Inhibition of phthalocyanine-sensitized photohemolysis of human erythrocytes by polyphenolic antioxidants: description of quantitative structure–activity relationships

Cancer Letters 157 (2000) 39±44

www.elsevier.com/locate/canlet

Inhibition of phthalocyanine-sensitized photohemolysis of human erythrocytes by polyphenolic antioxidants: description of quantitative structure±activity relationships AudroneÇ MarozieneÇ, Regina KliukieneÇ, Jonas SÏarlauskas, Narimantas CÏeÇnas* Institute of Biochemistry, Mokslininku 12, Vilnius 2600, Lithuania Received 8 March 2000; received in revised form 9 May 2000; accepted 10 May 2000

Abstract Polyphenolic antioxidants protected against Al-phthalocyanine tetrasulfonate-sensitized photohemolysis of human erythrocytes. A quantitative structure±activity relationship has been obtained to describe the protective effects of di- and trihydroxybenzenes: log cI50 (mM) ˆ (1.8620 ^ 1.5565) 1 (3.6366 ^ 2.8245) E71 (V) 2 (0.4034 ^ 0.0765) log P (r 2 ˆ 0:8367), where cI50 represents the concentrations of compounds for the 2-fold increase in the lag-phase of hemolysis, E71 represents the compound single-electron oxidation potential, and P represents the octanol/water partition coef®cient. The cI50 for quercetin and taxifolin were close, and cI50 for morin, kaempferol and hesperetin were lower than might be predicted by this equation. The protection from hemolysis by azide, a quencher of singlet oxygen ( 1O2) was accompanied by increase in cI50 of polyphenols, indicating that azide and polyphenols competed for the same damaging species, 1O2. These ®ndings point out to two factors, determining the protective ef®ciency of polyphenols against 1O2, namely, ease of electron donation and lipophilicity. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Polyphenol; Antioxidant; Singlet oxygen; Erythrocytes; Photosensitization

1. Introduction In photodynamic therapy of cancer, the cytotoxic or/and therapeutic effects of photosensitization by hematoporphyrins and metallophthalocyanines arise mainly from generation of singlet oxygen ( 1O2), and, possibly, from generation of superoxide and hydroxyl radicals [1]. Various antioxidants and radioprotectors, including thiols, disul®des, b-carotene and ¯avonoids [2±8] which act as scavengers of 1O2, may protect from the damaging effect of photosensitized irradia* Corresponding author. Tel.: 1370-2-729042; fax: 1370-2729196. Ï eÇnas). E-mail address: [email protected] (N. C

tion. It has long been suggested that antioxidants could play a useful role in minimizing adverse effects of photodynamic reactions, e.g. skin photosensitization, damage to normal tissues, and erythrocyte lysis. In addition to their antioxidant activity arising from scavenging of superoxide, hydroxyl and peroxyl radicals [9,10], ¯avonoids and phenolic compounds physically quench and chemically react with 1O2 [10±13]. However, the examples of their protection against 1O2 in biological systems are not numerous. Myricetin and other ¯avonoids, as well as the related polyphenol tannic acid, protected plasmid DNA from chemically generated 1O2 [14]. Quercetin and rutin prevented phthalocyanine- and hematoporphyrinsensitized photohemolysis of erythrocytes [7,8].

0304-3835/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0304-383 5(00)00469-9

40

A. MarozieneÇ et al. / Cancer Letters 157 (2000) 39±44

These data do not allow to de®ne the relationship between polyphenol structure and protective ef®ciency. In an initial attempt to obtain a quantitative structure±activity relationship between the physicochemical parameters and protective ef®ciency of ¯avonoid and phenolic antioxidants against the cytotoxicity of 1 O2, we have examined their activity against Alphthalocyanine tetrasulfonate (Al-PcS4)-sensitized photohemolysis of human erythrocytes.

