Dietary flavonoid iron complexes as cytoprotective superoxide radical scavengers

Free Radical Biology & Medicine, Vol. 34, No. 2, pp. 243–253, 2003 Copyright © 2003 Elsevier Science Inc. Printed in the USA. All rights reserved 0891-5849/03/$–see front matter

PII S0891-5849(02)01241-8

Original Contribution DIETARY FLAVONOID IRON COMPLEXES AS CYTOPROTECTIVE SUPEROXIDE RADICAL SCAVENGERS MAJID Y. MORIDANI,*† JALAL POURAHMAD,* HOANG BUI,* ARNO SIRAKI,*

and

PETER J. O’BRIEN*

†

*Faculty of Pharmacy, and Department of Pediatric Laboratory Medicine, The Hospital for Sick Children, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (Received 7 May 2002; Revised 9 September 2002; Accepted 10 October 2002)

Abstract—Superoxide radicals have been implicated in the pathogenesis of ischemia/reperfusion, aging, and inflammatory diseases. In the present work, we have shown that the Fe3⫹ complexes of flavonoids (polyphenols) were much more effective than the uncomplexed flavonoids in protecting isolated rat hepatocytes against hypoxia-reoxygenation injury. The 2:1 flavonoid-metal complexes of Cu2⫹, Fe2⫹, or Fe3⫹ were more effective than the parent compounds in scavenging superoxide radicals generated by xanthine oxidase/hypoxanthine (an enzymatic superoxide-generating system). The 2:1 [flavonoid:Fe3⫹] complexes but not the [deferoxamine:Fe3⫹] complex readily scavenged superoxide radicals. These results suggest that the initial step in superoxide radical scavenging (SRS) activity involves a redox-active flavonoid:Fe3⫹ complex. Flavonoid:Fe3⫹ complexes should, therefore, be tested as a therapy for the treatment of ischemia/reperfusion injury. © 2003 Elsevier Science Inc. Keywords—Iron, Copper, Superoxide dismutase mimics, Ischemia/reperfusion injury, Flavonoids, Antioxidants, Free radicals

INTRODUCTION

tocyte toxicity, which was attributed to ATP depletion and reductive stress, intracellular iron release, and a marked increase in oxygen activation [14,15]. The cytotoxicity and oxygen activation were prevented by catechols, such as caffeic acid, presumably as a result of their superoxide radical-scavenging (SRS) activity [14,15]. The SRS activity for a number of catechols (namely, protocatechuic acid, 4-t-butylcatechol, catechin, and tiron) was also found to be markedly enhanced when complexed with ferric ion. The [catechol:ferric] complexes were also much more cytoprotective than the uncomplexed catechols towards hepatocyte hypoxiareoxygenation injury [16]. Flavonoids as part of the human diet [17] are a group of phytochemicals with many antioxidant, antiviral, and antimutagenic effects [18 –20]. In the present work, we have investigated the SRS activity of the Fe2⫹, Fe3⫹, and Cu2⫹ chelates of flavonoids and found that the chelates were much more potent as superoxide radical scavengers than their corresponding uncomplexed flavonoids in a xanthine oxidase/hypoxanthine superoxide-generating system. The flavonoid metal complexes were also more effective than uncomplexed flavonoids at preventing hypoxia-reoxygenation hepatocyte injury caused by re-

The pharmacological administration of superoxide dismutase (SOD) or cell membrane-permeable superoxide dismutase mimics are suggested as a novel therapy to alleviate the degenerating action of superoxide radicals in the pathogenesis of ischemia/reperfusion injury, aging, cancer, and other metabolic, degenerative, and inflammatory diseases [1]. Various low-molecular-weight SOD-like complexes of manganese, copper, and iron have also been reported to exhibit SOD mimic activities [2–9]. At a physiological pH, catechols readily form thermodynamically stable bis complexes with ferric iron as bidentate ligands. Catechols are also biosynthesized and used as iron-sequestering agents by microorganisms [10,11]. These catechols have large stability constants for ferric iron (log K ⱖ 40) and low reduction potentials [12,13]. Previously, we showed that maintaining hepatocytes under a ⬍ 0.3% oxygen concentration resulted in hepaAddress correspondence to: Dr. Peter J. O’Brien, Faculty of Pharmacy, University of Toronto, 19 Russell Street, Toronto, Ontario M5S 2S2, Canada; Tel: (416) 978-2716; Fax: (416) 978-8511; E-Mail: [email protected]. 243

244

M. Y. MORIDANI et al.

Table 1. Distribution Coefficient Values at pH 7.4 for Flavonoids in Octan-1-o1/MOPS Buffer System

Flavonoid

Monitored wavelength (nm)

Dpart pH ⫽ 7.4

LogD7.4

C2¢C3 double bond

C3

C4 (C¢O)

C5

C7

C3⬘

C4d⬘

Rutin Catechin Taxifolin Quercetin Fisetin Luteolin Kaempferol

270, 280 280, 280, 250, 270, 265,

0.4 ⫾ 0.1 1.2 ⫾ 0.4 5.6 ⫾ 0.6 182 ⫾ 22 158 ⫾ 14 112 ⫾ 10 524 ⫾ 36

⫺0.40, ⫺0.43*, ⫺0.35† 0.08 0.75 2.26, 0.08*, 2.88† 2.20, 2.53† 2.05, 1.34,* 2.53† 2.72, 1.84,* 3.39†

double bond – – double bond double bond double bond double bond

OR OH OH OH OH – OH

C¢O – C¢O C¢O C¢O C¢O C¢O

OH OH OH OH OH OH OH

OH OH OH OH – OH OH

OH OH OH OH OH OH –

OH OH OH OH OH OH OH

365 300, 330 320, 380 320 365 310, 370

To a 25 ml MOPS buffer [(50 mM pH 7.4 and 1 mM DETAPAC) presaturated by octan-1-o1 overnight] containing the flavonoid (50 –200 ␮M), an aliquot of 5–10 ml octan-1-o1 (presaturated with MOPS buffer) was added in time intervals of 45 min, during which the solution was stirred over a magnetic stirrer. The absorbance of a sample (3 ml) from the aqueous phase after centrifugation was monitored at the given wavelength before and after each octan-1-o1 addition. The samples were returned to the flask before the next octan-1-o1 addition. The distribution coefficient values of the flavonoids between octan-1-o1 and water system were calculated from Dpart ⫽ [(A1 ⫺ A2)/A2 ⫻ Vw/Vo], where Dpart is the distribution coefficient value at a given pH (7.4); A1 and A2 are the absorbances of aqueous phase before and after the octan-1-o1 additions, respectively; and Vw and Vo are the volumes of aqueous and octan-1-o1 phases, respectively. Results were expressed as means ⫾ SD for three independent determinations for each flavonoid. R is a sugar moiety attached to the C3 of the rutin C-ring. * Values are taken from reference [21]. † Values are taken from reference [22].

