Sequential Analysis of the Releasing and Fuel Function of Glucose in Isolated Perifused Pancreatic Islets WALTER S. ZAWALICH AND FRANZ M. MATSCHINSKY Department of Pediatrics and Pharmacology, Washington University Medical School, St. Louis, Missouri 63110 bolic concentration dependency curves. Maximal and half maximal rates of glucose use were obtained at 30 and 8 mM, respectively, and lactate formation reached highest rates at 8 mM glucose. Physiological changes of glucose levels (from 5 to 10 mM) increased hormone release 4-fold (from 0.49 to 1.94 /u,U/islet/min) whereas glucose use was changed only slightly (from 52 to 75 pmol/islet/h), and lactate formation not at all. These data show that there is only limited association between metabolic and insulin releasing efficiency of glucose in pancreatic islets in vitro and also implicate a threshold phenomenon triggered by either a glucose metabolite or glucose itself. (Endocrinology 100: 1, 1977)
ABSTRACT. Isolated islets were continuously perifused with glucose to test their secretory capacity in a dynamic fashion, and were subsequently transferred to an incubation vial to measure their capacity for metabolizing glucose. Insulin release was measured by radioimmunoassay and metabolism of glucose by determining the rate of 3H2O formation from glucose tritiated on carbon atom 2 or 5 and by lactate accumulation. Insulin release was induced by glucose at a threshold of 5 mM, was half maximal at 8 mM, maximal at 15 mM and showed biphasic kinetics, which is consistent with published data. However, in contrast to most previous reports, utilization of glucose and lactate formation showed hyper-
M
ANY OF THE physiological substances that stimulate or inhibit hormone secretion from the pancreatic islet cells are fuel molecules, e.g., glucose, amino acids, fatty acids and ketone bodies. It is therefore crucial for an understanding of islet cell function to study the metabolism of fuel molecules and to establish whether and to what extent metabolism controls hormone secretion. One of the current approaches to this problem has been to incubate isolated islets in the presence of these fuels and to measure the rate of use of such molecules, usually indicated by the accumulation of one or several obligatory metabolic products. For example, islets are incubated in a small volume containing physiological levels (5-20 mM) of glucose labeled with tritium on C5 or uniformly Received April 26, 1976. Supported by USPHS Grant AM 10591 and through a postdoctoral fellowship to W.Z. (NS-05221) and through the Diabetes and Endocrinology Center #1P17-AM 17904. F.M.M. is an Established Investigator of the American Diabetes Association. Reprint requests to: Walter S. Zawalich, Children's Hospital, Department of Pediatric Metabolism, 500 S. Kingshighway, St. Louis, Missouri 63110, or Franz M. Matschinsky, Department of Pharmacology, Washington University School of Medicine, 660 S. Euclid, St. Louis, Missouri 63110.
labeled with 14C and the rates at which tritiated water, 14CO2, and lactate are formed are measured as indicators of glucose catabolism (1-4). It is usually implied that results obtained under such conditions are representative of the physiological fuel function of the molecule under investigation. Accordingly, extrapolations are readily made to the in vivo setting or to related in vitro studies which are designed to assess the releasing properties of a given fuel molecule. However, metabolic and release studies are, for methodological reasons, carried out under very different conditions; for measuring fuel utilization, larger numbers of islets need to be crowded into small volumes whereas for studying hormone release a few islets suffice to accumulate detectable hormone levels in relatively larger volumes of medium, or larger numbers of islets are perifused at flow rates of 1 ml/min or more (1-7). Countless investigations have been performed with this approach. There are problems with this general strategy. First, in most metabolic studies, islets of Langerhans are isolated by digestion of the pancreas with crude collagenase, but only limited efforts are made to remove thoroughly digestive enzymes accumulating 1
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 16:38 For personal use only. No other uses without permission. . All rights reserved.
