NMR-based metabonomic approach reveals changes in the urinary and fecal metabolome caused by resveratrol

Accepted Manuscript Title: NMR-based metabonomic approach reveals changes in the urinary and fecal metabolome caused by resveratrol Authors: Gabriela Torres Santiago, Jos´e Iv´an Serrano Contreras, Mar´ıa Estela Mel´endez Camargo, L. Gerardo Zepeda Vallejo PII: DOI: Reference:

S0731-7085(18)31277-9 https://doi.org/10.1016/j.jpba.2018.09.025 PBA 12218

To appear in:

Journal of Pharmaceutical and Biomedical Analysis

Received date: Revised date: Accepted date:

29-5-2018 2-9-2018 12-9-2018

Please cite this article as: Torres Santiago G, Serrano Contreras JI, Mel´endez Camargo ME, Zepeda Vallejo LG, NMR-based metabonomic approach reveals changes in the urinary and fecal metabolome caused by resveratrol, Journal of Pharmaceutical and Biomedical Analysis (2018), https://doi.org/10.1016/j.jpba.2018.09.025 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

NMR-based metabonomic approach reveals changes in the urinary and fecal metabolome caused by resveratrol Gabriela Torres Santiago†, José Iván Serrano Contreras†,*, María Estela Meléndez Camargo‡ and L. Gerardo Zepeda Vallejo†,* †Departamento

de Química Orgánica, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, Prolongación de Carpio y Plan de Ayala s/n, Col. Santo Tomas, C.P. 11340 Delegación Miguel Hidalgo, Ciudad de México, México.

SC RI PT

‡Departamento de Farmacia, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, Av. Wilfrido Massieu, Esq. Cda. Manuel Estampa s/n, Unidad Profesional Adolfo López Mateos, C.P. 07738 Delegación Gustavo A. Madero, Ciudad de México, México

D

M

A

N

U

Graphical Abstract



CC

 

Effects of pure resveratrol (RSV) on the urinary and fecal metabolome of normal female Wistar rats was studied by an NMR-based metabolomics approach. RSV causes significant variations of endogenous metabolites during the first 12 h after its administration. RSV modulates the composition and/or the function of the host gut microbiota. RSV may be involved in the control of energy homeostasis and downregulation of some biomarkers of oxidative stress. The effects of RSV are dose-dependent.

EP



TE

Highlights



A

Abstract

An untargeted NMR-based metabonomics approach was used to evaluate the effects of pure resveratrol (RSV, 50 and 250 mg/kg per os) on the urinary and faecal metabolome of normal female Wistar rats. Multivariate data analysis on both the endogenous and xenobiotic metabotype of RSV provided an insight into its metabolic fate and influence on endogenous metabolites. The xenobiotic trajectory shows that RSV is highly metabolized within the first 12 h, the period of the most significant variation of endogenous metabolites. The results reveal alterations in gut microbiota cometabolites, mainly at the high dose of RSV, such as hippurate, phenylacetyl glycine (PAG), p-cresyl

SC RI PT

glucuronide (p-CG), p-cresyl sulfate (p-CS) and 3-indoxylsulfate (3IS), as well as in osmolytes (creatine, creatinine, taurine and proline betaine). This metabolic variation could mean that RSV modulates the composition and/or function of the gut microbiota as well as its interaction with the host through the gut-microbiome-liver-kidney axis. For instance, RSV may interact with conjugating enzymes present in the intestine and liver. There were also modifications in metabolites of the tricarboxylic acid (TCA) cycle and energy metabolism (2-oxoglutarate, lactate and alanine), and dietderived metabolites (pantothenate and trans-aconitate). These effects of RSV are perhaps related to its capacity to control energy homeostasis, provide renal protection, and downregulate some biomarkers of oxidative stress (e.g., glycoproteins). Such changes contribute to reduced oxidative stress and inflammation, which are associated with RSV-induced biological activity to improve various conditions, including metabolic disorders, obesity, and chronic and cardiovascular diseases. Keywords: NMR; metabolomic profiling; resveratrol; multivariate analysis; urine; feces.

1. Introduction

TE

D

M

A

N

U

Resveratrol (RSV, 3,4',5-trans-trihydroxy-stilbene) a phenolic compound, has been investigated to explore several types of biological activity and is currently used for its antioxidant, anti-cancer and anti-inflammatory properties [1,2]. Despite the widespread use of RSV, the corresponding mechanisms of biological activity are still unclear. Several mechanism are propose including the capacity of RSV to modulated the gut microbiome composition [3]. RSV is rapidly metabolized in the intestine and liver. Although a short amount of this compound remains unreacted, much of it is converted into sulfate and glucuronide conjugates and in less extend into dihydroresveratrol (DHRSV) [4]. While biological activity is generally attributed to non-conjugated RSV, its metabolites are also known to play a relevant role. For instance, the RSV-3-sulfate metabolite inhibits proinflammatory mediators and produces antioxidant effects [5]. Moreover, glucuronide and the reduced metabolites (DH-RSV) have demonstrated equal or similar activity, acting as antioxidant and anti-inflammatory agents as well as inducing cytotoxicity in cancer cell lines [6].

CC

EP

The NMR-based metabonomics approach allows the characterization of the response to a treatment based on the xenometabolome profile and the correlation with the alteration of certain metabolic pathways [7]. Metabolic profiling of urine and feces affords information about the dependence of host health and disease stages on gut microbiota [8]. The metabonomics approach considers the host as a supraorganism, a holistic concept that encompasses changes in the host and the gut microbiota caused by diet, drugs, xenobiotic intake, or any other type of exposure known as the exposome.

A

Although multiple therapeutic claims have been made for RSV, ranging from the prevention of cell oxidation processes, improvements in the management of cancer, and the attainment of glucose homeostasis, its mechanisms are not clearly understood. A recent longitudinal metabolomics study examined urine from rats after administering extracts of P. cuspidatum, one of the major sources of RSV [9]. However, there is a need for research on the combined metabolomic analysis of urine and feces in a model of normal rats treated with pure RSV to discard the effects caused by other components. Hence, the aim of the present study was to use an NMR-based metabonomics approach to evaluate the effect of pure RSV on the metabolism of rats. This approach provides relevant information on the metabolic fate of RSV and its influence on both gut microbiota and host metabolism.

