The Pyridoindole Antioxidant Stobadine Prevents Alloxan-Induced Lipid Peroxidation by Inhibiting its

Propagation Milan Stefek and Zuzana Trnkova

Institute of Experimental Pharmacology, Slovak Academy of Sciences, Dubravska cesta, 842 I6 Bratislava. Slovak Republic

(Received May 9. 1995: Accepted August 3 . 1995)

Ahs t rc~c~ t ; Under in pifro conditions, the pyridoindole stobadine inhibited alloxan-induced lipid peroxidation in a model biological membrane with the efficacy comparable with that of the standard Trolox. Intermediary alloxan radicals and hydroxyl radicals were not directly involved in the process of lipid peroxidation, however, the presence of iron chelate was a necessary prerequisite. Since stobadine did not affect the kinetics of alloxan redox-cycling in the presence of GSH, we suggest that the protective action of stobadine against the alloxan-induced lipid peroxidation was mediated predominantly by its ability to quench peroxyl radicals. inhibiting thus the propagation stage of the oxidative damage. The results also indicate that toxic effects of alloxan may well be mediated by mechanism(s) not involving hydroxyl radicals.

The pyridoindole stobadine (( -)-cis-2.8-dimethyl-2,3,4,4a,5, 9b-hexahydro-lH-pyrido[4,3b]indole) proved efficient in protecting isolated rat hearts against functional damage after their exposure to periods of ischaemia followed by re- perfusion (Benes & Stolc 1989). The cardioprotective effect of stobadine was suggested to arise from its antioxidant properties (Horakova et a/. 1994) based on the pyridoindole structure of the molecule (fig. 1); the stobadine molecule traps free radicals, giving rise to the resonance stabilized pyridoindolyl radical (Steenken et a/. 1992). The free radical scavenging ability of stobadine was well documented under in vitro conditions (Stasko et ul. 1990; Stefek & Benes 1991; Horakova r t r d . 1992; Kagan et a/. 1993). To test the anti- oxidant properties of stobadine in vivo we recently used al- loxan-induced hyperglycaemia as a model of free radical pathology in intact animals. Stobadine was found to be an efficient inhibitor of the diabetogenic effect of alloxan (Ste- fek & Trtikova 1995).

This work was carried out to study possible molecular mechanisms involved in the observed ability of stobadine to protecl the animal against the diabetogenic action of allox- an. I n a series of in i i t ro experiments we investigated the effect of stobadine on the process of lipid peroxidation in- itiated in a model biological membrane by the reaction sys- tem or alloxan-GSH-Fe-EDTA. A preliminary account of this work was presented elsewhere (Stefek & Trnkova 1994).

Chemistry and Biochemistry. Czechoslovak Academy of Sciences. Prague, Czech Republic. and was used as a dihydrochloride. Alloxiin monohydrate, 1,1.3,3-tetra-ethoxypropane. 2-keto-4-methiolbutyric acid (KMBA), and diethylenetriaminepentaacetic acid (DETAPAC) were purchased from Sigma Chemical Co., St. Louis, MO, U.S.A. Glutathione (GSHj was obtained from Merck, Darmstadt. Ger- many and Trolox@ from Aldrich Chemical Co., Milwaukee. WI. U.S.A. Thiobarbituric acid was from Fluka AG, Buchs. Switzer- land. Other chemicals were obtained from local commercial sources and were of analytical grade.

Lipid peroxidation asstry. Liver microsomes were prepared from male Wistar rats (220-270 g) as described elsewhere (Stefek 1993). The liver preparations were heat-denatured for 10 min. in a boiling water bath and then resuspended in 1.15'X KC1. The incubation mixture consisted of liver microsomes (1.2-1.8 mg protein/ml) in 12.5 mM potassium phosphate buffer, pH 7.4. containing 5 m M GSH. 25 pM Fe(NH&(S0,j2, 30 pM EDTA. and stobadine or the standard antioxidants (Trolox@ or KMBA) as indicated below. The reaction was started by the addition of alloxan giving a final con- centration of 0. I mM. Incubations were conducted aerobically at 37" for different periods of time up to 90 min. The formation of thiobarbituric acid-reactive products was measured by the method of Buege & Aust (1978). The malondialdehyde formed in the in- cubation mixture by decomposition of standard tetraethoxypropane was used to prepare the calibration curve. Control experimenls showed that stobadine, Trolox or KMBA, at the concentrations used, did not interfere with the above mentioned assay of thiobarbi- turic acid-reactive material. The lag phase in thiobarbituric acid- reactive material accumulation was determined from A ( 5 3 2 mi) versus time curves by extrapolation of the most rapid linear phase

