Nutrients Induce Different Ca²⁺ Signals in Cytosol and Nucleus in Pancreatic β-Cells

Islets from adult (8–10 weeks old) Swiss albino male mice (OF1) killed by cervical dislocation in accord with our institutional guidelines were isolated and then dispersed into single cells following a procedure described elsewhere (14). Isolated cells were cultured in RPMI-1640 medium (Invitrogen, Barcelona, Spain) containing 5.6 mmol/l glucose for 12–16 h before experiments. Inmunocytochemistry analysis of cultured cells revealed that around 80% of the preparation was insulin-producing cells (16). In addition, we selected β-cells according to their large size and low nuclear/cytoplasmic ratio compared with other islet cell types (16). Thus, we expected the majority of cells that we studied in the experiments to be pancreatic β-cells.

Cells from the pancreatic β-cell line INS-1 were grown in monolayer cultures as follows (12). When cells reached 80% confluence, they were washed twice with Krebs-Ringer bicarbonate buffer at pH 7.4, containing 5.6 mmol/l glucose and 0.07% BSA. Cells were subsequently incubated 2 more days in RPMI-1640 medium containing 5 mmol/l glucose and 10% FCS, but the second-day cells were trypsinized and transferred to poly-l-lysine-treated coverslips.

Fatty acids were prepared as previously described (12). Defatted BSA was used as a control condition (12).

Cells were loaded with 2 μmol/l FLUO-3 acetoxymethyl ester (Molecular Probes, Leiden, the Netherlands) for 30 min at room temperature. Under these conditions, a homogenous distribution of the dye in the cytosol and in the nucleus was confirmed in the majority of cells (∼83%) by imaging cultures as described earlier (14). Cells were perfused at a rate of 0.5 ml/min in a Krebs-Ringer buffer (in mmol/l): 119 NaCl, 4.7 KCl, 1.2 MgSO₄, 1.2 KH₂PO₄, 25 NaHCO₃, 2.5 CaCl₂, 5 HEPES, and 3 glucose gassed with a mixture of 95% O₂, 5% CO₂ (pH = 7.4). Ca²⁺ concentration ([Ca²⁺]) changes were measured using spot confocal microscopy, which excels in measuring localized Ca²⁺-induced fluorescence elevations because of its remarkable signal-to-noise ratio over conventional systems (8,14,17). The spot illumination-detection configuration has been previously described in detail (14,17). Briefly, a laser-illuminated pinhole (10 μm) was focused onto a spot through the objective either on the nucleus or the bulk cytosol of an isolated islet cell (Fig. 1A). The full-width half-maximal dimension of the illumination spot is 0.6 μm, whereas the depth of field is 1.1 μm. Thus, the predicted detection volume is about 0.6 × 0.6 × 1.1 μm³, which is ∼1.2 and 0.075% of the nuclear and cell volume (assuming a radius of 2 and 5 μm, respectively). Changes in fluorescence in this small volume were detected with a photodiode (HR008; UDT, Hawthorne, CA), which was connected to an Axopatch-200A amplifier (50 GΩ feedback; Axon Instruments, Foster City, CA). The laser illumination time was 30 ms. We minimized the number of measurements on the same spot to prevent probe photo-bleaching. Under our experimental conditions and sampling rate, bleaching was insignificant. Autofluorescence was determined to be <2%. Nuclear Ca²⁺ was measured by positioning the spot in the center of the nucleus, identified easily by transmitted light microscopy, whereas cytosolic Ca²⁺ was recorded by placing the spot in a cytosolic area adjacent to the nuclear periphery (separation ≈ 1 μm). Distances were measured from acquired images using a charge coupled device (CCD) camera (14).

