Anthocyanin Metabolites in Human Urine after the Intake of New Functional Beverages

Sugar intake abuse is directly related with the increase of metabolic diseases such as type 2 diabetes, obesity, and insulin resistance. Along this line, the development of new beverages using alternative sweeteners could help with combatting the pathophysiological disorders associated to the consumption of sugar. To provide evidence on this issue, in the present work, the bioavailability of anthocyanins was evaluated after the acute ingestion of a new maqui-citrus-based functional beverage rich in polyphenols, and supplemented with a range of sweeteners including sucrose (natural high caloric), stevia (natural non-caloric), and sucralose (artificial non-caloric), as an approach that would allow reducing the intake of sugars while providing bioactive phenolic compounds (anthocyanins). This approach allowed the evaluation of the maximum absorption and the diversity of metabolites excreted through urine. The beverages created were ingested by volunteers (n = 20) and the resulting anthocyanin metabolites in their urine were analyzed by UHPLC-ESI-MS/MS. A total of 29 degradation metabolites were detected: Caffeic acid, catechol, 3,4-dihidroxifenilacetic acid, hippuric acid, trans-ferulic acid, 2,4,6-trihydroxybenzaldehyde, trans-isoferulic acid, and vanillic acid derivatives, where peak concentrations were attained at 3.5 h after beverage intake. Sucralose was the sweetener that provided a higher bioavailability for most compounds, followed by stevia. Sucrose did not provide a remarkably higher bioavailability of any compounds in comparison with sucralose or stevia. The results propose two sweetener alternatives (sucralose and stevia) to sucrose, an overused high calorie sweetener that promotes some metabolic diseases.


Introduction
The abuse of sugar intake in the population is based on the stimulation provided by the sweetness provided by compounds such as sucrose, glucose, and saccharine liquid, through the stimulation of the shared brain reward pathways [1][2][3][4]. This has boosted the consumption of sugar-sweetened beverages, critically contributing to the amount of unhealthy sugar consumed through the human diet [5]. Indeed, the raising consumption of sweeteners occurring during the last decades has been associated to the growing prevalence of metabolic diseases, namely type 2 diabetes, as a result of their contribution to obesity and insulin resistance [6]. In this regard, Bernstein et al., 2012 reported that sugar-sweetened beverages may be related with the increased risk of stroke in humans, as evidenced by epidemiological studies [7].
In order to prevent the deleterious effect of sugar intake in modern societies, global strategies have been focused on lifestyle, suggesting the need to change dietary habits, in parallel to the increase

Anthocyanin Content of Juices
In order to determine the bioavailability of the anthocyanins present in the maqui-citrus beverages created using three separate sweeteners (stevia, sucralose, and sucrose), the anthocyanin composition of the juices was characterized. Thus, the presence of 8 anthocyanins was observed, with the most abundant being Dp 3-O-sambubioside, Dp 3-glucoside, Dp 3,5-O-diglucoside, and Dp 3-O-glucoside, with average concentrations of 3.15, 3.49, and 2.93 mg/100 mL, on average, and respectively. The second most-abundant group of compounds according to their concentration in the juices was the co-eluting Cy 3-O-sambubioside-5-O-glucoside and Cy 3,5-O-diglucoside (1.48 mg/100 mL, on average) and Dp 3-O-sambubioside (1.10 mg/100 mL, on average). The lowest concentration corresponded to Cy 3-O-sambubioside (0.40 mg/100 mL, on average) and Cy 3-O-glucoside (0.55 mg/100 mL, on average). No significant differences between the anthocyanin content of the array of maqui-citrus beverages was observed neither when considering individual nor total anthocyanins (Table 1). Z Cy, cyanidin; Dp, delphinidin; Glc, glucoside; Sam, sambubioside. Y Concentration of anthocyanins in samples presented as means ± SD (n = 3). The quantification of anthocyanins was done on UV chromatograms recorded at 520 nm as cyanidin-3-O-glucoside at 520 nm. N.s., not significant.

