Resveratrol prevents apoptosis in K562 cells by inhibiting lipoxygenase and cyclooxygenase activity

Authors


A. Finazzi Agrò, Department of Experimental Medicine and Biochemical Sciences, University of Rome ‘Tor Vergata’, Via di Tor Vergata 135, I-00133 Rome, Italy. Fax: + 39 067 259 6468, Tel.: + 39 067 259 6468, E-mail: finazzi@uniroma2.it

Abstract

The natural polyphenolic compound resveratrol (trans-3,4′,5-trihydroxystilbene) is shown to prevent apoptosis (programmed cell death) induced in human erythroleukemia K562 cells by hydrogen peroxide and other unrelated stimuli. Resveratrol reversed the elevation of leukotriene B4 (from 6.40 ± 0.65 to 2.92 ± 0.30 pmol·mg protein−1) and prostaglandin E2 (from 11.46 ± 1.15 to 8.02 ± 0.80 nmol·mg protein−1), induced by H2O2 challenge in K562 cells. The reduction of leukotriene B4 and prostaglandin E2 correlated with the inhibition of the 5-lipoxygenase activity, and the cyclooxygenase and peroxidase activity of prostaglandin H synthase, respectively. Resveratrol also blocked lipoperoxidation induced by hydrogen peroxide in K562 cell membranes. Resveratrol was found to act as a competitive inhibitor of purified 5-lipoxygenase and 15-lipoxygenase and prostaglandin H synthase, with inhibition constants of 4.5 ± 0.5 µm (5-lipoxygenase), 40 ± 5.0 µm (15-lipoxygenase), 35 ± 4.0 µm (cyclooxygenase activity of prostaglandin H synthase) and 30 ± 3.0 µm (peroxidase activity of prostaglandin H synthase). Altogether, the results reported here suggest that the anti-apoptotic activity of resveratrol depends on the direct inhibition of the main arachidonate-metabolizing enzymes.

Abbreviations
cisplatin

cis-diaminedichloroplatinum(II)

FBS

fetal bovine serum

5-H(P)ETE

5-hydro(pero)xyeicosatetraenoic acid

LTB4

leukotriene B4

MEM

minimal essential medium

PCD

programmed cell death

PGE2

prostaglandin E2

PhCH2SO2F

phenylmethanesulphonyl fluoride

PHS

prostaglandin H synthase

resveratrol

trans-3,4′,5-trihydroxystilbene

ROS

reactive oxygen species

stilbene

trans-1,2-diphenylethylene

TMPD

N,N,N,N′-tetramethyl-p-phenylenediamine.

Resveratrol (trans-3,4′,5-trihydroxystilbene) occurs naturally in grapes and a variety of medicinal plants, where it functions as a phytoalexin that protects against infections and other stress factors [1]. Because of its high concentration in grape skin, a significant amount of resveratrol is present in red wine [2] and is thought to be responsible for the reduced risk of cardiovascular disease associated with moderate consumption of this beverage [2]. Resveratrol has been reported to have antiplatelet [1], anti-inflammatory [3,4] and anticarcinogenic [5] effects. Recently, it has also been reported to act as an agonist for the estrogen receptor [6] and to inhibit phorbol-ester-mediated activation of protein kinase C and AP-1-mediated gene expression in human mammary epithelial cells [7]. The various activities of resveratrol have been attributed to its antioxidant properties [8], which contribute to control the intracellular redox balance by inhibiting the formation of reactive oxygen species (ROS; [5]). Moreover, this polyphenolic compound interferes with the arachidonate metabolism, by reducing the levels of leukotrienes (generated by the ‘lipoxygenase pathway’ of the arachidonate cascade) and prostanoids (generated by the ‘cyclooxygenase pathway’), and this activity is also considered crucial for its biological effects [3,5,7]. Although the cyclooxygenase activity [5] and expression [7] of prostaglandin H synthase (PHS) have been shown to be reduced by resveratrol in cellular systems, a direct interaction of this compound with the enzyme has not been characterized to date. The effect of resveratrol on the other arachidonate-metabolizing enzyme, lipoxygenase, has not been investigated either.

