ΔF508 Mutation Results in Impaired Gastric Acid Secretion*

The cystic fibrosis transmembrane conductance regulator (CFTR) is recognized as a multifunctional protein that is involved in Cl– secretion, as well as acting as a regulatory protein. In order for acid secretion to take place a complex interaction of transport proteins and channels must occur at the apical pole of the parietal cell. Included in this process is at least one K+ and Cl– channel, allowing for both recycling of K+ for the H,K-ATPase, and Cl– secretion, necessary for the generation of concentrated HCl in the gastric gland lumen. We have previously shown that an ATP-sensitive potassium channel (KATP) is expressed in parietal cells. In the present study we measured secretagogue-induced acid secretion from wild-type and ΔF508-deficient mice in isolated gastric glands and whole stomach preparations. Secretagogue-induced acid secretion in wild-type mouse gastric glands could be significantly reduced with either glibenclamide or the specific inhibitor CFTR-inh172. In ΔF508-deficient mice, however, histamine-induced acid secretion was significantly less than in wild-type mice. Furthermore, immunofluorescent localization of sulfonylurea 1 and 2 failed to show expression of a sulfonylurea receptor in the parietal cell, thus further implicating CFTR as the ATP-binding cassette transporter associated with the KATP channels. These results demonstrate a regulatory role for the CFTR protein in normal gastric acid secretion.

A defect in or absence of the cystic fibrosis transmembrane conductance regulator (CFTR) 2 protein is responsible for the autosomal recessive, multiorgan disease cystic fibrosis. Defective or deficient copies of the protein on the cell surface result in abnormalities in viscosity and electrolyte content of various exocrine secretions. To date more than 1000 different mutations in the CFTR gene have been identified and associated with cystic fibrosis, the most common being the ⌬F508 mutation. 70% of defective alleles are affected by this mutation causing a deletion of a phenylalanine residue at position 508. The defective protein is synthesized but recognized as abnormal and targeted for degradation. When defective ⌬F508 protein is placed in a cell-free lipid bilayer, it retains a substantial part of its function (1)(2)(3)(4)(5).
The CFTR protein is a cAMP-regulated chloride channel that belongs to the superfamily of ATP-binding cassette transporters. It contains two membrane-spanning domains, each consisting of six helices, as well as two nuclear binding domains responsible for ATP hydrolysis, and a regulatory domain with phosphorylation sites for various protein kinases. Both amino (N) and carboxyl (C) tails of the protein are oriented cytoplasmically and mediate CFTR interaction with other ion channels, receptors, and the cytoskeleton (6 -10). Initially thought to be merely a chloride conductance channel, the cystic fibrosis transmembrane conductance regulator, as the name implies, has been shown to regulate several other cellular processes. CFTR interacts and plays a regulatory role in the activity of the outwardly rectifying chloride channel (11)(12)(13)(14), epithelial sodium channel (15,16), as well as the renal outer medullary K ϩ channel type 2 (ROMK2 or Kir1.1b) (17)(18)(19)(20)(21).
The acid secretory cell of the gastric gland is the parietal cell. Stimulation by either hormonal (histamine and gastrin) or neuronal (acetylcholine) secretagogues via their corresponding receptors on the basolateral plasma membrane results in translocation and insertion of the H,K-ATPase from its inactive state in the cytoplasmic tubulovesicle elements to the apical membrane of the secretory canaliculi. Proton secretion then occurs by exchanging an H ϩ ion for a K ϩ ion at the expense of 1 ATP molecule via the H,K-ATPase. This proton secretion is coupled with the extrusion of a Cl Ϫ ion via an apical Cl Ϫ channel. To maintain the exchange of H ϩ for K ϩ it is necessary to have an apical K ϩ channel that secretes K ϩ into the lumen of the gland to provide a continuing K ϩ gradient for the H,K-ATPase. This cyclic K ϩ movement and the concurrent H ϩ and Cl Ϫ secretion leads to the generation of 0.1 N HCl in the lumen of the gland, which is then secreted into the interior of the stomach (22)(23)(24)(25)(26)(27)(28). The molecular identity of both the K ϩ channel and the Cl Ϫ channel on the apical surface of the parietal cell is still in question, however, recently a variety of K ϩ channel proteins have been identified in the parietal cell so that the secretion of K ϩ may either be associated with multiple channel proteins, or complexes of these different K ϩ channels that have been identified (29 -33). Based on the putative role of CFTR in modulat-ing ROMK2 activity (a protein identified in the parietal cells), we postulated a probable role for CFTR in gastric acid secretion.
