*Dendropanax morbifera* **Leaf Extracts Improved Alcohol Liver Injury in Association with Changes in the Gut Microbiota of Rats**

## **Taekil Eom 1,**†**, Gwangpyo Ko 2,**†**, Kyeoung Cheol Kim 3, Ju-Sung Kim <sup>3</sup> and Tatsuya Unno 1,2,\***


Received: 21 August 2020; Accepted: 22 September 2020; Published: 24 September 2020

**Abstract:** This study evaluated the protective effects of *Dendropanax morbifera* leaf (DML) extracts in the liver due to excessive ethanol consumption. Our results showed that the ethanol extract had better antioxidant activity than the water extract, likely due to the higher levels of total flavonoid and phenolic compounds in the former. We found that the main phenolic acid was chlorogenic acid and the major flavonoid was rutin. Results from the animal model experiment showed concentration-dependent liver protection with the distilled water extract showing better liver protection than the ethanol extract. Gut microbiota dysbiosis induced by alcohol consumption was significantly shifted by DML extracts through increasing mainly *Bacteroides* and *Allobaculum*. Moreover, predicted metabolic activities of biosynthesis of beneficial monounsaturated fatty acids such as oleate and palmitoleate were enhanced. Our results suggest that these hepatoprotective effects are likely due to the increased activities of antioxidant enzymes and partially promoted by intestinal microbiota shifts.

**Keywords:** *Dendropanax morbifera*; hepatoprotective; leaf extract; liver damage; alcoholism; oxidative damage

### **1. Introduction**

The worldwide increase in alcohol consumption has led to alcoholic liver disease (ALD) accounting for more than 5% of all diseases, with more than three million deaths being estimated to be from ALD. ALDs include alcoholic fatty liver disease, alcoholic hepatitis, and alcoholic cirrhosis, and these can later develop into liver cancer [1,2].

The liver is an organ that absorbs, metabolizes, and stores various substances that enter the body, with 80% of the alcohol consumed by humans being detoxified by the liver. When alcohol is consumed, it is absorbed from the small intestine and metabolized by enzymes such as alcohol dehydrogenase (ADH) and aldehyde dehydrogenase (ALDH) in the liver, but when excessive alcohol consumption exceeds the metabolic ability of ADH and ALDH, the expression of Cytochrome P450 2E1 (CYP2E1) enzymes is induced, leading to alcoholic metabolism [3]. This can lead to production of acetaldehyde or reactive oxygen species (ROS) and toxic substances, causing oxidative damage and inflammation to liver cells, resulting in ALD [4]. ROS produced by excessive alcohol consumption are known to be removed by antioxidant enzymes to prevent oxidative damage caused by alcohol. Antioxidant enzymes such as catalase (CAT), superoxide dismutase (SOD), glutathione *S*-transferas (GST), and glutathione reductase (GR) play a direct role in removing ROS [5].

The intestines are normally home to various types of bacteria, with the term "gut microbiota" referring to the various microorganisms that exist in the human gut. Although the functions of the gut microbiota are not completely known, their effects on human health have recently been actively studied [6,7]. In particular, the gut microbiome is involved in inhibiting the growth of pathogenic bacteria, stimulating the production of total and pathogen-specific mucosal IgA, nutrient production in mucosal cells, the development and regulation of the immune system, and immunological resistance. Thus, the gut microbiome appears to play an essential role in maintaining human homeostasis by engaging in bidirectional interactions with the host [8]. Gut microbiomes vary depending on race, age, dietary habits, and drug use, and the impact of these factors on the health of the host due to changes in the gut microbiome is referred to as dysbiosis [9]. Several recent studies have shown that liver damage caused by alcohol consumption and gut microbiota interact very closely with each other, which is referred to as the gut-liver axis [10–12].

Recently, many studies using various medicinal plant resources have been conducted on foods and medicines that prevent liver damage due to excessive alcohol intake [13,14]. However, studies on changes in the gut microbiome caused by medicinal plants and their effects on the prevention of liver damage are insufficient. *Dendropanax morbifera*, which belongs to the Araliaceae family, is a medicinal plant that grows on the southern coast of Korea and is known to have various physiological functions such as anti-inflammatory, anti-cancer, anti-diabetic, and immune regulatory effects [15–18]. In addition, a previous study conducted by our research team showed that *D. morbifera* extract, with its maximal antioxidant activity, inhibited liver damage caused by oxidative stress [19]. Therefore, this study aims to verify the prevention of alcoholic liver damage by *D. morbifera* extract and to analyze its relationship with the gut microbiome, thereby examining its mechanism of action.

### **2. Materials and Methods**

### *2.1. Extraction*

The *D. morbifera* leaves and stems used in the experiment were purchased from Jeju Hwangchil (Jeju, Korea), and were dried in the shade, crushed, and stored in a refrigerator. To extract the *D. morbifera* leaves and stems, 1 L of 70% ethanol or distilled water was added per 100 g of the sample, followed by three rounds of extraction for 3 h at 80 ◦C. After the extraction, the sample was filtered under reduced pressure, concentrated using a rotary evaporator (Hei-VAP Precision, Heidolph, Schwabach, Germany), and dried using a freeze dryer (SCANVAC, Stockholm, Sweden), then used in the experiment.

### *2.2. Total Polyphenol and Flavonoid Contents*

The polyphenol and flavonoid compounds contained in the extracts were quantified using the Folin-Ciocalteu method and the aluminum chloride method as previously described [20].

### *2.3. Antioxidant Activity*

The antioxidant activity of the extracts was compared and evaluated using the 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical scavenging assay, as well as the total equivalent antioxidant capacities (TEAC) and ferric reducing antioxidant power (FRAP) methods as previously described [19].

### *2.4. HPLC Analysis of Phenolic Acid and Flavonoid Components of D. morbifera Extracts*

Flavonoid and phenolic acid contents were quantified using the prominence HPLC system with an SPD-M20A PDA detector (Agilent infinity 1260 series, Munich, Germany). The analysis was performed with a Triart C-18 column (250 mm × 4.6 mm, 5 μm) from YMC Co., Ltd. (Tokyo, Japan). The flavonoid analysis conditions are column temperature 35 ◦C, flow rate 0.8 mL/min, injection volume 10 μL and detection wavelength 280 nm. The gradient program was designed as follows: Distilled water containing 0.1% trifluoroacetic acid (Sigma-Aldrich, Steinheim, Germany) was mixed with acetonitrile (Sigma-Aldrich) containing 0.1% trifluoroacetic acid with the concentrations increased from 10% to 20% for the first 5 min, held for 20 min, 20% to 25% for 10 min, held for 15 min, 25% to 30% for 5 min, held for 10 min, 30% to 60% for 5 min, held for 5 min, 60% to 80% for 5 min, and held for 5 min (total 80 min).

The phenolic acid analysis conditions are column temperature 45 ◦C, flow rate 1.0 mL/min, injection volume 10 μl and detection wavelength 245 nm. The gradient program was designed as follows: Distilled water containing 0.1% formic acid (Sigma-Aldrich) for 4 min, and mixed with methanol (Sigma-Aldrich) containing 0.1% formic acid with the concentrations increasing from 0% to 15% for 10 min, held for 3 min, increased from 15% to 16.5% for 7 min, 16.5% to 18% for 4 min, 19% to 25% for 2 min, 25% to 28% for 6 min, 28% to 30% for 2 min, held for 3 min, increased from 30% to 40% for 5 min, 40% to 48% for 2 min, 48% to 53% for 5 min, 53% to 60% for 10 min, 60% to 70% for 2 min, and held for 5 min (total 70 min).

### *2.5. Animal Experiments*

Five-week old Sprague Dawley (SD) rats were purchased from Daehan Biolink (Eumsung, Korea) and used after a one-week acclimation period. The animals were kept in a room with a temperature of 20–22 ◦C, a humidity of 50%, and a 12 h light-dark cycle. All experiments were approved by the Jeju University Institutional Animal Care and Use Committee (IACUC) (Approval number: 2016-0056).

In order to see whether the *Dendropanax morbifera* leaf (DML) extracts protected against alcoholic acute liver damage, the experiment was conducted in 8 groups as follows. Dose of ethanol and DML extracts were determined as previously described [19,21]. An overview of the experimental design is outlined in Figure S1. Before the feeding trial, all mice were acclimated for 7 days by feeding normal diet ad libitum. Feeding trial was conducted for 10 days. At the first day, animals were divided into 3 groups and fecal materials were collected. (Control (CTL), fed normal diet during the trial; Ethanol, fed normal diet for 10 days and ethanol was supplied from day 8 to day 10; and DML extracts group, fed normal diet with various concentrations of DML extracts for 10 days and ethanol was supplied from day 8 to day 10). Ethanol (5 g/kg) and DML extracts (100, 300, 500 mg/kg) were administered orally. On the 10th day, 8 h after the last ethanol intake, fecal materials were collected, and rats were sacrificed for dissection.

### *2.6. Serum Biochemical Analysis*

The blood samples obtained were left on ice for 20 min and then centrifuged at 570× *g* for 10 min, after which the upper layer of the serum was collected. The separated serum was stored at −80 ◦C prior to use in the experiments. Serum ethanol and acetaldehyde concentrations were measured and quantified using an ethanol assay kit (Megazyme, Bray, Ireland) and an acetaldehyde assay kit (Megazyme, Bray, Ireland), respectively. Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) concentrations were measured using an automated hematology analyzer.

### *2.7. Antioxidant Enzyme Activity*

For the analysis of antioxidant enzyme activity in liver tissue, a sample of liver tissue was homogenized using a homogenizer in 50 mM phosphate buffer (pH 7.4). The homogenized suspension was centrifuged at 3000× *g* for 20 min at 4 ◦C, and the supernatant was used to measure the enzyme activity. CAT, SOD, GR, and GST were considered as oxidative stress markers [5] and measured as follows: SOD activity was analyzed using the McCord & Fridovich method [22]; and CAT activity was determined using the protocol by Aebi [23]; GST and GR activity was measured using a protocol by Koneru et al. [24]. The amount of protein was quantified using bicinchoninic acid (BCA) to measure the enzyme activity.

### *2.8. Hepatic Histopathological Observation*

For histopathological observations, a portion of the middle lobe of the liver was dissected and fixed in 10% buffered formalin solution (Sigma-Aldrich, Steinheim, Germany) for 24 h. After the fixation, the tissue was embedded in paraffin, cut to a thickness of 5 μm, and stained with hematoxylin (Sigma-Aldrich, Steinheim, Germany) and eosin (Sigma-Aldrich, Steinheim, Germany) (H&E) for observation using an optical microscope with 100-fold magnification.

### *2.9. Analysis of Gut Microbiota*

For microbial community analysis, the V4 hypervariable region of the 16S rRNA gene was amplified, and a library for Illumina Miseq (250 bp × 2) was constructed with two-step PCR. Briefly, first PCR was performed using a KAPA HiFi HotStart ReadyMix PCR kit (Roche, South San Francisco, CA, USA) as follows: 95 ◦C for 3 min, 25 cycles of 95 ◦C for 30 s, 55 ◦C for 30 s, and 72 ◦C for 30 s, and 72 ◦C for 5 min. The obtained PCR products were further purified using a HiAccuBead (AccuGene, Seoul, Korea). The purified PCR products were again subjected for PCR to attach barcode-sequences. After PCR products were purified in the same manner, an equimolar of the final PCR amplicons were pooled and sent for sequencing with MiSeq according to the manufacturer's instructions at Macrogen Inc. (Seoul, Korea).

Sequence data was analyzed using MOTHUR [25]. In brief, raw reads paired-end assembly was done with make.contigs, and then aligned to the SILVA Database [26]. After eliminating the singleton, the pre.cluster MOTHUR subroutine was performed to correct the error for rare sequences. Chimeric sequences were detected using VSEARCH [27]. Taxonomic classification was done using Ribosome database project (RDP version 16) database [28]. Sequences classified to undesired taxa (i.e., Chloroplast and Mitochondria) were removed using remove.lineages MOTHUR subroutine. Clustering was performed with 97% similarity using Opti.clust [29] and designated as operational taxonomic units (OTUs). Number of reads per sample was normalized to 20,000 for downstream analyses. Species richness and evenness were evaluated using Chao [30] and Shannon indices [31], respectively. The non-metric multidimensional scaling (NMDS) analysis was conducted based on the Bray-Curtis distance [32]. The linear discriminant analysis effect size (LEfSe) [33] was used to identify significantly increased or decreased taxa after the DML extracts treatments. Analysis of molecular variance (AMOVA) was applied to test significant difference between microbiota. Metabolic activities were predicted using Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) 2 [34] and abundance comparison was performed using the ALDEx2 package in R software (https://www.r-project.org/) [35]. Spearman correlation analysis was performed to estimate associations between the LEfSe-selected OTUs and ALDEx2-selected predicted metabolic activities.