active oxygen species. Iron complexes of the iron-chelating drug deferoxamine had no SRS activity and did not prevent hypoxia-reoxygenation hepatocyte injury. MATERIALS AND METHODS

Chemicals Flavonoids, xanthine oxidase, hypoxanthine, catechin, quercetin, rutin, fisetin, luteolin, taxifolin, kaempferol, caffeic acid, 4-t-butylcatechol, Tris(hydroxymethyl)aminomethane (Tris), nitro blue tetrazolium (NBT), sodium phosphate monobasic, sodium phosphate dibasic, diethylenetriaminepentaacetic acid (DETAPAC), ethylenediaminetetraacetic acid (EDTA), 3-(N-morpholino)propanesulfonic acid (MOPS), ferrous sulfate, ferric nitrate, ferric chloride, cupric acetate, dimethyl sulfoxide (DMSO), and 1-octanol were obtained from Sigma-Aldrich Corp. (St. Louis, MO, USA). The stock solutions of flavonoids (polyphenols) were prepared in DMSO. Other chemicals were dissolved in Millipore filtered water or buffer (Millipore, Etobicoke, Ontario, Canada). A Shimadzu uv-visible spectrophotometer (UV-240; Shimadzu, Kyoto, Japan) was used throughout the experiment. UV-VIS spectroscopy of flavonoids metal complexes Stock solutions of catechols, deferiprone, and flavonoids (polyphenols) were prepared (2.5 mM) using

DMSO as a solvent. A final concentration of 25 ␮M was then used in a cuvette that contained phosphate buffer (10 mM, pH 5.5 and 7.4) and the absorption spectra were recorded between 200 –700 nm. Scans with 50 ␮M ferrous sulfate, ferric nitrate, and cupric acetate were taken after 5 min and compared with the flavonoid alone. The reversibility of the complexes formed was determined by the addition of 2.5 or 25 molar equivalents of EDTA (125 ␮M and 1.25 mM, respectively). Distribution coefficient value determination The distribution coefficient measurement system used was a modified method of previously published work [21]. To a 25 ml MOPS buffer [(50 mM pH 7.4 and 1 mM DETAPAC) presaturated by octan-1-ol overnight] containing the flavonoid (50 –200 ␮M), an aliquot of 5–10 ml octan-1-ol (presaturated with MOPS buffer) was added and stirred for 45 min before letting the mixture stand for 15 min. The absorbance of a sample (3 ml) from the aqueous phase after centrifugation was monitored at the given wavelength (Table 1) before and after each octan-1-ol addition. The samples were returned to the flask before the next octan-1-ol addition. All solutions were stored and manipulated at room temperature. The distribution coefficient values of the flavonoids between octan-1-ol and water system were calculated from Dpart ⫽ [(A1 ⫺ A2)/A2 ⫻ Vw/Vo], where Dpart is the distribution coefficient value at pH 7.4; A1 and A2 are the

Flavonoid iron complexes

245

Table 2. Flavonoid:Fe2⫹, Fe3⫹, or Cu2⫹ Complexes as Superoxide Scavengers (%) in Xanthine/Hypoxanthine System

Compound None Deferoxamine 4-t-Butylcatechol Rutin Taxifolin Catechin Fisetin Quercetin Luteolin Kaempferol Histidine

Ligand (10 ␮M)

[Ligand:Fe2⫹] (10 ␮M:5 ␮M)

[Ligand:Fe3⫹] (10 ␮M:5 ␮M)

[Ligand:Cu2⫹] (10 ␮M:5 ␮M)

0 0 60 ⫾ 5 21 ⫾ 2 22 ⫾ 3 36 ⫾ 3 22 ⫾ 2 14 ⫾ 1 7⫾1 20 ⫾ 2 –

0 0 61 ⫾ 4 20 ⫾ 2 29 ⫾ 2 52 ⫾ 3 19 ⫾ 3 35 ⫾ 3 – – –

0 0 68 ⫾ 3 42 ⫾ 2 42 ⫾ 2 42 ⫾ 2 77 ⫾ 2 79 ⫾ 3 23 ⫾ 3 37 ⫾ 4 –

0 63 ⫾ 3 83 ⫾ 4 82 ⫾ 2 78 ⫾ 2 84 ⫾ 3 66 ⫾ 4 74 ⫾ 3 81 ⫾ 3 73 ⫾ 3 31 ⫾ 4

The reaction mixture contained the ligand (50 ␮M), xanthine oxidase XO (25 munit/ml), and NBT (50 ␮M) in Tris buffer (0.1 M, pH 7.4), to which hypoxanthine (35 ␮M) was added. The absorbance of the reaction mixture was monitored spectrophotometrically at 560 nm for 10 min. A 5 min premixed stock solution of the [ligand:metal] (100 ␮M:50 ␮M) complex was used for the experiment. Results are expressed as means ⫾ SD for three independent determinations. Fe3⫹ in the absence of ligand had no superoxide radical–scavenging activity.

absorbances of aqueous phase before and after the octan1-ol additions, respectively; and Vw and Vo are the volumes of aqueous and octan-1-ol phases, respectively. Superoxide radical-scavenging activity measurement To a mixture of flavonoids (50 ␮M), xanthine oxidase (25 munit/ml), and nitro blue tetrazolium (50 ␮M) in Tris buffer (0.1 M, pH 7.4), an aliquot of 10 ␮l hypoxanthine (3.5 mM) was added. The total volume of the mixture was 1 ml. The absorbance of the reaction mixture was monitored spectrophotometrically at 560 nm for 10 min using a Pharmacia Ultraspec model 1000 (Pharmacia LKB Biotechnology Inc., Piscataway, NJ, USA). To measure the SRS activity of the complex, a stock solution of the [flavonoid:metal] (100 ␮M:50 ␮M) complex was prepared 5 min before adding to the reaction mixture at a final concentration of (10 ␮M:5 ␮M). Xanthine oxidase inhibition assay A modified method described by Cotelle et al. [23] was used to measure the xanthine oxidase-catalyzed uric acid production. Briefly, hypoxanthine (35 ␮M) was added to a mixture of xanthine oxidase (25 munit/ml) and flavonoid (10 –50 ␮M) or [flavonoid:Fe3⫹] (10 ␮M:5 ␮M) in Tris buffer (pH 7.4, 0.1 M). The change in absorbance was monitored at 293 nm at room temperature for 5 min. Cytotoxicity in isolated rat hepatocytes Adult male Sprague-Dawley rats, 250 –300 g, were obtained from Charles River Canada Laboratories (Montreal, Quebec, Canada), fed ad libitum, and allowed to acclimatize for one week on clay chip bedding. Hepatocytes were isolated from rats by collagenase perfusion of

