Polyphenols are among the most abundant antioxidant compounds of the Mediterranean diet and may play a key role in the prevention of cardiovascular and neurodegenerative diseases, and cancer(Reference Arts and Hollman1). Health effects derived from polyphenol consumption depend on their bioavailability, a factor that greatly varies from one compound to another(Reference Manach, Williamson and Morand2). Among polyphenols, the oligomers and polymers of flavan-3-ols, also called proanthocyanidins, are not absorbed or degraded into monomers during their transit through the stomach(Reference Manach, Williamson and Morand2, Reference Kroon, Clifford and Crozier3). However, when they reach the colon, they are metabolised by the intestinal microbiota into various phenolic acids, including phenylpropionic, phenylacetic and benzoic acid derivatives(Reference Deprez, Brezillon and Rabot4). Hydroxycinnamic acid esters and polyphenols linked to rhamnose are also degraded into phenolic acids by the microbiota(Reference Manach, Williamson and Morand2). Recently, it has been reported that these metabolites may also exert several biological activities, such as the inhibition of platelet aggregation and activation function(Reference Rechner and Kroner5), inhibition of cyclo-oxygenase-2 in HT-29 colon cancer cells(Reference Karlsson, Huss and Jenner6), reduction in the synthesis of prostanoids in colon cells(Reference Russell, Drew and Scobbie7), antiproliferative activity in prostate and cancer cells(Reference Gao, Xu and Krul8) and, finally, influence cell proliferation, apoptosis and signalling pathways in human colon carcinoma cells(Reference Glinghammar and Rafter9).
Atherosclerosis is now considered to be a low-grade chronic inflammatory process resulting from the interactions between plasma lipoproteins, cellular components (monocyte/macrophages, T lymphocytes, endothelial cells and smooth muscle cells) and the extracellular matrix of arterial wall(Reference Tedgui and Mallat10). Pro-inflammatory cytokines are involved in all phases of the atherosclerotic process: they stimulate chemokines and adhesion molecules, leading to early recruitment of monocytes and lymphocytes in the arterial intima, and later exert potential noxious effects promoting weakening of plaques that are more prone to rupture(Reference Tedgui and Mallat10). However, data concerning the effect of phenolic compounds on the production of the inflammatory mediators from mononuclear cells are scarce and contradictory. Some researchers(Reference Sanbongi, Suzuki and Sakane11) reported that polyphenols from cocoa may reduce the expression of IL-2 mRNA in human lymphocytes, and others(Reference Mao, Powell and Van de Water12) found differential effects of isolated cocoa procyanidin fractions (monomer to decamers) on the expression and secretion of IL-1β from peripheral blood mononuclear cells (PBMC). In the same way, cocoa polyphenols produced an increase in TNF-α secretion(Reference Mao, van de Water and Keen13) or even a down-regulation of IL-2 secretion and IL-2 receptor surface expression on a lymphoid cell line(Reference Ramiro, Franch and Castellote14), whereas in another study, polyphenols from olive oil had no effect on the secretion of TNF-α, IL-1β or IL-6 in the human whole blood(Reference Miles, Zoubouli and Calder15).
Considering the lack of information regarding the anti-inflammatory properties of microbial-derived phenolic acids, the aim of the present study was to investigate the effect of some microbial phenolic metabolites on the modulation of the production of the most representative pro-inflammatory cytokines, i.e. TNF-α, IL-1β and IL-6, in lipopolysaccharide (LPS)-stimulated PBMC from healthy human volunteers.