Recently, the cancer chemopreventive activity of resveratrol has been related to its ability to trigger apoptosis (programmed cell death, PCD), a process which counteracts mitosis in regulating cell number [9,10]. However, when human leukemia HL-60 cells were used as a model system, contradictory results were reported [11,12]. In contrast, ROS and lipid peroxides have long been recognized as crucial elements in the execution of PCD [13–15]. In particular, interest has been attracted by the role of the lipoxygenase pathway in PCD, showing that such unrelated stimuli as cis-diaminedichloroplatinum(II) (cisplatin), transforming growth factor b1, retinoic acid and peroxides may induce apoptosis after lipoxygenase activation in mammalian cells [16–19]. Also alterations of membrane properties in apoptotic cells have been related to lipoxygenase activation [20]. Some of these changes at the membrane level might be the basis for specific recognition and elimination of apoptotic cells by macrophages [15,21].

In this study, we investigated the role of resveratrol in the execution of PCD induced by oxidative stress in human erythroleukemia K562 cells. In an attempt to shed light on the mechanism of action of resveratrol, the effect of this compound on membrane peroxidation and the activity of the arachidonate-metabolizing enzymes 5-lipoxygenase and PHS were measured. Moreover, the interaction of resveratrol with purified 5-lipoxygenase and 15-lipoxygenase and PHS was characterized in vitro, in order to support the hypothesis of a direct in vivo effect of this compound on the enzymes responsible for the arachidonate metabolism. The results reported here may give a biochemical background to the effect of resveratrol on the arachidonate cascade observed in the chemoprevention of coronary heart disease [4] and cancer [5].

Materials and methods

Materials

Chemicals were of the purest analytical grade. Hydrogen peroxide, resveratrol, trans-1,2-diphenylethylene (stilbene), stilbene oxide, arachidonic (eicosatetraenoic) acid, linoleic (octadecadienoic) acid, ATP, hemin, phenol, phenylmethanesulphonyl fluoride (PhCH2SO2F), soybean trypsin inhibitor, N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD), ribonuclease (RNase), propidium iodide, (cisplatin) and transforming growth factor b1 were purchased from Sigma Chemical Co. Trypsin, l-glutamine and nonessential amino acids were purchased from Flow Laboratories Ltd. Culture media Ham’s F-12 and minimal essential medium (MEM) were from Gibco, whereas fetal bovine serum (FBS) was from HyClone. Authentic leukotriene B4 (LTB4) and prostaglandin E2 (PGE2) were from Cayman Chemical Company and had specified purity > 98%. [1-14C]Arachidonic acid (specific activity, 55 mCi mmol−1; 2.04 GBq·mmol−1) was purchased from Amersham.

Enzymes

5-Lipoxygenase (EC 1.13.11.34) was purified to electrophoretic homogeneity from ungerminated barley (Hordeum vulgare), as reported previoulsy [22]. 15-Lipoxygenase (EC 1.13.11.12) was purified to electrophoretic homogeneity from soybean (Glycine max) seeds as reported previously [23]. Prostaglandin H synthase (EC1.14.99.1; PHS), purified from ram seminal vesicles, was purchased from ICN Biochemicals and migrated as a single band of 71 kDa in 12% SDS/PAGE [24].

Cell culture and treatment

Non-adherent K562 cells were grown in a 1 : 1 mixture of MEM and Ham’s F-12 media, supplemented with 12% heat-inactivated FBS, 1.2 mg mL−1 sodium bicarbonate, 1% nonessential amino acids, 2 mm l-glutamine and 15 mm Hepes buffer, pH 7.4, maintaining cell cultures at 37 °C in a humidified atmosphere with 5% CO2. Cells were grown in suspension at a density of 5 × 105 cells·mL−1, split 1 : 7 twice every week and fed routinely 24 h prior to each experiment. Cells were routinely tested for mycoplasma infection. K562 cells (6 × 105 cells·mL−1) were treated with 10 mm hydrogen peroxide for 15 min, as described previously [18]. Different amounts of resveratrol, stilbene or stilbene oxide (all dissolved in methanol) were added directly to the culture medium, and control cells were treated with the same volumes of vehicle alone, which never exceeded 1% of the total volume. After each treatment, cells were grown for 24 h at 37 °C, cell viability was then assessed by Trypan blue dye-exclusion.

