Targeted Deletion of Kcne2 Causes Gastritis Cystica Profunda and Gastric Neoplasia

Gastric cancer is the second leading cause of cancer death worldwide. Predisposing factors include achlorhydria, Helicobacter pylori infection, oxyntic atrophy and TFF2-expressing metaplasia. In parietal cells, apical potassium channels comprising the KCNQ1 α subunit and the KCNE2 β subunit provide a K+ efflux current to facilitate gastric acid secretion by the apical H+K+ATPase. Accordingly, genetic deletion of murine Kcnq1 or Kcne2 impairs gastric acid secretion. Other evidence has suggested a role for KCNE2 in human gastric cancer cell proliferation, independent of its role in gastric acidification. Here, we demonstrate that 1-year-old Kcne2 −/− mice in a pathogen-free environment all exhibit a severe gastric preneoplastic phenotype comprising gastritis cystica profunda, 6-fold increased stomach mass, increased Ki67 and nuclear Cyclin D1 expression, and TFF2- and cytokeratin 7-expressing metaplasia. Some Kcne2 −/−mice also exhibited pyloric polypoid adenomas extending into the duodenum, and neoplastic invasion of thin walled vessels in the sub-mucosa. Finally, analysis of human gastric cancer tissue indicated reduced parietal cell KCNE2 expression. Together with previous findings, the results suggest KCNE2 disruption as a possible risk factor for gastric neoplasia.


Introduction
Gastric cancer remains one of the major causes of mortality, and the second leading cause of cancer death, worldwide. Gastric carcinogenesis is a process that progresses through multiple stages including oxyntic atrophy, ulcer formation, foveolar hyperplasia, hypo-and achlorhydria and mucous cell metaplasia [1]. Despite advances in treatment of gastric cancer, relatively little is known about the molecular mechanisms involved in gastric carcinogenesis, the development of normal gastric mucosa into preneoplastic gastric lesions, or the potential role for ion channels in these processes.
Parietal cells achieve gastric acidification by virtue of an apical H + K + ATPase (HKA) which pumps protons into the stomach lumen in exchange for K + ions. To maintain this activity, K + ions must travel from the parietal cell into the stomach lumen via the apical membrane, to ensure continued substrate for the HKA [2]. This K + ion efflux occurs through one or more types of apical, K +selective channels. Several inward rectifier (Kir) channels are implicated in this process based on parietal cell expression and pharmacological evidence [3,4], but thus far there is genetic evidence for just one channel fulfilling this role: KCNE2-KCNQ1 [5][6][7]. KCNQ1 is a six-transmembrane domain a subunit from the S4 superfamily, which forms functional, voltage-gated, homotetrameric K + -selective channels in heterologous expression studies [8,9]. KCNQ1 can also form heteromeric channel complexes with ancillary subunits from the KCNE gene family, and all five known KCNE gene products have been shown to regulate KCNQ1 function in heterologous expression studies [10]. One, KCNE2, converts KCNQ1 to a voltage-independent, constitutively-active channel whose current is increased by low pH. KCNE2 and KCNQ1 are both expressed at or near to the parietal cell apical membrane, and targeted gene deletion of either subunit results in achlorhydria due to impaired gastric acid secretion [5][6][7]. Thus KCNE2-KCNQ1 channels are considered essential for gastric acid secretion and thought to provide luminal K + ions as a substrate for the gastric HKA.
Kcnq1 2/2 mice and Kcne2 2/2 mice show similar gastric phenotypes, characterized by achlorhydria, hypergastrinemia and gastric glandular hyperplasia [5][6][7]. Parietal cells from either null show ,10-fold reduced capacity to recover from proton loading, suggesting a primary defect in gastric acid secretion. Kcnq1 2/2 mice also develop metaplasia, dysplasia and premalignant adenomatous hyperplasia of the stomach independent of infection [11]. This suggests that KCNQ1 dysfunction could lead to gastric neoplasia independent of H. pylori due to severe achlorhydria, or represent an additional risk factor which in conjunction with H. pylori could predispose to gastric cancer.
