Introduction

Excessive alcohol consumption increases the risk of liver cirrhosis and cancers (mostly those of the upper digestive and respiratory tract), while low to moderate red wine consumption has been associated with health-promoting properties (Artero et al., 2015). Red wine is different from other alcoholic beverages due to its content in various phenolic compounds. Moderate consumption of red wine has been linked to a lower risk of cardiovascular disease (CVD), diabetes, osteoporosis, and maybe neurological diseases (Artero et al., 2015; Karatzi et al., 2004; Iriti and Varoni, 2014). Recent studies have suggested that the positive health effects of wine, particularly red wine, cannot be merely attributed to its ethanol content (Karatzi et al., 2004; Iriti and Varoni, 2014). Wine is a complex matrix that contains many compounds of biological interest, of which phenolic compounds are some of the most interesting, as they are all biologically active and may play a role in the health benefitsof wine (Waterhouse, 2002; Nash et al., 2018). With the growing interest in this area, understanding the absorption, bioavailability, and metabolism of phenolic compounds from wine in humans is of utmost importance (Cueva et al., 2017; Fernandes et al., 2017; Mosele et al., 2015; Nash et al., 2018).

In general terms, the human metabolome is defined as the complete set of small metabolites found in a biological sample (i.e., cells, tissue, organ, biological fluids), and is influenced by gut microbiota and genetics, as well as by environmental factors (diet, contaminants, pharmaceuticals, and other lifestyle factors) (Riedl et al., 2017). As a consequence, all these factors contribute to interindividual differences in the metabolome, which translate into distinct nutritional requirements and diverse responses to nutritional or medical interventions (Riedl et al., 2017; Morand and Tomás-Barberán, 2019). An emerging idea derived from this issue consists in studying these metabolic differences in order to group subjects with common metabolic profiles when studying the effects of a specific intervention. This has led to the definition of the metabotype, which refers to a group of individuals with similar metabolic profiles (Tomás-Barberán et al., 2016; Cueva et al., 2017). Although there is no global consensus regarding the parametres for the definition of metabotypes, three different approaches can be applied depending on the objective of the study: general fasting metabotypes (including all the possible biochemical parametres); specific fasting metabotypes (searching for differences only in certain parametres, such as lipid profile); or response clustering, which is defined after a previously designed intervention has been carried out. Metabolomics has emerged as a tool for classification of individuals in metabotytes, and can therefore be used to predict physiological response in a dietary intervention. A recent metabolomic study has revealed two main clusters of postmenopausal women after bread intake, based on the correlation of data obtained from fasting metabotyping and insulin plasma levels (Moazzami et al., 2014). In the same way, a metabolomic approach applied to major metabolites (bile acids, fatty acids, amino acids, carboxy acids, hydroxylic acids, and aromatic derivatives) has led to individual stratification into two clusters corresponding to contrasting dietary patterns (high meat and low vegetable intake; low meat and high vegetable intake) (Wei et al., 2018).

Dietary polyphenols belong to one of the compound classes in which metabolism seems to be associated with individual metabotypes (Selma et al., 2009; Bolca et al., 2013; Manach et al., 2017; Mena et al., 2018). For instance, there is evidence of stratification in ellagitannin-metabolizing phenotypes (urolithin-producer metabotypes) or isoflavone‐metabolizing phenotypes (equol‐producer metabotypes). In previous work, three different urolithin-producing metabotypes have been described (Tomás-Barberán et al., 2014): after a 6 months intervention with pomegranate capsules enriched in phenolics, a reduction in different blood lipid biomarkers - including total cholesterol, LDL-cholesterol or non-HDL cholesterol among others - was only perceived in metabotype B obese patients (González-Sarrias et al., 2017). In a further study, a consistent relationship between urolithin metabotypes and cardiometabolic risk biomarkers was reported; results demonstrated that overweight metabotype B patients are at risk of cardiovascular disease, whereas metabotype A patients are protected from these disorders (Selma et al., 2017). In those previous studies, the term ‘individual stratification’ was used to classify patients prior to an intervention and predict their response in terms of level of responsiveness/non-responsiveness.

Several studies on wine have applied metabolomics to study changes in metabolites after moderate wine consumption (Vázquez-Fresno et al., 2015; Boto-Ordonez et al., 2013; Urpi-Sarda et al., 2015; Muñoz-González et al., 2013). In particular, 1H-NMR and UHPLC-MS/MS approaches have been used to evaluate differences in the urinary metabolome of cardiovascular risk patients after moderate intake of red wine and dealcoholized red wine (Vázquez-Fresno et al., 2012; Boto-Ordoñez et al., 2013). More specifically, Vázquez-Fresno et al. (2016) clustered a specific population into four phenotypic groups according to their biochemical characteristics. The two most discriminating clusters (“obese and diabetic” metabotype and “healthier” metabotype) showed different metabolic responses to a wine polyphenol intervention.

The present study aims to investigate the differential responsiveness to moderate wine consumption in healthy subjects. For this purpose, urinary metabolomics of samples collected before and after a red wine intervention period were studied, and differences in the observed metabotypes were explored.