ZAWALICH AND MATSCHINSKY
during the process of isolation. Equally important, only seldom are normal physiological releasing functions established prior to metabolic testing. Variability in islet size and in the magnitude of the secretory response to glucose exhibited by different groups of islets may present additional problems. It would, therefore, seem most advantageous to examine metabolic capacity and insulin release by the same group of islets. One approach to accomplish this which was repeatedly used in the past relied on metabolic and cofactor measurements of islet material obtained by rapid sampling methods in the course of kinetic studies of insulin release (2,5,8). However, the potential of this approach is limited since it is often difficult or impossible to assess the actual use of a fuel or of a fuel combination by determining the levels of crucial metabolites in islets exposed to such fuels. We have made an attempt to control at least some of the factors that might influence the measurement of metabolic flux of fuel molecules. The influence of extensive preperifusion of the isolated islets on the metabolic response to glucose was assessed. Islets, isolated from the rat pancreas by a version of the collagenase procedure, were perifused with non-stimulatory levels of glucose until insulin output reached low basal levels. They were then briefly exposed to high glucose to test the secretory response in a dynamic fashion and were finally transferred to an incubation vial for measuring their capacity to utilize glucose. Metabolic data obtained with this procedure differ in many respects from the results obtained earlier with simpler procedures of batch incubation. This raises fundamental questions concerning the reliability of techniques commonly used in metabolic studies of isolated islets. Materials and Methods Animals and islet isolation Male Holtzman rats weighing between 300400 g were used in all studies. The animals had free access to food at all times. Islets were
Ehdo • 1977 Vol 100 • No 1
isolated according to the method of Lacy and Kostianovsky (10). Perifusion studies and insulin analysis For studying the kinetics of insulin release a perifusion system similar to that employed by Lacy, Walker and Fink (7) was used. One hundred islets were used in each perifusion chamber and 2 chambers were run in parallel in each experiment. The filters used in these experiments were of nylon cut into discs 13 mm in diameter (Tetko Inc., Elmsford, N.Y.; pore size: 10 /u,m). Two filters were used in each chamber: the top one supported the islets, while the bottom one served as a filter control when studying metabolism. Before use in the perifusion, the filters were boiled in 7% NaHCO3 for 30 min followed by 30 min of boiling in distilled water and drying at room temperature. It was essential to use this type of filter and this pretreatment protocol because the commonly employed Metricel filters (Gelman Instruments Co., Ann Arbor, Mich.) resulted in high blank values when studying tritiated water formation from tritiated glucose (see below). In all experiments, the islets were first perifused for 45 min with low glucose (2.75 niM) and then for an additional 20 min with various concentrations of glucose (2.75-40 DIM). The dead space of the system is approximately 4 ml, causing a delay of 4 min from switching to appearance of the perfusate in the collection vials. In some instances perifusion of islets with non-stimulatory and stimulatory glucose levels was repeated after measuring glucose utilization. The flow rate was 1 ml/min ± 10% in all perfusion studies. Therefore, insulin release rates are expressed in terms of fiU/ml x 100 islets. Insulin was measured using the method of Morgan and Lazarow (11) with porcine insulin as standard. Metabolic studies At the end of the perifusion period, the filter containing the islets was removed and placed in a glass vial, 15 mm in diameter and 1 inch high. The blank filter free of islets was similarly placed in another glass vial. To these vials was added 0.125 ml of incubation fluid, identical to the perifusion medium except for the presence of radioactive glucose. The medium containing glucose, tritiated on carbon 2 or 5, was prepared in the following manner: 3.5 /il of [5-3H]glucose or 7 /xl of [2-3H]glucose
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 16:38 For personal use only. No other uses without permission. . All rights reserved.
