Inhibition of Protein Synthesis Sequentially ... - DIAL@UCLouvain

0013-7227/01/$03.00/0 Endocrinology Copyright © 2001 by The Endocrine Society

Vol. 142, No. 1 Printed in U.S.A.

Inhibition of Protein Synthesis Sequentially Impairs Distinct Steps of Stimulus-secretion Coupling in Pancreatic ␤ Cells* MARIA-JOSE GARCIA-BARRADO†, MAGALIE A. RAVIER, JEAN-FRANC ¸ OIS ROLLAND, PATRICK GILON, MYRIAM NENQUIN, JEAN-CLAUDE HENQUIN

AND

Unite´ d’Endocrinologie et Me´tabolisme, University of Louvain Faculty of Medicine, B-1200 Brussels, Belgium ABSTRACT Proteins with a short half-life are potential sites of pancreatic ␤ cell dysfunction under pathophysiological conditions. In this study, mouse islets were used to establish which step in the regulation of insulin secretion is most sensitive to inhibition of protein synthesis by 10 ␮M cycloheximide (CHX). Although islet protein synthesis was inhibited approximately 95% after 1 h, the inhibition of insulin secretion was delayed and progressive. After long (18 –20 h) CHX-treatment, the strong (80%) inhibition of glucose-, tolbutamide-, and K⫹induced insulin secretion was not due to lower insulin stores, to any marked impairment of glucose metabolism or to altered function of K⫹-ATP channels (total K⫹-ATP currents were however decreased). It was partly caused by a decreased Ca2⫹ influx (whole-cell Ca2⫹ current) resulting in a smaller rise in cytosolic Ca2⫹ ([Ca2⫹]i). The situation was very different after short (2–5 h) CHX-treatment. In-

sulin secretion was 50 – 60% inhibited although islet glucose metabolism was unaffected and stimulus-induced [Ca2⫹]i rise was not (2 h) or only marginally (5 h) decreased. The efficiency of Ca2⫹ on secretion was thus impaired. The inhibition of insulin secretion by 15 h of CHX treatment was more slowly reversible (⬎4 h) than that of protein synthesis. This reversibility of secretion was largely attributable to recovery of a normal Ca2⫹ efficiency. In conclusion, inhibition of protein synthesis in islets inhibits insulin secretion in two stages: a rapid decrease in the efficiency of Ca2⫹ on exocytosis, followed by a decrease in the Ca2⫹ signal mediated by a slower loss of functional Ca2⫹ channels. Glucose metabolism and the regulation of K⫹-ATP channels are more resistant. Proteins with a short half-life appear to be important to ensure optimal Ca2⫹ effects on exocytosis, and are the potential Achille’s heel of stimulus-secretion coupling. (Endocrinology 142: 299 –307, 2001)

N

glucokinase (10 –12) or the sulfonylurea receptor of the K⫹ATP channel (13–15) profoundly perturb ␤ cell function. In contrast, virtually nothing is known about the stability of islet proteins. This is, however, a critical issue because changes in the synthesis or degradation rates of proteins with a short half-life might contribute to the ability or failure of ␤ cells to adapt to various pathophysiological conditions. Thus, distinct cell proteins may have vastly different halflives, from minutes to days. Proteins with a long half-life usually are structural components or perform housekeeping functions, whereas proteins with a short half-life play more specialized, often regulatory, roles. In the present study, we examined the consequences of a blockade of islet protein synthesis on the regulation of insulin secretion. A similar approach was already used 30 yr ago. Several groups showed that high concentrations of cycloheximide (CHX) and puromycin, two inhibitors of translation, caused a delayed (⬃60 min) inhibition of glucose-induced insulin secretion from the perfused rat pancreas (16 – 18) and isolated rat islets (19). The initial interpretation that insulin biosynthesis was required for sustained secretion (second phase) was rapidly refuted (20, 21), but no explanation for the inhibition of insulin secretion by protein synthesis inhibition was ever provided. Assuming that those steps of stimulus-secretion coupling involving the proteins with the smallest pool size and shortest half-life would be first affected, we evaluated the impact of different durations of protein synthesis inhibition on the major events regulating

ORMAL glucose homeostasis critically depends on the precise control of insulin secretion. This control involves two major pathways that both require metabolism of the sugar by ␤ cells (1). The first one serves to produce a triggering signal through the following sequence of events: the increase in the ATP/ADP ratio resulting from glucose metabolism closes ATP-sensitive K⫹ channels (K⫹-ATP channels) in the plasma membrane; the ensuing membrane depolarization opens voltage-dependent Ca2⫹ channels, allowing Ca2⫹ influx and rise of free cytosolic Ca2⫹ ([Ca2⫹]i) that eventually triggers exocytosis (2– 6). The second pathway serves to produce as yet incompletely identified amplifying signals that increase the efficiency of Ca2⫹ on exocytosis (1, 7–9). This elaborated stimulus-secretion coupling implicates a large number of proteins. Many of these proteins are identified and mutations of the gene coding for some of them, e.g. Received July 26, 2000. Address all correspondence and requests for reprints to: J. C. Henquin, Unite´ d’Endocrinologie et Metabolisme, UCL 55.30, avenue Hippocrate 55, B-1200 Brussels, Belgium. E-mail: [email protected]. * This work was supported by the Interuniversity Poles of Attraction Program (P4/21), Belgian State Prime Minister’s Office, Federal Office for Scientific, Technical, and Cultural Affairs; by Grant 3.4552.98 from the Fonds de la Recherche Scientifique Me´dicale, Brussels; and by Grant 95/00 –188 from the General Direction of Scientific Research of the French Community of Belgium. † Postdoctoral fellow on leave from the Department of Pharmacology, Faculty of Medicine, University of Salamanca, Spain.

299

300

PROTEIN SYNTHESIS INHIBITION AND INSULIN SECRETION

insulin secretion. Remarkably, glucose metabolism and the generation of the triggering signal were only slightly and slowly affected whereas the efficiency of Ca2⫹ on the secretory process was rapidly impaired. Materials and Methods Preparation and solutions Islets were aseptically isolated by collagenase digestion of the pancreas of fed female NMRI mice followed by hand selection (22). Except for electrophysiological experiments (see below), the islets were then cultured (generally for 18 h) in 2.5 ml RPMI 1640 medium (Life Technologies, Inc., Paisley, UK) containing 10 mm glucose, 10% heat-inactivated FCS, 2 mm glutamine, 100 IU/ml penicillin, and 100 ␮g/ml streptomycin. The control medium used for islet isolation was a bicarbonate-buffered solution containing (in mm): NaCl 120, KCl 4.8, CaCl2 2.5, MgCl2 1.2 and NaHCO3 24. It was gassed with O2/CO2 (94/6) to maintain a pH of 7.4, and contained 1 mg/ml BSA. The same medium was used for most experiments. When the concentration of KCl was increased to 30 mm, that of NaCl was decreased accordingly. All experiments, except patch-clamp recordings, were carried out at 37 C.

