Chlorotetracycline as a fluorescent Ca2+ probe in pancreatic islet cells.

Pancreatic islets, or suspensions of islet cells, from noninbred ob/ob-mice were incubated with chlorotetracycline and analyzed for Ca2+-dependent fluorescence in a microscope. Unless logarithmically transformed, signals from islets were asymmetrically distributed with unstable variance. Signals from cells pelleted in glass capillaries were more homogeneous and depended linearly on the thickness of the sample. The effect of sample thickness and a significant enhancement of fluorescence by alloxan suggest that beta-cells were involved in producing the signal from whole islets. The signal from dispersed cells was probably diagnostic of Ca2+ in beta-cell plasma membranes because it was suppressed by La3+ and had a spectrum indicative of an apolar micromilieu; fluorescent staining of cell surfaces was directly seen at high magnification. Fluorescence from cells was enhanced by 0.5-10 mM Ca2+ in a dose-dependent manner, whereas less than 0.5 mM Ca2+ saturated the probe alone in methanol. The signal from islets or dispersed cells was suppressed by 5 mM theophylline; that from cells was also suppressed by 0.5 mM 3-isobutyl-1-methylxanthine, 1.2 or 15 mM Mg2+, 3-20 mM D-glucose, and, to a lesser extent, 20 mM 3-O-methyl-D-glucose. D-glucose was more inhibitory in the absence than in the presence of Mg2+, as if Mg2+ and D-glucose influenced the same Ca2+ pool. L-glucose, D-mannopheptulose, or diazoxide had no noticeable effect and 20 mM bicarbonate was stimulatory. The results suggest that microscopy of chlorotetracycline-stained cells can aid in characterizing calcium pools of importance for secretion. Initiation of insulin release may be associated with an increas

Influence of the mutation "diabetes" on insulin release and islet morphology in mice of different genetic backgrounds.

Mice, 7-8-mo old, of the C57BL/KsJ-db strain and homozygotic for the mutant gene db, exhibited marked hyperglycemia and moderately elevated serum insulin levels. Light and electron microscopy provided evidence of a slightly decreased proportion of beta cells in the pancreatic islets, irregular islet architecture with intraislet ducts, and degenerative as well as hypertrophic changes in the individual beta cells. As a rule, islets microdissected from these mice did not release insulin in response to glucose, theophylline, iodoacetamide, or chloromercuribenzene-p-sulphonic acid. The absence of secretory responses was not simply due to lack of insulin. Although the islet content of insulin was decreased in C57BL/KsJ-db/db mice, the remaining amount was severalfold larger than that released from stimulated islets of normal controls. Another mutation, db(2J), an allele of db with identical phenotypic expressions in the C57BL/KsJ strain, was studied on the genetic background C57BL/6J. In contrast to the severely diabetic C57BL/KsJ-db/db animals, the C57BL/6J-db(2J)/db(2J) mice were characterized by highly elevated serum insulin levels and only moderate hyperglycemia. Their endocrine pancreas was enlarged and showed an increased proportion of beta cells. Like the islets of normal mice, those of C57BL/6J-db(2J)/db(2J) mice responded to glucose and chloromercuribenzene-p-sulphonic acid, the glucose-induced responses being potentiated by theophylline or iodoacetamide. C57BL/KsJ-db/db mice should provide a valuable model for studying defects in insulin secretion in relation to diabetes mellitus. Mice of the C57BL/6J strain offer a control material that may help to elucidate the dependence of the insulin secretory defect on the background genome.

Neural cell adhesion molecule (N-CAM) is required for cell type segregation and normal ultrastructure in pancreatic islets.