The achlorhydria and gastric hyperplasia we previously observed in Kcne2 2/2 mice was striking given that the poreforming subunit of the complex, KCNQ1, was still present, and in fact it was strongly expressed in double the number of cells per gastric gland in Kcne2 2/2 mice compared to Kcne2 +/+ mice [7]. In a recent study, KCNE2 expression was found to be expressed at relatively low levels in human gastric tumors and in gastric cancer cell lines; furthermore, forced over-expression of KCNE2 suppressed the growth of human gastric cancer cells in tissue culture, and in nude mice [12]. These findings raised the intriguing possibility that KCNE2 might be involved in gastric carcinogenesis, and potentially independently of either H. pylori infection or achlorhydria. Here, to further define the potential role of KCNE2 in regulation of normal gastric cell growth, we scrutinized the gastric pathology of Kcne2 2/2 mice up to 15 months of age.

Kcne2 2/2 mice exhibit progressive gastric hyperplasia
We previously demonstrated that Kcne2 is required for normal regulation of gastric mucosal cell growth: 3-month-old Kcne2 2/2 mice have gastric hyperplasia, achlorhydria and abnormal parietal cell morphology [7]. Here we compared the stomach mass of 3week, 3-month-and 12-15-month-old Kcne2 2/2 mice. The mass of stomachs at 3 weeks of age were similar for Kcne2 +/+ and Kcne2 2/2 mice, whereas by 3 months there was significant gastric hyperplasia in Kcne2 2/2 mice. This difference had increased by 12-15 months to the extent that Kcne2 2/2 mice had stomachs 6-fold larger than those of Kcne2 +/+ mice of the same age (Figure 1 A).
Kcne2 2/2 mice exhibit gastritis cystic profunda Even at 3 weeks of age, despite similar stomach mass to that of Kcne2 +/+ mice, gastric mucosa of Kcne2 2/2 mice already showed vacuolations close to the basolateral side (Figure 1 B; magnified in D). To investigate the progression of gastric hyperplasia and its consequences, a total of eleven 12-15 month-old Kcne2 2/2 mice and five Kcne2 +/+ mice were histologically examined. The gastric mucosa in all Kcne2 2/2 mice showed glandular hypertrophy and diffuse hyperplasia with increased numbers of mucous cells and parietal cells that were occasionally vacuolated (Figure 1 C). Strikingly, gastritis cystica profunda (GCP) was observed in 11/11 Kcne2 2/2 mice, but 0/5 of the Kcne2 +/+ mice (x 2 p = 0.002) (magnification of cyst in Figure 1 E). Dilated glands contained cell debris and, in some but not all Kcne2 2/2 mice, neutrophils. There was also gastric epithelial hyalinosis with occasional intraluminal crystals, and lymphoplasmacytic aggregates in the mucosa, submucosa and muscularis. In some 12-15-month-old Kcne2 2/2 mouse stomachs, prominent lamina proprial fibrosis was noted, especially in the superficial layers. GCP, also previously referred to as benign gastric pseudotumor, generally presents as a spectrum of hyperplastic and metaplastic changes following damage to the gastric mucosa. While GCP demonstrates histological features associated with malignancy, it is in itself generally considered functionally benign -although it has been suggested as a possible risk factor for gastric cancer [13,14].
Increased proliferative markers in gastric mucosa of Kcne2 2/2 mice Importantly, expression of established preneoplastic markers was increased in Kcne2 2/2 mice compared to Kcne2 +/+ mice. Ki67 antigen is a cell cycle related nuclear protein commonly used as a proliferation marker in proliferating and neoplastic tissues, including the stomach [15]. Ki67 staining in sections from Kcne2 +/+ mouse stomachs (3 months) indicated a small band of Ki67 positive cells within the isthmic regions of gastric glands, whereas this proliferative compartment was significantly expanded in age-matched Kcne2 2/2 mice (Figure 2 A-C). Epithelium lining cystic areas of GCP in 12-month-old Kcne2 2/2 mice also demonstrated prominent staining of nuclei with Ki67, indicative of a high proliferation rate (Figure 2 B).
