Cx36 is the predominant connexin isoform expressed by pancreatic β-cells. However, little is known about the role of this protein in the functioning of insulin-secreting cells. To address this question, we searched for a cell line expressing Cx36 and having glucose-induced insulin secretion comparable to that of primary β-cells. By evaluating Cx36 expression in MIN6, βTC3, RIN2A, INS1, and HIT cell lines, which differ in their sensitivity to glucose, we found that wild-type MIN6 cells fit these requirements. Therefore, we stably transfected MIN6 cells with a cDNA coding for a Cx36 antisense sequence to study the role of Cx36 in these cells. Independent clones of MIN6 cells were obtained that had a markedly reduced Cx36 expression. Loss of Cx36 decreased functional gap junctional conductance in these clones. This alteration impaired the synchronization of glucose-induced [Ca2+]i oscillations and insulin secretion in response to glucose, to secretagogues that increase [cAMP]i, and to depolarizing conditions. These data provide the first evidence that Cx36-made channels 1) mediate functional coupling in MIN6 cells, 2) provide for synchronous [Ca2+]i oscillations, and 3) are necessary for proper insulin secretion in response to metabolizable and nonmetabolizable secretagogues.

The well coordinated functioning of cells within tissues is ensured by multiple mechanisms, including the direct exchange of ions, second messengers, and other metabolites through connexin channels, clustered at gap junctions (1,2). These channels connect almost all types of vertebrate cells, including those of pancreatic islets (3). Previous studies suggest that connexin-mediated communication between β-cells is required for the control of insulin secretion. Thus, glucose stimulation of single β-cells is reduced compared with that of cell clusters (4), pharmacological blockade of gap junction channels markedly decreases insulin release (5), and transgenic mice whose β-cells express Cx32 show altered insulin secretion in response to glucose (6). While these findings provide evidence that connexin-dependent communication contributes to the control of insulin release, they do not specifically address the function of native connexins expressed by pancreatic islet cells.

We recently found that Cx36 is the predominant isoform expressed by β-cells (7). As yet, however, the specific role of this somewhat unusual connexin isoform (8,9) and, in particular, its possible contribution to insulin secretion have not been investigated. Moreover, no data are available to show how connexin-mediated signaling may influence β-cell function even though gap junctions have been implicated in the synchronization of [Ca2+]i oscillations in β-cells (1013).

To address these questions, we screened MIN6 cells, which, unlike some other cell lines, retain glucose-induced insulin secretion and express Cx36 like primary β-cells (1419). We decreased Cx36 expression via the stable transfection of an antisense construct. Here, we show that the loss of Cx36 resulted in impaired electrical coupling, desynchronization of [Ca2+]i oscillations, and altered insulin release of MIN6 cells. These results support the hypothesis that Cx36 plays a critical role in glucose-induced insulin secretion.

Cell lines and tissues.

RIN2A (14), INS1 (15), and HIT cells (16) were cultured in RPMI-1640 medium, whereas βTC3 (17) and MIN6 cells (18) were cultured in Dulbecco’s modified Eagle’s medium as indicated in the original reports. Control tissues were obtained from OFA rats (olfactory bulb) or C57BL/6 mice (liver and heart), which were frozen in liquid nitrogen and stored at −80°C until use.

RNA analysis.

Samples of total RNA were extracted, reverse-transcribed, and amplified as previously described (7). To detect native Cx36, we amplified either a 980-bp or a 559-bp fragment, using the oligonucleotide pairs 5′-CACAGCGATGGGGGAATGGA-3′/5′-TGCCCTTTCACACATAGGCA-3′ and 5′-GAGCCCAGGCCAAGAGGAAGTC-3′/5′-GGCATGCTGAAGGGGGAGAAAT-3′as sense and antisense primers, respectively. For cyclophilin, we amplified a 387-bp fragment, using the oligonucleotides 5′-GGTCAACCCCACCGTGTTCT-3′/5′-GCCATCCAGCCACTCAGTCT-3′. For the Cx36 antisense construct, we amplified a 554-bp fragment, using the oligonucleotides 5′- GTCTCCTTACTGGTGGTCTCTGTG-3′/5′-TAGAAGGCACAGTCGAGG-3′. Northern blotting analysis was performed as previously described (7).

Western blot and immunofluorescence.

Cell and tissue extracts were prepared as reported (20). Protein content was measured using a detergent compatible protein assay kit (Bio-Rad Laboratories). Samples of membrane preparations were equally loaded (10 μg) in each lane and fractionated by electrophoresis in a 12% polyacrylamide gel. Equal loading was further verified by densitometric analysis after Coomassie-blue staining of the gels, which showed comparable amounts of total proteins in each lane (βTC3 19 ± 2.3, MIN6 17 ± 2, INS1 18 ± 1, RIN2A 24 ± 0.6, and HIT 23 ± 1.9 [arbitrary units ± SD]; n = 3). Electrophoresed samples were transferred onto polyvinylidine flouride membranes (Immobilon-P; Millipore) using a constant current of 400 mA for 2 h, and in the presence of 0.01% SDS and 20% methanol. Membranes were saturated as described (20) and then incubated over night at 4°C with an affinity-purified antiserum against either Cx36 (7), diluted 1:800, or smooth muscle α-actin (21), diluted 1:1,000. After a 60-min incubation at room temperature with goat anti-rabbit Igs, conjugated to horseradish peroxidase (Bio-Rad Laboratories), diluted 1:3,000, membranes were developed by enhanced chemiluminescence according to the manufacturer’s instructions (Amersham Pharmacia Biotech). Densitometric analysis of the gels and films was performed with the Image Quant software (Molecular Dynamics).

For immunofluorescence, cells were incubated for 2 h at room temperature in the presence of either a polyclonal rabbit antiserum against residues 289–303 of rat Cx36 (7), diluted 1:200, or residues 154–168 of mouse Cx36 (22), diluted 1:1,000. After exposure to fluorescein-conjugated antibodies against rabbit Igs, diluted 1:500, cells were counterstained with Evans blue, coverslipped, and photographed using a fluorescence microscope (Axiophot; Carl Zeiss).

Insulin secretion and content.

