The role of Ca²⁺ in the pathophysiology of pancreatitis

Actions of physiological stimulants and their intracellular messengers

Although this review article predominantly deals with pancreatic acinar cells, the earliest mechanistic work on the role of Ca²⁺ in controlling exocrine secretion was carried out on salivary glands. Douglas & Poisner (2012), in experiments on perfused cat submandibular (submaxillary) glands, discovered that the presence of external Ca²⁺ was required to sustain acetylcholine (ACh)-evoked salivary secretion, but they also noted that the requirement for external Ca²⁺ was not as acute as in the case of endocrine glands, such as the adrenal medulla and the neurohypophysis, where hormone secretion is totally and immediately dependent on the presence of Ca²⁺ in the extracellular solution. Indeed, ACh continues to evoke salivary fluid secretion for quite some time after the removal of Ca²⁺ from the perfusion fluid. Douglas & Poisner (2012) understood that, ‘Calcium has clearly some important role in the stimulant action of ACh on submaxillary salivary secretion. But so little is known of the action of ACh on secretory cells or of the details of the secretory process it initiates that we can only speculate on the nature of this role.’.

Selinger et al. (2003) were the first to demonstrate the existence of ATP-dependent Ca²⁺ uptake into a microsomal fraction from parotid and submaxillary glands. A few years later, we showed that ACh and adrenaline evoked a marked increase in the rate of release of ⁴⁵Ca²⁺ from intracellular stores in preloaded perfused cat submandibular glands, and proposed – correctly as it turned out – that ACh (and adrenaline) acts by releasing Ca²⁺ from the endoplasmic reticulum (ER) (Nielsen & Petersen, 2013). Shortly thereafter, similar results were obtained in studies on superfused mouse and rat pancreatic fragments (Case & Clausen, 2000; Matthews et al. 1994). For many years thereafter, it was a major discussion point at all meetings in the field how interaction between a neurotransmitter and a hormone with a receptor site on the outside of the plasma membrane (Iwatsuki & Petersen, 1996b; Philpott & Petersen, 2001) could evoke Ca²⁺ release from an intracellular source. This key question was finally answered by experiments on permeabilized pancreatic acinar cells and isolated microsomal vesicles, in which it was shown that inositol 1,4,5-trisphosphate (IP₃) released Ca²⁺ from the ER (Streb et al. 1970, 2005).

The key experimental evidence that led to the now well-known concept of hormone- or neurotransmitter- evoked intracellular Ca²⁺ signalling by the release of Ca²⁺ from the ER came from experiments on exocrine gland cells, but there were also some important complicating issues that specifically arose from further work on these cells. In the earliest imaging studies, it was shown that the cytosolic Ca²⁺ signals evoked by ACh, in both pancreatic and lacrimal acinar cells, always started in the apical (granular) part of the cells (Kasai & Augustine, 2002; Toescu et al. 1992b). More importantly, it then became clear that, during sustained stimulation with a low (and therefore most probably physiological) concentration of ACh, or intracellular perfusion with IP₃, the cytosolic Ca²⁺ signals were confined to the apical granular area and did not spread out towards the base (Kasai et al. 1977; Thorn et al. 1992a; Gerasimenko et al. 2006bb). Moreover, application of ACh specifically at the base of the cell initiated a cytosolic Ca²⁺ signal at the opposite end of the cell, at the apical pole (Thorn et al. 1992a; Ashby et al. 2003). A detailed study of the distribution of organelles in living pancreatic acinar cells, confirming the general notion from many electron microscopic studies (Bolender, 1980), showed that the bulk of the ER was localized in the basolateral area, whereas the apical part of the cells was dominated by the secretory (zymogen) granules (ZGs). Nevertheless, the apical granular-rich area contained thin elements of ER that penetrated all the way to the apical membrane (Gerasimenko et al. 1996a). The conclusion from the early imaging studies (Kasai et al. 1977; Thorn et al. 1992a), namely that the apical area of the acinar cells contained the highest concentration or the most sensitive IP₃ receptors (IP₃Rs), was confirmed by Nathanson et al. (2013) and Lee et al. (2011), who showed by immunochemistry that the IP₃Rs were indeed concentrated in the apical region. The apical Ca²⁺ signals are physiologically important as they activate Ca²⁺-sensitive Cl⁻ channels, which are exclusively present in the apical membrane and are crucial for acinar fluid secretion (Park et al. 2011), as well as the exocytotic enzyme secretion, which can be monitored by capacitance measurements (Maruyama et al. 2005; Maruyama & Petersen, 2009).

