Stimulatory concentrations of glucose induce two patterns of cytosolic Ca2+ concentration ([Ca2+]c) oscillations in mouse islets: simple or mixed. In the mixed pattern, rapid oscillations are superimposed on slow ones. In the present study, we examined the role of the membrane potential in the mixed pattern and the impact of this pattern on insulin release. Simultaneous measurement of [Ca2+]c and insulin release from single islets revealed that mixed [Ca2+]c oscillations triggered synchronous oscillations of insulin secretion. Simultaneous recordings of membrane potential in a single β-cell within an islet and of [Ca2+]c in the whole islet demonstrated that the mixed pattern resulted from compound bursting (i.e., clusters of membrane potential oscillations separated by prolonged silent intervals) that was synchronized in most β-cells of the islet. Each slow [Ca2+]c increase during mixed oscillations was due to a progressive summation of rapid oscillations. Digital image analysis confirmed the good synchrony between subregions of an islet. By contrast, islets from sarco(endo)plasmic reticulum Ca2+-ATPase isoform 3 (SERCA3)-knockout mice did not display typical mixed [Ca2+]c oscillations in response to glucose. This results from a lack of progressive summation of rapid oscillations and from altered spontaneous electrical activity, i.e., lack of compound bursting, and membrane potential oscillations characterized by lower-frequency but larger-depolarization phases than observed in SERCA3+/+ β-cells. We conclude that glucose-induced mixed [Ca2+]c oscillations result from compound bursting in all β-cells of the islet. Disruption of SERCA3 abolishes mixed [Ca2+]c oscillations and augments β-cell depolarization. This latter observation indicates that the endoplasmic reticulum participates in the control of the β-cell membrane potential during glucose stimulation.
- electrical activity
- insulin-secreting cell
pancreatic β-cells are electrically excitable and transduce variations in the concentration of glucose and other nutrients into changes in the concentration of free cytosolic Ca2+ ([Ca2+]c), which eventually trigger insulin secretion. Glucose-stimulated mouse islets display simple [Ca2+]c oscillations or mixed [Ca2+]c oscillations characterized by rapid oscillations superimposed on slow ones (3, 16, 17, 34). Simple [Ca2+]c oscillations result from synchronous oscillations of the membrane potential that induce intermittent Ca2+ influx through voltage-dependent Ca2+ channels (15, 33). The origin of mixed [Ca2+]c oscillations is unclear. It has been suggested that the mixed pattern is due to the combination of distinct responses originating from separate cell populations within the islet (28). This proposal is not supported by studies showing that [Ca2+]c oscillations are generally similar and synchronous in different regions of the islet (15, 33). We instead hypothesize that these mixed [Ca2+]c oscillations are induced by the as yet unexplained compound bursting that is characterized by a series of membrane potential oscillations of changing duration that occur in bursts and are separated by prolonged silent intervals. This peculiar electrical activity has been documented in a number of glucose-stimulated islets (4, 9, 21).
The endoplasmic reticulum (ER) takes up cytosolic Ca2+ by sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA) (13, 41). Pancreatic β-cells express several SERCA isoforms, including SERCA2b and SERCA3 (12, 36) and possibly SERCA2a also (27). Whereas SERCA2 has a high affinity for Ca2+ and controls basal [Ca2+]c, SERCA3 has a low affinity for Ca2+ and influences depolarization-induced [Ca2+]c oscillations by taking up cytosolic Ca2+ during each period of Ca2+ influx (1). This latter isoform might therefore affect the [Ca2+]c oscillation pattern.
In the present study, we monitored [Ca2+]c in intact islets and the membrane potential of a β-cell simultaneously within the same islet to evaluate whether mixed [Ca2+]c oscillations are induced by compound bursting. Using digital image analysis, we determined whether mixed [Ca2+]c oscillations are present and synchronized in all islet regions. By measuring [Ca2+]c and insulin release from single islets simultaneously, we evaluated the influence of mixed [Ca2+]c oscillations on insulin secretion. We then tested the hypothesis that SERCA3 is implicated in the generation of mixed [Ca2+]c oscillations in experiments with islets from SERCA3-knockout (SERCA3−/−) mice.
MATERIALS AND METHODS
Solutions and drugs.
