Cell Side-specific Sensitivities of Intracellular Ca2+ Stores for Inositol 1,4,5-Trisphosphate, Cyclic ADP-ribose, and Nicotinic Acid Adenine Dinucleotide Phosphate in Permeabilized Pancreatic Acinar Cells from Mouse*

In pancreatic acinar cells hormonal stimulation leads to a cytosolic Ca2+ wave that starts in the apical cell pole and subsequently propagates toward the basal cell side. We used permeabilized pancreatic acinar cells from mouse and the mag-fura-2 technique, which allows direct monitoring of changes in [Ca2+] of intracellular stores. We show here that Ca2+ can be released from stores in all cellular regions by inositol 1,4,5-trisphosphate. Stores at the apical cell pole showed a higher affinity to inositol 1,4,5-trisphosphate (EC50 = 89 nm) than those at the basolateral side (EC50 = 256 nm). In contrast, cADP-ribose, a modifier of Ca2+-induced Ca2+ release, and nicotinic acid adenine dinucleotide phosphate (NAADP) were able to release Ca2+ exclusively from intracellular stores located at the basolateral cell side. Our data agree with observations that upon stimulation Ca2+ is released initially at the apical cell side and that this is caused by high affinity inositol 1,4,5-trisphosphate receptors. Moreover, our findings allow the conclusion that in Ca2+ wave propagation from the apical to the basolateral cell side observed in pancreatic acinar cells Ca2+-induced Ca2+ release, modulated by cADP-ribose and/or NAADP, might be involved.

In pancreatic acinar cells regulation of spatio-temporal changes in the free cytosolic Ca 2ϩ concentration ([Ca 2ϩ ] i ) plays an essential role in receptor-mediated stimulation of protein secretion. Hormonal stimulation of the cells leads to production of inositol 1,4,5-trisphosphate (IP 3 ) 1 (1), which is followed by a cytosolic Ca 2ϩ signal starting in the apical cell pole (2,3). This cytosolic Ca 2ϩ signal can either remain localized at the apical side in the form of Ca 2ϩ oscillations, or it can spread toward the basolateral cell side in form of a "Ca 2ϩ wave" (3,4).
The mechanism underlying generation and propagation of the Ca 2ϩ wave is still a matter of debate. Despite IP 3 generation by activation of phospholipase C located in the basolateral plasma membrane, the initial rise in [Ca 2ϩ ] i takes place at the opposite side of the cell. From results obtained by measurments of the [Ca 2ϩ ] i it was concluded indirectly that stores in the apical cell pole possess receptors with higher sensitivities for IP 3 than stores in the basolateral cell pole (3,(5)(6)(7). A structural basis for the assumption of different IP 3 sensitivities may be the existence of different IP 3 receptor isoforms in pancreatic acinar cells (types I, II, and III) (8,9). In IP 3 binding studies it has been shown that the type II IP 3 receptors have a higher IP 3 affinity than the type III receptors (10,11) and that the type II IP 3 receptors are located exclusively in the apical cell pole (12,13). Taken together, these findings suggest that Ca 2ϩ stores located in the apical region are equipped mainly with the higher affinity type II IP 3 receptors; however, these predicted sensitivity differences between the stores in the two cell poles have not yet been quantified as changes of the [Ca 2ϩ ] in the Ca 2ϩ stores itself which allow concentration-response measurements.
Recently, it has been shown that the speed of the Ca 2ϩ wave that propagates from the apical to the basolateral cell side depends on the type of stimulating hormone such as acetylcholine, bombesin, or cholecystokinin (14). Because ryanodine receptors are present in pancreatic acinar cells (15) and participate in the generation of Ca 2ϩ signals (16 -18), a potential regulator of hormone-induced Ca 2ϩ signals could be cADPribose, which modifies the ability of ryanodine receptors to mediate Ca 2ϩ -induced Ca 2ϩ release in many cell systems (19). Another candidate that may play a role in Ca 2ϩ release is nicotinic acid adenine dinucleotide phosphate (NAADP) (20,21). It has been shown that both cADP-ribose and NAADP can induce Ca 2ϩ oscillations in pancreatic acinar cells (22,23). This led to the hypothesis that in addition to IP 3 -induced Ca 2ϩ release, part of stored Ca 2ϩ is released via other mechanisms, probably by hormone-induced elevation of the cellular cADPribose and/or NAADP concentration.
