Role of Oxidative Stress Ausˇra Nemeikaite˙-Cˇ E˙Niene˙ A, Egle˙ Sergediene˙ B, Henrikas Nivinskasb and Narimantas Cˇ E˙Nasb* a Institute of Immunology, Mole˙Tu˛ Pl

Home , Carminic acid

Cytotoxicity of Natural Hydroxyanthraquinones: Role of Oxidative Stress Ausˇra Nemeikaite˙-Cˇ e˙niene˙ a, Egle˙ Sergediene˙ b, Henrikas Nivinskasb and Narimantas Cˇ e˙nasb* a Institute of Immunology, Mole˙tu˛ Pl. 29, Vilnius 2021, Lithuania b Institute of Biochemistry, Mokslininku˛ St. 12, Vilnius 2600, Lithuania. Fax: 370-2-729196. E-mail: [email protected] * Author for correspondence and reprint requests Z. Naturforsch. 57c, 822Ð827 (2002); received April 29/June 3, 2002 Hydroxyanthraquinones, Cytotoxicity, Oxidative Stress In order to assess the role of oxidative stress in the cytotoxicity of natural hydroxyanthraquinones, we compared rhein, emodin, danthron, chrysophanol, and carminic acid, and a series of model quinones with available values of single-electron reduction midpoint potential 1 at pH 7.0 (E 7), with respect to their reactivity in the single-electron enzymatic reduction, and their mammalian cell toxicity. The toxicity of model quinones to the bovine leukemia virus-transformed lamb kidney fibroblasts (line FLK), and HL-60, a human promyelocytic 1 leukemia cell line, increased with an increase in their E 7. A close parallelism was found between the reactivity of hydroxyanthraquinones and model quinones with single-electron transferring flavoenzymes ferredoxin: NADP+ reductase and NADPH: cytochrome P-450 reductase, and their cytotoxicity. This points to the importance of oxidative stress in the toxicity of hydroxyanthraquinones in these cell lines, which was further evidenced by the protective effects of desferrioxamine and the antioxidant N,NЈ-diphenyl-p-phenylene diamine, by the potentiating effects of 1,3-bis-(2-chloroethyl)-1-nitrosourea, and an increase in lipid peroxidation.

Introduction somerase II (Müller et al., 1996) and protein ki- Natural 1,8-dihydroxyanthraquinones rhein, nase C (Chan et al., 1993), and the inhibition of danthron, emodin (Fig. 1) are used in pharmacy as glycolysis and mitochondrial functions (Floridi laxatives, other hydroxylated anthraquinones are et al., 1989, 1990) are considered as potential used in industry as textile dyes or food colourants, mechanisms of the cytotoxicity of hydroxyanthra- e.g. carminic acid (Fig. 1). Although hydroxylated quinones. The relative importance of these mecha- anthraquinones were considered as potential anti- nisms is insufficently understood so far, although tumour agents (Koyama et al., 2001), their intake it may be important in view of the diverse fields may be associated with increased colon cancer and of hydroxy-anthraquinone use, as well as of some adenoma risk (Schörkhuber et al., 1998). These similarity of their structure to other important qui- compounds also possess immunosuppressive nones, anthracycline antibiotics adriamycin and (Huang et al., 1992), photosensitizing (Rahimipour et al., 2001), mutagenic and cytotoxic properties (Mian et al., 1991; Müller et al., 1996). The enzymatic redox cycling and subsequent oxidative stress (Mian et al., 1991; Bironaite˙ and Öllinger 1997, Ka˚gedal et al., 1999), the inhibition of topoi-

1 Abbreviations:E7, single-electron reduction midpoint potential at pH 7.0; cL50, the concentration of compound for 50% cell survival; FNR, ferredoxin:NADP+ reductase; P-450R, NADPH: cytochrome P-450 reductase; Q, quinone; kcat, catalytic constant; kkcat/Km, bimolecular rate constant; DPPD, N,NЈ-diphenyl-p-phenylene di- Fig. 1. The structures of hydroxyanthraquinones used in amine; BCNU, 1,3-bis-(2-chloroethyl)-1-nitrosourea. this study.

