1. Introduction

The human exocrine pancreas secretes 1-2 liters of pancreatic juice per day. When stimulated, the pancreas secretes alkaline pancreatic juice containing copious amounts of bicarbonate (HCO₃^-) (23, 74). HCO₃^- plays essential roles in the digestive system. HCO₃^- determines the pH of bodily fluids as a major buffer system that guards against toxic pH fluctuations (116). HCO₃⁻ in pancreatic juice neutralizes gastric acid, and provides an optimal pH environment for digestive enzymes to function in the duodenum (74). In addition, HCO₃⁻ acts as a moderate chaotropic ion that facilitates the solubilization of macromolecules, such as digestive enzymes and mucins (42). The importance of HCO₃^- is highlighted in the abnormal HCO₃^- secretion in cystic fibrosis (CF), which causes poor mucin hydration and solubilization leading to obstruction of ductal structures of the pancreas, intestine, vas deferens and the lung (112, 113).

The exocrine pancreas is composed of two major cell types, acinar and duct cells. Acinar cells secrete a small volume of isotonic, plasma-like, NaCl-rich fluid and digestive enzymes. Duct cells modify the ionic composition of the fluid and secrete the bulk of the fluid and HCO₃⁻of the pancreatic juice. According to the Henderson-Hasselbalch equation, at pH 7.4 and 5% CO₂, the HCO₃^- equilibrium concentration in plasma is approximately 25 mM. In humans, dogs, cats, and guinea pigs, HCO₃^- concentration in postprandial pancreatic juice is higher than 140 mM (23, 74). This remarkable transport performance has attracted much attention from pancreatologists and physiologists. Current understanding of the molecular mechanism of pancreatic HCO₃^- secretion was improved by the recent identification of ion transporters and channels, including the cystic fibrosis transmembrane conductance regulator (CFTR) (61), the electrogenic Na⁺-HCO₃^- co-transporter NBCe1-B (also known an pNBC1) (1), and the solute-linked carrier 26 (SLC26) transporters (25, 100), together with regulatory proteins, such as with-no-lysine kinase 1 (WNK1) (102), SPAK (30)and the inositol-1,4,5-triphosphate (IP₃) receptor binding protein released with IP₃ (IRBIT) (140).

2. Control of Pancreatic HCO₃^- Secretion

Pancreatic HCO₃^- secretion increases in response to ingestion of a meal, and is regulated by multiple neurohumoral inputs. Fluid and enzyme secretion by acinar cells are controlled predominantly by an increase in cytoplasmic free Ca²⁺ concentration ([Ca²⁺]_i) (88, 106, 107). Fluid and HCO₃^- secretion by duct cells are regulated by cAMP signals (74, 86) that synergizes with Ca²⁺ to generate the physiological response(105). Pancreatic ductal cells express receptors for a battery of hormones and neurotransmitters. The two major hormones controlling ductal fluid and HCO₃^- secretion are the G_s-coupled, cAMP generating hormone secretin and the G_q-coupled, Ca²⁺ mobilizing hormone cholecystokinin (CCK), which are released from neuroendocrine cells in the upper duodenum. Cholinergic vagal output via an enteropancreatic vagovagal reflex also plays an important role in controlling ductal fluid and HCO₃^- secretion. In addition to these classic stimulators, several other humoral agents are released by the pancreas for fine tuning its secretion, including insulin, somatostatin, purines, and prostaglandins (78). Additional information on hormonal control of pancreatic secretion can be found in a previous review (74) and the “Regulation of Pancreatic Secretion” section in Pancreapedia (18).

Humoral Control

Secretin: The low pH (below 4.5) gastric chyme stimulates the release of secretin from duodenal S cells into the blood (12, 19). Secretin stimulates ductal fluid and HCO₃^- secretion and synergizes with Ca²⁺ mobilizing agonists to potentiate enzyme secretion by acinar cells. Plasma secretin levels rise after a meal (19, 110) and correlate with HCO₃^- output (118). Secretin-stimulated fluid and HCO₃^- secretion is modulated by both peptide hormones, such as CCK and somatostatin, and by vagal stimulation (38, 68, 144).

CCK: CCK is a major stimulator of acinar cell enzyme and fluid secretion which is mediated by the Ca²⁺-dependent exocytosis of zymogen granules and activation of apical (luminal) Cl^- channels, respectively. CCK also acts on pancreatic duct secretion; however, the effects of CCK on pancreatic duct differ among species. In humans, the effect of CCK alone on ductal fluid secretion is weak; however, CCK greatly potentiates the effects of secretin (144).

Purines: Pancreatic duct cells express multiple purinergic receptor (P2Rs) types, including ionotropic P2X and metabotropic P2Y receptors at the apical and basolateral membranes (82). P2Rs are stimulated by purinergic ligands released from nerve terminals at the basolateral space, zymogen granules of acinar cells into the luminal space, or efflux by ductal ATP transporters to both the basal and luminal compartments. Stimulation of P2Rs increases [Ca²⁺]_i in duct cells (96, 98). Several studies have examined effects of P2Rs on ion transporters in ductal cell lines, but there are almost no studies on ductal fluid and HCO₃^- secretion. Ishiguro et al. demonstrated that luminal ATP stimulated, while basolateral ATP inhibited fluid and HCO₃^- secretion in guinea-pig pancreatic duct (51). 