Analysis of arachidonate metabolites

Incorporation and release of [14C]arachidonic acid was followed as reported by Maccarrone et al. [25]. K562 cells (8 × 106 in 4 mL), untreated or treated with 10 mm H2O2 for 15 min, in the presence or absence of 30 µm resveratrol, were washed and resuspended in serum-free medium, containing 1 µCi [14C]arachidonic acid per flask (4 mL). After 2 h of incubation at 37 °C, cells were washed again and resuspended in normal culture medium for an additional 22 h at 37 °C. Membrane lipids were then extracted from K562 cells as described previously [17], and radioactivity was measured in both lipid extracts and culture media by liquid scintillation counting. Under our experimental conditions, K562 cells incorporated 45 ± 5% of the [14C]arachidonic acid supplied. Arachidonate metabolites PGE2 and LTB4 were extracted from K562 cells on octadecyl/SPE columns (Baker) and were analyzed by RP-HPLC on a C18 3 × 3 CR column (SGE), as reported previously [26]. RP-HPLC was performed on a Perkin–Elmer 1022 LC Plus liquid chromatograph at a flow rate of 1.2 mL·min−1, using methanol/water/trifluoroacetic acid (70 : 30 : 0.07, by volume) as mobile phase. Chromatograms were recorded at 270 nm, assessing peak identity by comparison with authentic standards. Quantitative determinations were performed by integrating peak areas of each compound.

Analysis of membrane lipid spectra

Membrane lipids were extracted from 10 × 106 K562 cells, untreated or treated with 10 mm H2O2 for 15 min, in the presence or absence of 30 µm resveratrol, as described previously [17]. Absorption spectra were then recorded in the wavelength range 200–300 nm, in order to measure the oxidative index, i.e. the A234/205 ratio [27]. Differential spectra were obtained by subtracting the absorption spectrum of controls from that of H2O2-treated K562 cells. Spectra were recorded at room temperature in a UV-Visual spectrometer Lambda 18 (Perkin–Elmer).

Assay of enzyme activity

The activity of 5-lipoxygenase in K562 cells was measured by incubating cell extracts for 10 min at 37 °C in the presence of 1 mm ATP, 2 mm CaCl2 and 40 µm arachidonic acid [25]. 5-Lipoxygenase activity was expressed as pmol 5-hydroperoxyeicosatetraenoic acid (5-HPETE) formed per min per µg protein. Prostaglandin H synthase is a dual enzyme possessing cyclooxygenase and peroxidase activity on the same protein molecule [28]. Cyclooxygenase activity was measured polarographically, at 30 °C, in 0.1 m potassium phosphate, pH 7.2, 1 mm phenol and 75 µm arachidonic acid [29] and was expressed as nmol O2 consumed per min per µg protein. Peroxidase activity was assayed spectrophotometrically, at 30 °C, in 0.1 m Tris/HCl, pH 8.1, 170 µm TMPD, 1 µm hemin and 75 µm arachidonic acid [30] and was expressed as mU per µg protein (1 U = amount of enzyme causing a change of 1 A590 unit during 5 min incubation with the substrate). The cyclooxygenase and peroxidase activities of purified PHS were measured with the same methodology used for cell extract PHS. Progress curves of the oxygenation of 75 µm arachidonic acid by 35 nm PHS were analyzed by measuring oxygen consumption [24]. The activity of purified 5-lipoxygenase and 15-lipoxygenase was measured spectophotometrically by recording the formation of conjugated hydroperoxides from linoleic acid at 234 nm [31]. Progress curves of the dioxygenation of 90 µm linoleic acid by 15 nm 5-lipoxygenase or 15-lipoxygenase were also recorded by following product formation at 234 nm [32]. The effect of resveratrol on the formation of side products during the dioxygenation of 90 µm linoleic acid catalysed by 15 nm 5-lipoxygenase was assessed by following the formation of oxodienes (i.e. 13-oxo-octadeca-9,11-dienoic plus 13-oxo-trideca-9,11-dienoic acids) at 285 nm [32].