0939Ð5075/2002/0900Ð0822 $ 06.00 ” 2002 Verlag der Zeitschrift für Naturforschung, Tübingen · www.znaturforsch.com · D A. Nemeikaite˙-Cˇ e˙niene˙ et al. · Cytotoxicity of Natural Hydroxyanthraquinones 823 daunorubicin, whose mechanisms of cytotoxicity described previously (Pueyo and Gomez-Moreno, remain the subject of considerable controversy 1991) and was a generous gift of Dr. M. Martinez- (Gewirtz, 1999). Julvez and Professor C. Gomez-Moreno (Zara- Frequently, the aerobic cytoxicity of quinones or goza University, Spain). The enzyme concentra- ε Ð1 Ð1 nitroaromatic compounds, the other important tion was determined using 459 = 9.4 mm cm . group of prooxidants, increases with an increase in The activity of FNR using 1 mm ferricyanide as their single-electron reduction midpoint potential electron acceptor (concentration of NADPH, 1 ∆ µ µ Ð1 Ð1 at pH 7.0 (E 7) with a relationship log cL50/ 200 m) was equal to 330 mol mg min .The ∆ 1 ϳ Ð1 E 7 Ð10 V , where cL50 is the concentration catalytic constant (kcat) and the bimolecular rate of compound for 50% cell survival (Guissany constant (kcat/Km) of quinone reduction corre- et al., 1990; O’Brien, 1991). This points to the oxi- spond to the reciprocal intercepts and slopes of dative stress as to the main factor of cytotoxicity, plots [E]/v vs. 1/[Q], where [E] is the enzyme con- since, as a rule, the rates of single-electron reduc- centration, and [Q] is the concentration of qui- tion of quinones or nitroaromatics by flavoen- none. kcat is the number of NADPH molecules ox- 1 zymes initiating redox cycling increase with E 7 of idized by the single active center of an enzyme per oxidants (Butler and Hoey, 1993; Anusevicˇius second. The rate of oxygen consumption during et al., 1997). However, the studies of natural enzymatic reactions was monitored using a Clark hydroxyanthraquinones in this direction are ham- electrode. 1 pered by the absence of E 7 values for a number The culture of bovine leukemia virus-trans- of important compounds (Wardman, 1989; Rath formed lamb kidney fibroblasts (line FLK) was et al., 1996). grown and maintained in Eagle’s medium supple- In this work, we have compared the cytotoxicity mented with 10% fetal bovine serum at 37 ∞Cas of several natural hydroxyanthraquinones (Fig. 1) described previously (Nemeikaite˙ and Cˇ e˙nas, with their reactivity towards single-electron trans- 1993). HL-60, a human promyelocytic leukemia ferring flavoenzymes NADPH:cytochrome P-450 cell line, was cultured in RPMI-1640 medium with reductase (EC 1.6.2.4) and ferredoxin: NADP+ re- 10% fetal serum (Dicˇkancaite˙ et al., 1997). In the ductase (EC 1.18.1.2). Taken together with the cytotoxicity experiments, cells (3.0 ¥ 104/ml, FLK, analogous data of the model quinone compounds or 3.0 ¥ 105/ml, HL-60) were grown in the pres- 1 with available E 7 values, our results demonstrate ence of various amounts of quinones for 24 h, and that the oxidative stress may be the main factor of counted using a hematocytometer with viability the cytotoxicity of hydroxyanthraquinones. determined by exclusion of Trypan blue. Before the count, FLK cells were trypsinized. After the 24 h incubation of the cells with quinones, lipid Materials and Methods peroxidation was monitored according to the for- Hydroxyanthraquinones (Fig. 1) and other rea- mation of malondialdehyde, using the