Insights into the metabolism and microbial biotransformation of dietary flavan-3-ols and the bioactivity of their metabolites

Flavan-3-ols, occurring in monomeric, as well as in oligomeric and polymeric forms (also known as condensed tannins or proanthocyanidins), are among the most abundant and bioactive dietary polyphenols, but their in vivo health effects in humans may be limited because of their recognition as xenobiotics. Bioavailability of flavan-3-ols is largely influenced by their degree of polymerization; while monomers are readily absorbed in the small intestine, oligomers and polymers need to be biotransformed by the colonic microbiota before absorption. Therefore, phenolic metabolites, rather than the original high molecular weight compounds found in foods, may be responsible for the health effects derived from flavan-3-ol consumption. Flavan-3-ol phenolic metabolites differ in structure, amount and excretion site. Phase II or tissular metabolites derived from the small intestine and hepatic metabolism are presented as conjugated derivatives (glucuronic acid or sulfate esters, methyl ether, or their combined forms) of monomeric flavan-3-ols and are preferentially eliminated in the bile, whereas microbial metabolites are rather simple conjugated lactones and phenolic acids that are largely excreted in urine. Although the colon is seen as an important organ for the metabolism of flavan-3-ols, the microbial catabolic pathways of these compounds are still under consideration, partly due to the lack of identification of bacteria with such capacity. Studies performed with synthesized or isolated phase II conjugated metabolites have revealed that they could have an effect beyond their antioxidant properties, by interacting with signalling pathways implicated in important processes involved in the development of diseases, among other bioactivities. However, the biological properties of microbederived metabolites in their actual conjugated forms remain largely unknown. Currently, there is an increasing interest in their effects on intestinal infections, inflammatory intestinal diseases and overall gut health. The present review will give an insight into the metabolism and microbial biotransformation of flavan-3-ols, including tentative catabolic pathways and aspects related to the identification of bacteria with the ability to catabolize these kinds of polyphenols. Also, the in vitro bioactivities of phase II and microbial phenolic metabolites will be covered in detail.


I. Introduction
Proanthocyanidins or condensed tannins are polymers of flavan-3-ols and are among the most abundant polyphenols in our diet. Proanthocyanidins exhibit a wide range of biological activities, including antioxidant, anti-carcinogenic, cardioprotective, antimicrobial and neuro-protective activities, as has been demonstrated in many in vitro and ex vivo studies. 1 In the last decade, a large body of epidemiological data has been accumulated supporting the assumption that the consumption of flavan-3-olrich food such as cocoa, red wine or tea may reduce the risk of coronary heart disease (CHD). [2][3][4] Proanthocyanidins exhibit a high structural diversity and a wide range of degree of polymerization (DP), and their content varies considerably between the different plant sources. Procyanidins, consisting of (epi)catechin units, are the most abundant type of proanthocyanidins in nature. Propelargonidins and prodelphinidins contain (epi)afzelechin and (epi)gallocatechin units, respectively, and are usually mixed with procyanidins. With regard to the interflavanic bond nature, B-type procyanidins [C-4 (upper unit) /C-6 or C-8 (lower unit)] are more abundant than A-type procyanidins, which contain an additional ethertype bond [C-2 (upper unit)-O-C-7 (lower unit)]. Fruits (grapes, apples and pears), legumes, cocoa and beverages such as wine, cider and beer are among the most important sources of B-type proanthocyanidins. 5 Polymeric proanthocyanidins with DP >10 represent the largest amount in 21 kinds of food. 5 The daily intake of flavan-3-ols in the United States has been estimated to be around 60 mg/day for proanthocyanidins with a DP <2. 5 In the Spanish population it has been estimated to be 18-31 mg/day when considering proanthocyanidins with a DP up to 3, 6 and 450 mg/day when considering highly polymerized proanthocyanidins. 7 Polyphenols are recognized as xenobiotics (i.e. foreign or artificial substances, usually of synthetic origin) by the human organism, and therefore bioavailability is a factor that limits the health benefits derived from proanthocyanidin consumption. Bioavailability of proanthocyanidins is largely influenced by their DP. While monomeric flavan-3-ols are readily absorbed in the small intestine, oligomeric and polymeric forms pass intact through the gastrointestinal tract, reaching the colon where they are transformed by the intestinal microbiota before absorption. The scientific evidence accumulated during the last decade indicates that the beneficial effect of these phytochemicals could be attributed to the conjugated metabolites (formed during the phase II metabolism of monomeric flavan-3-ols), and mainly to metabolites derived from the microbial catabolism of proanthocyanidins, rather than to the original forms found in food which have been widely used in most bioactivity studies. [8][9][10] Recent studies have estimated that the amount of non-absorbable polyphenols reaching the colon is very high and that microbe-derived phenolic metabolites excreted in urine represent the largest proportion of polyphenol intake. This recognition is leading to a reformulation of estimated bioavailability values and the potential bioactivity of polyphenols. 11 Currently, there is an increasing interest in the determination of the possible health implications derived from the interaction between phenolic compounds and human microbiota, in particular concerning the effect on microbiota composition and gut health. However, the biological properties of microbial-derived metabolites are still largely unknown.
The aim of the present review is to provide updated information on metabolites formed from dietary flavan-3-ols, as well as their bioactivity and potential health effects. After giving a general overview about the bioavailability of dietary flavan-3ols in humans, structures of main phase II or tissular metabolites derived from small intestine and liver metabolism are presented.
A special section is dedicated to the microbial catabolism of monomeric flavan-3-ols and proanthocyanidins, describing possible catabolic pathways, microbial reactions, and characteristic metabolites derived from the biotransformation process. Intrisic characteristics of candidate catabolic bacteria and structural flavan-3-ol features limiting bacteria degradation are also discussed. Finally, the main biological activities reported for both phase II or tissular and microbial metabolites derived from flavan-3-ols are reviewed, taking into consideration the results of studies performed with conjugated metabolites at in vivo concentrations.

