Transepithelial ion transport across duct cells of the salivary gland
Epithelial Signaling and Transport Section, Molecular Physiology and Therapeutics Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD, USA
Correspondence: Ehud Ohana, Epithelial Signaling and Transport Section, Molecular Physiology and Therapeutics Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, 20892 MD, USA. Tel: +1 301 594 0332, Fax: +1 301 402 1228, E-mail: email@example.com or firstname.lastname@example.org
Fluid and electrolyte secretions are vital for all epithelia and when aberrant lead to numerous pathophysiological conditions. Electrolyte transport across epithelia generates the osmotic force for fluid movement and is mediated by several membrane proteins expressed on both apical and basolateral poles of epithelial cells. Sodium and chloride are crucial for regulation of fluid secretion, thus regulating salivary volume. Bicarbonate (), on the other hand, is the major pH buffer; hence, aberrant secretion is a major factor in diseases such as cystic fibrosis (CF) causing altered mucin hydration and solubilization. Here, the structure–function mechanisms of the major membrane transporters involved in salivary duct electrolyte transport are reviewed focusing on transepithelial movement of Cl− and .
Vectorial transepithelial ion movement in secretory glands is crucial for controlling the ion composition and volume of the secreted fluid. and fluid secretion and Cl− absorption that are essential for the normal function of all epithelia are also highly important for proper function of salivary glands. is the major pH buffer that protects glands from cytotoxic changes in pH and when secreted into the oral cavity prevents enamel erosion by acidic pH (Catalan et al, 2011) (Joiner, 2007). Being a chaotropic ion, it also facilitates the solubilization of mucins and enzymes in the secreted fluid.
Similar to all other exocrine glands, salivary glands consist of two major cell types, acinar and duct cells, each playing a different key role in gland function. The Cl− and Na+ secretions by acinar cells are utilized as the driving force for bulk fluid secretion. Subsequently, the duct cells absorb the Cl− and Na+ and secrete K+ and . Salivary glands share many characteristics with other exocrine glands such as the pancreas and the mammary glands in terms of morphology, function, and regulation. Nevertheless, other properties, for example, the identity of fluid-secreting cells, are different. In the salivary gland, the bulk of fluid secretion occurs in the acinar cells, and the duct cells are considered to be water impermeable. However, in the pancreas, fluid secretion is mainly mediated by ductal cells. All the processes described above involve the function of numerous cellular and molecular machineries that are tightly regulated.
Here, the transmembrane transport mechanisms that are involved in the orchestrated maintenance and regulation of electrolyte transport in salivary gland duct cells are reviewed focusing on Cl− and . The salivary duct cells are currently emerging as targets for clinical interventions as during certain pathophysiologies, for example, gland damage following radiation treatment, most of the damage is inflicted on acinar cells. Because duct cells functionally participate in secretion and absorption, the remaining duct cells can be manipulated to retain gland function.
Salivary gland morphology
The salivary gland consists mainly of acinar cells and various duct cells that form the ductal tree. The tree functions as a scaffold to support the entire gland structure and also as a vehicle for mediating the unidirectional movement of saliva. Acinar cells are the primary cells that secrete fluid rich in mucins, electrolytes, other proteins, and digestive enzymes (Lee et al, 2012). The fluid shuttles all other secreted contents further up the ductal tree where the ionic composition is modified by the duct cells. The fluid that is secreted by the acinar cells is rich in NaCl, and as it flows through the duct, Cl− is being actively absorbed in exchange for , resulting in a more alkaline fluid that contains less Cl−. This is achieved by the continuous activity of a large number of membrane transport proteins that are expressed on the luminal and basolateral poles of the gland epithelium (see Figure 1; Table 1). Many of these transporters, discussed here, exploit the initial high luminal Na+ and Cl− concentrations as a driving force for secreting other ions such as potassium and bicarbonate.