Materials and methods

6. Statistical analysis and data interpretation

A principal component analysis (PCA) provided an overview of data distribution (Mass Profiler Professional software v14.0, Agilent technologies) (Figure 1). The metabolites that had been significantly altered after wine consumption [26] were then selected and an unsupervised Hierarchical Clustering Analysis (HCA) using SIMCA-P software v.14 was carried out. Based on the classification, a non-parametric Kruskall-Wallis analysis was performed (Mass Profiler Professional software v14.0, Agilent Technologies) and a list of the metabolites (p<0.05) responsible for the stratification of the volunteers into the clusters was obtained. Only those with a Coefficient of Variance (CV)<20 % were selected for further analysis. After that, statistical differences between the stratified clusters, determined using Orthogonal Partial Least-Squares (OPLS) discriminant analysis (DA) was performed and metabolites with Variable Importance in Projection values (VIP)>1 were selected. Finally, the models were validated by the goodness-of-fit parameter (R²X), the proportion of the variance of the response variable that is explained by the model (R²Y) and the predictive ability parametre (Q²), calculated by seven-fold internal cross-validation. The validation of the OPLS-DA models was carried out with a permutation test (n=200).

A tentative identification of statistically different (p<0.05) metabolites was performed by carrying out a mass search in the HMDB within a mass accuracy window of 15 ppm. When several candidates were found for a given m/z, metabolite identification was sorted giving preference to metabolites normally found in urine or human fluids after wine intake. When available, co-injection of standards with urine samples was carried out for further confirmation.

Figure 1. Workflow and statistical approach.

Save View full size Expand inline Collapse inline

Results

1. Urine metabotypes associated with red wine consumption

A total of 33 healthy subjects were accepted for the trial for a 4-week intervention study during which subjects daily ingested 250 mL of red wine. A control group (n=8) followed the same pattern with the difference that no red wine was consumed. 24h-urine samples after a wash-out period (labeled as “pre-wine” samples), and after the intervention period (labeled as “post-wine” samples) were analysed by UHPLC-MS using a non-targeted metabolomic approach. After UHPLC-MS analysis and data pretreatment, 1825 entities were obtained after deconvolution, alignment and filtering. This data set was used for PCA analysis to evaluate the differences in the urinary metabolome between the experimental groups (Figure 1). PCA scores plot (Figure 2) shows a clear separation of subjects after intervention (“post-wine”), indicating that red wine consumption clearly induced perturbations in the urine metabolome. Moreover, as might be expected, an overlapping of samples from control subjects and case subjects before wine intervention (“pre-wine”) was observed. Hierarchical clustering analysis (HCA) was then performed with those metabolites whose responses were significantly altered (p<0.05) after wine intervention (n=182) (Esteban-Fernández et al., 2018; Jiménez-Girón et al., 2015).

Figure 2. PCA score plot of the first two components from non-targeted metabolomics data from urine belonging to all studied groups: “post-wine” case, “post-wine” control, “pre-wine” case and “pre-wine” control. The first component explains 10.1 % of variance and the second component explains 5.7 % of variance.

Save View full size Expand inline Collapse inline

A dendrogram showing the overall structural similarity determined by Ward's clustering based on Euclidean distance is represented in Figure 3. The clustering of individuals by HCA suggests the existence of three clusters: Cluster 1 (metabotype 1) was the most abundant (n=17) and showed a close link with cluster 2 (metabotype 2, n=7), while cluster 3 (metabotype 3, n=9) was clearly differentiated from clusters 1 and 2. With this new classification of volunteers, data of samples after wine intervention were subjected to a Kruskal-Wallis statistical analysis and a new set of 1460 significant signals (p<0.05) were generated. Only those signals with a Coefficient of Variance (CV)<20 % (n=59) were selected for further OPLS-DA analysis.

Figure 3. Dendrogram of HCA shows case “post-wine”samples clustered into three groups: cluster 1 (green), cluster 2 (red), cluster 3 (blue).

Save View full size Expand inline Collapse inline

Score plots from the supervised OPLS-DA (Figure 4) showed distinguishable separation between clusters 1, 2 and 3. To select the metabolites that were important for this model, VIP scores were used. VIP scores > 1 were considered as relevant and 29 metabolites were therefore considered significant in differentiating the three metabotypes (Table 1).

The model parametres were: R2X=0.39, R2Y=0.67 and Q2=0.51. The response permutation test (with n=200) was used to validate the predictive capability of the computed OPLS-DA model (Figure S1 of the Supporting Information).

Table 1. Metabolites identified or tentatively identified with VIP>1 and associated with principal component 1 and principal component 2 of OPLS-DA analysis of post-consumption samples.