Using fluorescence methods we have shown recently that Ca 2ϩ stores in distinct regions of pancreatic acinar cells in primary culture release Ca 2ϩ in response to cADP-ribose and IP 3 from different and in part from common Ca 2ϩ stores (24). Primary cultured acinar cells lack the typical polarized organization of a differentiated acinar cell. We therefore used freshly isolated acinar cells from mouse in the present study to obtain more information about the localization of functionally different Ca 2ϩ stores.
By use of the mag-fura-2 technique, which allows monitoring of changes in [Ca 2ϩ ] of intracellular Ca 2ϩ stores ([Ca 2ϩ ] store ) from permeabilized cells (25) we demonstrate directly here that IP 3 added to pancreatic acinar cells at low concentrations (0.1-1 M) releases Ca 2ϩ mainly from stores located in the apical cell pole. In contrast, cADP-ribose and NAADP release Ca 2ϩ exclusively from basolateral stores. These findings may help to explain the typical spatio-temporal pattern of the cytoplasmic Ca 2ϩ signal after hormonal stimulation of pancreatic acinar cells.

EXPERIMENTAL PROCEDURES
Cell Preparation-Adult male CD-1 mice were killed by cervical dislocation, and the pancreas was rapidly removed and placed into preparation buffer (composition in mM: 130 NaCl, 4.7 KCl, 1.3 CaCl 2 , 1 MgCl 2 , 1.2 KH 2 PO 4 , 10 HEPES, 10 glucose, 0.2% (w/v) albumin, and 0.01% (w/v) trypsin inhibitor; pH adjusted to 7.4 with NaOH). After removing adherent blood vessels and fat tissue we injected the pancreas with 1 ml of the preparation buffer supplemented with collagenase type V (30 units/ml) and incubated it for 10 min at 37°C. After enzymatic digestion the tissue was cut into small pieces of ϳ1 mm 3 and pipetted gently through tips of decreasing diameter to dissociate the tissue into cells. Cells were then centrifuged for 2 min at 30 ϫ g, and the pellet was resuspended in preparation buffer without collagenase. With this isolation procedure single cells as well as small clusters consisting of two up to five cells were obtained. 2) at 37°C to allow pore formation and to wash out cytoplasmatic dye.
During the experiments cells were perfused with intracellular buffer at room temperature. 0.5 mM ATP was always added to the perfusion solution to prevent depletion of intracellular Ca 2ϩ stores by insufficient activity of the endoplasmic reticulum Ca 2ϩ -ATPase. Ca 2ϩ Imaging-The [Ca 2ϩ ] store was monitored using the fluorescent low affinity Ca 2ϩ indicator mag-fura-2 (Molecular Probes) in combination with a microscopic imaging system (T.I.L.L.-Photonics, Germany) coupled to an inverted microscope (Carl Zeiss, Germany). Cells were excited alternately at 345 and 380 nm for 30 -50 ms, and the resultant emission was collected above 510 nm every 10 s. Experiments were done at room temperature (21-24°C), and cells were perfused continuously with intracellular buffer with different test compounds as indicated in the figures. To estimate the free [Ca 2ϩ ] store from the measured fluorescence ratio we used the Grynkiewicz relation (26) with a determined K D of 17 M for the Ca 2ϩ binding of mag-fura-2.
Confocal Microscopy-To check to what extent the subcellular organization of pancreatic acinar cells is perturbed by streptolysin O, permeabilized cells were loaded with the intracellular membrane marker 3,3Ј-dihexyloxacarbocyanine iodide (DiOC6, 1 g/ml for 2 min) (27) in the preparation buffer. After loading, cells were washed in preparation buffer, and confocal pictures of the cells were taken with a confocal microscope (Bio-Rad MRC1024) with an excitation wavelength of 488 nM before and after permeabilization.