Neuronal Control

Pancreatic secretion is regulated by the enteric nervous system, which is composed of a gut-brain axis and an intrapancreatic system. The major neurotransmitter acting on pancreatic duct cells is acetylcholine released by vagal parasympathetic fibers. Duct cells express both M1 and M3 muscarinic receptors, which act through changes in [Ca²⁺]_i, but the M3 receptors are likely the main receptors since their expression level is higher than the M1 receptors (31, 62). In humans, cholinergic stimulation enhances ductal secretion stimulated by secretin, likely by synergistic mechanism that is mediated by IRBIT (105). Vasoactive intestinal peptide (VIP) and ATP are also localized in parasympathetic nerve terminals (69, 97). Vagal stimulation causes VIP release that is coupled with fluid and HCO₃^- secretion (46, 69).

3. Key Transporters Involved in Pancreatic HCO₃^- Secretion

Pancreatic HCO₃^- secretion is mediated by a coordinated function of transporters expressed in the apical and basolateral membranes of duct cells. Pancreatic HCO₃^- secretion can be divided into 2 steps. The first step is uptake of HCO₃^- into the duct cells from the blood through the basolateral membrane. The second step is efflux of HCO₃^- across the apical membrane of duct cells. Regulatory mechanisms in the cytosol that include ions like Cl^- and several kinases and phosphatases, act on the transporters to coordinate and integrate the secretory process. Recent advances in molecular, cellular, and physiological techniques have enhanced our understanding of the molecular identity, localization, function, and regulatory mechanisms of ductal ion transporters (75). The major ion transporters expressed in the apical and basolateral membranes of the pancreatic duct cells are summarized in Table 1 and Figure 1.

Figure 1. A schematic diagram depicting the transporters and channels in the apical (luminal) and basolateral membranes of pancreatic duct cells.

The main driving force for HCO₃^- secretion is achieved by the Na⁺ gradient generated by the Na⁺/K⁺ ATPase pump and K⁺ Channels at the basolateral membrane, which generate the intracellular negative membrane potential. HCO₃^- is loaded mainly through the electrogenic (1Na⁺-2HCO₃^-) NBCe1-B, and partly by NHE1 located in the basolateral membrane. Basolateral AE2 may act to supply Cl^-_in to maintain the secretion. Apical HCO₃^- secretion is performed by the interacting and functionally interrelated CFTR and Slc26a6. Transcellular HCO₃^- movement generates a lumen-negative electrical potential that results in paracellular Na⁺ secretion through the paracellular pathway. Water follows Na⁺ and HCO₃^- osmotically via paracellular and transcellular (aquaporins) pathways. In the resting state, luminal NHE3 and NBCn1-A function as salvage luminal HCO₃^-. Modified from (75).

Table 1: Transporters of pancreatic duct

Na⁺/K⁺ ATPase, and K⁺ channels

The main driving force for fluid secretion is achieved by the Na⁺/K⁺ ATPase pump and K⁺ channels which generate the transmembrane Na⁺ and K⁺ gradients and the negative intracellular membrane potential (75, 101). The Na⁺/K⁺ ATPase pump is expressed in the basolateral membrane of the ducts (84, 117, 126, 131). The Na⁺/K⁺ ATPase pump generates the Na⁺ and K⁺ gradients by extruding 3 Na⁺ ions in exchange for 2 extracellular K⁺ ions which move inside using the energy of ATP hydrolysis. The K⁺ gradient generated by the pump, in conjunction with basolateral K⁺ channels, yields a negative membrane potential. The Na⁺ gradient is used to drive several Na⁺-coupled solutes, including HCO₃^- absorption by the basolateral Na⁺-HCO₃^- cotransporter NBCe1-B and basolateral and luminal Na⁺/H⁺ exchangers (NHEs). The negative membrane potential aids in HCO₃^- efflux through electrogenic transporters. MaxiK channels (KCNMA1) have been identified on the basolateral membrane of rat pancreatic duct cells, and are likely candidates for maintaining the membrane potential during agonist-stimulated HCO₃^- secretion (33). A Ba²⁺-sensitive channel of 82 pS conductance (KCNN4) appears to be a basolateral K⁺ channel, which is responsible for the resting K⁺ permeability (99).

Na⁺-HCO₃^- Co-transporters (NBCs)

The main ductal basolateral membrane HCO₃^- accumulation transporter is NBCe1-B (75). NBCe1-B was cloned from pancreas and was named pNBC1 (1). It was later re-named NBCe1-B as part of classification of the NBC family (11). NBCe1-B is an electrogenic transporter with a 1 Na⁺: 2 HCO₃^- stoichiometry in pancreatic duct cells (37). NBCe1-B can be regulated by cAMP-dependent protein kinase A (PKA) phosphorylation at Ser1026 and Thr49 (36). In principle, Na⁺/H⁺ exchangers in the basolateral membrane (e.g. NHE1) can also mediate HCO₃^- influx in duct cells. However, the electrogenic NBCe1-B utilizes the Na⁺ gradient more efficiently than the electroneutral NHE1 (1 Na⁺: 1 HCO₃^-). Indeed, NBCe1-B contributes up to ~75% of the HCO₃^- influx during secretin-induced ductal fluid and HCO₃^- secretion in guinea pig (54, 56). The activity of NBCe1-B is controlled by multiple inputs, including IRBIT and the WNK/Ste20-related proline/alanine-rich kinase (SPAK) pathway (124, 141) and most notably intracellular Cl^- (121). In addition to NBCe1-B, the duct expresses electroneutral NBCn1-A (NBC3) on the apical (luminal) membrane (103, 111). This transporter may mediate HCO₃^- salvage in the resting state to maintain acidified pancreatic juice (32, 85).

