- Email: [email protected]
Effects of cytotoxic drugs and inhibitors of insulin secretion on a serially transplantable rat insulinoma and cultured rat insulinoma cells
Gen. Pharmac. Vol. 18, No. 3, pp. 293-297, 1987 Printed in Great Britain. All rights reserved
0306-3623/87 $3.00 + 0.00 Copyright © 1987 Pergamon Journals Ltd
EFFECTS OF CYTOTOXIC DRUGS A N D INHIBITORS OF INSULIN SECRETION ON A SERIALLY TRANSPLANTABLE RAT INSULINOMA A N D CULTURED RAT INSULINOMA CELLS PETER R. FLATT, SARA K. SWANSTON-FLATT, KIM S. TAN and VINCENT MARKS Department of Biochemistry, University of Surrey, Guildford, Surrey GU2 5XH, U.K. [Tel: (0483) 571281] (Received 17 July 1986)
Abstract--1. The effects of cytotoxic drugs and inhibitors of insulin secretion were examined in vivo in rats with a radiation-induced transplantable insulinoma, and in vitro using cultured rat insulinoma cells and the derived RINm5F insulin-secreting cell line. 2. Administration of diazoxide to insulinoma-bearing rats resulted in a transient decrease of plasma insulin with a temporary rise of glucose concentrations. Mannoheptulose and somatostatin failed to affect the marked hyperinsulinaemia and hypoglycaemia. 3. Streptozotocin produced a rapid and sustained decrease of insulin concentrations in insulinomabearing rats, accompanied by a progressive elevation of plasma glucose. Administration of alloxan failed to affect circulating insulin or glucose concentrations. 4. In vitro, streptozotocin and alloxan exerted approximately equipotent time-dependent and concentration-dependent cytotoxic effects on insulinoma cells and RINm5F ceils as established by cell staining with trypan blue. The cytotoxic actions of both drugs were decreased by agents believed to scavenge free radicals or to act as inhibitors of poly(ADP-ribose) synthetase. 5. The results suggest that the cytotoxic actions of streptozotocin and alloxan on rat insulinoma cells and RINm5F ceils are mediated by the generation of hydroxyl free radicals and DNA strand breaks. The ineffectiveness of alloxan in insulinoma-bearing rats probably reflects the high rate of decomposition of the drug in vivo.
INTRODUCTION The serially transplantable New England Deaconess Hospital ( N E D H ) rat insulinoma developed in one of a pair of parabiont rats by X-ray irradiation (Chick et al., 1977) provides a useful animal model of insulinoma-hypoglycaemia. Routine subscapular transplantation of tumour fragments in inbred N E D H rats of the Surrey subline consistently produces an insulinoma syndrome characterized by rapid tumour growth and defective insulin secretion, resulting in marked hyperinsulinaemia, severe hypoglycaemia and neuroglycopeenic coma within one month (Flatt et al., 1986a,b, 1987). Studies in N E D H rats with the transplantable insulinoma have indicated the potential effectiveness of surgery in reversing hypoglycaemia, but tumour recurrence was invariably observed within 2 months (Flatt et al., 1986b). This clearly illustrates some of the potential problems associated with the treatment of malignant insulinomas. In addition to surgical intervention, treatment of hypoglycaemia due to insulin-secreting tumours in man involves the use of inhibitors of insulin secretion or the administration of cytotoxic drugs (Editorial, 1981; Marks and Rose, 1981; Friesen, 1982). In the present study, we have examined the effects of mannoheptulose, somatostatin and diazoxide (inhibitors of insulin srcretion), and of alloxan and streptozotocin (cytotoxic drugs) on plasma insulin and glucose concentrations of
insulinoma-bearing rats. The agents selected have all been used clinically to treat insulinoma, although with varying degrees of success (Editorial, 1981; Marks and Rose, 1981; Friesen, 1982). In addition to in vivo studies in insulinoma-bearing rats, the cytotoxic actions of alloxan and streptozotocin have been evaluated in vitro using cultured rat insulinoma cells and the R I N m 5 F cell line originally derived from the transplantable rat insulinoma (Gazdar et al., 1980). MATERIALS AND METHODS Animals Male inbred albino NEDH (New England Deaconess Hospital) rats (Rattus norvegicus albinus) from the colony at the University of Surrey carrying a serially transplantable radiation-induced insulinoma (Chick et al., 1977) were used at 14-16 weeks of age. The origin of these rats has been described elsewhere (Flattet al., 1986b). The rats were housed in an air-conditioned room at 22 + 2°C with a lighting schedule of 12hr light (0700--1900hr) and 12hr dark. A standard pellet diet (Spratts Laboratory Diet 1, Lillico Ltd, Reigate, U.K.) and tap water were supplied ad libitum. Transplantation and in vivo studies Single subcutaneous subscapular tumours (subline SFA/I) from donor NEDH rats were used as the source of tumour fragments for transplantation. Each tumour was rapidly excised, the capsule was removed and the contents finely minced. Recipient rats were lightly anaesthetized with
293
294
PETER R. FLATT et al.
ether, and 0.1 ml of minced tumour was implanted subcutaneously into the subscapular region using a 16 gauge needle. The effects of cytotoxic drugs and established inhibitors of pancreatic B-cell insulin secretion were examined in groups of insulinoma-bearing rats at 18-22 days after transplantation. The following agents were administered by intraperitoneal injection in a volume corresponding to 5 ml/kg body wt: diazoxide (Glaxo Group Research, Ware, U.K.; 100mg/kg body wt in 0.9% NaC1), somatostatin (Sigma Chemical Co. Ltd, Poole, U.K.; 1 mg/kg body wt in 0.9% NaCI), mannoheptulose (Sigma; 2 g/kg body wt in distilled water), alloxan monohydrate (Sigma; 200mg/kg body wt in 0.9% NaC1) and streptozotocin (Upjohn Chemical Co. Ltd, Kalamazoo, U.S.A.; 100mg/kg in 0.5M sodium citrate buffer, pH 4.5). Alloxan and streptozotocin were dissolved in chilled buffers immediately prior to rapid injection. A second dose of alloxan or streptozotocin (100 mg/kg or 50 mg/kg body wt, respectively) was administered to the rats at 50 hr after the initial injection. The doses and mode of administration of alloxan and streptozotocin were chosen on the basis of their established ability to induce pancreatic B-cell destruction in the rat (Rerup, 1970). Plasma glucose (Stevens, 1971) and insulin (Flatt and Bailey, 1981) were determined immediately before and at the times indicated in the Figures after administration of each test agent. Blood samples were taken from the tail tip of fed conscious rats at 0900-1400 hr. The plasma was separated and stored at - 2 0 ° C until analysis. Cell culture and cytotoxicity studies
Tumors weighing 2.2-7.9 g were excised from insulinomabearing rats, stripped of surrounding capsule and used to prepare tumour B-cell suspensions by gentle mechanical disruption. Isolated tumour cells were cultured in suspension for 3 days at 37°C in a humidified atmosphere of 5% CO 2. The culture medium was RPMI-1640 (Gibco Europe Ltd, Paisley, U.K.) containing 11.1 mmol/1 glucose, 10% foetal calf serum with added antibiotics (100 U/ml penicillin and 0.1 mg/ml streptomycin; Gibco Europe Ltd). After culture, cell viability assessed by trypan blue exclusion was increased from 21% in freshly isolated cells to 72%. In addition to cultured tumour cells, cytotoxicity studies were performed using the insulin-secreting RINm5F cell line established from the transplantable rat insulinoma (Gazdar et al., 1980). The cell line was kindly provided by Dr A. J. Bone (Edinburgh, U.K.) by courtesy of Professor C. Hellerst6m (Uppsala, Sweden). The RINm5F cells were maintained in monolayer culture under the same conditions as tumour B-cells. Confluent RINm5F cell cultures were rinsed in Ca2+-free and Mg2÷-free Hanks Balanced Salt Solution to detach the cells prior to experimentation. At this time, cell viability assessed by trypan blue exclusion was 80%.