The medium used for all experiments was a HCO3−-buffered solution containing (in mM) 120 NaCl, 4.8 KCl, 2.5 CaCl2, 1.2 MgCl2, 24 NaHCO3, and various concentrations of glucose (3–15 mM as indicated). This medium was supplemented with 1 mg/ml BSA (fraction V; Boehringer-Mannheim, Mannheim, Germany) and gassed with O2-CO2 (94:6 ratio) to maintain pH 7.4 at 37°C. For the preparation of islets, the medium contained 5 mM HEPES and 10 mM glucose. For membrane potential recordings, BSA was omitted. Thapsigargin was purchased from Alomone Laboratories (Jerusalem, Israel) or from Sigma (St. Louis, MO), and diazoxide was a gift from Schering-Plough Avondale (Rathdrum, Ireland).
Animals and cells.
The research project was approved by, and the experiments were conducted in accordance with, the guidelines of the Commission d’Ethique et d’Expérimentation Animale of the University of Louvain Faculty of Medicine. Mice were killed by cervical dislocation and decapitation. Islets of Langerhans were aseptically isolated after collagenase digestion of the pancreas from Naval Medical Research Institute (NMRI) mice (only for insulin secretion experiments), SERCA3−/− mice, or SERCA3+/+ mice. Both male and female SERCA3−/− mice had a lower weight (24.8 ± 0.6 g, n = 25) than their C57BL/6J controls (SERCA3+/+) bred from their wild-type littermates (30.7 ± 1 g, n = 21; P < 0.01). Blood glucose and plasma insulin concentrations from fed animals were lower in SERCA3−/− (5.25 ± 0.15 mM glucose, 1.14 ± 0.09 ng insulin/ml, n = 25) than in SERCA3+/+ mice (6.31 ± 0.15 mM glucose, 2.3 ± 0.3 ng insulin/ml, n = 21; P < 0.01).
Islets were cultured in RPMI 1640 medium containing 10% heat-inactivated FCS, 100 IU/ml penicillin, 100 μg/ml streptomycin, and 10 mM glucose. For the experiments shown in Fig. 8, islets were cultured on glass coverslips for 3–4 days. This period was necessary for the islets to attach firmly to coverslips and sustain frequent solution changes at a high flow rate. For all other experiments, islets were cultured for 1 day. Single cells from flexor digitorum brevis muscles and single cardiomyocytes of NMRI mice were prepared as described previously (11, 29).
Islets were loaded for 2 h at 37°C with 2 μM fura-PE3 AM (MoBiTec, Göttingen, Germany). The loading solution was a HCO3−-buffered solution containing 10 mM glucose. A glass coverslip was used as the bottom of an ∼1-ml temperature-controlled perifusion chamber mounted on the stage of an inverted microscope. The islets were held in place by suction using a glass micropipette. The temperature within the chamber was 37°C. [Ca2+]c was measured using dual-wavelength (340 and 380 mm) excitation microspectrofluorometry with a photometry-based system (Photon Technologies International, Princeton, NJ), a Photometrics Cascade 650 camera (Roper Scientific, Trenton, NJ) driven by MetaFluor software (Universal Imaging, Downingtown, PA), or an intensified charge-coupled device (CCD) video camera (Photonic Science, Tunbridge Wells, UK) driven by Tardis software (VisiTech International, Sunderland, UK) to capture emitted fluorescence at 510 nm. The sampling rate was 5 ratios/s with the photometry-based system, 1 ratio/s with the Cascade camera, and 1 ratio/2.56 s with the Photonic Science CCD camera. Measurements were performed with a ×20 magnification lens objective microscope (Zeiss, Jena, Germany). [Ca2+]c was calculated by comparing the ratio of the 510-nm signals successively acquired at 340 and 380 nm with a calibration curve based on the equation of Grynkiewicz et al. (19) and established by filling the chamber with an intracellular solution containing 10 μM fura-PE3 free acid and ∼10 mM or <1 nM free Ca2+. A Kd value of 290 nM was used for the fura-PE3-Ca2+ complex (38).
The membrane potential of a single β-cell within an islet was recorded continuously at 37°C using a high-resistance (100–300 MΩ) intracellular microelectrode simultaneously with [Ca2+]c measurement in the whole islet. β-cells were identified on the basis of the typical electrical activity that they display in the presence of a stimulatory concentration of glucose (21).
Insulin secretion measurements.