Experimental Approach
We used freshly isolated pancreatic acinar cells from mice loaded with the Ca 2ϩ -sensitive dye mag-fura-2. This dye not only stains the cytoplasm of the cell but is also taken up into intracellular compartments such as the endoplasmic reticulum. To estimate [Ca 2ϩ ] store not influenced by the cytoplasmic dye and to test the effects of IP 3 , cADP-ribose, and NAADP on these stores, cells had to be permeabilized with streptolysin O before each experiment as described in "Experimental Procedures." Using this method about 30% of the cells lost cytoplasmic dye, demonstrating successful permeabilization. The remaining cells that had been permeabilized showed a discrete distribution of fura dye with a high intensity at the apical cell Zymogen granules are present mainly on the apical cell side (left). DiOC6 fluorescence, which stains mitochondria and endoplasmic reticulum, is located mainly in a "ring" between the apical and basolateral cell poles and also near the plasma membrane of the basolateral cell side (right). Panel B, a triplet of pancreatic acinar cells after permeabilization (left, light microscopic picture; right, DiOC6 fluorescence). Zymogen granules are still present in the apical cell pole. DiOC6 fluorescence seems a bit more diffuse compared with unpermeabilized cells (panel B, right) but still forms a ring around the lumen and is also present near the plasma membrane at the basolateral cell pole (seen best in the top cell of the triplet). pool and a low intensity at the basolateral cell side (not shown). Light microscopy with differential interference contrast showed that the nucleus and the zymogen granules of these cells did not redistribute after permeabilization (Fig. 1, A and B). Confocal microscopy of DiOC6-stained cells before and after permeabilization verified that even the stained membranes (of mitochondria and endoplasmic reticulum) were largely kept in position ( Fig. 1, A and B).

IP 3 -induced Ca 2ϩ Release
Effect of IP 3 -When permeabilized cells were analyzed we found that a concentration of 5 M IP 3 induced a maximal Ca 2ϩ release in all regions of the cell. Further release of Ca 2ϩ could only be achieved by the application of the Ca 2ϩ ionophore ionomycin (10 M), by which Ca 2ϩ stores were completely depleted ( Fig. 2A). The maximal IP 3 -induced Ca 2ϩ release as estimated by the decrease in the fluorescence ratio was 91.8 Ϯ 7.0% of the ionomycin effect (n ϭ 8 experiments, 43 cells; Fig.  2B).
The so-called "quantal Ca 2ϩ release," which means that an increase in submaximal [IP 3 ] leads to stepwise release of discrete amounts of Ca 2ϩ , could also be observed by direct monitoring of [Ca 2ϩ ] store ( Fig. 2A).
The IP 3 effect was reversible. Washing the cells with intracellular buffer, subsequently to IP 3 -induced Ca 2ϩ release, resulted in complete refilling of the stores (data not shown; n ϭ 8 experiments, 25 cells).
Effects of IP 3 in Different Subcellular Regions-The use of video imaging allowed us to analyze simultaneously Ca 2ϩ release from different regions within a single cell. We found that all regions of a cell responded to 5.0 M IP 3 with a maximal release of Ca 2ϩ (Fig. 3A). However, the fluorescence ratio be-fore stimulation was always higher in the basolateral cell side compared with the apical cell side. This does not essentially mean that the [Ca 2ϩ ] store is higher basolaterally. It is more probable that the typical spatial distribution of organelles (high proportion of endoplasmic reticulum at the basolateral side versus high proportion of zymogen granules at the apical cell pole) is responsible for these different fluorescence ratios in the opposite cell poles. Assuming that compartments in the apical cell region take up fluorescence dye but no Ca 2ϩ , fluorescence from these stores will quench part of the mag-fura-2 signal from functional Ca 2ϩ stores in the apical cell pole. This problem of the so-called "silent compartments" has been considered earlier (28). Thus it did not seem appropriate to use the ratio signal itself for estimation of different IP 3 sensitivities in different cell regions. We therefore standardized the data for the IP 3 -induced decrease in the mag-fura-2 ratio by comparison with the maximal ratio decrease induced by ionomycin in the respective cell region (set at 100%). By this procedure it became obvious that a low concentration of IP 3 showed different efficiencies in the apical and in the basolateral side of the cell. At a submaximal [IP 3 ] of 0.1 M, Ca 2ϩ release was 55.3 Ϯ 23.8% of the maximal ionomycin-inducible ratio decrease in the apical side of the cell, whereas it was only 20.51 Ϯ 9.8% (Fig. 3, B-D) in the basolateral cell side.