Given that it was well established that IP₃ elicits Ca²⁺ release from the ER (Berridge, 2012), it might be regarded as surprising that the physiological cytosolic Ca²⁺ signals should occur in an area of the acinar cells that contains relatively little ER. With the discovery that IP₃ could release Ca²⁺ from single isolated ZGs (Gerasimenko et al. 2003a), it became necessary to consider the possibility that the physiological apical Ca²⁺ signals could have arisen from release from ZGs rather than from the ER. Although the finding that IP₃ could release Ca²⁺ from ZGs was regarded initially with great suspicion, similar results were obtained from studies of isolated secretory granules from tracheal goblet cells (Nguyen et al. 1994) and mast cells (Queseda et al. 1979). There is now no longer any doubt that Ca²⁺ can be released via IP₃Rs from an acid non-ER store, dominated by ZGs, as this has been documented in great detail (Gerasimenko et al. 2006a, 2013), but this does not necessarily mean that Ca²⁺ release from such stores plays an important role in normal stimulus–secretion coupling.

The so-called Ca²⁺ tunnel experiments (Mogami et al. 1973) showed that the ER could be refilled, after ACh-elicited emptying, from a point source at the base of an isolated acinar cell by a thapsigargin-sensitive process, and that re-stimulation with ACh would again cause a primary [Ca²⁺]_i rise in the apical pole, more than 10 μm away from the Ca²⁺ entry point at the base, and without any discernible rise in [Ca²⁺]_i during the refilling period. A few years later, we were able to demonstrate directly that a high ACh concentration caused a major reduction in [Ca²⁺] in the intracellular stores in the basal part of the cells (dominated by ER), but not in the apical part (dominated by ZGs), in spite of the fact that [Ca²⁺]_i rose primarily in the apical pole (Park et al. 2005; Petersen et al. 2008). It was also shown that the whole of the ER, including the fine extensions and terminals in the apical pole, is functionally connected, and that Ca²⁺ diffuses easily inside the lumen of the ER (Park et al. 2005; Petersen et al. 2008). These studies indicated that physiological stimuli, such as ACh, primarily release Ca²⁺ from the ER, and that the bulk of the Ca²⁺ comes from basal stores. However, the primary Ca²⁺ release into the cytosol occurs in the apical area because this is where the IP₃Rs are concentrated. Ca²⁺ tunnelling through the ER works because, in the pancreatic acinar cells, the Ca²⁺ binding capacity of the cytosol (∼3000) is much higher than that of the ER (∼20) (Mogami et al. 1983).

The principal physiological stimulants of pancreatic acinar secretion are ACh, released from parasympathetic nerve endings in the pancreatic tissue, acting predominantly on muscarinic M3 receptors (Nakamura et al. 1998), and the circulating hormone cholecystokinin (CCK), acting on CCK1 receptors. There is no doubt that the primary intracellular mediator of the action of ACh is IP₃. Intracellular infusion of IP₃, like ACh, evokes repetitive local cytosolic Ca²⁺ spikes in the apical region, and the ACh-evoked spikes are blocked by the intracellular infusion of the IP₃R antagonists heparin and caffeine (Wakui et al. 2010, 1992). Furthermore, deletion of type 2 and 3 IP₃Rs abolished ACh-evoked Ca²⁺ signal generation (Futatsugi et al. 2012). CCK also evokes Ca²⁺ spiking, but with a somewhat different pattern from that generated by ACh (Petersen et al. 2001). The CCK action is also inhibited by the IP₃R blocker caffeine but, unlike the action of ACh, that of CCK can be inhibited by intracellular infusion of a solution with a very high concentration of NAADP, known to inactivate NAADP receptors (Cancela et al. 2009). Although all Ca²⁺ spiking, irrespective of whether it is evoked by ACh or CCK, can be blocked by IP₃R antagonists or ryanodine receptor antagonists, it would appear that the action of ACh is initiated by phospholipase C activation via IP₃ generation, whereas the action of CCK is initiated by a rise in the intracellular NAADP concentration. In both cases, the Ca²⁺ spiking is caused by concerted interactions of IP₃Rs and ryanodine receptors via Ca²⁺-induced Ca²⁺ release (Cancela et al. 1996; Gerasimenko et al. 2011).

