The strategies available for treating salivary gland hypofunction are limited because relatively little is known about the secretion process in humans. An initial microarray screen detected ion transport proteins generally accepted to be critically involved in salivation. We tested for the activity of some of these proteins, as well as for specific cell properties required to support fluid secretion. The resting membrane potential of human acinar cells was near −51 mV, while the intracellular [Cl−] was ∼62 mM, about fourfold higher than expected if Cl ions were passively distributed. Active Cl− uptake mechanisms included a bumetanide-sensitive Na+-K+-2Cl− cotransporter and paired DIDS-sensitive Cl−/HCO3− and EIPA-sensitive Na+/H+ exchangers that correlated with expression of NKCC1, AE2, and NHE1 transcripts, respectively. Intracellular Ca2+ stimulated a niflumic acid-sensitive Cl− current with properties similar to the Ca2+-gated Cl channel BEST2. In addition, intracellular Ca2+ stimulated a paxilline-sensitive and voltage-dependent, large-conductance K channel and a clotrimazole-sensitive, intermediate-conductance K channel, consistent with the detection of transcripts for KCNMA1 and KCNN4, respectively. Our results demonstrate that the ion transport mechanisms in human parotid glands are equivalent to those in the mouse, confirming that animal models provide valuable systems for testing therapies to prevent salivary gland dysfunction.
- salivary glands
millions of americans suffer from decreased salivary gland output, often termed xerostomia. The most commonly diagnosed causes of salivary gland hypofunction include the autoimmune disease Sjögren's syndrome, iatrogenic therapies, such as medications and irradiation for head and neck cancers, and systemic diseases like diabetes mellitus and pernicious anemia (12, 29, 34, 43). In addition, the cause of xerostomia in ∼20% of subjects is idiopathic (10, 25). Irrespective of the etiology, the clinical consequences of loss of salivary gland function are the same. These include a greatly enhanced risk of dental caries, periodontal disease, candidiasis, and gastric and esophageal ulcers (29). Patients suffering from salivary gland hypofunction currently use saliva-stimulating agents, such as cholinergic receptor agonists or artificial salivas (2, 13). These treatments are not very effective, frequently produce adverse side effects and usually require lifelong use. Treatments that permanently correct or more specifically address salivary gland dysfunction would be preferred. However, because of restricted accessibility, relatively little is known about the secretion process in the major salivary glands of humans, and thus, much of our knowledge is limited to animal models. Consequently, an important step in making therapies a reality requires a thorough understanding of the comparative and molecular physiology of the secretion process in human and other mammalian salivary glands.
In an effort toward achieving this goal, we have performed a comprehensive evaluation of the functional and molecular properties of the ion transport proteins expressed in human parotid glands and have compared these with the transporters expressed in mouse salivary glands. Salivary gland acinar cells secrete most, if not all, of the fluid component of saliva. The current secretion model predicts that the primary driving force for basal to apical, transacinar fluid and electrolyte secretion is Cl− movement (5, 32) (see also Fig. 6). Such Cl− trafficking involves both uptake mechanisms located in the basolateral membrane to concentrate intracellular Cl− above its electrochemical equilibrium concentration, and apical efflux channels, which are activated by an increase in intracellular [Ca2+]. Other critical steps in fluid secretion include the movement of Na ions through the acinar cell tight junctions and the efflux of K+ from the acinar cells. Pathological defects that result in hyposalivation may occur at multiple steps in this fluid secretion process, including, for example, ion transporter activation, agonist-receptor interaction, and second messenger generation. Given the vital role of transepithelial movement of electrolytes in secretion, perturbation of ion transport function is likely to be involved in many such conditions. Here, we confirm that human parotid acinar cells employ the same repertoire of ion transport proteins found in other mammalian salivary glands. Given this high degree of similarity, animal models (especially mouse) will likely continue to provide valuable insight for understanding fluid secretion in humans and for developing strategies for averting the consequences of salivary gland dysfunction.
BCECF-AM [2′-7′-bis-(2-carboxyethyl)-5-(and 6)-carboxyfluorescein, acetoxymethyl ester], EIPA [5-(N-ethyl-N-isopropyl)amiloride], DIDS (4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid) and SPQ [6-methoxy-N-(3-sulfopropyl)quinolinium] were purchased from Molecular Probes (Eugene, OR). Liberase was from Roche (Indianapolis, IN), paxilline was from Biomol (Plymouth Meeting, PA), and all other chemicals were purchased from Sigma Chemical (St. Louis, MO) or as described in the text.
Human and Mouse Parotid Tissue
Human parotid tissue was obtained from healthy male and female subjects (30–70 years of age) scheduled to have parotid surgery because their gland contained a pleomorphic adenoma that required removal of all or a large portion of the gland. Much of the normal tissue surrounding the tumor is not used for diagnostic evaluation of the sample. This discarded tissue was collected immediately after surgical excision and transported in ice-cold physiological saline to the laboratory where the tissue was either frozen in liquid N2, or acinar cells were isolated for acute functional assays. Tissue was obtained as approved by the University of Rochester Institutional Review Board.
