The Ryanodine Receptor Calcium Channel of β-Cells: Molecular Regulation and Physiological Significance

IP₃Rs and RY receptors, the two main families of intracellular Ca²⁺ channels, share some structural and functional similarities. IP₃Rs of β-cells have been reviewed elsewhere and will be mentioned here only briefly (6). cDNAs for three RY receptors have been cloned. RY₁ is present mainly in skeletal muscle. RY₂ is abundant in heart but is also the major isoform in the brain. RY₃ is present at low levels in many cells. The three genes of human RY receptors RYR1, RYR2, and RYR3 have been mapped to chromosome positions 19q13.1, 1q42.1–1q43, and 15q14-q15, respectively. Two putative alternative splicing sites have been postulated for RY₂ mRNA. Homologues of mammalian RY receptors and IP₃Rs are present in C. elegans, D. melanogaster, and Zebrafish. IP₃Rs and RY receptors probably arose by a gene duplication event in invertebrates. The phylogenetic tree of RY receptor family suggests that the three vertebrate RY receptor genes were probably generated at the same time (7). However, one analysis suggests that RY₂ may be the original vertebrate form of RY receptors (8). The three isoforms probably arose by two gene duplication events in vertebrates (9).

Investigators considered the possibility that β-cells might have RY receptor-like channels when they found that theophylline releases Ca²⁺ from intracellular stores (10). After the discovery that IP₃ releases Ca²⁺ from the endoplasmic reticulum (ER) of β-cells, studies of this channel dominated the field (11). However, IP₃ can release only ∼50% of the Ca²⁺ sequestered into the ER (11). It was a possibility that the IP₃-insensitive ER Ca²⁺ pool may be equipped with the RY receptor (5). Thimerosal, a sulfhydryl-oxidizing agent that activates RY receptors, proved useful to test this hypothesis.

In RINm5F cells and β-cells obtained from ob/ob mice, thimerosal released Ca²⁺ from the IP₃-insensitive ER Ca²⁺ pool (5,12). The release was potentiated by caffeine, suggesting that a RY receptor might be involved. While thimerosal can activate some IP₃Rs, the type 3 IP₃R that is the predominant isoform in RINm5F cell and rat β-cells (13) is not activated by thimerosal (14). Subsequent studies that demonstrated Ca²⁺ release by ryanodine from islet microsomes (15) and by caffeine from ER of intact β-cells (16) strengthened the view that RY receptors are present in β-cell.

RY receptors have been demonstrated in a variety of insulin-secreting cells (Table 1). Such studies used pharmacological tools (5), endogenous ligands (15,17), molecular techniques (16,18,19), and quantitative ryanodine-binding (19). Some studies, however, imply total lack of RY receptors in β-cells (20,21). Contradictory reports as to whether β-cells have RY receptors are common even when the same investigators are involved (16,20,22,23). This may be due to low abundance of these channels, differences in cell types, or methods used. Thus, some clones of RINm5F cells have RY receptors (5,24) whereas others do not (25). β-Cells from some colonies of ob/ob mice have RY receptors (16) whereas others do not (18). When the same cell types are used, caffeine is more likely to release Ca²⁺ from the ER of intact cells (25,26) than of permeabilized cells (23), suggesting that the permeabilization results in the loss of one or more regulatory factors.

The most abundant RY receptor mRNA in β-cells is that of RY₂, as indicated by RNAse protection assay and RT-PCR analysis (16,18,19). The probes and primers used in these studies target different regions of RY₂ cDNA that code for the highly conserved membrane-spanning and COOH-terminus portions of RY₂. It is not known whether the β-cell RY₂ mRNA is identical to that of the heart or to one of its alternatively spliced transcripts. At the protein level, the presence of RY₂ receptors has been demonstrated by Western blot of membranes from INS-1 cells (25). In human β-cells, the receptor has been demonstrated by quantitative ryanodine-binding (19). Very low levels of RY₁ mRNA have been observed in βTC3 cells, and RY₃ mRNA is expressed in HIT-T15 cells (27). Thus, on the balance of current evidence, the existence of RY receptors in β-cells appears to be well documented.