### *2.10. Statistical Analysis*

Analysis of all data is expressed as ± standard deviation, and significance was determined using ANOVA and verified using the Duncan test. Minitab ver. 17 (Minitab Inc., IL, State College, PA, USA) was used for all statistical analysis.

#### **3. Results**

### *3.1. Total Polyphenol and Flavonoid Contents*

The total polyphenol and flavonoid contents were higher in *D. morbifera* leaf (DML) extracts than stem extracts, and higher in the 70% ethanol (EtOH) extracts than the water extracts (Table S1). Flavonoid content specifically was 1.5 times higher in 70% EtOH than in water for leaf extracts, and seven times higher in 70% EtOH than in water for stem extracts.

### *3.2. Antioxidant Activity*

Three widely used antioxidation experiments were used to measure the antioxidant activity of DML and stem extracts, and the results are shown in Table S2. The DPPH radical scavenging assay is the simplest method of measuring antioxidant activity, wherein electrons or protons are provided to unstable DPPH radicals, and the ability to scavenge the radicals is measured. DPPH radical scavenging assays showed similar radical scavenging activities for water and EtOH extracts, but leaf extracts showed more than twice the radical scavenging activity of stem extracts. FRAP activity measures antioxidant activity based on the reducing power of the sample, and samples with a large reducing power are considered to have excellent antioxidant activity, as they can effectively reduce ROS with high oxidation states. The results from the FRAP activity assays were consistent with those obtained from the DPPH radical scavenging assay, with the leaf extract demonstrating more than twice the reducing power of the stem extract, and the water and ethanol extracts having similar reducing powers. TEAC activity compares and evaluates the antioxidant activities of samples with an equal amount of Trolox, which is used as an antioxidant. The TEAC value showed a similar trend, but the difference between the leaf and the stem extracts was not as significant as observed with the DPPH and FRAP assays. The antioxidant activity evaluated was in the order of *D. morbifera* leaf ethanol extract (DMLEE), *D. morbifera* leaf distilled water extract (DMLDE), *D. morbifera* stem ethanol extract (DMSEE), and *D. morbifera* stem distilled water extract (DMSDE), which was consistent with the total phenolic and flavonoid contents.

### *3.3. Flavonoid and Phenolic Acid Analyses of D. morbifera Leaf Extract*

Content analysis was performed for 18 flavonoids and 12 phenolic acids using HPLC (Figure S2). A smaller number of flavonoids were identified in the water extracts compared to the ethanol extracts, and the flavonoid with the highest content was rutin, a quercetin glycoside (Table 1). The content of rutin was the highest in DMLEE (44.88 ± 0.436 mg/g), followed by DMLDE (15.723 ± 0.005 mg/g), DMSEE (1.085 ± 0.059 mg/g) and DMSDE (0.122 ± 0.095 mg/g). After rutin, taxifolin was the flavonoid compound with the highest content, which was the highest in DMLEE (11.705 ± 0.029 mg/g), followed by DMLDE (4.369 ± 0.014 mg/g), DMSEE (1.593 ± 0.023 mg/g) and DMSDE (1.253 ± 0.218 mg/g).


**Table 1.** Individual flavonoid contents in the *Dendropanax morbifer* leaf and stem extracts (unit: ug/g of extracts).


**Table 1.** *Cont.*

Means ± SD of determinations were made in triplicate experiments.

All 12 Phenolic acids were identified in the extracts, and the major phenolic acids were caffeic acid and chlorogenic acid, a caffeic acid glycoside. However, there was more caffeic acid in the ethanol extracts than the distilled water extracts, and more chlorogenic acid in the distilled water extracts than the ethanol extracts (Table 2). DMLDE (5.165 ± 0.004 mg/g) contained the most chlorogenic acid, followed by DMLEE (2.945 ± 0.119 mg/g), DMSDE (1.532 ± 0.637 mg/g), and DMSEE (0.811 ± 0.024 mg/g). Caffeic acid content was the highest in DMLEE (21.824 ± 1.356 mg/g), followed by DMLDE (13.850 ± 0.024 mg/g), DMSEE (17.446 ± 0.286 mg/g) and DMSDE (11.072 ± 0.178 mg/g).

**Table 2.** Individual phenolic acid contents in the *Dendropanax morbifer* leaf and stem extracts (unit: ug/g of extracts).


Means ± SD of determinations were made in triplicate experiments.

### *3.4. E*ff*ects of DML Extracts on Suppression of Liver Damage*

The effect of DML extracts on liver damage due to alcohol consumption was examined by orally administering DML extract for seven days and then excessive alcohol in a binge rat model [36]. After sacrificing the rats, the activity of serum ALT and AST enzymes were compared as indicators of liver damage, as was liver weight. During the experiment, rats in all groups showed weight gain, and recent studies reported that continuous consumption of excessive alcohol causes fat metabolism in the liver and formation of fatty liver tissue, which results in liver hypertrophy [37]. In this study, the alcohol-administered groups also showed decrease in weight compared to the group that was not administered alcohol. Body and liver weight gain for the group given DML extracts did not show significant difference from that of Ethanol groups (Table S3).

Serum AST and ALT activities were investigated and used as an index to judge the extent of direct liver damage, as AST and ALT are released from the liver to blood in response to liver damage, increasing their enzymatic activity. Results from Figure 1 show that repeated intake of high concentrations of alcohol was shown to increase serum AST and ALT activities (*p* < 0.05 compared to the Control group). However, the groups with DML extract intakes had inhibited AST and ALT activities after alcohol consumption. In particular, the group that consumed the 500 mg/kg ethanol extract had a 40% decrease in AST activity and a 58% decrease in ALT activity compared to the alcohol intake group. The group that consumed the 500 mg/kg distilled water extract showed AST activity that was reduced by 53% and ALT activity that was reduced by 65% compared to the alcohol intake group.

**Figure 1.** Effect of the *Dendropanax morbifer* leaf ethanol and D.W extracts on serum ALT and AST activities in alcohol-fed rat. (**A**) AST activity and (**B**) ALT activity. The data are expressed as the mean ± SD (*n* = 6), and different letters indicate (a > b > c > d) a significant difference at *p* < 0.05, as determined by a Duncan's multiple range test.

The damage to the liver tissue due to excessive ethanol consumption represented by histochemical staining of the liver is shown in Figure 2. The control group with no ethanol consumption showed a normal hepatocyte structure, whereas the group with ethanol consumption showed reduced boundaries between the hepatocytes, and enlargement of hepatocytes due to fat accumulation and infiltration of neutrophils were observed. However, the symptoms and the histopathological changes were improved in the groups that were administered with DML extract.

### *3.5. Reduction of Ethanol and Acetaldehyde in Blood with DML Extract Intake*

Oral administration of excessive alcohol increases absorption of ethanol in the small intestines, and it is metabolized to acetaldehyde through ADH and ALDH in the liver. After liver damage, the decomposition of alcohol by ADH and ALDH decreases, so the concentrations of ethanol and acetaldehyde in blood can be used as an indirect indicator for examining liver damage caused by ethanol. As shown in Figure 3, the group with excessive ethanol intake had 1.8 times higher serum ethanol level (*p* < 0.05) and 2.1 times higher acetaldehyde level than the ethanol group. However, the group being given DML extracts showed a concentration-dependent decrease in blood ethanol and acetaldehyde concentration, and for a given concentration, the administration of the distilled water extracts showed a significant reduction compared to the use of the ethanol extracts.

**Figure 2.** Histopathological evaluation (H&E staining) of rat liver. Liver section images of (**A**) control group, (**B**) ethanol group, (**C**) EtOH + DMLDE 100 mg/kg, (**D**) EtOH + DMLDE 300 mg/kg and (**E**) EtOH + DMLDE 500 mg/kg (**F**) EtOH + DMLEE 100 mg/kg, (**G**) EtOH + DMLEE 300 mg/kg and (**H**) EtOH + DMLEE 500 mg/kg display histopathological changes in liver.

**Figure 3.** Effect of the *Dendropanax morbifer* leaf ethanol and D.W extracts on serum concentration of ethanol and acetaldehyde. (**A**) Ethanol concentration and (**B**) acetaldehyde concentration. The data are expressed as the mean ± SD (*n* = 6), and different letters indicate (a > b > c > d) a significant difference at *p* < 0.05, as determined by a Duncan's multiple range test.

### *3.6. E*ff*ects of DML Extracts on Antioxidant Enzymes in Liver Tissues*

ROS produced by continuous alcohol intake weakens the defense mechanisms associated with oxidative stress in the body and induces oxidative damage. Therefore, the impact of DML extract intake on the activities of antioxidant enzymes related to other antioxidant defenses were evaluated. Figure 4 shows the activities of the antioxidant enzymes, CAT, SOD, GST, and GR. Among the ROS produced in the body, superoxide is converted to hydroxyl radicals by SOD, and the produced hydroxyl radicals are decomposed into water by CAT. CAT activity decreased in the control group compared to the control group, but with no significant difference. In addition, groups with 100 or 300 mg/kg intakes of DML extract did not show differences in CAT activity compared to the ethanol group, but the group with an intake of 500 mg/kg had significantly increased CAT activity. SOD activity was significantly decreased in the ethanol group compared to the control group, and the enzyme activity significantly increased with the intake of DML extracts.

**Figure 4.** Effects of the *Dendropanax morbifer* leaf D.W and ethanol extracts on antioxidant enzyme activity in alcohol-fed rat. (**A**) Catalase activity, (**B**) SOD activity, (**C**) GR activity and (**D**) GST activity. The data are expressed as the mean ± SD (*n* = 6), and different letters indicate (a > b > c > d) a significant difference at *p* < 0.05, as determined by a Duncan's multiple range test.

#### *3.7. Microbiota Shifted by DML Extracts*

Various concentrations of DML extracts affected microbiota similarly except for DMLEE at 300 mg/kg and DMLDE at 100 mg/kg (Figure S3). With a few exceptions, most of the DML treatments did not show significant differences between the different types of extracts nor concentrations of DML extracts (Table S4). Among the treatments, we selected rats treated with highest concentrations of DMLDE and DMLEE for the further gut microbiota analyses, because they showed most distinctive improvements from the liver damages. Results from Figure 5 show that species evenness between the CTL and DMLEE groups was significantly different, while there was no significant difference in species richness. AMOVA and NMDS indicated that no significant microbial community differences between each other across groups before the treatment (Figure S4), but significant differences has detected after the treatment between CTL and Ethanol as well as Ethanol and DML treated samples (Table S4, Figure 5C). These results suggest that alcohol ingestion and DML treatments can significantly shift the microbiota of rats, while extraction methods did not affect the total microbiota shifts.

**Figure 5.** The comparative analysis of gut microbiota using ecological indices and non-metric multidimensional scaling (NMDS). (**A**) Species richness, (**B**) species evenness, and (**C**) NMDS. \* indicates significant difference between CTL and DMLEE.

We investigated taxonomic compositions of rats' gut microbiota at the phylum, family and genus levels (Figure S5). Two major phyla, Bacteroidetes and Firmicutes were detected at the phylum level. At the lower taxonomic levels, the family Prevotellaceae and the genus *Prevotella* were the most abundant in the rats' gut microbiota. We observed the family Sutterellaceae and the genera *Prevotella* and *Romboutsia* were significantly decrease by alcohol consumption. However, taxonomic composition did not cluster samples according to the treatment groups.

Then differential abundance test was performed to investigate significantly increased or decreased OTUs. Results from Figure 6 show that alcohol consumption significantly increased the abundance of members in the family Porphyromonadaceae, the genus *Alloprevotella, Clostridium, Turicibacter* and *Romboutsia*, while decreased the abundance of other OTUs belonging to the family Porphyromonadaceae, the genus *Prevotella*, and *Parasutterella*. DMLDE increased *Allobaculum* and one unclassified OTU belonging to the phylum Candidatus Saccharibacteria, while decreased 12 OTUs including various unclassified OTUs mostly belonging to the family Porphyromonadaceae. On the other hand, there are seven and five OTUs increased and decreased by DMLEE treatments, respectively. Most of the OTUs were the unclassified members of the family Porphyromonadaceae. Two OTUs (Otu0022 and Otu0037) were decreased and one OTU (Otu0012) was increased by both DML treatments, but other OTUs were specific to each treatment. There were no OTUs whose abundance were changed due to alcohol consumption but recovered by DML treatments.

**Figure 6.** The relative abundance increased or decreased bacteria in OTU level in the comparison between ethanol and CTL or DML extracts. Abundance analysis was performed using linear discriminant analysis effect size (LEfSe). (**A**) Ethanol v/s CTL, (**B**) ethanol v/s DMLDE, (**C**) ethanol v/s DMLEE.

PICRUSt2 predicted several metabolic activity changes by alcohol consumption and DML extracts treatments. Results in Figure 7 show that seven and six metabolic activities were predicted to be enriched and depleted due to alcohol consumption, respectively. On the other hand, there are five metabolic activities (biosynthesis of oleic acid, (5Z)-dodec-5-enoate, palmitate, cytochrome c aerobic respiration, and CMP-legionaminate) predicted to be enriched by DML treatments. Beside these, there are four and 12 metabolic activities predicted to be enriched by DMLEE and DMLDE, respectively. Our results showed there were more metabolic activities predicted to be enriched by DMLDE than by DMLEE.

