Abstract: Background and Objective: Tea polyphenols are increasingly gaining the interest of the scientific community as they are strong anti-oxidants and it appears that they have many beneficial effects on human health. Here in a simple and rapid high pressure liquid chromatographic (HPLC) method with diode array detection has been developed for the simultaneous determination of five tea polyphenols: Three flavan-3-ols [(+)-catechin, (-)-epicatechin, (-)-epigallocatechin gallate], a phenolic acid (gallic acid), a hydroxycinnamic acid (chlorogenic acid) and three purine alkaloids (caffeine, theobromine and theophylline) in blood serum. Methodology: A phenolic acid (protocatechuic acid) was chosen as the internal standard. The developed method was validated in terms of sensitivity, linearity, accuracy, precision, stability and selectivity. Results and Conclusion: The proposed method is simple, rapid and requires neither special preparation procedure nor highly sophisticated instrumentation; therefore it can be proposed as a useful tool for the evaluation of the beneficial effects of these compounds in human health. All examined compounds and the internal standard were separated within 21 min. For spiked serum samples good linearity (R2>0.9902) was observed in the range of 0.5-20 ng μL-1. Relative standard deviation was lower than 15.4% while the limit of detection was 0.2 ng μL-1.
INTRODUCTION
Polyphenols are a large group of structurally related compounds that can be found in vegetables, fruits, grain, green or black tea, coffee and cacao. The most important property of polyphenols is their powerful antioxidative ability. In addition to this, a large number of studies have correlated the consumption of these compounds with many health beneficial effects due to their anti-carcinogenic, anti-microbial, anti-inflammatory and anti-diabetic properties1-4.
According to the nature of the carbon skeleton, polyphenols can be classified into four main classes: flavonoids, phenolic acids and the less common stilbenes and lignans1,5.
Flavonoids are the most abundant natural polyphenols in human diet. They are divided into nine main classes, among which flavonols and flavan-3-ols are the most common4,5.
The main flavan-3-ols are catechins which are very abundant in tea. They are characterized by di-or tri-hydroxyl group substitution of the B ring and meta-5,7-dihydroxy substitution of the A ring. Catechins can be divided into free catechins like Catechin (C) and epicatechin (EC) and esterified catechins like epigallocatechin gallate (EGCG). Among flavan-3-ols, EGCG has the highest activity1,3,4,6.
Natural phenols with more distinct acid properties often have an acceptor group as a substituent in the benzene ring. Tea is an important source of Gallic Acid (GA) which is a trihydroxybenzoic acid. Chlorogenic acid (ClGn), a derivative of coumaryl-quinic acid and protocatechuic acid (PCA) Internal standard (I.S.), a dihydroxybenzoic acid can also be found in tea1,6.
Finally, tea leaves contain caffeine (CAF) (1,3,7-trimethylxanthine) and considerably smaller amounts of theobromine (TB) (3,7-dimethylxanthine) and theophylline (TF) (1,3-dimethylxanthine). These purine alkaloids can be analyzed simultaneously with catechins or other polyphenols1,6,7.
The chemical structures of the examined analytes are given in Fig. 1.
Flavonoids that contain one or more sugar groups are referred to as glycosides (or glucosides in case of a glucose moiety). When no sugar group is present, they are referred to as aglycones.
In flavonoid analysis, it plays important role whether to determine the target analytes in their various conjugated forms or in the form of aglycones. In biological fluids flavonoids exist as glucuronide and sulphate conjugates. In order to determine the total aglycone content, a hydrolysis step is required otherwise the free content is measured1,3.
The stability of tea catechins depends on the pH and temperature. Tea catechins in aqueous solutions are very stable in acidic conditions (pH<4), whereas they are unstable in alkaline conditions (pH>6). In addition to this, in order to avoid epimerization or degradation of the target analytes, high temperatures should be avoided1,2.
