Resveratrol: One molecule, many targets

Abstract

Resveratrol is one of the numerous polyphenolic compounds found in several vegetal sources. In recent years, the interest in this molecule has increased exponentially following the major findings that resveratrol (i) is shown to be chemopreventive in some cancer models, (ii) is cardioprotective, and (iii) has positive effects on several aspects of metabolism, leading to increased lifespan in all the metazoan models tested thus far, including small mammals. Such remarkable properties have elicited a vast interest towards the identification of target proteins of resveratrol and have led to the identification of enzymes inhibited by resveratrol and others whose activation is enhanced. In the vast majority of cases, resveratrol displays inhibitory/activatory effects in the micromolar range, which is potentially attainable pharmacologically, although targets with affinities in the nanomolar range have also been reported. Here, we provide an overview of the various classes of enzymes known to be inhibited (or activated) by resveratrol. It appears that resveratrol, as a pharmacological agent, has a wide spectrum of targets. The biological activities of resveratrol may thus be dependent on its simultaneous activity on multiple molecular targets. © 2008 IUBMB IUBMB Life, 60(5): 323–332, 2008

INTRODUCTION

As a chemical entity, resveratrol (3,4′,5-trihydroxy-trans-stilbene) is known since the 40s when it was first isolated from the roots of white hellebore and later from Polygonum cupsidatum, a medicinal plant1, 2. Resveratrol is also present in appreciable amount in a variety of edible fruits including nuts, berries, and grape skin (and thereof derived red wine). From a botanical point of view, resveratrol acts as a phytoalexin, that is a toxic compound produced by a plant in response to a parasitic attack or under conditions of stress.

AMPK, AMP-activated protein kinase; COX, cyclooxygenase; IC₅₀, inhibitory concentration 50; IKK, IκB alpha kinase; LOX, lipooxygenase; MAPK, mitogen-activated protein kinase; NF-κB, nuclear factor κB; PGC-1α, peroxisome proliferator-activated receptor-γ coactivator 1α; PI3K, phosphoinositide 3-kinase; PKB, protein kinase B; PKC, protein kinase C; PKD, protein kinase D; RTK, receptor tyrosine kinase.

Starting from the 80s, a substantial amount of research was directed towards the identification of plant-derived compounds endowed with pharmacological properties. Extraction of bioactive compounds from Polygonum cupsidatum and other Poligonaceae yielded resveratrol (and other related stilbenes) as an inhibitor of enzymes involved in arachidonate metabolism in leukocytes3 and of partially purified kinases4. Although these two studies did not evaluate the therapeutic potential of those inhibitory effects in an in vivo model, they set the basis for further investigation on resveratrol.

The search for novel cancer chemopreventive agents led to the isolation of resveratrol from yet another plant species (Cassia quinquangulata) based on the inhibition of the cyclooxygenase COX-1, as identified by bioassay-guided fractionation5. The same study demonstrated that topic application of resveratrol decreased pedal oedema in a rat model of carrageenan-induced paw inflammation and decreased tumorigenesis on a mouse skin cancer model5, thus raising the interest in resveratrol as an anticancer compound. A further unbiased screening for modulators of the activity of sirtuin proteins, a class of histone deacetylases involved in lifespan determination, again identified resveratrol, this time as a sirtuin activator6. As sirtuin activator, resveratrol administration has been shown to increase yeast lifespan by 70%6, as well as to prolong lifespan in C. elegans, D. melanogaster7, lower vertebrates8 and mice9.

Speculations that uptake of resveratrol by red wine consumption could be behind the so-called French paradox—whereby the French population, in spite of a rather fatty diet, has a lower incidence of cardiovascular disease10—is supported by the fact that resveratrol administration has cardioprotective effects11, 12.

In spite of the obvious beneficial effects on lifespan and health, the molecular target(s) through which resveratrol acts have not yet been univocally determined. As a matter of fact, resveratrol has been shown to inhibit a plethora of enzymes belonging to different classes, including (but not limited to) kinases, lipo- and cyclooxygenases, sirtuins and other proteins. From a mechanistic point of view, both the m-hydroxyquinone and 4-hydroxystyryl moieties (Fig. 1) of the molecule have been shown to be important for the determination of resveratrol's inhibitory properties towards various enzymes (Table 1). We will summarise here the current evidence about the various resveratrol targets, which provide a framework of possible pathways mediating some of the end-point responses elicited by resveratrol treatment. Finally, we will discuss the idea that the effects of resveratrol may reflect its simultaneous action on several targets.

Figure 1. (A) Chemical structure of resveratrol, 3,4′,5-trihydroxy-trans-stilbene. The two moieties are shown, as well as the numbering of the hydroxyl groups. (B) From left: chemical structures of stilbene and the structurally similar flavone; the common phenyl rings are lettered. Chemical structures of the flavone-derivatives quercetin and fisetin, which share several common targets with resveratrol.

Table 1. Summary of the enzymes modulated by resveratrol

Inhibited enzymes	Year of first report	In vitro IC50 (unless otherwise stated)	Associated biological effects	References
* IC50 determined as inhibition of endothelin-1—induced MAPK activation in porcine coronary arteries-derived endothelium. † Concentration required to effectively decrease endothelin-1—induced p38 and JNK-1 phosphorylation in porcine coronary arteries-derived endothelium. ‡ Concentration of resveratrol topically applied to mouse skin. § Concentration required to effectively decrease Src and STAT3 tyrosine phosphorylation in v-Src-transformed mouse fibroblasts.
Cyclo- and lipooxygenase	1985	15 and 3.7 μM	Decreased inflammation and tumorigenesis in mice	3, 5
PKCs and p56^lck	1993	40 and 60 μM	Growth inhibition and induction of apoptosis in cancer cell lines	4, 13, 14
ERK1	1999	37 μM*	Reversal of endothelin-1 stimulation	15
JNK1	1999	50 μM†	Reversal of endothelin-1 stimulation	15
p38	1999	50 μM†	Reversal of endothelin-1 stimulation	15
IKKβ	2006	1 μM‡	Inhibition of phorbol ester-induced expression of COX2	16
Src	2006	20 μM§	Inhibition of viability of several human tumor cell lines	17
STAT3	2006	20 μM§	Inhibition of viability of several human tumor cell lines	17
Ribonucleotide Reductase	1998	50 μM	Inhibition of DNA synthesis in leukemia cell lines	18
DNA polymerases α and δ	1998	3.3 and 5 μM	Not investigated	19, 20
PKD	2000	35–50 μM	Cancer chemopreventive action	21, 22
PKC alpha	2003	<10 μM	Not investigated	23
Quinone reductase 2	2004	35–50 nM	Resistance to menadione-induced cell death	24
Aromatase	2007	25 μM	Not investigated	25
Activated enzymes	Year of first report	Resveratrol concentration tested	Associated biological effects	References
Sir2/SIRT1	2003	100 μM	Lifespan prolongation	6, 7
Adenylyl cyclase	2003	0.8 μM	Inhibition of MCF7 cells proliferation	26
AMPK	2006	50 μM	Improvement of metabolism	9, 27

PHYSIOLOGICAL EFFECTS OF RESVERATROL

C. elegans

Wild-type adult worms showed an increase of up to 14% in lifespan upon resveratrol treatment, while sir-2.1 null mutants did not exhibit a significant lifespan extension relative to untreated worms7. As previously mentioned, these observations have not been fully reproduced in an independent investigation28, therefore further work evaluating C. elegans lifespan upon resveratrol administration is warranted.

Resveratrol rescued neuronal dysfunction phenotypes induced by mutant polyglutamines in transgenic C. elegans. This activity of resveratrol was lost in Sir2.1 and forkhead transcription factor daf-16 mutants, suggesting that the mode of action for the rescue of mutant polyglutamines toxicity was through sir-2.1 and daf-1629.

D. melanogaster

In the wild-type strain Canton-S, lifespan was extended up to 23% with fisetin (Fig. 1B) and up to 29% with resveratrol. A calorie-restricted diet increased fly lifespan by 40% in females and by 14% in males and, under these conditions, neither fisetin nor resveratrol further increased longevity, indicating that caloric restriction and resveratrol administration regulate a common genetic program controlling ageing7. Resveratrol failed to extend lifespan in flies completely lacking functional Sir2 or in flies in which Sir2 is severely decreased7. An independent investigation confirming the lifespan-promoting properties of resveratrol appeared30 but a contradictory study disputes this notion28.

Rodents

In mice, continued resveratrol administration (22.4 ± 0.4 mg/kg/day for 110 weeks) shifted the physiology of middle-aged mice on high-calorie diet towards that of mice on a standard diet and significantly increased their survival9. Several physiological parameters were positively altered by resveratrol administration, including inhibition of high caloric diet-induced body weight and liver weight, inhibition of pancreatic damage as well as an improved morphology of the heart. In this experimental setting, the above-mentioned changes were associated with longer lifespan, increased insulin sensitivity, reduced insulin-like growth factor-1 (IGF-1) levels, and increases in AMPK and peroxisome proliferator-activated receptor-γ coactivator 1α (PGC-1α) activities, increased mitochondrial number, and improved motor function9. Baur et al., also reported having supplemented a second group of mice with 5.2 ± 0.1 mg/kg/day resveratrol. The effects observed at this dose were reportedly less prominent and have to date not yet been published. A concurrent study evaluated the impact of a shorter supplementation with resveratrol (200 or 400 mg/kg/day for 15 weeks) and similarly demonstrated substantial improvements in metabolic homeostasis, especially in high-fat-diet fed mice, including increased mitochondrial biogenesis and expression of genes involved in oxidative phosphorylation, increased aerobic capacity, decreased fasting and nonfasting plasma glucose and increased insulin sensitivity31. Being such improvements correlated with decreased PGC-1α acetylation (rendering PGC-1α active) a model whereby resveratrol—by activating SIRT1—induces PGC-1α deacetylation, has been proposed and demonstrated biochemically in cultured muscle cells and fibroblasts31. Additionally, resveratrol has been shown to improve insulin sensitivity by repressing transcriptionally the protein tyrosine phosphatase 1B, which acts to reverse RTK (including the insulin receptor) action32.

Humans

At present, a large body of evidence from in vitro and animal studies indicates that resveratrol may be beneficial to many aspects of human health. To translate such promising findings into clinical reality, studies on the potential toxicity, pharmacokinetics and bioavailability in humans, followed by interventional clinical trials are clearly needed. Although several studies have attempted to define the pharmacokinetics of resveratrol in rodent models (reviewed in ref.1 with rather contradictory results, only a few reports in humans have appeared. A detailed study on the metabolism of resveratrol in humans demonstrated that absorption after oral administration is very high but resveratrol is detected only in trace amounts (<5 ng/ml) in plasma, with the bulk of resveratrol being very quickly metabolized into sulphate and glucoronic acid resveratrol-derivatives, or hydrogenation of the aliphatic double bond33. As occurrence of such derivatives is also prominent in rodents1 it can be suggested that either (i) plasmatic resveratrol-derivatives are reconverted into resveratrol in target organs or (ii) resveratrol-derivatives per se are the active molecules, and this certainly deserves further investigation. In keep with this second idea, a phase I dose escalation pharmacokinetic study indicated that resveratrol may be administered in a single dose up to 5 g, resulting in a peak level of 2.4 μM 1.5-h postabsorption. Given that cancer chemopreventive effects of resveratrol in in vitro studies require concentrations of at least 5 μM, further investigation on resveratrol-derivatives, which are present in the plasma at higher levels, is warranted34. Whatever the active molecule (resveratrol or a derivative), phase I and II clinical trials, testing resveratrol administration to patients with colon cancer, have been started (http://clinicaltrials.gov/show/NCT00256334).

RESVERATROL TARGETS

Cyclooxygenases and Lipooxygenases

Cancer and inflammation are critically linked, and among the enzymes that synthesize proinflammatory mediators from arachidonic acid are the cyclo- and lipooxygenases35. Pioneering studies on arachidonate metabolism in rat peritoneal leukocytes demonstrated that resveratrol inhibited both 5-lipooxygenase (LOX, IC₅₀ of 2.7 μM) and cyclooxygenase (COX, IC₅₀ of <1 μM) activities, resulting in a reduced accumulation of the lipooxygenase product 5-HETE (5S-hydroxy-6,8,11,14-eicosatetraenoic acid) and COX products HHT (heptadecatrienoic acid) and thromboxane B23. Comparable inhibitory potencies have been found in in vitro assays using purified 5-lipooxygenase and COX36. Further investigation showed that both the cyclooxygenase and hydroperoxidase activities of COX-1 are inhibited by resveratrol (IC₅₀ of 15 and 3.7 μM, respectively), while the isoform COX-2 is scantly inhibited5. Inhibition of COX-1 cyclooxygenase activity decreased the establishment of preneoplastic lesions in a mouse mammary gland culture. Also, both tumour number and incidence in a mouse skin model of tumorigenesis were effectively decreased by topic application of 25 μM resveratrol twice a week5. The inactivation of COX-1 by resveratrol occurs through a “hit-and-run” mechanism. Resveratrol firstly binds to the peroxidase active site. Therein, the irreversible inactivation of the enzyme occurs concomitantly to the oxidation of one hydroxyl group on the resveratrol m-hydroquinone moiety, which is thus a necessary structural requirement for COX inhibition to occur37. This observation provided a structural basis for the design of more specific COX-1 inhibitors37. Very recently, stilbene derivatives containing the m-hydroquinone moiety and exhibiting submicromolar inhibitory potency towards COX-1 have been described and suggested to be potentially useful lead compounds for in vivo studies38. A rather general consensus on the specific inhibitory action of resveratrol towards COX-1 catalytic activity has been reached. Also, an effect on COX-2 transcriptional regulation by resveratrol has been revealed, an inhibition that can likewise affect prostaglandin levels by regulation of COX-2 protein expression. Resveratrol suppresses the transcriptional activation of the COX-2 gene induced by several protein kinases, including PKCα and the MAPK Erk1, thus implying the direct inhibition of these protein kinases as a chief mechanism directing the inhibition of COX-2 transcription39.

The synthesis of proinflammatory molecules by COX and LOX is an important step for initiation of tumorigenesis. Thus, the inhibitory effects of resveratrol on COX-1/LOX catalytic activities and COX-2 transcription, with consequent decrease of COX-2 protein levels, are highly relevant to cancer chemoprevention40.

Kinases

A second family of enzymes whose activity is modulated by resveratrol comprises the broad family of protein and lipid kinases. As early as 1993, bioassay-directed fractionation of Polygonum cupsidatum showed that resveratrol is capable to inhibit a partially purified tyrosine kinase (p56^lck) and serine/threonine kinase (PKC, with both α and β isoforms present in a mixture)4. The inhibitory potency was in the micromolar range—as observed for the COX/LOX enzymes—with an IC₅₀ of 60 μM and 40 μM towards p56^lck and PKC, respectively.

PKCs and PKD

Further work on purified PKC isozymes (and the distantly related PKD) demonstrated that the common PKCα, the novel PKCε and the atypical PKCζ are inhibited in vitro with an IC₅₀ < 100 μM. Thus, at least one isozyme from each of the PKC subfamilies displays sensitivity to resveratrol41. Inhibitory action towards PKCs is competed for by ATP, suggesting a catalytic domain-directed mechanism41. In addition, resveratrol has been reported to be a competitive inhibitor towards the C1A and C1B phorbol ester-binding domains of PKCα, as demonstrated by its interaction (maximal at 2.8 μM) with a fusion peptide containing both domains from PCKα23. Activation of PKC isozymes requires autophosphorylation in the catalytic activation loop. Such event was not significantly affected by resveratrol up to 100 μM for any of the PKCs tested (α, β1,γ,δ,ε,ζ) in an in vitro assay. On the contrary, PKD autophosphorylation was inhibited by resveratrol with an IC₅₀ = 52 μM, which correlated with an IC₅₀ = 36 μM towards PKD catalytic activity as tested in vitro21. However, studies on COS-7 cells overexpressing PKD showed that resveratrol inhibits PKD activatory phosphorylation with an IC₅₀ of 800 μM, suggesting that the pharmacological action of resveratrol on PKD in vivo is negligible22. Inhibition of PKC isoforms by resveratrol is related to growth inhibition and induction of apoptosis in various cancer cell models, including gastric cancer cells13 and prostate cancer cells14.

Tyrosine Kinases

The inhibition by resveratrol of the activities of receptor tyrosine kinases and Src kinase may be relevant to its antitumor activity. Src transformed mouse fibroblasts treated with 20 μM resveratrol displayed decreased Src tyrosine phosphorylation and catalytic activity and decreased tyrosine phosphorylation of the Src target molecule STAT-317. Similarly, STAT-3 (but not STAT-1) tyrosine phosphorylation and consequential DNA-binding activity was inhibited by resveratrol in the 30–50 μM range, leading to apoptotic death, in human cancer cell lines derived from breast, pancreatic and prostate carcinomas17. The concept that resveratrol may possess cancer chemopreventive activity by targeting tyrosine kinase cascades has been further expanded by a study on HER-2/neu transgenic mice, which are predisposed to the development of mammary tumours42. Administration of 4 μg/day resveratrol to HER-2/neu transgenic mice increased the incidence of tumour-free mice and reduced tumour number/diameter. Such effects were reported to depend on a resveratrol-induced decrease of HER-2/neu mRNA expression and subsequent induction of apoptosis43. Further studies are thus warranted to determine whether the HER-2/neu protein expression level and tyrosine kinase activity is likewise affected by resveratrol.

MAPK Family

The MAPK family includes the extracellular signal regulated kinases (Erk1/2) and the stress activated kinases JNK1/2 and p38. Activation of stress kinases is implicated in cardiovascular disorders44. In coronary artery smooth muscle, resveratrol (30–50 μM) inhibited endothelin-1 induced Erk1/2 enzymatic activity and Erk1/2, JNK and p38 activatory phosphorylations15, thus suggesting a partial mechanism for the beneficial cardiovascular effects of resveratrol. However, other reports have shown that cardioprotection mediated by resveratrol occurs through activation of the MAPK signalling45. Furthermore, in muscle-derived cell lines, occurrence of Erk1/2 inhibition by resveratrol—after insulin stimulation—is disputed, with reports indicating the occurrence of inhibition46, and others showing no major effects47. Such discrepancies may be because of the different cell types used to investigate the effects of resveratrol and further investigation are necessary to finally define the effects of resveratrol on the MAPK family.

Lipid Kinases and Akt/PKB Signalling

Since resveratrol acts as a cancer chemopreventive molecule it could be expected that it might have an impact on the PI3K/PKB signalling pathway, which is virtually always activated in tumours48. Indeed, recent evidence indicates that apoptosis in prostate cancer, uterine cancer, and lymphoblastic leukaemia cell lines is—at least in part—dependent on a negative modulation of PKB49–51. Being a broad-spectrum kinase inhibitor, resveratrol might decrease PKB activity by direct inhibition. Yet, to the best of our knowledge, such observation has not yet been reported. Alternative mechanisms have been proposed to explain the resveratrol-dependent decrease in PKB activation following agonist stimulation. In hepatocytes, resveratrol inhibited insulin-induced PKB (and also Erk1/2) activation through disruption of the interactions between activated insulin receptor substrates and its downstream binding proteins, including the p85 regulatory subunit of PI3K and Grb246. In human myotubes and muscle derived cell lines we reported that resveratrol inhibits PKB activatory phosphorylations with an IC₅₀ < 10 μM. Such inhibition was not direct, instead, resveratrol was found to be a competitive inhibitor of class IA (i.e., RTK-coupled) PI3K p110α and p110β with IC₅₀s in the 20–50 μM range. As reported for PKCs, resveratrol acts as a competitive inhibitor that targets the ATP binding site47. Interestingly class IB PI3K p110γ was inhibited with a much lower potency (IC₅₀ > 100 μM), indicating resveratrol as a PI3K subclass-specific inhibitor. Further studies on the inhibitory potential of resveratrol on cancer cells might eventually define PI3K as a target of resveratrol in cancer.

IκB Alpha Kinase

Activation of the NF-κB signal transduction pathway is one of the major events linking inflammation to tumorigenesis52. Resveratrol, by suppressing IKK-mediated phosphorylation of IκB, inhibited NF-κB, thus preventing TPA (12-O-tetradecanoylphorbol-13-acetate)-induced COX-2 overexpression53. Detailed investigation on the underlying mechanism identified IKK as the molecular target of resveratrol16. Such inhibition, obtained by topic application of 1 μM resveratrol, effectively inhibited TPA-induced COX-2 expression in mouse skin. Since low doses of resveratrol were required for effective inhibition, IKK is among the targets the most potently inhibited by resveratrol in an in vivo setting16.

AMPK

While most of the tested kinases are inhibited by resveratrol at concentrations ranging from 1 μM to 1 mM, AMPK stands as a unique protein kinase that is activated by resveratrol treatment. First evidence came from a study on HepG2 hepatocytes, in which resveratrol (in the 10–50 μM range) increased AMPK phosphorylation and activity and protected cells from glucotoxicity27. The improvement of insulin sensitivity in mice fed on a long-term treatment with resveratrol was associated to AMPK activatory phosphorylation9. Furthermore, in a recent report resveratrol was shown to stimulate glucose transport in C₂C₁₂ myotubes via AMPK activation54, acting in this respect similarly to the antidiabetic drug metformin55.

Whether resveratrol directly activates AMPK or exerts its positive action through modulation of metabolism, ultimately impinging on AMPK regulation remains to be determined.

Sirtuins

Perhaps the most striking biological effect induced by resveratrol is the prolongation of lifespan in yeast and the full spectrum of metazoans, including mice on a high-calorie diet6, 7, 9.

The search for yeast genes involved in the regulation of lifespan has led to the identification of ≈50 gene mutations involved in the regulation of yeast replicative lifespan (i.e., the maximal number of asymmetric divisions that can be obtained starting from a mother cell)56. Mostly, enhanced longevity is associated to gene deletion. In contrast, overexpression of the SIR2 gene increases yeast life span. The SIR2 gene product, Sir2, belongs to the silent information regulator family of proteins, also termed sirtuins, which comprises seven members in mammals (SIRT1 to 7)57. Sirtuins are NAD⁺-dependent deacetylases that contribute to the maintenance of genomic stability, transcriptional silencing and DNA repair. Also, they orchestrate lipid mobilisation from the adipocyte and gluconeogenesis58. The deacetylase activity of sirtuins is directed to histone proteins but also other targets including p53 and FOXO-family and PGC-1α transcription factors59–61.

Yeast lifespan is prolonged by growth on a low-glucose containing medium, a condition referred as caloric restriction (CR), which is the oldest known and still sole intervention known to prolong the lifespan in mammals62, 63. Therefore, Sir2 has been proposed to act as a caloric restriction mimetic (reviewed in ref.64, which prompted the search for sirtuin activating molecules. In 2003, resveratrol was identified as an activator of recombinant SIRT1 and Sir2 by a high-throughput screening approach. 100 μM resveratrol activated SIRT1 13.4 fold and less potently Sir26. In keep with the genetic role of Sir2-family proteins in conferring longevity, resveratrol administration prolonged lifespan in a variety of organisms6–9. Such findings sparked a great scientific interest; also they reached to the general public and made of resveratrol a sought-after “miracle drug.” However, while genetic evidence linking SIR2 and longevity is well established65, the link between resveratrol administration and lifespan extension as well as the biochemical activation of Sir2/SIRT1 by resveratrol have been disputed.

Failure to detect resveratrol-induced lifespan extension in yeast66 and more recently drosophila28 have been reported. Similarly, resveratrol administration to C. elegans did not uniformly increase lifespan in all the replicates performed28. More stable resveratrol derivatives that increase yeast replicative lifespan have recently been obtained67, and will be instrumental for further lifespan measurements in C. elegans, drosophila, and mice. As for the activatory action of resveratrol on Sir2/SIRT1, it has been shown by two independent groups that the originally reported resveratrol-induced increase of sirtuin enzymatic activity is dependent on the modified substrate used in the in vitro screening (a fluorescently-modified p53 acetylated substrate). A comparison of SIRT1 deacetylase activity was performed using a p53 acetylated peptide, a histone H4 acetylated peptide and the corresponding fluorescently modified peptides. This clearly demonstrated activation by resveratrol only towards SIRT1 deacetylase activity when the fluorescently-modified peptide was used, casting doubts on the physiological relevance of SIRT1 activation by resveratrol66, 68. Furthermore, proof that resveratrol activates yeast Sir2 or SIRT1 in vivo is still formally lacking69.

Other Proteins

Besides the action of resveratrol on COX/LOX, kinases and sirtuins, on which there is a substantial amount of biochemical and functional insights, several other resveratrol targets have been identified that certainly deserve further investigation to fully elucidate the underlying biochemical outcomes and the biological relevance. Such targets include ribonucleotide reductase, adenylyl cyclase, aromatase, DNA polymerases α and δ, and quinone reductase 2.

Ribonucleotide Reductase

The ribonucleotide reductase is a complex enzyme that catalyses the reduction of ribonucleotides into the corresponding deoxyribonucleotides and is targeted by phenol-derivatives used as anticancer drugs such as gemcitabine and hydroxyurea. Because of its structural similarity with these anticancer drugs, resveratrol was tested as a ribonucleotide reductase inhibitor and was found to inhibit the enzyme even more potently than hydroxyurea (which has an IC₅₀ = 100 μM). As for hydroxyurea, resveratrol destroys a catalytically essential tyrosil radical of ribonucleotide reductase in a stoichiometric amount. Furthermore, the IC₅₀ value towards inhibition of DNA synthesis in human erythroleukemia and mouse mastocytoma cells was found to be of 8–10 μM18. An analog of resveratrol with increased inhibitory potency towards ribonucleotide reductase has recently been reported, which could allow the initiation of preclinical studies70.

Adenylyl Cyclase

Resveratrol was found to increase cAMP levels in MCF-7 human breast cancer cells in a time and concentration dependent manner (EC₅₀ ≈ 1 μM). Concomitantly, cell proliferation was significantly lower than in control cells, suggesting adenylyl cyclase as a possible mediator of the observed effect26, since increased cAMP levels have a cytostatic and proapoptotic effect in breast cancer cells71. Although of potential interest, no further studies have to date substantiated this observation. The effect of resveratrol on MCF-7 proliferation may thus be mediated through other targets.

Quinone Reductase 2

Quinone reductases QR1 and QR2 are cytosolic flavoproteins catalyzing the reduction of quinine and quinone derivatives72. Initially, QR2 was identified as a resveratrol interacting protein by affinity chromatography73. QR2 deficient mice and cells, being unable to metabolise menadione—with simultaneous production of reactive oxygen species—displayed resistance to menadione-induced cytotoxicity72. Similarly, resveratrol administration (5 μM) to human erythroleukemic K562 cells or RNAi to QR2 decreased menadione-induced cell death24. Biochemical and crystallographic studies showed that resveratrol inhibits QR2 with a dissociation constant of 35 nM, making QR2 the resveratrol target with the highest affinity in vitro described to date24. Hence, QR2 is potentially one of the most interesting targets to explain the actions of resveratrol, although in vivo studies will be required to confirm this notion.

Aromatase

Estrogen biosynthesis is catalysed by cytochrome P450 (CYP) 19 enzyme or aromatase, which controls the rate-limiting reaction. Catalytic activity of recombinant aromatase as well as mRNA expression in human placental JEG-3 cells and human breast cancer cell lines (MCF-7 and SK-BR-3) were significantly reduced by 25–50 μM resveratrol treatment25, 74. Modulation of estrogen synthesis and estrogen receptor function by resveratrol may thus contribute to the protection against several types of cancer, including breast cancer75.

DNA Polymerases

Bioassay-directed purification of Psoralea corylifolia (a chinese herb) ethanol extracts led to the identification of corylifolin and bakuchiol as DNA polymerase inhibitors. As both molecules share a 4-hydroxystyryl moiety, also present in resveratrol (Fig. 1), DNA polymerase inhibiting activity of resveratrol was tested, and was found to occur at 10 μM in an SV40 viral DNA replication assay19. More detailed structure-function analysis subsequently showed that resveratrol inhibits DNA polymerases α and δ (IC₅₀ 3.3 and 5 μM, respectively) and, by comparison with structurally related resveratrol-derivatives, demonstrated the absolute requirement of the 4′-hydroxystyryl moiety in a trans conformation with respect to the m-hydroquinone for inhibition to occur20. Resveratrol treatment of both normal human fibroblasts and fibrosarcoma cells resulted in a slower progression of cells through S phase and in a significantly lower amount of BrdUrd incorporation, indicating that a DNA polymerase inhibitory activity can underline the observed data20. Resveratrol inhibition of DNA synthesis can thus provide an additional molecular mechanism for the chemopreventive activity of the molecule.

ESTABLISHING A LINK BETWEEN RESVERATROL- TARGET PROTEINS AND OBSERVED BIOLOGICAL EFFECTS

PAGE analysis (parametric analysis of gene set enrichment) indicated that resveratrol caused a significant alteration in 127 pathways, including the TCA cycle, glycolysis, the classic and alternative complement pathways, butanoate and propanoate metabolism, sterol biosynthesis, Stat3 signalling, insulin signalling, IGF-1 and mTOR signalling, oxidative phosphorylation and electron transport9. Such a diverse range of effects may depend on the interplay between resveratrol and a very specific cellular target that must be an upstream controller of many cellular events. In this line of reasoning is the idea that resveratrol effects may be mediated by the SIRT1-PGC1α axis. A second hypothesis is that resveratrol exerts its beneficial effects on metabolism, cardiovascular function and tumor chemoprotection by simultaneously interacting with several cellular targets. The fact that multiple resveratrol targets have been discovered (summarised in Table 1) may argue in favour of this second hypothesis. We note that, in the original discoveries of resveratrol inhibition of COX, kinases and QR2 or activation of sirtuin, other molecules structurally related to resveratrol have likewise been shown to act similarly. These molecules appear to act simultaneously on some or all the aforementioned resveratrol targets. As an example, quercetin, a flavonoid structurally related to resveratrol (Fig. 1B), has been shown to (i) activate SIRT16, (ii) inhibit QR224, (iii) inhibit the p56^lck tyrosine kinase4, and (iv) PI3K76. Structural data on the quercetin binding to QR224, p56^lck tyrosine kinase77 and PI3K78 have also been obtained and suggest similar binding modes of quercetin and resveratrol to their common targets. In summary, it is clear from the current literature that resveratrol has numerous targets, and possibly relies its biological effects through their simultaneous modulation. Thus, future studies on the properties of resveratrol should—within the reasonable experimental feasibility—evaluate the impact of resveratrol on the maximal number of reported targets. Overall, the complex physiological action of resveratrol and its capacity to modulate different pathways in the micromolar range support the hypothesis of a mechanism involving multiple molecular targets.

Acknowledgements

This work has been supported by an ARD/PNRD (Association pour la Recherche sur le Diabète and Programme National de Recherche sur le Diabète) grant (to L.P.). S. Fröjdö is recipient of a doctoral fellowship from the Ministère de l'Enseignement Supérieur et de la Recherche.