Article pubs.acs.org/JAFC
Detection of Maillard Reaction Product [5-(5,6-Dihydro‑4H‑pyridin-3ylidenemethyl)furan-2-yl]methanol (F3-A) in Breads and Demonstration of Bioavailability in Caco‑2 Intestinal Cells Xiu-Min Chen, Yue Dai, and David D. Kitts* Food, Nutrition, and Health, Faculty of Land and Food Systems, University of British Columbia, 2205 East Mall, Vancouver, British Columbia V6T 1Z4, Canada S Supporting Information *
ABSTRACT: [5-(5,6-Dihydro-4H-pyridin-3-ylidenemethyl)furan-2-yl]methanol, also called F3-A, has been isolated from hexose−lysine Maillard reaction (MR) models. Here we report on optimized conditions for the recovery of F3-A and concentrations found in bread. Recovery of F3-A was best achieved when samples were extracted with dichloromethane (DCM) at a solvent to sample ratio of 2:1 (v/v) after adjustment of the pH to 12. The amount of F3-A in whole wheat bread was significantly (P < 0.05) higher than that in white bread; bread crust contained a significantly (P < 0.05) higher amount of F3-A (0.9−7.8 μg/100 g) than the bread crumb (not detectable−3.5 μg/100 g); and toasting increased F3-A concentration with a range of not detectable to 6.0 μg/100 g in the control bread and 4.0 and 17.7 μg/100 g in the dark-toasted white sandwich bread and 100% whole wheat sandwich bread, respectively. The in vitro permeability of F3-A was measured using Caco-2 cell monolayer. The apparent permeability coefficient (Papp) of F3-A is (6.01 ± 0.35) × 10−5 cm/s, which is similar to that of propranolol, a highly passive transcellular absorbed drug. In conclusion, the concentration of F3-A recovered in bread varies with the type of bread and degree of toasting, and F3-A is bioavailable. KEYWORDS: F3-A, Maillard reaction, bread, extraction, permeability
■
INTRODUCTION Maillard reaction products (MRPs) are produced when reducing sugars react with amino acids, peptides, or proteins, particularly in heat-processed food systems. MRPs have both anti- and pro-inflammatory effects.1−4 Formerly, we characterized a MRP, [5-(5,6-dihydro-4H-pyridin-3-ylidenemethyl)furan-2-yl]methanol (F3-A), from a glucose−lysine Maillard reaction (MR) model, which exhibited inhibition of nitric oxide (NO) and inducible nitric oxide synthase (iNOS) in cytokineinduced Caco-2 cells.5,6 F3-A is produced in MR model systems that contain lysine and hexose or saccharides that comprise a hexose monomer, such as glucose, fructose, lactose, or sucrose.5,6 We have proposed the pathway of F3-A production in our previous paper,5 and the reaction scheme is shown in Figure 1. This product was first reported by Miller et al.7,8 in glucose and lysine MR mixture under acidic condition and then by Totsuka et al.9 in the mixture containing hydroxymethylfurfural (HMF) and lysine. However, there is no information on the F3-A content in food products. Bread is a high-temperature, heatprocessed bakery product that represents a major food system in Western and European diets. Wheat flour contains lysine and carbohydrate and undergoes Maillard reaction that could produce F3-A during baking. We hypothesize that F3-A can be detected in bread products once an optimal extraction procedure is developed to maximize recovery of this bioactive molecule. In addition, for F3-A to display bioactive antiinflammatory activity, it should also be bioavailable; however, at present there are no studies that have defined intestinal permeability and, thus, potential bioavailability. © XXXX American Chemical Society
The Caco-2 cell, a human colorectal adenocarcinoma cell line, is widely employed to measure drug and nutrient transport, with it exhibiting morphological and functional similarities to the small intestine when fully differentiated.10−14 The extent of nutrient or drug absorption is correlated with permeability across Caco-2 monolayers.15 This concept was used to determine the relative permeability across Caco-2 to predict F3-A bioavailability. The aim of the present study was to establish operational conditions for maximizing F3-A recovery, required to enable quantification of F3-A in different bread samples and with breads toasted to different degrees. We also measured intestinal permeability using the Caco-2 monolayer. The apparent permeability coefficient obtained for F3-A was compared to two known model drugs, propranolol and furosemide, that have contrasting solubility and permeability characteristics.
■
MATERIALS AND METHODS
Chemicals. [5-(5,6-Dihydro-4H-pyridin-3-ylidenemethyl)furan-2yl]methanol (F3-A) was synthesized by the Chemistry Department, Simon Fraser University, Canada (purity > 95% determined by NMR described by Chen et al.5). Ethyl acetate (EA), dichloromethane (DCM), chloroform (CF), and acetonitrile (ACN) were purchased from Fisher Scientific (Toronto, ON, Canada). Three types of breads, including white sandwich bread (WI), 100% whole wheat (sandwich) bread (WW), and 100% whole wheat home-style bread (WH), all Received: Revised: Accepted: Published: A
September 30, 2016 November 2, 2016 November 6, 2016 November 6, 2016 DOI: 10.1021/acs.jafc.6b04367 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Figure 1. Proposed reaction pathway for F3-A. produced at Dempster’s Bakery, ON, Canada, were purchased from Costco, Richmond, BC, Canada. Minimum essential medium (MEM) with Earle’s salts containing 1.0 g/L glucose, penicillin, streptomycin, lucifer yellow, propranolol, furosemide, dimethyl sulfoxide (DMSO), Hanks’ balanced salts solution (HBSS), NaOH, and HCl were purchased from Sigma-Aldrich (St. Louis, MO, USA). Fetal bovine serum (FBS) was purchased from Invitrogen (Burlington, ON, Canada). Effects of Extraction Conditions on the Recovery of F3-A. A two-factor and two one-factor factorial designs were used to optimize the extraction conditions, which included determining the extraction pH and solvent, the sample-to-solvent ratio (v/v), and time required for maximum extraction efficiency. F3-A standard was dissolved in acidic ddH2O (pH 6) at both a low and a high (1 μg/mL or 20 μg/ mL) concentration, with pH adjustment to alkaline conditions (e.g., 10 or 12, respectively) using 5.0 M NaOH. To determine the best solvent for extraction of F3-A from aqueous solution, the alkaline F3-A solution was extracted three times with EA, DCM, or CF solvents, at a solvent-to-sample ratio (VS:V0) of 0.67:1 (v/v). Solvent extracts were combined and evaporated in a rotary evaporator (Buchi water bath B480, Buchi, Switzerland), and residues were redissolved in acidic ddH2O and measured by high-performance liquid chromatography (HPLC). The percentage recovery of F3-A was calculated. Following the choice of the optimal pH and solvent, additional tests were conducted with different solvent-to-sample ratios (set at 1:1, 2:1, and 3:1) and number of repeated extractions (e.g., range = 1−3 times) to determine maximum recovery of F3-A. Synthesized, purified F3-A was used as the standard for determining the limit of determination (LOD) and limit of quantification (LOQ), which were calculated as 3.3 and 10 times the standard deviation of response/slope, respectively. Quantification of F3-A in Breads. Slices of both WI and WW breads were toasted in a household toaster (Proctor Silex, Hamilton
Beach Brands, Southern Pines, NC, USA) for either 2 or 3 min to yield a light-toasted (LT) or a dark-toasted (DT) bread, respectively. Bread without toasting was used as a control (C) bread. Slices of C, LT, and DT from both WI and WW breads, as well as C slices of WH, were separated into crust (the most outside thin brown layer) and crumb (the remaining part of the bread), followed by blending to fine particles using a coffee grinder (Hamilton Beach, Black and Decker) and kept in a 4 °C cold room until analyzed. Crumb or crust (10 g) was digested in 100 mL of 0.01, 0.1, or 1 N HCl, respectively, on an incubator shaker (Innova40, Newbrunswick Scientific, Enfield, CT, USA) set at 250 rpm and 37 °C for 1 h. Preliminary analysis informed us that detectable F3-A peaks in WI crust could be found only in samples digested with 1 N HCl; therefore, we routinely employed 1 N HCl in the following experiments. Ten grams of bread, sampled in triplicate, and another sample spiked with 20 μL of 5 mg/mL F3-A were digested with 100 mL of 1 N HCl for 1 h, followed by centrifuging at 10000g for 1 h. The supernatant was adjusted to pH 12 using 5 N NaOH, followed by extraction with the optimized extraction conditions and extraction with DCM at a solvent-to-sample ratio of 2:1 (v/v). The DCM fraction was evaporated and the residue redissolved in 1 mL of acidified ddH2O (pH 6). The concentration of F3-A was calculated as
⎛ C × C1 ⎞ C=⎜ 0 ⎟ ⎝ C2 − C1 ⎠
(1)
where C is the concentration of F3-A in bread, C0 is the known concentration (1 μg/mL) of F3-A added to the bread, C1 is the measured F3-A concentration in bread samples without spike, and C2 is the measured F3-A concentration in bread samples spiked with 1 μg/mL of F3-A. Lysine content in bread crust and crumb was analyzed at the University of Manitoba, Canada, using AOAC Official Method B
DOI: 10.1021/acs.jafc.6b04367 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
acidic conditions and hydrophobic under basic conditions;5 therefore, herein, we adjusted test samples to pH 10 and 12, respectively, to establish maximal recovery. At low concentration, the recoveries of F3-A ranged from only 4.1 to 57.4%, with DCM extracting F3-A most efficiently compared with CF and EA, respectively, when the solvent-to-sample ratio was set at 0.67:1 (Figure 2). The recoveries of F3-A at high
994.12 with an amino acid analyzer (model S2100, S4300, Sykam, Germany). Water activity was measured using AquaLab (Decagon Devices, Pullman, WA, USA). Briefly, 0.5 g of ground sample was placed in a plastic dish followed by equilibration in the chamber. After equilibrium, water activity was recorded using the AquaLab water activity meter. Permeability of F3-A across Caco-2 Monolayer. The permeability of F3-A was measured according to the method described by Chen et al.16 using Caco-2 cells (HTB-37, ATCC), cultured in MEM and supplemented with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C under a 5% CO2 atmosphere. Cells were subcultured weekly, and the medium was changed every 2−3 days. Caco-2 cells used in this experiment were from passages 22−28 and seeded on 24-well translucent transwell inserts (0.4 μm, highdensity polyethylene terephthalate membrane, BD Biosciences, San Jose, CA, USA) at a density of 2.5 × 105 cells/cm2 and grown for 3−4 weeks. Transepithelial electrical resistance (TEER) values were measured using a volt-ohmmeter (Millicell ERS, Millipore, Bedford, MA, USA). The integrity of the monolayers was checked before and after the permeability study by measuring the TEER value and also lucifer yellow. The amount of lucifer yellow transported to the basolateral side within 2 h was quantified using a fluorospectrometer (Fluoroskan Ascent FL, Thermo Labsystems, Helsinki, Finland) at an excitation wavelength of 425 nm and an emission wavelength of 530 nm. F3-A along with test model drugs, namely, propranolol and furosemide, were each dissolved in DMSO and diluted further in HBSS to a final concentration of 100 μM with 0.1% DMSO. Aliquots (300 μL) of F3-A, propranolol, and furosemide were placed on the apical side along with 700 μL of HBSS applied to the basolateral side. The inserts were moved to wells with fresh HBSS at 15, 30, 45, 60, 90, and 120 min. Concentrations of basolateral solutions were measured. The apparent permeability coefficients (Papp, cm/s) were calculated according to eq 2
Papp =
⎛ dQ ⎞ V ⎜ ⎟ × ⎝ dt ⎠ A × C 0
Figure 2. Recovery of F3-A with different pH and extraction solvents: (A) 1 μg/mL; (B) 20 μg/mL. Data expressed as the mean ± SD of three replicates. Different letters (a−c) above bars denote significant difference (P < 0.05) in different extraction solvents at the same pH. An asterisk denotes significant difference (P < 0.05) between pH 10 and pH 12 with the same extraction solvent. EA, ethyl acetate; CF, chloroform; DCM, dichloromethane.
(2)
where dQ/dt is the permeability rate (μM/s) calculated from the slope of the amount of tested compounds transported versus time, V is the volume of the basolateral chamber (mL), A is the surface area of the inset (cm2), and C0 is the initial concentration (μM) of the drugs and F3-A. Quantification of F3-A, Propranolol, and Furosemide by HPLC. F3-A, propranolol, and furosemide were determined using HPLC (Agilent 1100 Series) with an Agilent Zorbax SB-C18 column (4.6 × 150 mm, 3.5 μm) based on methods described in Chen and Kitts.5,16 The column temperature was set at 30 °C, and the flow rate was 1 mL/min. F3-A was separated with a linear gradient of solvent A, 5−20% of ACN, and solvent B, 0.2% formic acid, for 6 min and detected with a diode array detector (DAD) at 370 nm. The mobile phases for propranolol and furosemide comprised 55% ammonium acetate (50 mM, pH 3.5) and 45% ACN under isocratic conditions, and they are detected with a fluorescent detector at excitation at 283 and 235 nm and emission at 338 and 410 nm, respectively. Statistical Analysis. Data are presented as the mean ± standard deviation (SD) of triplicate experiments, and both one-way and twoway analysis of variance analyses were performed using the general linear model of MINITAB software (version 14, Minitab Inc., State College, PA, USA) to analyze data. Significant differences among treatments were compared using Tukey’s test (P < 0.05).
concentration were increased except when using EA at pH 10, and the extraction efficiency of DCM was similar to that of CF, but significantly (P < 0.05) greater than that of EA (Figure 2). According to the solvent miscibility table,17 the polarity indices for DCM, CF, and EA are 3.1, 4.1, and 4.4, respectively; thus, DCM and CF are more hydrophobic than EA. Therefore, our results indicated that F3-A was extracted more efficiently in the more hydrophobic solvent. The low recovery of F3-A at low concentration, approaching the LOQ, is partially due to experimental error. Increased pH produced a significantly (P < 0.05) greater recovery of F3-A at both low and high concentrations. F3-A contains a nitrogen, which can be protonated under acidic condition, whereas under basic condition, F3-A is deprotonated and more hydrophobic than when in the acidic condition. At pH 12, more F3-A exists as a deprotonated form, which explains the higher relative recovery compared to pH 10. DCM was therefore determined to be the best extraction solvent used due to the similar extraction efficacy as CF when the extraction pH was set at 12 and less toxic. The optimal solvent-to-sample ratio was evaluated to further increase the yield after selection of the pH and solvent type. The recoveries of F3-A were increased with the increase of solvent-to-sample ratio and reached maximum values when the ratio reached 2:1, whereas further increase of VS:V0 to 3:1 caused a decrease in recovery (Table 1). The recovery of high concentrations of F3-A was also significantly higher (P < 0.05) compared to extractions made of low concentration (107.4 vs
■
RESULTS AND DISCUSSION Effects of Extraction Conditions on the Recovery of F3-A. The LOD and LOQ of F3-A were determined to be 0.21 and 0.64 μg/mL, respectively, on the basis of a standard curve presented in the Supporting Information. We utilized both low (1 μg/mL) and high (20 μg/mL) concentrations to determine the extraction efficiency of F3-A. F3-A is highly hydrophilic at C
DOI: 10.1021/acs.jafc.6b04367 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry Table 1. Recoveries (%) of F3-A with Different Solvent-toSample Ratiosa
Table 3. F3-A Concentrations (μg/100 g) in Crust and Crumb of Three Different Types of Breads (WI, WW, and WH) with Three Different Toasting Levels (C, LT, and HT)a
Vs:V0 F3-A concentration (μg/mL)
0.67:1
1:1
2:1
3:1
1 20
57 ± 8bc 85 ± 3b*
62 ± 2ab 84 ± 11b*
68 ± 4a 107 ± 3a*
48 ± 6c 81 ± 6b*
toasting level bread type WI crust WI crumb WW crust WW crumb WH crust WH crumb
a
Samples with different concentrations were adjusted to pH 12 and extracted with DCM three times at different solvent-to-sample ratios. Data are expressed as the mean ± SD of three replicates. Different letters (a−c) denote significant difference (P < 0.05) in the same row. An asterisk denotes significant difference (P < 0.05) in the same column.
C 0.9 NDaα 6.0 1.3 7.8 3.5
LT
± 0.4a*α ± ± ± ±
0.8a*γ 0.7aα 0.6*δ 0.8β
0.8 ± 0.4a*α NDaα 13.5 ± 0.3b*γ 4.0 ± 0.4bβ N/A N/A
DT 4.0 6.6 17.7 12.1 N/A N/A
± ± ± ±
0.5b*α 0.7bβ 0.6c*δ 0.9cγ
68.4%). To investigate the cost efficiency of repeated extractions, the recoveries of three extractions at a solvent-tosample ratio of 2:1 were analyzed. The percentage recovery was calculated as the percentage of F3-A extracted in each extraction of the total amount of F3-A extracted in all three extractions. The percentage recovery of the first extraction was significantly higher (P < 0.05) than the percentage recoveries of the second and third extractions, and the initial concentration of F3-A had no significant effect on the percentage recoveries of all three extractions (Table 2). The first extraction recovered
Samples were digested in 1 N HCl for 1 h at 37 °C, followed by adjustment of the pH to 12 and extracted with DCM once at a solvent to sample ratio of 2:1. Data are expressed as the ean ± SD of three replicates. WI, white sandwich bread; WW, 100% whole wheat sandwich bread; WH, 100% whole wheat home-style bread; C, breads without toasting; LT, light-toasted bread (toasted for 2 min); DT, dark-toasted bread (toasted for 3 min); N/A, not determined; ND, not detectable. Different letters (a−c) denote significant difference (P < 0.05) in the same row. Different letters (α−δ) denote significant difference (P < 0.05) of different crust or crumb samples with the same toasting level. An asterisk denotes significant difference (P < 0.05) between crust and crumb of the same bread sample at the same toasting level.
Table 2. Recoveries (%) of Each Extraction at Different Concentrationsa
Table 4. Lysine Content (%) and Water Activity of Control Bread Samples
a
extraction number
bread type
lysine (%)
water activity
WI crust WI crumb WW crust WW crumb WH crust WH crumb
0.133 0.238 0.170 0.309 0.124 0.273
0.901 0.935 0.933 0.944 0.905 0.930
F3-A concentration (μg/mL)
1
2
3
1 20
99 ± 1a 96 ± 3a
1 ± 1b 4 ± 3b
0 ± 0b 0 ± 0b
a
Samples with different concentrations were adjusted to pH 12 and extracted with DCM three times at a solvent-to-sample ratio of 2:1. Data are expressed as the mean ± SD of three replicates. Different letters (a, b) denote significant difference (P < 0.05) in the same row.
pyrraline, to be higher in bread crust than in the crumb.18 Home-style 100% whole wheat bread had higher F3-A content compared with whole wheat sandwich bread and white sandwich bread, having the relatively lowest amount of F3-A. These results indicated that the composition of the bread dough is another important factor. We attribute this finding to the relatively higher lysine content in whole wheat flour used to bake this bread. On the basis of the U.S. Department of Agriculture (USDA) National Nutrient Database,19 enriched white bread wheat flour (e.g., 0.23 g/100 g flour) contains less lysine than whole grain wheat flour (e.g., 0.36 g/100 g flour). Lysine is an important reactant of the MR that includes the production of F3-A.5 Lysine content in bread crust was shown to be consistently lower than that of crumb in all breads tested (Table 4) and corresponded to a lower water activity and greater F3-A recovered in the crust. WH bread also had lower lysine compared to WW bread, which again supports the finding that a higher F3-A content was recovered in WH bread. The contents of MRPs in foods vary depending on the types of food and MRPs and the processing methods. For example, cereal and fruit contain 2.6 and 0.13 mg/100 g of CML, respectively.20 3-Deoxyglucosone content in cookies is around 38.5 mg/100 g food.21 HMF, a precursor of F3-A, is present in breakfast cereals (ranging from 3.5 to 13 mg/100 g)22 and in toasted bread (18 mg/100 g).23 Other flavor and aroma MRPs, such as 2-ethyl-3,5-dimethylpyrazine, 2,3-diethyl-5-methylpyr-
>96% of F3-A, which was satisfactory to conclude that one solvent extraction produced a good recovery while reducing the need for excessive solvent use. Quantification of F3-A in Bread. F3-A in bread sample was confirmed using LC-MS as described by Chen et al.5,6 after comparison with pure standard (Supporting Information). F3A in three different types of breads (WI, WW, and WH) and toasted to relatively light and dark extents was quantified using the optimized extraction method described above. In general, bread crust contained significantly (P < 0.05) higher amounts of F3-A (Table 3) compared to bread crumb. We attribute this result to a number of plausible reasons. First, the exterior, or crust, of the bread would receive higher temperature exposure during the baking time, hence enabling more F3-A to be produced as a result of greater extent of MR. The water activity of bread crust was also lower compared to crumb (Table 4), and toasting further decreased water activity (e.g., WI crust, C vs DT, 0.901 vs 0.757). During baking, the water activity on the surface of the dough would be reduced, thus providing conditions favorable for MR, whereas the water activity is relatively higher in the dough. The moisture content of crust being lower than that of crumb would also partially explain the higher F3-A in the crust. This finding agrees with observations made by other workers that reported concentrations of two known MRPs, namely, carboxymethyllysine (CML) and D
DOI: 10.1021/acs.jafc.6b04367 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Propranolol is noted as a high permeability−high solubility lipophilic drug that is mainly absorbed by the passive transcellular route with >90% bioavailability after oral consumption. Furosemide, on the other hand, is a low permeability−high solubility hydrophilic compound with an oral bioavailability ranging from 10 to 60%.15,26 On the basis of these two extremes for estimating bioavailability, we found that F3-A corresponded very much like propranolol, having a Papp (6.01 × 10−5 cm/s) that was similar to that of propranolol (6.24 × 10−5 cm/s) and considerably higher than that of furosemide Papp (0.22 × 10−5 cm/s). The percentage transport of these three compounds from apical to basolateral during a 2 h incubation period corresponded to the differences noted in Papp estimates, with both F3-A (42.8%) and propranolol (41.9%) being notably greater than furosemide (2.4%). Because compounds with Papp > 1.0 × 10−5 cm/s are completely absorbed in humans (>90%),15 our results would indicate that F3-A should be relatively bioavailable following oral administration. The physicochemical parameters of F3-A also indicated that its lipophilicity (log P = 0.294 ± 0.449) and polar surface area (45.7 Å2) could make it readily bioaccessible to most organ systems, such as the blood−brain barrier.27 This information is important for quantifying the MR bioavailability and excretion information that is needed to determine the potential exposure from oral intake, its physiological effects, and health implications of consuming low or high MRP-containing diets. In this study, we report the detection of F3-A in food samples for the first time. The extraction method that we optimized to recover F3-A worked well with both nontoasted and toasted breads. We will need to utilize this method to determine if F3-A can be detected in other MR-rich food systems. We also demonstrated that F3-A content in bread can be affected by the type of flour that is used to make the bread, the breadmaking operation, the parts of the bread (crumb and crust), and, finally, the extent of toasting. Of the breads tested in this study, toasted bread contained a relatively greater F3-A content than nontoasted bread. Finally, the observation that F3-A has a permeability similar to that of propranolol is further evidence to conclude that F3-A is permeable and that its high solubility character makes it bioavailable for potential antiinflammatory activity.
azine, and 2-acetyl-1-pyrroline, are present in a wide range of 0.1−1000 μg/100 g in different foods.24 Our results indicate that F3-A content is low in bread relative to its precursor, HMF, but with the introduction of heat, as is the case with toasting, can increase 2−10-fold. Table 3 shows that toasting significantly (P < 0.05) increased F3-A content in both WW bread crust and crumb. In WI bread, LT produced little change in F3-A content, whereas DT toasting significantly (P < 0.05) increased F3-A content. We previously showed that longer heating time will increase F3-A production in a glucose−lysine aqueous MR model,5 which corresponds to the observation herein that a greater amount of F3-A was found in DT. Another interesting observation was that the amount of F3-A in bread crumb increased to a greater extent compared to that recovered in the crust in both WW and WI breads after greater toasting. This may be explained by the limiting amount of lysine in the crust or, in part, by the influence of the positioning of the bread to the toasting wire. It is worth mentioning that the heating wires in the home-style toaster were positioned on two sides with the bread in the middle, thus enabling the crumb to receive greater exposure to heat than crust. F3-A Permeability Studies. Bioavailability of drugs or nutrients following oral intake can be predicted from permeability and solubility measurements that are performed using Caco-2 monolayer, the reason being that maximum absorption of target compounds is associated with maximum permeability and concentration at the site of absorption.15 A classification system has been used for drugs to predict bioavailability, and this system has four distinct categories, which are described as being high permeability−high solubility, high permeability−low solubility, low permeability−high solubility, and low permeability−low solubility.25 In general, compounds that have high permeability−high solubility and high permeability−low solubility are completely absorbed, whereas low permeability−high solubility drugs have lower absorption rates of 40−80%; the low permeability−low solubility drugs have comparatively very poor and incomplete absorption.15 The apparent permeability coefficient (Papp) of F3-A was determined in this study using the intestinal Caco-2 cell model, and values were compared with propranolol and furosemide, respectively. TEER values (Table 5) did not change before or after the permeability measurement. Less than 10% of lucifer yellow was recovered from the basolateral side, thus indicating that the monolayers retained good integrity during the experiment.
■
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jafc.6b04367. Standard curve of F3-A; LC-MS result of F3-A in bread sample and of standard F3-A; HPLC chromatogram of whole wheat bread crust extract (PDF)
Table 5. Apparent Permeability Coefficients, Percentage Transport, and TEER Valuesa
■
TEER (Ω·cm2) sample
before
after
F3-A propanolol furosemide
215 ± 8a 221 ± 15a 203 ± 8a
225 ± 5a 225 ± 18a 210 ± 10a
Papp −5
% transport
6.0 ± 0.4b 6.2 ± 0.6b 0.2 ± 0.0a
42.8 ± 1.7b 41.9 ± 4.4b 2.4 ± 0.5a
(×10
cm/s)
ASSOCIATED CONTENT
AUTHOR INFORMATION
Corresponding Author
*(D.D.K.) Phone: (604) 822-5560. Fax: (604) 822-5143. Email:
[email protected]. Funding
This study was supported by a NSERC-discovery grant to D.D.K.
