Clock gene expression in the liver of streptozotocin-induced and spontaneous type 1 diabetic rats.

Several investigations have shown a relation between diabetes and alterations of the liver circadian clock. We investigated the diurnal expression of clock genes and clock-controlled genes (CCGs) in 3-hour intervals for a 24-h period in the livers of male streptozotocin (STZ)-treated rats, male spontaneous type 1 diabetic LEW.1AR1-iddm (Iddm) rats, and Iddm rats treated for 10 days with insulin. Hepatic mRNA was extracted, and the relative expression of clock genes (Per1, Per2, Bmal1, Clock, Cry1), as well as CCGs (Dbp, E4bp4, RevErbα, Rorα, Pparγ), was analyzed by reverse transcription followed by real-time polymerase chain reaction. Diabetic STZ and Iddm rats, as well as insulin-substituted Iddm rats, exhibited a significant diurnal expression pattern of clock genes as determined by Cosinor analysis; however, the MESOR (midline estimating statistic of rhythm) of Bmal1, Per2, and Clock transcript expression was altered in Iddm and insulin-substituted Iddm rats. The hepatic expression of the CCGs Dbp and RevErbα revealed a diurnal rhythm in all investigated groups. Insulin administration to Iddm rats normalized the enhanced MESOR in the expression of Dbp, RevErbα, and E4bp4 to the levels of normoglycemic controls. Cosinor analysis indicated no diurnal rhythm of Pparγ expression in the livers of diabetic STZ or Iddm rats or in those of insulin-substituted Iddm rats. Also, insulin substitution could not reverse the decreased MESOR of Pparγ expression in Iddm rats. In consequence of the diabetic disease, changes in the expression of clock genes and CCGs suggest alterations in the hepatic peripheral clock mechanism.

GLUT4 in the endocrine pancreas--indicating an impact in pancreatic islet cell physiology?

The glucose transporter GLUT4 is well known to facilitate the transport of blood glucose into insulin-sensitive muscle and adipose tissue. In this study, molecular, immunohistochemical, and Western blot investigations revealed evidence that GLUT4 is also located in the mouse, rat, and human endocrine pancreas. In addition, high glucose decreased and insulin elevated the GLUT4 expression in pancreatic α-cells. In contrast, high glucose increased GLUT4 expression, whereas insulin led to a reduced expression level of the glucose transporter in pancreatic ß-cells. In vivo experiments showed that in pancreatic tissue of type 2 diabetic rats as well as type 2 diabetic patients, the GLUT4 expression is significantly increased compared to the nondiabetic control group. Furthermore, type 1 diabetic rats exhibited reduced GLUT4 transcript levels in pancreatic tissue, whereas insulin treatment of type 1 diabetic animals enhanced the GLUT4 expression back to control levels. These data provide evidence for the existence of GLUT4 in the endocrine pancreas and indicate a physiological relevance of this glucose transporter as well as characteristic changes in diabetic disease.

Catecholamines are the key for explaining the biological relevance of insulin-melatonin antagonisms in type 1 and type 2 diabetes.

In this paper, we analyze the biological relevance of melatonin in diabetogenesis. As has recently been demonstrated, melatonin decreases insulin secretion via specific melatonin receptor isoforms (MT1 and MT2) in the pancreatic ß-cells. In addition, type 2 diabetic rats, as well as patients, exhibit decreased melatonin levels, whereas the levels in type 1 diabetic rats are increased. The latter effects were normalized by insulin substitution, which signifies that a specific receptor-mediated insulin-melatonin antagonism exists. These results are in agreement with several recent genome-wide association studies, which have identified a number of single nucleotide polymorphisms in the MTNR1B gene, encoding the MT2 receptor, that were closely associated with a higher prognostic risk of developing type 2 diabetes. We hypothesize that catecholamines, which decrease insulin levels and stimulate melatonin synthesis, control insulin-melatonin interactions. The present results support this assertion as we show that catecholamines are increased in type 1 but are diminished in type 2 diabetes. Another important line of inquiry involves the fact that melatonin protects the ß-cells against functional overcharge and, consequently, hinders the development of type 2 diabetes. In this context, it is striking that at advanced ages, melatonin levels are reduced and the incidence of type 2 diabetes is increased. Thus, melatonin appears to have a protective biological role. Here, we strongly repudiate misconceptions, resulting from observations that melatonin reduces the plasma insulin level, that the blockage of melatonin receptors would be of benefit in the treatment of type 2 diabetes.

