampler coupled to a heated electrospray (H-ESI) source and a TSQ
Quantiva triple quadrupole mass spectrometer (Thermo Fisher Scientific, San Jose,
California, USA). Data was processed using Trace Finder 3.3 (Thermo Fisher Scientific,
San Jose, California, USA). The employed liquid chromatographic method was based
on previous work by Gouveia-Figueira, et al. with modifications [46]. Separation was
performed in gradient mode using water with 0.1% of acetic acid as mobile phase A and
acetonitrile/isopropanol (90:10, v/v) as mobile phase B at a 0.3 mL·min-1 flow rate and
40 ºC. The gradient mode employed for the CORTECS C18 superficially porous column
(100 mm x 2.1 mm – 1.6 µm) and the CORTECS C18 superficially porous guard column
(5 mm x 2.1 mm – 1.6 µm) consisted of 0.0-0.5 min (40% of B), 0.5-5.0 min (79% of
B), 5.0-6.5 min (75% of B), 6.5-8.1 min (40% of B) and 8.1-10.0 min (40% of B).
Endocannabinoids were analyzed in positive ion mode, with spray voltage set at 3.5 kV,
ion transfer tube temperature of 333 ºC, vaporizer temperature of 317 ºC, collision-
induced dissociation gas set at 1.5 mTorr, and a dwell time of 100 ms. The arbitrary
units for the sheath gas, auxiliary gas, and sweep gas were 40, 12, and 1, respectively.
MS transitions and other additional parameters are listed in Table 1.

2.5 Bio-SPME-nano-ESI-MS/MS

Direct coupling analyses were performed in four main steps:

extraction/preconcentration, rinse, desorption, and ionization. For extraction, a
preconditioned Bio-SPME fiber was inserted into a 750-µL glass vial containing 300 µL
of plasma sample and 200 µL of matrix modifier solution (50 µL of Gu-HCL 1 mol·L-1,
75 µL of ACN and 75 µL of LC-MS grade water). Extraction and rinse steps were
performed as described in the previous section. Next, the fiber was rapidly inserted into
a coated glasstip (1.0/0.58 mm, o.d./i.d) emitter (New Objective Inc. Woburn, MA.
USA) prefilled with 4 µL of desorption solution (methanol/isopropanol – 50:50, v/v
with 0.1% acetic acid), whereupon desorption of analytes took place under static
conditions (5 min). Finally, an electrical potential of 2.1 kV was applied between the
emitter and the entrance of the MS system to start the ionization of analytes via

7
electrospray ionization [37]. Detection was performed by MS/MS; the parameters
utilized for MS/MS detection are described in Table 1.

3. RESULTS AND DISCUSSION

3.1 Chromatographic separation (LC-MS/MS)

The endocannabinoids were separated in 10 min with the following

chromatographic elution: anandamide (tR = 6.98 min), 2-AG (tR = 7.55 min) and 1-AG
(tR = 7.75 min). Although 2-AG and 1-AG isomers have the same molecular mass, their
polarities are slightly different, leading to successful separation with a resolution of 1.3
[16]. This successful chromatographic separation was enabled with C18 superficially
porous particle (1.7 µm) column which led to a short diffusion path, fast mass transfer
(C-term in the Van Deemter equation), and sharper peaks (peak width ≅ 11 s) [30]. C18
fully porous column (3 µm) was also evaluated for endocannabinoids chromatographic
separation. Although the elution gradient mode was optimized for this column, the
isomers separation 2-AG and 1-AG was not achieved (Figure 1). Indeed, separation of
these isomers has proven to be a challenging task, as noted by several authors who
could not achieve this ambitious goal [15, 23, 24, 29, 47, 48].
Although Gong et al. achieved separation of the targeted endocannabinoids
using columns of a larger particles size (> 2µm), chromatographic run time and
efficiency (i.e. peak broadening) were highly affected [16]. Gouveia-Figueira et al.
reported successful separation of the isomers 2-AG and 1-AG with the use of a column
consisted of sub-2 µm particles [46]. However, the column switching sample
preparation technique chosen for subsequent work by other authors negatively impacted
the expected resolution due to peak broadening, consequently impairing isomer
separation even when columns with sub-2 µm particles were used [23, 24]. Yet, isomer
separation was not objective of previous work conducted in our group. In this work, a
column consisted of sub-2 µm particles was selected for separation of the selected
isomers, and the sample preparation technique was optimized so as to avoid peak
broadening, thus increasing analytical method sensitivity.

3.2 Stability of 2-AG and AEA in organic solvents, effect of LC column

temperatures, and 2-AG/1-AG isomerization in plasma samples.

8
 The stability of the targeted endocannabinoids in organic solvents was evaluated
in different solvent compositions. Stock solutions containing 100 ng·mL-1 of both AEA
and 2-AG, as well AEA-d4 and 2-AG-d5, were prepared in different solvents. 1-AG was
excluded from this experiment in order to assess the stability of 2-AG in different
mixtures without introduction of any confounding variables related to the isomerization
of 2-AG into 1-AG. It should be highlighted that the purity of the 2-AG reagent utilized
in this work was 90%; hence, it should not come as a surprise that even at time zero, the
instrumental response attained for all solvent mixtures corresponded to 90 ng·mL-1. The
following solvents mixtures were evaluated: acetonitrile (100%), methanol (100%),
isopropanol (100%), acetonitrile/methanol (50:50, v/v), acetonitrile/isopropanol (50:50,
v/v), methanol/isopropanol (50:50, v/v) and water/acetonitrile (90:10, v/v). Each
solution was prepared in triplicate and injected into the UHPLC-MS/MS system for 18
hours. The solutions remained in the autosampler at 5 ºC throughout the entire
experiment.
The isomerization process has been widely discussed in the literature, with
results demonstrating that polar protic solvents favor the conversion of 2-AG to 1-AG
[18, 28, 49, 50]. As expected, the isomerization process of 2-AG was more pronounced
when polar protic solvents were used (see Figure S1); consequently, a continuous
increase of 1-AG was observed over time in those solvents, even though this compound
was not spiked in the solvent (see Figure S2). No decreases in concentration were
observed for AEA, with the exception of the solution composed of water/acetonitrile
(90:10, v/v) (see Figure S3). In this composition, both 2-AG and AEA presented a
decrease in concentration, which can be explained by their adherence to lab materials
over time due to their high hydrophobicity [51].
Furthermore, the stability of 2–AG and AEA was evaluated through analysis of a
standard solution (100 ng·mL-1) in acetonitrile at different temperatures during the
chromatographic separation process (25 ºC to 40 ºC). According to Zoerner et al., the
isomerization process can be influenced by temperature; in other words, increasing the
temperature can accelerate the conversion of 2-AG into 1-AG. As can be seen in Figure
S4, an analysis of variance (ANOVA) revealed no significant differences (p > 0.05)
among the different temperatures evaluated, thus supporting that the isomerization
process is constant throughout chromatographic separation, and can thus be ignored.
Given this, a temperature of 40 ºC was selected for subsequent analysis so that sharper
peaks could be attained. The peak width achieved at this temperature was 11 s for all

