Alloxan-induced DNA Strand Breaks in Pancreatic Islets

THEJOURNAL

OF BIOLOGICAL CHEMISTRY W 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 266, No. 4, Issue of February 5, pp. 2112-2114,1991 Printed in U.5’. A.

Alloxan-induced DNA Strand Breaks in Pancreatic Islets EVIDENCE FOR Hz02 ASANINTERMEDIATE* (Received for publication, July 16, 1990)

Nobuyuki Takasu, Takayuki Asawa, Ichiro Komiya, Yoshitaka Nagasawa, and Takashi Yamada From the Departmentof Gerontology, Endocrinology and Metabolism, School of Medicine, Shinshu University, Asahi 3-1-1, Matsumoto, Nagano-ken 390, Japan

Alloxan exhibits the most potentdiabetogenicity and has been used for induction of experimental diabetes mellitus. Understanding the mechanisms of action of the typical diabetogenic agent is important for elucidating the causes of diabetes. Okamoto (Okamoto, H. (1985)BioEssays 2,15-21) proposed a model in which DNA fragmentation plays an important role for the development of diabetes. This DNA fragmentation is supposed toresult from the accumulationof superoxide or hydroxyl radicals. However, direct evidence for this accumulation is lacking. Using rat pancreatic islets, we demonstrated that alloxan stimulated HzOz generation, which induced DNA strand breaks. These findings support Okamoto’s proposal thatalloxan induces diabetes through the following biochemical events: alloxan 4 HzOzgeneration + DNA strand breaks + diabetes mellitus. Perhaps this report constitutes the first demonstration of alloxan-stimulated HzOz generation which could conceivably act as an intermediate for alloxan-induced DNA strand breaks.

presence of exogenously added peroxidase to convert homovanillic acid to a fluorescent metabolite. The washed islets were suspended in Krebs-Ringer bicarbonate-20 mM HEPES’ buffer (pH 7.4)with 0.5% bovine serum albumin. The islet suspension (100 islets/0.5 ml) was transferred to a thermostated quartz cuvette (37‘C) for measurement of H202generation; the cuvette contained horseradish peroxidase (type I 1 , O . l mg/ml, final, Sigma) andhomovanillic acid (0.44 mM, final, Sigma). H202 generation was measured in a Hitachi F4000 fluorometer (Hitachi, Tokyo). Fluorescence was recorded with excitation andemission wavelengths of 315 and 425 nm, respectively. The islet suspensions were stirred continuously.After reaching a steady state, alloxan was added to the cuvette using a microsyringe. The rates of alloxan-stimulated H202 generationwere obtained from the slopes after reaching steady state levels and expressed as nanomoles/islet/min. Standard curveswere produced from incubations in which the islets were replaced by known amounts of H202. Density Gradient Analysis of Islet DNA-Density gradient analysis of isletDNAwas done as before (5).Batches of 100 islets were incubated at 37 “C in 200 p1 of Krebs-Ringer bicarbonate medium (pH 7.4) containing 5 mM sodium pyruvate, 5 mM sodium fumarate, 5 mM sodium glutamate,2 mg/ml bovine serum albumin, and2.8 mM glucose in an atmosphere of95% 02 and 5% C 0 2 . After a 5-min preincubation, alloxan or H202was added to the medium, and the islets were incubated for the indicated periods a t 37 “C. After incubation, saline-washed islets were suspended in 50 p1 of cold saline and immediately layered over 0.5 ml of lysis solution (1.0 N NaOH, 0.01 M EDTA, 1% (v/v) Triton X-100) that bad been just layered Alloxan and streptozotocin selectively damage pancreatic over 5 ml of a 5-20% (w/v) linear sucrose gradient containing 0.3 N B-cells and exhibit the most potent diabetogenicity. They NaOH, 0.7 M NaC1, and 0.01 M EDTA.Onthebottom of each have been widely used for induction of experimental diabetes gradient was a 1-ml 80% (w/v) sucrose shelf. The loaded gradients at temperature for 30 min. The gradients mellitus (1, 2). Understanding the mechanisms of action of were placed in the dark room the typical diabetogenic agent is important for elucidating the were then centrifuged at 26,000 rpm a t 20 “C for 200 min in Beckman causes of diabetes. Thus, many mechanisms have been con- SW rotor. After centrifugation, fractions of 20 drops were collected the gradients. DNA in each fraction was precipitated by adding sidered over the last 40 years for the specific B-cell toxicity 2from ml of 20% cold trichloroacetic acid with 200 pg of bovine serum of alloxan and streptozotocin. Okamoto andco-workers (3-5) albumin as carrier. The precipitate was washed three times withcold had gathered the evidence from their own and others and trichloroacetic acid (20%) and then assayed for DNA content fluoproposed a model, which has been increasingly accepted and rometrically (9). Others-Experiments were conducted a t least seven times. Typical is now called “Okamoto’s model” (6). Central to thismodel is the fragmentation of nuclear DNA of pancreas B-cells. This data and the final concentrations of alloxan are shown in the text and figures. Data were statistically analyzed with analysis of variance DNA fragmentation seems to be important for the develop- or Student’s t test. p < 0.05 was considered statistically significant. ment of diabetes and issupposed to result from the accumu- Chemicals were of the highest purity available commercially.