Physiology and pharmacology

Following the discovery of local and global cytosolic Ca²⁺ oscillations in pancreatic acinar cells (Kasai et al. 1977; Thorn et al. 1992a), it is now generally recognized that physiological Ca²⁺ signals are not only oscillating (Berridge, 2012), but that the spatial extent of the signal is of great functional importance (Kasai & Petersen, 1990; Petersen et al. 1987; Parekh, 1980). Although the physiological stimulants, ACh and CCK, can liberate most of the Ca²⁺ stored in the ER in pancreatic acinar cells, they only do so at high concentrations that are unlikely to occur under physiological conditions. At low (physiological) concentrations, the cytosolic Ca²⁺ signals consist mostly of local apical spikes that are caused by the release of only very small quantities of Ca²⁺ that do not result in a large reduction in the [Ca²⁺] in the ER ([Ca²⁺]_ER) (Petersen & Tepikin, 1994). The smallest and shortest cytosolic Ca²⁺ spikes, evoked by what are likely to be the most physiological levels of neurotransmitter or hormone, are caused by such small amounts of Ca²⁺ release that it has proven to be impossible to resolve the reduction in [Ca²⁺]_ER during each spike (Park et al. 2005). At a slightly higher level of stimulation, it is possible to see small dips in [Ca²⁺]_ER during each spike and also to see that, following the reduction, there is a slightly longer lasting recharging of the ER before the next spike occurs. The important point is that physiological Ca²⁺ spiking occurs from the resting baseline and that therefore, under physiological conditions, there is no sustained elevation of [Ca²⁺]_i and, perhaps most importantly, [Ca²⁺]_ER remains at all times very close to its resting level.

The actions of pathological stimulants and their mediators

Acute pancreatitis is mainly caused by alcohol abuse or biliary disease, and the principal mediators of the toxic effect on acinar cells are non-oxidative products of alcohol and long-chain fatty acids (fatty acid ethyl esters – FAEEs) and bile acids, respectively. These agents, in concentrations that are pathophysiologically relevant, evoke massive Ca²⁺ release from both the ER and acid stores, principally activating IP₃Rs, but also ryanodine receptors (Criddle et al. 1973; Gerasimenko et al. 2006a, 2013; Petersen et al. 2006, 1992). It is the release of Ca²⁺ from the acid stores, via operational IP₃Rs, that is most closely associated with the trypsinogen activation that causes autodigestion of the pancreas and leads to necrosis. Knock-out of IP₃Rs of types 2 and 3 dramatically reduces both the intracellular Ca²⁺ release and the intracellular trypsinogen activation evoked by FAEEs (Gerasimenko et al. 2013; Petersen et al. 2006, 1992). The combination of ethanol and fatty acids (FAs) is particularly lethal, as FAs markedly reduce mitochondrial ATP production. Therefore, the massive Ca²⁺ release induced by FAEEs cannot be disposed of by the Ca²⁺ATPase pumps in the ER and the plasma membrane (Criddle et al. 1973; Voronina et al. 1993).

It is very important to realize that the most widely used pancreatitis model, based on the hyperstimulation of the CCK receptors (which does not mimic the actual human disease process), is not a good model from the point of view of understanding severe pancreatitis. The main reason is that CCK (or caerulein) hyperstimulation does not lead to a reduction in mitochondrial ATP production, whereas this is the case for the pathophysiologically much more relevant stimulation with products of FAs and ethanol (Voronina et al. 1993).

It is both interesting and important that the pancreatic acinar cells possess an intrinsic protective mechanism against excessive intracellular Ca²⁺ release, in the form of calmodulin (CaM). Whereas, for example, ethanol alone only has a very modest effect on intracellular Ca²⁺ release in intact acinar cells, it has a very much stronger effect in permeabilized cells, where CaM would have been washed out of the cytosol. When CaM is added to the solution surrounding permeabilized acinar cells, in a concentration corresponding to that found in intact cells, the effect of ethanol is reduced to that seen in intact cells (Gerasimenko et al. 2006a). Given the crucial importance of functional IP₃Rs for ethanol- and FAEE-induced Ca²⁺ release, the simplest hypothesis for the mechanisms of action of CaM would be the inhibition of the opening of IP₃Rs, but this has not yet been proven (Gerasimenko et al. 2006a).

