Development of a Hydrophilic Liquid Interaction Chromatography

Jan 14, 2010 - München, Gregor-Mendel-Strasse 2, D-85350 Freising, Germany, and Institute of Thermal Separation. ProcessessWorkgroup of Heat and Mass...
1 downloads 0 Views 646KB Size
Download PDF
Anal. Chem. 2010, 82, 1486–1497

Development of a Hydrophilic Liquid Interaction Chromatography-High-Performance Liquid Chromatography-Tandem Mass Spectrometry Based Stable Isotope Dilution Analysis and Pharmacokinetic Studies on Bioactive Pyridines in Human Plasma and Urine after Coffee Consumption Roman Lang,*,† Anika Wahl,† Thomas Skurk,‡ Erkan Firat Yagar,§ Ludger Schmiech,† Rudolf Eggers,§ Hans Hauner,‡ and Thomas Hofmann*,† Chair of Food Chemistry and Molecular Sensory Science, Technische Universita¨t Mu¨nchen, Lise-Meitner-Strasse 34, D-85354 Freising, Germany, Else Kroener-Fresenius-Center for Nutritional Medicine, Technische Universita¨t Mu¨nchen, Gregor-Mendel-Strasse 2, D-85350 Freising, Germany, and Institute of Thermal Separation ProcessessWorkgroup of Heat and Mass Transfer, Technische Universita¨t Hamburg Harburg, Eissendorfer Strasse 38, D-21073 Hamburg, Germany The paper reports on the development of an accurate hydrophilic liquid interaction chromatography tandem mass spectrometry (HILIC-MS/MS) based stable isotope dilution analysis for the simultaneous quantitation of the food-derived bioactive pyridines trigonelline, nicotinic acid, nicotinamide, and N-methylpyridinium, as well as their key metabolites nicotinamide-N-oxide, N-methylnicotinamide, N-methyl-2-pyridone-5-carboxamide, N-methyl4-pyridone-5-carboxamide, and N-methyl-2-pyridone-5carboxylic acid in human plasma and urine. Precision of the stable isotope dilution analysis (SIDA) was 1.9% and 11.9% relative standard deviation (n ) 6), and accuracy was between 92.4% and 113.0%. The lower limit of quantitation (LLOQ) was 50 fmol (10 pmol/mL) injected onto the column for all analytes with the exception of N-methyl-2-pyridone-5-carboxylic acid and N-methyl-2pyridone-5-carboxamide, for which an LLOQ of 100 fmol (20 pmol/mL) was found. The method was applied to monitor the plasma appearance and urinary excretion and to determine pharmacokinetic parameters of the bioactive pyridines as well as their metabolites in a clinical human intervention study with healthy volunteers (six women, seven men) after oral administration of 350 mL of a standard coffee beverage. Trigonelline plasma levels increased from 160 nmol/L to maximum concentrations of 5479 (males) or 6547 nmol/L (females), and Nmethylpyridinium plasma levels raised from virtually complete absence to maximum values of 777 (females) * To whom correspondence should be addressed. Phone: +49-8161-71-2902. Fax: +49-8161-71-2949. E-mail: [email protected] (T.H.). † Chair of Food Chemistry and Molecular Sensory Science, Technische Universita¨t Mu ¨ nchen. ‡ Else Kroener-Fresenius-Center for Nutritional Medicine, Technische Universita¨t Mu ¨ nchen. § Technische Universita¨t Hamburg Harburg.

1486

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

or 804 nmol/L (males) within 2-3 and 1-2 h after coffee consumption, respectively. The high plasma levels of N-methylpyridinium found after coffee consumption clearly demonstrate for the first time that this cation is entering the vascular system, which is the prerequisite for biological in vivo effects claimed for that compound. In contrast, the coffee intervention did not significantly influence the plasma concentrations of N-methyl-2-pyridone-5-carboxamide and N-methyl-4-pyridone-5-carboxamide, the major niacin metabolites. Within 8 h after coffee intervention, an urinary excretion of 57.4 ( 6.9% of trigonelline and 69.1 ( 6.2% of N-methylpyridinium was found for the male volunteers, whereas females excreted slightly less with 46.2 ( 7.4% and 61.9 ( 12.2% of these pyridines. Millions of people around the world appreciate coffee beverages whether at home, while on the move, or at work due to its pleasing overall aroma and typical taste as well as for its stimulatory effect. Although worldwide coffee consumption is constantly increasing, there is still a huge lack of information on the bioavailability of some of its bioactive ingredients. Beside caffeine, the betaine N-methylnicotinic acid, coined trigonelline (1 in Figure 1), is the second most abundant alkaloid in coffee beans. Unlike caffeine, trigonelline is known to be strongly degraded upon coffee roasting to give volatile pyridines and pyrroles as well as nonvolatile N-containing products such as nicotinic acid (2),1-3 which is together with nicotinamide (3) commonly referred to as niacin, and N-methylpyridinium ions (1) Viani, R.; Horman, I. J. Food Sci. 1974, 39, 1216–1217. (2) Lang, R.; Yagar, E. F.; Eggers, R.; Hofmann, T. J. Agric. Food Chem. 2008, 56, 11114–11121. (3) Smith, R. F. Nature 1963, 197, 1321. 10.1021/ac902616k  2010 American Chemical Society Published on Web 01/14/2010

Figure 1. Chemical structures of trigonelline (1), nicotinic acid (2), nicotinamide (3), N-methylpyridinium (4), N-methylnicotinamide (5), N-methyl4-pyridone-5-carboxamide (6), N-methyl-2-pyridone-5-carboxamide (7), N-methyl-2-pyridone-3-carboxamide (8), nicotinamide-N-oxide (9), and N-methyl-2-pyridone-5-carboxylic acid (10), respectively.