**Figure 7.** Heatmap analysis of predicted metabolic pathways that were significantly increased or decreased in the comparison between ethanol and CTL or DML extracts. Abundance analysis was performed using ALDEx2 (*p* < 0.05).

A total of 23 OTUs that were affected by either of both DML treatments in this study. We investigated associations of these OTUs with 21 predicted metabolic activities that were also affected by DML treatments. Results from Figure 8 show that seven OTUs were significantly correlated with one or more of the 17 predicted metabolic pathways. Three OTUs (Otu0005, Otu0012, and Otu0130) were associated with the most of the predicted metabolic pathways, while the other OTUs were associated with only one or two. The three OTUs are classified to be the genera *Bacteroides* and *Allobaculum* and one unclassified belonging to the phylum Candidatus Saccharibacteria. Predicted metabolic pathways associated with these OTUs were mostly related with monounsaturated fatty acids (oleate and palmitoleate) synthesis pathways.

**Figure 8.** Heatmap analysis of operational taxonomic units (OTUs) associated with the predicted metabolic pathways shifted by DML extracts (*p* < 0.01).

### **4. Discussion**

The liver is an organ that plays an important role in the metabolism of various toxic substances, including ethanol. In the body, 80–90% of alcohol is first decomposed into acetaldehyde by ADH present in liver cells, then metabolized by ALDH to form acetic acid, which finally undergoes complete decomposition via hydrolysis into carbon dioxide and water. Nicotinamide adenine dinucleotide phosphate (NADP), acetaldehyde, and various ROS produced during alcohol metabolism in the liver react with DNA and proteins to act as major mediators of liver damage [38]. Therefore, various antioxidants such as plant extracts have been recently suggested for use as medicines and functional foods to prevent liver damage caused by oxidative stress by inhibiting ROS production due to ethanol [39]. It was also found in several previous studies that DML and branch extracts have strong antioxidant activities, and so their function for reducing liver damage was studied. First, it was confirmed that the antioxidant activity of DML extract was higher than that of the stem extract. Antioxidant activity showed a close relationship to the total phenol and flavonoid content in the extract [17,19]. Total phenolic compounds have a hydroxyl group attached to an aromatic ring, while flavonoids have a basic C6-C3-C6 skeleton [40]. These compounds have a common aromatic hydroxyl group with excellent ability to donate electrons, with donation being stabilized by a resonance effect, leading to excellent antioxidant activity [41]. The antioxidant activity of DML extracts in this study may also be due to the effects of these total phenolics and flavonoids. Previous studies on the effect of harvesting time on the antioxidant activities of DML extracts showed that antioxidant activity was correlated with total phenolic and flavonoid content, and that DML extract had greater antioxidant activity than the branch or bark extracts, which was consistent with the findings of this study [17]. We measured the prevention of liver damage due to ethanol using DML extracts with excellent antioxidant

activity, and found that both ethanol and distilled water extracts prevented liver damage, with the latter showing a much greater effect as opposed to in vitro experiments.

Chronic ethanol intake promotes liver damage by reducing the activity of antioxidants such as glutathione (GSH) and various antioxidant enzymes in the body [5]. The effects of GST and GR related to the metabolism of GSH on the enzymatic activity were confirmed. GSH, which acts as an antioxidant in the body, reacts with radicals via GST to form inactive glutathione disulfide (GSSG), with the cycle then closing via conversion of GSSG back into active GSH by GR [42]. Both GST and GR activities were significantly decreased in the Ethanol group compared to the control group, but the intake of DML extract led to the recovery of their activities up to that of the control group. In addition, DMLDE showed a greater effect than DMLEE groups of the same doses.

Activities of antioxidant enzyme such as SOD, GST, and GR were reduced in the Ethanol group compared to the control group, and the groups that were administered DML extracts showed a concentration-dependent increase in the activities of all antioxidant enzymes. In a study by Bae et al. [21], subjects were given 100 or 300 mg/kg of DMLDE for four weeks and ethanol for the next three weeks, and it was found that the activities of antioxidant enzymes related to glutathione metabolism were reduced in the EtOH-treated group. For the treatment group that received 300 mg/kg of DMLDE, a similar recovery of antioxidant activity was observed to that seen for a positive control group treated with silymarin.

Plant-derived flavonoids and phenolic acids have strong antioxidant properties and are known to prevent inflammatory diseases, cancer, cardiovascular diseases, and various degenerative diseases associated with oxidative stress [13,43]. Previous studies have found that rutin and chlorogenic acid are the main components of *D. morbifera* leaves, which is consistent with the findings of this study [19]. Rutin in particular has been reported to exhibit anti-inflammatory activity against various inflammatory diseases and to protect the liver from various liver damage models [44]. Intake of chlorogenic acid, a type of caffeic acid glycoside that is present in coffee beans, is known to prevent liver damage caused by ethanol consumption by increasing inflammation and antioxidant enzymatic activities in liver [45]. In this study, the rutin, chlorogenic acid, and caffeic acid that are present in the extract are known to prevent liver damage.

In this study, we observed that alcohol consumption increased *Prevotella* and *Parasutterella* and decreased *Clostridium*, *Turcibacter* and *Romboutsia*, which is in line with the previous reports [46–49]. Moreover, alcohol consumption also increased the abundance of some of the gram-negative bacteria (i.e, Porphyromonadaceae and *Alloprevotella*). It has been reported that alcohol consumption may compromise intestinal permeability, allowing the intestinal lipopolysaccharide (LPS) to flow into the liver through the portal vein [50]. In contrast, treatment of DML extracts decreased relative abundance of these gram-negative bacteria. Moreover, it has been reported that DML extracts suppressed the production of LPS-induced pro-inflammatory mediators and cytokines [51] We observed that DMLDE increased the genus *Allobaculum*, known as a butyrate producer [52]. Butyrate was reported to improve alcohol-derived liver damage through enhancing intestinal epithelial barrier function in mice [53,54]. On the other hand, DMLEE increased the abundance of OTUs belonging to the family Porphyromonadaceae and *Bacteroides* and decreased the abundance of other members of the family Porphyromonadaceae. Previous study has reported that gut microbiota of mice ingested with pectin restored *Bacteroides* and completely prevented alcohol-derived liver damage, suggesting that abundance of *Bacteroides* may be associated with the recovery from alcohol-derived liver damage [55]. Although many members of the family Porphyromonadaceae include short chain fatty acids producers such as *Butyricimonas*, *Coprobacter* and *Macellibacteroides* [56–58], this family has been reported to be associated with alcohol consumption and resulting complications in chronic liver diseases [59,60]. To date, due to the lack of genetic information, further studies are required at the genus or species level to investigate roles of these OTUs belonging to the family Porphyromonadaceae.

Our data suggested that isopropanol biosynthesis and succinate fermentation to butanoate may have been largely enriched by alcohol consumption due to secondary alcohol dehydrogenase and succinate semialdehyde produced by *clostridium* strains. Ubiquitous bacteria produce ubiquinone which can be used to catalyze ethanol to a ubiquinol and acetaldehyde. Therefore, there is a possibility that the alcohol consumption may have enriched the metabolic activities of bacterial ubiquinol biosynthesis. Cmp-legionaminate biosynthesis was also depleted by alcohol consumption, but our data also suggested DML treatments may have recovered both of them. The legionaminic acid has been reported to be a virulence-associated cell-surface glycoconjugate in *Campylobacter*, an intestinal pathogenic bacteria [61], however another study has reported that this pathway was also predicted to be enriched in the cecum of probiotics treated broilers [62]. In our data, DMLEE showed enriched aromatic biogenic amine degradation which was depleted by alcohol consumption. Many of aromatic biogenic amines are involved in gastrointestinal pathology such as intestinal bowel syndrome and inflammatory bowel disease [63]. In this study, our results suggested that this biogenic amine degradation could be associated with the enrichment of L-tyrosine degradation. Besides Cmp-legionaminate biosynthesis, there were four more enriched metabolic pathways by both DML treatments, including aerobic respiration I (cytochrome c), oleate biosynthesis IV, (5z)-dodec-5-enoate biosynthesis, and palmitate biosynthesis II. Alcohol consumption is known to be negatively associated with activity of mitochondria [64], thus enriched metabolic activities of cytochrome c suggest that DML treatments may have reduced the damage caused by alcohol consumption. The rest of the three enriched metabolic activities by DML treatments were similar to each other, suggesting these might be predicted from the abundance of similar bacterial species. Oleate and palmitoleate are the major monounsaturated fatty acids in olive oils and its consumption has been reported to improve liver secretory activity, reduce inflammation and oxidative stress [65]. Our data indicated that DMLDE may have enhanced the metabolic activities of biosynthesis of stearate, palmitoleate, and biotin. Stearate is converted to oleate by stearoyl-CoA desaturase 1 (SCD1) and palmitoleate is produced from palmitate, therefore enhancement of these metabolic pathways may bring similar benefits as olive oils do. It has been reported that the amount of SCD1 is substantially low in germ free mice, suggesting that SCD1 is regulated by gut microbiota [66]. Biotin biosynthesis was also predicted to be enhanced by DMLDE. Biotin is a B-complex vitamin that acts as an essential coenzyme. Low biotin levels can occur in elderly individuals, excessive alcohol consumers and smokers. Glucokinase decreased by alcohol consumption can be increased by supplementation of biotin [67,68]. DMLDE also enriched fucose degradation, suggesting increased intestinal bacteria that consume fucose. Fucose is a monosaccharide abundant in mammalian gut, and a degradation of fucose plays an important role in maintaining gut homeostasis resulting in reduced inflammation and controlled hepatic bile acid synthesis [69]. Collectively, our data suggest that DML extracts, especially DMLDE may have enriched metabolic activities predicted to be beneficial to the gut and liver.

Furthermore, we investigated the association of the differentially abundant bacteria with the predicted metabolic activities. We observed that many of the predicted metabolic changes were positively associated with *Bacteroides*, *Allobaculum* and one OTU of Candidatus Saccharibacteria. Previously, it has been reported alcoholic patients have significantly low abundance of *Bacteroides* in human gut [70]. Xiao et al. [71] reported that chronic alcohol consumption shifts gut microbiota which may be a partial cause of alcohol withdrawal-induced anxiety. The decreased abundance of *Allobaculum* was the most obvious shift caused by excessive alcohol consumption. Our results suggest that DML extracts significantly increased the abundance of these genera, suggesting that DML extracts may partially recover the gut microbiota shifted by alcohol consumption. Millman et al. [72] reported that the ingestion of extra virgin olive oils significantly increased these two genera. In addition, they reported that the abundance of *Bacteroides* was found to be associated with loss of plasma triglyceride concentration. Houghton et al. [73] suggested that *Allobaculum* may improve age-related mitochondrial dysfunction through producing butyrate. With the increasing evidence of beneficial effects of these bacteria, our results suggest that DML extracts may improve intestinal functions through shifting gut microbiota impaired by alcohol consumption.

In this study, we investigated antioxidative effects of DML extracts on rats fed ethanol. Our results showed that DML extracts significantly improve the damage cause d by alcohol consumption. However, there are limitations in the study. Metabolomic shifts were only estimated by gut microbiota shifts using PICRUSt2, which could be more accurate if mass-spectrometry metabolomics was applied. In addition, further study should proceed for the identification of bioactive compounds in the DML extracts. Nevertheless, the study presents the effects of DML extracts not only on rats' physiology but also on gut microbiota, providing a fundamental information of how DML extracts may act as functional food.

### **5. Conclusions**

Summarizing these results, we believe that the large amounts of polyphenol compounds contained in DML extracts promote ethanol metabolism in the liver, preventing liver damage. Our results showed hot water extracts showed better improvement in vivo, suggesting that water extracts may provide more bioavailable materials, resulting in enriching beneficial intestinal metabolic activities. These results suggest that DML extracts, especially ones extracted with hot water, may be used as a material for new functional foods or medicines that can prevent and treat liver toxicity caused by ethanol.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-3921/9/10/911/s1, Figure S1: Animal experiment design, Table S1: Total polyphenol and total flavonoid contents of *Dendropanax morbifer* leaf and stem extracts, Table S2: Antioxidant activity of the *Dendropanax morbifer* leaf and stem extracts, Figure S2: HPLC chromatogram of *Dendropanax morbifer* leaf and stem extracts, Table S3: Body weight change and liver-body weight ratio of *Dendropanax morbifer* leaf ethanol and D.W extracts in alcohol-fed rats, Figure S3: Comparison of microbiota according to the concentrations of DML extracts. Numbers indicate the concentrations of DML treatments, Figure S4: Comparison of microbiota before DML extract treatment, Table S4: Analysis of molecular variances (AMOVA), Figure S5: Comparison of the taxonomic composition.