Assessment of the Concentration of Anthocyanins Metabolites in Urine Samples
In urine, the quantification of the anthocyanin metabolites excreted was performed on basal urine (0 h), as well as on urine excreted between 0 and 3.5 h, 3.5-12 h, and 12-24 h. The excretion kinetics for all the compounds matched between the volunteers ingesting the three types of juices evaluated, showing the highest concentration for all compounds identified at 3.5 h after the consumption of the beverages ( Figure 1). For this reason, all the concentrations described and analyzed statistically below refer to the timeframe 0-3.5 h after the ingestion.
An additional metabolite of anthocyanins, THBA, was found in urine after the intake of maqui-citrus juices. This metabolite is synthesized as a result of the anthocyanin degradation mediated by aldehyde [25]. Again, its highest total urine concentration, as the sum of THBA-glucuronide and THBA-sulfate, corresponded to the intake of sucralose-sweetened juices Content (mean ± SD, n = 20) of single anthocyanins metabolites (caffeic acid, caffeic acid-glucuronide, caffeic acid-sulfate, catechol-sulfate, 3,4-dihydroxyphenylacetic acid, 3,4-dihydrox yphenylacetic acid-glucuronide, 3,4-dihydroxyphenylacetic acid-sulfate, hippuric acid, hippuric acid-sulfate, trans-ferulic acid-glucuronide, trans-ferulic acid-sulfate, 3,4-dihydroxyphenylacetic acid -glucuronide, 3,4-dihydroxyphenylacetic acid-sulfate, trans-isoferulic acid-sulfate, vanillic acid, vanillic acid-glucuronide, vanillic acid-glucuronide-sulfate, and vanillic acid-sulfate) in basal urine and 3.5,12, and 24-h urine of healthy volunteers after ingesting 330 mL of maqui-citrus juices developed using as sweeteners stevia (∆), sucralose (O), and sucrose ( ). Significantly different bioavailabilities according to an analysis of variance (ANOVA) and duncan's multiple rank test were found at p < 0.001 (***). As mentioned before, no precursor anthocyanins (Dp or Cy) were found in urine. However, an array of metabolites was detected and quantified, being the focus of the bioavailability for these compounds. For example, CA, a common degradation product of anthocyanins that is formed in the small intestine as a result of a pH-mediated breakdown of anthocyanins [24], was identified. Indeed, this phase I reaction could be responsible, to a high extent, for the limited bioavailability of the precursor anthocyanin. The total urine concentration of CA (expressed as the sum of CA and its phase II esterified metabolites) was 4.75 µg/mg creatinine for sucralose, the sweetener that provided the highest bioavailability which surpassed the excretion recorded, when ingesting the beverages with stevia and sucrose added, by 6.7% and 22.4%, respectively. When analyzing the effect of the sweetener regarding the individual metabolites (unesterified CA, CA-glucuronide, and CA-sulfate), in all the cases the highest value corresponded to sucralose (0.03, 0.04, and 4.68 µg/mg creatinine, respectively), relative to the bioavailability when using stevia (a 16.0% lower bioavailability regarding CA), sucrose (a 24.7% lower bioavailability of caffeic acid-glucuronide), and both sucrose and stevia (a 14.6% lower bioavailability, on average, for CA-sulfate). These differences were statistically significant for CA-glucuronide (p < 0.001) and CA-sulfate (p < 0.01) (Figure 1).
In turn, CA is metabolized to TFA and TIFA via enterocyte-based methylation and sulfation/glucuronidation steps [22]. The urine concentration of TFA, expressed as the sum of the concentrations of the individual metabolites (unesterified and phase II metabolites), was 30.00 µg/mg creatinine, on average, for all beverages, with no significant differences found between them. Alternatively, TIFA-sulfate was the only metabolite of TIFA found in urine, and its concentration was 91.16 µg/mg creatinine in the urine of the volunteers who ingested the sucralose-sweetened beverage (which had the highest bioavailability). As for the beverages with stevia and sucrose, the TIFA concentration achieved was 28.7% (significantly) lower, on average. However, further studies are needed to learn more about the influence of sweeteners in the different metabolic pathways, since, to the best of our knowledge, no research on this issue has provided evidence suggesting the modulation of the diverse metabolic pathways as a consequence of the diverse sweeteners.