Analysis of apoptosis

K562 cells (1 × 106) were collected by centrifugation at 800 g for 5 min and were fixed in a 1 : 1 solution of NaCl/Pi (2.7 mm KCl, 137 mm NaCl, 10 mm sodium phosphate, pH 7.4) and methanol/acetone (4 : 1 v/v). Cells were washed in NaCl/Pi, incubated with 13 kU RNase for 20 min at 37 °C, and then with 40 µg·mL−1 propidium iodide for 15 min, at 37 °C. The number of apoptotic bodies was evaluated by flow cytometry using propidium iodide staining [33] on a FACScan flow cytometer (Becton Dickinson), and was expressed as apoptotic bodies counted every 100 cells. Samples were excited at 488 nm using a 15 mW argon laser, and the fluorescence emitted at 570 nm was measured. Five thousand events were evaluated by the Lysis II Programme (Becton Dickinson), using an electronic gating FSC-H/FSC-A/NaCl/Cit to eliminate cell aggregates. The data reported in this paper are the mean of three independent determinations, with the indicated SD values. Also, reported figures are representative of triplicate experiments. Statistical analysis was performed by the Student’s t-test, elaborating experimental data by means of the InStat programme (GraphPad Software).

Results

Effect of resveratrol and related compounds on K562 cell death

PCD was induced by H2O2-mediated oxidative stress, as described by Hockenbery et al. [34] and in K562 cells as described by Maccarrone et al. [18]. The induction of apoptosis was dose-dependent and reached a maximum in K562 cells treated with 10 mm hydrogen peroxide for 15 min and then grown for 24 h in culture medium [18]. Under these conditions, apoptotic body formation leveled off at 40%, compared with 3% of the untreated control (Fig. 1A). Resveratrol protected K562 cells against H2O2-induced PCD, in a dose-dependent manner, 30 µm resveratrol almost abolishing the effect of H2O2 (Fig. 1A). Therefore, treatment of K562 cells with H2O2 in the presence of 30 µm resveratrol was used to further investigate the mechanism of PCD prevention by this polyphenolic compound. Stilbene and stilbene oxide are chemically related to resveratrol, which is a trihydroxystilbene. Neither of these compounds was able to significantly prevent K562 cell death induced by H2O2 (Fig. 1B). Interestingly, resveratrol, but not its analogs, also prevented PCD induced in K562 cells by treatment with either 5 µm 5-hydroxyeicosatetraenoic acid (5-HETE) or a combination of 5 µm cisplatin plus 1 ng·mL−1 transforming growth factor b1 (data not shown).

Figure 1.

Figure 1.

Effect of resveratrol and related compounds on apoptosis induced by H2O2in K562 cells. Human erythroleukemia K562 cells were exposed to 10 mm H2O2 for 15 min, in the presence of different concentrations of resveratrol (A) or 30 µm of chemically related compounds (B). The formation of apoptotic bodies in K562 cells was evaluated by cytofluorimetric analysis after 24 h of culture. Data are the mean (± SD) of three independent determinations, each performed in duplicate. In (B) differences between samples untreated or treated with stilbene or stilbene oxide were not significant (P > 0.05).

Effect of resveratrol on arachidonate metabolism in K562 cells

Treatment with H2O2 did not significantly affect the ability of human erythroleukemia K562 cells to incorporate arachidonate into membrane lipids or to release it into the culture medium. Indeed, incorporation or release of arachidonate in K562 cells treated with 10 mm H2O2 for 15 min and then cultured for 24 h were 95% or 92% of the control values, respectively. Approximately 25% of incorporated archidonic acid was released during the 22 h following incubation with [14C]arachidonate, the remaining 75% being detected in the membrane lipid extracts. RP-HPLC analysis indicated that arachidonate was not present in membrane lipids as a free fatty acid (data not shown), suggesting that it was incorporated into the different (phospho)lipid classes present in human cell membranes [17]. Treatment of cells with 30 µm resveratrol did not affect arachidonate incorporation or release (not shown). In contrast, the early phase of PCD induced by oxidative stress was characterized by increased LTB4 content, which reached 326% of the untreated control in cells exposed for 15 min to 10 mm H2O2 (Table 1). Treatment of K562 cells with H2O2 in the presence of 30 µm resveratrol led to a significantly reduced LTB4 production, which was ≈ 150% of the untreated controls (Table 1). Prostaglandin E2 was also found to increase in K562 cells after 15 min of treatment with 10 mm H2O2, reaching 200% of the basal level (Table 1). Prostaglandin E2 increase was significantly counteracted by treatment of K562 cells with 30 µm resveratrol, down to 140% of untreated cells (Table 1). Resveratrol significantly reduced even the basal level of LTB4 and PGE2 in K562 cells, to 48% and 65% of the untreated control, respectively (Table 1).