Cycloheximide treatment CHX was added to the culture medium or the experimental solutions from a 5 mm stock solution prepared daily in sterile water. The treatment with CHX was applied for different periods of time which, for practical reasons (batching of the islets, transfer from culture to preincubation medium) slightly varied: 2 h (120 –135 min), 5 h (4.5–5 h) and 20 h (19 –20 h). Except when indicated otherwise, CHX was not present during the acute tests of islet functioning but was withdrawn just before the start of these tests.

Measurements of insulin secretion and islet insulin content Insulin secretion during the culture period was measured in aliquots of medium taken after 18 h. When the reversibility of the effects of CHX was studied, the culture medium was changed after 12 h and aliquots of the fresh medium were taken every 3 h. At the end of the culture, the islets were recovered from the Petri dishes, and their insulin content was determined after extraction in acid-ethanol (23). Insulin was measured by a double-antibody RIA (24) with rat insulin as the standard. Insulin secretion was also measured in acute experiments using islets cultured with or without CHX. At the end of the culture, the islets were first washed and preincubated in control medium containing CHX or not as appropriate. Batches of 20 islets were then placed in small chambers of a perifusion system (25). Effluent samples were collected every 2 or 6 min for insulin measurement. In one series of experiments batches of 3 islets were incubated for 1 h in 1 ml medium containing 15 mm glucose with or without 0.5 mm dibutyryl cAMP or 25 nm phorbol myristic acid (PMA) (Sigma, St Louis, MO). At the end of the incubation, a portion of medium was taken for insulin measurement.

Measurements of total protein synthesis The islets were cultured for 1–18 h in 1.25–2.5 ml RPMI medium supplemented with l-[3,5-3H] tyrosine (20 – 40 ␮Ci/ml depending on the duration of the experiment), and with or without CHX. At the end of the culture period with the tracer, the medium was removed and the islets were collected and washed 4 times with control medium containing 1 mm nonradioactive tyrosine. They were then distributed in batches of 15 in polyethylene conical tubes, to which 500 ␮l ice-cold trichloracetic acid (10%) was added to precipitate proteins. The tubes were centrifuged, the supernatant was discarded and the pellet was rinsed again with trichloracetic acid (3 times). The pellet was eventually solubilized in 200 ␮l 0.1 m NaOH, and its radioactive content measured by liquid scintillation spectrometry, using Flow-Safe F (Lumac, Groningen, The Netherlands) as scintillator and counting at an efficiency of approximately 60%. The results (cpm/islet) obtained in test groups were expressed as a percentage of those in control islets treated in the same way within the same experiment.

Endo • 2001 Vol. 142 • No. 1

Measurements of glucose oxidation After culture and preincubation with or without CHX, batches of 10 islets were incubated for 2 h in 50 ␮l control medium supplemented with 1 ␮Ci [U-14C] glucose. Oxidation of glucose was calculated from the production of [14C]CO2. Technical aspects of the method have been described previously (26). In these experiments CHX was still present during the incubation.

Measurements of [Ca2⫹]i and NAD(P)H After culture, the islets to be used for [Ca2⫹]i measurements were loaded with fura-PE3 during 2 h of preincubation in the presence of 2 ␮m fura-PE3 acetoxymethylester. The islets to be used for NAD(P)H measurements were preincubated without dye. The medium contained CHX (test islets) or not (control islets). After preincubation, the islets were transferred into the perifusion chamber of a microspectrofluorimeter system that has previously been described in detail (27, 28). NAD(P)H measurements were obtained from one test and one CHXtreated islet side by side.

Measurements of

Rb⫹ efflux

86

After culture, the islets were loaded with 86Rb⫹ (used as a tracer for K⫹) during 90 min of preincubation in control medium containing 80 ␮Ci 86 RbCl (0.37 mm). The medium contained CHX (test islets) or not (control islets). The islets were then washed with nonradioactive medium and placed in batches of 30 in the same perifusion system as that used to study insulin secretion. The radioactivity lost by the islets was measured in effluent fractions collected at 2 min intervals, and the fractional efflux rate was calculated for each period (25).

Electrophysiological experiments After isolation, the islets were dispersed into single cells (29) and cultured on glass coverslips in RPMI 1640 medium as described above. The cells were allowed to attach to the glass during 16 –18 h before addition of CHX to the culture medium for 19 –20 h. The electrical recordings were made using the amphotericine-perforated whole-cell mode of the patch-clamp technique at 21–23 C (30 –31). Perforation required a few min and the voltage-clamp was considered satisfactory when the series resistance had fallen below 20 m⍀. Membrane currents were measured using an EPC-9 patch-clamp amplifier (Heka Electronics, Lambrecht/Pfalz, Germany) and the software Pulsefit (version 8 –31). To measure K⫹-ATP currents, 100 ms depo- and hyperpolarizing pulses of 20 mV were applied alternatively from a holding potential of ⫺70 mV. The pipette solution was (mm): K2SO4 70, NaCl 10, KCl 10, MgCl2 3.7, HEPES 5 (adjusted to pH 7.1), and the bath solution was (mm): NaCl 120, KCl 4.8, CaCl2 2.5, MgCl2 1.2, NaHCO3 24, HEPES 5 (adjusted to pH 7.4). To measure Ca2⫹ currents, the holding potential was ⫺80 mV, and 100 ms depolarizing pulses were applied to various potentials to establish current-voltage (I/V) relationships. The pipette solution was (mm): Cs2SO4 76, NaCl 10, KCl 10, MgCl2 1, HEPES 5 (adjusted to pH 7.1), and the bath solution was (mm): NaCl 125, KCl 4.8, CaCl2 2.5, MgCl2 1.2, tetraethylammonium-Cl 10, HEPES 5 (adjusted to pH 7.4).