II. Bioavailability of monomeric flavan-3-ols and proanthocyanidins
Bioavailability is a key issue linking polyphenols and health effects. In the case of flavan-3-ols, the degree of polymerization (DP) and galloylation are factors affecting their bioavailability ( Fig. 1). Monomeric flavan-3-ols are absorbed in the small intestine and extensively metabolized into glucuronide conjugates by phase II enzymes. 12,13 These metabolites can reach the systemic circulation or be eliminated in the bile. Further metabolism into sulfate conjugates and methyl derivatives occurs in the liver. However, oligomers with DP >3 and polymers are not absorbed in the small intestine and reach the colon, where they are subjected to microbial catabolism. Microbial metabolites are further absorbed and metabolized by phase II enzymes, to finally enter the circulation or be eliminated in urine.

II.2 Absorption and metabolism of dimeric proanthocyanidins
In the last decade, the absorption and metabolism of dimeric proanthocyanidins have been a subject of speculation. It was first thought that procyanidins could be depolymerized into bioavailable monomers under the acidic conditions of the stomach, 39 but later studies failed to demonstrate this occurrence in vivo. [40][41][42] In contrast to monomers, glucuronidated or sulfated metabolites of dimeric procyanidins have not been detected in biological fluids, 43 although some methylated forms have been reported. 44 Procyanidins B1 [epicatechin-(4b/8)-catechin] and B2 [epicatechin-(4b/8)-epicatechin] have been detected in their intact form at very low levels in human plasma (nM range) after consumption of cocoa 45 or grape seeds, 46 and present the lowest C max in plasma among flavonoid compounds. 13 Besides dimer B2, procyanidin B5 (epicatechin-(4b/6)-epicatechin) has also been detected in the plasma of rats fed cocoa extracts, 47,48 but it was not detected in human plasma after cocoa consumption. 45 However, dimer B3 [catechin-(4a/8)-catechin] and trimer C2 [catechin-(4a/8)-catechin-(4a/8)-catechin] were not detected in the plasma of rats fed the corresponding purified compounds. 49 Recently, oligomers with DP 2-5 have been detected in rat plasma after the administration of apple procyanidin fractions (1 g/kg weight) with the same DP. 44

III. Microbial catabolism of monomeric flavan-3-ols and proanthocyanidins
It has been estimated that 90-95% of dietary polyphenols are not absorbed in the small intestine and therefore accumulate in the colon. 50 In the case of flavan-3-ols, in studies performed with ileostomy patients (i.e. patients whose colon has been removed surgically), it was calculated that approximately 70% of the ingested monomeric flavan-3-ols from green tea could pass from the small to the large intestine, with 33% corresponding to the intact parent compounds. 29 Recently, it has been reported that after oral administration of [ 14 C]procyanidin B2, 63% of the total radioactivity was excreted via urine, indicating that a large quantity of the parent compound is degraded by the gut microflora. 51 The recognition that the colon is a very active organ for the metabolism of flavan-3-ols, particularly proanthocyanidins, has led to a resurgence in the study of the biotransformation of these compounds and other polyphenols by the intestinal Fig. 3 Metabolic pathway tentatively proposed for the catabolism of monomeric flavan-3-ols and dimeric procyanidins by the intestinal microbiota. microbiota 8,10 and their implication in the overall bioavailability and bioactivity of polyphenols.

