Sodium Channel Subunit SCNN1B Suppresses Gastric Cancer Growth and Metastasis via GRP78 Degradation

Gastric cancer is one of the most common human cancers. Despite improvements in the surveillance and treatment of gastric cancer, it remains a devastating disease with poor prognosis (1). Epigenetic dysregulation plays an important role in gastric carcinogenesis. Previous studies have shown that the inactivation of tumor suppressor genes by promoter DNA methylation contributes to the pathogenesis of human gastric cancer (2–8). To unveil novel tumor suppressor genes that are silenced by epigenetic mechanisms in gastric cancer, we used genome-wide methylation array (Infinium Human Methylation 450 K) to comprehensively profile CpG site methylation in five gastric cancer cell lines (AGS, HGC27, MGC803, MKN1, and MKN45), an immortalized human gastric epithelial cell GES1 and two normal gastric tissue samples. Using this approach, we identified SCNN1B as a novel gene that is highly methylated in human gastric cancer, whose potential role in gastric cancer development is largely unknown.

SCNN1B is located on chromosome 16p12.2 and it encodes β-subunit of the epithelial sodium channel (ENaC). SCNN1B is a part of a multiprotein complex consisting of three subunits (α, β, and γ) that controls fluid and electrolyte transport across epithelia in diverse organs. SCNN1B is classified as a membrane channel, but accumulating evidence also indicates that ENaC subunits, including SCNN1B, participate in cellular differentiation (9–11). SCNN1A, which encodes the α-subunit of ENaC, has been shown to be silenced by promoter methylation in neuroblastoma and breast cancer (12, 13). However, the functional importance of SCNN1B in human cancer remains unexplored. In this study, we identified frequent silencing of SCNN1B in human gastric cancer, which was associated with promoter methylation. We demonstrated a significant correlation between the silence of SCNN1B protein expression and poor disease-specific survival of gastric cancer patients. We revealed that SCNN1B suppresses gastric cancer growth by inducing apoptosis and cell-cycle arrest and inhibiting metastasis abilities. The tumor-suppressive effect of SCNN1B was found to be mediated via (i) the direct interaction with GRP78, a chaperone with oncogenic properties; (ii) the reduction of GRP78 protein by inducing its polyubiquitination and proteasome-mediated degradation; and (iii) the induction of the unfolded protein response (UPR) response, which activates PERK, ATF4, XBP1s, and C/EBP homologous protein (CHOP), leading to caspase-dependent apoptosis and cell-cycle arrest. Moreover, tissue microarray (TMA) analysis of 245 gastric cancer patients revealed that high SCNN1B expression is an independent prognostic factor that predicts better survival of gastric cancer patients.

Sixteen gastric cancer cell lines and a normal gastric epithelial cell line were used in this study: AGS, KATOIII, MKN45, and NCI-N87 cells were obtained from the ATCC; MKN1, MKN74, SNU1, SNU638, SNU719, and YCC10 cells were obtained from the Korean Cell Line Bank; MKN7 cells were obtained from RIKEN Cell Bank; BGC823, HGC27, MGC803, SGC7901, and normal gastric epithelial cell line GES1 were obtained from the Cell Bank of Chinese Academy of Sciences (Shanghai, China). All cells were purchased between 2014 and 2015, and routinely cultured in DMEM containing 10% FBS and penicillin–streptomycin.

The full-length ORF of SCNN1B was cloned into pcDNA3.1, pCMV4-FLAG, and pEGFP-N1 vectors. Transfection was performed with Lipofectamine 2000 (Life Technologies). Cell lines stably expressing SCNN1B were obtained after selection with neomycin (G418, Life Technologies) for at least 2 weeks. SCNN1B siRNA were purchased from RiboBio Co. Ltd and transfected into MKN1 and NCI-N87 cells using Lipofectamine 2000.

Cells were plated in 6-well plates at 1,000 cells per well in complete DMEM. Medium was changed every 3 to 4 days. At the endpoint, cells were stained with 0.1% Crystal violet and the number of colonies consisting of >50 cells were counted. Cell growth curve was performed using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Sigma-Aldrich).

Cells were plated in 12-well plates and serum-starved overnight. Annexin V-PE/7-aminoactinomycin D (7-AAD) staining kit (BD Biosciences) was used to determine cell apoptosis. For cell cycle, cells were serum-starved for 24 hours and then stimulated with complete medium for 4 to 12 hours. Cell-cycle distribution was then assessed by flow cytometry after staining with propidium iodide (Life Technologies).