2. Materials and methods 2.1 Chemicals

SC RI PT

RSV was isolated from ResVitáleTM capsules, which were obtained from a local Pharmacy. The content of 30 capsules was poured into 200 mL of ethanol and stirred for 3 h. The extract was filtered and concentrated under reduced pressure in a rotary evaporator to a volume of 50 mL and then partitioned with hexane (3x100 mL). The ethanol layer was evaporated to dryness and purified by column chromatography on silica gel (hexane–ethyl acetate 20:80 v/v) to obtain a yellow solid (10 g). This in turn was dissolved in hot ethyl acetate and saturated with cold hexane to induce precipitation of RSV as a colorless powder (7 g, Rf = 0.26, hexane:ethyl ether 1:3); mp=253 °C [10].

U

After dissolving RSV (50 mg, 0.22 mmol) in THF (25 mL), 5% Pd/carbon catalyst (100 mg, Sigma Aldrich) was added and the mixture was stirred under hydrogen atmosphere at room temperature for 8 h. It was subsequently washed with ethyl acetate and filtered. The residue was evaporated in vacuum to yield a pale-yellow powder (80%), corresponding to DH-RSV. The purity, chemical shifts and multiplicity of RSV and DH-RSV were confirmed by recording and analyzing 1D and 2D NMR spectra (Figure S1).

N

2.2 Animals and sample collection

A

CC

EP

TE

D

M

A

All experiments with rats were carried out in accordance with the Mexican Official Standard (NOM062-ZOO-1999), which establishes technical specifications for the production, care and use of laboratory animals. Female Wistar rats (n=40, 230-280 g) were acclimated for seven days under controlled environmental conditions (temperature, 22-24 ºC; relative humidity, 50-55%; light/dark cycle, 12-12 h, lights on at 8 PM). Animals were fed a standard rodent diet (rodent laboratory chow 5001, Bertwood, MO, USA) and water was available ad libitum throughout the study. After the acclimation period, rats were randomly separated into 3 groups and administered by orally gavage with water (control group, n=14, dose volume of 2 ml/kg of body weight (bw)), and one of two doses of pure RSV dispersed in water, [at 50 (low dose group, n=6) or 250 (high dose group, n=14) mg/kg of bw, with a dose volume of 2 mL/kg of bw]. Rats were food deprived 12 h prior the administration of the vehicle or RSV and their feed was returned 4 h after dosing. After administration rats were housed in individual metabolic cages to collect feces and urine separately (3M12D100/3700M071, Tecniplast, Buguggiate, Va, Italy). Before placing the animals into the metabolic cage, their bladders were emptied by gentle compression on the lower part of the abdomen, and the voided urine was discarded. Subsequently, the total volume of excreted urine for each rat was collected and measured at 6, 12 and 24 h. Feces were collected (at 24 h) only from both the control and the high-dose RSV groups, and the total amount of feces was weighed. All samples were stored at -80 °C to await analysis by NMR. The amount of food consumed by animals was also quantified throughout the study.

2.3 Sample preparation Urine samples were thawed at room temperature and then centrifuged at 15,600 g for 10 min. For each sample, 400 μL of the supernatant were transferred to a 5 mm NMR tube and mixed with 140 μL phosphate buffer [pH 7.4, 0.2 M, containing 0.1 % (w/v) sodium azide (Sigma Aldrich)] and a solution of 60μL TSP ((3-(trimethylsilyl)-2,2,3,3-tetradeuteropropionic acid, Sigma Aldrich) in D2O

(Deuterium oxide, Sigma Aldrich) was added. D2O provided a field frequency lock and TSP a chemical shift reference (1H, δ 0.0).

SC RI PT

The fecal samples were homogenized with 3 mL of 0.2 M phosphate buffer per gram of stool. The homogenate was sonicated, vortexed and centrifuged at 15,600 g for 10 min. For each sample, 540 μL of the supernatant was transferred to a 5 mm NMR tube and mixed with 60 μL of TSP/D2O solution for NMR analysis. For both the urine and feces samples, there was a final TSP concentration of 1 mM in the NMR tube.

2.4 1D and 2D NMR spectroscopy

All the 1H NMR spectra of urine and fecal water samples were acquired at 298K on a Varian NMR System 500 MHz (now Agilent, Santa Clara, CA, US) spectrometer. A NOESY-PRESAT (recycle delay90°-t1-90°-tm-90°-acquisition) pulse sequence was recorded with 128 transients (32 dummy scans) and 64K data points in a spectral width of 6514.4 Hz. All the FIDS were zero filled to 64K data points and multiplied by an exponential weighting function corresponding to a line broadening of 0.3 Hz before Fourier transformation.

D

M

A

N

U

The identification of metabolites in urine and fecal water was performed by consulting databases such as the Chenomx reference library (Chenomx NMR Suite 8.0, Chenomx Inc., Edmonton, Alberta, Canada), the human metabolome database (HMBD, http://www.hmdb.ca), the biological magnetic resonance data bank (BMRB, http://www.bmrb.wisc.edu) and the Birmingham metabolite library (BML-NMR, http://www.bml-nmr.org). Data in the literature provided another resource for supporting metabolite assignments [11]. 2D NMR spectra were acquired from representative samples selected from the control and RSV groups to confirm the metabolite assignments. 1H-Jresolved (JRES) and 1H-13C heteronuclear multiple quantum correlation (HMQC) experiments were performed.

EP

TE

To trace the chemical shifts of the xenometabolome of RSV, a spike-in procedure was performed on a urine sample of an individual from an experimental group. For this purpose, a millimolar solution of DH-RSV and RSV aglycone were prepared in phosphate buffer (as aforementioned). The corresponding NMR spectra were acquired with the previously described protocol.