Classical cell dissociation/reaggregation experiments with embryonic tissue and cultured cells have established that cellular cohesiveness, mediated by cell adhesion molecules, is important in determining the organization of cells within tissue and organs. We have employed N-CAM-deficient mice to determine whether N-CAM plays a functional role in the proper segregation of cells during the development of islets of Langerhans. In N-CAM-deficient mice the normal localization of glucagon-producing alpha cells in the periphery of pancreatic islets is lost, resulting in a more randomized cell distribution. In contrast to the expected reduction of cell-cell adhesion in N-CAM-deficient mice, a significant increase in the clustering of cadherins, F-actin, and cell-cell junctions is observed suggesting enhanced cadherin-mediated adhesion in the absence of proper N-CAM function. These data together with the polarized distribution of islet cell nuclei and Na+/K+-ATPase indicate that islet cell polarity is also affected. Finally, degranulation of beta cells suggests that N-CAM is required for normal turnover of insulin-containing secretory granules. Taken together, our results confirm in vivo the hypothesis that a cell adhesion molecule, in this case N-CAM, is required for cell type segregation during organogenesis. Possible mechanisms underlying this phenomenon may include changes in cadherin-mediated adhesion and cell polarity.

Calcium and secretion: distinction between two pools of glucose-sensitive calcium in pancreatic islets.

D-Glucose, but not L-glucose or 3-O-methyl-D-glucose, stimulates 45Ca2+ uptake by both lanthanum-displaceable and lanthanum-nondisplaceable pools in pancreatic islets. The nondisplaceable pool probably represents secretory granules, while the displaceable pool may be located in the beta-cell membrane. Kinetic studies with isotopically labeled islets suggest that only the displaceable pool participates in the short-term coupling of the glucose stimulus with secretion.

Effects of glucose, chloromercuribenzene-p-sulphonic acid and 4-acetamido-4'-isothiocyanostilbene-2,2'-disulphonic acid on phosphate efflux from pancreatic islets.

Collagenase-isolated pancreatic islets of non-inbred ob/ob mice, containing more than 90% beta-cells, were labelled with radioactive orthophosphate (32P or 33P) and then subjected to non-recirculating perifusion. The basal D-glucose concentration in the perifusion medium was 2.8 mM. When the concentration was suddenly raised to 5.6, 8.3 or 16.7 mM, D-glucose promptly elicited a transient and dose-dependent release of radiophosphate. In the presence of 2.8 mM D-glucose, 0.1 mM of the poorly permeating sulphydryl blocker, chloromercuribenzene-p-sulphonic acid, also evoked a phosphate flush resembling the one induced by D-glucose. The basal radiophosphate release was partially inhibited by 1 mM 4-acetamido-4-'-isothiocyanostilbene-2,2'-disulphonic acid. However, the phosphate flush induced by 16.7 mM D-glucose was not noticeably inhibited by 4-acetamido-4'-isothiocyanostilbene-2,2'-disulphonic acid. It is concluded that the phosphate flush emanates from beta-cells and that membrane sulphydryl groups may participate in its regulation. Although at least the basal phosphate release may in part represent transmembrane transport through 4-acetamido-4'-isothiocyanostilbene-2,2'-disulphonic acid-sensitive anion channels, other mechanisms are also likely to participate in the glucose-induced phosphate flush.

Stimulation of insulin release by thiols.

The effects of thiol compounds on insulin release were studied in microdissected pancreatic islets of non-inbred ob/ob micemin control experiments the reactivity of thiols against 6,6'-dithiodinicotinic acid and the degradation of mouse insulin were measured. At a concentration of 0.1 mM, 1-thio-D-glucose or reduced glutathione potentiated the insulin-releasing action of 10 mM D-glucose without affecting glucose oxidation. When tested at a concentration equivalent to about 0.1 mM reactive thiol, dextran-linked L-cysteine also potentiated the glucose-induced insulin secretion. In microperifusion experiments the insulin-releasing action of 1-thio-D-glucose was found to exhibit a rapid onset followed by a decline of the secretory rate to values lower than those observed with 10 mMD-glucose alone. No thiol stimulated insulin release in the absence of glucose. It is suggested that thiol compounds stimulate insulin release by splitting membrane disulphides in the beta-cells.

Effects of dextran-linked chloromercuribenzoic acid on insulin release from microdissected pancreatic islets.