Confluent cultures in 6-well plates were scratched with sterile P-200 pipette tips, washed, and cultured in DMEM containing 2% FBS. Cells were photographed after 0, 12, 24, and 48 hours, respectively. Wound closure (%) was evaluated by the TScratch software.

Cell invasion was determined using BD BioCoat Matrigel Invasion Chamber (BD Biosciences). Cells (5 × 10⁴/well) were seeded onto the top chamber in serum-free DMEM. Complete DMEM (supplemented with 10% FBS) was added to the bottom chamber as a chemoattractant. After 48 hours, cells that had invaded through the membrane were stained with 0.1% Crystal violet and counted.

Cells (0.5–1 × 10⁵/well) were seeded onto 96-well plates. After 30 and 60 minutes, the medium was aspirated, then the cells were washed with PBS and stained with 0.1% Crystal violet. The Crystal violet was dissolved in 10% acetic acid overnight and absorbance was measured at 540 nm.

Cells were seeded onto coverslips in a 6-well plate and transfected with GFP-tagged SCNN1B and Myc-tagged GRP78. Twenty-four hours after transfection, cells were fixed with 4% paraformaldehyde and permeabilized with 0.2% Triton X-100, blocked in 5% BSA in PBS, and incubated with anti-Myc (1:4,000 dilution) overnight at 4°C, followed by anti-mouse IgG secondary antibody conjugated with Alexa Fluor 594 (1:400 dilution, Yeasen) in the dark for 1 hour. Cells were then mounted with ProLong Gold Antifade Mountant with DAPI (Life Technologies). Images were captured using a Carl Zeiss LSM 780 confocal laser scanning microscope (Carl Zeiss AG).

Paired primary gastric tumors and adjacent normal gastric tissues were collected immediately after surgical resection at the Prince of Wales Hospital (Hong Kong, China). The specimens were snap-frozen in liquid nitrogen and stored at −80°C and were also fixed in 10% formalin and embedded in paraffin for routine histologic examination. Biopsies from 3 cases of normal mucosa obtained during gastroscopy were recruited as healthy controls, which were confirmed by an experienced pathologist at the Prince of Wales Hospital. All patients gave informed consent, and the study protocol was approved by the Clinical Research Ethics Committee of the Chinese University of Hong Kong (Hong Kong, China).

TMA was generated from formalin-fixed, paraffin-embedded archived tissue samples of 245 patients with gastric cancer prior to radiotherapy/chemotherapy, which were collected at the Prince of Wales Hospital (Hong Kong, China; ref. 14), with a median follow-up time of 40.8 months. All subjects provided informed consent for obtaining the specimens. TMA was stained with a commercially available anti-SCNN1B antibody (HPA015612, Sigma-Aldrich). Anti-SCNN1B antibody (HPA015612) was confirmed by antibody specificity analysis with protein arrays, with single peak corresponding to interaction only with its own antigen. Cytoplasmic expression of SCNN1B was assessed by H-score. The proportion score was in the light of proportion of cancer cells with positive cytoplasmic staining (0, no positive staining; 1, in 10% or fewer cells; 2, in between 10% and 25% cells; 3, in between 25% and 50% cells; 4, in more than 50% cells). The intensity score was assigned for the average intensity of cancer cells with positive staining (0, none; 1, weak; 2, intermediate; 3, strong). The IHC score of SCNN1B was calculated by the following formula: IHC score = proportional score (0–4) × intensity score (0–3), ranging from 0 to 12. Finally, the cytoplasmic expression of SCNN1B in gastric cancer tissue was divided into 3 groups according to IHC score (low, ≤3; intermediate, 4–6; high, 7–12). The results were scored independently by two pathologists and the average of the two values was taken.

Cells transiently transfected with SCNN1B-FLAG or empty vector were lysed with ice-cold RIPA lysis buffer (50 mmol/L Tris-Cl, 150 mmol/L NaCl, 1% NP-40, 0.5% sodium deoxycholate, and 1% SDS, pH 8.0) supplemented with protease inhibitors (Roche). Total proteins were immunoprecipitated using 2 μg of anti-Flag (F1804, Sigma-Aldrich) and bound to 40 μL of Protein G-Agarose (Santa Cruz Biotechnology). After washing 5 times with RIPA buffer, bound proteins were eluted with loading buffer, separated by SDS-PAGE, and visualized by silver staining. Protein bands of interest in-gel were digested, and subjected to LC/MS-MS (ABI4800 MALDI TOF/TOF, Applied Biosystems). The MS fragment spectra were analyzed using Mascot software (Matrix Science). To confirm the interaction of SCNN1B with GRP78, immune complexes were precipitated by anti-Flag and analyzed by Western blot analysis using anti-Flag and anti-GRP78 (Sc-13968, Santa Cruz Biotechnology).