(4)4-6 inducing phase I/II detoxifying enzymes in the rat.7 In the human body, nicotinic acid (2) is quickly converted into nicotinamide (3), which is the functional part of the coenzymes nicotinamide adenine dinucleotide (NAD) and nicotinamide adenine dinucleotide phosphate (NADP) playing a key role in numerous dehydrogenase-catalyzed reactions in living organisms. Since niacin is known to be formed to some extent from L-tryptophan, nicotinic acid has been considered as a “semivitamin”.8 As trigonelline (1) and its thermal degradation products 2 and 4 are strongly hydrophilic molecules, they are almost quantitatively extracted from the roast coffee powder into the aqueous brew (data not published), which is then ingested upon coffee consumption. Metabolism of food-derived doses of nicotinic acid (2) primarily involves its conversion into its amide 3 which is N-methylated via S-adenosylmethionine giving rise to N-methylnicotinamide (5) and subsequently oxidized to yield N-methyl-4-pyridone-5-carboxamide (6) and N-methyl-2-pyridone-5-carboxamide (7).9-11 N-Methyl-2pyridone-3-carboxamide (8), the isomer of 6, was not yet identified as a metabolite. In contrast, pharmacologic doses of 1-3 g of nicotinamide or nicotinic acid have been shown to induce the formation and excretion of nicotinamide-N-oxide (9) besides nicotinuric acid.12 In addition to its pellagra preventive effect, nicotinamide (3) was found to inhibit endotoxin stimulated proinflammatory cytokine release in vitro.13 Its metabolite N-methylnicotinamide (5) showed anti-inflammatory properties possibly due to the ability to reduce adherence (4) Stadler, R. H.; Varga, N.; Hau, J.; Vera, F. A.; Welti, D. J. Agric. Food Chem. 2002, 50, 1192–1199. (5) Stadler, R. H.; Varga, N.; Milo, C.; Schilter, B.; Vera, F. A.; Welti, D. J. Agric. Food Chem. 2002, 50, 1200–1206. (6) Bosco, M.; Toffanin, R.; de Palo, D.; Zatti, L.; Segre, A. J. Sci. Food Agric. 1999, 79, 869–878. (7) Somoza, V.; Lindenmeier, M.; Wenzel, E.; Frank, O.; Erbersdobler, H. F.; Hofmann, T. J. Agric. Food Chem. 2003, 51, 6861–6869. (8) Heseker, H.; Stahl, A. Erna¨hr-Umsch. (in German) 2008, 8, 744–749. (9) Experts Group on Vitamins and Minerals, Review of Niacin. London Food Standard Agency 2002. http://www.food.gov.uk/multimedia/pdfs/evm-0111r.pdf (accessed Oct 24, 2008). (10) Chang, M. L. W.; Johnson, B. C. J. Biol. Chem. 1961, 236 (7), 2096–2098. (11) Wong, P.; Bachki, A.; Banerjee, K.; Leyland-Jones, B. J. Pharm. Biomed. Anal. 2002, 30, 773–780. (12) Mrochek, J. E.; Jolley, R. L.; Young, D. S.; Turner, W. J. Clin. Chem. 1976, 22, 1821–1827. (13) Ungerstedt, J. S.; Blomba¨ck, M.; So ¨derstro ¨m, T. Clin. Exp. Immunol. 2003, 131, 48–52.

of proinflammatory cells and molecules to the surface of vascular endothelium.14 Moreover, 5 was found to inhibit platelet-dependent thrombosis.15 Despite these highly beneficial properties, the diverse bioactivities of nicotinamide (3) have been critically reviewed with the conclusion that the compound holds great potential for treatment of diseases but needs more thorough investigations since it modulates numerous pathways of both cell survival and death.16 Ingested trigonelline (1) seems to be mostly excreted unchanged via the urine, although some amounts are likely to be formed from metabolism of nicotinic acid (2).17,18 N-Methyl-2pyridone-5-carboxylic acid (10) has been assumed to be a metabolite formed by trigonelline oxidation.19 Several healthrelated bioactivities of trigonelline have been reported in literature, e.g., the ability to induce axonal extension in human neuroblastoma SK-N-SH cells20 and to exhibit anti-invasive activity against cancer cells.21 Very recently, trigonelline and nicotinic acid were identified in pumpkin paste to be responsible for the early reduced glucose level in Goto Kakizaki rats when tested for oral glucose tolerance.22 A similar observation was done in a human intervention study, in which orally administered trigonelline (1) significantly delayed glucose and insulin responses in an oral glucose tolerance test, thus probably contributing to the putative beneficial effects of coffee on development of type 2 diabetes.23 Furthermore, the betaine 1 showed bacteriostatic activity against the cariogenic (14) Gebicki, J.; Sysa-Jedrzejowska, A.; Adamus, J.; Wo´zniacka, R. M.; Zielonka, J. Pol. J. Pharmacol. 2003, 55, 109–112. (15) Chlopicki, S.; Swies, J.; Mogielnicki, A.; Buczko, W.; Bartus, M.; Lomnicka, M.; Adamus, J.; Gebicki, J. Br. J. Pharmacol. 2007, 152, 230–239. (16) Maiese, K.; Chong, Z. Z.; Hou, J.; Shang, J. C. Molecules 2009, 14, 3446– 3485. (17) Sarett, H. P.; Perlzweig, W. A.; Levy, E. D. J. Biol. Chem. 1940, 135, 483– 485. (18) Perlzweig, W. A.; Levy, E. D.; Sarett, H. P. J. Biol. Chem. 1940, 136, 729– 745. (19) Lindenblad, G. E.; Kaihara, M.; Price, J. M. J. Biol. Chem. 1956, 219, 893– 901. (20) Tohda, C.; Nakamura, N.; Komatsu, K.; Hattori, M. Biol. Pharm. Bull. 1999, 22, 679–682. (21) Hirakawa, N.; Okauchi, R.; Miura, Y.; Yagasaki, K. Biosci., Biotechnol., Biochem. 2005, 69, 653–658. (22) Yoshinari, O.; Sato, H.; Igarashi, K. Biosci., Biotechnol., Biochem. 2009, 73, 1033–1041. (23) van Dijk, A. E.; Olthof, M. R.; Meeuse, J. C.; Seebus, E.; Heine, R. J.; van Dam, R. M. Diabetes Care 2009, 32, 1023–1025.