Table 1. Ductal transporters
Major Ion Selectivity
Sodium/Potassium ATPase (Na+/K+ ATPase)
Primary (pump-ATP hydrolysis)
Generating Na+ and K+ gradients
Na+ and K+
Sodium Proton Exchanger (NHE1-3, SLC9A1/2/3)
Secondary active (exchanger)
NHE1: Mediates basolateral H+ efflux that contributes to influx
Na+ and H+
Secondary active (exchanger)
Fluid and secretion
Cl−, , Oxalate
Sodium Potassium Chloride Cotransporter (NKCC)
Secondary active (cotransporter)
Epithelial Na+, K+, and Cl− uptake
Na+, K+, Cl−
Channel – water transport
Water (certain members can mediate urea and glycerol movement)
Channel (passive ion diffusion) MaxiK – voltage and Ca2+-activated, IK1 – Ca2+ activated
K+ secretion and membrane potential maintenance (during stimulation)
Na+/ Cotransporter (NBCn1-A, NBC3, SLC4A7)
Secondary active (cotransporter)
Channel (passive ion diffusion)
Calcium-activated Cl− Channels
Channel (passive ion diffusion)
Acinar cells are triangle-shaped polarized epithelial cells (illustrated in Figure 1) that have numerous secretory vesicles that can be easily visualized under the microscope in close proximity to the apical pole. These vesicles contain the enzymes, mucins, other proteins, and ions to be secreted into the lumen by vesicle fusion to the apical membrane in an exocytotic process. The centroacinar cells are found at the terminal of the salivary gland ductal tree where the intercalated duct ends. At this junction, the content of the acinar cell secretory vesicles is released to begin its journey along the ductal tree. The initial duct portion is termed the intercalated duct, which is followed by the larger intralobular duct together forming the major site of secretion and Cl− absorption in both salivary glands and the pancreas (Figure 1). This is achieved by cells of the striated ducts, which specialize in transcellular secretion. These cells have microvilli that increase their surface area for more efficient absorption as well as a large number of mitochondria to meet the high energy consumption of the secretion process. The fluid then proceeds to the interlobular and interlobar ducts, which, in rodents, are the sites of bulk K+, , and water secretion and NaCl absorption. This entire process generates a high volume of saliva rich in mucins, other proteins, and enzymes in a high and low Cl− milieu (Lee et al, 2012).
Membrane transporters that mediate electrolyte and fluid transport across salivary gland epithelia
*For a summary of all relevant transporters discussed, as well as a list of key terms with definitions used, please see Tables 1 and 2.
Table 2. Terms and definitions
A net movement of charge per transport cycle mediated by ion transporters across cellular membranes. For example, the electrogenic Slc26a6 mediates 1 Cl−, resulting in a net outward movement of 1 negative charge
Primary active transport (pump)
The mechanism of an ion transporter mediating ion transport against its electrochemical gradient by exploiting ATP hydrolysis as energy source. (examples from the text – The Na+/K+ATPase)
Secondary active transport (exchangers, cotransporters)
The mechanism of an ion transporter mediating ion transport against its electrochemical gradient by exploiting the electrochemical gradient energy of another ion that is transported down its gradient. (examples from the text – The SLC26 exchangers, the Na+/ cotransporters)
PDZ stands for the first three proteins that were shown to share the domain, PSD95, Dlg1, and ZO-1. The PDZ scaffold proteins are interacting with membrane proteins that have PDZ consensus amino acid sequences. These scaffold proteins mediate the trafficking of other proteins simultaneously to the plasma membrane and are therefore forming protein complexes and determining polarity
Kinases (PKA, SPAK)
Enzymes that transfer phosphate groups that are usually derived from ATP hydrolysis to other enzymes/proteins in a process termed phosphorylation. These include PKA (protein kinase A) and SPAK that are mentioned in the text
Cyclic AMP (cAMP)
Cyclic adenosine monophosphate (cAMP) is a molecule derived from ATP by the enzyme adenylyl cyclase. cAMP is a second messenger involved in several pivotal physiological roles, most notably, regulating cAMP-dependent protein kinases
ATP-binding cassette proteins are transmembrane transporters that mediate the transport of a wide variety of substrates across the membrane with its gradient (not pumps). These proteins exploit ATP hydrolysis for channel pore gating and have intracellular nucleotide-binding domains (NBDs). This family of transporters includes the CFTR and several multi-drug-resistance proteins
*Protein/gene designation human proteins – capital letters only; mouse/rat proteins – first letter capital and the rest lower case; bacterial – all lowercase. Gene designation is in Italics similar to the protein.
This ATP-dependent primary transport mechanism is ubiquitously expressed in all cells including salivary gland epithelium. Also termed the Na+/K+ pump, it is localized to the basolateral membrane of acinar and duct cells and exploits the energy of ATP hydrolysis to mediate an inward 2K+ (and 1H+) and outward 3Na+ transport against its electrochemical gradient. This generates the high extracellular low intracellular Na+ concentrations and low extracellular high intracellular K+ concentrations necessary to maintain the resting membrane potential of all cells. The Na+ and K+ gradients are then utilized by other membrane transporters on both the basolateral and apical membranes. An elaborated depiction of the Na+/K+ pump was achieved almost half a century after its discovery and is reviewed by Toyoshima et al, 2011.