View popup Expand inline Collapse inline

Retention time	Detected m/z	Assigned ion	Molecular Formula	Error (ppm)	Tentative ID	Level of annotation according to MSI		Group	VIP value		Relative response order
									PC1	PC2
2.29	147.07	M-H	C6H12O4	3.52	2,3-Dihydroxy-3-methylvalerate			Wine	1.87	1.64		3>2>1
6.86	613.11	M+Cl	C30H26O12	2.24	Procyanidin B-type dimer			Microbial	1.71	1.61		3>2>1
8.38	293.10	3M-H	C5H6O2	0.47	2-Furanmethanol			Wine	1.64	1.38		3>2>1
1.99	161.05	M-H	C6H10O5	3.43	3-Hydroxymethylglutaric acid			Wine	1.60	1.32		3>2>1
3.47	175.06	M-H	C7H12O5	6.30	2,3-Dimethyl-3-hydroxyglutaric acid			Wine	1.57	1.32		3>2>1
3.47	175.06	M-H	C7H12O5	6.30	Isopropylmalic acid			Wine	1.57	1.31		3>2>1
1.05	129.02	M-H	C5H6O4	8.27	γ-delta-Dioxovaleric acid			Wine	1.54	1.27		3>2>1
1.03	147.03	M-H	C5H8O5	7.50	Hydroxyglutaric acid			Wine	1.48	1.32		3>2>1
1.05	111.01	M-3H	C12H16O9S	0.83	4-Hydroxy-5-(dihydroxyphenyl)-valeric acid-O-methyl-O-sulphate			Microbial	1.46	1.20		3>2>1
2.09	144.07	M-2H	C15H14O6	5.16	Epicatechin/Catechin			Wine	1.46	1.21		3>2>1
1.06	191.02	M-H	C6H8O7	5.09	Citric acid		1	Wine	1.44	1.19		3>2>1
0.96	264.99	M+K-2H	C13H8O4	4.40	Urolithin A			Microbial	1.37	1.27		1>2>3
3.48	263.95	M+Br	C7H7NO5	12.91	2-Amino-3-carboxymuconic acid semialdehyde			endogenous	1.34	1.20		3>2>1
0.76	193.03	M+Cl	C7H10O4	4.72	Isopropylmaleate			Wine	1.30	1.07		3>2>1
4.37	277.00	M+Hac-H	C7H6O6S	0.51	Sulfosalicylic acid			Microbial	1.24	1.13		2>3>1
0.78	144.03	M-2H	C15H14O6	5.16	Epicatechin/Catechin		1	Wine	1.11	1.04		2>3>1
0.75	191.02	M-H	C7H12O6	8.84	Quinic acid			Wine	1.09	>1		3>2>1
2.66	161.05	M+FA-H	C5H8O3	0.90	2-Oxovaleric acid			Wine	1.09	>1		3>2>1
1.52	424.05	M+Cl	C12H23NO9S2	4.78	3-Methylbutyl glucosinolate			Diet	1.06	>1		3>2>1
0.75	227.00	M-H	C13H8S2	3.82	5-Ethynyl-5'-(1-propynyl)-2,2'-bithiophene			Diet	1.03	>1		3>2>1
2.07	641.21	M+TFA-H	C25H36O12	1.97	16,17-Dihydro-16a,17-dihydroxygibberellin A4 17-glucoside			Microbial	1.02	>1		3>2>1
2.77	269.11	M-H	C12H18N2O5	0.39	Hypoglycin B			Diet	>1	1.52		2>3>1
3.67	204.03	M-H20-H	C10H9NO5	2.05	(R)-2,3-Dihydro-3,5-dihydroxy-2-oxo-3-indoleacetic acid			Wine	>1	1.39		2>1>3
1.32	442.08	M+TFA-H	C17H15NO6	13.07	(Z)-N-Feruloyl-5-hydroxyanthranilic acid			Microbial	>1	1.30		2>1>3
2.20	365.14	M+Na-2H	C20H24O5	3.12	5-(6-Hydroxy-3,7-dimethyl-2,7-octadienyloxy)-7-methoxycoumarin			Wine	>1	1.27		2>3>1
9.30	233.15	M+Hac-H	C13H18	0.01	1-Methyl-4-(1-methyl-2-propenyl)-benzene			Diet	>1	1.14		1>3>2
2.29	245.11	M-H	C10H18N2O5	1.65	Aspartyl-Leucine/Leucyl-Aspartate			endogenous	>1	1.13		2>3>1
2.68	194.05	M-H	C9H9NO4	1.64	4-Hydroxyhippuric acid		1	Microbial	>1	1.12		3>1>2
1.35	293.03	M-H20-H	C13H12O9	12.82	Cis-Caffeoyl tartaric acid/Caftaric acid			Wine	>1	1.12		3>2>1

Figure 4. OPLS-DA score plot of “post-wine” urine samples. Cluster 1 is represented by green circles, cluster 2 by red circles and cluster 3 by blue circles. The ellipse shown in the model represents the Hotelling T2 with 95 % confidence.

Save View full size Expand inline Collapse inline

With the aim of detecting possible underlying determinants leading to the 3 metabotype clusters, we analysed the urinary metabolomic profile of samples collected before the dietary intervention. 328 metabolites were detected after Kruskall-Wallis analysis (p<0.05) and those with CV<20 % were selected for further identification. From them, 28 metabolites were tentatively identified and submitted to OPLS-DA analysis. When OPLS-DA was performed for the metabolomic data set obtained from urine samples before wine intervention, a limited separation between previously found clusters was observed (Figure 5). A value of VIP (Variable Important in Projection)  ≥ 1 was considered to identify the variables most important to group separation (Table 2). The OPLS-DA model parametres were: R2X=0.26, R2Y=0.44 and Q2=0.27. The volunteers’ metabolic phenotypes before wine intervention were more similar and thus appeared closer in the OPLS-DA. PC1 (13 %) differentiated between cluster 1 and cluster 3, whereas PC2 (13 %) separated Cluster 2.

Figure 5. OPLS-DA score plot of “pre-wine” urine samples. Cluster 1 is represented by green circles, cluster 2 by red circles and cluster 3 by blue circles. The ellipse shown in the model represents the Hotelling T2 with 95 % confidence.