the liver according to Molde´ us and coworkers [24]. Isolated hepatocytes (106 cells/ml) (10 ml) were suspended in Krebs-Henseleit buffer (pH 7.4) containing HEPES (12.5 mM) in continually rotating round-bottomed 50 ml flasks, under an atmosphere of 95% O2 and 5% CO2 in a water bath of 37°C for 30 min. Then flavonoids and their Fe3⫹ complexes were added to the flasks and the hepatocytes were incubated under an atmosphere of 95% N2 and 5% CO2 for 90 min before reoxygenation with 1% O2, 94% N2, and 5% CO2 atmosphere. Before reoxygenation, the first assessment of the cells viability was carried out by determining trypan blue exclusion from the hepatocytes [24]. The SRS activity of flavonoids (20 ␮M) were compared with their 2:1 Fe3⫹ complexes. A 5 min premixed [flavonoid:Fe3⫹] (20 mM:10 mM) complex was prepared before addition to the hepatocytes. Statistical analysis Statistically significant differences between control and test compounds were determined using Student’s t-test. The acceptable values were p ⱕ .05.

RESULTS

Superoxide radical-scavenging activity of flavonoid iron complexes The SRS activities of flavonoids with and without complexing with Fe2⫹ and Fe3⫹ were compared using an enzymatic superoxide radical-generating system (hypoxanthine/xanthine oxidase). As shown in Table 2, the SRS activity of all tested [flavonoid:Fe2⫹] complexes was greater than that of the uncomplexed flavonoids. The order of SRS effectiveness of the 2:1 [flavonoid:Fe2⫹]

246

M. Y. MORIDANI et al. Table 3. Prevention of Hypoxia-Reoxygenation Injury Cytotoxicity in Isolated Rat Hepatocytes by Deferoxamine and Flavonoids-Fe3⫹ Complexes Hepatocyte toxicity (% of trypan blue uptake) Treatment

90 min

2h

3h

Hypoxia-reoxygenation ⫹Superoxide dismutase (100 unit/ml) ⫹Fe2⫹ (10 ␮M) ⫹Deferoxamine (20 ␮M) ⫹[Deferoxamine:Fe3⫹] (20 ␮M:10 ␮M) ⫹4-t-Butylcatechol (20 ␮M) ⫹[4-t-Butylcatechol:Fe3⫹] (20 ␮M:10 ␮M) ⫹Taxifolin (20 ␮M) ⫹[Taxifolin:Fe3⫹] (20 ␮M:10 ␮M) ⫹Luteolin (20 ␮M) ⫹[Luteolin:Fe3⫹] (20 ␮M:10 ␮M) ⫹Quercetin (20 ␮M) ⫹[Quercetin:Fe3⫹] (20 ␮M:10 ␮M) ⫹ Fisetin (20 ␮M) ⫹ [Fisetin:Fe3⫹] (20 ␮M:10 ␮M) ⫹Rutin (20 ␮M) ⫹[Rutin:Fe3⫹] (20 ␮M:10 ␮M) ⫹Catechin (20 ␮M) ⫹[Catechin:Fe3⫹] (20 ␮M:10 ␮M)

45 ⫾ 2 31 ⫾ 3 43 ⫾ 3 44 ⫾ 3 41 ⫾ 2 43 ⫾ 4 34 ⫾ 3 47 ⫾ 3 25 ⫾ 2 45 ⫾ 3 22 ⫾ 5 40 ⫾ 5 32 ⫾ 6 45 ⫾ 2 22 ⫾ 4 49 ⫾ 3 29 ⫾ 4 38 ⫾ 4 31 ⫾ 3

75 ⫾ 3 48 ⫾ 4 72 ⫾ 4 72 ⫾ 3 69 ⫾ 2 65 ⫾ 5 51 ⫾ 4 62 ⫾ 4 33 ⫾ 3 55 ⫾ 2 32 ⫾ 6 69 ⫾ 7 48 ⫾ 5 54 ⫾ 4 34 ⫾ 3 54 ⫾ 4 36 ⫾ 5 66 ⫾ 6 53 ⫾ 6

93 ⫾ 4 52 ⫾ 4* 96 ⫾ 4 86 ⫾ 2 88 ⫾ 3* 93 ⫾ 5 62 ⫾ 4* 73 ⫾ 3 35 ⫾ 2* 66 ⫾ 3 35 ⫾ 3* 88 ⫾ 8 58 ⫾ 5* 65 ⫾ 3 36 ⫾ 3* 66 ⫾ 2 46 ⫾ 2* 95 ⫾ 6 66 ⫾ 4*

Isolated rat hepatocytes (106 cells/ml; 10 ml) were suspended in Krebs-Henseleit buffer (pH 7.4) containing 12.5 mM HEPES in continually rotating round-bottomed 50 ml flasks, under an atmosphere of 95% O2 and 5% CO2 for 30 min at 37°C. Then the ligands and their corresponding iron complexes were added to the hepatocytes and incubated under an atmosphere of 95% N2 and 5% CO2 for 90 min before reoxygenation with 1% O2, 94% N2, and 5% CO2 atmosphere. Before reoxygenation, the first assessment of the cells viability (90 min) was carried out by determining trypan blue exclusion by the hepatocytes. A 5 min premixed [ligand: Fe3⫹] complex was prepared before addition to hepatocytes. * Significantly different from hypoxia-reoxygenated cells, p ⬍ .05. Results are expressed as means ⫾ SD for three independent determinations.