Parotid tissue was also obtained from BlackSwiss-SvJ129 hybrid mice aged between 2 and 5 mo. Mice were housed in pathogen-free, microisolator cages with free access to laboratory chow and water ad libitum with a 12:12-h light-dark cycle. Mice were rendered unconscious by exposure to CO2 and killed by exsanguination prior to removal of the parotid glands. Animal protocols were approved by the Animal Resources Committee of the University of Rochester.
Salivary Cell Preparation
Parotid acinar cells were prepared by enzymatic digestion as previously described (3). Tissue was finely minced in Eagle's minimum essential medium (Biofluids) containing Liberase (0.3 mg/7.5 ml) and incubated at 37°C in a shaker with continuous agitation (100 cycles/min). After 20 min of incubation, the salivary gland tissue was dispersed by gentle pipetting (10 times) and centrifuged (210 g × 15 s). The cell pellet was resuspended in a further 7.5 ml of digestion medium for an additional 40 min, at the end of which time the salivary gland cells were rinsed and harvested by centrifugation. For patch- clamp studies, single acinar cells were isolated by an initial digestion for 5 min in a solution containing Liberase and 0.02% trypsin, followed by incubation in a Liberase-containing solution as described above.
RNA Isolation and Northern Blot Analysis
Total RNA was isolated from parotid gland tissues according to the manufacturer's protocol (RNeasy, Qiagen, Valencia, CA). Before Northern blot analysis, mRNA was extracted by chromatography (oligo-dT resin, Oligotex mRNA Mini Kit; Qiagen). Northern blots were prepared and hybridized using the cDNA probes described in Table 1.
One microgram of mRNA was fractionated by electrophoresis in 1% agarose/2.2 M formaldehyde gel and blotted on BrightStar-Plus positively charged nylon membrane (Ambion, Austin, TX) using 10× SSC buffer. Blots were UV-cross-linked, prehybridized with ULTRAhyb buffer (Ambion) for 60 min at 42°C, and then a P-labeled probe was added and hybridized overnight at 42°C. Probes were labeled using the Random Primers DNA Labeling System (Invitrogen, Carlsbad, CA) and purified with Centri-Sep columns (Princeton Separations, Adelphia, NJ). RNA size markers given in the figures were based on the position of the 18S and 28S bands (∼1.9 and 4.7 kb, respectively, for mouse RNA).
Targeted cDNA Array
The expression of transcripts for ion transporter proteins in human and mouse parotid glands was examined using a custom-designed “salivary gland secretion” cDNA array slide. A detailed description of the array can be found at http://www.urmc.rochester.edu/Aab/Oralbio/labpages/microarraycob/.
Probes for target genes were designed to include key water and ion transport proteins, as well as many secretion-associated signaling molecules and representative secretory proteins. Human and mouse cDNAs representing 187 “secretion” genes were obtained from either Open Biosystems (Huntsville, AL) or the “Mouse 15K cDNA Clone Set” (National Institute of Aging). PCR primers (Integrated DNA Technologies, Coralville, IA) were designed to amplify 200–1,200 bp products from the 3′ ends of highly homologous regions of the human and mouse genes. PCR products of the expected length were purified (PCR cleaning kit, Qiagen, CA), sequence verified, and dried (Eppendorf Speedvac, Hamburg, Germany). Products were resuspended at 200 ng/μl in Pronto printing buffer (Corning, Corning, NY) and printed onto UltraGap Gamma Amino Propyl Silane slides (Corning) using a Bio-Rad VersaArray arrayer and 8 SMP3B stealth pins (TeleChem International, Sunnyvale, CA). Each cDNA was printed (∼125 μm diameter) in duplicate at adjacent sites with spot-to-spot separation of 375 μm.
Twenty micrograms of total RNA was transcribed and labeled using the Superscript Indirect cDNA Labeling Kit (Invitrogen) with Cy3 or Cy5 dye (Amersham Biosciences). Labeled cDNA was mixed with hybridization buffer (0.5 mg/ml Cot 1 DNA, 0.2 mg/ml yeast tRNA, 4× SSC buffer, 50 mM pH 8 Tris, 0.3% SDS, 0.2 mg/ml BSA), incubated at 95°C for 5 min, and added directly to the array slide within the hybridization cassette (Corning). The cassette was submerged in a 58°C water bath for 18 h, at the end of which time, the slides were thoroughly washed (2× SSC/0.2% SDS for 5 min, 0.1× SSC/0.1% SDS for 2 min, 0.2× SSC for 30 s, 0.05× SSC for 30 s, and then H2O for 30 s), dried by centrifugation, and immediately scanned (Scan Array Express, Perkin Elmer, Cambridge, MA).