2. Materials and methods Freshly prepared suspensions of erythrocytes from healthy patients obtained from Vilnius Blood Transfusion Center, were washed twice by centrifugation, resuspended in 0.01 M K-phosphate (pH 7.0), containing 0.137 M NaCl, 0.0027 M KCl, 10 mM glucose and 1 mM EDTA, and stored at 48C. The stock suspension of erythrocytes was used within 7±10 days after preparation. For photosensitization experiments,

erythrocytes were diluted by the same solution to ®nal concentration, 2.6 £ 10 6/ml; afterwards 10 mM Al-PcS4 and polyphenolic antioxidant compounds (Fig. 1) were added to 2 ml of stirred suspension in dark, 30 min prior to irradiation. The irradiation was performed at 25 ^ 0.28C, under continuous stirring, in a 1 cm inner diameter thermostated glass cell. A light source was a 250 W Philips 7748S lamp with a cut-off ®lter, transmitting only light with l . 590 nm. The ¯uence rate at the surface of the cell was 25 W/m 2, as measured by IMO-2N radiometer (Russia). During continuous irradiation, aliquots of erythrocytes (0.1 ml) were taken and diluted by 20 times, the decrease in their apparent absorbance at 740 nm, which accompanied cell disruption [7], was measured. Complete hemolysis was established after the addition of 40 mg/ ml digitonin. Photosensitized oxidation of antioxidants (50 mM) by 10 mM Al-PcS4 was followed recording their absorbance spectra (240±500 nm) by Hitachi 557 spectrophotometer, at 258C. Reactions were carried out in the absence of erythrocytes, the reference cell contained

Fig. 1. Structural formules of polyphenolic antioxidants.

A. MarozieneÇ et al. / Cancer Letters 157 (2000) 39±44

10 mM Al-PcS4. The kinetic data at 3±4 speci®c wavelengths were analyzed for each compound according to a single-exponent ®t, giving suf®ciently close (^10%) pseudo®rst order rate constants (kox). The most pronounced absorbance changes were observed at wavelengths presented in Table 1. Al-PcS4 was obtained from Porphyrin Products (Logan, UT), while all other compounds came from Sigma or Aldrich. Bis(1,6-hexanediolgallate) (compound IIf, Fig. 1) and di-3,5-t-butylcatechol were synthesized as described previously [15,16]. The octanol/water partition coef®cients for polyphenolic compounds were calculated using ACD LogP (version 1) software, a generous gift of Advanced Chemistry Development Inc. (Toronto, Canada). The calculation of kinetic data and regression analysis were performed using Statistica (version 4.3) software (StatSoft Inc., 1993).

3. Results In accordance with previous data [7,17], the photosensitized erythrocyte hemolysis was characterized by lag-phase, de®ned as the intercept of the tangent to the

41

turning point of the curve with the time axis (Fig. 2A). The lag-phase was calculated from a linear approximation of kinetics of erythrocyte lysis between 10 and 80%. Catalase or superoxide dismutase (50 mg/ml) increased the lag-phase by #5%. On the other hand, azide markedly protected from erythrocyte hemolysis (Fig. 2A). After 24 h incubation of erythrocytes with iron chelator desferrioxamine (1 mM), and their subsequent washing by centrifugation before photolysis, the lag-phase was increased by 30±40% (data not shown). The same effect was observed in the presence of desferrioxamine and catalase. Polyphenolic compounds protected from erythrocyte hemolysis in a concentration-dependent manner (Fig. 2A,B). Their relative protective ef®ciency, expressed as concentrations of compounds for the 2fold increase in the lag-phase (cI50), was calculated from lag-phase vs. polyphenol concentration plots (Fig. 2B). The values of cI50 are given in Table 1. In the presence of 3 mM azide, quercetin further protected erythrocytes, however, its cI50 increased from 170 ^ 30 to 300 ^ 30 mM (Fig. 2B). Analogously, 3 mM azide increased cI50 of di-3,5-t-butylcatechol from 50 ^ 10 to 100 ^ 30 mM (data not shown).

Table 1 The cI50, kox (the wavelengths corresponding to the most pronounced absorbance changes given in parentheses), E71, and P No.

Compound

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Quercetin Morin Kaempferol Taxifolin Hesperetin Naringenin Gallic acid Methylgallate Ethylgallate Butylgallate Octylgallate Bis(1,6-hexanediolgallate) Caffeic acid Butylhydroquinone di-3,5-t-Butylcatechol

a b c d e

cI50 (mM)

kox (min 21)

E71 (V)

Log P

170 ^ 30 90 ^ 10 130 ^ 20 850 ^ 150 240 ^ 20 470 ^ 50 $ 2000 1200 ^ 300 1000 ^ 200 680 ^ 50 28 ^ 5 250 ^ 30 $ 2000 450 ^ 50 50 ^ 10

0.1 (330) a 0.09 (400) 0.03 (340) 0.025 (330) 0.008 (335) 0.015 (335) 0.055 (260) 0.065 (270) 0.1 (270) 0.08 (270) 0.175 (270) 0.05 (270) 0.06 (320) 0.33 (260) 0.033 (260)

0.33 b 0.60 c 0.75 c 0.50 b 0.72 b 2.59 0.91 0.56 b 2.07 3.13 5.26 4.44 0.54 b 0.46 d 0.48 e

2.74 1.97 2.69 1.22 2.3 1.54

2.47 2.33 4.26

Numbers in parentheses are in nm. Determined in Ref. [10]. Calculated in Ref. [10]. Assumed to be equal to E71 of methylhydro-quinone [18]. Calculated from E71 value of catechol, 0.53 V, assuming that the introduction of alkyl group into the benzene ring decreases E71 by 25 mV.