CFTR

The discovery of acidic pancreatic juice in patients with cystic fibrosis (CF) was a milestone in understanding the mechanism of pancreatic HCO₃^- secretion (57). The CF transmembrane conductance regulator (CFTR) was discovered as the protein mutated in patients with CF (61, 114, 115). Although CFTR is a member of the ATP-binding cassette (ABC) transporters superfamily that usually act as membrane pumps that transport their substrates against the electrochemical gradient (20), CFTR functions as an anion (Cl^- and HCO₃^-) channel through which ions diffuse down the electrochemical gradient. CFTR is located at the apical membrane of pancreatic ducts (17, 129, 146) (and all secretory epithelia), and is activated by the cAMP/PKA pathway. At [Cl^-]_i higher than 10 mM, CFTR functions as a Cl^- channel that has limited permeability to HCO₃^- (79, 109, 120). However, when [Cl^-]_i drops to below 10 mM, CFTR anionic selectivity changes to increase HCO₃^- permeability and mediate luminal HCO₃^- exit (59, 102). Indeed, as has been shown in patients with CF (17, 50, 129), CFTR is critically involved in epithelial HCO₃^- secretion. This leads to revision of the original model of ductal HCO₃^- secretion, in which Cl^-/HCO₃^- exchangers mediate apical HCO₃^- efflux and CFTR facilitates the apical Cl^-/HCO₃^- exchangers by recycling the Cl^- (9). This continues to be the case at high Cl^-_in. However, at low [Cl^-]_i, HCO₃^- efflux via CFTR driven by the membrane potential has essential role in HCO₃^- efflux and HCO₃^--driven fluid secretion in the pancreatic duct (55, 128). Interestingly, this indicates that HCO₃^- permeability of CFTR is not fixed but is dynamically modulated by the protein kinase WNK1 (102). The low [Cl^-]_i present during active HCO₃^- secretion triggers the activation of WNK1 and, in turn, WNK1 acts on CFTR to increase CFTR HCO₃^- permeability (102). Notably, CFTR mutations with altered WNK1-mediated increase in HCO₃^- permeability are associated with chronic pancreatitis in humans (71).

CFTR has a more global role in ductal fluid and HCO₃^- secretion. In addition to functioning as a Cl^- and HCO₃^- channel, CFTR functions as a scaffold forming macromolecular complexes with other transporters and regulatory proteins at the apical membrane (75). CFTR has a PSD95/Discs-large/ZO-1 (PDZ) ligand at the C-terminus and binds to PDZ domains of adapter proteins, such as Na⁺/H⁺ exchanger regulatory factors (NHERFs) and SH3 and multiple ankyrin repeat domains 2 (Shank2) (72, 125) through which CFTR interacts and regulates the activity of slc26a6, slc26a3 (66), NHE3 (3) and NBCn1-A (103). Another interaction of CFTR is with soluble NSF attachment protein receptor (SNARE) proteins, A-kinase anchor proteins (AKAPs), kinases and phosphatases (39) that may serve to regulate CFTR activity and the activity of the transporters interacting with CFTR.

Cl^-/HCO₃^- Exchangers

Cl^-/HCO₃^- exchangers mediate the bulk of HCO₃^- exit across the apical membranes of the pancreatic duct cells until that the last critical portion of HCO₃^- exit is mediated by CFTR once it gains HCO₃^- permeability. In humans, members of the solute-linked carrier 4 (SLC4) and SLC26 families function as Cl^-/HCO₃^- exchangers. Among the SLC4 transporters, duct cells express AE2 (SLC4A2) at the basolateral membrane that regulates pH_i and protects against alkaline load (101). However, our recent studies revealed essential role for AE2 in ductal fluid and HCO₃^- secretion (47). Intuitively, basolateral HCO₃^- efflux mechanism should inhibit rather than stimulate ductal HCO₃^- secretion. It is not clear why AE2 is essential for ductal fluid secretion, but one possibility is that AE2 may provide the duct with Cl^- that is needed to keep the luminal slc26a6 functioning in a face of limited Cl^- provided by acinar secretion (47).

Among the SLC26 family transporters, SLC26A3, and SLC26A6 are located on the apical membrane of the pancreatic duct cells and mediate Cl^-/HCO₃^- exchange. Interestingly SLC26A3 has a 2Cl^-/1HCO₃^- stoichiometry (67, 122), while SLC26A6 functions as a 2HCO₃^-/1Cl^- exchanger (63, 122). A persistent osmotic gradient is needed to support the copious fluid secretion by the pancreatic duct. This is satisfied by the coupled action of NBCe1-B and SLC26A6 that results in a continuous net HCO₃^- (osmolite) transcellular transport and thus transcellular H₂O flow (123, 130, 136). In addition, SLC26 transporters interact with CFTR through the sulfate transporter and anti-sigma factor antagonist (STAS) domain, and regulate pancreatic secretion by activating CFTR (67).