Materials and Methods

C/itwiiwls. Stobadine. (-)-cis-2,8-diinethyI-2,3,4,4a.5.9b-liex~iliydro- 1 H-p);rido[4,3b]indole was synthesize at the Institute of Organic

Author for correspondence: Milan Stefek. Institute of Experimental Pharmacology. Slovak Academy of Sciences, Dubravska cesta, 842 16 Bratislava. Slovak Republic (fax +42-7-375 928).

k Fig. I. Chemical structure of stobadine, (-)-cis-2.8-dimethyl-2,3. 4.4a,S,9b-hexahydro-l H-pyrido[4,3,b] indole.

78 MILAN STEFBK A N D ZUZANA TRNKOVA

4 0-

2 0-

TahL 1

Alloxan-induced lipid peroxidation of microsomal membrane in vi- tro. Characterization of the svstem.

Malondialdehyde formed Experimental conditions (nmolimg protein)

Complete system" 43.2i.0.3 (8) - alloxan 5.4i.0.5 (3) - alloxaniGSH 0.2+0.1 (3) - FeiEDTA 4.020.3 (3)

microsomes 0

'I Boiled microsomes (1.2-1.8 mgiml), alloxan (0.1 mM), GSH (5 mM), Fe'+ (25 pM), EDTA (30 pM) in 12.5 mM potassium phos- phate buffer, pH 7.5. Time of incubation, 60 min.

Results are mean valuest-S.E.M., with number of experiments in parentheses.

of increasing absorbance to the baseline level and taken as the x- intercept. Protein content was measured by the method of Geiger & Bessnian (1972).

Sprcrral chnges during ulloxan reduction. Spectral changes during the interaction between alloxan and GSH were monitored on Hew- lett Packard 8452A diode array spectrophotometer. A volume of 5.8 ml of GSH solution in 50 mM potassium phosphate buffer, pH 7.5, containing 50 pM DETAPAC, was divided equally between two cuvettes. The spectra between 240 and 370 nm were recorded in 2 min. intervals for 20 min. starting immediately after the addition of alloxan (dissolved in ice-cold isotonic KCI) into the sample cuvette: a corresponding volume of isotonic KCI was added to the reference cuvette. When appropriate, stobadine was added in equal concen- trations (up to 0.5 mM) to both cuvettes before starting the reac- tion. The reactions were conducted at ambient temperature.

De/errnincztion of' H,02. For the assay of H202 production, 0. I mM alloxan and 5 mM GSH in 12.5 m M potassium phosphate buffer, pH 7.5, were incubated in the absence or presence of stobadine (up to 2.5 mM) at 37" in the final volume of 1.0 ml. The reaction was started by addition of alloxan. In the selected time intervals the level of H 2 0 1 generated was assayed by the Fe(SCN)3 method of Hildebrandt & Roots (1975). Standard curves were prepared by using H 2 0 2 standardized spectrophotometrically a t 240 nm ( ~ = 4 3 . 6 M- ' cn- l ) . The presence of alloxan, GSH, or stobadine at the con- centrations used in the experiments did not affect the slopes of the standard lines.

Results

Lipid prroxidution. Rat liver microsomes were used as a model biological mem- brane to study the effect of stobadine on alloxan-induced lipid peroxidation. The liver preparations were heat-de- natured to eliminate specific enzyme activities of micro- somes. Lipid peroxidation was induced by the system of alloxaniGSHIFeiEDTA and evaluated by determination of the thiobarbituric acid-reactive material accumulated. The presence of all components of the model system studied was required for maximal activity (table 1). The key role of al- loxan redox cycling in the process was demonstrated by a profound attenuation of the thiobarbituric acid-reactive material formation in the absence of alloxan and alloxani glutathione. The presence of chelated iron was a prerequi- site for the toxic effect of alloxan. Increasing concentrations

2. 51

0 . 0 . 5~ 0 0 2 0 4 0 6 0 80

Time (min) Fig. 2. Time course of alloxan-induced lipid peroxidation in heat- denatured rat liver microsomal membrane. Effect of stobadine. Stobadine: 0 (-0-), 50 pM (-m-), Alloxan: 0.1 mM: GSH: 5 mM; Fe2+: 25 pM; EDTA: 30 pM. For other experimental conditions see Materials and Methods. Data from one representative experiment.

of alloxan, GSH, or iron chelate at constant concentrations of the other components resulted in increasing rates of lipid peroxidation and exhibited saturation kinetics. In the model system studied, saturating amounts of alloxan, GSH and iron chelate were used; doubling the amounts did not result in any significant increase in the rate of thiobarbituric acid- reactive material accumulation (data not shown).