9
endocannabinoids. In order to prevent any damage to the chemistry of the particles in
the LC column, temperatures above this range were not evaluated.
As a matter of fact, according to Rouzer et al., it is reasonable to postulate that
the binding interaction of 2-AG with albumin may lead to catalysis of the acyl
migration reaction [28]. This phenomena was also reported by Zoerner et al. [11].
According to publications reporting the use of the Roswell Park Memorial Institute
medium (RPMI) [28] and Hank’s Balanced Salt Solution medium (HBSS) [50], 2-AG is
rapidly converted to 1-AG (3-20 min depending on sample characteristics). For
instance, the RPMI halftime of conversion for a sample was around 10 min without
serum and 2.3 min with serum. Likewise for HBSS, the halftime of conversion was 16 ±
4 min without serum and 9 ± 3 min with serum. Furthermore, researchers have shown
that the isomer 1-AG is thermodynamically more stable than 2-AG, and that
isomerization takes place until an equilibrium ratio of 1:9 (2-AG:1-AG) is reached [50].
Consequently, several works have reported the amount of 2-AG as the sum of peak
areas of 2-AG and 1-AG [11, 15, 16, 23, 46]. In this study, the conversion of 2-AG to 1-
AG in plasma samples was also observed due to the amount of water in human plasma
(≅ 90%) [52, 53] and albumin [20, 25, 32] (Figure 2). Hence, the sum of 2-AG and 1-
AG was described as 2-AG(s).

3.3 Evaluation of coating, desorption solution, and matrix modifier

Based on the Log P of the evaluated endocannabinoids, namely AEA (≅ 5.7)

and 2-AG (≅ 6.2), it was expected that these compounds would have strong affinity for
non-polar coatings such as C18 and C30. However, as these endocannabinoids also
present polar chemistry groups such as amides (AEA) and esters (2-AG), the HLB
coating presented comparable affinity for these compounds. As can be seen in Figure 3,
among the evaluated coatings, C30 and HLB presented the best recoveries. Due to the
major non-polar feature of endocannabinoids, C30 presented the highest sorption and
consequently, the highest extraction efficiency. HLB, on the other hand, allowed for
more versatility due to its capacity to interact with both hydrophilic and lipophilic sites.
Different desorption solvents were also evaluated as part of this work, including
methanol (100%), acetonitrile (100%), isopropanol (100%), methanol/acetonitrile
(50:50, v/v), methanol/isopropanol (50:50, v/v) and acetonitrile/isopropanol (50:50,
v/v). Given the comparable affinities of C30 and HLB for the compounds under study,

10
both coatings were included in the solvent evaluation. As shown in Figure S5,
methanol/isopropanol (50:50, v/v) presented the best recoveries among the evaluated
solvents for both HLB and C30. Carryover values lower than 1% were obtained for all
compounds. An ANOVA test followed by a post hoc test (Tukey’s test) confirmed the
presence of significant differences between the methanol/isopropanol (50:50, v/v)
mixture and the other solvents at the 5% level (p < 0.05).
The optimization of matrix modifier was carried out by experimental design trial
tests using the two selected fibers (C30 and HLB). As illustrated in Figure S6, the C30
fiber presented lower extraction efficiency than HLB. A reasonable explanation for this
observed phenomenon would be that despite the observed increase in the free
concentration of analytes due to the addition of organic modifier, the higher rate of
organic content in the matrix also led to a significant decrease in affinity between the
analyte and the coating. In other words, it can be concluded that extraction via
hydrophobic interaction was impaired by matrix modifier addition. In contrast, given
that the HLB coating allows for both hydrophobic and hydrophilic interactions, addition
of a matrix modifier allowed for an overall increase in extracted amounts. Therefore,
HLB coated fibers were chosen for further CCD optimization experiments.
Figure 4 illustrates the optimum composition of matrix modifiers added in
plasma samples to release bound endocannabinoids from proteins. For AEA and 2-
AG(s), the independent variables TFA 10% (µL) and ACN (µL) were significant (p <
0.05). According to the experimental design, high amounts of organic solvent (from 120
µL to 150 µL of ACN) and TFA 10% (v/v) (0.0 µL to 50 µL - studied range) decreased
the recovery of endocannabinoids in plasma samples. Indeed, high levels of both were
shown to undermine the extraction of AEA and 2-AG(s) due to the faster protein
precipitation process observed for these concentrations as compared to other
experimental conditions, which may lead to the co-precipitation of the compounds
under study towards the bottom of the vial. On the other hand, addition of a reasonable
amount of ACN (0.0 µL to 120 µL) can improve recoveries, as shown in the
experimental design. While guanidine hydrochloride did not seem to have a significant
effect on the recoveries of the targeted compounds, its interactions with the other
independent variables were shown to be significant  a fact only observable through
application of an experimental design.
Significant effects were found for recoveries of AEA and 2-AG(s) in plasma
samples, as observable in the generated ANOVA tables (see Tables S1 and S2).
11
According to the experimental design, neither AEA nor 2-AG(s) presented significant
lack-of-fit, indicating that the models were well adjusted, and thus can be used for
prediction. Therefore, according to the mathematical model (equations 1 and 2 are given
in the supplementary material), the optimal matrix modifier composition to aid in the
extraction of endocannabinoids in plasma samples (300 µL) included no addition of
TFA 10% (0.0µL), 50 µL of Gu-HCL (1 mol·L-1), 75 µL of ACN, and 75 µL of H2O
(LC-MS grade) (Figure S7). The interaction of Gu-HCL (1 mol·L-1) with ACN appears
to increase the capability of the modifier to disrupt the biding between
endocannabinoids and proteins in plasma, thus leading to an increased free
concentration of analytes available for extraction, without causing perturbations to the
affinity constant.
Indeed, according to Zoerner et al., the presence of albumin can reduce
extraction efficiency in certain conditions due to the high binding affinity of
endocannabinoids for this compound. It has thus been previously recommended in such
cases that protocols such as solvent extraction and protein precipitation be used to
disturb protein binding so as to enrich the extraction of endocannabinoids from plasma
samples [11]. However, these protocols often require large volumes of solvents and can
be very time-consuming. In the current work, the use of the optimized matrix modifier
solution allowed for an increase in the free concentration of endocannabinoids while
also affording the superior sample clean-up typical of SPME. Through addition of the
optimized matrix modifier solution, extracted amounts increased by 4.7 and 8.1 fold for
AEA and 2-AG(s), respectively.
Additionally, aiming to achieve the highest sensitivity, an extraction time profile
was performed using the optimized conditions (extraction, desorption, and
chromatographic analysis). The attained results (presented in Figure S8) revealed that
HLB fibers reach extractive equilibrium for both AEA and 2-AG within 60 min. An
ANOVA test yielded no significant (p > 0.05) differences for extraction times above 60
min.

