The prominent biological effects of adrenaline (A), noradrenaline (NA) and dopamine (DA) as well as the clinical importance of their metabolites (such as dihydroxyphenylacetic acid (DOPAC), methoxy–4- hydroxyphenyl glycol (MHPG), dihydroxyphenylglycol (DHPG), metanephrine (M), normetanephrine (NM), vanillylmandelic acid (VMA), homovanillic acid (HVA)) have forced researchers to evaluate new analytical methodologies for their isolation and preconcentration from biological samples.
Trang 1Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/chroma
Natalia Mi ˛ekus∗, Alina Plenis, Marta Rudnicka, Natalia Kossakowska, Ilona Ol ˛edzka,
Piotr Kowalski, Tomasz B aczek ˛
Department of Pharmaceutical Chemistry, Medical University of Gda ´nsk, Hallera, 107, 80-416, Gda ´nsk, Poland
a r t i c l e i n f o
Article history:
Received 15 October 2019
Revised 9 March 2020
Accepted 10 March 2020
Available online 12 March 2020
Keywords:
Biogenic amines
Capillary electrophoresis
Hierarchical cluster analysis
Solid-phase microextraction
Solid-Phase Extraction
Dispersive Liquid-Liquid Microextraction
a b s t r a c t
The prominentbiologicaleffects of adrenaline(A),noradrenaline (NA)and dopamine(DA) as well as theclinical importanceoftheir metabolites(suchas dihydroxyphenylacetic acid(DOPAC), methoxy–4-hydroxyphenylglycol(MHPG),dihydroxyphenylglycol(DHPG),metanephrine(M),normetanephrine(NM), vanillylmandelicacid(VMA),homovanillicacid(HVA))haveforcedresearcherstoevaluatenew analyti-calmethodologiesfortheirisolationandpreconcentrationfrombiologicalsamples.Forthisreason,the threemostpopularextractiontechniques(dispersiveliquid-liquidmicroextraction(DLLME),solid-phase extraction(SPE), solid-phasemicroextraction(SPME))weretested.Micellarelectrokinetic chromatogra-phy(MEKC)– amodeofcapillaryelectrophoresis– withadiodearraydetector(DAD)wasappliedto assesstheextractionefficiency.Next,theenrichmentfactor(EF)ofeachappliedmethodwascalculated
inrespecttostandardmixturesoftheanalytesatthesameconcentrationlevels.TheEFresultsofseven selectedmetabolitesofbiogenicamines(BAs)fromurineaftersamplepreparationproceduresbasedon twenty-fivedifferent protocols(oneDLLME, thirteenSPEand eleven SPME)werecalculated and com-paredusinghierarchicalclusteranalysis(HCA).TheSPEaswellasSPMEprocedureswereproved tobe themosteffectiveapproachesforthesimultaneousextractionofthechosencompounds.Moreover,an ionicliquid(IL)– 1-ethyl-3-methylimidazoliumbis(trifluoromethylsulfonyl)imide– addedtomethanolin SPMEadditionally couldsuccessfully improvethe extraction efficiency.It was alsoconfirmedthat the HCAapproachcouldbeconsidered asupportivetool intheselection ofasuitablesamplepreparation procedureforthatgroupofendogenoussubstances
© 2020TheAuthors.PublishedbyElsevierB.V ThisisanopenaccessarticleundertheCCBY-NC-NDlicense
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
1 Introduction
The secretory cells of the adrenal medulla mainly produce cat-
echolamines: adrenaline (A), noradrenaline (NA) and dopamine
(DA) The metabolism of these relevant compounds takes place
mainly in the gastrointestinal tract (GI), intraneurally and in the
adrenal medulla owing to two enzymes: catechol-O-methyl trans-
ferase (COMT) and monoamine oxidase (MAO) ( Fig.1)
The main metabolites of DA are: dihydroxyphenylacetic acid
(DOPAC) and homovanillic acid (HVA), whereas A is converted into
metanephrine (M) and further into 3–methoxy–4-hydroxyphenyl
∗ Corresponding author
E-mail addresses: natalia.miekus-purwin@gumed.edu.pl , miekusn@gmail.com (N
Mi ˛ekus)
glycol (MHPG) By the actions of alcohol dehydrogenase in the liver, MHPG is metabolized to vanillylmandelic acid (VMA) The
NA metabolic pathway end products are also MHPG and VMA, but NA is also metabolized intraneurally to dihydroxyphenylgly- col (DHPG) and in the adrenal medulla to normetanephrine (NM) [1] The physiological metabolism of those monoamine neurotrans- mitters (NTs) could be interrupted (or their ratios of concentration visibly changed) in pathophysiological stages of the human organ- ism, which include neuroendocrine tumors (NETs) – pheochromo- cytoma (PHE) and neuroblastoma (NBL) [ 2, 3] As such, the deter- mination of the concentration of HVA and VMA in urine samples remains the gold standard for the biochemical diagnosis of NETs [4] Furthermore, the determination of DHPG and MHPG in plasma samples could provide reliable information regarding the effects of COMT and MAO on NA [5] Nevertheless, the determination of O- methylated metabolites (M and NM) in plasma or urine samples https://doi.org/10.1016/j.chroma.2020.461032
0021-9673/© 2020 The Authors Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
Trang 22 N Mi ˛ekus, A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032
Fig 1 Scheme of the main neuronal and non-neuronal pathways of dopamine, noradrenaline and adrenaline metabolism with pKa values for corresponding metabolites
Legend: DA, dopamine; NA, noradrenaline; A adrenaline; DOPAC, 3,4-dihydroxyphenylacetic acid; DHPG, 3,4-dihydroxyglycol; NM, normetanephrine; M, metanephrine; HVA, homovanillic acid; MHPG, 3–methoxy–4-hydroxyphenylglycol; VMA, vanillylmandelic acid; MAO, monoamine oxidase; COMT, catechol O-methyltransferase; ADH, alcohol dehydrogenase
has been shown to have higher sensitivity towards the diagnosis of
both NETs than the estimation of catecholamines or the concentra-
tion of VMA and HVA [6] The levels of M and NM were evaluated
not only to give more reliable data during the diagnosis of PHE, but
