Extraction and preconcentration of compounds from the l-tyrosine metabolic pathway prior to their micellar electrokinetic chromatography separation

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.

Journal of Chromatography A 1620 (2020) 461032 Contents lists available at ScienceDirect Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma Extraction and preconcentration of compounds from the l-tyrosine metabolic pathway prior to their micellar electrokinetic chromatography separation ∗ Natalia Miekus ˛ , Alina Plenis, Marta Rudnicka, Natalia Kossakowska, Ilona Oledzka, ˛ Piotr Kowalski, Tomasz Baczek ˛ ´ sk, Hallera, 107, 80-416, Gdan ´ sk, Poland Department of Pharmaceutical Chemistry, Medical University of Gdan a r t i c l e i n f o Article history: Received 15 October 2019 Revised 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 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–4hydroxyphenyl 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 For this reason, the three most popular extraction techniques (dispersive liquid-liquid microextraction (DLLME), solid-phase extraction (SPE), solid-phase microextraction (SPME)) were tested Micellar electrokinetic chromatography (MEKC) – a mode of capillary electrophoresis – with a diode array detector (DAD) was applied to assess the extraction efficiency Next, the enrichment factor (EF) of each applied method was calculated in respect to standard mixtures of the analytes at the same concentration levels The EF results of seven selected metabolites of biogenic amines (BAs) from urine after sample preparation procedures based on twenty-five different protocols (one DLLME, thirteen SPE and eleven SPME) were calculated and compared using hierarchical cluster analysis (HCA) The SPE as well as SPME procedures were proved to be the most effective approaches for the simultaneous extraction of the chosen compounds Moreover, an ionic liquid (IL) – 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide – added to methanol in SPME additionally could successfully improve the extraction efficiency It was also confirmed that the HCA approach could be considered a supportive tool in the selection of a suitable sample preparation procedure for that group of endogenous substances © 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/) Introduction The secretory cells of the adrenal medulla mainly produce catecholamines: 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 transferase (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 Miekus) ˛ 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 dihydroxyphenylglycol (DHPG) and in the adrenal medulla to normetanephrine (NM) [1] The physiological metabolism of those monoamine neurotransmitters (NTs) could be interrupted (or their ratios of concentration visibly changed) in pathophysiological stages of the human organism, which include neuroendocrine tumors (NETs) – pheochromocytoma (PHE) and neuroblastoma (NBL) [2,3] As such, the determination 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 Omethylated 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/) N Miekus, ˛ A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032 Fig 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 concentration 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 analysis 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 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 throughput 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 isolation methods should also have the advantage of preconcentrating analytes in the biological sample To address this issue, the three most popular and efficient extraction and preconcentration techniques: dispersive liquid-liquid microextraction (DLLME), solidphase 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 matrices, 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 consumption 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 approaches, the extraction efficiency was evaluated by the calculation of the enrichment factor (EF) This was done through a comparison 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 chromatography (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 liquids (ILs), namely: 1–butyl–3-methylimidazolium tetrafluoroborate (IL1), 1-ethyl-3-methylimidazolium tetrafluoroborate (IL2) and 1ethyl-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] N Miekus, ˛ A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032 In our earlier studies, experimental data demonstrated the usefulness 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 ltryptophan (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 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 