Cytokeratin (CK)-7 is a 54 kDa polypeptide expressed in a wide variety of epithelial tissues including lung, breast, and fetal human stomach; however, it is not expressed in normal adult gastrointestinal epithelia [16]. Upregulation of CK-7, indicative of dedifferentiation, was evident in Kcne2 2/2 gastric mucosa at 12 months; in age-matched Kcne2 +/+ mice, gastric mucosal positive epithelial cells were much rarer ( TFF2-expressing metaplasia in gastric mucosa of Kcne2 2/2 mice Trefoil-factor family (TFF)2-expressing metaplasia is associated with progression to gastric cancer in humans, and in Mongolian gerbils infected with H. pylori [17]. Gastric mucosa of Kcne2 2/2 mice exhibited areas of metaplasia with prominent TFF2 expression, including in epithelial cells lining cysts ( Figure 3 A-C). TFF2-expressing metaplasia is typically associated with oxyntic atrophy, characterized as a loss of parietal cells [1]. Here, we analyzed gastric mucosa from 12-month-old mice with TFF2expressing metaplasia for expression of the HKA b subunit (HKA b), a parietal cell marker. Strikingly, there was no evidence of oxyntic atrophy in terms of number of parietal cells (Figure 4  Increased gastric mucosal Cyclin D1 expression and gastric neoplasia in Kcne2 2/2 mice Reduced KCNE2 expression was previously suggested to enhance proliferation in the gastric cancer cell line SGC7901 via increased expression of Cyclin D1 [12]. Here, western blots suggested that Kcne2 gene deletion increased total Cyclin D1 expression in the gastric mucosa of 12-month-old mice (Figure 5 A). Cyclin D1 showed weak and predominantly cytoplasmic expression at the base of Kcne2 +/+ oxyntic glands (Figure 5 B,C, left panels), as previously described for normal, proliferating cells in mouse stomach [18]. In contrast, in Kcne2 2/2 mice, Cyclin D1 showed more widespread and nuclear staining in the neck isthmus region of gastric glands, and glandular pit (Figure 5 B,C, right panels). In two of eleven 1-year-old Kcne2 2/2 mice analyzed, pyloric polypoid adenomas were observed, extending into the duodenum (Figure 5 D). The same two Kcne2 2/2 mice also exhibited neoplastic growth in the form of thrombi composed of fibrin and glandular epithelium, in thin-walled vessels (identified using endothelial marker CD34) within the gastric submucosa ( Figure 5 E-H).

Reduced parietal cell KCNE2 expression in human gastric cancer tissue
Given that in a previous study reduced KCNE2 expression was found to enhance gastric cancer cell line proliferation, and KCNE2 expression was found to be reduced in human gastric cancer [12], we examined KCNE2 expression in normal human gastric mucosa, and gastric cancer tissue. Immunofluorescence studies revealed striking differences with respect to KCNE2 and KCNQ1 co-localization, between normal and malignant human gastric mucosae. As expected, KCNQ1 and KCNE2 co-localized strongly in parietal cells in normal human gastric mucosa (

Potassium channels in proliferative disorders
Potassium channels have emerged as a potential target for anti-cancer therapies [19], including KCNQ1, hERG and Kv2.1, all of which are a subunit partners of KCNE2. Disrupted imprinting caused by mutations in the KCNQ1 gene cause Beckwith-Wiedemann syndrome (BWS), which predisposes to cancer [20][21][22], and KCNQ1 knockout predisposes to gastric metaplasia [11]. The hERG potassium channel a subunit, which forms complexes with the KCNE2 ancillary subunit in human heart [23], is over-expressed in some tumors, and has been identified as a tumor survival factor [24][25][26][27] as has Kv2.1 [28].
KCNE2 was previously found to be expressed 2-fold lower in human gastric cancer tissue than in neighboring normal gastric cells, and forced upregulation of KCNE2 had anti-proliferative effects on gastric cancer cells in vitro and injected into nude mice [12]. The postulated mechanism for this anti-proliferative effect is down-regulation of Cyclin D1, delaying progression through the cell cycle [12] similar to what was observed with the hERG blocker cisapride [29]. Elevated expression of Cyclin D1 in human gastric tumors correlates with a particularly poor prognosis [30], therefore understanding factors that increase Cyclin D1 expression may lead to therapeutic avenues for this and other forms of cancer. Overexpression of Cyclin D1 has been suggested to contribute to oncogenesis by disturbing the cell cycle, and has been reported to be an important oncogenic factor in esophageal carcinoma [31], and associated with nuclear accumulation of b-catenin in ovarian endometrioid adenocarcinomas [32]; nuclear Cyclin D1 overexpression in gallbladder carcinomas is a critical event [33].
