Mechanisms and rescue strategies of calcineurin inhibitor mediated tolerance abrogation induced by anti-CD4 mAb treatment.

To ensure safety tolerance induction protocols are accompanied by conventional immunosuppressive drugs (IS). But IS such as calcineurin inhibitors (CNI), for example, cyclosporin A (CsA), can interfere with tolerance induction. We investigated the effect of an additional transient CsA treatment on anti-CD4mAb-induced tolerance induction upon rat kidney transplantation. Additional CsA treatment induced deteriorated graft function, resulting in chronic rejection characterized by glomerulosclerosis, interstitial fibrosis, tubular atrophy and vascular changes. Microarray analysis revealed enhanced intragraft expression of the B cell attracting chemokine CXCL13 early during CsA treatment. Increase in CXCL13 expression is accompanied by enhanced B cell infiltration with local and systemic IgG production and C3d deposition as early as 5 days upon CsA withdrawal. Adding different CNIs to cultures of primary mesangial cells isolated from glomeruli resulted in a concentration-dependent increase in CXCL13 transcription. CsA in synergy with TNF-α can enhance the B cell attracting and activating potential of mesangial cells. Transient B cell depletion or transfer of splenocytes from tolerant recipients 3 weeks after transplantation could rescue tolerance induction and did inhibit intragraft B cell accumulation, alloantibody production and ameliorate chronic rejection.

Permanent CNI treatment for prevention of renal allograft rejection in sensitized hosts can be replaced by regulatory T cells.

Recent data suggest that donor-specific memory T cells (T(mem)) are an independent risk factor for rejection and poor graft function in patients and a major challenge for immunosuppression minimizing strategies. Many tolerance induction protocols successfully proven in small animal models e.g. costimulatory blockade, T cell depletion failed in patients. Consequently, there is a need for more predictive transplant models to evaluate novel promising strategies, such as adoptive transfer of regulatory T cells (Treg). We established a clinically more relevant, life-supporting rat kidney transplant model using a high responder (DA to LEW) recipients that received donor-specific CD4(+)/ 8(+) GFP(+) T(mem) before transplantation to achieve similar pre-transplant frequencies of donor-specific T(mem) as seen in many patients. T cell depletion alone induced long-term graft survival in naïve recipients but could not prevent acute rejection in T(mem)(+) rats, like in patients. Only if T cell depletion was combined with permanent CNI-treatment, the intragraft inflammation, and acute/chronic allograft rejection could be controlled long-term. Remarkably, combining 10 days CNI treatment and adoptive transfer of Tregs (day 3) but not Treg alone also induced long-term graft survival and an intragraft tolerance profile (e.g. high TOAG-1) in T(mem)(+) rats. Our model allows evaluation of novel therapies under clinically relevant conditions.

Do insulinotropic glucose-lowering drugs do more harm than good? The hypersecretion hypothesis revisited.

Significant progress has been made in recent years in the characterisation of the signal pathways of beta cell dysfunction and death in the pathogenesis of type 2 diabetes. Glucolipotoxicity acts as an exogenous factor whereas oxidative stress and endoplasmic reticulum stress may result from the processes of signal recognition and stimulated secretion within the beta cell. The pharmacological stimulation of secretion may thus appear to be a double-edged sword: it counteracts hyperglycaemia, but may do so at the expense of beta cell mass. So, in the long run, insulinotropic glucose-lowering drugs might do more harm than good. However, much of this logic is derived by analogy from the long-held assumption that beta cell hypersecretion imposed by insulin resistance causes the absolute secretion deficit in the later course of type 2 diabetes. In this concept the beta cell has a secondary role and loss of beta cell mass is necessary for the manifestation of type 2 diabetes. Recent studies have shown that a secretion deficit can exist well before insulin resistance and that major genetic risk factors concern beta cell function. Also, the evidence for a beta cell toxic effect of insulinotropic drugs is currently inconclusive. Assuming that the insulin secretion deficit is of pathogenetic importance in a network with insulin resistance as an aggravating factor, an insulinotropic glucose-lowering drug may do more good than harm if it relieves the beta cell from the stress of glucose overstimulation and does so without inducing hypoglycaemia.

Prevention of spontaneous immune-mediated diabetes development in the LEW.1AR1-iddm rat by selective CD8+ T cell transfer is associated with a cytokine shift in the pancreas-draining lymph nodes.