FLUO-3 has higher absorbance and fluorescence emission in the nucleoplasm than in the cytoplasm (18). Consequently, a specific calibration curve of the fluorescence for each cellular compartment is required to compare Ca²⁺ signals. We estimated in situ the dissociation constant (K_d) for FLUO-3 in two different pools of cells where the illumination spot was focused either in the nucleoplasm (n = 8 for β-cells and n = 8 for INS-1 cells) or in the cytoplasm (n = 9 for β-cells and n = 8 for INS-1 cells). After a brief treatment with 0.5 μmol/l digitonin for 3 min (Sigma, Madrid, Spain), cells were then exposed to an intracellular solution containing different [Ca²⁺] buffered with EGTA using a commercial calibration kit (Molecular Probes), but K⁺ was adjusted to a concentration of 125 mmol/l. This procedure did not produce leakage of the fluorescent probe. Calibrations gave a K_d in the cytosol of 0.66 ± 0.05 and 0.61 ± 0.08 μmol/l and a K_d in the nucleoplasm of 0.51 ± 0.04 and 0.48 ± 0.05 μmol/l (means ± SE) for β-cells and INS-1 cells, respectively. [Ca²⁺] was calculated according to the calibration equation: [Ca²⁺] = K_d (F − F_min)/(F_max − F). Fluorescence was measured after each experiment either in the nucleoplasm or in the cytoplasm by exposure of cells to the extracellular solution with saturating [Ca²⁺] (F_max) or absence of Ca²⁺ (F_min) (0 mmol/l Ca²⁺ plus 2 mmol/l EGTA), respectively, in the presence of 10 μmol/l ionomycin.

Nuclei were isolated according to procedures published earlier (8). Briefly, isolated islet cells were suspended in an intracellular solution (in mmol/l: 125 KCl, 2 K₂PO₄, 40 HEPES, 0.1 MgCl₂, at pH 7.2, and 100 nmol/l Ca²⁺) and disrupted by brief sonication. Nuclei were separated by centrifugation, resuspended in the intracellular solution, and then allowed to attach onto glass chambers. Isolated nuclei were loaded with the Ca²⁺ probe Calcium Green-1 dextran (30 μg/ml; 30 min at 4°C; Molecular Probes). For further details, see Quesada et al. (8).

Spot confocal microscopy has been successfully applied in pancreatic β-cells, skeletal muscle cells, and neurons to monitor Ca²⁺ dynamics in local intracellular domains (8,14,17). We used this technique to measure [Ca²⁺] changes either in the cytosol or in the nucleus of INS-1 cells and pancreatic β-cells exposed for 5 min to 40 mmol/l K⁺, 16.7 or 22 mmol/l glucose, and 0.5 mmol/l oleate in 3 mmol/l glucose (Fig. 1A–C). All stimuli led to a [Ca²⁺] rise in both intracellular compartments. When we examined the Ca²⁺ responses, we found no differences in duration between those generated in the nucleus and in the cytosol in any of the two types of insulin-secreting cells (P > 0.05; unpaired t test). However, significant differences in amplitude were obtained when comparing nuclear and cytosolic [Ca²⁺] mean values in both kinds of cells (Fig. 1D and E). Whereas K⁺ induced similar elevations of [Ca²⁺] in both intracellular compartments, glucose and oleate produced Ca²⁺ signals of higher amplitude in the nucleus (Fig. 1D and E).

It has been shown that Ca²⁺ probes exhibit enhanced absorbance and fluorescence emission in the nucleoplasm (18). Accordingly, we performed two independent in situ calibrations in the nucleus and in the cytosol to get comparable values between both cell locations (9) (see research design and methods). The differences in amplitude observed between the two compartments cannot arise from calibration artifacts, otherwise such differences should be reflected concomitantly with all the stimuli and not just with glucose and oleate. Similar [Ca²⁺] values in the nucleus and in the cytosol were obtained at rest with 3 mmol/l glucose as well as with stimulatory conditions with 40 mmol/l K⁺. These data also support the reliability of the calibration, because Ca²⁺ is expected to be equilibrated between both compartments in such conditions in β-cells (19) (see discussion). Furthermore, the same observations have been confirmed in two different kinds of insulin-producing cells with similar phenotypes and physiological responses.