Samples isolated from the parotid glands of four human subjects and four mice were analyzed by array. Positive (β-actin and GAPDH) and negative controls, including blank spots and 10 alien genes (Array Validation Kit, Stratagene, La Jolla, CA), were arrayed in duplicate and used to normalize the sensitivity, signal linearity, and consistency of the assay. For “spot” identification and quantification of the fluorescent signal intensities, the microarray images were analyzed using Scan Array Express v2.1 software (Perkin Elmer). The fluorescence signal intensity for each DNA spot (average intensity of each pixel present within the spot) was calculated and subtracted using local background correction after normalization (52). A positive signal was accepted when the spot intensity was greater than the mean intensity + 2 SD of the negative controls (19, 50). Expression of a gene was considered “present” when at least 3 out of 4 samples were positive.
Measurements of the electrophysiological properties of human parotid acinar cells were made at room temperature (20–22°C) using the patch-clamp technique in various configurations. Data analysis was performed using pClamp (ver. 8.0, Axon Instruments, Sunnyvale, CA), Origin (version 7.0, Origin Software, Northampton, MA), or custom software.
Membrane potential measurements.
Membrane potential was determined using the perforated patch technique in current-clamp mode. Electrophysiological data were acquired using an Axopatch 200B amplifier and Digidata 1320A digitizer (Axon Instruments, Foster City, CA) and filtered at 2 kHz. Pipettes (Corning 8161 patch glass, Warner Instruments, Hamden, CT) were pulled to give a final resistance of 2–3 MΩ in the solutions described below. The pipette was filled with (in mM): 95 K-methanesulfonate, 45 KCl, 15 NaCl, 1 MgCl2, 5 BAPTA, 10 HEPES (pH 7.2), and then the pipette tip was back-filled with the same solution supplemented with 250 μg/ml nystatin and 2 mM Lucifer yellow. The nystatin stock solution (75 mg/ml in DMSO) was prepared daily. The liquid junction potential was minimized by briefly filling the bath with the pipette solution and zeroing the voltage. Immediately after obtaining the giga-seal, the recording chamber was perfused with solution A (in mM): 110 NaCl, 25 Na-gluconate, 5.4 KCl, 0.4 KH2PO4, 0.33 NaH2PO4, 0.8 MgSO4, 2.2 CaCl2, 10 glucose, 20 HEPES, pH 7.4 with NaOH. After the access resistance declined to 5–15 MΩ (less than 10 min), the membrane potential was recorded in current-clamp mode. Exclusion of Lucifer yellow fluorescence from the patched cells confirmed that the perforated patch remained intact throughout the experiment. Membrane potential was determined in resting cells and then during stimulation by superfusion with 0.3 μM carbachol. The arithmetic mean of the membrane potential was computed when sustained oscillations occurred during stimulation periods (excluding the initial “spike”).
K+ current measurements.
Whole-cell and single-channel patch- clamp recordings were done with an Axopatch 200B amplifier. Data acquisition was performed using a 12-bit analog/digital converter controlled by a personal computer. The current records were filtered at 5 kHz. Whole-cell patch pipettes were constructed from GC-150 glass (Warner Instruments) with resistance values between 4 and 6 MΩ. The pipette (internal) solution was 135 mM K-glutamate, 10 mM HEPES (pH 7.2), 5 mM EGTA, and with CaCl2 added to establish various Ca2+ concentrations (see also http://www.stanford.edu/∼cpatton/maxc.html). The external solution for whole cell patch recordings consisted of (in mM): 135 Na-glutamate, 5 K-glutamate, 2 CaCl2, 2 MgCl2, and 10 HEPES (pH 7.2). The use of glutamate instead of Cl− effectively eliminates Cl channel currents. The measured relevant junction potential in these recordings was less than 4 mV, sufficiently small that no correction was made.
Single-channel currents were obtained from inside-out patches with electrodes constructed from quartz (Garner Glass) and coated with sticky wax. The electrode tips were about 1–2 μm in diameter, and the current records were filtered at 2 kHz. These single-channel experiments used an external (pipette) solution that consisted of (in mM) 135-K glutamate, 2 CaCl2, 2 MgCl2, and 10 HEPES (pH 7.2). The internal (bath) solution was the same as used for the whole cell experiments.
Cl− current measurements.