thiobarbi- gents were obtained from Sigma or Aldrich, and turic acid test (Ramanathan et al., 1994). used as received. The kinetic measurements were carried out spectrophotometrically using a Hi- Results and Discussion tachi-557 spectrophotometer in 0.1 m K-phosphate buffer (pH 7.0) containing 1 mm EDTAat25∞C. The single-electron reduction of quinones by NADPH: cytochrome P-450 reductase (P-450R) flavoenzymes ferredoxin: NADP+ reductase and from pig liver was prepared as described pre- NADPH:cytochrome P-450 reductase is exten- viously (Yasukochi and Masters, 1976), the en- sively documented (Butler and Hoey, 1993; Anu- ε zyme concentration was determined using 460 = sevicˇius et al., 1997;). Therefore, we used FNR and 22 mmÐ1cmÐ1. The activity of P-450R using 50 µm P-450R as model systems for the evaluation of the cytochrome c as an electron acceptor (NADPH redox cycling activity of natural hydroxyanthraqu- concentration, 100 µm) was 77 µmol mgÐ1 minÐ1. inones. Their reduction by FNR and P-450R was The reduction of cytochrome c was monitored analogous to the single-electron reactions of other ∆ε Ð1 Ð1 + using 550 =20mm cm . Ferredoxin: NADP low-potential quinones under aerobic conditions, reductase (FNR) from Anabaena was prepared as i.e., quinones oxidized excess NADPH, with the 824 A. Nemeikaite˙-Cˇ e˙niene˙ et al. · Cytotoxicity of Natural Hydroxyanthraquinones consumption of a stoichiometric amount of O2 per mole of NADPH. For example, in the presence of 150 µm NADPH and 20 nM P-450R, 10 µm 5-hydroxy-, 5,8-dihydroxy-, or 2-methyl-1,4-naphthoquinone, or 9,10-phenanthrene quinone oxidized 50Ð70 µm NADPH in 1.5 min, whereas 10 µm danthron, rhein, emodin, chrysophanol, or carminic acid oxidized 30Ð60 µm NADPH in 50 min. FNR catalyzed the reduction of added cytochrome c (50 µm) by hydroxyanthraquinones, at the rate of 180Ð190% NADPH oxidation rate. The reduction of cytochrome c was inhibited by 80Ð90% by 30 µg/ml superoxide dismutase. Taken together, our data show that natural hydroxyanthraquinones undergo P-450R- and FNR-catalyzed redox cycling with the formation of superox- Fig. 2. The dependence of bimolecular rate constants of reduction (kcat/Km) of model quinones and danthron by ide (reactions (1,2), where Q is quinone, and Q.- NADPH:cytochrome P-450 reductase (A) and ferre- is semiquinone): doxin:NADP+ reductase (B) on their single-electron re- 1 duction midpoint potentials (E 7). The numbers of com- P-450R, FNR pounds are taken from Table I. 2Q + NADPH 444445 2Q.- + NADP+ +H+ (1) .- 444445 .- Q +O2 Q+O2 (2) line. Table I shows the concentrations of hydroxy-

We were unable to determine the kcat values of anthraquinones for 50% cell survival (cL50), and the reaction, since at their concentrations above the cL50 values for model compounds (partly de- 30 µm, danthron, chrysophanol and carminic acid termined in the present study, and partly taken inhibited P-450R, and emodin, chrysophanol and from our previous works (Nemeikaite˙ and Cˇ e˙nas, danthron inhibited FNR. In the other cases, the 1993; Dicˇkancaite˙ et al., 1997)). It is evident, that reaction rates followed the linear dependence on the cytotoxicity of model quinone compounds and the hydroxyanthraquinone concentrations up to danthron towards both cell lines increases with an µ 1 150Ð200 m. The molecular reduction rate con- increase in their E 7 (Fig. 