investigated and discussed, in order to select the most efficient procedure providing the largest recovery of analytes and the most effective purification of the sample matrix It was decided to use hierarchical 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 MEKC method The reference urine samples were prepared daily, just before use, by diluting the stock solution as appropriate with 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 °C in closed containers for a maximum of h Materials and methods 2.5 DLLME conditions 2.1 Chemicals and reagents 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 (laboratory shaker – Elpin 358S, Lubawa, Poland) and centrifuged for (12 0 g) (laboratory centrifuge – MPW-211 or MPW- 350R, Warsaw, Poland) Next, mL of supernatant was separated and placed in a clean glass tube and then mL of EtOH and 500 μL of DCM were added The samples were shaken mechanically for 10 and centrifuged for (4 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 Eppendorf tube and then evaporated to dryness at 45 °C (Labconco®, Kansas City, Missouri, USA) The residue was dissolved in 150 μL of mM sodium tetraborate using a long vortex time (2 for each sample) Then the samples were injected into the capillary and analyzed by the elaborated MEKC method Methanol (MeOH), hexane, ethanol 96% (EtOH) and acetone 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-ethyl3-methyl-imidazolium tetrafluoroborate (IL2) and 1-ethyl-3methylimidazolium bis(trifluoromethylsulfonyl)imide (IL3) were supplied by Sigma-Aldrich (Darmstadt, Germany) Sodium tetraborate decahydrate (borax), boric acid and sodium hydroxide (NaOH) were obtained from Merck (Darmstadt, Germany) Capillary Regenerator 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) 2.4 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 under a pressure of 50 psi and, subsequently, with the Milli-Q water for under a pressure of 50 psi The background electrolyte (BGE) consisted of 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 2.6 SPE conditions 2.2 Apparatus All separation studies were carried out using a capillary electrophoresis (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 detector Analysis of the obtained data was made using 32 Karat 8.0 software (Beckmann, Fullerton, CA, USA) The device was additionally 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 mL of MeOH Subsequently, they were shaken on an MS 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 mM sodium tetraborate decahydrate and separated by the SPE (Agilent Vac Elut SPS 24 Manifold, Santa Clara, United States) was carried out on hydrophilic-lipophilic balanced (HLB) (SupelTM -select HLB, Sigma Aldrich, Germany), octadecyl sorbent (C18) (Discovery® DSC-18, Sigma Darmstadt, Germany) and ´ cyanopropyl (CN) (Chromabond®-CN, VWR, Gdansk, Poland) cartridges, previously activated with mL of MeOH and mL of MilliQ water The mL of human urine was spiked with the working solution of analytes at a concentration of 10 μg/mL Next, the samples were applied to the SPE columns which were next washed with mL of Milli-Q water and dried in a vacuum for The analytes were desorbed from the SPE cartridges to clean glass tubes with mL of one of the tested eluents: MeOH, DCM, hexane, 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 mM sodium tetraborate and analyzed by the elaborated MEKC method 2.7 SPME conditions mL of the human urine sample was spiked with the working solution of analytes at a concentration of 10 μg/mL In the meantime, 96-well SPME brushes with polystyrene-divinylbenzene (PSDVB) resin, which was used as complementary to SPE HLB-type N Miekus, ˛ A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032 resin or C18 resin, were conditioned with mL of MeOH/H2 O (1:1; v/v) for 30 and washed with mL of deionized water for 10 s Then, mL of each sample was applied to the wells of a 96well plate and the BAs were extracted for 60 (shaking speed 850 rpm) Then the brush fibers were washed again with deionized water for 10 s to remove impurities and a 60 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 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 procedure was carried out, as well as the control sample containing all the analytes at a concentration of 10 μg/mL in mM borax, which was not undergoing the sample preparation procedure The value of the EF was calculated according to Eq (1): EF = H/H0 (1) where: H – the peak height of the analyte determined by the MEKC method in the human urine sample undergoing the extraction procedure, H0 – 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 procedure was evaporated and next, the residue