Novel Players in Pancreatic Islet Signaling: From Membrane Receptors to Nuclear Channels

Changes in the intracellular Ca²⁺ concentration ([Ca²⁺]_i) in response to different stimuli play a pivotal role in signaling numerous cellular processes. The versatility of Ca²⁺ to relay cellular information is achieved by different ways that manage intracellular Ca²⁺ signals, which include location and/or spatial constraint, amplitude, frequency, and duration (1). In β-cells, as in the majority of cells, Ca²⁺ signals play a key role as a messenger for multiple functions. The classic stimulus-secretion coupling that drives insulin release involves the closure of plasma membrane ATP-dependent K⁺ (K_ATP) channels by increasing the intracellular ATP/ADP ratio (2), diadenosine polyphosphates (DPs) (3,4), and cGMP (5) on account of glucose metabolism (Fig. 1). Channel closure induces membrane depolarization that activates voltage-operated Ca²⁺ (VOC) channels and Ca²⁺ influx (6). The set of channels of β-cell plasma membrane generates an oscillatory electrical activity that causes [Ca²⁺]_i to oscillate (7–9). Remarkably, the oscillatory Ca²⁺ pattern triggers a pulsatile insulin secretion (10–12). Membrane potential and [Ca²⁺]_i oscillations are built as a result of the precise interplay between gap-junction conductance and cell input conductance in the appropriate range (13,14).

The stimulus-secretion coupling is regulated not only by global Ca²⁺ oscillations, as described in the above paragraph, but by localized Ca²⁺ signals as well. These occur beneath the plasma membrane in areas associated with hot spots of exocytosis (15–17). There are several advantages of developing spatial-restricted Ca²⁺ signals rather than global increases in Ca²⁺ to control insulin release. Localized Ca²⁺ signals have only a limited spatial spread, and [Ca²⁺]_i declines sharply away from the site of origin (i.e., VOC channels) because of the buffering properties of the cytosol. This results in high [Ca²⁺]_i near the plasma membrane Ca²⁺ channels and adjacent Ca²⁺-sensitive targets in the exocytotic machinery. As a consequence, the Ca²⁺ signal has a highly specific and efficient effect on secretion, probably by ensuring the precise [Ca²⁺]_i required for the activation of the low-affinity Ca²⁺-dependent exocytotic enzymes (17,18) and preventing a massive activation of other Ca²⁺-dependent functions distant from the sites of exocytosis. In the pancreatic β-cell, as in other cell types, the secretory granules exist in distinct functional pools, with different proximity to the plasma membrane and probably with different exposure to [Ca²⁺]_i (19). One of them is the readily releasable pool (RRP). A group of these RRP granules is located in close proximity to VOC channels, and, consequently, they should be exposed to exocytotic levels of Ca²⁺ during the opening of Ca²⁺ channels. These granules are referred to as the immediate releasable pool and are situated ∼10 nm from Ca²⁺ channels. At this distance, the Ca²⁺ concentration can be expected to rise to high levels, probably in the micromolar range (15,16), required by proteins participating in membrane fusion and exocytosis (17,18). Both RRP and immediate releasable pool granules interact with the plasma membrane via exocytotic proteins, which mediate granule release and thus insulin secretion (20,21). Therefore, because release-competent granules are situated close to the points of Ca²⁺ entry, the development of localized submembranal Ca²⁺ gradients would dramatically increase the efficiency and specificity of Ca²⁺ to trigger the release of these granules. Another advantage of these local Ca²⁺ transients is that they can be removed from the cytoplasm at relatively low energy cost for the cells, in contrast to global Ca²⁺ changes. This is particularly relevant in the pancreatic β-cell because Ca²⁺ removal from the cytoplasm is mostly driven by Ca²⁺ ATPases. Consequently, ATP consumption would produce a fall in the ATP/ADP ratio, which could affect the K_ATP channel activity and thereby the membrane potential. Finally, because secretion in the pancreatic β-cell can be maintained over periods of several minutes, the spatial restriction of Ca²⁺ signals close to the plasma membrane for insulin release would prevent a broad activation of other Ca²⁺-dependent functions. In fact, prolonged high cytosolic Ca²⁺ levels may trigger other cellular processes including apoptosis, which would lead to cell death.