Channels are transmembrane proteins that mediate ion transport by diffusion down its electrochemical gradient and therefore require no direct energy source for active transport. By utilizing the K+ gradient generated by the Na+/K+ ATPase, the K+ channels mediate an outward current that determines the acinar cell membrane potential during stimulated K+ secretion, which is measured to be very close to the K+ diffusion potential. Two major types of Ca2+-activated K+ channels are expressed in acinar cells – the MaxiK and IK1 channels. The MaxiK channel (protein) is encoded by the Kca1.1 gene and is a Ca2+- and voltage-activated potassium channel of high conductance (100–200 pS) (Maruyama et al, 1983). The IK1 channel (protein) is encoded by the Kca3.1 gene and acts as a Ca2+-activated time- and voltage-independent potassium channel of intermediate conductance (20–40 pS) (Hayashi et al, 1996).
These channels are essential for stimulated secretion in parotid salivary glands, as stimulated secretion in double-knockout (KO) mice that lack both channels is dramatically reduced by more than 70% (Romanenko et al, 2007). Interestingly, however, individual deletion of either of these channels did not disrupt normal parotid gland secretion in mice (Begenisich et al, 2004) (Romanenko et al, 2006). This is explained by the observation that membrane potential of submandibular acinar cells from either Kca3.1-null or Kca1.1-null mice remains strongly hyperpolarized following strong stimulation to maintain normal secretion. Nevertheless, only in double-KO mice does the acinar cell membrane potential remain very close to the Cl− equilibrium potential. Therefore, each channel by itself can maintain hyperpolarization of acinar cells, which is sufficient to sustain normal secretion. In spite of the fact that both K+ channels can function alone, when co-expressed, IK1 channels have an inhibitory effect on the MaxiK current, yet IK1 activity seems to be unaffected by MaxiK (Thompson and Begenisich, 2006).
The Na+ bicarbonate cotransporters (NBCs)
All mammalian genomes encode 10 genes of the SLC4 family of transporters that include three Cl−/ exchangers and five Na+/ cotransporters. These transporters play a key role in intracellular pH regulation, which is maintained, on the one hand, by acid ‘loaders’ such as the electrogenic SLC4 Na+/ cotransporters that mediate ion transport with a stoichiometry of 1 Na+ and 3 per transport cycle. On the other hand, acid ‘extruders’ are necessary to preserve the delicate intracellular pH homeostasis, which is achieved by different electroneutral and electrogenic SLC4 proteins that have a Na+/ stoichiometry lower than 1:3. The terms acid ‘loaders’ and ‘extruders’ refer to the influx or efflux of protons (H+), respectively, which is equivalent to movement in the opposite direction. One of the electrogenic Na+-coupled bicarbonate transporters is the NBCe1 that has several known variants, namely NBCe1-A (also known as kNBC), which is predominantly expressed in the kidney proximal tubule, NBCe1-B (pNBC), which is expressed in the pancreas and salivary glands, and has a high sequence similarity to NBCe1-A except for a longer N terminus. Another variant, NBCe1-C, is expressed in the brain and is highly similar in sequence to NBCe1-B except for an extended and less conserved C-terminus (Boron et al, 2009).
The NBCe1-B isoform is ubiquitously expressed in almost all epithelia including the salivary gland. It is localized to the basolateral membrane of both duct and acinar cells and is capable of mediating the bulk entry at the basolateral pole during stimulated ductal fluid and secretions (Lee et al, 2012). In pancreatic ductal cells, NBCe1-B has a 1:2 stoichiometry for Na+/ transport and is, therefore, electrogenic. This transporter is absorbing at the basolateral pole, which can then be transported to the salivary gland lumen by the Cl−/ exchanger Slc26a6, which will be discussed here later. Surprisingly, an electroneutral NBC transporter, NBCn1-A, was shown to express at the luminal membrane of salivary and pancreatic duct cells. This means that is not only secreted into the salivary gland lumen but also absorbed from it. As suggested by Luo et al, 2001, the influx of mediated by NBCn1-A is important for salvage to maintain low salivary pH, thus preventing premature activation of salivary enzymes during periods of low luminal fluid flow.