Based on this result we tested different concentrations of IP 3 between 0.1 and 5.0 M IP 3 and found that the dose-response curve for IP 3 was shifted to the left for the apical compared with that for the basolateral cell regions (Fig. 3C). The halfmaximal effective concentration for IP 3 in the apical cell side was 89 nM compared with 256 nM in the basolateral cell side (n ϭ 10).

cADP-ribose-and NAADP-induced Ca 2ϩ Release
Effect of cADP-ribose and NAADP on the Whole Cell-The distribution pattern of different IP 3 -sensitive stores described in the last paragraph led us to assume that start and propagation of hormone-induced Ca 2ϩ waves from the apical to the basolateral cell sides (14,17,29,30) are related to the functions of Ca 2ϩ stores in these cell regions. In particular we addressed the question of whether cADP-ribose-and NAADP-induced Ca 2ϩ release from stores in pancreatic acinar cells could be involved in propagation of the Ca 2ϩ wave toward the basolateral cell side.
We therefore first superfused permeabilized cells with up to 100 M cADP-ribose, a concentration known to be supermaximal for Ca 2ϩ release from the stores of rat pancreatic acinar cells in primary culture (24). However, cADP-ribose was ineffective in freshly prepared cells.
The inability of cADP-ribose to release Ca 2ϩ from permeabilized cells could be caused by the loss of a cytoplasmic factor necessary for the releasing process. We therefore tried application of 5 g/ml calmodulin, described to be a cofactor for the cADP-ribose effect (31). But this did not result in cADP-riboseinduced Ca 2ϩ release from the stores either. A pharmacological tool often used to sensitize the Ca 2ϩ -induced Ca 2ϩ release mechanism is caffeine. We tested cADP-ribose in the presence of caffeine, which alone had no effect on Ca 2ϩ release from intracellular Ca 2ϩ stores (n ϭ 3 experiments, 14 cells). At 50 M cADP-ribose and in the presence of 10 mM caffeine, Ca 2ϩ was released from the stores (n ϭ 11 experiments, 39 cells; Fig.  4A), and the effect was maximal (33 Ϯ 15% of the ionomycininduced ratio decrease). Neither an increase in caffeine nor an increase in [cADP-ribose] further induced Ca 2ϩ release. Therefore the following cADP-ribose experiments were performed in the presence of 10 mM caffeine without further mention in this text.
Similar to cADP-ribose, NAADP was able to release Ca 2ϩ from permeabilized pancreatic acinar cells (Fig. 4B). In a concentration of 50 nM the ratio decrease that could be induced by NAADP was 35 Ϯ 18% of the ionomycin-induced ratio decrease (n ϭ 4 experiments, 8 cells). Increasing the NAADP concentration to 100 nM did not change this result (data not shown).
The effects of cADP-ribose-and NAADP-induced Ca 2ϩ release were reversible. Removal of cADP-ribose or NAADP from the intracellular buffer resulted in refilling of the stores (data not shown).