Cells were plated at the density of 5 × 104/cm2 within 6- or 12-well plates and tested 3 days later. After a preincubation for 30 min at 37°C in a Krebs-Ringer bicarbonate buffer containing no glucose, cells were incubated for 30 min in the same medium supplemented with the indicated concentration of glucose, with or without other secretagogues. The medium was collected, centrifuged 10 min at 3,000g, and the supernatant was frozen at −20°C for insulin assay. Cultures were then extracted for 24 h at 4°C in acid-ethanol and the extracts also frozen for determination of insulin and protein content. Insulin was measured by a radioimmunoassay (5,6). Protein content was measured by the DC assay kit (Bio-Rad Laboratories). Values of secreted insulin were normalized to basal release (in the presence of 2.8 mmol/l glucose) or protein content. Insulin extracts were expressed as percentage of protein content.

To measure glucose transport, cells were incubated for 30 min at 37°C in a Krebs-Ringer bicarbonate buffer containing 700 μmol/l 6-N-(7-nitrobenzen-2-oxa-1,3-diazol-4-yl)amino-6-deoxiglucose (6-NBDG) (Molecular Probes) (23) and observed by fluorescence microscopy.

Transfection of Cx36 antisense construct.

The sequence coding for rat Cx36 (−3/+973 bases) was inserted in an antisense orientation within plasmid pcDNA 3.1(+), which contains the cytomegalovirus early promoter region and a neomycin-resistance sequence. Subconfluent cultures of MIN6 cells were exposed for 24 h to 10 μg/ml Lipofectin reagent (Gibco BRL) mixed with 10 μg/ml of either the plasmid or the void pcDNA3.1(+). Stable transfectants were selected in presence of 350 μg/ml Geneticin G-418 sulfate (Gibco BRL). For cloning, selected cells were plated at a density of 1 per well in 96-well plates.

Electrophysiological and dye transfer measurements.

To evaluate electrical coupling, subconfluent cultures of MIN6 cells were continuously superfused with an external solution that contained 115 mmol/l NaCl, 3 mmol/l CaCl2, 5 mmol/l KCl, 2 mmol/l MgCl2, 10 mmol/l HEPES, and 11.1 mmol/l glucose with or without 15 mmol/l tetraethylammonium (TEA), pH 7.2. In the presence of glucose alone, MIN6 cells did not reliably exhibit robust oscillatory behavior, as reported for βTC3 (24). The addition of TEA favored the production of electrical spikes. Hence, we added the drug to improve the signal-to-noise ratio of both the electrical and calcium recordings. To analyze membrane potential under current-clamp conditions, a perforated patch was carried out using pipettes filled with a solution containing 28.4 mmol/l K2SO4, 63.7 mmol/l KCl, 11.8 mmol/l NaCl, 1 mmol/l MgCl2, 20.8 mmol/l HEPES, 0.5 mmol/l EGTA, 0.3–0.5 mg/ml amphotericin B (pH 7.2), and an Axopatch-200 B patch-clamp amplifier (Axon Instruments, Union City, CA). The voltage clamp mode of the amplifier was used to evaluate gap junction currents and current-voltage relations. Seal resistances ranged from 1 to 5 GΩ. Gap junctional conductance (Gc) was calculated as previously reported (2527).

To evaluate dye coupling, individual cells were microinjected as described (20) with a glass microelectrode filled with either 4% (wt/vol) Lucifer Yellow CH (448 Da, net charge −2; Sigma), 1% propidium iodide (414 Da, net charge +2; Sigma), or a mixture of 5% neurobiotin (N-[2-aminoethyl] biotinamide hydrochloride) (287 Da, net charge +1; Vector Laboratories) and 0.4% rhodamine 3-isothiocyanate dextran 10S (Sigma). After injection of neurobiotin, cells were fixed in 4% paraformaldehyde in phosphate buffer (0.1 mol/l, pH 7.4) for 20 min, exposed to 0.25% Triton X-100 in PBS for 30 min, and incubated with fluorescein-conjugated streptavidin (Jackson ImmunoResearch Laboratories), diluted 1:400 for 60 min. Mn2+ (55 Da, net charge +1) diffusion measurements was assessed as reported (28). Results were expressed as mean ± SE of the number of cells communicating with the microinjected cell.

[Ca2+]i measurements.

Clusters of MIN6 cells were loaded with 5 μmol/l Fluo-3/AM and 2 μmol/l pluronic acid (Molecular Probes, Eugene, OR) and incubated at 37°C for 30 min, in an O2/CO2 incubator. Cells were then incubated for 15–20 min in a recording solution which contained 11.1 mmol/l glucose with or without 15 mmol/l TEA. The concentration of cytoplasmic-free Ca2+ ([Ca2+]i) was assessed with a laser scanning confocal microscope system (Olympus FluoView, Tokyo, Japan) using an excitation wavelength of 488 nm and an emission wavelength of 526 nm. Fluorescence due to Fluo-3 excitation was simultaneously measured in 3–5 cells per cluster at 32–35°C. To estimate synchrony among neighboring cells, data were expressed in units of relative fluorescence and plotted as a function of time. Synchrony was defined as the percent of cells within a given cluster displaying simultaneous [Ca2+]i transients.

Data analysis.

Data were expressed as mean ± SE. Differences between means were assessed by Student’s t test and considered significant when P < 0.05.

Different insulin-producing cell lines express distinct levels of Cx36.

RT-PCR amplification of total RNA failed to detect the transcripts for Cx43 and Cx45 in RIN2A, INS1, HIT, βTC3, or MIN6 lines under conditions that easily amplified these RNAs in control heart tissue (Fig. 1A). In contrast, Cx36 transcript was found in all the lines investigated, as well as in olfactory bulb, which was used as a positive control (Fig. 1B). Further RT-PCR screening of MIN6 cells failed to detect Cx26, Cx32, or Cx30 transcripts (not shown).

Using Northern blot, a 2.9-Kb transcript was detected in the five cell lines and in the olfactory bulb, but not in the heart, which were used as positive and negative controls, respectively (Fig. 1C). Relative to ribosomal RNA, the levels of Cx36 transcript varied, being most abundant in βTC3 and MIN6 cells, but barely detectable in RIN2A and HIT cells (Fig. 1C).