Other Transporters, Channels, and Carbonic Anhydrases

Na⁺/H⁺ exchangers (NHEs): TheSLC9A NHEs family is electroneutral 1Na⁺/1H⁺ exchangers. The ubiquitous NHE1 (SLC9A1) is essential for pH_i homeostasis and supply Na⁺ to the Na⁺/K⁺ ATPase pump (147), including the basolateral membrane of pancreatic duct. Diffusion of CO₂ from the blood into the duct and CO₂ generated by metabolism is hydrated by the action of carbonic anhydrases to generate HCO₃^- and H⁺. NHE1 does not have a major role in basolateral HCO₃^- influx as revealed by minimal inhibition of fluid and HCO₃^- secretion by inhibition of NHE1 in pancreatic duct of most species (133, 139). Interestingly, the NHE3 isoform is expressed in the apical membrane of pancreatic duct and is thought to mediate HCO₃^- salvage at the resting state (73). At the resting state, the pancreatic juice is acidic, indicating an active H⁺ secretion (32, 85) that may be mediated by the combined action of NHE3 and NBCn1-A. Similar to NBCn1-A (103), NHE3 interacts with CFTR via PDZ domain containing proteins (3), and is regulated by IRBIT (43, 44).

Ca²⁺-activated Cl^- channels (CaCCs): Several members of the anoctamin (TMEM16/ANO) family function as CaCC (15, 119, 143). TMEM16A/ANO1, TMEM16B/ANO2, TMEM16F/ANO6, and TMEM16K/ANO10 are expressed in pancreas (75). However, ANO1 is expressed in acinar but not duct cells (119), ANO6 functions as a flipase (131). The function of ANO2 and ANO10 in the pancreas is not clear at this time. Nevertheless, ample evidence show that the pancreatic duct (and ducts of other secretory glands) has CaCC activity in the apical membrane (34, 35, 134, 145). The molecular identity of this channel is not known at present, nor its function in HCO₃^- secretion. Several other CaCCs are known and are candidates for the ductal CaCC. In pancreatic acinar cells and other serous cells, ANO1 may have a role in HCO₃^- transport. At physiological [Ca²⁺]_i concentrations ANO1 functions as a Cl^- channel. However, at high [Ca²⁺]_i and perhaps at high [Ca²⁺]_i microdomains, ANO1 HCO₃^- permeability is increased by Ca²⁺/calmodulin (58, 60), raising the possibility that ANO1 can provide an alternative Cl^- and HCO₃^- conduction in acinar cells.

Aquaporins: Although the paracellular pathway is permeable to H₂O, H₂O flows, at least in part, transcellularly via the water channels aquaporins (AQP) family. This is best illustrated in salivary glands, where knockout of AQP5 markedly reduces salivation (83). Among the 13 AQPs, AQP1 and AQP5 are the major aquaporins in pancreatic duct (14, 64, 65). The role of individual aquaporins in the duct has not been established yet.

Carbonic Anhydrases: A topic poorly studied that deserve more attention is the role of the ductal carbonic anhydrases (CAs) in fluid and electrolyte secretion, in particular with the emerging secretory epithelial diseases due to mutations in CAs. Mutations that affect the action of CA4 cause retinitis pigmentosa (5) and a mutation in CA12 causes salt wasting (28, 90). All transporters involved in fluid and HCO₃^- secretion are affected by the HCO₃^- concentration at the cellular compartments and microdomains that determine HCO₃^- availability at plasma membrane inner and outer surfaces. Hydration of CO₂ by CAs determines local HCO₃^- concentration both at the outer and inner plasma membrane surfaces (87). Several CAs are localized in the cytoplasm (such as CA2 and CA7) and several are anchored at the plasma membrane (such as CA4, CA12 and CA14) with the catalytic site at the extracellular surface and regulate HCO₃^- concentration at the basolateral (CA4 and CA12), or the luminal (CA4) membrane surfaces (29).

CAs localized in the plasma membrane and cytoplasm interact with H⁺ and HCO₃^- transporters that mediates ductal fluid and HCO₃^- secretion and regulate their activity. CA4 interacts with the C terminus of NBCe1-A to increase its activity (4). The C terminus of NBCe1-A and NBCe1-B are conserved and thus likely CA4 regulates NBCe1-B. NBCn1-A recruits the cytoplasmic CA2 to the plasma membrane, where CA2 increases the activity of NBCn1-A (81). CA2 is closely associated with NHE3 and increases NHE3 activity (70). CA2 interacts with apparently novel site at the C terminus of NHE1 to regulate NHE1 activity (77). CA2 was reported to interact with the C terminus of slc26a6 to increase its activity. However, the role of other CAs, in particular the plasma membrane anchored CAs, on the activity of the slc26a6 and other SLC26 transporters has not been investigated yet. Finally, CA2 also interacts with AQP1 to increase water flux by AQP1 by an unknown mechanism (135).