Insulin release in response to dextran-linked p-chloromercuribenzoic acid was studied in microdissected pancreatic islets of non-inbred ob/ob-mice. No contamination of the dextran-linked mercurial with free chloromercuribenzoic acid was detected before or after the incubation with islets. In comparison with free mercurial, of the same thiol-blocking activity, the dextran-linked compound had a weak insulin-releasing action with a different dose vs. response relationship. The dextran-linked mercurial had no demonstrable effect on the islet content of cyclic AMP. The results support the hypothesis that free organic mercurials mainly stimulate insulin release by blocking thiol ground that are embedded within the beta-cell plasma membranes beneath their surfaces.

Morphometry of Rambourg-positive and Rambourg-negative beta-cell granules after culture with low and high glucose concentrations.

Dispersed islet cells from noninbred ob/ob mice were cultured for 3 days with 3 or 20 mM D-glucose and silver stained according to Rambourg et al. Two tinctorial subsets of dark and light intracellular granules were analyzed by morphometry at the ultrastructural level. The two types of granules were similar in size and shape. However, with 3 mM glucose the dark granule cores were surrounded by larger vesicles than the light granules. With 20 mM glucose, both types of granule vesicles and cores became smaller and dark-granule cores became more rounded, compared with cultures with 3 mM glucose. The higher glucose concentration also induced a marked decrease in the number (-84%) and volume density (-90%) of dark granules. In contrast, the number of light granules increased (+60%) with maintenance of their volume density. We suggest that the dark Rambourg-positive and the light Rambourg-negative beta-cell granules are functionally distinct subsets. The dark granules are probably engaged in insulin discharge. We discuss the unclear role of the light granules with a view to previously postulated heterogeneities of the insulin granule pool and their significance for exocytosis and intracellular hormone degradation.

Cytotoxic activation of complement by mouse pancreatic islet cells.

Human serum, or serum proteins excluded by Sephadex G-25, irreversibly inhibited the ability of mouse pancreatic islet cells to accumulate Rb+. The same treatment reduced the capacity of serum to subsequently inhibit Rb+ uptake by fresh islet cells or to lyse sensibilized sheep erythrocytes. Serum-treated islet cells exhibited electron microscopic signs of damage, including ruptures of the plasma membrane, swelling of mitochondria, and reduced electron density of the cytoplasmic ground substance. Serum induced a prompt insulin release, which was not inhibited by epinephrine. The serum effects were prevented by mild heating (50 degrees C or 56 degrees C, 30 min) but not by treating serum with 10 mM EGTA and 10 mM MgCl2, or with soybean trypsin inhibitor. Inhibition of Rb+ accumulation in response to human serum was also observed with dispersed mouse exocrine pancrease, liver, and spleen cells but not with whole islets. Homologous mouse serum had no effect on mouse liver or spleen cells but significantly decreased the Rb+ uptake by mouse islet cells. Autologous serum had no noticeable effect. It is suggested that mouse islet cells can activate complement via the alternative pathway and that triggering of this pathway is controlled by cellular discriminators of species, organ, and self.

Further studies on the metabolism of D-glucose anomers in pancreatic islets.

The alpha and beta anomers of commercially available D-(5-3H) glucose were separated by miniaturized Hudson-Dale procedures based on precipitation with acetic acid. Reflectometric measurements of the reactivity with matrix-bound glucose oxidase showed that the preparations were about 90 per cent pure with respect to anomeric composition. Nonradioactive anomers separated by the same procedures were analyzed by optic polarimetry and gas chromatography. The preparations were about 90 per cent pure with respect to anomeric composition and produced no peaks but D-glucose on trimethylsilylation and chromatography. Microdissected pancreatic islets of noninbred ob/ob-mice exhibited a linear production of 3H2O for three to nine minutes when incubated with 6 mM alpha-D-(5-3H) glucose, beta-D-(5-3H) glucose, or D-(5-3H) glucose in anomeric equilibrium; the three glucose preparations did not differ in their rate of conversion to 3H2O. The rate of 3H2O production increased with glucose concentration (3-21 mM) during incubations for three minutes and, again, there was no evidence for the metabolic activity's being dependent on the anomeric composition of the labeled sugar. When microdissected islets were perifused without glucose and suddenly exposed to 5-6 mM alpha-D-glucose or beta-D-glucose, the concentration of glucose-6-phosphate rose within five minutes and did not differ significantly between experiments with alpha-D-glucose and beta-D-glucose. In the same perifusion experiments, only alpha-D-glucose caused a pronounced stimulation of insulin secretion, the difference from beta-D-glucose being significant. The results indicate that the recognition of glucose as an insulin secretagogue does not only involve metabolism by glucose-6-phosphate. The possible roles of the sorbitol pathway and of hypothetical regulatory sites for the glucose molecule ("receptors") are briefly discussed.

