當前位置:首頁 > 醫源資料庫 > 在線期刊 > 美國臨床營養學雜志 > 2007年86卷第6期 > Influence of acute phytochemical intake on human urinary metabolomic profiles

Influence of acute phytochemical intake on human urinary metabolomic profiles

來源:《美國臨床營養學雜志》 作者:Marianne C Walsh 2008-12-28
336*280 ads

摘要: Marianne C Walsh, Lorraine Brennan, Estelle Pujos-Guillot, Jean-Louis Sé。bé。dio, Augustin Scalbert, Ailí。 ABSTRACTBackground: Diversity in dietary intake contributes to variation in human metabolomic profiles and artifacts from acute dietary intake can affect metabolomics data。...


Marianne C Walsh, Lorraine Brennan, Estelle Pujos-Guillot, Jean-Louis Sébédio, Augustin Scalbert, Ailís Fagan, Desmond G Higgins and Michael J Gibney

1 From the Centre for Food and Health (MCW and MJG), the School of Biomolecular and Biomedical Sciences, Conway Institute (LB), and the School of Medicine and Medical Science, Conway Institute (AF and DGH), University College Dublin, Ireland, and the Unite de Nutrition Humaine, Institute National de la Recherche Agronomique, St-Genès-Champanelle, France (EPG, JLS, and AS)

2 Supported by The Irish Research Council for Science, Engineering and Technology.

3 Reprints not available. Address correspondence to MC Walsh, Room 3.02 Nutrition Unit, Centre for Food and Health, School of Agriculture, Food Science and Veterinary Medicine, University College Dublin, Dublin 4, Ireland. E-mail: marianne.walsh{at}ucd.ie.


ABSTRACT  
Background: Diversity in dietary intake contributes to variation in human metabolomic profiles and artifacts from acute dietary intake can affect metabolomics data.

Objective: We investigated the role of dietary phytochemicals on shaping human urinary metabolomic profiles.

Design: First void urine samples were collected from 21 healthy volunteers (12 women, 9 men) following their normal diet (ND), a 2-d low-phytochemical diet (LPD), or a 2-d standard phytochemical diet (SPD). Nutrient intake was assessed during the study. Urine samples were analyzed by using 1H nuclear magnetic resonance spectroscopy (1H NMR) and mass spectrometry (MS), which was followed by multivariate data analysis.

Results: Macronutrient intake did not change throughout the study. Partial least-squares-discriminant analysis indicated a clear distinction between the LPD samples and the ND and SPD samples, relating to creatinine and methylhistidine excretion after the LPD and hippurate excretion after the ND and SPD. The predictive power of the LPD versus the ND model was 74 ± 3% and 82 ± 6% with the 1H NMR and MS data sets, respectively. The predictive power of the LPD versus the SPD model was 83 ± 8% and 69 ± 4% for the 1H NMR and MS data sets respectively. A cross platform comparison of both data sets by co-inertia analysis showed a similar distinction between the LPD and SPD.

Conclusions: Acute changes in urinary metabolomic profiles occur after the consumption of dietary phytochemicals. Dietary restrictions in the 24 h before sample collection may reduce diversity in phytochemical intakes and therefore reduce variation and improve data interpretation in metabolomics studies using urine.

Key Words: Nutritional metabolomics • metabonomics • metabolic profiling • nutrigenomics


INTRODUCTION  
Metabolomics affords an efficient approach for assessing the overall metabolic profiles of tissues or biofluids and is increasingly being used within a range of scientific disciplines. One of the challenges in adopting a metabolomic approach, particularly in human nutrition research, is the ability to distinguish meaningful responses within metabolic profiles that are produced as a consequence of a specific stimulus rather than as an artifact of daily variation. In addition, when distinct metabolic profiles are identified that differentiate 2 sample populations, it is important to understand the contribution of any exogenous factors that may be unique to each group. These factors could lead to misinterpretation of the true differences in endogenous metabolism between the groups. Genetic and environmental variation in human populations contributes to considerable diversity among human metabolomes (1-6), and diet is one of the key environmental factors because of its dynamic input to metabolism.