3A). The linear depen- 1 2 stants (kcat/Km) of hydroxyanthraquinones by dences of log cL50 vs. E 7 are characterized by r = FNR and P-450R are given in Table I, together 0.9592 (FLK), and r2 = 0.9040 (HL-60). For the with kcat/Km of model quinones (partly deter- quantitative analysis of the cytotoxicity of hy- mined in the present study, and partly taken from droxyanthraquinones with unavailable values of 1 our previous work (Anusevicˇius et al., 1997)). E 7, we applied the approach used in our previous Among hydroxyanthraquinones used in this study, work for the description of the proxidant cytotox- 1 only the E 7 value of danthron has been available icity of nitroaromatic explosives with unavailable 1 ˇ so far (Table I). It is evident that in the reactions E 7 (Ce˙nas et al., 2001), namely the geometrical with P-450R and FNR, danthron follows the same mean of the reactivity of hydroxyanthraquinones 1 parabolic log kcat/Km dependence on E 7 as other and model quinone compounds in FNR- and quinones (Fig. 2). The kcat/Km of rhein, carminic P-450R-catalyzed reactions (log kcat/Km (FNR) + acid, chrysophanol, and emodin are close to or be- log kcat/Km (P-450R)/2 (Table I) as a correlation low the kcat/Km of danthron and above the kcat/ parameter. It is evident, that the low cytotoxicity 1 Km of 2-hydroxy-1,4-naphthoquinone with E 7 = of hydroxyanthraquinones is consistent with their Ð0.41 V (Table I). low redox cycling ability (Fig. 3B). For all the For the cytotoxicity experiments, we used the examined compounds, the linear dependences of bovine leukemia virus-transformed lamb kidney log cL50 vs. (log kcat/Km (FNR) + log kcat/Km fibroblast line FLK used in our previous studies (P-450R)/2 are characterized by r2 = 0.9492 (FLK), (Nemeikaite˙ and Cˇ e˙nas, 1993; Cˇ e˙nas et al., 2001), and r2 = 0.8838 (HL-60). The antioxidant N,NЈ- and HL-60, a human promyelocytic leukemia cell diphenyl-p-phenylene diamine (DPPD) (Miccadei A. Nemeikaite˙-Cˇ e˙niene˙ et al. · Cytotoxicity of Natural Hydroxyanthraquinones 825

1 Table I. Midpoint potentials of single-electron reduction of quinones (E 7), their bimolecular rate constants of + reduction (kcat/Km) by ferredoxin:NADP reductase (FNR) and NADPH:cytochrome P-450 reductase (P-450R), and quinone concentrations for 50% survival of FLK and HL-60 cells during a 24-h incubation (cL50).

1 a Ð1 Ð1 µ No. Compound E 7 [V] kcat/Km [M s ]cL50 [ m] a) FNRb b) P-450R a) FLK b) Hl-60

1. 5-Hydroxy-1,4-naphthoquinone Ð0.09 1.2 ð 0.1 ¥ 106 4.6 ð 0.3 ¥ 107 0.50 ð 0.10 1.70 ð 0.20c 2. 5,8-Dihydroxy-1,4-naphthoquinone Ð0.11 9.0 ð 1.0 ¥ 105 7.0 ð 0.6 ¥ 107 0.11 ð 0.02d 1.05 ð 0.20c 3. 9,10-Phenanthrene quinone Ð0.12 2.2 ð 0.1 ¥ 106 3.1 ð 0.4 ¥ 107 0.70 ð 0.08 3.00 ð 0.27 4. 1,4-Naphthoquinone Ð0.15 6.0 ð 0.7 ¥ 105 3.5 ð 0.2 ¥ 107 1.60 ð 0.10 1.13 ð 0.15 5. 2-Methyl-5-hydroxy-1,4-naphthoquinone Ð0.16 6.7 ð 0.7 ¥ 105 1.9 ð 0.1 ¥ 107 1.50 ð 0.10 3.60 ð 0.30c 6. 2-Methyl-1,4-naphthoquinone Ð0.20 4.6 ð 0.3 ¥ 105 1.1 ð 0.1 ¥ 107 3.50 ð 0.30 55.0 ð 12.0 7. Tetramethyl-1,4-benzoquinone Ð0.26 2.0 ð 0.2 ¥ 105 5.0 ð 0.2 ¥ 106 16.0 ð 3.0d 61.0 ð 11.0 8. Danthron Ð0.325e 2.5 ð 0.2 ¥ 105 1.7 ð 0.1 ¥ 105 120 ð 15.0 150 ð 12.0 9. 