Save View full size Expand inline Collapse inline

Table 2. Metabolites with VIP values >1 and associated with PC1 and PC2 of OPLS-DA analysis in pre-consumption urine samples.

View popup Expand inline Collapse inline

Retention time	Detected m/z	Assigned ion	Molecular Formula	Error (ppm)	Tentative ID	Group	VIP value		Relative response order
							PC1	PC2
1.98	385.10	M+Br	C18H26O4	3.13	Capsiate	Diet	1.83	1.42	2>3>1
1.32	160.06	M-H	C6H11NO4	0.42	Acetylhomoserine	Diet	1.79	1.63	3>2>1
0.69	177.04	M-H	C7H8N4O3	3.67	Dimethyluric acid	Endogenous	1.67	1.83	1>3>2
2.15	328.05	M-H	C10H12N5O6P	2.91	Adenosine 2',3'-cyclic phosphate	Endogenous	1.63	1.36	3>2>1
2.20	365.14	M-H	C17H22N2O7	1.30	Tetrahydropentoxyline	Diet	1.60	1.29	3>2>1
2.31	317.06	M+TFA-H	C7H12N2O5	3.44	Aspartyl-Alanine/Alanyl-Aspartate	Endogenous	1.31	1.01	3>2>1
9.00	194.08	M+Hac-H	C8H9NO	1.37	2,3-Dihydro-1H-pyrrolizine-5-carboxaldehyde	Diet	1.21	1.41	2>3>1
2.30	317.06	M+TFA-H	C7H12N2O5	2.81	Alanyl-Aspartate/Aspartyl-Alanine	Endogenous	1.15	>1	3>2>1
3.61	439.14	M+FA-H	C18H22N2O8	0.27	N-Acetylserotonin glucuronide	Endogenous	1.05	>1	2>3>1
0.93	248.03	M+Hac-H	C6H7NO4S	10.37	2-aminophenol sulphate	Diet	1.05	>1	2>3>1
3.99	263.10	M-H	C13H16N2O4	1.78	di-Hydroxymelatonin	Endogenous	1.02	1.76	2>3>1
3.99	835.29	2M-H	C18H26O11	1,98	Oleoside dimethyl ester	Diet	>1	1.13	2>1>3
0.71	147.03	M+H	C6H12S2	2,27	Propenyl propyl disulfide	Diet	>1	1.07	1>2>3

Discussion

Polyphenol-rich dietary interventions in humans have demonstrated high interindividual variability in polyphenol degradation patterns. A number of factors contribute to interindividual differences in the human metabolome (gut microbiota, and genetic and environmental factors). Metabotyping has been proposed to overcome the confounding effects of variation, and is becoming accepted for studying the effects of a specific intervention. In the nutritional field, metabotyping is closely linked to the concept of “personalised nutrition”, which has evolved from a nutrigenomic concept to an approach involving dietary, phenotypic and genotypic features (O'Donovan et al., 2016). Nowadays, nutritional issues haveprogressed from individual to group guidelines, optimising dietary interventions and reducing healthcare costs and diagnosis time (O'Donovan et al., 2016; Toro-Martín et al., 2016). In agreement with this, the “Food4me” study has shown that personalised nutrition can cause positive changes in the nutritional behaviour of patients (Celis-Morales et al., 2017). However, in order to deliver targeted dietary advice, the identification of metabolic phenotypes or “metabotypes” is of key importance.

The definition of robust metabotypes is a difficult task, since it involves taking into account a variety of parametres (Toro-Martín et al., 2017; Brennan, 2017). Nevertheless, the study of responses to a dietary intervention is of great interest since the differences between metabolic activities become more obvious. In this context, metabolomics has been successfully applied in the development of novel population classification methods after the consumption of a specific food, particularly polyphenol-rich foods, such as strawberry, pomegranate or walnuts (Bolca et al., 2007; González-Sarria et al., 2017; Spencer et al., 2008; Truchado et al., 2012). Regarding wine or grape-derived products, metabolomics has been mainly applied to the discovery of biomarkers associated with their intake (Vázquez-Fresno et al., 2015; Van Dorsten et al., 2010). Only Vazquez et al. have determined changes in urinary metabolites among different clinic phenotypes in an intervention study with dealcoholized red wine (Vázquez-Fresno et al., 2016). In the present study, after carrying out an unsupervised HCA, healthy volunteers were clustered into three different groups after moderate red wine consumption for a 4-week period. Since we did not use a quantitative approach, it is not possible to assume that these clusters correspond to individuals with great, moderate and low metabolic capacity. However, they are in good agreement with the results of a previous study in which we identified three distinct metabotypes (low, moderate and high metaboliser phenotypes), based on the total phenolic content of faeces from the same volunteers (Muñoz-González et al., 2013). Metabotypes observed in this previous study were evidenced after a dietary intervention, which, to our knowledge, has not been previously reported in relation to moderate red wine consumption. Regarding beneficial health outcomes, it is assumed that wine should only be consumed in a moderate and regular way by healthy people, within a framework of balanced habits. Therefore, the categorisation of individuals demonstrated in this work has revealed the relevance of wine moderate intake in a real and everyday nutritional intervention/situation.