42

A. MarozieneÇ et al. / Cancer Letters 157 (2000) 39±44

Both ¯avonoids and polihydroxybenzenes were photooxidized by Al-PcS4 with pseudo®rst order rate constants (kox), given in Table 1. Azide (3 mM) inhibited photooxidation of querce-tin, morin and methylgallate by 90%. On the contrary, superoxide dismutase and catalase (50 mg/ml) inhibited the reaction by #5%. The identi®cation of polyphenol photooxidation products was beyond the scope of the present work. Table 1 also contains the compiled values of single-electron oxidation potentials of polyphenols (E71) [10,18], and their octanol/water partition coef®cients, P, calculated in the present work. 4. Discussion The protective effect of azide (Fig. 2A) and the absence of marked protection by catalase and superoxide dismutase imply that the photogenerated 1O2 is the main causative factor of Al-PcS4-sensitized erythrocyte lysis. The increase of cI50 for quercetin in the presence of azide (Fig. 2B) indicates that the protection of erythrocytes by quercetin also arises mainly from its interaction with 1O2, i.e. that the quercetin and azide compete for the same damaging species. The Stern±Volmer analysis in the presence of two 1O2 acceptors, Q1 and Q2, yields the following

expression

F 0 =Fˆ 11k q…1† ‰Q1 Št 1 kq…2† ‰Q2 Št

…1†

where F 0 and F are quantum yield in the absence and presence of acceptors, kq(1) and kq(2) are the sums of rate constants of physical quenching and chemical reaction between 1O2 and acceptors Q1 and Q2, respectively, and t is the 1O2 lifetime [11]. One may assume, that the lag-phase of photohemolysis is inversely related to the quantum yield of photoproduction of 1 O2. Then, in the absence of Q2, the cI50 for acceptor Q1 is equal to 1/kq(1)t . At the ®xed concentration of Q2, the cI50 for Q1 should increase, becoming equal to 1/kq(1)t 1 kq(2)[Q2]/kq(1). The inhibition of polyphenol photooxidation by azide and the absence of protective effects of superoxide dismutase and catalase, imply that 1O2 is the main causative agent of polyphenol photooxidation. We have found that the correlation between log cI50 of polyphenols (Table 1) and the rate of their photooxidation, log kox, did not exist (r 2 ˆ 0:0039). In view of importance of lipophilicity in biological activity of polyphenols [19], we have introduced log P as a second parameter for correlation. The resulting rough correlation (r 2 ˆ 0:5170) demonstrated the increase in polyphenol protective ef®ciency upon the increase in their lipophilicity, and uncertain relation-

Fig. 2. Polyphenol protection against continuous photosensitized erythrocyte hemolysis. (A) Time course of photosensitized erythrocyte hemolysis. Additions: none (1), 100 mM butylgallate (2), 200 mM butylgallate (3), 3 mM sodium azide (4). (B) The dependence of lagphase of erythrocyte photosensitized hemolysis on quercetin concentration in the absence (1) and in the presence of 3 mM sodium azide (2).

A. MarozieneÇ et al. / Cancer Letters 157 (2000) 39±44

ship between log cI50 and log kox. The analogous separate analysis of protective effects of ¯avonoids (compounds 1±6, Table 1) and hydroxybenzenes (7± 15), also did not reveal the link between log cI50 and log kox. Since the 1O2 quenching rate constants by polyphenols may increase with the ease of their electron donation, i.e. with decrease in their potential of semiquinone/catechol or phenoxyl radical/phenol redox couple, E71 [10,12,13], we have tried to correlate log cI50 and E71 of compounds (Table 1). It was assumed, that E71 for all derivatives of gallic acid (compounds 7,9±12) was suf®ciently close or equal to E71 of ethylgallate, 0.56 V [10]. However, the correlation between log cI50 and E71 was absent (r 2 ˆ 0:0031). The introduction of log P as a second parameter resulted in a rough correlation (r 2 ˆ 0:5244) with uncertain dependence of log cI50 upon E71. On the other hand, log cI50 of tri- and dihydroxybenzenes (compounds 7±15, Table 1) clearly increased upon the increase in their E71 values (r 2 ˆ 0:8367) (Fig. 3)