As shown in fig. 2, A (532 nm) versus time profile of

T F

1 s 0 i I

1 0 1 0 0 1 0 0 0 c ( p M )

Fig. 3. Effect of stobadine (-O-), Trolox" (-W-) and 2-keto-4-me- thiol-butyric acid (KMBA) (-A-) on the duration of the lag phase of alloxan-induced lipid peroxidation in heat-denatured rat liver microsomes. Experimental conditions same as in Fig. 2. Each point represents the meani.S.E.M. of three experiments.

ALLOXAN-INDUCED LIPID PEROXIDATION 79

microsomes treated with alloxan to induce lipid peroxi- dation displayed a lag phase of approximately 4 min. fol- lowed by a period of rapidly increasing absorbance. The duration of the lag phase is a measure of susceptibility to peroxidation. In the presence of increasing concentrations of stobadine a profound prolongation of the lag phase was observed. Similar assays were conducted for TroIox@ and KMBA as reference antioxidants. As shown in fig. 3, ap- proximately the same concentrations of Trolox@ induced the same lag phase prolongation as did stobadine. On the other hand, KMBA up to 1 niM had no significant effect on the lag phase.

Rcrhrction ? f dloxcrn to dinhrric acid Further we studied the effect of stobadine on the reduction of alloxan to dialuric acid by monitoring the spectral changes in the range of 240-370 nm. Addition of alloxan (50 pM) to a solution of GSH (1 mM) at p H 7.5 led to an increase in absorption at 273 nm, characteristic of dialuric acid; the shoulder at 305 nm gradually changed into a peak as reported by other authors (Winterbourn & Munday 1989). At lower concentrations of GSH the reaction was correspondingly slower. The presence of stobadine (up to 0.5 mM) did not significantly affect the reduction of alloxan to dialuric acid (data not shown).

Hyrlrogen / l e ro .de generution

Fig. 4 shows the time course of H 7 0 2 production in the reaction of alloxan with GSH. During the first 10 min. of the reaction, up to 200 nmol/ml of H 2 0 2 was generated. The time course of Hz02 formation was not significantly affected by the presence of stobadine up to 2.5 mM.

250: T

I l\ T

Y

20 40 60 a o O L : ' ' ' " I I ' " ' " " ' 1 " 1 ' " " ' 1 " " " " ' 1 " ' 0

Time (min) Fig. 4 Time course of H 2 0 2 generation in the reaction of alloxan with GSH. Effect of stobadine. Stobadine: 0 (-O-), 2.5 mM (-m-). Alloxan: 0.1 mM. GSH: 5 mM. For other reaction conditions see Materials and Methods. Each point represents the mean5S.E.M. of three experiments.

Discussion

The site where alloxan exerts its toxicity in the cell is still unknown. Some authors have suggested that it is the cell or lysosomal membrane (Grankvist & Marklund 1986; Zhang rt r r l . 1992a & b), others stressed the ability of alloxan to cause DNA strand breaks (Uchida et a/. 1982; Sakurai et a/. 1992 & 1994). Several lines of evidence indicated that alloxan toxicity may involve lipid peroxidation - alloxan in combination with GSH was very effective in promoting lipid peroxidation providing catalytic amounts of iron were present (Reif et ( I / . 1989; Monteiro & Winterbourn 1989).

To investigate possible molecular mechanism(s) of the protective effect of stobadine against alloxan toxicity ob- served in mice (Stefek & Trnkova 1995), we conducted a series of experiments under in vitro conditions. We used rat liver microsomes as a model biological membrane. To elim- inate the activity of specific microsnmal enzymes, the micro- somal preparations were heat denatured. The model mem- brane was stressed by the reaction system of alloxan/GSH/ FelEDTA.

Sakurai et ul. (1990) used the reaction system of alloxanl GSHIFeIEDTA to study the mechanism of reactive free rad- ical species generation during the alloxan redox cycling; hy- droxyl radicals, superoxide anion radicals, hydrogen per- oxide, and intermediary alloxan radicals were detected in the model reaction mixture. All these species are potentially capable of altering membrane integrity.