also their concentration was positively correlated with the size and
adrenal or extra-adrenal location of a tumor Even though the anal-
ysis of the concentration of M and NM in plasma is considered as
more appropriate than urine samples for the diagnosis of PHE, the
concentration of those two compounds in tumor tissues is usually
3 orders of magnitude higher than in plasma samples [5]
The precise diagnosis and description of the localization and
size of a tumor of a neuroendocrine origin (PHE or NBL) requires
the simultaneous analysis of the main catecholamine metabolites
from the biological specimens of patients Modern, high through-
put analytical methods could provide a great tool for the fast
and sensitive profiling of metabolites However, the extraction of
trace amounts of metabolites needs to be evaluated since the
level of each metabolite is extremely low Therefore, applied iso-
lation methods should also have the advantage of preconcentrat-
ing analytes in the biological sample To address this issue, the
three most popular and efficient extraction and preconcentration
techniques: dispersive liquid-liquid microextraction (DLLME), solid-
phase extraction (SPE) and solid-phase microextraction (SPME)
were tested The DLLME procedure is simple and fast in terms of
sample preparation and is rarely used for complex biological ma-
trices, which require the removal of ballast substances, such as
proteins SPE is widely employed to concentrate and purify bio-
logical samples before analysis SPME holds some advantages over traditional sample preparation methods, such as little consump- tion of toxic and hazardous organic solvents and the relative ease
of online coupling to chromatographic systems However, for each extraction technique, the key extraction parameters affecting the extraction efficiency should be optimized For each of these ap- proaches, the extraction efficiency was evaluated by the calcula- tion of the enrichment factor (EF) This was done through a com- parison of the signal intensity of the analytes in respect to signals obtained for standard mixtures of the compounds of interest at the same concentration levels The signal intensities (peak heights) were determined using an optimized micellar electrokinetic chro- matography (MEKC) method coupled with a diode array detector (DAD) for the determination of DHPG, VMA, MHPG, HVA, NM, M and DOPAC in human urine samples Additionally, three ionic liq- uids (ILs), namely: 1–butyl–3-methylimidazolium tetrafluoroborate (IL1), 1-ethyl-3-methylimidazolium tetrafluoroborate (IL2) and 1- ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (IL3) were tested as the desorbent additives for SPME to evaluate their utility for a new class of compounds different to those tested in our previous studies [7] The decision to apply ILs was driven by the fact that the use of traditional organic solvents could lead to pollution of the natural environment and danger to human health ILs are believed to constitute alternative solvents, characterized
by low volatility, and chemical and physical stability [7–9] Some modern analytical methods have been optimized with the use of distinct ILs during the analysis of biogenic amines (BAs) [ 7, 8, 10, 11]
Trang 3In our earlier studies, experimental data demonstrated the useful-
ness of IL2 at a concentration of 20 ng/mL in SPME, followed by
MEKC for the determination of DA, A, NA, l-tyrosine (L-Tyr) and l
tryptophan (L-Tryp) from human urine samples [7] Owing to the
optimized method, the extraction yields of BA precursors together
with some basic BAs increased from nearly 9 times for l-Tryp up
to 21 times for A [7]
The main objective of the presented work was to develop and
optimize an effective extraction protocol for seven analyzed BAs
(DHPG, VMA, MHPG, HVA, NM, M and DOPAC) from human urine
samples In this study, different modifications of three sample
preparation procedures based on SPME, SPE and DLLME were in-
vestigated and discussed, in order to select the most efficient pro-
cedure providing the largest recovery of analytes and the most ef-
fective purification of the sample matrix It was decided to use hi-
erarchical cluster analysis (HCA) in order to check whether it could
be considered as a useful tool facilitating the selection of the most
effective sample preparation procedure for specific analytes This
powerful analytical platform, consisting of the most appropriate
sample preconcentration, enrichment and analysis methods, was
evaluated for each of the studied compounds
2 Materials and methods
2.1 Chemicals and reagents
Methanol (MeOH), hexane, ethanol 96% (EtOH) and ace-
tone were supplied by POCH (Gliwice, Poland) Reagents, such
as sodium dodecyl sulfate (SDS), DHPG, VMA, MHPG, HVA,
M, NM, DOPAC, acetonitrile (ACN), dichloromethane (DCM),
1–butyl–3-methylimidazolium tetrafluoroborate (IL1), 1-ethyl-
3-methyl-imidazolium tetrafluoroborate (IL2) and 1-ethyl-3-
methylimidazolium bis(trifluoromethylsulfonyl)imide (IL3) were
supplied by Sigma-Aldrich (Darmstadt, Germany) Sodium tetrabo-
rate decahydrate (borax), boric acid and sodium hydroxide (NaOH)
were obtained from Merck (Darmstadt, Germany) Capillary Re-
generator Basic Wash Solution was purchased from Beckman
Coulter (CA, USA) All chemicals were of analytical grade and were
applied without further purification The purified water used in
all experiments was obtained from Milli-Q equipment (Millipore,
Bedford, MA, USA)
2.2 Apparatus
All separation studies were carried out using a capillary elec-