Maldigestion is a well recognized problem in cystic fibrosis patients, generally ascribed to pancreatic exocrine insufficiency (34 -38). Decreased gastric acid secretion has not been directly implicated in the malabsorption of these patients. In fact, decreased pH of duodenal content due to impaired HCO 3 Ϫ secretion in these patients worsens the pancreatic insufficiency by decreasing survival of lipolytic activity of exogenous enzymes and precipitates bile acids causing malabsorption. Clearly the focus for patients to enhance nutrient absorption has been on the pancreas. In two studies, treatment with H2-receptor antagonists and proton pump inhibitors to decrease acid secretion and raise duodenal pH improved fat absorption in CF patients (39,40 (42)(43)(44).
It would also be reasonable to predict that patients with cystic fibrosis are at an increased risk for developing duodenal ulcers due to impaired HCO 3 Ϫ secretion and reduced mucosal defense. However, Akiba et al. demonstrated the contrary was true. He suggested that HCO 3 Ϫ -trapped intracellularly may protect the duodenal mucosa from acid injury and may explain why ulceration is a rare occurrence in CF patients (45)(46)(47). An alternative explanation could be reduced acidification of gastric content entering the duodenum. This has not been described as a possible protective factor. Alternatively, misregulation of the potassium recycling pathway by an absent or defective CFTR protein may prevent H,K-ATPase activity, also leading to elevations in gastric pH and reduced risk of ulcers.
Using intracellular pH measurements on parietal cells of isolated hand-dissected mouse gastric glands as described in previous studies by our laboratory (48 -59), we observed a significant reduction in acid secretion in cells treated with 100 M glibenclamide, a known CFTR channel inhibitor (60,61). CFTR-inh172 (62), a new specific inhibitor of the CFTR channel also markedly reduced acid secretion in streptolysin-O (SLO)-permeabilized mouse gastric glands (63). Furthermore, ⌬F508 knock-out mice demonstrated reduced parietal cell acid secretion from isolated glands as well as in whole stomach preparations when compared with their wild-type counterparts upon histamine stimulation. Our studies show that CFTR plays an important regulatory role in the mouse stomach in secretagogue-induced acid secretion.

EXPERIMENTAL PROCEDURES
Animals-Wild-type CD-1 and C57BL/6 mice were housed in climate-and humidity-controlled light-cycled rooms, fed standard rodent chow, and allowed free access to water. Data obtained from wild-type CD-1 and C57BL/6 mice are pooled in the results section. Gene-targeted mice homozygous for the ⌬F508 mutation (64), fed standard chow and maintained on Colyte solution (Schwarz Pharma) (65), were also used for the study. All animals were fasted with free access to water or water plus Colyte solution, for 12-18 h prior to sacrifice to ensure a consistent minimum of basal gastric acid secretion.

