Gallic Acid Antagonizes P-Selectin–Mediated Platelet–Leukocyte Interactions
http://www.100md.com
循环学杂志 2005年第1期
the Division of Biopharmaceutics (C.C.M.A., B.C.H.L., T.J.C.v.B., E.A.L.B.), Leiden/Amsterdam Center for Drug Research, Leiden University, Gorlaeus Laboratories, Leiden, the Netherlands,Center for Molecular and Vascular Biology (A.B., K.D., M.F.H), University of Leuven, Leuven, Belgium.
Abstract
Background— Current paradigm attributes the low incidence of cardiovascular disorders in Mediterranean countries despite a high saturated fat intake, the "French paradox," to the antioxidant capacity of red wine polyphenols. Conceivably, other antiinflammatory pathways may contribute to at least a similar extent to the atheroprotective activity of these polyphenols. We have investigated whether gallic acid (GA), an abundant red wine polyphenol, modulates the activity of P-selectin, an adhesion molecule that is critically involved in the recruitment of inflammatory cells to the vessel wall and thus in atherosclerosis.
Methods and Results— GA potently inhibited the binding of a peptide antagonist (IC50, 7.2 μmol/L) and biotin-PAA-Lea-SO3H, an established high-affinity ligand, to P-selectin (IC50, 85 μmol/L). Under dynamic flow conditions, GA markedly and dose dependently attenuated the rolling of monocytic HL60 cells over P-selectin-transfected Chinese hamster ovary cells (EC50, 14.5 μmol/L) while increasing the velocity of P-selectin-dependent rolling of human blood leukocytes over a platelet monolayer. In vivo tests established that GA administration to normolipidemic C57/Bl6 and aged atherosclerotic apolipoprotein E–deficient mice impaired the baseline rolling of conjugates between activated platelets and circulating monocytes over femoral vein endothelium, as judged by online video microscopy (ED50, 1.7±0.3 and 1.5±0.4 mg · kg–1 · h–1, respectively).
Conclusions— Our findings provide a solid mechanistic foundation through which GA intervenes in major inflammatory pathobiologies by binding and antagonizing P-selectin.
Key Words: atherosclerosis ; endothelium ; inflammation ; nutrition ; platelets
Introduction
Coronary artery disease (CAD) is the major cause of death in Western society. Intriguingly, in France, where there is a pattern of saturated fat intake similar to that of other Northern European countries and the United States, the incidence of CAD is 50% lower.1 This so-called "French paradox" has fuelled considerable debate on the actual mechanism underlying this phenomenon, implicating a role for diet components and for the consumption of alcohol-containing beverages, particularly (red) wine.2–4
Wine is an abundant source of polyphenolic antioxidants, including catechins and gallates. Indeed, wine was shown to inhibit atherosclerosis in hyperlipidemic mice,5 a finding that was tentatively attributed to an inhibitory effect on LDL oxidation.6,7 Although the antioxidant theory may be plausible, recent outcomes of 3 randomized large-scale trials failed to confirm a protective effect of the antioxidant vitamin E on CAD,8 despite its inhibitory effect on lipid peroxidation in vivo.9 In analogy, treatment with probucol, a bisphenolic antioxidant, yielded contradictory results, depending on animal species and strain. Red wine polyphenols may thus be atheroprotective at least in part through nonantioxidant mechanisms. Catechins were reported to inhibit platelet-derived growth factor receptor-; (PDGFR;)–mediated signaling in vitro,10 although inhibition of PDGF signaling may have an adverse effect on plaque stability, aggravating the pathobiology.11 Resveratrol was reported to prevent platelet-leukocyte interaction and to reduce the expression of Mac-1.12 Moreover, resveratrol and other polyphenols such as quercetin were also shown to enhance endothelial nitric oxide synthase13 and to reduce the expression of the procoagulant activity of vascular cells.14
Our attention was therefore drawn to gallic acid (GA), a major antioxidant constituent in red wine that has been shown to display in situ or as an alkyl derivative antioxidant15 and anticarcinogenic16 activity. Interestingly, we previously found that derivatization of peptide antagonists for human P-selectin with GA enhanced their affinity for P-selectin >500-fold.17 P-selectin is an adhesion molecule that is intrincally implicated in atherothrombosis by mediating leukocyte-endothelium, leukocyte-platelet, and platelet-platelet interactions.18,19 The absence of P-selectin was found to result in reduced atherosclerotic lesion development and neointimal growth.20 Recent evidence confirms that both endothelial and platelet P-selectin contribute to atherosclerosis, in agreement with the occurrence of P-selectin-dependent and platelet-assisted monocyte rolling over endothelial cells.21 In addition, P-selectin has been associated with thrombosis because P-selectin activation induces an improvement in platelet cohesion and mediates platelet-leukocyte aggregation.19 The surprisingly high affinity of the GA-modified peptides led us to investigate whether GA might bind and antagonize P-selectin by itself, thus providing an additional plausible explanation for the cardioprotective effect of GA and its derivatives.
Methods
Cell Culture
Chinese hamster ovary cells stably transfected with human P-selectin (CHO-P cells) were kindly donated by Dr P. Modderman (University of Amsterdam). Cells were grown in DMEM containing 10% FCS, 5 mmol/L L-glutamine, 20 000 U penicillin/streptomycin, and 5 mmol/L nonessential amino acids. Flasks with cells were incubated at 37°C in 5% CO2 for 3 or 4 days until cells had grown nearly confluent. HL60 cells were from ATCC and grown in RPMI with 10% FCS, 5 mmol/L L-glutamine, and 20 000 U penicillin/ streptomycin.
Competition ELISA
Compounds were assayed for their ability to inhibit TM11-PO binding to human P-selectin22 or that of biotin-PAA-Lea-SO3H binding to human and mouse P-selectin and to human E- and L-selectin.23
HL60 Adhesion Assay
HL60 cells were labeled with 5 μmol/L calcein acetoxymethyl ester (calcein-AM) for 30 minutes at 37°C. HL60 cells (50 000 per well) were added to CHO-P cells (seeded in 96 well plates) in the presence or absence of GA (1 hour at 4°C). After gentle washing, fluorescence was measured.
Platelet Isolation, Labeling, and Preparation of Platelet Monolayer
Platelet-rich plasma was prepared from human24 and murine blood25; then, washed platelets were incubated for 20 minutes at 37°C with 5 μmol/L calcein-AM.21 After the platelet suspension was centrifuged at 700g for 25 minutes, fluorescent platelets were resuspended in HEPES-Tyrode’s buffer. A human platelet carpet for subsequent perfusion studies was made by overnight coating of glass coverslips with calf skin collagen (1 mg/mL in 50 mmol/L acetic acid). Coverslips were then rinsed in complete Tyrode’s buffer (containing 2 mmol/L CaCl2, 1 mmol/L MgCl2, and 1% human albumin). Whole human blood anticoagulated with 20 μg/mL hirudin was then reperfused over the collagen-coated surface in a parallel-flow chamber at a shear rate of 300 s–1 in the presence of 1 mmol/L tirofiban (Merck). In these conditions, a homogeneous carpet of activated spread platelets formed on the coverslips without the formation of platelet aggregates.
Flow Chamber and Perfusion Studies
Dynamic interactions between HL60 cells and CHO-P cell monolayers grown on glass coverslips coated with 30 μg/mL collagen S (Boehringer) were analyzed in a parallel-plate perfusion chamber as described.21 HL60 cell translocation over CHO-P monolayers was measured at wall shear rates of 300 and 600 s–1. GA was added to the HL60 cell suspensions 2 minutes before the start of perfusion. To prevent oxidation, the GA solution was prepared fresh. The average velocity of HL60 cells rolling over the CHO-P cells was calculated from the rolling distance of the HL60 cells in a 1-second time frame with the NIH Image program. Adhered HL60 cells were counted from pictures taken during the same experiment.
Measurements of Intracellular Ca2+ Concentration
HL60 cells (3x106) were incubated for 10 minutes at 37°C with 100 μmol/L of Quin-2 acetoxymethyl ester (Quin 2-AM), centrifuged, and resuspended in 3 mL Hanks’-HEPES buffer supplemented with CaCl2, glucose, and BSA. Fluorescence was measured with a Perkin Elmer LS5 fluorimeter (ex=339 nm, em=492 nm) at 37°C under constant stirring (Finitial). After 1 minute, digitonin (1 μL, 5 mmol/L) was added to achieve Fmax. Fmin was measured by the addition of 200 μL of 500-mmol/L EGTA. GA (500 μmol/L) was added 1 minute before the initial fluorescence Finitial was measured. The intracellular calcium concentration (in nmol/L) was determined by (Finitial–Fmin)/(Fmax–Fmin)x115.