2-Hydroxy-1,4-naphthoquinone Ð0.41 6.7 ð 0.4 ¥ 103 4.4 ð 0.3 ¥ 104 700 ð 100 1000 ð 120c 10. Rhein Ð 3.3 ð 0.2 ¥ 105 6.7 ð 0.4 ¥ 104 150 ð 20.0 100 ð 10.0 11. Emodin Ð 1.3 ð 0.1 ¥ 104 1.3 ð 0.1 ¥ 105 155 ð 15.0 120 ð 11.0 12. Chrysophanol Ð 1.5 ð 0.1 ¥ 104 2.5 ð 0.3 ¥ 105 150 ð 15.0 178 ð 19.0 13. Carminic acid Ð 2.0 ð 0.2 ¥ 104 4.8 ð 0.3 ¥ 104 350 ð 52.0 720 ð 81.0 a From Wardman (1989). b Bimolecular rate constants of reduction of compounds 1Ð7,9 taken from Anusevicˇius et al. (1997). c From Dicˇkancaite˙ et al. (1997). d From Nemeikaite˙ and Cˇ e˙nas (1993). e From Rath et al. (1996). et al., 1988) and the iron-chelating agent desferri- oxidant flavoenzyme glutathione reductase (EC oxamine, the latter preventing the Fenton reac- 1.6.4.2) and depletes intracellular reduced gluta- tion, gave significant protection from the toxicity thione (Öllinger and Brunmark, 1991), potenti- of danthron to FLK cells (Fig. 4A), and rhein and ated the toxicity of danthron (Fig. 4B). The other emodin to HL-60 cells (data not shown). This is evidence of oxidative stress was an increase in the analogous to the protective effects of DPPD and intracellular content of the lipid peroxidation pro- desferrioxamine against the cytotoxicity of model duct malondialdehyde. After 24 h incubation with quinones and rhein to rat hepatocytes (Öllinger 150 µm danthron or 500 µm carminic acid that re- and Brunmark, 1991; Bironaite˙ and Öllinger, sulted in the death of 80Ð85% FLK cells, the 1997), and model quinones to HL-60 cells (Dicˇk- content of malondialdehyde was equal to 2.1 ð ancaite˙ et al., 1997). 1,3-Bis-(2-chloroethyl)-1-nit- 0.4 nmol/106 cells, whereas in the untreated cells it rosourea (BCNU), which inactivates the anti- was equal to 0.5 ð 0.1 nmol/106 cells.

Fig. 3. (A) The dependence of cytotoxicity of model quinones and danthron to FLK (a) and HL-60 cells (b) on the single-electron reduction midpoint potential of quinones 1 (E 7). (B) The dependence of cytotoxicity of model quinones and natural hydroxyanthraquinones to FLK (a) and HL-60 cells (b) on their reactivitity towards ferredoxin:NADP+ reductase and NADPH: cytochrome P-450 reductase ((log kcat/Km (FNR) + log kcat/Km (P-450R)/2). The numbers of compounds are taken from Table I. 826 A. Nemeikaite˙-Cˇ e˙niene˙ et al. · Cytotoxicity of Natural Hydroxyanthraquinones

desferrioxamine, the comparison with model quinones shows that daunorubicin is much more toxic than one may expect from its low reduction potential, Ð0.34 V (Dicˇkancaite˙ et al., 1997). It indicates that in addition to the oxidative stress (Powis, 1989), the other mechanisms of daunorubicin cytotoxicity, e.g. topoisomerase inhibition (Gewirtz, 1999), may be equally or even more important. In the present study, the correlations between the cytotoxicity and the redox cycling ability of natural hydroxyanthraquinones and model quinones with Fig. 4. (A) The protecting effects of DPPD (2 µm) and a wide range of E1 values (Fig. 3A,B), supported µ µ 7 desferrioxamine (300 m) in the cytotoxicity of 100 m by the effects of the antioxidants and BCNU danthron to FLK cells. Additions: danthron (1), danthron + DPPD (2), danthron + desferrioxamine (3), (Fig. 4A,B), show that the toxicity of hydroxy- danthron + DPPD + desferrioxamine. (B) The potentiat- anthraquinones to FLK and HL-60 cells is mainly ing effects of 25 µm BCNU in the cytoxicity of danthron determined by their redox cycling reactions with to FLK cells. Additions: 50 µm danthron (1), 50 µm danthron + BCNU (2), 100 µm danthron (3), 100 µm