Here, we describe extensive metaplastic changes in the gastric muscosa due to genetic disruption of Kcne2. This is associated with increased nuclear Cyclin D1 expression in the gastric mucosa, reminiscent of results from previous in vitro studies of KCNE2 [12]. Neither the current report nor the previous study, however, delineate the mechanism for Cyclin D1 upregulation in Kcne2 2/2 mucosa -is it a consequence of metaplastic changes secondary to achlorhydria in these mice, or is there a more direct association between Kcne2 and Cyclin D1? The study by Yanglin and colleagues argues for the latter at least in part, because in their study the Cyclin D1 changes occurred in isolated cells without the influence of achlorhydria [12]. We consider that changes secondary to achlorhydria are the most likely dominant factor, but a direct link cannot be excluded. Interestingly, hERG, another partner of KCNE2, is expressed in gastric cancer cells but not normal gastric epithelia, and the hERG channel blocker cisapride was previously found to suppress gastric cancer cell growth by inhibiting entry into S phase from G(1) phase in the cell cycle, increasing apoptosis [29]. Because KCNE2 partially suppresses hERG currents by reducing unitary conductance and speeding deactivation [23], it is an intriguing possibility that the observed anti-proliferative effects of KCNE2 overexpression in vitro [12] are due to hERG current suppression promoting apoptosis, and conversely that KCNE2 down-regulation or genetic disruption may favor gastric cellular proliferation by increasing hERG current density.   GCP and TFF2-expressing metaplasia (SPEM) in Kcne2 2/2 mice GCP can present in human beings with intermittent epigastric pain, bloating, gastric obstruction or upper gastrointestinal bleeding, and is most commonly seen in patients with a history of either gastrectomy or gastrostomy [34]. However, a few cases have been reported with no association to prior gastric surgery [35,36]. The pathogenesis of human GCP is believed to arise from an injury of the muscularis mucosae, which may lead to ectopic entrapping of gastric glands in the submucosa, muscularis mucosae or serosa, or from chronic inflammation. GCP is commonly considered a benign clinical entity, although association with gastric cancer has been reported [14,37,38].
In laboratory animals, GCP secondary to Helicobacter sp. infection has been described as a predecessor of intramucosal dysplasia and neoplasia [39], and associated with development of gastric carcinoma. GCP was frequently found in H. pylori-infected Mongolian gerbils that also developed gastric adenocarcinoma and lymphoid hyperplasia [40]. In a previous murine genetic model of GCP, TGF beta 1 +/2 mice developed glandular hyperplastic lesions that shared morphologic features with human GCP, with mixed inflammatory infiltration of the surrounding mucosa and chronic vasculitis in the tissues adjacent to these lesions [41]. Here, in Kcne2 2/2 mice, GCP occurs in the absence of a primary inflammatory lesion, and independent of vascular obliteration. Instead, we propose that the primary defect is loss of the KCNE2 b subunit from channel complexes with KCNQ1, thereby impairing an important apical K + recycling pathway in parietal cells and preventing gastric acidification by the parietal cell HKA. While this represents a 'functional atrophy' in the parietal cell population, in previous studies of intestinal metaplasia an actual loss of parietal cells, termed 'oxyntic atrophy', was considered an important process in the etiology of intestinal metaplasia, associated with acid reflux disease or H. pylori infection, and is the most common metaplastic process observed in the upper gastrointestinal tract [42].
