| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
MUCOSAL BIOLOGY
1Division of Biomedical Sciences, University of California, Riverside, California; 2Department of Medicine, University of North Carolina, Chapel Hill, North Carolina; and 3Department of Gastroenterology, Hepatology and Endocrinology, Medical School of Hannover, Hannover, Germany
Submitted 2 March 2005 ; accepted in final form 26 April 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
mouse; NKCC1; cystic fibrosis transmembrane conductance regulator; CFTR; pepsin
The undisputed origin of gastric acid is the parietal cell (32). When stimulated, the parietal cell secretes HCl through a concerted operation of apical membrane H+/K+ exchange pumps, a Cl– exit pathway (possibly ClC-2), and a K+ recycling pathway (probably Kir4.1, Kir2.1, and/or KCNQ1), together with basolateral mechanisms for Cl– uptake and H+ uptake (HCO3– export). It is generally accepted that functional coupling between apical HCl exit and basolateral HCl uptake is accomplished, largely if not wholly, by a DIDS-sensitive Cl–/HCO3– exchange process (54, 59) that displays a high transport capacity (74) and strong allosteric activation by intracellular alkalinity (68). This process appears to be mediated by anion exchanger-2 (AE2), the main anion exchanger isoform in the stomach (37, 64, 73). In the mouse, AE2 is essential not only for gastric acid secretion but also for normal parietal cell development (22). The nonacidic component of Cl– secretion has been observed in a variety of gastric preparations as a transmucosal movement of Cl– in excess of H+ or a short-circuit current (Isc) that is stimulated by histamine and inhibited by serosal replacement of Na+ or by addition of serosal barium (7, 12, 19, 47, 52, 66). Recent experiments on the perfused rat stomach suggested that the nonacidic component of Cl– secretion is independent of proton pumping and is stimulated, together with HCO3– secretion, by PGE2 and luminal acidification (12).
Na-K-2Cl cotransporter-1 (NKCC) is a membrane glycoprotein that harnesses the chemical gradients of Na+ and Cl– to electroneutrally move salt and osmotically obliged water into the cell. In many fluid-secreting epithelial cells, basolaterally situated NKCC mediates salt uptake in concert with Na+-K+-ATPase and K+ channels for the purpose of transepithelial Cl– secretion (45, 65). Exceptionally high levels of NKCC have been found in the gastric mucosa of a variety of animal species (18, 40, 49, 60, 71). In a previous study (49) of the rat stomach, we established that NKCC is localized to parietal cells inhabiting the base of glands in the gastric corpus and to mucous gland cells forming the extreme base of antral glands.
The breadth of functions that Na-K-2Cl cotransport might play in gastric secretion has been poorly understood. Previous studies of amphibian (71) and guinea pig (5) gastric mucosa concluded that NKCC is involved in acid secretion based on evidence that a major component of HCl output is dependent on serosal Na+ and is blocked by NKCC inhibitors such as bumetanide or furosemide. However, a recent study with NKCC-deficient mice indicated that Na-K-2Cl cotransport is not required for the stomach to acidify its contents after stimulation with histamine (18). In addition, studies on chambered gastric mucosa isolated from the eel (76) and bullfrog (28) and on isolated rabbit gastric glands (50) have found no inhibitory effect of bumetanide on gastric acid secretion.
We have proposed that gastric NKCC might instead contribute importantly to nonacidic electrolyte secretion (49). To explore this hypothesis, we compared gastric H+, Na+, K+, Cl–, pepsinogen, and fluid secretions in NKCC-deficient mice and their normal littermates using the pylorus ligation technique. To assess the electrogenic character of NKCC-dependent ion flow, we measured the effect of bumetanide on Isc and H+ secretion by chambered mucosal sheets isolated from the mouse gastric corpus. Our results indicate that NKCC contributes importantly to both nonacidic electrogenic ion secretion and pepsinogen secretion in the mouse stomach.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Mouse lines.