Isolation of Gastric Glands and Digital
Imaging for pH i -Wild-type and transgenic mice were sacrificed by overdose of isoflurane anesthesia. Following laparotomy a total gastrectomy was performed, and the stomach was immediately rinsed with ice-cold HEPES-buffered Ringer solution (125 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 1.2 mM MgCl 2 , 32.2 mM HEPES, and 5 mM glucose, pH 7.4) ( Table 1) to eliminate any residual food matter, and kept in HEPES solution on ice until use. 0.5-cm square sections were cut from the corpus and transferred to the stage of a dissecting microscope where individual gastric glands were hand-dissected using a previously described technique (66). After isolation, the glands were transferred to coverslips precoated with adhesive Cell-Tak (BD Cell-Tak Cell and Tissue Adhesion, BD Biosciences) and mounted in a thermostatically controlled chamber maintained at 37°C on an inverted microscope (Olympus IX50) attached to a digital imaging system (Universal Imaging, Downingtown, PA) for the duration of the experiment. Isolated gastric glands were loaded with 10 M of the pH-sensitive dye 2Ј,7Ј-bis-(2-carboxyethyl)-5-(and 6)-carboxyfluorescein acetomethyl ester (BCECF-AM, Molecular Probes, Eugene, OR) for 10 min in HEPES at 37°C. After the glands were loaded, the chamber was flushed with HEPES-buffered Ringer solution to remove non-de-esterfied dye. BCECF was excited at 490 Ϯ 10 and 440 Ϯ 10 nm, and the emitted fluorescence light was measured at 535 Ϯ 10 nm using an intensified charge-coupled device camera. Data were collected every 12 s, and the ratio at 490/440 nm was initially recorded as arbitrary pH units, which were then converted to absolute intracellular pH (pH i ) using a high K ϩ /nigericin calibration technique (67). Data are expressed as changes in pH i (⌬pH i ) per minute. Acid extrusion was monitored in the absence of HCO 3 Ϫ as intracellular alkalinization after the removal of Na ϩ from the bath and using the NH 4 Cl prepulse technique, which caused reproducible and sustained intracellular acidification. Alkalinization rate (⌬pH i /min) for the calculation of Na ϩ -independent pH i recovery (H,K-ATPase activity) was measured in the pH range of 6.50 -7.00. Na ϩ -free solution eliminated the possibility of contribution to alkalinization rates by the Na/H-exchanger.
Secretagogues-100 M histamine, 100 M carbachol, or 100 M pentagastrin were used to induce acid secretion by wildtype mouse gastric glands. 100 M histamine was used to assess acid secretory capability of isolated gastric glands from ⌬F508defective mice. Pharmacological Inhibitors of CFTR-CFTR inhibition was carried out using 100 M glibenclamide. All drugs were present in combination with BCECF during incubation of the glands for 10 min before the experiment, as well as in all perfusate solutions excluding the K ϩ /nigericin calibration solution. In addition to glibenclamide, the recently developed more CFTR-specific inhibitor CFTR-inh172 was also used at a dose of 10 M in SLO-permeabilized mouse gastric glands as described by Ammar et al. (63). During the 10-min incubation period, 1 g/ml SLO was used to permeabilize the cell in the presence of histamine and CFTR-inh172. CFTR-inh172 and histamine were then included in all of the above-mentioned perfusate solutions. Experiments with similar doses of histamine, SLO, and vehicle (Me 2 SO) in the absence of CFTR-inh172 were also performed as controls. Isolated Perfused Whole Stomach pH Measurements-A total gastrectomy was performed on wild-type and ⌬F508 (Ϫ/Ϫ) mice, following anesthesia with isoflurane and ligation of both the gastroesophageal and gastroduodenal junctions. Once excised, 0.2 cc of non-buffered saline was injected into the isolated whole stomach. The entire stomach was placed in a bath of oxygenated HEPES-buffered Ringer solution at 37°C, pH 7.4, for 60 min in the absence of secretagogue (control) or in the presence of 200 M histamine. Following incubation, the injected non-buffered saline solution was aspirated from the stomach, and its pH was measured.