The insulin-melatonin antagonism: studies in the LEW.1AR1-iddm rat (an animal model of human type 1 diabetes mellitus).

AIMS/HYPOTHESIS: It is well documented that melatonin influences insulin secretion mediated by G-protein-coupled melatonin receptor isoforms MT1 and MT2, which are present in rat and human pancreatic islets, as well as in rat insulinoma cells. Recent investigations have proven that hyperinsulinaemic Goto-Kakizaki (GK) rats, which are a rat model of type 2 diabetic rats, and humans have decreased melatonin plasma levels, whereas a streptozotocin-induced rat model of diabetes developed reduced insulin levels combined with increased melatonin levels. METHODS: Plasma levels of glucose, insulin and melatonin as well as RNA expression of pineal Aanat, Hiomt (also known as Asmt), insulin receptor, adrenoceptor ß1 and the clock genes Per1 and Bmal1 (also known as Arntl) were determined in male and female LEW.1AR1-iddm rats as well as in insulin-substituted LEW.1AR1-iddm rats. RESULTS: Severe hypoinsulinaemia in diabetic LEW.1AR1-iddm rats was associated with decreased body weight and increased melatonin plasma levels combined with mainly elevated expression of Aanat, Hiomt, pineal insulin receptor and adrenoceptor ß1. The changes were normalised by insulin substitution. Diurnal profiles of plasma melatonin and of antagonistic clock genes Per1 and Bmal1 were maintained in diabetic and insulin-substituted rats. CONCLUSIONS/INTERPRETATION: The assumed causal relation between elevated melatonin and reduced insulin levels in LEW.1AR1-iddm rats is supported by the observation that insulin substitution normalised these changes. Further support for this interpretation comes from the observation that in GK rats an increase of plasma insulin was combined with a decrease of plasma noradrenaline (norepinephrine), the most important activator of melatonin synthesis. These relationships between the noradrenergic and insulin pathway support the existence of melatonin-insulin antagonism.

Transcripts of calcium/calmodulin-dependent kinases are changed after forskolin- or IBMX-induced insulin secretion due to melatonin treatment of rat insulinoma beta-cells (INS-1).

The objective of the present study was to examine the effects of melatonin on transcripts of isoforms of calcium/calmodulin-dependent protein kinases in rat insulinoma beta-cells INS-1. Investigations show that calcium/calmodulin-dependent kinase IV and calcium/calmodulin-dependent kinase 2d are expressed in human and rat pancreatic islets and INS-1 cells. By application of either forskolin or 3-isobutyl-1-methylxanthine for 6 hours, calcium spiking was evoked and the release of insulin was increased. The expression of the calcium/calmodulin-dependent kinase IV and calcium/calmodulin-dependent kinase 2d transcripts was significantly increased due to forskolin or 3-isobutyl-1-methylxanthine. Acute melatonin treatment (6 h) in the presence of either forskolin or 3-isobutyl-1-methylxanthine caused a significant decrease in insulin release and induced significant downregulation of calcium/calmodulin-dependent kinase IV and calcium/calmodulin-dependent kinase 2d transcripts in INS-1 batch cultures. The attenuating effect of melatonin on transcripts could be almost completely reversed by preincubation with the melatonin receptor antagonist luzindole. Thus, the insulin-inhibiting effect of melatonin in INS-1 cells is associated with significant changes in transcripts of calcium-signaling components suggesting that melatonin influences gene expression of components, which are known to be involved in insulin secretion or insulin gene expression.