III.1 First steps of the catabolism of flavan-3-ols: formation of hydroxyphenylvalerolactones and valeric acids
The complex catabolism of B-type proanthocyanidins involves C-ring opening, followed by lactonization, decarboxylation, dehydroxylation, and oxidation reactions, among others. 10 Although numerous in vitro fermentation and in vivo studies have been carried out in recent years, the accumulated knowledge has only led to partial elucidation of the catabolic route of monomeric and B-type dimeric structures 49,52-55 (Fig. 3). In the case of galloylated monomeric flavan-3-ols (ECG and EGCG), the microbial catabolism usually starts with the rapid cleavage of the gallic acid ester moiety by microbial esterases, giving rise to gallic acid which is further decarboxylated into pyrogallol. [56][57][58] The C-ring is subsequently opened, giving rise to diphenylpropan-2ol, which is later converted into 5-(3 0 ,4 0 -dihydroxyphenyl)-gvalerolactone (in the case of (epi)catechins) or 5-(3 0 ,4 0 ,5 0 -trihydroxyphenyl)-g-valerolactone (in the case of (epi)gallocatechins). 56,58,59 The valerolactone ring later breaks, giving rise to 5-(3 0 ,4 0 -dihydroxyphenyl)valeric acid and/or 4-hydroxy-5-(3 0 ,4 0 -dihydroxyphenyl)valeric acid. The identification of this latter compound was firstly proposed by Khori et al. 57 and recently confirmed by Llorach et al. 60 in urine samples collected after cocoa consumption in humans, as well as by Stoupi et al. 61 after in vitro fermentations carried out with human faeces in the presence of (À)-epicatechin and procyanidin B2.
Although it was first proposed that 4-hydroxy-5-(hydroxyphenyl)valeric acids could arise from the degradation of diphenylpropan-2-ols, concurrently with hydroxyphenyl-gvalerolactones 57 (Fig. 3), it has recently been suggested that they are formed instead from hydroxyphenyl-g-valerolactones, and that an interconversion between both forms [4-hydroxy-5-(hydroxyphenyl)valeric acids and 5-(hydroxyphenyl)-g-valerolactones] may exist, but is largely displaced towards the formation of the formers. 61 Subsequent biotransformations of these valeric acids give rise to hydroxyphenylpropionic and hydroxybenzoic acids by successive loss of carbon atoms from the side chain through b-oxidation. 56

III.2 Metabolites arising from the catabolism of dimeric procyanidins
The possible formation of 3,4-dihydroxyphenylacetic acid via aoxidation of 3,4-dihydroxyphenylpropionic acid (as described for tyrosine 49,62 ) in the microbial catabolism pathway of monomeric flavan-3-ols, has been widely debated. Firstly, it was thought that 3,4-dihydroxyphenylacetic acid was only characteristic of the catabolism of dimeric procyanidins; 63 however, other authors have recently proposed a-oxidation as a possible pathway for the formation of this compound in the case of both monomers and dimers, 61 without discarding other possible pathways, as proposed by Appeldoorn et al. 54 in the case of dimers. According to these latter authors, 3,4-dihydroxyphenylacetic acid results from the cleavage of the upper unit of dimeric procyanidins, whereas the lower unit gives rise to 5-(3 0 ,4 0dihydroxyphenyl)-g-valerolactone and to the triggering of the rest of the previously described route (Fig. 3). The possible depolymerization of dimeric structures into monomeric units, firstly proposed by Groenewoud et al., 64 has been recently confirmed to occur but to a lesser extent, 54,61 representing less than 10% in the case of procyanidin B2. 61 Other microbial metabolites arising exclusively from the catabolism of dimeric procyanidins have recently been identified, such as 5-(2 0 ,4 0dihydroxyphenyl)-2-ene-valeric acid, as well as other compounds which have been tentatively identified as derivatives from the A-ring of the upper unit, including the interflavanic bond. 61

III.3 Last steps of the catabolism of flavan-3-ols
Finally, the last steps of the microbial catabolism of (epi)catechin involve dehydroxylation of 3,4-dihydroxylated phenolic acids at C-4 0 (preferentially), and C-3 0 , resulting in 3-and 4-monohydroxylated phenolic acids, respectively. 53,61 In the case of (epi)gallocatechins, dehydroxylation preferentially occurs at C-5, resulting in 3,4-dihydroxylated phenolic acids which undergo further dehydroxylation at C-4 and C-3, as mentioned above. However, in the case of hydroxyphenylvalerolactones, the 3,5-dihydroxylated derivative arising from the dehydroxylation of 5-(3 0 ,4 0 ,5 0 -trihydroxyphenyl)-g-valerolactone has also been identified, indicating that dehydroxylation at C-4 0 occurs. 20 Once absorbed, the microbial metabolites from flavan-3-ols are mainly metabolized in the liver by phase II enzymes as conjugated derivatives that are subsequently eliminated in urine. At the same time, a portion of microbial metabolites (non-conjugated microbial metabolites) is eliminated in the faeces.
Several microbe-derived metabolites that have been detected in urine in their actual conjugated form by targeted analysis including: monoglucuronide and monosulfate of 5-(3 0 ,4 0 -and 3 0 ,5 0 -dihydroxyphenyl)-g-valerolactone, in addition to the methyl-sulfate derivatives of 5-(3 0 ,4 0 ,5 0 -trihydroxyphenyl)-gvalerolactone. 20,22,30,65 In the case of phenolic acids, monoglucuronide and monosulfate conjugates of mono-and di-hydroxyphenylpropionic and p-coumaric acids have been reported. 66 Other reactions occurring in the liver and kidney include: glycine conjugation, dehydrogenation, hydroxylation and methylation. 53 The excretion of microbial metabolites varies markedly between subjects, and for some individuals it may also vary with the substrate, reaching a very high proportion (up to 50%) of the intake of polyphenols. 8