An important issue is whether the density of RY receptors in β-cells is high enough to be of functional significance and how it compares with that of IP₃Rs in these cells. The level of RY₂ in rodent β-cells or cell lines as compared with that in heart or brain is low. In RNAse protection assay, mouse βTC3 cells show a band corresponding to RY₂ mRNA, which is ∼1,000-fold less than that in heart (16). In INS-1 cells, RY₂ protein is ∼10 times less than that in brain (25). However, comparisons with heart or brain may be misleading since these tissues contain high amounts of the protein. RY₂ level in β-cells may be comparable to that in pancreatic acinar cells, kidney, endothelial cells, and adrenal chromaffin cells (28,29). The level of RY receptors in the glucose-insensitive RINm5F cells as evidenced from RT-PCR (30), Western blot, and functional studies is low (25). Nevertheless, these cells were useful for providing the first indications of the existence of RY receptors in β-cells (5). INS-1 cells, which are related to RINm5F cells but are glucose-responsive, have more RY receptors (25,26).

In ob/ob mice, leptin deficiency leads to profound disturbances with accompanying changes in the islets. These mice are used as a model for studies in obesity and diabetes. Paradoxically, islets from ob/ob mice are used for “physiological” studies. This is because ob/ob islets are large and consist of 90–95% β-cells. These β-cells respond “normally” to elevated glucose with a release of insulin. However, islet-phenotype in these mice depends on the genetic background on which the ob gene is expressed. Islets from a noninbred colony of ob/ob mice express RY₂ at low level (16). In islets of ob/ob mice obtained from the Jackson Laboratories (an inbred colony), RY₂ message was not detected even on PCR amplification (18). Thus, the level of RY receptors in different strains of ob/ob mice varies. This may be one reason why caffeine and ryanodine are either not effective (21) or only marginally effective (20) in ob/ob β-cells. One study compared the levels of IP₃Rs and RY receptors in mouse islets and concluded that mouse islets contain only RY₂ and almost no IP₃Rs. However, these authors did not use efficient primers for IP₃R-1, as they did not detect IP3R-1 mRNA in the brain (18).

The relative densities of RY receptors and IP₃Rs in β-cells may vary in different species. However, evidence that human β-cells have RY receptors is convincing (19). Finally, it must be remembered that there is often no direct relationship between channel densities and the magnitude of a functional response. Thus, despite relatively low densities, strategic location of these channels at intracellular sites may be important for cell function.

RY receptors are made of four ∼560-kDa RY receptor protomers and four associated molecules of FKBP12 or FKBP12.6 (31). The latter are isoforms of the 12-kDa binding protein for the drug FK506. Each subunit of the RY receptor may form a pore and FKBP enables the four subunits to gate as one unit. The channel is a macromolecular complex with >20 associated proteins including calmodulin, calsequestrin, anchoring proteins, kinases, and phosphatases. The subunits of RY receptor have an enormous NH₂-terminal cytosolic domain followed by 4–10 highly conserved transmembrane segments, which are followed by another short cytosolic domain. Cryoelectron microscopy and 3d-reconstruction reveal that RY receptor is a fourfold symmetrical mushroom-like structure with a large cytosolic assembly and a short transmembrane region. The cytosolic domain is the modulatory region and contains binding sites for Ca²⁺, adenine nucleotides, calmodulin, FKBPs as well as the phosphorylation sites.

FKBP12.6 binds to isoleucine-proline sequence of each subunit of RY₂ receptors and thereby stabilizes the RY receptor tetramer and facilitates coordinated gating of the channel. Immunosuppressants used in islet transplantation, e.g., sirolimus (rapamycin) and tacrolimus (FK506), bind to FKBP12.6. Islets contain FKBP12.6 and FK506 releases Ca²⁺ from islet microsomes. According to one report, cADPR activates RY receptor by binding to FKBP12.6, but this view is not supported by other studies (32,33).

RY receptor is a cation-selective channel that allows permeation of Ca²⁺, many other divalent and monovalent cations, and under certain experimental conditions, even large molecules like glucose (34). Because of its short and wide channel region, RY receptor is suitable for sudden and large release of Ca²⁺ from intracellular stores. Ca²⁺ conductance of RY receptor is on the order of 100 pS, which compares to about 10 pS for voltage-gated Ca²⁺ channels. Physiological regulators of RY receptors include Ca²⁺, Mg²⁺, and ATP, which act by binding to the cytosolic sites. An important property of RY receptors is that they are regulated by ER Ca²⁺ load: the open probability of RY receptor is reduced as the luminal [Ca²⁺] is reduced. Ca²⁺ may be released spontaneously and cyclically when the ER Ca²⁺ load is high (35–37).