Structural analysis of the Na+/ cotransporters remains elusive; however, certain determinants within the structure of the SLC4 family members have been extensively studied. The transmembrane domain of all SLC4s is highly conserved and is most likely to harbor the ion-binding site and transport pathway. The variability is mostly at the N- and C-termini of the proteins, which are mainly regulatory domains involved in ion transport regulation and protein–protein interaction. The sequence variability at these sites determines the different characteristics of each SLC4 protein in terms of stoichiometry, cellular localization, and protein–protein interactions. A relevant example is the two Na+/ transporters of the salivary gland epithelium mentioned above. The sequence of both the N- and C-termini of NBCe1-B and NBCn1-A are generally similar, yet have some major differences. 1) The different localization of the two proteins can be attributed to the fact that NBCn1-A has a PDZ consensus sequence at its C-terminal, which is absent in NBCe1-B (Boron et al, 2009). PDZ proteins were named PDZ as the first three proteins that were shown to share the domain were PSD95, Dlg1, and ZO-1. The PDZ scaffold proteins are interacting with membrane proteins that have PDZ consensus amino acid sequences usually at the extreme end of the protein C-terminal. These scaffold proteins mediate the trafficking of several other proteins simultaneously to the plasma membrane and are therefore forming protein complexes and determining polarity. 2) The extreme N-terminal of the two proteins is mostly similar and includes two highly important domains. The first is known as the autoinhibitory domain (AID), which has an inhibitory effect on the transporter activity. That was discovered when constructs lacking the AID showed more than twofold increase in transport activity (Boron et al, 2009). Overlapping or in close proximity to the AID is the IRBIT (IP3-binding protein released with inositol 1,4,5-triphosphate) protein-binding site. IRBIT is a newly discovered highly ubiquitous protein first isolated from the brain and shown to bind the IP3 receptor, which is a crucial component for cellular calcium signaling (Ando et al, 2003). IP3 binding to the receptor subsequently releases IRBIT that was shown to bind, in turn, to several membrane proteins including NBCn1-A, NBCe1-B, Cftr, and Slc26a6 (Yang et al, 2011). IRBIT dramatically upregulates the activities of NBCn1-A, NBCe1-B as well as Slc26a6 and Cftr by a dual mechanism. On the one hand, IRBIT inhibits SPAK kinase (Table 2) that under resting conditions downregulates the activity of these proteins by phosphorylation and, on the other hand, activates them through direct binding (Yang et al, 2011). IRBIT is therefore suggested to act as a master regulator of transepithelial transport. Taken together, the sequence domains that were mentioned above provide insight as to how sequence differences between proteins of the same family lead to diversity in localization and regulation, specifically in salivary gland epithelium and generally in polarized cells.
Solute carrier 26 member 6 (SLC26A6)
The Slc26 family of transporters consists of ten members that show very diverse transport mechanisms and ion selectivity (Ohana et al, 2009). A relevant example would be Slc26a6 that is expressed in salivary gland duct cells. Slc26a6 has a very unique feature, which it shares with Slc26a3 only. These transporters can function as both ion channels and chloride-dependent anion exchangers mediating the transport of and carboxylic acids, depending on the transported ion (Ohana et al, 2011). Other members of the family show very different transport mechanisms. Slc26a2 is a /Cl−/OH− exchanger (Ohana et al, 2012), Slc26a4 is a Cl−/I−/ exchanger (Shcheynikov et al, 2008), while Slc26a7 and Slc26a9 are Cl− channels per se.
The importance of the SLC26 transporters is further emphasized by the large number of human diseases that have been linked with SLC26 protein mutations. Pathophysiologies such as cystic fibrosis, congenital Cl− diarrhea, and pancreatitis involve aberrant transepithelial Cl− absorption and secretion that are mediated by both Slc26a3 and Slc26a6 (Dorwart et al, 2008) (Ohana et al, 2009). Slc26a6 is, to date, the only member of the family that was detected in salivary glands (Shcheynikov et al, 2008). It is, however, highly plausible that other members of the family such as Slc26a3 are endogenously expressed in salivary glands as well. The function of Slc26a6 in salivary gland is not well established; however, its pivotal role in pancreatic physiology and in oxalate transepithelial transport in the kidney and intestine has been described previously (Aronson, 2006) (Ohana et al, 2013). Interestingly, a link between kidney function and salivary gland pathophysiology was described recently in chronic kidney failure patients. Analysis of salivary calculi composition in these patients surprisingly indicated a Ca2+ oxalate component (Davidovich et al, 2009). This was the first time Ca2+ oxalate deposits were reported to be linked to salivary glands, yet the mechanism leading to its formation remains unknown.