III.4 Main microbial phenolic metabolites found in urine
Several feeding studies have revealed significant changes in the urinary excretion of microbe-derived phenolic acids after the intake of rich sources of flavan-3-ols. Among phenolic acids, mono-and di-hydroxylated phenylpropionic and phenylacetic acids, together with hydroxyhippuric acids, have been found as main urinary microbial phenolic acids derived from flavan-3-ol intake.
With regard to cocoa and cocoa-derived products, Gonthier et al. 49 reported an increase in 3,4-dihydroxyphenylacetic and 3-hydroxyphenylacetic acids in urine after the administration of procyanidin B3 to rats. Similarly, Rios et al. 67 reported a significant increase in the urinary excretion of these compounds, as well as in 3-hydroxyphenylpropionic and 3-hydroxybenzoic acids in healthy humans after acute consumption of flavanol-rich chocolate. Recently, Urpi-Sarda et al. 68 also found increased urinary levels of 3,4-dihydroxyphenylacetic and 3-hydroxyphenylacetic acids in humans after chronic consumption of cocoa powder with milk. Other studies have reported an increased urinary excretion of 3-hydroxypropionic and 3-hydroxyphenylacetic acids after human consumption of grape seed polyphenols. 69 In the case of green tea, 3-hydroxyphenylpropionic and 4-hydroxyphenylacetic and 4-hydroxyphenylacetic acids significantly increased in human urine. 58 Finally, 3-hydroxyphenylpropionic and 3-hydroxybenzoic acids were also reported to increase in the urine of rats fed wine polyphenols. 53 Besides these phenolic acids, which are also common to the microbial catabolism of other flavonoids, 10 5-(3 0 ,4 0 ,5 0 -trihydroxyphenyl)-g-valerolactone and 5-(3 0 ,4 0 -dihydroxyphenyl)-g-valerolactone are considered important microbial metabolites and potential biomarkers of flavan-3-ol consumption in humans, as has been confirmed after the intake of green tea, 20,22,23,58 cocoa products 30 and almond skins. 65,70 IV. Intestinal bacteria with ability to catabolize flavan-3-ols It is important to mention that the above difference of opinions concerning the possible catabolic route of monomeric and dimeric flavan-3-ols could be partly attributed to differences in the microbiota composition of faecal samples used in the different studies, suggesting that different pathways could coexist or one predominate over the others, depending on the catabolic capacity of the microbiota. An important limitation in this area is that bacteria belonging to human microbiota with the capacity to catabolize flavan-3-ols have still not been identified. To date, only bacteria with the capacity to catabolize other types of flavonoid compounds, mainly flavonols and flavones, have been described. These bacteria, in general, belong to the Clostridium and Eubacterium groups. 10 Among the factors that may limit the identification of flavan-3-ol catabolic bacteria, it is important to highlight the wellknown growth inhibitory effects of proanthocyanidins. Another factor that deserves consideration is the structural features of flavan-3-ols as complex non-planar molecules.

IV.1 Inhibitory effects of proanthocyanidins and ''tanninresistant'' bacteria
The growth-inhibitory effects of proanthocyanidins on bacteria have been reviewed by Smith et al. 71 Tannins are capable of complexing with polymers and minerals, making nutrients unavailable. In addition, they could have a direct effect by interacting with membranes, cell walls, and/or extracellular proteins. ''Tannin-resistant'' bacteria have been defined as those bacteria that are able to withstand the inhibitory effect of tannins. ''Resistance'' implies that some action is required on the part of the organism to withstand the inhibitory effect of tannins, including inducible adaptation or even gene transfer. 71 Tanninresistance may also depend on the tannin concentration, structural composition and DP. It is important to highlight the fact that bacteria which are predominant in tannin-rich mediums may not be resistant per se, but are less affected by nutrient limitations or are better able to access limiting nutrients.
''Tannin-resistant'' Gram-negative species (Enterobacteriaceae and Bacteriodes) have been isolated from rat faecal samples after prolonged administration of condensed tannins from Acacia angustissima, a forage legume. 72 ''Tannin-resistant'' Gram-positive bacteria have also been identified. Brooker et al. 73 isolated a Streptococcus strain (named S. caprinus and close to S. bovis) from the rumen of goats which was able to grow at 2.5% of condensed tannins. A Streptococcus strain (close to S. bovis and S. gallolyticus) has also been isolated from the rumen of sheep, goats and deer. 74 Later, Molina et al. 75 has also isolated a Eubacterium strain (close to E. cellulosolvens) from the rumen of moose, able to tolerate 0.5 g/L of condensed tannins.
Some mechanisms by which bacteria can overcome inhibition by tannins include: modification/degradation of the substrate, dissociation of tannin-substrate complexes, cell membrane modification/repair and metal ion sequestration. It has been reported that Bifidobacterium infantis and Lactobacillus acidophilus are not inhibited by tannins because lactic acid bacteria do not require iron as they do not depend on metal-chelating enzymes, in particular heme enzymes. 76 Moreover, in vivo studies have revealed that consumption of grape seed extract, containing 40% of condensed tannins, produced an increase in the bifidobacteria population in healthy individuals. 77 Although tannin resistance is the first step in order for bacteria to metabolize condensed tannins, resistance does not guarantee metabolic activity, and the biodegradation pathway of ''tannin-resistant'' bacteria has not yet been described.