Cl− currents were recorded in whole cell configuration using a PC-501A amplifier (Warner Instruments, Holliston, MA). Pipettes fabricated with Corning 8161 glass had a resistance of 2–4 MΩ when filled with the internal solution. Calcium-activated Cl− currents were recorded from cells bathed with a solution containing (in mM): 139 TEA-Cl, 20 HEPES, 0.5 CaCl2, and 100 d-mannitol (pH 7.3). TEA was used as the monovalent cation which essentially eliminates K channel currents. To test the Cl− dependency of the whole cell current, 139 mM bath TEA-Cl was replaced with equimolar TEA-glutamate. The intracellular solution contained (in mM): 80 NMDG-glutamate, 50 NMDG-EGTA, 30 CaCl2, and 20 HEPES (pH 7.3). This latter solution contained an estimated free [Ca2+] of 250 nM (1). Currents were recorded from 2-s test pulses from −80 to +100 mV in 20-mV increments applied every 7 s. At the end of each test pulse, the voltage was repolarized to −80 mV for 700 ms. The holding potential was 0 mV. Blockade of the calcium-activated Cl− current by niflumic acid was assessed in cells dialyzed with an intracellular solution that contained (in mM): 9.7 TEA-Cl, 30 EGTA, 21 CaCl2, 20 HEPES (pH 7.3) and an estimated free [Ca2+] of 250 nM. In these experiments, currents were recorded using the voltage protocol described above except for the holding potential, which was set at −50 mV.
Intracellular [Ion] Measurements
Acinar cells were loaded with either pH- or Cl−-sensitive fluoroprobe by incubation for 15–20 min at room temperature with 2 μM BCECF-AM (7) or 1 mM SPQ (11), respectively. The fluorescence of dye-loaded acinar cells was monitored in a superfusion chamber mounted on a Nikon Diaphot inverted epifluorescence microscope interfaced with an Imago Sensicam (TILL Photonics, Pleasanton, CA).
BCECF-loaded acinar cells were excited at 490 and 440 nm, and emitted fluorescence was measured at 530 nm. Cells were superfused with a physiological, HCO3− containing solution B (in mM): 110 NaCl, 25 NaHCO3, 5.4 KCl, 0.4 KH2PO4, 0.33 NaH2PO4, 0.8 MgSO4, 1.2 CaCl2, 10 glucose, and 20 HEPES. When NH4Cl was used to monitor Na+-K+-2Cl− cotransporter activity, 30 mM NaCl was replaced with equimolar NH4Cl. Chloride salts were replaced with equimolar gluconate in Cl−-free solutions, and additional calcium was added to compensate for chelation. Solutions were gassed with 5% CO2 and 95% O2 for at least 30 min before the pH was adjusted to 7.4 with NaOH. Intracellular pH data were expressed as a fluorescence ratio F490/F440 (46).
SPQ-loaded cells were superfused with solutions A or B (see Membrane potential measurements or Intracellular pH, respectively) and excited at 360 nm and emitted fluorescence was measured at 510 ± 40 nm. HCO3−-free solutions were gassed with 100% O2. Intracellular [Cl−] was estimated by in situ calibration of the fluorescence, as previously described (11). The calibration solutions contained (in mM): 80 KCl, 70 K-gluconate, 10 glucose, 0.005 nigericin, and 0.01 tributyltin (pH 7.4). The [Cl−] was adjusted from 0 to 80 mM by replacement of KCl with K-gluconate.
Data analyses and presentation.
Reported values are the means ± SE for the number of acinar cells or aggregates examined. Statistical analyses were performed using Student's t-test; P values of <0.05 were considered statistically significant. All experiments were performed with three or more separate preparations. The figures show results from a single representative experiment unless otherwise noted.
Screening of Human and Mouse Parotid Gland RNA
As an initial step in defining the fluid secretion mechanism in human parotid glands, gene expression was screened using a targeted cDNA array slide. This custom-designed array contained probes for 187 secretion-related genes that encode for ion/water transporters (75 genes) and receptors/regulators (101 genes), proteins potentially involved in the fluid secretion mechanism, as well as 11 secretory protein genes (see methods). Table 2 shows representative examples of relevant genes expressed in the samples isolated from human and mouse parotid glands. Of the 75 probes on the array representing ion/water transporter proteins, 59 were detected in human parotid tissue, whereas 61 of the probes hybridized with the mouse parotid gland RNA samples. Of the 59 genes expressed in human parotid tissue, 51 were also expressed in mouse parotid glands (86%).
Important examples of ion transporter genes found in both species included the Na+-K+-2Cl+ cotransporter NKCC1, the anion exchanger AE2, the Na+/H+ exchanger NHE1, the Ca2+-dependent K channels KCNN4 and KCNMA1, and the Cl channels CLC2 and CFTR. Receptors detected in human or mouse parotid glands included the β2 adrenergic receptor and those linked to an increase in intracellular [Ca2+], such as muscarinic (M1-M5) and numerous P-type nucleotide receptors. An increase in the intracellular [Ca2+] is thought to be the primary signal responsible for activating fluid secretion. For that reason, it is significant that numerous Ca2+ regulatory genes were also observed such as plasma membrane and Serca Ca2+ pumps (PMCA2 and SERCA1, respectively), phospholipase C (PLCα and β), and the calmodulin/Ca2+-dependent kinase CamK2A. There do not appear to be any major differences in gene expression; thus these results demonstrate that the salivary glands from these two species express a similar set of ion transporter and regulatory proteins to generate fluid and electrolyte secretion.