Circadian and age-dependent expression patterns of GLUT2 and glucokinase in the pancreatic beta-cell of diabetic and nondiabetic rats.

Alterations in glucose sensing are well-known in both humans and animal models of non-insulin-dependent diabetes mellitus. However, the circadian- and age-dependent expression of glucose-sensing genes has not previously been investigated in vivo. In the present paper, we show a progressive loss of beta-cell GLUT2-mRNA and, by immunocytochemistry, a gain of soluble, cytoplasmic GLUT2-protein in Goto-Kakizaki rat islets. We report that GLUT2-mRNA shows significant diurnal variation, which is stronger in metabolically healthy rats. We also demonstrate the significant diurnal variation of glucokinase-mRNA, with higher levels in the pancreas of 6-week-old Goto-Kakizaki rats than in Wistar rats. This leads to a maximum glucose phosphorylation capacity in-phase with food intake, enhanced glucose-stimulated insulin secretion, and prevents postprandial hyperglycemia. Perfusion experiments showed a reduction in glucose-stimulated insulin secretion in Goto-Kakizaki rat islets with an impaired first phase. Hyperglycemia and hypoinsulinemia in newborn and up to 3-week-old Goto-Kakizaki rats are thus probably due to reduced pancreatic beta-cell content, reduced beta-cell insulin content and impaired glucose sensing. The de-compensation of the metabolic situation in 42-week-old Goto-Kakizaki rats is likely to be caused by beta-cell destruction accompanied by negligible accumulation of GLUT2 in the cell membrane and further reduction of glucokinase expression.

Circadian changes of ether-a-go-go-related-gene (Erg) potassium channel transcripts in the rat pancreas and beta-cell.

Evidence has previously been presented that circadian rhythms play a role in islet hormone secretion. Here, RT-PCR was used to monitor the circadian expression of ether-a-go-go-related gene (Erg) potassium channel isoforms and Erg1 splice variants. Immunohistochemistry was used to identify the pancreatic distribution patterns of ERG1a and ERG1b, as well as ERG2 and ERG3. The influence of ERG on insulin secretion was monitored by perfusion of rat INS-1 beta-cells with the blockers E-4031 and rBeKm-1. We identified Erg1a, Erg1b, Erg2 and Erg3 transcripts in islets and INS-1 cells. Immunohistochemistry showed differential expression of ERG isoforms in the islet. Ca(2+) imaging and electrophysiological recordings of INS-1 cells during ERG blocking by E-4031 indicated functional ERG channels. Serum shock treatment of INS-1 cells elicited a time-dependent expression response for Erg transcripts. These results add to the current understanding of the function of ERG channels in beta-cells and the circadian secretion processes of insulin.

Did the gradual loss of GLUT2 cause a shift to diabetic disorders in the New Zealand obese mouse (NZO/Hl)?

The New Zealand obese mouse (NZO/Hl) is characterised by hereditary obesity and type-2 diabetes, including insulin resistance, hyperinsulinaemia, and glucose intolerance. In other diabetic models, it has been revealed that the proper functioning of the glucose transporter isoform 2 (GLUT2) is essential for adequate secretion of insulin. The aim of this study was to compare the distribution of islet cells and GLUT2, as well as the expression of GLUT2-mRNA, in the pancreas of NZO mice and metabolically unimpaired NMRI (Naval Medical Research Institute) mice. Pancreas tissue was obtained from different stages of development. For molecular determination of the expression level of GLUT2-mRNA, total-RNA was extracted from the pancreas and analysed by quantitative real-time RT-PCR. All investigated NZO mice displayed increased weight, elevated hyperinsulinaemia, and slightly enhanced blood glucose levels compared with the NMRI control mice. By means of immunofluorescence microscopy drastically reduced insulin levels were detected, which might be compensated by the observed islet cell hyperplasia and hypertrophy. Furthermore, the normally peripheral localisation of the alpha-cells within islets was disturbed. By contrast, there were no changes in somatostatin cell distribution. However, considerable differences appeared with regard to GLUT2: whereas the beta-cells of NMRI mice showed dense immunostaining of the GLUT2 transporter on the cell surface, in all age groups of NZO mice, GLUT2 on the plasma membranes was reduced and dispersed in the cytoplasm. These findings agree with the molecular biological results, which displayed decreased mRNA-expression of GLUT2. In summary, the observed alteration of islet morphology and of GLUT2 expression in diabetic mice complements our previous results from a superfusion protocol and further clarifies the mechanisms of diabetogenesis in NZO mice.