complexes at inhibiting superoxide anion generation by the hypoxanthine/xanthine oxidase system was: 4-t-butylcatechol ⬎ catechin ⬎ quercetin ⬎ taxifolin ⬎ fisetin, rutin ⬎⬎ luteolin and kaempferol, whereas deferoxamine was ineffective. The [ligand:Fe2⫹] complexes of quercetin and catechin inhibited between 35–50% of NBT reduction caused by the hypoxanthine/xanthine oxidase-catalyzed superoxide radicalgenerating system. The superoxide-scavenging activity of the 2:1 [flavonoid:Fe3⫹] complexes was, however, greater than that of the 2:1 [flavonoid:Fe2⫹] complexes. The order of SRS activity of the [flavonoid:Fe3⫹] (2:1) complexes in the hypoxanthine/xanthine oxidase system was: quercetin ⬎ fisetin ⬎ 4-t-butylcatechol ⬎ catechin, taxifolin, rutin ⬎ kaempferol ⬎ luteolin, but not deferoxamine (Table 2). A 5-fold increase in SRS activity was observed for the 2:1 [quercetin:Fe3⫹] complex. Generally, there was an increase in SRS activity of the flavonoid when complexed with either Fe2⫹ or Fe3⫹ ions. On the other hand, the [deferoxamine:Fe2⫹] or the [deferoxamine: Fe3⫹] complexes had no SRS activity. There was also no SRS activity observed for uncomplexed Fe2⫹ and Fe3⫹ ions.

Superoxide radical-scavenging activity of flavonoid copper complexes The SRS activity of flavonoid:Cu2⫹ complexes was greater than uncomplexed flavonoids or histidine:Cu2⫹ complex (the plasma form) using the hypoxanthine/xanthine oxidase system for generating superoxide radicals. All tested flavonoids were less effective ligands for scavenging superoxide radicals than their Cu2⫹ complexes. The order of effectiveness of the Cu2⫹ complexes for SRS activity was: catechin, rutin, 4-t-butylcatechol, luteolin ⬎ taxifolin ⬎ quercetin, kaempferol ⬎ fisetin, deferoxamine. The order of the effectiveness of the uncomplexed ligands for SRS activity was: 4-t-butylcatechol, catechin ⬎ rutin, taxifolin, fisetin, kaempferol ⬎ quercetin, luteolin, whereas deferoxamine was inactive. Xanthine oxidase inhibition assay The flavonoid:Fe3⫹ complexes at the concentrations used did not inhibit the xanthine oxidase-catalyzed oxidation of xanthine to uric acid. Uncomplexed rutin, taxifolin, and catechin at 50 ␮M and their flavonoid:Fe3⫹ (10 ␮M:5 ␮M) complexes did not inhibit xanthine oxidase, whereas kaempferol at 10 ␮M inhibited the enzyme

Flavonoid iron complexes

247

Table 4. Spectra Shift and the Ratio of Peak Recovery for Flavonoids Interaction with Cu2⫹, Fe2⫹, or Fe3⫹ Cations Bathochromic shift in peak position and peak recovery band I (320–420 nm) Fe2⫹

Fe3⫹

Cu2⫹

Flavonoids

⫹ Fe2⫹* (nm)

EDTA* (recovery)

⫹ Fe3⫹* (nm)

⫹EDTA* (recovery)

⫹ Cu2⫹ (nm)

⫹ EDTA* (recovery)

Taxifolin Luteolin Rutin Quercetin Kaempferol Fisetin

0 0 0 15 2 50

100% 100% 100% 90% 73% 54%

0 22 18 27 0 0

100% 100% 100% 76% 87% 71%

0 53 5 35 50 35

100% 91% 90% 63% 58% 83%

Flavonoid solutions (25 ␮M) were prepared in a cuvette containing phosphate buffer (10 mM, pH 5.5) and the absorption spectra were recorded between 200 and 700 nm before and after the addition of 50 ␮M ferrous sulfate or cupric acetate followed by 1.25 mM EDTA. Alternatively, ferric ammonium sulfate was added to the flavonoids in phosphate buffer (10 mM, pH 7.4). * The spectra were recorded 5 min after each addition.

15%. However, uncomplexed fisetin, quercetin, and luteolin at 10 ␮M inhibited the xanthine oxidase activity by 75–95%. Despite the fact that fisetin, quercetin, luteolin, and kaempferol at 10 ␮M had an inhibitory effect on xanthine oxidase, they showed little or no effect (0 –10%) on xanthine oxidase activity when complexed with Fe3⫹ (5 ␮M). Cytotoxicity in isolated rat hepatocytes As shown in Table 3, superoxide dismutase prevented hypoxia-reoxygenation injury in isolated rat hepatocytes. Flavonoid:Fe3⫹ complexes were much more effective than the uncomplexed flavonoids at preventing reactive oxygen species-mediated hypoxia-reoxygenation hepatocyte injury. The order of effectiveness for inhibiting the reactive oxygen species-mediated cytotoxicity by the [ligand:Fe3⫹] complexes was: taxifolin, luteolin, fisetin ⬎ rutin ⬎ quercetin ⬎ catechin, t-butylcatechol, but not deferoxamine. The Fe3⫹ complexes of luteolin and fisetin were much more effective at inhibiting hypoxiareoxygenation hepatocyte injury than the uncomplexed species. By contrast, the uncomplexed flavonoids were much less effective and the order of effectiveness was: luteolin, fisetin, rutin ⬎ taxifolin ⬎ quercetin ⬎ catechin, t-butylcatechol. The [deferoxamine:Fe3⫹] complex did not protect the hepatocytes from reactive oxygen species-mediated hypoxia-reoxygenation injury. UV-VIS spectroscopy of flavonoid metal complexes The effects of Cu2⫹, Fe2⫹, and Fe3⫹ ion metals on the spectral characteristics of flavonoids are described in terms of shifts in band I (320 – 420 nm) and band II (250 –320 nm), which relate to B and A ring absorption, respectively. The interactions of Fe2⫹ ion with luteolin (348 nm), quercetin (371 nm), and fisetin (355 nm) at a 2:1 metal-