3.4 SPME-UHPLC-MS/MS

The method validation was performed according to the international guidelines

issued EMA (European Medicines Agency) and FDA (Food and Drug Administration).

12
Linearity, limit of quantitation, precision and accuracy were the figures of merit used
for validation.
The calibration curves obtained for all endocannabinoids were linear with
correlation coefficients (r2) higher than 0.999. In addition, the lack-of-fit test was not
significant at the 5% level (p > 0.05), indicating that well-adjusted models were
obtained for all endocannabinoids (see Table 2). The calibration curves were attained on
basis of the signal ratio of the analyte and tis isotopologue (A/IS) for five concentration
levels in three independent replicates (n = 3) covering the range between 1 and 200 ng
mL-1 (Figure S9). Since the evaluated compounds are endogenous, the elaboration of
calibration curves followed the standard addition procedure.
Precision and accuracy were expressed as coefficient of variation (CV) and
relative standard error (RSE), respectively; For AEA, CV values for precision ranged
from 1.4 to 6.8% (Intra-assay), and from 0.7 to 12.1 (Inter-assay). For accuracy, RSE
values oscillated from -19.5 to 6.1% (Intra-assay) and from -16.7 to 9.1% (Inter-assay).
In the case of 2-AG(s), precision ranged from 1.2 to 6.1% (Intra-assay), and from 0.9 to
11.3% (Inter-assay) while for accuracy, values fluctuated from -14.9 to 12.6% (Intra-
assay) and from -13.5 to 14.4% (Inter-assay). Precision and accuracy values obtained
for both endocannabinoids can be considered satisfactory. In addition, attained RSE
values higher than 15% were related to the LOQ of the studied compounds, and as such
are deemed acceptable in accordance with issued international guidelines [52].
A calibration curve was prepared in phosphate-buffered saline solution (PBS)
(pH 7.4) mixed with human albumin to mimic the sample matrix (plasma), and a
comparison among the slopes of the calibration curves in different matrices (i.e. plasma
and PBS with albumin) was drawn out using analysis of variance. According to the
statistical test (ANOVA), the slopes presented no significant differences (p > 0.05),
indicating that the sensitivities of both calibration curves (plasma and PBS with Human
albumin) were similar. The calibrations curves for the analytes under study in plasma
and in PBS with human albumin can be seen in Figure S9. Internal standards for both
endocannabinoids (AEA and 2-AG(s)) were added to the matrices in a concentration of
50 ng·mL-1 with the aim of obtaining the correction of random error through the ratio
between the area of the compound and the internal standard.
The limit of quantitation (LOQ) was calculated as the first point of the calibration curve
with a back calculated concentration with less than 20% of RSD. In addition, precision

13
and accuracy were confirmed for each LOQ. The developed method yielded a LOQ of 1
ng·mL-1 for both target analytes.