2.5 Processing of 1H NMR spectra

A

CC

All 1H NMR spectra were phased, baseline corrected and referenced to TSP at δ 0.0 with Agilent VnmrJ 4.2 software. Full resolution 1D 1H NMR spectra (∼17 k data points) were imported into MatLab (R2014a, The MathWorks Inc., Natick, MA) and normalized using the probabilistic quotient method (PQN). To remove the drug-related peaks, the statistical total correlation spectroscopy (STOCSY) algorithm [12] was applied on the RSV data using the driver peak at 1H = 6.8824 ppm. The urinary spectral regions containing RSV-related resonances (δ 2.74-2.88, 3.6-123 3.68, 5.15-5.24, 6.28-6.32, 6.41-6.59, 6.65-7.15, 7.320-7.527) were separated from the rest of the spectra to form RSV and endogen metabotype spectral data sets. These signals were confirmed by data from the spike-in experiments, and the 1D 1H NMR spectra from experimental groups at the first time point. The residual water and urea resonances from urine (δ 4.12–6.47 ppm), and the water resonance from stool (δ 4.07–5.75 ppm) were also removed. For both matrices, the region corresponding to TSP (δ −0.20 to 0.50) was also removed.

2.6. Multivariate statistical analyses

SC RI PT

Multivariate data analysis was performed with SIMCA software (version 14.1 Umetrics, Sweden). Principal component analysis (PCA) and orthogonal partial least square-discriminant analysis (OPLSDA) were performed using Pareto-scaled NMR data. PCA analyses were performed separately on the endogenous and xenobiotic data sets in order to observe their relationship throughout the study. The OPLS-DA models were constructed to investigate the statistical relationship between the NMR data (X) and the dose (Y = 0, 50, 250). To determine the effects on the endogenous metabolome in dependence to the dose, OPLS-DA models were performed to compare the two doses. The OPLS-DA models were built per each timepoint in the following manner: -

All three groups were analyzed simultaneously (control, high dose and low dose).

-

A two-wise comparison between the control and each of the treated groups

-

A two-wise comparison between the low and high dose of RSV

U

In addition, Urine analyses of PCA were performed separately on the endogenous and xenobiotic data sets to observe their relationship throughout the study.

N

For stool, PCA and OPLS-DA were performed only for control rats and rats treated with high dose.

TE

D

M

A

OPLS-DA models were validated via 7-fold cross-validation, analysis of variance of the cross-validated residuals (CV-ANOVA). R2X corresponds to the variation of X correlated with Y, while R2Y indicates the percentage of variation in Y explained by the supervised method. Q2 is an estimate of the predictive ability of the model and goodness of fit, calculated by cross-validation. The CV-ANOVA diagnostic is based on CV for the estimation of independent predictors (OPLS-scores) and predictive residuals to assess the significance of the model given by a p-value lower than 0.05 [13]. The selection of significant metabolites was based on the correlation-scaled predictive loading vector with a cutoff of |p(corr)|≥0.6. The relevant information of the metabotype is stored in p(corr) for OPLS models, whose values range from −1.0 to 1.0. A high absolute p(corr) value indicates that a given metabolite is more abundant in one group than in another and viceversa [14].

EP

3. RESULTS AND DISCUSSION

CC

An untargeted NMR-based metabonomics approach was presently employed. Given the widespread acceptance of RSV as a nutritional supplement, we set out to design an approach that could reveal changes in the urinary and fecal metabotypes of normal rats after dosing them with pure RSV. In contrast to previously described metabolic profiles [9], where RSV was examined as a raw extract, pure RSV was herein administered to discard the likely effects caused by other components.

A

A non-invasive protocol was developed, based on the analysis of the metabolic profile of urine and feces. This strategy allowed for the evaluation of the biomarker variations induced by RSV treatment. Several reports have documented a significant therapeutic potential for RSV at low doses (10-50 mg /kg bw), without toxicological effects at high doses in model animals (200-500 mg/kg bw) [15]. The doses employed presently were selected in accordance with previous descriptions of RSV administration.

3.1. Analysis of urine sample 1H NMR spectra to determine the RSV metabotype

1

SC RI PT

H NMR spectra of rat urine provided relevant information about the metabolites excreted after RSV treatment. From a simple visual inspection of the normalized complete set of 1H NMR spectra of urine (corresponding to the period 0-6 h post-administration of RSV), some predominant signals can be appreciated in the aromatic region and near 3-5 ppm (Figure S2A). This result is compatible with previous studies of RSV metabolism [16], finding that the main biotransformation products are glucuronide and sulfate conjugates. The 1H NMR chemical shifts of the latter compounds may dominate the aromatic region of our data set.

M

A

N

U

RSV is absorbed in the small intestine and partially metabolized to glucuronide and sulfate metabolites. Conjugated metabolites circulating in plasma are reabsorbed by enterohepatic recirculation and deconjugated by bacterial enzymes in the gut, affording non-conjugated RSV and DH-RSV (Figure S3) [16]. Renal excretion is one of the major ways of elimination resveratrol and it metabolites, only a small amount are excreted in feces [4,16]. The most abundant signals of RSV metabolites appeared from 0-6 h post-administration for both doses, although some of the signals were also present in the spectra corresponding to urine samples collected after the 6-12 h interval, especially those from animals given the high dose. NMR signals of the RSV xenometabolome had almost disappeared in the 12-24 h period, showing that biotransformation and excretion of RSV occurs primarily within 12 h after intake. The presence of non-conjugated RSV and DH-RSV was confirmed by spike-in experiments (Figure S4). 1H NMR signal assignments of conjugated RSV glucuronides and sulfates were supported by the literature [5,9]. Moreover, a STOCSY algorithm of the 1H NMR urine spectral data set was constructed to remove RSV-related peaks (Figure S2B). The complete set of signals from the metabotype of RSV was separated to recover the information reflecting the endogenous metabolism.

3.2. Metabolic profiling of xenobiotic and endogenous metabolism

A

CC

EP

TE

D

Based on the urine samples taken at the three timepoints (6, 12 and 24 h), 1H NMR spectrum of the endogenous metabolome and the xenometabolome were subjected to PCA analysis separately to study its dynamic change. The metabolic trajectory of endogenous and xenobiotic metabolites, depicted by changes in the mean PCA score values for 24 hours, allowed for an appreciation of the development of the metabolome profile over time (Figure 1). The xenobiotic trajectory of the high dose (Figure 1A) experienced a great displacement compared to the low dose in the first interval (0-6 h) suggesting that metabolism and excretion of the RSV metabolites was dose-dependent. This can be due the glucuronide conjugates are formed by the phase II enzymes in the liver and small intestine [17] and there is an important contribution from enterohepatic recirculation of RSV, in which the metabolites formed in the small intestine are re-absorbed and metabolized in situ in the liver, and then are excreted in urine [16]. With the high dose of RSV, a metabolic shift from glucuronidation to sulfation occurs, owing to the saturation of the glucuronide pathway. Glucuronidation is the first step of metabolism, followed by sulfation [4,16]. The latter pathway is activated when the capacity of glucuronidation is exceeded, resulting in the metabolism of RSV mainly by a sulfation reaction via sulfonyl transferases (Figure S3). In other words, formation of conjugated metabolites depends on the amount of RSV administered, and the observed displacement in the xenometabolome trajectory is due to the balance of metabolites formed (Figure 1A).