Immunohistochemistry-Male C57BL/6 mice (20 -25 g) were anesthetized with pentobarbital intraperitoneal and perfused through the left ventricle with PBS buffer, pH 7.4, followed by paraformaldehydelysine periodate fixative (68). Following perfusion for 5 min with fixative, the stomachs were removed, cleaned from food residues, and fixed overnight at 4°C by immersion in paraformaldehydelysine periodate. Stomachs were then washed three times with PBS, and the sections were cut at a thickness of 5 m after cryoprotection with 2.3 M sucrose and 50% polyvinylpyrrolidone in PBS for at least 12 h. Immunostaining was carried out as described previously (53). Briefly, sections were incubated with 1% SDS for 5 min, washed three times with PBS, and incubated with PBS containing 1% bovine serum albumin for 15 min prior to the primary antibody. The primary antibodies mouse monoclonal anti-human ␤ gastric H ϩ / K ϩ -ATPase (Affinity Bioreagents, Golden, CO) and goat anti-SUR-1/2 (Santa Cruz Biotechnology, Santa Cruz, CA), were diluted at 1:1000 and 1:100, respectively, in PBS and applied overnight at 4°C. Sections were then washed twice for 5 min with high NaCl PBS (PBS plus 2.7% NaCl), once with PBS, and incubated with the secondary antibodies (donkey-anti-mouse Alexa 546, donkey-anti-goat Alexa 488, or donkey-anti-mouse Alexa 488, Molecular Probes, Eugene, OR) at a dilution of 1:200 for 1 h at room temperature. Sections were then washed twice with high NaCl PBS and once with PBS before mounting with Vecta-Mount (Vector Laboratories, Burlingame, CA). The specimens were viewed with a Nikon E-800 microscope or a Zeiss LSM-410 confocal microscope.
Statistics-All data are summarized as means Ϯ S.E. Significance was determined using an unpaired Student's t test with p Ͻ 0.05 considered to be statistically significant.

Secretagogue-induced H,K-ATPase Activity in Isolated Gastric Glands from Wild-type Mice in the Presence and Absence of
100 M Glibenclamide-The first series of experiments in this study assessed CFTR involvement in parietal cell secretagogueinduced acid secretion by H,K-ATPase using the CFTR inhibitor glibenclamide at a dose of 100 M. To confirm that stimulation of the Na ϩ -independent pH i recovery represented H,K-ATPase activity, glands were preincubated with 100 M of the specific inhibitor of gastric H,K-ATPase omeprazole for 10 min in the presence of the specific secretagogue used. Omeprazole prevented the stimulatory effect of all secretagogues on the Na ϩ -independent pH i recovery rate.

Secretagogue-induced H,K-ATPase Activity in Isolated SLOpermeabilized Gastric Glands from Wild-type Mice in the Presence and Absence of 10 M CFTR-inh172-
To specifically target the CFTR protein, the inhibitor CFTR-inh172 was used in SLO-permeabilized mouse gastric glands. Histamine-induced, Na ϩ -independent intracellular recovery (0.135 Ϯ 0.015 pH unit/min, n ϭ 11 parietal cells from 2 glands from 2 mice) was significantly reduced in the presence of 10 M CFTR-inh172 (0.038 Ϯ 0.005 pH unit/min, n ϭ 38 parietal cells from 4 glands from 2 mice) (Fig. 3).
Immunohistochemistry-Immunohistochemistry confirmed the presence of the ␤ subunit of H,K-ATPase inside the parietal cells of mouse gastric glands from the corpus (Fig. 4B). However, no signal for SUR 1 and 2 could be detected inside the parietal cells of the same section (see Fig. 4A).
Histamine-induced H,K-ATPase Activity in Isolated Gastric Glands from Gene-targeted Mice Homozygous for the ⌬F508 Mutation-To confirm a role for CFTR in H,K-ATPase-mediated acid secretion by the parietal cell, 100 M histamine was used to stimulate Na ϩ -independent H ϩ extrusion in isolated gastric glands from cystic fibrosis mice homozygous for the ⌬F508 mutation. In comparison to alkalinization rates of histamine-induced isolated gastric glands from wild-type mice (0.088 Ϯ 0.006 pH unit/min, n ϭ 65 parietal cells from 11 glands from 11 mice), glands isolated from transgenic ⌬F508 cystic fibrosis mice (Fig. 5) had a substantially diminished response to    100 M histamine (0.022 Ϯ 0.002 pH unit/min, n ϭ 102 parietal cells from 12 glands from 9 mice) (Fig. 6).