3.5 Bio-SPME-nano-ESI versus Bio-SPME-LC-MS/MS

Validation of the Bio-SPME-nano-ESI method followed the same guidelines

applied to the SPME-UHPLC-MS/MS method. The calibration curves presented
linearity for both endocannabinoids, with a correlation coefficient higher than 0.996.
The lack-of-fit test was insignificant at the 5% level (p > 0.05), supporting that the
developed linear models are well adjusted for all endocannabinoids (Table 2). The
calibration curves were attained on basis of the signal ratio of the analyte and tis
isotopologue (A/IS) for five concentration levels in three independent replicates (n = 3)
covering the range between 50 and 800 ng mL-1 (Figure S10).
Precision and accuracy values are in accordance with limits recommended by
international guidelines [52]. CV values for AEA ranged from 0.55 to 7.2% (Intra-
assay) and from 1.2 to 10.3% (Interassay), while RSE values alternated from -16.5 to
11.2% (Intra-assay) and from 1.2 to 10.3% (Interassay). For 2-AG(s), CV values ranged
from 2.5 to 8.1% (Intra-assay) and from 1.9 to 9.4% (Interassay), while RSE values
varied from -17.7 to 11.8% (Intra-assay) and from 14.9 to 9.8% (Interassay). Calibration
curves for both endocannabinoids are presented in Figure S10. The method achieved an
LOQ of 50 ng·mL-1 for both compounds.
One of the main differences in workflow between both approaches entails their
distinctive desorption steps (see Figure 5). Essentially, in SPME-UHPLC-MS/MS,
desorption of analytes was performed for a period of 60 min into a vial containing 17
µL of solvent (Methanol/Isopropanol 50:50, v/v), whereupon 10 µL of the solution was
injected into the UHPLC-MS/MS system. In the nano-ESI workflow, desorption was
directly performed into the emitters (1.0/0.58 mm, o.d./i.d) for a period of 5 min with
the use of 4 µL of solvent (Methanol/Isopropanol 50:50, v/v with 0.1% Acetic acid).
As can be concluded from the above, the developed methods yield largely
different total times of analysis; in addition to the large differences in time needed for
desorption, in SPME-UHPLC-MS/MS, a total of 10 min is required for the
chromatographic run, while the direct coupling approach only requires 30 s for
detection and separation (see Figure 6).

14
 Although nano-ESI is a more efficient ionization technology [53], and despite
the fact that only a fraction of the desorption solvent (10 µL) is injected into the liquid
chromatograph, the SPME-UHPLC-MS/MS method presented better LOQs than nano-
ESI. The higher LOQ obtained via direct coupling to MS can be explained by the co-
extraction of some interferences that presented the same mass transition as the
endocannabinoids of interests (see Figure 7). While the direct coupling method does not
incorporate a chromatographic separation step, use of strategies such as ion mobility
may allow for separation of co-extracted interferences, which may improve signal-to-
noise ratios and thus enable the attainment of lower LOQs [54].
The SPME-UHPLC-MS/MS method herein presented provides lower LOQs for
AEA [13, 15, 17] and 2-AG [13, 15, 16, 47, 48] in comparison to some studies in the
literature, while providing rapid analysis in high throughput mode, yielding a total
analysis time per sample of 11.25 min, with total analysis time understood to be the
total time needed from analyte extraction to data attainment. For the Bio-SPME-nano-
ESI, the total analysis per sample was around 5.40 min. Furthermore, among the
methods reported in the literature, only a few of them were able to separate the isomers
2-AG and 1-AG [16, 18, 22, 46, 49]. Although other methods presented LOQs [18, 19,
22, 23, 29, 55] lower than the current SPME-UHPLC-MS/MS method, the linear range
is adequate to determine endocannabinoids in the plasma of healthy humans [11]. While
the total AG concentration (individual 2-AG and 1-AG analytical signals) in healthy
humans’ plasma ranges between 5 to 20 nmol·L-1 (≅ 1.9 to 7.6 ng·mL-1), when only the
2-AG peak is considered, then the calculated 2-AG concentrations are about 2-fold
lower, and similar to AEA concentrations (≅ 1 to 3.8 ng·mL-1) [11]. Indisputably, the
LOQ of the developed SPME-UHPLC-MS/MS method is extremely close to the
concentrations limits of endocannabinoids in plasma of healthy humans; however,
further improvements in analytical sensitivity are highly recommended to evaluate
lower concentrations of AEA and 2-AG [11, 56, 57].
In contrast, the method attained a wide linear range (1 – 200 ng·mL-1), which
allows for investigations of higher endocannabinoids concentrations correlated with
anomalies [58], osteoarthritis pain [59], multiple sclerosis [2], post-traumatic stress
disorder [60], neuropathic pain [61], and Parkinson’s disease [62]. In addition, the
method allows for monitoring of high endocannabinoids concentration levels proposed
for preventative treatments of cardiovascular events [63], and Alzheimer’s disease [64].

15
 Although the Bio-SPME-Nano-ESI method presented a wider linear range (from
50 to 800 ng mL-1), it could not achieve satisfactory endocannabinoids LOQs in plasma
due to the presence of interferences. The lower sensitivity attained is associated with
potential ion suppression. Other analytes (endogenous and exogenous) that are co-
extracted by the SPME fiber and desorbed into the emitter can contribute to ion
suppression providing lower sensitivity. On the contrary, when applying the SPME-
UHPLC-MS/MS method, the interferences are separated during the chromatographic
step thus, the sensitivity is higher .
Overall, the main advantage of SPME over other sample preparation
technologies is its capability to perform in vivo sampling. Thus, the developed methods
extend the potentiality of SPME for analysis of plasma samples and provide
encouragement for future in vivo research involving complex biomatrices.

4. CONCLUSIONS

The current work introduces the development of two methods for analysis of
endocannabinoids: SPME-UHPLC-MS/MS and SPME-nano-ESI-MS/MS. Both of them
were optimized and successfully applied to determine endocannabinoids in denatured
plasma samples. SPME-UHPLC-MS/MS method presented lower LOQ values than
SPME-Nano-ESI-MS/MS. The additional separation provided by the LC-MS/MS
approach favored the detectability of endocannabinoids by LC-MS/MS over nano-ESI-
MS/MS, although the SPME-nano-ESI-MS/MS method decreased the total analysis
time due to significant reductions in desorption and detection times.
Undoubtedly, the SPME technique opens up new perspectives for in vivo
analysis as compared with traditional sample preparation techniques, a possibility that is
strongly attributed to SPME’s unique features of minimum invasiveness and minimum
system disturbance [40, 65]. The use of microfluidic open interface (MOI) coupled to
high-resolution ion mobility is envisioned as a powerful combination that might
increase the sensitivity, speed of analysis, and robustness of the SPME method for
direct coupling analysis [39].