A fundamental property of RY receptors is that they can be both activated and inhibited by cytosolic Ca²⁺. Ca²⁺ at nanomolar to micromolar concentration increases the open probability of RY₂ receptor by acting on the high-affinity Ca²⁺-binding sites. The concentration and speed of delivery of Ca²⁺ are critical determinants for activation of RY receptors (35). Activation of l-type Ca²⁺ channels at more negative potentials (e.g. −40 to −10) is more effective in activating RY receptor because of larger driving force of Ca²⁺ at more negative voltages. Brief opening of l-type Ca²⁺ channels may be more effective in activating RY receptors than long-lasting opening. A micromolar to millimolar concentration of Ca²⁺ decreases open probability of RY receptor by acting on the low-affinity Ca²⁺-binding sites. Mg²⁺ competes with Ca²⁺ at both the activation and inhibition sites of RY receptors.

The tetrameric RY₂ has ∼80 free cysteines, some of which are critical for gating the channel. Reduced glutathione inhibits, and oxidized glutathione activates RY receptors (38). Redox-active molecules including NO and free radicals may thus affect RY receptors, and such processes are likely to be of physiological or pathological significance. Thiol oxidation has been shown to activate β-cell RY receptor (5), leading to release of Ca²⁺ from intracellular stores and increase of [Ca²⁺]_c (39).

RY receptors have binding sites for many agents reflected in the huge size of these channels. Methylxanthines, imidazoles and imidazolines, perchlorates, suramin, and volatile anesthetics activate RY receptor. Inhibitors of RY receptors include ruthenium red, procaine, tetracaine, ryanodine, dantrolene, and octanol. Some of these agents have additional concentration-dependent effects on other channels or pumps. On the other hand, drugs such as verapamil and D600, commonly used as blockers of l-type Ca²⁺ channel, also inhibit RY receptors (40). Familiarity with the usage and mechanism of action of these pharmacological tools is important for RY receptor studies.

Millimolar concentrations of methylxanthines activate RY receptors directly and may increase cAMP level indirectly, i.e., by inhibiting phosphodiesterases (PDEs). Caffeine is the most commonly used RY receptor-activator in muscle and neuronal research. The methylxanthines extensively used in islet research are theophylline and 3-isobutyl-1-methylxanthine (IBMX). The effects of theophylline and caffeine as activators of RY receptors are highly comparable (41). Millimolar concentrations of IBMX, commonly used as a PDE inhibitor, can activate RY receptors directly (42). In this respect, the efficacies of 1.5 mmol/l caffeine, theophylline, and IBMX may be almost equal (43).

Many imidazoles and imidazolines activate RY receptors (44). The imidazole ring is part of the xanthine structures and appears to be necessary for activating RY receptors (45). cAMP has an imidazole moiety as part of its structure. Dibutyryl cAMP, a lipophilic analog of cAMP, when used at millimolar concentration can activate RY receptor directly by binding to the caffeine site (44). Theophylline, IBMX, and imidazole release Ca²⁺ from intracellular stores of β-cells (10,46–48). Ca²⁺ release from ER stores can be detected by measuring Ca²⁺-activated plasma membrane conductance. In glucose-primed islets, theophylline increases ⁸⁶Rb⁺ efflux (49,50) and hyperpolarizes the β-cell membrane (50), suggesting Ca²⁺ release from the ER and consequent activation of calcium-activated K⁺ channels (K⁺_Ca). This Ca²⁺ release and consequent increased ⁸⁶Rb⁺ efflux cannot be entirely due to inhibition of PDEs. The structurally related activator of RY receptors, imidazole, which activates PDEs (51,52), also increases ⁸⁶Rb⁺ efflux (indicating Ca²⁺ release from ER) in glucose-primed islets (53).

Compared with theophylline and IBMX, fewer studies tested Ca²⁺ release by caffeine in β-cells. The main regulator of RY₂ receptor is Ca²⁺, which activates or inhibits the channel, depending on concentration of the ion and the rate at which it is delivered. Ca²⁺ binds to separate high-affinity activation and low-affinity inhibition sites on the RY receptor. There is thus a competition between the two, and it appears that activation can only take place if the increase in [Ca²⁺]_c is fast (35,54). The full-blown Ca²⁺ release by caffeine, as occurs in situ is a two-step process; in the first step, there is a small Ca²⁺ release by caffeine. If the released Ca²⁺ is mopped up, for instance, by binding to high intracellular buffers or fura-2, the second step will not be engaged (54). The second step is a regenerative phenomenon in which released Ca²⁺ acts on a cluster of RY receptors, triggering large Ca²⁺ release. Caffeine increases the Ca²⁺ affinity of the Ca²⁺ activation site of RY receptor. For caffeine to activate RY receptor optimally, it must increase the sensitivity of the activation site to Ca²⁺ quickly, so that there is little time for inhibition of the channel by the released Ca²⁺ (54). Slow application of caffeine by conventional perfusion systems driven by peristaltic pumps may not elicit a rise in [Ca²⁺]_c (20). Furthermore, when channel density is low, a modest Ca²⁺ release by caffeine is counteracted by extrusion and uptake mechanisms (55). Caffeine applied rapidly by a puffer pipette or a U-tube (56) typically induces transient increase in [Ca²⁺]_c (16,19,29,57–59). Dose-response studies of Ca²⁺ release by caffeine in β-cells have not been reported but the most commonly used concentration is 5–10 mmol/l. From studies in other cells, it is known that the threshold concentration of caffeine for Ca²⁺ release is ∼250 μmol/l (43).