A reticular variant of elastolytic giant cell granuloma.

An 80-year-old man presented with a distinctive reticular form of elastolytic giant cell granuloma (EGCG). This represents a rare subset of EGCG, thus belonging to a larger group of granulomatous skin diseases with poorly understood pathogenesis. The patient exhibited numerous erythematous to brownish, sharply demarcated patches of reticulate appearance with partial central atrophy and nonelevated margins, which involved the trunk and especially chest, shoulders, upper back and arms. Skin biopsy revealed a granulomatous infiltrate with multinucleate histiocytes containing remnants of elastic fibres in conjunction with a significant loss of elastic fibres throughout the dermis. There is little published information on treatment, which remains unsatisfactory.

Influence of oxygen concentration on redox cycling of alloxan and dialuric acid.

Alloxan, a chemical diabetogen, decays in the absence of reductants into alloxanic acid. In the presence of glutathione, it is reduced via the alloxan radical into dialuric acid, which autoxidizes back to alloxan. During this redox cycling process, reactive oxygen species are formed that destroy beta-cells in islets of Langerhans. Previous experiments were conducted with oxygen concentrations about ten times as high as within cells. The aim of our in vitro study was to evaluate the impact of different oxygen concentrations (0, 25, 250 micromol/l) at a given initial ratio of glutathione and alloxan on this redox cycling. Reduction of alloxan, oxidation of glutathione, and the formation of glutathiol (GSSG) were continuously recorded by HPLC for 90 minutes at 25 degrees C in air, calibration gas, or argon. In the absence of reductants, alloxan irreversibly decomposed into alloxanic acid regardless of oxygen presence. When the reaction system contained glutathione, decomposition was significantly retarded and therefore influenced by oxygen. In argon, decay could not be observed due to its reduction and the absence of oxygen. Increasing oxygen concentration enabled a redox cycling and therefore an ongoing decay. The highest decomposition along with the highest consumption of glutathione occurred at 250 micromol/l oxygen. The lower the oxygen, the more dialuric acid could be detected. After calculation, about 33 redox cycles per hour generates an amount of reactive oxygen species sufficient to damage pancreatic beta cells and induce insulin deficiency.

[Corneal wound healing after hyperopic PRK and LASIK]. / Korneale Wundheilungsreaktionen nach hyperoper PRK und LASIK.

BACKGROUND: The purpose of this study was to investigate the influence of different postoperative treatments on the wound healing reaction in the anterior stroma after PRK and in the interface area after LASIK. METHODS: Seventy-two corneal buttons of refractively treated rabbit eyes underwent different postoperative eyedrop regimens with antibiotics and/or steroids or additional UV-B irradiation. Morphological and immunohistological investigations were performed 6 months postoperatively by light and transmission electron microscopy. RESULTS: PRK eyes showed interdigitations between the epithelia and the anterior stroma. LASIK-treated eyes showed only minor changes between epithelia and stroma in the incisional region. Only a slight increase in deposits of fibrillar extracellular matrix components were detectable in the interface region. CONCLUSIONS: The clinically important problem of haze after PRK is caused by the interdigitations between epithelia and anterior stroma. The delicate wound healing reactions in the interface region in LASIK eyes corresponded to the clinically visible minor changes in these corneas.