Histamine-induced Acid Secretion in Whole Stomach Preparations from Gene-targeted Mice Homozygous for the ⌬F508 Mutation-Luminal gastric pH measurements from whole stomach preparations were performed to evaluate gastric acid secretion at the whole organ level in ⌬F508 cystic fibrosis and wild-type mice. Control experiments in the absence of bath secretagogue yielded a mean pH of 3.6 Ϯ 0.1 (n ϭ 5) in wildtype mouse whole stomach and 4.5 (n ϭ 3) in ⌬F508 CF mouse stomach. In the presence of 200 M histamine in the bath, mean pH in wild-type stomach was 3.4 Ϯ 0.2 (n ϭ 5). The intragastric pH of histamine-induced luminal acid secretion in ⌬F508 CF stomach was significantly higher (4.3 Ϯ 0.3 (n ϭ 3)) (Fig. 7).

DISCUSSION
Since being cloned in 1989 (7), CFTR has been shown to be present extensively along the gastrointestinal tract. Despite evidence for its presence in the parietal cell in the stomach (69 -71), it has not been assigned a role in parietal cell physiology as yet, nor have there been convincing reports of gastric pathology in patients with cystic fibrosis. In this study, we present direct evidence for a functional role for the CFTR protein in acid secretion by the parietal cell. Furthermore, mice with cystic fibrosis (⌬F508) were found to have significantly blunted acid secretion in response to histamine in both isolated gastric glands as well as whole stomach preparations. Recently a study conducted on CFTR (Ϫ/Ϫ) mouse antral gland base cells using pH-Stat titration and the pylorus ligation technique failed to identify a role for CFTR in gastric acid secretion (44). In contrast in the present study, individual gastric glands were dissected from the corpus of wild-type and ⌬508 CFTR-deficient mice that were loaded with a pH-sensitive dye, and measured with a digital real-time imaging device to allow direct measurements of intracellular pH and acid secretion as calculated from proton efflux from the cell. Mice with cystic fibrosis, homozygous for the most common allele causing the disease in humans, the ⌬F508 mutation, demonstrated defective gastric acid secretion in this study. These results compliment our measurements of impaired proton efflux (acid secretion) following application of the CFTR-specific inhibitor CFTR-inh172 or glibenclamide. One possible explanation for the divergent results may be due to the complete ablation of the CFTR gene potentially causing activation of compensatory mechanisms (i.e. up-regulation of non-CFTR regulated K ϩ channels) thereby preventing reductions in acid secretion and giving an apparently normal profile. We chose to use mice with the ⌬F508 mutation, because it is the most common mutation in humans.
Our data demonstrate a convincing role for CFTR in parietal cell acid secretion, raising the issue as to what may be the cellular and molecular mechanisms behind its involvement. The multifunctional nature of the CFTR protein is well established as are its interactions with several other apical channels and transporters (9). As a chloride conductance channel, however, FIGURE 6. Histamine stimulation of gastric glands isolated from cystic fibrosis mice homozygous for the ⌬F508 mutation demonstrate poor acid secretory response to histamine compared with wild-type gastric glands. A, pH i tracing from a wild-type gastric gland stimulated by 100 M histamine (dashed line represents intracellular alkalinization rate). B, pH i tracing from a gastric gland isolated from a cystic fibrosis mouse (⌬F508 Ϫ/Ϫ ) stimulated by 100 M histamine (dashed line represents intracellular alkalinization rate). C, bar graph depicting histamine-induced acid secretion in wild-type and ⌬F508-defective cystic fibrosis mouse gastric glands. Data are presented as means Ϯ S.E. (wild-type: n ϭ 65 cells from 11 glands from 11 mice; ⌬F508 Ϫ/Ϫ : n ϭ 102 cells from 12 glands from 9 mice). *, p value Ͻ 0.0001. CFTR, having a conductance of 8ps (72), is unlikely to be a candidate for the 28pS (73) chloride conductance reported as that responsible for chloride secretion by the parietal cell into the gastric lumen. This current has characteristics similar to ClC-2, and this has been postulated as the Cl Ϫ current (74,75). Considering its pleiotropic nature, CFTR most likely participates in gastric acid secretion via interaction and modulation of other ion channels or transporters. CFTR may interact with H,K-ATPase in either trafficking from the tubulovesicles to the secretory canaliculi, or as a cAMP provider for protein kinase A-dependent phosphorylation and activation. However, an interaction between the two proteins has not been described despite their common expression in several other organs, including the kidney and colon. An association with parietal cell apical K ATP channels seems to be a likely role for CFTR-modulated H,K-ATPase-mediated secretagogue-induced acid secretion. The presence of K ATP channels, including Kir2.1, Kir4.1, and Kir7.1, has been shown in parietal cells (30 -32), and our laboratory has demonstrated a significant role for K ATP channels in the recycling of K ϩ for acid secretion. 3 Our laboratories have also previously shown by coexpression studies in oocytes that CFTR interacts with and modulates ROMK2 and could likely be the ATP-binding cassette transporter associated with ROMK2 in the kidney, taking into account the absence of a sulfonylurea (SUR) receptor in renal epithelia and the relative abundance of CFTR (18). Similarly, we demonstrate the absence of SUR expression in the parietal cell. Therefore the presence of CFTR as the ATP-binding cassette transporter regulating K ATP activity may characterize its role in acid secretion. In addition, Malinowska et al. (32) observed a discrepancy in gating kinetics between gastric vesicles and Kir2.1-expressed in oocytes. Gribble et al. (77) demonstrated that Kir2.1 is unaffected by SUR1 when coinjected in Xenopus oocytes. Taken together with our results from the present study, it is likely that CFTR may be the ATP-binding cassette transporter associated with Kir2.1 in the parietal cell, thus explaining the discrepancy observed by Malinowska et al.
Despite significant inhibition of acid secretion by glibenclamide and CFTR-inh172 and substantially reduced secretagogue-induced acid secretion in gastric glands isolated from cystic fibrosis mice as compared with wild-type in our experiments, acid secretion is not entirely eliminated. Residual acid secretion may be due to cAMP-independent Ca 2ϩ -dependent intracellular pathways that activate H,K-ATPase in a CFTRindependent fashion. This residual acid secretion would be expected to be sensitive to conventional inhibitors and may explain the data with H2-receptor antagonists and proton pump inhibitors that were shown clinically to decrease acid secretion and raise duodenal pH and improved fat absorption in CF patients (4,35). Ca 2ϩ -dependent compensatory pathways could be up-regulated in these two instances illustrating the diverse complex intracellular mechanisms that converge at H,K-ATPase-mediated acid secretion in the parietal cell. In fact, Roch et al. (78) demonstrated the regulation of K ATP chan-nels by intracellular Ca 2ϩ in addition to cAMP in cystic fibrosis epithelial cells.
Our data support a role for a CFTR-modulated cAMP-dependent pathway for acid secretion. In this model apical CFTR acts by regulating apical K ϩ channels and cAMP-dependent phosphorylation of H,K-ATPase thereby activating H ϩ exchange for K ϩ in addition to cAMP-dependent phosphorylation of ClC-2 and CFTR. The former will mediate Cl Ϫ conductance, whereas the latter, when present, activates K ϩ channels to recycle K ϩ for H,K-ATPase. In contrast to other organs affected in cystic fibrosis, we describe a role for CFTR in acidification rather than alkalinization. In the duodenum (79 -81) and airways (76,82,83), the absence of CFTR-mediated HCO 3 Ϫ secretion in cystic fibrosis accounts for the reduced pH observed in the intestinal lumen and airway surface liquid, respectively. It would seem counterintuitive for gastric CFTR, on the other hand, to mediate HCO 3 Ϫ secretion into the gastric lumen in conjunction with cAMP-activated secretagogue-induced acid secretion. Raising intracellular pH in the face of activated acid secretion is circumvented through anion exchangers present basolaterally in the parietal cell.
In conclusion, we have provided evidence supporting a role for CFTR in gastric acid secretion. The elucidation of a functional role for CFTR in the stomach may have important implications as an additional target for the suppression of acid secretion in patients with problems of non-controlled hyperacidity.