trophoresis (CE) apparatus (P/ACE MDQ Capillary Electrophoresis
System, Beckman Coulter, Fullerton, CA, USA) The appliance was
equipped with an automatic sample dispenser and a DAD detec-
tor Analysis of the obtained data was made using 32 Karat 8.0
software (Beckmann, Fullerton, CA, USA) The device was addition-
ally equipped with a capillary thermostat system by means of a
coolant, which allowed the temperature to remain constant during
the analysis
2.3 Preparation of stock and working solutions
Stock solutions were prepared by accurately weighing 1.0 mg
of each analyte on an electronic scale (Ohaus, PA, USA) Then
the weighed analytes, namely: DHPG, VMA, MHPG, HVA, M, NM,
DOPAC were dissolved separately in 1 mL of MeOH Subsequently,
they were shaken on an MS 3 Basic, IKA® shaker (USA) Standard
human urine samples were enriched with each of the analytes
to a final concentration of 10 μg/mL and then put aside for one
of the extraction methods After the isolation procedure, the dry
residue containing the extracted analytes was dissolved with 50
μL of 2 mM sodium tetraborate decahydrate and separated by the
MEKC method The reference urine samples were prepared daily, just before use, by diluting the stock solution as appropriate with
2 mM of sodium tetraborate decahydrate (to a final concentration
of each analyte of 10 μg/mL) The stock standard solutions were kept in a freezer ( −20 °C), in closed containers and new solutions were prepared once every two weeks The working solutions were stored at 4 °C in closed containers for a maximum of 7 h
2.4 MEKC conditions
For MEKC separation, the following parameters were applied: uncoated fused silica capillary with an effective/total length of 50/60.2 cm and 75 μm i.d.; wavelength of UV detection 200 nm; hydrodynamic injection (15 s, at 0.5 psi); total analysis time
20 min; applied voltage 22 kV; temperature 25 ( ± 0.1) °C Between each run, the capillary was rinsed with 0.1 M NaOH for 1 min un- der a pressure of 50 psi and, subsequently, with the Milli-Q wa- ter for 1 min under a pressure of 50 psi The background elec- trolyte (BGE) consisted of 5 mM of sodium tetraborate decahydrate,
150 mM of boric acid, 50 mM of SDS and 15% ( v/v ) of MeOH The apparent pH value of the BGE equalled 7.3
2.5 DLLME conditions
1 mL of the human urine sample was spiked with the working solution of analytes at a concentration of 10 μg/mL Subsequently,
130 μL of cold acetone was added and the samples were shaken for
5 min (laboratory shaker – Elpin 358S, Lubawa, Poland) and cen- trifuged for 5 min (12 0 0 0 g ) (laboratory centrifuge – MPW-211
or MPW- 350R, Warsaw, Poland) Next, 1 mL of supernatant was separated and placed in a clean glass tube and then 1 mL of EtOH and 500 μL of DCM were added The samples were shaken me- chanically for 10 min and centrifuged for 4 min (4 0 0 0 g ) In order
to separate the organic phase, 400 μL of the solution was taken from the bottom of the glass tube and transferred to a clean Ep- pendorf tube and then evaporated to dryness at 45 °C (Labconco®, Kansas City, Missouri, USA) The residue was dissolved in 150 μL of
2 mM sodium tetraborate using a long vortex time (2 min for each sample) Then the samples were injected into the capillary and an- alyzed by the elaborated MEKC method
2.6 SPE conditions
SPE (Agilent Vac Elut SPS 24 Manifold, Santa Clara, United States) was carried out on hydrophilic-lipophilic balanced (HLB) (Supel TM-select HLB, Sigma Aldrich, Germany), octadecyl sor- bent (C18) (Discovery® DSC-18, Sigma Darmstadt, Germany) and cyanopropyl (CN) (Chromabond®-CN, VWR, Gda ´nsk, Poland) car- tridges, previously activated with 1 mL of MeOH and 1 mL of Milli-
Q water The 1 mL of human urine was spiked with the working solution of analytes at a concentration of 10 μg/mL Next, the sam- ples were applied to the SPE columns which were next washed with 1 mL of Milli-Q water and dried in a vacuum for 5 min The analytes were desorbed from the SPE cartridges to clean glass tubes with 1 mL of one of the tested eluents: MeOH, DCM, hex- ane, acetone and ACN:MeOH (1:1, v/v ) The solvent was evaporated
to dryness at 45 °C The dry residue was dissolved in 150 μL of
2 mM sodium tetraborate and analyzed by the elaborated MEKC method
2.7 SPME conditions
1 mL of the human urine sample was spiked with the working solution of analytes at a concentration of 10 μg/mL In the mean- time, 96-well SPME brushes with polystyrene-divinylbenzene (PS- DVB) resin, which was used as complementary to SPE HLB-type
Trang 44 N Mi ˛ekus, A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032
resin or C18 resin, were conditioned with 1 mL of MeOH/H 2O (1:1;
v/v ) for 30 min and washed with 1 mL of deionized water for
10 s Then, 1 mL of each sample was applied to the wells of a 96-
well plate and the BAs were extracted for 60 min (shaking speed
850 rpm) Then the brush fibers were washed again with deionized
water for 10 to remove impurities and a 60 min desorption step
with one of the four tested desorbents: MeOH, acetone, ACN:MeOH
(1:1; v/v ) or DCM was carried out Afterwards, the samples were
evaporated to dryness on a centrivap (45 °C, 1.5 h) and the residue
was dissolved in 150 μL of 2 mM sodium tetraborate and analyzed
by the elaborated MEKC method
2.8 Data analysis
In order to calculate the EF of each tested method, the MEKC
separation of the sample undergoing the sample extraction proce-
dure was carried out, as well as the control sample containing all
the analytes at a concentration of 10 μg/mL in 2 mM borax, which