**Author Contributions:** T.E., G.K. and T.U. conceptualization; T.E., G.K. and T.U. data curation; T.E., G.K. and T.U. formal analysis; J.-S.K. and T.U. funding acquisition; T.E., G.K. and K.C.K. investigation; T.E., G.K., K.C.K., J.-S.K. and T.U. methodology; T.U. project administration; T.E., G.K. and T.U. resources; G.K. and T.U. software; T.U. supervision; T.E., G.K. and K.C.K. validation; T.E., G.K. and T.U. visualization; T.E., G.K. and T.U. writing—original draft; T.E., G.K. and T.U. writing—review & editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2016R1A6A1A03012862). This research was also partially supported by the Cooperative R&D between Industry, Academy, and Research Institute funded by Korea Ministry of SMEs and Startups in 2015 (grant number C0297060).

**Acknowledgments:** We are grateful to Sustainable Agriculture Research Institute (SARI) in Jeju National University for providing the experimental facilities.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Review*

## **Insight into Polyphenol and Gut Microbiota Crosstalk: Are Their Metabolites the Key to Understand Protective E**ff**ects against Metabolic Disorders?**

## **Mireille Koudoufio 1,2,3, Yves Desjardins 3, Francis Feldman 1,2,3, Schohraya Spahis 1,2,3, Edgard Delvin 1,4 and Emile Levy 1,2,3,5,\***


Received: 16 September 2020; Accepted: 30 September 2020; Published: 13 October 2020

**Abstract:** Lifestyle factors, especially diet and nutrition, are currently regarded as essential avenues to decrease modern-day cardiometabolic disorders (CMD), including obesity, metabolic syndrome, type 2 diabetes, and atherosclerosis. Many groups around the world attribute these trends, at least partially, to bioactive plant polyphenols given their anti-oxidant and anti-inflammatory actions. In fact, polyphenols can prevent or reverse the progression of disease processes through many distinct mechanisms. In particular, the crosstalk between polyphenols and gut microbiota, recently unveiled thanks to DNA-based tools and next generation sequencing, unravelled the central regulatory role of dietary polyphenols and their intestinal micro-ecology metabolites on the host energy metabolism and related illnesses. The objectives of this review are to: (1) provide an understanding of classification, structure, and bioavailability of dietary polyphenols; (2) underline their metabolism by gut microbiota; (3) highlight their prebiotic effects on microflora; (4) discuss the multifaceted roles of their metabolites in CMD while shedding light on the mechanisms of action; and (5) underscore their ability to initiate host epigenetic regulation. In sum, the review clearly documents whether dietary polyphenols and micro-ecology favorably interact to promote multiple physiological functions on human organism.

**Keywords:** dietary polyphenols; metabolites; oxidative stress; inflammation; epigenetics; microflora; cardiometabolic complications

### **1. Introduction**

Polyphenols, synthesized in a wide variety of fruits, legumes, and herbs, interestingly serve to protect against biotic and abiotic stresses caused by insects and parasites [1–3]. However, much attention is presently given to dietary polyphenols in view of their evident health benefits. These phenolic compounds are intrinsically strong free radical scavengers [4] and exhibit obvious potential in alleviating oxidative stress (OxS)-related illnesses. In this context, epidemiological and animal food intervention studies have established a solid association between the consumption of polyphenol-rich foods or beverages and their preventive influence on complex diseases, including insulin resistance (IR) [5–8], type 2 diabetes (T2D) [9–12], obesity [5,9,13–16], cardiovascular diseases (CVD) [17–22], neurodegenerative disorders [23–26], and cancer [22,27–29]. Recently, polyphenols have also been suggested as plausible adjunctive therapeutic agents for the COVID-19-induced inflammatory storm [30–32].

The growing evidence as to the prevention and management of different disorders by dietary polyphenols has led to multiple studies focusing on their dietary sources [33], efficacy [20], and bioavailability [34–36]. However, much remains to be learned about their potential mechanisms of action. In particular, their dynamic interaction with intestinal microbiota is of utmost importance as ~90% of the amount ingested reach the colon to directly influence microbial ecology and, at the same time, undergo microflora-induced metabolic modifications [37,38]. These interactions represent the specific topic of the present critical review. First, we will discuss the prebiotic action of polyphenols. Second, we will focus on their transformation and physiological impact on the gastrointestinal (GI) tract. Third, we will emphasize the role of their metabolites resulting from their catabolism by colonic bacteria. Fourth, we will analyze their cardiometabolic effects while highlighting their mechanisms of action and underscoring epigenetic regulation. Clearly, the ultimate goal of this review is to show that the health benefits of dietary polyphenols are through the action of their bioactive metabolites. Nonetheless, for the reader's understanding, it is essential to provide a brief introduction, which describes the structure, classification, and bioavailability of polyphenols.

### **2. Polyphenols: Classification, Structure, and Bioavailability**

### *2.1. Classification and Structure of Polyphenols*

Concisely, polyphenols are characterized by at least one aromatic ring and one hydroxyl functional group with evolving structure from a simple molecule to a complex polymer. They are classified into flavonoids and non-flavonoids, according to the intricacy of their structure, number of phenol rings, and carbon skeleton [32,39–41].

Flavonoids present a benzo-γ-pyrone structure containing two aromatic rings (A and B) bound by a 3-carbon bridge (C6–C3–C6). Based on the differences in the C ring, flavonoids can be subdivided into six sub-classes, namely: (1) flavonols (e.g., quercetin, kaempferol); (2) flavones (e.g., luteolin, apigenin); (3) isoflavones (e.g., daidzein, genistein); (4) flavanones (e.g., naringenin, hesperetin); (5) flavanols (e.g., catechins, epigallocatechins); and (6) anthocyanidins (e.g., malvidin, cyanidin). Although polyphenols may be encountered in plants as aglycones, they are generally found as glycoside derivatives, glucoside, galactoside, rhamnoside, xyloside, rutinoside, arabinopyranoside, and finally arabinofuranoside, being the most common [39]. Interestingly, the various tannin functional groups (e.g., hydroxyls) allow them to create non-covalent (hydrophobic and hydrogen) bounds and covalent bonds with proteins and carbohydrates [42–44]. Moreover, tannins are polyphenols that are most involved in binding to proline-rich proteins [45].

### *2.2. Bioavailability of Polyphenols*

### 2.2.1. Factors Affecting Polyphenol Absorption

Polyphenol bioavailability (e.g., alimentary proportion delivered to blood circulation) is dependent on diverse conditions such as their stability, transport, and metabolic behavior. Several intrinsic and extrinsic factors influence the content of plant polyphenols [35]. It is worth mentioning that their relative composition and levels vary extensively among species and between varieties of the same species in association with their genetic background and state of ripeness, which is related to the time of harvest. For example, the content of olive secoiridoid phenolic derivatives decreased when irrigation and ripening increased [35,46–48]. Furthermore, the concentration and variety of phenolic derivatives depend on storage conditions as illustrated by the variability of the phenolic content and total antioxidant capacity of 14 apple and 6 pear cultivars harvested at different periods [46]. The chemical structure is another important intrinsic factor when considering bioavailability. Polyphenols vary widely in molecular weight, secondary and tertiary structures, as well as glycosylation level [49–51]. For instance, resveratrol and quercetin glycosides are absorbed to a lesser extent than the respective aglycones, which influences their metabolic fate through the digestive tract, their absorption and release in the portal circulation, and excretion in urine or feces [52].

With regards to extrinsic aspects, short time storage at low temperature also affects polyphenol content as exemplified by the broccoli loss of its caffeoyl-quinic and sinapic acid contents [53]. Food processing is yet another external factor when considering polyphenol bioavailability. Thermal processing causes diverging effects on polyphenol content and absorption. For instance, Xu et al. [54] reported that thermal processing significantly decreased bean total phenolic content and antioxidant properties, while Khatun et al. [55] reported the opposite by documenting an increase in the same properties when heating spices. These dissimilarities may be explained both by the difference in polyphenol moieties, matrix, and cooking process. The role of the food matrix on the disposition of polyphenols should not be under-evaluated given the interaction between food components ingested simultaneously (fat, proteins, complex carbohydrates) and different polyphenolic compounds likely influencing their interaction with the microbiota and ultimately their absorption. In a rat model, the combination of lecithin and soya bean oil or emulsifiers (sucrose, fatty ester, polyglycerol fatty acid ester, and sodium taurocholate) increased intestinal absorption efficiency of water-dissolved quercetin when these constituents had no effect given separately [56]. Similarly, the administration of hydroxytyrosol as a sole natural product to humans or rats provides a better bioavailability than when combined with refined oil or yogurt, although its handling differed between the two species [57]. Differing digestive tract possesses such as luminal machinery, bile acid secretion, and enterohepatic cycle may explain the differences. The above examples suffice to warrant precautions when interpreting results obtained from different protocols. The several factors mentioned, and others have to be carefully taken into account when conducting studies using cellular models, animals, or clinical investigations, as they could explain variability in outcomes results from numerous reports. In general, polyphenols exhibit a low bioavailability with maximal plasma concentrations reached within 2 to 4 h post ingestion, and an apparent short elimination time with a return to baseline levels within 8 to 12 h [58]. Hence, 24 h-urine collections generally provide a more accurate evaluation of total polyphenol absorption, metabolism, and excretion [59]. These pharmacokinetic properties suggest that only long-term consumption of a variety of polyphenols will affect the health trajectory and support the role of the "Mediterranean diet" in improving population health perspectives.

### 2.2.2. Absorption of Polyphenols and Derivatives

Polyphenol intestinal absorption was indirectly estimated by raised antioxidant defense in the circulation in response to their consumption. The basic principles for their transport and local metabolism are partway established. Figure 1 schematizes the digestion and absorption processes as generally accepted today for polyphenols. Distinction must be made between the digestive and absorption processes of low and high molecular weight polyphenols along the digestive tract as the sites and the efficacy may differ. In decreasing order of absorption kinetics, isoflavones, caffeic and gallic acids lead, followed by catechins, flavanones, and quercetin glucosides. High molecular weight polyphenols such as proanthocyanidins, galloylated catechins, and anthocyanins come last [60]. For instance, the degree of polymerization of procyanidins (dimer B3, trimer C2, and polymers), isolated from willow tree catkins, decreased their absorption through the rat gut barrier and can limit their metabolism by the intestinal microbiota when compared to catechins [61].

**Figure 1.** Integrated view of intestinal dietary polyphenol absorption, luminal transformation, and actions of their relevant metabolites on cardiometabolic disorders. As dietary polyphenols have limited absorption in the stomach and the small intestine, the unabsorbed polyphenols continue their transit to the colon where they are hydrolyzed, demethylated, decarboxylated, dehydroxylated, and ring fissioned by microbiota. Following these processes, microbial metabolites are subjected to phase II metabolism in the colon and liver, and enter the bloodstream to exert their biological effects, which extend to peripheral organs. Unabsorbed polyphenols and metabolites are excreted in feces, and absorbed microbial metabolites are mostly excreted in the urine. Noteworthy, whereas polyphenols improve microbiota composition, diversity and functions through their prebiotic actions, gut microbiota transform them into efficient bioactive regulators capable of optimizing cardiometabolic health in healthy individuals, while alleviating and mitigating the metabolic syndrome in patients. \*Created with Servier Medical Art (A service to medicine provided by Les Laboratoires Servier, 50, rue Carnot—92284 Suresnes Cedex—France.

### 2.2.3. Gastric Uptake

It is well established that plant polyphenols predominantly present as glycosides undergo deglycosylation to their respective lipophilic aglycones, thereby enhancing their absorption through the GI [62,63]. Whether the stomach plays a significant role remains uncertain until today. To our knowledge, there is little direct in vivo evidence on the contribution of the human stomach to the assimilation of phenolic compounds. It needs to be underlined that in vitro simulation or animal models are the most common sources of information. Table 1 summarizes the results obtained in in vitro and in vivo models as well as in humans.


**Table 1.** Gastric handling of polyphenols.





Purified cocoa procyanidin oligomers, incubated in simulated gastric juice, led to the appearance of dimers and monomers in a time-dependent fashion [64,65]. Similarly, olive oil polyphenol glycosides, including oleuropein (glycoside of elenolic acid linked to hydroxytyrosol) underwent rapid and time-dependent hydrolysis, yielding appreciable amounts of free hydroxytyrosol and tyrosol, probably through nucleophilic attack [66]. On the contrary, other studies, using the same in vitro model, along with oral administration of polyphenols to humans, showed stability of a variety of polyphenols in the acidic milieu [67–73]. Interestingly, the incubation of apple phloretin and quercetin for 5 min with native saliva resulted in the production of the respective aglycones, likely by oral bacterial flora as the process was blocked by antibiotics [68]. In the same study, hydroxycinnamic acid derivatives, flavonols, dihydrochalcones, and monomeric flavan-3-ols were stable when incubated with simulated gastric juice at pH 1.8, but procyanidin B2 was degraded. Obviously, the in vitro models suffer from being an inadequate reflection of the gastric fluid, which lacks mucus normally secreted by parietal cells. This is supported by the notable stability of cocoa procyanidin polymers in vivo in the human stomach environment [73]. In terms of in vivo studies, the in situ digestion rat pylorus ligated model shows that polyphenol aglycones are absorbed by the stomach, and suggests the limited role of the stomach in flavonoid glycosides metabolism [74–77]. These results reveal inequality in the behavior of polyphenols in different experimental models and warrant careful analysis to obtain a complete representation.