Oxyntic atrophy in the absence of significant inflammation can be induced by DMP-777, a neutrophil elastase inhibitor which also targets parietal cells due to its action as a protonophore [43]. Oxyntic atrophy in gastric-deficient mice, following treatment with DMP-777, leads rapidly to TFF2-expressing metaplasia, also referred to as spasmolytic peptide expressing metaplasia (SPEM) [44]. SPEM is notable because of its strong association with human gastric adenocarcinoma, and it is observed in the majority of fundic gland biopsies from human H. pylori-infected gastritis patients [45]. The appearance of SPEM in Kcne2 2/2 mice in the absence of either H. pylori or classic oxyntic atrophy suggests that the 'functional atrophy' caused by Kcne2 deletion may be enough to trigger SPEM. One major difference between Kcne2 2/2 mice and those in the Nozaki study is that Kcne2 2/2 mice are hypergastrinemic, not gastrin-deficient. Another difference is that DMP-777 disrupts a gastric tubulovesicle proton gradient without impairing H + K + -ATPase function, whereas targeted Kcne2 disruption prevents H + /K + -ATPase function by removing its luminal substrate, K + ions [7,44]. Perhaps this latter mechanism may in the future reveal clues regarding the critical events in oxyntic atrophy that result in SPEM: it is currently suggested that SPEM induction results from a deficiency in parietal cell-secreted regulators of normal gastric mucosal renewal, which include sonic hedgehog and TGF-a [46,47] due to loss of parietal cells. Further studies of Kcne2 2/2 parietal cells -which are still present but malfunctioning -may reveal regulators which they can still secrete, and those which they cannot.
The gastric mucosa of Kcne2 2/2 mice also show increased expression of the established proliferation marker Ki67, manifesting as an increase in the width of the mucosal proliferative band, and as Ki67-positive cells lining the cysts in regions of GCP. Similarly, the de-differentiation marker CK7, not expressed in 1-year-old Kcne2 +/+ gastric mucosa, was widely expressed in that of Kcne2 2/2 mice. Taking these and the markers described above together, the Kcne2 2/2 gastric mucosa exhibits metaplasia with multiple features of preneoplasia, and in some cases neoplasia, notably in a specific pathogen-free environment with no evidence of gastric Helicobacter infection or oxyntic atrophy, and in the absence of chemical inhibitors of gastric acidification. Together with the finding here that parietal cell KCNE2 expression appears to be reduced in human gastric cancer tissue (Figure 4) and the previous report that KCNE2 inhibits gastric cancer cell proliferation [12], the data suggest KCNE2 disruption is associated with gastric cancer progression. Future studies will involve examining whether Kcne2 disruption increases predisposition to gastric cancer within pathogen and carcinogen-based protocols, the mechanisms behind these possible differences, and the potential mechanistic links between Kcne2, Cyclin D1 and cell cycle perturbation outside the realm of achlorhydria-associated disease etiology. Furthermore, as KCNQ1 and KCNE2 are also co-expressed in thyroid epithelial cells, where they are important for thyroid hormone biosynthesis [48], it will be of interest to examine a potential role for KCNE2 in abnormal thyrocyte proliferation.

Materials and Methods
Generation and use of gene-targeted mice All mice described in this study were housed, utilized and euthanized according to NIH and Cornell University Institutional Animal Care and Use Committee guidelines. All mice described in this study were housed, utilized and euthanized according to NIH and Weill Medical College Institutional Animal Care and Use Committee guidelines. Ethical approval to breed and harvest tissue from wild-type and Kcne2 2/2 mice for biomedical research was approved by Weill Medical College Institutional Animal Care and Use Committee (protocol 0704-610A). Kcne2 2/2 mice were generated as previously described from C57BL/6 Kcne2 +/2 x Kcne2 +/2 crosses [7]. Numerical data were analyzed with EXCEL software (Microsoft) using one-way analysis of variance (ANOVA) with statistical significance set at P,0.05.
For Cyclin D1 detection, stomach slides were heated at 58-60uC for 30 min and then deparaffinized. Endogenous peroxidase activity was blocked by incubating sections in 1% hydrogen peroxide in PBS for 15 min (8.3 ml 30% H 2 O 2 / 241.7 ml PBS) followed by unmasking of the antigenic epitope by microwaving in 10 mM Citrate Buffer at high power for 15 min. Slides were cooled for 20 min, washed in distilled water for 5 min, transferred to PBS and incubated in 10% normal horse serum (in 2% BSA-PBS diluent) for 30 min in a humid chamber. Mouse monoclonal anti-cyclin D1-antibody (Cell Signaling) was added at 1:1000 with 2% BSA in PBS and incubated overnight at 4uC in a humid chamber. Slides were rinsed in PBS and the secondary biotinylated horse anti-mouse IgG antibody (Vector Labs) was applied at 1:500 in PBS for 30 min at RT. Slides were washed and an Avidin-Biotin Complex (Vectastain ABC Elite Kit, Vector Labs) diluted at 1:25 in PBS was applied for 30 min. The signal was detected by incubation in 3,3-Diaminobenzidine (Sigma, Batch No. 026K3767) until the desired color intensity was reached. After final washes in water, slides were lightly counterstained in hematoxylin. KCNQ1 and HKA b single labeling and detection were performed as previously described [7]. CD34 detection was with 1:50 anti-CD34 antibody (AbCam). Slides were viewed with a Nikon Eclipse E600 microscope and photographed using a RT Color Camera and SPOT software (Diagnostic Instruments, Inc.).

