- Email: [email protected]
Hypoxanthine levels in human urine serve as a screening indicator for the plasma total cholesterol and low-density lipoprotein modulation activities of fermented red pepper paste
Available online at www.sciencedirect.com
Nutrition Research 30 (2010) 455 – 461 www.nrjournal.com
Hypoxanthine levels in human urine serve as a screening indicator for the plasma total cholesterol and low-density lipoprotein modulation activities of fermented red pepper paste Yujin Kim a,1 , Youn-Je Park b,1 , Seung-Ok Yang a , So-Hyun Kim a , Sun-Hee Hyun a , Sayeon Cho a , Young-Suk Kim c , Dae Young Kwon d , Youn-Soo Cha e , Soowan Chae f , Hyung-Kyoon Choi a,⁎ a
College of Pharmacy and Research Center for Biomolecules and Biosystems (WCU), Chung-Ang University, Seoul 156-756, Republic of Korea b Department of Applied Bioscience, CHA University, Seoul 135-081, Republic of Korea c Department of Food Science and Technology, Ehwa Womans University, Seoul 120-750, Republic of Korea d Korea Food Research Institute, Seongnam 463-746, Republic of Korea e Department of Food Science and Human Nutrition, Chonbuk National University, Jeonju 561-756, Republic of Korea f Clinical Trial Center for Functional Foods, Chonbuk National University, Jeonju 561-756, Republic of Korea Received 20 April 2010; revised 7 June 2010; accepted 23 June 2010
Abstract Fermented red pepper paste (FRPP) is one of the most well-known traditional foods in Korea. The effects of FRPP in experimental animals and adipocytes have been previously reported. However, the biochemical effects have not yet been validated in humans with various genetic backgrounds and environmental factors. In this study, 28 female volunteers (body mass index, more than 23 kg/m2) aged 19 to 60 years were treated with either FRPP or a placebo for 12 weeks. Marked cholesterol modulation was observed in the FRPP-treated group compared with the placebo group. Although the baseline (pretreatment) total cholesterol and low-density lipoprotein levels and body mass index of the volunteers did not differ significantly between the placebo- and FRPP-treated groups, FRPP caused a modulation of cholesterol levels not seen in the placebo group, causing either no variation or a decrease in low-density lipoprotein and total cholesterol levels. Thus, urinary metabolomic profiling of pretreatment samples was carried out in these 2 FRPP-treated groups using 1H-nuclear magnetic resonance–based metabolomic techniques. These 2 groups, with their opposing cholesterol-modulation tendencies, could be clearly differentiated by orthogonal projections to latent structures–discriminant analysis-derived score plots. In addition, their levels of hypoxanthine differed markedly. We propose that urinary hypoxanthine levels can be used as a screening biomarker to predict the efficacy of the cholesterol-modulating activity of FRPP in human subjects. © 2010 Elsevier Inc. All rights reserved. Keywords: Abbreviations:
Capsicum annuum; Fermented red pepper paste; Cholesterol-modulation activity; Metabolomics; 1H-NMR; Women BMI, body mass index; FRPP, fermented red pepper paste; LDL, low-density lipoprotein; OPLS-DA, orthogonal projections of latent structure–discriminant analysis; TC, total cholesterol.
1. Introduction ⁎ Corresponding author. Fax: +82 2 812 3921. E-mail address: [email protected] (H.-K. Choi). 1 These authors contributed equally to this work. 0271-5317/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.nutres.2010.06.014
Worldwide, the obese population is increasing, and obesity has become a serious socioeconomic problem because it can result in hypertension, diabetes, osteoarthritis, stroke, and cancer [1,2]. Previous studies have found that
456
Y. Kim et al. / Nutrition Research 30 (2010) 455–461
obesity exerts strong effects on lipoprotein metabolism, leading to elevated levels of low-density lipoprotein (LDL) and total cholesterol (TC) [3]. Fermented red pepper paste (FRPP; Gochujang, CODEX standard number 294R-2009) is one of the most well-known traditional foods in Korea [4]; it is made with glutinous rice, Meju (fermented soybean blocks), salt, and red pepper (Capsicum annuum L.) powder and is fermented for several months or longer. Its unique taste, flavor, and color are due to the actions of microorganisms such as bacteria, yeasts, and molds during the fermentation process [5,6]. Several studies have found antiobesity effects of FRPP in animal models and cell cultures. Rhee et al [7] investigated the reduction in lipid levels in adipose tissue and serum after FRPP feeding, Choo [8] discovered that FRPP increases energy expenditure and results in a reduction in body fat gain, and Ahn et al [9] added FRPP to 3T3-L1 adipocytes and found that it decreases fat accumulation by modulating adipogenesis and lipolysis. However, these studies have focused mainly on the analysis of data from other animals or cells, rather than on humans. Humans have diverse metabolic regulation systems as a result of genetic and environmental differences; thus, the effects of nutritional interventions can be specific to the individual [10]. Metabolomic approaches have been applied to toxicity screening, drug metabolism, and identifying the biochemical effects of several foods [11-13]. In those reports, urine was found to be a useful sample for metabolomics studies [14]. Kim et al [15] categorized rats into 4 groups—normal-diet low gainers, normal-diet high gainers, high-fat-diet low gainers, and high-fat-diet high-gainers—and studied the metabolomic differences in their urine, whereas Akira et al [16] characterized the urinary metabolite profiles of young, spontaneously hypertensive rats in the early stages of hypertension, and Williams et al [17] determined the metabolic fingerprints of male and female obese rats using 1H-nuclear magnetic resonance (NMR)–based metabolomic techniques. In addition, various examples of the successful application of metabolomics in studies of the metabolic fingerprinting of other biologic fluids (eg, serum) and tissues, combined with the use of multivariate statistical analysis [18,19], exist. The main hypothesis of this study was that FRPP might have different antiobesity and cholesterol-modulating effects in individuals, and the effects could be predicted using urinary metabolic profiling. In the current study, the antiobesity and cholesterol-modulating effects of FRPP were assessed in humans. The FRPP-treated cohort was subsequently divided into 2 groups according to whether the FRPP caused a modulation in their TC and LDL levels, and their pretreatment (baseline) urine samples were subjected to metabolic profiling by 1 H-NMR–based metabolomic techniques. The objectives of the study were to assess antiobesity and cholesterol-modulating effects of FRPP and to determine whether urinary metabolites can be used as screening indicators for selecting patients for whom FRPP will be efficacious as an antiobesity and cholesterol-modulating treatment.