An additional metabolite of anthocyanins, THBA, was found in urine after the intake of maqui-citrus juices. This metabolite is synthesized as a result of the anthocyanin degradation mediated by aldehyde [25]. Again, its highest total urine concentration, as the sum of THBA-glucuronide and THBA-sulfate, corresponded to the intake of sucralose-sweetened juices (7.71 µg/mg creatinine), while when ingesting juices created using sucrose and stevia, 23.5% and 29.9% lower concentrations were obtained. When analyzing the effect of the sweetener in regard to the bioavailability of individual metabolites (glucuronide and sulfate), in both cases the highest value corresponded to sucralose (0.03 and 7.68 µg/mg creatinine, respectively), which significantly surpassed the urine concentration provided by stevia and sucrose, are related to THBA-glucuronide (by 34.9%, on average; p < 0.001) and relative to THBA-sulfate (by 29.8% and 23.4%, respectively; p < 0.001) (Figure 1).
Another CA degradation pathway gives rise to 3, 4-dihydroxyphenil propionic acid, an intermediate metabolite that ultimately provides DHPAA. This reaction is developed with the participation of mammalian enzymes and colonic microbiota [22] and, according to the results obtained in the present work, it is significantly affected by the separate sweeteners. Hence, the highest total concentration of DHPAA in urine corresponded to volunteers who ingested maqui-citrus juices sweetened with stevia (41.35 µg/mg creatinine), while the intake of beverages created using sucralose and sucrose provided urine concentrations which were 12.2% and 23.4% lower relative to stevia. The analysis of the relative contribution of the diverse DHPAA metabolites found (unesterified DHPAA and DHPAA-sulfate), showed that the highest value also corresponded to stevia (1.35 and 39.06 µg/mg creatinine, respectively), which significantly surpassed the bioavailability provided by sucralose and sucrose, in regard to DHPAA (by 17.3%, on average; p < 0.001) and the bioavailability provided by sucrose, concerning DHPAA-sulfate (by 24.7%; p < 0.001) (Figure 1).
Lastly, the metabolic pathway of anthocyanins transforms DHPAA to CAT, HA, and VA, as well as their phase II conjugates [21,22]. With respect to CAT, the metabolite detected was CAT-sulfate, and its urinary excretion did not show significant differences between sweeteners although the highest values corresponded to sucralose (8934.8 µg/mg creatinine), which showed 22.3% higher values than stevia and sucrose, on average. The urine excretion of HA and HA-sulfate gives rise to the highest concentrations for sucralose (180.02 µg/mg). Regarding these anthocyanin metabolites, stevia and sucrose showed significantly lower amounts (16.9%, on average; p < 0.001) than sucralose. This value was provided almost entirely by HA. Focusing on HA-sulfate, the excretion value was 3.20 µg/mg creatinine for stevia. The remaining sweeteners showed a 41.9% lower excretion, on average. In the case of VA, the total urine concentration calculated as the sum of its unesterified metabolites and their phase II conjugates was 595.98 µg/mg creatinine for stevia. However, when ingesting the juices created using sucralose and sucrose, 16.3% and 33.2% lower amounts were excreted. In this case, this value was provided almost entirely by VA-sulfate. Focusing on VA and VA-glucuronide excretion rates, no significant differences were found between sweeteners. However, the VA-glucuronide-sulfate excretion value was 3.46 µg/mg creatinine for sucralose, which significantly surpassed the bioavailability provided by stevia (a 52.4% lower; p < 0.001) (Figure 1).
The assessment of the urinary excretion of CAT, HA, and VA indicated much higher concentrations relative to the level of their precursors in the juices ingested. This prompted us to hypothesize that the presence of CAT, HA, and VA in urine could be the result of other dietary and endogenous sources, as demonstrated for HA [24]. This could contribute to the detection of very high urine concentrations that could not be otherwise explained by the single ingestion of the anthocyanins provided by maqui to the beverages.