Many methods have been developed for the determination and quantification of tea polyphenols in biological fluids. Most of them use High Pressure Liquid Chromatography (HPLC) with different detection modes. HPLC with UV detection8-10 is the preferred method for the analysis of tea polyphenols in food and beverages while HPLC with coulometric electrochemical detection11-15 due to its selectivity and sensitivity and HPLC with fluorescence detection8,10 seems to be the most sensitive method. HPLC-MS or HPLC-tandem MS16-19 and HPLC-DAD-ESI/MS20, 21 are the most promising methods for the quantification of these antioxidants in biological fluids, however, these type of detectors are not usually available in common laboratories as they are very expensive. Another technique that has been used for the analysis of polyphenols in the biological area and showed good sensitivity is GC-MS, however this technique cannot overcome LC/MS(/MS) since the derivatization step which is necessary makes the method time-consuming and labor-intensive3,8,22.
In this study, a simple methodology for the simultaneous determination of three flavonoids, one phenolic acid, one hydroxycinnamic acid and three purine alkaloids with HPLC-DAD has been developed. Another phenolic acid has been chosen as the internal standard. No special sample preparation procedure was used and good recovery, low detection limits and good sensitivity were achieved.
To the best of our knowledge, this is the first direct method for the simultaneous determination of these polyphenols and purine alkaloids in blood serum.
MATERIALS AND METHODS
Chemicals and materials: (+)-Catechin hydrate (C) (≥96%), (-)-epigallocatechin gallate (EGCG) (≥95%), (-)-epicatechin (EC) (≥90%), caffeine (CAF) (purity not specified), theobromine (TB) (≥99%), theophylline anhydrous (TF) (≥99%), gallic acid (GA) (≥97.5%), chlorogenic acid (ClCn) hemihydrate (≥98%) were purchased from Sigma-Aldrich (St. Louis, MO, USA) while protocatechuic acid (PCA) (internal standard, purity not specified) was purchased from Alfa Aesar (Karlsruhe, Germany).
HPLC-grade methanol (MeOH), sodium acetate trihydrate (CH3COONa·3H2O), acetic acid (100%) (CH3CO2H), ethyl acetate (CH3CO2CH2CH3) and acetone ((CH3)2CO) were purchased from Merck KGaA (Darmstadt, Germany). HPLC-grade acetonitrile (ACN) was purchased from Fisher Scientific (Loughborough, UK). 2-Propanol (CH2CH(OH)CH3) was purchased from Panreac Quamica Sau (Barcelona, Spain). Trifluoroacetic acid (99%, extra pure) (CF3CO2H) was purchased from Acros Organics (New Jersey, USA). The ultrapure water used during the study was provided by a Milli-Q purification system (Millipore, Bedford, MA, USA).
Oasis® HLB 60 mg/3 mL extraction cartridges were purchased from Waters (MA, USA), LiChrolut® RP-18 200 mg/3 mL by Merck (Darmstadt, Germany) and Abselut Nexus 30 mg L-1 cm3 by Varian (CA, USA).
Instrumentation and chromatographic conditions: The HPLC system consisted of a Rheodyne 7725i injection valve (Rheodyne, Cotati, California, USA), an LC-10ADVP pump by Shimadzu (Kyoto, Japan) and a FCV-10ALVP solvent mixing system. For the acquisition of the data, a SCL-10AVP computing integrator and an SPD-M10AVP photodiode array detector were used.
A Lichrospher® 100, RP-18e (250x4 mm, 5 μm) analytical column purchased from Merck KGaA (Darmstadt, Germany) was used for the chromatographic separation at room temperature. The analytes were separated by gradient elution and the mobile phase consisted of acetic acid 1% (A), acetonitrile (B) and methanol (C). Acetic acid solution was filtered using a glass vacuum-filtration apparatus, purchased from Alltech Associates (Deerfield, IL, USA) and Whatman cellulose nitrate 0.2 μm membrane filters (Whatman Laboratory Division, Maidstone, England). Degassing of solvents was achieved by using a DGU-10B system. A flow rate of 1 mL min-1 was programmed during the analytical run and the injection volume was 20 μL. Monitoring of the analytes was performed at the wavelength of maximum absorbance which was 330 nm for ClGn and 270 nm for the rest of the analytes. Peak identification was performed both by retention times and by spectral information provided by the detector.
A VisiprepTM SPE vacuum manifold by Supelco (Bellefonte, PA, U.S.A.), a nine port Reacti-VapTM (model 18780) by PIERCE (Rockford, IL, USA) and a Glas-col small vortexer (Terre Haute, IN, USA) were used for sample preparation. Centrifugations were carried out on a model SD micro-centrifuge by Sigma-Aldrich (St. Louis, MO, USA).