The role of diet in shaping metabolic profiles is not fully elucidated, but it is clear that diet will have both an acute and chronic effect. Understanding the chronic effects of diet is the most relevant in terms of nutrition research, but in terms of the interpretation of nutritional metabolomics data, an understanding of acute dietary effects is also important. Foods in the human diet are not merely a source of nutrients and are abundant in many other compounds that contribute to their taste, color, aroma, texture, and shelf-life. Plant-derived foods and drinks are a rich source of such compounds, namely phytochemicals, and 5000–10 000 are present in human food. It is estimated that the average diet corresponds to a daily dose of 1.5 g phytochemicals (7).

The role of these phytochemicals in shaping the metabolic profiles obtained by techniques based on nuclear magnetic resonance (NMR) and mass spectrometry (MS) is not yet clear. Previous studies that investigated the metabolic fate of polyphenols have shown their appearance in plasma as early as 1 h and in urine within 24 h after ingestion (8-10). Hence, to further our understanding of the role that diet plays in shaping metabolic profiles, it is important to consider these acute dietary influences that contribute to normal variation in metabolic profiles. This is particularly relevant for urine because it accumulates many end products of metabolism. Urine is a complex biofluid with an extensively varying composition, whereas the composition of other biofluids can be controlled more tightly by varying the excretion of many metabolites into urine. In a previous study we reported that dietary standardization for 24 h before sample collection reduced the extent of variation in urinary metabolic profiles, but not in fasting plasma or salivary profiles (6).

In an attempt to further our understanding of the effect that acute dietary intakes have on NMR and MS profiles, the current study was designed to investigate the role of phytochemicals in shaping urinary metabolic profiles. In the long term, understanding the influence of acute dietary intakes will lead to an improved interpretation of dietary intervention outcomes.


SUBJECTS AND METHODS  
Study design
Ethical approval was received from the Faculty of Health Sciences Ethics Committee, Trinity College Dublin, in accordance with the Declaration of Helsinki. Twenty-one healthy, free-living subjects were recruited from the Dublin metropolitan area, and each participant provided written informed consent. Exclusion criteria were a body mass index (in kg/m2) <18.5 or >30.0, being <18 or >35 y of age, use of prescribed medication (oral contraceptive use was permitted), and having had a urinary tract infection within 1 mo of commencing the study.

The study duration was 6 d (days 0–5). Samples of first void urine were collected chilled on study days 1, 3, and 5. Dietary intake was recorded on days 0–4. Subjects were instructed to follow their normal diet (ND) on day 0, with the result that the day 1 urine samples represented normal urinary metabolic profiles. On days 1–2, subjects followed a low-phytochemical diet (LPD), with the result that the day 3 urine samples represented urinary metabolic profiles in the absence of phytochemicals. On days 3–4, subjects continued the LPD with the addition of fruit and vegetable drinks, with the result that the day 5 urine samples represented urinary profiles after a standard fruit and vegetable intake (ie, standardized phytochemical intake). Two days was an adequate time for both the phytochemical washout and dosage period, given that the plasma elimination half-life for major phytochemicals is between 1.1 and 28.1 h (10). For the LPD the subjects were provided with a list of allowed and forbidden foods. There were no restrictions on the quantities of food eaten. All fruit and vegetables were forbidden (this included all products that contain fruit and vegetables, such as jams, sauces, soup, tea, coffee, and chocolate). The allowed list included meat, dairy products, and some low-pigmented plant products, such as white breads, non-whole-grain breakfast cereals, potatoes, white rice, and white pasta. For the standardized phytochemical diet (SPD), the fruit and vegetable source was 4 x 100 mL apple, carrot, and strawberry drinks (Knorr Vie, produced by Unilever; Internet: www.knorr-vie.com).

Nutrient intake
Nutrient intakes were assessed with the use of 5-d weighed food records. The participants were given detailed instructions regarding the completion of the records and were tutored on the estimation of food portion sizes for instances when the food scales could not be used. The records were reviewed after each sample collection, and clarification of food portions and preparation was made. Total energy, protein, carbohydrate, fat, fiber, vitamin C, and carotene intakes were estimated by using WISP© (Weighed Intake Software Program; Tinuviel Software, Anglesey, United Kingdom). The WISP database of food composition included the manufacturer's nutritional information for the standard drinks.