AIMS/HYPOTHESIS: The LEW.1AR1-iddm rat is an animal model of spontaneous type 1 diabetes mellitus. This study analysed how adoptive transfer of selective T cell subpopulations affects the incidence of diabetes. METHODS: CD4(+) or CD8(+) T cells were isolated from diabetic LEW.1AR1-iddm rats or diabetes-resistant LEW.1AR1 rats. Cells were selectively transferred into athymic LEW.1AR1-Whn ( rnu ) or prediabetic LEW.1AR1-iddm rats. The animals were monitored for blood glucose, islet infiltration and immune cell composition of pancreas-draining lymph nodes. RESULTS: After adoptive transfer of CD4(+) T cells from diabetic LEW.1AR1-iddm rats into athymic LEW.1AR1-Whn ( rnu ) rats, 50% of the recipients developed diabetes. Transfer of CD8(+) T cells failed to induce diabetes. Only 10% of the athymic recipients became diabetic after co-transfer of CD4(+) and CD8(+) T cells. Adoptive transfer of CD8(+) T cells from LEW.1AR1 or diabetic LEW.1AR1-iddm rats into prediabetic LEW.1AR1-iddm rats significantly reduced the incidence of diabetes. In protected normoglycaemic animals regulatory CD8(+)/CD25(+) and CD4(+)/CD25(+) T cell subpopulations that were also FOXP3-positive accumulated in the pancreas-draining lymph nodes. In this lymphatic organ, gene expression of anti-inflammatory cytokines was significantly higher than in diabetic rats. CONCLUSIONS/INTERPRETATION: Our results show that adoptive transfer of CD4(+) but not CD8(+) T cells from diabetic LEW.1AR1-iddm rats induced diabetes development. Importantly, CD8(+) T cells from diabetic LEW.1AR1-iddm rats and diabetes-resistant LEW.1AR1 rats provided protection against beta cell destruction. The accumulation of regulatory T cells in the pancreas-draining lymph nodes from protected rats indicates that transferred CD8(+) T cells may have beneficial effects in the control of beta cell autoimmunity.

Cell loss during pseudoislet formation hampers profound improvements in islet lentiviral transduction efficacy for transplantation purposes.

Islet transplantation is a promising treatment in type 1 diabetes, but the need for chronic immunosuppression is a major hurdle to broad applicability. Ex vivo introduction of agents by lentiviral vectors-improving beta-cell resistance against immune attack-is an attractive path to pursue. The aim of this study was to investigate whether dissociation of islets to single cells prior to viral infection and reaggregation before transplantation would improve viral transduction efficacy without cytotoxicity. This procedure improved transduction efficacy with a LV-pWPT-CMV-EGFP construct from 11.2 +/- 4.1% at MOI 50 in whole islets to 80.0 +/- 2.8% at MOI 5. Viability (as measured by Hoechst/PI) and functionality (as measured by glucose challenge) remained high. After transplantation, the transfected pseudoislet aggregates remained EGFP positive for more than 90 days and the expression of EGFP colocalized primarily with the insulin-positive beta-cells. No increased vulnerability to immune attack was observed in vitro or in vivo. These data demonstrate that dispersion of islets prior to lentiviral transfection and reaggregation prior to transplantation is a highly efficient way to introduce genes of interest into islets for transplantation purposes in vitro and in vivo, but the amount of beta-cells needed for normalization of glycemia was more than eightfold higher when using dispersed cell aggregates versus unmanipulated islets. The high price to pay to reach stable and strong transgene expression in islet cells is certainly an important cell loss.

Modulation of human glucokinase intrinsic activity by SH reagents mirrors post-translational regulation of enzyme activity.