Intracellular [Cl−] and Membrane Potential
The current secretion model states that fluid and electrolyte transport is driven by transacinar Cl− movement. This process requires the intracellular [Cl−] of acinar cells to be accumulated to a level greater than its electrochemical equilibrium. With a resting membrane voltage between −50 and −60 mV (see below) and an external Cl− concentration of 120 mM, the expected equilibrium values for intracellular Cl− would be 12 to 18 mM. To test whether the intracellular Cl− level in human parotid cells is greater than the equilibrium level, as required for chloride-based fluid secretion, we used the Cl−-sensitive dye SPQ to estimate the intracellular [Cl−]. From experiments like the one shown in A of Fig. 1, we found that the intracellular [Cl−] in human parotid acinar cells was 62.4 ± 2.5 mM (n = 5) in a HCO3−-free solution, four to five times the predicted equilibrium value for the intracellular [Cl−]. Changing the bath solution from a HCO3−-free to a HCO3−-containing solution did not significantly change the intracellular [Cl−] (n = 4: HCO3−-containing = 56.3 ± 5.5 mM Cl−; HCO3−-free = 57.7 ± 3.4 mM Cl−). Thus, as in other mammalian salivary gland cells (6, 11, 44, 54), including those from mice (8, 36), human parotid acinar cells possess mechanism(s) for concentrating the intracellular [Cl−] well above its electrochemical equilibrium.
Using the perforated-patch technique and an intracellular [Cl−] of 62 mM (as determined above, Fig. 1A), the membrane potential (Vm) was recorded in the current-clamp mode at rest and during stimulation with the muscarinic receptor agonist carbachol (0.3 μM; CCh). Fig. 1B shows an example of such an experiment. The average resting membrane potential under these conditions was −51 ± 2 mV (n = 19), approximately midway between the equilibrium potentials for K and Cl ions (−81 and −17 mV, respectively). Carbachol stimulation produced a rapid depolarization that approached the Cl− equilibrium potential in nearly all of the cells (−27 ± 2 mV, 17 out of 19), suggesting the opening of Cl channels. This initial transient depolarization was followed by a hyperpolarizing shift, consistent with activation of K channels. The magnitude of the hyperpolarizing shift was independent of the oscillation behavior observed in ∼60% of the cells (oscillating cells: −64 ± 2 mV, n = 11; nonoscillating cells: −63 ± 2 mV, n = 8; P > 0.7).
Na+-Dependent Cl− Uptake Mechanisms
The observation that the intracellular [Cl−] of human parotid acinar cells is four or fivefold greater than its electrochemical equilibrium (Fig. 1) indicates that these cells express a mechanism for concentrating Cl− and is consistent with the prediction that fluid and electrolyte secretion is driven by transacinar Cl− movement. The two Cl− uptake mechanisms previously described in rodent salivary gland acinar cells, Na+-K+-2Cl− cotransport and paired Na+/H+ and Cl−/HCO3− exchange (5, 32) were detected by cDNA array analysis in human parotid tissue (Table 2). To test for the functional presence of these three electroneutral ion transport mechanisms in human acinar cells, the intracellular pH-sensitive dye BCECF was used to monitor the activity of these transporters.
Na+-K+-2Cl− cotransporter activity was examined by monitoring the transport of the K+ surrogate NH4+ (9) in a HCO3−-containing solution. Fig. 2A shows that addition of NH4Cl caused a very rapid intracellular alkalinization, as uncharged NH3 equilibrated across the plasma membrane, consuming intracellular protons and raising the intracellular pH. Subsequently, the intracellular pH decreased more slowly as NH4+ entered the acinar cell primarily via the Na+-K+-2Cl− cotransporter. The muscarinic receptor agonist carbachol (CCh, 0.5 μM) was used to enhance cotransporter activity (9). The agonist-induced acidification was blocked greater than 90% by the specific Na+-K+-2Cl− cotransport inhibitor bumetanide (100 μM), such that the rate of acidification was not significantly different from that observed in unstimulated acinar cells (Fig. 2, A and B). Bumetanide had no significant effect on the acidification rate of resting cells (in the absence of CCh), suggesting that during unstimulated conditions, the Na+-K+-2Cl− cotransporter is relatively inactive (Fig. 2B). We have previously shown that the mouse salivary acinar cell Na+-K+-2Cl− cotransporter is encoded by the Slc12a2 gene (8). Northern blot analysis (Fig. 2B, inset) detected transcripts consistent with the expected size of the human and mouse transcripts from this gene. These results confirm the array data (Table 2) and demonstrate the presence of Na+-K+-2Cl− cotransporter NKCC1 transcripts in both human and mouse parotid glands.
Na+/H+ and Cl−/HCO3− exchangers.