Cytotoxicity studies were performed by incubating approximately 2.5 x 106 tumour B-cells or RINm5F cells in 250#1 of modified Krebs Ringer bicarbonate buffer (pH 7.4), containing 20mM N-2-hydroxyethylpiperazineN'-2-ethanesulfonic acid (HEPES), 115 mM NaCI, 24 mM NaHCO3, 4.7 mM KCI, 2.56 mM CaC12, 1.2 mM MgSO 4, 1.2mM KH2PO4, 1.4mM D-glucose and 5 mg/ml bovine serum albumin. The buffer was supplemented with alloxan or streptozotocin and the test agents at the concentrations indicated in the Figures and Table. Alloxan and streptozotocin were added from stock solutions of 250mM (0.4-12.8/21) which were prepared in 10mM HCI and 10 mM citrate buffer (pH 4.5) respectively immediately prior to the commencement of cell incubations. The addition of alloxan or streptozotocin did not noticeably alter the pH of the incubation medium. The cells were incubated for 10-60 rain in a water bath at 3T'C. At the end of incubation, a 50/21 aliquot of suspended cells was added to 150/~1 unsupplemented buffer, followed by an equal volume of chilled trypan blue solution (British Drug Houses Ltd, Poole, U.K.; 2mg/ml in 0.9% NaC1). After thorough mixing, the cells were counted using a haemocytometer. The frequency of cells with trypan blue-stained nuclei was assessed by counting approximately 200 randomly selected cells under a 40 × objective lens in a light microscope. The validity of this counting procedure has been demonstrated previously (Grankvist et al., 1977, 1979). It was checked in control experiments without cytotoxic drugs that addition of 10raM HC1 or 10mM citrate buffer (pH4.5) to the incubation medium did not affect cell viability as assessed by exclusion of trypan blue. Statistical analysis
Groups of data are presented as means + SEM. Statistical evaluation was performed using Student's paired and unpaired t-tests. Differences were considered to be significant for P < 0.05. RESULTS Studies on transplantable rat insulinorna in vivo
I m p l a n t a t i o n of small fragments of the serially t r a n s p l a n t a b l e i n s u l i n o m a resulted in the develo p m e n t of a single s u b s c a p u l a r t u m o u r in each rat by 18-22 days. As s h o w n in Figs 1 a n d 2, this was a c c o m p a n i e d by a m a r k e d increase o f p l a s m a insulin (range, 11.3-18.9 ng/ml) a n d decrease o f plasma glucose (range, 1.9-2.9 mmol/1) c o m p a r e d with control rats (2.3 4- 0.5 ng/ml a n d 6.7 4- 0.3 mmol/1, n = 8; d a t a not illustrated). A d m i n i s t r a t i o n of m a n n o heptulose or s o m a t o s t a t i n to i n s u l i n o m a - b e a r i n g rats 24
8
4
2b~ 2'-.
16
I
1
I
0
-
-0
1
60
120
0
I
J
60 120 Minutes r~inutes Fig. 1. Effects of diazoxide (O O; 100 mg/kg), mannoheptulose (E] Fq; 2 g/kg) and somatostatin (O O; 1 mg/kg) on plasma glucose and insulin concentrations of insulinoma-bearing rats at 18-22 days after implantation of tumour fragments. Values are mean___ SEM of groups of 6-8 rats. *P < 0.05 - 0.001 compared with time zero. O
Rat insulinoma 20
295
24
15 16 10
0--0-0 I I 0 2
0
0---0"'0""-'0 I ~ I I 4 24 48
1 72
,. 0 I 96
0
I
I ~
I
I
I
I
2
4
24
48
72
96
Hours
Hours
Fig. 2. Effects of streptozotocin (O O; 100 mg/kg at time zero and 50 mg/kg at 50 hr) and alloxan monohydrate (O O; 200 mg/kg at time zero and 100 mg/kg at 50 hr) on plasma glucose and insulin concentrations of insulinoma-bearing rats at 21 days after implantation of tumour fragments. Values are mean + SEM of groups of 6 rats. *P < 0.05-0.001 compared with time zero. did not affect circulating insulin or glucose concentrations (Fig. 1). However, diazoxide produced a transient decrease of plasma insulin associated with a temporary rise of glucose concentrations (Fig. 1). As shown in Fig. 2, administration of alloxan to insulinoma-bearingrats failed to affect plasma insulin or glucose concentrations, although there was a tendency towards a less severe hyperinsulinaemia after the drug. In contrast, streptozotocin produced a rapid, marked and sustained decrease of circulating insulin concentrations accompanied by a progressive elevation of plasma glucose (Fig. 2). By 48-96 hr, the glucose concentrations of streptozotocin-treated rats were significantly greater than those of untransplanted control rats. Studies on cultured rat tumour cells a n d R I N m S F cell line in vitro
Exposure of isolated cultured tumour cells and RINm5F cells to 2 m M alloxan or 2 m M streptozotocin in vitro resulted in a marked and timedependent decrease of cell viability as assessed by cell staining with trypan blue (Fig. 3). The effects of the two cytotoxic drugs were similar irrespective of the use of tumour cells or RINm5F cells. As shown in Fig. 4, incubation of either cell type for 20 min with
different concentrations of alloxan or streptozotocin revealed that the cytotoxic effects of the drugs were dose-dependent. There were no significant differences in the potency of alloxan and streptozotocin over the concentration range tested. Table 1 shows the effects of various agents on the cytotoxic action of 8 mM alloxan or 8 mM streptozotocin on tumour cells and RINm5F cells during 20 rain incubations. These agents were selected on the basis of their established protective effects against alloxan- and/or streptozotocin-cytotoxicity of insulin-secreting pancreatic B-cells (Malaisse, 1982; Okamoto and Yamamoto, 1983). Each agent tested appeared to offer some degree of protection against the toxic effects of alloxan or streptozotocin on either tumour cells or RINm5F cells. There were no substantial differences between protective effects using the two cell types. In contrast to the action of streptozotocin, the effects of alloxan on cell viability were not diminished by benzamidine, butanol or catalase at the concentrations employed. DISCUSSION
Consistent with previous reports (Flatt et al., 1986a,b, 1987), transplantation of small fragments of
Turnout cells
RI NmSF cells
24
8~ 16
t.~
8
.5
O.