The low-affinity glucose phosphorylating enzyme glucokinase plays a key role in the process of glucose recognition in pancreatic B-cells. To evaluate mechanisms of intrinsic regulation of enzyme activity human pancreatic B-cell and liver glucokinase and for comparison rat liver glucokinase were expressed in E. coli bacteria. A one-step purification procedure through metal chelate affinity chromatography revealed 58 kDa proteins with high specific activities in the range of 50 U/mg protein and K(m) values around 8 mM for the substrate D-glucose with a preference for the alpha-anomer. There were no tissue specific differences, no species differences in the electrophoretic mobility, and no differences of the kinetic properties of these well conserved enzymes. The deletion of the 15 tissue-specific NH2-terminal amino acids of the human glucokinase resulted in a catalytically active enzyme whose kinetic properties were not significantly different from those of the wild-type enzymes. The human and rat glucokinase isoforms were non-competitively inhibited by the sulfhydryl group reagents alloxan and ninhydrin with Ki values in the range of 1 microM. The inhibition of glucokinase enzyme activity was reversed by dithiothreitol with an EC50 value of 9 microM for alloxan and of 50 microM for ninhydrin. D-Glucose provided protection against alloxan-induced inhibition of human and rat glucokinase isoenzymes with half-maximal effective concentrations between 11 and 16 mM. The enzyme inhibition by alloxan was accompanied by a change in the electrophoretic mobility with a second lower molecular 49 kDa glucokinase band which can be interpreted as a compact glucokinase molecule locked by disulfide bonds. Quantification of free sulfhydryl groups revealed an average number of 3.6 free sulfhydryl groups per enzyme molecule for the native human glucokinase isoforms. Alloxan decreased the average number of free sulfhydryl groups to 1.9 per enzyme molecule indicating that more than one SH side group is oxidized by this compound. The extraordinary sensitivity of the SH side groups of the glucokinase may be a possible mechanism of enzyme regulation by interconversion of stable (active) and unstable (inactive) conformations of the enzyme. In pancreatic B-cells the glucose-dependent increase of reduced pyridine nucleotides may stabilize the enzyme in the 58 kDa form and provide optimal conditions for glucose recognition and glucose-induced insulin secretion.

Differential regulation of [Ca2+]i oscillations in mouse pancreatic islets by glucose, alpha-ketoisocaproic acid, glyceraldehyde and glycolytic intermediates.

Glucose induces slow oscillations of the cytoplasmic Ca2+ concentration in pancreatic beta-cells. In order to elucidate the mechanisms responsible for the slow [Ca2+]i oscillations the effects of various nutrient insulin secretagogues on glucose-induced [Ca2+]i oscillations in intact mouse pancreatic islets and single beta-cells were studied. These were the glycolytic intermediates, glyceraldehyde and pyruvate, and the mitochondrial substrate, alpha-ketoisocaproic acid (KIC). Glucose, at a 10 or 15 mM concentration, induced the typical slow oscillations of [Ca2+]i (0.4 min(-1)). At higher glucose concentrations the frequency of these oscillations decreased further (0.2 min(-1)). Glyceraldehyde, an insulin secretagogue like glucose, did not cause slow oscillations of [Ca2+]i in the absence of glucose. However, it exhibited a synergistic action with glucose. Glyceraldehyde, at 3 or 10 mM concentration, induced slow [Ca2+]i oscillations at a substimulatory concentration of 5 mM glucose (0.3-0.4 min(-1)) and reduced the frequency of the glucose-induced [Ca2+]i oscillations at stimulatory concentrations of 10 or 15 mM glucose (0.2 min(-1)). KIC (5 or 10 mM) as well as pyruvate (10 mM), the end product of glycolysis, and its ester methyl pyruvate (10 mM), did not cause slow oscillations of [Ca2+]i in the absence of glucose. In contrast to glyceraldehyde, however, all three compounds were capable of preventing the slow [Ca2+]i oscillations induced by glucose. Mannoheptulose (2 mM), an inhibitor of glucokinase and glucose-induced insulin secretion, reversibly blocked any kind of [Ca2+]i oscillation and returned the [Ca2+]i to a basal level through its ability to inhibit glycolytic flux. It can be concluded therefore that only substrates which generate a glucokinase-mediated metabolic flux through glycolysis and produce glycolytic ATP can induce slow [Ca2+]i oscillations in pancreatic beta-cells.

Metabolic regulation, activity state, and intracellular binding of glucokinase in insulin-secreting cells.