Immunoblots of cell extracts carried out with a polyclonal antiserum against Cx36 revealed bands of ∼36 KDa in all the lines tested and the olfactory bulb, but not in the heart (Fig. 2A). Levels of Cx36 were measured by densitometry and evaluated relative to those of α-actin in five independent experiments to control for the small unavoidable variations in immunolabelled signals between different experiments. The highest level of Cx36 expression was seen in MIN6 cells (Fig. 2A).

Immunofluorescence labeling of cultures with antibodies against two different epitopes of Cx36 revealed the protein at sites of membrane contact between MIN6 cells and, to a lesser extent, between βTC3 cells (Fig. 2B). Staining was barely detectable between INS1 cells (Fig. 2B) and was absent in cultures of RIN2A and HIT cells (not shown).

MIN6 cells retain responsiveness to glucose.

Cells were incubated for 30 min in the presence of a basal (2.8 mmol/l) and a maximal stimulatory concentration of glucose (22.4 mmol/l). We found that MIN6 cells showed a 4.5-fold increase in insulin secretion in 22.4 mmol/l glucose (Fig. 3). Under the same conditions, the RIN2A, INS1, βTC3, and HIT clones we used failed to respond to glucose (Fig. 3).

MIN6 cell membranes lack Cx36 after transfection of an antisense construct.

MIN6 cells were stably transfected with a Cx36 cDNA in the antisense orientation, selected for acquired neomycin resistance, and four independent lines (A–D) were cloned. Using RT-PCR amplification, we found that the endogenous Cx36 transcript was expressed in both wild-type and transfected cells (Fig. 4A). In contrast, the Cx36 antisense construct was only detected in the four antisense-transfected clones (Fig. 4B).

By Western blot, comparable levels of immunolabelled Cx36 were detected in wild-type and neomycin-resistant cells (Fig. 5A). In contrast, Cx36 expression was decreased by >85% (relative to α-actin levels) in the four antisense-transfected clones (Fig. 5A).

Immunofluorescence labeling confirmed this marked decrease (Fig. 5B). Thus, whereas Cx36 was abundantly distributed at cell interfaces of wild-type and neomycin-resistant MIN6 cells, it was undetectable in the four antisense-transfected clones (Fig. 5B).

Loss of Cx36 decreases junctional coupling and synchronization of [Ca2+]i oscillations.

By patch-clamping individual MIN6 cells in the presence of 11 mmol/l glucose and 15 mmol/l TEA, we observed prolonged action potentials with amplitudes >50 mV (Fig. 6). Using current-clamp to record electrical activity, the membrane potential and electrical pattern of wild-type and antisense-transfected cells were found to be similar. Upon voltage clamping the cells to −65 mV, spontaneous inward junctional currents were seen in the wild-type MIN6 cells (Fig. 6). These currents, which were not seen in single MIN6 cells (not shown), due to cell-cell coupling, were infrequent and small in the Cx36-antisense clones (Fig. 6). Once the voltage clamp was released, cells returned to their typical firing pattern (Fig. 6), confirming that electrical activity was not altered by the voltage-clamp protocol. By measuring the amplitude of the junctional current spikes, we determined that junctional conductance was significantly smaller (P < 0.05) in antisense clones A and B (0.15 ± 0.04, NS, n = 16) than in wild-type MIN6 cells (0.71 ± 0.16, NS, n = 12). After microinjection of a tracer, only one wild-type MIN6 cell was seen to communicate with the injected cell, irrespective of the type of the tracer used (Lucifer Yellow 1 ± 0.3, n = 35; propidium iodide 1 ± 0.3, n = 20; neurobiotin 0 ± 0.1, n = 16; and Mn2+ 1 ± 0.3, n = 24).

To monitor [Ca2+]i oscillations, clusters of MIN6 cells were loaded with Fluo-3 and fluorescence changes were recorded (Fig. 7A). We observed [Ca2+]i oscillations in the presence of 11.1 mmol/l glucose (not shown) that were markedly enhanced in amplitude by exposure to 15 mmol/l TEA. Under the latter conditions, most MIN6 cells exhibited [Ca2+]i oscillations (Fig. 7B) having both fast and slow components, with the faster signals occurring at a frequency of ∼0.8 min−1. [Ca2+]i oscillations were well synchronized in cells of wild-type clusters, but showed little or no synchrony in clusters of the antisense-transfected clones (Fig. 7B). Quantitative analysis showed that, in the latter clones, fractional synchrony (0.63 ± 0.07, clones A and B pooled, n = 24) was reduced (P < 0.05) compared with the control value (0.95 ± 0.25, n = 17). Control [Ca2+]i spikes were drastically reduced by the addition of either 200 μmol/l diazoxide or 10 μmol/l nifedipine (not shown), indicating that they were mostly accounted for by Ca2+ influx through voltage-gated channels.

MIN6 cells lacking Cx36 show impaired insulin secretion.

Wild-type MIN6 cells showed a progressive increase in insulin secretion as a function of glucose concentration (Fig. 8A). Thus, their insulin output was 183 ± 18 and 612 ± 78 ng insulin/mg protein in the presence of 2.8 and 22.4 mmol/l glucose, respectively (n = 21 dishes from seven independent experiments). When these cells were exposed to 100 μmol/l 3-isobutyl-1-methylxanthine (IBMX) plus 1 μmol/l forskolin (FSK) or 20 mmol/l KCl in the presence of 2.8 mmol/l glucose, their insulin secretion was further stimulated up to 946 ± 93 ng insulin/mg protein (n = 27) and 1,344 ± 69 ng insulin/mg protein (n = 9), respectively (Fig. 8A). Wild-type cells displayed maximal stimulation in the presence of 11 mmol/l glucose plus 15 mmol/l TEA (1,633 ± 142 ng insulin/mg protein, n = 10 dishes from four independent experiments). Neomycin-transfected cells showed a comparable secretion response (Fig. 8A). In contrast, all of the four antisense-transfected clones failed to significantly increase insulin secretion when exposed to a stimulatory concentration of glucose (Fig. 8A). In the presence of IBMX and FSK, these clones increased their insulin release, but this response (266 ± 27 and 358 ± 31 ng insulin/mg protein, for clones B and D, respectively; n = 18 dishes from six independent experiments) was significantly (P < 0.001) lower than that of control MIN6 cells (Fig. 8A). When exposed to either KCl or 11.2 mmol/l glucose plus 15 mmol/l TEA, the antisense clones were further stimulated to release insulin. Still, this increase was significantly (P < 0.05) lower than in control cells (KCl 444 ± 61, 465 ± 71, 629 ± 72, and 895 ± 66 ng insulin/mg protein, for clones A, B, C, and D, respectively, n = 9; TEA 507 ± 119, 414 ± 49, 934 ± 158, and 559 ± 73 ng insulin/mg protein, n = 10).