was dissolved in 150 μL of 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 calculated according to Eq (1) could be above The comparative study of the EF results of seven BAs from urine samples obtained under the 25 tested sample preparation procedures 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 Table 1, was retained unchanged in the chemometric calculation 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 liquidliquid 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 concentrations in biological matrices are extremely low (ng, pg or less), and BAs have a hydrophilic nature and are characterized by photoand thermo-lability [13] A detailed description of the aforementioned sample preparation 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 omitted, while the focus was on the isolation of seven BA metabolites with three different analytical approaches based on DLLME, SPE or SPME (fully described in Sections 2.5–2.7) 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 l-tyrosine metabolic pathway The extraction efficiency for each of the method was confirmed trough the analysis of electropherograms obtained by the MEKC-DAD method supported by the HCA chemometric analysis The signal recorded on the electropherogram from each isolated analyte was compared with the signal 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 conditions 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 challenge 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 contain 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 surfactants and an organic modifier was examined For this purpose, the influence of an anionic surfactant such as SDS in the range of – 50 mM was tested A borate buffer without SDS and one at a concentration below 50 mM SDS did not allow satisfactory separation; 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 – 20%, v/v) were tested as organic components of the running buffer in order to increase sensitivity and improve the resolution for BAs The experimental 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 resolution 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 preconcentration techniques: the first was stacking – accomplished by placing the sample in a solution the ionic strength of which is significantly less than that of the separation buffer, and the second – sweeping, based on the interaction between analytes in the matrix free of the SDS and the surfactant molecule-formed pseudostationary phase in the BGE In this study, for the selection of the sample buffer (injection medium), different borax concentrations (in the range of – 10 mM) were tested Our research showed the best sharpness and symmetry of peaks for a sample containing a mM borax solution Ultimately, experimental data revealed that Table Mean EF-values for seven BAs extracted from urine samples with twenty five different sample preparation procedures (n = 3) SPE_2 SPE_3 SPE_4 SPE_5 SPE_6 SPE_7 SPE_8 SPE_9 SPE_10 SPE_11 SPE_12 — HLB HLB HLB HLB C18 C18 C18 C18 CN CN DCM MeOH Acetone Hexane Acetone MeOH DCM Hexane C18 MeOH/ACN (1:1, v/v) MeOH Acetone 0.06 ± 0.005 0.1 ± 0.002 0.03 ± 0.002 0.4 ± 0.003 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.02 ± 0.001 0.03 ± 0.002 0.02 ± 0.002 0.0 ± 0.0 0.1 ± 0.01 0.03 ± 0.002 0.01 ± 0.002 0.2 ± 0.01 0.2 ± 0.02 0.1 ± 0.01 0.0 ± 0.0 0.07 ± 0.005 0.2 ± 0.01 0.0 ± 0.0 3.4 2.5 0.6 7.4 0.3 1.7 4.1 Eluting/desorbing agent — Analytes Mean EF-values (n = 3) DHPG VMA MHPG HVA NM M DOPAC 0.4 0.2 0.1 0.8 0.2 0.6 0.2 Type of solid phase in SPE/SPME ± ± ± ± ± ± ± 0.03 0.02 0.01 0.07 0.02 0.05 0.01 SPE_13 CN Eluting/desorbing agent Hexane 0.1 ± 0.01 0.2 ± 0.02 0.05 ± 0.004 0.3 ± 0.02 0.1 ± 0.01 0.02 ± 0.002 0.04 ± 0.003 1.2 1.8 0.2 3.1 0.4 1.7 0.5 SPE_14 CN SPME_15 C18 SPME_16 C18 SPME_17 C18 SPME_18 C18 SPME_19 PS-DVB SPME_20 PS-DVB SPME_21 PS-DVB SPME_22 PS-DVB SPME_23 PS-DVB DCM DCM Acetone MeOH MeOH/ACN (1:1, v/v) MeOH/ACN (1:1, v/v) Acetone MeOH MeOH with IL1 MeOH with MeOH with IL2 IL3 0.0 ± 0.0 0.02 ± 0.002 0.0 ± 0.0 0.3 ± 0.002 0.0 ± 0.0 0.06 ± 0.004 0.02 ± 0.002 0.1 ± 0.01 0.2 ± 0.02 0.08 ± 0.01 0.8 ± 0.05 0.0 ± 0.0 0.1 ± 0.01 0.4 ± 0.03 0.2 ± 0.01 0.2 ± 0.01 0.1 ± 0.01 0.8 ± 0.05 0.4 ± 0.03 0.7 ± 0.02 0.04 ± 0.003 0.2 ± 0.01 0.2 ± 0.01 0.06 ± 0.005 0.8 ± 0.04 0.6 ± 0.04 2.8 ± 0.2 0.3 ± 0.02 0.0 ± 0.0 0.02 ± 0.001 0.0 ± 0.0 0.03 ± 0.002 0.3 ± 0.02 0.05 ± 0.004 0.5 ± 0.02 0.7 1.4 0.1 2.0 0.4 1.2 1.2 Analytes Mean EF-values (n = 3) DHPG VMA MHPG HVA NM M DOPAC 0.1 ± 0.03 0.1 ± 0.02 0.1 ± 0.01 0.06 ± 0.005 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.03 ± 0.002 0.02 ± 0.001 0.01 ± 0.002 0.01 ± 0.001 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 ± ± ± ± ± ± ± 0.1 0.2 0.01 0.4 0.03 0.2 0.03 ± ± ± ± ± ± ± 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.3 1.7 0.5 4.0 2.8 1.1 2.5 ± ± ± ± ± ± ± 0.1 0.2 0.03 0.3 0.2 0.1 0.2 3.5 4.9 1.8 7.8 4.5 0.6 5.6 ± ± ± ± ± ± ± 0.3 0.3 0.2 0.6 0.4 0.05 0.4 ± ± ± ± ± ± ± 0.05 0.2 0.01 0.2 0.03 0.2 0.1 1.9 1.8 1.9 3.9 4.1 5.2 1.9 ± ± ± ± ± ± ± 0.2 0.1 0.1 0.3 0.3 0.4 0.1 0.8 0.7 1.0 1.8 4.1 5.6 1.1 ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0.03 0.2 0.005 0.6 0.02 0.2 0.3 0.06 0.05 0.1 0.2 0.3 0.4 0.1 1.5 0.6 0.6 0.6 0.0 0.0 0.8 0.6 0.7 1.0 1.8 3.3 5.7 1.0 ± ± ± ± ± ± ± ± ± ± ± ± ± ± 1.6 ± 0.1 0.9 ± 0.05 0.6 ± 0.04 0.8 ± 0.03 0.5 ± 0.02 0.1 ± 0.01 0.6 ± 0.04 0.2 0.04 0.05 0.03 0.0 0.0 0.06 0.04 0.05 0.1 0.1 0.3 0.4 0.1 SPME_24 PS-DVB 1.8 0.9 1.1 2.4 4.2 6.2 1.1 ± ± ± ± ± ± ± 0.1 0.1 0.1 0.2 0.3 0.4 0.1 SPME_25 PS-DVB DCM 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ± ± ± ± ± ± ± 0.0 0.0 0.0 0.0 0.0 0.0 0.0 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-methylimidazolium 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 N Miekus, ˛ A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032 Type of solid phase in SPE/SPME DLLME_1 N Miekus, ˛ A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032 Fig 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: mM sodium tetraborate decahydrate, 150 mM boric acid, 50 mM SDS and 15% (v/v) MeOH, pH = 7.3 Legend: – DHPG, – VMA, – MHPG, – HVA, – NM, – M, – DOPAC Fig Electropherogram obtained for the standard sample of seven analytes (each at the concentration of 10 μg/mL) dissolved in mM sodium tetraborate (water solution) under the optimized MEKC conditions Separation parameters and legend as in Fig Legend: – DHPG, – VMA, – MHPG, – HVA, – NM, – M, – DOPAC the 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 analytes In the case of SPE, knowing the physicochemical properties of DHPG, VMA, MHPG, HVA, M, NM and DOPAC, SPE with HLB, N Miekus, ˛ A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032 CN or C18 cartridges was applied for the isolation of these analytes 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, lTryp and l-Tyr [7] The elaborated DLLME, SPE and SPME extraction 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 Table 1, 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 eluting 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 solvents 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 results, 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 extraction 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 effective 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 analysis in order to find the most effective sample preparation procedure for the simultaneous isolation of the analytes HCA offers graphic data visualization of the relationships between the variables and/or objects without losing any significant information This statistical approach is based on an algorithm that groups similar objects into groups called clusters The endpoint is a set of clusters, 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 Dendrograms 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 (Table 1, Fig 4A), the tested extraction procedures were located in clusters I, II and III Taking into account the distance between the variables observed 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 effective isolation of DOPAC (EF = 4.1) and VMA (EF = 2.5) was obtained, while the lowest isolation was found for MHPG (EF = 0.6) and NM (EF = 0.3) The EF-values of the analytes were more comparable after using SPE_6 with acetone, although this solute modification 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 Summarizing, 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 addition of three different ILs (IL1, IL2 and IL3) as desorbing solvents (SPME_22–24, respectively) were located together, while the protocol 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 effective 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 analytes 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 preparation 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 desorbing 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 almost 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 protocols 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, especially SPME_25 and SPE_5, which were not able to isolate any compound 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- N Miekus, ˛ A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032 Fig 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 different 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 chemical structures, which defined the different