Comparing the effect of cADP-ribose, NAADP, and IP 3 on Ca 2ϩ release it became evident that IP 3 released Ca 2ϩ with higher efficiency than either cADP-ribose or NAADP. After a supermaximal concentration of IP 3 (5.0 -10 M), the addition of 50 M cADP-ribose or 50 nM NAADP did not result in a further Ca 2ϩ release (Fig. 4, C and D). However, the addition of supermaximal doses of IP 3 subsequently to supermaximal effective concentrations of either cADP-ribose or NAADP further decreased the [Ca 2ϩ ] store (Fig. 4, A and B). Therefore, although IP 3 is additive to cADP-ribose and NAADP, these two substances are not additive to each other. As shown in Fig. 4E application of NAADP subsequently to cADP-ribose did not increase Ca 2ϩ release further. This led us to conclude that there are at least two types of Ca 2ϩ stores in pancreatic acinar cells: one type with IP 3 receptors and another one with all three IP 3 , cADP-ribose, and NAADP receptors. Heparin, an IP 3 receptor antagonist, should discriminate between both IP 3 -and cADP-ribose-induced Ca 2ϩ release. As shown in Fig. 5, IP 3 -induced Ca 2ϩ release could be completely inhibited by 100 g/ml heparin (n ϭ 4 experiments, 22 cells; Fig. 5, A-C), whereas cADP-ribose-induced release of Ca 2ϩ was unaffected by heparin (n ϭ 3 experiments, 13 cells; Fig. 5, B and C).
Effect of cADP-ribose and NAADP at the Subcellular Level-As was shown in the second last paragraph, Ca 2ϩ stores highly sensitive to IP 3 are located in the apical cell region, where the hormone-induced Ca 2ϩ wave starts. We then analyzed discrete regions in a cell with respect to their Ca 2ϩ releasing properties in response to cADP-ribose and NAADP. We found that release of Ca 2ϩ in response to the application of 50 M cADP-ribose and 50 nM NAADP occurred only from basolateral Ca 2ϩ stores (Fig. 6, A and B). In a series of 10 similar experiments with 37 cells for cADP-ribose and 4 experiments with 8 cells for NAADP, we never observed any Ca 2ϩ release from stores at the apical side of the cell. We therefore assume that the distribution of Ca 2ϩ stores that are highly sensitive to IP 3 in the apical cell pole and which respond to cADP-ribose and NAADP in the basolateral cell area could explain the start and propagation of Ca 2ϩ waves from the apical to the basolateral cell side.

DISCUSSION
Ca 2ϩ stores play an important role in the regulation of free [Ca 2ϩ ] i , a trigger for different cellular events. In pancreatic acinar cells an increase in [Ca 2ϩ ] i in the apical cell pole is necessary for fusion of zymogen granules with the apical plasma membrane and for opening of Cl Ϫ channels, which leads to exocytosis of proteins and secretion of Cl Ϫ , respectively. An increase in [Ca 2ϩ ] i at the basolateral cell side leads to opening of cation channels, which results in Na ϩ influx and in sustained NaCl and water secretion from the cell at the apical side (32). Opening and closing of these Ca 2ϩ -sensitive channels is coordinated spatially and temporally by the Ca 2ϩ signal itself (push-pull model (2)). Hypotheses concerning this phenomenon predict different Ca 2ϩ stores (i.e. different affini- FIG. 5. Effect of heparin on cADP-ribose-and IP 3 -induced Ca 2؉ release. Panel A, ratio signal of Ca 2ϩ stores from whole pancreatic acinar cell after permeabilization. Conditions are indicated by horizontal bars. 100 g/ml heparin inhibits 10 M IP 3 -induced Ca 2ϩ release in permeabilized cells (representative experiment of four similar ones with a total of 22 cells). Application of heparin in the presence of IP 3 led to refilling of the Ca 2ϩ stores (0.5 mM ATP was present in the bath solution). Application of heparin before IP 3 prevented Ca 2ϩ release from the stores. iono., ionomycin. Panel B, heparin did not inhibit 50 M cADP-ribose (cADPr)-induced Ca 2ϩ release compared with control cells (33 Ϯ 15% of the ionomycin-induced effect). In contrast, IP 3 was able to induce Ca 2ϩ release after heparin had been washed away. Shown is a representative experiment (n ϭ 3, 13 cells). Panel C, data of all experiments with the protocols shown in panels A and B are summarized. Mean values Ϯ S.D. of IP 3 (n ϭ 4, 22 cells) and cADP-ribose effects (n ϭ 3, 13 cells) in the presence of heparin as normalized to the maximal effect without heparin (100%) are given. cADP-ribose was given together with 10 mM caffeine. The heparin effect on IP 3 -induced Ca 2ϩ release was highly significant (p Ͻ 0.001), whereas the heparin blockade of cADP-ribose-induced Ca 2ϩ release was insignificant (p Ͼ 0.1). ties for agonists or different receptor densities) within a single cell (5)(6)(7)33).