Presentation of results Results are presented as means ⫾ sem. The statistical significance of differences between means was assessed by ANOVA followed by a Newman-Keuls test.

Results Concentration-dependence, onset and reversibility of cycloheximide effects

Addition of CHX to the culture medium for 18 h inhibited insulin secretion and islet protein synthesis in a parallel, concentration-dependent manner with, however, a slightly stronger effect on synthesis than secretion (Fig. 1a). At the end of the culture period, the insulin content of control islets


FIG. 1. Insulin secretion and protein synthesis by, and insulin content of mouse islets cultured without or with CHX for 18 h. a, The amount of insulin released in the culture medium and the incorporation of [3,5-3H] tyrosine into proteins by islets cultured with CHX were expressed as a percentage of values obtained in control islets within the same experiments. Absolute values for these controls were: 133 ⫾ 6 ng insulin/islet and 2098 ⫾ 209 cpm/islet. b, The insulin content of the islets used to study insulin secretion was determined at the end of the culture. It is shown by the lower, filled portion of the columns, together with that of fresh, noncultured islets from the same preparations. The upper, open portion of the columns show the amount of insulin released by the same islets during the 18 h of culture. The whole columns thus correspond to the sum of insulin release and content. Values are means ⫾ SEM for 17 batches of islets from 4 experiments (insulin secretion and content) or 14 batches of islets from 3 experiments (protein synthesis).

(cultured without CHX) averaged 100 ⫾ 6 ng, which was 47% less than that of fresh, noncultured islets (P ⬍ 0.001), but the sum of insulin content and secretion (233 ⫾ 6 ng) was 23% larger (P ⬍ 0.05) (Fig. 1b). Insulin synthesis has thus partially compensated for insulin secretion. This was not the case after culture with CHX: the insulin content was slightly lower than that of fresh islets (P ⬍ 0.01), but the sum of insulin content and secretion was not different. Importantly, the insulin content of islets cultured with CHX was higher (P ⬍ 0.05 or less) than that of control islets (Fig. 1b). This rules out the trivial possibility that the defects of insulin secretion caused by CHX are simply due to insufficient insulin stores. Because maximum effects were obtained with 10 ␮m CHX, this concentration was used in all subsequent experiments. To test the reversibility of CHX effects, islets were cultured with or without 10 ␮m CHX for 12 h. This inhibited insulin secretion by 80% (Fig. 2a). The islets were then transferred to fresh culture medium. When CHX remained present, both insulin secretion and protein synthesis were strongly depressed as compared with controls. When CHX was withdrawn, a significant synthesis of proteins was noted after 3 h (Fig. 2b), whereas insulin secretion did not significantly increase before 6 h (Fig. 2a). Islet protein synthesis was inhibited by 93 and 94% after 1 and 4 h of treatment with 10 ␮m CHX. The effect of the drug was thus rapid and complete at the concentration used. In

301

FIG. 2. Reversibility of the inhibition of insulin secretion and protein synthesis by CHX. a, Culture of the islets with 10 ␮M CHX for 12 h (0 –12 h) markedly inhibited insulin secretion as compared with control islets. At 12 h, the islets were transferred to a similar medium (E and F) or changed from a CHX-containing to a control medium (䡺). Insulin secretion was then monitored every 3 h between 12 and 21 h. Values are means ⫾ SEM for 21 batches of islets from three experiments. b, Other batches of islets were similarly treated except that [3,5-3H] tyrosine was added to the fresh culture medium at 12 h, and its incorporation into newly synthesized proteins was studied at 15 and 18 h. Values are means ⫾ SEM for eight batches of islets from two experiments.

FIG. 3. Onset of CHX-inhibition of insulin secretion and protein synthesis by mouse islets. After 18 h of culture without CHX, two batches of 25 islets were perifused in parallel with a medium containing 15 mM glucose. At time 0, following a 30-min stabilization period, 10 ␮M CHX was added to the medium perifusing test islets (F). Values are means ⫾ SEM for 6 experiments. To measure protein synthesis, batches of 20 islets were incubated for 1 or 4 h in the presence of [3,5-3H] tyrosine, with or without 10 ␮M CHX. The results were expressed as a percentage of controls run in parallel. Absolute values for these controls were 388 ⫾ 19 and 1154 ⫾ 85 cpm/islet after 1 and 4 h, respectively. Values are means ⫾ SEM for 14 batches of islets from three experiments.

contrast, no inhibition of glucose-induced insulin secretion occurred before about 1 h (Fig. 3). There was thus a clear temporal sequence in the effects of CHX: upon addition of the drug, the inhibition of insulin secretion was delayed and progressive compared with that of protein synthesis; upon withdrawal of CHX, recovery of insulin secretion followed resumption of protein synthesis. Insulin secretion by islets pretreated with cycloheximide

Islets from the same preparation were treated with 10 ␮m CHX for 2, 5, or 20 h, and their ability to secrete insulin in response to various stimuli was then tested in a perifusion system. In control islets, 15 mm glucose induced biphasic

302


Endo • 2001 Vol. 142 • No. 1

FIG. 4. Inhibition of glucose-, tolbutamide-, or high K⫹-induced insulin secretion in mouse islets pretreated for 2, 5, or 20 h with 10 ␮M CHX. An initial stabilization period of 30 min is not shown. CHX was withdrawn at 10 min. a, The concentration of glucose (G) was increased from 3 to 15 mM and 100 ␮M tolbutamide (Tolb) was then added as indicated. b, The medium contained 15 mM glucose and 250 ␮M diazoxide (Dz) throughout, and the concentration of K⫹ was increased from 4.8 to 30 mM as indicated. c– e, Average secretion rates during application of each stimulus to islets pretreated with 10 ␮M CHX for different periods of time. Values are means ⫾ SEM for 6 experiments. *, P ⬍ 0.001 vs. controls without CHX pretreatment.