IV.2 Structural features of flavan-3-ols limiting bacterial catabolism
There is some evidence that the structural characteristics and stereochemistry of flavan-3-ols could be limiting factors for intestinal bacteria to be able to degrade these types of compounds. It has been reported that Eubacterium ramulus was unable to degrade (+)-catechin because of the absence of a functional group at C-4 in this flavonoid structure. 78 Similarly, the human bacterium Eubacterium sp. (SDG-2) was able to open the ring of the 3R [(À)-catechin and (À)-epicatechin] and the 3S [(+)-catechin and (+)-epicatechin] forms of monomeric flavan-3ols into 1,3-diphenylpropan-2-ols (Fig. 3), but was incapable of producing the same fission in their galloylated esters. 79 However, in no instance was this bacteria able to continue the catabolism up to the formation of 5-(3 0 ,4 0 -dihydroxyphenyl)-g-valerolactone. Another characteristic of this bacterium was the ability to dehydroxylate the OH groups in the B ring of 1,3-diphenylpropan-2-ols, but only of the R forms. 79 This fact, together with the inability to catabolize the gallate esters, suggests that the spatial configuration of both the original flavan-3-ol molecule and intermediate metabolites may limit the microbial degradation of flavan-3-ols. In fact, in a recent in vitro fermentation study with human faeces it was found that (+)-catechin (2R,3S) was firstly converted into (+)-epicatechin (2S,3S) by intestinal microbiota in order for the biotransformation process to proceed. 80 Taken together, these findings suggest that it may be difficult to identify a single bacterium capable of exhibiting the whole catabolic pathway proposed in Fig. 3, but rather the catabolism may be carried out by different bacteria with specific catabolic activities that work in sequential form on the appearance of the different intermediate metabolites. Among the different phases of the catabolic pathway, formation of 5-(3 0 ,4 0 -dihydroxyphenyl)-g-valerolactone seems to be a limiting step.