GLUCOSE METABOLISM AND INSULIN RELEASE (both obtained from Amersham-Searle) were placed in 10 x 75 mm glass tubes. (The specific activities of the radioactive sugars were 1 Ci/ mmol and 0.5 Ci/mmol for 5 and 2-T-glucose, respectively.) The tubes were dried overnight in a desiccator and the next day the tritiated glucose was redissolved with 300 /x\ of perifusion fluid containing the desired glucose concentration. The medium thus prepared was added to the incubation vials, which were then stoppered, gassed for 30 sec with a mixture containing 95% O2 and 5% CO2 and incubated at 37 C in a water bath with shaking at 100 strokes/min. After 1 h, the vials were removed, unstoppered, and aliquots of 50 /A of incubation fluid were transferred in duplicate to 6 x 30 Him glass tubes to which 5 /x\ IN HC1 was added. After gentle mixing these glass tubes were placed in 20 ml glass scintillation vials (Wheaton Scientific, Millville, N.J.) which contained 0.5 ml distilled water and were kept on ice. In addition, 50 /-il of a tritiated water (3H2O) standard (New England Nuclear) was treated in the same way to allow correction for incomplete equilibration during the diffusion step. The scintillation vials were stoppered, wrapped individually in aluminum foil to insure equal heat distribution, and kept for 18 h in an oven maintained at 50 C. Subsequently, the foil was removed, the vials unstoppered and the small 6 x 30 mm tube taken out of the scintillation vial. The outside was washed with distilled water and the entire tube dropped into a 20 ml scintillation vial to which 0.5 ml water was added. Scintillation mixture (10.0 ml) was then added to all vials. The mixture consisted of 190 g naphthalene (Eastman Kodak Co., Rochester, N.Y.), 30.4 g scintillator (2a70, Research Products International, Elk Grove Village, Illinois) added to 1 gallon of 1,4-dioxane (Fisher Scientific, Fair Lawn, N.J.). The vials were then counted for 10 min in a Beckman Scintillation Counter. Treatment of the data Usually 70-75% of the 3H2O was recovered in the larger vial. From this the diffusion correction factor was derived. This factor varied between 1.33 and 1.45. Correction for the filter blank was carried out in the following manner: the filter control, i.e., the lower filter which contained no islets, resulted in 400-600 cpm in the corresponding outer vials. This was divided by the number of counts representing [2 or 53 H]glucose remaining in the inner vial (approximately 450,000 cpm). The number of counts
representing 3H-glucose in the inner tube from the islet sample was multiplied by this percentage figure obtained in the control filter. The resulting value was noted as the true filter blank and was subtracted from the number of counts accumulating in the outer vial from the islet samples. This number was multiplied by the diffusion correction factor to allow for 100% recovery. Finally, the formula employed by Ashcroft et al. (1) was used to calculate the rate of glucose utilization. The utilization is ultimately expressed as picomoles glucose utilized/islet/h. Lactate measurements Lactate was measured in duplicates in 1 /xl aliquots of the incubation fluid by the enzymatic fluorometric method of Matschinsky and Ellerman (3). High sensitivity was achieved by enzymatic cycling of DPNH (12). The following description details the modifications of the published procedures. Samples (1 /u,l) of incubation medium (not acidified) were introduced into the oil wells (13). At this step lactate standards ranging from 1 to 5 x 10" n moles were also introduced. To this was added 1.5 /u.1 of the following solution: 2-amino-2-methylpropanol buffer, pH 9.9, 0.1M; DPN + , 660 /LtM; glutamate, 4 mM; bovine serum albumin, 0.06%; lactic acid dehydrogenase, 75 /ng/ml; glutamatepyruvate transaminase, 75 /Ltg/ml. After 45 min at 22-25 C, 0.13 /ul of 1.26N NaOH was added followed by heating of the oil well rack at 75 C (20 min). Aliquots of 0.5 fi\ were then transferred to 50 ix\ of ice cold cycling reagent. The cycling reagent was basically the same as in the published procedure. Alcohol-dehydrogenase was 25 /ig/ml and malate dehydrogenase, 2.5 /Ltg/ml, resulting in a cycling rate of about 1000 times/h. After 1 h at 37 C the reaction was stopped by brief boiling (3 min) and the malate formed was measured as described in the original procedure. Great care was exercised to avoid contamination of reagents and glassware with lactate. To accomplish this, gloves had to be worn throughout the procedure (including the perifusion experiment, the metabolism studies, and the actual lactate assay) and the glassware had to be rinsed thoroughly with hot tap water and distilled water just prior to use.