2. Methods and materials 2.1. Preparation of FRPP samples Fermented red pepper paste was prepared following a traditional fermentation process starting with the preparation of Meju added to glutinous rice. Soybeans (Glycine max L.) containing 10.5% water, 5.4% ash, 17.9% fat, 38.7% crude protein, and 26.2% carbohydrate were used. The soybeans were digested for 9 hours at 20°C, dried in baskets, autoclaved at 121°C for 30 minutes, cooled until the temperature decreased to 30°C, and then ground into fine powder. Glutinous rice was digested for 12 hours, autoclaved, and cooled in the same conditions. The obtained soybean powder and glutinous rice were mixed in a 6:4 ratio, 0.5% (wt/wt) koji rice prepared with Aspergilus sojae, and 0.5% (wt/wt) Bacillus subtilis culture broth was added, and then these were molded in 18 × 10 × 3-cm boxes to prepare Meju. The Meju was then stored in a room at 35°C with 90% relative humidity for 3 days. The fermented Meju was dried for 3 to 4 days and crushed with a flat roller (Daeyul Industry Co., LTD, Daejun, Republic of Korea) for the preparation of FRPP. Fermented red pepper paste contains 18.0% glutinous rice, 6.0% Meju powder, 4.0% dried barley sprouts, 10.5% salt, 17.5% FRPP powder, and 44.0% distilled water according to the standardized protocols of FRPP preparation by Soonchang Moonokrae Foods (Soonchang, Republic of Korea). Fermented red pepper paste was distributed into eight 100-kg jars and was stored for fermentation and maturation for 3 months in the sauce manufacturing facility of Soonchang Moonokrae Foods. Fermented red pepper paste pills were prepared by mixing the glutinous rice powder, FRPP powder, Meju powder, dried barley sprouts, salt, soy sauce, black-eyed pea powder, and cocoa powder
Table 1 Composition of FRPP pills and placebo pills Components
FRPP
Placebo
Contents (g) Contents (g) Main components
Enhancer
Additive Total weight Freeze-dried weight Energy (kJ)
FRPP powder Glutinous rice powder Dried barley sprouts Salt Meju powder Soy sauce Black-eyed pea powder Cocoa powder Flour Fat powder Honey Pepper scent Spicy powder Caramel pigment Red pigment
11.9 10.9 5.2 5.2 4.7 2.1 7.5 2.5 – – – – – – – 50.0 32.0 477.3
– – – 1.1 1.4 6.7 9.3 1.4 13.1 6.3 1.4 0.3 0.1 0.1 1.5 41.3 32.0 473.1
Y. Kim et al. / Nutrition Research 30 (2010) 455–461
in the mixing ratios shown in Table 1. The pills for the placebo group were also prepared using the formulation shown in Table 1, and these contained spicy-flavored powder instead of FRPP, pepper scent, Meju powder, soy sauce, fat powder, flour, black-eyed pea powder, cocoa powder, salt, honey, caramel pigment, and red pigment. All samples were stored at −70°C until administration to the volunteers after thawing in room temperature. 2.2. Study subjects Twenty-eight female volunteers with body mass index (BMI) more than 23 kg/ m2 (19-60 years of age) were recruited and randomly divided into 2 groups as they were treated with FRPP pills or placebo pills. The energy content of both pills were adjusted to 477.3, 473.1 kJ/g, respectively. They were fed either a FRPP pill or a placebo pill 3 times a day for 12 weeks. All 28 volunteers had previously consented to the treatment. The protocol was approved by the Functional Foods Institutional Review Board of Chonbuk National University Hospital. 2.3. Plasma and urine sample preparation and cholesterol analysis Blood was drawn from the volunteers after a 12-hour fast. Plasma samples were then prepared by the Chonbuk National University Hospital. The concentrations of LDL and TC in plasma were analyzed on a Hitachi 7600-110 analyzer (Hitachi High-Technologies Corporation, Tokyo, Japan) by standard methods at the biochemical laboratory of Chonbuk National University Hospital. Twenty-four-hour urine samples were collected from the volunteers at Chonbuk National University. Ten milliliters of the samples was then transferred into a 15-mL conical tube with 1% sodium azide and were stored at −70°C until assay. 2.4. NMR measurements Pretreatment (baseline) urine samples were thawed in a water bath at 40°C and then vortexed for 5 seconds; 0.3 mL of each urine sample was transferred to an Eppendorf tube. Then, 0.1 mol D2O [containing 0.05% 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid, sodium salt, as an internal standard for D2O] was made by adding 1.232 g of KH2PO4 as a buffering agent to 100 g of D2O. The pH was adjusted to 6.0 (as measured using a pH meter; model 720P; Istek, Seoul, Korea) by adding 1 N NaOD. A 0.2-mL aliquot of D2O was added to a 0.3-mL aliquot of the urine sample and vortexed for several seconds to ensure thorough mixing. The mixture thus prepared was pipetted into a 5-mm NMR tube (Norell, Landisville, NJ) [20]. Samples were analyzed at 600.13 MHz with an NMR spectrometer (Avance 600 FT-NMR, Bruker, Germany) at a temperature of 298 K. To suppress the residual water signal, a noesygpprld pulse sequence was used. With a spectral width of 10 775.9, a total of 64 transients were collected into 64 data
457
points. The acquisition time per scan of 3.04 seconds was used with a relaxation delay of 1 second. The exponential linebroadening function of 0.30 Hz was applied to free induction decays before Fourier transform. 2.5. Statistical analyses The spectral region from δ = 0.00 to 10.00 was segmented into 0.04-ppm using Amix software (Version 3.7, Bruker Biospin). In aqueous extracts, the region from δ = 4.60 to 4.90 was excluded due to the residual signal of water, along with the urea region from δ = 5.60 to 6.00 [21]. All data were mean centered, with scaling to unit variance as a preprocessing method, and orthogonal projections of latent structure– discriminant analysis (OPLS-DA) were conducted using the SIMCA-P software (version 12.0; Umetrics, Umeå, Sweden). An analysis of variance and independent t test were performed using SPSS software (version 17; SPSS Inc, Chicago, Ill); the power analysis of the t test was conducted using Power and Precision software (version 3.2.0; Biostat, Englewood, NJ). The mean differences, standard deviations, level of significance (P = .05), and sample size were used for the calculation of the power value. 