In the complete reaction system used we observed time dependent accumulation of thiobarbituric acid-reactive ma- terial, indicating oxidative damage of the model membrane. Stobadine was found to inhibit the alloxan-induced lipid peroxidation with the efficacy comparable to that of the standard Trolox@.

Our results indicate that the intermediary alloxan rad- icals were not directly involved in the initiation of micro- somal lipid peroxidation. As shown in table 1, the inconi- plete reaction system in which iron chelate was omitted, leaving alloxaniGSH only, was uneffective in inducing lipid peroxidation. This finding was even more remarkable with regard to the fact that the yield of alloxan radicals was re- ported to be increased in the absence of iron (Reif Pt d. 1989). In the light of these results, alloxan radicals d o not seem to be directly responsible for the toxic effects of allox- an, a t least in the reaction system studied, which is not in line with the hypothesis put forward by Nukatsuka e l ti/.

( 1989). Superoxide anion radicals and hydrogen peroxide, gener-

ated during redox-cycling of alloxan, lack sufficient reac- tivity to initiate lipid peroxidation. They may react, how- ever, to produce a more reactive oxidant capable of hydro- gen abstraction from the lipid molecule, such as hydroxyl radicals or an iron-oxygen complex (Halliwell & Gutteridge 1992). Regardless of the nature of the initiating species, we found iron to be a prerequisite for the prooxidant effects of alloxan to develop (table I ) , which is in agreement with li t- erary data (Reif et a/. 1989; Sakurai er I / / . 1990; Fischer &

80 MILAN STEFEK A N D ZUZANA TRNKOVA

Hamburger 1980). Accordingly, strong metal chelating agents binding iron to inert chelate would be protective against alloxan toxicity. This, however, does not apply to stobadine as it was shown to lack metal chelating properties (Misik 1991).

Hydroxyl radicals are believed to be the terminal toxic agents in alloxan diabetes (Malaise 1982; Oberley 1988). This concept is based upon observations that hydroxyl rad- ical scavengers ameliorate alloxan toxicity both in isolated cells (Grankvist et al. 1979; Harman & Fischer 1982) and in the whole animal (Heikkila 1977; Tibaldi et a/. 1979). The absence of any inhibitory effect of the efficient hydroxyl radical scavenger KMBA, as shown in fig. 3, suggests that under our reaction conditions, hydroxyl radicals were not responsible for the damaging effects of alloxan. This finding is in accordance with the opinion that hydroxyl radicals play only a marginal role in iron catalyzed lipid peroxidation (Gutteridge & Halliwell 1990). 0x0-complexes of higher oxidation forms of iron were suggested as alternative spe- cies involved in the initiation of the oxidative damage in the system of alloxanlGSHIFelEDTA (Sakurai et a/. 1992) and other Fenton systems (Goldstein et al. 1993). However, it is likely that true (primary) initiation contributes little to microsomal peroxidation, since membrane fractions iso- lated from disrupted cells invariably contain trace amounts of lipid hydroperoxides. Added metal chelates may largely stimulate peroxidation by decomposing the preformed per- oxides (secondary initiation; Gutteridge & Halliwell 1990).

Our results further showed that stobadine affected nei- ther the rate of alloxan reduction to dialuric acid nor HzOz production in the system of alloxaniGSH. These results in- dicate that stobadine did not affect the redox cycling of alloxan in the reaction system of alloxaniGSHIFeiEDTA, the process during which potentially toxic free radicals are formed.

Since stobadine has been characterized to have only a minor affinity to superoxide anion radicals and hydrogen peroxide (Horakova et al. 1994) it is not likely that direct scavenging of these reactive oxygen species would partici- pate significantly in the observed inhibitory effect of stoba- dine.

Conclusions

On balance then, it can be summarized that in the reaction system studied: 1) stobadine inhibited alloxan-induced lipid peroxidation with the efficacy comparable to that of the standard Trolox@; 2) intermediary alloxan radicals or hy- droxyl radicals were not directly involved in the process of lipid peroxidation; 3) stobadine did not affect either GSH- mediated reduction of alloxan to dialuric acid or hydrogen peroxide production via alloxan redox cycling in the pres- ence of GSH: 4) the presence of iron chelate was found to be a prerequisite for the prooxidant effect of alloxan; and 5 ) stobadine is known not to chelate iron and is character- ized with only a minor affinity to superoxide anion radicals and hydrogen peroxide.