Immunofluorescence
All human tissue was obtained from ProSci, Poway, CA, and was certified as having been obtained from qualified and licensed pathology services, using appropriate conformed consent for use in biomedical research and with donor anonymity protection; therefore, additional Weill Medical College Institutional Review Board ethics approval was not required. Immunofluorescence detection of HKA b, KCNQ1 and KCNE2 in human gastric tissue (ProSci) was performed using a Discovery XT processor (Ventana Medical Systems). The following human tissues were used: normal stomach tissue from a 50-year-old woman (PSC-10-809-XB1); gastric carcinoma tissue, patient information unavailable (PSC-10-814-CA1); gastric adenocarcinoma tissue from a 51-year-old woman (PSC-10-809-XA1). The primary antibody concentrations used were: 0.5 mg/ml anti-HKA b (mouse monoclonal, Affinity Bioreagents), 1 mg/ml anti-KCNQ1 (rabbit or goat polyclonal, Chemicon), and in-house anti-KCNE2 serum was used at a 1:500 dilution after column-enriching IgG. Preceding the primary antibody incubation, the tissue sections were blocked for 30 min in 10% normal goat serum, 2% BSA in PBS, followed by 8 min Avidin/Biotin block. The primary antibody incubation (3 hr) was followed by 32 min incubation with biotinylated anti-mouse IgG (ABC kit from Vector labs), 60 min incubation with biotinylated anti-goat IgG (ABC kit from Vector labs), or biotinylated anti-rabbit antibody at 1:200 dilution (Vectastain ABC kit). The secondary detection was performed with Streptavidin-HRP D (Ventana Medical Systems), followed by incubation with Tyramide-Alexa Fluor 488 (Invitrogen) or Tyramide Alexa Fluor 568 (Invitrogen). Stained slides were viewed with a Zeiss Axiovert 200 widefield microscope and pictures were acquired using MetaMorph software 7.1 (Molecular Devices).

Western blotting
For western blotting, gastric membrane fractions were prepared as previously described [7]. Three stomachs from 12-month-old Kcne2 +/+ and Kcne2 2/2 mice were removed postmortem, cut open along the curvatura ventriculi major, cleared of ingesta and immediately snap-frozen in liquid nitrogen. Stomachs were then homogenized in a buffer containing 50 mM Tris-HCl, 150 mM NaCl, 100 ug/ml PMSF, 1 ug/ml Nonidet P40, 0.5% SDS and 0.5% sodium orthovanadate, then incubated on ice for 30 min and centrifuged at 120006g for 20 min at 4uC. Protein concentration of the supernatant was measured according to the Bradford method. Total protein (40 mg/lane) was loaded into a pre-cast tris-glycine 4-20% gel (Jule Inc, Milford, CT) and separated by electrophoresis. Proteins were then transferred onto a PVDF membrane (Bio-Rad, Hercules, CA), and blocked with 5% milk, 0.05% Tween-20 in PBS for 12 hours at 4uC on a rocker. Primary antibody incubations (4 hr, RT in 1% milk, 0.05% Tween-20, PBS, 'Buffer A') were: 1:2000 anti-Cyclin-D1 (Abcam); 1:5000 anti-GAPDH (Abcam). Membranes were washed 4 times, 20 min each with antibody incubation buffer then incubated with the appropriate secondary antibodies (BioRad) diluted 1:10000 in Buffer A for 2 hr at RT then washed 4620 min each with Buffer A and once for 5 min with PBS. Membranes were incubated for 1 min with the SuperSignal ECL reagent (Pierce) then exposed on BioMax Light Film (Kodak) and developed using an RP X-OMAT Processor (Kodak).