5. ACKNOWLEDGMENTS
The authors are grateful to Supelco (Sigma Milipore), Thermo Scientific (USA)
and the Natural Science and Engineering Research Council (NSERC) of Canada for the

16
financial support provided through the Industrial Research Chair program. V.R.A.Jr and
M.E.C.Q are grateful to the Fundação de Amparo à Pesquisa do Estado de São Paulo
(FAPESP, nº 2016/16180-6 and 2017/02147-0) for the financial support.

6. REFERENCES

[1] O. Haugh, J. Penman, A.J. Irving, V.A. Campbell, The emerging role of the
cannabinoid receptor family in peripheral and neuro-immune interactions, Curr Drug
Targets, (2016).
[2] L. Jean-Gilles, S. Feng, C. Tench, V. Chapman, D. Kendall, D. Barrett, C.
Constantinescu, Plasma endocannabinoid levels in multiple sclerosis, Journal of the
Neurological Sciences, 287 (2009) 212-215.
[3] G. Bedse, A. Romano, A.M. Lavecchia, T. Cassano, S. Gaetani, The role of
endocannabinoid signaling in the molecular mechanisms of neurodegeneration in
Alzheimer's disease, J Alzheimers Dis, 43 (2015) 1115-1136.
[4] K.-M. Jung, G. Astarita, S. Yasar, V. Vasilevko, D.H. Cribbs, E. Head, C.W.
Cotman, D. Piomelli, An amyloid β 42-dependent deficit in anandamide mobilization is
associated with cognitive dysfunction in Alzheimer's disease, Neurobiology of Aging,
33 (2012) 1522-1532.
[5] C. Criscuolo, P. Mancini, V. Menchise, F. Saccà, G. De Michele, S. Banfi, A. Filla,
Very late onset in ataxia oculomotor apraxia type I, Annals of Neurology, 57 (2005)
777-777.
[6] A. Pisani, F. Fezza, S. Galati, N. Battista, S. Napolitano, A. Finazzi‐Agrò, G.
Bernardi, L. Brusa, M. Pierantozzi, P. Stanzione, High endogenous cannabinoid levels
in the cerebrospinal fluid of untreated Parkinson's disease patients, Annals of
Neurology, 57 (2005) 777-779.
[7] B. Kluger, P. Triolo, W. Jones, J. Jankovic, The therapeutic potential of
cannabinoids for movement disorders, Movement Disorders, 30 (2015) 313-327.
[8] N. Battista, M. Sergi, C. Montesano, S. Napoletano, D. Compagnone, M.
Maccarrone, Analytical approaches for the determination of phytocannabinoids and
endocannabinoids in human matrices, Drug Testing and Analysis, 6 (2014) 7-16.
[9] M.W. Manseau, D.C. Goff, Cannabinoids and Schizophrenia: Risks and Therapeutic
Potential, Neurotherapeutics, 12 (2015) 816-824.
[10] F. Leweke, D. Piomelli, F. Pahlisch, D. Muhl, C. Gerth, C. Hoyer, J. Klosterkötter,
M. Hellmich, D. Koethe, Cannabidiol enhances anandamide signaling and alleviates
psychotic symptoms of schizophrenia, Translational Psychiatry, 2 (2012) e94.
[11] A.A. Zoerner, F.-M. Gutzki, S. Batkai, M. May, C. Rakers, S. Engeli, J. Jordan, D.
Tsikas, Quantification of endocannabinoids in biological systems by chromatography
and mass spectrometry: a comprehensive review from an analytical and biological
perspective, Biochimica et Biophysica Acta (BBA)-Molecular and Cell Biology of
Lipids, 1811 (2011) 706-723.
[12] C. Marchioni, I.D. de Souza, V.R. Acquaro, J.A. de Souza Crippa, V. Tumas,
M.E.C. Queiroz, Recent advances in LC-MS/MS methods to determine
endocannabinoids in biological samples: Application in neurodegenerative diseases,
Analytica Chimica Acta, 1044 (2018) 12-28.
[13] S. Gouveia-Figueira, M.L. Nording, Development and validation of a sensitive
UPLC-ESI-MS/MS method for the simultaneous quantification of 15 endocannabinoids
and related compounds in milk and other biofluids, Anal Chem, 86 (2014) 1186-1195.