Prevention of detrimental effect of cyclosporin A on vascular ingrowth of transplanted pancreatic islets with verapamil.

The revascularization of pancreatic islet clusters transplanted beneath the renal capsule was studied in a syngeneic mouse model. The degree of vascular ingrowth was visualized by in vivo fluorescence microscopy (fluorescein isothiocyanate-dextran) and judged by a semiquantitative method from coded video recordings. The recipients of isografts were divided into four groups, depending on their daily immunosuppressive treatment: 1) none (controls), 2) 15 mg/kg cyclosporin A (CsA), 3) 0.4 mg/kg verapamil + 15 mg/kg CsA, and 4) 20-30 mg/kg methylprednisolone. In control animals, capillary ingrowth was first demonstrated on day 6, followed by progressive vascularization up to day 34. After 6 mo, the vascular architecture was similar to that seen in normal islets in situ. CsA alone significantly decreased vascular ingrowth on day 14 compared with controls (P less than .02). Verapamil prevented the detrimental effect of CsA (P less than .01), probably by improving renal subcapsular blood flow. Methylprednisolone did not affect revascularization compared with control animals at day 14. We conclude that CsA inhibits vascular ingrowth into transplanted pancreatic islets, which is likely to have clinical implications. The prevention of CsA vascular ingrowth inhibition by a calcium antagonist indicates a possible approach to the correction of this problem, particularly when the renal capsule is used as the recipient's transplant site.

Effects of ouabain on insulin release, adenosine 3',5'-monophosphate and guanine 3',5'-monophosphate in pancreatic islets.

Isolated pancreatic islets of noninbred ob/ob mice were used to test the hypothesis that adenylate cyclase responds to changes of the transmembrane milieu or electric field in intact beta-cells. In the presence of a phosphodiesterase inhibitor, ouabainstimulated both the release of insulin and the islet content of cAMP. Ouabain had no noticeable effect on the islet content of cGMP. These results support the hypothesis at test. However, because ouabain also had some stimulatory effect on cAMP in islet homogenates, a direct action of ouabain on adenylate cyclase cannot be ruled out.

On the possible role of thiol groups in the insulin-releasing action of mercurials, organic disulfides, alkylating agents, and sulfonylureas.

The thiol activity of pancreatic islets was spectrophotometrically assayed as the formation of 6-mercaptonicotinic acid from the organic disulfide, 6,6'-dithiodinicotinic acid. Islets containing more than 90% beta-cells were microdissected from non-inbred ob/ob-mice. Comparisons of intact with homogenized islets indicated that the organic disulfide penetrates relatively slowly into the beta-cells. When tested at concentrations know to enhance insulin release, p-chloromercuribenzene-sulfonic acid almost completely blocked the thiol activity of intact islets, whereas no significant effect was observed with iodoacetamide, D-glucose, or glibenclamide. Although glibenclamide had no demonstrable effect on the thiol activity of free L-cysteine, the binding of glibenclamide to serum albumin was decreased by blocking the albumin thiols with azobenzene-2-sulfenyl bromide. The uptake of glibenclamide by pancreatic islets was inhibited by cysteine or reduced glutathione. Cysteine, as well as 6,6'-dithiodinicotinic acid, also seemed to interact negatively with glibenat organic mercurials and disulfides stimulate insulin release by blocking thiol groups in the beta-cell plasma membranes. The thiol groups involved in iodoacetamide-induced secretion may escape detection by the assay employed, or target groups other than thiols may be involved. The data on glibenclamide are compatible with, but do not unequivocally support, the notion that thiol groups may play a role in sulfonylurea-induced insulin release.

Effects of 5-hydroxytryptamine on rubidium ion fluxes and insulin release in cultured pancreatic islets.

We have incubated pancreatic islets isolated from noninbred ob/ob mice and NMRI mice for 3 days with or without 5-hydroxytryptamine (5-HT) in the medium and tested the effect of such long term treatment on subsequent insulin release and 86Rb+ accumulation and efflux. Two tenths millimolars of 5-HT abolished insulin release in response to 20 mM glucose. Two tenths millimolars of 5-HT also diminished the ability of islets to accumulate 86Rb+ and the effect of 10 mM glucose on 86Rb+ efflux. One one-hundredth millimolars of 5-HT had no effect on insulin release or 86Rb+ fluxes. Clearly, islets subjected to 5-HT for 3 days at concentrations that do not elicit demonstrable effects in short term incubations show a reduced secretory response. However, the physiological role of the high affinity uptake system for 5-HT in islet cells [Michaelis-Menten constant (Km) = 1.6 microM] remains unknown.

Further studies on the relationship between insulin release and lanthanum-nondisplaceable 45Ca2+ uptake by pancreatic islets: effects of fructose and starvation.

Relationships between the release of insulin and the incorporation of 45Ca2+ into a lanthanum-nondisplaceable (intracellular) pool were studied in islets microdissected from the pancreatic glands of non-inbred ob/ob mice. In comparison with D-glucose, D-fructose was slowly oxidized and had only marginal effects on insulin release. However, fructose was as effective as glucose in stimulating the lanthanum-nondisplaceable 45Ca2+ uptake. The 45Ca2+ uptake was dose-dependent on the concentration of fructose in the range 0-20 mM; the same dose-dependence was obtained with glucose. Fasting the mice for 3 days caused a total block of the insulin secretory response to 20 mM glucose, but it produced an enhancement of the glucose-induced 45Ca2+ uptake. Both the inhibition of insulin release and the enhancement of 45Ca2+ uptake were counteracted by pretreating the isolated islets with 20-40 mM D-glucose; pretreatment with L-glucose or fructose could not counteract the effects of fasting. Although some functional relationship may exist between the lanthanum-nondisplaceable uptake of 45Ca2+ and the insulin secretory apparatus, it is concluded that the uptake of Ca2+ is not simply the result of stimulated insulin release.

Loss of cholinergic potentiating responsiveness in mouse islets transplanted to the kidney.

Pancreatic islets from BALB/c mice were transplanted to the kidney of syngenic hosts. After 1-40 weeks, the grafts were removed, perifused in vitro, and extracted. Fresh islets were similarly examined. The graft insulin content fell by 70% in 1 week and remained low throughout the observation period. In contrast, rates of basal or glucose-stimulated insulin release were not much, if at all, decreased. In fresh islets and grafts removed after 3 or 28 weeks, 2 consecutive pulses of glucose stimulation, 20-25 min long and separated by 20 min at basal glucose concentration (2 or 2.8 mmol/L), elicited the same insulin secretory response. When 10 mumol/L acetylcholine and 10 mumol/L eserine were present during the second pulse, the glucose-stimulated insulin release from fresh islets was potentiated as much as 11-fold. This potentiation was reduced by one half during the first week of transplantation, and subsequently by 80-90%. It is concluded that vagal deprivation rapidly induces a state of persistent cholinergic refractoriness in transplanted beta-cells, despite morphological signs of autonomic reinnervation of the grafts.

Dynamics of glucose-induced insulin release from mouse islets transplanted under the kidney capsule.