flavonoid ratio at pH 5.5 produced 22, 15, and 50 nm bathochromic shifts in the band I absorbance, respectively (Table 4). On addition of EDTA, the original spectra were fully recovered from the luteolin-Fe2⫹ complexes, whereas the recoveries of band I absorbance after EDTA addition to the fisetin and quercetin-Fe2⫹ complexes were 54 and 90%, respectively. The spectra of Fe2⫹ interaction with quercetin and luteolin were shown as examples in Fig. 1. There was no shift in band I and II absorbance for taxifolin (330 nm) and rutin (350 nm) on addition of Fe2⫹, however, there was a 2 nm red shift for kaempferol (360 nm) upon interaction with 2 molar equivalents of Fe2⫹ at pH 5.5. On EDTA addition to the [taxifolin:Fe2⫹] and [rutin:Fe2⫹] complexes, the extent of absorbance for taxifolin and rutin original peaks were fully recovered, whereas the recovery for the kaempferol band I absorbance after EDTA addition was 73% with an absorbance maxima at 355 nm under similar conditions. No significant shifts (shifts generally fell between 1–9 nm) at band II absorbance were observed for interactions between the flavonoids and Fe2⫹ ion. Catechin (278) lacks the band I absorbance and there was no shift in catechin band II absorbance when complexed with Fe2⫹ ion. However, upon EDTA addition the peak shifted to 272 nm. Briefly, Fe2⫹ oxidized kaempferol, quercetin, and fisetin about 10 –50%. Like the Fe2⫹-flavonoid interaction, EDTA fully restored the taxifolin, kaempferol, and catechin flavonoidFe3⫹ interaction absorbance peaks. Similarly, Fe3⫹ interaction with flavonoids resulted in fisetin, quercetin, and kaempferol oxidation (15–30%). Upon treatment with Cu2⫹, fisetin, quercetin, and kaempferol (20– 40%) were oxidized to a larger extent than rutin and luteolin (5–10%). Distribution coefficient measurement The distribution coefficient values of the flavonoids were measured in the octan-1-ol/MOPS buffer mixtures

248

M. Y. MORIDANI et al.

Scheme 1. The chemical structures of quercetin, fisetin, and kaempferol and their oxidized forms. R1, R2⫽OH, quercetin; R1⫽H, R2⫽OH, fisetin; R1⫽OH, R2⫽H, kaempferol.

DISCUSSION

Fig. 1. Effect of EDTA on the Fe2⫹ complex with (A) quercetin and (B) luteolin. Solutions were prepared (25 ␮M) in a cuvette containing phosphate buffer (10 mM, pH 5.5) and the absorption spectra were recorded between 200 –700 nm. Scans with 50 ␮M ferrous sulfate were recorded after 5 min and compared with flavonoids alone. Finally, scans with 1.25 mM EDTA were recorded.

at pH 7.4 (Table 1). The flavonoid water solubility at pH 7.4 in order of decreasing solubility was: rutin ⬎ catechin ⬎ taxifolin ⬎ luteolin ⬎ fisetin ⬎quercetin ⬎ kaempferol. The lowest distribution coefficient observed was 0.43 for rutin, likely due to the presence of a hydrophilic sugar moiety attached to the 3-OH group on the C-ring of the flavonoid. Rutin had the greatest solubility in water, whereas quercetin and kaempferol demonstrated the highest distribution coefficient values, 182 and 524, respectively, in octan-1-ol/water system at pH 7.4. Thus, kaempferol and quercetin demonstrated the greatest solubilities in a hydrophobic environment.

In the present work, we have selected a number of flavonoids along with one clinically available iron chelator deferoxamine to investigate their SRS properties and their abilities to prevent the hypoxic hepatocyte injury. We have found that superoxide dismutase or flavonoids prevented hypoxia-reoxygenation hepatocyte injury. Furthermore, flavonoid:Fe3⫹ complexes were much more effective than the uncomplexed flavonoids. At a physiological pH, most of the catecholic compounds form 2:1 complexes with Fe3⫹, whereas catechols with lower pKa values bind Fe3⫹ in 3:1 fashion [10,25,26]. We used a 2:1 ratio iron:flavonoid mixture for the spectrophotometric study only, to ensure that the observed spectral changes were mainly due to an interaction between iron and flavonoid. This was not possible unless an excess amount of iron was used. The reason for using such a ratio in this experiment was that, with a 2:1 flavonoid:iron ratio, several species such as a free flavonoid, 1:1 flavonoid:iron complex, and 2:1 flavonoid: iron complex could be formed in the solution, which otherwise could be avoided by using a 2:1 iron:flavonoid ratio at pH 5–7. However a 2:1 flavonoid:iron complex was used in other experiments in this report to avoid having a free iron species in the solution. With a 3:1 flavonoid:iron ratio, a large proportion of flavonoid would have been present in its free form in a solution with pH 5–7. This would have made it difficult to distinguish and assign the cytoprotective effect to the ligand: iron complex only. The changes in the flavonoid spectra that occurred when Fe2⫹, Cu2⫹, and Fe3⫹ were added to the flavonoids indicated the formation of a flavonoid:Fe3⫹ complex that was reversed by EDTA. Metal complexation probably occurred with the catechol of the B-ring and/or the 4-carbonyl/3-hydroxy group of the C-ring, and/or the 5-hydroxy group of the A-ring. Fisetin and kaempferol share a general chemical structure with quercetin (Scheme 1) except that kaempferol lacks a hydroxy group at position 3' of the B-ring and fisetin lacks a hydroxy group at

Flavonoid iron complexes

249

Scheme 2. The chemical structures of taxifolin, catechin, and luteolin. Taxifolin has a higher distribution coefficient value than catechin, due to an intramolecular hydrogen bonding between 3-hydroxy and 4-carbonyl groups of the taxifolin C-ring, but a lower distribution coefficient value than luteolin, due to the absence of a double bond on the C2⫽C3 of the C-ring.

position 5 of the A-ring. Rutin resembles quercetin except that the 3-OH group is blocked with a disaccharide glucorhamnoside. Luteolin lacks the free 3-OH group on the C-ring, whereas catechin and taxifolin lack the double bond (Scheme 2). Briefly, the major differences in the chemical structures of these flavonoids are in their C-rings. Spectroscopic studies indicated that quercetin and fisetin were capable of chelating Fe2⫹ and Cu2⫹ cations; however, on addition of EDTA the spectra reverted to the original quercetin and fisetin spectra. Although kaempferol has a similar chemical structure to quercetin and fisetin, it lacks the 3'-hydroxy group on the B-ring. Despite this, the kaempferol complexes with Cu2⫹, Fe2⫹, or Fe3⫹ also resulted in a partial oxidation of kaempferol, suggesting that the interaction of kaempferol with metals initially occurred with the 3-hydroxy and 4-carbonyl groups of the C-ring. This makes it possible for kaempferol to chelate and transfer electrons to the metal ions and further suggests that fisetin and quercetin complex metals via the C-ring rather than the B-ring, which results in their partial oxidation (Scheme 3). We suggest that the 3-OH and 4'-OH groups and C2⫽C3 double bond are essential for the partial oxidation of quercetin, fisetin, and kaempferol when complexed with a redox-cycling metal (Scheme 1). This does not occur with catechin, taxifolin, or luteolin because: (i) catechin and taxifolin lack the C2⫽C3 double bond and instead possess a 3-OH group; (ii) luteolin has a C2⫽C3 double bond and lacks the 3-OH group; and, (iii) all three compounds have 4'-OH group in their chemical structures. Rutin was also not oxidized when complexed with Fe2⫹, Cu2⫹, or Fe3⫹, presumably because the 3-OH