17
[14] M.S. Gachet, P. Rhyn, O.G. Bosch, B.B. Quednow, J. Gertsch, A quantitiative LC-
MS/MS method for the measurement of arachidonic acid, prostanoids,
endocannabinoids, N-acylethanolamines and steroids in human plasma, J Chromatogr B
Analyt Technol Biomed Life Sci, 976-977 (2015) 6-18.
[15] C. Garst, M. Fulmer, D. Thewke, S. Brown, Optimized extraction of 2-arachidonyl
glycerol and anandamide from aortic tissue and plasma for quantification by LC-
MS/MS, European Journal of Lipid Science and Technology, 118 (2016) 814-820.
[16] Y. Gong, X. Li, L. Kang, Y. Xie, Z. Rong, H. Wang, H. Qi, H. Chen, Simultaneous
determination of endocannabinoids in murine plasma and brain substructures by
surrogate-based LC-MS/MS: Application in tumor-bearing mice, J Pharm Biomed Anal,
111 (2015) 57-63.
[17] D.J. Liput, E. Tsakalozou, D.C. Hammell, K.S. Paudel, K. Nixon, A.L.
Stinchcomb, Quantification of anandamide, oleoylethanolamide and
palmitoylethanolamide in rodent brain tissue using high performance liquid
chromatography-electrospray mass spectroscopy, J Pharm Anal, 4 (2014) 234-241.
[18] M.G. Balvers, H.M. Wortelboer, R.F. Witkamp, K.C. Verhoeckx, Liquid
chromatography-tandem mass spectrometry analysis of free and esterified fatty acid N-
acyl ethanolamines in plasma and blood cells, Anal Biochem, 434 (2013) 275-283.
[19] X. Xiong, L. Zhang, L. Cheng, W. Mao, High-throughput salting-out assisted
liquid-liquid extraction with acetonitrile for the determination of anandamide in plasma
of hemodialysis patients with liquid chromatography tandem mass spectrometry,
Biomed Chromatogr, 29 (2015) 1317-1324.
[20] É.A. Souza-Silva, N. Reyes-Garcés, G.A. Gómez-Ríos, E. Boyacı, B. Bojko, J.
Pawliszyn, A critical review of the state of the art of solid-phase microextraction of
complex matrices III. Bioanalytical and clinical applications, TrAC Trends in Analytical
Chemistry, 71 (2015) 249-264.
[21] J. Peng, F. Tang, R. Zhou, X. Xie, S. Li, F. Xie, P. Yu, L. Mu, New techniques of
on-line biological sample processing and their application in the field of
biopharmaceutical analysis, Acta Pharm Sin B, 6 (2016) 540-551.
[22] M. Sergi, N. Battista, C. Montesano, R. Curini, M. Maccarrone, D. Compagnone,
Determination of the two major endocannabinoids in human plasma by mu-SPE
followed by HPLC-MS/MS, Anal Bioanal Chem, 405 (2013) 785-793.
[23] C. Marchioni, I.D. de Souza, C.F. Grecco, J.A. Crippa, V. Tumas, M.E.C. Queiroz,
A column switching ultrahigh-performance liquid chromatography-tandem mass
spectrometry method to determine anandamide and 2-arachidonoylglycerol in plasma
samples, Anal Bioanal Chem, 409 (2017) 3587-3596.
[24] I.D. Souza, L.W. Hantao, M.E.C. Queiroz, Polymeric ionic liquid open tubular
capillary column for on-line in-tube SPME coupled with UHPLC-MS/MS to determine
endocannabinoids in plasma samples, Analytica Chimica Acta, 1045 (2019) 108-116.
[25] B.H. Fumes, M.R. Silva, F.N. Andrade, C.E.D. Nazario, F.M. Lanças, Recent
advances and future trends in new materials for sample preparation, TrAC Trends in
Analytical Chemistry, 71 (2015) 9-25.
[26] H. Trufelli, P. Palma, G. Famiglini, A. Cappiello, An overview of matrix effects in
liquid chromatography–mass spectrometry, Mass Spectrometry Reviews, 30 (2011)
491-509.
[27] A. Pastor, M. Farre, M. Fito, F. Fernandez-Aranda, R. de la Torre, Analysis of ECs
and related compounds in plasma: artifactual isomerization and ex vivo enzymatic
generation of 2-MGs, J Lipid Res, 55 (2014) 966-977.
[28] C.A. Rouzer, K. Ghebreselasie, L.J. Marnett, Chemical stability of 2-
arachidonylglycerol under biological conditions, Chem Phys Lipids, 119 (2002) 69-82.

18
[29] V. Kantae, S. Ogino, M. Noga, A.C. Harms, R.M. van Dongen, G.L. Onderwater,
A.M. van den Maagdenberg, G.M. Terwindt, M. van der Stelt, M.D. Ferrari, T.
Hankemeier, Quantitative profiling of endocannabinoids and related N-
acylethanolamines in human CSF using nano LC-MS/MS, J Lipid Res, 58 (2017) 615-
624.
[30] V.R. Acquaro, F.M. Lanças, M.E.C. Queiroz, Evaluation of superficially porous
and fully porous columns for analysis of drugs in plasma samples by UHPLC–MS/MS,
Journal of Chromatography B, 1048 (2017) 1-9.
[31] E. Gionfriddo, E. Boyacı, J. Pawliszyn, New Generation of Solid-Phase
Microextraction Coatings for Complementary Separation Approaches: A Step toward
Comprehensive Metabolomics and Multiresidue Analyses in Complex Matrices,
Analytical Chemistry, 89 (2017) 4046-4054.
[32] N. Reyes-Garces, E. Gionfriddo, G.A. Gomez-Rios, M.N. Alam, E. Boyaci, B.
Bojko, V. Singh, J. Grandy, J. Pawliszyn, Advances in Solid Phase Microextraction and
Perspective on Future Directions, Analytical Chemistry, 90 (2018) 302-360.
[33] E.A.S. Silva, S. Risticevic, J. Pawliszyn, Recent trends in SPME concerning
sorbent materials, configurations and in vivo applications, Trac-Trends in Analytical
Chemistry, 43 (2013) 24-36.
[34] V. Bessonneau, Y. Zhan, I.A.M. De Lannoy, V. Saldivia, J. Pawliszyn, In vivo
solid-phase microextraction liquid chromatography-tandem mass spectrometry for
monitoring blood eicosanoids time profile after lipopolysaccharide-induced
inflammation in Sprague-Dawley rats, Journal of Chromatography A, 1424 (2015) 134-
138.
[35] M. Walles, Y. Gu, C. Dartiguenave, F.M. Musteata, K. Waldron, D. Lubda, J.
Pawliszyn, Approaches for coupling solid-phase microextraction to nanospray, Journal
of Chromatography A, 1067 (2005) 197-205.
[36] J. Deng, Y. Yang, M. Xu, X. Wang, L. Lin, Z.P. Yao, T. Luan, Surface-coated
probe nanoelectrospray ionization mass spectrometry for analysis of target compounds
in individual small organisms, Anal Chem, 87 (2015) 9923-9930.
[37] G.A. Gomez-Rios, N. Reyes-Garces, B. Bojko, J. Pawliszyn, Biocompatible Solid-
Phase Microextraction Nanoelectrospray Ionization: An Unexploited Tool in
Bioanalysis, Anal Chem, 88 (2016) 1259-1265.
[38] B.-C. Yang, S.-F. Fang, X.-J. Wan, Y. Luo, J.-Y. Zhou, Y. Li, Y.-J. Li, F. Wang,
O.-P. Huang, Quantification of monohydroxylated polycyclic aromatic hydrocarbons in
human urine samples using solid-phase microextraction coupled with glass-capillary
nanoelectrospray ionization mass spectrometry, Analytica Chimica Acta, 973 (2017)
68-74.
[39] G.A. Gómez-Ríos, M.F. Mirabelli, Solid Phase Microextraction-mass
spectrometry: Metanoia, TrAC Trends in Analytical Chemistry, 112 (2019) 201-211.
[40] T. Vasiljevic, V. Singh, J. Pawliszyn, Miniaturized SPME tips directly coupled to
mass spectrometry for targeted determination and untargeted profiling of small samples,
Talanta, 199 (2019) 689-697.
[41] G.A. Gómez-Ríos, M. Tascon, N. Reyes-Garcés, E. Boyacı, J.J. Poole, J.
Pawliszyn, Rapid determination of immunosuppressive drug concentrations in whole
blood by coated blade spray-tandem mass spectrometry (CBS-MS/MS), Analytica
Chimica Acta, 999 (2018) 69-75.
[42] A. Roszkowska, M. Tascon, B. Bojko, K. Goryński, P.R. dos Santos, M. Cypel, J.
Pawliszyn, Equilibrium ex vivo calibration of homogenized tissue for in vivo SPME
quantitation of doxorubicin in lung tissue, Talanta, 183 (2018) 304-310.