When [Ca²⁺]_c is measured in single β-cells, only ∼20–50% of cells respond to caffeine (16,17,58). This suggests that several conditions must be fulfilled for activation of RY receptors. The number of cells responding to caffeine and magnitude of Ca²⁺ release is increased by various maneuvers. Adequate Ca²⁺ loading of the ER, slightly elevated basal [Ca²⁺]_c, low cytosolic [Mg²⁺], and synchronous activation of many RY receptors seem to be necessary to elicit a Ca²⁺ transient with caffeine. Experimentally, the Ca²⁺ pools can be filled, for instance, by depolarizing the cell transiently, allowing Ca²⁺ entry through the voltage-gated Ca²⁺ channels, before applying caffeine (19,54). The number of excitable RY receptors can be increased by protein kinase A (PKA) phosphorylation (16,60). Cytosolic [Mg²⁺] can be reduced by increasing intracellular [ATP], for instance, by providing high glucose.

A difficulty in interpreting Ca²⁺ data obtained with caffeine arises from the fact that the xanthine drug inhibits PDEs. Furthermore, caffeine and theophylline can cause modest depolarization of β-cell plasma membrane (20) by inhibiting K_ATP (ATP-sensitive potassium) channel and possibly by activating a nonselective cation channel in the plasma membrane (61). The resulting Ca²⁺ entry through the plasma membrane Ca²⁺ channels may obscure caffeine-induced Ca²⁺ release (20). Experimentally such difficulties can be avoided if caffeine is used under conditions in which membrane potential is clamped at −70 mV (19).

9-methyl-7-bromoeudistomin D is a novel activator of RY receptors (62). It is useful since it is not a methylxanthine and does not inhibit PDEs. 4-chloro-m-cresol and 4-chloro-3-ethylphenol are also useful since they do not inhibit PDEs, activate RY receptors, and do not activate IP₃Rs (16,22,63,64). However, they inhibit islet metabolism (65).

The molecular basis of ryanodine action is still poorly understood. RY receptors have a single high-affinity site and a separate low-affinity site for ryanodine. The alkaloid can both activate and inhibit RY receptor, making it sometimes difficult to interpret the results obtained with ryanodine. In general, nanomolar ryanodine sensitizes RY receptor to activation by Ca²⁺ and micromolar ryanodine inhibits it. Some analogs of ryanodine (e.g., β alanyl ryanodine) only activate the channel, but we have not found them useful in experiments with intact β-cells (16). In cell-free systems, prolonged exposure to a high concentration of ryanodine (e.g., 10 μmol/l) locks the RY receptor in a partially open state. Such treatment eventually depletes the ryanodine-sensitive Ca²⁺ pools because of drainage of Ca²⁺. Very high concentrations of ryanodine (e.g., 100 μmol/l to 10 mmol/l) completely block the channel. Inhibition of cellular processes by high concentrations of ryanodine is taken as evidence for the involvement of RY receptors. However, ryanodine binding is use-dependent; the alkaloid binds only to the open conformation of the channel. For demonstrating inhibition by ryanodine, the RY channel must first be opened in the presence of the alkaloid and maintained open for a long time to allow ryanodine-binding to the inhibitory binding sites. Thus, when cells are pretreated in the presence of caffeine (or elevated [Ca²⁺]_c) and ryanodine for 5–30 min, the RY receptor of RINm5F cells (24) and β-cells is inhibited (19). If such protocols are not used and ryanodine is added to the perfusion for a brief period, the RY receptor of β-cells may appear to be insensitive to the inhibitory action of the alkaloid (59,64). This may explain why some reports demonstrate clear effect of ryanodine on β-cells (19,66,67), whereas others find little or no effect (2,3). It must be emphasized that it is often difficult to inhibit RY receptors by high concentration of ryanodine in intact cells and such lack of inhibition alone should not preclude conclusions as to the existence or involvement of RY receptors in intact cells (68).