3. Results and discussion 3.1. Biochemical parameters As indicated in Table 2, the BMI did not differ significantly between pretreatment and posttreatment conditions in both the FRPP- and placebo-treated groups. Thus, FRPP had no significant effect on body weight. However, the cholesterol-modulating effect was markedly greater in the FRPP-treated group than in the placebotreated group. The TC and LDL levels of the participants treated with the placebo pills did not differ significantly during the 12-week experimental period. However, the subjects in the FRPP-treated group exhibited either a significant decrease of more than 10.7% and 15.7% in TC and LDL or no significant variation in TC and LDL after the 12-week experimental period. In other words, FRPP exerted 2 different effects in our human cohort, whose baseline (pretreatment) TC/LDL levels and BMIs did not significantly different from each other. This effect was investigated further by subjecting the pretreatment urine samples of the FRPP-treated group to metabolomic profiling using 1 H-NMR-based metabolomic techniques. 3.2. 1H-NMR spectral analysis Fig. 1 shows the assignment of each peak in the representative 1 H-NMR spectra from urine sample of pretreatment of FRPP. The following signals were assigned based on comparisons with the chemical shifts of standard compounds using the Chenomx NMR suite software (version 5.1; Chenomx, Inc, Edmonton, Canada): leucine at δ = 0.94 (d, J = 6.78 Hz); isoleucine at δ = 0.94 (d, J = 6.78 Hz); valine at δ = 0.98 (d, J = 6.9 Hz) and 1.02 (d, J = 6.24 Hz);
458
Y. Kim et al. / Nutrition Research 30 (2010) 455–461
Table 2 Pretreatment and posttreatment plasma TC and LDL levels and BMIs in the FRPP- and placebo-treated groups FRPP (n = 14) TC/LDL-decreased group (n = 8)
Placebo (n = 14) TC/LDL-unvaried group (n = 6)
Pretreatment
Posttreatment
Pretreatment
Posttreatment
Pretreatment
Posttreatment
TC (mg/dL)
197.0 ± 16.8
207.0 ± 32.8
224.2 ± 23.8
202.2 ± 27.2
197.7 ± 25.9
LDL(mg/dL)
134.3 ± 17.3
175.0 ± 14.2 ⁎ 10.7% a 112.8 ± 14.7 ⁎ 15.7% a 26.4 ± 2.3
134.8 ± 25.3
149.5 ± 24.8
126.3 ± 27.8
122.9 ± 27.8
26.3 ± 2.6
26.5 ± 2.2
27.3 ± 3.1
26.4 ± 3.4
BMI
26.4 ± 2.4
Values are means ± SD (n = 6 or n = 8). a TC and LDL decrease ratio (%) = (posttreatment TC/LDL − pretreatment TC/LDL) × 100/(pretreatment TC/LDL). ⁎ Represent significant differences between predose group and postdose group at P b .05, and the power values regarding the t test results of TC and LDL in the TC/LDL-decreased group were 0.75 and 0.70, respectively.
methylsuccinate at δ = 1.06 (d, J = 7.1 Hz); 2-methylglutarate at δ = 1.06 (d, J = 7.1 Hz); lactate at δ = 1.34 (d, J = 7.3 Hz); alanine at δ = 1.46 (d, J = 7.9 Hz); acetate at δ = 1.94 (s); citrate at δ = 2.54 (d, J = 14.6 Hz) and δ = 2.74 (d, J = 17.7 Hz); creatine at δ = 3.02 (s) and δ = 3.90 (s); creatinine at δ = 3.06 (s) and δ = 4.10; taurine at δ = 3.42 (t, J = 6.0 Hz); malonate at δ = 3.34 (s); glycine at δ = 3.54 (s); mannitol at δ = 3.66 (dd, J1 = 6.5 Hz, J2 = 6.0 Hz), δ = 3.74 (m), δ = 3.78 (d, J = 9.4 Hz), and δ = 3.86 (dd, J1 = 2.8 Hz, J2 = 3.4 Hz);
τ-methylhistidine at δ = 3.86 (s), δ = 7.34 (s), and δ = 8.34 (s); histidine at δ = 7.34 (s) and δ = 8.34 (s); hippurate at δ = 7.54 (t, J = 7.2 Hz), δ = 7.62 (t, J = 7.1 Hz), and δ = 7.82 (d, J = 7.9 Hz); hypoxanthine at δ = 8.18 (s); and formate at δ = 8.42 (s). 3.3. OPLS-DA of predose samples of the FRPP treatment Orthogonal projection to latent structures–discriminant analysis is a preprocessing method that provides better
Fig. 1. Representative 1H-NMR spectra of FRPP pretreatment urine samples.
Y. Kim et al. / Nutrition Research 30 (2010) 455–461
visualization and interpretation of the model than partial least squares–discriminant analysis. The predictive component, which stands for the variation between the groups, is predicted based on the known orthogonal component, which stands for the variation within the groups [22]. Therefore, OPLS-DA is a useful tool for discriminating between compounds that contribute to the separation between groups of interest. Orthogonal projections of latent structure–discriminant analysis was applied to investigate the metabolic differences in the pretreatment urine between those in the FRPP-treated group that exhibited no variation in TC/LDL (TC/LDLunvaried group) and those that exhibited a decrease in TC/ LDL (TC/LDL-decreased group). Fig. 2A shows the OPLS-
459
DA–derived score plots, which reveal that the groups can be clearly separated. This means that the effect of FRPP differed according to the individual's pretreatment metabolic conditions. To identify the compounds contributing to the separation of each sample in the OPLS-DA model, a loading plot analysis was performed (Fig. 2B). The locations of loading plots correspond to those of score plots. Therefore, the peaks in the positive position of the loading plots are dominant in the samples, which are located in the positive position along the x axis of the score plots in this study. In the pretreatment condition, the levels of creatine, taurine, malonate, τ-methylhistidine, histidine, and hippurate were higher in the TC/LDL-decreased group and are located
Fig. 2. A, OPLS-DA-derived score plot of the pretreatment samples (b, pretreatment TC/LDL-unvaried group; 0, pretreatment TC/LDL-decreased group). B, OPLS-DA-derived loading plot of the pretreatment samples of the TC/LDL-unvaried and TC/LDL-decreased groups. 1, leucine; 2, isoleucine; 3, valine; 4, methylsuccinate; 5, 2-methylglutarate; 6, lactate; 7, alanine; 8, acetate; 9, citrate; 10, creatine; 11, taurine; 12, malonate; 13, glycine; 14, mannitol; 15, τ-methylhistidine; 16, histidine; 17, hippurate; 18, hypoxanthine; 19, formate.