Metabotype 3 subjects produced the highest levels of the most detected metabolites, particularly valeric acid and valerolactones derivatives, which have already been proposed as urinary biomarkers associated with wine consumption (Urpi-Sarda et al., 2015; Vázquez-Fresno et al., 2012). Valeric acids are short-chain fatty acids (SCFA) produced by the microbial catabolism of protein-derived branched chain amino acids. They have been linked to several health benefits, including anti-inflammatory (Huda-Faujan et al., 2010), anti-tumorigenic and antimicrobial effects (Tan et al., 2014). This group of individuals also showed a metabotype marked with increased SCFAs production, highlighting the effects of moderate wine consumption produced on the gut microbial communities, as already observed for gut bacteria, such as Flavonifractor plautii (Navarro-Peran et al., 2008). Metabotype 3 individuals showed the highest levels of procyanidins and 4-hydroxy-5-(dihydroxyphenyl)-valeric acid-O-methyl-O-sulphate, the latter being a phenolic conjugate previously observed in human urine after tea intake. Metabotype 3 subjects had the largest increase in epicatechin/catechin oligomers (procyanidin B-type dimers). The oligomers and monomers of epicatechin and catechin and their metabolites exert several health-promoter activities including antioxidant, antibacterial, antiadhesive and cardio protective effects (Navarro-Peran et al., 2008; Escandón et al., 2016). The metabotype 3 group was also characterised by significantly higher 2-amino-3-carboxymuconic acid semialdehyde, an intermediate on the tryptophan-niacin catabolic pathway; it is therefore related to quinic acid, which is also produced by the different metabotypes. In agreement with our observations, an increase of urine nicotinic acid, as well as changes in amino acids related to the tryptophan metabolic pathway, have been previously observed in healthy volunteers after short-term intake of red wine and grape polyphenol extract (Jacobs et al., 2012). Moreover, quinic acid is metabolised by microbiota into hippuric acid, and it has been used as a dietary supplement due to its ability to increase nicotinamide and tryptophan in urine, suggesting a positive effect of this acid in nutrient bioavailability. Hydroxyhippuric acid, its product, has been previously reported in metabolomics studies concerning moderate wine consumption (Vázquez-Fresno et al., 2012; van Dorsten et al., 2010).

Interestingly, metabotype 1 subjects produced the highest levels of urolithin A, a metabolite produced in the gut following consumption of ellagitannin. This type of compound is not only known for its cardioprotective (Graf et al., 2015) and anti inflammatory (David et al., 2014) properties, but also for its protective effects on the ageing process (Ryu et al., 2016). The benefits implied from this metabolite seems to be structure-dependent and its production has been associated with the composition of colonic microbiota, concretely with the presence of Gordinobacter spp. (González-Sarrias et al., 2017).

In spite of all the above-mentioned advantages of personalised nutrition and metabotyping, there are some limiting factors. An important challenge is the translation of basic research into clinically relevant dietary guidelines, which could be achieved by establishing a common framework for scientist, clinicians and health professionals. It would be necessary to properly define the metabotypes and to select the metabolites to be analysed for population stratification in wine intervention studies. This selection should be wide enough to be representative of a physiological condition and be meaningful for the appropriate clustering of individuals/patients. However, the most notable bottleneck is the scale of these observations regarding the overall population, since there is huge diversity in the parametres and observations of each study, thus making it difficult to establish universal criteria.

Conclusions

Urine, an easily accessible biofluid, can be used to reveal different metabotypes associated with moderate wine consumption. This study has successfully demonstrated the stratification of a population into three different clusters of individuals/metabotypes after moderate and regular wine consumption. This novel clustering has been performed on healthy volunteers, highlighting the potential applications of tailored nutrition, not only for patients at risk of disease, but also as a promoter/indicator of healthy life habits. Metabolomics constitutes a useful tool in metabotyping; however, in order to optimise dietary interventions or medical treatments, more effort should be made to homogenise the studies and facilitate access to these tools by general population.

Acknowledgments

This work was funded by the Spanish MINECO (AGL2015-64522-C2-R and AGL2017-89055-R projects) and Comunidad de Madrid (ALIBIRD2020-CM P2018/BAA-4343). A.E-F is a recipient of a fellowship from the FPI-MINECO programme.