The NKCC-deficient mice used in this study were generated by targeted gene disruption as described previously (57). This Slc12a2
506–621 mutant manifests several developmental and physiological defects; in addition to impairments in intestinal (26) and airway (24, 27) Cl– secretion, these mice exhibit growth retardation, deafness, ataxia, abnormal spermatogenesis (57), and inner ear histopathology (58). Studies on a different NKCC-deficient mouse line (18) have revealed similar phenotypic alterations, along with reduced mean arterial pressure (18, 53) and impaired saliva secretion (17). Measurements were performed on litter-matched male and female adult (2–8 mo of age) mice bred on a heterogeneous strain background (C57Bl/6J + DBA/2J). Mice were housed with a 12:12-h light-dark cycle with access to standard mouse chow and water ad libitum and euthanized by CO2 asphyxiation. All animal studies were approved by the University of California Riverside Institutional Animal Care and Use Committee.
We also compared gastric secretion in cystic fibrosis transmembrane conductance regulator (CFTR)-null (CFTRtm1UNC) mice with their heterozygous (CFTR+/–) littermates. CFTRtm1UNC mice were bred in a B6D2/129 genetic background as described previously (70). Both CFTR–/– and control CFTR+/– mice were maintained on standard chow and Colyte solution (Schwarz Pharma) ad libitum. Colyte, an osmotic laxative composed of 6% polyethylene glycol and electrolytes, prevents intestinal impaction in CFTR-deficient mice maintained on a solid diet without altering intestinal histomorphology (11).
Antibodies.
For immunolocalization of NKCC, we employed affinity-purified rabbit antibodies directed against the carboxy-terminal domain (N1c) or the amino-terminal domain (NT) of human NKCC. Pepsinogen was labeled with a rabbit antiserum directed against porcine pepsinogen. For immunolocalization of AE2, we employed a previously characterized (73) affinity-purified rabbit antibody (SA6), which was kindly supplied by Drs. A. Stuart-Tilley and S. Alper (Beth Israel-Deaconess Medical Center). A rabbit antibody (R3195) raised against the 13-amino acid carboxy terminus of rodent CFTR (20) was provided by Dr. C. Marino (University of Tennessee). Parietal cells were identified by labeling with mouse antibody 2G11 against H+-K+-ATPase 
-subunit (Affinity Bioreagents) or with rhodamine-conjugated Dolichos biflorus lectin (Vector Laboratories).
Gastric gland isolation.
Glands were isolated from mouse gastric corpus for confocal immunocytochemistry by a modification of the collagenase method originally described by Berglindh and Obrink (9). Stomachs were excised, opened along the greater curvature, and pinned to a wax dissection plate on ice. The muscle layers overlaying the corpus region of the stomach were separated from the mucosa by a blistering technique (19). We injected oxygen-saturated gland medium (HCO3–-free minimal essential medium supplemented with 20 mM HEPES, pH 7.4) at multiple sites just below the mucosa using a 27-gauge needle; the muscularis externa was then dissected away with fine scissors. The mucosal sheet was minced into small pieces (
2 mm3) with a razor blade, washed with ice-cold gland medium, and transferred to a 50-ml flask containing 15 ml gland medium, 5 mg collagenase type-1A (Sigma C9891), and 15 mg BSA. The flask was incubated at 37°C with orbital mixing and continuous oxygenation. After 20–30 min, the suspension was diluted 1:1 with gland medium and the mucosal fragments were dispersed by gentle pipetting. Glands were attached to polylysine-coated glass slides and circumscribed with a thick hydrophobic ring (ImmEdge pen, Vector). The ring was filled with 0.3 ml of medium and incubated on a thermostated (37°C) aluminum block mounted inside an oxygenated humidified chamber atop an oscillating table (Thermolyne Roto-Mix, 30 rpm). After 20–30 min, the slide was submerged in neutral-buffered formalin for 30 min, rinsed in PBS containing azide, and stored at 4°C for up to 3 wk.