Regulation of glucose-induced insulin secretion is crucially dependent on glucokinase function in pancreatic beta-cells. Glucokinase mRNA expression was metabolically regulated allowing continuous translation into enzyme protein. Glucokinase enzyme activity in the beta-cell was exclusively regulated by glucose. Using a selective permeabilization technique, different intracellular activity states of the glucokinase enzyme in bioengineered glucokinase-overexpressing RINm5F tissue culture cells were observed. These results could be confirmed in analogous experiments with dispersed islet cells. A diffusible glucokinase fraction with high enzyme activity could be distinguished from an intracellularly bound fraction with low activity. Glucose induced a significant long-term increase of the active glucokinase fraction. This effect was accomplished through the release of glucokinase enzyme protein from an intracellular binding site of protein character. The inhibitory function of this protein factor was abolished through proteolytic digestion or heat inactivation. Northern blot analyses revealed that this binding protein was not identical to the well-known liver glucokinase regulatory protein. This hitherto unknown new protein factor may have the function of a glucokinase regulatory protein in the pancreatic beta-cell, which may regulate glucokinase enzyme activity in a glucose-dependent manner.

Relation between antioxidant enzyme gene expression and antioxidative defense status of insulin-producing cells.

Antioxidant enzyme expression was determined in rat pancreatic islets and RINm5F insulin-producing cells on the level of mRNA, protein, and enzyme activity in comparison with 11 other rat tissues. Although superoxide dismutase expression was in the range of 30% of the liver values, the expression of the hydrogen peroxide-inactivating enzymes catalase and glutathione peroxidase was extremely low, in the range of 5% of the liver. Pancreatic islets but not RINm5F cells expressed an additional phospholipid hydroperoxide glutathione peroxidase that exerted protective effects against lipid peroxidation of the plasma membrane. Regression analysis for mRNA and protein expression and enzyme activities from 12 rat tissues revealed that the mRNA levels determine the enzyme activities of the tissues. The induction of cellular stress by high glucose, high oxygen, and heat shock treatment did not affect antioxidant enzyme expression in rat pancreatic islets or in RINm5F cells. Thus insulin-producing cells cannot adapt the low antioxidant enzyme activity levels to typical situations of cellular stress by an upregulation of gene expression. Through stable transfection, however, we were able to increase catalase and glutathione peroxidase gene expression in RINm5F cells, resulting in enzyme activities more than 100-fold higher than in nontransfected controls. Catalase-transfected RINm5F cells showed a 10-fold greater resistance toward hydrogen peroxide toxicity, whereas glutathione peroxidase overexpression was much less effective. Thus inactivation of hydrogen peroxide through catalase seems to be a step of critical importance for the removal of reactive oxygen species in insulin-producing cells. Overexpression of catalase may therefore be an effective means of preventing the toxic action of reactive oxygen species.

Complementary action of antioxidant enzymes in the protection of bioengineered insulin-producing RINm5F cells against the toxicity of reactive oxygen species.

To determine the importance of different antioxidative enzymes for the defense status of insulin-producing cells, the effects of stable overexpression of glutathione peroxidase (Gpx), catalase (Cat), or Cu/Zn superoxide dismutase (SOD) in insulin-producing RINm5F cells on the cytotoxicity of hydrogen peroxide (H2O2), hypoxanthine/xanthine oxidase (H/XO), and menadione have been investigated. Single overexpression of Cat or Gpx provided less protection than the combined expression of Cat plus SOD or Cat plus Gpx, while single overexpression of SOD either had no effect on the toxicity of the test compounds or increased it. RINm5F cells were also susceptible to butylalloxan, a lipophilic alloxan derivative that is selectively toxic to pancreatic beta-cells. Overexpression of enzymes, both alone and in combination, did not protect against butylalloxan-induced toxicity while SOD overexpression increased it, as evident from a half maximally effective concentration (EC50) value. The addition of Cat to the culture medium completely prevented the toxic effects of H2O2 and H/XO but had no significant effect on the toxicity of menadione or butylalloxan. Extracellular SOD had no effect on the toxicity of any of the test compounds. The results of this study show the importance of a combination of antioxidant enzymes in protecting against the toxicity of reactive oxygen species. Thus, overexpression of Cat and Gpx, alone or in combination with SOD, by use of molecular biology techniques can protect insulin-producing cells against oxidative damage. This may represent a strategy to protect pancreatic beta-cells against destruction during the development of autoimmune diabetes and emphasizes the importance of optimal antioxidative enzyme equipment for protection against free radical-mediated diseases.

Protection of insulin-producing RINm5F cells against cytokine-mediated toxicity through overexpression of antioxidant enzymes.