Urine collection
First void midstream urine samples were collected on the mornings of days 1, 3, and 5. The subjects were given insulated ice packs in which they were asked to store the samples immediately until they were received by the study investigator. On arrival at the laboratory, the samples were centrifuged at 2500 x g for 10 min at 4 °C to remove any solid debris. Fractions (500 µL) of the urine supernatants were then stored at –80 °C until 1H NMR analysis.

NMR spectroscopy
For NMR spectroscopy the urine samples were buffered with a phosphate buffer (0.2 mol KH2PO4/L, 0.8 mol KH2PO4/L). To each 350 µL urine, 180 µL phosphate buffer (pH 7.4), TSP (sodium trimethylsilyl [2,2,3,3-2H4] proprionate), and 10% D2O were added. NMR spectra were acquired at 298 K on a 500-MHz DRX NMR spectrometer (Bruker Biospin, Karlsruhe, Germany) using a noesypresat pulse sequence. Spectra were acquired with 32-k data points and 128 scans over a spectral width of 8 kHz. Water suppression was achieved during the relaxation delay (2.5 s) and the mixing time (100 ms). All 1H NMR spectra were referenced to TSP at 0.0 ppm and processed manually with the Bruker software with the use of a line broadening of 0.2 Hz. All spectra were baseline corrected. The spectra were then reduced by integrating into bins across spectral regions of 0.02 ppm with an AMIX (Bruker Biospin). The water region (4.2–6.0 ppm) was excluded. The data were normalized to the sum of the spectral integral to account for differences between urinary concentrations.

Mass spectrometry
The urine samples (500 µL) were defrosted at room temperature, centrifuged at 7000 x g for 5 min at 4 °C, and then diluted 4-fold with distilled water. Chromatography was performed with a Waters Alliance 2695 HPLC system (Waters Corporation, Manchester, United Kingdom). The HPLC system was coupled to a Waters Qtof-Micro equipped with an electrospray source and a lockmass sprayer. The source temperature was set to 120 °C with a cone gas flow of 50 L/h, a desolvation temperature of 300 °C, and a nebulization gas flow of 400 L/h. The capillary voltage was set at 3000 V and the cone voltage to 30 V. The mass spectrometric data were collected in continuum full-scan mode with a mass-to-charge ratio (m/z) of 100–1000 from 0 to 10 min, in positive mode. All analyses were acquired by using the lockspray with a frequency of 5 s to ensure accuracy. Leucine-enkephalin was used as the lock mass ([M+H]+ m/z 556.2771) at a concentration of 0.5 ng/µL (in MeOH/water, 50/50 by vol with 0.1% formic acid).

To avoid possible differences between sample batches, a Latin square was carried out to obtain a randomized list of samples for analysis. The 2.1 x 150 mm SymmetryShieldRP18 5 µm column was injected with 10 µL diluted urine at 30 °C. Mobile phase components were A = 1% formic acid and B = acetonitrile with 1% formic acid. The column was eluted with a gradient of 0–20% B over 0–4 min, followed by an increase from 20% to 95% over 4–8 min. The mobile phase was then held at this composition for 1 min and then returned to 100% A at 9 min for 5 min reequilibration. The flow rate was set to 300 µL/min.

The raw data were transformed to centroid mode and mass corrected before being analyzed with MarkerLynx Applications Manager v1.0. The liquid chromatography–MS data were peak-detected and noise-reduced for both the liquid chromatography and MS components. Each peak in the resulting 3-dimensional data set was represented by retention time m/z and its ion intensity in each sample. The matrix obtained was then exported for statistical analysis.

Statistical analysis
The nutrient intake data were analyzed by using SPSS for WINDOWS (version 12: SPSS Inc, Chicago, IL). One-factor analysis of variance (ANOVA) was used to assess the differences between the ND, LPD, and SPD. Repeated-measures ANOVA was used to check for diet x sex interactions. Differences were considered significant at P < 0.05, and Tamhane's T2 post hoc tests were carried out to assess the significant differences indicated by the ANOVA results.