was not undergoing the sample preparation procedure The value
of the EF was calculated according to Eq.(1):
where: H – the peak height of the analyte determined by the MEKC
method in the human urine sample undergoing the extraction pro-
cedure, H 0 – the peak height of the analyte determined by MEKC
in the control sample without sample pretreatment
Due to the fact that the sample undergoing the extraction pro-
cedure was evaporated and next, the residue was dissolved in 150
μL of 2 mM sodium tetraborate, the effect was the concentration
of the analyte in the sample Therefore, the calculated height of
the peak could be higher than that of the control sample without
sample pretreatment In consequence, the value of the EF calcu-
lated according to Eq.(1)could be above 1
The comparative study of the EF results of seven BAs from urine
samples obtained under the 25 tested sample preparation pro-
cedures was conducted under HCA using the Euclidean distance
method and the single linkage method Statistica 13.3 software
(StatSoft, Tulsa, USA) was used to achieve the task The numbering
of the extraction procedures, as presented in Table1, was retained
unchanged in the chemometric calculation
3 Results and discussion
Analyte pre-concentration procedures are essential, particularly
when less sensitive detection methods (spectrophotometric, e.g
DAD) are employed The most common purification techniques that
allow a differentiated degree of isolation of analytes are liquid-
liquid extraction (LLE), DLLME, SPE, SPME and their variants and
combinations Each of them provides a different degree of analyte
concentration (initial off-line concentration) and each is differently
useful for the isolation of specific analytes from the matrix [12] In
the case of BAs, this is particularly important because their concen-
trations in biological matrices are extremely low (ng, pg or less),
and BAs have a hydrophilic nature and are characterized by photo-
and thermo-lability [13]
A detailed description of the aforementioned sample prepara-
tion protocols has recently been described in our previous papers
[ 7, 10, 11] For this reason, in the presented study, the advantages
and disadvantages of each of the isolation methods were omit-
ted, while the focus was on the isolation of seven BA metabo-
lites with three different analytical approaches based on DLLME,
SPE or SPME (fully described in Sections2.5–2.7) To the best of
our knowledge, there have been no other studies to date for the
simultaneous determination of such a large group of compounds
from the -tyrosine metabolic pathway The extraction efficiency
for each of the method was confirmed trough the analysis of elec- tropherograms obtained by the MEKC-DAD method supported by the HCA chemometric analysis The signal recorded on the electro- pherogram from each isolated analyte was compared with the sig- nal from the reference sample with the same concentration of the test compound At the beginning, the sample buffer and the BGE composition were optimized to ensure the best separation condi- tions for multiple biomolecules
3.1 The BGE and sample buffer for the simultaneous separation of analytes
Because of the diversity of pKa values (data in Fig 1) and the amphoteric nature of the selected panel of BAs, their simultaneous determination by conventional capillary zone electrophoresis (CZE)
is relatively difficult Moreover, their chemical structures contain a few functional groups which could be ionized in a wide pH range, hence finding the optimal ingredients of the BGE is a real chal- lenge In our study, a borate buffer was selected because catechol compounds and some substituted catechols (like BAs) can become charged in a weakly alkaline electrolyte In effect, the analytes con- tain vicinal hydroxyl groups which after becoming charged are able
to form complexes with borate ions [ 14, 15], which allows adequate electrophoretic mobility to be obtained However, due to the lack
of a satisfactory separation of all analytes, the addition of surfac- tants and an organic modifier was examined For this purpose, the influence of an anionic surfactant such as SDS in the range of 0 – 50 mM was tested A borate buffer without SDS and one at a concentration below 50 mM SDS did not allow satisfactory separa- tion; however, a concentration at 50 mM SDS gave full separation
of seven compounds of interest ( Fig.2) Moreover, ACN and MeOH (in different volume proportions in the range of 5 – 20%, v/v ) were tested as organic components of the running buffer in order to in- crease sensitivity and improve the resolution for BAs The experi- mental results indicated that the effective separation of the peaks
of interest could be observed when 15% of MeOH ( v/v ) was added
to the BGE Therefore, a mixture of sodium tetraborate (5 mM), boric acid (150 mM), SDS (50 mM) and MeOH (15%, v/v ) (apparent
pH 7.3) was selected to separate all analytes at the highest resolu- tion without any interferences ( Fig.3) In these MEKC conditions, DHPG, VMA, MHPG and HVA were cations and therefore did not interact with the hydrophobic interior of SDS micelles, and their migration times (MTs) were shorter than NM, M and DOPAC, which were anions
To obtain an increase in the analyte signal, it was also necessary for the injection parameters and the sample buffer composition to
be optimized The ionic strength of the sample plays a significant role in CE and can positively or negatively affect the separation, the detectability of analytes and the migration time The viscosity and