Table 1. Effect of resveratrol on the content of LTB4 and PGE2 in human cells. Erythroleukemia K562 cells were treated with 10 mm H2O2 for 15 min, in the absence or in the presence of 30 µm resveratrol, or with 30 µm resveratrol alone. Values in parentheses represent percentage of the control, arbitrarily set to 100. Data are the mean (± SD) of three independent experiments, each performed in duplicate.
SampleLTB4 content
(pmol·mg·protein−1)
PGE2 content
(nmol·mg·protein−1)
  1. a P < 0.01 compared with K562 (control). b P < 0.05 compared with K562 (control). c P < 0.01 compared with K562 (H2O2). dP < 0.05 compared with K562 (H2O2).

K562 (control)1.96 ± 0.205.73 ± 0.61
 (100%)(100%)
K562 (H2O2)6.40 ± 0.6511.46 ± 1.15
 (326%)a(200%)a
K562 (resveratrol)0.95 ± 0.103.72 ± 0.40
 (48%)b(65%)b
K562 (H2O2 + resveratrol)2.92 ± 0.308.02 ± 0.80
 (149%)b,c(140%)b,d

Effect of resveratrol on 5-lipoxygenase and PHS activity in K562 cells

Leukotriene B4 is the main product of 5-lipoxygenase [35], whereas PGE2 is the main product of PHS in vivo[28]. Therefore, the activity of both enzymes was measured in K562 cells treated with 10 mm H2O2 for 15 min, in the presence or in the absence of 30 µm resveratrol. Both cyclooxygenase and peroxidase activities of PHS were assayed, because they catalyze two different steps of prostanoid biosynthesis [28,29]. Table 2 shows that treatment with H2O2 roughly doubled both 5-lipoxygenase and PHS activities in K562 cells. Resveratrol (30 µm) also reduced the basal level of 5-lipoxygenase and PHS cyclooxygenase and peroxidase activity in K562 cells (Table 2). Remarkably, it abolished the enhancement of enzyme activity induced by H2O2 treatment (Table 2). Unlike resveratrol, stilbene and stilbene oxide at the same concentration (30 µm) were unable to reverse the hydrogen peroxide-enhanced 5-lipoxygenase or PHS activity in K562 cells (data not shown).

Table 2. Effect of resveratrol on 5-lipoxygenase and PHS activity of human cells. Erythroleukemia K562 cells were treated with 10 mm H2O2 for 15 min, in the absence or in the presence of 30 µm resveratrol, or with 30 µm resveratrol alone. Values in parentheses represent percentage of the control, arbitrarily set to 100. Data are the mean (± SD) of three independent experiments, each performed in duplicate.
Sample5-Lipoxygenase activity
(pmol 5-HPETE·min−1·µg protein−1)
Cyclooxygenase activity of PHS
(nmol O2·min−1·µg protein−1)
Peroxidase activity of PHS
(mU·µg protein−1)
  1. a P < 0.01 compared with K562 (control). b P < 0.05 compared with K562 (control). c P > 0.05 compared with K562 (control). d P < 0.01 compared with K562 (H2O2). e P < 0.05 compared with K562 (H2O2).

K562 (control)0.36 ± 0.040.54 ± 0.051.40 ± 0.15
 (100%)(100%)(100%)
K562 (H2O2)0.86 ± 0.090.93 ± 0.102.52 ± 0.26
 (239%)a(172%)a(180%)a
K562 (resveratrol)0.18 ± 0.020.35 ± 0.040.84 ± 0.09
 (50%)b(65%)b(60%)b
K562 (H2O2 + resveratrol)0.39 ± 0.040.60 ± 0.061.65 ± 0.17
(108%)c,d(111%)c,e(118%)c,e 