insulin secretion, and subsequent addition of 100 ␮m tolbutamide strongly potentiated the response (Fig. 4a). By opening K⫹-ATP channels and preventing ␤ cell depolarization (32, 33), diazoxide abolished glucose-induced insulin secretion, but a rise in extracellular K⫹ to 30 mm triggered a large monophasic response (Fig. 4b) (7). After 20 h of CHX treatment, the stimulatory effects of glucose, tolbutamide, and high K⫹ were markedly (60 – 80%) impaired. The inhibition increased with the duration of CHX pretreatment, but the major impairment (P ⬍ 0.001) occurred early, during the first two hours, regardless of the stimulus (Fig. 4, c– e). In another series of experiments the islets were treated with 10 ␮m CHX for 2 h before being incubated in batches of 3 in 1 ml of medium containing 15 mm glucose. Insulin secretion averaged 3.8 ⫾ 0.15 and 1.7 ⫾ 0.16 ng/islet/h in control and CHX-treated islets, respectively (n ⫽ 16, P ⬍ 0.01). In control islets insulin secretion was markedly potentiated by 0.5 mm dibutyryl cAMP (342 ⫾ 18%) or 25 nm PMA (208 ⫾ 8%). Similar potentiations were observed when CHXtreated islets were stimulated with dibutyryl cAMP (348 ⫾ 29%) or PMA (284 ⫾ 16%). The protein kinase A and protein kinase C pathways are thus functional after 2 h of treatment with CHX. Glucose metabolism by islets pretreated with cycloheximide

The effect of glucose on ␤ cell metabolism was first measured by recording the NAD(P)H fluorescence of the islets (34, 35). Raising the concentration of glucose from 3 to 15 mm markedly and reversibly increased the signal (Fig. 5a). The response was not different from that of controls after 5 h of treatment with 10 ␮m CHX, but it was partially impaired (P ⬍ 0.01) after 20 h of treatment. This slight alteration of glucose metabolism was confirmed by measurements of glucose oxidation (Fig. 5b). In both low and high glucose, CO2 pro-

FIG. 5. Glucose metabolism by islets pretreated with 10 ␮M CHX for 5 or 20 h. a, Changes in reduced pyridine nucleotides (NAD[P]H fluorescence) evoked by an increase in glucose (G) concentration from 3 to 15 mM. CHX was present until transfer of the islets to the recording system. Values are means ⫾ SEM for 10 islets from five experiments. b, Glucose oxidation by islets incubated for 2 h in the presence of 3 or 15 mM glucose CHX was present during the incubation. Values are means ⫾ SEM for 10 –15 batches of islets from two to three experiments. *, P ⬍ 0.01 vs. controls without CHX pretreatment.

duction from glucose was slightly reduced by 20 h CHX treatment (20 and 16%, respectively, P ⬍ 0.01). The inhibition was not larger when the concentration of CHX was raised to 100 ␮m during the 20 h of treatment (not shown), indicating that the small change in glucose metabolism is secondary to inhibition of protein synthesis (already maximal with 10 ␮m CHX) rather than to a side effect of the drug. K⫹ efflux in islets and K⫹-ATP currents in ␤ cells pretreated with cycloheximide

Loading the islets with 86Rb⫹ and monitoring the efflux of this tracer of K⫹ provides instructive information concerning


the regulation of K⫹ permeability in ␤ cells under basal and stimulating conditions (36). In control islets perifused with a glucose-free medium, the rate of 86Rb⫹ efflux was high, reflecting a high permeability of the membrane to K⫹ (Fig. 6). Addition of 3 and then 15 mm glucose to the medium (Fig. 6a) or addition of 25 and then 100 ␮m tolbutamide to the glucose-free medium (Fig. 6b) induced a graded and reversible inhibition of 86Rb⫹ efflux that is known to reflect closure of K⫹-ATP channels (4, 36). Similar changes were produced by glucose and tolbutamide when islets had been pretreated with 10 ␮m CHX for 5 h. After 20 h of treatment, however, the basal rate of 86Rb⫹ efflux was lower than in controls (P ⬍ 0.01), but glucose and tolbutamide still decreased it to a similar minimum rate as in control islets. The magnitude of the inhibitory effect was thus smaller, but the absolute final rate was not different. The perforated whole-cell configuration of the patch clamp technique was used to study total K⫹-ATP currents in ␤ cells. A representative experiment is shown in Fig. 7a. K⫹-ATP currents were initially small because of the presence of 10 mm glucose in the medium. Lowering ␤ cell ATP with 3 mm azide and direct opening of the channels with diazoxide caused a marked increase in K⫹-ATP currents, which was reversed by tolbutamide. In 10 mm glucose alone, K⫹-ATP currents were not significantly different in ␤ cells pretreated with 10 ␮m CHX for 20 h and in controls (Fig. 7b). However, the increase produced by azide and diazoxide was smaller

FIG. 6. 86Rb⫹ efflux from islets pretreated with 10 ␮M CHX for 5 or 20 h. An initial stabilization period of 30 min is not shown. CHX was withdrawn at 10 min. a, The concentration of glucose was increased to 3 and 15 mM as indicated. b, Tolbutamide was added to a glucosefree medium at the indicated concentrations. Values are means ⫾ SEM for five experiments.

303

after CHX (P ⬍ 0.01), which suggests that the total number of functional K⫹-ATP channels was decreased. The relative inhibitory action of tolbutamide was similar in control and CHX treated ␤ cells, so that a significant difference (P ⬍ 0.01) in total current persisted in the presence of the three drugs. Ca2⫹ currents in ␤ cells and [Ca2⫹]i changes in islets pretreated with cycloheximide

Depolarization of ␤ cells to ⫺40 mV or above elicited typical, partially inactivating, Ca2⫹ currents (Fig. 7c), that increased in amplitude up to 0 mV and then decreased at more positive potentials (Fig. 7d). After pretreatment of the cells with 10 ␮m CHX for 20 h, both peak and sustained currents were inhibited by about 40% (P ⬍ 0.001). The percent block was similar at all voltages between ⫺30 and ⫹50 mV. Stimulation of control islets with a rise in the glucose concentration from 3 to 15 mm induced [Ca2⫹]i changes in three phases: an initial small decrease (seen in 20/24 islets) was followed by a large increase and eventually by oscillations. These oscillations, illustrated for a representative islet in the inset, are masked by averaging of the traces (Fig. 8a). After treatment of the islets with 10 ␮m CHX for 20 h, basal [Ca2⫹]i was not significantly changed but 15 mm glucose only evoked a monophasic rise: the initial drop was consistently abolished, the increase in [Ca2⫹]i started sooner than in controls, and no oscillations occurred in any of the 16 tested islets. Average [Ca2⫹]i was not significantly decreased, but the difference between basal and steady-state [Ca2⫹]i in 15 mm glucose was 35% smaller after 20 h CHX than in controls (P ⬍ 0.001). After treatment of the islets with CHX for 5 h, the initial [Ca2⫹]i decrease persisted in only 6/20 islets, and the subsequent rise was attenuated, although the difference with controls was significant (P ⬍ 0.05) only for the change above basal levels. During steady-state stimulation, oscillations of [Ca2⫹]i undistinguishable from control ones were present in all islets (not shown). Treatment of the islets with CHX for only 2 h had no quantitative or qualitative impact on glucose-induced [Ca2⫹]i changes. Addition of 100 ␮m tolbutamide to the medium containing 15 mm glucose caused a prompt and sustained rise in [Ca2⫹]i in control islets (Fig. 8a). This rise was smaller in islets treated with 10 ␮m CHX for 5 h (P ⬍ 0.05) and 20 h (P ⬍ 0.001), but it was unaltered after 2 h of CHX treatment (Fig. 8, a and d). When diazoxide was present in the medium containing 15 mm glucose, [Ca2⫹]i was low and stable, but a large and sustained rise occurred when extracellular K⫹ was increased to 30 mm. This rise was inhibited (P ⬍ 0.001) by pretreatment with 10 ␮m CHX for 20 h (Fig. 8, b and e). Two hours of treatment had no effect, and the small inhibition by 5 h of treatment was significant (P ⬍ 0.05) only when calculated above baseline. Links between CHX-induced [Ca2⫹]i and insulin secretion changes