19
[43] E. Boyacı, B. Bojko, N. Reyes-Garcés, J.J. Poole, G.A. Gómez-Ríos, A. Teixeira,
B. Nicol, J. Pawliszyn, High-throughput analysis using non-depletive SPME: challenges
and applications to the determination of free and total concentrations in small sample
volumes, Scientific Reports, 8 (2018) 1167.
[44] N. Kumar, A. Bansal, G.S. Sarma, R.K. Rawal, Chemometrics tools used in
analytical chemistry: An overview, Talanta, 123 (2014) 186-199.
[45] M. Otto, Chemometrics: statistics and computer application in analytical chemistry,
John Wiley & Sons2016.
[46] S. Gouveia-Figueira, M.L. Nording, Validation of a tandem mass spectrometry
method using combined extraction of 37 oxylipins and 14 endocannabinoid-related
compounds including prostamides from biological matrices, Prostaglandins Other Lipid
Mediat, 121 (2015) 110-121.
[47] C. Muguruza, M. Lehtonen, N. Aaltonen, B. Morentin, J.J. Meana, L.F. Callado,
Quantification of endocannabinoids in postmortem brain of schizophrenic subjects,
Schizophr Res, 148 (2013) 145-150.
[48] B. Han, R. Wright, A.M. Kirchhoff, J.A. Chester, B.R. Cooper, V.J. Davisson, E.
Barker, Quantitative LC-MS/MS analysis of arachidonoyl amino acids in mouse brain
with treatment of FAAH inhibitor, Anal Biochem, 432 (2013) 74-81.
[49] A.A. Zoerner, S. Batkai, M.T. Suchy, F.M. Gutzki, S. Engeli, J. Jordan, D. Tsikas,
Simultaneous UPLC-MS/MS quantification of the endocannabinoids 2-arachidonoyl
glycerol (2AG), 1-arachidonoyl glycerol (1AG), and anandamide in human plasma:
minimization of matrix-effects, 2AG/1AG isomerization and degradation by toluene
solvent extraction, J Chromatogr B Analyt Technol Biomed Life Sci, 883-884 (2012)
161-171.
[50] K. Docs, Z. Meszar, S. Gonda, A. Kiss-Szikszai, K. Hollo, M. Antal, Z. Hegyi, The
Ratio of 2-AG to Its Isomer 1-AG as an Intrinsic Fine Tuning Mechanism of CB1
Receptor Activation, Front Cell Neurosci, 11 (2017) 39.
[51] J. Pawliszyn, Handbook of solid phase microextraction, Elsevier2011.
[52] F.D.a. Administration, Analytical procedures and methods validation for drugs and
biologics, Guidance for Industry, 2015.
[53] S.R. Needham, G.A. Valaskovic, Microspray and microflow LC–MS/MS: the
perfect fit for bioanalysis, Bioanalysis, 7 (2015) 1061-1064.
[54] C. Liu, G.A. Gómez-Ríos, B.B. Schneider, J.Y. Le Blanc, N. Reyes-Garcés, D.W.
Arnold, T.R. Covey, J. Pawliszyn, Fast quantitation of opioid isomers in human plasma
by differential mobility spectrometry/mass spectrometry via SPME/open-port probe
sampling interface, Analytica chimica acta, 991 (2017) 89-94.
[55] L. Lin, H. Yang, P.J. Jones, Quantitative analysis of multiple fatty acid
ethanolamides using ultra-performance liquid chromatography-tandem mass
spectrometry, Prostaglandins Leukot Essent Fatty Acids, 87 (2012) 189-195.
[56] A. Molvarec, G. Fügedi, E. Szabó, B. Stenczer, S. Walentin, J. Rigó Jr, Decreased
circulating anandamide levels in preeclampsia, Hypertension Research, 38 (2015) 413.
[57] D. Tsikas, A.A. Zoerner, Analysis of eicosanoids by LC-MS/MS and GC-MS/MS:
A historical retrospect and a discussion, Journal of Chromatography B, 964 (2014) 79-
88.
[58] V. Di Marzo, S. Petrosino, Endocannabinoids and the regulation of their levels in
health and disease, Curr Opin Lipidol, 18 (2007) 129-140.
[59] C. La Porta, S.A. Bura, J. Llorente-Onaindia, A. Pastor, F. Navarrete, M.S. García-
Gutiérrez, R. De la Torre, J. Manzanares, J. Monfort, R. Maldonado, Role of the
endocannabinoid system in the emotional manifestations of osteoarthritis pain, Pain,
156 (2015) 2001-2012.