formation of superoxide (reactions (1,2), and, sub- danthron + BCNU (4). Cell viability in control experi- sequently, other reactive oxygen species like H2O2 ments, 96 ð 3%; DPPD, desferrioxamine and BCNU de- and hydroxyl radical, OH. (Powis, 1989; Hippeli creased cell viability by 1Ð3%, n =3Ð4. and Elstner, 1997). It appears that the rate constants of the single-electron enzymatic reduction Although the involvement of oxidative stress in of natural hydroxyanthraquinones as well as of the cytotoxicity of rhein and other dihydroxy- other groups of prooxidant compounds with un- 1 anthraquinones has been demonstrated in previ- available E 7 values may serve as a useful tool for ous studies (Mian et al. 1991; Bironaite˙ and the quantitative description of their cytotoxity Öllinger 1997, Ka˚gedal et al., 1999), their redox with the involvement of oxidative stress. cycling properties and cytotoxicity were not com- 1 Acknowledgements pared to other quinones with available E 7 values. This leads to the uncertainty of the relative impor- This work was supported in part by EC grant tance of oxidative stress and other potential No. CT961004. We thank Professor C. Gomez- cytotoxicity mechanisms. For example, although Moreno and Dr. M. Martinez-Julvez (Zaragoza daunorubicin-induced apoptosis and necrosis in University, Spain) for their generous gift of HL-60 cells are partly prevented by DPPD and ferredoxin:NADP+ reductase. A. Nemeikaite˙-Cˇ e˙niene˙ et al. · Cytotoxicity of Natural Hydroxyanthraquinones 827

Anusevicˇius Zˇ ., Martinez-Julvez M., Genzor C. G., Niv- Koyama J., Morita I., Ogata M., Mukainaka T., Tokuda inskas H., Gomez-Moreno C. and Cˇ e˙nas N. (1997), H. and Nishino H. (2001), Inhibitory effects of Electron transfer reactions of Anabaena PCC 7119 anthraquinones and bianthraquinones on Epstein- ferredoxin:NADP+ reductase with nonphysiological Barr virus activation. Cancer Lett. 170,15Ð18. oxidants. Biochim. Biophys. Acta 1320, 247Ð255. Main M., Fratta, D., Rainaldi G., Simi S., Mriani T., Be- Bironaite˙ D. and Öllinger K. (1997), The hepatotoxicity netti D. and Gervasi P. G. (1991), Superoxide anion of rhein involves impairment of mitochondrial func- production and toxicity in V79 cells of six hydroxy- tions. Chem.-Biol. Interact. 103,35Ð50. anthraquinones. Anticancer Res. 11, 1071Ð1076. Butler J. and Hoey B. M. (1993), The one-electron re- Miccadei S., Kyle M. E., Gilford D. and Farber J. L. duction potential of several substrates can be related (1988), Oxidative cell injury in the killing of cultured to their reduction rates by cytochrome P-450 reduc- hepatocytes by allyl alcohol. Arch. Biochem. Biophys. tase. Biochim. Biophys. Acta 1161,73Ð78. 265, 311Ð320. Cˇ e˙nas N., Nemeikaite˙-Cˇ e˙niene˙ A., Sergediene˙ E., Nivin- Müller S. O., Eckert I., Lutz W. K. and Stopper H. skas H., Anusevicˇius Zˇ . and Sˇ arlauskas J. (2001), (1996), Genotoxicity of the laxative drug components Quantitative structure-activity relationships in enzy- emodin, aloe-emodin and danthron in mammalian matic single-electron reduction of nitroaromatic ex- cells: topoisomerase II mediated? Mutat. Res. 371, plosives: implications for their cytotoxicity. Biochim. 165Ð173. Biophys. Acta 1528,31Ð38. Nemeikaite˙ A. and Cˇ e˙nas N. (1993), The changes of pro- Chan T. C. K., Chang C.