V. Bioactivity of flavan-3-ol metabolites
As a consequence of their extensive metabolism in the human organism, the original flavan-3-ol structures present in food are not present in plasma and urine (with the exception of small amounts of gallate ester of monomeric flavan-3-ols and dimeric procyanidins that appear unmetabolized, as mentioned above) but rather appear as a complex series of phase II or tissular metabolites and, particularly, of microbe-derived phenolic metabolites. Therefore, both types of circulating metabolites should be responsible for the health benefits associated with the consumption of dietary sources rich in flavan-3-ols.
V.1 Bioactivity of phase II or tissular metabolites derived from small-intestine and liver metabolism One of the limitations of many in vitro and ex vivo studies which have tried to unravel the health effects of flavan-3-ols has been the use of unconjugated structures, as well as the use of test concentrations (mM) at a much higher range than that found in biological fluids (mM range). Taking this into consideration, this section will only try to cover the results of studies performed with conjugated metabolites in the micromolar or submillimolar range (0.5-30 mM) found in plasma (Table 1).
In general, the conjugation process (glucuronidation, sulfation and methylation) affects the physico-chemical properties of flavan-3-ols and, in turn, their residence in plasma, their excretion rate, and finally the bioactive properties of the parent compound. 13 In particular, sulfation and glucuronidation involve a considerable attenuation of biological activity. The case of methylation seems to be more complex because the incorporation of methyl groups reduces the number of available OH groups, but at the same time increases the lipophilic nature of the compound, which can be advantageous for cellular uptake by passive diffusion. 13 Antioxidant activity. The antioxidant activity of flavonoid metabolites has been widely studied, considering the fact that oxidative stress is implicated in the initiation and progression of chronic diseases. In the case of flavonoid compounds (i.e. quercetin), it has been observed that glucuronidation at C-3 0 and C-4 0 of the B ring (catechol-type structure) produces a greater loss of antioxidant capacity than when it occurs at C-3 of the C ring. [81][82][83] In contrast, glucuronidation at C-7 (A ring) seems to produce a slight increase in antioxidant activity. 83 In the case of flavan-3-ol metabolites, (À)-epicatechin and its 7-Oglucuronide presented a similar delay of Cu 2+ -induced LDL oxidation, whereas the activities of the 3 0 -O-glucuronide and the 4 0 -O-methyl-3 0 -O-glucuronide were significantly lower. 84 However, in the case of galloylated (epi)gallocatechins, the position of glucuronidation affected the anti-radical capacity against DPPH differently to the other flavonoids, since EGCG-7-O-glucuronide and EGCG-4 00 -O-glucuronide (galloylation in the gallic acid ring) were less active than the aglycone, whereas the 3 0 -and 3 00 -O-glucuronides showed the same activity as the aglycone. 24 For non-galloylated (epi)gallocatechins, EGC-7-Oglucuronide and -3 0 -O-glucuronide were more active than the aglycone. 24 In the case of O-methylation, Cren-Oliv e et al. 85 also reported that the catechol B-ring was also the active moiety of (+)-catechin, since the 3 0 -and 4 0 -O-methyl ethers and 3 0 ,4 0 -di-O-methyl ether showed a much lower inhibition of Cu 2+ -induced LDL oxidation than the aglycone, but the activity was recovered when these positions were free, as in the 5,7-di-O-methyl analogue. The C-3 0 and C-4 0 -O-methyl ethers of (À)-epicatechin also showed a lower inhibition of peroxynitrite-induced tyrosine nitration than the parent compound. 86 Similarly, O-methylation at position C-3 0 in (À)-epicatechin, (À)-epigallocatechin and (À)-epicatechin-3-O-gallate elicited a potential inhibition of lipid oxidation of canola oil in comparison to the aglycone. 87 In a recent study, C-3 0 and C-4 0 -O-methyl ethers of (+)-catechin and (À)-epicatechin showed a lower antioxidant capacity than the parent compound, as measured by the ferric-reducing power (FRAP) and by the ability to scavenge the ABTS + radical cation. 88 Moreover, the antioxidant activity of these metabolites was found to be pH dependent, but significant radical scavenging activity was found to be retained at pH 7.4, suggesting that they could act as potential antioxidants under physiological conditions. 88 Vascular effects. Epicatechin and its metabolite, epicatechin-7-O-glucuronide, have been identified as independent predictors of the vascular effects observed after flavanol-rich cocoa intake. 19 Anti-inflammatory effects. In the case of EGCG metabolites, glucuronidation at C-7 affected the ability to inhibit the production of NO or the arachidonic acid metabolism in HT29 cells compared to the aglycone, but it was not affected in the case of glucuronidation at C-3 0 , C-3 00 , C-4 00 . 24 Conversely, in the case of ECG, glucuronidation at C-3 0 decreased such capacity by 20% compared to the aglycone, but it was not affected in the case of the 7-O-glucuronide. 24 Inhibition of cellular growth. The effectiveness of (À)-epicatechin metabolites on the inhibition of cellular growth has been studied in various types of cell lines. In the case of neuronal cells, it has been reported that 3 0 -O-methyl-epicatechin was as effective as (À)-epicatechin in the inhibition of apoptosis induced by oxidized LDL. 89 Similarly, it has been reported that 3 0 -O-methylepicatechin was as efficient as (À)-epicatechin in protecting human fibroblasts against cell death induced by oxidative stress. 90 In the case of galloylated flavan-3-ol metabolites, methylation at C-4 0 and C-4 00 in (À)-epigallocatechin-3-O-gallate (EGCG) produced a 50% decrease in the growth-inhibitory and pro-apoptotic activities of murine osteoclasts, compared to EGCG. 91 In another study, methylated derivatives of EGCG at positions C-4 00 and C-4 0 -4 00 (dimethyl derivative) presented less inhibitory capacity than EGCG of the enzyme 20S proteasome, which catalyzes the degradation of intracellular proteins and is associated with cancer. 92 Table 1 Biological activity of phase II or tissular metabolites of flavan-3-ols.     inhibited growth of KYSE150 cells by 20-40% at 50mM, but had no effect on HT-29 cells.  For Lactobacilli and Staphylocccus aureus strains, the order of activity was:
-Phenolic acids failed to inhibit the growth of the Gram-negative bacterium P. aeruginosa  and Polyphenol mix 1 (+1.93 mM).
Other compounds tested showed no significant effect on the platelets' sensitivity towards the agonist TRAP.
None of the tested polyphenol metabolites affected ADP-and collagen-induced platelet aggregation at concentrations up to 100 mM.

TRAP-induced platelet activation
Activation of platelet with TRAP increased the P-selectin expression (from 0.5 to 15%). The activation was reduced with dihydroferulic acid (À20
Epinephrine-induced glycoprotein CD63   The three compounds increased the tolerance to the applied pressure of the inflamed paw, but 3,4-dihydroxyphenylpropionic acid was the most potent.