Methods
Six healthy volunteers (two men and four women) with an average age of 29·3 (sd 2·1) years (range 27–33 years), weight of 62·8 (sd 18·5) kg (range 50–100 kg), height of 1·7 (sd 0·1) m (range 1·6–1·9 m) and BMI of 21·5 kg/m2 (sd 3·4) (range 17·9–27·7 kg/m2) participated in the study. None of them reported a history of heart disease, homeostatic disorder or any other medical disease. None were receiving any medication or taking any vitamin supplements. A 24 h food recall questionnaire was used to assess their habitual nutrient intake. This information was converted into dietary data using the Professional Diet Balancer software (Cardinal Health Systems, Inc., Edina, MN, USA) Their habitual diet (mean of six volunteers) included an intake of 9789·81 kJ/d (2339·82 kcal/d); 117·41 g/d protein; 257·35 g/d carbohydrates; 24·99 g/d dietary fibre; 5·27 g/d soluble fibre; 76·43 g/d total sugar (14·93 g/d monosaccharides, 31·27 g/d disaccharides); 92·48 g/d of total fat (26·03 g/d saturated fat, 43·30 g/d monounsaturated fat, 15·93 g/d polyunsaturated fat, 1·22 g/d trans-fatty acids, 221·87 mg/d cholesterol); 863·77 retinol equivalents/d vitamin A; 2·33 mg/d vitamin B1; 2·39 mg/d vitamin B2; 40·79 mg/d vitamin B3; 2·49 mg/d vitamin B6; 7·77 μg/d vitamin B12; 84·91 mg/d vitamin C; 4·89 μg/d vitamin D; 13·08 mg/d vitamin E; 144·47 mg/d estimated polyphenol intake; 181·65 mg/d phytoesterols.
Peripheral blood from the volunteers was collected and PBMC were isolated by density gradient centrifugation over Ficoll-Hypaque (Pharmacia, Uppsala, Sweden)(Reference Sacanella, Estruch and Gaya16). Harvested cells were washed with PBS 10 × buffer (Roche Diagnostics GmbH, Mannheim, Germany) and then counted in a haemocytometer chamber. Cell viability was estimated with trypan blue. PBMC were resuspended in RPMI-1640 (Biowhittaker, Verviers, Belgium) containing fetal bovine serum (10 %) and gentamicin (0·05 mg/ml) (RPMI-10 % fetal) up to a concentration of 1 × 106 viable cells/ml.
For each of the following phenolic acids, a 3 μm solution was prepared in RPMI-10 % fetal: 3,4-dihydroxyphenylpropionic acid (3,4-DHPPA); 3-hydroxyphenylpropionic acid (3-HPPA); 3,4-dihydroxyphenylacetic acid (3,4-DHPAA); 3-hydroxyphenylacetic acid (3-HPAA); 4-hydroxybenzoic acid (4-HBA; Sigma-Aldrich, St Louis, MO, USA); 4-hydroxyhippuric acid (4-HHA; PhytoLab GmbH & Co. KG, Vestenbergsgreuth, Germany).
PBMC were cultured with the different phenolic acid solutions in the presence of LPS (Sigma-Aldrich). Five hundred microlitres of the 1 × 106 cells/ml suspension (5 × 105 cells, total number of cells) were pre-treated (16 h at 37°C, 5 % CO2) with 250 μl of the 3 μm-phenolic acid solution (1 μm, final concentration with cells) in twenty-four-well plates. After the pre-treatment period, the cell viability was estimated with trypan blue and was higher than 95 %. LPS (1 μg/ml) was then added to the culture followed by incubation for 72 h at 37°C. Unstimulated and LPS-stimulated polyphenol-free cells were also cultured under the same conditions. Experiments were performed in duplicate. After the incubation period, the cultures were centrifuged and the supernatant collected and stored at − 80°C until analysis. Pro-inflammatory cytokines IL-1β, IL-6 and TNF-α were determined in the culture supernatants by ELISA (Bender Med Systems GmbH, Vienna, Austria). Detection limits were as follows: 0·7 pg/ml for IL-1β; 0·92 ng/ml for IL-6; 1·65 pg/ml for TNF-α.
For the statistical treatment of the data, t test for paired samples were performed using the PC software package SPSS version 14.0 (SPSS Inc., Chicago, IL, USA). Differences between values were expressed as the percentage of enhancement or inhibition. All statistical tests were two-tailed, and the significance level was 0·05.
Results
The effects of tested phenolic acids on the secretion of TNF-α, IL-1β and IL-6 from the PBMC of healthy subjects are shown in Fig. 1 and Table 1. Stimulation of PBMC with LPS significantly increased the levels of the three pro-inflammatory cytokines up to 295·48 pg/ml (min = 48·34, max = 773·45) for TNF-α, 128·38 pg/ml (min = 44·24, max = 209·91) for IL-1β and 386·58 pg/ml (min = 331·36, max = 475·57) for IL-6.