Paired Na+/H+ and Cl−/HCO3− exchangers also contribute to Cl− uptake in many, but not all, salivary glands. For example, human labial (40) and rat sublingual (55) acinar cells do not express detectable anion exchanger activity, whereas mouse sublingual (36) and submandibular (23) acini and rat parotid acinar cells (31) express robust Cl−/HCO3− exchanger activity. Functional and molecular support for expression of paired Na+/H+ and Cl−/HCO3− exchangers in human parotid acinar cells is shown in Fig. 3. In Fig. 3A, acinar cells were acid loaded by the NH4Cl prepulse technique (7) in a HCO3−-containing solution. This maneuver stimulated a Na+-dependent (left) and EIPA-sensitive (right) intracellular pH recovery, consistent with a Na+/H+ exchange mechanism. There appeared to be little functional Na+/HCO3− cotransporter activity in human parotid acinar cells because EIPA blocked nearly all of the recovery from an acid load in the HCO3−-containing solution (initial rate inhibited 92 ± 3%, n = 5). The effectiveness of the relatively low concentration of the specific Na+/H+ exchange inhibitor used in these studies (2 μM EIPA) suggests expression either the NHE1 or NHE2 isoform (53). We have previously demonstrated that Nhe1 is the dominant EIPA-sensitive Na+/H+ exchanger in mouse salivary acinar cells and is encoded by the Slc9a1 gene (7, 38). The Northern blot shown in Fig. 3B (right) verifies that Na+/H+ exchanger NHE1 transcripts are expressed in human parotid glands (see also Table 2).
The DIDS-sensitive anion exchanger AE2 that is encoded by the Slc4a2 gene is considered to be the most likely basolateral Cl−/HCO3− exchanger expressed in salivary acinar cells (8, 36). Functional evidence for a Cl−/HCO3− exchange mechanism is demonstrated in Fig. 3B (left). Here, the removal of extracellular Cl− produced a DIDS-sensitive alkalinization (300 μM DIDS; 45.3 ± 7.3% inhibition, n = 6), indicative of Cl−/HCO3− exchanger activation. Consistent with this prediction, Northern blot analysis (Fig. 3B, middle) shows that appropriate-sized message for the anion exchanger AE2 is expressed in human parotid glands and thus also confirms the array results (Table 2). However, the DIDS concentration used in these experiments would be expected to block more than 90% of AE2 anion exchanger activity. Therefore, AE2 is not likely the only anion exchanger in human parotid acinar cells. Several members of the SLC26A gene family, some of which have been reported to act as anion exchangers, were expressed in human parotid glands (Table 2).
Ca2+-Dependent Cl− and K+ Currents
The membrane potential (Vm) of human parotid acinar cells at rest (−51 mV) and during stimulation (−63 mV) was approximately midway between the equilibrium potentials for K and Cl ions (−81 and −17 mV, respectively; Fig. 1B). These results suggest that both K+ and Cl− currents contribute to the Vm during resting and stimulated conditions. Moreover, these currents are likely due to the activation of Ca2+-gated Cl and K channels (1, 35).
Ca2+-activated Cl− currents.
The current secretion model states that fluid production requires transacinar Cl− movement and is thus associated with Cl− efflux across the apical membrane. The model further predicts that a Ca2+-gated Cl channel is the source of this efflux (5, 32). In agreement with this model, electrophysiological experiments performed in human parotid acinar cells confirmed the presence of a Ca2+-activated Cl− current. Fig. 4A, left, shows time-dependent, outwardly rectifying Cl− currents in response to 2-s voltage pulses in cells dialyzed with ∼250 nM intracellular [Ca2+]. Large tail currents were seen when the Vm was changed to a potential of −80 mV at the end of the test pulse. In contrast, no current was recorded in cells dialyzed with a Ca2+-free solution (not shown; 20 mM EGTA and 0 Ca2+; n = 3), suggesting that these currents were due to activation of a Ca2+-dependent Cl channel and that relatively little voltage-activated Cl− current is present in human parotid acinar cells. Moreover, the outward currents shown in Fig. 4A, left were nearly abolished and the reversal potential shifted +49 ± 26 mV (n = 3) in acinar cells bathed in 139 mM glutamate/1 mM chloride (Fig. 4A, right), indicating that the current was Cl− selective. Further support for the presence of Ca2+-activated Cl channels was obtained using niflumic acid (NFA), a chloride channel antagonist, which is relatively specific for this channel type in salivary gland acinar cells (30). The Ca2+- and time-dependent Cl− currents observed at positive voltages were blunted by 100 μM NFA. Fig. 4B shows current-voltage relationships obtained before (solid squares) and after (open circles) exposure to NFA (n = 4). The Ca2+-dependent Cl− current measured at the end of the 2-s voltage step to +100 mV was blocked 88 ± 2% by 100 μM niflumic acid. The above properties are hallmarks of Ca2+-gated Cl channels (1, 16, 22, 30). Recent results indicate that the molecular identity of the channel responsible for the Ca2+-gated Cl− current may be a member of the BEST gene family (16, 17). Consistent with this possibility, Northern blot analysis detected BEST2 transcripts in both human and mouse parotid tissues (Fig. 4C). However, the array probe detected BEST2 message in all mouse samples but failed to detect significant levels of BEST2 transcript in three out of the four human samples. This array probe was generated from a mouse BEST2 cDNA, which was 77% identical to the homologous human sequence; this difference in sequence likely explains the less robust BEST2 signal for human samples when using the standard array hybridization protocol. Indeed, optimization of the Northern blot analysis conditions for this BEST2 probe detected transcripts without difficulty in human parotid tissue.