20 Minutes
40
60
0
20
40
60
Minutes
Fig. 3. Effects of time of exposure to streptozotocin (O O; 2retool/l) and alloxan ( 0 ....... O; 2 mmol/1)on the frequency of tumour cells or RINm5F cells stained with trypan blue. Values are presented as mean + SEM of 5 separate experiments. A significantincrease (P < 0.05--0.01)in the percentage of cells stained was evident by 10 min. O.P. 18/~-F
PETER R. FLATT et al.
296
RINm5F
Ttrnour cells
cells
40
i .~
lO
O
4
8
I
I
12
16
Drug concentration (mmol/l)
O Drug
4 concentration
8
12
16
(m~i/l)
Fig. 4. Effects of 20 min exposure to different concentrations of streptozotocin (O 0 ) and alloxan (O O) on the frequency of tumour cells or RINm5F cells stained with trypan blue. Values are presented as mean + SEM of 5 experiments. A significant increase (P < 0.05) in the percentage of cells stained was evident by 1 mmol/1. a serially transplantable insulinoma into N E D H rats o f the Surrey subline, consistently resulted in m a r k e d hyperinsulinaemia and hypoglycaemia by 18-22 days. A t this time, the t u m o u r B-cells do not r e s p o n d to glucose, arginine or glibenclamide, although responsiveness to agents such as glucagon and epinephrine which m o d u l a t e insulin release t h r o u g h the adenylate cyclase-vyclic A M P system appears to be retained (Flatt et al., 1987). In the present study, m a n n o h e p t u l o s e and somatostatin which are established inhibitors o f pancreatic B-cell insulin release (Hedeskov, 1980; Samols et al., 1983) failed to affect plasma insulin and glucose concentrations o f insulinoma-bearing rats. This is in general agreement with the limited effectiveness o f these agents in patients with insulinoma ( J o h n s o n and Wolff, 1979; M a r k s and Rose, 1981), although a somatostatin analogue has been used with some success (Long et al., 1979). Diazoxide is by far the most effective inhibitor o f insulin secetion employed for the management o f clinical insulinoma (Editorial, 1981; M a r k s and Rose, 1981). As d e m o n s t r a t e d by the present study, an inhibitory action on insulin release with an a c c o m p a n y i n g elevation o f circulating glucose was observed in insulinoma-bearing rats. How-
ever, as in man, the effect was only transient indicating that successful treatment o f insulinoma with diazoxide can only be achieved by repeated dosage with the drug. A d m i n i s t r a t i o n o f the B-cytotoxic drugs alloxan and streptozotocin revealed a m a r k e d divergence in their effectiveness in reversing the hyperinsulinaemia and hypoglycaemia o f insutinoma-bearing rats. Thus whereas streptozotocin rapidly reversed the metabolic abnormalities, alloxan was without effect on plasma insulin and glucose concentrations. This finding is in agreement with the results o f clinical studies which have s h o w n that streptozotocin but not alloxan is highly effective in the treatment o f malignant endocrine pancreatic t u m o u r s (Marks and Rose, 1981). Since b o t h drugs are established pancreatic B-cell toxins in the rat (Rerup, 1970), their different effectiveness may be a consequence o f the altered function o f the t u m o u r B-cells. However, as indicated by the results o f in vitro cytoxicity studies, cultured rat insulinoma cells and R I N m 5 F cells were equally sensitive to alloxan and streptozotocin. This suggests that the different effectiveness o f alloxan and streptozotocin in insulinoma-bearing rats is due to differences in the stability and in vivo clearance o f the
Table 1. Effectsof various agents on the frequency of tumour cells or RINm5F cells stained with trypan blue after 20 min exposure to 8 mmol/l streptozotocin or 8 mmol/l alloxan Change in the percentage of cells stained with trypan blue Agent Tumour cells RINm5F cell line (concentration) Streptozotocin Alloxan Streptozotocin Anoxan Glucose (16.7 mM) - 17.3 ± 5.2* - 15.2 + 6.6 - 18.4 _+4.2* - 15.7 ± 4.7* Theophylline (5 mM) -30.2 + 5.8*** -25.1 ± 6.6** - 16.2 ± 7.2 -21.5 + 4.4*** Nicotinamide (10 mM) -29.1 + 6.0*** -20.3 + 3.3*** -22.4 ± 5.4** 13.4 ± 3.9* Sodium benzoate (50mM) -31.8 _+6.1"** - 16.1 ± 8.5 - 17.7 ± 4.9* 15.7 + 3.3*** Benzamide (1 mM) -31.0_+4.3"*** -21.7 ± 9.0 17.7+6.4" -3.2+7.7 Dimethylsulphoxide (28 rnM) -25.4 ± 9.8* - 19.8 ± 6.3* - 16.0 ± 7.5 -5.4 + 5.4 Butanol (22 mM) -24.2 _ 9.1" -- 12.3 ± t0.7 - 19.2 ± 5.7* --3.9 ± 5.4 Catalase (250 U/ml) 30.7 + 3.2**** 19.7 ± 9.2 18.8 + 3.6*** 13.4 ± 10.0 Superoxide dismutase (2000 U/ml) -24.5 ± 8.4* -21.4 ± 5.2** -21.8 ± 2.2**** - I 1.8 ± 8.1 Values are presented as mean ± SEM of 5 separate experiments. *P < 0.05, **P < 0.02, ***P < 0.01 and ****P < 0.001 compared with the frequency of cells stained with trypan blue after exposure to streptozotocinor alloxan in the absenceof other agents. The percentageof tumour cellsand RINm5F cells stained followingincubations with the drugs alone were: Streptozotocin, 53.3 ± 6.7% and 45.8 ± 4.9%, respectively: Alloxan, 48.2 ± 3.9% and 51.7 ± 5.3%, respectively.