AGS cells stably transfected with SCNN1B expression vector or empty vector were lysed with RIPA buffer supplemented with protease inhibitors. Immunoprecipitation was performed using anti-GRP78 or control IgG, respectively. Immunoprecipitated proteins were analyzed by Western blot analysis using anti-ubiquitin (3936, Cell Signaling Technology).

BGC823 (1 × 10⁷ cells/0.1 mL PBS) and MKN45 (1 × 10⁶ cells/0.1 mL PBS) cells stably expressing the control vector or SCNN1B were injected subcutaneously into the left and right dorsal flank of 4- to 6-week-old female Balb/c nude mice (n = 6/group), respectively. Tumor size was measured every 2 days for 2–3 weeks using a digital caliper. Tumor volume (V) was estimated by measuring the longest diameter (L) and shortest diameter (W) of the tumor and calculated by formula V = 0.5 × L × W². At the endpoint, tumors were harvested and weighted. All experimental procedures were approved by the Animal Ethics Committee of the Chinese University of Hong Kong (Hong Kong, China).

All the results were expressed as mean ± SEM (continuous variables) or described as frequency and percentage (categorical data). To compare the difference between two groups, independent sample t test or Mann–Whitney U test was used. The difference between growth rates was determined by ANOVA with repeated-measures ANOVA. The Pearson χ²test or Fisher exact test was used for analysis of the associations between patient clinicopathologic characteristics and SCNN1B expression. Kaplan–Meier analysis and log-rank test were performed to evaluate the association between SCNN1B expression and disease-specific survival. Cox proportional hazards regression model was performed to assess the prognostic value of SCNN1B expression. All the statistical analyses were performed using GraphPad Prism, version 6.0 (GraphPad Software) or SPSS, version 20.0 (SPSS Inc.). P < 0.05 was considered statistically significant.

Using the Infinium Human Methylation 450 K array, we interrogated genome-wide CpG methylation in five human gastric cancer cell lines (AGS, HGC27, MGC803, MKN1, and MKN45) as compared with normal gastric epithelial cell line GES1 and normal gastric tissues (Fig. 1A). Using stringent criteria, we identified SCNN1B to be preferentially methylated at its promoter in gastric cancer (Fig. 1B).

We initially examined SCNN1B mRNA expression in human normal tissues, and found that SCNN1B was widely expressed in most human normal tissues with strong expression in the stomach (Supplementary Fig. S1). On the other hand, SCNN1B mRNA expression was silenced in 13 of 16 (81.3%) gastric cancer cell lines (Fig. 1C); only MKN1, MKN7, and NCI-N87 cells lines expressed significant levels of SCNN1B mRNA (Fig. 1C; Supplementary Fig. S2). To determine the role of promoter methylation in silencing of SCNN1B, we evaluated its promoter methylation by methylation-specific PCR (MSP) and bisulfite genomic sequencing (BGS). MSP analysis revealed dense SCNN1B promoter methylation in all gastric cancer cell lines with silenced SCNN1B expression (Fig. 1C). BGS analysis of 38 CpG sites in SCNN1B promoter and the first exon showed dense methylation (average methylation > 50%) in the SCNN1B-silenced gastric cancer cell lines examined, but not in SCNN1B-expressing MKN1 cells and normal gastric tissues (Fig. 1D; Supplementary Fig. S3). To test whether promoter methylation directly mediates the silencing of SCNN1B, six gastric cancer cell lines with silenced SCNN1B expression were treated with DNA methyltransferase inhibitor, 5-Aza-2′-deoxycytidine (5-Aza). 5-Aza restored SCNN1B expression in all six cell lines, indicating that promoter methylation contributes to the transcriptional silencing of SCNN1B (Fig. 1E). In addition, treatment with 5-Aza plus histone deacetylase inhibitor trichostatin A (TSA) could fully restore SCNN1B expression in MKN45 cells with moderate promoter methylation (Fig. 1E).