43

kaempferol, taken from Ref. [10] (Table 1), might be overestimated. Indeed, a large difference between the values of E71 for quercetin, morin, and kaempferol (Table 1), strongly contrasts with their suf®ciently close voltammetric midpoint potentials given in the same Ref. [10], namely, 0.31 V (quercetin), 0.386 V (morin), and 0.416 V (kaempferol). However, the analysis of the redox potential calculation methods, used in Ref. [10], is beyond the scope of the present work. Nevertheless, our data indicate that a quantitative structure±activity relationship exists for the protective action of a large group of polyphenolic antioxidants (Fig. 3): (i) the protection increases with the increase in electron-donating properties of compound which is most probably related to the ef®ciency of 1O2 quenching [10±13]; (ii) the protection increases upon the increase in compound lipophilicity, i.e. their ability to penetrate into erythrocytes [20]. The more thorough studies on this regularity involving various tumour cell lines, are underway.

log cI50 …mM† ˆ …1:8620 ^ 1:5565† 1 …3:6366 ^ 2:8245†E17 …V† 2 …0:4034 ^ 0:0765†log P

…2†

It is evident, that cI50 for quercetin and taxifolin are close to those predicted by Eq. (2), however, the protective ef®ciency of morin, kaempferol and hesperetin is markedly higher (Fig. 3). At present, we cannot precisely explain this phenomenon. It is possible that another mechanism of erythrocyte protection from photolysis takes place simultaneously with 1O2 quenching by polyphenols, namely, polyphenols and ¯avonoids may also chelate intraerythrocyte iron ions, which are released from hemoglobin under oxidizing conditions [20]. The data of the present work indicate that the iron ion chelator desferrioxamine can slow down the photosensitized hemolysis, which is in line with previous ®ndings on the protection by desferrioxamine against oxidative damage to erythrocytes [21]. Thus, it is possible that polyphenols and ¯avonoids comprise separate series of compounds, differing in their ability to chelate iron ions [20,22]. This may result in deviation from the expected dependence of protective ef®ciency upon E71 and log P (Fig. 3). Another possible explanation is that the calculated values of E71 for morin and

Fig. 3. The dependence of protection ef®ciency of polyphenols and ¯avonoids on their single-electron oxidation potential (E71) and octanol/water partition coef®cient (P). The numbers of compounds are taken from Table 1. The regression describing the protective ef®ciency of di- and trihydroxybenzenes (compounds 7±15, Table 1), was plotted according to a multiparameter Eq. (2), retaining the proportion between coef®cients at variables E71 and log P.

44

A. MarozieneÇ et al. / Cancer Letters 157 (2000) 39±44

Acknowledgements This work was supported in part by Lithuanian Science and Studies Foundation Program `Laser Photosensitization in Tumour Therapy' and by EC grant No. CT961004. References [1] B.W. Henderson, T.J. Dougherty, How does photodynamic therapy work? Photochem. Photobiol. 55 (1992) 145±157. [2] J. Dillon, J.C. Kennedy, R.H. Pottier, J.E. Roberts, In vitro and in vivo protection against phototoxic side effects of photodynamic therapy by radioprotective agents WR-2721 and WR77913, Photochem. Photobiol. 48 (1988) 235±238. [3] D.A. Bellnier, T.J. Dougherty, Protection of murine foot tissue and transplantable tumor against photofrin II mediated photodynamic sensitization with WR-2721, J. Photochem. Photobiol. B 4 (1989) 219±225. [4] C.P. Sigestad, V.H. Fingar, T.J. Wieman, R.D. Lindberg, Chemical modi®cation of normal tissue damage induced by photodynamic therapy, Br. J. Cancer 74 (1996) S89±S92. [5] R. KliukieneÇ, A. MarozieneÇ, H. Nivinskas, N. CÏeÇnas, V. KirvelieneÇ, B. Juodka, The protective effects of dihydrolipoamide and glutathione against photodynamic damage by Alphthalocyanine tetrasulfonate, Biochem. Mol. Biol. Int. 41 (1997) 707±713. [6] M.M. Mathews-Roth, M.A. Pathak, T.B. Fitzpatrick, L.C. Harber, F.H. Kass, Beta carothene therapy for erythropoietic protoporphyria and other diseases, Arch. Dermatol. 113 (1977) 1229±1232. [7] Y. Sorata, U. Takahama, M. Kimura, Cooperation of quercetin with ascorbate in the protection of photosensitized lysis of human erythrocytes in the presence of hematoporphyrin, Photochem. Photobiol. 48 (1988) 195±199. [8] E. Ben-Hur, I. Rosenthal, Y. Granot, Inhibition of phtalocyanine-sensitized photohemolysis of human erythrocytes by quercetin, Photochem. Photobiol. 57 (1993) 984±988. [9] C.A. Rice-Evans, N.J. Miller, G. Paganga, Structure-antioxidant activity relationships of ¯avonoids and phenolic acids, Free Radic. Biol. Med. 20 (1996) 933±956. [10] S.V. Jovanovic, S. Steenken, M.G. Simic, Y. Hara, Antioxidant properties of ¯avonoids: reduction potentials and electron transfer reactions of ¯avonoid radicals, in: C.A. Rice-Evans,