### 2.2.4. Small Intestine Uptake

### Duodenum

Using the in vitro three-step model simulating the digestive process from the mouth to the small intestine [78], it was shown that the duodenal digestion phase either with static or continuous-flow cellulose membrane-based dialysis containing bile salts and pancreatin resulted in dialyzable (chyme-available for passive absorption into the systemic circulation) and non-dialyzable (digested fraction available for colon) fractions [79]. Profiling the chyme and the non-dialyzable procyanidin content of cocoa liquor post gastric digestion revealed that the high-molecular weight procyanidins (pentamers to nonamers) were hydrolyzed essentially into monomers and dimers. Similar profiles were observed for the duodenal digest. Interestingly, the cocoa liquor duodenal digest contained much higher procyanidin concentrations than the cocoa powder counterpart that the authors attributed to the protective effect of fat micellar structures present in the liquor. More recently, using the same dynamic duodenal model, most of the intact procyanidins isolated from chocolate nibs were retained in the non-dialyzable fraction, which would likely be available for colonic digestion by the microbiota, while smaller molecular weight dimers and trimers were quantified in the dialyzable fraction [67]. On the other hand, the addition of carbohydrate-enriched food resulted in lower duodenal digestibility of procyanidin. Last, with the ex vivo reverted duodenal sac model, the bioavailability of curcumin was found to be significantly enhanced when encapsulated in low- and high-molecular weight polylactic-co-glycolic acid nanoparticles, possibly attributable to faster dissolution following the administration [80]. Overall, these results underpin the effect of the food matrix on the metabolism of polyphenols.

### Jejunum and Ileum

It is well known that under physiological conditions, the small intestine, particularly the jejunum and ileum segments, represents the principal location for the digestion and absorption of dietary lipids. This holds for polyphenols, although with differing efficiency according to their molecular weight and type of glycosylation. Upon reaching the jejunum and ileum, glycosylated polyphenols are either hydrolyzed into their respective aglycones by the membrane-bound brush-border lactase phlorizin hydrolase [81] or transported intact by the enterocyte sodium-glucose co-transporter and hydrolyzed by the cytosolic β-glucosidase [82–84]. The nature of carbohydrate moieties affects the absorption of polyphenols through the small intestine. For instance, whereas glucoside conjugates and their aglycones

are absorbed in the small intestine, those containing rhamnose molecules, [e.g., the flavonols hesperidin (hesperidin-7-*O*-glucosyl-rhamnose) and rutin (quercetin-3-*O*-glucosyl-rhamnose)], must proceed to the colon where rhamnose will be removed by bacterial rhamnosidase [85]. On the other hand, catechin and epicatechin, monomers of the flavonol family, which are often acylated by gallic acid, are readily absorbed by enterocytes without any deconjugation or hydrolysis [86,87].

Caco-2 cells, a human colon carcinoma derived cell line, which upon confluence develops enterocyte-like characteristics, have extensively been used as a model for studying lipid and drug metabolism [88] and transport [89]. They also have served as a model for studying the absorption, transport, and metabolism of polyphenols as detailed in Table 2.

Caco-2 cells have been shown to absorb major dietary hydroxycinnamates and diferulates after de-esterification (phase 1 transformation) and conversion into glucuronate, methyl, and sulfate conjugates (Phase 2 transformation) [90]. In 2006, Corona et al. [66] reported that Caco-2 cells transported hyroxytyrosol from the apical to the basolateral compartment, with the appearance of 3-*O*-methyl–hydroxytyrosol and glutathionyl-hydroxytyrosol in both compartments. They observed a similar transport for tyrosol, without any evidence of further metabolism or transport for oleuropein. The same year, trans-epithelial transport of trans-piceid (3-β trans-resveratrol glucoside) and deglycosylation in trans-resveratrol were reported using human intestinal Caco-2 cell monolayers [82]. The important question in the context of nutrition is whether polyphenols behave similarly when evaluated as single compounds as when assessed in complex mixtures [93,94]. Su et al. [92] partially answered this question by showing that, in Caco-2 cells, (+)-catechin significantly enhanced the cellular uptake and transport of the isoflavone daidzein-8-*C*-glucoside, whereas this compound significantly abated that of (+)-catechin, and that inhibition of efflux pumps increased their uptake. Although valuable for screening polyphenol bioavailability, caution should be used when interpreting the data obtained from the Caco-2 cellular model in terms of predictability. Indeed, even if in vitro smaller phenolic derivatives of hesperidin 2S were absorbed, hesperidin conjugates were the main bioavailable moieties in a randomized human controlled study [69].

The intestinal porcine enterocyte cell line (IPEC), a non-transformed non-tumorigenic permanent intestinal cell line derived from the jejunum is another model that merits attention for two main reasons. First, it maintains its differentiated characteristics, and second, there is a close analogy in results obtained in vivo with the porcine model, the closest to the human GI system [95,96]. To our knowledge, it has seldom been used to study polyphenol absorption. However, a good correlation for the uptake of grape pomace polyphenols was observed between IPEC cells and the duodenum and colon of piglets [97]. In addition, Wan et al. [98] proved functionality of the same in vitro model by showing enhanced the secretion of α-defensins-1 and 2 and reduced the E. coli translocation across a confluent IPEC cells by epigallocatechin-3-gallate, the major polyphenol specie in green tea.






### Colon

The case of the colon bears some singularity in the sense that there is a reciprocal relationship between polyphenols and colonic microbiota. It has been clearly established that the gut microbiome is a key actor in modulating the production, bioavailability, and biological activities of complex low molecular weight phenolic metabolites grouped under the term polyphenol metabolome [99,100]. Rarely, the transport of polyphenols by the colon has directly been evaluated as most of the studies employed intestinal cell models displaying features of the small intestine [101]. It is, however, worthy to mention that T84 colon carcinoma cell line monolayers, incubated with ferulic, isoferulic, cinnamic, and hydroxycinnamic acids, as well as with flavonoids, showed appreciable transport from the luminal to the basolateral side [102]. Measurable levels of ferulic glucuronide and sulfate were detected in the apical and basolateral compartments when supplied at supra-physiological concentrations. Other reports bear mostly on the polyphenol anticancerogenic properties without addressing the uptake and transport per se [103–106].

The enterohepatic cycle is an important mechanism for the bioavailability and disposition of polyphenol aglycones and their metabolites. This is true whether absorbed by the small intestine or the colon (where they undergo in situ phase II enzymatic conversion to their respective methyl, glucuronide, and sulfate conjugates) or delivered to the liver via the portal circulation for further conjugation (before entering the blood stream and being excreted in urine). Alternatively, they enter the bile duct where mixed with bile salts, re-enter the small intestine for either a second-round absorption by enterocytes and transport to the liver, or excretion in the stools [37,107–109].

#### **3. Interaction between Polyphenols and the Colon Microbiota**

Humans are colonized by a wide array of microorganisms referred to as the microbiota that consists of obligate and facultative anaerobes, mainly targeting the colon [110,111]. Their main assignment is to provide assistance to their host in digesting food complex polysaccharides and proteins unprocessed in the upper GI tract [112]. The saccharolytic pathway, mainly active in the proximal colon, yields short chain fatty acids, of which acetic, propionic, and butyric acids are the most abundant, as well as lactic acid, CO2, methane, and ethanol. The second catabolic pathway responsible for protein fermentation, mainly present in the distal colon, leads to metabolite species such as ammonia, various amines, thiols, phenols, and indoles [113,114]. The healthy gut ecosystem consists of the anaerobes *Bacteroidetes* and *Firmicutes*, the latter contributing to more than 90% of the total bacterial species, and *Proteobacteria*, *Verrucomicrobia*, and *Actinobacteria* accounting for the balance [115].

Diet composition has a definite role in the taxonomic and functional profile of the colon microbiota. For example, several reports have shown that high carbohydrate and fiber intakes are related to higher abundance of 3 enterotypes: *Lachnospiraceae*, *Ruminococcaceae*, and *Bifidobacteria* [116–118]. For their part, omnivorous women were shown to have a higher abundance of fecal butyrate producing taxa (*Clostridium cluster XIVa* and *Roseburia*/*Eubacterium*) than vegetarians [119]. Intervention studies have also established that changing from a carnivorous diet to a plant-based diet resulted in a gradual decrease in abundance of *Bacteroides*, a bile-tolerant symbiotic microorganism, and an increase in *Firmicutes* that preferentially metabolizes plant polysaccharides [120]. Furthermore, population food tradition-related differences in microbiota constitution have been reported; those consuming meat-based diets had higher abundances of *Bacteroides* than those traditionally consuming plant-based diets [118,121]. A controlled 10-week-diet intervention, involving overweight volunteers, demonstrated a rapid modification in microbiota species composition despite the inter-individual variation in initial composition [122]. Overall, the results indicate that *Bacteroides*, together with *Alistipes* and *Parabacteroides*, may be the primary proteolytic taxon in the human colon [112].

### *3.1. Polyphenol–Microbiota Interaction*

Of interest, although still poorly understood, evidence is accruing on the influence of dietary polyphenols, perceived as xenobiotics, on the modulation of the colonic microflora and health [123]. They interfere with bacterial cell-to-cell communication and coordinate pathogenic behaviors, two clinically important characteristics with regards to virulence control and wound healing [124,125]. They also sensitize bacteria to xenobiotics and alter membrane permeability. This is exemplified by the sensitization of methicillin-resistant *Staphylococcus aureus* to β-lactam through the (−)-epicatechin gallate-mediated alterations of the bacterium cell wall bilayer, and enhanced release of lipotechoic acid from the cytoplasmic membrane [126]. Polyphenols, acting as prebiotics, have also been shown to modulate gut metabolism and immunity as well as inflammatory pathways through the change of T-cell functions, inhibition of mast cell degranulation, and down regulation of inflammatory cytokine responses [127–130]. Studies in animals and in humans have shown modification of colon microflora by polyphenolic compounds, resulting in growth inhibition of certain bacterial groups while permitting others to flourish [131–133]. Valuable effects of polyphenols, such as those present in red wine, include among others, enhanced abundance of beneficial bacterium taxa such as *Bifidobacterium* and *Lactobacillus*, capable of improving gut barrier protection, *Faecalibacterium prausnitzii* having anti-inflammatory properties, and *Roseburia* as a butyrate producer. Interestingly, the proliferation of these bacteria occurred at the expense of the less desirable *Escherichia coli* and *Enterobacter cloacae* [134]. Most of the important biological actions resulting from phenolic compounds and microbiota are depicted in Figure 2.

### *3.2. Impact of Microbiota on Polyphenol Metabolism*

The transformation of polyphenols into bioactive metabolites is tributary to the colonic microflora species. For example, the *Firmicutes*: *Eubacterium ramulus* and *oxidoreducens*, *Clostridium orbiscidens* and *Flavonifractor plautii* metabolize the flavonols Kaempferol, Quercetin, and Myricetin by *O*-deglycosylation and C-ring fission into a series of metabolites comprising protocatechuic, 2-(3,4-dihydroxyphenyl-acetic, 2-(4-hydroxyphenil)-propionic, 3-hydroxyphenylacetic, 3-(3-Hydroxyphenyl)-propionic, 2-(3-hydroxyphenil)-acetic, 2-(3-dihydroxyphenil)-acetic, and 3-(3,4-dihydroxyphenil)-acetic acids as well as short-chain fatty acids (SCFAs) [83,135,136] (Figure 2). Furthermore, *Enterococcus casseliflavus (Firmicutes)* hydrolyzes sugar moieties from quercetin-3-*O*-glucoside, a process releasing the aglycone and ultimately producing lactic, formic, and acetic acids, together with ethanol [137]. The *Actinobacteria*: *Slackia equolifaciens*, *Slackia isoflavoniconvertens*, *Eggerthella lenta*, *Adlercreutzia equolifaciens*, and *Bifidobacterium* spp metabolize the flavonones Naringenin and Hesperidin through C-ring fission yielding 3-(4-hydroxyphenyl)-propionic, 3-phenylpropionic and hydroxyphenylpropionic acids [135,136]. *Adlercreutzia equolifaciens*, *Paraeggerthella hongkongensis*, *Slackia equolifaciens*, *Slackia isoflavoniconvertens*, *Bacteroides ovatus*, *S. intermedius*, *R. productus*, *E. sp. Julong*, *E. faecium EPI1*, *L. mucosae EPI2*, and *F. magna,* have also been shown convert some the isoflavone daidzein to (*S*)-equol, a nonsteroidal estrogen-like molecule [((3*S*)-3-(4-hydroxyphenyl)-7-chromanol)] [108,137].