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1487

bacterium Streptococcus mutans.24 In contrast to these beneficial aspects, trigonelline (1) was found to stimulate the growth of estrogen-dependent breast cancer cells MCF-7 in vitro, thus leading to the suggestion that a trigonelline-rich diet might play a role in cancer development.25 To evaluate the biological impact of the food-derived compounds 1-4 as well as their putative metabolites 5-10, the development of an appropriate method for the analysis of these pyridine derivatives is mandatory. Although nicotinic acid derivatives can be easily visualized by UV detection,26,27 many quantitative analytical methods utilize electrospray mass spectrometry in the MS28,29 or the tandem mass spectrometry (MS/MS) mode2,30-32 for their sensitive and selective quantitation in foods and biofluids.32 Various attempts were undertaken to chromatographically separate such pyridines involving normal-phase HPLC,31 hydrophilic liquid interaction chromatography (HILIC),32 and reversedphase HPLC.2,25,27,29 Although the use of stable isotope labeled internal standards was found to successfully overcome matrix effects during LC-MS/MS analysis of selected target compounds complex biological samples such as plasma and urine,2,31 a multimethod for the accurate quantitative analysis of the target molecules 1-10 in a single chromatographic run by means of a stable isotope dilution analysis (SIDA) is as yet not available. Therefore, the aim of the study was to develop and validate a multimethod SIDA enabling the simultaneous quantitation of the bioactive alkaloid trigonelline (1) and its thermal degradation products 2 and 4, as well as their putative metabolites 3 and 5-10 in biofluids by means of LC-MS/MS detection. The SIDA developed should then be applied for the first time to quantitatively study plasma appearance and urinary excretion and to determine pharmacokinetic parameters for these bioactive pyridines in a clinical human intervention study with healthy volunteers after oral administration of a standard coffee beverage. MATERIALS AND METHODS Materials. Trigonelline hydrochloride, nicotinic acid, d4nicotinic acid, nicotinamide, nicotinamide-N-oxide, N-methylnicotinamide iodide, formic acid, and all chemicals for synthesis were obtained from Sigma-Aldrich (Steinheim, Germany). Solvents for chromatography were HPLC-grade, water was Millipore-grade. d3-Trigonelline hydroiodide (d31) was prepared by methylation of nicotinic acid using d3methyl iodide;33 d4-nicotinamide (d4-3) was obtained by amidation of d4-nicotinic acid.2 N-Methylpyridinium iodide (4) and d3-N-methylpyrdinium iodide (d3-4) were synthesized according to the literature.4,5 The syntheses of d3-N-meth(24) Antonio, A. G.; Morales, R. S.; Perrone, D.; Maia, L. C.; Santos, K. R. N.; Io´rio, N. L. P.; Farah, A. Food Chem. 2010, 118, 782–788. (25) Allred, K. F.; Yackley, K. M.; Vanamala, J.; Allred, C. D. J. Nutr. 2009, 139, 1833–1838. (26) Stratford, M. R. L.; Dennis, M. F. J. Chromatogr., B 1992, 582, 145–151. (27) Creeke, P. I.; Seal, A. J. J. Chromatogr., B 2004, 817, 247–253. (28) Pfuhl, P.; Ka¨rcher, U.; Ha¨ring, N.; Baumeister, A.; Tawab, M. A.; SchubertZsilavecz, M. J. Pharm. Biomed. Anal. 2005, 36, 1045–1052. (29) Perrone, D.; Donangelo, C. M.; Farah, A. Food Chem. 2008, 110, 1030– 1035. (30) Chen, P.; Wolf, W. R. Anal. Bioanal. Chem. 2007, 387, 2441–2448. (31) Li, A. C.; Chen, Y.-L.; Junga, H.; Shou, W. Z.; Jiang, X.; Naidong, W. Chromatographia 2003, 58, 723–731. (32) Hsieh, Y.; Chen, J. Rapid Commun. Mass Spectrom. 2005, 19, 3031–3036. (33) Ciusa, W.; Nebbia, G. Gaz. Chim. Ital. 1950, 80, 98–99.

1488

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

ylnicotinamide (d3-5), N-methyl-2-pyridone-5-carboxamide (7), d3-N-methyl-2-pyridone-5-carboxamide (d3-7), N-methyl2-pyridone-3-carboxamide (8), d3-N-methyl-2-pyridone-3-carboxamide (d3-8), N-methyl-2-pyridone-5-carboxylic acid (10), and d3-N-methyl-2-pyridone-5-carboxylic acid (d3-10) are detailed in the Supporting Information. d6-DMSO and d4MeOD for NMR spectroscopy were purchased from Eurisotop (Giv-sur Yvette, France). Raw coffee beans (Arabica Brazil, harvested in 2008) were roasted for 165 s in a laboratory-scale fluidized bed roaster (Novopack, Germany) operated with a gas temperature of 257 °C and were then cooled with air (150 m3 h-1 for 120 s) to afford a medium dark coffee (67 SKT). The coffee samples obtained from five individual roast batches (300 g each) were blended, ground, packed in polyethylene bags, and stored at -20 °C until use. Preparation of the Coffee Brew. An aliquot of coffee powder (48 g) was placed in a coffee filter (size 4, Melitta, Germany) and percolated with distilled water (900 mL) by means of a household filter coffee machine (TCM, Germany). The freshly prepared coffee beverage was used for the experiments. Nuclear Magnetic Resonance (NMR) and Mass Spectrometry (MS). NMR spectra were recorded on a Bruker 500 MHz Avance III and a Bruker 400 MHz DRX spectrometer, respectively. MS and MS/MS data were acquired on a API4000QTRAP triple quadrupole by constant infusion via a syringe pump at a flow rate of 10 µL min-1. The chemical structure of each purified, synthetic compound was confirmed by MS, MS/MS, and NMR data. Synthesis of N-Methyl-4-pyridone-5-carboxamide (6) and d3-N-Methyl-4-pyridone-5-carboxamide (d3-6). 4-Methoxynicotinic Acid Methylester. 4-Chloronicotinic acid (6.6 mmol) was refluxed in methanolic HCl (1.25 mol/L, 25 mL) for 5 h, the solvent was removed by evaporation, and the residue was taken up in water (20 mL). After adjusting the pH value to 10 by the addition of solid sodium carbonate, the solution was extracted with chloroform (5 × 20 mL); the combined organic layers were dried over sodium sulfate, filtered, and evaporated yielding 4-methoxynicotinic acid methylester as a white crystalline solid (5.9 mmol, 91%) in a purity of 99%. 1H NMR (400 MHz, MeOD-d4): δ (ppm) 3.91 (s, 3H, H-C(9)), 4.00 (s, 3H, H-C8), 7.23 (d, J ) 6.09 Hz, H-C(5)), 8.55 (d, J ) 6.09 Hz, H-C(6)), 8.79 (s, 1H, H-C(2)). 13C NMR (100 MHz, MeOD-d4, HMBC, HMQC): δ (ppm) 51.5 (CH3, C(8)), 55.4 (CH3, C(9)), 107.9 (CH, C(5)), 116.5 (C, C(3)), 151.4 (CH, C(6)), 153.5 (CH, C(2)), 164.9 (C, C(7)), 165.6 (C, C(4)). 4-Methoxynicotinamide. 4-Methoxynicotinic acid methylester (820 mg, 4.9 mmol) was dissolved in a solution of ammonia in methanol (7 mol/L, 20 mL) and incubated with stirring in a closed reaction vessel at 90 °C for 48 h. The solution was reduced in a stream of nitrogen until the target compound crystallized. The solid was collected and dried under vacuum overnight yielding 4-methoxynicotinamide as a white solid (650 mg, 4.3 mmol, 87.3%) in a purity of 99%. 1H NMR (400 MHz, DMSO-d6): δ (ppm) 3.94 (s, 3H, H-C(8)), 7.16 (d, J ) 5.87 Hz, H-C(5)), 7.63 (brs, 1H, NHa), 7.66 (brs, 1H, NHb), 8.52 (d, J ) 5.87 Hz, H-C(6)), 8.72 (s, 1H, H-C(2)). 13C NMR (100 MHz, DMSO-d6, HMBC, HMQC): δ (ppm) 55.9 (CH3, C(8)), 107.5 (CH, C(5)), 119.0 (C, C(3)), 151.0 (CH, C(6)), 153.2 (CH, C(2)), 162.7 (C, C(7)), 164.9 (C, C(4)).