460
Y. Kim et al. / Nutrition Research 30 (2010) 455–461
and to reveal potential biomarkers for the personalized treatment of various diseases, including obesity. Acknowledgment This study was made possible by a National Research Foundation of Korea grant, funded by the Korean government (MEST; 2009-0065538) and by the World Class University program through the National Research Foundation of Korea, funded by the Ministry of Education, Science and Technology (grant number: R33-10029). References Fig. 3. Relative intensities of hypoxanthine (normalized to the creatinine intensity) in pretreatment urine samples from the 2 FRPP-treated groups (TC/LDL-unvaried in one and decreased in the other). An independent t test was performed to assess any statistically significant differences between the 2 groups. The error bars represent the standard deviation (n = 6 or 8). *Significantly different at P b .05; the power value was 0.79 (obtained by power analysis).
in the negative position of the loading plot, whereas relatively higher levels of leucine, isoleucine, valine, methylsuccinate, 2-methylglutarate, lactate, alanine, acetate, citrate, glycine, mannitol, hypoxanthine, and formate were observed in the TC/LDL-unvaried group, which are located in the positive position of the loading plot. An independent t test was performed to validate the significance of differences between the relative intensities of the selected compounds between the 2 groups (data not shown). With the exception of hypoxanthine, there were no significant differences in the relative levels of these selected compounds. Therefore, only the relative intensities of hypoxanthine are shown in Fig. 3; these were significantly higher in the TC/LDL-unvaried group than in the TC/LDLdecreased group. It was previously reported that serum hypoxanthine levels tend to increase in obese and oxidatively stressed conditions [23]. The variable effects of FRPP between the TC/LDLunvaried and TC/LDL-decreased groups may be attributable to differences in lifestyle, diet, physical constitution, and/or genetic background among the subjects. In the pretreatment condition, the urinary level of hypoxanthine was significantly lower in the TC/LDL-decreased group than in the TC/LDL-unvaried group. Therefore, we accept the hypothesis for this study that the urinary metabolite can be used as a screening indicator for cholesterolmodulating effects of FRPP. Further study is needed to determine the correlation between pretreatment urinary hypoxanthine levels and other environmental or genetic factors among individuals. We propose that hypoxanthine could be used as a potential biomarker for the screening of subjects to ensure the cholesterol-modulation activity of FRPP. Thus, our findings suggest that this methodology can be applied to assess the bioactivities of other natural resources in humans
[1] Kopelman PG. Obesity as a medical problem. Nature 2000;404: 635-43. [2] Grundy SH. Multi-factorial causation of obesity: implication for prevention. Am J Clin Nutr 1998;67:563-72. [3] Krauss R, Winston M, Fletcher B, Grundy S. Obesity: impact on cardiovascular disease. Circulation 1998;98:1472-6. [4] http://www.codexalimentarius.net/web/standard_list.do?lang=en, (accessed February 15, 2010). [5] Lee TS, Chun MS Choi JY, Noh BS. Changes of free sugars and free amino acids in kochujang with different mashing method. Food Biotechnol 1993;2:102-8. [6] Chun MS, Lee TS, Noh BS. The changes in capsaicin, dihydrocapsaicin and cansanthin in kochujang with different mashing methods. Food Biotechnol 1994;3:104-8. [7] Rhee SH, Kong KR, Jung KO, Park KY. Decreasing effect of kochujang on body weight and lipid levels of adipose tissues and serum in rats fed a high-fat diet. J Food Sci Nutr 2003;32:882-6. [8] Choo JJ. Anti-obesity effects of kochujang in rats fed on a high-fat diet. Korean J Nutr 2000;33:787-93. [9] Ahn IS, Do MS, Kim SO, Jung HS, Kim YI, Kim HJ, Park KY. Antiobesity effect of kochujang (Korean fermented red pepper paste) extract in 3T3-L1 adipocytes. J Med Food 2006;9:15-21. [10] German JB, Roberts MA, Watkins SM. Personal metabolomics as a next generation nutritional assessment. J Nutr 2003;133:4260-6. [11] Nicholson JK, Connelly J, Lindon JC, Holmes E. Metabonomics: a platform for studying drug toxicity and gene function. Nat Rev Drug Discov 2002;1:153-61. [12] Bales JR, Sadler PJ, Nicholson JK, Timbrell JA. Urinary excretion of acetaminophen and its metabolites as studied by proton NMR spectroscopy. Clin Chem 1984;30:1631-6. [13] Solanky KS, Bailey NJ, Beckwith-Hall BM, Bingham S, Davis A, Holmes E, et al. Biofluid 1H NMR-based metabonomic techniques in nutrition research-metabolic effects of dietary isoflavones in humans. J Nutr Biochem 2005;16:236-44. [14] Bollard ME, Holmes E, Lindon JC, Mitchell SC, Branstetter D, Zhang W, et al. Investigations into the biochemical changes due to diurnal variation and estrus cycle in female rats using high resolution 1H NMR spectroscopy of urine and pattern recognition. Anal Biochem 2001; 295:194-202. [15] Kim SH, Yang SO, Kim HS, Kim Y, Park T, Choi HK. 1H-nuclear magnetic resonance spectroscopy-based metabolic assessment in a rat model of obesity induced by a high-fat diet. Anal Bioanal Chem 2009; 395:1117-24. [16] Akira K, Masu S, Imachi M, Mitome H, Hashimoto M, Hashimoto T. 1 H NMR-based metabonomic analysis of urine from young spontaneously hypertensive rats. J Pharm Biomed Anal 2008;46:550-6. [17] Williams R, Lenz E, Evans J, Wilson I, Granger J, Plumb R, et al. A combined 1H NMR and HPLC-MS-based metabonomic study of urine from obese (fa/fa) Zucker and normal Wistar-derived rats. J Pharm Biomed Anal 2005;38:465-71.
Y. Kim et al. / Nutrition Research 30 (2010) 455–461 [18] Lindon JC, Nicholson JK, Holmes E, Antti H, Bollard ME, Keun H, et al. Contemporary issues in toxicology the role of metabonomics in toxicology and its evaluation by the COMET project. Toxicol Appl Pharmacol 2003;187:137-46. [19] Nicholson JK, Wilson ID. High resolution proton magnetic resonance spectroscopy of biological fluids. Prog Nucl Magn Reson Spectrosc 1989;21:449-501. [20] Webb-Robertson BJM, Lowry DF, Jarman KH, Harbo SJ, Meng QR, Fuciarelli AF, et al. A study of spectral integration and normalization in NMR-based metabonomic analyses. J Pharm Biomed Anal 2005;39:830-6.
461
[21] Wei L, Liao P, Wu H, Li X, Pei F, Li W, et al. Toxicological effects of cinnabar in rats by NMR-based metabolic profiling of urine and serum. Toxicol Appl Pharmacol 2008;227:417-29. [22] Bylesjo M, Rantalainen M, Cloarec O, Nicholson JK, Holmes E, Trygg J. OPLS discriminant analysis: combining the strengths of PLS-DA and SIMCA classification. J Chemometrics 2006;20: 341-51. [23] Saiki S, Sato T, Kohzuki M, Kamimoto M, Yosida T. Changes in hypoxanthine levels by exercise in obese subjects. Metabolism 2001; 50:627-30.