Mouse pancreatic islet grafts under the kidney capsule of syngeneic hosts were removed and perifused in vitro 1-40 weeks after the transplantation. In comparison with fresh islets, 12- to 40-week-old grafts exhibited an attenuated first phase of glucose-stimulated insulin release. In grafts 1, 12, 28, or 40 weeks old, but not in fresh islets, the mean secretory rate during the initial 10 min of stimulation was significantly lower than that during the subsequent 15 min. When expressed in relation to insulin content, the insulin output in response to 11 mmol/L glucose was no less from grafts than from fresh islets; in grafts 12 or 40 weeks old at 16.7 mmol/L glucose, the fractional output above baseline was significantly diminished during the initial 10 min, but not subsequently. Immediately on switching from basal to stimulatory glucose concentration, there was a transient drop in insulin secretion from the grafts, especially after more than 12 weeks of transplantation and in response to 16.7, as compared with 11, mmol/L glucose. When glucose was switched back from stimulatory to basal concentration, grafts also frequently exhibited a transient increase in the insulin secretory rate. Neither initial drops nor "off responses" were seen in untransplanted islets. The modifications of the secretory dynamics in islet grafts suggest that transplantation influences the balance between the stimulatory and inhibitory influences of glucose on the beta-cell's secretory machinery.

The protective effect of verapamil on mouse islets transplanted under the kidney capsule.

Islets from normal NMRI mice were transplanted under the kidney capsule of syngeneic recipients. The graft-bearing mice were divided into 4 groups treated daily with cremophor alone (control), cyclosporine (25 mg/kg body wt), CsA in combination with the calcium antagonist verapamil (0.4 mg/kg), or verapamil alone. After 3 weeks the grafts were removed, analyzed for insulin secretory dynamics in a perifusion system, and extracted for their contents of insulin. The graft insulin content was significantly decreased by CsA, an effect counteracted by verapamil. As compared with controls, all treatments increased the basal insulin at 2.8 mmol/L glucose. CsA together with verapamil enhanced the biphasic secretory response to 16.7 mmol/L glucose, whether expressed per graft or per unit of insulin content. The glucose-stimulated insulin release per graft was greater after combining CsA with verapamil than after CsA alone. It is concluded that CsA has adverse effects on islets transplanted to the kidney, and that these effects can be ameliorated by combining the immunosuppressant with verapamil.

Protection against cyclosporine-induced impairment of renal microcirculation by verapamil in mice.

Fluorescence microscopy was used to examine the effect of cyclosporine (CsA) infusion on renal subcapsular (cortical) blood flow in 53 living mice, using FITC-dextran (MW: 156,000) as a fluorescent marker. CsA (8-19 mg/kg body weight) given i.v. for 1 min induced complete inhibition of blood flow. A complete standstill of flow was also obtained during a continuous infusion with a rate of 0.8-2 mg/kg/min. With lower infusion rates (0.15-0.23 mg/kg/min), blood flow was partially impaired. In all experiments, the decrease in flow occurred after a 15-25 min delay, suggesting a CsA metabolite or exhaustion of a protective mechanism as the causative agent. Pretreatment with an alpha-blocking agent, phentolamine (1.0 mg/kg), did not prevent the CsA-induced inhibition of blood flow. In contrast, pretreatment with a calcium antagonist, verapamil (0.3-0.4 mg/kg), prevented the impairment of blood flow at low (0.15-0.23 mg/kg/min), and partially at higher (0.8-2.4 mg/kg/min) rates of CsA infusion. Clinical studies are warranted to explore the role of calcium antagonists in the prevention of posttransplant acute cyclosporine-induced nephrotoxicity.

Characteristics of 5-hydroxytryptamine transport in pancreatic islets.

1 Transmembrane transport of 3H-labelled 5-hydroxytryptamine (5-HT) by isolated pancreatic islets of non-inbred ob/ob mice was studied. 2 5-HT was vigorously accumulated in a temperature-dependent way by the islet cells. 3 Studies of the concentration-dependence of [3H]-5-HT uptake revealed complex kinetics with one component being saturated at 1 to 3 microM 5-HT (apparent association constant 0.6 x 10(6) M(-1) and the other non-saturated up to 1 mM 5-HT. 4 The saturable uptake was inhibited by Na+-deficiency and metabolic poisoning with 2,4-dinitrophenol and antimycin A, whereas the non-saturable component was not affected. 5 Omission of K+, Ca2+ or Mg2+ did not affect the uptake rate. 6 It is concluded that 5-HT is taken up by pancreatic beta-cells by mechanisms very similar to those observed in thrombocytes and neurones.