N-Methyl-4-pyridone-5-carboxamide Hydroiodide. A solution of 4-methoxynicotinamide (290 mg, 1.9 mmol) and methyl iodide (1 mL, 14 mmol) in water/methanol (1/9, v/v; 1 mL) was stirred for 24 h at 60 °C in the dark. After cooling, the yellow precipitate formed was collected; the supernatant was evaporated, redissolved in methanol (1 mL), and again heated with methyl iodide (1 mL, 14 mmol) for 24 h at 60 °C. The precipitates formed after cooling were combined and crystallized from acetone to afford crude N-methyl-4-pyridone-5-carboxamide hydroiodide as a yellow powder (400 mg, purity ∼80% based on 1H NMR). 1H NMR (500 MHz, DMSO-d6): δ (ppm) 3.87 (s, 3H, H-C(8)), 6.66 (d, J ) 7.32 Hz, 1H, H-C(3)), 6.96 (brs, 1H, NHa), 7.66 (brs, 1H, NHb), 7.96 (dd, J ) 2.32 Hz, 7.32 Hz, 1H, H C(2)), 8.63 (d, 1H, J ) 2.32 Hz, H-C(6)). 13C NMR (125 MHz, HMBC, HMQC, DMSO-d6): δ (ppm) 44.2 (CH3, C(8)), 118.3 (C, C(5)), 118.6 (CH, C(3)), 143.34 (CH, C(2)), 146.2 (CH, C(6)), 164.8 (C, C(7)), 174.3 (C, C(4)). For further purification and to obtain the free base of the target compound, N-methyl-4-pyridone-5-carboxamide hydroiodide (400 mg) was dissolved in water and purified by preparative HPLC on RP18 material (Hypersil, 250 mm × 21.5 mm i.d., 5 µm, Thermo Hypersil GmbH, Kleinostheim, Germany) using water (eluent A) and methanol (eluent B) as the mobile phase at a flow rate of 18 mL min-1. While monitoring the effluent at 240 nm, chromatography was performed starting with 100% solvent A for 1 min, then increasing solvent B to 50% within 15 min, followed by 12 min of isocratic elution. Freeze-drying of the collected peak effluent yielded the free base of N-methyl-4-pyridone-5-carboxamide as a white amorphous material (1 mmol, 55%) in a purity of 99%. N-Methyl-4-pyridone-5-carboxamide (6). MS (ESI+): m/z (%) 153 ([M + H]+, 100). MS/MS (ESI+, CE 52 V): m/z (%) 153 (100), 136 (30), 92 (26), 52 (20). 1H NMR (400 MHz, DMSOd6): δ (ppm) 3.75 (s, 3H, H-C(8)), 6.38 (d, 1H, J ) 7.49 Hz, H-C(3)), 7.41 (brs, 1H, Ha-N), 7.74 (dd, 1H, J ) 2.44 Hz, 7.49 Hz, H-C(2)), 8.44 (d, 1H, J ) 2.44 Hz, H-C(6)), 9.56 (brs, 1H, Hb-N). 13C NMR (100 MHz, DMSO-d6, HMBC, HMQC): δ (ppm) 43.4 (CH3, C(8)), 118.5 (C, C(5)), 119.8 (CH, C(3)), 141.9 (CH, C(2)), 145.6 (CH, C(6)), 165.2 (C, C(7)), 176.1 (C, C(4)). For the synthesis of d3-N-methyl-4-pyridone-5-carboxamide (d3-6), 4-methoxynicotinamide was methylated using deuteromethyl iodide, followed by crystallization and RP-HPLC purification to afford d3-N-methyl-4-pyridone-5-carboxamide (0.8 mmol, 42%) as the free base in a purity of 99%. d3-N-Methyl-4-pyridone-5-carboxamide (d3-6). MS (ESI+): m/z (%) 156 ([M + H]+, 100). MS/MS (ESI+, CE 52 V): m/z (%) 156 (100), 139 (55), 111 (15), 95 (50), 52 (20). 1H NMR (400 MHz, DMSO-d6): δ (ppm) 6.39 (d, 1H, J ) 7.49 Hz, H-C(3)), 7.42 (brs, 1H, Ha-N), 7.75 (dd, 1H, J ) 2.44 Hz, 7.49 Hz, H-C(2)), 8.45 (d, 1H, J ) 2.44 Hz, H-C(6)), 9.58 (brs, 1H, Hb-N). 13C NMR (100 MHz, DMSO-d6, HMBC, HMQC): δ (ppm) 119.0 (C, C(5)), 120.4 (CH, C(3)), 142.5 (CH, C(2)), 146.2 (CH, C(6)), 165.8 (C, C(7)), 176.7 (C, C(4)). Hydrophilic Liquid Interaction Chromatography/Mass Spectrometry. The Agilent 1200 series HPLC system consisted of a pump, a degasser, and an autosampler (Agilent, Waldbronn, Germany) and was interfaced to a 4000 QTRAP triple quadrupole