cAMP is not primarily a Ca²⁺-releasing messenger in the sense that inositol 1,4,5-trisphosphate is. cAMP per se does not release Ca²⁺ from ER of β-cells when the [Ca²⁺]_c or ER luminal [Ca²⁺] is low (16). cAMP, through PKA-mediated phosphorylation, ensures the in situ excitability of RY receptor. In resting β-cells, [Ca²⁺]_c is low and cytosolic [Mg²⁺] is high, which keeps the RY receptors inhibited. cAMP-dependent phosphorylation per se does not activate the channel; instead it releases the channel from Mg²⁺ inhibition (60). PKA phosphorylation brings the RY₂ receptor to an excitable state: the channel can then be excited either by the Ca²⁺ entering through the voltage-gated Ca²⁺ channels or simply by high loading of the ER (37). Furthermore, PKA phosphorylation favors dissociation of FKBP12.6 from RY₂ and thus increases open probability of the channel (31).

Kang et al. (26) have described a PKA-independent effect of cAMP on the RY receptor of β-cells. In this mode, cAMP promotes Ca²⁺ release through RY receptor as a consequence of increased filling of the ER by a mechanism that involves cAMP-regulated guanine nucleotide exchange factor and its interaction with Rap1b (26,69). There is no consensus as to the consequences of phosphorylation of IP₃Rs in terms of Ca²⁺ release. Xestospongic C, an inhibitor of IP₃Rs, does not inhibit CICR in β-cells, suggesting that IP₃Rs do not play a major role in mediating CICR in these cells (26). Some reports indicate that phosphorylation of type 1 and type 3 IP₃Rs decreases Ca²⁺ release (70,71). Detailed studies in β-cells suggest that, when cAMP releases Ca²⁺ from ER, it is likely that the release is through RY receptor rather than through IP₃Rs (19,72).

In β-cells, nitric oxide (NO) releases Ca²⁺ from ER (17,73). There is evidence that this is due to activation of the RY receptors (17). NO reversibly activates RY receptors by oxidation or by poly-S-nitrosylation (74) of critical thiols associated with the channel (5,39,75). The effects of NO (and cGMP-mediated phosphorylation) on IP₃Rs are largely inhibitory (76). Another pathway by which NO activates RY receptor is via cGMP, which in turn activates ADP-ribosyl cyclase, leading to formation of cADPR (77). This pathway is not involved in NO-induced Ca²⁺ release in β-cells since specific inhibitors of cADPR do not block this release (57). NO induced Ca²⁺ release in β-cells has been implicated in exocytosis (17), synchronization of Ca²⁺ signal (73), and apoptosis (78).

The NAD⁺ metabolite cADPR releases Ca²⁺ through a RY receptor–like channel (79). cADPR, however, does not bind to RY receptor or FKBPs associated with the channel (32), and some investigators do not see any effect of cADPR on RY receptors (33). In β-cells, cADPR releases Ca²⁺ from the ER (18), but this has not been a universal finding (12,80,81). Inappropriate experimental conditions as well as low levels of RY receptors in some insulin-secreting cells may partly account for such differences. More recently, Mitchell et al. (22) and Varadi and Rutter (82) used techniques that are sensitive in detecting local Ca²⁺ release from secretory vesicles and ER. Their data provide convincing evidence that micromolar cADPR releases Ca²⁺ from both the secretory vesicles and ER of MIN6 cells. It is not clear whether this is due to specific interaction of cADPR with its channel or due to interaction of cADPR with the adenine-nucleotide binding site of the RY receptors (83). In this regard, information on whether Ca²⁺ release by cADPR could be inhibited by inhibitors, e.g., 7-deaza-8-bromo-cyclic ADP-ribose, would be useful. It is noteworthy that cADPR antagonists cannot inhibit Ca²⁺ release through RY receptors of β-cells, suggesting that cADPR receptor may be different from RY receptors (57). Nevertheless, these recent data, together with the fact that glucose increases cADPR in β-cells, should renew interest in search for the elusive cADPR receptor (18,84).

β-cells act as fuel sensors by virtue of having K_ATP channels (85). However, the cells continue this function even under conditions in which K_ATP channels are clamped by diazoxide (86). It is likely that ATP and some other molecules arising from nutrient metabolism act on other ion channels or exocytotic processes. Molecules that arise from nutrient metabolism and activate RY receptors include ATP, glycolytic intermediates, palmitoyl CoA, and cADPR. Moreover, a physiological alkaline shift of intracellular pH, such as that occurs on glucose stimulation, favors activation of the RY receptor (87). In in vitro experiments, it is possible to demonstrate that glucose releases Ca²⁺ from the ER, suggesting that some metabolites of glucose may favor activation of intracellular Ca²⁺ channels (88).