Insulin secretion was measured simultaneously with [Ca2+]c (experiments shown in Fig. 5) as previously described (18). Briefly, a single islet was placed into a 110-μl temperature-controlled perifusion chamber. The flow rate was 1.8 ml/min, and the effluent fractions were collected in fractions of 30 s just downstream from the islet. Insulin was measured in duplicate in 400-μl aliquots of the effluent fraction. The RIA characteristics using rat insulin (Novo Research Institute, Bagsvaerd, Denmark) as the standard were reported elsewhere (23).
Total RNA was extracted, quantified, and reverse transcribed into cDNA exactly as described previously (24). The sense and antisense primers were chosen in the coding region of mRNA gene sequences, and their specificity was checked via a blast search of the GenBank database. The primers for SERCA1a and SERCA1b were similar (forward primer, 5′-TTCCATCTGCCTGTCCATGTC-3′; reverse primer, 5′-CTGGTTACTTCCTTCTTTCGTCTT-3′) but could be distinguished by the size of the amplicons (see below) (40, 43). The forward primer was common for SERCA2a and SERCA2b: 5′-AAATCTCCTTGCCTGTG-3′. The reverse primers were 5′-GATGGCTTCTGTTCTTG-3′ for SERCA2a and 5′-ATCGCTAAAGTTAGTGTCTGT-3′ for SERCA2b (37). Polymerization reactions were performed using a PerkinElmer 9700 ThermoCycler in a 25-μl reaction volume containing 3 μl of cDNA (20 ng of total RNA equivalents), 80 μM cold 2-deoxynucleotide 5′-triphosphate, 1.25 μCi of [α-32P]dCTP (3,000 Ci/mM), 100 ng of appropriate oligonucleotide primers, GeneAmp Gold PCR buffer, and 1.25 U of AmpliTaq Gold DNA polymerase (PerkinElmer, Foster City, CA). The thermal cycle profile was a 5-min denaturing step at 94°C, followed by amplification cycles [30 s at 94°C, 30 s at 58°C (for SERCA2a and SERCA2b) or 62°C (for SERCA1a and SERCA1b), and 1 min at 72°C each], and a final extension step of 10 min at 72°C. The amplicons were then separated on a 6% polyacrylamide gel in Tris-borate-EDTA buffer. The gel was dried, and [α-32P]dCTP-labeled amplicons were revealed using the Cyclone Storage Phosphor System (Packard, Meriden, CT). The sizes of amplicons were as expected (248 bp for SERCA1a, 206 bp for SERCA1b, 385 bp for SERCA2a, and 209 bp for SERCA2b).
Real-time fluorescence PCR.
These experiments were performed on the same batches of cDNA used for radioactive PCR (see above). Real-time PCR was performed with the iCycler iQ Real Time PCR detection system (Bio-Rad Laboratories, Hercules, CA) using the fluorescent dye SYBR Green I (Molecular Probes, Leiden, The Netherlands) to monitor double-stranded DNA accumulation during the annealing step of each cycle of amplification (excitation/emission ratio, 490/530 nm) (7). Amplifications of SERCA2a, SERCA2b, and TATA-box binding protein (TBP) cDNA were performed in duplicate in a 25-μl reaction volume containing 12.5 μl of iQ SuperMix (Bio-Rad Laboratories), cDNA (2–10 ng of total RNA equivalents) or water, SYBR Green I at 10−5 dilution, 10 nM fluorescein for fluorescence normalization between wells, and 300 nM primers specific for SERCA2a (forward, 5′-GGAACAACCCGCAATACTGG-3′; reverse, 5′-CTTTTCCCCAACCTCAGTCATG-3′), SERCA2b (forward, 5′-AAATCTCCTTGCCTGTG-3′; reverse, 5′-ATCGCTAAAGTTAGTGTCTGT-3′), or TBP (forward, 5′-ACCCTTCACCAATGACTCCTATG-3′; reverse, 5′-ACTTCGTGCCAGAAATGCTGA-3′). The thermal cycle profile was a 3-min denaturing step at 95°C to release DNA polymerase activity, followed by 40 cycles of amplification, each of which comprised a 15-s denaturation step at 95°C, a 45-s annealing step at 60°C, and a 15-s step at 82°C. After amplification, the specificity of PCR products was verified using melting curve analysis (31), and the threshold cycle (Ct) was determined using iCycler iQ software 3.0a (Bio-Rad Laboratories). Under these conditions, PCR efficiencies for amplification of SERCA2a, SERCA2b, and TBP were similar. Because each gene of interest was amplified in parallel with TBP using the same cDNA dilution, ΔCt = Ct gene − Ct TBP was calculated for each sample before averaging. The gene-to-TBP mRNA ratio can be calculated using the formula 2 − ΔCt.