In a recent study, our group identified molecular determinants that control the Slc26a3 and Slc26a6 modes of operation (Ohana et al, 2011). To this end, we exploited computer software to predict the 3D structure of Slc26a6 transmembrane domain (see Figure 2a). Interestingly, the software indicated the highest score of predicted transmembrane sector similarity between Slc26a6 and the crystal structure of the bacterial Cl− channel, ClC-ec1. This result was very intriguing as these are the only two known protein families that mediate a Cl−-dependent ion exchange activity as well as anion channel activity by the same protein. Using the model, we have identified a highly conserved Slc26 glutamic acid (Glu) residue that is spatially oriented similar to Glu 148 of the Escherichia Coli Cl− channel (ClC-ec1), which determines the ClC-ec1 mode of operation. When the ClC-ec1 Glu 148 residue was mutated to a neutral residue, the ClC-ec1 mode of operation changed from a Cl−/H+ exchanger to Cl− channel (Dutzler et al, 2002). The results of our study indicated that the conserved Slc26 Glu residue determines whether Slc26a3 and Slc26a6 would function as both exchanger and channel or exclusively as a channel, thus controlling their mode of operation and playing a similar role as in the ClC transporters. We also utilized the Slc26 putative model to identify more determinants related to ion binding and transport in another Slc26 member, Slc26a2, which suggests that our putative 3D model may be used as a general model for all Slc26 members. Additional studies further strengthen the hypothesis that Slc26 transporters and the ClCs have very similar architecture. A structural study that utilized our putative 3D model of Slc26 to refine a low-resolution dimer-like structure of a bacterial Slc26 homologue showed good spatial fit to our model (Compton et al, 2011). Furthermore, several recent studies describe a similarity between the cytoplasmic C-termini of Slc26 and the ClC transporters, the STAS (sulfate transporters antisigma) domain, and the CBS (cystathionine beta-synthase) domain, respectively. Both domains are at the C-terminus of the protein, they are globular structures that mediate protein–protein interaction, and both were shown to harbor a nucleotide-binding domain. Hence, based on several observations, a similarity is emerging between the ClC family of Cl− transporters and the Slc26 transporters, among them Slc26a6. These similarities cover structural and functional aspects related to these proteins.
In salivary gland physiology, Slc26a6 may, therefore, play a dual role – (i) being a Cl−/ exchanger, given a well-established contribution to the homeostasis of these ions as in the pancreas, with Slc26a6 playing a crucial role in maintaining salivary Cl−/ concentrations. This role is of high significance considering the mutual reciprocal regulation between Slc26a6 and Cftr. (ii) As Slc26a6 is the major oxalate transporter that determines oxalate concentrations in the blood and the urine, one can speculate that it may directly or indirectly determine salivary oxalate concentrations.
CTFR is a Cl− channel expressed on the apical membrane of duct cells of numerous glands, among them the salivary gland (Lee et al, 2012). Mutations in CFTR have been linked to impaired Cl− transport by the duct cells of exocrine glands, which is one of the hallmarks of the cystic fibrosis disease. Moreover, it has been suggested that Cftr is permeable to (Poulsen et al, 1994). However, its transport capability is likely to have a very limited role in ductal secretion, mainly because in the presence of physiological Cl− concentrations, the Cftr conductance, which is strongly regulated by Cl−, is extremely low. In addition, Slc26a6 is a potent Cl−/ exchanger and was shown to regulate Cftr activity (Ko et al, 2004) and determine the content of pancreatic secretions (Wang et al, 2006), and most likely those in saliva as well. To this end, a well-established role of Slc26a6 is to utilize the local inward Cl− gradient generated by Cftr to mediate efflux into the lumen. This is based on the following findings: (i) Deletion of Slc26a6 impairs pancreatic fluid and secretion. (ii) In sealed duct cultures derived from the parotid gland of Slc26a6-null mice, both basal and stimulated Cl−/ exchange activities were affected by Slc26a6 deletion, further strengthening the hypothesis that Slc26a6 and not Cftr is the major transporter in salivary glands (Shcheynikov et al, 2008).
CFTR (also known as ABCC7) is a member of the ATP-binding cassette (ABC) transporters superfamily; however, unlike many of the ABC transporters, CFTR is not a primary transport system. More specifically, CFTR does not exploit the energy of ATP hydrolysis to transport ions against its electrochemical gradient, but rather acts as an anion channel that transports ions down its electrochemical gradient. The CFTR channel is activated by the cAMP/PKA pathway and functions as a low-conductance Cl− channel (5–10 pS).