Effect of phenolic acids (1 μm) derived from microbial metabolism on the secretion of (a) TNF-α, (b) IL-1β and (c) IL-6, after stimulation with lipopolysaccharide (LPS) in peripheral blood mononuclear cells from six subjects (■, 1; □, 2; 
, 3; 
, 4; ▧, 5; 
, 6). 3,4-DHPPA, 3,4-dihydroxyphenylpropionic acid; 3-HPPA, 3-hydroxyphenylpropionic acid; 3,4-DHPAA, 3,4-dihydroxyphenylacetic acid; 3-HPAA, 3-hydroxyphenylacetic acid; 4-HBA, 4-hydroxybenzoic acid; 4-HHA, 4-hydroxyhippuric acid.
(Mean values with their standard deviations)
3,4-DHPPA, 3,4-dihydroxyphenylpropionic acid; 3-HPPA, 3-hydroxyphenylpropionic acid; 3,4-DHPAA, 3,4-dihydroxyphenylacetic acid; 3-HPAA, 3-hydroxyphenylacetic acid; 4-HBA, 4-hydroxybenzoic acid; 4-HHA, 4-hydroxyhippuric acid.
* Significant differences between LPS-stimulated cells and those in the presence of phenolic acids were determined by the t test.
3,4-DHPPA, 3,4-DHPAA and 4-HHA significantly reduced (P < 0·01 for 3,4-DHPPA and 3,4-DHPAA, P < 0·05 for 4-HHA) TNF-α secretion in LPS-stimulated PBMC by 84·9, 86·4 and 30·4 %, respectively (Table 1). Contrarily, 4-HBA significantly increased TNF-α secretion by 9·9 %. No significant changes in TNF-α secretion were recorded after the addition of the remaining tested phenolic acids, 3-HPPA and 3-HPAA. In the case of IL-1β, only 3,4-DHPPA and 3,4-DHPAA significantly (P < 0·001) reduced the secretion of this cytokine in LPS-stimulated PBMC by 93·1 and 97·9 %, respectively (Fig. 1; Table 1). Similarly, a significant reduction (P < 0·001) in IL-6 secretion from LPS-stimulated PBMC was also observed after treatment with 3,4-DHPPA and 3,4-DHPAA, resulting in an inhibition of 88·8 and 92·3 %, respectively. However, no significant changes were found in IL-1β and IL-6 levels after the addition of the remaining tested phenolic acids.
Discussion
In the present study, we observed that dihydroxylated phenolic acids derived from microbial metabolism presented marked in vitro anti-inflammatory properties, reducing the secretion of TNF-α, IL-1β and IL-6 in LPS-stimulated PBMC from healthy subjects. Six different phenolic acids (3,4-DHPPA, 3-HPPA, 3,4-DHPAA, 3-HPAA, 4-HBA and 4-HHA) derived from the microbial metabolism of polyphenols, in particular from monomeric flavanols and proanthocyanidins, were tested at a concentration level (1 μm) within the range (0·1–10 μm) found in plasma samples after the intake of a polyphenol-rich meal and recommended for in vitro studies(Reference Kroon, Clifford and Crozier3). Tested compounds were chosen on the basis of their structural features (i.e. hydroxylation pattern and side-chain length of the functional group) and on their abundance in biological fluids after the ingestion of polyphenol-rich foods. To our knowledge, the effects of microbial-derived phenolic acids on the production and release of pro-inflammatory cytokines from human PBMC have not been published previously.