Tissue preparation. NKCC-deficient mice and their wild-type littermates were given free access to food and water. After CO2 asphyxiation, the stomach was isolated, rinsed with PBS, and immersed in ice-cold PLP fixative (2% paraformaldehyde, 75 mM lysine, 10 mM sodium periodate, and 45 mM sodium phosphate, pH 7.4) (48). One hour later, the tissue was transferred to fresh fixative and incubated for an additional 3 h on ice. The tissue was stored for up to 3 mo in cryoprotectant (30% sucrose in PBS). For sectioning, tissues were embedded in freezing medium (Triangle Biomedical Sciences) and frozen at –25°C. Sections of 5 µm thickness were cut on a cryostat microtome (HM 500 OM, Microm Lamborgeraete) and mounted on glass slides (Superfrost Plus, Fisher). Every tenth consecutive slide was stained with hematoxylin and eosin for histological assessment.
Immunoperoxidase staining. After antigen retrieval (1% SDS in K+-free PBS for 10 min), sections were incubated sequentially with blocking solution (PBS containing 20% goat serum, 0.2% BSA, and 25 mM glycine, pH 7.4), biotin blocker (Vector, 10 min), avidin blocker (Vector, 10 min), peroxidase inhibitors (PBS containing 1% H2O2 and 0.05% sodium azide, 15 min), primary antibody (2 h), and biotinylated secondary antibody. Bound antibody was detected with a peroxidase-conjugated streptavidin system (Vectastain Elite ABC kit, Vector Laboratories). Vector-SG or Vector-DAB were employed as substrate with eosin or hematoxylin counterstaining, respectively.
Immunofluorescence labeling was performed on 5-µm cryosections of PLP-fixed tissues as described previously (49) or on formalin-fixed glands attached to glass slides. Tissue was exposed successively to antigen retrieval solution (1% SDS in K+-free PBS, 10 min), blocking solution (5% normal goat serum, 5% BSA, 25 mM glycine, and 0.2% Triton X-100, in PBS, pH 7.4, 40 min), primary antibodies (diluted in blocking solution, 1 h), and fluorophore-conjugated secondary antibodies (diluted in PBS), with three rinses between each step. Coverslips were mounted over Vectashield medium (Vector). Fluorescence images were captured with a Zeiss LSM-510 confocal microscope and assembled with Adobe Photoshop software.
Pylorus ligation was performed as described previously (55, 69). Mice were fasted for 15 h in wire-mesh cages (to prevent caprophagia) with free access to tap water. After mice were anesthetized with 2% isoflurane vapor in oxygen (Vasco), gastric secretion was stimulated by administration of 16 µg/kg iv pentagastrin in sterile saline or 2 mg ip histamine in 200 µl of sterile saline. After laparotomy, the pylorus was ligated with 3-0 silk. To counter postsurgical dehydration, the peritoneal cavity was infused with 1 ml (8% of total body water) of a sterile physiological salt solution, and the abdominal incision was repaired with 3-0 silk sutures. Two hours later, the mice were anesthetized, the esophagus was ligated, and the stomach was removed. Its contents were collected in preweighed tubes, weighed, and centrifuged (5 min, 12,000 g). Occasionally, samples were discolored with small amounts of blood, and these data were discarded. The volume of supernatant was determined gravimetrically, and its pH was measured with a miniature spear-tip electrode (Orion). Acid equivalents were measured by back titration to pH 7.0 with 10 mM NaOH. K+ and Na+ were measured by flame photometry (Beckman Klina-Flame) and Cl– by coulometric titration (Buchler-Cotlove). The pellet, consisting mainly of insoluble mucus, was weighed before and after drying for 12 h at 60°C. The volume of gastric fluid was calculated from the sum of the supernatant mass and pellet volatile mass.