Nitric oxide (NO) and reactive oxygen species (ROS) are crucial elements in cytokine-mediated beta-cell destruction. In insulin-producing RINm5F cells, overexpression of cytoprotective enzymes provides significant protection against the synergistic toxicity of NO and ROS. We therefore examined whether overexpression of catalase (Cat), glutathione peroxidase (Gpx), and Cu/Zn superoxide dismutase (SOD) can provide protection for bioengineered RINm5F cells against cytokine-mediated toxicity. A 72-h exposure of RINm5F control cells to interleukin-1beta (IL-1beta) alone or a combination of IL-1beta, tumor necrosis factor-alpha, and gamma-interferon resulted in a time- and concentration-dependent decrease of cell viability in the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT) cytotoxicity assay. Although IL-1beta alone caused only a moderate reduction of viability in the range of 25%, the cytokine mixture induced a significant loss of viability of >75%. This increased toxicity of the cytokine mixture compared with that of IL-1beta alone could be explained by a higher rate of NO generation within the early 24-48 h incubation period that would favor the toxic synergism of NO and oxygen free radicals. Overexpression of Cat, Gpx, and Cu/Zn SOD protected against toxicity of the cytokine mixture but not against that of IL-1beta alone. The reduction of cytokine-mediated toxicity was evident also because of an increased proliferation rate and a drastic decrease in the cell death rate. The improved antioxidant defense status did not prevent the activation of iNOS after cytokine exposure. However, RINm5F cells overexpressing cytoprotective enzymes showed a significantly lower level of ROS-damaged protein residues. Thus, protection through Cat, Gpx, and Cu/Zn SOD overexpression was apparently because of an inactivation of ROS generated in the signal cascades of the cytokines. Overexpression of cytoprotective enzymes thus represents a feasible strategy to protect insulin-producing cells against cytokine-mediated cytotoxicity.

Engineering of a glucose-responsive surrogate cell for insulin replacement therapy of experimental insulin-dependent diabetes.

Glucose responsiveness in the millimolar concentration range is a crucial requirement of a surrogate pancreatic beta cell for insulin replacement therapy of insulin-dependent diabetes. Novel insulin-secreting GK cell clones with millimolar glucose responsiveness were generated from an early-passage glucose-unresponsive RINm5F cell line. This line expressed constitutively both the K(ATP) channel and the GLUT2 glucose transporter; but it had a relative lack of glucokinase. Through overexpression of glucokinase, however, it was possible to generate glucose-responsive clones with a glucokinase-to-hexokinase ratio comparable to that of a normal pancreatic beta cell. This aim, on the other hand, was not achieved through overexpression of the GLUT2 glucose transporter. Raising the expression level of this glucose transporter into the range of rat liver, without correcting the glucokinase-to-hexokinase enzyme ratio, did not render the cells glucose responsive. These glucokinase-overexpressing RINm5F cells also stably maintained their molecular and insulin secretory characteristics in vivo. After implantation into streptozotocin diabetic immunodeficient rats, glucokinase-overexpressing cells retained their insulin responsiveness to physiological glucose stimulation under in vivo conditions. These cells represent a notable step toward the future bioengineering of a surrogate beta cell for insulin replacement therapy in insulin-dependent diabetes mellitus.

Low antioxidant enzyme gene expression in pancreatic islets compared with various other mouse tissues.

Using a sensitive Northern blot hybridization technique, gene expression of superoxide dismutase (SOD), catalase, and glutathione peroxidase was studied in pancreatic islets and for comparison in various other mouse tissues (liver, kidney, brain, lung, skeletal muscle, heart muscle, adrenal gland, and pituitary gland). Gene expression of the antioxidant enzymes was usually in the range of +/- 50% of that in the liver. Only in pancreatic islets gene expression was substantially lower. The levels of the cytoplasmic Cu/Zn SOD and the mitochondrial Mn SOD gene expression were in the range of 30-40% of those in the liver. Glutathione peroxidase gene expression was 15%, and catalase gene expression was not at all detectable in pancreatic islets. These low levels of antioxidant enzyme gene expression may provide an explanation for the extraordinary sensitivity of pancreatic beta cells towards cytotoxic damage by diabetogenic compounds and during the development of human and animal diabetes.

Effect of superoxide dismutase, catalase, chelating agents, and free radical scavengers on the toxicity of alloxan to isolated pancreatic islets in vitro.