Nutrition Research 30 (2010) 455 – 461 www.nrjournal.com
Hypoxanthine levels in human urine serve as a screening indicator for the plasma total cholesterol and low-density lipoprotein modulation activities of fermented red pepper paste Yujin Kim a,1 , Youn-Je Park b,1 , Seung-Ok Yang a , So-Hyun Kim a , Sun-Hee Hyun a , Sayeon Cho a , Young-Suk Kim c , Dae Young Kwon d , Youn-Soo Cha e , Soowan Chae f , Hyung-Kyoon Choi a,⁎ a
College of Pharmacy and Research Center for Biomolecules and Biosystems (WCU), Chung-Ang University, Seoul 156-756, Republic of Korea b Department of Applied Bioscience, CHA University, Seoul 135-081, Republic of Korea c Department of Food Science and Technology, Ehwa Womans University, Seoul 120-750, Republic of Korea d Korea Food Research Institute, Seongnam 463-746, Republic of Korea e Department of Food Science and Human Nutrition, Chonbuk National University, Jeonju 561-756, Republic of Korea f Clinical Trial Center for Functional Foods, Chonbuk National University, Jeonju 561-756, Republic of Korea Received 20 April 2010; revised 7 June 2010; accepted 23 June 2010
Abstract Fermented red pepper paste (FRPP) is one of the most well-known traditional foods in Korea. The effects of FRPP in experimental animals and adipocytes have been previously reported. However, the biochemical effects have not yet been validated in humans with various genetic backgrounds and environmental factors. In this study, 28 female volunteers (body mass index, more than 23 kg/m2) aged 19 to 60 years were treated with either FRPP or a placebo for 12 weeks. Marked cholesterol modulation was observed in the FRPP-treated group compared with the placebo group. Although the baseline (pretreatment) total cholesterol and low-density lipoprotein levels and body mass index of the volunteers did not differ significantly between the placebo- and FRPP-treated groups, FRPP caused a modulation of cholesterol levels not seen in the placebo group, causing either no variation or a decrease in low-density lipoprotein and total cholesterol levels. Thus, urinary metabolomic profiling of pretreatment samples was carried out in these 2 FRPP-treated groups using 1H-nuclear magnetic resonance–based metabolomic techniques. These 2 groups, with their opposing cholesterol-modulation tendencies, could be clearly differentiated by orthogonal projections to latent structures–discriminant analysis-derived score plots. In addition, their levels of hypoxanthine differed markedly. We propose that urinary hypoxanthine levels can be used as a screening biomarker to predict the efficacy of the cholesterol-modulating activity of FRPP in human subjects. © 2010 Elsevier Inc. All rights reserved. Keywords: Abbreviations:
Capsicum annuum; Fermented red pepper paste; Cholesterol-modulation activity; Metabolomics; 1H-NMR; Women BMI, body mass index; FRPP, fermented red pepper paste; LDL, low-density lipoprotein; OPLS-DA, orthogonal projections of latent structure–discriminant analysis; TC, total cholesterol.
1. Introduction ⁎ Corresponding author. Fax: +82 2 812 3921. E-mail address: [email protected] (H.-K. Choi). 1 These authors contributed equally to this work. 0271-5317/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.nutres.2010.06.014
Worldwide, the obese population is increasing, and obesity has become a serious socioeconomic problem because it can result in hypertension, diabetes, osteoarthritis, stroke, and cancer [1,2]. Previous studies have found that
456
Y. Kim et al. / Nutrition Research 30 (2010) 455–461
obesity exerts strong effects on lipoprotein metabolism, leading to elevated levels of low-density lipoprotein (LDL) and total cholesterol (TC) [3]. Fermented red pepper paste (FRPP; Gochujang, CODEX standard number 294R-2009) is one of the most well-known traditional foods in Korea [4]; it is made with glutinous rice, Meju (fermented soybean blocks), salt, and red pepper (Capsicum annuum L.) powder and is fermented for several months or longer. Its unique taste, flavor, and color are due to the actions of microorganisms such as bacteria, yeasts, and molds during the fermentation process [5,6]. Several studies have found antiobesity effects of FRPP in animal models and cell cultures. Rhee et al [7] investigated the reduction in lipid levels in adipose tissue and serum after FRPP feeding, Choo [8] discovered that FRPP increases energy expenditure and results in a reduction in body fat gain, and Ahn et al [9] added FRPP to 3T3-L1 adipocytes and found that it decreases fat accumulation by modulating adipogenesis and lipolysis. However, these studies have focused mainly on the analysis of data from other animals or cells, rather than on humans. Humans have diverse metabolic regulation systems as a result of genetic and environmental differences; thus, the effects of nutritional interventions can be specific to the individual [10]. Metabolomic approaches have been applied to toxicity screening, drug metabolism, and identifying the biochemical effects of several foods [11-13]. In those reports, urine was found to be a useful sample for metabolomics studies [14]. Kim et al [15] categorized rats into 4 groups—normal-diet low gainers, normal-diet high gainers, high-fat-diet low gainers, and high-fat-diet high-gainers—and studied the metabolomic differences in their urine, whereas Akira et al [16] characterized the urinary metabolite profiles of young, spontaneously hypertensive rats in the early stages of hypertension, and Williams et al [17] determined the metabolic fingerprints of male and female obese rats using 1H-nuclear magnetic resonance (NMR)–based metabolomic techniques. In addition, various examples of the successful application of metabolomics in studies of the metabolic fingerprinting of other biologic fluids (eg, serum) and tissues, combined with the use of multivariate statistical analysis [18,19], exist. The main hypothesis of this study was that FRPP might have different antiobesity and cholesterol-modulating effects in individuals, and the effects could be predicted using urinary metabolic profiling. In the current study, the antiobesity and cholesterol-modulating effects of FRPP were assessed in humans. The FRPP-treated cohort was subsequently divided into 2 groups according to whether the FRPP caused a modulation in their TC and LDL levels, and their pretreatment (baseline) urine samples were subjected to metabolic profiling by 1 H-NMR–based metabolomic techniques. The objectives of the study were to assess antiobesity and cholesterol-modulating effects of FRPP and to determine whether urinary metabolites can be used as screening indicators for selecting patients for whom FRPP will be efficacious as an antiobesity and cholesterol-modulating treatment.