Presentation of results.
Representative results are shown in traces obtained with the indicated number of islets from at least three different cultures. The statistical significance between means was assessed using an unpaired t-test, a χ2-test, or ANOVA, followed by the Newman-Keuls test as appropriate. Differences were considered significant at P < 0.05.
Effects of various glucose concentrations on [Ca2+]c oscillations in islets.
We first determined the optimal conditions in which to analyze mixed [Ca2+]c oscillations. Islets were submitted to stepwise increases in the glucose concentration of the perifusion medium. In the presence of low glucose (≤5 mM), [Ca2+]c was low and stable (Fig. 1). The threshold glucose concentration at which [Ca2+]c oscillations appeared varied between 6 and 8 mM (39%, 72%, and 100% of the islets were active in 6, 7, and 8 mM glucose, respectively; n = 39) and was 6 mM in the experiment shown in Fig. 1. Just above the threshold, [Ca2+]c oscillations almost always showed a mixed pattern (Fig. 1). Therefore, 8 mM glucose was used in subsequent experiments aimed at testing the origin of mixed oscillations. At higher glucose concentrations (10–15 mM) the pattern was rarely mixed, but usually simple, either slow or rapid.
Mixed [Ca2+]c oscillations result from compound bursting that is synchronized between different islet regions.
To assess the relationship between oscillations of membrane potential and [Ca2+]c, we simultaneously measured [Ca2+]c in a whole islet and the membrane potential of a β-cell within the islet using an intracellular microelectrode. As previously reported (15, 33), in 10 mM glucose, simple [Ca2+]c oscillations, either rapid (Fig. 2A) or slow (Fig. 2B), resulted from synchronous membrane potential oscillations. As shown in Fig. 2C, mixed [Ca2+]c oscillations were induced by compound bursting: membrane potential oscillations of variable duration occurred in bursts separated by prolonged silent intervals. In some (3 of 10) islets, rapid [Ca2+]c oscillations coincided with changes in the frequency of action potentials on top of a depolarized plateau (Fig. 2D, see asterisk). These results suggest that the slow ascending phase of each mixed [Ca2+]c oscillation results from a progressive summation of the rapid [Ca2+]c oscillations triggered by synchronous oscillations of the bursting electrical activity.
Oscillations of [Ca2+]c and membrane potential were synchronous in the vast majority of islets, and only a few exceptions to the rule were observed. In 2 of 42 mixed oscillations (10 islets), a rapid [Ca2+]c oscillation occurred in the absence of concomitant membrane potential depolarization (Fig. 3A, see asterisk). In 4 of 42 mixed oscillations, a membrane potential oscillation was recorded in the impaled cell without detection of a significant [Ca2+]c oscillation in the islet (Fig. 3B, see asterisk). Because each mixed oscillation comprises ∼10 shorter oscillations, the incidence of dissociations between membrane potential and [Ca2+]c changes was ∼1%. The rarity of these dissociations implicates that all regions of the islets are well synchronized, even when the pattern is irregular.
We also used digital image analysis to compare glucose-induced [Ca2+]c oscillations in different small regions of the islets. In islets perifused with 8 mM glucose for 30 min, most mixed [Ca2+]c oscillations were perfectly synchronous between all regions (Fig. 4A). However, two types of minor desynchronization were sometimes detected. The first type affected 6.9% of the slow oscillations: [Ca2+]c remained low in one region, whereas it increased and oscillated in another region (data not shown). The second type affected 8.4% of the slow oscillations: [Ca2+]c remained steadily elevated in one region, but it oscillated in another (Fig. 4B, see asterisk). These desynchronizations are thus a rare phenomenon that affects only a few rapid oscillations during some mixed [Ca2+]c oscillations and are restricted to small regions of some islets. Synchronization of the mixed oscillations throughout the islet is thus the rule.
Mixed [Ca2+]c oscillations trigger insulin secretion oscillations.
We next investigated the influence of the mixed [Ca2+]c oscillations on insulin secretion by simultaneous monitoring of [Ca2+]c and insulin secretion in single islets. Because insulin secretion by one islet is undetectable in the presence of 8–10 mM glucose, the experiments had to be performed in 12–15 mM glucose. Only few islets displayed mixed [Ca2+]c oscillations under these conditions. Another constraint is the collection of enough secretion for the assay; the perifusion medium was thus sampled at 30-s intervals, which do not permit the detection of fast oscillations. As shown in Fig. 5, each mixed [Ca2+]c oscillation triggered a synchronous slow oscillation of insulin secretion.