lation of superoxide or hydroxyl radicals inthe caseof alloxan (5). However, direct evidence forthis accumulation is lacking. Using rat pancreatic islets,we present evidence that alloxan stimulates H,Oz generation, which induces DNA breaks. EXPERIMENTAL PROCEDURES

Pancreatic Islets-Pancreatic islets were obtained from male Wistar rats weighing 180-210 g by collagenase (Wako Pure Chemical Industries, Osaka, Japan) digestion as described by Lacy and Kostianovsky (7). HLO2 Generation-HeOz generation was recorded continuously as described before (8). This method utilizes the medium H20, in the

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

RESULTS

Effect of Alloxan on H 2 0 zGeneration-As shown in Fig. lA, addition of 1 mM alloxan to islet suspension resulted in a gradual increase in HzOZ generation with an initial 1.5-min lag. This alloxan-stimulated H202generation increased linearlyup to 60 min (datanot shown). The rate of H202 generation was 0.92 nmol/islet/min when calculated fromthis linear slope. Graded doses of alloxan stimulated H202generation (Fig. 1B); the effect of alloxan on the rate of HZ0z generation was evident at a concentration of 0.1 mM and maximal at 1 mM.

’ The abbreviation used is: HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid.

2112

Alloxan-stimulated H202and DNA Breaks Pancreatic in Islets

2113

peak,indicatingthatalloxangenerates isletDNAbreaks. Graded doses of alloxan induced DNA fragmentation (Fig. 20). The effect of alloxan on DNA fragmentation was observed at a concentration of 0.1 mM and was maximal at 1

Effect of Alloxan on DNA Fragmentation-Islets were incubated with alloxan for 7-20 min, andvelocity sedimentation of DNA was examined in an alkaline sucrose density gradient. DNA of isletsincubatedwithoutalloxan for 20 min was recovered as a single peak near the bottom of the gradient, the position at which undamaged DNA sediments (Fig. 2 A ) . However, after only a 7-min incubation with 1 mM alloxan, a considerable amount of DNA sedimented as a broad peak in the middle of the gradient with a concomitant decrease in 1 mM undamagedDNA (Fig. 2B). Afterincubationwith alloxan for 20 min, the DNA was almost completely fragmented (Fig. 2C); islet DNA sedimented slower as a broad

mM.