SC RI PT

On the other hand, the endogenous metabolic PCA trajectory shows variation between control rats and those treated with each of the two doses of RSV. With neither dose does the basal metabolic state return, not even during the 12-24 h post-administration interval (Figure 1B). The greatest deviation of the trajectory is during the first 6 h and is more evident at the high dose. This indicates that RSV has a more pronounced metabolic effect at the high dose, which is still evident from 12-24 h. The t1/2 of RSV is known to take place in rats between 1 to 4 h, with an almost negligible excretion of RSV metabolites after 12 h [2]. This finding suggests that the effects on the endogenous metabolome are dominant during the first 12 h and is consistent with the observations of the xenometabolome trajectory, indicating that the greatest formation and excretion of metabolites are in this interval. The accumulation of metabolites and free RSV in the liver and intestine may have prevented the recovery of homeostasis during the current investigation. The formation of different metabolites according to dose may activate different metabolic pathways and promote the corresponding responses, either by interacting with conjugating enzymes or by modifying the composition of the microbiome that can lead to an indirect biological activity of RSV.

U

3.3 Identification of discriminant metabolites in the endogenous metabotype

D

M

A

N

The 1H NMR spectra of urine samples from RSV-treated rats showed metabolic perturbations throughout the study. Differences existed in the levels of urine metabolites and co-metabolites between the control and experimental groups when comparing the three distinct collection times. Such differences were determined by two-wise OPLS-DA analyses by comparing (for each time interval): a) each of the experimental groups (50 or 250 mg/kg) to the control group, b) the low dose to the high dose, and c) the control and both RSV treated groups analyzed at once. The OPLS crossvalidated scores show a clear segregation between the control and RSV-treated groups (Figure 2, Figures S5-S7).

A

CC

EP

TE

According with the three-wise OPLS-DA model (Figure 2), the greatest changes in the urinary profile in response to RSV intake occurred during the first 12 h post-administration. During this period, the urinary levels of hippurate, glycoproteins and taurine were downregulated (Figure 2, Table 1). In the 0-6 h interval, the levels of pantothenate, 2-oxoglutarate, alanine and TMAO were significantly lower for RSV-treated versus control rats, while the levels of DMG and proline-betaine increased. The diminution of creatinine, p-CG, p-CS and 3-IS was seen only with the high dose (Figure S5, Table 1). For the 6-12 h interval, lower levels of creatinine, phenylacetyl glycine (PAG), p-CG and p-CS were found in RSV-treated rats (Figure 2, Table 1). The reduction of 2-oxoglutarate, 3-IS, alanine, p-CG, pCS and pantothenate was detected in the urine of the rats given the high dose (Figure S5, Table 1). For the 12-24 h interval, only the pairwise comparison of the control vs the 250 mg dose of RSV revealed significant differences: a decrease in the levels of hippurate and 2-oxoglutarate, as well as an increase in dimethylglycine (DMG), trans-aconitate and taurine. The pairwise OPLS model comparing the high versus low dose (Figure S6, Table 1) showed the doseeffects where the urinary levels of 2-oxoglutarate, alanine, pantothenate and PAG were significantly lower in the high dose group, while creatine and trans-aconitate were significantly higher. The extracted metabolites displaying the greatest changes are summarized in Table 1. With the two-way OPLS model of the low dose versus control, no significant differences were detected in the 12-24 h interval (Figure S7), meaning that metabolism and the greatest excretion of RSV metabolites occurs

in the first 12 h post-administration. At the dose of 250 mg/kg, RSV continued to influence metabolism even during the 12-24 h interval. Compared to the urine samples, PCA score plots of fecal water portray less separation between baseline versus post-administration data (Figure 3). The fecal levels of lactate and alanine were significantly lower for the RSV-treated versus control animals (Figures 3A and 3B, Figure S8).

SC RI PT

3.4 Metabolomic changes promoted by the administration of RSV treatment

U

Compared to control rats, RSV-treated animals exhibited metabolic changes, evidenced by the biological matrices of urine and fecal water established by their respective 1H NMR spectra. The endogenous metabolites in the samples were identified (Figure S9, Table S1), and the most significant changes were determined by statistical analysis of the 1H NMR data sets (Table 1). Excretion of B-vitamins, osmolytes and metabolites related to the tricarboxylic acid (TCA) cycle diminished in RSV-treated animals, as did the co-metabolites produced by the gut microbiota. In feces, the main change was the downregulation of lactate, pyruvate, and alanine. These results are useful for understanding the biological pathways altered by the administration of RSV and the interaction of this compound with the gut microbiota-host axis.

N

3.4.1. Discriminant urinary metabolites

CC

EP

TE

D

M

A

The comparison of the control and experimental groups reveals that RSV treatment induced a significant decrease in co-metabolites, such as hippurate, PAG, 3IS, p-CG and p-CS, which are produced by the interaction between the host and its gut microbiota during the metabolism of food and xenobiotics. The downregulation of hippurate in rats treated with RSV was the same for both doses (50 and 250 mg/g) during the first 12 h post-administration (Table 1). A different behavior was recently described in the urine samples of rats supplemented with 100 mg/kg of fry extract of P. cuspidatum containing 20% of RSV[9]. Therefore, high levels of hippurate in the urine of rats can be attributed to the high content in the P. cuspidatum extract of polyphenols, including quercetin, emodin and chlorogenic acids [18]. The above polyphenols are transformed into benzoate by both the gut microbiome and the host, and then conjugated with glycine to form hippurate in the liver and to a lesser extent in the kidney. Contrary to the previous reports, the current investigation detected a downregulation of urinary hippurate during the first 12 h post-administration of RSV (Figure 2, Table 1). This behavior can be explained by the fact that RSV is metabolized by bacteria in the intestinal microbiota and by phase II enzymes such as glucuronidases and sulfatases, which represent an alternative pathway for the conjugation of polyphenols with glycine [2].