[11]

[12] [13] [14]

[15] [16]

[17]

[18] [19]

[20]

[21]

[22]

L. Packer (Eds.), Flavonoids in Health and Disease, Marcel Dekker Inc, New York, 1998, pp. 137±161. C. Tournaire, S. Croux, M.-T. Maurette, I. Beck, M. Hocquaux, A.M. Braun, E. Oliveros, Antioxidant activity of ¯avonoids: ef®ciency of singlet oxygen ( 1Dg) quenching, J. Photochem. Photobiol. B. 19 (1993) 205±212. A.P. Darmanian, W.J. Jenks, P. Jardon, Charge-transfer quenching of singlet oxygen O2( 1Dg) by amines and aromatic hydrocarbons, J. Phys. Chem. A 102 (1998) 7420±7426. R.V. Bensasson, A. Jossang, A. Zahir, B. Bodo, E.J. Land, Redox regulation of tumor cell toxicity by ¯avones from Lethedon tannaensis, Free Radic. Biol. Med. 27 (1999) 95±99. T.P.A. Devasagayam, M. Subramanian, B.B. Singh, R. Ramanathan, N.P. Das, Protection of plasmid pBR322 DNA by ¯avonoids against single-strand breaks induced by singlet molecular oxygen, J. Photochem. Photobiol. B 30 (1995) 97±103. W.G. Christiansen, Some derivatives of gallic acid and pyrogallol, J. Am. Chem. Soc. 48 (1926) 1360±1363. N.B. Nevrekar, S.R. Savardekar, T.S. Pandit, N.A. Kudav, Use of urea and its N-substituted tert-butyl derivatives in tert-butylation of aromatic compounds, Chem. Indust. (London) 5 (1983) 206±207. R. KersÏieneÇ, A. MarozieneÇ, R. KliukieneÇ, ZÏ. AnusevicÏius, J. Ï eÇnas, Modulation of erythrocyte photoDidzÏiapetrieneÇ, N. C hemolysis rate by glutathione reductase inactivating alkylating agents, Biochem. Mol. Biol. Int. 45 (1998) 709±716. P. Wardman, Reduction potentials of one-electron couples involving free radicals in aqueous solutions, J. Phys. Chem. Ref. Data 18 (1989) 1637±1755. E. SergedieneÇ, K. Jonsson, H. Szymusiak, B. Tyrakowska, Ï eÇnas, Prooxidant toxicity of polyphenoI.M.C.M. Rietjens, N. C lic antioxidants to HL-60 cells: description of quantitative structure±activity relationships, FEBS Lett. 462 (1999) 392±396. M. Ferrali, C. Signorini, B. Caciotti, L. Sugherini, L. Ciccoli, D. Giachetti, M. Comporti, Protection against oxidative damage of erythrocyte membrane by the ¯avonoid quercetin and its relation to iron chelating activity, FEBS Lett. 416 (1997) 123±129. M. Ferrali, C. Signorini, L. Ciccoli, M. Comporti, Iron release and membrane damage in erythrocytes exposed to oxidizing agents, phenylhydrazine, divicine and isouramil, Biochem. J. 285 (1992) 295±301. S.A. van Acker, G.P. van Balen, D.J. van den Berg, A. Bast, W.J. van der Vijgh, In¯uence of iron chelation on the antioxidant activity of ¯avonoids, Biochem. Pharmacol. 56 (1998) 935±943.