(Applied Biosystems/MDS Sciex, Darmstadt). Separation was carried out on a TSKgel Amide 80 3 µm column (150 mm × 2 mm, TOSOH, Bioscience Stuttgart, Germany) equipped with guard column (10 mm × 2 mm) of the same type using acetonitrile containing 1% formic acid (eluent A) and aqueous 1% formic acid (eluent B) as the mobile phase at a flow rate of 200 µL min-1. After sample injection (5 µL), chromatography was performed by increasing eluent B from 20% to 50% within 5.5 min, then to 97% within 1 min, followed by an isocratic period (2.5 min) for column rinsing. The starting conditions were re-established within 1 min, and the column was equilibrated for 8 min prior to the next injection. The effluent was channeled into the ion source 2 min after injection and to waste 8 min after injection by means of a Valco valve. The ion source was operated in the positive electrospray mode, nitrogen was used as the nebulizer gas (50 psi) and curtain gas (15 psi), respectively, heater gas was set at 50 psi, and the temperature was 380 °C. In the multiple reaction monitoring (MRM) mode, the transitions from the parent ion [M]+ and [M + H]+, respectively, to the fragment ions formed after collision-induced dissociation (CID) were recorded (Table S-1 in the Supporting Information). The CAD gas was set at “medium”, the dwell time for each mass transition was 15 ms, and the pause between the mass ranges was 5 ms. The quadrupoles operated at unit mass resolution, and the ion spray voltage was +5500 V. For instrumentation control and data acquisition Sciex Analyst software v1.4.2 (Applied Biosystems, Darmstadt, Germany) was used. Quantitative Analysis and Method Validation. Stock Solutions. Individual stock solutions of the analytes and the internal standards trigonelline (1), d3-trigonelline (d3-1), nicotinic acid (2), d4-nicotinic acid (d4-2), nicotinamide (3), d4-nicotinamide (d4-3), N-methylpyridinium (4), d3-N-methylpyridinium (d3-4), N-methylnicotinamide (5), d3-N-methylnicotinamide (d3-5), Nmethyl-4-pyridone-5-carboxamide (6), d3-N-methyl-4-pyridone5-carboxamide (d3-6), N-methyl-2-pyridone-5-carboxamide (7), d3-N-methyl-2-pyridone-5-carboxamide (d3-7), N-methyl-2-pyridone-3-carboxamide (8), d3-N-methyl-2-pyridone-3-carboxamide (d3-8), nicotinamide-N-oxide (9), N-methyl-2-pyridone-5-carboxylic acid (10), and d3-N-methyl-2-pyridone-5-carboxylic acid (d3-10) were prepared in DMSO (1-5/d3-1-d3-5) or acetonitrile/water (9/1, v/v; 6-10/d3-6-d3-10) at a concentration 5 µmol/mL each and stored at -20 °C until use. Internal Standard (IS) Working Solutions. Appropriate aliquots of the internal standard stock solutions were combined and diluted with acetonitrile to give a multicomponent stock solution with a concentration of 10 nmol/mL of each labeled compound. Aliquots of this multicomponent stock solution were stored at -20 °C until use. After thawing, the internal standards stock solution was diluted 1:10 with aqueous hydrochloric acid (0.1 mol/L) to give the internal standard working solution with concentrations of 1 nmol/mL of each labeled compound. Calibration Curve, Linear Range, and Lower Limit of Quantitation. Standard solutions containing a fixed concentration of the labeled internal standards (100 pmol/mL) and the analytes in concentrations ranging from 10 to 1000 pmol/mL in acetonitrile/ water (80/20, v/v) were analyzed by HILIC-MS/MS. Peak area ratios of analyte and corresponding internal standard obtained from six replicates per standard solution were plotted against the Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1489

analyte concentration followed by linear regression with 1/x2 weighing. The concentrations of the standards were backcalculated using the calibration curves, and precision and accuracy were calculated as relative standard deviation (RSD, %) and deviation from nominal value, respectively (Table S-2 in the Supporting Information). The lower limit of quantitation (LLOQ) was defined as the lowest standard concentration with precision < ±15% relative standard deviation and accuracy >85% of the nominal value. Quantitative Analysis in Plasma, Urine, and Coffee Brew. An aliquot (50 µL) of the plasma sample or the urine sample, which was 1:20 diluted with water, was placed into an Eppendorf cup (1.5 mL) containing the internal standard working solution (1 nmol/mL of d3-1-d3-10, 50 µL). After vortexing for 30 s, acetonitrile (300 µL) was added, followed by centrifugation (12 000 rpm, 4 °C, 20 min), and an aliquot (5 µL) of the supernatant was directly analyzed by means of HILIC-MS/ MS as detailed above. For quantitation of trigonelline (1), nicotinic acid (2), nicotinamide (3), and N-methylpyridinium (4) in coffee beverage, aliquots (1 mL) of separately prepared coffee brews (n ) 8) were pooled, and after dilution with water (1/10, 1/100, 1/1000, v/v), portions (50 µL) were spiked with the internal standard working solution (1 nmol/mL of d3-1-d3-10, 50 µL), diluted with acetonitrile (300 µL), centrifuged, and aliquots (5 µL) were analyzed by means of HILIC-MS/MS as detailed above. Precision and Accuracy in Plasma Samples. Plasma samples were collected from individuals who did not consume coffee or any coffee-containing foods for at least 10 days and were pooled to give the pooled blank plasma. Spiked plasma samples were prepared by adding an aliquot (100 µL) of serial dilutions of a mixture of the analytes 1-10 to aliquots (900 µL) of the pooled blank plasma to yield plasma samples spiked with a concentration of 0.10, 0.25, and 1.00 nmol/mL plasma. After an equilibration for 60 min at 6 °C, these samples were frozen and stored at -20 °C until use. Blank and spiked plasma samples were processed and analyzed in replicates (n ) 6) as detailed above to assess precision and accuracy. Human Intervention Study. The coffee drinking study was approved by the ethical commission of the faculty of medicine of the Technische Universita¨t Mu¨nchen, Germany. The study was performed at the Human Study Center (HSC) of the Else Kro¨nerFresenius-Center for Nutritional Medicine at the Technische Universita¨t Mu¨nchen, Germany. The line-up consisted of six female (age 24-28, body weight 61.0 kg ± 7.2%) and seven male (age 27-30, body weight 73.7 kg ± 12.9%) healthy, nonsmoking volunteers who were used to consume coffee on a daily base. All volunteers gave written consent to the study. The clinical study started with a run-in phase (10 days) without consumption of tea, coffee, cocoa, cola, and chocolate, as well as any vitamin supplements. At the day of admission to the study center, the volunteers were fasted for at least 12 h. A venous catheter was inserted into the anticubital vein, and a blood sample was taken as a base level; thereafter a single dose of a freshly prepared coffee beverage (350 mL) was ingested within 5 min. Additional blood sampling was performed 15, 30, 45, 60, 120, 240, 360, and 480 min after coffee ingestion. The volunteers consumed a bolus of tap water (250 mL) every 2 h, and urine was quantitatively collected in 2 h intervals. 1490