References

• Artero A., A. Artero J.J. Tarin and A. Cano, 2015. The impact of moderate wine consumption on health, Maturitas, 80, 3-13. doi:10.1016/j.maturitas.2014.09.007
• Beger R.D., W. Dunn, M.A. Schmidt, S.S. Gross, J.A. Kirwan, M. Cascante, L. Brennan, D.S. Wishart, M. Oresic, T. Hankemeier, D.I. Broadhurst, A.N. Lane, K. Suhre, G. Kastenmuller, S.J. Sumner, I. Thiele, O. Fiehn and R. Kaddurah-Daouk, 2016. Metabolomics enables precision medicine: "A White Paper, Community Perspective", Metabolomics, 12, 149. doi:10.1007/s11306-016-1094-6
• Bolca S., S. Possemiers, A. Herregat, I. Huybrechts, A. Heyerick, S. De Vriese, M. Verbruggen, H. Depypere, D. De Keukeleire, M. Bracke, S. De Henauw, W. Verstraete and T. Van de Wiele, 2007. Microbial and dietary factors are associated with the equol producer phenotype in healthy postmenopausal women. Journal Nutr, 137, 2242-2246. doi:10.1093/jn/137.10.2242
• Bolca S., T. Van de Wiele and S. Possemiers, 2013. Gut metabotypes govern health effects of dietary polyphenols. Curr Opin Biotechnol, 24, 220-225. doi:10.1016/j.copbio.2012.09.009
• Bondia-Pons I., T. Barri, K. Hanhineva, K. Juntunen, L.O. Dragsted, H. Mykkanen and K. Poutanen, 2013. UPLC-QTOF/MS metabolic profiling unveils urinary changes in humans after a whole grain rye versus refined wheat bread intervention. Mol Nutr Food Res, 57, 412-422. doi:10.1002/mnfr.201200571
• Boto-Ordonez M., M. Urpi-Sarda, M.I. Queipo-Ortuno, D. Corella, F.J. Tinahones, R. Estruch and C. Andres-Lacueva, 2013. Microbial metabolomic fingerprinting in urine after regular dealcoholized red wine consumption in humans. J Agric Food Chem, 61, 9166-9175. doi:10.1021/jf402394c
• Brennan L., 2017. Use of metabotyping for optimal nutrition. Curr Opin Biotechnol, 44, 35-38.
• Celis-Morales C., K.M. Livingstone, C.F. Marsaux, A.L. Macready, R. Fallaize, C.B. O'Donovan, C. Woolhead, H. Forster, M.C. Walsh, S. Navas-Carretero, R. San-Cristobal, L. Tsirigoti, C.P. Lambrinou, C. Mavrogianni, G. Moschonis, S. Kolossa, J. Hallmann, M. Godlewska, A. Surwillo, I. Traczyk, C. A. Drevon, J. Bouwman, B. van Ommen, K. Grimaldi, L. D. Parnell, J.N. Matthews, Y. Manios, H. Daniel, J.A. Martinez, J.A. Lovegrove, E.R. Gibney, L. Brennan, W.H. Saris, M. Gibney and J.C. Mathers, 2017. Effect of personalized nutrition on health-related behaviour change: evidence from the Food4Me European randomized controlled trial, Int J Epidemiol, 46, 578-588. doi:10.1093/ije/dyw186
• Cueva C., Gil-Sánchez I., Ayuda-Durán B., González-Manzano S., González-Paramás A.M., Santos-Buelga C., Moreno-Arribas M., 2017. An integrated view of the effects of wine polyphenols and their relevant metabolites on gut and host health. Molecules,, 22(1), 99. doi:10.3390/molecules22010099
• David L.A., C.F. Maurice, R.N. Carmody, D.B. Gootenberg, J.E. Button, B.E. Wolfe, A.V. Ling, A.S. Devlin, Y. Varma, M.A. Fischbach, S.B. Biddinger, R.J. Dutton and P.J. Turnbaugh, 2014. Diet rapidly and reproducibly alters the human gut microbiome, Nature, 505, 559-563. doi:10.1038/nature12820
• Escandón R.A., M. del Campo, R. López-Solis, E. Obreque-Slier and H. Toledo, 2016. Antibacterial effect of kaempferol and (−)-epicatechin on Helicobacter pylori. Eur Food Res Technol, 242, 1495-1502. doi:10.1038/nature12820
• Esteban-Fernandez A.,C. Ibanez C. Simo B. Bartolome and M.V. Moreno-Arribas, 2018. An ultrahigh-performance liquid chromatography-time-of-flight mass spectrometry metabolomic approach to studying the impact of moderate red-wine consumption on urinary metabolome, J Proteome Res, 17, 1624-1635. doi:10.1021/acs.jproteome.7b00904
• Fernandes I., Pérez-Gregorio R., Soares S., Mateus N. and de Freitas V., 2017. Wine Flavonoids in Health and Disease Prevention. Molecules, 22, 292. doi:10.3390/molecules22020292
• Gonzalez-Sarrias A., R. Garcia-Villalba, M. Romo-Vaquero, C. Alasalvar, A. Orem, P. Zafrilla, F.A. Tomas-Barberan, M.V. Selma and J.C. Espin, 2017. Clustering according to urolithin metabotype explains the interindividual variability in the improvement of cardiovascular risk biomarkers in overweight-obese individuals consuming pomegranate: A randomized clinical trial, Mol Nutr Food Res, 61, 1600830- 1600882. doi:10.1002/mnfr.201600830
• Graf D., R. Di Cagno, F. Fak, H.J. Flint, M. Nyman, M. Saarela and B. Watzl, 2015. Contribution of diet to the composition of the human gut microbiota, Microb Ecol Health Dis, 26, 26164. doi:10.3402/mehd.v26.26164