The Cftr gating mechanism has been extensively studied by several groups based on high-resolution crystal structures of the nucleotide-binding domains (NBD) (Lewis et al, 2004) (Lewis et al, 2005, 2010; Thibodeau et al, 2005) and computer models based on bacterial homologues such as the sav1866 (Dawson et al, 2007). The initial stage of the transport process involves Cl− ions that tend to accumulate in close proximity to the channel near positively charged residues on both sides of the membrane. The opening process of the channel is triggered by intracellular ATP binding to one of the NBD of the channel (Figure 2c). However, this binding is interrupted by an interaction between the regulatory (R) domain and the NBD, which sterically prevents ATP molecules from binding to the NBD. Therefore, phosphorylation of the R domain by PKA eliminates the interaction between the R domain and the NBD causing them to separate and allowing the binding of ATP to both NBDs. Subsequently, NBD1 and NBD2 approach each other, and this movement initiates a conformational change transmitted through cytoplasmic linkers to open the CFTR gate. The gate would then close 1 s after the release of ATP and maintains the closed state until an ATP molecule would bind to resume an open state and restart the cycle.
The Sodium proton exchangers (NHEs)
The robust vectorial transport of protons or equivalents such as across the membrane of salivary gland epithelium may dramatically change intracellular pH and have cytotoxic effects. Therefore, while regulating the salivary pH, epithelial cells must simultaneously maintain pH homeostasis. For this purpose, both acinar and duct cells of the salivary gland express the sodium proton exchangers, NHEs. Additional roles of the NHEs are regulation of organellar pH homeostasis and cell volume. In salivary glands, muscarinic stimulation of the gland triggers fluid secretion and upregulates NHE expression, underscoring the importance of these transporters during stimulation.
The NHE family of transporters includes 9 members, and only five of them, namely NHE1-5, are expressed on the plasma membrane of cells. The mammalian NHE transporters are ubiquitously expressed in the large majority of, if not all, cells and are electroneutral ion exchangers that mediate the extrusion of one proton in exchange for the influx of a single sodium ion. Therefore, by exploiting the inward Na+ gradient generated by the Na+/K+ ATPase, NHEs protect cells from H+ accumulation and cytoplasmic acidification. Nhe1, 2, and 3 were shown to be expressed in salivary gland epithelia; Nhe1 is expressed on the basolateral membrane of both acinar and duct cells of mouse and rat salivary gland (Evans et al, 1999) and Nhe2 and Nhe3 at the luminal membrane of interlobular and main ducts of mouse and rat salivary gland (He et al, 1997). Deletion of Nhe in mice led to the conclusion that Nhe1 is the most prominent pH regulator in acinar cells during resting and stimulated conditions. This is obviously crucial for Cl− absorption and secretion as lower intracellular pH would result in increased i leading to increased Cl−/ exchange activity (Evans et al, 1999). Moreover, Cftr regulates Nhe3 via interaction at the apical membrane (Lee et al, 2001), representing another aspect of NHE's effect on Cl− and transport. The role of Nhe2 remains unknown as deletion of this transporter did not affect pH or metabolism in salivary glands and other relevant tissues. To date, only the crystal structure of the bacterial NHE homologue,nhaA, is available (Hunte et al, 2005). It has 12 transmembrane domains, N- and C-termini both facing the cytoplasm and an overall antiparallel architecture. This architecture describes proteins that have two internal and structurally similar halves with opposite orientations relative to the membrane (Figure 2b). This architecture resembles that of the chloride channel ClC transporters and can also be found in the aquaporins. As our putative model of Slc26a6 is based on the ClC structure, Slc26 transporters may also share this general architecture.