After acid ethanol extraction, we found that the insulin content of the antisense-transfected clones (0.6 ± 0.1, 1.3 ± 0.1, 0.8 ± 0.1, and 1.2 ± 0.1% of protein content in clones A, B, C, and D, respectively; n = 45 from three independent experiments) was in the range of that measured in wild-type (1.2 ± 0.1% of protein content; n = 90 dishes from six independent experiments) and neomycin-resistant MIN6 cells (0.7 ± 0.1% of protein content; n = 90 dishes from six independent experiments) (Fig. 8B).

Loss of glucose-induced insulin secretion in the antisense-transfected clones did not correlate with a loss of glucokinase, which was detected by immunostaining in all clones (not shown). In contrast, we found that the incorporation of the nonhydrolyzable glucose analog 6-NBDG was extensive in wild-type and neomycin-resistant cells, but was variably reduced in the antisense-transfected clones (two independent experiments). Thus, no incorporation of 6-NBDG was found in clones A and C, and little labeling by the analog was observed in clones B and D (data not shown).

The high degree of homology and the restricted distribution of Cx36, which are conserved among different animal species (79,29), suggest that this protein may play an essential role in the function of pancreatic β-cells. To test this idea, we first investigated its pattern of expression in different lines of insulin-producing cells. We found that all the lines we studied expressed Cx36 and that the protein was detectable at the cell membranes of only those cell lines (βTC3, MIN6, and INS1) that have been reported to retain some ability to increase insulin release when challenged by high glucose concentrations (30,31). At the cell passages we used only MIN6 cells that expressed the largest amounts of Cx36 responded to sugar stimulation, suggesting that the level of Cx36 at cell-to-cell contacts is somehow related to glucose-induced insulin secretion. This relationship is certainly not linear and/or the same for all types of cells, since the clone of βTC3 cells we used, which expressed slightly less Cx36 than MIN6 cells, failed to increase insulin release in response to glucose. These data are in full agreement with the multifactorial regulation of the glucose-induced insulin secretion pathway, in which the connexin-dependent signaling is but one relevant factor (6,3133). Our data indicate that Cx36 should be added to the list of proteins that differentiate glucose-sensitive from glucose-insensitive cells, since loss of this protein and/or the resulting blockade of the cell-to-cell communication it permits (5) are common defects of cell lines that feature impaired insulin release.

This observation prompted us to study whether Cx36 is causally implicated in the functioning of insulin-secreting cells. Unfortunately, the acute inhibition of junctional communication with pharmacological agents cannot yet be achieved in a specific manner (5,11,19,34,35). However, taking into account that Cx36 is expressed in limited amounts (7) and that connexins have a rather short half life (36), we considered whether it was feasible to decrease connexin expression using an antisense strategy (37,38). By transfection of an antisense Cx36 construct, we found that the reduced expression of the protein markedly inhibited the electrical coupling of MIN6 cells that, under control conditions, have a junctional conductance similar to that reported in mouse islets (27,39). We did not evaluate the eventual parallel decrease in the permeability of junctional channels of the antisense-transfected clones, since the native Cx36 channels permit only a limited diffusion of a variety of gap junction permeant tracers (40) between wild-type MIN6 cells. This finding is in agreement with the low conductance and open probability reported for Cx36 channels (22,41).

Previous studies have suggested that junctional coupling may provide the basis for coordinating the glucose-induced oscillatory increase of [Ca2+]i in β-cells (1013). As yet, however, the lack of specific inhibitors (11,19) has not permitted to experimentally test this hypothesis. Our data clearly show that a marked decrease in the Cx36 level was associated with profound alterations in the synchronization of [Ca2+]i oscillations in MIN6 cells. Since both wild-type and antisense-transfected cells retained the ability to increase [Ca2+]i in response to secretagogues, these data provide the first direct evidence for an essential role of gap junction-mediated coupling in controlling cell-to-cell synchronization of [Ca2+]i changes. The electrical studies show that this synchronization was due to ion current flow through gap junctions from the cytoplasm of one MIN6 cell to that of its coupled neighbor. It remains to be assessed whether the same mechanism applies to other types of insulin-producing cell lines (e.g., RIN2A and HIT), which we did not investigate with regard to [Ca2+]i oscillations. Furthermore, the presence of a Cx36-dependent synchronization of Ca2+ oscillations between MIN6 cells does not exclude the possibility that, under certain conditions, extracellular messengers may also participate to this control (19,34,42).

In view of the evidence implicating that gap junction channels play a role in the regulation of β-cell secretion (3,6,43), we monitored insulin secretion of the antisense-transfected MIN6 cells. The data show that, compared with controls, cells lacking Cx36 failed to respond to an elevation of glucose and had much reduced responses to nonmetabolizable secretagogues known to either increase [cAMP]i or depolarize the cells. These major secretory alterations could not be attributed to the transfection procedure itself, inasmuch as the secretory ability of neomycin-resistant and wild-type cells was comparable. Furthermore, the glucose unresponsiveness could not be accounted for by a loss of glucokinase, which was expressed in all antisense-transfected clones, nor by an impairment in insulin biosynthesis or storage, since the hormone content of these clones was similar to that of control cells. We found that the ability to incorporate glucose was markedly reduced in some but not all of the antisense-transfected clones, possibly in keeping with the well-documented loss of the GLUT2 transporter, which is a usual event after multiple cell passage (4446). However, in view of both the highly efficient sugar transport permitted by GLUT2 (47) and the persistence of glucose-induced [Ca2+]i increase in clones (e.g., clone A) that failed to incorporate a detectable amount of a glucose analog, it is unlikely that a partial decrease in glucose uptake could account for the almost complete loss of glucose-induced insulin release. Rather, since reduced glucose-induced secretion was also observed in clones B and D, which more easily incorporated glucose, the data imply that Cx36 contributes to control glucose-induced insulin secretion by acting at a step independent of glucose uptake. Additional experiments are required to elucidate the mechanism involved in reducing secretion. Most likely, this mechanism does not specifically affect the glucose-dependent pathway, inasmuch as clones lacking Cx36 also showed a reduced response to secretagogues that activate stimulus-secretion coupling pathways independently of glucose metabolism (48). The data are compatible with a Cx36-dependent regulation affecting proximal as well as distal steps of the secretory pathway.