A particularly interesting CA is the basolateral membrane anchored CA12 with a catalytic site at the extracellular membrane surface. A human mutation in CA12(E143K) is the cause of an autosomal recessive form of salt wasting, which leads to hyponatremia with hyperkalemia, high sweat Cl^-, dehydration and failure to thrive. (27, 28, 90). A recent work to understand the cause of the disease established a prominent role for CA12 in ductal fluid and HCO₃^- secretion. Thus, CA12 increased, while CA12 (E143K) markedly reduced, ductal fluid secretion in isolated ducts and in vivo. This could be attributed to a potent stimulation of ductal and topically expressed AE2 by CA12. The E143K mutation is a folding mutation that resulted in retention of CA12(E143K) in the ER (47). How exactly CA12 with external catalytic site activates AE2 is not obvious. CA12 may clear the extruded HCO₃^- from the membrane surface to prevent its buildup at the mouth of the AE2. If this can be established, it will be a new mode of regulating HCO₃^- transporters by CAs.            

4. Regulation and Mechanism of Pancreatic HCO₃^- Secretion

Intracellular Signaling Pathways: cAMP and Ca²⁺

The cAMP/PKA pathway is central in inducing ductal HCO₃^- secretion. Secretin is the major hormone that activates the cAMP pathway. VIP also signals to increase cAMP via VIP receptors (VPAC1) (26, 132). At maximal receptor stimulation, the cAMP/PKA pathway can fully activate fluid and HCO₃^- secretion by activation of the apical CFTR and the basolateral Na⁺-HCO₃^- cotransporter, NBCe1-B (142). However, at physiological conditions the cAMP/PKA pathway synergizes with the Ca²⁺ signaling pathway to activate the secretory process (see below).

Several agonists that act on the pancreatic duct engage the Ca²⁺ signaling pathway. These include CCK, cholinergic stimuli, P2Rs, and protease-activated receptor 2 (PAR2) receptors (62, 107). Activated CCK and muscarinic receptors activate PLC to generate IP₃ that releases Ca²⁺ from intracellular stores and activates the membrane Ca²⁺ influx channels, Orai1 and TRPC. P2Rs (82, 94) and PAR2 (6, 91, 93, 95) also act through activation of the Ca²⁺ signaling pathway.

At physiological stimulus intensity, the cAMP and Ca²⁺ signaling pathways synergize to activate ductal secretion (59). Early studies in vivo already noted the synergistic action of ductal stimuli. Application of secretin at a level observed in the postprandial state only produces modest HCO₃^- and fluid output (24, 40). Application of CCK and stimulation of M1 and M3 receptors markedly augmented secretin-stimulated pancreatic fluid secretion, although alone CCK and muscarinic stimulation have minimal effect on ductal secretion (74, 144). The molecular mechanism of synergism was resolved with the discovery of regulation of ductal secretion by IRBIT which is discussed below. The cAMP and Ca²⁺ signaling pathways crosstalk on several additional levels to modulate the activity of each (59, 104). cAMP/PKA phosphorylates IP₃R2 to augment Ca²⁺ release from the ER (13). Ca²⁺ influx through the Orai1 channels activates the Ca²⁺-dependent adenylyl cyclase (AC) AC8 (137). Ca²⁺ can also activate the CFTR-dependent Cl⁻/HCO₃⁻ exchanger in CAPAN-1 human pancreatic duct cells (92), which may involve activation by IRBIT.  

Regulation by IRBIT

Activation of NBCe1-B, slc26a6, and CFTR: IRBIT was isolated as a protein that interacts with the IP₃ receptors (IP₃Rs) and it can be dissociated from the IP₃Rs by IP₃ (21).  IRBIT competes with IP₃ for binding to the IP₃Rs (8) to inhibit Ca²⁺ release. In fact, the IP₃Rs appear to function as IRBIT buffers to prevent IRBIT access to many transporters and targets regulated by IRBIT (75). The C-terminal region of IRBIT family proteins shows ~ 50% homology with the ubiquitous housekeeping enzyme S-adenosyl-l-homocysteine hydrolase (AHCY), with IRBIT having additional N terminal sequence while it lacks the hydrolase activity (7). The main known domains of IRBIT are PP1 and calcineurin binding motif, a PEST domain, a coiled-coil domain, and a PDZ ligand at the end of C terminus (75).

IRBIT plays an important role in pancreatic ductal secretion by regulating multiple transporters and mediating the synergistic action of the cAMP/PKA and Ca²⁺ signaling pathways (Figure 2). Knockdown of IRBIT in ducts and knockout in mice modestly inhibit fully stimulated pancreatic duct fluid and HCO₃^- secretion (142), and eliminate the physiological synergistic action of the cAMP/PKA and Ca²⁺ signaling pathways (105). IRBIT accumulates at the apical pole where IP₃Rs are highly expressed, but it can be found all over the cell where IP₃Rs are present (76). A search for IRBIT binding proteins identified NBCe1-B as a binding partner, where IRBIT binds to the autoinhibitory domain N terminus of NBCe1-B to activate it by removing the autoinhibition (124). Subsequent detailed studies, in particular with the pancreatic duct revealed that IRBIT at the apical pole potently activates CFTR (140, 142), SLC26A6 (105), and possibly NHE3 (44). At the basal side, IRBIT regulates NBCe1-B (124, 140, 142). IRBIT activates the transporters by multiple mechanisms. First, IRBIT recruits protein phosphatase 1 (PP1) to the transporters to dephosphorylate NBCe1-B and CFTR at sites that are phosphorylated by the kinase SPAK activated by phosphorylation mediated by the kinases WNK1 and WNK4. This enhances the plasma membrane relocation of NBCe1-B, CFTR (140) and slc26a6 (105) from intracellular vesicular pools. At the plasma membrane, IRBIT directly interacts with the transporters to further increase their activity. The activation mechanism is not known in all cases. However, information exists for NBCe1-B in which IRBIT eliminates autoinhibition (124). The PDZ binding motif of IRBIT is required for assembling the IRBIT-NBCe1-B and IRBIT-CFTR complex (142).