The results presented in Figs. 4 and 8 showed that 2 h of CHX treatment inhibited insulin secretion without significantly affecting [Ca2⫹]i, whereas longer treatment decreased the Ca2⫹ signal as well. Figure 9 displays the relationship

304


Endo • 2001 Vol. 142 • No. 1

FIG. 7. K⫹-ATP and Ca2⫹ currents in ␤ cells pretreated with 10 ␮M CHX for 20 h. a, Representative control ␤ cell: K⫹-ATP currents, initially small because of the presence of 10 mM glucose in the medium, were markedly increased by addition of 250 ␮M diazoxide (Dz) and 3 mM azide; this increase was reversed by 100 ␮M tolbutamide (Tolb). b, Quantification of these changes in 10 –18 control and CHX-treated cells from 5 separate cultures. *, P ⬍ 0.01 vs. controls without CHX pretreatment. c, Mean Ca2⫹ current evoked by depolarization of control and CHX-treated ␤ cells from ⫺80 to 0 mV. d, Current-voltage relationships for peak Ca2⫹ currents recorded in control and CHX-pretreated ␤ cells. Values are means ⫾ SEM for 31 control and 21 CHX cells from 4 different cultures.

between [Ca2⫹]i and insulin secretion during stimulation by 15 mm glucose alone, with tolbutamide or with high K⫹ and diazoxide. For each stimulus are shown, linked by the broken lines, the results obtained in control islets (open symbols) and in islets treated with CHX for 2, 5 and 20 h (filled symbols from right to left). Two excellent correlations with distinct slopes were found for control and CHX-treated islets. Fitting all data with a single regression line resulted in a much weaker coefficient of correlation (R ⫽ 0.79). For each stimulus, the inhibition of insulin secretion by CHX treatment appeared to involve two successive changes: an initial decrease in the efficiency of Ca2⫹ (after 2 h) followed by a decrease in [Ca2⫹]i. We finally tested whether the reversibility of the inhibition by CHX might also occur in two stages. As shown in Fig. 10, a and b, 5 h of CHX washing after 15 h pretreatment was followed by a greater reversibility of insulin secretion than [Ca2⫹]i during stimulation with high K⫹. The values of these new control and test experiments fell close to the regression lines calculated from the first experimental series (Fig. 10c). The critical observation was that the value for the reversibility group did not move along the lower curve of CHXtreated islets, but jumped to the upper curve of control islets. The recovery of secretion thus mainly involved an increase in Ca2⫹ efficiency. Discussion

Our study shows that inhibition of protein synthesis in pancreatic islets inhibits insulin secretion by interfering with distinct steps of stimulus-secretion coupling in a time-dependent manner, compatible with the sequential disappearance of distinct key proteins of short half-life. Rapid and virtually complete (⬃95% after 1 h) blockade of protein synthesis was achieved by 10 ␮m CHX, an inhibitor of translation. The sensitivity of our preparation of mouse islets to CHX was similar to that of rat islets in which 1–10 ␮m CHX inhibited protein synthesis by 75–95% (37–39). The

present results confirm the observations (16 –19) that the inhibition of insulin secretion by CHX starts only after about 1 h of drug application. They further show that there is a time lag between the reversibility of the inhibition of protein synthesis and insulin secretion. Thus, both onset and offset of the inhibition of secretion are delayed compared with those of protein synthesis. Several observations also indicate that the changes in ␤ cell function produced by CHX are not secondary to a crude toxic effect: their reversibility, the minor alteration of glucose metabolism even after 20 h, and the longer delay (2–3 days) before appearance of apoptosis in CHX-treated ␤ cells (40). It seems reasonable to propose that the time-dependent changes in stimulus-secretion coupling produced by addition and removal of CHX result from disappearance and reappearance of proteins with distinct half-lives. The inhibition of insulin secretion measured after 2 h of CHX treatment cannot be explained by insufficient total insulin stores, but could theoretically result from the depletion of a finite, small pool of insulin that cannot be replenished because of the suppression of insulin biosynthesis. Newly synthesized insulin can be released rapidly without the newly formed granules mixing with the older ones (20, 21, 41, 42). However, this only corresponds to a small (⬍10%) fraction of total insulin secretion (20, 21, 43), much smaller than the approximately 50% inhibition observed here already after 2 h. After longer treatment of the islets with CHX (18 h) insulin reserves were higher than in untreated controls. We, therefore, conclude that neither the initial nor the sustained effect of CHX on insulin secretion is mediated by the loss of a particular pool of insulin. Insulin secretion is tightly dependent on glucose metabolism in ␤ cells. A slight (⬃15%) decrease in glucose-induced rise in NAD(P)H fluorescence and in glucose oxidation was observed after 20 h of protein synthesis inhibition. This inhibition of glucose metabolism may contribute to the delayed


305

FIG. 8. Inhibition of glucose-, tolbutamide-, or high K⫹-induced [Ca2⫹]i rise in mouse islets pretreated with 10 ␮M CHX for 2, 5, or 20 h. The recordings were preceded by a 4 –5 min stabilization period. CHX was not present during the experiments. a, The concentration of glucose (G) was increased from 3 to 15 mM and 100 ␮M tolbutamide (Tolb) was then added as indicated. The inset shows typical [Ca2⫹]i oscillations in one control islet (same time scale as for main panel). b, The medium contained 15 mM glucose and 250 ␮M diazoxide (Dz) throughout, and the concentration of K⫹ was increased from 4.8 to 30 mM as indicated. c– e, Average [Ca2⫹]i during application of each stimulus to islets pretreated with 10 ␮M CHX for different periods of time. Values are means ⫾ SEM for 16 –24 islets from 4 – 6 cultures (a, c, d) or 16 islets from 4 cultures (b, e). *, P ⬍ 0.05 or less vs. controls without CHX pretreatment.