Rat insulinoma drugs. Indeed, alloxan is believed to undergo a high rate of decomposition in vivo with an estimated half-life of less than 60 sec compared with approximately 5 min for streptozotocin (Rerup, 1970). It is of interest in this respect that one of the few clinical case reports using alloxan for the treatment of islet-cell carcinoma stressed the likely importance of using a rapid infusion rate (Zimmer, 1964). Using cultured t u m o u r cells and the R I N m 5 F cell line originally derived from the rat insulinoma, it was shown that alloxan and streptozotocin exerted approximately equipotent cytotoxic effects which were both time and concentration dependent. Although strictly comparable studies have not been performed, it appears that the effects and mode of action of alloxan and streptozotocin on t u m o u r B-cells are quite similar to those recently established for pancreatic B-cells (Malaisse, 1982; Okamoto and Yamamoto, 1983; Bailey and Flatt, 1986). For example, the cytotoxic actions of the drugs on pancreatic B-cells are equally rapid and apparent within the concentration range tested in the present study. More notably, the cytotoxic actions of alloxan and streptozotocin on t u m o u r B-cells and R I N m 5 F cells were decreased in the presence of various agents which are believed to facilitate the scavenging of hydroxyl free radicals (glucose, benzoate, dimethyl sulphoxide, butanol, catalase and superoxide dismutase) or to act as poly (ADP-ribose) synthetase inhibitors (theophylline, nicotinamide and benzamide) in the pancreatic B-cells. Thus it would appear from the present study that the cytotoxic actions of alloxan and streptozotocin on t u m o u r B-cells and R I N m 5 F cells are mediated through the generation of hydroxyl free radicals and D N A strand breaks as established for pancreatic B-cells (Malaisse, 1982; Okamoto and Yamamoto, 1983; Bailey and Flatt, 1986). Acknowledgements--These studies were supported by a
grant from the Cancer Research Campaign (SP 1630). The authors would like to acknowledge the technical assistance of N. Abdullah and S. J. Taylor.
REFERENCES
Bailey C. J. and Flatt P. R. (1986) Animal models of diabetes. In Recent Advances in Diabetes, (Edited by Nattrass M.), Vol. 2, pp. 71-89. Churchill Livingstone, Edinburgh. Chick W. L., Warren S., Chute R. N., Like A. A., Lauris V. and Kitchen K. C. (1977) A transplantable insulinoma in the rat. Proc. Natn. Acad. Sci. U.S.A. 74, 628~532. Editorial (1981) Insulinomas. Br. med. J. 282, 927-928. Friesen S. R. (1982) Tumours of the endocrine pancreas. New Engl. J. Med. 306, 580-590. Flatt P. R. and Bailey C. J. (1981) Abnormal plasma glucose
297
and insulin responses in heterozygous (ob/+ ) mice. Diabetologia 7,0, 573-577. Flatt P. R., Bailey C. J., Gray C. and Swanston-Flatt S. K. (1986b) Metabolic effects of radiation-induced rat insulinoma at pancreatic, hepatic and subscapular transplantation sites. Comp. Biochem. Physiol. 8.~A, 183-186. Flatt P. R., Tan K. S., Bailey C. J., Powell C. J., SwanstonFlatt S. K. and Marks V. (1986a) Effects of transplantation and resection of a radiation-induced rat insulinoma on glucose homeostasis and the endocrine pancreas. Br. J. Cancer 54, 685-692. Flatt P. R., Tan K. S., Swanston-Flatt S. K., Webster J. D. and Marks V. (1987) Metabolic effects and secretory properties of a radiation-induced transplantable rat insulinoma. Comp. Biochem. Physiol. In press.
Gazdar A. F., Chick W. L., Oie H. K., Sims H. L., King D. L., Weir G. C. and Lauris V. (1980) Continuous, clonal, insulin- and somatostatin-secreting cell lines established from a transplantable rat islet cell tumour. Proc. Natn. Acad. Sci. U.S.A. 77, 3519-3523. Grankvist K., Lernmark A. and T/iljedal I.-B. (1977) Alloxan cytotoxicity in vitro: Microscope photometric analyses of trypan blue uptake by pancreatic islet ceils in suspension. Biochem. J. 162, 19-24. Grankvist K., Marklund S., Sehlin J. and T/iljedal I.-B. (1979) Superoxide dismutase, catalse and scavengers of hydroxyl radical protect against the toxic action of alloxan on pancreatic islet cells in vitro. Biochem. J. 182, 17-25. Hedeskov C. J. (1980) Mechanism of glucose-induced insulin secretion. Physiol. Rev. 60, 44~509. Johnson B. F. and Wolff F. W. (1970) Trial of mannoheptulose in man. Metabolism 19, 354-362. Long R. S., Adrian T. E., Brown M. R., Rivier J. E., Barnes A. J., Mallinson C. N., Vale W., Christofides N. D. and Bloom S. R. (1979) Suppression of pancreatic endocrine tumour secretion by long acting somatostatin analogue. Lancet 2, 764-767. Malaisse W. J. (1982) Alloxan toxicity to the pancreatic B-cell. A new hypothesis. Biochem. Pharmac. 31, 3527-3534. Marks V. and Rose F. C. (1981) Hypoglycaemia, 2nd edn. Blackwell Scientific Publications, Oxford. Okamoto H. and Yamamoto H. (1983) DNA strand breaks and poly (ADP-ribose) synthetase activation in pancreatic islets--a new aspect to development of insulindependent diabetes and pancreatic B-cell tumours. In ADP Ribosylation, DNA Repair and Cancer, (Edited by Miwa, M.), pp. 297-308. Tokyo/VNU Science Press, Utrecht. Rerup C. C. (1970) Drugs producing diabetes through damage of the insulin secreting ceils. Pharmac. Rev. 22, 485-518. Samols E., Weir G. C. and Bonner-Weir S. (1983) Intraislet insulin-glucagon-somatostatinrelationships. In Glucagon II, (Edited by Lef6bvre, P. J.), pp. 133-173, SpringerVerlag, Berlin. Stevens J. F. (1971) Determination of glucose by an automatic analyser. Clinica chim. Acta 32, 199-201. Zimmer F. E. (1964) Islet cell carcinoma treated with alloxan: Associated with calcified hepatic metastases and thyrotoxic myopathy. Ann. Intern. Med. 61, 543-549.