How were submembranal local Ca²⁺ signals determined in β-cells? In the early 1990s, two important findings suggested the existence of localized Ca²⁺ gradients. On the one hand, agonists such as carbachol were not able to initiate insulin release when applied in the absence of glucose, even when they induced higher global [Ca²⁺]_i signals than glucose (22). On the other hand, there were discrepancies between the Ca²⁺ sensitivity of exocytotic responses and the levels of [Ca²⁺]_i reached in the cytosol during glucose stimulation (22–25). Therefore, it was suggested that high [Ca²⁺] changes restricted to discrete domains beneath the cell membrane could have principal control on the secretory machinery and thus could be responsible for triggering exocytosis of insulin granules. In light of these new ideas, novel digital imaging methods that drastically increase the image spatial resolution by removing the out-of-focus fluorescence, have allowed us to measure the development of submembranal [Ca²⁺]_i gradients in pancreatic islet cells in response to glucose and carbachol (22). That study showed that 1) stimulatory glucose concentrations induced steep spatial gradients of [Ca²⁺]_i in the vicinity of the plasma membrane, 2) Ca²⁺ levels in response to glucose reached values in the micromolar range, 3) these local Ca²⁺ signals were polarized in the cell, and 4) carbachol induced [Ca²⁺]_i changes that originated in the center of the cell and declined toward the cell membrane. These results indicated that insulin release might be principally regulated by [Ca²⁺]_i in the vicinity of the exocytotic sites. New evidence has been accumulated in favor of this idea. Actually, in the pancreatic β-cell, there is a faint coupling of Ca²⁺ channels to the secretory machinery, with exocytosis being triggered by peaks of [Ca²⁺]_i of ∼10 μmol/l at the release sites (16,26). In a new study, the effects of several exogenous Ca²⁺ chelators on insulin release were analyzed to know how cytosolic Ca²⁺ distribution affected the dynamics of insulin exocytosis (27). These experiments were also supported by modeling Ca²⁺ diffusion in the vicinity of channels and release sites in conditions that simulated trains of depolarizations that resembled those induced by stimulatory glucose concentrations. The findings obtained in this study suggested that the first phase of glucose-induced insulin release is due to exocytosis of a primed pool of secretory granules located in close proximity to Ca²⁺ channels (around 50 nm), whereas the second phase is due to mobilization of a reserve pool of granules placed at ∼300 nm from the plasma membrane. In addition, this model predicted that [Ca²⁺]_i in the outermost shells of pancreatic β-cells was in the micromolar range but not higher than 14 μmol/l. These results were consistent with other reports (16,19,26) and provided additional evidence in favor of the existence of two different vesicle types responsible for glucose-induced insulin release: primed vesicles, located close to the Ca²⁺ channels, and reserve vesicles, not strictly co-localized with the Ca²⁺ channels.

Later on, a new report using a Monte Carlo simulation of three-dimensional buffered Ca²⁺ diffusion (28,29) demonstrated again that a large free Ca²⁺ concentration was localized near the ion channels, with steeper gradients in the submembrane domain. Theoretical predictions were then further supported by experimental data obtained using spot confocal microscopy, a technique that excels in measuring spatial-restricted intracellular Ca²⁺ signals because of its remarkable high signal-to-noise ratio (15,30). This technique enabled us to measure and characterize the glucose-induced [Ca²⁺]_i gradients in the first two microns close to the plasma membrane to verify the values simulated by models (15). These results demonstrated that, in pancreatic islet cells, nutrient secretagogues induce Ca²⁺ microgradients, which dissipate within the first micron from the membrane, and that [Ca²⁺]_i can readily reach around 10 μmol/l (Fig. 2). The finding that these gradients were maintained during the period that the stimulus was present also indicated that pancreatic islet cells might have a specialized and compartmentalized Ca²⁺ homeostatic machinery, mainly controlled by the mitochondrias and the endoplasmic reticulum (Fig. 1). These intracellular organelles are likely to be in close proximity to the plasma membrane, shaping these Ca²⁺ gradients and restricting the action of the Ca²⁺ signal to the exocytotic site (15).