Ablation of SERCA3 does not alter the expression of other SERCA subtypes.
The known influence of SERCA3 on [Ca2+]c oscillations triggered by depolarization with high K+ concentration ([K+]) (1) led us to test the possible impact of the ablation of SERCA3 on mixed [Ca2+]c oscillations.
Islets from SERCA3−/− mice have previously been shown to express a short nonfunctional SERCA3 (1). We verified that ablation of SERCA3 was not compensated by the overexpression of another SERCA isoform. Expression of SERCA1a, SERCA1b, SERCA2a, and SERCA2b mRNA in control tissues (cardiomyocytes and skeletal muscles) and islets of SERCA3+/+ and SERCA3−/− mice was first assessed using radioactive RT-PCR (36 cycles). SERCA1a and SERCA1b mRNA could not be detected in islets from SERCA3+/+ and SERCA3−/− mice, whereas both isoforms were expressed in skeletal muscles and cardiomyocytes (Fig. 6A). SERCA2a was not detected in islets, weakly expressed in skeletal muscle, and strongly expressed in cardiomyocytes (Fig. 6B), and SERCA2b was well expressed in all tissues (Fig. 6C).
The expression of SERCA2a and SERCA2b mRNA from islets of SERCA3+/+ and SERCA3−/− mice and from cardiomyocytes was also assessed using real-time PCR (Fig. 6D). Ablation of SERCA3 did not significantly affect the mRNA expression of both SERCA2 isoforms in islets. These results demonstrate that the absence of SERCA3−/− is not compensated by the overexpression of other SERCA isoforms.
Ablation of SERCA3 alters the pattern of [Ca2+]c oscillations and electrical activity.
As in control islets, the threshold glucose concentration that triggered [Ca2+]c oscillations in islets from SERCA3−/− mice varied between 6 and 8 mM (39%, 92%, and 100% of the islets were active in 6, 7, and 8 mM glucose, respectively, n = 36; not significantly different vs. controls, χ2-test), and was 6 mM in the experiment shown in Fig. 7A. [Ca2+]c oscillations of islets of SERCA3−/− mice were clearly different from those of SERCA3+/+ mice. As previously reported (1), their amplitude was larger because the nadir was close to basal levels and both the ascending and descending phases were more abrupt (note the difference in scale between Figs. 1 and 7A). In addition, an as yet unrecognized feature was observed. Contrary to observations in SERCA3+/+ islets, SERCA3−/− islets never displayed characteristic mixed [Ca2+]c oscillations with several rapid oscillations on top of slow ones. They displayed simple oscillations and, occasionally, sporadic groups of two or three irregular oscillations.
Simultaneous measurement of [Ca2+]c and membrane potential in SERCA3−/− islets perifused with 8 mM glucose showed that [Ca2+]c and membrane potential oscillations were well synchronized (Fig. 7B). As in control islets, however, rare dissociations between [Ca2+]c and membrane potential were observed (data not shown). In 1.5% of all examined oscillations (n = 136 in 12 islets), brief elevation of [Ca2+]c was observed while the plasma membrane was hyperpolarized. In 0.7% of the oscillations, [Ca2+]c did not increase during oscillations of the membrane potential. Importantly, unlike SERCA3+/+ β-cells, SERCA3−/− β-cells never displayed compound bursting. Quantification of electrical activity in both types of islets at 8 mM glucose also showed that the percentage of active phases was higher in SERCA3−/− than in SERCA3+/+ β-cells [51 ± 3% (n = 12) vs. 29 ± 3% (n = 10); P < 0.01], whereas the frequency of membrane potential oscillation was lower in SERCA3−/− than in SERCA3+/+ β-cells [1.2 ± 0.2 oscillations/min (n = 12) vs. 3.2 ± 0.4 oscillations/min (n = 10); P < 0.01]. This clearly indicates that SERCA3 influences the membrane potential.
We used image analysis to compare [Ca2+]c oscillations in different regions of the islets from SERCA3−/− mice. As in control islets, the oscillations were synchronous in all regions of the islets, although their amplitude was sometimes variable; transient desynchronizations were observed only rarely in a few islets (data not shown).