pH of the sample buffer as well as the potential amount of organic component in the sample are of great importance The appropriate selection of the sample buffer allows for significant narrowing of the analyte band, which promotes the simultaneous separation of many components of the analyzed mixture
In our research, we simultaneously developed two online pre- concentration techniques: the first was stacking – accomplished by placing the sample in a solution the ionic strength of which is sig- nificantly less than that of the separation buffer, and the second – sweeping, based on the interaction between analytes in the ma- trix free of the SDS and the surfactant molecule-formed pseudo- stationary phase in the BGE In this study, for the selection of the sample buffer (injection medium), different borax concentrations (in the range of 2 – 10 mM) were tested Our research showed the best sharpness and symmetry of peaks for a sample containing a
2 mM borax solution Ultimately, experimental data revealed that
Trang 5Mean EF-values for seven BAs extracted from urine samples with twenty five different sample preparation procedures ( n = 3)
Type of solid phase in
MeOH/ACN (1:1,
Type of solid phase in
SPE/SPME
(1:1, v/v )
MeOH/ACN (1:1, v/v )
IL2
MeOH with IL3
DCM
Legend: EF – enrichment factor; BAs – biogenic amines; DLLME – dispersive liquid-liquid microextraction; SPE – solid phase extraction; SPME – solid phase microextraction; HLB – hydrophilic-lipophilic balanced sorbent; DVB – divinylbenzene resin; C18 – octadecyl sorbent; CN – cyanopropyl sorbent; MeOH – methanol; ACN – acetonitrile; DCM – dichloromethane; IL1 – 1–butyl–3-methylimidazolium tetrafluoroborate; IL2 – 1-ethyl-3-methyl- imidazolium tetrafluoroborate; IL3 – 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide; DHPG – dihydroxyphenylglycol; VMA – vanillylmandelic acid; MHPG – methoxy–4-hydroxyphenyl glycol; HVA – homovanillic acid; NM – normetanephrine; M – metanephrine; DOPAC – dihydroxyphenylacetic acid
Trang 66 N Mi ˛ekus, A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032
Fig 2 Effect of SDS concentration in the electrolyte on the MEKC separation of BAs metabolites Separation parameters: applied voltage 22 kV, an effective/total capillary
length 50/60.2 cm and 75 μm i.d., λ= 200 nm, hydrodynamic injection 15 s at 0.5 psi, temp 25 ( ± 0.1) °C; BGE: 5 mM sodium tetraborate decahydrate, 150 mM boric acid,
50 mM SDS and 15% ( v/v ) MeOH, pH = 7.3
Legend: 1 – DHPG, 2 – VMA, 3 – MHPG, 4 – HVA, 5 – NM, 6 – M, 7 – DOPAC
Fig 3 Electropherogram obtained for the standard sample of seven analytes (each at the concentration of 10 μg/mL) dissolved in 2 mM sodium tetraborate (water solution)
under the optimized MEKC conditions Separation parameters and legend as in Fig 2
Legend: 1 – DHPG, 2 – VMA, 3 – MHPG, 4 – HVA, 5 – NM, 6 – M, 7 – DOPAC
the 2 mM sodium tetraborate (pH 8.7) solution was optimal as a
sample buffer for the studied compounds
The developed MEKC-DAD method with the optimized sample
buffer, BGE, injection time, pressure, current and capillary length
allowed limits of detection (LODs) to be obtained of less than
0.1 μg / ml for all BAs
3.2 Verification of the isolation and preconcentration methods
The DLLME approach for DA, A, NA, l-Tryp, l-Tyr, 5-HT, l-DOPA [ 10, 16] as well as HVA and VMA [11]was previously evaluated by our team [ 10, 11, 16] Here, it was applied for a new group of ana- lytes In the case of SPE, knowing the physicochemical properties
of DHPG, VMA, MHPG, HVA, M, NM and DOPAC, SPE with HLB,
Trang 7CN or C18 cartridges was applied for the isolation of these an-
alytes from the standard urine samples In turn, the SPME-based
methods were carried out with the application of similar solvents
for desorption as those used in the SPE-based methods described
here Also, based on our previous research – ionic liquids (IL)
were tested as new SPME desorbent additives which could improve
the isolation and preconcentration of analytes, e.g DA , A , NA , l
Tryp and l Tyr [7] The elaborated DLLME, SPE and SPME extrac-
tion methods were compared by the calculation of their EF-values
on the basis of the peak heights of the analytes obtained by the
MEKC-DAD analysis The obtained results, summarized in Table1,
indicate that the final extractions of BAs from urine depend on the
physicochemical nature of the analytes, the type of sorbents used,
and the eluting/desorbing agents The extraction results indicate
that in the case of SPE procedures based on MeOH as the elut-
ing agent, the most effective solid phase to retain the analytes was
the non-polar sorbent (C18), followed by HLB, while CN offered the
worst retention However, the efficiency of the tested organic sol-
vents for the simultaneous elution of the analytes was varied Thus,
the most effective simultaneous elution of the analytes, except for
M, was obtained using SPE-C18 with MeOH as the eluting agent
For HLB cartridges, the use of MeOH gave also the best elution re-
sults, although these values were significantly lower than for C18
In the case of CN, acetone was the most effective eluting agent In
the case of SPME procedures, the properties of PS-DVB sorbent in
combination with MeOH as the desorbent allowed very good ex-
traction efficiency to be obtained with EF-values of 5.2 for M, 4.1
for NM, 3.9 for HVA, 1.9 for DHPG, MHPG and DOPAC, and 1.8 for
VMA On the other hand, the EF value for M increased when ILs
were added to MeOH, and obtained the highest value for MeOH