In summary, we have shown that the loss of Cx36 affects the electrical coupling, synchronization of [Ca2+]i oscillations, and insulin secretion evoked by a variety of secretagogues that activate different intracellular pathways. The data provide evidence that Cx36 contributes to regulate some still-unknown steps of the stimulus-secretion coupling mechanism of MIN6 cells. In view of the numerous characteristics that MIN6 cells share with primary β-cells, the data also provide the first insight that Cx36 is likely to be important for the physiological secretion of primary islet cells in vivo.

FIG. 1.

Cx36 is the sole connexin detected in insulin-producing cell lines. A: RT-PCR amplification identified the transcript for Cx43 and Cx45 in heart but not in the insulin-secreting cell lines βTC3, MIN6, INS1, RIN2A, and HIT. B: When the RT-PCR was run with primers for Cx36, the connexin transcript was detected in olfactory bulb and in all the cell lines we investigated, but not in liver. Amplification of the same samples for cyclophilin was used to control the loading of each lane. Left lanes: size standards; right lanes: blank control (H2O). C: Northern blot analysis confirmed the expression of a Cx36 transcript in all the lines tested. However, the levels of this expression differed in the five lines. RNA from olfactory bulb and heart were used as positive and negative control, respectively. The methylene blue staining of ribosomal RNAs indicates a comparable loading of all samples.

FIG. 1.

Cx36 is the sole connexin detected in insulin-producing cell lines. A: RT-PCR amplification identified the transcript for Cx43 and Cx45 in heart but not in the insulin-secreting cell lines βTC3, MIN6, INS1, RIN2A, and HIT. B: When the RT-PCR was run with primers for Cx36, the connexin transcript was detected in olfactory bulb and in all the cell lines we investigated, but not in liver. Amplification of the same samples for cyclophilin was used to control the loading of each lane. Left lanes: size standards; right lanes: blank control (H2O). C: Northern blot analysis confirmed the expression of a Cx36 transcript in all the lines tested. However, the levels of this expression differed in the five lines. RNA from olfactory bulb and heart were used as positive and negative control, respectively. The methylene blue staining of ribosomal RNAs indicates a comparable loading of all samples.

Close modal
FIG. 2.

Different insulin-producing cell lines express distinct levels of Cx36 protein. A: A few bands of ∼36 KDa were immunodetected in membrane extracts of the five lines tested, as well as in the olfactory bulb (upper panel). Loading of each lane was controlled by immunolabeling of α-smooth muscle actin (middle panel) or Coomassie blue staining (lower panel). The histogram shows the mean ± SE ratio of the Cx36 and α-actin signals, as evaluated by densitometric analysis of five independent experiments. The highest ratio (MIN6 cells) was given the value 100%. B: Immunostaining of monolayers with antibodies against Cx36 resulted in abundant fluorescent labeling of many MIN6 cell interfaces. In contrast, the same antibodies labeled weakly βTC3 cultures and barely INS1 cultures. Bar: 10 μm.

FIG. 2.

Different insulin-producing cell lines express distinct levels of Cx36 protein. A: A few bands of ∼36 KDa were immunodetected in membrane extracts of the five lines tested, as well as in the olfactory bulb (upper panel). Loading of each lane was controlled by immunolabeling of α-smooth muscle actin (middle panel) or Coomassie blue staining (lower panel). The histogram shows the mean ± SE ratio of the Cx36 and α-actin signals, as evaluated by densitometric analysis of five independent experiments. The highest ratio (MIN6 cells) was given the value 100%. B: Immunostaining of monolayers with antibodies against Cx36 resulted in abundant fluorescent labeling of many MIN6 cell interfaces. In contrast, the same antibodies labeled weakly βTC3 cultures and barely INS1 cultures. Bar: 10 μm.

Close modal
FIG. 3.

The five lines tested have a different ability to secrete insulin in response to glucose. The βTC3, INS1, RIN2A, and HIT cells used in our study did not respond to a stimulatory concentration of glucose (22.4 mmol/l). In contrast, MIN6 cells showed a 4.5-fold increase in insulin secretion. Columns represent mean ± SE of 12 culture dishes from three independent experiments.

FIG. 3.

The five lines tested have a different ability to secrete insulin in response to glucose. The βTC3, INS1, RIN2A, and HIT cells used in our study did not respond to a stimulatory concentration of glucose (22.4 mmol/l). In contrast, MIN6 cells showed a 4.5-fold increase in insulin secretion. Columns represent mean ± SE of 12 culture dishes from three independent experiments.

Close modal
FIG. 4.

Antisense-transfected clones of MIN6 cells express both the endogenous Cx36 transcript and the antisense construct. A: The endogenous Cx36 transcript was detected in wild-type MIN6 (WT), cells transfected for neomycin resistance (neo), and four clones (A–D) selected after antisense transfection. B: Using specific primers, the Cx36 antisense construct was detected only in the antisense-transfected clones. The plasmid used for transfection (pl) served as both negative (for Cx36) and positive control (for the antisense construct). Another negative control was liver. Lanes labeled “-” show samples that were not reverse-transcribed before PCR processing. Left lanes: blank control (H2O); right lanes: size standards.

FIG. 4.

Antisense-transfected clones of MIN6 cells express both the endogenous Cx36 transcript and the antisense construct. A: The endogenous Cx36 transcript was detected in wild-type MIN6 (WT), cells transfected for neomycin resistance (neo), and four clones (A–D) selected after antisense transfection. B: Using specific primers, the Cx36 antisense construct was detected only in the antisense-transfected clones. The plasmid used for transfection (pl) served as both negative (for Cx36) and positive control (for the antisense construct). Another negative control was liver. Lanes labeled “-” show samples that were not reverse-transcribed before PCR processing. Left lanes: blank control (H2O); right lanes: size standards.