20
[60] D. Hauer, G. Schelling, H. Gola, P. Campolongo, J. Morath, B. Roozendaal, G.
Hamuni, A. Karabatsiakis, P. Atsak, M. Vogeser, I.-T. Kolassa, Plasma Concentrations
of Endocannabinoids and Related Primary Fatty Acid Amides in Patients with Post-
Traumatic Stress Disorder, PLoS ONE, 8 (2013) e62741.
[61] I. Kaufmann, D. Hauer, V. Huge, M. Vogeser, P. Campolongo, A. Chouker, M.
Thiel, G. Schelling, Enhanced Anandamide Plasma Levels in Patients with Complex
Regional Pain Syndrome following Traumatic Injury: A Preliminary Report, European
Surgical Research, 43 (2009) 325-329.
[62] V. Pisani, G. Madeo, A. Tassone, G. Sciamanna, M. Maccarrone, P. Stanzione, A.
Pisani, Homeostatic changes of the endocannabinoid system in Parkinson's disease,
Movement Disorders, 26 (2011) 216-222.
[63] R.B. Mounsey, S. Mustafa, L. Robinson, R.A. Ross, G. Riedel, R.G. Pertwee, P.
Teismann, Increasing levels of the endocannabinoid 2-AG is neuroprotective in the 1-
methyl-4-phenyl-1,2,3,6-tetrahydropyridine mouse model of Parkinson's disease,
Experimental Neurology, 273 (2015) 36-44.
[64] J. Mulder, M. Zilberter, S.J. Pasquaré, A. Alpár, G. Schulte, S.G. Ferreira, A.
Köfalvi, A.M. Martín-Moreno, E. Keimpema, H. Tanila, M. Watanabe, K. Mackie, T.
Hortobágyi, M.L. de Ceballos, T. Harkany, Molecular reorganization of
endocannabinoid signalling in Alzheimer’s disease, Brain, 134 (2011) 1041-1060.
[65] S. Lendor, S.-A. Hassani, E. Boyaci, V. Singh, T. Womelsdorf, J. Pawliszyn, A
solid phase microextraction-based miniaturized probe and protocol for extraction of
neurotransmitters from brains in vivo, Analytical Chemistry, (2019).

21
Figure captions
Figure 1. TIC chromatograms for endocannabinoids standard solution (ACN –100
ng·mL-1) using two different analytical columns. (a) Kinetex C18 column and (b)
CORTECS C18 column. Both endocannabinoids responses were normalized to 100%.
Figure 2. Total Ion Chromatogram (TIC) chromatogram of (a) standard solution (ACN
– 100 ng mL-1) and (b) plasma sample fortified with standard solution (100 ng·mL-1).
Figure 3. Comparison among C18, C30, and HLB coatings. Both endocannabinoids
responses were normalized to 100%.
Figure 4. Optimum response surface (Dering and Suich) AEA and 2-AG(s). The red
region entails the maximum amount of endocannabinoids extracted. (a) Gu-HCL 1
mol·L-1 (µL) x TFA 10% (µL); (b) ACN (µL) x TFA 10% (µL); (c) ACN (µL) x Gu-
HCL 1 mol·L-1 (µL).
Figure 5. Analytical workflows for the developed SPME-UHPLC-MS/MS and SPME-
nano-ESI-MS/MS methods.
Figure 6. An Ion-chronogram representative of the SPME-nano-ESI-MS/MS method (a)
and a chromatogram (TIC) representative of the SPME-UHPLC-MS/MS (b).
Extractions using SPME-UHPLC-MS/MS and SPME-nano-ESI-MS/MS methods were
performed in plasma samples at concentrations of 100 ng·mL-1 for both
endocannabinoids. Both endocannabinoids responses were normalized to 100%.
Figure 7. Chromatograms of monitored transitions for both endocannabinoids at LOQ
concentrations (1 ng mL-1) in plasma samples.

22
Table captions
Table 1. Retention time and MS parameters for all transitions monitored for SPME-LC-
MS/MS.
Table 2. Linear equation, linear range, internal standard, Pearson’s coefficient and lack
of fit of SPME-UHPLC-MS/MS and SPME-nano-ESI-MS/MS analytical methods.

23
 Retention time Precursor Product Collision RF lens
Analyte
(min) (m/z) (m/z) Energy (V) (V)
287.2 13.1
AEA 6.68 348.2 62.1 17.5 66
269.2 14.9
AEA-d4 6.67 358.2 287.2 13.1 65
287.2 13.3
2-AG 7.55 379.2 269.2 17.2 78
203.2 21.1
2-AG-d5 7.55 384.2 287.2 14.9 85
287.2 13.3
1-AG 7.75 379.2 269.2 17.2 78
203.2 21.1
1-AG-d5 7.74 384.2 287.2 14.9 85
 Target analyte Range Internal
Linear equation -1
Pearson’s coefficient (r) Lack of fit (p-value)
SPME-UHPLC-MS/MS (ng·mL ) standard
AEA Y = 0.0445x + 0.0249 1-200 AEA-d4 0.9999 0.9907
Plasma
2-AG(s) Y = 0.0453x + 2.0932 1-200 2-AG-d5 0.9990 0.8202
AEA Y = 0.0453x - 0.0007 1-200 AEA-d4 0.9995 0.8477
PBS
2-AG(s) Y = 0.0437x - 0.0088 1-200 2-AG-d5 0.9999 0.9994

SPME-nano-ESI-MS/MS
AEA Y = 0.0046x + 0.0868 50-800 AEA-d4 0.9969 0.1169
Plasma
2-AG(s) Y = 0.0201x + 0.0941 50-800 2-AG-d5 0.9985 0.6755
Highlights
1. Direct coupling of biocompatible SPME with Mass Spectrometry.
2. Comparison between direct SPME-nano-ESI-MS/MS with SPME-UHPLC-MS/MS
methods.
2. DI- SPME with modified plasma sample.
3. The chromatographic separation favored the detectability of SPME-UHPLC-MS/MS
method.
4. The SPME-nano-ESI-MS/MS approach decrease the total analysis time.
Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests: