Chlorogenic Acid Compounds from Coffee Are Differentially Absorbed and Metabolized in Humans 1,2

The Journal of Nutrition Biochemical, Molecular, and Genetic Mechanisms Chlorogenic Acid Compounds from Coffee Are Differentially Absorbed and Metabolized in Humans 1,2 Mariana Monteiro, Adriana Farah,* Daniel Perrone, Luiz C. Trugo, 3 and Carmen Donangelo Laboratório de Bioquímica Nutricional e de Alimentos, Departamento de Bioquímica, Instituto de Química, Universidade Federal do Rio de Janeiro, Ilha do Fundão, RJ 21944, Brazil Abstract Chlorogenic s (CGA) are abundant phenolic compounds in coffee, with caffeoylquinic (CQA), feruloylquinic (FQA), and dicaffeoylquinic (dicqa) s being the major subclasses. Despite the potential biopharmacological properties attributed to these compounds, little is known about their bioavailability in humans. In this study, we evaluated the distribution profile of the major CGA isomers and metabolites in plasma and urine of 6 healthy adults for 4 h after brewed coffee consumption. Three CQA isomers and 3 dicqa isomers were identified in the plasma of all subjects after coffee consumption, whereas 2 FQA were identified in only 1 subject. Two plasma concentration peaks were observed, the first at 0.5 1.0 h and the second at 1.5 4.0 h after coffee consumption. The molar ratio CQA:diCQA was 12.2 in the brewed coffee, whereas in plasma it ranged from 0.6 2.9. The molar ratios 5-CQA:3-CQA and 5-CQA:4-CQA were consistently higher in plasma than in the brew. The main CGA metabolites identified in urine after coffee consumption were: dihydrocaffeic, gallic, isoferulic, ferulic, vanillic, caffeic, 5-CQA, sinapic, r-hydroxybenzoic, and r-coumaric s (gallic and dihydrocaffeic s being the major ones). This study indicates that the major CGA compounds present in coffee are differentially absorbed and/or metabolized in humans, with a large inter-individual variation. Moreover, urine does not appear to be a major excretion pathway of intact CGA compounds in humans. J. Nutr. 137: 2196 2201, 2007. Introduction Coffee is the most consumed food product in the world. In the last few years, a series of epidemiological studies have associated coffee consumption with health benefits, such as reduction of the relative risk of diabetes type 2 and Parkinson and Alzheimer diseases (1 3). In vitro and in vivo studies have attributed these beneficial properties of coffee mainly to the antioxidant capacity of the chlorogenic s (CGA) 4 (4 6), which are usually responsible for 2 5 g/100 g of roasted coffee composition (7,8). CGA are formed by the esterification of hydroxycinnamic s, such as caffeic, ferulic, and r-coumaric, with quinic. Quantitatively speaking, the main CGA subclasses in coffee are caffeoylquinic s (CQA), dicaffeoylquinic s (dicqa), and feruloylquinic s (FQA) with at least 3 isomers per group (9). Among these compounds, 5-CQA alone accounts for ;35% of total CGA in roasted coffee, with all CQA and dicqa isomers being together responsible for 92 95% of CGA (8,10). 1 Author disclosures: M. Monteiro, A. Farah, D. Perrone, L. C. Trugo, C. Donangelo, no conflicts of interest. 2 Supported by Consórcio Brasileiro de Pesquisa e Desenvolvimento do Café (CBP&D Café) - EMBRAPA, CNPq and FAPERJ (Brazil). 3 In memoriam. 4 Abbreviations used: AUC, area under the curve; C max, maximum plasma concentration; CGA, chlorogenic ; CQA, caffeoylquinic; dicqa, dicaffeoylquinic; FQA, feruloylquinic; T max, time corresponding to C max. * To whom correspondence should be addressed. E-mail: afarah@iq.ufrj.br. Despite the potential biopharmacological properties attributed to different CGA compounds, studies on their bioavailability in humans are scarce, mainly due to analytical limitations. Caffeic, a primary metabolite of CQA, and dicqa have been detected in rat and human plasma and urine, especially in conjugated forms (11 16). The presence of free and conjugated forms of 5-CQA in rat plasma has been reported after 5-CQA intraperitoneal administration (11). Lafay et al. (17) identified the presence of 5-CQA and caffeic in rat plasma 1.5 h after the consumption of a diet supplemented with 5-CQA. Recently, Farah et al. (18) identified 3-CQA, 4-CQA, 5-CQA, and other cinnamates in human digestive fluids after 12 h fasting, indicating their circulation in the blood stream. Until present, no CQA, FQA, or dicqa isomers have been identified in human plasma. The objective of this study was to determine the distribution profile of the main CGA compounds and metabolites in human plasma and urine after acute coffee consumption. Subjects and Methods Subjects. Six nonsmoking subjects, 22 55 y of age, 2 male and 4 female, were recruited. They were healthy as judged by a medical questionnaire, with normal blood values for hemoglobin and hematocrit and were not taking any medication or nutritional supplements. The study protocol was approved by the Comissão de Ética para Análises de Projeto de 2196 0022-3166/07 $8.00 ª 2007 American Society for Nutrition. Manuscript received 17 May 2007. Initial review completed 9 June 2007. Revision accepted 11 July 2007.

Pesquisa- CAPPesq/ Faculdade de Medicina/ Universidade de São Paulo Ethical Committee and fully explained to the subjects, who gave their written informed consent prior to participation. Coffee brew preparation. Beans of decaffeinated Coffea canephora cv. Conillon were roasted to light medium roasting degree for preparation of coffee brew. Coffee brew was prepared by adding 250 ml of freshly boiled water (90 95 C) into 40 g of coffee powder placed in a paper filter. An aliquot of the coffee brew offered to each subject was separated for subsequent analysis of the CGA content. Study design and sample collection. The subjects were instructed to avoid consumption of phenolic-containing foods during the 24 h prior to the study. They were asked to eat only animal foods, refined cereal foods, and artificial beverages. On the day of the study, after 8 10 h overnight fasting, an i.v. catheter was inserted into the antecubital vein and a baseline heparinized blood sample was obtained. A standard amount (190 ml) of brewed coffee was offered to each subject and sequential blood draws were obtained at 0.5, 1, 1.5, 2, 3, and 4 h after coffee consumption. Blood samples were collected into heparin-containing tubes. Plasma samples were obtained by centrifugation of the blood samples immediately after being drawn. Urine samples were also collected at baseline and at intervals of 0 2 h and 2 4 h after coffee consumption into appropriate plastic containers. Plasma and urine aliquots for determination of CGAwere ified with HCl and kept frozen in liquid nitrogen until analysis. Urine aliquots for determination of creatinine were ified with HCl and kept at 220 C until analysis. Two hours after coffee consumption, subjects ate a CGAfree snack composed of white bread (50 g) and cream cheese (40 g). Clarification of brewed coffee for chromatographic analysis. The brewed coffee was clarified using an adaptation of the method described by Trugo and Macrae (20). Extraction of chlorogenic s from plasma for chromatographic analysis. Each plasma sample was deproteinized with addition of ethanol and vortexed. The sample was then centrifuged at 17,500 3 g; 5 min at 4 C. The protein pellet was resuspended twice in ethanol and recentrifuged in the same conditions. The pooled ethanol phases were dried under nitrogen flow and the residue obtained was resuspended in sodium acetate buffer, vortexed, and incubated at 37 C for 2 h in the presence of 4000 units of b-glucuronidase and 116 units of sulfatase from Helix Pomatia (Sigma-Aldrich). This enzyme treatment was used to quantify the sum of free and conjugated forms of CGA and metabolites. At the end of incubation, the sample was ified with HCl and methanol aqueous solution (40%, v:v) was added, as in Farah (21). The mixture was centrifuged at 12,000 3 g; 10 min at 4 C and the supernatant was analyzed by both HPLC and liquid chromatography MS as described below. The extraction reproducibility was tested using aliquots of the same plasma sample collected 1.0 h after coffee consumption and stored in liquid nitrogen. Analyses were carried out by HPLC (see below) on 2 consecutive days, with triplicate analyses carried out each day. The CV for all CGA compounds and caffeic analysis in plasma was #5.5%. The extraction recovery was tested in triplicate by adding known amounts of 5-CQA to plasma aliquots not containing CGA. We recovered 88% 5-CQA, 4% 4-CQA, and 7% caffeic, totaling in moles 99% of the initial amount of 5-CQA. Based on these results, analyses of plasma samples were conducted in duplicates. Extraction of chlorogenic s from urine for chromatographic analysis. The ph of each urine aliquot was adjusted to 5.0 with sodium acetate, 4000 units of b-glucuronidase and 116 units of sulfatase from Helix Pomatia were added, and the mixture was incubated in a water bath at 37 C for 2 h. At the end of incubation, the sample was ified with HCl and methanol aqueous solution (40%, v:v) was added with subsequent vortexing. The content was transferred to an Eppendorf tube with a cellulose filter (Microcon YM-10, Milipore) and centrifuged at 17,000 3 g; 30 min at 4 C, as in Farah (21). The supernatant was analyzed by both HPLC and LCMS as described below. Standards. Chlorogenic (5-CQA), caffeic, r-coumaric, ferulic, isoferulic, gallic, dihydrocaffeic, vanillic, syringic, 3,4-dihydroxybenzoic, r-hydroxybenzoic, sinapic, n-metil hippuric, o-metil hippuric, and r-metil hippuric were purchased from Sigma-Aldrich. Hippuric was purchased from Merck. A mixture of 3-CQA, 4-CQA, and 5-CQA was prepared from 5-CQA using the isomerization method of Trugo and Macrae (19,20), also described in Farah et al. (19). For dicqa, a mixture of 3,4-diCQA, 3,5-diCQA, and 4,5 dicqa (Roth) was used. 4-FQA and 5-FQA were identified according to Farah et al. (19,22). HPLC analysis. Brewed coffee, plasma, and urine extracts were analyzed by a HPLC gradient system as described in detail by Farah et al. (19) using a high precision pump (LC-10AD, Shimadzu) and a UV detector (model SPD-10AVp) operating at 325 nm for detection of chlorogenic and other hydroxycinnamic s and at 280 nm for gallic, 3,4- dihydroxybenzoic, dihydrocaffeic, r-hydroxybenzoic, syringic, sinapic, and hippuric s. Fifty microliters of plasma and urine extracts and 20 ml of clarified coffee samples were injected. For the urine analyses, small changes in the gradient were introduced for a better separation of phenolic compounds. Chromatographic data were recorded and integrated in the Class Vp computer software (Shimadzu). The identification of CQA, FQA, and dicqa was primarily performed by comparison of the retention time of the investigated peaks with those of the respective standards (see above). Samples were also spiked with small amounts of the available standards. We confirmed their identity by LCMS analyses as described by Farah et al. (22). Due to the low purity of some of the standards, the quantification of all CGA was performed using the area of the 5-CQA standard combined with the molar extinction coefficients of the respective CGA as previously described (19,20). The quantification of all phenolic s (caffeic, ferulic, isoferulic, r-coumaric, gallic, dihydrocaffeic, vanillic, syringic, 3,4-dihydroxybenzoic, r-hydroxybenzoic, and sinapic ) was performed by comparing their peak areas with those of the respective standards. The detection limit for 5-CQA (4-fold baseline noise) under the conditions used in this study was 0.002 mmol/l. Results of CGA and other phenolic compounds in urine were expressed as micromoles per millimoles creatinine. Urine creatinine was determined by the Jaffe reaction (23). Pharmacokinetic calculations. For each subject, plasma concentrations of caffeic and total and specific isomers of CQA and dicqa were plotted over time after coffee consumption for the 4 h of the study. From this plotting, the following plasma pharmacokinetic parameters were calculated: area under the curve (AUC) of total and individual components using GraphPad Prisma software (version 4.0); maximum plasma concentration (C max ) and time corresponding to C max (T max )of total and individual components. We also calculated molar ratios between AUC of specific components. Statistical analysis. Results are presented as means with corresponding SD. Associations between plasma AUC or C max and urinary concentration of specific compounds were tested by simple correlation using Statistica (version 7). Molar ratios of specific CGA components were compared between brewed coffee (calculated as concentration ratios) and plasma (calculated as AUC ratios) by unpaired t test using GrapPad Prisma software. Differences were considered significant at P, 0.05. Results Brewed coffee. The content of CGA compounds determined in the brewed coffee portion (190 ml) offered to the subjects is presented in Table 1. The CV for each CGA component in the brewed coffee aliquots analyzed was consistently lower than 5%, indicating that the brewing process was reproducible in terms of CGA composition. Eight major CGA compounds (3-CQA, 4-CQA, 5-CQA, 3,4-diCQA, 3,5-diCQA, 4,5-diCQA, 4-FQA, and 5-FQA) were identified in the coffee brew. CQA represented most of CGA (;86%) and 5-CQA contributed Absorption and metabolism of chlorogenic s 2197

TABLE 1 Content of the main CGA compounds in the brewed coffee portion offered to the subjects 1 Compound mmol/190 ml 3-CQA 858 6 35 4-CQA 1002 6 46 5-CQA 1068 6 49 Total CQA 2928 6 99 415-FQA 227 6 6 3,4-diCQA 77 6 3 3,5-diCQA 73 6 3 4,5-diCQA 90 6 4 Total dicqa 240 6 6 Total CGA 3395 6 105 1 Values are means 6 SD, n ¼ 6. 36.5%, 4-CQA 34.2%, and 3-CQA 29.3% of the CGA composition. The mean total amount of CGA in the 190 ml of brew was 3.4 mmol. The mean molar ratios of CQA:diCQA, 5-CQA:3-CQA, and 5-CQA:4-CQA in the brewed coffee were 12.2, 1.25, and 1.07, respectively. Plasma samples. No CGA, caffeic, or other phenolic compound was detected in plasma before coffee consumption. After coffee consumption, 3-CQA, 4-CQA, 5-CQA, 3,4-diCQA, 3,5-diCQA, and 4,5-diCQA were identified in the plasma of all subjects. In only 1 subject were 2 additional CGA (4-FQA and 5-FQA) identified. Although caffeic, an immediate metabolite of CQA and dicqa, was not detected in the brew, it was present in all plasma samples after coffee consumption, contributing on average to 14% of total plasma hydroxycinnamates. A large inter-individual variation was observed in the pharmacokinetic profile of all CGA compounds and caffeic in plasma (Fig. 1). However, the plasma kinetic profile of the individual isomers within the subclasses was very similar for each subject, following the same pattern. Therefore, for clarity purposes, only the plasma pharmacokinetic profiles of total CQA, total dicqa, total CGA, and caffeic at baseline and at 4 h after coffee consumption are given (Fig. 1). Two plasma concentration peaks were consistently identified in all subjects: peak 1, 0.5 1 h after coffee consumption, and peak 2, 1.5 4 h after coffee consumption. It should be noted that the C max within the 4 h of the study corresponded to the earliest peak in subjects 2 and 5 and to the latest peak in subjects 1, 4, and 6. C max, T max, and AUC of the CGA compounds identified in the plasma of the 6 subjects after coffee consumption are shown (Table 2). C max of total CGA varied from 4.7 to 11.8 mmol/l; C max of total CQA varied from 1.4 to 8.3 mmol/l, whereas C max of total dicqa varied from 1.1 to 4.0 mmol/l among the individuals, with mean concentrations of 7.7, 4.9, and 3.0 mmol/l, respectively. Although T max for the different CGA compounds was, on average, close to 2 h, T max for total CGA, total CQA, and total dicqa varied considerably in the individuals (from 1 to 4 h). 4-FQA and 5-FQA were identified in only 1 subject (no. 3) at 1, 1.5, and 2 h after coffee consumption. In this subject, the C max of 415-FQA was 0.38 mmol/l and T max was 1.0 h. C max of caffeic (mean ¼ 1.6 mmol/l) occurred at 1.4 h after coffee consumption, earlier than for total CQA (2.25 h) and for total dicqa (2.3 h). 5-CQA, the main CGA in the coffee brew, was also the major CGA identified in the plasma of all subjects at all FIGURE 1 Pharmacokinetic profile of total CGA (s), CQA (:),dicqa (h), and caffeic (CA) (d) in plasma samples of the individual subjects for 4 h after coffee consumption. time points after coffee consumption, as indicated by both C max and AUC of 5-CQA (Table 2). Considering mean values, 5-CQA plasma AUC contributed 40.7% of AUC of total hydroxycinnamates in plasma. AUC of caffeic corresponded to 14.1% of TABLE 2 Pharmacokinetic parameters of CGA and caffeic identified in plasma for 4 h after coffee consumption 1 Compound C max T max AUC mmol/l h mmolhl 21 3-CQA 1.00 6 0.75 1.75 6 0.99 1.65 6 0.96 4-CQA 1.04 6 0.68 2.08 6 1.20 1.85 6 1.24 5-CQA 3.14 6 1.64 2.33 6 1.17 8.10 6 5.05 Total CQA 4.89 6 2.53 2.25 6 1.25 11.48 6 7.12 3,4-diCQA 0.92 6 0.32 2.25 6 1.25 1.75 6 0.58 3,5-diCQA 1.17 6 0.95 2.33 6 1.17 1.85 6 0.83 4,5-diCQA 1.11 6 0.36 2.33 6 1.17 2.04 6 0.67 Total dicqa 3.03 6 1.28 2.33 6 1.17 5.63 6 1.83 Total CGA 7.66 6 2.50 2.25 6 1.25 17.11 6 8.41 Caffeic 1.56 6 1.52 1.42 6 0.38 2.81 6 1.75 Total hydroxycinnamates 19.92 6 10.16 1 Values are means 6 SD, n ¼ 6. 2198 Monteiro et al.

AUC of total hydroxycinnamates. The molar ratio CQA:diCQA in plasma (2.06 6 0.96) was lower than in the brewed coffee (12.24 6 0.15). Moreover, the plasma molar ratios 5-CQA: 3-CQA and 5-CQA:4-CQA were 4.91 6 2.16 and 4.38 6 2.85, respectively, ;4 times greater in plasma than in the brewed coffee (1.25 6 0.01 and 1.07 6 0.01, respectively). Urine samples. The urinary concentrations (mmol/mmol creatinine) of CGA metabolites were identified in each subject before and after coffee consumption (Table 3). Small amounts of 5-CQA, caffeic, ferulic, isoferulic, p-coumaric, gallic, and vanillic were identified in the urine of different subjects collected before coffee consumption. From these compounds, only isoferulic was identified in the urine of all subjects. As with plasma, after coffee consumption, a large inter-individual variation was observed in the concentration and urinary pharmacokinetic profile of all compounds. The only intact CGA compound identified in urine after coffee intake was 5-CQA. Dihydrocaffeic, gallic, isoferulic, vanillic, caffeic, ferulic, and sinapic were identified in the urine of all subjects. Para-hydroxybenzoic was identified in 3 subjects and r-coumaric was identified in 5 of the 6 subjects, respectively. Hippuric, n-metil hippuric, o-metil hippuric, and r-metil hippuric were not detected in the urine of any of the subjects. Gallic was the phenolic compound with the highest increment in concentration after coffee consumption in the urine of all the subjects, followed by dihydrocaffeic. On average, gallic and dihydrocaffeic concentrations together represented ;56% of the total concentration of phenolic compounds identified in urine at 4 h after coffee consumption. No significant correlations were observed between the urinary concentrations of 5-CQA or specific phenolic s and plasma AUC or C max of the total and individual phenolic compounds identified. Discussion The high content of CGA in the brewed decaffeinated C. canephora used in this study was expected and was in accordance with contents observed in the literature for both regular and decaffeinated coffees (6,8,19,20,22). This is the first study to our knowledge to identify 8 intact CGA compounds in human plasma after acute coffee consumption: 3 CQA isomers, 2 FQA isomers, and 3 dicqa isomers. While most studies investigating the bioavailability of CGA from coffee and other food sources identified only caffeic in plasma and not intact CGA compounds (11,12,14,16), our study demonstrates that all major CGA compounds in coffee are bioavailable in humans. Moreover, 5-CQA, the major CGA in coffee brew, was alone responsible for only 40% of total hydroxycinnamates identified in plasma for the 4 h of the study, with a considerable contribution of other CGA to total plasma hydroxycinnamates. 3-CQA and 4-CQA together were responsible for ;18% and dicqa responsible for ;28%. It should be noted that, in our study, plasma caffeic contributed to only 14% of the total plasma hydroxycinnamates. Considering that no nonesterified caffeic was present in the coffee brew and that during the analytical recovery test in plasma ;7% of 5-CQA was hydrolyzed into caffeic, our results suggest that part of caffeic identified in plasma might be originated from hydrolysis of CGA, probably as an artifact in the analytical procedure. Recovery tests showed that the addition of the Helix Pomatia extract to plasma or water containing 5-CQA caused partial hydrolysis of CGA with production of TABLE 3 Urinary concentration of CGA metabolites identified in each subject for 4 h after acute coffee consumption Subject 5- CQA Caffeic Dihydrocaffeic Ferulic Isoferulic r-coumaric Gallic r-hydroxy-benzoic Vanillic Sinapic Total Phenolics 1 mmol/mmol creatinine Baseline 0.35 Nd Nd 1.25 0.05 0,50 Nd Nd Nd Nd 2.15 0 2 h 0.37 1.15 4.96 3.21 0.19 2.08 7.63 Nd 3.85 2.47 25.91 2 4 h 0.51 0.66 8.52 2.83 0.20 2.14 9.67 Nd 4.33 4.56 33.42 2 Baseline Nd Nd Nd Nd 0.02 Nd 3.82 Nd 0.94 Nd 4.78 0 2 h 0.58 1.57 6.60 3.65 0.19 1.60 12.20 Nd 4.41 2.14 32.94 2 4 h 0.42 0.93 9.50 2.61 0.22 0.57 13.30 Nd 7.02 4.94 39.51 3 Baseline Nd 1 0.07 Nd 0.07 0.29 Nd Nd Nd Nd Nd 0.43 0 2 h 1.32 0.25 28.14 3.11 1.50 0.38 31.60 7.45 15.68 2.65 92.08 2 4 h 0.54 0.19 17.25 2.24 1.08 0.27 18.45 2.64 4.40 0.51 46.57 4 Baseline Nd Nd Nd Nd 0.96 Nd Nd Nd Nd Nd 0.96 0 2 h 1.35 1.35 19.55 7.59 4.56 1.18 65.38 Nd 16.55 1.18 118.69 2 4 h 0.95 0.74 40.27 4.78 3.40 Nd 42.97 18.28 6.83 1.68 119.90 5 Baseline 0.04 Nd Nd 0.05 2.07 Nd Nd Nd Nd Nd 2.16 0 2 h 0.85 0.98 47.87 7.86 4.40 Nd 140.57 24.92 4.21 1.40 233.06 2 4 h 0.34 2.36 23.47 13.72 9.59 Nd 65.60 9.94 1.75 1.37 128.14 6 Baseline 0.05 0.02 Nd Nd 0.42 Nd Nd Nd Nd Nd 0.49 0 2 h 0.66 0.37 24.45 3.35 1.34 1.44 41.98 Nd 21.72 14.49 109.80 2 4 h 0.89 0.76 15.33 12.91 3.87 1.22 35.03 Nd 14.42 11.61 96.04 1 Nd, Not detected. Absorption and metabolism of chlorogenic s 2199

caffeic. This is in accordance with Manach et al. (24), who observed that besides b-glucuronidase and sulfatase activities, this type of extract contains other esterases able to degrade CGA into caffeic. In addition to analytical artifacts, CGA hydrolysis during digestion, absorption, and/or metabolism should be considered. In fact, nonesterified caffeic was produced when 5-CQA was incubated with human intestinal digestive fluids (25). Also, esterases in rat and human intestinal mucosa with the ability to hydrolyze hydroxycinnamates have been reported (26,27). The 2 plasma concentration peaks of the different CGA compounds observed after coffee consumption suggest a complex process of absorption and metabolism consistent with metabolism of xenobiotic compounds. Considering that a liquid food may take up to 1 h to reach the small intestine (28,29), our results suggest an early absorption of CGA in the stomach or absorption in the initial intestinal portion, followed by absorption throughout the small intestine. Absorption of 5-CQA and other phenolic compounds, such as ferulic, p-coumaric, gallic, and caffeic s in the stomach has recently been reported in rats (17,30). In addition, it has been shown that 5-CQA and caffeic are also absorbed in the small intestine of rats (31,32), preferentially in the jejunum (31). Considering that enterohepatic circulation of phenolic compounds has been observed for up to 48 h after phenolics intake (12,33), we cannot exclude the possibility that part of CGA compounds identified in plasma a few hours after intake originate from recycling. Interestingly, in our study, the relative magnitude of CGA plasma peaks 1 and 2 varied considerably among the subjects. Moreover, a large inter-individual variability was observed in the overall plasma kinetics of all CGA compounds. This variability may be attributed to inter-individual differences in digestive transit time, preferential site of absorption, and metabolism of 5-CQA and caffeic as reported in the literature for other phenolic compounds (34 36). According to Manach et al. (34), important inter-individual differences in the capacity to metabolize phenolic compounds originate from differences in the activity of the cytochrome P450 and carrier systems that may be influenced by genetic polymorphisms. In this study, the lower average CQA:diCQA molar ratio in plasma compared with the coffee brew suggests different mechanisms of absorption and/or metabolism for CQA and dicqa, with favored dicqa absorption compared with CQA and/or favored tissue uptake of CQA compared with dicqa. Moreover, despite similar amounts of FQA and dicqa in the brew, FQA was detected in the plasma of only 1 subject in a concentration 5.5 times lower than that of dicqa in the same subject. The low or undetectable concentration of FQA in human plasma following coffee consumption may be explained by a poor absorption of these compounds compared with CQA and dicqa and/or by a rapid uptake and storage by organs such as the liver. Liver uptake and storage of 5-CQA has been previously suggested (31,37,38). Also, Farah (36) observed a favored uptake of FQA over CQA and dicqa in an in vitro study using human hepatoma cells. Another possibility for low plasma concentrations of FQA would be the demethylation of the ferulic moiety of the ester and conversion of FQA into CQA, because the only difference between both classes of compounds is the substitution of a metoxyl group by a hydroxyl group in the 3 position of the aromatic ring of the hydroxycinnamic moiety. Whether the demethylation would occur in the intestinal brush border cells or in the liver cells remains unanswered. In this study, the molar ratios 5-CQA:3-CQA and 5-CQA: 4-CQA were higher in plasma than in the coffee brew, suggesting either a higher absorption of 5-CQA compared with 3-CQA and 4-CQA isomers or that 3-CQA and 4-CQA are rapidly metabolized and/or stored in organs such as liver. This last possibility is in agreement with a previous study (36) where a favored uptake of compounds with the esterification of cinnamic s in the 3 and 4 positions of the quinic was observed compared with those in the 5 position. Although the subjects were advised not to consume food sources of phenolic compounds on the day prior to the study, low concentrations of 5-CQA, caffeic, ferulic, isoferulic, r-coumaric, gallic, and vanillic s were identified in the baseline (fasting) urine of the different subjects. Cremin et al. (12) also detected 5-CQA in urine of subjects even after 2 d consuming a low phenolic diet, possibly because 5-CQA is slowly metabolized and excreted. 3-CQA, 4-CQA, and 5-CQA have also been found in human gastrointestinal fluids after 12-h fasting (18). The identification of phenolic compounds in fasting urine and gastrointestinal fluids supports the hypothesis of gradual CGA utilization and excretion in humans (12,37). After coffee consumption, in addition to 5-CQA and the phenolic s present in the baseline urine, dihydrocaffeic, r-hydroxybenzoic, and sinapic s were also identified, consistent with previous human studies evaluating urinary metabolites after consumption of different sources of 5-CQA or hydroxycinnamates (12,13,15,16,33,39 42). Different authors (13,21, 39,42) suggest that ferulic, isoferulic, dihydrocaffeic, and vanillic s are the main human caffeic metabolites. In our study, dihydrocaffeic, an immediate caffeic metabolite, and gallic were the major metabolites excreted in urine after coffee consumption. Moreover, the urinary excretion of ferulic, isoferulic, and vanillic s also increased after coffee consumption. From the different CQA and dicqa compounds identified in plasma after coffee consumption, only 5-CQA, the major CGA in plasma, was also identified in urine, although in small amounts. Therefore, it appears that bile and other digestive fluids, rather than urine, are the preferential excretion pathways for circulating intact CGA compounds, as suggested for flavonoids (18,43,44). In conclusion, our study clearly indicates that at least 6 intact major CGA compounds are found in human plasma after acute coffee consumption and also that CQA and dicqa isomers are probably differentially absorbed and/or metabolized. Moreover, urine does not appear to be a major excretion pathway of intact CGA compounds. A large inter-individual variation clearly exists in CGA absorption and/or metabolism. Acknowledgment The authors thank COCAM Company (São Paulo, Brazil) for providing the coffee samples for this study. Literature Cited 1. Lindsay J, Laurin D, Verreault R, Hebert R, Helliwell B, Hill GB, McDowell I. Risk factors for Alzheimer s disease: a prospective analysis from the Canadian study of health and aging. Am J Epidemiol. 2002; 156:445 53. 2. Salazar-Martinez E, Willett WC, Ascherio A, Manson JE, Leitzmann MF, Stampfer MJ, Hu FB. Coffee consumption and risk for type 2 diabetes mellitus. Ann Intern Med. 2004;140:1 8. 3. Ranheim T, Halvorsen B. Coffee consumption and human health: beneficial or detrimental? Mechanisms for effects of coffee consumption on different risk factors for cardiovascular disease and type 2 diabetes mellitus. Mol Nutr Food Res. 2005;49:274 84. 4. Natella F, Nardini M, Giannetti I, Dattilo C, Scaccini C. Coffee drinking influences plasma antioxidant capacity in humans. J Agric Food Chem. 2002;50:6211 6. 2200 Monteiro et al.

5. Pellegrini N, Serafini M, Colombi B, Del Rio D, Salvatore S, Bianchi M, Brighenti F. Total antioxidant capacity of plant foods, beverages and oils consumed in Italy assessed by three different in vitro assays. J Nutr. 2003;133:2812 9. 6. Moreira DP, Monteiro MC, Ribeiro-Alves MR, Donangelo CM, Trugo LC. Contribution of chlorogenic s to the iron-reducing activity of coffee beverages. J Agric Food Chem. 2005;53:1399 402. 7. Clifford MN. Chlorogenic s and other cinnamates: nature, occurrence, dietary burden, absorption and metabolism. J Sci Food Agric. 2000;80:1033 43. 8. Farah A, Donangelo CM. Phenolic compounds in coffee. Braz J Plant Physiol. 2006;18:23 36. 9. Clifford MN, Johnston KL, Knight S, Kuhnert N. Hierarchical scheme for LC-MS n identification of chlorogenic s. J Agric Food Chem. 2003;51:2900 11. 10. Monteiro MC, Trugo LC. Determination of bioactive compounds in Brazilian roasted coffee. Quim Nova. 2005;28:637 41. 11. Azuma K, Ippoushi K, Nakayama M, Ito H, Higashio H, Terao J. Absorption of chlorogenic and caffeic in rats after oral administration. J Agric Food Chem. 2000;48:5496 500. 12. Cremin P, Kasim-Karakas S, Waterhouse ALLC. ES-MS detection of hydroxycinnamates in human plasma and urine. J Agric Food Chem. 2001;49:1747 50. 13. Rechner AR, Spencer JPE, Kuhnle G, Harn U, Rice-Evans CA. Novel biomarkers of the metabolism of caffeic derivatives in vivo. Free Radic Biol Med. 2001;30:1213 22. 14. Nardini M, Cirillo E, Natella F, Scaccini C. Absorption of phenolic s in humans after coffee consumption. J Agric Food Chem. 2002;50: 5735 41. 15. Olthof MR, Hollman PCH, Buijsman MNCP, Van Amelsvoort JMM, Katan MB. Chlorogenic, quercetin-3-rutinoside and black tea phenol are extensively metabolized in humans. J Nutr. 2003;133:1806 14. 16. Wittemer SM, Ploch M, Windeck T, Muller SC, Drewelow B, Derendorf H, Veit M. Bioavailability and pharmacokinetics of caffeoylquinic s and flavonoids after oral administration of artichoke leaf extracts in humans. Phytomedicine. 2005;12:28 38. 17. Lafay S, Gil-Izquierdo A, Manach C, Morand C, Besson C, Scalbert A. Chlorogenic is absorbed in its intact form in the stomach of rats. J Nutr. 2006;136:1192 7. 18. Farah A, Guigon F, Trugo LC. Chlorogenic from coffee are excreted in human digestive fluids. Proceedings of the 21 st International Conference on Coffee Science; 2007 Sep 11 15; Montpellier (France): ASIC. Available from: www.asic-cafe.org/htm/eng/proceedings.htm. 19. Farah A, De Paulis T, Trugo LC, Martin PR. Effect of roasting on the formation of chlorogenic lactones. J Agric Food Chem. 2005;53: 1505 13. 20. Trugo LC, Macrae R. Chlorogenic composition of instant coffees. Analyst. 1984;109:263 6. 21. Farah A. Distribuicxão nos grãos, importância na qualidade da bebida e biodisponibilidade dos ácidos clorogêncios do café [dissertation]. Rio de Janeiro: Universidade Federal do Rio de Janeiro; 2004. 22. Farah A, De Paulis T, Trugo LC, Martin PR. Chlorogenic s and lactones in regular and water-decaffeinated arabica coffee. J Agric Food Chem. 2006;54:374 81. 23. Husdan H, Rapoport A. Estimation of creatinine by the Jaffee reaction. A comparison of three methods. Clin Chem. 1968;14:222 38. 24. Manach C, Williamson G, Morand C, Scalbert A, Rémésy C. Bioavailability and bioefficacy of polyphenols in humans. I. Review of 97 bioavailability studies 1 3. Am J Clin Nutr. 2005;81:S230 42. 25. Farah A, Guigon F, Trugo LC. The effect of human digestive fluids on chlorogenic isomers in coffee. Proceedings of the 21 st International Conference on Coffee Science; 2007 Sep 11 15; Montpellier (France): ASIC. Available from: www.asic-cafe.org/htm/eng/proceedings.htm. 26. Andreasen MF, Kroon PA, Williamson G, Garcia-Conesa MT. Esterase activity able to hydrolyze dietary antioxidant hydroxycinnamates is distributed along the intestine of mammals. J Agric Food Chem. 2001;49:5679 84. 27. Kern SM, Bennett RN, Needs PW, Mellon FA, Kroon PA, Garcia- Conesa MT. Characterization of metabolites of hydroxycinnamates in the in vitro model of human small intestinal epithelium Caco-2 cells. J Agric Food Chem. 2003;51:7884 91. 28. Minami H, McCallum RW. The physiology and pathophysiology of gastric emptying in humans. Gastroenterology. 1984;86:1592 610. 29. Moore JG, Christian PE, Coleman RE. Gastric emptying of varying meal weight and composition in man. Evaluation by dual liquid and solid-phase isotopic method. Dig Dis Sci. 1981;26:16 22. 30. Konishi Y, Zhao Z, Shimizu M. Phenolic s are absorbed from the rat stomach with different absorption rates. J Agric Food Chem. 2006;54: 7539 43. 31. Spencer JP, Chowrimootoo G, Choudhury R, Debnam ES, Srai SK, Rice-Evans C. The small intestine can both absorb and glucuronidate luminal flavonoids. FEBS Lett. 1999;458:224 30. 32. Lafay S, Morand C, Manach C, Besson C, Scalbert A. Absorption and metabolism of caffeic and chlorogenic in the small intestine of rats. Br J Nutr. 2006;96:39 46. 33. Bourne LC, Rice-Evans C. Bioavailability of ferulic. Biochem Biophys Res Commun. 1998;253:222 7. 34. Manach C, Scalbert S, Morand C, Rémésy C, Jiménez L. Polyphenols: food sources and bioavailability. Am J Clin Nutr. 2004;79:727 47. 35. Rechner AR, Smith MA, Kuhnle G, Gibson GR, Debnam ES, Srai AKS, Moore KP, Rice-Evans CA. Colonic metabolism of dietary polyphenols: influence os structure on microbial fermentation products. Free Radic Biol Med. 2004;36:212 25. 36. Farah A, Monteiro MC, Guigon F, Moreira DP, Donangelo CM, Trugo LC. Chlorogenic s from coffee are absorbed and excreted in human digestive fluids. Proceedings of the 21 st International Conference on Coffee Science; 2007 Sep 11 15; Montpellier (France): ASIC. Available from: www.asic-cafe.org/htm/eng/proceedings.htm. 37. Booth AN, Emerson OH, Jones FT, Deeds F. Urinary metabolites of caffeic and chlorogenic s. J Biol Chem. 1957;229:51 9. 38. Baer-Dubowska W, Szaefer H. Inhibition of murine hepatic cytochrome p-450 activities by natural and synthetic compounds. Xenobiotica. 1998;28:735 43. 39. Hodgson JM, Chan SY, Puddey IB, Devine A, Wattanapenpaiboon N, Wahlqvist ML, Lukito W, Burke V, Ward NC, et al. Phenolic metabolites as biomarkers for tea- and coffee-derived polyphenol exposure in human subjects. Br J Nutr. 2004;91:301 5. 40. Ito H, Gonthier MP, Manach C, Morand C, Mennen L, Rémésy C, Scalbert A. Polyphenol levels in human urine after intake of six different polyphenol-rich beverages. Br J Nutr. 2005;94:500 9. 41. Rechner AR, Kuhnle G, Hu H, Roedig-Penman A, Van MH, Braak D, Moore KP, Rice-Evans CA. The metabolism of dietary polyphenols and the relevance to circulating levels of conjugated metabolites. Free Radic Res. 2002;36:1229 41. 42. Rechner AR, Kuhnle G, Bremner P, Hubbard GP, Moore KP, Rice-Evans CA. The metabolic fate of dietary polyphenols in humans. Free Radic Biol Med. 2002;33:220 35. 43. Gonthier MP, Verny MA, Besson C, Rémésy C, Scalbert A. Chlorogenic bioavailability largely depends on its metabolism by the gut microflora in rats. J Nutr. 2003;133:1853 9. 44. Donovan JL, Crespy V, Manach C, Morand C, Besson C, Scalbert A, Rémésy A. Catechin is metabolized by both the small intestine and the liver in rats. J Nutr. 2001;131:1753 7. Absorption and metabolism of chlorogenic s 2201