2. Methods and materials 2.1. Preparation of FRPP samples Fermented red pepper paste was prepared following a traditional fermentation process starting with the preparation of Meju added to glutinous rice. Soybeans (Glycine max L.) containing 10.5% water, 5.4% ash, 17.9% fat, 38.7% crude protein, and 26.2% carbohydrate were used. The soybeans were digested for 9 hours at 20°C, dried in baskets, autoclaved at 121°C for 30 minutes, cooled until the temperature decreased to 30°C, and then ground into fine powder. Glutinous rice was digested for 12 hours, autoclaved, and cooled in the same conditions. The obtained soybean powder and glutinous rice were mixed in a 6:4 ratio, 0.5% (wt/wt) koji rice prepared with Aspergilus sojae, and 0.5% (wt/wt) Bacillus subtilis culture broth was added, and then these were molded in 18 × 10 × 3-cm boxes to prepare Meju. The Meju was then stored in a room at 35°C with 90% relative humidity for 3 days. The fermented Meju was dried for 3 to 4 days and crushed with a flat roller (Daeyul Industry Co., LTD, Daejun, Republic of Korea) for the preparation of FRPP. Fermented red pepper paste contains 18.0% glutinous rice, 6.0% Meju powder, 4.0% dried barley sprouts, 10.5% salt, 17.5% FRPP powder, and 44.0% distilled water according to the standardized protocols of FRPP preparation by Soonchang Moonokrae Foods (Soonchang, Republic of Korea). Fermented red pepper paste was distributed into eight 100-kg jars and was stored for fermentation and maturation for 3 months in the sauce manufacturing facility of Soonchang Moonokrae Foods. Fermented red pepper paste pills were prepared by mixing the glutinous rice powder, FRPP powder, Meju powder, dried barley sprouts, salt, soy sauce, black-eyed pea powder, and cocoa powder
Table 1 Composition of FRPP pills and placebo pills Components
FRPP
Placebo
Contents (g) Contents (g) Main components
Enhancer
Additive Total weight Freeze-dried weight Energy (kJ)
FRPP powder Glutinous rice powder Dried barley sprouts Salt Meju powder Soy sauce Black-eyed pea powder Cocoa powder Flour Fat powder Honey Pepper scent Spicy powder Caramel pigment Red pigment
11.9 10.9 5.2 5.2 4.7 2.1 7.5 2.5 – – – – – – – 50.0 32.0 477.3
– – – 1.1 1.4 6.7 9.3 1.4 13.1 6.3 1.4 0.3 0.1 0.1 1.5 41.3 32.0 473.1
Y. Kim et al. / Nutrition Research 30 (2010) 455–461
in the mixing ratios shown in Table 1. The pills for the placebo group were also prepared using the formulation shown in Table 1, and these contained spicy-flavored powder instead of FRPP, pepper scent, Meju powder, soy sauce, fat powder, flour, black-eyed pea powder, cocoa powder, salt, honey, caramel pigment, and red pigment. All samples were stored at −70°C until administration to the volunteers after thawing in room temperature. 2.2. Study subjects Twenty-eight female volunteers with body mass index (BMI) more than 23 kg/ m2 (19-60 years of age) were recruited and randomly divided into 2 groups as they were treated with FRPP pills or placebo pills. The energy content of both pills were adjusted to 477.3, 473.1 kJ/g, respectively. They were fed either a FRPP pill or a placebo pill 3 times a day for 12 weeks. All 28 volunteers had previously consented to the treatment. The protocol was approved by the Functional Foods Institutional Review Board of Chonbuk National University Hospital. 2.3. Plasma and urine sample preparation and cholesterol analysis Blood was drawn from the volunteers after a 12-hour fast. Plasma samples were then prepared by the Chonbuk National University Hospital. The concentrations of LDL and TC in plasma were analyzed on a Hitachi 7600-110 analyzer (Hitachi High-Technologies Corporation, Tokyo, Japan) by standard methods at the biochemical laboratory of Chonbuk National University Hospital. Twenty-four-hour urine samples were collected from the volunteers at Chonbuk National University. Ten milliliters of the samples was then transferred into a 15-mL conical tube with 1% sodium azide and were stored at −70°C until assay. 2.4. NMR measurements Pretreatment (baseline) urine samples were thawed in a water bath at 40°C and then vortexed for 5 seconds; 0.3 mL of each urine sample was transferred to an Eppendorf tube. Then, 0.1 mol D2O [containing 0.05% 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid, sodium salt, as an internal standard for D2O] was made by adding 1.232 g of KH2PO4 as a buffering agent to 100 g of D2O. The pH was adjusted to 6.0 (as measured using a pH meter; model 720P; Istek, Seoul, Korea) by adding 1 N NaOD. A 0.2-mL aliquot of D2O was added to a 0.3-mL aliquot of the urine sample and vortexed for several seconds to ensure thorough mixing. The mixture thus prepared was pipetted into a 5-mm NMR tube (Norell, Landisville, NJ) [20]. Samples were analyzed at 600.13 MHz with an NMR spectrometer (Avance 600 FT-NMR, Bruker, Germany) at a temperature of 298 K. To suppress the residual water signal, a noesygpprld pulse sequence was used. With a spectral width of 10 775.9, a total of 64 transients were collected into 64 data
457
points. The acquisition time per scan of 3.04 seconds was used with a relaxation delay of 1 second. The exponential linebroadening function of 0.30 Hz was applied to free induction decays before Fourier transform. 2.5. Statistical analyses The spectral region from δ = 0.00 to 10.00 was segmented into 0.04-ppm using Amix software (Version 3.7, Bruker Biospin). In aqueous extracts, the region from δ = 4.60 to 4.90 was excluded due to the residual signal of water, along with the urea region from δ = 5.60 to 6.00 [21]. All data were mean centered, with scaling to unit variance as a preprocessing method, and orthogonal projections of latent structure– discriminant analysis (OPLS-DA) were conducted using the SIMCA-P software (version 12.0; Umetrics, Umeå, Sweden). An analysis of variance and independent t test were performed using SPSS software (version 17; SPSS Inc, Chicago, Ill); the power analysis of the t test was conducted using Power and Precision software (version 3.2.0; Biostat, Englewood, NJ). The mean differences, standard deviations, level of significance (P = .05), and sample size were used for the calculation of the power value. 