**Figure 2.** Beneficial actions resulting from the interaction between polyphenols and intestinal microbiota. Polyphenolic compounds exhibit prebiotic ability, which modifies bacterial composition and function. On the other hand, intestinal microbiota ameliorates intestinal homeostasis and cardiometabolic health by producing polyphenolic metabolites. \* Created with Servier Medical Art.

### **4. Polyphenols and Metabolic Syndrome**

The polyphenol metabolites, produced by the gut microbiota once absorbed and directed to the target tissues and organs, contribute to the metabolic health, through their antioxidant and anti-inflammatory properties, by preventing or reducing the risk of developing several cardiometabolic disorders (CMD), notably the metabolic syndrome (MetS).

The increasing prevalence of childhood MetS along with overweight and obesity is a major public health concern in both developed and developing countries. MetS is a key risk factor for the development of T2D, non-alcoholic fatty liver disease, and CVD in early adulthood [138–140]. MetS is a cluster of interrelated risk factors, including abdominal obesity, dyslipidemia, hypertension, hyperglycemia, and IR, which is triggered by cellular redox imbalance and inflammation as cardinal features [141–143].

In the last two decades, strategies to limit the deleterious effects of the MetS have been focused on diet regimen containing natural fruits, green vegetables, whole grains, legumes, probiotics, vitamin C, vitamin E, and ω-3 polyunsaturated-fatty acids [20,144–148]. Likewise, prebiotics, including polyphenols with their prebiotic potential, have also demonstrated promising effects on IR, glycemia, lipid profile, and CVD [8,134,149]. Other multiple studies, involving animal models, have consistently shown that administration of polyphenols, extracted from various sources, increased insulin sensitivity (via measurement of homeostatic model assessment for insulin resistance index, glucose intolerance, or insulin tolerance tests) and decreasing circulating free-fatty acid, triglyceride, cholesterol, C-reactive protein, resistin, and leptin concentrations [150–154]. At the clinical level, a parallel, double-blinded controlled and randomized 6-week dietary intervention demonstrated that strawberry and cranberry polyphenols improved insulin sensitivity in overweight and obese nondiabetic, insulin-resistant human subjects [8]. There was not a significant improvement of other cardiometabolic risk factors, which is probably due to the length of the intervention. Nevertheless, these in vivo results show the potential clinical preventive and therapeutic impact

of polyphenols on human diseases. Considering that impaired redox potential and inflammatory processes are the basis of MetS, polyphenols, identified for their antioxidant and anti-inflammatory properties [155,156], appear as suitable candidates for preventing its development. The review of the mechanisms linking polyphenols and alleviation of these features follows.

### *4.1. Antioxidative E*ff*ects of Polyphenols*

Polyphenols are known as major contributors to the fruit total antioxidant activity [157]. They do so by donation of an electron or hydrogen atom to reactive oxygen, nitrogen, and chlorine species based either on hydrogen atom transfer or single electron transfer by proton transfer [158]. Reacting with the inner side of the plasma membrane hydrophobic compounds, polyphenols impede lipid and protein oxidation, thus protecting the structure, fluidity, and function of the cell membrane [159]. Polyphenols also chelate transition metals as Fe2+, thus directly reducing the Fenton reaction and preventing oxidation by highly reactive hydroxyl radicals [160].

Reactive oxygen species (ROS) play leading roles in provoking pan-cellular inflammation. Their action is mediated by the activation of powerful transcription factors such as nuclear redox factor-2 (Nrf2), nuclear factor-kappa B (NF-κB), and activator protein 1 [161]. Under unstressed conditions, inactive Nrf2 is linked to its cytoplasmic regulator actin-anchored Kelch-like ECH-associating protein 1. Due to reactive species accumulation in cells under OxS, Nrf2 dissociates from Kelch-like ECH-associating protein 1, translocates into the nucleus where it modulates antioxidant-responsive elements-mediated transcription cytoprotective genes [162,163]. Phenolic compounds are able to enhance protective pathways by activating Nrf2. For example, flavonoids induce Nrf2 translocation to the nucleus where it forms heterodimers with the leucine zipper-like transcription factors small musculo-aponeurotic fibrosarcoma proteins, then bind to antioxidant-responsive elements to activate target genes such as Phase II enzymes involved in detoxification [164]. In addition, Zhang et al. [165], using human breast cells, hepatic human liver cancer cells, and mouse Hepa-1 cells, demonstrated cell-dependent agonist/antagonist effects of several flavonoids on the regulation of aryl-hydrocarbon receptor-mediated signalling pathways. Polyphenols may also suppress pro-oxidative response by down-regulating the synthesis of the pro-inflammatory interleukin-(IL)-1β, tumor necrosis factor-alpha (TNFα), and interferon-gamma. They do so by modulating NF-κB and mitogen-activated protein kinase signaling pathways [166]. Moreover, phenolic compounds can display genoprotective effects, thereby preventing/reversing the progression of various CMD diseases [167]. Through the protection of DNA against OxS damages, some types of polyphenols can promote metabolic health [168–171]. For example, epigallocatechin gallate and resveratrol exhibit high capacity to decrease DNA double strand breaks, chromosome loss, and DNA deterioration response in H2O2-induced genotoxicity in cells [172,173]. A human supplementation trial has also shown the ability of green tea polyphenolic antioxidants to increase the resistance of DNA to oxidation damages [174].

### *4.2. Anti-Inflammatory E*ff*ects of Polyphenols*

The crosstalk between oxidative and inflammation signalling pathways is important for cell homeostasis and survival. Indeed, overproduction of ROS has been shown to stimulate the release of pro-inflammatory cytokines. Reciprocally, NF-κB controlled genes are able to regulate ROS production [175,176]. Various studies, particularly those oriented toward cancer, revealed that polyphenols are able to control chronic inflammatory responses at the level of cytokine production, NF-κB-mediated gene expression, and the release of the tumor-suppressing transforming growth factor-beta [177–179].

Furthermore, closely related to the MetS, our group has more recently demonstrated, in post-confluent Caco-2/15 cells, that polyphenols extracted from dried apple peels (DAPP), containing phenolic acids, flavonol glycosides, flavan-3-ols, and procyanidins prevented iron-ascorbate-mediated lipid peroxidation and counteracted lipopolysaccharide-mediated inflammation as shown by the down regulation of the cytokines TNFα and IL-6, lessening of

cyclooxygenase-2 expression, and production of prostaglandin E2 via the decline of NF-κB. Concomitantly, the alleviation of inflammation was accompanied by the induction of Nrf2 (orchestrating cellular antioxidant defenses and maintaining redox homeostasis) and peroxisome proliferator-activated receptor gamma coactivator 1-alpha (the "master controller" of mitochondrial biogenesis) [180]. We later demonstrated that DAPP prevented and alleviated dextran sodium sulfate-induced intestinal inflammation in mice, stimulated antioxidant transcription factors, and improved mitochondrial dysfunction. They also decreased lipid peroxidation and up-regulated antioxidant enzymes, while decreasing the activity of myeloperoxidase, the expression of cyclooxygenase-2, and the production of prostaglandin E2. Moreover, DAPP partially restored mitochondrial redox homeostasis, fatty acid β-oxidation, ATP synthesis, apoptosis, and regulatory mitochondrial transcription factors [181]. As DAPP in parallel decreased the relative abundance of *Peptostreptococcaceae* and *Enterobacteriaceae* bacteria [182], it is possible that the modifications of intestinal microbiota brought about the amelioration of oxidative and inflammatory processes.

### *4.3. Epigenetic Control of Polyphenols*

Epigenetic modification of gene expression is extremely important in the modulation of the OxS and inflammatory pathways briefly described above. Such alterations have been associated with abnormalities in various metabolic pathways [183]. Epigenetic control includes methylation of CpG rich regions, post-transcriptional modulation of chromatin histone/non-histone, and micro-RNAs (miRNAs) regulating gene expression [184–186]. Importantly, while these alterations may persist for cells lifespan and may have inter-generational effects, they are reversible and have thus become prime targets for interventions in the treatment ofMetS [187]. Unfortunately, data on these mechanisms, particularly those bearing on DNA methylation and histone acetylation, are scarce. Table 3 shows that, in both animal models and human studies, polyphenol consumption decreases obesity, IR, and hypertension. These metabolic improvements are associated to epigenetic modifications, including increased DNA methylation, and histone methylation and acetylation.

Interestingly, reports involving miRNAs are more profuse. Table 4 summarizes the major findings obtained regarding the modulation of miRNAs and their effect on the expression of specific genes. It can be appreciated that polyphenols, independently of their origin, modulate miRNAs, especially controlling OxS and inflammation pathways in different cell models. In animal models, they also regulate miRNAs expression involved in the control of inflammation, obesity, lipogenesis, and energy expenditure. One human study shows that polyphenol consumption modulates miRNAs expression related to inflammation processes.



Tx: treatment; HFD: high-fat diet; HFHSD: high-fat high-sucrose diet; miR: micro-RNA; IR: insulin resistance; AT: adipose tissue; Leuk: leukocytes; Mtases: methylases; ND: no data; ↑: increase; ↓: decrease.






Tx:line-19; HUVEC:oxide synthase;hydrolase; HO-1: heme oxygenase-1; Ref-1: redox factor-1; TLR-2: toll like receptor-2; ARE: antioxidant response element; CRP: C-reactive protein; MCP-1: monocyte chemo-attractant protein-1; T2D: type-2 diabetes; HT: hypertensive; HPBMCs: human peripheral blood mononuclear cells; LRRFIP-1: leucine-rich repeat flightless-interacting protein-1; HaCat: human keratinocyte cell line; IG: intra-gastric; ASC: apoptosis-associated Speck-like protein Casp1: caspase-1; SGBS: Simpson–Golabi–Behmel syndrome; ALP: alkaline phosphatase; CCL-3: chemokine (C-C motif) ligand-3; TC: total cholesterol, TAG: triacylglycerol; LDL-C: low-density lipoprotein cholesterol; TAC: total antioxidant capacity; MDA: malondialdehyde; SOD: superoxide dismutase; CAT: catalase; GPX: glutathione peroxidase; GSH: glutathion, OA: osteoarthritis; COX2: cyclooxygenase-2; PGE2: prostaglandin-E2; EGCG: epigallocatechin-3- *O*-gallate; MMP-2: matrix-degrading enzyme metalloproteinase; NADPH oxidase: nicotinamide adenine dinucleotide phosphate oxidase; CXCL-10: C-X-C motif ligand 10; M-CSF: macrophage colony-stimulating factor; PPARγ: peroxisome proliferator-activated receptor; NFE2L2: nuclear factor, erythroid 2 like 2; TRXR2: thioredoxin reductase 2; MnSOD: manganese superoxide dismutase; TXNIP: thioredoxin-interacting protein; NLRP3: NOD-like receptor (NLR) family, pyrin domain containing 3; PPARα: peroxisome proliferator activated receptor-α; CPT-1: carnitine palmitoyl transferase-1; SREBP-1: sterol regulatory element binging protein 1; SCD-1: stearoyl-CoA desaturase-1; vEGF: vascular endothelial growth factor; eNOS: endothelial nitric oxide synthase; PGC-1α: peroxisome proliferator-activated receptor coactivator 1α; GLUT-4: glucose transporter-4; VCAM-1: endothelial vascular cell adhesion molecule-1; PXR: pregnane X receptor; GST: glutathione S-transferase; MDRP1: multidrug-resistant protein 1, CYP1A1: cytochrome P450; SIRT1: sirtuin 1; ↑: increase; ↓: decrease.

### **5. Polyphenol Metabolites**

In the first part of the present review, we have defined the polyphenols, described their structure, reported their digestion and absorption, elaborated on their mode of action and potential benefits, emphasized their epigenetic regulation, and pointed out their interaction with intestinal microbiota. The goal of the second part is to focus on the polyphenol metabolites, derived from their colonic microbial metabolism and biotransformation. As mentioned above, parent or native polyphenols have glycosidic linkages, which limit their absorption in the small intestine, forcing them to continue their way to the colon. It is in this part of the large intestine that glycosides are cleaved and further metabolized by microbiota to potentially generate more active and better-absorbed metabolites. Using transporters and passive diffusion, the small molecular weight end-products have easy access to the circulation. Hence, the second part of this review is dedicated to specify the biological activity and consequential functional effects of polyphenol metabolites on CMD. Table 5 provides information on the different metabolites issued from colonic digestion and on their pleiotropic effects.

### *5.1. Flavonoid Metabolites and Their Antioxidant and Anti-Inflammatory E*ff*ects*

Bacterial flavonoid catabolism in the colon follows a generic pattern yielding non-specific metabolites such as hydroxylated phenylpropionic acid (HPPA) and phenylacetic acids (HPAA) that ultimately may be subjected to β-oxidation and glycination to yield hippuric acid and its hydroxylated counterpart [34,217]. It can therefore be appreciated that a relatively small number of degradation products emanate from extremely diverse parent flavonoids.