Leukocyte Rolling Over Collagen-Bound Platelet Monolayers
Coverslips were coated overnight with 1 mg/mL calf skin collagen in 50 mmol/L acetic acid at 4°C and then rinsed in complete Tyrode’s buffer (containing 2 mmol/L CaCl2, 1 mmol/L MgCl2, and 1% human albumin) in the presence of 1 mmol/L Aggrastat. Under these conditions, a homogeneous carpet of spread platelets formed on the coverslips. Whole blood anticoagulated with 20 μg/mL hirudin was reperfused in a flow chamber at a shear rate of 300 s–1 for 5 minutes at 37°C over collagen-coated glass coverslips. Blood was then progressively washed out with complete Tyrode’s buffer containing 1% human albumin and perfused at a constant shear rate of 150 s–1 in the presence or absence of either 500 μmol/L GA or 25 μg/mL of the anti-P-selectin antibody WAPS12.2. After 3 minutes of washing, leukocytes translocating over the platelet carpet were visualized with an inverted microscope, digital movies were captured, and translocation velocity was calculated by image analysis.
In Vivo Inhibition of Endothelial Cell Inflammation
All animal experiments were reviewed and approved by the Institutional Review Board of the University of Leuven and performed in compliance with the guidelines of the International Society on Thrombosis and Hemostasis.26 For the study of platelet-assisted leukocyte rolling over the endothelial surface of blood vessels in vivo, online video microscopy was performed in C57/Bl6 mice. Healthy wild-type (WT) mice (6 to 12 weeks old) were anesthetized with Nembutal (70 mg/kg IP), and the jugular vein was catheterized. After the femoral vein was exposed, mice were mounted on the table of an inverted epifluorescent microscope so that the blood circulation in the femoral vein could be visualized through a Cohu CCD video camera. Calcein-AM–labeled fluorescent murine platelets (500x106 platelets in 200 μL) were then injected into the catheterized jugular vein, and baseline rolling was recorded for the labeled platelets. After 10 minutes, 50 μg/kg collagen and anti-IIb/;3 antagonist G4120 (1 mg/kg) were injected to activate circulating platelets while preventing their aggregation. After 5 minutes, the collagen-induced rolling and tumbling of leukocyte-platelet conjugates were then recorded for 10 minutes. The number of rosettes rolling was counted in the registered movies. The effect of GA on rosette rolling was investigated in 2 complementary manners. In WT mice, it was injected 5 minutes before collagen administration, and rosette rolling was analyzed as in the controls. Interactions of rosettes with the vessel wall were also studied in the femoral vein of aged atherosclerotic apolipoprotein E–deficient (ApoE–/–) mice (>1 year). In these mice, tumbling rosettes were counted 10 minutes after collagen injection to measure the steady-state rosette rolling; then, a GA bolus (0.75 to 7.5 mg/kg) was injected, followed by continuous GA infusion for up to 3 hours at 0.75 to 7.5 mg · kg–1 · h–1, during which time GA inhibition of vessel wall inflammation was measured in real time. For clarity, unless otherwise stated, "infusion of" indicates the use of a bolus injection of a given dose followed by infusion of the same dose per hour.
Statistical Analysis
Comparisons between groups were analyzed via 2-tailed Student’s t test. The ED50 was calculated via a 4-parameter logistic curve-fitting logarithm with Grafit 3.0 software.
Results
The ability of GA to inhibit human P-selectin binding was studied in a competition assay based on a P-selectin-binding peptide (TM11, H2N-CDVEWVDVSSLEWDLPC-COOH). TM11-PO, a tetrameric complex between streptavidin-PO and TM11-biotin, has previously been shown to bind with high affinity and specificity to human P-selectin.22 Interestingly, GA-derived peptides containing the consensus sequence of this peptide, like GA-(E)WVDV, displayed a considerably higher affinity for P-selectin in this competition assay (IC50, 15 to 37 nmol/L)17 than the parental peptide, which bound P-selectin at micromolar concentration (IC50, 8 μmol/L). GA by itself appeared to be a potent inhibitor of TM11-PO binding to human P-selectin at low micromolar concentrations (IC50, 7.2 μmol/L) (Figure 1A).
The specificity of GA was further assessed in a competition assay based on a biotinylated sulfo-Lewis A derivatized polyacrylamide (ie, biotin-PAA-Lea-SO3H, an established selectin ligand)23 because TM11-PO does not bind to E-selectin. Although GA potently inhibited human P-selectin (85 μmol/L), it was ineffective against E-selectin binding (20% at a concentration of 1 mmol/L; Figure 1B), and it was a moderate inhibitor of L-selectin (IC50, 241 μmol/L). GA binding to P-selectin appeared to be rather species independent because GA also antagonized mouse P-selectin binding with a similar affinity (IC50, 199 μmol/L). Besides GA, n-dodecyl gallate, monophenols (4-hydroxy benzoic acid), and polyphenols [caffeic acid, and (-)-epigallocatechine gallate (EGCG)] were also tested for their capacity to antagonize human P-selectin binding (the Table). Only GA and EGCG were able to inhibit biotin-PAA-Lea-SO3 binding to human P-selectin.
Competition by Various Compounds of the Biotin-PAA-Lea-SO3 Binding to Human P-Selectin
Subsequently, GA was tested in a static cellular competition assay for its ability to inhibit the adhesion of HL60 cells to CHO-P cells. These macrophage-derived HL60 cells have a high expression of P-selectin glycoprotein ligand-1 (PSGL-1), the endogenous high-affinity P-selectin ligand.27,28 Indeed, GA potently inhibited this interaction (IC50, 14 μmol/L; Figure 1C). Under dynamic flow conditions, with GA added to the HL60 cells just before perfusion, GA dose dependently increased the rolling velocity of HL60 cells across a CHO-P monolayer (wall shear rate, 150 s–1) at an EC50 of 14.5 μmol/L, which is essentially similar to the capacity of GA to inhibit HL60 cell adhesion (Figure 2A). In the same experimental setup, the number of adhering HL60 cells was determined. Already at 50 μmol/L, GA significantly decreased the HL60 cell adherence to the CHO-P monolayer compared with the untreated control (–34% and –43%, respectively; Figure 2B), regardless of the wall shear rate (300 and 600 s–1). At 250 μmol/L GA, inhibition was even more pronounced (–41% and –54%, respectively). Calcium influx in HL60 monocytes was not influenced by the addition of GA (500 μmol/L), suggesting that the observed inhibitory activity of GA is not caused by modulating signal transduction (Figure 2C).
Flow cytometry analysis of rosette formation between activated human platelets and monocytes in human blood revealed extensive platelet binding to monocytes when platelets were activated by collagen or the tromboxane A2 analog U46619 (not shown). This binding was entirely P-selectin dependent because the neutralizing anti-P-selectin antibody WAPS12.2 prevented conjugate formation entirely (Figure 3A). When the rolling of resting leukocytes over a carpet of activated platelets was analyzed, GA induced a sharp rise in the rolling speed of those cells retained on the platelet surface (Figure 3B), reflecting impaired adhesion. However, the platelet-endothelium interaction was not disturbed because the number of activated platelets adhering to activated EAhy-926 endothelial cells remained unaltered after the addition of GA (Figure 3C). Similarly, GA did not affect the adherence of calcein-labeled THP-1 cells to activated EAhy-926 endothelial cells (Figure 3D). This implies that intercellular/vascular cell adhesion molecule function is undisturbed in the presence of GA.
Injecting fluorescently labeled resting platelets into C57/Bl6 mice did not trigger platelet rolling or platelet adhesion on the femoral vein endothelium (not shown). When circulating platelets were activated by the intravenous administration of collagen, with the concomitant injection of anti-IIb/;3 antagonist G4120 to avoid platelet aggregation, conjugates formed in situ between activated platelets and leukocytes. These rosettes were found to tumble along the vessel wall, as illustrated in Figure 4, which shows in each box an individual rosette representing 2 to 3 labeled platelets attached to a single unlabeled, and thus invisible, leukocyte (individual spots in Figure 4). Each box consists of an overlay of a series of consecutive pictures taken at 0.1-second intervals. Calculation of the number of rolling conjugates revealed an average of 46±7 rosettes rolling across the field (0.067 mm2) per minute (n=30 fields).