3. Results and discussion 3.1. Biochemical parameters As indicated in Table 2, the BMI did not differ significantly between pretreatment and posttreatment conditions in both the FRPP- and placebo-treated groups. Thus, FRPP had no significant effect on body weight. However, the cholesterol-modulating effect was markedly greater in the FRPP-treated group than in the placebotreated group. The TC and LDL levels of the participants treated with the placebo pills did not differ significantly during the 12-week experimental period. However, the subjects in the FRPP-treated group exhibited either a significant decrease of more than 10.7% and 15.7% in TC and LDL or no significant variation in TC and LDL after the 12-week experimental period. In other words, FRPP exerted 2 different effects in our human cohort, whose baseline (pretreatment) TC/LDL levels and BMIs did not significantly different from each other. This effect was investigated further by subjecting the pretreatment urine samples of the FRPP-treated group to metabolomic profiling using 1 H-NMR-based metabolomic techniques. 3.2. 1H-NMR spectral analysis Fig. 1 shows the assignment of each peak in the representative 1 H-NMR spectra from urine sample of pretreatment of FRPP. The following signals were assigned based on comparisons with the chemical shifts of standard compounds using the Chenomx NMR suite software (version 5.1; Chenomx, Inc, Edmonton, Canada): leucine at δ = 0.94 (d, J = 6.78 Hz); isoleucine at δ = 0.94 (d, J = 6.78 Hz); valine at δ = 0.98 (d, J = 6.9 Hz) and 1.02 (d, J = 6.24 Hz);
458
Y. Kim et al. / Nutrition Research 30 (2010) 455–461
Table 2 Pretreatment and posttreatment plasma TC and LDL levels and BMIs in the FRPP- and placebo-treated groups FRPP (n = 14) TC/LDL-decreased group (n = 8)
Placebo (n = 14) TC/LDL-unvaried group (n = 6)
Pretreatment
Posttreatment
Pretreatment
Posttreatment
Pretreatment
Posttreatment
TC (mg/dL)
197.0 ± 16.8
207.0 ± 32.8
224.2 ± 23.8
202.2 ± 27.2
197.7 ± 25.9
LDL(mg/dL)
134.3 ± 17.3
175.0 ± 14.2 ⁎ 10.7% a 112.8 ± 14.7 ⁎ 15.7% a 26.4 ± 2.3
134.8 ± 25.3
149.5 ± 24.8
126.3 ± 27.8
122.9 ± 27.8
26.3 ± 2.6
26.5 ± 2.2
27.3 ± 3.1
26.4 ± 3.4
BMI
26.4 ± 2.4
Values are means ± SD (n = 6 or n = 8). a TC and LDL decrease ratio (%) = (posttreatment TC/LDL − pretreatment TC/LDL) × 100/(pretreatment TC/LDL). ⁎ Represent significant differences between predose group and postdose group at P b .05, and the power values regarding the t test results of TC and LDL in the TC/LDL-decreased group were 0.75 and 0.70, respectively.
methylsuccinate at δ = 1.06 (d, J = 7.1 Hz); 2-methylglutarate at δ = 1.06 (d, J = 7.1 Hz); lactate at δ = 1.34 (d, J = 7.3 Hz); alanine at δ = 1.46 (d, J = 7.9 Hz); acetate at δ = 1.94 (s); citrate at δ = 2.54 (d, J = 14.6 Hz) and δ = 2.74 (d, J = 17.7 Hz); creatine at δ = 3.02 (s) and δ = 3.90 (s); creatinine at δ = 3.06 (s) and δ = 4.10; taurine at δ = 3.42 (t, J = 6.0 Hz); malonate at δ = 3.34 (s); glycine at δ = 3.54 (s); mannitol at δ = 3.66 (dd, J1 = 6.5 Hz, J2 = 6.0 Hz), δ = 3.74 (m), δ = 3.78 (d, J = 9.4 Hz), and δ = 3.86 (dd, J1 = 2.8 Hz, J2 = 3.4 Hz);
τ-methylhistidine at δ = 3.86 (s), δ = 7.34 (s), and δ = 8.34 (s); histidine at δ = 7.34 (s) and δ = 8.34 (s); hippurate at δ = 7.54 (t, J = 7.2 Hz), δ = 7.62 (t, J = 7.1 Hz), and δ = 7.82 (d, J = 7.9 Hz); hypoxanthine at δ = 8.18 (s); and formate at δ = 8.42 (s). 3.3. OPLS-DA of predose samples of the FRPP treatment Orthogonal projection to latent structures–discriminant analysis is a preprocessing method that provides better
Fig. 1. Representative 1H-NMR spectra of FRPP pretreatment urine samples.
Y. Kim et al. / Nutrition Research 30 (2010) 455–461
visualization and interpretation of the model than partial least squares–discriminant analysis. The predictive component, which stands for the variation between the groups, is predicted based on the known orthogonal component, which stands for the variation within the groups [22]. Therefore, OPLS-DA is a useful tool for discriminating between compounds that contribute to the separation between groups of interest. Orthogonal projections of latent structure–discriminant analysis was applied to investigate the metabolic differences in the pretreatment urine between those in the FRPP-treated group that exhibited no variation in TC/LDL (TC/LDLunvaried group) and those that exhibited a decrease in TC/ LDL (TC/LDL-decreased group). Fig. 2A shows the OPLS-
459
DA–derived score plots, which reveal that the groups can be clearly separated. This means that the effect of FRPP differed according to the individual's pretreatment metabolic conditions. To identify the compounds contributing to the separation of each sample in the OPLS-DA model, a loading plot analysis was performed (Fig. 2B). The locations of loading plots correspond to those of score plots. Therefore, the peaks in the positive position of the loading plots are dominant in the samples, which are located in the positive position along the x axis of the score plots in this study. In the pretreatment condition, the levels of creatine, taurine, malonate, τ-methylhistidine, histidine, and hippurate were higher in the TC/LDL-decreased group and are located
Fig. 2. A, OPLS-DA-derived score plot of the pretreatment samples (b, pretreatment TC/LDL-unvaried group; 0, pretreatment TC/LDL-decreased group). B, OPLS-DA-derived loading plot of the pretreatment samples of the TC/LDL-unvaried and TC/LDL-decreased groups. 1, leucine; 2, isoleucine; 3, valine; 4, methylsuccinate; 5, 2-methylglutarate; 6, lactate; 7, alanine; 8, acetate; 9, citrate; 10, creatine; 11, taurine; 12, malonate; 13, glycine; 14, mannitol; 15, τ-methylhistidine; 16, histidine; 17, hippurate; 18, hypoxanthine; 19, formate.