In addition, although the Dp class of anthocyanins was the most-commonly found in the drinks, representing up to 81.0% of the total anthocyanins, on average, for all beverages, its direct metabolite (GA), a fragmentation product of delphinidin glycosides, was not detected [26,27]. This could be the result of the degradation of GA, the most relevant DP metabolite in the human body, to form different molecules such as 4-O-methylgallic acid, pyrogallol, 2-O-methylpyrogallol, and resorcinol [28]. With respect to the relative contribution of the phase II metabolites of the compounds monitored, from the results retrieved, it was observed that sulfated compounds were the most frequently found of the detected compounds, representing about 83.6% of the total excretion, on average.
Moreover, sucralose was the sweetener that provided the highest bioavailability for most compounds (CA, CAT, THBA, and TIFA derivatives, DHPAA-sulfate, HA and VA-glucuronide-sulfate), followed by stevia (DHPAA, HA-sulfate, and VA-sulfate). Sucrose did not provide a remarkably higher bioavailability of any compound as compared with sucralose or stevia, due to intestinal sugar carriers, which may play an important role in flavonoid absorption [29]. A study carried out by Mülleder et al. showed that the addition of sucrose to juices decreased the total excretion of anthocyanins [30].
The urinary metabolites profile, unesterified and their phase II conjugates, found in the volunteer samples after the intake of the maqui-citrus based beverage matched with the degradation metabolites of cyanidin after intake, previously reported by Kay et. al. [22].

Principal Component Analysis
After assessing the urinary metabolites profile detecting, the data were examined further using PCA, a statistical technique that simplifies a complex data set by reducing the number of variables ( Figure 2). This (theoretical) classification contributed to the identification of sweeteners according to the urine metabolites profile of volunteers, thus indicating actual differences between them that will be much useful for the design of future nutritional trials. Hence, 18 components representing the total variance (100.0%) were extracted, with the first six explaining 68.2% of the variance. Of the total variance, 26.2% was explained by PC1, being mostly attributable to CA, CA-Glucuronide, CA-Sulfate, HA-Glucuronide, TFA-Glucuronide, and TIFA-Sulfate. With respect to PC2, the concentrations of DHPAA-Glucuronide, DHPAA-Sulfate, VA-Glucuronide, VA-Sulfate, and VA-Glucuronide-Sulfate were the most relevant contributors. The percentage of the variance explained by PC2 was 24.8%. Finally, PC 3, 4, 5, and 6 explained 7.0%, 4.4%, 3.8%, and 3.1% of variance, respectively. variance, 26.2% was explained by PC1, being mostly attributable to CA, CA-Glucuronide, CA-Sulfate, HA-Glucuronide, TFA-Glucuronide, and TIFA-Sulfate. With respect to PC2, the concentrations of DHPAA-Glucuronide, DHPAA-Sulfate, VA-Glucuronide, VA-Sulfate, and VA-Glucuronide-Sulfate were the most relevant contributors. The percentage of the variance explained by PC2 was 24.8%. Finally, PC 3, 4, 5, and 6 explained 7.0%, 4.4%, 3.8%, and 3.1% of variance, respectively. The PCA developed using the concentrations of bioavailable metabolites of anthocyanins provided clusters alongside the PC1 vs PC2 plot (Figure 2A). Indeed, three clusters were clearly separated, with intermediate (positive and negative) scores for PC1 (stevia and sucralose, -clusters 1 and 2-), negative scores for PC1 (sucrose -cluster 3-). Regarding the ordinates axis (PC2), clusters 1-3 displayed intermediate scores that were more evident for clusters 1 and 2, whilst cluster 3 presented mainly negative values.
The analysis of the samples' distribution, jointly with the eigenvectors informing on the weight of each variable and the correlation between them, indicated the relationship of the observed classification with the abundance of compounds in the samples under evaluation ( Figure 2B). In this sense, the positive contribution of VA to PC1 suggests a close correlation between the samples distribution and clusters through the abscises axis, whilst the growing content of all additional individual metabolites contributed to negative values along this axis ( Figure 2B).