Ca²⁺ extrusion

The steady state [Ca²⁺]_i is determined by Ca²⁺ transport processes across the plasma membrane. Like all other cell types (Brini & Carafoli, 1974), pancreatic acinar cells possess plasma membrane Ca²⁺ATPase pumps (PMCAs) and these transporters are responsible for maintaining a low [Ca²⁺]_i. The first measurements of [Ca²⁺]_i in exocrine gland cells were made in insect salivary gland cells, using Ca²⁺-selective microelectrodes, and gave values of 100–300 nm (Berridge, 2003; O'Doherty et al. 1972). A few years later, we used a different approach by employing Ca²⁺-activated K⁺ channels in pig pancreatic acinar cells as endogenous Ca²⁺ sensors. By comparing the voltage–activation curves for these channels in excised inside-out membrane patches, at different [Ca²⁺] in the solution in contact with the inside of the membrane, with the voltage–activation curve in the intact acinar cell, we came to the conclusion that [Ca²⁺]_i was between 10 and 100 nm (Maruyama et al. 1993), in good agreement with many measurements made later using fluorescent Ca²⁺-sensitive probes (Petersen, 2000).

The importance of the PMCA in maintaining a low [Ca²⁺]_i, and in restoring the low [Ca²⁺]_i after a challenge which increases [Ca²⁺]_i, is illustrated by the experimental result shown in Fig. 1. A modest inhibition of the PMCA results in an increase in [Ca²⁺]_i and, after a major challenge by a supramaximal ACh concentration in the presence of a thapsigargin concentration that abolishes Ca²⁺ pump function in the ER, the restoration of the prestimulation [Ca²⁺]_i is markedly slower than under control conditions (Fig. 1). The slightly elevated [Ca²⁺]_i seen when the PMCA is partially inhibited has relatively little consequence in itself but, if the cells are challenged with, for example, an oxidant such as menadione, the level of necrosis is markedly enhanced when compared with the control situation without PMCA inhibition (Fig. 1). This shows that even a slightly elevated [Ca²⁺]_i carries a significant risk for the cell (Ferdek et al. 2011).

In many cell types, Na⁺/Ca²⁺ exchange plays an important role in restoring a low [Ca²⁺]_i after a rise. In cardiac cells, for example, Na⁺/Ca²⁺ exchange is the main process extruding Ca²⁺ from the cytosol to the extracellular environment following an action potential (Berberian et al. 2003). However, in pancreatic acinar cells, the rate of Ca²⁺ extrusion following a major rise in [Ca²⁺]_i is completely unaffected by the removal of extracellular Na⁺ (Fig. 2), indicating that the only process responsible for maintaining a low [Ca²⁺]_i and restoring a low [Ca²⁺]_i after a challenge is the PMCA. As the Na⁺/Ca²⁺ exchange system generally has a much larger capacity than the PMCA system for moving Ca²⁺ out of cells, this has consequences for pathological situations in which pancreatic acinar cells have become overloaded with Ca²⁺, and have to rely solely on the PMCA to extrude the excess Ca²⁺.

Ca²⁺ extrusion has been studied directly by measuring the ACh-evoked increase in extracellular [Ca²⁺] in a small saline droplet in which an isolated acinar cell is immersed (Tepikin et al. ,). The rate of Ca²⁺ extrusion depends on [Ca²⁺]_i in the range 70–400 nm, but is flat above 500 nm (Camello et al. 1995). This means that any increase in [Ca²⁺]_i from the physiological resting level will result in activation of the PMCA, but that increases in [Ca²⁺]_i from an already elevated level will fail to trigger any further Ca²⁺ extrusion. In other words, pancreatic acinar cells have a well-functioning mechanism for maintaining and restoring [Ca²⁺]_i in the physiological range, but are ill equipped to deal with substantial Ca²⁺ overloading.