Scheme 3. Fe3⫹/Cu2⫹-mediated oxidation of kaempferol and its analogues, quercetin and fisetin. After tautomerization, the 3,4-dihydroxy groups of kaempferol chelate and transfer an electron to Fe3⫹ or Cu2⫹ ion.

group of rutin is blocked with a sugar moiety so that the existing C2⫽C3 double bond should behave similarly to that of luteolin. In one study, Brown et al. [21] described the UV spectra of the complexes formed between Cu2⫹ and luteolin, rutin, quercetin, and kaempferol and reported that Cu2⫹ was able to oxidize quercetin and kaempferol. They attributed this to quercetin, kaempferol, and fisetin having a lower oxidation potential than luteolin and rutin [27]. However, they did not acknowledge that Cu2⫹ could oxidize a small fraction (5–10%) of luteolin and rutin. A structural activity relationship between the distribution coefficient values of the flavonoids was observed with their chemical structures. The solubilities of the flavonoids in the octanol/water system at pH 7.4 in decreasing order of solubility was: rutin ⬎ catechin ⬎ taxifolin ⬎ luteolin ⬎ fisetin ⬎ quercetin ⬎ kaempferol.

250

M. Y. MORIDANI et al.

The absence of the C2⫽C3 double bond in catechin and taxifolin is likely largely responsible for their greater water solubilities. Catechin possesses a lower distribution coefficient value than that of taxifolin despite the presence of an extra oxygen on the C4 of taxifolin. This is probably due to an intramolecular hydrogen binding between 3-OH and 4-carbonyl groups of the taxifolin C-ring, which prevents efficient solvolysation of the 3-OH and 4-carbonyl groups (Scheme 2). Thus, the 3-OH group in catechin is more accessible to surrounding water molecules and in principle this will diminish catechin’s distribution coefficient value in comparison to that of taxifolin. Kaempferol has the highest distribution coefficient value among flavonoids that contain a C2⫽C3 double bond, as it is a phenolic compound, whereas quercetin, fisetin, and luteolin are B-ring catechols. The B-ring catechol-containing flavonoids have a lower pKa value than the corresponding B-ring phenol-containing flavonoids. Therefore, the B-ring catechol-containing flavonoids (quercetin, fisetin, and luteolin) dissociate more readily in an aqueous solution than the B-ring phenol-containing flavonoid (kaempferol). The hydrogen binding between 3-OH and 4-carbonyl group of fisetin and quercetin makes the compounds more lipophilic with a lower distribution coefficient than luteolin. The complexation of flavonoids with ferrous, ferric, and cupric ions was found to increase their superoxide dismutase activity in a xanthine oxidase/hypoxanthine superoxide-generating system. In the present study, it was also discovered that the iron complexes of flavonoids were much more effective at preventing hypoxia-reoxygenation injury when incubated with the isolated hepatocytes than the corresponding compounds alone. The cytoprotectiveness of the tested iron complexes was similar despite the differences in the partition coefficient of the flavonoids, and this in turn suggested that the iron complexes may protect the hepatocyte cell membrane extracellularly [16,28]. The cell impairment superoxide dismutase also protected the hepatocytes. On the other hand, deferoxamine and its metal complexes demonstrated almost no SRS activities and did not protect the hepatocytes. This is probably due to the strong sequestering property of deferoxamine towards ferrous and ferric ions. Deferoxamine is a hexadentate ligand while the flavonoids are bidentate ligands. Therefore, when deferoxamine binds the metals there would not be any coordinate site available or open to a readily displaceable ligand, such as water molecules or superoxide radicals. Previously, we reported that the iron complexes of catechols were much more effective than the uncomplexed ligands at preventing hypoxia-reoxygenation injury [16,29]. Recently we showed that catecholamine:iron complexes were more cytoprotective

against hypoxia-reoxygenation injury than uncomplexed catecholamines [28]. Previously it was suggested by Graf and coworkers [30] that the availability of an iron coordinate site that is open or occupied by a readily displaceable ligand such as water facilitates the reaction of iron chelates with superoxide radicals. The 2:1 [bidentate:Fe3⫹] complexes have at least two coordinate sites that are available to a superoxide radical, whereas in the 3:1 [catechol:Fe3⫹] and 1:1 [deferoxamine:Fe3⫹] complexes all the coordinate sites are occupied by the ligands, which in principle hinders a free electron transfer between the complex and superoxide radical. Dismutation of superoxide radicals by the [flavonoid:iron] complexes is depicted in Scheme 4, in which a 1:1 [luteolin:Fe3⫹] complex is used for simplicity. Generally, it is assumed that the ability of flavonoids to chelate metals is very important for their antioxidant activity; as examples, if iron or copper are still catalytically active, flavonoids can scavenge the superoxide radicals formed. Therefore, flavonoids would have a double synergistic action by activating the iron/copper center, while having the ability to scavenge the superoxide radicals. Afanas’ev et al. reported that a copper rutin chloride complex was much more effective as an antioxidant in preventing microsomal lipid peroxidation than rutin, which may be due to presence of the copper superoxide dismutating center [31,32], whereas the antioxidant activity of an iron rutin chloride complex was much lower and in some cases approached that of rutin [32]. Van Acker et al. also suggested that chelation with iron is the minor factor in the antioxidant activity of potent flavonoid antioxidants (i.e., preventing microsomal lipid peroxidation) but that it is of major importance in the antioxidant activity of flavonoids, which are less active [27,33]. The enzyme xanthine oxidase catalyzes the oxidation of hypoxanthine to xanthine and finally to uric acid [18]. During the reoxygenation of xanthine oxidase, the oxygen molecule is reduced to superoxide radical, which ultimately disproportionates to hydrogen peroxide. Consequently, xanthine oxidase is considered an important biological source of superoxide. One should note that flavonoid:iron complexes can not only scavenge free radicals but can also catalyze hydroxyl radical formation. Flavonoid can also inhibit xanthine oxidase enzyme; thereby, one may claim that the scavenging effect that was observed could be in reality an enzymatic inhibition effect rather than a superoxide-scavenging activity. This possibility was ruled out in this report at least for rutin, taxifolin, and catechin when they were tested for xanthine oxidase inhibition. In addition, we have clearly demonstrated in this report that flavonoid:iron complexes greatly diminished the hypoxia-reoxygenation injury to-