Effect of H,O, on DNA Fragmentation-Islets were incubated with 500 p~ H 2 0 2for 20 min, and velocity sedimentation of DNA was examined in an alkaline sucrose density gradient. The DNA of islets incubated without Hz02for 20 min was recovered as a single peak near the bottom of the gradient, the position at which undamaged DNA sediments (Fig. 3 A ) . However, after incubation with 500 p~ H202for 20 min, the DNA was almost completely fragmented (Fig. 3 B ) ; the islet DNA sedimented slower as a broad peak, indicating that H,Oz induces islet DNA breaks. Graded doses of H202 induced DNA strand breaks (Fig. 3C). The effect of H202on DNA fragmentation was evident at a concentration of 30 p~ and maximal at500 PM. DISCUSSION

We demonstrated that alloxan stimulated Hz02generation that induced DNA strandbreaks. These findings may support Okamoto’sproposal that alloxan induces diabetes mellitus through the following biochemical events: alloxan + H202 generation -+ DNA strand breaks+ diabetes mellitus. rnin Alloxan has been known to be diabetogenic and induces FIG. 1. Effect of alloxan on HzOz generation. A , time course DNA strand breaks inisolated rat pancreas isletsin vitro and of alloxan-stimulated H,Oz generation. HzOz generation was recorded in vivo to cause diabetes mellitus (5, 10). Catalase has been continuously as described under “Experimental Procedures.” Addishown to protect against alloxan-induced islet DNA strand tion of 1 mM alloxan (arrow) to pancreatic islet suspension resulted in a gradual increase in H202 generation with a 1.5-min initial lag. breaks (51, indicating thatH,O, generation might mediate the This alloxan-stimulatedHz02 generation increasedlinearly. The rate action of alloxan.The H,O, generationand DNA strand of HzOz generationis 0.92 nmol/islet/minfromthelinear slope breaks may play some role for the development of diabetes (0-r -O+). B, effects of graded doses of alloxan on HZO, gener- mellitus. ation. The ratesof alloxan-stimulated Hz02generation were obtained Generation of Hz02 appears tobe a natural process and is from the slopes after getting steady state. Graded doses of alloxan stimulated Hz02 generation; theeffect of alloxan on Hz02 generation of universal occurrence in a variety of cells. However, there to generation of H,O, in pancreatic was evident at a concentration of 0.1 mM ( p < 0.05) and maximal at have been no reports show 1 mM. Each point is the mean & S.E. of 5-7 determinations. islets. Although direct evidence was lacking for the presence

C

5 bottom

10

15 1

5

10

fraction number

15 1

5

10

15 top

Alloxan

mM

FIG. 2. Effect of alloxanon DNA fragmentation. A , B, C, effects of alloxanon DNA fragmentation (velocity sedimentation of DNA). Islets were incubated with alloxan for 7-20 min, and velocity sedimentation of DNA was examined in an alkaline sucrose density gradient. ofDNA islets incubated without alloxan for 20 min wasrecovered as a single peak near the bottomof the gradient ( 4 , fraction 4), the position at which undamaged DNA sediments ( A ) .However, after a 7-min incubation with 1 mM alloxan, a considerable amount of DNA sedimented asa broad peak in the middle of the gradient with a concomitant decrease in undamaged DNA ( B ) .After incubation with 1 mM alloxan for 20 min, the DNA was almost completely fragmented ( C ) ;islet DNA sedimented slower as a broad peak, indicating that alloxan generates isletDNA breaks. D, effects of a 20-min exposure to graded doses of alloxan on undamaged DNA contents. Effects of graded doses of alloxan on DNA fragmentation were studied as in C. DNA in fraction 4 ( 4 in panel A ) was considered undamaged. DNA contents in fraction4 in the absence of alloxan was considered to be control (100%).Graded doses of alloxan decreased the amountsof undamaged DNA contents; the effect was evident a t a concentration of 0.1 mM ( p < 0.05) and maximal at 1 mM. Each point is the mean f S.E. of 5-7 determinations.