The effect of superoxide dismutase, catalase, metal-chelating agents and hydroxyl radical scavengers on the toxicity of alloxan to isolated ob/ob mouse pancreatic islets in vitro has been compared with the reported ability of such substances to protect against alloxan diabetes in vivo. Superoxide dismutase and catalase protected beta-cells of isolated pancreatic islets against alloxan cytotoxicity, as did the hydroxyl radical scavengers dimethyl sulfoxide (DMSO) and butanol. However, 1,3-dimethylurea and thiourea, that are recognised as effective hydroxyl radical scavengers and that protect animals against the diabetogenic effects of alloxan, were without effect. Similarly, desferrioxamine, that inhibits hydroxyl radical formation from alloxan in chemically defined systems, did not protect against alloxan toxicity. Diethylenetriamine pentaacetic acid, which does not inhibit hydroxyl radical formation from alloxan, also gave no significant protection. The results indicate a role for superoxide radical and hydrogen peroxide in the mechanism of toxicity of alloxan but do not support the involvement of the hydroxyl radical in this process. Alternative explanations must be sought for the ability of hydroxyl radical scavengers and metal-chelating agents to protect against alloxan toxicity in vivo.

Effects of sodium butyrate on glucose transporter and glucose-phosphorylating enzyme gene expression in RINm5F insulinoma cells.

RINm5F insulinoma cells show a defective physiological insulin secretory response to glucose stimulation. The short chain carbonic acid sodium butyrate induced a growth arrest during a 72-h tissue culture period. In contrast to control RINm5F cells, 2 mM glucose increased insulin secretion by more than 70% in these sodium butyrate-treated cells (1 mM) without any further increase of the secretory rate between 2 and 20 mM glucose. This effect of sodium butyrate on insulin secretion was assessed in comparison with its effect on gene expression of the GLUT1 and GLUT2 glucose transporter, hexokinase type I and type II, glucokinase and insulin. Sodium butyrate at a 1 mM concentration decreased GLUT1 gene expression by nearly 50%, but did not induce gene expression of the low-affinity GLUT2 glucose transporter above the detection limit. Furthermore, sodium butyrate increased glucokinase gene expression by more than 50% and hexokinase type II gene expression by more than 100%, while insulin gene expression was increased only by 24%. Hexokinase type II enzyme activity was increased by more than 100% without a concomitant significant change of the glucokinase enzyme activity. Sodium butyrate (2 mM) caused effects comparable with those of 1 mM sodium butyrate. Thus the improved insulin secretory responsiveness of RINm5F insulinoma cells after sodium butyrate treatment at low non-physiological millimolar glucose concentrations can be interpreted as a result of an increased hexokinase-mediated metabolic flux rate through the glycolytic chain.

Cytokines and nitric oxide inhibit the enzyme activity of catalase but not its protein or mRNA expression in insulin-producing cells.

Pancreatic beta-cells have low activities of the antioxidant enzyme catalase. Nitric oxide interacts with the haem group of catalase inhibiting its activity. We have studied the activity of catalase in beta-cells under conditions mimicking prediabetes and in which nitric oxide is generated from cytokine treatment in vitro. We also studied whether there is regulation of catalase enzyme activity by nitric oxide at the protein or gene expression level. RINm5F insulin-producing cells, treated for 24 h with cytokines, showed increased medium nitrite production (17+/-2.2 vs 0.3+/-0.2 pmol/ micro g protein) and significantly decreased cellular catalase activity (42.4+/-4.5%) compared with control cells. A similar reduction was seen in catalase-overexpressing RIN-CAT cells and in rat or human pancreatic islets of Langerhans. Catalase activity was also suppressed by the long-acting nitric oxide donor diethylenetriamine/nitric oxide adduct (Deta-NO) and this inhibition was reversible. The inhibition of catalase activity by cytokines in RINm5F cells was significantly reversed by the addition of the nitric oxide synthase 2 (NOS2) inhibitors nitro monomethylarginine or N-(3-(aminomethyl)benzyl)acetamidine (1400W). Protein expression was found to be unchanged in cytokine- or Deta-NO-treated RINm5F cells, while mRNA expression was marginally increased. We have shown that inhibition of catalase activity by cytokines is nitric oxide dependent and propose that this inhibition may confer increased susceptibility to cytokine- or nitric oxide-induced cell killing.