Flavonoid iron complexes

251

Scheme 4. Mechanisms of superoxide radical scavenging by flavonoid iron complexes [16].

wards isolated rat hepatocytes. In the light of these results, we conclude that the sum effect of flavonoid:iron complexes under our experimental conditions is due to scavenging free radicals rather than generating them or the inhibition of xanthine oxidase. In one study by Cos et al. [18], it was reported that taxifolin and catechin at 100 ␮M had little inhibitory effect on xanthine oxidase, whereas fisetin, luteolin, quercetin, and kaempferol had inhibitory effects with IC50 values of 4.3, 0.6, 2.6, and 1.1 ␮M, respectively, which is partly in good agreement with our findings. We have found that rutin, taxifolin, and catechin did not inhibit xanthine oxidase at 50 ␮M. Although fisetin, quercetin, luteolin, and kaempferol exhibited inhibition on xanthine oxidase at 10 ␮M, surprisingly their iron complexes [flavonoid:Fe3⫹] (10 ␮M:5 ␮M) did not inhibit xanthine oxidase, indicating that the uncomplexed phenolic groups were required for xanthine oxidase inhibition. Thus, the SRS activity values measured for [flavonoid:Fe3⫹] complexes were not the result of any inhibitory effect on xanthine oxidase activity. Previously, we showed that oxypurinol, an xanthine oxidase inhibitor, did not protect against the hepatocyte hypoxia:reoxygenation injury [15] and could be explained if mitochondria rather than xanthine oxidase

were the source of the oxygen radicals. It should also be noted that the differences in flavonoid inhibitory effects on xanthine oxidase that were previously published may result from the different source of xanthine oxidase used—Escherichia coli [18] vs. milk. CONCLUSION

Our studies suggest that iron complexes of flavonoids readily scavenged superoxide radicals generated extracellularly. In the absence of iron, a lower degree of cytoprotection against hypoxic injury was achieved when a flavonoid was incubated alone. We have found that the Fe3⫹ complexes of flavonoids were much more effective than the uncomplexed flavonoids in protecting isolated hepatocytes against hypoxia-reoxygenation injury. This conclusion is valid regardless of what complex form is present (2:1 or 1:1 flavonoid complex) or even when a small fraction of flavonoid is present in its free form in the 2:1 flavonoid:iron mixture. Furthermore, the 2:1 flavonoid-metal complexes of Cu2⫹, Fe2⫹, or Fe3⫹ were much more effective than the parent compounds in scavenging superoxide radicals generated by a xanthine oxidase/hypoxanthine system. However, the [deferoxamine:Fe3⫹] complex did not scavenge superox-

252

M. Y. MORIDANI et al.

ide radicals, suggesting that the initial step for SRS activity requires the formation of a redox complex between Fe3⫹ and the flavonoids. The clinically used iron-chelating drug deferoxamine has proved highly successful at preventing ischemia or reperfusion injuries in animal experiments [34]. Flavonoids in contrast to deferoxamine may have an additional therapeutic role in scavenging superoxide radicals once they chelate iron and should be tested as a novel therapeutic strategy for ischemia/reperfusion injury. Acknowledgements — The authors wish to thank the Natural Sciences and Engineering Research Council of Canada for providing a research grant to support this work. The authors also are grateful to Mr. Ford Barker for the isolated rat hepatocytes preparation.

[16]

[17]

[18]

[19]

[20] REFERENCES [1] Michelson, A. M. Clinical use of superoxide dismutase and possible pharmacological approaches. In: Autor, A. P., ed. Pathology of oxygen. New York: Academic Press; 1982:277–302. [2] Benov, L.; Fridovich, I. Functional significance of Cu, Zn-SOD in Escherichia coli. Arch. Biochem. Biophys. 327:249 –253; 1996. [3] Wada, K.; Fujibayashi, Y.; Tajima, N.; Yokoyama, A. Cu-ATSM, an intracellular-accessible superoxide dismutase (SOD)-like copper complex: evaluation in an ischemia/reperfusion injury model. Biol. Pharm. Bull. 17:701–704; 1994. [4] Darr, D. J.; Yanni, S.; Pinnell, S. R. Protection of Chinese hamster ovary cells from paraquat-mediated cytotoxicity by a low-molecular-weight mimic of superoxide dismutase (DF-Mn). Free Radic. Biol. Med. 4:357–363; 1988. [5] Weiss, R. H.; Flickinger, A. G.; Rivers, W. J.; Hardy, M. M.; Aston, K. W.; Ryan, U. S.; Riley, D. P. Evaluation of activity of putative superoxide dismutase mimics. Direct analysis by stopped flow kinetics. J. Biol. Chem. 268:23049 –23054; 1993. [6] Archibald, F. S.; Fridovich, I. The scavenging of superoxide radical by manganous complexes: in vitro. Arch. Biochem. Biophys. 214:452– 463; 1982. [7] Wada, K.; Fujibayashi, Y.; Yokoyama, A. Copper(II) [2,3-butanedionebis(N4-methylthiosemicarbazone)], a stable superoxide dismutase-like copper complex with high membrane penetrability. Arch. Biochem. Biophys. 310:1–5; 1994. [8] Iuliano, L.; Pedersen, J. Z.; Chilselli, A.; Pratico, D.; Rotilio, G.; Violi, F. Mechanism of reaction of a suggested superoxide dismutase mimic, Fe(II)-N, N, N', N'-tetrabis(2-pyridylmethyl)ethylenediamine. Arch. Biochem. Biophys. 293:153–157; 1992. [9] Nagano, T.; Hirano, T.; Hirobe, M. Superoxide dismutase mimics based on iron in vivo. J. Biol. Chem. 264:9243–9249; 1989. [10] Avdeef, A.; Sofen, S. R.; Bregante, T. L.; Raymond, K. N. Coordination chemistry of microbial iron transport compounds. 9. Stability constants for catechol models of enterobactin. J. Am. Chem. Soc. 100:5362–5370; 1978. [11] Neilands, J. B. Microbial iron transport compounds (sidrochromes). In: Eichorn, G., ed. Inorganic biochemistry. New York: Elsevier Science Inc.; 1973:167–202. [12] O’Brien, I. G.; Cox, G. B.; Gibson, F. Enterochelin hydrolysis and iron metabolism in Escherichia coli. Biochem. Biophys. Acta 237:537–549; 1971. [13] Cooper, S. R.; McArdle, J. V.; Raymond, K. N. Siderophore electrochemistry: relation to intracellular iron release mechanism. Proc. Natl. Acad. Sci. USA 75:3551–3554; 1978. [14] Niknahad, H.; Khan, S.; O’Brien, P. J. Hepatocyte injury resulting from the inhibition of mitochondrial respiration at low oxygen concentrations involves reductive stress and oxygen activation. Chem. Biol. Interact. 98:27– 44; 1995. [15] Khan, S.; O’Brien, P. J. Modulating hypoxia-induced hepatocyte