Paracellular water permeability has long been considered a spontaneous activity that was driven and regulated by the osmotic gradient across the epithelium. Indeed, this process has a significant role in many epithelial tissues including the salivary gland; however, the identification of the aquaporins that were first reported by Preston and Agre (Preston and Agre, 1991) indicated that water transport is also mediated by specialized water channel proteins and can, therefore, be regulated by intracellular components in epithelial cells. The AQP family of proteins includes 13 members, yet AQP1, AQP3, AQP5, and AQP8 are the only members of the family that were identified in salivary glands thus far. (mRNA)AQP1 was identified in microvascular endothelial cells (Gresz et al, 2001) and Aqp8 protein in rat myoepithelial cells (Wellner et al, 2006). Expression of Aqp3 was detected in mouse submandibular gland during development, but not postnatally (Larsen et al, 2009). In contrast, AQP3 expression was detected on the basolateral membrane of acinar cells in 3D cultures of human parotid gland (Chan et al, 2011). As hypothesized, Aqp5-deficient mice show defective secretion of saliva (Ma et al, 1999), and as Aqp5 (structure illustrated in Figure 3 a and b from (Horsefield et al, 2008)) is also expressed in the airway submucosal glands, these mice suffer airway mucus aberrations as well (Song and Verkman, 2001). Surprisingly, however, Aqp1 knockout mice have completely normal salivary gland function, while their kidney water reabsorption, which also involves Aqp1 activity, is compromised. This may be due to the adaptation of the salivary gland to Aqp1 deletion that plays a more significant role in kidney water reabsorption and therefore cannot be compensated for. More insight regarding the structure–function aspects of Aqp is derived from studies that resolved the crystal structure of several Aqp members. These studies show that all Aqp water transporters are expressed as tetramers (Figure 3a and b) where each monomer has its own water transport pathway. One of the most intriguing and functionally crucial properties of an Aqp is its high water selectivity and conductance (3 × 109 molecules/s.). Although a subgroup of the Aqp family is also permeable to other molecules such as urea and glycerol, the water-specific aquaporins are generally not permeable to ions, including protons. This allows Aqps to transport water across the epithelium without changing the intra- and extra-cellular ion concentrations. The water specificity of the Aqp1 is achieved by a 3Å diameter at the narrowest region of the 20Å long pore as reported for the Aqp1 crystal structure (Sui et al, 2001). The narrow constriction site is formed by four residues in the human AQP1, namely F58, H182, C191, and R197 (Figure 3c). This constriction site is only slightly larger than the 2.8 Å size of the water molecule. The water selectivity of AQP is highly important in the salivary gland as the ion composition, and more specifically Na+, Cl−, , K+, and H+ concentrations will affect fluid volume and enzymatic activity and are, therefore, subjected to tight regulation. If the robust activity of water flux through an Aqp would have been accompanied by a non-specific ion leak, it could affect the ion composition of the saliva resulting in aberrant saliva production. As the main site of fluid secretion in salivary glands is the acinar cells, when damaged, it can dramatically hamper the volume of saliva. Therefore, by utilizing gene therapy methods, overexpression of AQP proteins in the duct cells may restore normal volume of saliva (Gao et al, 2011). Recently, clinical trial results have been reported by Baum et al. showing an increase in parotid gland saliva flow and relieved symptoms of xerostomia in treated patients (Baum et al, 2012). In this trial, patients who recovered from head and neck cancer following radiotherapy were treated by adenoviral-mediated AQP1 cDNA transfer to an irradiated parotid gland. Remarkably, the results attest for the safety and feasibility of gene therapy techniques to successfully manipulate gene expression and, specifically, AQP expression in salivary glands to treat disease.
The sodium/potassium/Cl− cotransporters (NKCCs)
Although Nkcc (Na+/K+/Cl− cotransporter) is mainly found in acinar cells (Evans et al, 2000), it was also reported to express in rat intralobular duct epithelia (He et al, 1997); hence, its pivotal role in fluid and ion secretion in the salivary gland has to be acknowledged here. NKCC1 is a ubiquitously expressed protein and a member of the Na+/K+/Cl− cotransporter family. NKCC1 is inhibited by the diuretics furosemide and bumetanide (Haas and Forbush, 2000) and is localized on the basolateral membrane of parotid acinar cells. The NKCC transporters exploit the Na+ gradient generated by the Na+/K+ ATPase to mediate the cotransport of K+ and Cl− ions into the cell. Thus, NKCC activity along with the net transport activity of both Na+/H+ and Cl−/ transporters utilize the Na+ gradient to increase intracellular Cl−. Moreover, Nkcc1 is activated by cell shrinkage to retrieve cell volume back to normal. Nkcc mediates 70% of the influx of Cl− ions that will be subsequently secreted to the lumen, that is, into the forming saliva. This is based on elegant studies in Nkcc1-null mice, showing a dramatic reduction in acinar Cl− influx and saliva secretion in KO mice compared with wild-type mice (Evans et al, 2000). Reduced Cl− secretion was also observed in the intestine and trachea of Nkcc1-null mice (Flagella et al, 1999).
NKCC1 and NKCC2 are electroneutral transporters and have a 1Na+/1K+/2Cl− stoichiometry (Haas and Forbush, 2000). NKCCs have a suggested topology of 12 transmembrane domains (TMDs) with large intracellular hydrophilic N- and C-termini (Gerelsaikhan et al, 2006). The structure of NKCC has not yet been resolved; however, several domains that are important for ion binding and transport or docking sites of inhibitory compounds were identified based on biochemical and functional assays in several NKCC1 variants including a human–shark NKCC chimeras. Na+ binding seems to involve TMDs 2 and 7, while TMDs 4 and 7 are important for Cl− transport. Bumetanide binding is coordinated by residues in the TMD 2–7 region, further strengthening the notion that these regions harbor the ion transport pathway.