Anti-proliferative activity and cytotoxicity
Tanaka et al.   Interaction with cellular signalling pathways. In recent years, it has been suggested that polyphenols may exert their health effects via a mechanism of action beyond their antioxidant activity, and which is more related to its ability to generate an adaptive response at the cellular level that involves interaction with certain key proteins in triggering cell signalling pathways of oxidative stress and exposure to environmental toxins. 9 In the case of flavan-3-ols, most studies have been performed mainly with the non-conjugated forms. It has been reported that EGCG induces apoptosis and causes cell-cycle arrest in tumor cells -but not in non-transformed normal cells -through the modulation of nuclear factor kappa-B (NF-kB). NF-kB is a redox-sensitive transcription factor which regulates the expression of proinflammatory cytokines, iNOS, COX-2, growth factors and inhibitors of apoptosis, and is related to inflammatory diseases (atherosclerosis, ulcerative colitis and rheumatoid arthritis), as well as neurodegenerative diseases and cancer. 93,94 In another study, EGCG was also found to down-regulate NF-kB-inducing kinase (NIK), death-associated protein kinase (DPAK 1), and rho B and tyrosine protein kinase in PC-9 human lung cancer cells. 95 A down-regulation of genes involved in a wide range of physiological functions was found in the mucosa of rats with induced colon carcinogenesis that had been fed wine polyphenols for 16 weeks, being the major pathways down-regulated those involved in the inflammatory response and steroid metabolism. 96 With regards to genes involved in relevant process of atherosclerosis, red wine polyphenols were also found to significantly inhibit the proliferation of human vascular smooth muscle cellsbut not of human vascular endothelial cells -by reducing the promoter activity and expression of the cyclin A gene. 97 Green tea polyphenols have been shown to modulate the regulation of the transcriptional expression of proatherogenic molecules, including the sterol-response element binding protein (SREBP), PPAR-g, IL-8, and apoprotein-E. 98

V.2 Bioactivity of microbe-derived phenolic metabolites
The biological activities of microbial metabolites derived from the catabolism of flavan-3-ols are still largely unknown, but in recent years those of hydroxyphenyl-g-valerolactones, and especially of phenolic acids (di-and mono-hydroxylated phenylproponic, phenylacetic, benzoic acids and derivatives) formed from the subsequent catabolism of the former, have started to be unravelled. In contrast to phase II or tissular metabolites derived from small-intestine and liver metabolism as described above, to date, in vitro studies performed with microbe-derived phenolic metabolites have been carried out with unconjugated metabolites (with the exception of hippuric acids) ( Table 2).

V.2.2 Phenolic acids
Effects on intestinal microbiota. Some phenolic acids, including 3-O-methyl gallic, gallic, caffeic, 4-hydroxyphenylpropionic, phenylpropionic, and 4-hydroxyphenylacetic acids derived from the microbial degradation of tea catechins, were able to inhibit the growth of several pathogenic and nonbeneficial intestinal bacteria without significantly affecting the growth of beneficial bacteria (Lactobacillus spp. and Bifidobacterium spp.). 101 Other studies have revealed that dihydroxylated forms (i.e. 3,4-dihydroxyphenylacetic and 3,4-dihydroxyphenylpropionic acids) efficiently destabilize the outer membrane of Salmonella. 102 Recently, Cueva et al. 103 found that the number and position of substitutions in the benzene ring of phenolic acids and the saturated side chain length influenced the antimicrobial potential of phenolic acids against different microorganisms (Escherichia coli, Lactobacillus spp., Staphylococcus aureus, Pseudomonas aeruginosa and Candida albicans), although it was strain-dependent. In general, non-hydroxylated and monohydroxylated phenolic acids were more potent than dihydroxylated or disubstituted phenolic acids. With regard to the saturated side chain, the order of potency, for the same benzene ring-substitution, was benzoic > phenylacetic > phenylpropionic acid. Moreover, Lactobacillus spp. and S. aureus (Gram-positive) appeared more susceptible to the action of a series of microbial phenolic acids than Gram-negative bacteria, such as E. coli and P. aeruginosa. 103 Antioxidant activity. Among a series of microbe-derived phenolic acids, 3,4-dihydroxyphenylacetic and 3,4-dihydroxytoluene exhibited the highest radical scavenging activity against DPPH in cultured rat hepatocytes. 104 However, only the latter metabolite was found to be effective against the lipid peroxidation of rat hepatocytes challenged with tert-butyl-hydroperoxide. 104 Anti-thrombotic activity. Rechner et al. 105 studied the effect of several microbe-derived phenolic acids and their mixture on platelet function through several tests, including: platelet aggregation, P-selectin expression on resting platelets, effect on TRAP-induced platelet activation and epinephrine-stressed platelets. Dihydrocaffeic acid (3,4-dihydroxyphenylpropionic acid), dihydroferulic acid (4-hydroxy-3-methoxyphenylpropionic acid) and 3-hydroxyphenylpropionic acid, as well as the polyphenol mixture, were among the metabolites with the best activity in all tests performed. 105 Anti-inflammatory activity. Studies carried out by Karlsson et al. 106 showed that faecal samples containing microbial phenolic acids affected cyclooxygenase-2 (COX-2) protein levels in colon cancer cells (HT-29) stimulated with TNF-a. Recently, Russell et al. 107 reported that phenolic acids presenting 4-hydroxy-3-methoxy substitution and a one-carbon side chain such as vanillic acid and its derivatives (vanillin, vanillyl alcohol and acetovanillone), as well as a three-carbon side chain (cinnamic, o-, m-and p-coumaric acid, and caffeic acid), inhibited cytokine-induced prostanoid biogenesis in human colonic fibroblasts. A structure-activity relationship has been observed between phenolic acids and their anti-inflammatory effects, since only dihydroxylated phenolic acids (i.e. 3,4-dihydroxyphenylpropionic and 3,4-dihydroxyphenylacetic acids) significantly inhibited the production of pro-inflammatory cytokines TNF-a, IL-1b, IL-6 in peripheral blood mononuclear cells (PBMC) stimulated with LPS, whereas no significant effect was found for the monohydroxylated ones. 108 Similarly, Larrosa et al. 109 recently found that these dihydroxylated phenolic acids provided the best inhibition of prostaglandin E2 production in cancer cells of fibroblast (CCD-18) stimulated with IL-1b. In vivo experiments with rats have also shown that 3,4-dihydroxypropionic acid was the most potent metabolite in writhing and paw pressure tests in rodents and reduced the expression of cytokines TNF-a, IL-1b, IL-8, as well as the levels of malonaldehyde and oxidative damage to DNA in the distal mucosa of rats with dextran sodium sulfate (DSS)-induced colitis. 109 Anti-proliferative activity and cytotoxicity. Among a series of microbial phenolic metabolites, 3,4-dihydroxyphenylacetic acid presented anti-proliferative activity in prostate (LNCaP) and, in particular, in colon cancer (HCT116) cells. 110 In vivo studies have also revealed that protocatechuic acid reduces the incidence and multiplicity of cancerous tumors in the colon of rats. 111 The modulation of cytochrome P450 and enzymes involved in xenobiotic activation and/or detoxification pathways (phase II enzymes) by protocatechuic acid in mouse liver and kidney has also been reported. 112 Moreover, protocatechuic acid affected the level of rat hepatic and renal glutathione S-transferase (GST) isoenzymes. 113 Cytotoxicity assays have also shown that protocatechuic acid effectively kill the HepG2 hepatocellular carcinoma cells by stimulating the c-Jun N-terminal kinase (JNK) and p38 subgroups of the mitogen-activated protein kinase (MAPK) family. 114 A similar signalling pathway has been reported to be involved in the apoptosis of human gastric adenocarcinoma cells by protocatechuic acid. 115 In a recent study, protocatechuic acid has also shown significant neuroprotective effects on retenoneinduced apoptosis in PC12 cells by ameliorating the mitochondrial dysfunction. 116 Modulation of lipid metabolism. It has been reported that 3,4dihydroxytoluene acid inhibits the synthesis of heptocellular cholesterol by inhibiting the incorporation of acetate into HepG2 liver cells. 104