FIG. 9. Correlations between insulin secretion and [Ca2⫹]i in control islets (open symbols) and islets pretreated with 10 ␮M CHX for different periods of time (filled symbols). For each stimulus (glucose, tolbutamide or high K⫹), 4 data points are shown and linked by broken lines: one open symbol corresponding to control values in the absence of CHX and three filled symbols corresponding, from right to left, to values measured after 2, 5, and 20 h of CHX treatment. Separate regression lines were calculated for control and test islets. Values are taken from Figs. 4 and 8.

inhibition of insulin secretion, but it is too small to explain it entirely. In contrast, we did not observe any impairment of the NAD(P)H signal after 5 h of CHX treatment, and others have reported that glycolysis and glucose oxidation in mouse islets are unaffected by 1 mm CHX after 2–3 h (44, 45). The inhibition of insulin secretion occurring during the first hours of protein synthesis inhibition cannot be ascribed to any major defect in glucose metabolism. It can also be concluded that the proteins involved in glucose metabolism have a relatively long half-life, which may explain why ␤ cells can survive a few days without protein synthesis (40). K⫹-ATP channels serve as transduction units between nutrient metabolism and biophysical events in ␤ cells, and are

the direct target of sulfonylureas (4 – 6, 15). An inhibition of insulin secretion could result from an impairment of the ability of glucose or tolbutamide to close the channels. Neither 86Rb⫹ efflux measurements nor patch-clamp recordings suggested that this might be the case after 5 or 20 h of protein synthesis inhibition. The maximum K⫹-ATP current activated by metabolic poisoning and diazoxide was decreased by approximately 50% in ␤ cells treated with CHX for 20 h. Because over 90% K⫹-ATP channels must be closed before the ␤ cell membrane depolarizes (46, 47), a 50% decrease in available channels is not expected to cause major effects on ␤ cell function. It may explain, why basal 86Rb⫹ efflux was slightly decreased in test islets, but it is important to note that this decrease was less than that produced in control islets by 3 mm glucose, a concentration that is insufficient to depolarize ␤ cells (36). Upon membrane depolarization, voltage-dependent Ca2⫹ channels open and allow an influx of Ca2⫹ that raises [Ca2⫹]i in ␤ cells (3– 6, 27). Not only glucose and tolbutamide, but also high K⫹, that depolarizes independently of K⫹-ATP channels (by shifting the equilibrium potential for K⫹), were less effective in increasing [Ca2⫹]i after 20 h of protein synthesis inhibition. This reduced efficacy is entirely compatible with the commensurate decrease in Ca2⫹ current measured in similarly treated ␤ cells. Whether the actual number of Ca2⫹ channels is decreased by 40% under these conditions or whether the availability of the channels is reduced because of the loss of a regulatory protein cannot be answered by these experiments. Two other aspects of the specific [Ca2⫹]i response to glucose were perturbed after 20 h of CHX treatment: the small decrease preceding the initial rise, and the oscillations occurring during sustained stimulation were abrogated. The significance of these changes is difficult to establish because the inhibition of glucose-induced insulin se-

306


FIG. 10. Comparison of the reversibility of CHX inhibition of [Ca2⫹]i and insulin secretion. Three groups of islets were cultured and preincubated in parallel: control islets (C) were never treated with CHX; the second group of islets remained in the presence of 10 ␮M CHX for 20 h (CHX); the third group (R) was treated with 10 ␮M CHX for 15 h before being transferred to control medium for 5 h (including the preincubation for loading with fura-PE3). a and b, During measurements of insulin secretion or [Ca2⫹]i the medium contained 15 mM glucose (G) and 250 ␮M diazoxide (Dz) throughout, and the concentration of K⫹ was raised from 4.8 to 30 mM as indicated. Values are means ⫾ SEM for 6 experiments (secretion) and 21 islets from 3 preparations ([Ca2⫹]i). c, Insulin secretion and [Ca2⫹]i measured in the 3 groups of islets were superimposed on the regression lines calculated in Fig. 9.

cretion after 5 h of CHX was accompanied by an absence of the initial [Ca2⫹]i decrease in 70% of the islets but a persistence of [Ca2⫹]i oscillations in all islets. Anyhow it is intriguing that the small initial decrease, which reflects glucoseinduced Ca2⫹ sequestration in the endoplasmic reticulum (48), is lost in ␤ cells from diabetic db/db mice (49). From our data, one may conclude that a lesser rise in ␤ cell [Ca2⫹]i, secondary to a decrease in functional Ca2⫹ channels, contributes to the inhibition of insulin secretion probably after 5 and certainly after 20 h of protein synthesis inhibition. However, no similar mechanism explains the inhibition of insulin secretion after 2 h. A major observation of the present study was that short (2 h) treatment of the islets with CHX inhibited insulin secretion without impairing the rise in [Ca2⫹]i produced by three distinct agents, glucose, tolbutamide and high K⫹. This dissociation indicates that the efficacy of Ca2⫹ on secretion was decreased. Thus, the relationship between [Ca2⫹]i and insulin secretion in response to the three agents could be fitted by distinct, tight correlations corresponding to the absence and presence of CHX, respectively. Moreover, after resumption of protein synthesis (after CHX washing), the recovery of insulin secretion was more dependent on restoration of the

Endo • 2001 Vol. 142 • No. 1

action of Ca2⫹ than elevation of [Ca2⫹]i. Altogether our data indicate that protein synthesis inhibition in ␤ cells inhibits insulin secretion in sequential steps: a decrease in the efficiency of Ca2⫹, followed by a decrease in Ca2⫹ influx and glucose metabolism. The synthesis of a distinct subset of islet proteins other than proinsulin is stimulated by glucose (50 –52). Among these proteins, other secretory peptides and proteolytic enzymes involved in prohormone processing have been identified. Many others are unknown and might include proteins whose loss rapidly leads to impairment of insulin secretion. We have not identified these proteins that seem to be distinct from protein kinases A and C and their targets because activation of these kinases normally potentiated insulin secretion after short treatment with CHX. We speculate that these proteins might be implicated in the amplification pathway of glucose-induced insulin secretion (1, 7–9), the pathway through which glucose increases the efficiency of Ca2⫹ on exocytosis of insulin granules. It is, however, not necessary to postulate that these proteins with a short half-life are glucose dependent. They could also be important in other secretory systems. In conclusion, islet protein inhibition impairs insulin secretion by a sequential alteration of distinct steps of stimulussecretion coupling. Glucose metabolism is not readily affected, which indicates that the involved enzymes have a relatively long half-life. K⫹-ATP and Ca2⫹ channels may be similarly affected but only the decrease in Ca2⫹ channels exerts a negative impact on insulin secretion. Unexpectedly, among all ␤ cell proteins implicated in the regulation of insulin secretion, that or those with the shortest half-life appear to modulate (directly or indirectly) the action of Ca2⫹ on exocytosis. These characteristics make them the potential Achille’s heel of stimulus-secretion coupling in ␤ cells. Acknowledgments We are grateful to Dr J.C. Jonas for advice and comments on the manuscript, to Fabien Knockaert for technical assistance and to Ste´phanie Roiseux for editorial help.

References 1. Henquin JC 2000 The triggering and amplifying pathways of the regulation of insulin secretion by glucose. Diabetes 49:1751–1760 2. Prentki M, Matschinsky FM 1987 Ca2⫹, cAMP, and phospholipid-derived messengers in coupling mechanisms of insulin secretion. Physiol Rev 67:1185–1248 3. Newgard CB, McGarry JD 1995 Metabolic coupling factors in pancreatic ␤-cell signal transduction. Annu Rev Biochem 64:689 –719 4. Ashcroft FM, Rorsman P 1989 Electrophysiology of the pancreatic ␤-cell. Prog Biophys Mol Biol 54:87–143 5. Henquin JC, Jonas JC, Sato Y, Detimary P, Gilon P 1999 Signal transduction: regulation of insulin secretion by changes in Ca2⫹ concentration and action in pancreatic ␤ cells. In: Howell SL (ed) Advances in Molecular and Cell Biology, The Biology of the Pancreatic ␤ Cell. JAI Press, Greenwich, vol 29:247–275 6. Seino S, Iwanaga T, Nagashima K, Miki T 2000 Diverse roles of KATP channels learned from Kir6.2 genetically engineered mice. Diabetes 49:311–318 7. Gembal M, Gilon P, Henquin JC 1992 Evidence that glucose can control insulin release independently from its action on ATP-sensitive K⫹ channels in mouse B-cells. J Clin Invest 89:1288 –1295 8. Sato Y, Henquin JC 1998 The K⫹-ATP channel-independent pathway of regulation of insulin secretion by glucose. In search of the underlying mechanism. Diabetes 47:1713–1721 9. Aizawa T, Komatsu M, Asanuma N, Sato Y, Sharp GWG 1998 Glucose action ‘beyond ionic events’ in the pancreatic ␤ cell. Trends Pharmacol Sci 19:496 – 499 10. Froguel P, Zouali H, Vionnet N, Velho G, Vaxillaire M, Sun F, Lesage S, Stoffel M, Takeda J, Passa P, Permutt MA, Beckman JS, Bell JI, Cohen D 1993


11.

12.

13. 14.

15. 16. 17. 18. 19. 20. 21. 22. 23.

24. 25. 26. 27. 28. 29. 30. 31.

Familial hyperglycemia due to mutations in glucokinase. Definition of a subtype of diabetes mellitus. N Engl J Med 328:697–702 Glaser B, Kesavan P, Heyman M, Davis E, Cuesta A, Buchs A, Stanley CA, Thornton PS, Permutt MA, Matschinsky FM, Herold KC 1998 Familial hyperinsulinism caused by an activating glucokinase mutation. N Engl J Med 338:226 –230 Davis EA, Cuesta-Munoz A, Raoul M, Buettger C, Sweet I, Moates M, Magnuson MA, Matschinsky FM 1999 Mutants of glucokinase cause hypoglycaemia- and hyperglycaemia syndromes and their analysis illuminates fundamental quantitative concepts of glucose homeostasis. Diabetologia 42:1175–1186 Thomas PM, Cote GJ, Wohlik N, Haddad B, Mathew PM, Rabl W, AguilarBryan L, Gagel RF, Bryan J 1995 Mutations in the sulfonylurea receptor gene in familial hyperinsulinemic hypoglycemia of infancy. Science 268:426 – 429 Kane C, Sheperd RM, Squires PE, Johnson PR, James RF, Milla PJ, AynsleyGreen A, Lindley KJ, Dunne MJ 1996 Loss of functional KATP channels in pancreatic ␤-cells causes persistent hyperinsulinemic hypoglycemia of infancy. Nat Med 2:1344 –1347 Aguilar-Bryan L, Bryan J 1999 Molecular biology of adenosine triphosphatesensitive potassium channels. Endocr Rev 20:101–135 Curry DL, Bennett LL, Grodsky GM 1968 Dynamics of insulin secretion by the perfused rat pancreas. Endocrinology 83:572–584 Basabe JC, Lopez N, Viktora J, Wolff F 1970 Studies of insulin secretion in the perfused rat pancreas. Effect of diazoxide and A025. Diabetes 19:271–281 Curry DL 1971 Insulin secretory dynamics in response to slow-rise and squarewave stimuli. Am J Physiol 221:324 –328 Morris GE, Korner A 1970 The effect of glucose on insulin biosynthesis by isolated islets of Langerhans of the rat. Biochim Biophys Acta 208:404 – 413 Sando H, Borg J, Steiner DF 1972 Studies on the secretion of newly synthesized proinsulin and insulin from isolated rat islets of Langerhans. J Clin Invest 51:1476 –1485 Sando H, Grodsky GM 1973 Dynamic synthesis and release of insulin and proinsulin from perifused islets. Diabetes 22:354 –360 Jonas JC, Gilon P, Henquin JC 1998 Temporal and quantitative correlations between insulin secretion and stably elevated or oscillatory cytoplasmic Ca2⫹ in mouse pancreatic ␤ cells. Diabetes 47:1266 –1273 Detimary P, Jonas JC, Henquin JC 1995 Possible links between glucoseinduced changes in the energy state of pancreatic B-cells and insulin release: unmasking by decreasing a stable pool of adenine nucleotides in mouse islets. J Clin Invest 96:1738 –1745 Hales CN, Randle PJ 1963 Immunoassay of insulin with insulin-antibody precipitate. Biochem J 88:133–146 Henquin JC 1978 D-glucose inhibits potassium efflux from pancreatic islet cells. Nature 271:271–273 Henquin JC, Lambert AE 1976 Bicarbonate modulation of glucose-induced biphasic insulin release by rat islets. Am J Physiol 231:713–721 Gilon P, Henquin JC 1992 Influence of membrane potential changes on cytoplasmic Ca2⫹ concentration in an electrically excitable cell, the insulinsecreting pancreatic B-cell. J Biol Chem 267:20713–20720 Ravier MA, Gilon P, Henquin JC 1999 Oscillations of insulin secretion can be triggered by imposed oscillations of cytoplasmic Ca2⫹ or metabolism in normal mouse islets. Diabetes 48:2374 –2382 Jonkers FC, Jonas JC, Gilon P, Henquin JC 1999 Influence of cell number on the characteristics and synchrony of Ca2⫹ oscillations in clusters of mouse pancreatic islet cells. J Physiol 520:839 – 849 Rae J, Cooper K, Gates P, Watsky M 1991 Low access resistance perforated patch recordings using amphotericin B. J Neurosci Methods 37:15–26 Mariot P, Gilon P, Nenquin M, Henquin JC 1998 Tolbutamide and diazoxide influence insulin secretion by changing the concentration not the action of cytoplasmic Ca2⫹ in ␤ cells. Diabetes 47:365–373

307

32. Henquin JC, Meissner HP 1982 Opposite effects of tolbutamide and diazoxide on 86Rb⫹ fluxes and membrane potential in pancreatic B-cells. Biochem Pharmacol 31:1407–1415 33. Trube G, Rorsman P, Ohno Shosaku T 1986 Opposite effects of tolbutamide and diazoxide on the ATP-dependent K⫹ channel in mouse pancreatic ␤-cells. Pflugers Arch 407:493– 499 34. Panten U, Christians J, Kriegstein EV, Poser W, Hasselblatt A 1973 Effect of carbohydrates upon fluorescence of reduced pyridine nucleotides from perifused isolated pancreatic islets. Diabetologia 9:477– 482 35. Duchen MR, Smith PA, Ashcroft FM 1993 Substrate-dependent changes in mitochondrial function, intracellular free calcium concentration and membrane channels in pancreatic ␤-cells. Biochem J 294:35– 42 36. Henquin JC, Meissner HP 1984 Significance of ionic fluxes and changes in membrane potential for stimulus-secretion coupling in pancreatic ␤-cells. Experientia 40:1043–1052 37. Permutt MA 1974 Effect of glucose on initiation and elongation rates in isolated rat pancreatic islets. J Biol Chem 248:2738 –2742 38. Welsh M, Scherberg N, Gilmore R, Steiner DF 1986 Translational control of insulin biosynthesis. Biochem J 235:459 – 467 39. Hughes JH, Colca JR, Easom RA, Turk J, McDaniel ML 1990 Interleukin 1 inhibits insulin secretion from isolated rat pancreatic islets by a process that requires gene transcription and mRNA translation. J Clin Invest 86:856 – 863 40. Hoorens A, Van de Casteele M, Klo¨ppel G, Pipeleers D 1996 Glucose promotes survival of rat pancreatic ␤ cells by activating synthesis of proteins which suppress a constitutive apoptotic program. J Clin Invest 98:1568 –1574 41. Halban PA 1982 Differential rates of release of newly synthesized and of stored insulin from pancreatic islets. Endocrinology 110:1183–1188 42. Rhodes CJ, Halban PA 1987 Newly synthesized proinsulin/insulin and stored insulin are released from pancreatic ␤ cells predominantly via a regulated, rather than a constitutive, pathway. J Cell Biol 105:145–153 43. Schuit FC, Kiekens R, Pipeleers DG 1991 Measuring the balance between insulin synthesis and insulin release. Biochem Biophys Res Commun 178:1182–1187 44. Ashcroft SJH, Hedeskov CJ, Randle PJ 1970 Glucose metabolism in mouse pancreatic islets. Biochem J 118:143–154 45. Ashcroft SJH, Weerasinghe LCC, Bassett JM, Randle PJ 1972 The pentose cycle and insulin release in mouse pancreatic islets. Biochem J 126:525–532 46. Cook DL, Satin LS, Ashford MLJ, Hales CN 1988 ATP-sensitive K⫹ channels in pancreatic ␤-cells: spare-channel hypothesis. Diabetes 37:495– 498 47. Ashcroft FM, Ashcroft SJH, Harrison DE 1988 Properties of single potassium channels modulated by glucose in rat pancreatic ␤-cells. J Physiol 400:501–527 48. Roe MW, Mertz RJ, Lancaster ME, Worley III JF, Dukes ID 1994 Thapsigargin inhibits the glucose-induced decrease of intracellular Ca2⫹ in mouse islets of Langerhans. Am J Physiol 266:E852–E862 49. Roe MW, Philipson LH, Frangakis CJ, Kuznetsov A, Mertz RJ, Lancaster ME, Spencer B, Worley III JF, Dukes ID 1994 Defective glucose-dependent endoplasmic reticulum Ca2⫹ sequestration in diabetic mouse islets of Langerhans. J Biol Chem 269:18279 –18282 50. Nagamatsu S, Bolaffi JL, Grodsky GM 1987 Direct effects of glucose on proinsulin synthesis and processing during desensitization. Endocrinology 120:1225–1231 51. Guest PC, Bailyes EM, Rutherford NA, Hutton JC 1991 Insulin secretory granule biogenesis: coordinate regulation of the biosynthesis of the majority of constituant proteins. Biochem J 274:73–78 52. Collins H, Najafi H, Buettger C, Rombeau J, Settle RG, Matschinsky FM 1992 Identification of glucose response proteins in two biological models of ␤-cell adaptation to chronic high glucose exposure. J Biol Chem 267:1357–1366