The above mentioned data indicate that stobadine does

not interfere with alloxan-induced lipid peroxidation at the stage of initiation (primary or secondary). We suggest that the protective effect of stobadine against alloxan toxicity is based predominantly on its demonstrated ability to scav- enge peroxyl radicals (Steenken et al. 1992; Kagan et al. 1993; Stefek et a/. 1992), thus inhibiting the propagation phase of the oxidative damage. We realize, however, that additional mechanism(s) may be involved in the protective effect of stobadine against the diabetogenic action of allox- an observed in vivo. The present results further indicate that processes not involving hydroxyl radicals may well partici- pate on the toxic effects of alloxan.

References

Benes. L. & S. Stolc: Stobadine. Drugs Future 1989. 14. 135- 137.

Buege, J. A. & S. D. Aust: Microsomal lipid peroxidation. In: Methods in eri:ymology, Vol. 52, Part C. Eds.: S. Fleischen & L. Packer. Academic Press 1978, pp. 302-310.

Fischer, L. J. & S. A. Hamburger: Inhibition of alloxan action in isolated pancreatic islets by superoxide dismutase, catalase and a metal chelator. Diubetes 1980, 29, 213-216.

Geiger, I? J. & S. I? Bessman: Protein determination by Lowry’s method in the presence of sulfhydryl reagents. Antr l~ , t . Biorliem. 1972, 49, 467473.

Goldstein, S., D. Meyerstein & G. Czapski: The Fenton reagents. Free Radical B i d . Med. 1993, 15, 435445.

Grankvist, K., S. Marklund, J. Sehlin & I. B. Fdljedal: Superoxide dismutase, catalase, and scavengers of hydroxyl radical protect against the toxic action of alloxan on pancreatic islet cells in vitro. Biochem. J . 1979, 182, 17-25.

Grankvist, K. & S. L. Marklund: Effect of extracelluarly generated free radicals on the plasma membrane permeability of isolated pancreatic beta cells. Int. J. Biocliem. 1986, 18, 109-113.

Gutteridge, J. M. C. & B. Halliwell: The measurement and mechan- ism of lipid peroxidation in biological systems. TlBS 1990, 15.

Halliwell, B. & J. M. C. Gutteridge: Biologically relevant metal ion- dependent hydroxyl radical generation. FEBS Lett. 1992, 307,

Harman. A. W. & L. J. Fischer: Alloxan toxicity in isolated rat hepatocytes and protection by sugars. Biorhem. Pkurnirrcol. 1982,

Heikkila, R. E.: The prevention of alloxan-induced diabetes in inice by dimethyl sulfoxide. Eur. J . fharnmcol. 1977. 44. 191-193.

Hildebrandt, A. G. & I. Roots: Reduced nitocinamide adenine dinu- cleotide phosphate (NADPH)-dependent formation and break- down of hydrogen peroxide during mixed-function oxidase reac- tions in liver microsornes. Arch. Biochem. Biophys. 1975, 171. 385-397.

Horakova, L., K. Briviba & H. Sies: Antioxidant activity of the pyridoindole stobadine in liposomal and microsomal lipid pcroxi- dation. Chem-Biol. Interactions 1992, 83. 85-93.

Horakova, L., H. Sies & S. Steenken: Antioxidant action of stobadi- ne. In: Methods in enzymology, Vol. 234, Part D. Ed.: L. Packer. Academic Press, 1994, pp. 572-581.

Kagan, V. E., M. Tsuchiya, E. Serbinovd, L. Packer & H. Sies: Interaction of the pyridoindole stobadine with peroxyl. superox- ide and chromanoxyl radicals. Bioclwm. Plzczrmrrcol. 1993. 45. 393400.

Malaise, W. J.: Alloxan toxicity to the pancreatic B-cells. A new hypothesis. Biochent Phannacol. 1982, 31, 352773534,

Misik, V.: E/’ect of antioxidunts on memhrunc, lipid perosirltrriori. Ph.D. Thesis, Institute of Experimental Pharmacology. Slovak Academy of Sciences, Bratislava 1991.

129- 135.

108-1 12.

31, 3731-3736.

ALLOXAN-INDUCED LIPID PEROXIDATION X I

Monteiro. H. I? & C. C. Winterbourn: Release of iron from ferritin by divicine, isouramyl, acid-hydrolyzed vicine, and dialuric acid and initiation of lipid peroxidation. Arch. B ~ O C ~ O H . Biophys. 1989. 271, 53C-545.