Previous studies based on the consumption of catechin by human subjects resulted in an increase in 3-HPPA(Reference Das17). Consumption of catechins and proanthocyanidins from chocolate by human subjects resulted in an increased urinary excretion of 3-HPPA, 3,4-DHPAA, 3-HPAA and 3-HBA(Reference Rios, Gonthier and Remesy18). More recently, Ward et al. (Reference Ward, Croft and Puddey19) described that 3-HPPA was the main metabolite in urine when the human subjects were supplemented with grape-seed polyphenols. Studies on laboratory animals have shown similar results. Main urinary metabolites formed from rats fed a catechin diet were 3-HPPA, 3-HBA and 3-HPA(Reference Gonthier, Cheynier and Donovan20). Besides these metabolites, 3-HPAA and 4-HBA were detected in the urine of rats fed wine polyphenols(Reference Gonthier, Cheynier and Donovan20). In vitro experiments also confirmed that 3-HPPA was the most abundant metabolite produced from proanthocyanidin polymers(Reference Deprez, Brezillon and Rabot4) by human colonic microflora, whereas 3,4-DHPAA was the main metabolite from rutin, which was further dehydroxylated to 3-HPAA(Reference Aura, O'Leary and Williamson21). This compound was also found to be the major end product of the colonic metabolism of chlorogenic acid by human faecal microbiota in vitro (Reference Gonthier, Remesy and Scalbert22). The yield of microbial metabolites could be high, in particular for the polyphenols that are poorly absorbed in the small intestine. For example, for chlorogenic and caffeic acids, it represented 57·4 and 28·1 % of the total intake, respectively(Reference Gonthier, Remesy and Scalbert22). However, in order for these metabolites to be effective at a physiological level, they need to be absorbed and reach target tissues. In fact, once produced by the microbiota, some colonic metabolites could be further absorbed and reach the liver and the kidney where they could be methylated, hydroxylated or conjugated with glycine(Reference Gonthier, Cheynier and Donovan20, Reference Scheline23). Recently, the absorption mechanism of some of these colonic metabolites is beginning to be elucidated. Metabolites such as ferulic, p-coumaric, m-coumaric and 3-HPPA are absorbed by the monocarboxylic acid transporter, whereas caffeic acid and 3,4-DHPPA permeate across Caco-2 cells via the paracellular pathway(Reference Konishi and Shimizu24–Reference Konishi and Kobayashi26). Using immunohistochemical tests, Kawai et al. (Reference Kawai, Nishikawa and Shiba27) have recently confirmed that polyphenol metabolites could penetrate the tissues. Quercetin-3-glucuronide, a major metabolite of quercetin, was permeable in LPS-stimulated macrophages, and was converted into the more active aglycone, a part of which was further converted into the methylated form. These data suggest that microbial phenolic metabolites could also undergo a similar pathway in injured cells.
Atherosclerosis is now considered as an inflammatory disease(Reference Ross28). Recent epidemiological and clinical studies have shown that the Mediterranean diet or its main components, rich in polyphenols, are associated with a lower inflammatory status(Reference Estruch, Sacanella and Badia29). However, in epidemiological and even in clinical studies, it is difficult to control the effects of the diet consumed and physical activity performed(Reference Borodulin, Laatikainen and Salomaa30). Thus, in vitro studies allow us to obtain additional information in relation to the direct effect of some compounds (i.e. polyphenol metabolites) in biochemical pathways related to cardiovascular health, such as the production of pro-inflammatory cytokines that participate in the first stages of atherosclerosis. LPS is a bacterial protein and is used as a method to challenge immune cells to produce cytokines, including the inflammatory cytokines. Some of the inflammatory cytokines that are produced are those that have been associated with chronic inflammation and atherosclerosis risk.
The results found in the present study indicate that the effects of the tested phenolic acids on cytokine secretion by PBMC were structure dependent. With the exception of the effects of 4-HHA on TNF-α secretion, only the dihydroxylated phenolic acids, 3,4-DHPPA and 3,4-DHPAA, caused a statistically significant decrease in the secreted levels of the three different cytokines from LPS-stimulated PBMC (Table 1). The standard deviation of the data reflects large inter-individual difference in cytokine secretion among the volunteers. In addition, the degree of inhibition was found to be influenced by the cytokine family. The inhibition of IL-1β by both compounds was slightly higher than that for TNF-α and IL-6 (Table 1). Monohydroxylated phenolic acids (3-HPPA, 3-HPAA, 4-HBA and 4-HHA) did not produce significant changes in cytokine secretion with the exception of 4-HHA on TNF-α secretion, which produced a significant increase in this cytokine.