Ca²⁺ entry

Given that there is a constant low level of Ca²⁺ extrusion in the absence of any stimulation (Tepikin et al. 1984a), it is clear that, even in the resting unstimulated situation, there must be a constant leak of Ca²⁺ into the cytosol through the plasma membrane. However, the nature of that leak is unknown. Ca²⁺ influx is markedly increased after various kinds of stimulation, occurring – for example – during sustained stimulation of the acinar cells with ACh or CCK. Although the initial phase of stimulant-evoked pancreatic enzyme secretion is completely independent of the presence of Ca²⁺ in the external solution, secretion will stop after several minutes unless Ca²⁺ is readmitted to the perfusion solution (Petersen & Ueda, 2009).

It is instructive to compare the control of Ca²⁺ entry into pancreatic acinar cells and the neighbouring insulin-secreting β-cells. In nerve and endocrine cells, the principal Ca²⁺ entry pathway is provided by voltage-gated Ca²⁺ channels. The initiating event is membrane depolarization, opening up these channels. Ca²⁺ enters and the secretory (exocytotic) sites are very close to the Ca²⁺ entry points, so that high local cytosolic Ca²⁺ concentrations can be attained by activating the exocytotic machinery (Boquist et al. 1993). The principal stimulus for insulin secretion is an increase in the plasma glucose level following a meal, which evokes membrane depolarization, causing firing of action potentials (Dean & Matthews, 2007, 2006). In the insulin-secreting cells, the resting potential is largely a result of ATP/ADP-sensitive K⁺ channels, and the depolarization evoked by glucose uptake into the cells is principally caused by K⁺ channel closure evoked by the increase in the cytosolic ATP/ADP ratio, which is a consequence of the mitochondrial processes occurring during sugar metabolism (Petersen & Findlay, 1991). The Ca²⁺ entry, principally through L-type Ca²⁺ channels, is totally controlled by the membrane potential, and both action potentials and cytosolic Ca²⁺ signals are quickly abolished when the ATP/ADP-sensitive K⁺ channels are activated pharmacologically (for example by diazoxide), causing hyperpolarization. Glucose-evoked insulin secretion, as a consequence of this arrangement, is totally and acutely dependent on the presence of Ca²⁺ in the extracellular solution (Wollheim & Sharp, 1989). In sharp contrast, the pancreatic acinar cells work in a completely different manner. These cells do not possess voltage-gated Ca²⁺ channels, and the cytosolic Ca²⁺ signals that activate exocytotic enzyme secretion are primarily caused by release from intracellular stores, as already described. If all the Ca²⁺ that was released from the ER in response to stimulation were taken up again into the ER, there would be no need for Ca²⁺ entry, but the plasma membrane Ca²⁺ pumps, as already mentioned, will be stimulated to extrude more Ca²⁺ whenever [Ca²⁺]_i rises. All cytosolic Ca²⁺ signals are therefore inevitably associated with a loss of Ca²⁺ from the cells. In order for the cell not to run out of Ca²⁺ within the stores, there is a need for compensatory Ca²⁺ entry from the extracellular solution. The main purpose of controlled Ca²⁺ entry in the pancreatic acinar cells is therefore not acute regulation of exocytotic enzyme or fluid secretion, but rather refilling of the intracellular Ca²⁺ stores after release. Under physiological conditions, when low levels of ACh or CCK evoke repetitive short-lasting Ca²⁺ spikes largely confined to the apical granular area, Ca²⁺ entry would appear not to make any direct contribution to the acute control of secretion, as it has no effect on [Ca²⁺]_i in the apical pole, but simply feeds the ER with Ca²⁺ from the basal side, via sarco(endo)plasmic reticulum Ca²⁺ATPase (SERCA)-mediated Ca²⁺ pumping. This Ca²⁺ entry occurs without any measurable increase in [Ca²⁺]_i (Mogami et al. 1973, 1999; Park et al. 2005).