Pretreatment of mice with GA before platelet activation dose dependently reduced conjugate tumbling with an ED50 of 1.5±0.4 mg · kg–1 · h–1 (representing the dose of a combined injection/infusion protocol as described in Methods) (Figure 5A). Conjugate-vessel wall interactions were almost abrogated at an infusion dose of 2.25 mg GA · kg–1 · h–1 (to 6±2 conjugates per field per minute; n=16 fields). Tethering of individual platelets was marginal (Figure 6 and Data Supplement Movies).
Injection of labeled platelets into aged atherosclerotic ApoE–/– mice and subsequent in situ activation with collagen revealed increased rosette-endothelium interactions, as expected for endothelium activated by hyperlipidemia, ie, overexpressing von Willebrand factor,24,29 and upregulating other adhesion molecules such as intercellular adhesion molecule-1, vascular cell adhesion molecule-1, or E-selectin.30 In addition to increased rosette tumbling, the femoral venous endothelium of ApoE–/– mice uncovered increased tethering of single individual platelets (Figure 6), in agreement with in vitro and ex vivo perfusion studies over endothelial cells and rabbit aorta endothelium.24 In analogy to the studies in normolipidemic mice, in atherosclerotic ApoE–/– mice injected with collagen to activate platelets and preform conjugates, the infusion of 3.75 mg GA · kg–1 · h–1 also led to a sharp decrease in the number of preformed tumbling conjugates down to 34±13% of control value (P<0.01) without affecting single platelet tethering (Figure 5B).
Discussion
Polyphenols, (red) wine-derived antioxidants, have received considerable attention during the last decade because they are likely to reduce the risk of CAD (the French paradox).1 The atheroprotective effect of polyphenols has been tentatively attributed to their capacity to inhibit LDL oxidation6,7,31 and to inhibit the expression of proinflammatory proteins, including PDGFR;.10 However, because large-scale trials have been largely inconclusive in showing an effect of other antioxidants, additional nonoxidant pathways may be at least equally prominent. The notion that red wine polyphenols act as a cardioprotective by inhibiting PDGFR; activity is contradicted by recent studies showing that the PDGFR; axis is instrumental in maintaining plaque stability and in preventing aneurysm formation by promoting vascular smooth muscle cell migration to and their proliferation in the cap.11,32 The present report shows that GA, one of the most abundant polyphenols in red wine, can potently intervene in P-selectin function, an adhesion molecule that is intrincally involved in the initial rolling of monocytes along the inflamed endothelium33 and in thrombosis.19,34
Previously, we reported that derivatization of specific human P-selectin-binding peptides with GA led to a dramatically enhanced affinity of the peptide toward P-selectin.17 This led to the presumption that GA by itself may specifically interact with P-selectin. Indeed, the polyphenol was found to inhibit the binding of 2 different selectin ligands, TM11-PO22 and biotin-PAA-Lea-SO3H,23 to P-selectin at micromolar affinity. Conceivably, the TM11-based assay is most reflective of the functional activity of GA, which is corroborated by the finding that static and dynamic interactions between PSGL-1 displaying HL60 cells and CHO-P monolayers were inhibited by GA. The interaction appears to be rather specific for P-selectin, although GA also displayed some affinity for L-selectin, which is in concordance with previous studies showing that the ligand binding profiles of P- and L-selectin largely overlap.35
In vitro, GA does not disturb platelet-endothelium or leukocyte-endothelium interaction. However, GA induced a sharp rise in the rolling speed of leukocytes retained on the platelet surface, reflecting impaired adhesion. In vivo studies of platelet-assisted inflammation indicate that GA can interrupt platelet-assisted rolling and tumbling of platelet-leukocyte rosettes over the endothelium of normocholesterolemic and hypercholesterolemic mice without affecting the deposition of single platelets on activated endothelium. This effect could be observed when GA was administered before and even after platelets had been activated.
GA effects in vitro and in vivo are not attributable to its antioxidant activity. n-Dodecyl gallate displays an antioxidant activity similar to that of GA, although it is unable to inhibit P-selectin binding at low concentrations. Similarly, other equally potent and structurally related antioxidants such as caffeic acid and dihydroxy benzoic acid are completely ineffective in inhibiting P-selectin. Finally, precautions were taken to prevent oxidation of the gallate during the experiments (which can easily be appreciated because oxidized gallate is dark green/brown).
GA is abundantly available not only in wine (grapes) but also in (green) tea. The total phenolic content in various red wines was analyzed as 1100 to 3165 mg/L, of which GA attributes 35 to 70 mg/L and the epicatechine gallate derivatives attribute 120 to 360 mg/L.36 The GA content of green tea is even higher (2.2 mg/mL, corresponding to 149 mg/g dry leaves).37 A study by Lee et al38 showed that cocoa contained higher total gallate levels than green tea or red wine (611 versus 165 and 340 mg per serving, respectively). Oral administration of only 50 mg GA to a healthy volunteer resulted in plasma levels of GA and its major metabolite 4-O-methylgallic acid in the low micromolar range,39 implying that the observed EC50 values of 15 to 40 μmol/L can be readily reached after consumption of only 100 to 500 mg GA.
In addition to being involved in atherosclerosis and thrombosis, P-selectin was reported to play a role in metastasis by promoting tumor growth and to facilitate the metastatic seeding of mucin-producing carcinomas.40 P-selectin deficiency was shown to be accompanied by reduced implanted carcinoma cells growth and reduced metastasis formation.40 In addition, P-selectin is expressed by several metastatic pancreatic tumor cells.41 Thus, it is interesting that Ohno et al42 recently demonstrated that GA administration (50 mg/kg) could inhibit metastasis formation and growth.
In conclusion, GA inhibits P-selectin-mediated inflammation both in vitro and in vivo. This study provides for the first time a molecular mechanism through which GA can exert a beneficial effect in several pathobiologies such as CAD, thrombosis, and cancer by binding and antagonizing P-selectin under static and dynamic conditions already at concentrations readily achieved after moderate wine, (green) tea, or cocoa consumption.
Acknowledgments
This study was partly funded by grants from Yamanouchi Japan (C.C.M.A, B.C.H.L), the Netherlands Organization of Scientific Research (NWO grant 016.026.019 to E.A.L.B), and the Dutch Heart Foundation (NHS, 2003T201). The support of the FWO Vlaanderen (project G.0376.01) is recognized. During this study, A.B. was the holder of a Marie Curie fellowship. K.D. is the recipient of a junior fellowship from the University Hospital Leuven. The skillful assistance of Petra Vandervoort during the cell culture and coverslip preparation and of Hans de Bont during Ca flux measurements is highly appreciated.
Footnotes
Drs Appeldoorn and Bonnefoy contributed equally to this work.
The online-only Data Supplement, which contains 4 movies that supplement the Results, can be found with this article at http://www.circulationaha.org.
References
Renaud S, de Lorgeril M. Wine, alcohol, platelets, and the French paradox for coronary heart disease. Lancet. 1992; 339: 1523–1526.
Pearson TA. Alcohol and heart disease. Circulation. 1996; 94: 3023–3025.
Gronbaek M. Alcohol, type of alcohol, and all-cause and coronary heart disease mortality. Ann N Y Acad Sci. 2002; 957: 16–20.
Vogel RA. Alcohol, heart disease, and mortality: a review. Rev Cardiovasc Med. 2002; 3: 7–13.
Aviram M, Hayek T, Fuhrman B. Red wine consumption inhibits LDL oxidation and aggregation in humans and in atherosclerotic mice. Biofactors. 1997; 6: 415–419.
Frankel EN, Kanner J, German JB, Parks E, Kinsella JE. Inhibition of oxidation of human low-density lipoprotein by phenolic substances in red wine. Lancet. 1993; 341: 454–457.
Kondo K, Matsumoto A, Kurata H, Tanahashi H, Koda H, Amachi T, Itakura H. Inhibition of oxidation of low-density lipoprotein with red wine. Lancet. 1994; 344: 1152–1159.
Clarke R, Armitage J. Antioxidant vitamins and risk of cardiovascular disease: review of large-scale randomised trials. Cardiovasc Drugs Ther. 2002; 16: 411–415.