The endocrine pancreas is not considered to be a classic estrogen target, although the effects of 17β-estradiol on some physiological aspects of the islets of Langerhans have been known for a long time. On the one hand, the level of plasma insulin is increased in pregnant rats in response to increased levels of sex steroids (31–33). Moreover, 17β-estradiol at concentrations comparable to those of pregnancy enhances insulin secretion in perfused rat pancreas (34). In humans, 17β-estradiol reverses the effect of menopause on glucose and insulin metabolism, resulting in an increase of pancreatic insulin secretion as well as an improvement of insulin resistance (35,36). On the other hand, in glucagon-releasing α-cells, it was described that 17β-estradiol produces an inhibitory effect on glucagon secretion (37). In spite of this evidence, the mechanism of action used by 17β-estradiol on α- and β-cells is just being discovered.

Recently, an estrogen membrane receptor was visualized in pancreatic β-cells. It is responsible for a rapid insulinotropic effect of 17β-estradiol when applied at physiological concentrations (38). Once bound to its membrane receptor, estrogen triggers the synthesis of cGMP, which in turn activates protein kinase G (PKG). Then, the K_ATP channels close in a PKG-dependent manner, causing the plasma membrane to depolarize, enhancing intracellular calcium signals (5). As a result, insulin secretion is increased (38). This receptor not only exists in pancreatic β-cells, but also in α-cells. When 17β-estradiol acts through this receptor in α-cells, it inhibits low glucose-induced [Ca²⁺]_i oscillations, and therefore it abolishes glucagon release (39).

The cGMP-PKG pathway is not exclusive of the membrane effect of 17β-estradiol, because it is also used by glucose to induce insulin secretion. It has been reported that glucose induces a dose-response increase on cGMP levels in islets of Langerhans (40). Furthermore, inhibitors of cGMP synthesis block both the insulin secretion (41,42) and the intracellular calcium oscillatory pattern (Fig. 3) induced by glucose. Although it is not known whether the increase in cGMP levels after glucose metabolism directly affects the K_ATP activity, estradiol certainly produces a decrease in its activity through a PKG-dependent mechanism (5,38) (Fig. 4.). This points to the potassium channel as the target for the cGMP-PKG pathway activated by glucose and estradiol, which in turn produces the depolarization of the plasma membrane, the increase in the intracellular calcium, and the insulin release. However, whether the K_ATP is phosphorylated by PKG or is closed by an indirect mechanism is still unknown. The fact that the estradiol action is mediated by the same intracellular pathway as the glucose highlights the importance of the hormone as a potentiator of the insulinotropic effect of the nutrient.

The nature of the membrane estrogen receptor responsible for the rapid effects known to be produced by estradiol in a large number of cell types has been a matter of controversy in recent years. Whereas some groups support the view that the intracellular estrogen receptor (ER)-α can be located at the plasma membrane, it still remains unclear how this soluble intracellular receptor can be attached to the membrane (43). However, pharmacological studies of the islets of Langerhans demonstrate that the membrane ER responsible for the physiological effects of 17β-estradiol is not related to any of the intracellular ERs described so far. This was shown by the use of several antibodies against ER-α and ER-β, which failed to recognize any structure on the plasma membrane of the cells (44). Instead, the membrane ER seems to be an undefined receptor that binds adrenaline, norepinephrine, and dopamine (39) but does not interact with either specific anti-estrogens such as ICI182,780 (39) or other steroids such as estriol (44). This nonclassical membrane estrogen receptor is not related to adrenergic or dopaminergic receptors, as demonstrated by the lack of interaction with specific ligands for these receptors. Many chemicals were tested in binding experiments, and a structural feature was obtained from those able to block the interaction between 17β-estradiol and nonclassical membrane ER in nonpermeabilized cells. Surprisingly, it was found that the nonclassical membrane ER binds molecules with either a catechol ring or the A-ring of the 17β-estradiol within its structure, such as 2-hydroxyestradiol (39,44).