Chemicals and Reagents
The standards used for quantification purposes, caffeic (CA), gallic (GA), 3,4-dihidroxifenilacetic (DHPAA), hippuric (HA), trans-ferulic (TFA), trans-isoferulic (TIFA), vanillic acids (VA), 2,4,6-trihidroxi benzaldehid (THBA), and catechol (CAT), were obtained from Sigma Aldrich (St. Louis, MO, USA). Cyanidin (Cy) 3-O-glucoside and delphinidin (Dp) 3-O-glucoside were purchased from TransMIT (Geiben, Germany). Formic acid and acetonitrile were obtained from Fisher-Scientific (Loughborough, UK). All solutions were prepared with ultrapure water from a Milli-Q Advantage A10 ultrapure water purification system (Millipore, MA, USA). For preparing the maqui-citrus beverages, maqui powder was mixed with citrus juices to obtain the base beverage. Then, the three selected sweeteners were added, in different concentrations, in order to obtain an acceptable taste and to obtain the different beverages analyzed in the present work. The beverages underwent a pasteurization treatment through the application of 85 • C for 58 s. Afterwards, the mixtures were bottled and stored at 5 • C until being consumed by the volunteers.

Experimental Design and Samples Collection
A double-blind, randomized, cross-over clinical study was conducted on overweight individuals (n = 20). The study and protocol were approved by the official Ethical Committee of Clinical Studies (CEIC) from the General University Hospital Morales Meseguer (Murcia, MU, Spain) (Murcia), as well as the Ethical Committee from the Catholic University of Murcia (UCAM) (Murcia, MU, Spain) and registered at ClinicalTrials.gov (NCT04016337). The inclusion criteria were: Individuals who were healthy, non-smoking, overweight (between 24.9 and 29.9 kg/m 2 following WHO criteria), aged 35-55 years, non-dyslipidemic and normotense, with no chronic illnesses and not taking any medication.
After interviewing 46 individuals to assess their eligibility, 26 were not included in the study, as 4 did not fulfil the inclusion/exclusion criteria and 22 declined to participate. A total number of 20 volunteers were included in the study, as can be observed in Figure 3, according to the CONSORT 2010 statement [32]. There were no drop-outs. The basal characteristics of the volunteers are described as mean ± standard deviation (SD) for the following variables: Age (43 ± 8; years), Sex (16 men and 4 women), Weight (85.4 ± 10.4; kg), Height (1.77 ± 0.09; meters), BMI (27.25 ± 2.43; kg/m 2 ) and Fat mass percentage (29.6 ± 8.1).
WHO criteria), aged 35-55 years, non-dyslipidemic and normotense, with no chronic illnesses and not taking any medication.
After interviewing 46 individuals to assess their eligibility, 26 were not included in the study, as 4 did not fulfil the inclusion/exclusion criteria and 22 declined to participate. A total number of 20 volunteers were included in the study, as can be observed in Figure 3, according to the CONSORT 2010 statement [32]. There were no drop-outs. The basal characteristics of the volunteers are described as mean ± standard deviation (SD) for the following variables: Age (43 ± 8; years), Sex (16 men and 4 women), Weight (85.4 ± 10.4; kg), Height (1.77 ± 0.09; meters), BMI (27.25 ± 2.43; kg/m 2 ) and Fat mass percentage (29.6 ± 8.1). Volunteers were provided with written information about the study and all of them signed the informed consent form. All the participants had to follow a strict menu that was low in polyphenols and added sugars, created by a dietitian, for the 2 days before the intervention as well as the day of the intervention. Volunteers were provided with written information about the study and all of them signed the informed consent form. All the participants had to follow a strict menu that was low in polyphenols and added sugars, created by a dietitian, for the 2 days before the intervention as well as the day of the intervention.
The day of the intervention, each volunteer received, on fasting conditions, one of the test drinks (330 mL), with one particular sweetener added (either stevia, sucralose or sucrose). Bottles were codified, thus the composition of each bottle was unknown to the volunteer, as well as to the researchers who provided the drinks and collected and processed the samples. The compliance was controlled as all volunteers were allocated to the same room for the intake of the drink, at the Catholic University of Murcia (UCAM).
After the intake of the test drink, volunteers were transferred to a clinical setting of the Nursing Department at the University, where they remained for the first 3.5 h after the intervention. Afterwards, urine containers were provided with instructions on collection of samples, volume measurement and storing. The day after, the volunteers provided the urine samples.