Multivariate data analyses were applied to both the 1H NMR and MS data by using Simca-P+ software (version 10.0; Umetrics, Umeå, Sweden) and the R Statistics package with modules MADE4 (11) and ade4 (12). The data sets were mean centered and Pareto scaled (each variable was weighted according to 1/SD). Principal component analysis (PCA), an unsupervised pattern recognition technique, was performed initially to assess variation and expose any trends or outlying data. Partial least-squares-discriminant analysis (PLS-DA) was then performed to define the maximum separation between the LPD versus the ND or the standard SPD. The data were visualized by constructing principal component scores and loadings plots, where each point on the score plot represented an individual urine sample and each point on the loadings plot represented a single 1H NMR spectral region or MS reading. The PLS-DA models were cross-validated by randomly removing each one-third of the data to be used as a test data set, and a training data set was constructed with the remaining two-thirds of the data. The class of each sample in the test data set was then predicted based on the model built from the training set. The quality of all models was judged by the goodness-of-fit parameter (R2) and the predictive ability parameter (Q2), which is calculated by a 7-fold internal cross-validation of the data.

Co-inertia analysis (CIA) (13) was applied to the LPD and SPD samples from the 1H NMR and MS data. CIA is a multivariate statistical method used to identify patterns in parallel data sets. In brief, it first carries out a simple ordination such as PCA on each of the data sets. CIA then finds pairs of axes from the 2 data sets that have maximum covariance. The first few axes from the CIA are used to create 2 dimensional plots. On these plots, the relations between the sample data points and variables from both data sets are visible.


RESULTS  
Twenty-one healthy volunteers (12 women, 9 men) aged 20–34 y completed the study. The subjects' demographic characteristics ( ± SD) are shown in Table 1
View this table:
TABLE 1. Demographic characteristics1

 
Nutrient intake
The mean (±SD) nutrient intake data for all subjects for each of the diets are shown in Table 2. The LPD was followed on days 1 and 2, and the SPD was followed on days 3 and 4; therefore, the mean intakes for both diets are presented. Repeated-measures ANOVA showed no diet-by-sex interactions, but indicated that energy intake in men was significantly higher during the LPD (P < 0.05). No significant differences were found between the 3 diets for total energy, protein, carbohydrate, or fat intakes. One-factor ANOVA (P < 0.05) with post hoc tests found that fiber intake was significantly lower during the LPD than during the ND (P < 0.05) and the SPD (P < 0.001). In addition, one-factor ANOVA (P < 0.001) with post hoc tests found that vitamin C intake was significantly lower during the LPD than during the ND (P < 0.05) and the SPD (P < 0.001). Finally, one-factor ANOVA (P < 0.001) with post hoc tests found that carotene intake was significantly higher during the SPD than during both the ND (P < 0.001) and the LPD (P < 0.001) and that carotene intake was significantly higher with the ND than with the LPD (P < 0.05).


View this table:
TABLE 2. Nutrient intakes1

 
Multivariate data analysis of the 1H NMR urinary profiles
Initial PCA of the 1H NMR urinary data showed 5 outlying samples (positioned outside the Hotelling's T2 elipse on the score plot). NMR spectra of these outlying samples were inspected. Three of these outliers were subsequently removed from the data set, one because of intense signal intensities in the spectral regions 0.87–0.89, 1.06, 1.22, and 1.37–1.42 ppm, which corresponded to a drug metabolite (drug use was confirmed by the dietary records). Another 2 samples showed enhanced signal intensities in the spectral region 2.17–2.19 ppm, which corresponded to a high p-cresol glucuronide concentration (p-cresol glucuronide originates as p-cresol, a metabolic product of Clostridium difficile in the large intestine and is conjugated with glucuronide in the liver before urinary excretion). The remaining 2 outliers were retained because their spectra presented no unusual features. PCA was then repeated, and the score plot is shown in Figure 1A. The first 2 components accounted for 31% of variation in the data, and the samples from the LPD tended to cluster on the right side of the plot. To probe further the differences between the LPD and the ND, a PLS-DA model was constructed. The first 2 components accounted for 30% (R2X value) of the variation in the model and had a Q2 value of 42%. Interrogation of the loadings plot and the NMR spectra showed that the discrimination between the LPD and the ND samples was mainly dominated by a higher level of hippurate (spectral regions: 3.96–3.98, 7.54–7.56, 7.56–7.58, and 7.84–7.86 ppm) in the ND samples and a higher level of creatinine (spectral regions: 3.04–3.06 and 4.06–4.08 ppm) and methyl histidine (spectral region: 3.74–3.76 ppm) in the LPD samples. Validation of these models, as described in Subjects and Methods, indicated that 74 ± 3% of the samples were classified correctly.