The present results on TNF-α, IL-1β and IL-6 secretion are in agreement with other studies performed with other polyphenols and cell types. Quercetin inhibited the expression of IL-8 and MCP-1 in TNF-α-stimulated synovial cells(Reference Sato, Miyazaki and Kambe31). Small oligomeric procyanidin fractions (monomer to tetramer) isolated from cocoa reduced the secretion of IL-1 from PHA-stimulated PBMC, whereas polymers (pentamer to decamer) produced an increase in the secreted levels(Reference Mao, Powell and Van de Water12). However, the same fractions promoted the secretion of TNF-α(Reference Mao, van de Water and Keen13). Dimeric flavanols isolated from pine bark also enhanced TNF-α levels in stimulated macrophages, while monomers strongly inhibited its secretion(Reference Park, Rimbach and Saliou32). Ramiro et al. (Reference Ramiro, Franch and Castellote33) found that epicatechin, isoquercitrin and cocoa extracts decrease the secretion of TNF-α by macrophages in a dose-dependent manner. Also in this line, it has also been reported that several flavones and flavanols inhibited TNF-α secretion by LPS-stimulated macrophages(Reference Wang and Mazza34).
The superior effect of dihydroxylated phenolic acids in comparison with the monohydroxylated ones on the inhibition of pro-inflammatory cytokines has also been reported for other tested biological properties. Thus, 3,4-DHPAA showed more potent cytotoxicity against tumour cell lines than 4-hydroxyphenylacetic acid(Reference Gao, Xu and Krul8). 3,4-DHPPA was also among the phenolic acids inhibiting the expression of P-selectin in resting platelets(Reference Rechner and Kroner5). According to Russell et al. (Reference Russell, Scobbie and Chesson35), dihydroxylated phenolic acids present a better antioxidant capacity than monohydroxylated ones due to their stabilisation into quinones.
The mechanism associated with the inhibitory or stimulatory activities of polyphenols on cytokine production may result from transcriptional and post-transcriptional events(Reference Ramiro, Franch and Castellote14). In fact, NF-κB, a transcription factor responsible for the activation of a series of cytokines, including TNF-α and IL-1β, is redox sensitive, and it is well known that the antioxidants such as flavonoids can inhibit its activation(Reference Park, Rimbach and Saliou32). Other authors have recently suggested that, besides their antioxidant effects, polyphenols could also function as signalling molecules(Reference Williams, Spencer and Rice-Evans36).
In this sense, Tedgui & Mallat(Reference Tedgui and Mallat10) have suggested that future therapeutic approaches to treat atherosclerosis may include agents that block pro-inflammatory cytokine signalling or the transcription of inflammatory-mediating molecules, among others. The results found in the present study demonstrate that due to their down-regulating effect on the production of pro-inflammatory cytokines TNF-α, IL-1β and IL-6, polyphenols such as dihydroxylated phenolic acids derived from microbial metabolism could be among the new generation of therapeutic agents for the management of immunoinflammatory diseases such as atherosclerosis.
Acknowledgements
This research was supported by national grants: CICYT's (AGL: 2004-08 378-C02-01/02 and 2006-14 228-C03-02/01); CIBER 06/03 Fisiopatología de la Obesidad y la Nutrición is an initiative of Instituto de Salud Carlos III, Spain; Ingenio-CONSOLIDER programme, Fun-c-food (CSD2007-063). M. U.-S. and N. K. thank the FPI and FPU fellowship programmes, respectively, and M. M. thank the post-doctoral programme, Juan de la Cierva, all from the Ministry of Science and Innovation. R. E. is a recipient of a grant from Fondo de Investigación Sanitaria, Madrid, Spain. The authors are not aware of any personal, financial, political or academic conflict of interest. The authors' contributions were as follows: M. M., R. E. and C. A.-L.: conception and design; M. M., N. K., M. U.-S. and M. V.-A.: analysis and interpretation of the data; M. M., N. K., R. E. and C. A.-L.: drafting of the article; M. M., N. K., M. U.-S., R. M. L.-R., R. E. and C. A.-L.: critical revision and final approval; M. M., R. E. and C. A.-L.: initiated and designed the study and obtained the funding.