We evaluated mRNA and protein expression of SCNN1B in gastric tissues from 74 primary gastric cancer patients. SCNN1B mRNA was significantly downregulated in gastric cancer as compared with paired adjacent normal gastric tissues (P < 0.0001; Fig. 2A and B). SCNN1B mRNA expression was also downregulated in gastric cancer in the The Cancer Genome Atlas (TCGA) cohort (n = 34; P < 0.0001; Fig. 2B). Western blot analysis and IHC confirmed the reduced expression of SCNN1B in gastric cancer as compared with adjacent normal tissues (n = 10; P < 0.001; Fig. 2A and B). We next examined the methylation status of SCNN1B in primary gastric cancer. MSP and BGS analysis demonstrated that SCNN1B promoter methylation was significantly higher in gastric cancer as compared with adjacent normal tissues (Fig. 2A and C). None of the normal gastric biopsies showed SCNN1B promoter methylation. These data implied that SCNN1B is silenced by promoter methylation in gastric cancer. Consistent with our data, analysis of the TCGA dataset revealed an inverse correlation between SCNN1B mRNA expression and promoter methylation in gastric cancer (P < 0.001; Fig. 2D).

To evaluate the association of SCNN1B expression with clinicopathologic features and clinical outcomes, we assessed the SCNN1B protein expression in gastric cancer utilizing a gastric cancer TMA (n = 245). SCNN1B cytoplasmic expression showed a significant correlation with TNM stage (P < 0.001) and lymphatic metastasis (P = 0.036), but had no correlation with age, gender, H. pylori infection, histologic Lauren classification, or tumor grade (Supplementary Table S1). In univariate Cox regression analysis, an intermediate or high cytoplasmic SCNN1B score was associated with better disease-specific survival [intermediate: HR, 0.482; 95% confidence interval (CI), 0.320–0.726, P < 0.001; high: HR, 0.247, 95% CI, 0.091–0.674, P = 0.006]. Apart from SCNN1B expression, age (P = 0.048), histologic Lauren classification (P < 0.001), tumor grade (P = 0.024), and TNM stage (P < 0.001) were also correlated with survival by univariate analysis. After adjustment for potential confounding factors such as age, gender, histologic Lauren classification, tumor grade, and TNM stage, SCNN1B expression was found to be an independent prognostic factor for disease-specific survival [intermediate: HR, 0.547; 95% confidence interval (CI), 0.360–0.829; P = 0.005; and high: HR, 0.353; 95% CI, 0.128–0.971; P = 0.044] by multivariate Cox proportional hazards regression analysis (Supplementary Table S2). As shown by Kaplan–Meier curves, gastric cancer patients with intermediate or high SCNN1B protein expression had significantly longer survival (P < 0.001; Fig. 2E). Further stratification of the TMA cohort into early stage (TNM stage I/II) and late stage (TNM stage III/IV) revealed that intermediate or high protein expression of SCNN1B was associated with better survival in late-stage gastric cancer (P = 0.011; Fig. 2F). Analysis of another two independent gastric cancer cohorts (GSE62254 and GSE14210) also showed that high SCNN1B mRNA expression was associated with better survival in late-stage gastric cancer (Supplementary Fig. S4; Supplementary Tables S3 and S4). These results indicate that high SCNN1B expression predicts a favorable prognosis in patients with gastric cancer.

The frequent silencing of SCNN1B in gastric cancer and its association with patient survival prompted us to hypothesize that SCNN1B functions as a tumor suppressor. To this end, we generated four gastric cancer cell lines (AGS, BGC823, MGC803, and MKN45) with stable SCNN1B expression. Ectopic expression of SCNN1B was validated by RT-PCR and Western blot analysis (Fig. 3A), which was comparable with that of normal gastric tissues (Supplementary Fig. S5). SCNN1B overexpression suppressed colony formation ability by 55%–80% as compared with empty vector–transfected cells in all four gastric cancer cell lines (Fig. 3B; P < 0.01). Consistently, cell growth curve assay revealed that ectopic SCNN1B expression inhibited viability in these cell lines (Fig. 3C, P < 0.001).