0306-3623/87 $3.00 + 0.00 Copyright © 1987 Pergamon Journals Ltd
EFFECTS OF CYTOTOXIC DRUGS A N D INHIBITORS OF INSULIN SECRETION ON A SERIALLY TRANSPLANTABLE RAT INSULINOMA A N D CULTURED RAT INSULINOMA CELLS PETER R. FLATT, SARA K. SWANSTON-FLATT, KIM S. TAN and VINCENT MARKS Department of Biochemistry, University of Surrey, Guildford, Surrey GU2 5XH, U.K. [Tel: (0483) 571281] (Received 17 July 1986)
Abstract--1. The effects of cytotoxic drugs and inhibitors of insulin secretion were examined in vivo in rats with a radiation-induced transplantable insulinoma, and in vitro using cultured rat insulinoma cells and the derived RINm5F insulin-secreting cell line. 2. Administration of diazoxide to insulinoma-bearing rats resulted in a transient decrease of plasma insulin with a temporary rise of glucose concentrations. Mannoheptulose and somatostatin failed to affect the marked hyperinsulinaemia and hypoglycaemia. 3. Streptozotocin produced a rapid and sustained decrease of insulin concentrations in insulinomabearing rats, accompanied by a progressive elevation of plasma glucose. Administration of alloxan failed to affect circulating insulin or glucose concentrations. 4. In vitro, streptozotocin and alloxan exerted approximately equipotent time-dependent and concentration-dependent cytotoxic effects on insulinoma cells and RINm5F ceils as established by cell staining with trypan blue. The cytotoxic actions of both drugs were decreased by agents believed to scavenge free radicals or to act as inhibitors of poly(ADP-ribose) synthetase. 5. The results suggest that the cytotoxic actions of streptozotocin and alloxan on rat insulinoma cells and RINm5F ceils are mediated by the generation of hydroxyl free radicals and DNA strand breaks. The ineffectiveness of alloxan in insulinoma-bearing rats probably reflects the high rate of decomposition of the drug in vivo.
INTRODUCTION The serially transplantable New England Deaconess Hospital ( N E D H ) rat insulinoma developed in one of a pair of parabiont rats by X-ray irradiation (Chick et al., 1977) provides a useful animal model of insulinoma-hypoglycaemia. Routine subscapular transplantation of tumour fragments in inbred N E D H rats of the Surrey subline consistently produces an insulinoma syndrome characterized by rapid tumour growth and defective insulin secretion, resulting in marked hyperinsulinaemia, severe hypoglycaemia and neuroglycopeenic coma within one month (Flatt et al., 1986a,b, 1987). Studies in N E D H rats with the transplantable insulinoma have indicated the potential effectiveness of surgery in reversing hypoglycaemia, but tumour recurrence was invariably observed within 2 months (Flatt et al., 1986b). This clearly illustrates some of the potential problems associated with the treatment of malignant insulinomas. In addition to surgical intervention, treatment of hypoglycaemia due to insulin-secreting tumours in man involves the use of inhibitors of insulin secretion or the administration of cytotoxic drugs (Editorial, 1981; Marks and Rose, 1981; Friesen, 1982). In the present study, we have examined the effects of mannoheptulose, somatostatin and diazoxide (inhibitors of insulin srcretion), and of alloxan and streptozotocin (cytotoxic drugs) on plasma insulin and glucose concentrations of
insulinoma-bearing rats. The agents selected have all been used clinically to treat insulinoma, although with varying degrees of success (Editorial, 1981; Marks and Rose, 1981; Friesen, 1982). In addition to in vivo studies in insulinoma-bearing rats, the cytotoxic actions of alloxan and streptozotocin have been evaluated in vitro using cultured rat insulinoma cells and the R I N m 5 F cell line originally derived from the transplantable rat insulinoma (Gazdar et al., 1980). MATERIALS AND METHODS Animals Male inbred albino NEDH (New England Deaconess Hospital) rats (Rattus norvegicus albinus) from the colony at the University of Surrey carrying a serially transplantable radiation-induced insulinoma (Chick et al., 1977) were used at 14-16 weeks of age. The origin of these rats has been described elsewhere (Flattet al., 1986b). The rats were housed in an air-conditioned room at 22 + 2°C with a lighting schedule of 12hr light (0700--1900hr) and 12hr dark. A standard pellet diet (Spratts Laboratory Diet 1, Lillico Ltd, Reigate, U.K.) and tap water were supplied ad libitum. Transplantation and in vivo studies Single subcutaneous subscapular tumours (subline SFA/I) from donor NEDH rats were used as the source of tumour fragments for transplantation. Each tumour was rapidly excised, the capsule was removed and the contents finely minced. Recipient rats were lightly anaesthetized with
293
294
PETER R. FLATT et al.
ether, and 0.1 ml of minced tumour was implanted subcutaneously into the subscapular region using a 16 gauge needle. The effects of cytotoxic drugs and established inhibitors of pancreatic B-cell insulin secretion were examined in groups of insulinoma-bearing rats at 18-22 days after transplantation. The following agents were administered by intraperitoneal injection in a volume corresponding to 5 ml/kg body wt: diazoxide (Glaxo Group Research, Ware, U.K.; 100mg/kg body wt in 0.9% NaC1), somatostatin (Sigma Chemical Co. Ltd, Poole, U.K.; 1 mg/kg body wt in 0.9% NaCI), mannoheptulose (Sigma; 2 g/kg body wt in distilled water), alloxan monohydrate (Sigma; 200mg/kg body wt in 0.9% NaC1) and streptozotocin (Upjohn Chemical Co. Ltd, Kalamazoo, U.S.A.; 100mg/kg in 0.5M sodium citrate buffer, pH 4.5). Alloxan and streptozotocin were dissolved in chilled buffers immediately prior to rapid injection. A second dose of alloxan or streptozotocin (100 mg/kg or 50 mg/kg body wt, respectively) was administered to the rats at 50 hr after the initial injection. The doses and mode of administration of alloxan and streptozotocin were chosen on the basis of their established ability to induce pancreatic B-cell destruction in the rat (Rerup, 1970). Plasma glucose (Stevens, 1971) and insulin (Flatt and Bailey, 1981) were determined immediately before and at the times indicated in the Figures after administration of each test agent. Blood samples were taken from the tail tip of fed conscious rats at 0900-1400 hr. The plasma was separated and stored at - 2 0 ° C until analysis. Cell culture and cytotoxicity studies
Tumors weighing 2.2-7.9 g were excised from insulinomabearing rats, stripped of surrounding capsule and used to prepare tumour B-cell suspensions by gentle mechanical disruption. Isolated tumour cells were cultured in suspension for 3 days at 37°C in a humidified atmosphere of 5% CO 2. The culture medium was RPMI-1640 (Gibco Europe Ltd, Paisley, U.K.) containing 11.1 mmol/1 glucose, 10% foetal calf serum with added antibiotics (100 U/ml penicillin and 0.1 mg/ml streptomycin; Gibco Europe Ltd). After culture, cell viability assessed by trypan blue exclusion was increased from 21% in freshly isolated cells to 72%. In addition to cultured tumour cells, cytotoxicity studies were performed using the insulin-secreting RINm5F cell line established from the transplantable rat insulinoma (Gazdar et al., 1980). The cell line was kindly provided by Dr A. J. Bone (Edinburgh, U.K.) by courtesy of Professor C. Hellerst6m (Uppsala, Sweden). The RINm5F cells were maintained in monolayer culture under the same conditions as tumour B-cells. Confluent RINm5F cell cultures were rinsed in Ca2+-free and Mg2÷-free Hanks Balanced Salt Solution to detach the cells prior to experimentation. At this time, cell viability assessed by trypan blue exclusion was 80%.
Cytotoxicity studies were performed by incubating approximately 2.5 x 106 tumour B-cells or RINm5F cells in 250#1 of modified Krebs Ringer bicarbonate buffer (pH 7.4), containing 20mM N-2-hydroxyethylpiperazineN'-2-ethanesulfonic acid (HEPES), 115 mM NaCI, 24 mM NaHCO3, 4.7 mM KCI, 2.56 mM CaC12, 1.2 mM MgSO 4, 1.2mM KH2PO4, 1.4mM D-glucose and 5 mg/ml bovine serum albumin. The buffer was supplemented with alloxan or streptozotocin and the test agents at the concentrations indicated in the Figures and Table. Alloxan and streptozotocin were added from stock solutions of 250mM (0.4-12.8/21) which were prepared in 10mM HCI and 10 mM citrate buffer (pH 4.5) respectively immediately prior to the commencement of cell incubations. The addition of alloxan or streptozotocin did not noticeably alter the pH of the incubation medium. The cells were incubated for 10-60 rain in a water bath at 3T'C. At the end of incubation, a 50/21 aliquot of suspended cells was added to 150/~1 unsupplemented buffer, followed by an equal volume of chilled trypan blue solution (British Drug Houses Ltd, Poole, U.K.; 2mg/ml in 0.9% NaC1). After thorough mixing, the cells were counted using a haemocytometer. The frequency of cells with trypan blue-stained nuclei was assessed by counting approximately 200 randomly selected cells under a 40 × objective lens in a light microscope. The validity of this counting procedure has been demonstrated previously (Grankvist et al., 1977, 1979). It was checked in control experiments without cytotoxic drugs that addition of 10raM HC1 or 10mM citrate buffer (pH4.5) to the incubation medium did not affect cell viability as assessed by exclusion of trypan blue. Statistical analysis
Groups of data are presented as means + SEM. Statistical evaluation was performed using Student's paired and unpaired t-tests. Differences were considered to be significant for P < 0.05. RESULTS Studies on transplantable rat insulinorna in vivo
I m p l a n t a t i o n of small fragments of the serially t r a n s p l a n t a b l e i n s u l i n o m a resulted in the develo p m e n t of a single s u b s c a p u l a r t u m o u r in each rat by 18-22 days. As s h o w n in Figs 1 a n d 2, this was a c c o m p a n i e d by a m a r k e d increase o f p l a s m a insulin (range, 11.3-18.9 ng/ml) a n d decrease o f plasma glucose (range, 1.9-2.9 mmol/1) c o m p a r e d with control rats (2.3 4- 0.5 ng/ml a n d 6.7 4- 0.3 mmol/1, n = 8; d a t a not illustrated). A d m i n i s t r a t i o n of m a n n o heptulose or s o m a t o s t a t i n to i n s u l i n o m a - b e a r i n g rats 24
8
4
2b~ 2'-.
16
I
1
I
0
-
-0
1
60
120
0
I
J
60 120 Minutes r~inutes Fig. 1. Effects of diazoxide (O O; 100 mg/kg), mannoheptulose (E] Fq; 2 g/kg) and somatostatin (O O; 1 mg/kg) on plasma glucose and insulin concentrations of insulinoma-bearing rats at 18-22 days after implantation of tumour fragments. Values are mean___ SEM of groups of 6-8 rats. *P < 0.05 - 0.001 compared with time zero. O
Rat insulinoma 20
295
24
15 16 10
0--0-0 I I 0 2
0
0---0"'0""-'0 I ~ I I 4 24 48
1 72
,. 0 I 96
0
I
I ~
I
I
I
I
2
4
24
48
72
96
Hours
Hours
Fig. 2. Effects of streptozotocin (O O; 100 mg/kg at time zero and 50 mg/kg at 50 hr) and alloxan monohydrate (O O; 200 mg/kg at time zero and 100 mg/kg at 50 hr) on plasma glucose and insulin concentrations of insulinoma-bearing rats at 21 days after implantation of tumour fragments. Values are mean + SEM of groups of 6 rats. *P < 0.05-0.001 compared with time zero. did not affect circulating insulin or glucose concentrations (Fig. 1). However, diazoxide produced a transient decrease of plasma insulin associated with a temporary rise of glucose concentrations (Fig. 1). As shown in Fig. 2, administration of alloxan to insulinoma-bearingrats failed to affect plasma insulin or glucose concentrations, although there was a tendency towards a less severe hyperinsulinaemia after the drug. In contrast, streptozotocin produced a rapid, marked and sustained decrease of circulating insulin concentrations accompanied by a progressive elevation of plasma glucose (Fig. 2). By 48-96 hr, the glucose concentrations of streptozotocin-treated rats were significantly greater than those of untransplanted control rats. Studies on cultured rat tumour cells a n d R I N m S F cell line in vitro
Exposure of isolated cultured tumour cells and RINm5F cells to 2 m M alloxan or 2 m M streptozotocin in vitro resulted in a marked and timedependent decrease of cell viability as assessed by cell staining with trypan blue (Fig. 3). The effects of the two cytotoxic drugs were similar irrespective of the use of tumour cells or RINm5F cells. As shown in Fig. 4, incubation of either cell type for 20 min with
different concentrations of alloxan or streptozotocin revealed that the cytotoxic effects of the drugs were dose-dependent. There were no significant differences in the potency of alloxan and streptozotocin over the concentration range tested. Table 1 shows the effects of various agents on the cytotoxic action of 8 mM alloxan or 8 mM streptozotocin on tumour cells and RINm5F cells during 20 rain incubations. These agents were selected on the basis of their established protective effects against alloxan- and/or streptozotocin-cytotoxicity of insulin-secreting pancreatic B-cells (Malaisse, 1982; Okamoto and Yamamoto, 1983). Each agent tested appeared to offer some degree of protection against the toxic effects of alloxan or streptozotocin on either tumour cells or RINm5F cells. There were no substantial differences between protective effects using the two cell types. In contrast to the action of streptozotocin, the effects of alloxan on cell viability were not diminished by benzamidine, butanol or catalase at the concentrations employed. DISCUSSION
Consistent with previous reports (Flatt et al., 1986a,b, 1987), transplantation of small fragments of
Turnout cells
RI NmSF cells
24
8~ 16
t.~
8
.5
O.
20 Minutes
40
60
0
20
40
60
Minutes
Fig. 3. Effects of time of exposure to streptozotocin (O O; 2retool/l) and alloxan ( 0 ....... O; 2 mmol/1)on the frequency of tumour cells or RINm5F cells stained with trypan blue. Values are presented as mean + SEM of 5 separate experiments. A significantincrease (P < 0.05--0.01)in the percentage of cells stained was evident by 10 min. O.P. 18/~-F
PETER R. FLATT et al.
296
RINm5F
Ttrnour cells
cells
40
i .~
lO
O
4
8
I
I
12
16
Drug concentration (mmol/l)
O Drug
4 concentration
8
12
16
(m~i/l)
Fig. 4. Effects of 20 min exposure to different concentrations of streptozotocin (O 0 ) and alloxan (O O) on the frequency of tumour cells or RINm5F cells stained with trypan blue. Values are presented as mean + SEM of 5 experiments. A significant increase (P < 0.05) in the percentage of cells stained was evident by 1 mmol/1. a serially transplantable insulinoma into N E D H rats o f the Surrey subline, consistently resulted in m a r k e d hyperinsulinaemia and hypoglycaemia by 18-22 days. A t this time, the t u m o u r B-cells do not r e s p o n d to glucose, arginine or glibenclamide, although responsiveness to agents such as glucagon and epinephrine which m o d u l a t e insulin release t h r o u g h the adenylate cyclase-vyclic A M P system appears to be retained (Flatt et al., 1987). In the present study, m a n n o h e p t u l o s e and somatostatin which are established inhibitors o f pancreatic B-cell insulin release (Hedeskov, 1980; Samols et al., 1983) failed to affect plasma insulin and glucose concentrations o f insulinoma-bearing rats. This is in general agreement with the limited effectiveness o f these agents in patients with insulinoma ( J o h n s o n and Wolff, 1979; M a r k s and Rose, 1981), although a somatostatin analogue has been used with some success (Long et al., 1979). Diazoxide is by far the most effective inhibitor o f insulin secetion employed for the management o f clinical insulinoma (Editorial, 1981; M a r k s and Rose, 1981). As d e m o n s t r a t e d by the present study, an inhibitory action on insulin release with an a c c o m p a n y i n g elevation o f circulating glucose was observed in insulinoma-bearing rats. How-
ever, as in man, the effect was only transient indicating that successful treatment o f insulinoma with diazoxide can only be achieved by repeated dosage with the drug. A d m i n i s t r a t i o n o f the B-cytotoxic drugs alloxan and streptozotocin revealed a m a r k e d divergence in their effectiveness in reversing the hyperinsulinaemia and hypoglycaemia o f insutinoma-bearing rats. Thus whereas streptozotocin rapidly reversed the metabolic abnormalities, alloxan was without effect on plasma insulin and glucose concentrations. This finding is in agreement with the results o f clinical studies which have s h o w n that streptozotocin but not alloxan is highly effective in the treatment o f malignant endocrine pancreatic t u m o u r s (Marks and Rose, 1981). Since b o t h drugs are established pancreatic B-cell toxins in the rat (Rerup, 1970), their different effectiveness may be a consequence o f the altered function o f the t u m o u r B-cells. However, as indicated by the results o f in vitro cytoxicity studies, cultured rat insulinoma cells and R I N m 5 F cells were equally sensitive to alloxan and streptozotocin. This suggests that the different effectiveness o f alloxan and streptozotocin in insulinoma-bearing rats is due to differences in the stability and in vivo clearance o f the
Table 1. Effectsof various agents on the frequency of tumour cells or RINm5F cells stained with trypan blue after 20 min exposure to 8 mmol/l streptozotocin or 8 mmol/l alloxan Change in the percentage of cells stained with trypan blue Agent Tumour cells RINm5F cell line (concentration) Streptozotocin Alloxan Streptozotocin Anoxan Glucose (16.7 mM) - 17.3 ± 5.2* - 15.2 + 6.6 - 18.4 _+4.2* - 15.7 ± 4.7* Theophylline (5 mM) -30.2 + 5.8*** -25.1 ± 6.6** - 16.2 ± 7.2 -21.5 + 4.4*** Nicotinamide (10 mM) -29.1 + 6.0*** -20.3 + 3.3*** -22.4 ± 5.4** 13.4 ± 3.9* Sodium benzoate (50mM) -31.8 _+6.1"** - 16.1 ± 8.5 - 17.7 ± 4.9* 15.7 + 3.3*** Benzamide (1 mM) -31.0_+4.3"*** -21.7 ± 9.0 17.7+6.4" -3.2+7.7 Dimethylsulphoxide (28 rnM) -25.4 ± 9.8* - 19.8 ± 6.3* - 16.0 ± 7.5 -5.4 + 5.4 Butanol (22 mM) -24.2 _ 9.1" -- 12.3 ± t0.7 - 19.2 ± 5.7* --3.9 ± 5.4 Catalase (250 U/ml) 30.7 + 3.2**** 19.7 ± 9.2 18.8 + 3.6*** 13.4 ± 10.0 Superoxide dismutase (2000 U/ml) -24.5 ± 8.4* -21.4 ± 5.2** -21.8 ± 2.2**** - I 1.8 ± 8.1 Values are presented as mean ± SEM of 5 separate experiments. *P < 0.05, **P < 0.02, ***P < 0.01 and ****P < 0.001 compared with the frequency of cells stained with trypan blue after exposure to streptozotocinor alloxan in the absenceof other agents. The percentageof tumour cellsand RINm5F cells stained followingincubations with the drugs alone were: Streptozotocin, 53.3 ± 6.7% and 45.8 ± 4.9%, respectively: Alloxan, 48.2 ± 3.9% and 51.7 ± 5.3%, respectively.
Rat insulinoma drugs. Indeed, alloxan is believed to undergo a high rate of decomposition in vivo with an estimated half-life of less than 60 sec compared with approximately 5 min for streptozotocin (Rerup, 1970). It is of interest in this respect that one of the few clinical case reports using alloxan for the treatment of islet-cell carcinoma stressed the likely importance of using a rapid infusion rate (Zimmer, 1964). Using cultured t u m o u r cells and the R I N m 5 F cell line originally derived from the rat insulinoma, it was shown that alloxan and streptozotocin exerted approximately equipotent cytotoxic effects which were both time and concentration dependent. Although strictly comparable studies have not been performed, it appears that the effects and mode of action of alloxan and streptozotocin on t u m o u r B-cells are quite similar to those recently established for pancreatic B-cells (Malaisse, 1982; Okamoto and Yamamoto, 1983; Bailey and Flatt, 1986). For example, the cytotoxic actions of the drugs on pancreatic B-cells are equally rapid and apparent within the concentration range tested in the present study. More notably, the cytotoxic actions of alloxan and streptozotocin on t u m o u r B-cells and R I N m 5 F cells were decreased in the presence of various agents which are believed to facilitate the scavenging of hydroxyl free radicals (glucose, benzoate, dimethyl sulphoxide, butanol, catalase and superoxide dismutase) or to act as poly (ADP-ribose) synthetase inhibitors (theophylline, nicotinamide and benzamide) in the pancreatic B-cells. Thus it would appear from the present study that the cytotoxic actions of alloxan and streptozotocin on t u m o u r B-cells and R I N m 5 F cells are mediated through the generation of hydroxyl free radicals and D N A strand breaks as established for pancreatic B-cells (Malaisse, 1982; Okamoto and Yamamoto, 1983; Bailey and Flatt, 1986). Acknowledgements--These studies were supported by a
grant from the Cancer Research Campaign (SP 1630). The authors would like to acknowledge the technical assistance of N. Abdullah and S. J. Taylor.
REFERENCES
Bailey C. J. and Flatt P. R. (1986) Animal models of diabetes. In Recent Advances in Diabetes, (Edited by Nattrass M.), Vol. 2, pp. 71-89. Churchill Livingstone, Edinburgh. Chick W. L., Warren S., Chute R. N., Like A. A., Lauris V. and Kitchen K. C. (1977) A transplantable insulinoma in the rat. Proc. Natn. Acad. Sci. U.S.A. 74, 628~532. Editorial (1981) Insulinomas. Br. med. J. 282, 927-928. Friesen S. R. (1982) Tumours of the endocrine pancreas. New Engl. J. Med. 306, 580-590. Flatt P. R. and Bailey C. J. (1981) Abnormal plasma glucose
297
and insulin responses in heterozygous (ob/+ ) mice. Diabetologia 7,0, 573-577. Flatt P. R., Bailey C. J., Gray C. and Swanston-Flatt S. K. (1986b) Metabolic effects of radiation-induced rat insulinoma at pancreatic, hepatic and subscapular transplantation sites. Comp. Biochem. Physiol. 8.~A, 183-186. Flatt P. R., Tan K. S., Bailey C. J., Powell C. J., SwanstonFlatt S. K. and Marks V. (1986a) Effects of transplantation and resection of a radiation-induced rat insulinoma on glucose homeostasis and the endocrine pancreas. Br. J. Cancer 54, 685-692. Flatt P. R., Tan K. S., Swanston-Flatt S. K., Webster J. D. and Marks V. (1987) Metabolic effects and secretory properties of a radiation-induced transplantable rat insulinoma. Comp. Biochem. Physiol. In press.
Gazdar A. F., Chick W. L., Oie H. K., Sims H. L., King D. L., Weir G. C. and Lauris V. (1980) Continuous, clonal, insulin- and somatostatin-secreting cell lines established from a transplantable rat islet cell tumour. Proc. Natn. Acad. Sci. U.S.A. 77, 3519-3523. Grankvist K., Lernmark A. and T/iljedal I.-B. (1977) Alloxan cytotoxicity in vitro: Microscope photometric analyses of trypan blue uptake by pancreatic islet ceils in suspension. Biochem. J. 162, 19-24. Grankvist K., Marklund S., Sehlin J. and T/iljedal I.-B. (1979) Superoxide dismutase, catalse and scavengers of hydroxyl radical protect against the toxic action of alloxan on pancreatic islet cells in vitro. Biochem. J. 182, 17-25. Hedeskov C. J. (1980) Mechanism of glucose-induced insulin secretion. Physiol. Rev. 60, 44~509. Johnson B. F. and Wolff F. W. (1970) Trial of mannoheptulose in man. Metabolism 19, 354-362. Long R. S., Adrian T. E., Brown M. R., Rivier J. E., Barnes A. J., Mallinson C. N., Vale W., Christofides N. D. and Bloom S. R. (1979) Suppression of pancreatic endocrine tumour secretion by long acting somatostatin analogue. Lancet 2, 764-767. Malaisse W. J. (1982) Alloxan toxicity to the pancreatic B-cell. A new hypothesis. Biochem. Pharmac. 31, 3527-3534. Marks V. and Rose F. C. (1981) Hypoglycaemia, 2nd edn. Blackwell Scientific Publications, Oxford. Okamoto H. and Yamamoto H. (1983) DNA strand breaks and poly (ADP-ribose) synthetase activation in pancreatic islets--a new aspect to development of insulindependent diabetes and pancreatic B-cell tumours. In ADP Ribosylation, DNA Repair and Cancer, (Edited by Miwa, M.), pp. 297-308. Tokyo/VNU Science Press, Utrecht. Rerup C. C. (1970) Drugs producing diabetes through damage of the insulin secreting ceils. Pharmac. Rev. 22, 485-518. Samols E., Weir G. C. and Bonner-Weir S. (1983) Intraislet insulin-glucagon-somatostatinrelationships. In Glucagon II, (Edited by Lef6bvre, P. J.), pp. 133-173, SpringerVerlag, Berlin. Stevens J. F. (1971) Determination of glucose by an automatic analyser. Clinica chim. Acta 32, 199-201. Zimmer F. E. (1964) Islet cell carcinoma treated with alloxan: Associated with calcified hepatic metastases and thyrotoxic myopathy. Ann. Intern. Med. 61, 543-549.