A

Urinary co-metabolites, hippurate and PAG are modulated not only by the type and amount of food consumption but also by the composition of the intestinal microbiome [19]. In the present study, food consumption was almost constant throughout the time. The increase in food intake observed at 12 h for the high-dose group (Figure S10) can be linked to the upregulation of trans-aconitate that is contained in the cane molasses of the rat chow (Figure S5 and S11). Since the effect on food intake has no relevance to the reduction in urinary co-metabolites of our biological model, RSV and/or its metabolites may have a direct effect on the bidirectional communication between the gut microbiota and the host.

SC RI PT

It is known that the production of phenolic compounds in the intestine directly depends on the type and activity of the gut microbiota. The consumption of antibiotics suppresses bacterial activity and/or changes the microbial composition, thus limiting the excretion of the aforementioned urinary co-metabolites. A similar behavior could be attributed to RSV due to its antimicrobial and antibacterial activity against gram-positive bacteria, and to its regulation of the growth of several types of microorganisms [20]. In this sense, PAG shows reduced urinary levels of phenolic compounds in RSV-treated rats, though in the first 6 h post-administration the effect was only significant at the high dose (Figure S6, Table 1). In the 6-12 h interval, the decrease was significant for both doses compared to the control animals (Figure 2, Table 1). This may owe itself to an alteration in the bacterial ecology caused by RSV and/or its conjugated metabolites, which in the 0-6 h interval are present in greater concentrations at the high dose. Hence, it is possible that RSV influences the gut-microbiota from the beginning of treatment. Moreover, in RSV-treated rats TMAO herein was found at lower levels in the 0-6 h interval witch can be correlated with the observations were RSV was used to treat atherosclerosis by ameliorating levels of TMAO and increasing hepatic neosynthesis of bile acid by regulating intestinal microbiota [21].

A

CC

EP

TE

D

M

A

N

U

The end products of the microbial metabolism of tyrosine are p-CG and p-C, and these cometabolites are conjugated during their passage through the liver and the mucous membrane of the colon [22]. This metabolites are normally excreted by the kidneys and represents a great risk for uremic toxicity and it is associated with cardiovascular disease and oxidative injury[22]. They were herein downregulated during the 24 h of the study, an effect found to be dose-dependent because the decrease was observed in the 6-12 h interval only with the high dose (Table 1). Since glucuronidase and sulfatase (the conjugating enzymes in the intestine and liver) become saturated, RSV and its metabolites accumulate in the large intestine at high doses. Furthermore, these enzymes are responsible for the conjugation of aromatic amino acids to yield co-metabolites such as p-CG, pCS and 3-IS. The competition in the production of RSV metabolites results in a minor excretion of the co-metabolites. 3-IS is an endogenous ligand that at very low concentrations is able to activate the aryl hydrocarbon receptor (AHR), which in turn allows for the gene expression involved in phases I and II of drug metabolism (e.g., CYP1A1/2, CYP1B1, UGT1A1/6 and sulfonylureas (SULT)1A1). Activation of this receptor is related to immune function, inflammatory signaling, drug metabolism, development of cancer and inflammatory bowel disease [23]. In the present study, the decrease in 3IS observed with RSV treatment should limit the activation of the AHR receptor, thereby inhibiting the transcription of CYP1A1 and consequently blocking binding with some potentially carcinogenic ligands. This inhibition of CYP1A1 is known to be dependent on the concentration of liver cell [17]. The fact that the lower levels of 3IS were encountered only in the high-dose group indicates a dosedependent effect (figure S5). Additionally, RSV may inhibit the glucuronidation of therapeutic drugs or dietary components that are substrates for glucuronidase and sulfatase, as polyphenols of the diet [17]. For example, UGTs enzymes, particularly the UGT1A1 and UGT1A9 isoforms, are responsible for the formation of RSV glucuronide in the intestine. Hence, the decrease in p-CG could possibly be related to the inhibitory capacity of RSV over UGT enzymes, thus limiting the metabolism of the dietary components in the intestine [17,24]. RSV treatment also proved to change urinary levels of proline betaine, creatine and creatinine, which may be generated from the metabolism of amino acids (e.g., tyrosine, arginine and glycine) or obtained directly from food (Figure S11). Their concentration is also mediated by the composition of

the intestinal microbiota. These observations together show a strong interaction between resveratrol and the intestinal microbiome, which significantly changes the levels of the aforementioned metabolites.

SC RI PT

Pantothenate (vitamin B5) is another metabolite that underwent a decrease in RSV-treated rats. It is partially decomposed in the intestinal microbiota and then reabsorbed to participate in β-alanine metabolism. Hence, the gut microbiota should certainly play a key role in the regulation of its urinary levels. Pantothenate is a precursor of the synthesis of acetyl coenzyme A (CoA), which is involved in the transformation of pyruvate to citrate in the tricarboxylic acid (TAC) cycle. If the production of CoA declines, the level of succinate and 2-oxoglutarate in urine is diminished and the metabolism of citrate and glyoxylic acid is downregulated [25]. As can be appreciated, RSV acts as an energy metabolism regulator.

N

U

N-Acetyl glycoproteins and O-acetyl glycoproteins (NACs and OACs) were downregulated in RSVtreated rats, suggesting a decrease in oxidative stress. Elevated levels of these metabolites are present in chronically stress-stimulated rats in a model of forced cold-water swimming, indicative of an inflammatory response. High levels of glycoproteins also exist in inflammatory diseases (e.g., liver diseases and cancer) [26]. The decline in glycoproteins could indicate reduced oxidative injury promoted by RSV.