Close modal
FIG. 5.

Antisense-transfected clones of MIN6 cells lack Cx36 at the cell membrane. A: A few bands of ∼36 KDa were immunolabelled in membrane proteins (upper panel) of either wild-type (WT) or neomycin-resistant cells (neo). Under the same conditions, lanes loaded with proteins of antisense-transfected cells showed weak (clones B and D) or no signal (clones A and C). Protein loading was evaluated by immunolabeling of α-smooth muscle actin (middle panel) and Coomassie blue staining (lower panel). The columns show the ratio of the Cx36 and α-actin signals. The ratio of wild-type cells was given the value 100%. B: By immunofluorescence labeling, bright fluorescent spots located Cx36 at interfaces of control MIN6 cells (WT). In contrast, no labeling was observed in cells of clone C, which is representative of all the antisense clones. Bar: 10 μm.

FIG. 5.

Antisense-transfected clones of MIN6 cells lack Cx36 at the cell membrane. A: A few bands of ∼36 KDa were immunolabelled in membrane proteins (upper panel) of either wild-type (WT) or neomycin-resistant cells (neo). Under the same conditions, lanes loaded with proteins of antisense-transfected cells showed weak (clones B and D) or no signal (clones A and C). Protein loading was evaluated by immunolabeling of α-smooth muscle actin (middle panel) and Coomassie blue staining (lower panel). The columns show the ratio of the Cx36 and α-actin signals. The ratio of wild-type cells was given the value 100%. B: By immunofluorescence labeling, bright fluorescent spots located Cx36 at interfaces of control MIN6 cells (WT). In contrast, no labeling was observed in cells of clone C, which is representative of all the antisense clones. Bar: 10 μm.

Close modal
FIG. 6.

Clones that had lost Cx36 show reduced junctional conductance. Both wild-type and clone A cells showed action potentials, when measured in current clamp mode (C-clamp). When cells potential was clamped at −65 mV (V-clamp), inward currents were seen in wild-type cells, indicating that these cells were electrically coupled. In contrast, these currents were minimal in the antisense-transfected clone A, indicating a marked decrease in coupling. After return to current clamp mode (C-clamp), all cells showed normal electrical activity.

FIG. 6.

Clones that had lost Cx36 show reduced junctional conductance. Both wild-type and clone A cells showed action potentials, when measured in current clamp mode (C-clamp). When cells potential was clamped at −65 mV (V-clamp), inward currents were seen in wild-type cells, indicating that these cells were electrically coupled. In contrast, these currents were minimal in the antisense-transfected clone A, indicating a marked decrease in coupling. After return to current clamp mode (C-clamp), all cells showed normal electrical activity.

Close modal
FIG. 7.

Loss of Cx36 and coupling reduced the synchronization of [Ca2+]i oscillations. A: Five cells were monitored in parallel within clusters of wild-type and clone A cells loaded with Fluo-3. B: In the presence of 11 mmol/l glucose plus 15 mmol/l TEA, changes in the levels of [Ca2+]i were simultaneously induced in wild-type cells (different colors indicate recordings from distinct cells). In contrast, cells of clone A exhibited oscillations that were markedly asynchronous.

FIG. 7.

Loss of Cx36 and coupling reduced the synchronization of [Ca2+]i oscillations. A: Five cells were monitored in parallel within clusters of wild-type and clone A cells loaded with Fluo-3. B: In the presence of 11 mmol/l glucose plus 15 mmol/l TEA, changes in the levels of [Ca2+]i were simultaneously induced in wild-type cells (different colors indicate recordings from distinct cells). In contrast, cells of clone A exhibited oscillations that were markedly asynchronous.

Close modal
FIG. 8.

Antisense-transfected clones of MIN6 cells lacking Cx36 show altered insulin secretion. A: Cells were incubated in the presence of different glucose concentrations (0–22.4 mmol/l) ± 100 μmol/l IBMX and 1 μmol/l FSK, 20 mmol/l KCl, or 15 mmol/l TEA. Wild-type cells (▪) and cells transfected with a neomycin-resistance gene () increased insulin secretion with glucose concentration and were similarly potentiated by IBMX plus FSK, KCl, and TEA. In contrast, the four antisense-transfected clones (□, , [cjs2106], and ) did not respond to glucose stimulation and, when compared with controls, showed a reduced secretory response to both IBMX plus FSK, KCl, and TEA. B: Insulin content of the four antisense-transfected clones was in the range of that observed in wild-type and neomycin-transfected cells.

FIG. 8.

Antisense-transfected clones of MIN6 cells lacking Cx36 show altered insulin secretion. A: Cells were incubated in the presence of different glucose concentrations (0–22.4 mmol/l) ± 100 μmol/l IBMX and 1 μmol/l FSK, 20 mmol/l KCl, or 15 mmol/l TEA. Wild-type cells (▪) and cells transfected with a neomycin-resistance gene () increased insulin secretion with glucose concentration and were similarly potentiated by IBMX plus FSK, KCl, and TEA. In contrast, the four antisense-transfected clones (□, , [cjs2106], and ) did not respond to glucose stimulation and, when compared with controls, showed a reduced secretory response to both IBMX plus FSK, KCl, and TEA. B: Insulin content of the four antisense-transfected clones was in the range of that observed in wild-type and neomycin-transfected cells.

Close modal

The Meda team was supported by grants from the Swiss National Foundation (31-67788.02), the Juvenile Diabetes Research Foundation International (1-2001-622), the European Union (QLRT-2001-01777 and QLGI-1999-00516), and the Fondation Romande pour la Recherche sur le Diabète and the National Institute of Health (1RO1 DK-63443-01). L.S.S. was supported by the National Institute of Health (DK-46409).

We thank J. Bauquis, A. Charollais, D. Caille, and P. Severi-De Marco for excellent technical assistance; D.F. Condorelli for providing the antisense-transfected MIN6 cells; G. Gabbiani and B. Teubner for the gift of the anti-actin and anti-Cx36 sera; N. Demaurex and A. Maturana for help in the Mn2+ diffusion experiments; and D. Bosco for helpful suggestions.