Results
It proved feasible to study in sequence the kinetics of insulin release and the metabo-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 16:38 For personal use only. No other uses without permission. . All rights reserved.
ZAWALICH AND MATSCHINSKY
40 45 50 55 6065
60min
40 45 50 55 60 65
DURATION OF PERIFUSION (A&C) OR INCUBATION (B),MIN
FlG. 1. Sequential analysis of insulin release and glucose use by isolated pancreatic islets. Two groups of 100 islets each were preperifused with low glucose (2.75 HIM, G2.75) for 45 min. One group was then stimulated with 27.5 mM glucose (G27.5) and the other group was maintained on 2.75 mM glucose (A). Following this, the ability to utilize glucose was determined with corresponding low and high glucose concentrations (B). Reperifusion, similar to A except that both sets of islets were ultimately exposed to high glucose, was performed in a last step (C). The means of experiments are plotted in each releasing profile. For reasons of clarity standard errors are left out but are similar to those presented in Fig. 2. In the case of glucose use the means ± SEM of the corresponding samples are given.
lism of glucose with the same group of islets (Fig. 1). This was possible, since in the process of perifusion islets become tightly attached to the nylon grid, allowing rapid transfer of the islet sample from one experimental setting to the other, i.e., from the perifusion chamber to the incubation vial and back. Inspection of the incubation fluid and counting of the islets on the nylon grid indicated that all islets were carried along. This was also supported by the insulin profiles which were virtually the same whether islets were first stimulated during Phase A or Phase C. The shape of the profiles was typically biphasic, the first phase observed during late stimulation (C) reached 80% of the initial peak observed during the earlier stimulation (A) and the maximal release rates were indistinguishable during Phase A and Phase C. When islets were exposed to high glucose successively during
Endo* 1977 Vol 100 • Nb 1
all three phases of the experiment, insulin release was blunted in Phase C. We have no satisfactory explanation for this phenomenon. Partial depletion of the insulin stores may be a contributing factor. The rates of glucose metabolism using high substrate levels measured in Phase B were within the range of previously published data (1,6,14) but results obtained here with basal glucose were higher than the corresponding results published earlier. Using the method of sequential functional and metabolic testing of isolated islets a dose-dependency study of glucose stimulated-insulin release and glycolysis was carried out (Figs. 2 and 3). In agreement with the results of many laboratories the concentration dependency of glucoseinduced insulin release was clearly sigmoidal. The threshold for stimulation was close to 5 mM glucose, the half maximal response was achieved at 8-10 mM, and the system was saturated at 15 mM. However, clearly hyperbolic curves were obtained when the rate of glycolysis, as measured by 3 H2O or lactate formation, was plotted as a function of glucose concentration. The half maximal rate of glucose use was reached at 8 mM glucose and the system subserving glucose metabolism was saturated at concentrations between 30 and 40 mM glucose. In addition, experiments conducted with 0.1 mM and 1 mM glucose (not shown in either figure) gave usage rates of 3.2 ± 0.8 and 16.4 ± 2pmol/islet/h, respectively. The halfmaximal rate of lactate formation was observed with 2 mM glucose and the system was saturated with approximately 5 mM glucose in the incubation fluid. The results differ from earlier findings in our laboratory (3,4) using freshly isolated islets and differ also from the results of Ashcroft et al. (1). In the studies referred to, the plots relating glucose use, CO2 production, and lactate formation to the glucose level of the medium, all exhibited clearly sigmoidal shapes. There is no reason to believe that the discrepancies are due to the analytical procedures. The lactate assay used here is highly specific, extremely sensitive, and
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 16:38 For personal use only. No other uses without permission. . All rights reserved.