To determine the cytokinetic effect of SCNN1B on gastric cancer cells, we analyzed apoptosis and cell-cycle distribution by flow cytometry. Overexpression of SCNN1B led to a significant increase in the total apoptotic cell population in AGS (P < 0.05), BGC823 (P < 0.01), MGC803 (P < 0.001), and MKN45 (P < 0.001) cells, as determined by Annexin V-PE/7-AAD dual staining (Fig. 3D). Induction of apoptosis by SCNN1B was confirmed by the elevated expression of key apoptosis markers such as cleaved forms of caspase-9, caspase-8, caspase-7, and PARP, as determined by Western blot analysis (Fig. 3D). We also observed an increased accumulation of gastric cancer cells in the G₁ phase (P < 0.05) and a reduction of S-phase population (P < 0.05) following ectopic SCNN1B expression (Fig. 3E). Consistent with G₁ arrest, we found that SCNN1B increased the expression of G₁-phase gatekeepers, p27^Kip1 and p53, while reducing expression of cyclin D1 and CDK2, both of which are important for G₁ progression (Fig. 3E). Next, we performed loss-of-function experiments using two independent SCNN1B-targeted siRNAs to knockdown endogenous SCNN1B in MKN1 and NCI-N87 cells (Fig. 3F; Supplementary Fig. S6A). The knockdown of SCNN1B increased colony formation ability (P < 0.001) and promoted cell-cycle progression in MKN1 and NCI-N87 cells (P < 0.01; Fig. 3F; Supplementary Fig. S6B and S6C). These data indicate that SCNN1B suppresses gastric cancer cell proliferation.

In light of the association between SCNN1B expression and metastasis of gastric cancer patients, we next ask whether SCNN1B has an effect on cell migration, adhesion, and invasion. SCNN1B overexpression markedly suppressed cell migration in AGS, BGC823, MGC803, and MKN45 cell lines by wound-healing assay. Quantitative analysis demonstrated a significant impairment in wound closure at different time points (P < 0.001) in SCNN1B-overexpressing cells, thereby suggesting that SCNN1B negatively regulates cell migration (Fig. 4A). SCNN1B also promoted cell adhesion in all four gastric cancer cell lines (Fig. 4B). In addition, Matrigel invasion assay revealed that ectopic expression of SCNN1B suppressed cell invasion in AGS, BGC823, and MGC803 cells by over 50% (P < 0.001; Fig. 4C). Conversely, siRNA-mediated SCNN1B silencing in MKN1 and NCI-N87 cells resulted in enhanced wound closure (P < 0.001), but decreased cell adhesion (P < 0.05) as compared with control (Fig. 4D and E; Supplementary Fig. S6D and S6E). Thus, SCNN1B reduces the metastatic ability of gastric cancer cells by inhibiting cell migration and invasion, while promoting cell adhesion.

In light of our in vitro results, we evaluated the impact of ectopic SCNN1B expression in the nude mice tumorigenicity assay. MKN45 and BGC823 cell lines with stable expression of empty vector or SCNN1B were injected into the left and right flanks of nude mice, respectively. As shown in Fig. 5, tumor growth was significantly slower in mice injected with MKN45-SCNN1B cells than those with MKN45-emtpy vector cells (P < 0.01; Fig. 5A) and in mice with BGC823-SCNN1B cells, than those with BGC823-emtpy vector cells (P < 0.01; Fig. 5B). The average tumor weight at sacrifice was significantly lower in MKN45-SCNN1B mice (P < 0.05; Fig. 5A) and in BGC823-SCNN1B mice (P < 0.05; Fig. 5B), as compared with their corresponding control mice. Ectopic SCNN1B expression in the tumor xenografts of MKN45-SCNN1B and BGC823-SCNN1B were confirmed by RT-PCR and Western blot analysis (Fig. 5A and B; Supplementary Fig. S7). Ki-67 staining revealed a significant reduction in cell proliferation in MKN45 tumors expressing SCNN1B (P < 0.05; Fig. 5A). These results supported a tumor-suppressive role for SCNN1B.