1
Bruzzone R, White TW, Paul DL: Connections with connexins: the molecular basis of direct intercellular signaling.
Eur J Biochem
238
:
1
–27,
1996
2
Richard G: Connexins: a connection with the skin.
Exp Dermatol
9
:
77
–96,
2000
3
Meda P: The role of gap junction membrane channels in secretion and hormonal action.
J Bioenerg Biomembr
28
:
369
–377,
1996
4
Bosco D, Meda P: Actively synthesizing beta-cells secrete preferentially after glucose stimulation.
Endocrinology
129
:
3157
–3166,
1991
5
Meda P, Bosco D, Chanson M, Giordano E, Vallar L, Wollheim C, Orci L: Rapid and reversible secretion changes during uncoupling of rat insulin-producing cells.
J Clin Invest
86
:
759
–768,
1990
6
Charollais A, Gjinovci A, Huarte J, Bauquis J, Nadal A, Martin F, Andreu E, Sanchez-Andres JV, Calabrese A, Bosco D, Soria B, Wollheim CB, Herrera PL, Meda P: Junctional communication of pancreatic beta cells contributes to the control of insulin secretion and glucose tolerance.
J Clin Invest
106
:
235
–243,
2000
7
Serre-Beinier V, Le Gurun S, Belluardo N, Trovato-Salinaro A, Charollais A, Haefliger JA, Condorelli DF, Meda P: Cx36 preferentially connects β-cells within pancreatic islets.
Diabetes
49
:
727
–734,
2000
8
Condorelli DF, Parenti R, Spinella F, Trovato Salinaro A, Belluardo N, Cardile V, Cicirata F: Cloning of a new gap junction gene (Cx36) highly expressed in mammalian brain neurons.
Eur J Neurosci
10
:
1202
–1208,
1998
9
Sohl G, Degen J, Teubner B, Willecke K: The murine gap junction gene connexin36 is highly expressed in mouse retina and regulated during brain development.
FEBS Lett
428
:
27
–31,
1998
10
Gylfe E, Grapengiesser E, Hellman B: Propagation of cytoplasmic Ca2+ oscillations in clusters of pancreatic beta-cells exposed to glucose.
Cell Calcium
12
:
229
–240,
1991
11
Jonkers FC, Jonas JC, Gilon P, Henquin JC: Influence of cell number on the characteristics and synchrony of Ca2+ oscillations in clusters of mouse pancreatic islet cells.
J Physiol
520
:
839
–849,
1999
12
Bergsten P: Slow and fast oscillations of cytoplasmic Ca2+ in pancreatic islets correspond to pulsatile insulin release.
Am J Physiol
268
:
E282
–E287,
1995
13
Gilon P, Henquin JC: Distinct effects of glucose on the synchronous oscillations of insulin release and cytoplasmic Ca2+ concentration measured simultaneously in single mouse islets.
Endocrinology
136
:
5725
–5730,
1995
14
Gazdar AF, Chick WL, Oie HK, Sims HL, King DL, Weir GC, Lauris V: Continuous, clonal, insulin- and somatostatin-secreting cell lines established from a transplantable rat islet cell tumor.
Proc Natl Acad Sci U S A
77
:
3519
–3523,
1980
15
Asfari M, Janjic D, Meda P, Li G, Halban PA, Wollheim CB: Establishment of 2-mercaptoethanol-dependent differentiated insulin-secreting cell lines.
Endocrinology
130
:
167
–178,
1992
16
Santerre RF, Cook RA, Crisel RM, Sharp JD, Schmidt RJ, Williams DC, Wilson CP: Insulin synthesis in a clonal cell line of simian virus 40-transformed hamster pancreatic beta cells.
Proc Natl Acad Sci U S A
78
:
4339
–4343,
1981
17
Efrat S, Linde S, Kofod H, Spector D, Delannoy M, Grant S, Hanahan D, Baekkeskov S: Beta-cell lines derived from transgenic mice expressing a hybrid insulin gene-oncogene.
Proc Natl Acad Sci U S A
85
:
9037
–9041,
1988
18
Miyazaki J, Araki K, Yamato E, Ikegami H, Asano T, Shibasaki Y, Oka Y, Yamamura K: Establishment of a pancreatic beta cell line that retains glucose-inducible insulin secretion: special reference to expression of glucose transporter isoforms.
Endocrinology
127
:
126
–132,
1990
19
Squires PE, Hauge-Evans AC, Persaud SJ, Jones PM: Synchronization of Ca(2+)-signals within insulin-secreting pseudoislets: effects of gap-junctional uncouplers.
Cell Calcium
27
:
287
–296,
2000
20
Meda P, Chanson M, Pepper M, Giordano E, Bosco D, Traub O, Willecke K, el Aoumari A, Gros D, Beyer EC, et al: In vivo modulation of connexin 43 gene expression and junctional coupling of pancreatic B-cells.
Exp Cell Res
192
:
469
–480,
1991
21
Hinz B, Celetta G, Tomasek JJ, Gabbiani G, Chaponnier C: Alpha-smooth muscle actin expression upregulates fibroblast contractile activity.
Mol Biol Cell
12
:
2730
–2741,
2001
22
Teubner B, Degen J, Sohl G, Guldenagel M, Bukauskas FF, Trexler EB, Verselis VK, De Zeeuw CI, Lee CG, Kozak CA, Petrasch-Parwez E, Dermietzel R, Willecke K: Functional expression of the murine connexin 36 gene coding for a neuron-specific gap junctional protein.
J Membr Biol
176
:
249
–262,
2000
23
Yamada K, Nakata M, Horimoto N, Saito M, Matsuoka H, Inagaki N: Measurement of glucose uptake and intracellular calcium concentration in single, living pancreatic beta-cells.
J Biol Chem
275
:
22278
–83,
2000
24
Roe MW, Worley JF 3rd, Qian F, Tamarina N, Mittal AA, Dralyuk F, Blair NT, Mertz RJ, Philipson LH, Dukes ID: Characterization of a Ca2+ release-activated nonselective cation current regulating membrane potential and [Ca2+]i oscillations in transgenically derived beta-cells.
J Biol Chem
273
:
10402
–10,
1998
25