with IL3 (6.2) For SPME procedures, the possibilities for the effec-
tive desorption of the analytes by the tested organic solvents were
also different, and they were dependent on the type of SPME fiber
used However, among the tested organic solvents, DCM gave the
worst results
Next, these data were evaluated by the HCA multivariate anal-
ysis in order to find the most effective sample preparation pro-
cedure for the simultaneous isolation of the analytes HCA offers
graphic data visualization of the relationships between the vari-
ables and/or objects without losing any significant information
This statistical approach is based on an algorithm that groups simi-
lar objects into groups called clusters The endpoint is a set of clus-
ters, where each cluster is distinct from any other cluster, and the
objects within each cluster are broadly similar to each other The
main output of HCA is a dendrogram which shows the hierarchical
relationship between the clusters [17] In this study, the Euclidean
distance was used for measuring the dissimilarity between each
pair of observations, while average linkage clustering was applied
to determine which clusters should be joined at each stage Den-
drograms calculated on the basis of the established EF-values are
illustrated in Fig.4A (variables) and 4B (objects), respectively
3.2.1 Relationships between the tested sample preparation protocols
established by HCA
According to the HCA results for the variables ( Table1, Fig.4A),
the tested extraction procedures were located in clusters I, II and
III Taking into account the distance between the variables ob-
served on the dendrogram, the biggest differences can be noticed
for the SPE-C18 procedures included in cluster I Among them,
SPE_7 with MeOH as the eluting agent was found as an outlier
This procedure offered the highest EFs for DHPG (3.5), VMA (4.9),
NM (4.5), DOPAC (5.6) and HVA (7.8) Slightly lower values were
measured for MHPG in comparison to those calculated after using
SPME_21 (1.8 vs 1.9) Only one analyte – M – was poorly extracted
from urine samples (EF = 0.6) The application of the mixture of
MeOH/ACN (1:1, v/v ) as the eluting agent (SPE_10) provided a more
effective extraction of M (EF = 1.7) and a comparable extraction of DHPG (EF = 3.4) and HVA (EF = 7.4) On the other hand, a less ef- fective isolation of DOPAC (EF = 4.1) and VMA (EF = 2.5) was ob- tained, while the lowest isolation was found for MHPG (EF = 0.6) and NM (EF = 0.3) The EF-values of the analytes were more com- parable after using SPE_6 with acetone, although this solute mod- ification caused a further decrease in efficiency for DHPG, VMA, MHPG, HVA, M and DOPAC However, in the case of NM, this effect was contrary to using SPE_10 The EF-value for NM after SPE_10 was 0.3, while after the procedure with SPE_6, it was 2.8 Sum- marizing, among the procedures positioned in cluster I, the SPE_7 protocol located on the left of the dendrogram offered the best EF results for six of the tested analytes Only M was poorly extracted using this protocol
Taking into account the sample preparation protocols located in cluster II, it can be noticed that each of them was based on SPME with PS-DVB coatings The methods using MeOH with the addi- tion of three different ILs (IL1, IL2 and IL3) as desorbing solvents (SPME_22–24, respectively) were located together, while the proto- col with pure MeOH (SPME_21) was found as an outlier of cluster
II In fact, SPME_22 and 23 gave comparable EF results for all tested analytes, but they were lower than after using SPE_7 The only EF parameter achieved for M was almost 10 times higher (5.2) On the other hand, this value was slightly lower than that calculated for SPME_24 with IL3 (EF = 6.2) This protocol was also more ef- fective than SPME_22 and 23 for other tested analytes Thus, the position of SPME_24 at a small distance to the above-mentioned procedures is fully justified As it was mentioned above, SPME_21 was positioned on the right of these procedures, which offered the more effective isolation of DHPG, VMA, MHPG, HVA and DOPAC This confirms that the modification of the desorbing solvent by the addition of ILs, especially IL1 and IL2, should be avoided for the an- alytes containing acid groups (VMA, HVA and DOPAC) or more than two hydroxide groups in the side chain of the molecule (DHPG, MHPG)
It can also be noticed that most of the tested sample prepara- tion procedures with more complicated structures were located in cluster III Therefore, SPME_20 based on PS-DVB and SPE_3 based
on HLB coatings were positioned in subcluster IIIA located closely
to cluster II For them, acetone (SPME_20) and pure MeOH (SPE_3) were used for the desorption/elution of the compounds of interest The EF parameters were higher for the analytes, except for DOPAC, after SPE_3 SPME_C18 with the mixture of MeOH:ACN as the des- orbing solvent was located on the right of subcluster IIIA as an outlier This protocol offered a significantly less effective extraction
of the analytes than that calculated for the protocols included in clusters I and II, except for M For this analyte, EF parameters al- most 5 times higher were calculated with respect to SPE_7 Two SPE procedures based on CN coatings were placed in subcluster III B1 These protocols gave low EF results for all analytes except DHPG