460
Y. Kim et al. / Nutrition Research 30 (2010) 455–461
and to reveal potential biomarkers for the personalized treatment of various diseases, including obesity. Acknowledgment This study was made possible by a National Research Foundation of Korea grant, funded by the Korean government (MEST; 2009-0065538) and by the World Class University program through the National Research Foundation of Korea, funded by the Ministry of Education, Science and Technology (grant number: R33-10029). References Fig. 3. Relative intensities of hypoxanthine (normalized to the creatinine intensity) in pretreatment urine samples from the 2 FRPP-treated groups (TC/LDL-unvaried in one and decreased in the other). An independent t test was performed to assess any statistically significant differences between the 2 groups. The error bars represent the standard deviation (n = 6 or 8). *Significantly different at P b .05; the power value was 0.79 (obtained by power analysis).
in the negative position of the loading plot, whereas relatively higher levels of leucine, isoleucine, valine, methylsuccinate, 2-methylglutarate, lactate, alanine, acetate, citrate, glycine, mannitol, hypoxanthine, and formate were observed in the TC/LDL-unvaried group, which are located in the positive position of the loading plot. An independent t test was performed to validate the significance of differences between the relative intensities of the selected compounds between the 2 groups (data not shown). With the exception of hypoxanthine, there were no significant differences in the relative levels of these selected compounds. Therefore, only the relative intensities of hypoxanthine are shown in Fig. 3; these were significantly higher in the TC/LDL-unvaried group than in the TC/LDLdecreased group. It was previously reported that serum hypoxanthine levels tend to increase in obese and oxidatively stressed conditions [23]. The variable effects of FRPP between the TC/LDLunvaried and TC/LDL-decreased groups may be attributable to differences in lifestyle, diet, physical constitution, and/or genetic background among the subjects. In the pretreatment condition, the urinary level of hypoxanthine was significantly lower in the TC/LDL-decreased group than in the TC/LDL-unvaried group. Therefore, we accept the hypothesis for this study that the urinary metabolite can be used as a screening indicator for cholesterolmodulating effects of FRPP. Further study is needed to determine the correlation between pretreatment urinary hypoxanthine levels and other environmental or genetic factors among individuals. We propose that hypoxanthine could be used as a potential biomarker for the screening of subjects to ensure the cholesterol-modulation activity of FRPP. Thus, our findings suggest that this methodology can be applied to assess the bioactivities of other natural resources in humans
[1] Kopelman PG. Obesity as a medical problem. Nature 2000;404: 635-43. [2] Grundy SH. Multi-factorial causation of obesity: implication for prevention. Am J Clin Nutr 1998;67:563-72. [3] Krauss R, Winston M, Fletcher B, Grundy S. Obesity: impact on cardiovascular disease. Circulation 1998;98:1472-6. [4] http://www.codexalimentarius.net/web/standard_list.do?lang=en, (accessed February 15, 2010). [5] Lee TS, Chun MS Choi JY, Noh BS. Changes of free sugars and free amino acids in kochujang with different mashing method. Food Biotechnol 1993;2:102-8. [6] Chun MS, Lee TS, Noh BS. The changes in capsaicin, dihydrocapsaicin and cansanthin in kochujang with different mashing methods. Food Biotechnol 1994;3:104-8. [7] Rhee SH, Kong KR, Jung KO, Park KY. Decreasing effect of kochujang on body weight and lipid levels of adipose tissues and serum in rats fed a high-fat diet. J Food Sci Nutr 2003;32:882-6. [8] Choo JJ. Anti-obesity effects of kochujang in rats fed on a high-fat diet. Korean J Nutr 2000;33:787-93. [9] Ahn IS, Do MS, Kim SO, Jung HS, Kim YI, Kim HJ, Park KY. Antiobesity effect of kochujang (Korean fermented red pepper paste) extract in 3T3-L1 adipocytes. J Med Food 2006;9:15-21. [10] German JB, Roberts MA, Watkins SM. Personal metabolomics as a next generation nutritional assessment. J Nutr 2003;133:4260-6. [11] Nicholson JK, Connelly J, Lindon JC, Holmes E. Metabonomics: a platform for studying drug toxicity and gene function. Nat Rev Drug Discov 2002;1:153-61. [12] Bales JR, Sadler PJ, Nicholson JK, Timbrell JA. Urinary excretion of acetaminophen and its metabolites as studied by proton NMR spectroscopy. Clin Chem 1984;30:1631-6. [13] Solanky KS, Bailey NJ, Beckwith-Hall BM, Bingham S, Davis A, Holmes E, et al. Biofluid 1H NMR-based metabonomic techniques in nutrition research-metabolic effects of dietary isoflavones in humans. J Nutr Biochem 2005;16:236-44. [14] Bollard ME, Holmes E, Lindon JC, Mitchell SC, Branstetter D, Zhang W, et al. Investigations into the biochemical changes due to diurnal variation and estrus cycle in female rats using high resolution 1H NMR spectroscopy of urine and pattern recognition. Anal Biochem 2001; 295:194-202. [15] Kim SH, Yang SO, Kim HS, Kim Y, Park T, Choi HK. 1H-nuclear magnetic resonance spectroscopy-based metabolic assessment in a rat model of obesity induced by a high-fat diet. Anal Bioanal Chem 2009; 395:1117-24. [16] Akira K, Masu S, Imachi M, Mitome H, Hashimoto M, Hashimoto T. 1 H NMR-based metabonomic analysis of urine from young spontaneously hypertensive rats. J Pharm Biomed Anal 2008;46:550-6. [17] Williams R, Lenz E, Evans J, Wilson I, Granger J, Plumb R, et al. A combined 1H NMR and HPLC-MS-based metabonomic study of urine from obese (fa/fa) Zucker and normal Wistar-derived rats. J Pharm Biomed Anal 2005;38:465-71.
Y. Kim et al. / Nutrition Research 30 (2010) 455–461 [18] Lindon JC, Nicholson JK, Holmes E, Antti H, Bollard ME, Keun H, et al. Contemporary issues in toxicology the role of metabonomics in toxicology and its evaluation by the COMET project. Toxicol Appl Pharmacol 2003;187:137-46. [19] Nicholson JK, Wilson ID. High resolution proton magnetic resonance spectroscopy of biological fluids. Prog Nucl Magn Reson Spectrosc 1989;21:449-501. [20] Webb-Robertson BJM, Lowry DF, Jarman KH, Harbo SJ, Meng QR, Fuciarelli AF, et al. A study of spectral integration and normalization in NMR-based metabonomic analyses. J Pharm Biomed Anal 2005;39:830-6.
461
[21] Wei L, Liao P, Wu H, Li X, Pei F, Li W, et al. Toxicological effects of cinnabar in rats by NMR-based metabolic profiling of urine and serum. Toxicol Appl Pharmacol 2008;227:417-29. [22] Bylesjo M, Rantalainen M, Cloarec O, Nicholson JK, Holmes E, Trygg J. OPLS discriminant analysis: combining the strengths of PLS-DA and SIMCA classification. J Chemometrics 2006;20: 341-51. [23] Saiki S, Sato T, Kohzuki M, Kamimoto M, Yosida T. Changes in hypoxanthine levels by exercise in obese subjects. Metabolism 2001; 50:627-30.