Alloxan-stimulated H202and DNA Breakin Pancreatic Islets

2114 40

A

"

1

fraction bottom

5

10

I

15 1

B

5

number

10

15

top

"202

PM

FIG. 3. Effect of HzOzon DNA fragmentation. A and B, effect of H20, on DNA fragmentation (velocity sedimentation of DNA). Islets were incubated without ( A ) or with ( B ) 500 PM Hz02for 20 min, and velocity sedimentation of DNA was examined in an alkaline sucrose density gradient as in Fig. 2. DNA of islets incubated without alloxan for 20 min was recovered as a single peak near the bottom of the gradient ( 4 , fraction 4), the position at which undamaged DNA sediments ( A ) .However, after incubation with 500 PM H,O, for 20 min, the DNAwas almost completely fragmented ( B ) ; islet DNA sedirnented slower as a broad peak, indicating that H20, generates islet DNA breaks. C , effects of graded doses of H,O, on undamaged DNA contents. Effects of a 20-min exposure to graded doses of alloxan on DNA fragmentation were studied as in Fig. 2. DNA in fraction 4 ( 4 in panel A ) was considered undamaged. DNA contents in fraction 4 in the absence of Hz02was considered to be control (100%).Graded doses of H,O, decreased the amounts of undamaged DNA contents; the effect was evident at a concentration of 30 p M ( p < 0.05) and maximal at 500 PM. Each point is the mean ? S.E. of 5-7 determinations.

of H2O2 in B-cells, Heikkila et al. (11) presented data to support the view that hydroxyl radicals, including H202, play a primary role in the diabetogenic action of alloxan. We clearly demonstrated the presence of H202 generation in pancreatic islets. Alloxan stimulates this HzOz generation. Further studies will be required to answer the questions of how this H202is generated and how alloxan stimulates this HzOzgeneration. Alloxan induces DNA strand breaks in pancreatic islets.

After several minutes incubation with alloxan, a considerable amount of DNA sedimented as a broad peak in themiddle of the gradient with a concomitant decrease in undamaged DNA. The results indicate that alloxan produces strand breaks in islet DNA. We also demonstrated that H202itself produces DNA strand breaks. From the above results, it is reasonable to propose that alloxan-induced DNA strand breaks are through H202 generation. The DNA strand breaks induce diabetes mellitus (1,5,6). According to Okamoto's model (5), alloxan induces islet DNA strand breaks, and poly(ADPribose) synthetase acts to repair the DNA breaks, consuming islet NAD; this rapid and marked depletion of islet NAD is assumed to induce diabetes mellitus. This seems to be of special importance in understanding the pathogenesis of insulin-dependent diabetes mellitus. The H202-induced DNA strand breaks are not specific for pancreatic islets and have been reported in other cells (12). We demonstrated itspossible significance in the alloxan-induced diabetes mellitus. We demonstrated that H202generation might mediate the alloxan-induced diabetes mellitus. However, it has been suggested that other hydroxyl radicals or superoxide also could be responsible for the destruction of the pancreatic islets by alloxan. Further study will be required to solve this. REFERENCES 1. Rerup, C. C. (1970) Pharmacol. Reu. 2 2 , 485-518 2. Agarwal, M. K. (1980) FEBS Lett. 1 2 0 , 1-1 3. Yamamoto, H., Uchigata, Y., and Okamoto, H. (1981) Nature 294,284-286 4. Uchigata, Y., Yamamoto, H., Kawamura, A., and Okamoto, H. (1982) J. Biol. Chem. 257,6084-6088 5. Okamoto, H. (1985) BioEssays 2 , 15-21 6. Renold, A. E. (1988) Diabetes Annu. 4, 592-608 7. Lacy, P. E., and Kostianovsky, M. (1967) Diabetes 16,35-39 8. Takasu, N., Yamada, T., Shimizu, Y., Nagasawa, Y., and Komiya, I. (1989) J. E n d o c r i d . 120,503-508 9. Fiszer-Szafarz, B., Szafarz, D., and Geuvarda de Murillo, A. (1981) Anal. Biochem. 110,165-170 10. Yamamoto, H., Uchigata, Y., and Okamoto, H. (1981) Biochem. Biophys. Res. Commun. 103,1014-1020 11. Heikkila, R. E., Winston, B., Cohen, G., and Barden, H. (1976) Biochem. Phurmacol. 2 5 , 1085-1092 12. Hoffmann, M. E., Mello-Filho, A.C., and Meneghini, R. (1984) Biochim. Biophys. Acta7 8 1 , 234-238