Kaikkonen J, Porkkala-Sarataho E, Morrow JD, Roberts LJ 2nd, Nyyssonen K, Salonen R, Tuomainen TP, Ristonmaa U, Poulsen HE, Salonen JT. Supplementation with vitamin E but not with vitamin C lowers lipid peroxidation in vivo in mildly hypercholesterolemic men. Free Radical Res. 2001; 35: 967–978.
Rosenkranz S, Knirel D, Dietrich H, Flesch M, Erdmann E, Bohm M. Inhibition of the PDGF receptor by red wine flavonoids provides a molecular explanation for the "French paradox." FASEB J. 2002; 16: 1958–1960.
Sano H, Sudo T, Yokode M, Murayama T, Kataoka H, Takakura N, Nishikawa S, Nishikawa SI, Kita T. Functional blockade of platelet-derived growth factor receptor-; but not of receptor- prevents vascular smooth muscle cell accumulation in fibrous cap lesions in apolipoprotein E–deficient mice. Circulation. 2001; 103: 2955–2960.
Rotondo S, Rajtar G, Manarini S, Celardo A, Rotilio D, de Gaetano G, Evangelista V, Cerletti C. Effect of trans-resveratrol, a natural polyphenolic compound, on human polymorphonuclear leukocyte function. Br J Pharmacol. 1998; 123: 1691–1699.
Wallerath T, Deckert G, Ternes T, Anderson H, Li H, Witte K, F;rstermann U. Resveratrol, a polyphenolic phytoalexin present in red wine, enhances expression and activity of endothelial nitric oxide synthase. Circulation. 2002; 106: 1652–1658.
Pendurthi UR, Williams JT, Mohan Rao LV. Resveratrol, a polyphenolic compound found in wine, inhibits tissue factor expression in vascular cells: a possible mechanism for the cardiovascular benefits associated with moderate consumption of wine. Arterioscler Thromb Vasc Biol. 1999; 19: 419–426.
Landrault N, Poucheret P, Ravel P, Gasc F, Cros G, Teissedre PL. Antioxidant capacities and phenolic levels of French wines from different varieties and vintages. J Agric Food Chem. 2001; 49: 3341–3348.
Ohno Y, Fukuda K, Takemura G, Toyota M, Watanabe M, Yasuda N, Xinbin Q, Maruyama R, Akao S, Gotou K, Fujiwara T, Fujiwara H. Induction of apoptosis by gallic acid in lung cancer cells. Anticancer Drugs. 1999; 10: 845–851.
Appeldoorn CCM, Molenaar TJM, Bonnefoy A, van Leeuwen SH, Vandervoort PAH, Hoylaerts MF, van Berkel TJC, Biessen EAL. Rational optimization of a short human P-selectin binding peptide leads to nanomolar affinity antagonists. J Biol Chem. 2003; 278: 10201–10207.
Vestweber D, Blanks JE. Mechanisms that regulate the function of the selectins and their ligands. Physiol Rev. 1999; 79: 181–213.
Merten M, Chow T, Hellums JD, Thiagarajan P. A new role for P-selectin in shear-induced platelet aggregation. Circulation. 2000; 102: 2045–2050.
Johnson RC, Chapman SM, Dong ZM, Ordovas JM, Mayadas TN, Herz J, Hynes RO, Schaefer EJ, Wagner DD. Absence of P-selectin delays fatty streak formation in mice. J Clin Invest. 1997; 99: 1037–1043.
Theilmeier G, Lenaerts T, Remacle C, Collen D, Vermylen J, Hoylaerts MF. Circulating activated platelets assist THP-1 monocytoid/endothelial cell interaction under shear stress. Blood. 1999; 94: 2725–2734.
Molenaar TJM, Appeldoorn CCM, de Haas SAM, Michon IN, Bonnefoy A, Hoylaerts MF, Pannekoek H, van Berkel TJC, Kuiper J, Biessen EAL. Specific inhibition of P-selectin mediated cell adhesion by phage display derived peptide antagonists. Blood. 2002; 100: 3570–3577.
Weitz Schmidt G, Stokmaier D, Scheel G, Nifant’ev NE, Tuzikov AB, Bovin NV. An E-selectin binding assay based on a polyacrylamide-type glycoconjugate. Anal Biochem. 1996; 238: 184–190.
Theilmeier G, Michiels C, Spaepen E, Vreys I, Collen D, Vermylen J, Hoylaerts MF. Endothelial von Willebrand factor recruits platelets to atherosclerosis-prone sites in response to hypercholesterolemia. Blood. 2002; 99: 4486–4493.
Oury C, Kuijpers MJ, Toth-Zsamboki E, Bonnefoy A, Danloy S, Vreys I, Feijge MA, De Vos R, Vermylen J, Heemskerk JW, Hoylaerts MF. Overexpression of the platelet P2X1 ion channel in transgenic mice generates a novel prothrombotic phenotype. Blood. 2003; 101: 3969–3976.
Giles AR. Guidelines for the use of animals in biomedical research. Thromb Haemost. 1987; 58: 1078–1084.
Sako D, Xiao Jia C, Barone KM, Vachino G, White HM, Shaw G, Veldman GM, Bean KM, Ahern TJ, Furie B, Cumming DA, Larsen GR. Expression cloning of a functional glycoprotein ligand for P-selectin. Cell. 1993; 75: 1179–1186.
Wilkins PP, McEver RP, Cummings RD. Structures of the O-glycans on P-selectin glycoprotein ligand-1 from HL-60 cells. J Biol Chem. 1996; 271: 18732–18742.
De Meyer GR, Hoylaerts MF, Kockx MM, Yamamoto H, Herman AG, Bult H. Intimal deposition of functional von Willebrand factor in atherogenesis. Arterioscler Thromb Vasc Biol. 1999; 19: 2524–2534.
Molenaar TJ, Twisk J, de Haas SA, Peterse N, Vogelaar BJ, van Leeuwen SH, Michon IN, van Berkel TJ, Kuiper J, Biessen EA. P-selectin as a candidate target in atherosclerosis. Biochem Pharmacol. 2003; 66: 859–866.
Kerry NL, Abbey M. Red wine and fractionated phenolic compounds prepared from red wine inhibit low density lipoprotein oxidation in vitro. Atherosclerosis. 1997; 135: 93–102.
Boucher P, Gotthardt M, Li WP, Anderson RG, Herz J. LRP: role in vascular wall integrity and protection from atherosclerosis. Science. 2003; 300: 329–332.
Lasky LA. Selectin-carbohydrate interactions and the initiation of the inflammatory response. Annu Rev Biochem. 1995; 64: 113–139.
Freedman JE, Loscalzo J. Platelet-monocyte aggregates: bridging thrombosis and inflammation. Circulation. 2002; 105: 2130–2132.
Nelson RM, Cecconi O, Roberts WG, Aruffo A, Linhardt RJ, Bevilacqua MP. Heparin oligosaccharides bind L- and P-selectin and inhibit acute inflammation. Blood. 1993; 82: 3253–3258.
Burns J, Gardner PT, O’Neil J, Crawford S, Morecroft I, McPhail DB, Lister C, Matthews D, MacLean MR, Lean MEJ, Duthie GG, Crozier A. Relationship among antioxidant activity, vasodilation capacity, and phenolic content of red wines. J Agric Food Chem. 2000; 48: 220–230.
Chandra S, De Mejia Gonzalez E. Polyphenolic compounds, antioxidant capacity, and quinone reductase activity of an aqueous extract of Ardisia compressa in comparison to mate (Ilex paraguariensis) and green (Camellia sinensis) teas. J Agric Food Chem. 2004; 52: 3583–3589.
Lee KW, Kim YJ, Lee HJ, Lee CY. Cocoa has more phenolic phytochemicals and a higher antioxidant capacity than teas and red wine. J Agric Food Chem. 2003; 51: 7292–7295.
Shahrzad S, Bitsch I. Determination of gallic acid and its metabolites in human plasma and urine by high-performance liquid chromatography. J Chromatogr B Biomed Sci Appl. 1998; 705: 87–95.
Kim YJ, Borsig L, Varki NM, Varki A. P-selectin deficiency attenuates tumor growth and metastasis. Proc Natl Acad Sci U S A. 1998; 95: 9325–9330.
Iwamura T, Caffrey TC, Kitamura N, Yamanari H, Setoguchi T, Hollingsworth MA. P-selectin expression in a metastatic pancreatic tumor cell line (SUIT-2). Cancer Res. 1997; 57: 1206–1212.
Ohno T, Inoue M, Ogihara Y. Cytotoxic activity of gallic acid against liver metastasis of mastocytoma cells P-815. Anticancer Res. 2001; 21: 3875–3880.(Chantal C.M. Appeldoorn, )
Abstract
Background— Current paradigm attributes the low incidence of cardiovascular disorders in Mediterranean countries despite a high saturated fat intake, the "French paradox," to the antioxidant capacity of red wine polyphenols. Conceivably, other antiinflammatory pathways may contribute to at least a similar extent to the atheroprotective activity of these polyphenols. We have investigated whether gallic acid (GA), an abundant red wine polyphenol, modulates the activity of P-selectin, an adhesion molecule that is critically involved in the recruitment of inflammatory cells to the vessel wall and thus in atherosclerosis.
Methods and Results— GA potently inhibited the binding of a peptide antagonist (IC50, 7.2 μmol/L) and biotin-PAA-Lea-SO3H, an established high-affinity ligand, to P-selectin (IC50, 85 μmol/L). Under dynamic flow conditions, GA markedly and dose dependently attenuated the rolling of monocytic HL60 cells over P-selectin-transfected Chinese hamster ovary cells (EC50, 14.5 μmol/L) while increasing the velocity of P-selectin-dependent rolling of human blood leukocytes over a platelet monolayer. In vivo tests established that GA administration to normolipidemic C57/Bl6 and aged atherosclerotic apolipoprotein E–deficient mice impaired the baseline rolling of conjugates between activated platelets and circulating monocytes over femoral vein endothelium, as judged by online video microscopy (ED50, 1.7±0.3 and 1.5±0.4 mg · kg–1 · h–1, respectively).
Conclusions— Our findings provide a solid mechanistic foundation through which GA intervenes in major inflammatory pathobiologies by binding and antagonizing P-selectin.
Key Words: atherosclerosis ; endothelium ; inflammation ; nutrition ; platelets
Introduction
Coronary artery disease (CAD) is the major cause of death in Western society. Intriguingly, in France, where there is a pattern of saturated fat intake similar to that of other Northern European countries and the United States, the incidence of CAD is 50% lower.1 This so-called "French paradox" has fuelled considerable debate on the actual mechanism underlying this phenomenon, implicating a role for diet components and for the consumption of alcohol-containing beverages, particularly (red) wine.2–4
Wine is an abundant source of polyphenolic antioxidants, including catechins and gallates. Indeed, wine was shown to inhibit atherosclerosis in hyperlipidemic mice,5 a finding that was tentatively attributed to an inhibitory effect on LDL oxidation.6,7 Although the antioxidant theory may be plausible, recent outcomes of 3 randomized large-scale trials failed to confirm a protective effect of the antioxidant vitamin E on CAD,8 despite its inhibitory effect on lipid peroxidation in vivo.9 In analogy, treatment with probucol, a bisphenolic antioxidant, yielded contradictory results, depending on animal species and strain. Red wine polyphenols may thus be atheroprotective at least in part through nonantioxidant mechanisms. Catechins were reported to inhibit platelet-derived growth factor receptor-; (PDGFR;)–mediated signaling in vitro,10 although inhibition of PDGF signaling may have an adverse effect on plaque stability, aggravating the pathobiology.11 Resveratrol was reported to prevent platelet-leukocyte interaction and to reduce the expression of Mac-1.12 Moreover, resveratrol and other polyphenols such as quercetin were also shown to enhance endothelial nitric oxide synthase13 and to reduce the expression of the procoagulant activity of vascular cells.14
Our attention was therefore drawn to gallic acid (GA), a major antioxidant constituent in red wine that has been shown to display in situ or as an alkyl derivative antioxidant15 and anticarcinogenic16 activity. Interestingly, we previously found that derivatization of peptide antagonists for human P-selectin with GA enhanced their affinity for P-selectin >500-fold.17 P-selectin is an adhesion molecule that is intrincally implicated in atherothrombosis by mediating leukocyte-endothelium, leukocyte-platelet, and platelet-platelet interactions.18,19 The absence of P-selectin was found to result in reduced atherosclerotic lesion development and neointimal growth.20 Recent evidence confirms that both endothelial and platelet P-selectin contribute to atherosclerosis, in agreement with the occurrence of P-selectin-dependent and platelet-assisted monocyte rolling over endothelial cells.21 In addition, P-selectin has been associated with thrombosis because P-selectin activation induces an improvement in platelet cohesion and mediates platelet-leukocyte aggregation.19 The surprisingly high affinity of the GA-modified peptides led us to investigate whether GA might bind and antagonize P-selectin by itself, thus providing an additional plausible explanation for the cardioprotective effect of GA and its derivatives.
Methods
Cell Culture
Chinese hamster ovary cells stably transfected with human P-selectin (CHO-P cells) were kindly donated by Dr P. Modderman (University of Amsterdam). Cells were grown in DMEM containing 10% FCS, 5 mmol/L L-glutamine, 20 000 U penicillin/streptomycin, and 5 mmol/L nonessential amino acids. Flasks with cells were incubated at 37°C in 5% CO2 for 3 or 4 days until cells had grown nearly confluent. HL60 cells were from ATCC and grown in RPMI with 10% FCS, 5 mmol/L L-glutamine, and 20 000 U penicillin/ streptomycin.
Competition ELISA
Compounds were assayed for their ability to inhibit TM11-PO binding to human P-selectin22 or that of biotin-PAA-Lea-SO3H binding to human and mouse P-selectin and to human E- and L-selectin.23
HL60 Adhesion Assay
HL60 cells were labeled with 5 μmol/L calcein acetoxymethyl ester (calcein-AM) for 30 minutes at 37°C. HL60 cells (50 000 per well) were added to CHO-P cells (seeded in 96 well plates) in the presence or absence of GA (1 hour at 4°C). After gentle washing, fluorescence was measured.
Platelet Isolation, Labeling, and Preparation of Platelet Monolayer
Platelet-rich plasma was prepared from human24 and murine blood25; then, washed platelets were incubated for 20 minutes at 37°C with 5 μmol/L calcein-AM.21 After the platelet suspension was centrifuged at 700g for 25 minutes, fluorescent platelets were resuspended in HEPES-Tyrode’s buffer. A human platelet carpet for subsequent perfusion studies was made by overnight coating of glass coverslips with calf skin collagen (1 mg/mL in 50 mmol/L acetic acid). Coverslips were then rinsed in complete Tyrode’s buffer (containing 2 mmol/L CaCl2, 1 mmol/L MgCl2, and 1% human albumin). Whole human blood anticoagulated with 20 μg/mL hirudin was then reperfused over the collagen-coated surface in a parallel-flow chamber at a shear rate of 300 s–1 in the presence of 1 mmol/L tirofiban (Merck). In these conditions, a homogeneous carpet of activated spread platelets formed on the coverslips without the formation of platelet aggregates.
Flow Chamber and Perfusion Studies
Dynamic interactions between HL60 cells and CHO-P cell monolayers grown on glass coverslips coated with 30 μg/mL collagen S (Boehringer) were analyzed in a parallel-plate perfusion chamber as described.21 HL60 cell translocation over CHO-P monolayers was measured at wall shear rates of 300 and 600 s–1. GA was added to the HL60 cell suspensions 2 minutes before the start of perfusion. To prevent oxidation, the GA solution was prepared fresh. The average velocity of HL60 cells rolling over the CHO-P cells was calculated from the rolling distance of the HL60 cells in a 1-second time frame with the NIH Image program. Adhered HL60 cells were counted from pictures taken during the same experiment.
Measurements of Intracellular Ca2+ Concentration
HL60 cells (3x106) were incubated for 10 minutes at 37°C with 100 μmol/L of Quin-2 acetoxymethyl ester (Quin 2-AM), centrifuged, and resuspended in 3 mL Hanks’-HEPES buffer supplemented with CaCl2, glucose, and BSA. Fluorescence was measured with a Perkin Elmer LS5 fluorimeter (ex=339 nm, em=492 nm) at 37°C under constant stirring (Finitial). After 1 minute, digitonin (1 μL, 5 mmol/L) was added to achieve Fmax. Fmin was measured by the addition of 200 μL of 500-mmol/L EGTA. GA (500 μmol/L) was added 1 minute before the initial fluorescence Finitial was measured. The intracellular calcium concentration (in nmol/L) was determined by (Finitial–Fmin)/(Fmax–Fmin)x115.
Leukocyte Rolling Over Collagen-Bound Platelet Monolayers
Coverslips were coated overnight with 1 mg/mL calf skin collagen in 50 mmol/L acetic acid at 4°C and then rinsed in complete Tyrode’s buffer (containing 2 mmol/L CaCl2, 1 mmol/L MgCl2, and 1% human albumin) in the presence of 1 mmol/L Aggrastat. Under these conditions, a homogeneous carpet of spread platelets formed on the coverslips. Whole blood anticoagulated with 20 μg/mL hirudin was reperfused in a flow chamber at a shear rate of 300 s–1 for 5 minutes at 37°C over collagen-coated glass coverslips. Blood was then progressively washed out with complete Tyrode’s buffer containing 1% human albumin and perfused at a constant shear rate of 150 s–1 in the presence or absence of either 500 μmol/L GA or 25 μg/mL of the anti-P-selectin antibody WAPS12.2. After 3 minutes of washing, leukocytes translocating over the platelet carpet were visualized with an inverted microscope, digital movies were captured, and translocation velocity was calculated by image analysis.
In Vivo Inhibition of Endothelial Cell Inflammation
All animal experiments were reviewed and approved by the Institutional Review Board of the University of Leuven and performed in compliance with the guidelines of the International Society on Thrombosis and Hemostasis.26 For the study of platelet-assisted leukocyte rolling over the endothelial surface of blood vessels in vivo, online video microscopy was performed in C57/Bl6 mice. Healthy wild-type (WT) mice (6 to 12 weeks old) were anesthetized with Nembutal (70 mg/kg IP), and the jugular vein was catheterized. After the femoral vein was exposed, mice were mounted on the table of an inverted epifluorescent microscope so that the blood circulation in the femoral vein could be visualized through a Cohu CCD video camera. Calcein-AM–labeled fluorescent murine platelets (500x106 platelets in 200 μL) were then injected into the catheterized jugular vein, and baseline rolling was recorded for the labeled platelets. After 10 minutes, 50 μg/kg collagen and anti-IIb/;3 antagonist G4120 (1 mg/kg) were injected to activate circulating platelets while preventing their aggregation. After 5 minutes, the collagen-induced rolling and tumbling of leukocyte-platelet conjugates were then recorded for 10 minutes. The number of rosettes rolling was counted in the registered movies. The effect of GA on rosette rolling was investigated in 2 complementary manners. In WT mice, it was injected 5 minutes before collagen administration, and rosette rolling was analyzed as in the controls. Interactions of rosettes with the vessel wall were also studied in the femoral vein of aged atherosclerotic apolipoprotein E–deficient (ApoE–/–) mice (>1 year). In these mice, tumbling rosettes were counted 10 minutes after collagen injection to measure the steady-state rosette rolling; then, a GA bolus (0.75 to 7.5 mg/kg) was injected, followed by continuous GA infusion for up to 3 hours at 0.75 to 7.5 mg · kg–1 · h–1, during which time GA inhibition of vessel wall inflammation was measured in real time. For clarity, unless otherwise stated, "infusion of" indicates the use of a bolus injection of a given dose followed by infusion of the same dose per hour.
Statistical Analysis
Comparisons between groups were analyzed via 2-tailed Student’s t test. The ED50 was calculated via a 4-parameter logistic curve-fitting logarithm with Grafit 3.0 software.
Results
The ability of GA to inhibit human P-selectin binding was studied in a competition assay based on a P-selectin-binding peptide (TM11, H2N-CDVEWVDVSSLEWDLPC-COOH). TM11-PO, a tetrameric complex between streptavidin-PO and TM11-biotin, has previously been shown to bind with high affinity and specificity to human P-selectin.22 Interestingly, GA-derived peptides containing the consensus sequence of this peptide, like GA-(E)WVDV, displayed a considerably higher affinity for P-selectin in this competition assay (IC50, 15 to 37 nmol/L)17 than the parental peptide, which bound P-selectin at micromolar concentration (IC50, 8 μmol/L). GA by itself appeared to be a potent inhibitor of TM11-PO binding to human P-selectin at low micromolar concentrations (IC50, 7.2 μmol/L) (Figure 1A).
The specificity of GA was further assessed in a competition assay based on a biotinylated sulfo-Lewis A derivatized polyacrylamide (ie, biotin-PAA-Lea-SO3H, an established selectin ligand)23 because TM11-PO does not bind to E-selectin. Although GA potently inhibited human P-selectin (85 μmol/L), it was ineffective against E-selectin binding (20% at a concentration of 1 mmol/L; Figure 1B), and it was a moderate inhibitor of L-selectin (IC50, 241 μmol/L). GA binding to P-selectin appeared to be rather species independent because GA also antagonized mouse P-selectin binding with a similar affinity (IC50, 199 μmol/L). Besides GA, n-dodecyl gallate, monophenols (4-hydroxy benzoic acid), and polyphenols [caffeic acid, and (-)-epigallocatechine gallate (EGCG)] were also tested for their capacity to antagonize human P-selectin binding (the Table). Only GA and EGCG were able to inhibit biotin-PAA-Lea-SO3 binding to human P-selectin.
Competition by Various Compounds of the Biotin-PAA-Lea-SO3 Binding to Human P-Selectin
Subsequently, GA was tested in a static cellular competition assay for its ability to inhibit the adhesion of HL60 cells to CHO-P cells. These macrophage-derived HL60 cells have a high expression of P-selectin glycoprotein ligand-1 (PSGL-1), the endogenous high-affinity P-selectin ligand.27,28 Indeed, GA potently inhibited this interaction (IC50, 14 μmol/L; Figure 1C). Under dynamic flow conditions, with GA added to the HL60 cells just before perfusion, GA dose dependently increased the rolling velocity of HL60 cells across a CHO-P monolayer (wall shear rate, 150 s–1) at an EC50 of 14.5 μmol/L, which is essentially similar to the capacity of GA to inhibit HL60 cell adhesion (Figure 2A). In the same experimental setup, the number of adhering HL60 cells was determined. Already at 50 μmol/L, GA significantly decreased the HL60 cell adherence to the CHO-P monolayer compared with the untreated control (–34% and –43%, respectively; Figure 2B), regardless of the wall shear rate (300 and 600 s–1). At 250 μmol/L GA, inhibition was even more pronounced (–41% and –54%, respectively). Calcium influx in HL60 monocytes was not influenced by the addition of GA (500 μmol/L), suggesting that the observed inhibitory activity of GA is not caused by modulating signal transduction (Figure 2C).
Flow cytometry analysis of rosette formation between activated human platelets and monocytes in human blood revealed extensive platelet binding to monocytes when platelets were activated by collagen or the tromboxane A2 analog U46619 (not shown). This binding was entirely P-selectin dependent because the neutralizing anti-P-selectin antibody WAPS12.2 prevented conjugate formation entirely (Figure 3A). When the rolling of resting leukocytes over a carpet of activated platelets was analyzed, GA induced a sharp rise in the rolling speed of those cells retained on the platelet surface (Figure 3B), reflecting impaired adhesion. However, the platelet-endothelium interaction was not disturbed because the number of activated platelets adhering to activated EAhy-926 endothelial cells remained unaltered after the addition of GA (Figure 3C). Similarly, GA did not affect the adherence of calcein-labeled THP-1 cells to activated EAhy-926 endothelial cells (Figure 3D). This implies that intercellular/vascular cell adhesion molecule function is undisturbed in the presence of GA.
Injecting fluorescently labeled resting platelets into C57/Bl6 mice did not trigger platelet rolling or platelet adhesion on the femoral vein endothelium (not shown). When circulating platelets were activated by the intravenous administration of collagen, with the concomitant injection of anti-IIb/;3 antagonist G4120 to avoid platelet aggregation, conjugates formed in situ between activated platelets and leukocytes. These rosettes were found to tumble along the vessel wall, as illustrated in Figure 4, which shows in each box an individual rosette representing 2 to 3 labeled platelets attached to a single unlabeled, and thus invisible, leukocyte (individual spots in Figure 4). Each box consists of an overlay of a series of consecutive pictures taken at 0.1-second intervals. Calculation of the number of rolling conjugates revealed an average of 46±7 rosettes rolling across the field (0.067 mm2) per minute (n=30 fields).
Pretreatment of mice with GA before platelet activation dose dependently reduced conjugate tumbling with an ED50 of 1.5±0.4 mg · kg–1 · h–1 (representing the dose of a combined injection/infusion protocol as described in Methods) (Figure 5A). Conjugate-vessel wall interactions were almost abrogated at an infusion dose of 2.25 mg GA · kg–1 · h–1 (to 6±2 conjugates per field per minute; n=16 fields). Tethering of individual platelets was marginal (Figure 6 and Data Supplement Movies).
Injection of labeled platelets into aged atherosclerotic ApoE–/– mice and subsequent in situ activation with collagen revealed increased rosette-endothelium interactions, as expected for endothelium activated by hyperlipidemia, ie, overexpressing von Willebrand factor,24,29 and upregulating other adhesion molecules such as intercellular adhesion molecule-1, vascular cell adhesion molecule-1, or E-selectin.30 In addition to increased rosette tumbling, the femoral venous endothelium of ApoE–/– mice uncovered increased tethering of single individual platelets (Figure 6), in agreement with in vitro and ex vivo perfusion studies over endothelial cells and rabbit aorta endothelium.24 In analogy to the studies in normolipidemic mice, in atherosclerotic ApoE–/– mice injected with collagen to activate platelets and preform conjugates, the infusion of 3.75 mg GA · kg–1 · h–1 also led to a sharp decrease in the number of preformed tumbling conjugates down to 34±13% of control value (P<0.01) without affecting single platelet tethering (Figure 5B).
Discussion
Polyphenols, (red) wine-derived antioxidants, have received considerable attention during the last decade because they are likely to reduce the risk of CAD (the French paradox).1 The atheroprotective effect of polyphenols has been tentatively attributed to their capacity to inhibit LDL oxidation6,7,31 and to inhibit the expression of proinflammatory proteins, including PDGFR;.10 However, because large-scale trials have been largely inconclusive in showing an effect of other antioxidants, additional nonoxidant pathways may be at least equally prominent. The notion that red wine polyphenols act as a cardioprotective by inhibiting PDGFR; activity is contradicted by recent studies showing that the PDGFR; axis is instrumental in maintaining plaque stability and in preventing aneurysm formation by promoting vascular smooth muscle cell migration to and their proliferation in the cap.11,32 The present report shows that GA, one of the most abundant polyphenols in red wine, can potently intervene in P-selectin function, an adhesion molecule that is intrincally involved in the initial rolling of monocytes along the inflamed endothelium33 and in thrombosis.19,34
Previously, we reported that derivatization of specific human P-selectin-binding peptides with GA led to a dramatically enhanced affinity of the peptide toward P-selectin.17 This led to the presumption that GA by itself may specifically interact with P-selectin. Indeed, the polyphenol was found to inhibit the binding of 2 different selectin ligands, TM11-PO22 and biotin-PAA-Lea-SO3H,23 to P-selectin at micromolar affinity. Conceivably, the TM11-based assay is most reflective of the functional activity of GA, which is corroborated by the finding that static and dynamic interactions between PSGL-1 displaying HL60 cells and CHO-P monolayers were inhibited by GA. The interaction appears to be rather specific for P-selectin, although GA also displayed some affinity for L-selectin, which is in concordance with previous studies showing that the ligand binding profiles of P- and L-selectin largely overlap.35
In vitro, GA does not disturb platelet-endothelium or leukocyte-endothelium interaction. However, GA induced a sharp rise in the rolling speed of leukocytes retained on the platelet surface, reflecting impaired adhesion. In vivo studies of platelet-assisted inflammation indicate that GA can interrupt platelet-assisted rolling and tumbling of platelet-leukocyte rosettes over the endothelium of normocholesterolemic and hypercholesterolemic mice without affecting the deposition of single platelets on activated endothelium. This effect could be observed when GA was administered before and even after platelets had been activated.
GA effects in vitro and in vivo are not attributable to its antioxidant activity. n-Dodecyl gallate displays an antioxidant activity similar to that of GA, although it is unable to inhibit P-selectin binding at low concentrations. Similarly, other equally potent and structurally related antioxidants such as caffeic acid and dihydroxy benzoic acid are completely ineffective in inhibiting P-selectin. Finally, precautions were taken to prevent oxidation of the gallate during the experiments (which can easily be appreciated because oxidized gallate is dark green/brown).
GA is abundantly available not only in wine (grapes) but also in (green) tea. The total phenolic content in various red wines was analyzed as 1100 to 3165 mg/L, of which GA attributes 35 to 70 mg/L and the epicatechine gallate derivatives attribute 120 to 360 mg/L.36 The GA content of green tea is even higher (2.2 mg/mL, corresponding to 149 mg/g dry leaves).37 A study by Lee et al38 showed that cocoa contained higher total gallate levels than green tea or red wine (611 versus 165 and 340 mg per serving, respectively). Oral administration of only 50 mg GA to a healthy volunteer resulted in plasma levels of GA and its major metabolite 4-O-methylgallic acid in the low micromolar range,39 implying that the observed EC50 values of 15 to 40 μmol/L can be readily reached after consumption of only 100 to 500 mg GA.
In addition to being involved in atherosclerosis and thrombosis, P-selectin was reported to play a role in metastasis by promoting tumor growth and to facilitate the metastatic seeding of mucin-producing carcinomas.40 P-selectin deficiency was shown to be accompanied by reduced implanted carcinoma cells growth and reduced metastasis formation.40 In addition, P-selectin is expressed by several metastatic pancreatic tumor cells.41 Thus, it is interesting that Ohno et al42 recently demonstrated that GA administration (50 mg/kg) could inhibit metastasis formation and growth.
In conclusion, GA inhibits P-selectin-mediated inflammation both in vitro and in vivo. This study provides for the first time a molecular mechanism through which GA can exert a beneficial effect in several pathobiologies such as CAD, thrombosis, and cancer by binding and antagonizing P-selectin under static and dynamic conditions already at concentrations readily achieved after moderate wine, (green) tea, or cocoa consumption.
Acknowledgments
This study was partly funded by grants from Yamanouchi Japan (C.C.M.A, B.C.H.L), the Netherlands Organization of Scientific Research (NWO grant 016.026.019 to E.A.L.B), and the Dutch Heart Foundation (NHS, 2003T201). The support of the FWO Vlaanderen (project G.0376.01) is recognized. During this study, A.B. was the holder of a Marie Curie fellowship. K.D. is the recipient of a junior fellowship from the University Hospital Leuven. The skillful assistance of Petra Vandervoort during the cell culture and coverslip preparation and of Hans de Bont during Ca flux measurements is highly appreciated.
Footnotes
Drs Appeldoorn and Bonnefoy contributed equally to this work.
The online-only Data Supplement, which contains 4 movies that supplement the Results, can be found with this article at http://www.circulationaha.org.
References
Renaud S, de Lorgeril M. Wine, alcohol, platelets, and the French paradox for coronary heart disease. Lancet. 1992; 339: 1523–1526.
Pearson TA. Alcohol and heart disease. Circulation. 1996; 94: 3023–3025.
Gronbaek M. Alcohol, type of alcohol, and all-cause and coronary heart disease mortality. Ann N Y Acad Sci. 2002; 957: 16–20.
Vogel RA. Alcohol, heart disease, and mortality: a review. Rev Cardiovasc Med. 2002; 3: 7–13.
Aviram M, Hayek T, Fuhrman B. Red wine consumption inhibits LDL oxidation and aggregation in humans and in atherosclerotic mice. Biofactors. 1997; 6: 415–419.
Frankel EN, Kanner J, German JB, Parks E, Kinsella JE. Inhibition of oxidation of human low-density lipoprotein by phenolic substances in red wine. Lancet. 1993; 341: 454–457.
Kondo K, Matsumoto A, Kurata H, Tanahashi H, Koda H, Amachi T, Itakura H. Inhibition of oxidation of low-density lipoprotein with red wine. Lancet. 1994; 344: 1152–1159.
Clarke R, Armitage J. Antioxidant vitamins and risk of cardiovascular disease: review of large-scale randomised trials. Cardiovasc Drugs Ther. 2002; 16: 411–415.
Kaikkonen J, Porkkala-Sarataho E, Morrow JD, Roberts LJ 2nd, Nyyssonen K, Salonen R, Tuomainen TP, Ristonmaa U, Poulsen HE, Salonen JT. Supplementation with vitamin E but not with vitamin C lowers lipid peroxidation in vivo in mildly hypercholesterolemic men. Free Radical Res. 2001; 35: 967–978.
Rosenkranz S, Knirel D, Dietrich H, Flesch M, Erdmann E, Bohm M. Inhibition of the PDGF receptor by red wine flavonoids provides a molecular explanation for the "French paradox." FASEB J. 2002; 16: 1958–1960.
Sano H, Sudo T, Yokode M, Murayama T, Kataoka H, Takakura N, Nishikawa S, Nishikawa SI, Kita T. Functional blockade of platelet-derived growth factor receptor-; but not of receptor- prevents vascular smooth muscle cell accumulation in fibrous cap lesions in apolipoprotein E–deficient mice. Circulation. 2001; 103: 2955–2960.
Rotondo S, Rajtar G, Manarini S, Celardo A, Rotilio D, de Gaetano G, Evangelista V, Cerletti C. Effect of trans-resveratrol, a natural polyphenolic compound, on human polymorphonuclear leukocyte function. Br J Pharmacol. 1998; 123: 1691–1699.
Wallerath T, Deckert G, Ternes T, Anderson H, Li H, Witte K, F;rstermann U. Resveratrol, a polyphenolic phytoalexin present in red wine, enhances expression and activity of endothelial nitric oxide synthase. Circulation. 2002; 106: 1652–1658.
Pendurthi UR, Williams JT, Mohan Rao LV. Resveratrol, a polyphenolic compound found in wine, inhibits tissue factor expression in vascular cells: a possible mechanism for the cardiovascular benefits associated with moderate consumption of wine. Arterioscler Thromb Vasc Biol. 1999; 19: 419–426.
Landrault N, Poucheret P, Ravel P, Gasc F, Cros G, Teissedre PL. Antioxidant capacities and phenolic levels of French wines from different varieties and vintages. J Agric Food Chem. 2001; 49: 3341–3348.
Ohno Y, Fukuda K, Takemura G, Toyota M, Watanabe M, Yasuda N, Xinbin Q, Maruyama R, Akao S, Gotou K, Fujiwara T, Fujiwara H. Induction of apoptosis by gallic acid in lung cancer cells. Anticancer Drugs. 1999; 10: 845–851.
Appeldoorn CCM, Molenaar TJM, Bonnefoy A, van Leeuwen SH, Vandervoort PAH, Hoylaerts MF, van Berkel TJC, Biessen EAL. Rational optimization of a short human P-selectin binding peptide leads to nanomolar affinity antagonists. J Biol Chem. 2003; 278: 10201–10207.
Vestweber D, Blanks JE. Mechanisms that regulate the function of the selectins and their ligands. Physiol Rev. 1999; 79: 181–213.
Merten M, Chow T, Hellums JD, Thiagarajan P. A new role for P-selectin in shear-induced platelet aggregation. Circulation. 2000; 102: 2045–2050.
Johnson RC, Chapman SM, Dong ZM, Ordovas JM, Mayadas TN, Herz J, Hynes RO, Schaefer EJ, Wagner DD. Absence of P-selectin delays fatty streak formation in mice. J Clin Invest. 1997; 99: 1037–1043.
Theilmeier G, Lenaerts T, Remacle C, Collen D, Vermylen J, Hoylaerts MF. Circulating activated platelets assist THP-1 monocytoid/endothelial cell interaction under shear stress. Blood. 1999; 94: 2725–2734.
Molenaar TJM, Appeldoorn CCM, de Haas SAM, Michon IN, Bonnefoy A, Hoylaerts MF, Pannekoek H, van Berkel TJC, Kuiper J, Biessen EAL. Specific inhibition of P-selectin mediated cell adhesion by phage display derived peptide antagonists. Blood. 2002; 100: 3570–3577.
Weitz Schmidt G, Stokmaier D, Scheel G, Nifant’ev NE, Tuzikov AB, Bovin NV. An E-selectin binding assay based on a polyacrylamide-type glycoconjugate. Anal Biochem. 1996; 238: 184–190.
Theilmeier G, Michiels C, Spaepen E, Vreys I, Collen D, Vermylen J, Hoylaerts MF. Endothelial von Willebrand factor recruits platelets to atherosclerosis-prone sites in response to hypercholesterolemia. Blood. 2002; 99: 4486–4493.
Oury C, Kuijpers MJ, Toth-Zsamboki E, Bonnefoy A, Danloy S, Vreys I, Feijge MA, De Vos R, Vermylen J, Heemskerk JW, Hoylaerts MF. Overexpression of the platelet P2X1 ion channel in transgenic mice generates a novel prothrombotic phenotype. Blood. 2003; 101: 3969–3976.
Giles AR. Guidelines for the use of animals in biomedical research. Thromb Haemost. 1987; 58: 1078–1084.
Sako D, Xiao Jia C, Barone KM, Vachino G, White HM, Shaw G, Veldman GM, Bean KM, Ahern TJ, Furie B, Cumming DA, Larsen GR. Expression cloning of a functional glycoprotein ligand for P-selectin. Cell. 1993; 75: 1179–1186.
Wilkins PP, McEver RP, Cummings RD. Structures of the O-glycans on P-selectin glycoprotein ligand-1 from HL-60 cells. J Biol Chem. 1996; 271: 18732–18742.
De Meyer GR, Hoylaerts MF, Kockx MM, Yamamoto H, Herman AG, Bult H. Intimal deposition of functional von Willebrand factor in atherogenesis. Arterioscler Thromb Vasc Biol. 1999; 19: 2524–2534.
Molenaar TJ, Twisk J, de Haas SA, Peterse N, Vogelaar BJ, van Leeuwen SH, Michon IN, van Berkel TJ, Kuiper J, Biessen EA. P-selectin as a candidate target in atherosclerosis. Biochem Pharmacol. 2003; 66: 859–866.
Kerry NL, Abbey M. Red wine and fractionated phenolic compounds prepared from red wine inhibit low density lipoprotein oxidation in vitro. Atherosclerosis. 1997; 135: 93–102.
Boucher P, Gotthardt M, Li WP, Anderson RG, Herz J. LRP: role in vascular wall integrity and protection from atherosclerosis. Science. 2003; 300: 329–332.
Lasky LA. Selectin-carbohydrate interactions and the initiation of the inflammatory response. Annu Rev Biochem. 1995; 64: 113–139.
Freedman JE, Loscalzo J. Platelet-monocyte aggregates: bridging thrombosis and inflammation. Circulation. 2002; 105: 2130–2132.
Nelson RM, Cecconi O, Roberts WG, Aruffo A, Linhardt RJ, Bevilacqua MP. Heparin oligosaccharides bind L- and P-selectin and inhibit acute inflammation. Blood. 1993; 82: 3253–3258.
Burns J, Gardner PT, O’Neil J, Crawford S, Morecroft I, McPhail DB, Lister C, Matthews D, MacLean MR, Lean MEJ, Duthie GG, Crozier A. Relationship among antioxidant activity, vasodilation capacity, and phenolic content of red wines. J Agric Food Chem. 2000; 48: 220–230.
Chandra S, De Mejia Gonzalez E. Polyphenolic compounds, antioxidant capacity, and quinone reductase activity of an aqueous extract of Ardisia compressa in comparison to mate (Ilex paraguariensis) and green (Camellia sinensis) teas. J Agric Food Chem. 2004; 52: 3583–3589.
Lee KW, Kim YJ, Lee HJ, Lee CY. Cocoa has more phenolic phytochemicals and a higher antioxidant capacity than teas and red wine. J Agric Food Chem. 2003; 51: 7292–7295.
Shahrzad S, Bitsch I. Determination of gallic acid and its metabolites in human plasma and urine by high-performance liquid chromatography. J Chromatogr B Biomed Sci Appl. 1998; 705: 87–95.
Kim YJ, Borsig L, Varki NM, Varki A. P-selectin deficiency attenuates tumor growth and metastasis. Proc Natl Acad Sci U S A. 1998; 95: 9325–9330.
Iwamura T, Caffrey TC, Kitamura N, Yamanari H, Setoguchi T, Hollingsworth MA. P-selectin expression in a metastatic pancreatic tumor cell line (SUIT-2). Cancer Res. 1997; 57: 1206–1212.
Ohno T, Inoue M, Ogihara Y. Cytotoxic activity of gallic acid against liver metastasis of mastocytoma cells P-815. Anticancer Res. 2001; 21: 3875–3880.(Chantal C.M. Appeldoorn, )