Preparation of standards: Stock solutions of 200 ng μL-1 were prepared in a mixture of methanol-MQ water (30:70, v/v) for GA, TB, TF, ClGn, CAF, EGCG, EC and in methanol for C and PCA (internal standard) by dissolving the appropriate amount of each analyte. Working standard solutions were prepared by dissolving the appropriate amount in a mixture of methanol-MQ water (30:70, v:v) within the range of 0.5-20 ng μL-1. Each one contained 3 ng μL-1 of PCA (internal standard). Calibration standards were prepared by spiking working standards into serum samples. Solutions were stored at 4°C.
Sample preparation: For the extraction of the analytes from blood serum, a simple method was used. The blood serum (50 μL) was mixed with 200 μL of water for blank samples or 200 μL of working solution (at five concentration levels) for spiked samples. In order to prevent tea polyphenols from oxidation and maintain their stability an aliquot of 20 μL of an acetate buffer solution (0.2 M CH3COOH, 0.2 M CH3COONa, pH 3.5) was added. The mixtures were vortexed and centrifuged for 10 min at 6000 rpm. The supernatant was removed and 750 μL of ethyl acetate were added in the pellet. The new mixture was centrifuged for another 10 min. The two supernatants were mixed and after adding 20 μL of the acetate buffer (0.2 M CH3COOH, 0.2 M CH3COONa, pH 3.5). They were evaporated to dryness under a gentle stream of N2. The residues were reconstituted in 200 μL of the internal standard by vortexing. The sample was centrifuged for 10 min and 20 μL of the sample was injected onto the HPLC column.
Method validation: The developed analytical method was fully validated in terms of selectivity, linearity, sensitivity, within-day and between-day precision, accuracy and stability. For the demonstration of the selectivity blank serum samples were subjected to the extraction procedure and injected onto the HPLC column. The resulting chromatograms were evaluated for possible interferences.
Linearity was evaluated by constructing the calibration curves by analyzing mixtures of the standard solutions and blood serum samples spiked with the standard solutions. The calibration curves were constructed by plotting the peak area of the analyte (at five concentration levels), to the peak area of the internal standard versus the nominal concentration. Calibration curves with respective correlation coefficients, slopes and intercepts resulted from the linear regression analysis. Limit of Detection (LOD) and Limit of Quantification (LOQ) values, were calculated by Signal (S) to Noise (N) ratio criterion, using the equations:
LOD = 3.3 S/N |
and
LOQ = 10 S/N |
respectively.
The precision of the method was assessed by calculating the Relative Standard Deviation (RSD) at three concentration levels (3, 5 and 10 ng μL-1). Intra-day (within-day) precision was estimated by five replicate measurements for each concentration level while intermediate (between-day) precision was estimated in duplicate measurements of freshly prepared samples during a period of 5 consecutive days.
The accuracy was expressed as the percentage recovery was measured by the comparison of peak area ratios of extracted samples of spiked urine samples, to standard solutions spiked with the same amount of analytes that were not extracted.
The long-term and short term stability of the analytes in human serum was also expressed as percentage recovery and was measured by spiking serum samples with a mixture of the analytes at a concentration of 5 ng μL-1 and analyzing after 1, 2 and 24 h of storage at 20°C. Stability was also assessed after 24 h and three days of storage in a freezer at -4°C. Furthermore, stability was studied after 1, 7 and 14 day of storage in a deep-freezer at -18°C. Finally, stability during four freeze-thaw cycles was investigated, using the 10% degradation criterion. Aliquots of spiked serum samples spiked with the analytes were deep-frozen (at -18°C). After freezing, they were left at room temperature to thaw. One aliquot was measured while the rest of them were refrozen.
RESULTS AND DISCUSSION
Chromatographic conditions: In order to achieve good peak resolutions of the 8 compounds at short time of analysis, gradient elution was chosen. After many trials, the optimal eluent system was chosen which consisted of a mixture of (A) acetic acid (1%), (B) acetonitrile and (C) methanol at initial composition of 90:5:5, at a flow rate of 1 mL min-1. The gradient program of the proposed method is shown in Table 1. The operating backpressure was between 135-155 bar and the total analysis time was approximately 20 min.