To further elucidate the molecular mechanism underlying the tumor-suppressive effect of SCNN1B, we performed coimmunoprecipitation (co-IP) of Flag-tagged SCNN1B in HEK293 cells, followed by mass spectrometry to identify its binding partners (Fig. 6A). The 78-kDa glucose-regulated protein (GRP78) was identified as a potential interacting partner for SCNN1B. GRP78 is a stress-inducible chaperone that normally resides in endoplasmic reticulum (ER); however, recent advances have shown that GRP78 plays an oncogenic role in cancer via supporting cell proliferation, invasion, and metastasis, and inhibition of apoptosis (15–17). We validated the interaction between SCNN1B and GRP78 in AGS and BGC823 cells (Fig. 6A), in which GRP78 was coimmunoprecipitated by Flag-tagged SCNN1B in both cell lines. We next evaluated colocalization of SCNN1B and GRP78 by immunofluorescence in AGS, BGC823, and MKN45 cell lines (Fig. 6B). Confocal microscopy images showed that GFP-tagged SCNN1B was found in the membrane and cytoplasm; whereas Myc-tagged GRP78 was expressed in membrane and cytoplasm. Colocalization of SCNN1B and GRP78 was observed mainly in the cytoplasm. These results indicate that SCNN1B is a binding partner of GRP78.

We next assessed the interplay between SCNN1B and GRP78. GRP78 mRNA levels were not altered following the ectopic expression of SCNN1B (Supplementary Fig. S8) On the other hand, SCNN1B expression strongly attenuated expression of GRP78 at protein level in gastric cancer cell lines (Fig. 6C). Moreover, SCNN1B decreased the expression of GRP78 in the cytoplasm and membrane that was consistent with their colocalization, thus implying that SCNN1B might regulate GRP78 via protein degradation (Fig. 6C).

Protein degradation in eukaryotic cells is mediated by two major pathways: ubiquitin–proteasome and autophagy–lysosomal pathways. To pinpoint the mechanism of SCNN1B-induced GRP78 degradation, we treated SCNN1B-overexpressing AGS cells with inhibitors of the proteasome (MG132) and lysosome (chloroquine). MG132, but not chloroquine, restored GRP78 protein levels in SCNN1B-overexpressing AGS cells (Fig. 6D), implying that the ubiquitin–proteasome pathway was involved in the degradation of GRP78. To validate this, we examined polyubiquitination of GRP78 with or without ectopic SCNN1B expression (Fig. 6D). Indeed, SCNN1B increased polyubiquitination of GRP78. Collectively, these data indicated that SCNN1B directly interacts with GRP78 and mediates GRP78 degradation via ubiquitination.

GRP78 is a central regulator of the UPR by sequestration of three canonical branches, PERK-eIF2α-ATF4, IRE1α-XBP1s, and ATF6 pathways. Given that GRP78 is downregulated by SCNN1B, we evaluated the activation of the three UPR signaling pathways. Increased expression of PERK, XBP1s, and ATF4 were demonstrated in SCNN1B-overexpressing gastric cancer cell lines by Western blot analysis (Fig. 6E). In addition, nuclear abundance of ATF4 and XBP1s was simultaneously induced in SCNN1B-expressing gastric cancer cells, suggesting that these transcription factors were activated (Fig. 6E). We also observed the upregulation of CHOP, a key mediator of UPR-mediated apoptotic pathway, in SCNN1B-overexpressing gastric cancer cell lines (Fig. 6E). To test whether induction of UPR plays an important role in tumor-suppressive effect of SCNN1B, we coincubated control and SCNN1B-expressing gastric cancer cells with the UPR stress inducer tunicamycin. We found that overexpression of SCNN1B sensitized AGS and MKN45 cells to the cytotoxic effect of tunicamycin as compared with controls. IC₅₀ values of tunicamycin in AGS-empty vector and AGS-SCNN1B cells were 368 and 234 ng/mL, respectively. A similar trend was observed in the MKN45 cell line, where empty vector cells (IC₅₀: 1,175 ng/mL) were less sensitive than SCNN1B-expressing cells (IC₅₀: 886 ng/mL) to tunicamycin (Fig. 6F). Collectively, induction of UPR plays an important role in tumor-suppressive function of SCNN1B in gastric cancer.