VI. Concluding remarks
Over the last decade, a large number of epidemiological and interventional studies have demonstrated that there may be an association between flavonoid consumption and human health. Mechanistic studies trying to determine flavan-3-ol health effects have revealed that these polyphenols exhibit a wide range of biological effects. Despite the enormous effort devoted to this area, some results may be misleading, since polyphenol metabolism as xenobiotics has not been considered in a large number of studies which employed structural forms and concentration ranges not found in vivo. Therefore, polyphenol bioavailability is a key issue in the link between polyphenol and human health. In comparison to other micronutrients, knowledge about polyphenol bioavailability is advancing with the progress of analytical instrumentation which allows the identification of new metabolites in vivo. The recognition that some polyphenols, in particular proanthocyanidins, are extensively metabolized by the intestinal microbiota into low molecular weight compounds, and that these metabolites represent a very large percentage of the amount ingested, is bringing into consideration the inclusion of microbial metabolism as part of the bioavailability concept currently adopted for polyphenols. On the basis of these facts, interest is now focused on the study of the bioactivity of microbe-derived metabolites, in addition to phase II or tissular metabolites, as compounds responsible for the health effects of flavan-3-ols. Although advances are being made in the determination of the bioactivity of microbe-derived metabolites, most studies carried out until now have failed, again by not testing the conjugated forms found in vivo. With regards to the bioactivity of actual conjugated forms derived from flavan-3-ol in vivo metabolism, research carried out in the last decade has revealed that flavan-3-ols are multifunctional compounds that may display effects by mechanism(s) of action beyond their antioxidant activity.
The health effects derived from the interaction between flavan-3-ols and the intestinal microbiota should be a subject of increasing interest. Although some authors have pointed out that polyphenols may be beneficial to gut health by increasing the population of potentially beneficial bacteria or exerting prebiotic actions, the effects that the interaction between flavan-3-ols and intestinal microbiota may have on the functionality of the metabolic activity of the microbiota and overall gastrointestinal health still remains largely unknown. In fact, for flavan-3-ols to function as a prebiotic, intestinal bacteria with such metabolic capacity should exist in the colon, but they are difficult to identify due to direct or indirect factors inherent in flavan-3-ols. The identification of flavan-3-ol-metabolizing bacteria and their possible use as a probiotic could be a good strategy for increasing the bioavailability and potential bioactivity of proanthocyanidins.