Effect of resveratrol on membrane lipid peroxidation

Exposure of K562 cells to 10 mm H2O2 for 15 min significantly increased membrane lipid peroxidation, as shown by a remarkable increase in UV absorption (Fig. 2) at 235 nm (Fig. 2, inset), due to the formation of conjugated fatty acid hydroperoxides. Consistently, the oxidative index of membrane lipids, i.e. the A234/205 ratio [27], was increased by H2O2 treatment from 0.060 ± 0.007 to 0.148 ± 0.015 (P < 0.01). When K562 cells were treated with H2O2 in the presence of 30 µm resveratrol, a significantly lower absorption of the membrane lipids was observed (Fig. 2). The oxidative index of membrane lipids in the presence of resveratrol was also reduced to 67% of the H2O2-treated samples (P < 0.05). It is noteworthy that hydroperoxyeicosatetraenoic acids, which are products of lipoxygenase activity, are among the most important conjugated hydroperoxides in cellular biomembranes [27].

Figure 2.

Figure 2.

Effect of resveratrol on membrane lipid peroxidation of K562 cells. Absorption spectra were recorded on membrane lipids extracted from K562 cells (10 × 106), untreated (a), treated with 10 mm H2O2 for 15 min (b) or treated with 10 mm H2O2 for 15 min in the presence of 30 µm resveratrol (c). Inset: differential spectrum obtained by subtracting the absorption spectrum of controls from that of H2O2-treated cells (b–a).

Interaction of resveratrol with purified 5-lipoxygenase and 15-lipoxygenase and PHS

Resveratrol inhibited the dioxygenation of linoleic acid by purified 5-lipoxygenase in a dose-dependent manner (Fig. 3A). Resveratrol (10 µm) almost completely inhibited 5-lipoxygenase and 15-lipoxygenase to 60% of the control (Fig. 3A). In fact, the respective half-inhibitory concentrations (IC50) were 2.5 ± 0.3 µm for the 5-lipoxygenase and 25 ± 3 µm for the 15-lipoxygenase (Table 3). Kinetic analysis of purified lipoxygenases inhibition by resveratrol showed that it acted as a complete, competitive inhibitor of both isozymes, with inhibition constants (Ki) of 4.5 ± 0.5 µm and 40 ± 5.0 µm, for 5-lipoxygenase and 15-lipoxygenase, respectively (Table 3). In order to obtain a better insight into the mechanism of resveratrol inhibition, the progress curves of the lipoxygenase-catalysed dioxygenation of 90 µm linoleic acid were analyzed. Figure 3C shows that 4 µm resveratrol prolonged the lag phase of the reaction catalysed by 5-lipoxygenase, and superimposable results were obtained with the dioxygenation of 90 µm linoleic acid catalysed by the 15-isozyme, alone or in the presence of 40 µm resveratrol. Remarkably, this compound did not affect oxodiene formation during the 5-lipoxygenase or 15-lipoxygenase catalysed dioxygenation of linoleic acid (data not shown).

Figure 3.

Figure 3.

Interaction of resveratrol with purified 5-lipoxygenase, 15-lipoxygenases and PHS. (A) Dioxygenation of 90 µm linoleic acid catalysed by 15 nm 5-lipoxygenase (empty bars) or 15 nm 15-lipoxygenase (hatched bars), recorded spectrophotometrically at 234 nm. Lipoxygenase activity was expressed as percentage of the untreated control, arbitrarily set to 100 (100% = 1.94 ± 0.20 nm·min−1, for 5-lipoxygenase, and 2.55 ± 0.25 nm·min−1, for 15-lipoxygenase). (B) Cyclooxygenase (empty bars) and peroxidase (hatched bars) activity of 35 nm PHS, expressed as percentage of the untreated control, arbitrarily set to 100 (100% = 30 ± 3 nmol O2·min−1, for the cyclooxygenase, and 10 ± 1 mU, for the peroxidase). (C) Progress curves of the dioxygenation of 90 µm linoleic acid catalysed by 15 nm 5-lipoxygenase, alone (a) or in the presence of 4 µm resveratrol (b). (D) Progress curves of the oxygenation of 75 µm arachidonic acid catalysed by the cyclooxygenase activity of 35 nm PHS, alone (a) or in the presence of 50 µm resveratrol (b). Progress curves of the dioxygenation of 90 µm linoleic acid catalysed by 15 nm 15-lipoxygenase, alone or in the presence of 40 µm resveratrol, were superimposable to those shown in (C) and were omitted for the sake of clarity. Data are the mean (± SD) of three independent determinations, each performed in duplicate.