The tumor-suppressive effect of SCNN1B is dependent on downregulation of GRP78. Given that SCNN1B abrogated GRP78 expression through polyubiquitination, we next conducted rescue experiments in which we cotransfected the control and SCNN1B-expressing gastric cancer cells with empty vector or GRP78. We first evaluated the effect of ectopic GRP78 expression on cell proliferation. As shown in Fig. 7A, ectopic expression of GRP78 restored GRP78 protein levels in SCNN1B-overexpressing cells. Moreover, colony formation assay showed that GRP78 restored the number of cell colonies in SCNN1B-expressing cells to baseline levels in AGS cells (P < 0.01), whereas GRP78 overexpression did not promote colony formation in control cells (Fig. 7A). This indicated that growth-suppressive effect of SCNN1B is mediated by downregulation of GRP78. We next investigated the effect of GRP78 overexpression on metastatic capacity of SCNN1B-overexpressing AGS cells. While control AGS cells had comparable wound closure rate irrespective of GRP78 expression, GRP78 promoted wound closure in SCNN1B-overexpressing AGS cells (P < 0.001; Fig. 7B), implying that SCNN1B-mediated degradation of GRP78 contributed to its antimetastatic effect in gastric cancer cells. In contrast, siRNA-mediated knockdown of GRP78 was additive with ectopic SCNN1B expression to suppress GRP78 expression and inhibit AGS cell growth as compared with empty vector control (Fig. 7C). These findings pointed to a pivotal role of GRP78 modulation in the tumor-suppressive effect of SCNN1B in gastric cancer.

In this study, we identified that SCNN1B is readily expressed in normal gastric tissues, but is frequently silenced in gastric cancer cell lines and primary gastric cancer. Silencing of SCNN1B is associated with promoter methylation. Demethylation treatment with 5-Aza restored expression of SCNN1B, confirming that promoter hypermethylation mediates the transcriptional silencing of SCNN1B in gastric cancer.

SCNN1B gene silencing in gastric cancer suggests that SCNN1B may possess tumor-suppressive function and its downregulation may contribute to the development and progression of gastric cancer. Consistent with our hypothesis, the ectopic expression of SCNN1B in four gastric cancer cell lines (AGS, BGC823, MGC803, and MKN45) significantly suppressed cell proliferation in vitro, while its knockdown in MKN1 and NCI-N87 cells, which express endogenous SCNN1B, promoted cell viability. Tumor-suppressive effect of SCNN1B was validated in vivo, as evidenced by the diminished growth of SCNN1B-expressing MKN45 and BGC823 cells in nude mice. SCNN1B suppressed gastric cancer cell proliferation through apoptosis induction and inhibition of cell-cycle progression. SCNN1B overexpression induced apoptosis in gastric cancer cells by activating both intrinsic and extrinsic apoptosis pathways, leading to the cleavage of caspase-8, caspase-9, and that of the downstream effectors, caspase-7 and PARP. Moreover, SCNN1B inhibited cell-cycle progression at G₀–G₁ phase, which was associated with upregulation of p27^kip1 and suppression of cyclin D1 and CDK2. Cyclin D1/CDK2 forms an active complex that promotes G₁–S transition, while p27^kip1 binds to and inhibits cyclin D1/CDK2 activity (18). SCNN1B hence tips the balance of gene expression toward that of cell-cycle arrest. Inhibition of cell growth in vivo by SCNN1B was confirmed by the reduced Ki-67 index in SCNN1B-expressing MKN45 xenografts. Metastasis is a major cause of cancer-related deaths. Here, we revealed that SCNN1B functions as a metastasis suppressor in gastric cancer by inhibiting cell migration and invasion, and concomitantly promoting cell adhesion. Taken together, these results indicate that SCNN1B functions as a tumor suppressor by inhibiting cell growth and metastasis in gastric cancer.

To further elucidate the mechanism of action of SCNN1B, we performed co-IP and mass spectrometry, which led to the identification of GRP78 as an interacting partner. The direct interaction between SCNN1B and GRP78 was validated by co-IP in gastric cancer cells and their colocalization by confocal immunofluorescence microscopy. Moreover, ectopic SCNN1B expression reduced the protein expression of GRP78 without altering its mRNA expression. Instead, we showed that SCNN1B expression induced polyubiquitination of GRP78, which tagged the protein for degradation in the proteasome system (19). Reexpression of GRP78 in SCNN1B-expressing gastric cancer cells abrogated the inhibitory effects of SCNN1B on cell proliferation and cell migration, whereas GRP78 knockdown further aggravated SCNN1B-mediated growth inhibition. These data indicated that SCNN1B mediates its tumor-suppressive effect by regulating GRP78, which is a master regulator of the UPR response. GRP78 is frequently upregulated during cancer progression to counter UPR, maintain ER homeostasis, and promote cell survival (15, 16).Taken together, interplay between SCNN1B and GRP78 regulates the stability of GRP78, which has serious repercussion on cell survival and migration/invasion.