Indication of circadian oscillations in the rat pancreas.

The central circadian oscillator of the suprachiasmatic nucleus controls diurnal rhythmicity of the body with light as its dominant zeitgeber. Recently, peripheral oscillators have been detected in liver and heart, which follow as yet unidentified cues. In this study real-time reverse transcription-polymerase chain reaction (RT-PCR) was used in analysis of the expression of the major clock genes Per1, Per2, Bmal1, Cry1, Tim (timeless) and Clock, as well as of the output genes Dbp and Rev-erbalpha in the pancreatic tissue of rats. The results presented here indicate a robust circadian expression of clock genes (e.g. Per1 and Bmal1) and the probable existence of a peripheral oscillator in the pancreas. Whether this oscillator regulates the diverse functions of the islets of Langerhans remains to be elucidated.

Transformation of barbituric acid into alloxan by hydroxyl radicals: interaction with melatonin and with other hydroxyl radical scavengers.

Barbituric acid (2,4,6-pyrimidinetrione) can be transformed by a non-enzymatic hydroxylation into alloxan (2,4,5,6-pyrimidinetetrone). This transformation can be used as a reaction indicating the formation of hydroxyl radicals (.OH). This conversion was detected using HPLC. Formation of .OH was demonstrated by electron spin resonance (ESR) spectroscopy combined with spin-trapping techniques. It was shown that .OH generated via the Fenton reaction abstracts first a hydrogen atom from barbituric acid (BA) and forms intermediately a paramagnetic derivative of BA. After a second attack by another .OH, the BA radical is transformed into dialuric acid (DA), which autoxidizes via the alloxan radical (.ALX) to ALX. Superoxide radicals (.O2-) are formed during autoxidation of DA and.ALX. They are able to regenerate ferrous ions. As a result, traces of iron salts are capable of catalyzing the conversion of large amounts of BA into ALX. Several scavengers of .OH were tested with regard to their efficiency in preventing the transformation of BA into ALX. Of all the scavengers analyzed, melatonin was shown to be one of the most potent compounds.

Formation of compound 305 requires the simultaneous generation of both alloxan and GSH radicals.

This in vitro study investigates the conditions under which "compound 305" is formed. Using HPLC, ESR as well as UV spectroscopy, "compound 305" was largely separated and characterized. It has an absorption peak at 314 nm, which changes after reoxygenation to shorter wavelengths within hours. The retention time of "compound 305" amounts to 10.93 +/- 0.042 min. The formation of "compound 305" does not depend on alloxan (ALX) or reduced glutathione (GSH), but most likely on the steady-state concentration of the paramagnetic derivatives of both reactants (ALX* and GS*). The alloxan radical (ALX*) is formed by either a one-electron transfer from e. g. GSH to alloxan or oxidation of dialuric acid. The concentration of the ALX* was determined to be 12 +/- 3.6 micromol/l using the stable ultramarine radical as an ESR standard. ALX* is stable only under anaerobic conditions. It disappears within 2 min in air. Since formation of "compound 305" needs both ALX* as well as GS*, which are also necessary for the generation of reactive oxygen species (ROS), it is assumed that formation of "compound 305" diminishes the toxicity of alloxan.

Simultaneous quantitative determination of alloxan, GSH and GSSG by HPlc. Estimation of the frequency of redox cycling between alloxan and dialuric acid.