View larger version (12K):
FIGURE 1.. A: Principal component analysis of nuclear magnetic resonance (NMR) urinary data. B: Partial least-squares-discriminant analysis (PLS-DA) of NMR urinary data. C: PLS-DA of mass spectrometry urinary data. , normal diet; , low-phytochemical diet; , standard phytochemical diet.

 
To investigate the effect of the controlled addition of phytochemicals to the diet, a PLS-DA model was constructed with the data from the LPD and the SPD (Figure 1B). The first 2 components of the model accounted for 29% (R2X value) of the variation in the data and had a Q2 value of 60%. Inspection of the loadings plot and the NMR spectra showed that the discrimination between the 2 sample groups was dominated by higher levels of hippurate after the SPD and higher levels of creatinine and methyl histidine after the LPD. Validation of this model indicated that 83 ± 8% of the samples were correctly classified.

The PCA and PLS-DA analyses were also carried out separately for men and women to ensure that sex effects did not influence the results. These analyses showed similar results, which indicated that sex did not affect the study findings.

Multivariate data analysis of the mass spectrometry urinary profiles
Two outlying samples were revealed after initial PCA of the MS urinary data (one of these was also an outlier in the NMR data). The PCA loadings plot and inspection of the individual outputs indicated that these samples had high peak intensities corresponding to masses 313.078 and 447.107, which relate to genistein acetate and genistein glucuronide, respectively. These metabolites result from soy isoflavones, but there was no evidence to confirm the consumption of soy isoflavones before sampling. These samples were subsequently removed from the data set, and PCA was repeated. The first 2 components of this model accounted for 24% of variation in the data, and, although there were no distinct clustering trends, the samples from the ND tended to locate in the bottom half of the plot. A PLS-DA model was constructed to assess any differences between the LPD samples and the ND samples. A 3-component model was generated, accounting for 28% (R2X value) of variation in the data and it had a Q2 value of 59%. Inspection of the PLS-DA loadings plot indicated that discrimination of the samples was mainly dominated by the ions with an m/z of 180.068, 105.028 (both corresponding to hippurate), 312.217, 197.07, and 169.036 (unidentified). These ions were associated with the ND samples, and validation of the model indicated that 82 ± 6% of the samples were classified correctly.

Another 3-component PLS-DA model was constructed to assess the differences between the LPD samples and the SPD samples (Figure 1C). This model accounted for 30% (R2X value) of the variation in the data, and the Q2 value was 44%. The corresponding loadings plot indicated that discrimination resulted from high peak intensities associated with the SPD samples, for m/z values of 180.068, 105.028 (both relating to hippurate), 413.045, 312.217, and 169.036 (unidentified). The model was validated, giving correct classification for 69 ± 4% of the samples.

Combined analysis of NMR and MS data
CIA was applied to the LPD and SPD samples of 1H NMR and MS data to visualize patterns and to identify metabolites with concentrations that changed in each group. It was carried out on 18 matched samples; 3 samples were removed from the analysis as they were considered outliers in either the 1H NMR or MS data or both. The initial ordinations showed that the sample distribution of 1H NMR split reasonably well. The MS data, however, did not separate so clearly (data not shown). CIA was applied to these initial ordinations. The combined sample distribution from the first 2 axes is shown in Figure 2A. Axis 1 and axis 2 from the CIA explain 46% and 20% of the variance, respectively. The base of each arrow represents one 1H NMR sample, and the tip represents the equivalent MS sample. The lengths of the arrows indicate how dissimilar the samples are. Although a perfect separation of the sample groups was not achieved, the SPD samples (in black) dominate the right side of the plot and the LPD samples (in gray) dominate the left side. The metabolites from the 1H NMR and MS data are plotted in Figures 2B and C. Metabolites that project in the same direction of both 1H NMR and MS plots have similar influence in both sets of data. For example, the top right quadrant of Figure 2B and Figure 2C are dominated by regions that correspond to hippurate. Because the top right quadrant is dominated in the sample plot by SPD, it can be interpreted that the SPD samples have a notable increase in hippurate compared with LPD.