physicochemical characteristics of these molecules Therefore, M and NM which possess the same chemical structure of the main molecule, but with a different type of amino group in the side chain (-NHCH3 and -NH2 , 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 between the analyte and the molecules of the solvents used as extraction/desorption agents or mobile-/solid-phase components de- N Miekus, ˛ A Plenis and M Rudnicka et al / Journal of Chromatography A 1620 (2020) 461032 pending 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 extraction protocols were able to isolate HVA more effectively than other analytes, especially SPE_3, 6, 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 significantly 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 comparably 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 isolation 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-3methylimidazolium bis(trifluoromethylsulfonyl)imide) to MeOH in SPME increased the efficiency for M, did not change the extraction 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 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 analysis, including the simultaneous or specific determination of the selected BAs Similarly, the HCA results showed that SPE-C18 with MeOH as the eluting agent and SPME-PS-DVB with the same solvent for the desorption should be considered as alternative tools for the extraction of the seven tested BAs The addition of 1-ethyl3-methylimidazolium bis(trifluoromethylsulfonyl)imide to MeOH in SPME offered a more effective extraction of M but can decrease efficiency for VMA, MHPG, HVA and DOPAC Thus, the application of multivariate data processing, i.e HCA can be considered as a valuable starting point for improving the reliable evaluation of sample 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 performed the experiments and collected the raw data, I.O., P.K optimize 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 described 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 University Press, NA, USA, 2010 [3] K Pacak, Phaeochromocytoma: a catecholamine and oxidative stress disorder, Endocr Regul 45 (2011) 65–90 [4] N Miekus, ˛ T Baczek, ˛ Non-invasive screening for neuroendocrine tumors - Biogenic 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 laboratory diagnosis of catecholamine-producing tumors, Trends Anal Chem 84 (2016) 106–116, doi:10.1016/j.trac.2016.01.027 [7] N Miekus, ˛ I Oledzka, ˛ N Kossakowska, A Plenis, P Kowalski, A Prahl, T Baczek, ˛ Ionic liquids as signal amplifiers for the simultaneous extraction of several neurotransmitters determined by micellar electrokinetic chromatography, Talanta 186 (2018) 119–123, doi:10.1016/j.talanta.2018.04.041 [8] N Kossakowska, I Oledzka, ˛ A Kowalik, N Miekus, ˛ P Kowalski, A Plenis, E ´ A Kaczorowska, M.A Krawczyk, E Adamkiewicz-Drozy ˙ nska, ´ Bien, T Baczek, ˛ Application of SPME supported by ionic liquids for the determination of biogenic 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 neurotransmitters at cyclodextrin/ionic liquid crystal/graphene composite electrode, Electrochim Acta 199 (2016) 319–331, doi:10.1016/j.electacta.2016.02.078 ´ A Miekus, [10] N Miekus, ˛ I Oledzka, ˛ A Plenis, P Kowalski, E Bien, ˛ M.A Krawczyk, ˙ nska, ´ E Adamkiewicz-Drozy T Baczek, ˛ 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 ´ A Miekus, M Krawczyk, [11] N Miekus, P Kowalski, I Oledzka, A Plenis, E Bien, ´ E Adamkiewicz-Drozynska, 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 Sciences 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 Oledzka, ˛ P Kowalski, N Miekus, ˛ T Baczek, ˛ Recent trends in the quantification of biogenic amines in biofluids as biomarkers of various disorders: a review, J Clin Med (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, Determination of pyridine and adenine nucleotide metabolites in bacillus subtilis cell extract by sweeping borate complexation capillary electrophoresis, Journal of Chromatography A 989 (2003) 293–301, doi:10.1016/s0021-9673(03) 0 031-1 [16] N Miekus, ˛ I Oledzka, ˛ D Harshkova, I Liakh, A Plenis, P Kowalski, T Baczek, ˛ Comparison of three extraction approaches for the isolation of neurotransmitters 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 analysis in clinical research with heterogeneous study population: highlighting its visualization with R, Ann Transl Med (2017) 75–75, doi:10.21037/atm.2017 02.05  ... 2.5–2.7) 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 l-tyrosine metabolic pathway The extraction. .. 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. .. structures of the BAs 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