Urine samples were collected 24 h earlier (0 point), as well as in the following intervals: 0-3.5 h, 3.5-12 h, 12-24 h. For each time-point, urine volumes were recorded and the samples were immediately aliquoted and stored at −80 • C until processing with the analytical procedures developed in the present work, according to the procedure described later in Section 3.4. (Urine sample processing and analysis by UHPLC-ESI-MS/MS). After 15 days, the process was repeated again, with each volunteer ingesting another drink, until all the drinks were consumed by all the volunteers (3 rounds). Figure 4 shows the cross-over study design. Analyses were performed once each period was finished and in the same batch to minimize analytical variations. developed in the present work, according to the procedure described later in Section 3.4. (Urine sample processing and analysis by UHPLC-ESI-MS/MS). After 15 days, the process was repeated again, with each volunteer ingesting another drink, until all the drinks were consumed by all the volunteers (3 rounds). Figure 4 shows the cross-over study design. Analyses were performed once each period was finished and in the same batch to minimize analytical variations. The total volume of each urine interval was recorded to calculate the absolute amounts of the compounds and metabolites excreted in the study period. Also, creatinine content was determined to normalize the concentrations of metabolites in urine as µg compound/mg creatinine, to control for differences in urine volumes.
The creatinine concentration in the urine samples was determined with the A-15 auto-analyzer (Biosystems ® , Pleasenton, CA, USA). To verify the accuracy of the measurement procedure, the Biochemistry Control Urine and calibration of the device was performed, following our internal quality control scheme. The creatinine in the sample reacts with picrate in an alkaline medium, forming a colored complex (Jaffé method), that is monitored and quantified at 505 nm by the auto-analyzer. Concentrations are expressed as mg/dL [33].
The identification and quantification of anthocyanin metabolites was performed by applying the method previously reported by Ludwig et al. (2015) with some modifications [24]. The analysis of samples for the profile and concentration of anthocyanin metabolites was carried out on an The total volume of each urine interval was recorded to calculate the absolute amounts of the compounds and metabolites excreted in the study period. Also, creatinine content was determined to normalize the concentrations of metabolites in urine as µg compound/mg creatinine, to control for differences in urine volumes.
The creatinine concentration in the urine samples was determined with the A-15 auto-analyzer (Biosystems ® , Pleasenton, CA, USA). To verify the accuracy of the measurement procedure, the Biochemistry Control Urine and calibration of the device was performed, following our internal quality control scheme. The creatinine in the sample reacts with picrate in an alkaline medium, forming a colored complex (Jaffé method), that is monitored and quantified at 505 nm by the auto-analyzer. Concentrations are expressed as mg/dL [33].

Statistical Analysis
Quantitative data are presented as mean ± SD of 20 volunteers. Specific differences were examined by an analysis of variance (ANOVA) and a multiple range test (Duncan's test). The data were processed using the SPSS 21.0 software package (SPSS Inc., Chicago, IL, USA) and the level of significance was set at p < 0.05.

Conclusions
The results show an extensive degradation of Cy and Dp glycosides from maqui as they pass through the digestion system with a great variety of metabolites, including their phase II conjugates. However, none of the parental anthocyanins were found.
The results place sucralose as the sweetener which provided the greatest bioavailability for most metabolites. It significantly surpassed the bioavailability provided by stevia and sucrose. However, there are some cases where stevia also had a significantly higher bioavailability than sucralose and sucrose. Nevertheless, sucrose does not provide a higher bioavailability, nor sucralose or stevia, in any of the cases.
Considering the differences of bioavailability between sweeteners, this study proposes two non-caloric sweetener alternatives (sucralose and stevia) in order to reduce the consumption of sucrose, a high calorie sweetener that is directly related with some metabolic diseases such as type 2 diabetes and obesity. This research study contributes new information for the beverage industry, by providing evidence that supports the development of alternative juices, sources of bioactive compounds, using diverse sweeteners that more adequately fit with the current market demands. On the other hand, the different bioavailability described in regard to widely-recognized bioactive compounds would allow further exploring the current interest in them for the prevention of health disturbances with additional nutritional trials.