The SPME_19 protocol based on the PS-DVB cartridge and the mixture of MeOH/ACN as a desorbing agent was positioned on the right as an outlier of subcluster III B.2 This method offered very low efficiency of the analytes (EFs from 0.0 to 0.5) Five proto- cols with various solid phases were located in subcluster III B 2.1, whereas SPE_5 using hexane was positioned on the right of this group as an outlier These procedures were also described by very low EF parameters in comparison to other tested protocols, espe- cially SPME_25 and SPE_5, which were not able to isolate any com- pound of interest Slightly more effective protocols with respect
to those located in subcluster III B.1 were located in subcluster III B2.2, where DCM in combination with C18 (SPME_15) and the HLB cartridge (SPE_2) as well as HLB and acetone as the eluting agent (SPE_4) were applied Unfortunately, these methods also offered poor extraction of the analytes On the left of the dendrogram, sub-
Trang 88 N Mi ˛ekus, A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032
Fig 4 HCA results obtained for the variables (A) and the objects (B) on the basis of EF-values obtained for the analytes after DLLME, SPME and SPE protocols following the
MEKC method
cluster III B3 was distinguished, which contained two procedures
based on C18 (SPME_16 and 17) and the DLLME_1 method Their
positions with respect to cluster III B2.2 suggest that slightly dif-
ferent EF results were calculated In fact, these protocols offered
more effective isolation of all tested analytes, especially for HVA
Unfortunately, each of them offered significantly worse results than
those determined for other protocols included in clusters I and II
3.2.2 Relationships between the tested analytes established by HCA
Taking into account the HCA results for the objects, it can be
observed that seven tested analytes were located in clusters I and
II ( Fig.4B) Their positions were clearly correlated with their chem- ical structures, which defined the different physicochemical char- acteristics of these molecules Therefore, M and NM which possess the same chemical structure of the main molecule, but with a dif- ferent type of amino group in the side chain (-NHCH 3 and -NH 2, respectively, Fig.1) were included in cluster I On the other hand, the relatively high distance between them indicates that relatively different EF results were calculated from them This confirms that the type of amino group can decide about the final interaction be- tween the analyte and the molecules of the solvents used as ex- traction/desorption agents or mobile-/solid-phase components de-
Trang 9pending on specific experimental conditions The analytes included
in cluster II possess carboxyl groups (HVA, DOPAC, VMA) or two
hydroxide groups in the side chain of the molecule (MHPG, DHPG)
( Fig.1) Among them, HVA was located as an outlier of cluster II,
while MHPG was positioned at a small distance to DOPAC, VMA
and DHPG This can be correlated with the fact that twenty ex-
traction protocols were able to isolate HVA more effectively than
other analytes, especially SPE_3, 6, 7 and 10, and SPME_16 and
20 Moreover, pure MeOH and the mixture of MeOH/ACN were the
best eluting solvents for HVA, whereas the addition of ILs caused
a decrease of this parameter In contrast to HVA, MHPG was sig-
nificantly less effectively isolated from urine than other analytes,
e.g after the application of DLLME_1, SPE_3 and 6, and SPME_18–
20 The most effective protocols for this compound were SPME_21
and SPE_7 DOPAC and VMA possessing carboxyl groups were com-
parably isolated using most tested protocols SPE_7 offered the
best conditions for the extraction of these BAs On the other hand,
SPE_10, especially for DOPAC, can be considered as an interesting
alternative
Summarizing, the obtained HCA results indicated that SPE-C18
with MeOH as the eluting agent offered the most effective isola-
tion of DHPG, VMA, HVA, NM and DOPAC, whereas SPME-PS-DVB
with the same solvent for the desorption was the best choice for M
and MHPG This SPME approach also guaranteed a relatively high
extraction for other BAs Moreover, the addition of IL3 (1-ethyl-3-
methylimidazolium bis(trifluoromethylsulfonyl)imide) to MeOH in
SPME increased the efficiency for M, did not change the extrac-
tion parameters for DHPG and NM, but should be avoided for VMA,
MHPG, HVA and DOPAC It was also confirmed that the success or
failure of the tested extraction procedures was dependent on the
specific chemical structures of the BAs
4 Conclusions
The present research is the first example of the comparison of
three different extraction approaches based on DLLME, SPME and
SPE for the isolation of compounds from the l-tyrosine metabolic
pathway from human urine samples In this study, one DLLME,
thirteen SPE and eleven different SPME protocols were performed,
then the separation of the analytes based on the developed MEKC
method was performed, and next, the EF values for each analyte in
specific extraction conditions were calculated Finally, the EF values
were compared by HCA The obtained results confirmed that the
use of HCA increases the probability of the selection of the most
appropriate sample preparation procedure for the specific analy-
sis, including the simultaneous or specific determination of the se-
lected BAs Similarly, the HCA results showed that SPE-C18 with
MeOH as the eluting agent and SPME-PS-DVB with the same sol-
vent for the desorption should be considered as alternative tools
for the extraction of the seven tested BAs The addition of 1-ethyl-
3-methylimidazolium bis(trifluoromethylsulfonyl)imide to MeOH in
SPME offered a more effective extraction of M but can decrease ef-