Table 3. Inhibition of purified 5-lipoxygenase, 15-lipoxygenase and PHS by resveratrol. The median inhibitory concentration (IC50) values were calculated from dose-dependency profiles. The type of inhibition and the inhibition constant (Ki) values were obtained by Lineweaver–Burk analysis of double reciprocal plots, in the substrate concentration range 0–100 µm (linoleic acid was used for 5-lipoxygenase and 15-lipoxygenases, and arachidonic acid for PHS activity, respectively). Data are the mean (± SD) of three independent experiments, each performed in duplicate.
EnzymeType of inhibitionIC50m)Kim)
5-LipoxygenaseCompetitive2.5 ± 0.34.5 ± 0.5
15-LipoxygenaseCompetitive25 ± 3.040 ± 5.0
Cyclooxygenase activity of PHSCompetitive20 ± 2.035 ± 4.0
Peroxidase activity of PHSCompetitive15 ± 2.030 ± 3.0

The interaction between purified PHS and resveratrol was also investigated, showing that this compound reduced both the cyclooxygenase and peroxidase activity of the enzyme, in a dose-dependent manner (Fig. 3B). The inhibition of PHS showed IC50 values of 20 ± 2.0 µm and 15 ± 2.0 µm, for the cyclooxygenase and peroxidase activity, respectively (Table 3). Kinetic analysis showed that resveratrol was a competitive inhibitor of PHS, with inhibition constants (Ki) of 35 ± 4.0 µm and 30 ± 3.0 µm, for the cyclooxygenase and peroxidase activity, respectively (Table 3). Progress curves of PHS cyclooxygenase activity with 75 µm arachidonic acid as substrate showed that 50 µm resveratrol prolonged the lag phase of the reaction (Fig. 3D). It is noteworthy that neither stilbene nor stilbene oxide inhibited significantly 5-lipoxygenase and 15-lipoxygenase or PHS activity (data not shown).