### 5.1.1. Flavonol Quercetin Metabolites

Quercetin (2-(3,4-dihydroxyphenyl)-3,5,7-trihydroxychromen-4-one) is supplied by apples, tea, red wine, berries, tomatoes, and onions. This polyphenol has attracted a lot of interest from the biomedical milieu in view of its beneficial properties in prevention-diseases (e.g., T2D, CVD) despite its limited bioavailability because of its low solubility, weak stability in the upper gastrointestinal tract, rapid fast metabolism, and short biological half-life [98]. However, quercetin molecules unable to be absorbed in the small intestine reach the colon, and are metabolized into a series of phenolic acids by luminal bacteria [265]. In fact, the health-promoting properties observed with quercetin could account for by microbiota-mediated metabolites.

Quercetin, quercetin 3-*O*-rutinoside (rutin) and their colon-derived flavonoid metabolites 3,4-dihydroxyphenylacetic acid (3,4-HPAA), 4-hydroxyphenylacetic acid (4-HPAA), 3-hydroxyphenylacetic acid (3-HPAA), and hippuric acid (*N*-benzoylglycine) have been shown to possess antioxidant and anti-inflammatory properties. On the other hand, when in vitro studies using primary cultures of rat liver parenchymal cells or the human HepG2 hepatoma cell line were treated with t-butylhydroperoxide to produce OxS, the radical scavenging properties were noted only for the parent molecules quercetin and rutin, as well as for 3,4-dihydroxytoluene (3,4-DHT). The metabolites 3-HPAA, 4-HPAA, and hippuric acid remained almost ineffective [208]. Similarly, quercetin and 3,4-DHT showed the potency to inhibit cholesterol synthesis. Another in vitro study, testing the radical scavenging capacity of 3,4-DHPAA, 3-HPAA, and 4-HPPA in human polymorphonuclear cells (stimulated with opsonized zymosan or N-formyl-methionyl-leucyl phenylalanine) revealed that out of the 3 metabolites investigated, only 3,4-DHPAA was active [209]. This radical scavenging capacity is of utmost importance when considering the cell protection from cytotoxicity. Indeed as an adjunct to its radical scavenging properties, 3,4-DHPAA activates phase 2 cytoprotective enzymes (i.e., hemeoxynegese-1, glutathione *S*-transferase), increases the gene and protein expression of aldehyde dehydrogenase isozymes in mouse hepatoma Hepa1c1c7 cells, and induces nuclear translocation of Nrf2 and aryl hydrocarbon receptor [210,211], all involved in controlling cellular REDOX status and in inhibiting cell cytotoxicity.

induced-endothelial

 dysfunction

(3-HPPA)

**5.**Polyphenolmetabolitesandfunctions.


**Table5.***Cont.*


BB-12

↑Mitochondria

homeostasis



**Table5.***Cont.*


 3,4-dihydroxyphenylpropionic acid; 3-HPPA, 3-hydroxyphenylpropionic acid; 4-HPPA, 4-hydroxyphenylpropionic acid; 4-HCA, 4-hydroxycinnamic acid; 3,4-DHPAA, 3,4-dihydroxyphenylacetic acid; 3-HPAA, hydroxyphenylacetic acid; 4-HPAA,4-hydroxyphenylacetic acid. *O*-DMA, O-demethylangolensin, 3,4-DHPVA, 3,4-dihydroxyphenyl-γ-valeric 3,4,5-THPVL, 3,4,5-trihydroxyphenyl-γ valerolactone; 3,4-DHPVL, 3,4-dihydroxyphenyl-γ-valerolactone; 3,5-DHPVL, 3,5-Dihydroxyphenyl-γ-valerolactone 3,-HPVL, 3-hydroxyphenyl-γ-valerolactone; DHR, dihydroresveratrol, NO: nitric oxide; ↑: increase; ↓: decrease.

acid;

In vitro studies have also established the anti-inflammatory properties of quercetin metabolites. For example, experiments with lipopolysaccharide-stimulated human peripheral blood mononuclear cells showed the modulatory effect of 3,4-DHPPA, 3-HPPA, 3,4-DHPAA, 3-HPAA, and 4-hydroxybenzoic acid on the expression of the central pro-inflammatory cytokines; TNFα, IL-1β, and IL-6. TNFα expression was also significantly decreased by the metabolites 3,4-DHPPA, 3,4-DHPAA, and 4-hydroxy-hippuric acid (4-HHA), whereas 3,4-DHPPA and 3,4-DHPAA only suppressed secretion of IL-1β and IL-6 [212]. In another study, in which lipopolysaccharide was used to induce inflammation in RAW 264.7 murine macrophage cell line, 3,4-DHT disclosed an anti-inflammatory capacity by modulating the I κB/NF-κB signaling pathway [266].

Surprisingly, very little research has highlighted the potential anti-diabetic properties of colonic metabolites derived flavonoids. However, when the ability of DHPAA and HPPA was examined on β cell function, the findings clearly emphasized an elevated glucose-stimulated insulin secretion, high protection against tert-butyl hydroperoxide-induced β cell toxicity, hence better survival and function of β cells [215]. The last functions were mediated by the activation of protein kinase C and the extracellular regulated kinases (ERKs) pathways. Collectively, these interesting data suggest that flavonol quercetin/rutin metabolites exert anti-diabetic actions.

### 5.1.2. Flavones and Flavanones Metabolites

Flavones (e.g., apigenin) are present in foods as cereals, parsley, thyme, celery, and citrus fruits under their *O*-glycosidic or C-glycosides derivatives. Once glucosides are hydrolyzed at the intestinal level by digestive enzymes, unabsorbed aglycons undergo further reactions in the large intestine by specific micro-organisms (*Clostridium orbiscindens*, *Enterococcus avium*) causing C-ring fission. Following the metabolism of apigenin, metabolites [e.g., phloretin chalcon, 3-(4-hydroxyphenyl)-propionic, 3-(3-hydroxyphenyl)-propionic, and 4-hydroxycinnamic acids] are acquired [221].

Flavanones (e.g., naringenin) are abundant in tomatoes and citrus fruits, and seem to have higher bioavailability in comparison to flavonols and flavan-3-ols [267]. Flavanone deglycosylation and further degradation by colonic microbiota pathways follow the similar fate of flavonols [267]. *Clostridium* species and *Eubacterium ramulus* are largely involved in these transformations in the colon [217,268]. Metabolites, including 3-(3,4-dihydroxyphenyl)-propionic, 3-(4-hydroxyphenyl)-propionic, and 3-(3-hydroxyphenyl)-propionic acids are produced following naringenin metabolism. However, as for quercetin metabolites, hydroxyphenyl propionic acids (derived from apigenin and naringenin metabolism) display antioxidant and anti-inflammatory capacities (Table 5).

### 5.1.3. Isoflavone Daidzein and Daidzin Metabolites

Isoflavone, as a phytoestrogen precursor, is another flavonoid class of interest. Like many types of polyphenols, isoflavones exert advantageous effects on intestinal health, menopause symptoms, hormone-mediated syndromes, and CVD [269]. They are also transformed into their active metabolites by enzymes from the gut microbiota to generate compounds endowed with elevated estrogenic activity (e.g., 4 ,7-isoflavandiol; equol) or inactive compounds (e.g., *O*-desmethylangolensin; *O*-DMA) [270]. Importantly, the majority of humans are competent to yield *O*-DMA, but only 25–50% among them engender equol [271,272].

The potential health properties of equol and *O*-DMA, both derived from daidzein and its 7-*O*-glycoside daidzin, are the subject of continuing debate. Some studies claim their beneficial effects on the risk of coronary heart disease events [273,274] and others have shown a limited action [275]. The diverging conclusions stems in part from the inter-individual variability in the gut microbiota [276] or from a characteristic that is influenced by ethnicity [277]. Interestingly enough, prehypertensive postmenopausal women producing equol/*O*-DMA had more favorable cardiovascular risk profiles than non-producers [278]. Furthermore, equol/*O*-DMA producers exhibited lower serum uric acid, triglycerides, and waist/hip ratio, as well as a tendency to have higher high

density lipoprotein-cholesterol compared to equol non-producers [279]. These effects are probably exerted through the control of cell redox potential as both equol and *O*-DMA have antioxidant capacity, exemplified by the stimulation of catalase and total superoxide dismutase activities [226]. In a population-based, cross-sectional investigation, urinary concentrations of isoflavones and equol/*O-*DMA were related to well-disposed insulin sensitivity and cardiometabolic markers in pregnant women [280]. Other studies emphasized the beneficial role of *O*-DMA in obesity [281]. However, caution should be exerted when concluding on the effects of isoflavone and related metabolite interventions, as sufficiently randomized controlled trials with statistical are needed [275].

#### 5.1.4. Flavanol Catechin Metabolites

Flavanols or flavan-3-ols are among the most consumed phenolic compounds in Western populations [282]. They are composed of monomers (e.g., catechin, epicatechin, epigallocatechin) and oligomers/polymers (known as proanthocyanidins) [283]. Several groups have reported their ability to prevent chronic diseases (e.g., T2D and CVD) [283–285]. Hydroxy-phenyl-γ-valerolactones and their related hydroxy-phenylvaleric acids are the main metabolites of dietary flavan-3-ols [87,286]. Although a number of studies based on different models aiming at assessing hydroxy-phenyl-γ-valerolactones bioactivity have been reported in the past decade, those involving human subjects are scarce. Inflammatory and OxS processes were particularly studied. For example, (3 4 -dihydroxyphenyl)-γ-valerolactone significantly decreased TNFα-induced NF-kB transcriptional activity in HepG2 cells [92]. Furthermore, a decrease was noted in nitrous oxide production and inducible nitrous oxide synthase expression in the murine macrophage RAW264.7 cells and freshly isolated human monocytes, both treated with (δ-(3,4-dihydroxyphenyl)-γ-valerolactone) a metabolite of the maritime pine bark Pycnogenol [233].

The accumulation of pro-inflammatory cytokine-producing T-lymphocytes and macrophages with the concomitant secretion of the vascular cell-adhesion molecule-1 by the endothelium is key to the early development of the atherosclerotic plaque [287]. Inhibition of this process was observed following the incubation of cultured human umbilical vein endothelial cells with 3 ,4 -dihydroxyphenyl) γ-valerolactone [234], thereby providing the evidence that this compound is a potential therapeutic target. Transferability of the concentrations (quantities) of natural compounds to in vivo or clinical environment (cytotoxicity) represents a major limit of these in vitro experiments. The polyphenolic metabolites 3 ,4 ,5 -trihydroxyphenyl)-γ-valerolactone and 5-(3 ,5 -dihydroxyphenyl)-γ-valerolactone also displayed, in spontaneously hypertensive rats, the capacity to lower angiotensin-1 converting enzyme activity and blood pressure, both contributing to improving endothelial function and arterial elasticity [235]. Paradoxically, the flavan-3-ol metabolites 5-(3 ,4 ,5 -trihydroxyphenyl)-γ-valerolactone, 5-(3 ,4 -dihydroxyphenyl)-γ-valerolactone, and 5-(3 -methoxy, 4 -hydroxyphenyl)-γ-valerolactone did not improve vascular endothelial plasticity of freshly isolated mouse saphenous arteries [236]. The dissimilarity in data could partially be explained by the different models used, especially that isolated saphenous artery endothelium is less reactive than the vessels of the spontaneously hypertensive rat. The in vivo versus the in vitro situations and the experimental time frame are also factors that have to be considered.

The decrease in obesity, a major element in the development of MetS, in response to the consumption of foods rich in flavan-3-ols (e.g., green tea), has been attributed to the role of their content on polyphenolic metabolites in increasing energy expenditure via the modulation of cell differentiation and thermogenesis in adipose tissues [231]. However, when tested in vitro, the flavan-3-ol metabolites [5-(3',4'-dihydroxyphenyl)-γ-valerolactone, 5-(3'-hydroxyphenyl)-γ-valerolactone-4'-O-sulfate, and 5 phenyl-γ-valerolactone-3',4'-di-O-sulfate] did not stimulate cell differentiation and the activation of immortalized pre-adipocytes. In addition, none of the metabolites regulated the expression of the uncouple protein 1, nor the main transcription factors implicated in brown adipocyte genesis. Nevertheless, both 5-(3',4'-dihydroxyphenyl)-γ-valerolactone and 5-(3'-hydroxyphenyl)-γ-valerolactone-4'-O-sulfate protected immortalized pre-adipocytes from

H2O2 generated OxS [231]. In a recent clinical trial, Rodriguez-Mateos et al. [288] showed that cranberry juice given to healthy volunteers improved, in a concentrationand time-dependent manner, the flow-mediated vasodilatation, an early clinical sign in the development of atherosclerosis. Interestingly, the plasma level of the flavan-3-ols metabolite, 5-(3 -hydroxyphenyl)-γ-valerolactone-4 -sulfate, was significantly positively correlated to flow-mediated vasodilatation. Noteworthy, mice on high-fat diet improved non-alcoholic fatty liver disease by lessening IR and endotoxin-toll-like receptor 4/NF-κB pathway in response to raised hepatic catechin metabolites (namely phenyl-γ-valerolactones) following catechin-rich green tea extract [289].