SERCA3 is responsible for mixed [Ca2+]c oscillations.
We previously reported that during a train of [Ca2+]c oscillations imposed by repetitive depolarization with KCl, the ER could be responsible for a progressive summation of [Ca2+]c oscillations (2). In the present study, we have investigated the possibility that SERCA3 is responsible for this summation, thereby contributing to the mixed pattern of [Ca2+]c oscillations.
To mimic the electrical activity occurring spontaneously in control islets stimulated by 8 mM glucose, we applied three series of five cycles of depolarization and repolarization of different durations with 30 and 10 mM [K+], respectively. The ATP-sensitive K+ channel (KATP) opener, diazoxide, was added to the medium to avoid that changes in SERCA activity influence the membrane potential (14, 39). In SERCA3+/+ islets, the three trains of depolarization and repolarization elicited a progressive summation of the rapid [Ca2+]c oscillations, producing a mixed [Ca2+]c pattern (Fig. 8A). The summation was particularly clear during the last train, in which the first depolarizations were of increasing duration as sometimes observed during spontaneous compound bursting (see Fig. 2, C and D). Figure 8E shows the difference in [Ca2+]c between the nadirs (during intervals separating pulses) and baseline level. As expected, the difference was larger during the second train, when the relative pulse duration was longer (Fig. 8E; compare 2 bars labeled A). When control SERCA3+/+ islets were pretreated with thapsigargin, an inhibitor of SERCA (22), summation of [Ca2+]c oscillations was largely abolished: [Ca2+]c returned to nearly basal levels after each rapid oscillation (Fig. 8, B and E; compare bars A and B), demonstrating that the ER is involved in the summation of the [Ca2+]c signal as previously described (2). Application of the same protocol to SERCA3−/− islets elicited a pattern of [Ca2+]c oscillations similar to that of control islets pretreated with thapsigargin (Fig. 8, C and E; compare bars B and C). Pretreatment of SERCA3−/− islets with thapsigargin did not affect their [Ca2+]c response (Fig. 8, D and E; compare bars C and D). These results show that SERCA3 is responsible for a progressive summation of the rapid [Ca2+]c oscillations during the mixed pattern.
The present study demonstrates that mixed [Ca2+]c oscillations elicited by mildly stimulatory glucose concentrations are induced by compound bursting in all β-cells of an islet and that SERCA3 plays a role in the generation of these mixed [Ca2+]c oscillations.
Origin of mixed [Ca2+]c oscillations elicited by glucose.
Whereas simple [Ca2+]c oscillations have been shown to result from synchronous oscillations of the membrane potential (see Refs. 18, 33; confirmed in this study), the origin of mixed [Ca2+]c oscillations is the subject of debate. It has been suggested that they result from the mixture of two different patterns of oscillation in distinct populations of β-cells within the islet (28). However, our combined measurements demonstrate that the mixed pattern is present in all β-cells of an islet and induced by bursts of membrane potential oscillations separated by long, silent intervals (i.e., compound bursting) (9, 21). Moreover, digital image analysis revealed that mixed [Ca2+]c oscillations were present most of the time in all regions of an islet and displayed synchrony between the different regions. In a few islets, [Ca2+]c oscillations were rarely desynchronized, with some regions oscillating while others remained steadily elevated or at the basal level. This finding is in accord with the observation that membrane potential and [Ca2+]c oscillations were occasionally dissociated in combined measurements. However, it is important to emphasize that these rare dissociations do not explain the mixed pattern, which appears to be an inherent property of individual β-cells as also suggested by its occurrence in single β-cells or in small islet cell clusters (25, 30).
Pancreatic β-cells express SERCA2b and SERCA3 only.
The presence of SERCA2b in islets has been established (27, 35, 36), but whether islets also express SERCA2a remains unclear (27, 36). By using specific primers for SERCA2a and SERCA2b in radioactive and real-time PCR experiments, we have confirmed that SERCA2b mRNA is abundantly expressed in islets and have shown that SERCA2a mRNA is absent or expressed at a much lower level than in cardiomyocytes. The presence of SERCA3 was established previously by several groups (12, 27, 36, 42), and, on the basis of immunocytochemistry, we also previously demonstrated that it is restricted to β-cells in islets (1). Importantly, we have demonstrated herein that SERCA3−/− mice did not compensate for the lack of SERCA3 by overexpressing another SERCA isoform.