ficiency for VMA, MHPG, HVA and DOPAC Thus, the application of
multivariate data processing, i.e HCA can be considered as a valu-
able starting point for improving the reliable evaluation of sam-
ple preparation protocols in pharmaceutical practice Moreover, it
was confirmed that the developed MEKC method, supported by
SPE or SPME, can be used as an off-line preconcentration technique
for the simultaneous isolation and determination of seven catechol
compounds in urine samples for diagnostic purposes
Author’s contribution
N.M coordinated the manuscript writing and submission, N.M.,
I.O planned all the experiments, wrote the experimental part of
the manuscript and introduction section, N.K., M.R., N.M per- formed the experiments and collected the raw data, I.O., P.K opti- mize the BGE for MEKC separation, wrote parts focused on the BGE optimization in “Results and Discussion” section, A.P performed the HCA chemometric analysis of raw experimental data and de- scribed them in the manuscript, T.B obtained financial support for the experiments and manuscript publication, T.B., I.O supervised the experimental procedures, All authors prepared and approved all the files related to Manuscript (figures, responses to Reviewers’ comments, tables)
Acknowledgement
The authors acknowledge the support of the MTB Korea V4 joint project from the following sources: National Center for Research and Development in Poland (DZP/V4-Korea- I/20/2018)
References
[1] H Lehnert , Pheochromocytoma : Pathophysiology and Clinical management, Frontiers of Hormone Research, Karger, Basel, Switzerland, 2004
[2] James Mike , Anaesthesia For Patients With Endocrine Disease, Oxford Univer- sity Press, NA, USA, 2010
[3] K Pacak , Phaeochromocytoma: a catecholamine and oxidative stress disorder, Endocr Regul 45 (2011) 65–90
[4] N Mi ˛ekus, T B ˛aczek, Non-invasive screening for neuroendocrine tumors - Bio- genic amines as neoplasm biomarkers and the potential improvement of "gold standards, J Pharm Biomed Anal 130 (2016) 194–201, doi: 10.1016/j.jpba.2016 06.013
[5] Catecholamines, Bridging Basic Science With Clinical Medicine, 1st Edition, Academic Press, NY, USA, 1998
[6] G Eisenhofer, M Peitzsch, B.C McWhinney, Impact of LC-MS/MS on the lab- oratory diagnosis of catecholamine-producing tumors, Trends Anal Chem 84 (2016) 106–116, doi: 10.1016/j.trac.2016.01.027
[7] N Mi ˛ekus, I Ol ˛edzka, N Kossakowska, A Plenis, P Kowalski, A Prahl,
T B ˛aczek, Ionic liquids as signal amplifiers for the simultaneous extraction of several neurotransmitters determined by micellar electrokinetic chromatogra- phy, Talanta 186 (2018) 119–123, doi: 10.1016/j.talanta.2018.04.041
[8] N Kossakowska, I Ol ˛edzka, A Kowalik, N Mi ˛ekus, P Kowalski, A Plenis, E Bie ´n, A Kaczorowska, M.A Krawczyk, E Adamkiewicz-Dro ˙zy ´nska, T B ˛aczek, Application of SPME supported by ionic liquids for the determination of bio- genic amines by MEKC in clinical practice, J Pharm Biomed Anal 173 (2019) 24–30 10.1016/j.jpba.2019.05.021
[9] N.F Atta, E.H El-Ads, Y.M Ahmed, A Galal, Determination of some neurotrans- mitters at cyclodextrin/ionic liquid crystal/graphene composite electrode, Elec- trochim Acta 199 (2016) 319–331, doi: 10.1016/j.electacta.2016.02.078 [10] N Mi ˛ekus, I Ol ˛edzka, A Plenis, P Kowalski, E Bie ´n, A Mi ˛ekus, M.A Krawczyk,
E Adamkiewicz-Dro ˙zy ´nska, T B ˛aczek, Determination of urinary biogenic amines’ biomarker profile in neuroblastoma and pheochromocytoma patients
by MEKC method with preceding dispersive liquid–liquid microextraction, J Chromatogr B 1036–1037 (2016) 114–123, doi: 10.1016/j.jchromb.2016.10.007 [11] N Miekus, P Kowalski, I Oledzka, A Plenis, E Bie ´n, A Miekus, M Krawczyk,
E Adamkiewicz-Drozy ´nska, T Baczek, Cyclodextrin-modified MEKC method for quantification of selected acidic metabolites of catecholamines in the presence
of various biogenic amines application to diagnosis of neuroblastoma, Journal
of Chromatography B: Analytical Technologies in the Biomedical and Life Sci- ences 1003 (2015) 27–34, doi: 10.1016/j.jchromb.2015.09.003
[12] N Drouin, S Rudaz, J Schappler, Sample preparation for polar metabolites in bioanalysis, Analyst 143 (2018) 16–20, doi: 10.1039/c7an01333g
[13] A Plenis, I Ol ˛edzka, P Kowalski, N Mi ˛ekus, T B ˛aczek, Recent trends in the quantification of biogenic amines in biofluids as biomarkers of various disor- ders: a review, J Clin Med 8 (2019) 640, doi: 10.3390/jcm8050640 [14] J Cao, B Li, Y.-.X Chang, P Li, Direct on-line analysis of neutral analytes
by dual sweeping via complexation and organic solvent field enhancement
in nonionic MEKC, Electrophoresis 30 (2009) 1372–1379, doi: 10.1002/elps
20 080 0523 [15] M.J Markuszewski, P Britz-McKibbin, S Terabe, K Matsuda, T Nishioka, De- termination of pyridine and adenine nucleotide metabolites in bacillus subtilis cell extract by sweeping borate complexation capillary electrophoresis, Jour- nal of Chromatography A 989 (2003) 293–301, doi: 10.1016/s0021-9673(03)
0 0 031-1 [16] N Mi ˛ekus, I Ol ˛edzka, D Harshkova, I Liakh, A Plenis, P Kowalski, T B ˛aczek, Comparison of three extraction approaches for the isolation of neurotrans- mitters from rat brain samples, Int J Mol Sci 19 (2018) 1–11, doi: 10.3390/ ijms19061560
[17] S Lin, S Van Poucke, Z Zhang, P Lan, F Murtagh, Hierarchical cluster analy- sis in clinical research with heterogeneous study population: highlighting its visualization with R, Ann Transl Med 5 (2017) 75–75, doi: 10.21037/atm.2017 02.05