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

The amount of urine excreted was measured by weight and aliquots were stored at -20 °C until analysis. Blood was collected into EDTA tubes (1 mg/mL) and was immediately centrifuged at 3000g for 10 min at 4 °C. Two plasma aliquots (50 µL, see workup above) of each sampling point was directly used for analysis; the remaining material was aliquoted in 500 µL portions and immediately frozen at -80 °C. Urine samples were kept at 4 °C during the study day and frozen as aliquots (40 mL) at the end of the study day. Calculation of Pharmacokinetic Data. Basal pharmacokinetic data were calculated using Graphpad Prism 5.02 for Windows, GraphPad Software, San Diego, CA, www.graphpad. com. “One-phase exponential decay” was chosen for nonlinear curve fitting of the last three and four quantified plasma concentrations of trigonelline and N-methylpyridinium, respectively, and for calculation of the elimination constant Kel and plasma halflife time. Area under curve from 0 to 8 h (AUC0-8h) was calculated from time zero to 8 h post coffee ingestion. Area under curve 0 h-∞ (AUC0h-∞) was calculated from AUC0-8h + (C8h/Kel), both by the linear trapezoidal rule. Collected urine was measured by weight and was converted to volume based on a density of about 1 kg/L. RESULTS AND DISCUSSION To accurately determine the amounts of the alkaloid trigonelline (1, Figure 1), the pyridine derivatives nicotinic acid (2), nicotinamide (3), and N-methylpyridinium (4) generated upon thermal degradation of 1 during coffee bean roasting, as well as their known metabolites N-methylnicotinamide (5), N-methyl-4pyridone-5-carboxamide (6), N-methyl-2-pyridone-5-carboxamide (7), nicotinamide-N-oxide (9), and the putative trigonelline metabolite N-methyl-2-pyridone-5-carboxylic acid (10), in human biofluids, a stable isotope dilution assay (SIDA) with HILIC-MS/ MS detection should be developed. Therefore, reference material of the not commercially available analytes 4, 6-8, and 10 as well as the corresponding stable isotope labeled analogues of 1-10 needed to be synthesized as internal standards. Synthesis of Reference Compounds and Stable Isotope Labeled Internal Standards. According to literature protocols, the iodide salts of d3-trigonelline (d3-1), N-methylpyridinium (4), d3-N-methylpyridinium (d3-4), and d3-N-methylnicotinamide (d35) were synthesized by alkylation of nicotinic acid,33 pyridine,4,5 and nicotinamide,34 respectively. d4-Nicotinamide (d4-3) was prepared by amidation of d4-nicotinic acid with ammonia.2 N-Methyl-2-pyridone-5-carboxamide (7), d3-N-methyl-2-pyridone5-carboxamide (d3-7), N-methyl-2-pyridone-3-carboxamide (8), and d3-N-methyl-2-pyridone-3-carboxamide (d3-8) were obtained by oxidation of N-methylnicotinamide iodide and d3-N-methylnicotinamide iodide, respectively, with iron(III) in aqueous alkaline solution. Using a similar strategy, N-methyl-2-pyridone5-carboxylic acid (10) and d3-N-methyl-2-pyridone-5-carboxylic acid (d3-10) were prepared by iron(III)-catalyzed oxidation of 1 and d3-1, respectively, under alkaline conditions.34 In contrast to the preparation of the pyridones 7 and 8, rather laborious strategies involving multiple synthetic steps are reported for the synthesis of the metabolite N-methyl-4-pyridone-5-carboxa(34) Holman, W. I. M.; Wiegand, C. Biochem. J. 1948, 43, 423–426.

Figure 2. Reaction sequence used for the synthesis of natural deuterium-abundant 6 and d3-labeled N-methyl-4-pyridone-5-carboxamide (d3-6); “b” indicates the d3-labeled methyl group in d3-6. Table 1. Overview on Mass Transitions, Chromatographic Properties, Calibration Curves, and Lower Limit of Quantitation of the HILIC-HPLC-MS/MS Method Developed for the Analysis of the Pyridines 1-10 mass transitionsa

compound (no./internal standard) trigonelline (1/d3-1) nicotinic acid (2/d4-2) nicotinamide (3/d4-3) N-methylpyridinium (4/d3-4) N-methylnicotinamide (5/d3-5) N-methyl-4-pyridone-5-carboxamide (6/d3-6) N-methyl-2-pyridone-5-carboxamide (7/d3-7) N-methyl-2-pyridone-3-carboxamide (8/d3-8) nicotinamide-N-oxide (9/d4-3) N-methyl-2-pyridone-5-carboxylic acid (10/d3-10)

m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z m/z

138 f 94*, 92 (1) 141 f 97*, 94 (d3-1) 124 f 80*, 78 (2) 128 f 84*, 81 (d4-2) 123 f 80*, 78 (3) 127 f 84*, 83 (d4-3) 94 f 79*, 78 (4) 97 f 79*, 78 (d3-4) 137 f 94*, 79 (5) 140 f 97*, 94 (d3-5) 153 f 136, 95, 92*, (6) 156 f 139, 110, 95*, (d3-6) 153 f 110*, 108 (7) 156 f 113*, 111 (d3-7) 153 f 136*, 108 (8) 156 f 139*, 111 (d3-8) 139 f 106*, 78, 51(9) 127 f 84*, 83 (d4-3) 154 f 126*, 110, 108 (10) 157 f 129*, 113, 111 (d3-10)

retention time (min)b

linear range (nM)

R2

precision (%)

accuracy (%)