View larger version (24K):
FIGURE 2.. A: Co-inertia analysis (CIA) of nuclear magnetic resonance (NMR) and mass spectrometry (MS) urinary data. Arrows are numbered with the subjects' identification number. The gray arrows represent samples from the low-phytochemical-diet (LPD) group, and the black arrows represent samples from the standard-phytochemical-diet (SPD) group. B: CIA metabolite plot for NMR urinary data. Each closed triangle represents an NMR bin. Hippurate metabolites are highlighted in black on the upper right, and creatinine is highlighted on the lower left. C: CIA metabolite plot for MS urinary data. Each closed triangle represents an MS ion. Again, hippurate ions are highlighted on the upper right.

 

DISCUSSION  
The application of metabolomics to nutritional studies holds great promise for future research. However, if it is to reach its full potential, we must understand the influence of all aspects of dietary composition on metabolic profiles. In the current study, the effect of acute changes in dietary phytochemicals on human urinary metabolic profiles was investigated. Modulation of the phytochemical intake led to significant changes in the metabolic profiles, and diets with varying phytochemical contents were distinguished and characterized by using pattern recognition analysis. Previous research has shown a notable effect of diet on urinary composition, which was detectable by metabolomic analysis. A recently published study investigated changes in human urinary profiles resulting from dietary intervention. Discrimination between the dietary treatments was dominated by an elevated excretion of creatinine, creatine, trimethylamine-N-oxide (TMAO), taurine and 1- and 3-methylhistidine associated with a high meat diet and p-hydroxyphenylacetate associated with a vegetarian diet (14). In another human study, a dietary intervention with soy isoflavones found subtle changes in urinary metabolic profiles that were associated with osmolyte fluctuation and energy metabolism (15). Differentiation between the urinary profiles of populations from distinct geographic locations has also been studied, and some of the differences between the groups have been attributed to diet (2, 4).

Dietary intervention studies carried out in animal models have also shown distinct changes resulting from the diet. A crossover metabolomic study in pigs successfully distinguished between a whole-grain diet and a non-whole-grain diet. In urine, differentiation was dominated by betaine and hippurate excretion associated with the whole-grain diet, and creatinine excretion was associated with the non-whole-grain diet (16). In another study, urine samples from rats that received 3 different 24-h dietary treatments (normal, overnight fast, or turkey diet) were discriminated on the basis of their NMR and extractive electrospray ionization mass spectra (17).

In addition to the changes that may result in metabolic profiles after a chronic dietary intervention, acute dietary intake can also influence the spectral outputs in metabolomics studies. These influences may be strong enough to produce outlying samples on PCA score plots and may influence the results of some studies if not identified and removed from the data. In a study investigating the metabolic effects of chamomile tea consumption, several outliers were identified after initial PCA analysis because of high urinary TMAO excretion, which resulted from fish consumption within 16 h of sampling (18). Alcohol consumption on the day before sampling has also been reported as a cause of outlying samples (6). The effect of other dietary components, such as phytochemicals, is not yet known; therefore, the main focus of the current study was to further our understanding of the effect of phytochemicals on urinary composition.

Assessment of nutrient intakes for the 21 volunteers found no significant differences in total energy, protein, carbohydrate, and fat intakes between the 3 dietary groups (ND, LPD, and SPD). Fiber intake was lower during the LPD than during the ND and SPD, but only significantly so for women. This was not unusual because of the removal of all fruit- and vegetable-type foods during the LPD. Fruit and vegetables were forbidden during the LPD and the SPD, but 6 g fiber/d was provided by the standard apple, carrot, and strawberry drinks during the SPD. Vitamin C and carotene intakes for the women were significantly lower during the LPD than during the ND (P < 0.001) and SPD (P < 0.001). Potatoes were permitted during the LPD and may have been the main source of vitamin C when other fruit and vegetables were removed from the diet. Carotene intakes were significantly higher during the SPD than during the ND (men: P < 0.05; women: P < 0.001) and LPD (men: P < 0.001; women: P < 0.001), and it is likely that the standard drinks were the source of carotene.