Data are expressed as the mean ± SD of three replicates. Different letters (a, b) denote significant difference (P < 0.05) in the same column. There are no significant differences in TEER values before and after permeability measurement.
a
Notes
The authors declare no competing financial interest. E
DOI: 10.1021/acs.jafc.6b04367 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
■
(11) Artursson, P.; Palm, K.; Luthman, K. Caco-2 monolayers in experimental and theoretical predictions of drug transport. Adv. Drug Delivery Rev. 2012, 64, 280−289. (12) Pinto, M.; Robineloeon, S.; Appay, M. D.; Kedinger, M.; Triadou, N.; Dussaulx, E.; Lacroix, B.; Simonassmann, P.; Haffen, K.; Fogh, J.; Robineleon, S.; Robine-Leon, S.; Simon-Assmann, P.; Robine-Léon, S.; Robin-Leon, S.; Simon-Assman, P.; Robine Leon, S.; Zwwibaum, A.; Zweibaum, A.; Simon-Assmami, P. Enterocyte-like differentiation and polarization of the human colon carcinoma cell line Caco-2 in culture. Biol. Cell. 1983, 47, 323−330. (13) Hilgers, A. R.; Conradi, R. A.; Burton, P. S. Caco-2 cell monolayers as a model for drug transport across the intestinal mucosa. Pharm. Res. 1990, 7, 902−910. (14) Sambuy, Y.; De Angelis, I.; Ranaldi, G.; Scarino, M.; Stammati, A.; Zucco, F. The Caco-2 cell line as a model of the intestinal barrier: influence of cell and culture-related factors on Caco-2 cell functional characteristics. Cell Biol. Toxicol. 2005, 21, 1−26. (15) Pade, V.; Stavchansky, S. Link between drug absorption solubility and permeability measurements in Caco-2 cells. J. Pharm. Sci. 1998, 87, 1604−1607. (16) Chen, X.; Elisia, I.; Kitts, D. D. Defining conditions for the coculture of Caco-2 and HT29-MTX cells using Taguchi design. J. Pharmacol. Toxicol. Methods 2010, 61, 334−342. (17) Solvent miscibility table; available at https://www.erowid.org/ archive/rhodium/pdf/solvent.miscibility.pdf (accessed Oct 28, 2016). (18) Hellwig, M.; Henle, T. Baking, ageing, diabetes: a short history of the Maillard reaction. Angew. Chem., Int. Ed. 2014, 53, 10316− 10329. (19) USDA National Nutrient Database; available at https://ndb.nal. usda.gov/ndb/nutrients/report?nutrient1=505&nutrient2= &nutrient3=&fg=20&max=25&subset=0&offset=125&sort= f&totCount=145&measureby=g (accessed July 20, 2016). (20) Hull, G. L.; Woodside, J. V.; Ames, J. M.; Cuskelly, G. J. Nε(Carboxymethyl) lysine content of foods commonly consumed in a Western style diet. Food Chem. 2012, 131, 170−174. (21) Degen, J.; Hellwig, M.; Henle, T. 1,2-Dicarbonyl compounds in commonly consumed foods. J. Agric. Food Chem. 2012, 60, 7071− 7079. (22) Delgado-Andrade, C.; Seiquer, I.; Navarro, M. P.; Morales, F. J. Estimation of hydroxymethylfurfural availability in breakfast cereals. Studies in Caco-2 cells. Food Chem. Toxicol. 2008, 46, 1600−1607. (23) Ameur, L. A.; Trystram, G.; Birlouez-Aragon, I. Accumulation of 5-hydroxymethyl-2-furfural in cookies during the backing process: validation of an extraction method. Food Chem. 2006, 98, 790−796. (24) Bastos, D. M.; Monaro, É; Siguemoto, É; Séfora, M. Maillard Reaction Products in Processed Food: Pros and Cons; INTECH Open Access Publisher: 2012; pp 281−300. (25) Amidon, G. L.; Lennernäs, H.; Shah, V. P.; Crison, J. R. A theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. Pharm. Res. 1995, 12, 413−420. (26) Walter, E.; Janich, S.; Roessler, B. J.; Hilfinger, J. M.; Amidon, G. L. HT29-MTX/Caco-2 cocultures as an in vitro model for the intestinal epithelium: In vitro−in vivo correlation with permeability data from rats and humans. J. Pharm. Sci. 1996, 85, 1070−1076. (27) SciFinder; available at https://scifinder.cas.org/scifinder/view/ scifinder/scifinderExplore.jsf (accessed July 20, 2016).
ACKNOWLEDGMENTS We thank Kaiwen Mu and Dr. Zhili Ma for help with sample extractions.
■
ABBREVIATIONS USED C, control; CF, chloroform; CML, carboxymethyllysine; DAD, diode array detector; DCM, dichloromethane; DMSO, dimethyl sulfoxide; DT, dark toast; EA, ethyl acetate; F3-A, [5-(5,6-dihydro-4H-pyridin-3-ylidenemethyl)furan-2-yl]methanol; HBSS, Hanks’ balanced salts solution; HMF, hydroxymethylfurfural; HPLC, high-performance liquid chromatography; iNOS, inducible nitric oxide synthase; LOD, limit of determination; LOQ, limit of quantification; LT, light toast; MEM, minimum essential medium; MR, Maillard reaction; MRPs, Maillard reaction products; NO, nitric oxide; SD, standard deviation; TEER, transepithelial electrical resistance; USDA, U.S. Department of Agriculture; VS:V0, solvent to sample ratio; WH, 100% whole wheat home-style bread; WI, sandwich bread; WW, 100% whole wheat (sandwich) bread
■
REFERENCES
(1) Chung, J. H.; Choi, S. Y.; Kim, J. Y.; Kim, D. H.; Lee, J. W.; Choi, J. S.; Chung, H. Y. 3-Methyl-1,2-cyclopentanedione down-regulates age-related NF-κB signaling cascade. J. Agric. Food Chem. 2007, 55, 6787−6792. (2) Oh, J.; Hong, C.; Nam, M.; Seol, H. M.; Lee, K. Antiinflammatory effects of lysine-galactose Maillard reaction products in Caco-2 and RAW264.7 co-culture systems. FASEB J. 2013, 27, lb216. (3) Chang, P.; Chen, T.; Chang, C.; Hou, C.; Chan, P.; Lee, H. Advanced glycosylation end products induce inducible nitric oxide synthase (iNOS) expression via a p38 MAPK-dependent pathway. Kidney Int. 2004, 65, 1664−1675. (4) Liu, J.; Zhao, S.; Tang, J.; Li, Z.; Zhong, T.; Liu, Y.; Chen, D.; Zhao, M.; Li, Y.; Gong, X. Advanced glycation end products and lipopolysaccharide synergistically stimulate proinflammatory cytokine/ chemokine production in endothelial cells via activation of both mitogen-activated protein kinases and nuclear factor-κB. FEBS J. 2009, 276, 4598−4606. (5) Chen, X.; Chen, G.; Chen, H.; Zhang, Y.; Kitts, D. D. Elucidation of the chemical structure and determination of the production conditions for a bioactive Maillard reaction product,[5-(5,6-dihydro4H-pyridin-3-ylidenemethyl)furan-2-yl]methanol, isolated from a glucose−lysine heated mixture. J. Agric. Food Chem. 2015, 63, 1739− 1746. (6) Chen, X.; Kitts, D. D. Evidence for inhibition of nitric oxide and inducible nitric oxide synthase in Caco-2 and RAW 264.7 cells by a Maillard reaction product [5-(5,6-dihydro-4H-pyridin-3ylidenemethyl)furan-2-yl]methanol. Mol. Cell. Biochem. 2015, 406 (1−2), 205−215. (7) Miller, R.; Olsson, K.; Pernemalm, P. Formation of aromatic compounds from carbohydrates. IX. Reaction of D-glucose and L-lysine in slightly acidic, aqueous solution. Acta Chem. Scand. 1984, 38B, 689− 694. (8) Miller, R. Synthesis and stereochemistry of (E)-5-(3,4,5,6tetrahydropyrid-3-ylidenemethyl)-2-furanmethanol, a product of the reaction between D-glucose and L-lysine. Acta Chem. Scand. 1987, 41B, 208−209. (9) Totsuka, H.; Tokuzen, K.; Ono, H.; Murata, M. A novel yellow compound and furpipate derivatives formed from furfural or 5hydroxymethylfurfural in the presence of lysine. Food Sci. Technol. Res. 2009, 15, 45−50. (10) Artursson, P.; Karlsson, J. Correlation between oral drug absorption in humans and apparent drug permeability coefficients in human intestinal epithelial (Caco-2) cells. Biochem. Biophys. Res. Commun. 1991, 175, 880−885. F
DOI: 10.1021/acs.jafc.6b04367 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