FIG. 6. Spatial pattern of cADP-ribose-and NAADP-induced
In the present study we have characterized different Ca 2ϩ pools at the apical and basolateral cell side of permeabilized pancreatic acinar cells from mouse. Permeabilization of the cells was necessary for two reasons: first, cytoplasmic magfura-2 dye had to be removed from the cytoplasm; and second, the cytoplasmic region of the cell had to be accessible for membrane-impermeant compounds such as IP 3 , cADP-ribose, NAADP, and heparin. As indicated in Fig. 1 and also suggested by others (34), the permeabilization procedure did not change the distribution of the major compartments of the cell (nucleus, mitochondria, zymogen granules) dramatically. However it is likely to assume that permeabilization favors the loss of some smaller cytoplasmic constituents that are also necessary for a normal Ca 2ϩ release process. This may be why we had to use 50 M cADP-ribose, which is a high concentration compared with other systems (31) to obtain maximal effects. We cannot exclude the possibility, however, that the integrity of the endoplasmic reticulum had been disturbed by permeabilization.
Recently it has been suggested that in pancreatic acinar cells the endoplasmic reticulum is continuous (35,36). This would make determination of discrete Ca 2ϩ -releasing structures difficult if not impossible because Ca 2ϩ release at any site from a tubular system with a continuous space should finally empty it. This was obviously not the case in our experiments. But even if the endoplasmic reticulum should have been disconnected by permeabilization our main conclusion that in different cell regions the distribution of IP 3 , cADP-ribose, and NAADP re-ceptors is different, remains unaffected.
We conclude from our studies that Ca 2ϩ pools highly sensitive for IP 3 but not for cADP-ribose or NAADP are located in the apical cell pole, whereas Ca 2ϩ pools less sensitive for IP 3 and sensitive for cADP-ribose and NAADP are located at the basolateral cell side. In an earlier study on the distribution of IP 3 receptors it had been concluded that Ca 2ϩ release in response to IP 3 is spatially homogeneous in rat pancreatic acinar cells (34). Here we used mouse cells that differ from rat pancreatic acinar cells in some features concerning Ca 2ϩ homoeostasis (37).
Because of limited spatial resolution we could not identify the structural source for Ca 2ϩ release in the apical cell pool. Gerasimenko et al. (38) have demonstrated that isolated zymogen granules can release Ca 2ϩ in response to both IP 3 and cADP-ribose. This result had been questioned by Yule et al. (13), who could not find IP 3 receptors in highly purified granular preparations. Because we could not detect any cADPribose-induced Ca 2ϩ release from the apical cell side where zymogen granules are located, the present study does not support the observation of Gerasimenko et al. on cADP-riboseinduced Ca 2ϩ release from zymogen granules (38). Spatial resolution of our technique does not allow a decision as to whether IP 3 , cADP-ribose, and NAADP, which are all effective in the basolateral cell side, release Ca 2ϩ from one structural pool or if there are different pools. However, our data, which show that the actions of NAADP and cADP-ribose are not additive, argue for the hypotheses that receptors for both substances are present in the same basolaterally located pools. The finding that IP 3 alone could induce maximal Ca 2ϩ release, which could not be enhanced by either cADP-ribose or NAADP, suggests that in addition to Ca 2ϩ pools only sensitive to IP 3 in the apical pool, basolateral located pools contain all three types of receptor, for IP 3 , cADP-ribose, and NAADP.