Ca2+-dependent K+ currents.
The hyperpolarization of the membrane potential seen in Fig. 1 during muscarinic stimulation was likely caused by activation of a K+ conductance, thus maintaining the electrochemical driving force for apical Cl− efflux. Two distinct types of Ca2+-dependent K+ currents are generally detected in mammalian salivary gland acinar cells (35). Fig. 5 demonstrates that human parotid acinar cells express both instantaneous, clotrimazole-sensitive (A) as well as time- and voltage-dependent, paxilline-sensitive (B) Ca2+-activated K+ currents. These results suggest that little, if any, other voltage- or Ca2+-dependent K+ current is expressed in human parotid acinar cells. The unitary conductance of the paxilline-sensitive current was 162 pS, whereas the single-channel conductance of the clotrimazole-sensitive Ca2+-dependent K+ current was 22 pS (Fig. 5C). Mouse parotid acinar cells express two types of Ca2+-activated K channels with these exact properties (Refs. 3 and 41). We have previously shown that the large-conductance, paxilline-sensitive, voltage- and time-dependent channel in mouse parotid acinar cells is encoded by the Kcnma1 gene (41). The smaller-conductance, clotrimazole-sensitive channel is encoded in mouse parotid acinar cells by the Kcnn4 gene (3). Northern blot (Fig. 5D) and array (Table 2) analyses confirmed the expression of these two genes in both human and mouse parotid tissues.
Very little is known regarding the ion transport mechanisms or the corresponding genes involved in the fluid secretion process in human major salivary glands, that is, the parotid, submandibular, and sublingual glands, because they are not readily available for study. Minor labial gland biopsies have yielded some information (40, 48, 49), but as their name implies, they provide little tissue, and it is not clear that observations made on the minor glands can be generalized to the major salivary glands. In contrast, considerable experimental support for the fluid secretion mechanism has been amassed in the major salivary glands of several mammalian model systems. The genes that encode many of the water and ion transport proteins involved in this process have recently been confirmed in the mouse (3, 7, 8, 32, 41). Nevertheless, it is critical to define the fluid secretion process in human tissue to develop specific clinical strategies to treat salivary gland dysfunction. The present study was thus designed to determine which ion transport proteins are functionally expressed in human parotid acinar cells and to compare these with those expressed in the more experimentally accessible mouse parotid gland. An initial genomic screen using a cDNA array representing 187 “secretion” (ion/water transporter, receptor/regulatory and secretory protein) genes detected essentially the same inventory of proteins in both human and mouse parotid glands. The cDNA array detected 121 genes in human parotid tissue, whereas 135 of the probes hybridized with the mouse parotid samples. Of the genes expressed in human parotid glands, ∼80% were also expressed in mouse parotid glands (97 out of 121), consistent with human and mouse salivary glands relying on a similar set of ion transporter and regulatory proteins to generate fluid and electrolyte secretion.
A fluid secretion model is proposed in Fig. 6 on the basis of the results of our molecular and functional analyses of the human parotid gland and from other model systems (5, 32). This model predicts that transepithelial Cl− movement acts as the driving force for fluid secretion in human parotid acinar cells. Transepithelial Cl− movement requires that the intracellular [Cl−] is elevated above its electrochemical equilibrium and that the membrane potential remains more hyperpolarized than the Cl− equilibrium potential during stimulation to maintain the driving force for apical Cl− efflux. Indeed, the intracellular [Cl−] of human parotid acinar cells was ∼62 mM, more than fourfold higher than predicted from the Cl− electrochemical equilibrium (62 mM vs. the predicted 15 mM if passive diffusion were operative), and the membrane potential remained hyperpolarized during muscarinic receptor activation (−63 ± 2 mV). These functional measurements are similar to those previously made in other mammalian model systems (11, 41, 54), suggesting that human acinar cells rely on the same ion transport mechanisms to generate saliva by transepithelial Cl− movement. Although not shown in the secretion model (Fig. 6), numerous aquaporin (AQP) water channels are expressed in human salivary glands (14, 15, 33). For example, AQP5 has been localized to the apical surface of human and rat salivary gland acinar cells (15, 18, 28), where it has been demonstrated to play an important part in stimulated transcellular water movement in mouse salivary glands (21, 26). The importance of water permeability to salivary gland function is reflected in the number of aquaporin isoforms detected in human and mouse parotid tissues, including AQP5 (see Table 2; also positive by Northern blot analysis, not shown).