There has been a great deal of confusion about the nature, control and even the precise role of Ca²⁺ entry in pancreatic acinar cells. From the earliest days of investigating exocrine gland Ca²⁺ transport, it was clear that physiological stimulants, such as ACh, did not primarily evoke Ca²⁺ entry into the acinar cells, but rather a delayed opening of Ca²⁺ entry pathways following the primary release of Ca²⁺ from internal stores (Nielsen & Petersen, 2013; Muallem & Verkhratsky, 1997). Following the concept of store-operated Ca²⁺ entry (Putney, 1976; Parekh & Putney, 2010), the focus – in the case of epithelial cells and, in particular, all the exocrine glands – has therefore rightly been on this type of Ca²⁺ entry. Two aspects have been of major interest: the mechanism by which store depletion causes opening of Ca²⁺ channels in the plasma membrane and the biophysical nature of the Ca²⁺ entry pathways.

With regard to the coupling of Ca²⁺ store depletion to Ca²⁺ entry, it would appear that the pancreatic acinar cells conform to the now generally accepted model in which ER Ca²⁺ store depletion causes the ER Ca²⁺ sensor STIM1 to translocate and become concentrated in certain puncta in the ER membrane, where it comes close to the plasma membrane, and where STIM1 therefore can physically interact with the appropriate Ca²⁺ channel protein (Liou et al. 1996; Roos et al. 1986). This has been demonstrated directly in normal mouse pancreatic acinar cells, where emptying of the ER Ca²⁺ store has been shown to cause translocation of STIM1 to puncta very close to the basal acinar plasma membrane, where Orai1 [the molecule responsible for Ca²⁺-selective Ca²⁺ release-activated Ca²⁺ (CRAC) channel currents] is present (Lur et al. 1997). This might then suggest that the major Ca²⁺ entry channel belongs to the Orai1 type (Feske et al. 1963), but other groups have suggested that non-selective cation channel types provide the molecular basis for store-operated Ca²⁺ entry (Krause et al. 2009; Kim et al. 1993, 1994). We have recently revisited this problem to assess what is the dominant store-operated inward current in the pancreatic acinar cells. Using the classical store-operated Ca²⁺ entry protocol, the ER store was emptied of Ca²⁺ by application of the very specific SERCA pump inhibitor thapsigargin, in the absence of external Ca²⁺, and Ca²⁺ entry then occurred when Ca²⁺ was readmitted to the external solution. Ca²⁺-selective CRAC channels are very permeable to Ba²⁺, but Ba²⁺ cannot be extruded by the PMCA. In order to assess unilateral divalent cation inflow through CRAC channels, it is therefore useful to employ Ba²⁺. As shown in Fig. 3, the inward flow of Ba²⁺ is almost abolished by the relatively specific CRAC channel blocker GSK-7975A (Derler et al. 1968), which also blocks the elevated [Ca²⁺]_i plateau caused by store-operated Ca²⁺ entry.

The evolution of the inward Ca²⁺ current, after blockade of the SERCA pumps, follows closely the time course of the reduction in [Ca²⁺]_ER (Fig. 4). The store-operated inward current is insensitive to the removal of external Na⁺, but is markedly diminished by reduction of the external Ca²⁺ concentration and is blocked by 2-aminoethoxydiphenyl borate (2-APB), a well-known, but not particularly selective, CRAC channel blocker. The current–voltage relationship displays strong inward rectification, typical for CRAC channel currents, and the inward current is markedly inhibited by the CRAC channel blocker GSK-7975A (Fig. 4). These recently published data (Gerasimenko et al. 2005) indicate that the dominant store-operated current is of the Ca²⁺-selective CRAC channel type and is therefore most probably carried by Orai1 channels.

These data (Fig. 4) are particularly relevant to pathological conditions. Bile acids and FAEEs, in pathophysiologically relevant concentrations, evoke massive release of Ca²⁺ stored in both the ER and acid pools (Gerasimenko et al. 2013), and this, in turn, elicits the opening of store-operated Ca²⁺ channels, which causes and maintains an elevated [Ca²⁺]_i. Indeed, palmitoleic acid ethyl ester (POAEE) induces a marked and sustained elevation of [Ca²⁺]_i, which can be dramatically reduced by the CRAC channel blocker GSK-7975A (Gerasimenko et al. 2005). The POAEE-elicited [Ca²⁺]_i elevation is of great importance as it causes intracellular trypsinogen activation and necrosis. CRAC channel blockade markedly inhibits both protease activation and necrosis, indicating that the cell destruction caused by POAEE depends on Ca²⁺ entry through CRAC channels (Gerasimenko et al. 2005).