ATP favors Ca²⁺ release through RY receptors by 1) filling the Ca²⁺ stores, 2) binding Mg²⁺ and thereby reducing cytosolic [Mg²⁺], and 3) allosteric regulation of RY receptor. In the presence of subactivating [Ca²⁺], physiological levels of ATP directly activates RY₂ receptor by acting on the ATP-binding site, whereas ADP acts as a partial agonist (89). In resting β-cells, high cytosolic [Mg²⁺] keeps the RY receptor inhibited. Glucose stimulation reduces cytosolic [Mg²⁺] by increasing ATP production and [ATP] (90) and thereby facilitates RY receptor activation.

Fructose 1,6-diphosphate (FDP), a product of the rate-limiting enzyme phosphofructokinase, activates RY₂ receptor and sensitizes the receptor to other activators (91,92). Consistent with this, FDP stimulates insulin secretion from HIT-T15 cells (27,93). It should be noted that FDP and several other glycolytic intermediates inhibit binding of IP₃ to IP₃Rs, thus making it unlikely that IP₃Rs participate in this response (92).

Glucose metabolism increases level of cytosolic long-chain acyl CoA (94). Such long-chain acyl CoAs sensitize RY₂ receptors of β-cells (22). Palmitoyl CoA sensitizes RY₂ receptors to activation by Ca²⁺ by mechanisms that include relief of Mg²⁺ inhibition of the channel.

Several groups have demonstrated acute increase of cADPR in glucose-stimulated β-cells (84,95), but this is not a universal observation (96). Measurement of cADPR in β-cells seems to be technically difficult. Recent availability of a more convenient and sensitive method for cADPR measurement may be helpful in elucidating the signaling role of cADPR in β-cells (97).

Mitochondrial ATP production is essential for stimulus-secretion coupling in β-cells. An increase in [Ca²⁺]_c increases mitochondrial Ca²⁺ concentration ([Ca²⁺]_m) and enhances mitochondrial ATP production (98). Because ER and mitochondria appear to be in close opposition, some RY receptors may be located close to the mitochondria. Studies in many cells have demonstrated a functional coupling between RY receptors and mitochondria. In β-cells, an important function of RY receptors may be to amplify Ca²⁺ signals by CICR to produce microdomains of high [Ca²⁺]_c, which would act as sources for elevating [Ca²⁺]_m and thus accelerate ATP production.

The effects of Ca²⁺ on cellular processes depend on the subcellular location of the [Ca²⁺]_c increase. Furthermore, the integrity of the [Ca²⁺]_c oscillatory process is important for secretion. Ca²⁺ release through RY receptor may affect secretion positively or negatively. By increasing [Ca²⁺]_c near the secretory sites, it may trigger exocytotic fusion of the secretory granules with the plasma membrane. On the other hand, by increasing [Ca²⁺]_c near K⁺_Ca channels, it can result in hyperpolarization of the β-cell membrane and inhibit secretion. Intermittent release of Ca²⁺ through RY receptor will have an intermittent hyperpolarizing effect, which may not inhibit net secretion and may rather increase it. Persistent release of Ca²⁺ will hyperpolarize for a prolonged period and will lead to inhibition of secretion. Whether Ca²⁺ released through RY receptor is more readily available to the exocytotic sites or to the site of K⁺_Ca channels is not known. In adrenal chromaffin cells, Ca²⁺ released through RY receptors is predominantly for driving the exocytotic processes (29,99). In β-cells, Ca²⁺ released through the RY receptors appear to be available to the exocytotic sites for stimulating secretion (17).

The role of RY receptors can be examined by testing the effects of methylxanthines. Because these tools have side effects, results need to be interpreted with careful reflection. By sensitizing RY receptors to incoming Ca²⁺, methylxanthines enhance CICR. By this means and the consequent effect on K⁺_Ca channels, methylxanthines are expected to terminate a burst earlier and to shorten the duration of the slow waves. By affecting another group of K⁺ channels, i.e., inhibiting the K_ATP channels directly, methylxanthines shorten the interval between the slow waves. The net effect is an increase in the number of bursts per minute. Such effects on electrical activity of β-cells are seen when theophylline is applied to β-cells stimulated by glucose (50,100). Consistent with this, theophylline and caffeine markedly potentiate glucose-induced insulin secretion (101,102). At high concentrations, caffeine inhibits glucose transport (103) but still stimulates insulin secretion (20,104,105). Stimulation of insulin secretion by caffeine and theophylline is tacitly assumed to be solely due to cAMP. However, an examination of the quantitative aspects of secretion by theophylline and cAMP suggests involvement of RY receptors. Thus, in rat islets, in Ca²⁺-free medium and 16.7 mmol/l glucose, 1.4 mmol/l theophylline stimulates insulin secretion, which is three times more than that induced by 1 mmol/l dibutyryl cAMP (106). Imidazole, which does not inhibit islet PDEs (51), but sensitizes RY receptors (44), stimulates secretion (107). Another way to dissect the RY receptor-activating and the PDE-inhibitory effect of methylxanthines is to use RY receptor inhibitors. In rat islets, tetracaine inhibits insulin secretion stimulated by theophylline (106). It should be emphasized that RY₂ receptor can be sensitized by cAMP-dependent phosphorylation (60) as well as by direct binding of methylxanthines to the channel. In either case, the main trigger for activation of RY receptor is still the Ca²⁺ that enters through the voltage-gated Ca²⁺ channels. Thus, despite reservations about nonspecific actions of these drugs, the available data strongly support a role for RY receptors in glucose-induced exocytosis of insulin.