The model shown in Fig. 6 includes basolateral Na+-K+-ATPase (51), which pumps Na+ out of the cell at the expense of ATP hydrolysis and consequently creates a large inward-directed Na+ chemical gradient (see Table 2). Na+-dependent Cl− uptake mechanisms would necessarily be located in the basolateral membrane of acinar cells, where they exploit the Na+ gradient to elevate intracellular Cl− above its electrochemical equilibrium. Consistent with this model, we detected bumetanide-sensitive Na+-K+-2Cl− cotransporter activity and NKCC1 gene expression in human parotid acinar cells (Fig. 2). We previously found that a null mutation in Nkcc1 (SLC12A2) eliminated cotransporter activity in mice and reduced in vivo stimulated secretion greater than 60%, thus demonstrating that this gene encodes for the basolateral Na+-K+-2Cl− cotransporter in mouse salivary gland acinar cells (8). The residual saliva produced in Nkcc1 null mice has been associated with NaCl uptake (in exchange for HCO3− and H+) that is mediated by the paired Cl−/HCO3− and Na+/H+ antiporters. In agreement with this possibility, molecular and functional evidence in the current as well as prior studies in rodents (8, 36) suggests that the DIDS-sensitive anion exchanger AE2 (SLC4A2) is most likely responsible for much of this basolateral exchanger activity in acinar cells (Fig. 3). However, the anion exchanger activity was only modestly DIDS-sensitive in human parotid acinar cells; thus there is the distinct possibility that AE2 is not the exclusive anion exchanger in this tissue. In fact, transcripts for AE4 and several members of the SLC26A gene family, some of which can carry out anion exchange (45), were detected by microarray. The Na+/H+ exchanger NHE1 is the primary regulator of acinar cell intracellular pH, as verified in Nhe1-3 (Slc9a1, Slc9a2, and Slc9a3) null mice (7). NHE1 is likely to be functionally coupled to the anion exchanger activity. Indeed, Nhe1−/− mice secrete significantly less saliva (38), demonstrating the importance of the Na+/H+ exchanger Nhe1 in salivary gland function. Consistent with the functional significance of this Na+/H+ exchanger in human salivary acinar cells as well, Na+/H+ exchanger activity with an NHE1-like sensitivity to the amiloride-derivative EIPA (53), and NHE1 messenger RNA, were detected in human parotid glands.
Iwatsuki et al. (20) first demonstrated the presence of Ca2+-dependent K+ and Cl− conductances in rat and mouse salivary gland acinar cells, but this study did not determine the nature of these currents. Both muscarinic and P-type nucleotide receptors are coupled to an increase in [Ca2+]i in human parotid acini (4). This increase in [Ca2+]i is thought to trigger the activation of both K and Cl channels involved in fluid secretion. In agreement with this premise, we found in human parotid acinar cells that intracellular Ca2+ stimulated a paxilline-sensitive and voltage-dependent, large-conductance K channel and a clotrimazole-sensitive, intermediate-conductance K channel, consistent with the detection of transcripts for KCNMA1 (maxi-K) and KCNN4 (IK1), respectively (Fig. 5). Similar to our results, a Ca2+- and voltage-activated K channel with a large-unit conductance of 160–165 pS (27, 37) and a Ca2+-activated intermediate K+ conductance (37) were previously detected in human salivary cells. However, in the present study, we did not detect a Na+-permeable current (27). In addition, an increase in the intracellular Ca2+ also stimulated a niflumic acid-sensitive Cl− current, and transcripts were identified by Northern blot analysis for the BEST2 Ca2+-gated Cl channel (Fig. 4). On the basis of the current literature, the BEST2 gene most likely encodes the Ca2+-dependent Cl channel expressed in salivary gland acinar cells (16, 17); however, there are other candidate Ca2+-gated Cl channel genes. Indeed, our microarray screen of the human and mouse salivary glands detected the expression of other BEST genes, as well as several members of the CLCA family of putative Ca2+-gated Cl channel genes (24).
In summary, the present study provides a comprehensive evaluation and confirmation of the ion transport proteins thought to be involved in the fluid secretion process in salivary gland acinar cells. Our results demonstrate that the ion transport mechanisms in human parotid glands are equivalent to those detected in mouse and most other mammalian salivary glands, and thus, confirm that animal models provide valuable systems for developing and testing clinical therapies to alleviate salivary gland dysfunction.
This work was supported in part by National Institutes of Health Grants DE09692, DE08921 (J. E. Melvin), and DE016960 (to T. Begenisich).
The authors thank Jill Thompson, Laurie Koek, Jennifer Scantlin, Mark Wagner and Pam McPherson for expert technical assistance. We also thank Dr. John Coniglio for his assistance in obtaining human parotid tissue.
↵* T. Nakamoto, A. Srivastava, and V. G. Romanenko contributed equally to this article.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society