Previous observations that hormone-induced Ca 2ϩ release starts in the apical cell pole (2,3,5,14,29,30) and is followed by spreading of the Ca 2ϩ signal to the basolateral cell pole via Ca 2ϩ -induced Ca 2ϩ release can now be interpreted more substantially by our present data. An arrangement of IP 3 -and cADP-ribose-and/or NAADP-sensitive Ca 2ϩ pools in sequence should explain initiation of the apical Ca 2ϩ signal with low [IP 3 ] and the spreading of the Ca 2ϩ wave to the basolateral cell side when [IP 3 ] increases and cADP-ribose and/or NAADP is produced probably during hormonal stimulation. IP 3 receptors act as a trigger for the apical Ca 2ϩ signal, whereas cADP-ribose and NAADP receptors are amplifiers in the Ca 2ϩ wave propagation. This view is in part similar to the conclusion of Cancela et al. (39), who claim that acetylcholine initially activates IP 3 receptors and that this signal is modified by cADP-ribose and NAADP. On the other hand these authors conclude from their data that the initial Ca 2ϩ release in the apical cell pole in response to another secretagog, cholecystokinin, is mediated by NAADP and not by IP 3 (39). Even though no data are available to show hormone-induced production of NAADP in pancreatic acinar cells so far, our present study does not support the idea that NAADP receptors are present in the apical cell pole.
In summary we speculate that IP 3 receptors in Ca 2ϩ stores with decreasing sensitivities for IP 3 toward the basolateral cell side lead to a Ca 2ϩ response from pool to pool in form of a Ca 2ϩ wave in the direction opposite that of the IP 3 gradient. This Ca 2ϩ wave is modulated by cADP-ribose-and NAADP-induced Ca 2ϩ release. Because both IP 3 receptors and cADP-ribose receptors are regulated by [Ca 2ϩ ] i (19,40,41) and the IP 3 receptor in addition by the [Ca 2ϩ ] store (41), local changes in [Ca 2ϩ ] in both the stores and in the cytoplasm because of IP 3 -, NAADPand cADP-ribose-induced Ca 2ϩ release should also contribute FIG. 7. Model for the localization of IP 3 -, NAADP-, and cADPribose-sensitive Ca 2؉ stores in pancreatic acinar cells. Hormoneinduced activation of phospholipase C (PLC) results in production of IP 3. This IP 3 diffuses from the basolateral cell pole to the apical cell side where Ca 2ϩ pools most sensitive to IP 3 are located (trigger zone). Ca 2ϩ that is released from this trigger zone diffuses to the opposite cell side (Ca 2ϩ wave). The speed of this wave can be accelerated by different processes. During continuous activation of phospholipase C the [IP 3 ] increases and releases Ca 2ϩ also from lower sensitive Ca 2ϩ stores on the way to the basolateral cell pole. Ca 2ϩ -induced Ca 2ϩ release (CICR) occurs when the wave reaches more basolaterally located Ca 2ϩ stores. The magnitude of Ca 2ϩ -induced Ca 2ϩ release and with this also the speed of the Ca 2ϩ wave may be modulated by cADP-ribose and/or NAADP released into the cytoplasm probably during hormonal stimulation. PIP 2 , phosphatidylinositol bisphosphate; G, G protein; DAG, diacylglycerol.
to the regulation of Ca 2ϩ signals in pancreatic acinar cells.
The knowledge of different Ca 2ϩ pools in the cell opens the question of whether activation of capacitative Ca 2ϩ entry (42) is activated by one (solely IP 3 -sensitive) or both (IP 3 -sensitive as well as IP 3 -, cADP-ribose-and NAADP-sensitive) types of stores. Further studies on the characteristics of different Ca 2ϩ pools in the cell could help to identify and to characterize the Ca 2ϩ pool that generates the message to activate capacitative Ca 2ϩ entry.