LLOQ (fmol)c

5.55

10-1000

0.999

1.4-6.8

96.2-103.4

50

3.19

10-1000

0.997

5.2-13.7

98.1-102.1

50

3.89

10-1000

0.995

6.6-11.2

97.7-103.2

50

4.84

10-500

0.999

1.3-10.7

96.6-102.2

50

5.75

10-1000

0.999

1.7-3.5

93.9-105.7

50

3.45

10-1000

0.999

1.0-11.2

95.4-102.1

50

3.17

20-1000

0.997

5.7-11.5

93.9-107.3

100

2.70

10-1000

0.998

2.6-11.7

92.3-106.4

50

4.32

10-1000

0.995

4.2-8.4

98.0-102.9

50

2.52

20-1000

0.997

4.5-14.4

95.2-101.6

100

a Quantifier mass transitions are marked by an asterisk. b Retention time of the compound. c Lower limit of quantitation is defined as lowest standard concentration of linear range and injected volume of 5 µL.

mide.35-37 Therefore, compound 6 as well as its isotopologue d3-6 were prepared in three steps starting with an acid-catalyzed methanolysis of 4-chloronicotinic acid, followed by its amidation with ammonia in methanol to give 4-methoxynicotinamide. Using a literature protocol,35 4-methoxynicotinamide was readily hydrolyzed in aqueous methanol to give the 4-hydroxy derivative which was reacted with methyl iodide and d3-methyl iodide, respectively, to yield the N-methylated 4-oxonicotinamides 6 as well as its isotopologue d3-6 as shown in Figure 2. Development of an SIDA and Quantitation of Analytes in Plasma and Urine Samples. In order to analyze the target pyridines with high selectivity by using MS/MS operating in the MRM mode, solutions of the analytes 1-10 as well as their corresponding isotopologues were individually infused into the electrospray ionization (ESI) source of the MS/MS system with a constant flow by means of a syringe pump to optimize ionization parameters and collision-induced fragmentation. At least two mass transitions of each compound were selected for the final acquisition method; typically the most intensive mass transition was used for quantitation, and a second transition was selected for unequivocal identification of the target analytes (Table 1). Preliminary studies with a series of different stationary phases revealed HILIC to meet the demand for the simultaneous analysis of the charged and noncharged pyridines 1-10 in a single chromatographic run (data not shown). (35) Wieland, T.; Fest, C.; Pfleiderer, G. Annalen (in German) 1961, 163–173. (36) Ross, W. C. J. J. Chem. Soc. 1966, 1816–1821. (37) Bernofsky, C. Anal. Biochem. 1979, 96, 189–200.

In order to verify the suitability of the HILIC-MS/MS analysis for the determination of the target compounds in human plasma, a plasma sample was collected from healthy volunteers 60 min after coffee consumption, pooled and spiked with the stable isotope labeled internal standards. The sample was then separated from plasma proteins by acetonitrile precipitation and, then, injected into the HILIC-MS/MS system monitoring selected mass transitions of the analytes 1-10 and the corresponding labeled internal standards. Final optimization of the chromatographic conditions enabled sufficient separation of all compounds of interest from matrix compounds with gradient elution within 9 min including column rinsing (Figure 3). Using the optimized chromatographic system, calibration curves were recorded for each analyte from 10 to 1000 pmol/mL to compensate differences in ionization parameters, intrinsic instrumental background noise of the respective MRM trace, and to assess linearity of the method (Table 1). For all compounds analyzed, excellent linearity was found with R2 g 0.995. The response of N-methylpyridinium was linear between 10 and 500 pmol/mL. The absolute amount at the lower end of the linear range, calculated from the lowest standard concentration included in the calibration curve (precision < ±15% relative standard deviation, accuracy >85%) and the injected sample volume of 5 µL, was defined as the lower limit of quantitation (LLOQ). The LLOQ for compounds 1-6, 8, and 9 was 50 fmol, and for compounds 7 and 10 it was 100 fmol injected onto the column (Table 1) corresponding to concentrations of 10 pmol/mL and 20 pmol/mL, respectively, in the processed sample. Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1491

Figure 3. Quantifier-MRM traces of the analytes (solid lines) and respective internal standards (dotted lines) in a human plasma sample collected 60 min after coffee consumption: trigonelline (1, m/z 138 f 94; d3-1, m/z 141 f 97), nicotinic acid (2, m/z 124 f 80; d4-2, m/z 128 f 84), nicotinamide (3, m/z 123 f 80; d3-3, m/z 127 f 84), N-methylpyridinium (4, m/z 94 f 79; d3-4, m/z 97 f 79), N-methylnicotinamide (5, m/z 137 f 94; d3-5, m/z 140 f 97), N-methyl-4-pyridone-5-carboxamide (6, m/z 153 f 92; d3-6, m/z 156 f 95), N-methyl-2-pyridone-5carboxamide (7, m/z 153 f 110; d3-7, m/z 156 f 113), N-methyl-2-pyridone-3-carboxamide (8, m/z 153 f 136; d3-8, m/z 156 f 129), nicotinamide-N-oxide (9, m/z 139 f 106, d4-3, m/z 127 f 84), and N-methyl-2-pyridone-5-carboxylic acid (10, m/z 154 f 126; d3-10, m/z 157 f 129).

Analysis of a sample of pooled plasma collected from volunteers who did not consume coffee or any coffee-containing food products for a period of at least 10 days demonstrated that the pyridine derivatives nicotinic acid (2) and N-meth1492

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

ylpyridinium (4) present in coffee and N-methyl-2-pyridone-5carboxylic acid (10) were not detectable in plasma when keeping a coffee-free diet (Table 2). On the other hand, compounds 1, 3, and 5-7 were found even in this plasma

Table 2. Results from Performed Standard Addition Experiments to Determine Precision and Accuracy of the Quantitative Analysis of 1-7 and 10 in Human Plasmaa plasma analyte spiking level (nmol/mL) 1 2 3 4 5 6 7 10 none 0.10 0.25 1.00

a

found (±SD) precision (RSD, found (±SD) precision (RSD, accuracy (%) found (±SD) precision (RSD, accuracy (%) found (±SD) precision (RSD, accuracy (%)

0.12 %) 9.0 0.20 %) 7.8 93.5 0.38 %) 1.9 103 1.14 %) 5.8 102

(±0.01) n.d.b

0.12 4.8 (±0.02) 0.11 (±0.01) 0.21 7.8 4.7 113 96.8 (±0.01) 0.26 (±0.01) 0.35 3.4 2.7 104 95.5 (±0.07) 0.95 (±0.03) 1.05 3.5 7.9 94.9 93.6