This in vitro study compares the frequency of redox cycling between alloxan and dialuric acid at different initial ratios of glutathione and alloxan. Alloxan oxidizes GSH to GSSG. The rate of GSH oxidation at a given initial GSH concentration of 2.0 mmol/L depends on the initial concentration of alloxan added. The higher the concentration of alloxan in relation to the initial concentration of GSH, the faster GSH oxidation proceeds, as well as oxygen consumption, and therefore, formation of reactive oxygen species. The highest rates of GSH oxidation, i.e. GSSG formation, were found at concentration ratios of between 2.0 mmol/L GSH and 0.2 and 0.04 mmol/L alloxan, respectively. Because 0.04 mmol/L alloxan oxidizes 2.0 mmol/L GSH completely, a frequency of at least 25 cycles between alloxan and dialuric acid within 3 hours can be assumed. During each redox cycle, two molecules of GSH are oxidized to one molecule of GSSG, and during each cycle one molecule of oxygen is reduced simultaneously to one molecule of hydrogen peroxide. In total, therefore, one molecule of alloxan oxidizes at least 50 molecules of GSH and forms about 25 molecules of hydrogen peroxide.

Comparison between xanthine oxidases from buttermilk and microorganisms regarding their ability to generate reactive oxygen species.

Xanthine oxidase (XO) forms uric acid from xanthine. It is assumed that at the same time oxygen is reduced by the XO to reactive oxygen species (ROS), mainly to .O2- and to H2O2. Under certain conditions such ROS can be highly damaging to cellular structures. Therefore, XO was frequently used as a model system, in which the impact of ROS on cellular compounds and structures has been investigated. In this in vitro study xanthine oxidases from buttermilk and from microorganisms were compared regarding their ability to generate ROS. It could be shown that both enzymes are able to transform xanthine to uric acid but differ significantly in their reductive properties to oxygen. XO from buttermilk reduces oxygen to both .O2- and H2O2 whereas XO from microorganisms generates H2O2, but fails to form .O2-. Since .O2- are involved in maintaining transition metal-mediated formation of hydroxyl radicals (.OH) from H2O2, we conclude that XO from microorganisms is therefore largely unsuitable in studies investigating just the interaction of .O2- with other ROS on cellular compounds.

'Classical' and 'new' diabetogens--comparison of their effects on isolated rat pancreatic islets in vitro.

This study compares functional and morphological alterations caused by application of alloxan, streptozotocin, xanthine oxidase/hypoxanthine (generation of reactive oxygen species), or S-nitroso-N-acetyl-D,L-penicillamine (SNAP, liberation of nitric oxide) to isolated rat pancreatic islets in vitro. In perifusion experiments, membrane leakage--detected by non-stimulated insulin release--was found after application of all drugs, but showed a substance-specific time pattern. Twenty-four hours after application of the classical diabetogens (alloxan or streptozotocin), potassium chloride- and glucose-stimulated insulin secretion were markedly reduced, while a persistent reduction was observed neither after exposure to xanthine oxidase/hypoxanthine, nor to SNAP. Morphological analysis of the islets revealed that nearly all beta-cells were destroyed following alloxan or streptozotocin treatment, while the majority of beta-cells were configured regularly after application of xanthine oxidase/hypoxanthine or SNAP. Necrotic cells found after xanthine oxidase/hypoxanthine usually differed in morphology from those observed after application of the classical diabetogens. While the former cells were characterised by swollen nuclei, the latter had shrunken nuclei with irregular condensed chromatin. Apoptosis was found only following nitric oxide exposure. Due to these differences, it seems unlikely that alloxan, streptozotocin, xanthine oxidase/hypoxanthine, and nitrix oxide have a common major feature in their toxic action.

A neuroendocrine releasing effect of melatonin in the brain of an insect, Periplaneta americana (L.).