The NKCC–/– mice used in this study were 33% smaller than their sibling controls, like those used in previous studies (26, 57, 58); however, their empty stomachs averaged only 15% smaller (34.9 ± 2.2 mg for NKCC–/– mice compared with 40.9 ± 2.5 mg for age-matched NKCC+/+ mice; n = 16). To compensate for these size differences, secretion data were normalized to stomach dry mass, determined by weighing the empty stomach after drying for 12 h at 60°C.
We measured pepsin activity using a standard hemoglobin digestion method (3). Briefly, 100 µl of 10-fold diluted gastric juice were mixed with 400 µl of 2.5% hemoglobin and 100 µl of 0.3 N HCl and incubated at 37°C. Under these acidic conditions, all secreted pepsinogen is converted to pepsin. After 10 min, 1 ml of 0.3 N TCA was added, and the samples were mixed and filtered using Whatman no. 3 paper. The absorbance of the filtrate was measured at 280 nm (Lambda 3A UV/VIS spectrophotometer, Perkin-Elmer), and the pepsin activity was calculated from a standard curve obtained with 100 µl of 5, 10, or 20 µg/ml purified porcine pepsin in 10 mM HCl. The quantity of pepsin accumulated in the gastric juice 2 h after pylorus ligation was expressed as micrograms of pepsin per milligram of dry empty stomach.
Hydration measurements. Total body water was measured gravimetrically as described previously (10). NKCC-deficient mice and their sibling controls were allowed free access to food and water. After mice were euthanized with pentobarbital sodium, a midline incision was made through the abdominal wall and chest cavity. Samples of blood and urine were collected from the exposed heart and bladder with syringe needles. The gastrointestinal tract was removed, cleared of its contents by rinsing with PBS, and returned to the carcass. The carcass was then weighed, dried to a constant mass (60°C for 7 days), and weighed again. The ratio of volatile mass to total mass was recorded as fractional body water. Plasma and urine osmolalities were measured with a vapor pressure osmometer (Wescor 5500).
Electrophysiology and H+ transport measurements.
The mucosal layer was dissected from the mouse gastric corpus under a stereomicroscope and mounted between two Lucite half chambers of a water-jacketed Ussing system equipped with a gas-lift system. The exposed surface area was 0.625 cm2. The serosal solution contained (in mM) 108 NaCl, 22 NaHCO3, 3 KCl, 1.3 MgSO4, 2 CaCl2, 1.5 KH2PO4, 8.9 glucose, and 10 sodium pyruvate and was gassed with 95% O2-5% CO2, pH 7.4. Indomethacin (1 µM) and tetrodotoxin (0.1 µM) were included in the serosal solution to minimize variation due to intrinsic prostanoid and neural tone, respectively. The mucosal solution (140 mM NaCl) was gassed with 100% O2 and maintained at pH 7.4 by the controlled addition (in 0.1-µl steps) of dilute (1 mM) NaOH using a pH-Stat titration system (Radiometer, Copenhagen, Denmark). Transepithelial voltage (Vt) and resistance (Rt) and acid secretion were monitored under open-circuit conditions with series resistance compensation using a current-voltage clamp device (World Precision Instruments, Sarasota, FL). Vt was measured with Ag/AgCl electrodes connected to the chambers via 3 M KCl bridges and referenced to the serosal side of the epithelium. Rt was measured by recording voltage deflections resulting from periodic current pulses (25 µA, 200 ms) passed across the tissue. Equivalent Isc was calculated from Ohm's law (Isc = Vt/Rt). Acid secretion, measured under open-circuit conditions, was calculated from the amount of NaOH needed to maintain the mucosal bath at a constant pH of 7.4. Unlike many previously described in vitro mammalian preparations, mouse gastric mucosa prepared in this manner remained stable and highly responsive to secretory stimuli for several hours, with acid secretion rates (7 µmol·cm–2·h–1) comparable with those observed in the stimulated lumen-perfused mouse.