GLUCOSE METABOLISM AND INSULIN RELEASE exhibits excellent linearity and reproducibility (Table 1). The biochemical basis of the method applied here for measuring glucose use seems to be sound too.
T Second Phase
/> 280 l/l
G
G275
' 2 75
2.75
100
\ 50 J
'5.5
2.75
INSULIN KbLIbASb,
100
150
5.5
200
100
100
50
300
10.0
'10.0
O
100
UJ
rh 100
50
ABSTRACT. Isolated islets were continuously perifused with glucose to test their secretory capacity in a dynamic fashion, and were subsequently transferred to an incubation vial to measure their capacity for metabolizing glucose. Insulin release was measured by radioimmunoassay and metabolism of glucose by determining the rate of 3H2O formation from glucose tritiated on carbon atom 2 or 5 and by lactate accumulation. Insulin release was induced by glucose at a threshold of 5 mM, was half maximal at 8 mM, maximal at 15 mM and showed biphasic kinetics, which is consistent with published data. However, in contrast to most previous reports, utilization of glucose and lactate formation showed hyper-
M
ANY OF THE physiological substances that stimulate or inhibit hormone secretion from the pancreatic islet cells are fuel molecules, e.g., glucose, amino acids, fatty acids and ketone bodies. It is therefore crucial for an understanding of islet cell function to study the metabolism of fuel molecules and to establish whether and to what extent metabolism controls hormone secretion. One of the current approaches to this problem has been to incubate isolated islets in the presence of these fuels and to measure the rate of use of such molecules, usually indicated by the accumulation of one or several obligatory metabolic products. For example, islets are incubated in a small volume containing physiological levels (5-20 mM) of glucose labeled with tritium on C5 or uniformly Received April 26, 1976. Supported by USPHS Grant AM 10591 and through a postdoctoral fellowship to W.Z. (NS-05221) and through the Diabetes and Endocrinology Center #1P17-AM 17904. F.M.M. is an Established Investigator of the American Diabetes Association. Reprint requests to: Walter S. Zawalich, Children's Hospital, Department of Pediatric Metabolism, 500 S. Kingshighway, St. Louis, Missouri 63110, or Franz M. Matschinsky, Department of Pharmacology, Washington University School of Medicine, 660 S. Euclid, St. Louis, Missouri 63110.
labeled with 14C and the rates at which tritiated water, 14CO2, and lactate are formed are measured as indicators of glucose catabolism (1-4). It is usually implied that results obtained under such conditions are representative of the physiological fuel function of the molecule under investigation. Accordingly, extrapolations are readily made to the in vivo setting or to related in vitro studies which are designed to assess the releasing properties of a given fuel molecule. However, metabolic and release studies are, for methodological reasons, carried out under very different conditions; for measuring fuel utilization, larger numbers of islets need to be crowded into small volumes whereas for studying hormone release a few islets suffice to accumulate detectable hormone levels in relatively larger volumes of medium, or larger numbers of islets are perifused at flow rates of 1 ml/min or more (1-7). Countless investigations have been performed with this approach. There are problems with this general strategy. First, in most metabolic studies, islets of Langerhans are isolated by digestion of the pancreas with crude collagenase, but only limited efforts are made to remove thoroughly digestive enzymes accumulating 1
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 16:38 For personal use only. No other uses without permission. . All rights reserved.