Overall, these findings offer promising beneficial health effects of polyphenols metabolites. However, additional clinical studies are necessary to validate the data in humans and to ensure their transferability from in vitro and animal models to in vivo human situations.

### 5.1.5. Tannin Ellagitannin and Urolithin Metabolites

Hydrolysable tannins, named ellagitannins, encompass one or more gallic acid units and one or more hexahydroxydiphenoic acid units, ester-connected with a sugar residue. In view of their difficulty to be absorbed in the small intestine, ellagitannins reach the colon where they are hydrolyzed into mainly ellagic acid and urolithin (3,4-benzocoumarin derivatives). Supporting data propose a stringent relationship between the consumption of ellagitannins-rich foods and healthy effects based on animal and human models.

The rat under high carbohydrate high-fat diet-induced MetS model has demonstrated the ability of ellagic acid to improve hepatic and cardiometabolic functions, while normalizing glucose tolerance, blood lipid components, central obesity, and body weight [290]. The mechanisms involved in the modulation of OxS and inflammation are through the protein level regulation of Nrf2 and NF-κB, with an impact on metabolic pathways such as fatty acid β-oxidation [290]. Other studies confirm the health benefit properties of ellagic acid, for example, by showing that it dose-dependently repressed the IL-1β-induced production of ROS [248]. This was associated with a significant reduction of the attachment of the established human monocyte cell line to IL-1β-treated human umbilical vein endothelial cells through the suppression of NF-κB, thereby downregulating the expression of vascular cell-adhesion molecule-1 and E-selectin and resulting in decreased monocyte adhesion. Along the same line of evidence, ellagic acid significantly inhibited oxidized low-density lipoprotein and platelet-derived growth factor-BB-induced proliferation of rat aortic smooth muscle cells via inactivation of ERK and marked decrease of proliferating cell nuclear antigen. [249]. More evidence for ellagic acid effects in preventing the development of atherosclerotic events has been provided by showing its capacity of reducing platelet growth factor (PDGF) receptor-β-phosphorylation, diminishing ROS production, and lessening downstream activation of ERK, and finally blocking platelet-derived growth factor-induced expression of cyclin D1 in primary cultures of rat aortic smooth muscle cells [291]. Moreover, these in vitro results were further supported by the ability of ellagic acid to block T2D-induced medial thickness, and lipid and collagen deposition in the arch of aorta in streptozotocin-induced diabetic rats [291]. Collectively, these data suggest that ellagic acid can inhibit OxS and inflammation response, which thus favorably contributes to the regulation of vascular homeostasis.

Urolithins, formed primarily from the oxidative linkage of galloyl groups, have extensively emerged as novel natural bioactive compounds, being currently the focus of extensive investigations. The type of urolithin produced depends on the colon microbiome. Three ellagitanninsmetabolizing metabotypes have been identified: 'metabotype A' (producers of the dihydoxylated urolithin-A conjugates), 'metabotype B' (producers of urolithin-A/isourolithin-A conjugates and/or monohydoxylated urolithin-B conjugates), and 'metabotype 0 (no urolithins produced) [292]. Contrary to the parent compounds, which bloodstream concentrations are negligible, these ellagic acid derivatives may reach micromolar concentrations and thus act as surrogate biomarkers of the inter-individual variation in the colon metabotype [245]. Among the numerous metabolites, urolithin-A

has been particularly studied in terms of its anti-inflammatory and anti-oxidative properties both in vitro and in vivo [250,293,294], as well as anti-atherosclerotic and anti-angiogenic activities [293,295]. In line with these features, adhesion model involving established human aortic endothelial cells and human leukemia monocytic THP-1 cells showed that urolithin-A inhibited TNFα-stimulated adhesiveness and endothelial cell migration [257]. To delve deeper and uncover the mechanisms, the authors could further demonstrate that these effects were concomitant to a down regulation of the levels of chemokine ligand 2, plasminogen activator inhibitor-1, and IL-8. Even more interesting, a clinical trial provided evidence that overweight-obese individuals consuming ellagitannin-containing food with the urolithin metabotype producing urolithin-A were protected against CMD when compared to those with a urolithin-B metabotype [296]. Despite the small number of participants, this clinical trial pointed out the important inter-individual disparity in metabolizing polyphenols, the regulation by microbiota, and indicated that urine metabolite profiling is a valuable tool in assessing potential effects. Fecal polyphenol metabolite profiling could also shed some light on the mechanisms of action. In this context, the profile of fecal phenolic-derived metabolites correlated to blood biomarkers of OxS and inflammation, thus providing support to this avenue [297].

Urolithin may constitute a central regulator of CMD as IR decreased in response to urolithin administration to mice fed high-fat, high-sucrose diet [298]. Concomitantly, urolithin lowered glucose, fatty acids, and triglycerides, while ameliorating adiponectin and mitochondrial function. Furthermore, in the same animal model, through thyroid hormone modulation, urolithin increased energy expenditure by enhancing thermogenesis in brown adipose tissue and inducing browning of white adipose tissue [299]. Therefore, this first-in-class natural compound contributed to metabolic homeostasis via its capacity to normalize OxS, inflammation, obesity, and IR, which overall contribute to MetS.

### *5.2. Non-Flavonoids*

### 5.2.1. Lignan Metabolites

Plant lignans are non-flavonoids, but polyphenolic, phytoestrogenic compounds. They exhibit several biological functions with potent protective benefits. Anti-estrogenic, antioxidant, anti-inflammatory, anti-atherosclerotic, and anti-carcinogenic effects are among their properties, thereby conferring therapeutic values [300,301]. These properties explain the impact of lignan-rich diets on the prevention of hyperlipidemia, IR and hyperglycemia, T2D, and CMD [302–304]. The most frequent forms of dietary lignans are secoisolariciresinol, lariciresinol, pinoresinol, and matairesinol, which are metabolized by gut microflora enzymes into enterolignans such as enterodiol and enterolactone undergoing enterohepatic circulation [305,306]. At least, 28 bacterial species belonging to 12 different genera (e.g., *Bacteroides*, *Clostridium*, *Bifidobacterium*, and *Ruminococcus*) have been identified to be involved in the transformation of lignans to enterolignans [307]. Several groups have quantified the concentrations of the enterolactone metabolite in blood and urine, and reported a protective association in view of the reduction (30–45%) of cardiovascular mortality [308]. Similarly, in a prospective investigation, enterolactone (more than enterodiol) was associated with a lower risk of T2D in U.S. women [309]. A cross sectional trial of the general U.S. population found that subjects with high excretion of enteractone had lower odds of having MetS [310]. Moreover, in a representative sample of U.S. adults, the cross-sectional analysis of data detected that urinary enterolignan concentrations are associated with lower serum triglyceride and greater high-density lipoprotein-cholesterol concentrations in U.S. adults [311]. Additional efforts are obviously needed to confirm these interesting cardiometabolic findings and to explore potential mechanisms of action behind diverse enterolignan metabolites.

### 5.2.2. Resveratrol Metabolites

Resveratrol (3,5,4 -trihydroxystilbene) belongs to the stilbene derivative and is a dietary polyphenolic compound accessible in 70 plant species such as grapes, mulberries, and peanuts [312]. Its pleiotropic effects have made it so famous that it is called a star molecule. It displays numerous biological activities, including antioxidant, anti-inflammatory, cardioprotective, anti-aging, and cancer-chemopreventive features [313,314]. Accordingly, the wealth of scientific reports revealed the effectiveness and beneficial actions of resveratrol in MetS and related disorders such as obesity, IR, and T2D [315,316]). For example, a meta-analysis evaluating 11 studies involving 388 T2D subjects mentioned that resveratrol markedly lessened fasting glucose, insulin, haemoglobin A1c, and IR [278]. Resveratrol was also able to improve hypertension by acting on vascular nitric oxide production, reducing endothelial dysfunction and arteriolar remodeling [317,318]. However, other studies could not record advantageous data on MetS in response to resveratrol [319].

Today, it is well established that resveratrol is easily absorbed by the intestine and promptly metabolized to produce glucuronide and sulfate conjugates [320,321]. Although many gaps are still persistent in the metabolic route of resveratrol, intestinal bacteria transform resveratrol into dihydroresveratrol, known to be partially absorbed and subsequently metabolized to conjugated forms, which can be excreted in urine [322,323]. Besides dihydroresveratrol, two additional bacterial *trans*-resveratrol metabolites were identified: 3,4 -dihydroxy-*trans*-stilbene and 3,4 -dihydroxybibenzyl (lunularin) [264]. Large inter-individual variation was noticed in different subjects: some of them were lunularin producers, others were dihydroresveratrol producers, and the last category was mixed producers, according to levels of these metabolites [264]. Clearly, there was a close relationship between lunularin producers and the abundance of *Bacteroidetes*, *Actinobacteria*, *Verrucomicrobia*, and*Cyanobacteria*, along with lower abundance of *Firmicutes* compared to dihydroresveratrol or mixed producers [264]. Whether resveratrol metabolites are physiologically relevant to tackle metabolic disorders still remains an enigma.

### **6. Conclusions**

Consuming a diet rich in fruits and vegetables is associated with a reduced risk of several noncommunicable lifestyle-related diseases, including CVD. They contribute several essential and bioactive micronutrients such as polyphenols. The first goal of this review was to focus on the available and pertinent literature, which assessed the health-promoting potential of polyphenols to modulate OxS and inflammation. These two major components are powerful inducers of IR, leading to CMD, especially MetS and related complications such as T2D and atherosclerosis.

For the reader's benefit, our review first summarized the most widely distributed classes, sources, and properties of the naturally occurring polyphenols, separated into two main groups, flavonoids and non-flavonoid compounds. Second, attention was paid to their transformation in the intestinal lumen, absorption in different regions of the GI tract, and their bioavailability in the bloodstream. Clearly, most of polyphenols undergo complex intraluminal transformation during digestive and absorptive processes. Data obtained from critical literature revision underline the poor bioavailability, aqueous solubility, permeability, and instability of especially polyphenols with high molecular weights. Third, considerable emphasis was devoted to the essential polyphenol-microbiota interaction as polyphenols directly exert their regulatory effects on intestinal bacteria with concomitant biotransformation of polyphenols into metabolites by gut microbiota. This fundamental interplay provides the central explanation as to the stupefying pleiotropic biological/clinical actions and functions of "parent" polyphenols whereas their systemic bioavailability is incredibly limited, which highlights the prominent contribution of polyphenol metabolites. Fourth, the multifaceted actions and functions of ''parent" polyphenols were summed up. Not only do polyphenols show the great ability to regulate cellular, molecular, and physiological pathways in relation with OxS and inflammation, but they also exhibit a remarkable effectiveness in improving cardiometabolic biomarkers to the advantage of human health. Thanks to their anti-obesogenic, -hyperlipidemic, -hypertensive, -IR, -atherosclerotic, as well as -diabetic effects, they may help to prevent and treat CMD as global public health problems with vast mortality and economic consequences. Fifth, although polyphenol metabolites were also shown to play a similar beneficial role in metabolic complications, there are various enigmatic issues to deal

with: inter-individual differences in the microbial conversion of parent polyphenols, largely unknown microbial metabolites from distinct polyphenols, unidentifiable bacterial strains involved, metabolite health-promoting effects of a multitude of metabolites, elusive metabolite doses, the underlying mechanisms of their specific actions, and the lack of clinical trials.

**Author Contributions:** Conceptualization, E.L. and E.D.; methodology, E.L.; validation, E.L. and E.D.; resources, E.L.; data curation, M.K., F.F. and E.D.; writing—original draft preparation, E.D. and E.L.; writing—review and editing, S.S., Y.D. and E.L.; supervision, E.L.; project administration, S.S.; funding acquisition, E.L. and Y.D. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by a grant from the JA deSève Research Chair in nutrition (E.L.) and the NSERC-Diana Food Industrial Chair on prebiotic effects of polyphenols (401240871) (E.L. & Y.D.).

**Conflicts of Interest:** The authors declare no conflict of interest.

### **Abbreviations**

CMD: cardiometabolic disorders; CVD: cardiovascular disease; DAPP: dried apple peels; ERK: extracellular regulated kinases; GI: gastrointestinal; HPAA: phenylacetic acids; HPPA: phenylpropionic acid; IL: interleukin; IPEC: intestinal porcine enterocyte cell line; IR: insulin resistance; MetS: Metabolic syndrome; miRNAs: micro-RNAs; NF- κB: nuclear factor-kappa B; Nrf2: nuclear redox factor-2; OxS: oxidative stress; ROS: reactive oxygen species; TNFα: tumor necrosis factor-alpha.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