(±0.01) n.d.b

0.13 4.2 (±0.01) 0.10 (±0.01) 0.22 11.9 5.2 103 96.1 (±0.01) 0.24 (±0.01) 0.37 4.3 3.8 97.2 97.5 (±0.08) 1.02 (±0.04) 1.07 3.6 7.2 102 95.4

(±0.01) 0.25 7.5 (±0.01) 0.32 8.5 94.0 (±0.01) 0.47 5.7 95.5 (±0.08) 1.21 4.7 98.0

(±0.02) 1.17 9.1 (±0.03) 1.23 5.3 96.7 (±0.03) 1.33 9.1 93.9 (±0.07) 2.00 6.2 92.4

(±0.11) n.d.b (±0.07) 0.11 (±0.02) 9.7 108 (±0.12) 0.24 (±0.01) 5.4 95.9 (±0.12) 1.04 (±0.05) 4.4 104

Concentrations given in nmol/mL plasma, means of replicates (n ) 6). b n.d. not detected because signal too low (S/N < 10).

sample, thus confirming these compounds as naturally abundant metabolites in human plasma as reported earlier.2,31 Not even trace amounts of N-methyl-2-pyridone-3-carboxamide (8) were detectable in any plasma sample taken within the study. As can be seen in Figure 3, only matrix compounds appeared in the MRM trace of compound 8. Nicotinamide-N-oxide (9) on the other hand could be detected, but concentrations were below the LLOQ in the processed plasma samples. As the cleanup of biofluid samples spiked with stable isotope labeled internal standards involved a protein precipitation step and irreproducible analyte/protein adsorption phenomena were not to be excluded, accuracy and precision of the method were evaluated by corresponding standard addition experiments. To achieve this, defined amounts of the analytes 1-7 and 10 were added to the pooled plasma sample, which was prepared from the “zero” samples collected after the coffee wash-out period, thus delivering spiked plasma samples containing the analytes in levels of 0.1-1.0 nmol/mL (Table 2). The precision of the SIDA, expressed by the relative standard deviation (RSD, %) obtained by replicate analysis (n ) 6) of each spiking level ranged between 1.9% and 11.9%, and the accuracy, defined by agreement (%) of calculated and measured concentration in the spiked plasma samples, was between 92.4% and 113.0%. These data clearly demonstrate the developed SIDA as a reliable tool enabling a rapid, simultaneous, and accurate quantitative determination of the pyridine derivatives 1-7 and 10 in human plasma samples (Table 2). Although being not validated for other matrixes, the developed SIDA is expected to be applicable also for the analysis of the target compounds in other matrixes such as, e.g., coffee and urine, because the workup of these sample types holds no critical steps which might lead to discrimination or loss of analyte. Influence of Coffee Consumption on the Plasma Appearance and Urinary Excretion of Pyridines in Humans. In order to quantitatively determine the influence of coffee consumption on the bioappearance of coffee derived pyridines in humans, a clinical human intervention study was conducted with a group of six female and seven male healthy volunteers who were used to consume coffee on a daily base. After keeping a coffee-free diet for 10 days, the volunteers were fasted for at least 12 h and, then, ingested a single dose of a freshly prepared coffee beverage within 5 min. HILIC-MS/MS-SIDA analysis of this coffee sample revealed that the ingested dose of 350 mL of the coffee beverage

Table 3. Concentrations of the Pyridines 1-4 in a Freshly Prepared Coffee Brew and in the Ingested Dose of Coffee Used in the Intervention Study compound

1

2

3

4

coffee brew 2310 (±7.2%) 71.5 (±5.4%) 1.67 (±12.5%) 491 (±6.5%) (nmol/mL)a ingested dose 808500 25025 585 171850 (nmol)b a Data a means of eight separate beverages pooled and analyzed in replicates (n ) 4). b Calculated on the basis of the quantitative data and a volume of 350 mL of the coffee brew.

contained 808 500 nmol of trigonelline (1), 25 025 nmol of nicotinic acid (2), 585 nmol of nicotinamide (3), and 171 850 nmol of N-methylpyridinium (4), respectively (Table 3). Quantitative analysis of the target compounds in the plasma samples taken 15, 30, 45, 60, 120, 240, 360, and 480 min after coffee ingestion revealed that compounds 1 and 4 were rather quickly absorbed into the circulation as shown in the plasma concentration-time plots in Figure 4. Starting from similar initial trigonelline plasma concentrations of about 160 nmol/L with large interindividual differences in both volunteer groups (Table 4), a rapid increase within the first 2-4 h after ingestion of the coffee brew was observed, and interestingly, no significant differences between male and female subjects were observed within this first period after coffee ingestion. However, the maximum plasma concentration (Cmax) of 6547 (±17.3%) nmol/L found for compound 1 was higher and was reached later (tmax ) 3.17 h) in the female group when compared to the group of the male volunteers (5479 ± 13.6%, tmax ) 2.29 h). This phenomenon is expected to be due to the difference in body weight and blood volume and, in consequence, in the higher dilution in the blood of the male volunteers. In fact, the mean body weight of the female group was 17.5% lower than that of the male group, and the mean AUC0-8h was 21.5% higher. Although the interindividual differences in the plasma samples taken after the coffee wash-out period were largely different (±57.3% in females, ±95.0% in males), the calculated AUC0-8h values within each group showed rather similar values with only low variance with 40 430 (±15.2%) nmol h L-1 found for the female group and 31 716 (±12.3%) nmol h L-1 found for the male group (Table 4). After reaching the maximum concentration, the plasma levels of 1 Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1493

Figure 4. Concentration-time plots of trigonelline (1, left plot) and N-methylpyridinium (4, right plot) in plasma samples collected after coffee consumption. Table 4. Pharmacokinetic Plasma Parameters and Urinary Excretion of Trigonelline (1) and N-Methylpyridinium (4) in Males and Females after Coffee Ingestion trigonelline (1) pharmacokinetic parameters in human plasma

females

N-methylpyridinium (4) males

females

males

Cmin (nmol/L) Cmax (nmol/L) tmax (h) range observed for tmax

157 (±57.3%) 6547 (±17.3%) 3.17 (±32.5%) 2.0-4.0

161 (±95.0%) 5479 (±13.6%) 2.29 (±43.5%) 1.0-4.0