GRP78 controls the UPR via the sequestration of IRE1α, PERK, and ATF6 (20, 21). While UPR is initiated as a prosurvival mechanism, sustained activation of this pathway induces apoptotic cell death and cell-cycle arrest (22). We demonstrated that the ectopic expression of SCNN1B, through suppressing GRP78 expression, triggered the proapoptotic arm of the UPR response. This was exemplified by the increased PERK expression, which, in turn, induced expression and nuclear localization of transcription factors ATF4 and CHOP in SCNN1B-expressing gastric cancer cells. Tunicamycin, an UPR inducer, exacerbated the inhibitory effect of SCNN1B on gastric cancer cell proliferation, implying that modulation of UPR response plays an important role in the tumor-suppressive effect of SCNN1B. ATF4 and CHOP have been shown to induce apoptosis following prolonged stress, in part, by increasing protein load and ATP depletion (23). CHOP also initiates expression of proapoptotic genes such as DR5 (24), BIM (25), and PUMA (26). UPR induction has also been associated with G₁–S cell-cycle arrest via downregulation of cyclin D1 (27) and upregulation of p27^kip1 (28). GRP78 is also known to promote cancer metastasis, independent of UPR signaling (17). Collectively, our findings suggested that SCNN1B exerts a tumor-suppressive effect through its involvement in regulating the expression of GRP78 and the UPR response signaling pathway.

Finally, we investigated the clinical importance of SCNN1B expression and promoter methylation in primary gastric cancer. Using a TMA cohort with 245 primary gastric cancer patients, we found that expression of SCNN1B was an independent prognostic factor of favorable patient survival by multivariate Cox regression analysis. Moreover, SCNN1B protein expression was significantly associated with survival benefit in late-stage (TNM stages III/IV) gastric cancer. This was further validated by the association of high SCNN1B mRNA expression with improved survival of late-stage gastric cancer patients in another two gastric cancer cohorts (29, 30). Gastric cancer varies greatly in clinical outcome depending on the aggressiveness of individual tumors. Currently, TNM staging is clinically the most important predictor of patient survival in gastric cancer, and additional prognostic markers are necessary to provide a more accurate assessment of disease outcomes. SCNN1B was identified to be associated with patient outcome, and therefore our data suggested that SCNN1B may serve as a novel prognostic marker for gastric cancer patients. Moreover, SCNN1B was found to suppress gastric cancer cell growth; therefore, it may serve as a therapeutic target of gastric cancer.

In summary, we identified for the first time that SCNN1B acts as a tumor suppressor through induction of apoptosis and cell-cycle arrest, and inhibition of cell migration and invasion. We also uncovered that SCNN1B exerts its effect by direct interaction with GRP78, which led to its degradation and subsequent induction of UPR (Fig. 7D). The expression status of SCNN1B may serve as prognostic markers in primary gastric cancer.

F.K.L. Chan has received speakers' bureau honoraria from AstraZeneca, Pfizer, Takeda, and is a consultant/advisory board member for Eisai. No potential conflicts of interest were disclosed by the other authors.

Conception and design: Y. Qian, C.C. Wong, F.K.L. Chan, J. Yu

Development of methodology: J. Xu, Y. Zhang, P.W.Y. Chiu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C.C. Wong, H. Chen, W. Kang, H. Wang, L. Zhang, W. Li, M.Y.Y. Go, P.W.Y. Chiu, E.K.W. Ng

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Qian, H. Chen, W. Li

Writing, review, and/or revision of the manuscript: Y. Qian, C.C. Wong, E.K.W. Ng, F.K.L. Chan, J. Yu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Xu, H. Wang, E.S.H. Chu, M.Y.Y. Go, J.J.Y. Sung, J. Yu

Study supervision: C.C. Wong, J. Si, J. Yu

The authors thank Dr. Zhinong Jiang, Sir Run Run Shaw Hospital (Hangzhou, China) and Dr. Ye Cheng, Zhejiang Cancer Hospital (Hangzhou, China). These two pathologists both read and scored the TMA slides of gastric cancer patients.

This project was supported by research funds from RGC-GRF (14114615 and 766613) from Hong Kong; Shenzhen Municipal Science and Technology R & D Fund (JCYJ20130401151108652), and Shenzhen Virtual University Park Support Scheme to CUHK Shenzhen Research Institute; National Natural Science Foundation of China (NSFC; 81502064); and Direct grant for Research 2013/2014, CUHK (4054100).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.