Discussion

Resveratrol was shown to protect K562 cells against oxidative-stress-induced PCD (Fig. 1A), as well as that induced by 5-HETE, cisplatin and transforming growth factor b1 (not shown). The anti-apoptotic activity of this polyphenolic compound required the presence of hydroxy groups, because stilbene and stilbene oxide under the same experimental conditions were ineffective (Fig. 1B). This anti-apoptotic activity of resveratrol correlated with its ability to reverse H2O2-induced increase in LTB4 and PGE2 concentration (Table 1). This effect was, in turn, a result of the inhibition of 5-lipoxygenase and the cyclooxygenase and peroxidase activity of PHS in the same cells (Table 2). In previous studies we have shown that the early phase of hydroperoxide-induced apoptosis in human cancer cells is characterized by increased levels of LTB4[18]. Here, we also showed that the level of PGE2 was enhanced by oxidative stress, although to a lower extent than LTB4 (Table 1). Moreover, the increased arachidonate metabolites in H2O2-challenged K562 cells were not due to an increased availability of the arachidonate substrate, because neither its incorporation nor its release were affected by H2O2. Instead, the activation of the enzymes responsible for the synthesis of eicosanoids, i.e. 5-lipoxygenase which generates LTB4[35], and PHS which produces PGE2[28], may be responsible for their increase. Interestingly, resveratrol, but not stilbene or stilbene oxide, prevented 5-lipoxygenase and PHS activation at the same concentration (30 µm) that protected K562 cells against H2O2-induced PCD (Table 2 and data not shown). The observation that unrelated apoptotic stimuli, such as H2O2, 5-HETE, cisplatin and transforming growth factor b1, all induce PCD through the activation of 5-lipoxygenase [17,18] may explain the anti-apoptotic effect of resveratrol. This compound significantly reduced membrane lipoperoxidation (Fig. 2), which modifies the membrane properties [20] and might be a signal for recognition of apoptotic cells by macrophages [15,21]. Therefore, the inhibition of 5-lipoxygenase and the associated oxidative damage to membrane lipids might explain, at least in part, the anti-apoptotic effect of resveratrol. This compound was already found to inhibit peroxidation of porcine low-density lipoproteins [8]. The inhibition of cyclooxygenase and peroxidase activity of PHS by resveratrol might further contribute to the reduction of the peroxide level in human cells [5,7]. As a matter of fact, the cancer chemopreventive activity of resveratrol [5] and its ability to inhibit a typical redox-regulated system such as phorbol-ester-mediated activation of protein kinase C and AP-1-mediated gene expression [7], have been attributed to the inhibition of cyclooxygenase expression and activity in cells. Nevertheless, an interaction of resveratrol with lipoxygenase has never been investigated, whereas that with PHS has been only preliminarly demonstrated [36,37]. Here, we used purified enzymes to investigate such an interaction in detail. To this aim, in vitro experiments were performed using 5-lipoxygenase purified from barley (Hordeum vulgare), and 15-lipoxygenase purified from soybean (Glycine max). It is worth reminding that these plant enzymes are widely used as models of the hardly available mammalian counterparts for structural and mechanistic studies [38–40]. PHS purified from ram seminal vesicles was used as a model of the human enzyme [28]. Resveratrol is shown for the first time to inhibit both 5-lipoxygenase and 15-lipoxygenases (Fig. 3A), as well as cyclooxygenase and peroxidase activity of PHS (Fig. 3B), in a dose-dependent manner. Kinetic analysis indicated that resveratrol acted as a competitive inhibitor in all cases (Table 3). Remarkably, resveratrol inhibited 5-lipoxygenase 5–10 times more efficiently than the 15-isozyme. Resveratrol prolonged the lag phase of both isozymes, indicating a possible reduction of Fe(III) to Fe(II) at the catalytic site [39]. Furthermore, the observation that resveratrol did not affect oxodiene formation during the 5-or 15-lipoxygenase-catalysed dioxygenation of linoleic acid (data not shown) suggests that this compound does not scavenge the free radicals formed in the course of enzyme activity [32]. Altogether, these results suggest that the mechanism of action of resveratrol on either lipoxygenase was the same. As far as PHS is concerned, both cyclooxygenase and peroxidase activities depend on ferriprotoporphyrin IX [28,29]. Again, the prolonged lag phase of the cyclooxygenase reaction was indicative of a reduction of Fe(III) to Fe(II) [24,29]. Interestingly, both cyclooxygenase and peroxidase activities of PHS, which together catalyse the first committed step in prostanoid formation [28], were inhibited to the same extent by resveratrol (Fig. 3B and Table 3). Suggestive is the finding that the IC50 value found for the inhibition of PHS activity by resveratrol is quite similar to its effective dose (ED50) for antitumor promotion and inhibition of ROS formation in human HL-60 cells treated with phorbol esters [5]. This resveratrol concentration has been also found to inhibit phorbol-ester-mediated activation of protein kinase C and AP-1-mediated gene expression [7]. Thus, the anticancer and anti-inflammatory properties of resveratrol might actually be due to a direct interaction with cellular PHS.

In conclusion, the present results indicate that resveratrol, unlike other stilbenes, prevents the death program induced in K562 cells by oxidative stress and other unrelated stimuli. Seemingly, they do not favour the hypothesis that cancer chemoprevention by resveratrol occurs through a pro-apoptotic activity of this antioxidant. The inhibition of the arachidonate-metabolizing enzymes 5-lipoxygenase, 15-lipoxygenase and PHS by resveratrol may give a biochemical background to the effect of this compound in coronary heart disease [4] and cancer [5] prevention.

Acknowledgements

The authors wish to thank Drs Cristina Fantini and Monica Bari for their excellent technical assistance. Financial support from Ministero dell’ Università e della Ricerca Scientifica e Tecnologica (MURST-PRIN 1997) and Istituto Superiore di Sanità (XI AIDS Project 1998) is gratefully acknowledged as well.

Footnotes

  1. Enzymes: 5-lipoxygenase (EC 1.13.11.34); 15-lipoxygenase (EC 1.13.11.12); prostaglandin H synthase (EC 1.14.99.1).

Ancillary