ZAWALICH AND MATSCHINSKY
during the process of isolation. Equally important, only seldom are normal physiological releasing functions established prior to metabolic testing. Variability in islet size and in the magnitude of the secretory response to glucose exhibited by different groups of islets may present additional problems. It would, therefore, seem most advantageous to examine metabolic capacity and insulin release by the same group of islets. One approach to accomplish this which was repeatedly used in the past relied on metabolic and cofactor measurements of islet material obtained by rapid sampling methods in the course of kinetic studies of insulin release (2,5,8). However, the potential of this approach is limited since it is often difficult or impossible to assess the actual use of a fuel or of a fuel combination by determining the levels of crucial metabolites in islets exposed to such fuels. We have made an attempt to control at least some of the factors that might influence the measurement of metabolic flux of fuel molecules. The influence of extensive preperifusion of the isolated islets on the metabolic response to glucose was assessed. Islets, isolated from the rat pancreas by a version of the collagenase procedure, were perifused with non-stimulatory levels of glucose until insulin output reached low basal levels. They were then briefly exposed to high glucose to test the secretory response in a dynamic fashion and were finally transferred to an incubation vial for measuring their capacity to utilize glucose. Metabolic data obtained with this procedure differ in many respects from the results obtained earlier with simpler procedures of batch incubation. This raises fundamental questions concerning the reliability of techniques commonly used in metabolic studies of isolated islets. Materials and Methods Animals and islet isolation Male Holtzman rats weighing between 300400 g were used in all studies. The animals had free access to food at all times. Islets were
Ehdo • 1977 Vol 100 • No 1
isolated according to the method of Lacy and Kostianovsky (10). Perifusion studies and insulin analysis For studying the kinetics of insulin release a perifusion system similar to that employed by Lacy, Walker and Fink (7) was used. One hundred islets were used in each perifusion chamber and 2 chambers were run in parallel in each experiment. The filters used in these experiments were of nylon cut into discs 13 mm in diameter (Tetko Inc., Elmsford, N.Y.; pore size: 10 /u,m). Two filters were used in each chamber: the top one supported the islets, while the bottom one served as a filter control when studying metabolism. Before use in the perifusion, the filters were boiled in 7% NaHCO3 for 30 min followed by 30 min of boiling in distilled water and drying at room temperature. It was essential to use this type of filter and this pretreatment protocol because the commonly employed Metricel filters (Gelman Instruments Co., Ann Arbor, Mich.) resulted in high blank values when studying tritiated water formation from tritiated glucose (see below). In all experiments, the islets were first perifused for 45 min with low glucose (2.75 niM) and then for an additional 20 min with various concentrations of glucose (2.75-40 DIM). The dead space of the system is approximately 4 ml, causing a delay of 4 min from switching to appearance of the perfusate in the collection vials. In some instances perifusion of islets with non-stimulatory and stimulatory glucose levels was repeated after measuring glucose utilization. The flow rate was 1 ml/min ± 10% in all perfusion studies. Therefore, insulin release rates are expressed in terms of fiU/ml x 100 islets. Insulin was measured using the method of Morgan and Lazarow (11) with porcine insulin as standard. Metabolic studies At the end of the perifusion period, the filter containing the islets was removed and placed in a glass vial, 15 mm in diameter and 1 inch high. The blank filter free of islets was similarly placed in another glass vial. To these vials was added 0.125 ml of incubation fluid, identical to the perifusion medium except for the presence of radioactive glucose. The medium containing glucose, tritiated on carbon 2 or 5, was prepared in the following manner: 3.5 /il of [5-3H]glucose or 7 /xl of [2-3H]glucose
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 16:38 For personal use only. No other uses without permission. . All rights reserved.
GLUCOSE METABOLISM AND INSULIN RELEASE (both obtained from Amersham-Searle) were placed in 10 x 75 mm glass tubes. (The specific activities of the radioactive sugars were 1 Ci/ mmol and 0.5 Ci/mmol for 5 and 2-T-glucose, respectively.) The tubes were dried overnight in a desiccator and the next day the tritiated glucose was redissolved with 300 /x\ of perifusion fluid containing the desired glucose concentration. The medium thus prepared was added to the incubation vials, which were then stoppered, gassed for 30 sec with a mixture containing 95% O2 and 5% CO2 and incubated at 37 C in a water bath with shaking at 100 strokes/min. After 1 h, the vials were removed, unstoppered, and aliquots of 50 /A of incubation fluid were transferred in duplicate to 6 x 30 Him glass tubes to which 5 /x\ IN HC1 was added. After gentle mixing these glass tubes were placed in 20 ml glass scintillation vials (Wheaton Scientific, Millville, N.J.) which contained 0.5 ml distilled water and were kept on ice. In addition, 50 /-il of a tritiated water (3H2O) standard (New England Nuclear) was treated in the same way to allow correction for incomplete equilibration during the diffusion step. The scintillation vials were stoppered, wrapped individually in aluminum foil to insure equal heat distribution, and kept for 18 h in an oven maintained at 50 C. Subsequently, the foil was removed, the vials unstoppered and the small 6 x 30 mm tube taken out of the scintillation vial. The outside was washed with distilled water and the entire tube dropped into a 20 ml scintillation vial to which 0.5 ml water was added. Scintillation mixture (10.0 ml) was then added to all vials. The mixture consisted of 190 g naphthalene (Eastman Kodak Co., Rochester, N.Y.), 30.4 g scintillator (2a70, Research Products International, Elk Grove Village, Illinois) added to 1 gallon of 1,4-dioxane (Fisher Scientific, Fair Lawn, N.J.). The vials were then counted for 10 min in a Beckman Scintillation Counter. Treatment of the data Usually 70-75% of the 3H2O was recovered in the larger vial. From this the diffusion correction factor was derived. This factor varied between 1.33 and 1.45. Correction for the filter blank was carried out in the following manner: the filter control, i.e., the lower filter which contained no islets, resulted in 400-600 cpm in the corresponding outer vials. This was divided by the number of counts representing [2 or 53 H]glucose remaining in the inner vial (approximately 450,000 cpm). The number of counts
representing 3H-glucose in the inner tube from the islet sample was multiplied by this percentage figure obtained in the control filter. The resulting value was noted as the true filter blank and was subtracted from the number of counts accumulating in the outer vial from the islet samples. This number was multiplied by the diffusion correction factor to allow for 100% recovery. Finally, the formula employed by Ashcroft et al. (1) was used to calculate the rate of glucose utilization. The utilization is ultimately expressed as picomoles glucose utilized/islet/h. Lactate measurements Lactate was measured in duplicates in 1 /xl aliquots of the incubation fluid by the enzymatic fluorometric method of Matschinsky and Ellerman (3). High sensitivity was achieved by enzymatic cycling of DPNH (12). The following description details the modifications of the published procedures. Samples (1 /u,l) of incubation medium (not acidified) were introduced into the oil wells (13). At this step lactate standards ranging from 1 to 5 x 10" n moles were also introduced. To this was added 1.5 /u.1 of the following solution: 2-amino-2-methylpropanol buffer, pH 9.9, 0.1M; DPN + , 660 /LtM; glutamate, 4 mM; bovine serum albumin, 0.06%; lactic acid dehydrogenase, 75 /ng/ml; glutamatepyruvate transaminase, 75 /Ltg/ml. After 45 min at 22-25 C, 0.13 /ul of 1.26N NaOH was added followed by heating of the oil well rack at 75 C (20 min). Aliquots of 0.5 fi\ were then transferred to 50 ix\ of ice cold cycling reagent. The cycling reagent was basically the same as in the published procedure. Alcohol-dehydrogenase was 25 /ig/ml and malate dehydrogenase, 2.5 /Ltg/ml, resulting in a cycling rate of about 1000 times/h. After 1 h at 37 C the reaction was stopped by brief boiling (3 min) and the malate formed was measured as described in the original procedure. Great care was exercised to avoid contamination of reagents and glassware with lactate. To accomplish this, gloves had to be worn throughout the procedure (including the perifusion experiment, the metabolism studies, and the actual lactate assay) and the glassware had to be rinsed thoroughly with hot tap water and distilled water just prior to use.
Results
It proved feasible to study in sequence the kinetics of insulin release and the metabo-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 16:38 For personal use only. No other uses without permission. . All rights reserved.
ZAWALICH AND MATSCHINSKY
40 45 50 55 6065
60min
40 45 50 55 60 65
DURATION OF PERIFUSION (A&C) OR INCUBATION (B),MIN
FlG. 1. Sequential analysis of insulin release and glucose use by isolated pancreatic islets. Two groups of 100 islets each were preperifused with low glucose (2.75 HIM, G2.75) for 45 min. One group was then stimulated with 27.5 mM glucose (G27.5) and the other group was maintained on 2.75 mM glucose (A). Following this, the ability to utilize glucose was determined with corresponding low and high glucose concentrations (B). Reperifusion, similar to A except that both sets of islets were ultimately exposed to high glucose, was performed in a last step (C). The means of experiments are plotted in each releasing profile. For reasons of clarity standard errors are left out but are similar to those presented in Fig. 2. In the case of glucose use the means ± SEM of the corresponding samples are given.
lism of glucose with the same group of islets (Fig. 1). This was possible, since in the process of perifusion islets become tightly attached to the nylon grid, allowing rapid transfer of the islet sample from one experimental setting to the other, i.e., from the perifusion chamber to the incubation vial and back. Inspection of the incubation fluid and counting of the islets on the nylon grid indicated that all islets were carried along. This was also supported by the insulin profiles which were virtually the same whether islets were first stimulated during Phase A or Phase C. The shape of the profiles was typically biphasic, the first phase observed during late stimulation (C) reached 80% of the initial peak observed during the earlier stimulation (A) and the maximal release rates were indistinguishable during Phase A and Phase C. When islets were exposed to high glucose successively during
Endo* 1977 Vol 100 • Nb 1
all three phases of the experiment, insulin release was blunted in Phase C. We have no satisfactory explanation for this phenomenon. Partial depletion of the insulin stores may be a contributing factor. The rates of glucose metabolism using high substrate levels measured in Phase B were within the range of previously published data (1,6,14) but results obtained here with basal glucose were higher than the corresponding results published earlier. Using the method of sequential functional and metabolic testing of isolated islets a dose-dependency study of glucose stimulated-insulin release and glycolysis was carried out (Figs. 2 and 3). In agreement with the results of many laboratories the concentration dependency of glucoseinduced insulin release was clearly sigmoidal. The threshold for stimulation was close to 5 mM glucose, the half maximal response was achieved at 8-10 mM, and the system was saturated at 15 mM. However, clearly hyperbolic curves were obtained when the rate of glycolysis, as measured by 3 H2O or lactate formation, was plotted as a function of glucose concentration. The half maximal rate of glucose use was reached at 8 mM glucose and the system subserving glucose metabolism was saturated at concentrations between 30 and 40 mM glucose. In addition, experiments conducted with 0.1 mM and 1 mM glucose (not shown in either figure) gave usage rates of 3.2 ± 0.8 and 16.4 ± 2pmol/islet/h, respectively. The halfmaximal rate of lactate formation was observed with 2 mM glucose and the system was saturated with approximately 5 mM glucose in the incubation fluid. The results differ from earlier findings in our laboratory (3,4) using freshly isolated islets and differ also from the results of Ashcroft et al. (1). In the studies referred to, the plots relating glucose use, CO2 production, and lactate formation to the glucose level of the medium, all exhibited clearly sigmoidal shapes. There is no reason to believe that the discrepancies are due to the analytical procedures. The lactate assay used here is highly specific, extremely sensitive, and
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 16:38 For personal use only. No other uses without permission. . All rights reserved.
GLUCOSE METABOLISM AND INSULIN RELEASE exhibits excellent linearity and reproducibility (Table 1). The biochemical basis of the method applied here for measuring glucose use seems to be sound too.
T Second Phase
/> 280 l/l
G
G275
' 2 75
2.75
100
\ 50 J
'5.5
2.75
INSULIN KbLIbASb,
100
150
5.5
200
100
100
50
300
10.0
'10.0
O
100
UJ
rh 100
50