Plasma membrane depolarization, either by a rise in extracellular K⁺ or by stimulatory glucose as a result of an ATP/ADP increase and K_ATP channel closure, leads to massive Ca²⁺ influx in the cytosol by activation of voltage-dependent Ca²⁺ channels (10). Likewise, oleate can produce a cytosolic increase of [Ca²⁺], probably by direct action on plasma membrane K_ATP channels or voltage-dependent Ca²⁺ channels (20,21). However, this classic pathway is not sufficient to explain the higher values of amplitude of the Ca²⁺ signals induced by both nutrients in the nucleus. In Fig. 2, we show that tolbutamide, which closes K_ATP channels, can trigger Ca²⁺ transients in isolated β-cell nuclei. This result confirms previous findings that propose the existence of a new signaling pathway, linking extracellular messages with nuclear function by means of nuclear K_ATP channels. Their blockage results in depolarization of the nuclear membrane and the release of Ca²⁺ from the nuclear envelope to the nucleoplasm, probably through ryanodine channels or a voltage-sensitive mechanism (8).

Ca²⁺ dynamics in the nucleus and the cytosol were monitored in isolated INS-1 cells and pancreatic β-cells exposed to K⁺, glucose, and oleate. Each stimulus produced a [Ca²⁺] rise of similar duration in both intracellular compartments, suggesting that, in our experimental conditions, there is no apparent restriction to Ca²⁺ diffusion between nucleus and cytosol, in agreement with previous reports in pancreatic β-cells (19) and in other systems (6). This idea is also supported by the fact that the elevation in cytosolic [Ca²⁺] produced by K⁺-induced Ca²⁺ influx was equilibrated in the nucleoplasm. Conversely, significant differences in the amplitude of the Ca²⁺ signals induced by glucose and oleate were reported between the nucleoplasm and the cytoplasm. Unlike the nutrients, K⁺ brought about similar values of amplitude in both cell locations, suggesting that the cytosolic [Ca²⁺] increase resulting from extracellular Ca²⁺ entry by plasma membrane depolarization is rapidly transmitted throughout the cell. However, an extracellular Ca²⁺ influx or a source of Ca²⁺ located in the cytosol cannot provide an explanation for the higher nuclear levels of the ion induced by nutrients. Thus, glucose and oleate should involve additional pathways and Ca²⁺ sources that lead to higher [Ca²⁺] in the nucleus. On account of nutrient metabolism, the basal ATP/ADP ratio, the concentration of diadenosine polyphosphates, and other messengers are augmented in the pancreatic β-cell, which in turn drive the closure of K_ATP channels, activating the well-known voltage-dependent Ca²⁺ influx through the plasma membrane (10,20,22,23). In addition to this classic pathway, it has been found that ATP, AP₄A (diadenosine tetraphosphate), and tolbutamide block K_ATP channels located on the nuclear membrane, causing a Ca²⁺ discharge in the nucleoplasm by mechanisms sensitive to nuclear transmembrane potential (8). Thus, the metabolism of glucose and oleate may result in a concurrent action on plasma membrane and nuclear K_ATP channels, causing a Ca²⁺ rise from two different pools. Ca²⁺ influx from the extracellular pool would be equilibrated throughout the whole cell, producing equivalent [Ca²⁺] in the nucleus and the cytosol, whereas Ca²⁺ release from the nuclear pool would account for the higher amplitude observed within the nucleoplasm. Although our results are consistent with similar observations in other systems (6), some investigators have failed to report differences between nuclear and cytosolic [Ca²⁺] levels in pancreatic β-cells (19). These discrepancies may be explained by the use of different methods to monitor [Ca²⁺] changes, or different culture conditions, which have altered Ca²⁺ responses in β-cells (24,25).

Given that gene transcription is sensitive to amplitude-modulated [Ca²⁺] changes (1–3), we propose that nutrients could signal nuclear functions by ensuring an appropriate transitory level of [Ca²⁺] within the nucleoplasm higher than that possibly required for the activation of cytosolic functions. An exciting challenge in the future would be to find whether these amplitude-modulated Ca²⁺ signals in the nucleus are able to regulate gene expression in the pancreatic β-cell, which still remains elusive, and how these signals are integrated within the islets of Langerhans.

This work was supported in part by grants from the Ministerio de Ciencia y Tecnología (PM99-0142 and GEN 2001-4748-CO5-05), Fundació Marató TV3 (99-1210), European Foundation for the Study of Diabetes, Instituto Carlos III (RETIp GO3/171, RGDM GO3/212, TERCEL GO3/210, and RCMN C03/08), and Generalitat Valenciana (GV99-139-1-04).

The authors thank E. Pérez, N. Illera, and A. Pérez for technical assistance.