A

3.4.2. Discriminant fecal metabolites

CC

EP

TE

D

M

OPLS-DA of fecal water demonstrated the depletion of alanine and lactate (Figure S8). Lactate is formed from pyruvate by gut microbiota metabolism. A change in the gut microbiota composition can lead to disruption of lactate and fatty acid production. For instance, a microbial imbalance that gives rise to an accumulation of lactate might cause bowel syndrome (SBS) or Crohn’s disease (CD) [27]. Similarly, dysbiosis of the normal bacterial ecology increases the generation of alanine, which perhaps brings about inflammation in colorectal cancer [28]. Presently, the depletion of alanine was detected in urine and stool. Alanine is a product of pyruvate and thereby has a close relation to metabolic pathways such as glycolysis, gluconeogenesis and the citric acid cycle. All these metabolic switches induced by RSV are associated with its anticancer activity through a reduction in the glycolysis of cells. RSV promotes a decrease in lactate dehydrogenase (LDH) activity and mimics the effects of caloric restriction, thus inhibiting glucose absorption and consequently diminishing the generation of lactate [29]. Since RSV and its metabolites mostly accumulate in the intestine, its major anti-cancer and antioxidant activity occurs in the colon.

A

Administration of pure RSV changes the urinary and fecal metabolomes by changing metabolic pathways (Figure 4), the level of osmolytes, the metabolism of vitamins and food components, and the level of some gut microbiota metabolites (e.g., hippurate, p-CG, p-CS, 3-IS and PGA). The persistent decrease in the excretion of co-metabolites may be due to the increased accumulation of RSV and its metabolites in the small intestine [16], possibly leading to changes in the gut microbiota. The xenobiotic trajectory shows that RSV and its metabolites were present in the urine through the 24 hours of the study, during which time resveratrol interacts with the microbiome-gut-liver-kidney axis and promote changes in the endogenous metabolome. Considering the accumulation of RSV and its metabolites in the intestine, the changes observed in co-metabolite production are consistent with the participation of the microbiome in response to RSV treatment. On the other hand, RSV may reduce the generation of Firmicutes bacteria in the intestine, thereby weakening the ability of the

SC RI PT

microbiome to produce tryptophan decarboxylases and diminishing the activity of tyrosinedecarboxylase [3]. This might lead to a decline in the production of co-metabolites in the intestine due to the metabolism of aromatic amino acids (e.g., p-CG, p-CS and 3-IS). Gut microbiota is also recognized as an important regulator of energy metabolism, a function that depends on its microbiome composition. A decrease was found in lactate and alanine as well as urinary metabolites of the TCA cycle. Hence, the influence of RSV on the intestinal microbiota allows it to mimic caloric restriction [30]. Since RSV can regulate energy metabolism and inhibit oxidative stress and inflammation, it might prevent several diseases, especially those involving the intestinal microbiota (e.g., metabolic disorders, obesity, Crohn’s disease, and even cardiovascular diseases).

4.0 Conclusion

U

The current NMR-based metabolomics approach allowed for an in-depth examination of the influence of RSV consumption on the metabolomic profiles of urine and fecal matter. Although the mechanisms of action of RSV have been studied, this is the first report, to our knowledge, of urinary and fecal biomarkers in response to the administration of pure RSV, thus eliminating the biological effects of other compounds (e.g., herbal extracts) in any given natural source of RSV. The present results provide the basis for future research to explore the possible effect of RSV administered over an extended period to prevent some chronic diseases.

N

Acknowledgements

Appendix A. Supplementary data

M

A

This research received financial support from SIP-IPN (Grants 20150758, 20160607, 20170808 and 20171355) and a doctoral scholarship for GT-S (280296) and JIS-C (with international mobility 219509/318260) from CONACyT. GZ thanks CONACYT-Mexico for Grant INFRA-2016 No. 269012.

A

CC

EP

TE

D

Supplementary data associated with this article can be found, in the online version, at http://

REFERENCES

S.S. Kulkarni, C. Cantó, The molecular targets of resveratrol, Biochim. Biophys. Acta - Mol. Basis Dis. 1852 (2015) 1114–1123. doi:10.1016/j.bbadis.2014.10.005.

[2]

J. Gambini, M. Inglés, G. Olaso, R. Lopez-Grueso, V. Bonet-Costa, L. Gimeno-Mallench, C. MasBargues, K.M. Abdelaziz, M.C. Gomez-Cabrera, J. Vina, C. Borras, Properties of Resveratrol: In Vitro and In Vivo Studies about Metabolism, Bioavailability, and Biological Effects in Animal Models and Humans, Oxid. Med. Cell. Longev. 2015 (2015). doi:10.1155/2015/837042.

[3]

J.K. Bird, D. Raederstorff, P. Weber, R.E. Steinert, Cardiovascular and Antiobesity Effects of Resveratrol Mediated through the Gut Microbiota, Adv. Nutr. An Int. Rev. J. 8 (2017) 839–849. doi:10.3945/an.117.016568.

[4]

J.-F. Marier, Metabolism and disposition of Resveratrol in rats: extent of absorption, glucuronidation, and enterohepatic recirculation evidenced by a linked-rat model, J. Pharmacol. Exp. Ther. 302 (2002) 369–373. doi:10.1124/jpet.102.033340.

[5]

Hoshino J,Park EJ, Kondratyuk TP, Marler L, Pezzuto JM, van Breeme RB, Mo s, Li Y, Cushman M, Selective synthesis and biological evaluation of sulfate-conjugated resveratrol metabolites, J Med Chem. 53 (2010) 5033–43. doi:10.1021/jm100274.

[6]

D. Lu, D.D. Ding, W. Yan, R. Li, D.F. Dai, Q. Wang, S. Yu, Y. Li, D.X. Jin, P.D.B. Zhou, Influence of glucuronidation and reduction modifications of resveratrol on its biological activities, Chembiochem. 14 (2013) Pages 1094-1104. https://doi.org/10.1002/cbic.201300080.

[7]

K. Cunningham, S.P. Claus, J.C. Lindon, E. Holmes, J.R. Everett, J.K. Nicholson, M. Coen, Pharmacometabonomic characterization of xenobiotic and endogenous metabolic phenotypes that account for inter-individual variation in isoniazid-induced toxicological response, J. Proteome Res. 11 (2012) 4630–4642. doi:10.1021/pr300430u.

[8]

L.E. Romick-Rosendale, A.M. Goodpaster, P.J. Hanwright, N.B. Patel, E.T. Wheeler, D.L. Chona, M.A. Kennedy, NMR-basedmetabonomics analysis ofmouse urine and fecal extracts following oral treatment with the broad-spectrum antibiotic enrofloxacin (Baytril), Magn. Reson. Chem. 47 (2009). doi:10.1002/mrc.2511.