Protocatechuic acid (3,4-dihydroxybenzoic acid) was chosen as the internal standard as it was eluted at a short time and was well resolved from the other compounds.
A typical chromatogram obtained under the above conditions is shown in Fig. 2a and b. Retention times were 4.041 min for GA, 6.276 min for PCA (internal standard), 7.178 min for TB, 9.195 min for TF, 12.246 min for C, 13.993 min for ClGn, 15.582 min for CAF, 19.799 min for EGCG and 20.764 min for EC.
Method optimization: In order to prevent polyphenols from oxidation, since they are susceptible in high temperatures and alkaline conditions, the spiked and blank serum samples were mixed with 20 μL of acetate buffer solution (0.2 M CH3COOH, 0.2 M CH3COONa, pH 3.5). Different Solid-Phase Extraction (SPE) protocols were tested but none yielded satisfactory recovery as shown in Table 2. Direct Liquid/Liquid Extraction (LLE) of serum by ethyl acetate was also tested but the recoveries were also low. Since SPE and LLE failed to provide the required rates an alternative sample treatment technique was tested. The mixture was vortexed and centrifuged in order to remove the proteins from the blood in the form of a pellet. After this step, non protein bound compounds were extracted by ethyl acetate at a volume of 750 μL which was added to the supernatant and the mixture was centrifuged again for the extraction of polyphenols from blood serum. This procedure yielded the higher recovery rates for all examined analytes except for EGCG. The results from the optimization study are given in Table 2.
The procedure described above was very simple and rapid with minimum consumption of organic solvents and very clean background signal.
Method validation
Selectivity: The good resolution between the analytes, the absence of interferences in the spiked serum samples and the good background signal indicate that a good selectivity was achieved. Typical chromatograms of blank and spiked serum samples are given in Fig. 2c-f.
Linearity and sensitivity: Linearity was evaluated by the use of calibration curves for each compound. In standard solutions, all the compounds were linear in the range of 1-10 ng μL-1 and the correlation coefficients were between 0.9914 and 0.9974. In spiked serum samples, GA, TB, TF, CAF, ClGn and EC were linear in the range of 1-10 ng μL-1, C was linear in the range of 1-20 ng μL-1 and EGCG was linear in the range of 0.5-20 ng μL-1. The correlation coefficients were between 0.9902 and 0.9979. The LOD was 0.2 ng μL-1 while the LOQ was 0.5 ng μL-1. Linearity and sensitivity data are shown in Table 3.
Accuracy and precision: The accuracy of the method was assessed b y means of recovery percentage. The results were between 83.3 and 104.9%. Precision was evaluated by means of Relative Standard Deviation (RSD). For within-day repeatability RSDs were lower than 15.4% while for between-day repeatability RSDs were lower than 15.2%. Accuracy and precision data are given in Table 4.
Stability: The stability experiments showed that at room temperature (20°C) of storage EC was stable for 1 h, GA and C for 2 h and TB, TF, ClGn, CAF and EGCG for 24 h. At -4°C of storage, GA, C, ClGn and EC were stable for one day while TB, TF, CAF and EGCG were stable for three days.
At -18°C, GA, C and EC were stable for one day, TF was stable for one week and TB, ClGn, CAF and EGCG were stable for two weeks. Finally, EC was stable for 1 freeze-thaw cycle, C and EGCG for 2 cycles, GA, TB and TF for 3 cycles and ClGn and CAF for 4 cycles.
The advantages of the developed method in comparison to the already published include the simplicity and rapidity of sample preparation step which is very important in the routine analysis of a serum samples for example in epidemiological studies. Furthermore, the fact that no sophisticated equipment is necessary, it makes the method applicable to any laboratory.
CONCLUSION
Herein a rapid analytical method has been developed for the simultaneous determination of 5 polyphenols and 3 purine alkaloids in human serum using HPLC-DAD.
The precise determination of these phenolic compounds makes this method a useful tool for bioavailability and pharmacokinetic studies that links the consumption of these polyphenols with the prevention of various diseases. Furthermore, it can be used in epidemiological studies in order to investigate the benefits of tea consumption in the prevention of cardiovascular and cancer diseases due to the anti-oxidant properties of the flavonoids.