Distinct differences in the urinary metabolic profiles were observed after the removal of fruit and vegetables from the diet (LPD) compared with the ND or the SPD. Consultation of the 1H NMR PLS-DA loadings plots indicated that the separation in both models was associated with hippurate excretion during the SPD and ND and by creatinine and methylhistidine excretion during the LPD. This indicates that hippurate excretion is associated with phytochemical intake and that the appearance of these 3 metabolites in urine is subject to variation based on recent dietary intake. Hippurate is produced by the conjugation of benzoic acid with glycine in the liver, and it is then excreted in urine. There are 2 main dietary routes that yield the production of hippurate. One is by the consumption of foods containing benzoic acid, which may be present naturally or added as a preservative, and the other is through the metabolism of plant phenols by the gut microflora. Phenylpropionic acids are produced by the microbial degradation of polyphenols in the colon, which are further metabolized to benzoic acids and finally hippurate in the liver (8). Consumption of polyphenol-rich foods (8) and drinks such as black tea, green tea (19-20), and chamomile tea (18) have all been associated with an increased excretion of urinary hippurate.

Methylhistidine is produced by the catabolism of the myofibrillar proteins actin and myosin. Urinary 3-methylhistidine concentration has been used as a marker of myofibrillar protein turnover in human subjects (21), and, although intestinal smooth muscle and dietary intake also contribute to urinary excretion, skeletal muscle turnover is thought to contribute to 50–80% of urinary 3-methylhistidine excretion (22). Urinary creatinine is also related to muscle metabolism because it is produced from the breakdown of creatine phosphate. Creatinine excretion is related to body weight and has been proposed as a predictor of fat-free mass (23). One metabolomics study identified methylhistidine and creatinine as metabolites that contribute to differentiate between a high-meat and vegetarian diet (14). The dietary records in the current study did not indicate that there were any differences in meat consumption, and the nutrient analysis showed that there were no significant differences in protein or energy intakes across the study days. Therefore, it is likely that the only difference between the diets was in their phytochemical content. It is possible that other dietary factors, such as the glycemic load, influenced energy metabolism and muscle proteolysis.

Multivariate data analysis of the MS data produced parallel results, although the distinction between the LPD and SPD was not as strong (69 ± 4% for the MS data compared with 83 ± 8% for the NMR data). The MS models showed a distinction between the diet containing phytochemicals and the LPD, and this separation was also dominated by an elevated hippurate excretion after phytochemical consumption. Other ions remain unidentified.

The cross platform comparison of the NMR and MS data showed that both methods produced compatible results and identified the same metabolites in the differentiation between the LPD and the SPD as when the data sets were analyzed independently. Use of this type of analysis in the future will provide a powerful means of analyzing data from multiple platforms. In addition, it may aid in the identification of unknowns from different platforms.

This study has provided insight into the effects of recent dietary intakes on the outcome of urinary metabolomics analysis. The results indicate that a varying phytochemical consumption can contribute to differences in urinary metabolic profiles. Therefore, dietary standardization or specific dietary restrictions during the 24 h before urine collection may improve data interpretation by reducing the confounding effects caused by diverse dietary intakes among study populations.


ACKNOWLEDGMENTS  
We offer our sincere thanks to the volunteers for their commitment and patience during the study.

The authors' responsibilities were as follows—MCW: contributed to the study design and responsible for conducting the experiment, data interpretation and writing the manuscript; LB: responsible for conducting the experiment, data interpretation, and manuscript editing; EP-G, J-LS, and AS: responsible for MS analysis and manuscript editing; AF and DGH: responsible for data interpretation and manuscript editing; MJG: responsible for the conception and design of the experiment, data interpretation, and manuscript editing. None of the authors had any conflicts of interest.


REFERENCES  

Received for publication April 19, 2007. Accepted for publication July 26, 2007.



醫學百科App—醫學基礎知識學習工具


頁:




關閉

Copyright © 2008 39kf.com All rights reserved. 醫源世界 版權所有
醫源世界所刊載之內容一般僅用于教育目的。您從醫源世界獲取的信息不得直接用于診斷、治療疾病或應對您的健康問題。如果您懷疑自己有健康問題,請直接咨詢您的保健醫生。醫源世界、作者、編輯都將不負任何責任和義務。
本站內容來源于網絡,轉載僅為傳播信息促進醫藥行業發展,如果我們的行為侵犯了您的權益,請及時與我們聯系我們將在收到通知后妥善處理該部分內容
聯系Email: