the Harwood Academic Publishers imprint, part of The Gordon and Breach Publishing Group. Printed in Malaysia. Aminoguanidine Exerts a fi-cell Function-preserving Effect in High Glucose-cultured fi-cells (INS-l)

We investigated the effects of aminoguanidine (AG) on ]/-cell functions in an insulin secreting cell line (INS-l). Culture with 27mM glucose for one week markedly decreased both insulin release and insulin content compared to culture in 0.8 mM or 3.3 mM glucose. Relative to culture at 27mM glucose alone, the co-exposure to lmM AG almost doubled basal as well as glucose or 25mM KCl-stimulated insulin release and increased insulin content by 42%. AG failed to affect release and content in cells cultured at 0.8 or 3.3mM glucose. Preproinsulin mRNA content in 27mM glucose-cultured cells was 52% suppressed compared to 0.8mM glucose-cultured cells, and AG treatment partially counteracted this decline. Advanced glycosylation end product (AGE)-associated fluorescence (370nm excitation and 440 nm emission) of cells’ extracts did not differ between 27mM and 0.8mM glucose-cultured cells after 1 week of culture and fluorescence was unaffected by AG. Accumulation of nitrite into culture media was markedly increased from 27mM glucose-cultured cells, and this accumulation was 33% suppressed by AG. In conclusion, AG partially protects against glucotoxic effects in INS-1 cells. These beneficial effects may involve a decrease in early glycation products and/or nitric oxide synthase (NOS) activity. The effects which were obtained after one week of high glucose exposure may supplement AGE-associated effects seen after chronically elevated glucose.


INTRODUCTION
Aminoguanidine (AG) is a nucleophilic hydrazine compound which potently inhibits AGE formation. [1,2] Beneficial effects of AG on diabetic complications have been reported in vivo and in vitro 2 and are associated with inhibitory effects on AGE formation. I3, 4] These effects of AG are considered strong evidence for an important role of AGEs in diabetic complications.
In our own experiments we have used AG as a probe to test for the possible influence of AGE formation in hyperglycemia-induced desensitization of pancreatic fl-cells. In these experiments, pancreatic fl-cells were cultured for a period of 6 weeks at high glucose with or without AG. Under these conditions AG suppressed AGE formation in rat pancreatic islet and concomitantly enhanced fl-cell functions such as insulin release and insulin biosynthesis. 5 These effects *Corresponding author. Tel.: +47-73868386, Fax: +47-73867546, e-mail: valdemar.grill@medisin.ntnu.no 111 were time dependent, since they were not grown in monolayer cultures as described observed after 1 but only after 6 weeks of expopreviously [12] in RPMI-1640 medium containsure to high glucose and AG. This time depending 11 mM glucose supplemented with 10 mM ence is in accord with previous short term HEPES, 10% heat-inactivated fetal calf serum, studies in which exposure to AG leads to no 2mM L-glutamine, I mM sodium pyruvate, effect [6] or inhibition [7] of insulin secretion. 50 btM fl-mercaptoethanol, 100 IU/ml penicillin AG has also been reported to exert effects and 100btg/ml streptomycin at 37C in a hurelated to the inhibition of nitric oxide synthase midified (5% CO2, 95% air) atmosphere. A pasactivity Is'91 as well as the inhibition of free sage number of around 20 were used for the radical formation from early glycation propresent experiments. Cells were seeded in 24 ducts. I10, 11] It seemed possible that AG could multi-well tissue culture dish (1.0 x 105 cells in exert these types of effects on fl-cells which are I ml of medium per well, 2 cm of diameter) and more metabolically active and dividing than cultured for 1 week in RPMI-1640 containing fl-cells in islets of adult animals. In the present 0.8, 3.3 or 27 mM glucose and with or without study we have therefore tested for effects of AG (final concentration; I mM). Before being AG on insulin secretion and biosynthesis in an added to the culture media, AG was dissolved insulinoma cell line (INS-l) during one week in redistilled water at acidic conditions (pH 3 Batch-type incubations were carried out as Louis, MO). Dextran T 70 were from Pharmacia previously described. I121 Briefly, the cells in each (Uppsala, Sweden). RNase-A and RNase-T1 group were washed in KRB medium [13] with the were from Boehringer Mannheim (Mannheim, following composition: Na + 143 mM, K + 5.8 mM, Germany Branson Ultrasonics Corp., Danbury, CT). The sonicates were centrifuged at 10,000g for 10 minutes at 4C. Fluorescence in the supernatants was measured, using excitation at 370nm+ emission at 440 nm [16] by SPEX-1681 0.22 m spectrometer (SPEX industries, Inc. Edison, N.J.). 50 tl of the sonicate was used for assay of protein contents.

Preproinsulin mRNA
Triplicate wells of cells in each group were harvested by trypsinization and total RNA was prepared as described by Chomczynski et al. [17] The quantitative analysis of preproinsulin mRNA was achieved by a solution hybridization assay using a RNA probe radiolabeled with 3SS-UTP. [18] An in vitro synthesized 58 bp oligonucleotide corresponding to the last part of exon 3 of the rat preproinsulin II gene and flanked by sulphanilamide and 25% concentrated HBPO4.
The reaction was carried out at 60C for I min.
Nitrite concentrations were measured as the absorbance at 546 nm in a spectrophotometer. [19] Insulin Assay Insulin was measured by RIA using rat insulin as standard, monoiodinated porcine insulin as tracer and antibody raised in our laboratory against porcine insulin. Antibody-bound insulin was separated from free insulin using Dextran T70-coated charcoal. 21  Danbury, CT) as previously described. [21] Presentation of Results All results are expressed as mean+SEM. Analyses between groups were carried out as appropriate by Student's t-test or one-way analysis of variance (ANOVA) with Student-Newman-Keuls' test. P value of < 0.05 was considered significant.

Cell Proliferation
The incorporation of [methyl-3H]thymidine into cells cultured at 0.8mM glucose was not significantly higher than in cells cultured at 27 mM glucose ( One week of culture with 27 mM glucose markedly reduced basal as well as 27 mM glucoseor 25 mM KCl-stimulated insulin release compared to culture with 0.8 or 3.3 mM glucose (Fig. 1). Co-culture with AG did not affect insulin release from 0.8 or 3.3 mM glucose-cultured cells. Culture with AG considerably enhanced release from 27 mM glucose-cultured cells. The effect amounted to a doubling of basal and 27 mM glucoseor 25 mM KCl-stimulated secretion (Fig. 1). However, AG failed to revive an insulin response to acute stimulation with 27mM glucose, seen after culture at 3.3mM glucose.   Insulin content of 27mM glucose-cultured cells was 90% decreased compared to cells cultured at 0.8 mM glucose and 87% decreased compared to 3.3 mM glucose. Addition of AG to cells cultured at 27 mM glucose significantly (P < 0.05) enhanced, but far from normalized, insulin content (by 42%). In contrast, addition of AG failed to affect insulin content in cells cultured at 0.

Insulin mRNA Content
Insulin mRNA content in 27mM glucose-cultured cells was 52% reduced compared to cells cultured at 0.8 mM glucose (Fig. 2). AG treatment significantly (P<0.05) increased insulin mRNA content up to 67% of that of 0.8mM glucose-cultured cells.

Medium Nitrite Concentration
Medium nitrite was measured as an indicator of NO synthesis from INS-1 cells. Nitrite in culture media from 3.3 mM glucose-cultured cells was not detectable in our assay system. In contrast, 27mM glucose induced marked accumulation of nitrite. Co-culture with AG I mM significantly (P<0.05) suppressed nitrite accumulation by 33% (Fig. 3). week's culture with or without AG. Cells in each group were harvested and total RNA was prepared as described. [17] The quantitative analysis of preproinsulin mRNA was achieved by a solution hybridization assay using a RNA probe radiolabeled with 35S-UTP. Mean+SEM of three experiments. *P < 0.05 vs. AG (-).

DISCUSSION
The present study demonstrates that AG exerts beneficial effects on fl-cell function during shortterm culture of clonal fl-cells. The beneficial effects were seen specifically in high glucosecultured cells. In those cells AG partially restored culture-induced deficiencies in insulin release, insulin content, insulin accumulation into culture media and insulin mRNA. Previous studies on a similar time scale [6,7] failed to observe beneficial AG effects. This failure may be due to the fact that AG was not tested during high glucose conditions deleterious to fl-cell function.
Although the effects of AG were seen only in high glucose-cultured cells, the beneficial effects of AG were not linked to glucose-regulated modalities of fl-cell function. Rather, they were generalized since, for instance, basal insulin secretion was affected together with glucoseand KCl-stimulated insulin secretion. It seems possible that beneficial effects on secretion were secondary to increased insulin content and biosynthesis.
Cell proliferation estimated by [3H]-thymidine incorporation did not differ between 0.8mM and 27mM glucose-cultured cells in agreement with a previous study. [22] Neither was an effect of AG on cell growth apparent in the present study. Hence, the beneficial effects of AG could not be explained by effects on proliferation.
Effects of AG in biological systems were previously assigned to the inhibitory effects of the compound on AGE formation. [1,2] In a previous study 51 we found evidence for such an effect being operative also in pancreatic islets. Hence, in 6 week-cultures of rat pancreatic islets we found that high-glucose culture increased the formation of AGEs and that AG partly inhibited this formation along with beneficial effects on insulin secretion and insulin biosynthesis. However, in the present study we found neither an increase by high glucose culture of AGE-associated fluorescence, nor an effect of AG on this parameter. These findings were not surprising considering the long period necessary for the build-up of AGE products. [3"4] Indeed, in our experimental system of rapidly growing cells (doubling time 45 to 84 h, unpublished results) cells in new generations are exposed to high glucose for considerably shorter time than the one week total period of culture. From these observations we conclude that other effects of AG, not related to AGE formation, are responsible for the beneficial effects in high-glucose-cultured INS-1 cells. Studies  of insulin release from isolated islets. E25 It thus seems possible that NO synthesis during high glucose culture is a distinct mechanism for "glucotoxicity" supplementing other mechanisms, including AGE formation in islets.
It is also possible that AG exerts its beneficial effects on fl-cell functions by inhibition of free radicals generated by early glycation products (Amadori products or Schiff base). I101 Matsuoka et al. reported that suppression of insulin gene transcription through decrease of DNA-binding activity of PDX-1 linked to ribose-induced glycation in HIT-T15 cells was prevented by both AG and an antioxidant, N-acethylcysteine, suggesting the importance of consequent increase of reactive oxygen species (ROS) for the glucotoxicity induced by glycation. [11] Similar effects have been further documented in vitro and in vivo. [26][27][28] Our results on insulin mRNA would agree with a transcription effect. In cells cultured at 0.8 mM glucose the change to 27mM glucose paradoxically inhibited insulin secretion. It is known that a low glucose concentration can induce insensitivity to glucose by mechanisms alike to those operative during in vivo fasting. [29] Such mechanisms may have been operative also during the present conditions.
In summary, we have demonstrated that AG partially protects against glucoseinduced fl-cell dysfunction in short term culture of clonal flcells. This effect may be linked to the inhibition of free radical formation from early glycation products and/or NO synthase inhibition. These effects may supplement the beneficial effects of AG-induced inhibition of AGE formation which were evidenced in a previous study .tsl