Nukatsuka. M.. H. Sakurai & J. Kawada: Generation of alloxan free radicals in chemical and biological systems: implication in the diabetogenic action of alloxan. Biochem. Biophj'.r. Rex Com- mun 1989. 165, 278-283.

Oberley. L. W.: Free radicals and diabetes. Fret Rtmical Bid. Mrd. 19x8. 5, 113-124.

Reif. D. W.. V. M. Samokyszyn. D. M. Miller & S. D. Aust: Alloxan- and glutathione-dependent ferritin iron release and lipid peroxi- dation. Arch. Bioc/wnz. Biophys. 1989. 269, 407414.

Sakurai. K . , T. Miura & T. Ogiso: Role of alloxan radical in gener- ation of hydroxyl radical by reaction of alloxan with glutathione in the presence of Fe3'-ethylene-diaminetetraacetic acid. Chon. Phrrn. Bull. 1990, 38, 993-997.

Sakurai. K . , K. Haga L T. Ogiso: Effect of glutathione on DNA strand breaks in the reaction system of glutathione-alloxan in the presence of Fe' t-ethylene-diaminetetrdacetic acid. Clienz. Pharm.

Sakurai. K.. K. Haga & T. Ogiso: A role of iron in DNA strand breaks in the reaction system of alloxan with reduced glutathione: iron (111) binding to the DNA. Bid. Plinmi. Bull. 1994, 17, 227- 231.

Stasko. A,. K. Ondrias, V. Misik. H. Szocsova & D. Gergel: Stobad- ine ~ a novel scavenger of free radicals. Chenz. Papers 1990, 44, 493 500.

Steenken. S., A. R. Sundquist, S. V. Jovanovic. R. Crockett & H. Sies: Antioxidant activity of the pyridoindole stobadine. Puls radiolytic characterization of one-electron-oxidized stobadine and quenching of singlet molecular oxygen. Cheni. Res. Tosicol. 1991. 5, 355-360.

Bull. 1992, 40, 2147-2150.

Stefek. M. & L. Benes: Pyridoindole stobadine is a potent scavenger of hydroxyl radicals. FEBS Lett. 1991, 294, 264266.

Stefek, M., M. Masarykova L L. Benes: lnhibition of cumene hy- droperoxide - induced lipid peroxidation by a novel pyridoindole antioxidant in rat liver microsomes. fhurniacology & TO.V~CO~O~.V 1992, 70, 407-41 I .

Stefek, M.: In v i m studies on the interaction of the pyridoindole antioxidant stobadine with rat liver microsomal P450. Xenobiut- icrr 1993, 23, 983-993.

Stefek, M. & Z. Trnkova: Stobddine - an inhibitor of alloxan toxic- ity. Can. J. Ph)).siol. Pl~armtrcol. 1994, 72 (Suppl. I ) , 591.

Stefek, M. & Z. Trnkova: Inhibition of alloxan-induced hypergly- caemia in mice by the pyridoindole stobadine. Phrrrmucologj~ & To.ricology 1995, 77, 320-322.

Tibaldi, J. J., F. Benjamin, S. Cabbat & R. E. Heikkila: Protection against alloxan-induced diabetes by various urea derivatives: re- lationship between protective effects and reactivity with hydroxyl radicals. J. Pharm. E.tp. Therup. 1979, 211, 415-418.

Uchida, Y., H. Yamamoto, A. Kawamura & H. Okamoto: Protec- tion by superoxide dismutase, catalase, and poly(ADP-ribose)- synthetase inhibitors against alloxan- and streptozotocin-induced islets DNA strand breaks and against the inhibition of proinsulin synthesis. J. Bid. Cfiein. 1982, 257. 60846088.

Winterbourn, C. C. & R. Munday: Glutathione-mediated redox cycling of alloxan. Mechanism of superoxide dismutase inhibition and of metal-catalyzed .OH formation. Biochem. f/iunnrrcol. 1989, 38, 271-277.

Zhang, H.. G. Gao & U. T. Brunk: Extracellular reduction of allox- an results in oxygen radical-mediated attach on plasma and lys- osomal membranes. APMIS 1992a, 100. 317-325.

Zhang, H., J. M. Zdolsek & U. T. Brunk: Alloxan cytotoxicity in- volves lysosomal damage. APMIS 1992b, 100, 309-316.