Statistics. Data are presented as means ± SE. The significance of the difference between means of paired data was assessed by Student's t-test. Differences were regarded as significant at P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
400 µm) than in the distal corpus and greater curvature (where the glands are shortest, 175 µm) (51). The overall distribution of AE2 in normal mice and NKCC-deficient mice was similar (Fig. 2, d and h), indicating that the developmental absence of one Cl– entry pathway (NKCC) is not associated with a compensatory overexpression of another (AE2) in parietal cells.
|
The only other region of the mouse stomach in which very high NKCC expression was detected was the antrum (Fig. 3). Labeling was exceptionally strong in the mucous cells (44) that form the extreme base of the antral gland. These same cells contained, in addition to a basolateral pathway for rapid Cl– entry (NKCC), apical pathways for Cl– exit (CFTR) and for water movement [aquaporin-5 (AQP5)] (Fig. 3). Another apical constituent of these cells was pepsinogen (Fig. 3). Genetic ablation of NKCC produced no conspicuous alterations in antral histology, CFTR expression, or pepsinogen expression by antral mucous gland cells (not shown).
|
In a series of early experiments, we found that peritoneal administration of histamine (3 mg) elevated acid secretion from 0.3 ± 0.1 to 1.6 ± 0.4 µmol·mg–1·2 h–1 and fluid secretion from 10.5 ± 1.9 to 19 ± 2.7 µl·mg–1·2 h–1 (n = 5). Comparable responses were observed after intravenous injection with pentagastrin (16 µg/kg): acid secretion increased to 1.1 ± 0.3 µmol·mg–1·2 h–1 and fluid secretion increased to 21 µl·mg–1·2 h–1; thus pentagastrin was used in all subsequent in vivo experiments as a standard gastric secretagogue. Basal secretory tone, i.e., in the absence of any external stimulus, was not routinely measured, because pylorus ligation itself is known to elevate gastric acid secretion through a gastrin-independent vagovagal reflex (1, 29).
Whereas genetic ablation of NKCC had no effect on the secretion of proton equivalents, other components of gastric secretion were markedly impaired: NKCC-deficient mice secreted 38% less Na+, 70% less K+, 27% less Cl–, and 32% less fluid over a 2-h interval than their sibling controls (Fig. 4). Measurements of the major cations (H+ + Na+ + K+) and the main anion (Cl–) in gastric fluid indicated that these cations amounted to 78% of the Cl– in both control and NKCC-deficient mice. The source of this small "cation gap" is not known. The amount of Cl– secreted in excess of H+, a component customarily designated "nonacidic" Cl– secretion, was substantially reduced in NKCC-deficient mice (Fig. 4). Thus, although NKCC is not required for HCl secretion, it plays a critical role in the secretion of Na+, K+, and "nonacidic" Cl–, along with the fluid osmotically associated with these ion flows. Gastric juice collected from pentagastrin-treated mice 2 h after pylorus ligation was slightly hyposmotic to plasma in control mice (290 ± 6.6 mosmol/kgH2O, n = 10) yet nearly isosmotic in NKCC-deficient mice (319 ± 6.3 mosmol/kgH2O, n = 7). The osmolality of plasma collected from control mice (332 ± 3 mosmol/kgH2O, n = 5) and from NKCC-deficient mice (333 ± 12 mosmol/kgH2O, n = 5) did not differ.
|
|
|
Electrogenic transport and acid secretion by isolated gastric mucosa. The gastric corpus of most animal species actively secretes Cl– in substantial excess of H+ and absorbs Na+ through electrogenic mechanisms that give rise to a lumen-negative Vt (16, 21, 46, 66). To characterize these processes in the mouse model and their dependence on NKCC activity, we isolated the mucosal layer from the corpus region (to circumvent possible antral contributions) and monitored electrical parameters and acid secretion in an Ussing chamber (Figs. 7 and 8).