Sherman A, Xu L, Stokes CL: Estimating and eliminating junctional current in coupled cell populations by leak subtraction. A computational study.
J Membr Biol
143
:
79
–87,
1995
26
Mears D, Sheppard NF, Jr., Atwater I, Rojas E: Magnitude and modulation of pancreatic beta-cell gap junction electrical conductance in situ.
J Membr Biol
146
:
163
–176,
1995
27
Gopel S, Kanno T, Barg S, Galvanovskis J, Rorsman P: Voltage-gated and resting membrane currents recorded from B-cells in intact mouse pancreatic islets.
J Physiol
521
:
717
–728,
1999
28
Niessen H, Harz H, Bedner P, Kramer K, Willecke K: Selective permeability of different connexin channels to the second messenger inositol 1,4,5-trisphosphate.
J Cell Sci
113
:
1365
–1372,
2000
29
Belluardo N, Trovato-Salinaro A, Mudo G, Hurd YL, Condorelli DF: Structure, chromosomal localization, and brain expression of human Cx36 gene.
J Neurosci Res
57
:
740
–752,
1999
30
Poitout V, Olson LK, Robertson RP: Insulin-secreting cell lines: classification, characteristics and potential applications.
Diabete Metab
22
:
7
–14,
1996
31
Newgard CB, Clark S, BeltrandelRio H, Hohmeier HE, Quaade C, Normington K: Engineered cell lines for insulin replacement in diabetes: current status and future prospects.
Diabetologia
40 (Suppl. 2)
:
S42
–S47,
1997
32
Efrat S, Tal M, Lodish HF: The pancreatic beta-cell glucose sensor.
Trends Biochem Sci
19
:
535
–538,
1994
33
Efrat S, Leiser M, Wu YJ, Fusco-DeMane D, Emran OA, Surana M, Jetton TL, Magnuson MA, Weir G, Fleischer N: Ribozyme-mediated attenuation of pancreatic beta-cell glucokinase expression in transgenic mice results in impaired glucose-induced insulin secretion.
Proc Natl Acad Sci U S A
91
:
2051
–2055,
1994
34
Bertuzzi F, Davalli AM, Nano R, Socci C, Codazzi F, Fesce R, Di Carlo V, Pozza G, Grohovaz F: Mechanisms of coordination of Ca2+ signals in pancreatic islet cells.
Diabetes
48
:
1971
–1978,
1999
35
Srinivas M, Hopperstad MG, Spray DC: Quinine blocks specific gap junction channel subtypes.
Proc Natl Acad Sci U S A
98
:
10942
–10947,
2001
36
Beardslee MA, Laing JG, Beyer EC, Saffitz JE: Rapid turnover of connexin43 in the adult rat heart.
Circulation Research
83
:
629
–635,
1998
37
Denhardt DT: Mechanism of action of antisense RNA: sometime inhibition of transcription, processing, transport, or translation.
Ann N Y Acad Sci
660
:
70
–76,
1992
38
Bennett CF, Cowsert LM: Application of antisense oligonucleotides for gene functionalization and target validation.
Curr Opin Mol Ther
1
:
359
–371,
1999
39
Zhang M, Bertram R, Sherman A, Satin LS: Contribution of gap junctional currents to islet electrical activity (Abstract).
Diabetes
51 (Suppl. 2)
:
A374
,
2002
40
Meda P: Assaying the molecular permeability of connexin channels.
Methods Mol Biol
154
:
201
–224,
2001
41
Srinivas M, Rozental R, Kojima T, Dermietzel R, Mehler M, Condorelli DF, Kessler JA, Spray DC: Functional properties of channels formed by the neuronal gap junction protein connexin36.
J Neurosci
19
:
9848
–9855,
1999
42
Squires PE, Persaud SJ, Hauge-Evans AC, Gray E, Ratcliff H, Jones PM: Coordinated Ca(2+)-signalling within pancreatic islets: does beta-cell entrainment require a secreted messenger.
Cell Calcium
31
:
209
–219,
2002
43
Calabrese A, Guldenagel M, Charollais A, Mas C, Caton D, Bauquis J, Serre-Beinier V, Caille D, Sohl G, Teubner B, Le Gurun S, Trovato-Salinaro A, Condorelli DF, Haefliger JA, Willecke K, Meda P: Cx36 and the function of endocrine pancreas.
Cell Adhes Commun
8
:
387
–391,
2001
44
Thorens B, Sarkar HK, Kaback HR, Lodish HF: Cloning and functional expression in bacteria of a novel glucose transporter present in liver, intestine, kidney, and beta-pancreatic islet cells.
Cell
55
:
281
–290,
1988
45
Knaack D, Fiore DM, Surana M, Leiser M, Laurance M, Fusco-DeMane D, Hegre OD, Fleischer N, Efrat S: Clonal insulinoma cell line that stably maintains correct glucose responsiveness.
Diabetes
43
:
1413
–1417,
1994
46
Clark SA, Quaade C, Constandy H, Hansen P, Halban P, Ferber S, Newgard CB, Normington K: Novel insulinoma cell lines produced by iterative engineering of GLUT2, glucokinase, and human insulin expression.
Diabetes
46
:
958
–967,
1997
47
Johnson JH, Newgard CB, Milburn JL, Lodish HF, Thorens B: The high Km glucose transporter of islets of Langerhans is functionally similar to the low affinity transporter of liver and has an identical primary sequence.
J Biol Chem
265
:
6548
–6551,
1990
48
Ashcroft FM, Proks P, Smith PA, Ammala C, Bokvist K, Rorsman P: Stimulus-secretion coupling in pancreatic beta cells.
J Cell Biochem
55
:
54
–65,
1994

Address correspondence and reprint requests to Alessandra Calabrese, Department of Morphology, University of Geneva, C.M.U., 1 rue Michel Servet, 1211 Geneva 4, Switzerland. E-mail: [email protected].

Received for publication 2 August 2002 and accepted in revised form 5 November 2002.

[Ca2+]i, cytoplasmic-free Ca2+; FSK, forskolin; Gc, gap junctional conductance; IBMX, 3-isobutyl-1-methylxanthine; TEA, tetraethylammonium.