• Huda-Faujan N., A.S. Abdulamir, A.B. Fatimah, O.M. Anas, M. Shuhaimi, A.M. Yazid and Y.Y. Loong, 2010. The impact of the level of the intestinal short chain fatty acids in inflammatory bowel disease patients versus healthy subjects, Open Biochem J, 4, 53-58. doi:10.2174/1874091X01004010053
• Iriti M. and E.M. Varoni, 2014. Cardioprotective effects of moderate red wine consumption: Polyphenols vs. ethanol, J Appl Biomed, 12, 193-202. doi:10.2174/1874091X01004010053
• Jacobs D.M., J.C. Fuhrmann, F.A. van Dorsten, D. Rein, S. Peters, E.J. van Velzen, B. Hollebrands, R. Draijer, J. van Duynhoven and U. Garczarek, 2012. Impact of short-term intake of red wine and grape polyphenol extract on the human metabolome, J Agric Food Chem, 60, 3078-3085. doi:10.1021/jf2044247
• Jiménez-Girón A., C. Ibáñez, A. Cifuentes, C. Simó, I. Muñoz-González, P.J. Martín-Álvarez, B. Bartolomé and M.V. Moreno-Arribas, 2015. faecal metabolomic fingerprint after moderate consumption of red wine by healthy subjects, J Proteome Res, 14, 897-905. doi:10.1021/pr500960g
• Karatzi K., C. Papamichael, K. Aznaouridis, E. Karatzis, J. Lekakis, C. Matsouka, G. Boskou, A. Chiou, M. Sitara, G. Feliou, D. Kontoyiannis, A. Zampelas and M. Mavrikakis, 2004. Constituents of red wine other than alcohol improve endothelial function in patients with coronary artery disease, Coron Artery Dis, 15, 485-490. doi:10.1097/00019501-200412000-00005
• Manach C., D. Milenkovic, T. Van de Wiele, A. Rodriguez-Mateos, B. de Roos, M.T. Garcia-Conesa, R. Landberg, E.R. Gibney, M. Heinonen, F. Tomas-Barberan and C. Morand, 2017. Addressing the inter-individual variation in response to consumption of plant food bioactives: Towards a better understanding of their role in healthy aging and cardiometabolic risk reduction, Mol Nutr Food Res, 2017, 61, 1600557- 1600568. doi:10.1002/mnfr.201600557
• Mena P., I.A. Ludwig, V.B. Tomatis, A. Acharjee, L. Calani, A. Rosi, F. Brighenti, S. Ray, J. L. Griffin, L. J. Bluck and D. Del Rio, 2018. Inter-individual variability in the production of flavan-3-ol colonic metabolites: preliminary elucidation of urinary metabotypes, Eur J Nutr, DOI: 10.1007/s00394-018-1683-4. doi:10.1007/s00394-018-1683-4
• Moazzami A.A., A. Shrestha, D.A. Morrison, K. Poutanen and H. Mykkanen, 2014. Metabolomics reveals differences in postprandial responses to breads and fasting metabolic characteristics associated with postprandial insulin demand in postmenopausal women. J Nutr 144, 807-814. doi:10.3945/jn.113.188912
• Morand C. and Tomás-Barberán F.A. Contribution of plant food bioactives in promoting health effects of plant foods: why look at interindividual variability?. European journal of nutrition. 2019, 1-7. doi:10.1007/s00394-019-02096-0
• Mosele J.I., Macià A. and Motilva M. Metabolic and microbial modulation of the large intestine ecosystem by non-absorbed diet phenolic compounds: A review. Molecules 2015, 20, 17429–17468. doi:10.3390/molecules200917429
• Muñoz-González I., A. Jiménez-Girón, P.J. Martín-Álvarez, B. Bartolomé and M.V. Moreno-Arribas, 2013. Profiling of Microbial-Derived Phenolic Metabolites in Human Feces after Moderate Red Wine Intake, J Agric Food Chem, 61, 9470-9479. doi:10.1021/jf4025135
• Nash V., C.S. Ranadheera, E.N. Georgousopoulou, D.D. Mellor, D.B. Panagiotakos, A.J. McKune, J. Kellett and N. Naumovski,. The effects of grape and red wine polyphenols on gut microbiota – A systematic review. Food Research International 2018, 113, 277–287. doi:10.1016/j.foodres.2018.07.019
• Navarro-Peran E., J. Cabezas-Herrera, L. Sanchez-Del-Campo, F. Garcia-Canovas and J.N. Rodriguez-Lopez, 2008. The anti-inflammatory and anti-cancer properties of epigallocatechin-3-gallate are mediated by folate cycle disruption, adenosine release and NF-kappaB suppression, Inflamm Res, 2008, 57, 472-478. doi:10.1007/s00011-008-8013-x
• O'Donovan C.B., M.C. Walsh, M.J. Gibney, E. R. Gibney and L. Brennan, 2016. Can metabotyping help deliver the promise of personalised nutrition?, Proc Nutr Soc, 75, 106-114. doi:10.1017/S0029665115002347
• Riedl A., C. Gieger, H. Hauner, H. Daniel and J. Linseisen, 2017. Metabotyping and its application in targeted nutrition: an overview, Br J Nutr, 117, 1631-1644. doi:10.1017/S0007114517001611
• Ryu D., L. Mouchiroud, P.A. Andreux, E. Katsyuba, N. Moullan, A.A. Nicolet-dit-Felix, E. G. Williams, P. Jha, G. Lo Sasso, D. Huzard, P. Aebischer, C. Sandi, C. Rinsch and J. Auwerx, 2016. Urolithin A induces mitophagy and prolongs lifespan in C. elegans and increases muscle function in rodents, Nat Med, 22, 879-888. doi:10.1038/nm.4132