There is now compelling evidence that cAMP-linked incretin hormones stimulate glucose-dependent insulin secretion by mechanisms that include CICR through RY receptors (19,26,66). Secretion is a complex process requiring more than just an elevation of [Ca²⁺]_c. Since Ca²⁺ released through RY receptor is available to both exocytotic sites and the K⁺_Ca channel sites, it is not surprising that both stimulation and inhibition of secretion have been reported with RY receptor ligands. The results depend on the experimental conditions used and it is often so that the experimental protocols employed do not mimic the physiological situation. Thus NO, which activates RY receptor, can stimulate (108) or inhibit (109) insulin secretion. Imidazole, an activator of RY receptors can stimulate (107) or inhibit (110) secretion. Dantrolene, an inhibitor of RY receptor, has also been shown to inhibit (111) or stimulate (112) glucose-induced insulin secretion.

Depolarization-induced [Ca²⁺]_c increase in β-cells is not just due to Ca²⁺ entry through the voltage-gated Ca²⁺ channels. There is evidence that this Ca²⁺ signal is modulated by CICR and uptake of Ca²⁺ into ER (2,3). In patch clamp experiments, Ca²⁺ entry through voltage-gated Ca²⁺ channels may be overestimated. Most voltage clamp experiments in β-cells are done in the presence of very high extracellular [Ca²⁺] and a highly buffered and altered cytosolic environment. In such experiments, unlimited extracellular space permits a large influx of Ca²⁺ through the channel. The [Ca²⁺]_c increase obtained on depolarization in human β-cells is usually not very high (∼300 nmol/l) (88). However, when conditions are created where incoming Ca²⁺ can trigger the RY receptors, a very large increase in [Ca²⁺]_c is seen (2,19,113). It is possible that under physiological conditions, there is relatively small Ca²⁺ entry through the voltage-gated Ca²⁺ channels, which in turn is amplified by CICR.

Even if one accepts that β-cells do not have large amounts of RY receptors, they may be important because of their large conductance and strategic location within the cell. Some RY receptors drain a nonmitochondrial Ca²⁺ pool that does not utilize a thapsigargin-sensitive pump (5). This Ca²⁺ pool may represent the secretory vesicles (22). RY receptors located on secretory vesicles (22,114) will allow a highly localized increase of [Ca²⁺]_c at exocytotic sites promoting exocytotic fusion. CICR may allow brief bout of Ca²⁺ entry through voltage-gated Ca²⁺ channels to generate a wave of [Ca²⁺]_c changes, which is required in the early steps of exocytosis, e.g., vesicle transport to exocytotic sites (115).

Another myth that has been propagated over the years is that glucose alone is the most important stimulus for insulin secretion. In fact, in the physiological range of excursion of glucose concentration, glucose is a rather poor stimulator of insulin secretion from pure β-cells. This is because glucose per se is not very effective in engaging CICR, which can be engaged efficiently if the process is sensitized, for instance, by cAMP-linked hormones (116). It needs to be emphasized that CICR is a multistep process that is facilitated by many factors (2,26). RY receptors, together with the ER-associated Ca²⁺ apparatus, act as a functional unit for co-incidence detection. A critical feature of RY receptor is that its gating is context-dependent (Fig. 1). Thus, optimal amplification of Ca²⁺ signaling occurs only when several conditions are simultaneously satisfied. Such factors include an ER full of Ca²⁺, elevated levels of FDP, cAMP, ATP, NO, low [Mg²⁺], alkaline pH, and others. The amplified Ca²⁺ signals often take the form of regenerative spike-like oscillations and presumably leads to very high local [Ca²⁺]. Many Ca²⁺-dependent processes, including exocytosis, require very high [Ca²⁺]_c, and CICR may be a molecular process to achieve this (117). The significance of this mode of signaling is illustrated by the fact that cAMP-linked incretin hormones, e.g., GLP-1, utilizes this process to stimulate insulin secretion in a typical context-dependent manner.