IRBIT and Synergism: An important action of IRBIT is mediating the synergistic action of the cAMP/PKA and Ca²⁺ signaling pathways (105) (see Figure 2). Physiological stimulus intensity must be quite weak to prevent cell toxicity that occurs under strong stimulation of all signaling pathways. Indeed, at physiological stimulus intensity the secretory process is activated only by about 10% or less of maximal stimulation. Synergism between weakly stimulated signaling is used to generate the maximal response while avoiding cell toxicity. IRBIT mediated the synergism between the cAMP/PKA and Ca²⁺ signaling pathways by translocation between cellular compartments and transporters. At the resting state, IRBIT is sequestered by the high level of IP₃Rs at the ductal ER apical pole and is not available for interaction with the transporters. The affinity of the IP₃Rs for IRBIT and IP₃ is regulated by PKA-mediated phosphorylation of specific IP₃Rs serine residues. Phosphorylation of the serine residues increases the affinity for IP₃ and at the same time decreases the affinity for IRBIT. Now a small increase in IP₃ evoked by weak stimulation of the Ca²⁺ signaling pathway is sufficient to dissociate IRBIT from the IP₃Rs (105). The released IRBIT can now bind to CFTR and slc26a6 first in intracellular vesicles to dephosphorylate them by the IRBIT recruited PP1 (and perhaps calcineurin) and promote their translocation to the luminal membrane. At the luminal membrane, IRBIT activates the transporters to initiate ductal fluid and HCO₃^- secretion (105). Of note, the synergistic action of the cAMP/PKA and Ca²⁺ signaling pathways is eliminated by the knockout of IRBIT (105), highlighting the key role of IRBIT in the synergistic action of the cAMP/PKA and Ca²⁺ signaling pathways, which is the physiological way that ductal fluid and HCO₃^- secretion take place.

Figure 2. A model for IRBIT associated pathway of pancreatic ductal fluid and HCO₃^- secretion. Key domains of IRBIT related to HCO₃^- secretion are illustrated at the top of the figure. In the resting state, IRBIT is bound to IP₃Rs, and SPAK phosphorylates NBCe1-B, Slc26a6, and CFTR located at intracellular organelle. When the duct cells are stimulated, IP₃ is released and bound to IP₃Rs, while IRBIT is disengaged from IP₃Rs. PP1 recruited to IRBIT dephosphorylates transporters located at the plasma membrane. IRBIT also binds to the autoinhibitory domain of NBCe1-B to activate it. Increased surface expression of the transporters also aids pancreatic ductal HCO₃^- secretion. Modified from (104). See text for details. 

Regulation by [Cl^-]_i

WNK1 and dynamic regulation of CFTR HCO₃^- permeability: The WNK proteins consist of four members (WNK1 – WNK4) with a conserved kinase domain that is noted for the unique position of the catalytic lysine residue (89). The discovery that mutations in WNK1 and WNK4 cause hypertension in humans has attracted much attention to the kinases function and regulation (138). The main function of the WNK kinases is the regulation of Na⁺, K⁺, Cl^-, HCO₃^-, and Ca²⁺ transporters in epithelia and brain (48, 49, 104). The WNKs act either by regulating surface expression of membrane transporters through modulation of their endocytosis or by phosphorylating the transporters and other target proteins directly or indirectly through affecting the effect of other kinases (49). Several functions of WNKs are mediated by phosphorylating and activating the downstream oxidative stress-responsive kinase 1 (OSR1) and SPAK (22). WNK1, WNK3, WNK4, SPAK, and OSR1 are expressed in the pancreatic duct (102, 140) and participate in the regulation of HCO₃^- transporters and channels (75). Accordingly, knockdown of WNK4 alone or a combined knockdown of WNK1, WNK3 and WNK4 increase pancreatic duct fluid secretion by removing a tonic negative effect of ductal HCO₃^- transporters (140). However, the role of the WNKs, in particular WNK1 changes at the terminal portion on the duct when [Cl^-]_i is reduced to below 10 mM. WNK1, and perhaps the other WNKs, binds [Cl^-]_i and its activity is regulated by [Cl^-]_i (108).

The role of the WNKs and SPAK at [Cl^-]_i above 10 mM is illustrated in the upper portion of the left model of Figure 3. The role of WNK1 in pancreatic HCO₃^- secretion at [Cl^-]_i below 10 mM is illustrated in the bottom portion of the figure. Osmotic stress or low [Cl^-]_i activates WNK1 (108). Notably, activation of WNK1 by low [Cl^-]_i greatly increases the HCO₃^- permeability of CFTR (58, 102). During active pancreatic HCO₃^- secretion, Cl^- concentration in the pancreatic juice progressively reduces Cl^- absorption. Because of the low basolateral and high luminal Cl^- permeability (52, 102), [Cl^-]_i rapidly decreases in response to the reduction in luminal duct Cl^- concentration. By the Nernst equation, at a membrane potential of -60 mV, [Cl^-]_i will be less than 1/10 of luminal Cl^- concentration. Indeed, ductal [Cl^-]_i was estimated to be about 5 mM during cAMP-induced active secretion (52, 102).