|
|
50% but did not alter acid output, indicating that a major source of current in the corpus epithelium is active amiloride-sensitive Na+ absorption. In other experiments, we noted that the amiloride-sensitive component of Isc became progressively smaller after 8-Br-cAMP stimulation (data not shown), suggesting that electrogenic Na+ absorption is downregulated under acid-secreting conditions, consistent with previous observations with piglet gastric mucosa (19). Addition of a concentration of bumetanide that fully inhibits Na-K-2Cl cotransport (100 µM) to the serosal bath decreased Vt and Isc further without affecting acid secretion. Finally, addition of DIDS reduced both Vt and Isc further and abolished acid secretion. It should be noted that the concentration of DIDS added (1 mM) would be sufficient to fully inhibit the three members of the SLC4 anion exchanger family (AE1, AE2, and AE3) (34, 41) known to exist in the rodent gastric mucosa (78) as well as SLC26A7, a Cl–/HCO3– exchanger (62) or channel (39) found in the basolateral membrane of parietal cells. A second series of experiments focused on the bumetanide-sensitive component of Isc (Fig. 8). Rather than waiting for the Na+ conductance to develop fully, we stimulated secretion as soon as a steady low rate of acid secretion was observed. Addition of the cAMP-dependent secretagogue forskolin evoked concomitant increases in acid output (Fig. 8A) and Isc (Fig. 8B). As in the previous series of experiments, inhibition of NKCC with serosal bumetanide reduced electrogenic ion flow to near basal levels without affecting acid secretion. Together, these results suggest that, in the mouse corpus, nonacidic Cl– secretion is an electrogenic process that involves Na-K-2Cl cotransport, whereas H+-linked Cl– secretion is a macroscopically electroneutral process that does not.
We conducted a limited set of preliminary experiments to assess whether the NKCC-mediated Cl– transport also contributes to Vt and Isc in the resting state. Exposing the unstimulated mouse corpus to luminal amiloride (10 µM) reduced both Vt and Isc significantly (data not shown). Serosal bumetanide (100 µM) also reduced Vt and Isc, although to a lesser extent. Thus NKCC-dependent Cl– secretion appears to be active in resting mucosa and further activated by cAMP-dependent stimulation.
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
The most probable source of this nonacidic electrogenic ion secretion is parietal cells. These cells, individually and collectively, are the most prominent site of NKCC expression in the gastric mucosa of most animal species thus far examined, including the mouse (Figs. 1 and 2), rat (49), and rabbit and human (C. Lytle, unpublished observations). The concept that NKCC functions as a dominant basolateral pathway for Cl– uptake in parietal cells has recently been substantiated by preliminary experiments on isolated rat gastric glands (42). Although mucous neck cells could be a second source of NKCC-dependent ion secretion in the mouse stomach, their contribution is probably of lesser importance because parietal cells outnumber mucous neck cells 5:1 (33); in some animal species, including the rat and rabbit, mucous neck cells are devoid of NKCC (Ref. 49 and unpublished observations).
Our finding that Cl– taken up by NKCC is not utilized for the purpose of acid secretion (HCl) corroborates earlier observations that gastric acidification is not impaired in NKCC knockout mice (18) and that histamine-stimulated aminopyrine accumulation by isolated rabbit gastric glands (50) and acid secretion by chambered eel gastric mucosa (76) are not inhibited by doses of bumetanide that abolish Na-K-2Cl cotransport. Contrasting results have been obtained in two previous studies with different animal models. First, in chambered mucosa isolated from the guinea pig stomach, serosal application of 1 mM furosemide inhibited acid secretion along with transmucosal current and Cl– secretion (5). This effect, however, seems unrelated to Na-K-2Cl cotransport because we have been unable to detect NKCC protein in guinea pig parietal cells (unpublished observations). Second, in gastric mucosa isolated from the amphibian Necturus, serosal application of 50 µM bumetanide blocked acid secretion (71). In Necturus, both acidic and nonacidic Cl– secretions appear to originate from the oxyntopeptic cell (21). Thus, unlike its mammalian counterpart, the amphibian oxyntopeptic cell appears to utilize Na-K-2Cl cotransport, rather than Cl–/HCO3– exchange, to import Cl– destined for secretion with H+. Evidently, this is not a feature of all amphibian oxyntopeptic cells, because acid secretion by bullfrog gastric mucosa was found to be insensitive to bumetanide (28).
When the corpus of the mouse stomach is separated from the antrum, stripped of its serosal layer, and incubated in an Ussing chamber, it generates a mucosal negative potential difference and an Isc comprising two major components: one inhibited by mucosal amiloride and the other by serosal bumetanide (Figs. 7 and 8). The amiloride-sensitive component, now recognized as electrogenic Na+ absorption, has been observed in gastric preparations from the rat (15), pig (19), monkey (75), and lizard (31). This absorptive process appears to be a property of mucus-secreting surface epithelial cells that line the corpus and antrum; the strongest evidence for spatial segregation of Na+ absorptive and Cl– secretory functions has come from electrophysiological studies of amphibian (Necturus) gastric mucosa (16). The second component of Isc largely reflects nonacidic Cl– secretion driven by Na-K-2Cl cotransport, as discussed below.
Our results support earlier speculation (38, 47, 49, 52, 76) that parietal cells function in two operationally distinct modes of Cl– secretion: one coupled to H+ secretion and the other nonacidic and electrogenic. What remains uncertain is whether both modes coexist in all parietal cells or whether they occur in spatially or functionally distinct populations. We (49) have proposed that HCl secretion predominates in parietal cells that migrate from the isthmus toward the gastric pit (which express abundant AE2 but no NKCC), whereas nonacidic Cl– transport predominates in parietal cells that migrate into the gland base (which express abundant NKCC but diminished levels of AE2). The concept that parietal cells nearest the surface are far more active in HCl secretion than those deep in the gland fits evidence that base parietal cells are morphologically unresponsive to acid secretagogues (35, 36, 38) with H+-K+-ATPase units that appear to be catalytically dormant (13) and refractory to physiological stimulation (14). This concept does not preclude the possibility that NKCC-expressing parietal cells in the neck and base retain some capacity, albeit diminished, to secrete acid. Indeed, acidification of the secretory canaliculus and the adjacent luminal compartment can be visualized with fluorescent dyes (2, 4, 8, 23, 61) in parietal cells inhabiting all segments of the rabbit gastric gland after stimulation with histamine. On the other hand, in preliminary experiments on isolated rabbit gastric glands, we have noted spatial and temporal differences in the onset, rate, and extent of histamine-stimulated acidification consistent with the idea that acid secretion in the gland base is generally weaker. Further experiments are needed to test the possibility that base parietal cells operate in two modes of Cl– secretion: acidic and nonacidic.
Another site of NKCC-dependent electrolyte and pepsinogen secretion might be the gland cells that inhabit the base of each antral mucous unit (44). Because they are richly equipped with pathways for basolateral Cl– entry (NKCC), apical Cl– exit (CFTR), and osmosis (AQP5), it seems likely that their principal function is electrogenic Cl– secretion. We reasoned that this component of Cl– secretion might be selectively absent from CFTR-deficient mice, given that antral base cells are the only detectable site of CFTR expression in the mouse stomach and that CFTR function is required for active Cl– secretion in similarly configured intestinal crypt cells (25). Indeed, a severe impairment in nonacidic gastric fluid secretion has been reported in patients with cystic fibrosis (30). However, we found no difference in the character of gastric secretion between mice with or without CFTR function. Evidently, antral contributions to gastric electrolyte secretion in the mouse are either negligible or not dependent on CFTR.
Our observation that pepsinogen secretion depends partly on NKCC is especially intriguing because chief cells do not themselves express the cotransporter. Evidently, the observed dependency involves events beyond the formation and exocytotic release of pepsinogen by the zymogenic cells that inhabit the neck and base of the gastric glands. We (49) have postulated that the NKCC-dependent secretion of neutral fluid by base parietal cells serves to facilitate bulk transport of pepsinogen from its site of secretion (base chief cells) to its site of action (stomach lumen). Thus the nonacidic secretory stream would not only flush pepsinogen through the narrow gland lumen but also delay its activation until the proenzyme reaches the highly acidic environment adjoining superficial parietal cells. This delayed activation would presumably safeguard deeper gland cells (which lack a protective apical mucous coating) from proteolytic attack. It follows from the "flushing" concept that impairment of the secretory stream would impede pepsinogen clearance from the gland lumen. As predicted, gastric pepsinogen secretion was reduced in NKCC-deficient mice. Impairment of pepsinogen clearance, in turn, might allow premature activation of pepsinogen followed by local proteolytic injury. However, no focal damage or inflammation was evident in NKCC-deficient mice. An alternative possibility is that the NKCC-dependent component of pepsinogen secretion originates from antral base cells, rather than chief cells, through processes that do not depend on CFTR.
If parietal cells situated deep in the gland base are in fact a significant source of gastric acid, the NKCC-dependent secretory stream might also serve to flush this acid from the gland lumen. However, elimination of this stream by genetic ablation or inhibition of NKCC did not impair acid secretion (Figs. 4, 7, and 8). This result suggests that luminal convection/diffusion does not limit the rate of H+ secretion, as surmised previously (63), and is compatible with the notion that acid is mainly secreted by superficial parietal cells in close proximity to the gastric pit.
Early studies on the gastric secretory responses to histamine and acetylcholine in humans and dogs showed that acid secretion is accompanied by a marked increase in K+ movement to the gastric fluid (77). The mechanisms and cellular sources of this K+ secretion remain poorly understood (21). It might reflect "slippage" in apical K+ recycling by parietal cells, i.e., imperfect functional coupling between the H+/K+ exchange pump and adjacent K+ channels. Our results indicate that in the mouse, gastric K+ secretion depends almost entirely on NKCC (Fig. 4). This implicates parietal cells or antral base cells as potential sources of K+ secretion. It has been suggested that diffusional restrictions within the gastric gland lumen limit the rate of K+ (but not H+) flux between the epithelial cells and the stomach lumen (63). If this proves correct, the impairment of bulk K+ secretion that we observed in NKCC-deficient mice could reflect a loss of NKCC-dependent convective flushing of the gland lumen.
Although NKCC-dependent Cl– secretion is not associated with H+ secretion, it is likewise stimulated by cAMP-dependent secretagogues such as histamine (76) and forskolin (Fig. 8). There is some evidence that it may be subject to regulation by other signals: in the rat gastric corpus, for example, peripheral-type benzodiazepine receptor agonists have been found to stimulate bumetanide-sensitive electrogenic Cl– secretion without influencing acid secretion (56). This mitochondrial receptor is expressed only in surface epithelial cells and basilar parietal cells of the rat stomach (56); the coexistence of this receptor and NKCC only in basilar parietal cells supports the notion that bumetanide-sensitive electrogenic Cl– secretion originates from this subset of parietal cells.
Together, our results suggest that parietal cells somehow selectively utilize the inward flow of Cl– via Na-K-2Cl cotransport (or the outward electrochemical Cl– gradient generated by this transporter) for the purpose of nonacidic electrogenic ion secretion.
|
GRANTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
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.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
F508 mutation in mouse cystic fibrosis transmembrane conductance regulator results in a temperature-sensitive processing defect in vivo. J Clin Invest 98: 1304–1312, 1996.[ISI][Medline]
) or was not (
) added to the serosal bath. Inhibition of NKCC was without significant effect on acid output. B: equivalent Isc monitored at 5-min intervals (n = 6).