CC

EP

TE

D

M

A

N

U

SC RI PT

[1]

A

[9]

G. Peron, J. Uddin, M. Stocchero, S. Mammi, E. Schievano, S. Dall’Acqua, Studying the effects of natural extracts with metabolomics: A longitudinal study on the supplementation of healthy rats with Polygonum cuspidatum Sieb. et Zucc., J. Pharm. Biomed. Anal. 140 (2017) 62–70. doi:10.1016/j.jpba.2017.03.015.

[10]

G. Solladié, Y. Pasturel-Jacopé, J. Maignan, A re-investigation of resveratrol synthesis by Perkins reaction. Application to the synthesis of aryl cinnamic acids, Tetrahedron. 59 (2003) 3315–3321. doi:10.1016/S0040-4020(03)00405-8.

[11]

J.I. Serrano-Contreras, I. García-Pérez, M.E. Meléndez-Camargo, L.G. Zepeda-Vallejo, NMRbased pharmacometabonomic analysis of normal rat urine and faeces in response to (±)venlafaxine treatment, J. Pharm. Biomed. Anal. 123 (2016) 82–92.

doi:10.1016/j.jpba.2016.01.044. E. Holmes, R.L. Loo, O. Cloarec, M. Coen, H. Tang, E. Maibaum, S. Bruce, Q. Chan, P. Elliott, J. Stamler, I.D. Wilson, J.C. Lindon, J.K. Nicholson, Detection of urinary drug metabolite (xenometabolome) signatures in molecular epidemiology studies via statistical total correlation (NMR) spectroscopy, Anal. Chem. 79 (2007) 2629–2640. doi:10.1021/ac062305n.

[13]

L. Eriksson, J. Trygg, S. Wold, CV-ANOVA for significance testing of PLS and OPLS® models, J. Chemom. 22 (2008) 594–600. doi:10.1002/cem.1187.

[14]

I. Surowiec, E. Johansson, F. Torell, H. Idborg, I. Gunnarsson, E. Svenungsson, P.J. Jakobsson, J. Trygg, Multivariate strategy for the sample selection and integration of multi-batch data in metabolomics, Metabolomics. 13 (2017) 1–12. doi:10.1007/s11306-017-1248-1.

[15]

W.D. Johnson, R.L. Morrissey, A.L. Usborne, I. Kapetanovic, J.A. Crowell, M. Muzzio, D.L. McCormick, Subchronic oral toxicity and cardiovascular safety pharmacology studies of resveratrol, a naturally occurring polyphenol with cancer preventive activity, Food Chem. Toxicol. 49 (2011) 3319–3327. doi:10.1016/j.fct.2011.08.023.

[16]

E. Wenzel, V. Somoza, Metabolism and bioavailability of trans-resveratrol, Mol. Nutr. Food Res. 49 (2005) 472–481. doi:10.1002/mnfr.200500010.

[17]

S.S. Brill, A.M. Furimsky, M.N. Ho, M.J. Furniss, Y. Li, A.G. Green, W.W. Bradford, C.E. Green, I.M. Kapetanovic, L. V Iyer, Glucuronidation of trans-resveratrol by human liver and intestinal microsomes and UGT isoforms., J. Pharm. Pharmacol. 58 (2006) 469–479. doi:10.1211/jpp.58.4.0006.

[18]

F. Ardelean, E.A. Moacă, C. Păcurariu, D.S. Antal, C. Dehelean, C.C. Toma, S. Drăgan, Invasive Polygonum cuspidatum: Physico-chemical analysis of a plant extract with pharmaceutical potential, Stud. Univ. Vasile Goldis Arad, Ser. Stiint. Vietii. 26 (2016) 415–421.

[19]

H.J. Lees, J.R. Swann, I.D. Wilson, J.K. Nicholson, E. Holmes, Hippurate: The Natural History of a Mammalian − Microbial Cometabolite, J. Proteome Res. 12 (2013) 1527–1546. doi:/10.1021/pr300900b.

[20]

L. Paulo, S. Ferreira, E. Gallardo, J.A. Queiroz, F. Domingues, Antimicrobial activity and effects of resveratrol on human pathogenic bacteria, World J. Microbiol. Biotechnol. 26 (2010) 1533– 1538. doi:10.1007/s11274-010-0325-7.

CC

EP

TE

D

M

A

N

U

SC RI PT

[12]

A

[21]

M.L. Chen, L. Yi, Y. Zhang, X. Zhou, L. Ran, J. Yang, J.D. Zhu, Q.Y. Zhang, M.T. Mi, Resveratrol attenuates trimethylamine-N-oxide (TMAO)-induced atherosclerosis by regulating TMAO synthesis and bile acid metabolism via remodeling of the gut microbiota, MBio. 7 (2016) 1–14. doi:10.1128/mBio.02210-15.

[22]

T. Gryp, R. Vanholder, M. Vaneechoutte, G. Glorieux, P-Cresyl Sulfate, Toxins (Basel). 9 (2017) 1–24. doi:10.3390/toxins9020052.

[23]

J.C. Schroeder, B.C. DiNatale, I.A. Murray, C.A. Flaveny, Q. Liu, E.M. Laurenzana, J.M. Lin, S.C. Strom, C.J. Omiecinski, S. Amin, G.H. Perdew, The uremic toxin 3-indoxyl sulfate is a potent endogenous agonist for the human aryl hydrocarbon receptor, Biochemistry. 49 (2010) 393–

400. doi:10.1021/bi901786x. O.F. Iwuchukwu, S. Nagar, Resveratrol (trans-resveratrol, 3,5,4’-trihydroxy-trans-stilbene) glucuronidation exhibits atypical enzyme kinetics in various protein sources, Drug Metab. Dispos. 36 (2008) 322–330. doi:10.1124/dmd.107.018788.

[25]

M. Disorders, L. Back, U.E. Panel, S. Sciences, C. Isbn, T. Pdf, N.A. Press, N. Academy, Dietary Reference Intakes for Thiamin, Riboflavin, Niacin, Vitamin B6, Folate, Vitamin B12, Pantothenic Acid, Biotin, and Choline, 2001.