-J., Koonchanok N. M. and oxidant and antioxidant enzyme activities in bovine Geahlen R. L. (1993), Selective inhibition of the leukemia virus transformed cells: their influence on growth of ras-transformed human bronchial epithelial quinone cytotoxicity. FEBS Lett. 326,65Ð68. cells by emodin, a protein-tyrosine kinase inhibitor. O’Brien P. J. (1991), Molecular mechanisms of quinone Biochem. Biophys. Res. Commun. 193, 1152Ð1158. cytotoxicity. Chem.-Biol. Interact. 80,1Ð41. Dicˇkancaite˙ E., Cˇ e˙nas N., Kalvelyte˙ A. and Serapiniene˙ Öllinger K. and Brunmark A. (1991), Effect of hydroxy N. (1997), Toxicity of daunorubicin and naphthoqui- substituent position on 1,4-naphtho-quinone toxicity nones to HL-60 cells: an involvement of oxidative to rat hepatocytes. J. Biol. Chem. 266, 21496Ð21503. stress. Biochem. Mol. Biol. Int. 41, 987Ð994. Powis G. (1989), Free radical formation by antitumour Floridi A., Castiglione S. and Bianchi C. (1989), Sites of quinones. Free Rad. Biol. Med. 6,63Ð101. inhibition of mitochondrial electron transport by Pueyo J. J. and Gomez-Moreno C. (1991), Purification of rhein. Biochem. Pharmacol. 38, 743Ð751. ferredoxin-NADP+ reductase, flavodoxin and fer- Floridi A., Castiglione S., Bianchi C. and Mancini A. reeoxin from a single batch of the cyanobacterium (1990), Effect of rhein on the glucose metabolism of Anabaena PCC7119. Prep. Biochem. 21, 191Ð204. Ehrlich ascites tumor cells. Biochem. Pharmacol. 40, Ramanathan R., Das N. P. and Tan C. H. (1994), Effects 217Ð222. of γ-linoleic acid, flavonoids, and vitamins on cytotox- Gewirtz D. A. (1999), A critical evaluation of the mech- icity and lipid peroxidation. Free Rad. Biol. & Med. anisms of action proposed for the antitumour effects 16,43Ð48. of the anthracycline antibiotics adriamycin and daun- Rath M. C., Pal H. and Mukherjee T. (1996), Pulse- orubicin. Biochem. Pharmacol. 57, 727Ð741. radiolytic one-electron reduction of anthraquinone Guissany A., Henry Y., Lougmani N. and Hickel B. and chloro-anthraquinone in aqueous-isopropanol-ac- (1990), Kinetic studies of four types of nitrohetero- etone mixed solvent. Radiat. Phys. Chem. 47, 221Ð cyclic radicals by pulse radiolysis: correlation of phar- 227. macological properties to decay rates. Free Rad. Biol. Schörkhuber M., Richter M., Dutter A., Sontag G. and Med. 8, 173Ð189. Marian B. (1998), Effect of anthraquinone-laxatives Hippeli S. and Elstner E. F. (1997), OH-radical-type re- on the proliferation and urokinase secretion of nor- active oxygen species: a short review on the mecha- mal, premalignant and malignant colonic epithelian nisms of OH-radical- and peroxynitrite toxicity. Z. cells. Eur. J. Cancer 34, 1091Ð1098. Naturforsch. 52c, 555Ð563. Wardman P. (1989), Reduction potentials of one-elec- Huang H. C., Chang J. H., Tung S. F., Wu R. T., Foegh tron couples involving free radicals in aqueous solu- M. L. and Chu S. H. (1992), Immunosuppressive ef- tion. J. Phys. Chem. Ref. Data 18, 1637Ð1755. fect of emodin, a free radical generator. Eur. J. Phar- Yasukochi Y. and Masters B. S. S. (1976), Some proper- macol. 211, 359Ð364. ties of a detergent-solubilized NADPH-cytochrome c Ka˚gedal K., Bironaite˙ D. and Öllinger K. (1999), Anthra- (cytochrome P-450) reductase purified by biospecific quinone cytotoxicity and apoptosis in primary cultures affinity chromatography. J. Biol. Chem. 251, 5337Ð of rat hepatocytes. Free Rad. Res. 31, 419Ð428. 5344.