Figure 3. A model depicting WNK1-mediated regulation of CFTR in pancreatic ductal function. During active pancreatic HCO₃^- secretion, Cl^- concentration in the pancreatic juice is progressively reduced due to Cl^-/HCO₃^- exchange activities at the apical membrane of duct cells. Because the basolateral membrane of duct cells has poor Cl^- permeability but the apical Cl^- permeability is very high due to activation of CFTR, [Cl^-]_i rapidly decreases in response to the reduction in luminal Cl^- concentration. Activation of WNK1 by low [Cl^-]_i increases the P_HCO3/P_Cl of CFTR to over 1.0, which greatly augments HCO₃^- flux through the CFTR pore. Simultaneously, WNK1/SPAK pathway inhibits Slc26a6 to prevent HCO₃^- reabsorption. This mechanism enables an increase to as much as 140 mM HCO₃^- in pancreatic juice. See text for details. Modified from (75).

WNK1 modulates the anion selectivity of CFTR by changing the pore size (58). Stimulation by WNK1 increases the pore size of CFTR from 4.8 Å to 5.3 Å, which facilitates the passage of larger anions, including HCO₃^- (4.3 Å, diameter), more than the smaller anion, Cl^- (3.7 Å, diameter) by reducing the energy barriers of size-exclusion, and ion dehydration. Changes in pore size affect the energy barrier of ion dehydration by altering the electric permittivity of the water-filled cavity in the pore. Dielectric constant (relative permitivity, ε) is a unit of electric permittivity, and the dielectric constant of water (ε_w) is approximately 80 at room temperature. Water molecules in confined geometries like ion channels exhibit a space-dependent reduction in the pore water ε_w down to 20, due to the restriction of the translational and rotational mobility of water molecules (2). Pore dilation relieves this restriction of water molecule movement and increases ε_w, which eventually leads to an increase in the overall ε of the anion channel pore. Indeed, the pore dilation induced by WNK1 activation increased the ε of the CFTR pore from 16 to 43 (58). In general, ions pass through the channel after dehydration (at least partial dehydration). Asymmetrically charged ions, such as HCO₃^-, show lower permeability than the symmetrically charged ions, such as Cl^-, due to the high hydration/dehydration energy barrier. The increase in anion channel pore ε greatly alleviates the dehydration penalty of the asymmetrically charged HCO₃^-, and increases P_HCO3/P_Cl (Figure 4).

Figure 4. WNK1 modulates the anion selectivity of CFTR by changing the pore size. Stimulation by WNK1 increases the pore size of CFTR from 4.8 Å to 5.3 Å and the pore dilation increases the dielectric constant (ε) of the CFTR pore from 16 to 43. The increase in pore size facilitates the passage of the larger anion, HCO₃^- (4.3 Å, diameter), more than the smaller anion, Cl^- (3.7 Å, diameter) by reducing the energy barriers of size-exclusion. More importantly, the dielectric constant increase enhances the HCO₃⁻ permeability of CFTR by reducing energy barriers required for ion dehydration of HCO₃⁻ (58). See text for details.

Interestingly, activated WNK1 while shifting CFTR anion selectivity, does not lose the inhibitory effect on SLC26A6 and SLC26A3 (102). When the luminal HCO₃^- concentration is greater than 140 mM, continuous activation of apical Cl^-/HCO₃^- exchange would reverse to absorb HCO₃^- from the lumen. This is more of a problem for the 2Cl^-/1HCO₃^- exchange by slc26a3 and less, if at all, for the 1Cl^-/2HCO₃^- slc26a6, especially at membrane potential of -60 mV across the luminal membrane. However, inhibition of the apical Cl^-/HCO₃^- exchangers is required to prevent the reverse mode of Cl^-/HCO₃^- exchange activity if slc26a3 dominates the exchange when ductal [Cl^-]_i is below 10 mM and ultimately achieves HCO₃^- concentration above 140 mM in pancreatic juice (127, 129).

Cl^-_in as a signaling ion: [Cl^-]_i emerged as a signaling ion by regulating several HCO₃^- transporters. Regulation of the WNK kinases by [Cl^-]_i and its role in CFTR HCO₃^- permeability was discussed above. By virtue of regulating the function of the WNK kinases, [Cl^-]_i may affect other transporters regulated by these kinases. A significant recent discovery is that [Cl^-]_i profoundly regulates the activity of several Na⁺-HCO₃^- cotransporters (NBCs) at the [Cl^-]_i physiological range (121). [Cl^-]_i regulates the activity of all NBCs tested NBCe1-B, NBCe1-C, and NBCe1-A. The IRBIT-independent activity of NBCe1-B is inhibited by [Cl^-]_i between 60-140 mM that is outside the physiological range and may function to inhibit NBCe1-B activity under pathological conditions. Most notably, when activated by IRBIT, NBCe1-B activity is reduced by [Cl^-]_i in the range of 5-20 mM, where at 20 mM [Cl^-]_i, NBCe1-B activity is reduced to the basal, IRBIT-independent level. Molecular analysis identified two Cl^- interacting motifs at the N terminus of NBCe1-B that mediate high and low affinity inhibition by [Cl^-]_i. Regulation of NBCe1-B is mediated by sites that contain the GXXXP motif. The first site mediates the high [Cl^-]_i affinity (5-20 mM) regulation of NBCe1-B and the second site mediates the low [Cl^-]_i affinity (60-140 mM) regulation of NBCe1-B (121). NBCe2-C activity is not regulated by IRBIT and in this case regulation of NBCe1-C is mediated by a single site containing the GXXXP motif and takes place at [Cl^-]_i between 10-30 mM. Regulation of NBCe1-A by [Cl^-]_i is mediated by a cryptic Cl^- interacting site containing the GXXXP motif. The cryptic NBCe1-A [Cl^-]_i interacting sites was unmasked by deletion of residues 29-41.