Comparative toxicity of alloxan, N-alkylalloxans and ninhydrin to isolated pancreatic islets in vitro.

The in vitro toxicity of the diabetogenic agent alloxan as documented by the induction of beta cell necrosis was studied in isolated ob/ob mouse pancreatic islets. The effect of alloxan has been compared with that of a number of N-alkyl alloxan derivatives and with that of the structurally related compound, ninhydrin. Alloxan and its derivatives were selectively toxic to pancreatic beta cells, with other endocrine cells and exocrine parenchymal cells being well preserved, even at high concentration. In contrast, ninhydrin was selectively toxic to pancreatic beta cells only at comparatively low concentration, destroying all islet cell types at high concentrations. The ultrastructural changes induced by all the test compounds in pancreatic beta cells in vitro were very similar to those observed during the development of alloxan diabetes in vivo. The relative toxicity of the various compounds to pancreatic beta cells in vitro was not, however, related to their ability to cause diabetes in vivo. Indeed, the non-diabetogenic substances ninhydrin, N-butylalloxan and N-isobutylalloxan were very much more toxic to isolated islets than the diabetogenic compounds alloxan and N-methylalloxan. These results suggest that the differences in diabetogenicity among alloxan derivatives are not due to intrinsic differences in the susceptibility of the pancreatic beta cells to their toxicity, but may reflect differences in distribution or metabolism. High concentrations of glucose protected islets against the harmful effects of alloxan and its derivatives, but not those of ninhydrin. Low levels of glucose, and non-carbohydrate nutrients, afforded little protection, indicating that the effect of glucose is not due to the production of reducing equivalents within the cell, 3-O-Methylglucose, which protects against alloan diabetes in vivo, did not protect against alloxan toxicity in vitro. Since 3-O-methylglucose is known to prevent uptake of alloxan by pancreatic beta cells, it appears that uptake of alloxan by the cell is not a prerequisite for the induction of beta cell necrosis.

Effect of metformin on SGLT1, GLUT2, and GLUT5 hexose transporter gene expression in small intestine from rats.

The effect of the antihyperglycaemic agent metformin was studied on gene expression of the energy-dependent sodium-hexose cotransporter (SGLT1) and the facilitative hexose transporters GLUT2 and GLUT5 in rat intestine. Metformin treatment (125 mg/kg body wt. twice daily for a period of 3 days) significantly increased SGLT1 gene expression in duodenum and jejunum. GLUT5 gene expression was increased by metformin treatment only in the jejunum. Gene expression of GLUT2 in the intestine was not significantly affected by metformin treatment. This increase in transporter gene expression offers the potential for increases in hexose uptake at the brush border membrane, and may compensate for other effects of the drug that have been suggested to decrease glucose uptake by SGLT1, as well as for metformin stimulation of glucose utilization by the intestinal mucosa.

Nutrient-dependent distribution of insulin and glucokinase immunoreactivities in rat pancreatic beta cells.

Functional heterogeneity among pancreatic beta cells is a characteristic feature of the islets of Langerhans. Under physiological conditions, beta cells in the pancreas of fed rats exhibited heterogeneous immunohistochemical staining for insulin and glucokinase. Intracellular beta cell glucokinase staining was either faint or dense. In the pericapillary space beta cell glucokinase immunoreactivity had a polar orientation, with the highest density in cytoplasmic regions close to the blood vessels. Starvation resulted in a loss of heterogeneity with homogeneous insulin staining in all beta cells of the islets, and this was accompanied by a loss of heterogeneous glucokinase staining. The intracellular polarity of glucokinase staining in contact to blood vessels also disappeared after starvation. Refeeding resulted in the reappearance of intercellular heterogeneity. In dependence on the functional demand, the endocrine pancreas recruited insulin from beta cells according to a well-defined hierarchy, with an initial preferential mobilization of medullary beta cells. In the course of this process intracellular polarity of glucokinase staining reappeared in areas of the beta cell with functional contact to the GLUT2 glucose transporter in the plasma membrane. This can be regarded as the morphological correlate of an activation of the glucose signal recognition apparatus. Interestingly, the study also provides evidence that the changes in glucokinase distribution apparently preceded those in insulin distribution, which is in keeping with the central role of glucokinase as the glucose sensor of the pancreatic beta cell.