Vectorial transport by salivary gland epithelial transporters
As described in Figure 1, Cl−, , and fluid secretions are regulated by the orchestrated activity of the membrane transporters reviewed here. In both acinar and ductal cells, the basolateral Na+/K+ ATPase fuels the transport by generating the Na+ and K+ gradients across the membrane by utilizing the energy of ATP hydrolysis. As salivary gland ducts are generally water impermeable, this creates a transepithelial Na+ gradient that is large enough to mainly be utilized as motive force for transport of other ions against its electrochemical gradient. For example, NKCC1, NBCe1-B (in duct cells), and NHE1 would utilize the inward Na+ gradient to transport Cl− together with K+ in, in, and protons out, respectively. The Cl− and that were inserted into the cell are now available for the apical membrane transporters to facilitate its vectorial transport to the salivary gland lumen. In acinar cells, the apical Ca2+-activated Cl− channel ANO1 would mediate the outward transport of Cl− down its electrochemical gradient. In duct cells, where the bulk of secretion and Cl− absorption occurs, Ca2+-activated Cl− channels and CFTR are responsible for local Cl− extrusion to the lumen. The high concentrations of local luminal Cl− secreted by acinar cells combined with the Cl− secreted locally by the duct cells is exploited by the SLC26A6 Cl−/ exchanger to mediate Cl− influx coupled with efflux. Therefore, SLC26A6 is the major transporter that mediates Cl− reabsorption in the duct. This results in high low Cl− concentrations in the saliva. NHE3 activity at the luminal membrane compensates for the net acid gain due to efflux by mediating Na+-dependent proton secretion. As mentioned earlier, NBCn1-A is important for salvage to prevent over alkalinization of the saliva during periods of low flow when excessive secretion is not necessary. This net vectorial transport of electrolytes generates the transepithelial osmotic gradient, which may subsequently drive a very small AQP-mediated water transport, because the duct is considered to be water impermeable. Many other proteins, hormones, and neurotransmitters are involved in the modulation and regulation of electrolyte and fluid secretion (Lee et al, 2012), yet are not in the scope of this review.
Impaired Ion transport in pathophysiologies of salivary gland
Aberrant ion transport by the mechanisms described above is associated with several human diseases of salivary glands. Different mutations in CFTR were linked to a wide range of phenotypes in patients with cystic fibrosis. The severity of the disease, especially in the pancreas, was related to CFTR contribution to transport, most likely through regulation of Slc26a6 (Ko et al, 2004). This leads to increased ductal Cl−, which is a major phenotype of cystic fibrosis. The reduced ductal diminishes the solubility of enzymes and mucins resulting in a thick obstructive mucous, which is also a hallmark of the disease.
Another major salivary gland pathophysiology occurs in the autoimmune disorder, Sjögren's syndrome, where patients suffer mainly from dry mouth and dry eye due to gland inflammation (Fox et al, 1998) (Almstahl and Wikstrom, 2003). Studies show that patients with Sjögren's syndrome have aberrant secretion; however, the cause remains unknown. It is plausible that CFTR and Slc26a6 expression is reduced in this disease much like in pancreatitis. This may occur as treatment of autoimmune pancreatitis with corticosteroids relieves symptoms in patients (Ko et al, 2010) due to improved CFTR (and possibly SLC26A6) membrane expression. Therefore, a similar cause may underlie the phenotypes observed in Sjögren's syndrome as symptoms of the disease were also shown to improve by prednisolone treatment (Miyawaki et al, 1999).
Finally, aberrant secretion and dry mouth as a result of oxidative damage have been observed in patients with malignancies of the head and neck following radiation treatment. These treatments were shown to severely and permanently damage salivary gland acinar cells while sparing a significant portion of duct cells. Still, the reduced secretion indicated that although some of the ductal cells are intact, their transporter activity may be impaired. This makes the remaining duct cells a perfect target for molecular manipulation as part of future treatments. Such treatments may involve gene therapy methods to induce expression of ion and water transporters, such as AQP, SLC26A6, and CFTR, which may help to regain normal gland function when acinar cells are damaged. As mentioned earlier, this concept has been successfully implemented for expression of AQP1 in patients suffering from xerostomia following radiotherapy for head and neck cancer (Baum et al, 2012).
The author's research is supported by the National Institute of Dental and Craniofacial Research (NIDCR) intramural program.