• Selma M.V., A. Gonzalez-Sarrias, J. Salas-Salvado, C. Andres-Lacueva, C. Alasalvar, A. Orem, F. A. Tomas-Barberan and J. C. Espin, 2017. The gut microbiota metabolism of pomegranate or walnut ellagitannins yields two urolithin-metabotypes that correlate with cardiometabolic risk biomarkers: Comparison between normoweight, overweight-obesity and metabolic syndrome, Clin Nutr, DOI: 10.1016/j.clnu.2017.03.012.
• Selma M.V., J.C. Espín and F.A. Tomás-Barberán, 2009. Interaction between phenolics and gut microbiota: role in human health, J Agric Food Chem, 57, 6485-6501. doi:10.1021/jf902107d
• Spencer J.P., M.M. Abd El Mohsen, A.M. Minihane and J.C. Mathers, 2008. Biomarkers of the intake of dietary polyphenols: strengths, limitations and application in nutrition research, Br J Nutrition, 99, 12-22. doi:10.1017/S0007114507798938
• Takagaki A. and F. Nanjo, 2015. Bioconversion of (-)-epicatechin, (+)-epicatechin, (-)-catechin, and (+)-catechin by (-)-epigallocatechin-metabolizing bacteria, Biol Pharm Bull, 38, 789-794. doi:10.1248/bpb.b14-00813
• Tan J.C. McKenzie, M. Potamitis, A. N. Thorburn, C. R. Mackay and L. Macia, 2014. The role of short-chain fatty acids in health and disease, Adv Immunol, 121, 91-119. doi:10.1016/B978-0-12-800100-4.00003-9
• Tomas-Barberan F.A., M.V. Selma and J.C. Espin, 2016. Interactions of gut microbiota with dietary polyphenols and consequences to human health, Curr Opin Clin Nutr Metab Care, 19, 471-476. doi:10.1097/MCO.0000000000000314
• Tomas-Barberan F.A., R. Garcia-Villalba, A. Gonzalez-Sarrias, M. V. Selma and J. C. Espin, 2014. Ellagic acid metabolism by human gut microbiota: consistent observation of three urolithin phenotypes in intervention trials, independent of food source, age, and health status, J Agric Food Chem, 62, 6535-6538. doi:10.1021/jf5024615
• Toro-Martin J. de, B. J. Arsenault, J. P. Despres and M. C. Vohl, 2017. Precision nutrition: a review of personalized nutritional approaches for the prevention and management of metabolic syndrome, Nutrients, 9, DOI: 10.3390/nu9080913. doi:10.3390/nu9080913
• Truchado P., M. Larrosa, M.T. Garcia-Conesa, B. Cerda, M.L. Vidal-Guevara, F.A. Tomas-Barberan and J.C. Espin, 2012. Strawberry processing does not affect the production and urinary excretion of urolithins, ellagic acid metabolites, in humans, J Agric Food Chem, 60, 5749-5754. doi:10.1021/jf203641r
• Urpi-Sarda M., M. Boto-Ordonez, M.I. Queipo-Ortuno, S. Tulipani, D. Corella, R. Estruch, F.J. Tinahones and C. Andres-Lacueva, 2015. Phenolic and microbial-targeted metabolomics to discovering and evaluating wine intake biomarkers in human urine and plasma, Electrophoresis, DOI: 10.1002/elps.201400506. doi:10.1002/elps.201400506
• van Dorsten F.A., C.H. Grun, E.J. van Velzen, D.M. Jacobs, R. Draijer and J.P. van Duynhoven. 2010. The metabolic fate of red wine and grape juice polyphenols in humans assessed by metabolomics, Mol Nutr Food Res, 54, 897-908. doi:10.1002/mnfr.200900212
• Vázquez-Fresno R., R. Llorach, A. Perera, R. Mandal, M. Feliz, F. J. Tinahones, D. S. Wishart and C. Andres-Lacueva, 2016. Clinical phenotype clustering in cardiovascular risk patients for the identification of responsive metabotypes after red wine polyphenol intake, J Nutr Biochem, 28, 114-120. doi:10.1016/j.jnutbio.2015.10.002
• Vázquez-Fresno R., R. Llorach, F. Alcaro, M.A. Rodríguez, M. Vinaixa, G. Chiva-Blanch, R. Estruch, X. Correig and C. Andrés-Lacueva, 2012. 1H-NMR-based metabolomic analysis of the effect of moderate wine consumption on subjects with cardiovascular risk factors, Electrophoresis, 33, 2345-2354. doi:10.1002/elps.201100646
• Vázquez-Fresno R., R. Llorach, M. Urpi-Sarda, O. Khymenets, M. Bulló, D. Corella, M. Fitó, M.A. Martínez-González, R. Estruch and C. Andres-Lacueva, 2015. An NMR metabolomics approach reveals a combined-biomarkers model in a wine interventional trial with validation in free-living individuals of the PREDIMED study, Metabolomics, 11, 797-806. doi:10.1007/s11306-014-0735-x
• Waterhouse A.L., 2002. Wine phenolics, Annals of the New York Academy of Sciences, 957, 21-36. doi:10.1111/j.1749-6632.2002.tb02903.x
• Wei R., A.B. Ross, M. Su, J. Wang, S.P. Guiraud, C.F. Draper, M. Beaumont, W. Jia and F.P. Martin, 2018. Metabotypes related to meat and vegetable intake reflect microbial, lipid and amino acid metabolism in healthy people, Mol Nutr Food Res, 62, doi: 10.1002/mnfr.201800583.