[21]

[22]

[23]

[24] [25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

injury by affecting intracellular redox state. Biochim. Biophys. Acta 1269:153–161; 1995. Zhao, Z. S.; Khan, S.; O’Brien, P. J. Catecholic iron complexes as cytoprotective superoxide scavengers against hypoxia: reoxygenation injury in isolated hepatocytes. Biochem. Pharmacol. 56: 825– 830; 1998. Hertog, M. G.; Hollman, P. C.; Katan, M. B.; Kromhout, D. Intake of potentially anticarcinogenic flavonoids and their determinants in adults in the Netherlands. Nutr. Cancer 20:21–29; 1993. Cos, P.; Ying, L.; Calomme, M.; Hu, J. P.; Cimanga, K.; Poel, B. V.; Pieters, L.; Vlietinck, A. J.; Berghe, V. Structure-activity relationship and classification of flavonoids as inhibitors of xanthine oxidase and superoxide scavengers. J. Nat. Prod. 61:71–76; 1998. Vanden Bergh, D. A. R. M.; Haemers, A.; Vleitinck, A. J. Chapter 17. In: Colegate, A. M.; Molyneux, R. J., eds. Bioactive natural products: detection, isolation and structural determination. London: CRC Press; 1993:405– 440. Bors, W.; Heller, W.; Michel, C.; Saran, M. Flavonoids as antioxidants: determination of radical-scavenging efficiencies. Methods Enzymol. 186:343–355; 1990. Brown, J. E.; Khodr, H.; Hider, R. C.; Rice-Evans, C. A. Structural dependence of flavonoid interactions with Cu2⫹ ions: implications for their antioxidant properties. Biochem. J. 330:1173– 1178; 1998. Perrissoud, D.; Testa, B. Inhibiting or potentiating effects of flavonoids on carbon tetrachloride-induced toxicity in isolated rat hepatocytes. Arzneim. Forsch. 26:1249 –1253; 1986. Cotelle, N.; Bernier, J. L.; Catteau, J. P.; Pommery, J.; Wallet, J. C.; Gaydou, E. M. Antioxidant properties of hydroxy-flavones. Free Radic. Biol. Med. 20:35– 43; 1996. Mold´eus, P.; H¨ogberg, J.; Orrenius, S. Isolation and use of liver cells. Methods Enzymol. 52:60 –71; 1978. Kawabata, T.; Schepkin, V.; Karamaki, N.; Phadke, R. S.; Packer, L. Iron coordination by catechol derivative antioxidants. Biochem. Pharmacol. 51:1569 –1577; 1996. Krishna, C. M.; Liebmann, J. E.; Kaufman, D.; DeGraff, W.; Hahn, S. M.; McMurry, T.; Mitchell, J. B.; Russo, A. The catecholic metal sequestering agent 1, 2-dihydroxybenzene-3,5-disulfonate confers protection against oxidative cell damage. Arch. Biochem. Biophys. 294:98 –106; 1992. van Acker, S. A.; van den Berg, D. J.; Tromp, M. N.; Griffioen, D. H.; van Bennekom, W. P.; van der Vijgh, W. J.; Bast, A. Structural aspects of antioxidant activity of flavonoids. Free Radic. Biol. Med. 20:331–342; 1996. Siraki, A. G.; Smythies, J.; O’Brien, P. J. Superoxide radical scavenging and attenuation of hypoxia-reoxygenation injury by neurotransmitters ferric complexes in isolated rat hepatocytes. Neurosci. Lett. 296:37– 40; 2000. Moridani, M. Y.; O’Brien, P. J. Iron complexes of deferiprone and dietary plant catechol as cytoprotective superoxide radical scavengers. Biochem. Pharmacol. 62:1579 –1585; 2001. Graf, E.; Mahoney, J. R.; Bryant, R. G.; Eaton, J. Iron catechol (o-dihydroxybenzene). Part I. Equilibria and kinetics of complex formation in aqueous acid solution. J. Biol. Chem. 259:3620 – 3624; 1984. Afanas’ev, I. B.; Ostrachovich, E. A.; Korkina, L. G. Effect of rutin and its copper complex on superoxide formation and lipid peroxidation in rat liver microsomes. FEBS Lett. 425:256 –258; 1998. Ostrakhovitch, E. A.; Afanas’ev, I. B.; Oxidative stress in rheumatoid arthritis leukocytes: suppression by rutin and other antioxidants and chelators. Biochem. Pharmacol. 62:743–746; 2001. van Acker, S. A.; van Balen, G. P.; van den Berg, D. J.; Bast, A.; van der Vijgh, W. J. Influence of iron chelation on the antioxidant activity of flavonoids. Biochem. Pharmacol. 56:935–943; 1998. Hershko, C. Control of disease by selective iron depletion: a novel therapeutic strategy utility iron chelators. Baillieres Clin. Haematol. 7:965–1000; 1994.

Flavonoid iron complexes ABBREVIATIONS

DETAPAC— diethylenetriaminepentaacetic acid DMSO— dimethyl sulfoxide EDTA— ethylenediaminetetraacetic acid HEPES— 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

MOPS—3-(N-morpholino)propanesulfonic acid NBT—nitro blue tetrazolium ROS—reactive oxygen species SRS—superoxide radical scavenging Tris—Tris(hydroxymethyl)aminomethane UV-VIS— ultraviolet visible XO—xanthine oxidase

253