Hence, cells have [Cl^-]_i sensing mechanism that plays an important role in the regulation of Na⁺ and HCO₃^- transporters that mediated the critical step of HCO₃^- influx in the process of ductal fluid and HCO₃^- secretion. At [Cl^-]_i of up to 20 mM, CFTR functions mostly as a Cl^- channel and slc26a6 mediates most ductal HCO₃^- secretion. As [Cl^-]_i is reduced below 20 mM and additional HCO₃^- secretion takes place in the face of unfavorable Cl^- and HCO₃^- gradients across the apical membrane, there is an increased demand for HCO₃^- entry across the basolateral membrane. Pancreatic duct cells achieve this by [Cl^-]_i-mediated regulation of NBCe1-B and CFTR, at which NBCe1-B activity and CFTR HCO₃^- permeability gradually increase as [Cl^-]_i is reduced towards 5 mM.    

A Model for Pancreatic HCO₃^- Secretion

Electrogenic HCO₃⁻ transporters can secrete higher concentrations of HCO₃⁻ than electroneutral transporters when the electrorepulsive force generated by the negative membrane potential is coupled to the efflux of HCO₃⁻. The electrogenic SLC26A6 exchanger with the stoichiometry of 1 Cl^-: 2 HCO₃⁻,can achieve luminal HCO₃⁻ concentration of up to about 120 mM at apical membrane potential of -60 mV (129). To drive luminal HCO₃^- concentration to 140 mM, the physiologic HCO₃^- concentrations in pancreatic juice, another mechanism is needed (129). Such a mechanism should be Cl^- independent, since significant fraction of pancreatic HCO₃⁻ secretion is retained in the absence of luminal Cl^- (53, 55). The WNK1 activated CFTR satisfy these requirements. At HCO₃^-_in in stimulated duct cells above 25 mM and membrane potential of -60 mV CFTR mediated HCO₃^- efflux even at luminal HCO₃^- concentrations of above 140 mM. The transcellular basal to luminal electrogenic HCO₃^- transport by both slc26a6 and CFTR generates a lumen-negative electrical potential that results in paracellular Na⁺ secretion. Water flows down the osmotic gradient generated by the Na⁺ and HCO₃^- fluxes via paracellular and transcellular (aquaporins) pathways. Overall, these processes generate an efficient mechanism for HCO₃^-–driven ductal fluid secretion to generate the volume and HCO₃^- content of the pancreatic juice.

5. Conclusions

The mechanism by which the human pancreatic duct secretes nearly isotonic HCO₃^- solution has long been an enigmatic question for both physiologists and clinicians (74, 129). When Bayliss and Starling first noticed that the exocrine pancreas secretes alkaline fluid, they assumed that carbonate is responsible for the strong alkalinity of the pancreatic juice (10). Later, with better understanding of the carbonate/HCO₃^-/CO₂ buffer system (45), it became clear that the exocrine pancreas secretes fluid in which the dominant anion is HCO₃^-, and HCO₃⁻ secretion is coupled to fluid secretion (16, 24, 41). Current understanding indicates that activation of three key transporters, the basolateral NBCe1-B (and likely AE2), and the luminal SLC26A6 and CFTR, and their synergistic regulation by the cAMP and Ca²⁺ signaling pathways through IRBIT and WNK1 perform for vectorial pancreatic HCO₃⁻ secretion that drives fluid secretion. NBCe1-B, with a 1 Na⁺/2 HCO₃^- stoichiometry, is the main HCO₃^- concentrating transporter in the basolateral membrane, and can achieve the necessary HCO₃^- influx (1, 54, 148). Basolateral AE2 activity is also required to support ductal HCO₃^- fluid and HCO₃^- secretion, although AE2 exact role is not known at present. The electrogenic SLC26A6, with a 1 Cl^-/2 HCO₃^- stoichiometry is the major apical Cl^-/HCO₃^- exchanger, which mediates most HCO₃^- efflux in the early step of pancreatic HCO₃⁻ secretion (67, 80). Activated WNK1 increases HCO₃^- permeability of CFTR, allowing further apical HCO₃^- efflux and setting pancreatic juice HCO₃^- concentrations above 140 mM (102). Our understanding of the mechanism of pancreatic fluid and HCO₃^- secretion will continue to improve as our knowledge of existing pathways increases and new mechanisms are identified and delineated, to provide a better scientific basis for therapeutic approaches to treat diseases like cystic fibrosis and acute and chronic pancreatitis.

6. Acknowledgments

We thank Dong-Su Jang for assistance in illustrations. This work was supported by grant, 2013R1A3A2042197, from the National Research Foundation, the Ministry of Science, ICT & Future Planning, Republic of Korea and by Intramural NIH/NIDCR grant DE000735-06.