Knockout mice for each of the RY receptors have been generated. However RY₁- and RY₂-deficient mice die at embryonic stage or after birth. It may be useful to have mice in which RY receptors are knocked out specifically in the β-cells. Knockout of CD38, which catalyzes synthesis of cADPR, impairs glucose-induced formation of cADPR, elevation of [Ca²⁺]_c, as well as insulin secretion (95).

From studies in various animal models of type 2 diabetes, it appears that impairment of mechanisms that ensure optimal secretory response of β-cells is an important component in the pathogenesis of type 2 diabetes (118). Impaired function of intracellular Ca²⁺ pools of β-cells has been described in several rodent models of type 2 diabetes (4,119). From the evidence discussed above, it seems fair to speculate that RY receptor-mediated CICR may be one of the many mechanisms that normally amplify insulin secretion following a trigger by nutrients (86). Impaired glucose metabolism in β-cells will fail to trigger insulin secretion because of impaired ATP production. At the same time, such impairment may also impair CICR because of impaired production of molecules such as FDP, long-chain acyl CoA, cADPR, cAMP, etc.—molecules that normally sensitize the RY receptors. One of the mechanisms utilized by cAMP-linked incretin hormones to amplify insulin secretion is clearly the RY receptor-mediated CICR, and such hormones are potential antidiabetic drugs (19). RY receptor may be relevant to the pathogenesis of type 1 diabetes also since this channel is a prototypic redox-sensitive Ca²⁺ channel and may thus mediate damaging actions of NO and free radicals (78).

The existence of RY receptors in β-cells is well documented. Any controversies about RY receptors of β-cells, including those involving cADPR, may be intrinsic to the complex mode of regulation of these channels and difficulties associated with the usage of the pharmacological tools. One example of this is the distinct context-dependence of activation of RY receptors. Despite its low abundance, RY receptors of β-cells may play important roles in stimulus-secretion coupling by virtue of their strategic locations within the cell, their ability to mediate CICR in a context-dependent manner, and their large conductance. Being Ca²⁺-activated ion channels, RY receptors have the unique ability to interact with neighboring Ca²⁺ channels and thereby amplify Ca²⁺ signals. Such amplification is engaged when the channel is sensitized by a set of messenger molecules generated from nutrient metabolism or ligand-binding. RY receptors are thus suitable for integration of signaling, co-incidence detection, and context-dependent signaling for insulin secretion. Secretagogues may modulate insulin secretion by affecting CICR mediated by RY receptor. Such processes may be involved in amplification of insulin secretion and may be a target for development of therapeutic agents that may stimulate insulin secretion in a context-dependent manner.

Explicit recognition of the fact that β-cells possess a robust mechanism for amplification of Ca²⁺ signaling may advance our understanding of stimulus-secretion coupling. Rigorous attention to experimental protocols, a clear understanding of the mechanism of action of RY receptor ligands as well as context-dependence of activation of these channels will help reduce controversies and move the field forward. Nevertheless, the issues involved need to be critically examined. In this respect, quantitative data on the relative densities of RY receptors and IP₃Rs and their relative contribution in Ca²⁺ signaling in β-cells, especially in human β-cells, will be helpful. β-cells may be heterogeneous in terms of the level of RY receptors and IP₃Rs. Hormonal and metabolic factors may alter RY receptor level in β-cells. The nature of coupling of voltage-gated Ca²⁺ channels to RY receptors, local exchange of signals between these channels, and their roles in local Ca²⁺ signaling need to be elucidated. Ca²⁺ release through IP₃Rs may also be a trigger for RY receptor-mediated CICR. In β-cells, this possibility is supported by the observation that dantrolene inhibits [Ca²⁺]_c response by cholinergic agonists in β-cells (88). Finally, the role of RY receptor-mediated CICR as a general mechanism for amplification of insulin secretion and factors that affect this process need to be explored.

Financial support was obtained from the Juvenile Diabetes Research Foundation International, Swedish Medical Research Council (K2001-32X-13469-02B), The Swedish Council for Natural Science Research, Swedish Fund for Research Without Animal Experiments, and the Karolinska Institutet.

M.S.I. is a recipient of a career development award from the Juvenile Diabetes Research Foundation International.

The author is grateful to Drs. C. J. Barker, J. D. Bruton, and T. Nilmon for comments on the manuscript, and gratefully acknowledges anonymous reviewers for the useful comments.