Melatonin exists in nearly all organisms, but little is known of its function in non-vertebrates. Long-term perifusions as well as short-term batch incubations of brains and molting glands of the cockroach Periplaneta americana were used to test the influence of melatonin on the prothoracicotropic hormone, a glandotropic neuropeptide in the brain, which stimulates the production of the molting hormone ecdysone in the molting gland. Changes of ecdysteroid production in molting glands were determined by radioimmunoassay as ecdysone equivalents. Melatonin (10 nmol/L) was without effect on the prothoracic gland but stimulated the prothoracicotropic effect of brains in both in vitro investigations, long-term perifusions and short-term batch incubations. The effect was dose-dependent. The melatonin effect on the release of prothoracicotropic hormone in the brain was suppressed by luzindole (10 nmol/L), a pre-synaptic receptor antagonist of melatonin. The retro-cerebral complex (corpora cardiaca-corpora allata) did not seem to be involved in the effect of melatonin on the brain. Serotonin (10 nmol/L) suppressed the release of prothoracicotropic hormone. This is the first experimental evidence of a neurohormonal releasing effect of melatonin in the insect nervous system.

Evidence for a melatonin receptor within pancreatic islets of neonate rats: functional, autoradiographic, and molecular investigations.

In a recent perifusion investigation, we showed that the pineal secretory product melatonin reduces insulin secretion from isolated pancreatic islets of neonate rats stimulated with potassium chloride (KCl), glucose, and forskolin. This effect of melatonin was reproduced with doses ranging from 200 pmol/L to 5 micromol/L. Because it is generally accepted that melatonin exerts some of its biological effects through specific, high-affinity pertussis-toxin-sensitive G-protein-coupled receptors, we blocked the putative melatonin receptor of pancreatic islets using both the non-hydrolyzable guanosine triphosphate analog guanosine 5'-O-(3-thiotriphosphate) (GTPgammaS, 30 micromol/L) and the melatonin antagonist luzindole (10 micromol/L). Both GTPgammaS and luzindole caused a near normalization of the melatonin-induced inhibition of the forskolin-stimulated insulin secretion. To localize putative melatonin receptors within the pancreatic islets autoradiographic studies were additionally carried out. These investigations showed specific binding of 2-[125I]iodomelatonin, which were in exact correspondence with the localization of the islets. In addition, gray-level analysis showed that unlabeled melatonin was able to reduce the binding of 2-[125I]iodomelatonin in a dose-dependent manner. Concentrations of unlabeled melatonin of 10(-9) mol/L produced a 50% reduction in specific binding, whereas concentrations of 10(-6) mol/L displaced the binding completely. Likewise, the results of molecular investigations showed that the rat pancreas contains a melatonin receptor, since reverse transcription polymerase chain reaction (RT-PCR) experiments, using specific primers for the rat melatonin receptor Mel1a, showed that mRNA for this melatonin receptor type is expressed in pancreatic tissue of newborn rats. In summary, it may be said that our functional. autoradiographic, and molecular results indicate that the Mel1a receptor is located on the pancreatic islets, possibly in the beta cells.

Influence of melatonin on free radical-induced changes in rat pancreatic beta-cells in vitro.

Free radicals may produce cytotoxicity to pancreatic islets under pathophysiological conditions. The aim of our in vitro investigations was to compare functional and morphological changes in pancreatic beta-cells induced by reactive oxygen species (ROS) generated by alloxan or xanthine oxidase/hypoxanthine (XO/HX), respectively. We demonstrate that short-term exposure to alloxan or to XO/HX leads to a temporarily elevated insulin release from isolated pancreatic islets. On application of alloxan, this effect is caused by beta-cell necrosis and can be prevented by administration of melatonin, while in contrast, XO/HX did not lead to long-term morphological changes in the majority of the cells. Among the cells destroyed by alloxan, only necrosis could be detected, while in contrast, some apoptotic cells were identified by the terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) reaction and electron microscopic examinations of cells treated with XO/HX. Melatonin was able to prevent the changes caused by alloxan, but failed to influence the alterations caused by XO/HX. Using electron spin resonance and lipid peroxidation assay, respectively, it was confirmed that melatonin effectively detoxifies hydroxyl radicals. Therefore, we believe that hydroxyl radicals are the toxic principle of alloxan, but not of XO/HX toxicity.