[26]

C. Xiao, P. Jia, M. Wu, Y. Zhang, S. Wang, X. Zhao, X. Zheng, Cold water forced swimming stress induced metabolic alterations in rats, Anal. Methods. 6 (2014) 4144. doi:10.1039/c4ay00374h.

[27]

D. Bustos, S. Pons, J.C. Pernas, H. Gonzalez, M.I. Caldarini, K. Ogawa, J.A. De Paula, Fecal lactate and short bowel syndrome, Dig. Dis. Sci. 39 (1994) 2315–2319. doi:10.1007/BF02087644.

[28]

Y. Lin, C. Ma, C. Liu, Z. Wang, J. Yang, X. Liu, Z. Shen, R. Wu, NMR-based fecal metabolomics fingerprinting as predictors of earlier diagnosis in patients with colorectal cancer, Oncotarget. 16 (2016) 16. doi:http://dx.doi.org/10.18632/oncotarget.8762.

[29]

A. Ferraresi, R. Titone, C. Follo, A. Castiglioni, G. Chiorino, D.N. Dhanasekaran, C. Isidoro, The protein restriction mimetic Resveratrol is an autophagy inducer stronger than amino acid starvation in ovarian cancer cells, Mol. Carcinog. 56 (2017) 2681–2691. doi:10.1002/mc.22711.

[30]

E. Morselli, M.C. Maiuri, M. Markaki, E. Megalou, A. Pasparaki, K. Palikaras, A. Criollo, L. Galluzzi, S.A. Malik, I. Vitale, M. Michaud, F. Madeo, N. Tavernarakis, G. Kroemer, Caloric restriction and resveratrol promote longevity through the Sirtuin-1-dependent induction of autophagy, Cell Death Dis. 1 (2010) e10-10. doi:10.1038/cddis.2009.8.

A

CC

EP

TE

D

M

A

N

U

SC RI PT

[24]

FIGURE AND TABLE CAPTIONS Figure 1. The mean PCA trajectory plots for (A) the RSV xenometabolome and (B) endogenous urinary metabotype. The error bars represent the standard error of the mean of the PCA scores per component at each time point. The numbers refer to the sampling time. Key: control group (blue); RSV groups, 50 mg/kg (orange) and 250 mg/kg (red).

SC RI PT

Figure 2. A, C, E) OPLS cross-validated scores for 0-6, 6-12 and 12-24 h post-administration, respectively, and B, D, F) loading plots resulting from the three-way comparison of the 1D 1H NMR spectra of urine samples after 6, 12 and 24 h of treatment. Significant variables are color-coded on the correlation scaled predictive loading vector with |p(corr) | ≥ 0.6. Triangles represent changes observed in both pairwise comparison models that were significant and with the same direction (control vs 50 mg/kg and control vs 250 mg/kg). Blue denotes a higher relative concentration and red a lower relative concentration versus the control. The pairwise comparison of the control vs 250 mg was the only one showing significant changes at 24 h. Key: 2, 2-Oxoglutarate; 6, alanine; 10, creatine; 13, creatinine; 13, DMG; 18, Hippurate; 29, PAG; 30, pantothenate; 31, p-CG; 32, p-CS; 33, prolinebetaine; 36, taurine; 27-28, O-Ac; 24-26, N-acetyl glycoproteins.

N

U

Figure 3. The A) PCA score plot and B) OPLS-DA cross-validated score plot (right), from the fecal water of control rats (blue dots) and animals administered with RSV at 250 mg/kg (red dots).

EP

TE

D

M

A

Figure 4. After oral administration, RSV reaches the colon and is partially metabolized by the gut microbiome. Then it is absorbed and passes through enterohepatic recirculation, where is transformed into its conjugated forms by phase II enzymes in the liver, kidney, and gut. However, a kind of aglycone is accumulated in the small intestine and is still active. These conjugated forms are the dominant RSV metabolites detected in urine. The endogenous metabolic changes can be related with the interaction of RSV and its metabolites with the gut microbiota during the enterohepatic recirculation. Upon consumption, RSV produces changes in both the urinary and fecal metabotypes, decreasing the levels of gut-microbiome-produced metabolites and osmolytes, and the metabolism of food-derived metabolites. This suggests and important effect on the gut microbiome and kidney. There is a downregulation of citric acid cycle metabolites that may be related to the modulation of energy homeostasis.

A

CC

Table 1. Summary of the urinary metabolites perturbed by resveratrol extracted from de OPLS analyses

Figure 1

D

TE

EP

CC

A

SC RI PT

U

N

A

M

D

TE

EP

CC

A

SC RI PT

U

N

A

M

Figure 2

D

TE

EP

CC

A

SC RI PT

U

N

A

M

Figure 3

D

TE

EP

CC

A

SC RI PT

U

N

A

M

Figure 4

Table 1

No.

Control 250 mg/kg dose 6 12 24

Significant metabolites

Control vs 50 mg/kg 6

12

24

250 vs 50 mg/kg dose 6 12 24

↓ ↓

↓

A

CC

EP

TE

D

M

A

N

U

SC RI PT

2-Oxoglutarate ↓* ↓ ↓ ↓ ↓ ↓ 2 3-Indoxylsulfate ↓ ↓ 4 Alanine ↓* ↓ ↓ ↓ ↓ ↓ 6 cis-Aconitate 8 Creatine ↑* 10 Creatinine ↓* ↓ 11 DMG ↑* ↑ ↑ ↑ 13 Hippurate ↓* ↓* ↓ 18 NAC ↓* ↓ 24 OACs ↓* ↓* ↓ 27 Phenylacetyl glycine ↓ ↓* ↓ ↓ ↓ ↓ 29 Pantothenate ↓* ↓* ↓ ↓ ↓ ↓ 30 p-CG ↓ ↓* ↓ ↓ 31 p-CS ↓ ↓* ↓ ↓ 32 Proline-betaine ↑* ↓* ↑ 33 Taurine ↓* ↓* ↑ ↑ ↓ ↑ ↓ 36 trans-Aconitate ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ 37 trigonelline ↓ ↓ 38 Trimethylamine-N-oxide ↓* ↓ ↓ 39 * Significant metabolites observed by OPLS analysis of the three groups (control, and both doses).

↑