In many different types of cells, including the pancreatic β-cell, Ca²⁺ changes in the nucleus are believed to be an important regulator of gene expression (45,46). Because nuclear processes exhibit a significant Ca²⁺-dependent responsiveness, the regulation of Ca²⁺ within the nucleus should be strictly controlled. However, the mechanisms whereby nuclear Ca²⁺ signals are induced and regulated are still controversial. Whereas it has been shown that functional channels in the nuclear membrane can regulate nuclear Ca²⁺ (46,47), some groups have argued against an independent nuclear Ca²⁺ homeostasis because of the presence of nuclear pores, which would allow for a rapid Ca²⁺ equilibration between nucleus and cytoplasm (47).

In pancreatic β-cells, besides the well-known effect on insulin release, Ca²⁺ influx by membrane depolarization induced by nutrient metabolism can also activate gene transcription (48). In fact, high glucose and fatty acids can regulate the expression of immediate-early response genes such as c-fos, c-jun, junB, zif-268, and nur-77, which are also coding for transcription factors implicated in several functions such as cell proliferation and differentiation (49,50). It has also been shown that long exposure of β-cells to high glucose levels increase the expression of genes coding for some metabolic enzymes (51). The expression of most of these genes, whose induction is mediated by nutrients, has in common the participation of a Ca²⁺ signal (52). However, the mechanisms that link these extracellular messages with nuclear function have been explored little. Recently, a novel pathway was reported by which nutrients could transmit Ca²⁺ messages to the nucleus that activate nuclear function (46). In this study, it was shown that a K_ATP channel with similar properties to that found on the plasma membrane is also present on the nuclear envelope of pancreatic β-cells. Blockade of nuclear K_ATP channels with the sulfonylurea tolbutamide or with DPs triggered nuclear Ca²⁺ transients by a voltage-sensitive mechanism (possibly ryanodine receptors), which induced phosphorylation of the transcription factor CREB (cAMP response element binding protein) (Fig. 4). In whole cells, fluorescence in situ hybridization also revealed that these Ca²⁺ signals might trigger c-myc expression. Thus, in the pancreatic β-cell, the increase of the ATP/ADP ratio, DPs, and possibly cGMP, as a result of nutrient metabolism, may act not just on plasma membrane K_ATP channels inducing extracellular Ca²⁺ influx, but also may affect those channels located in the nucleus producing nuclear Ca²⁺ transients. These novel findings indicate that functional K_ATP channels in the nucleus could be critical to link nutrient metabolism and nuclear function by Ca²⁺ signals. In view of the fact that, in several systems, gene expression is modulated by the code established by the amplitude, duration, and/or frequency of Ca²⁺ signals (53), it would be particularly interesting to analyze in future studies the relation between the dynamics of Ca²⁺ in the nucleus and its effect on nuclear function in the pancreatic β-cell. Actually, preliminary results of our group indicate that variations in the amplitude and duration of Ca²⁺ nuclear levels could play a role in the regulation of the expression of genes such as c-fos and c-myc.

This study was partially supported by grants from Ministerio de Ciencia y Tecnología, the Juvenile Diabetes Research Foundation International, Generalitat Valenciana, Fundació Marató TV3, the European Foundation for the Study of Diabetes, and Instituto Carlos III (RET:p G03/171, RGDM G03/212, TERCEL G03/210, RCMN C03/08).