Role of SERCA3 in the mixed pattern of [Ca2+]c oscillations.
That SERCA3 is restricted to β-cells makes the interpretation of our results easier, because any difference between SERCA3+/+ and SERCA3−/− islets can be attributed specifically to β-cells. SERCA3 plays a major role in the control of mixed [Ca2+]c oscillations, because SERCA3−/− islets never displayed characteristic mixed [Ca2+]c oscillations and compound bursting. Because SERCA3 takes up Ca2+ during each period of Ca2+ influx and shapes depolarization-induced [Ca2+]c oscillations (1), we hypothesized that it might contribute to the pattern of mixed [Ca2+]c oscillations by inducing a progressive summation of rapid [Ca2+]c oscillations. We have shown that this is indeed the case by comparing the [Ca2+]c responses of islets from SERCA3+/+ and SERCA3−/− mice to series of cycles of depolarization and repolarization mimicking compound bursting that occurred during mild glucose stimulation. The summation was observed even when the depolarizing phases were shorter than the repolarizing phases (see beginning of slow [Ca2+]c oscillation in Figs. 2, C and D, and 3A), which reinforces our previous proposal that the summation results from different kinetics of Ca2+ uptake and release by the ER, with the uptake being much faster than the release (14). These characteristics have recently been integrated into a mathematical model explaining the filtering of [Ca2+]c transients by the ER in β-cells (6).
The lack of compound bursting in SERCA3−/− islets could be linked to the decrease in frequency of membrane potential oscillations and the increase in duration of the active phases. SERCA3 could affect the membrane potential by several mechanisms, which would explain why SERCA3 disruption depolarizes β-cells. By refilling the ER with Ca2+, SERCA3 could control a depolarizing store-operated current (5, 8, 14, 39). Alternatively or in addition, SERCA3 could modulate the KATP current (26, 32) by consuming ATP.
Our results therefore indicate that SERCA3 is involved in the generation of mixed [Ca2+]c oscillations via at least two mechanisms: a progressive summation of the Ca2+ signal during fast [Ca2+]c oscillations and induction of compound bursting. Whether the first effect is the cause of the latter remains to be established.
Influence of mixed [Ca2+]c oscillations on insulin secretion.
In vivo studies and ex vivo experiments with the perfused pancreas have established the pulsatility of insulin secretion with a periodicity of 5–10 min (for review, see Ref. 17). The role of the oscillations of electrical activity or [Ca2+]c in individual islets for the generation of in vivo insulin oscillations is unclear. Nevertheless, our combined insulin and [Ca2+]c measurements, demonstrating that mixed [Ca2+]c oscillations can induce synchronous slow oscillations of insulin secretion, suggest that they could have a physiological impact.
In conclusion, we have demonstrated herein for the first time that mixed [Ca2+]c oscillations result from compound bursting synchronized in all β-cells within the islet and induce synchronous oscillations of insulin secretion. Our experiments with SERCA3−/− islets also demonstrate that SERCA3 plays an important role in the mixed pattern of [Ca2+]c oscillations and the control of the membrane potential, particularly regarding the generation of compound bursting. Although SERCA3 seems dispensable in normal glucose homeostasis, its normal functioning might have other important roles in β-cells. The cyclic elevations in [Ca2+] in the ER brought about by SERCA3 could influence synthesis, modifications, and folding of proteins, processes that are known to be affected by the [Ca2+] of the ER (10, 20).
This work was supported by Grant 3.4552.04 from the Fonds de la Recherche Scientifique Médicale (Brussels, Belgium), Grant ARC (00/05-260) from the General Direction of Scientific Research of the French Community of Belgium, and by the Interuniversity Poles of Attraction Programme (P5/17)-the Belgian Science Policy.
We thank Dr. R. Macianskiene and Prof. K. Mubagwa (University of Leuven, Leuven, Belgium) for the preparation of cardiomyocytes, Prof. G. Shull (University of Cincinnati, Cincinnati, OH) for the supply of SERCA3−/− mice, and Dr. Ziad Zeinoun and Prof. Frank Wuytack for advice.
M. C. Beauvois holds a research fellowship from Fonds pour la Formation à la Recherche dans l'Industrie et l'Agriculture (Brussels, Belgium), and J.-C. Jonas is Senior Research Associate and P. Gilon is Research Director of the Fonds National de la Recherche Scientifique, Brussels, Belgium.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society