ORAI2 Promotes Gastric Cancer Tumorigenicity and Metastasis through PI3K/Akt Signaling and MAPK-Dependent Focal Adhesion Disassembly

Gastric cancer is an aggressive disease with heterogeneous features that contributes to a notable proportion of cancer mortality worldwide (1). Most of the gastric cancers are adenocarcinomas with various molecular and histologic subtypes influenced by numerous genetic and environmental factors (2). Although evidence suggests a decline in incidence over the past few years, which is thought to be the result of improved environmental hygiene and lifestyle, gastric cancer still ranks as the world's third leading cause of cancer-related death due to ineffective treatment, and the lack of early diagnostic markers with high specificity and sensitivity (2, 3). Owing to its heterogeneity and complexity, gastric cancer is associated with poor prognosis, and patients with gastric cancers are often diagnosed at the advanced stage (3). Even for those who were newly diagnosed, regional metastasis is still found in the majority (4, 5). Lymph node status is one of the most important indicators of prognosis and recurrence in the treatment of gastric cancer (4). It was revealed that the number of metastatic lymph nodes was negatively correlated with overall survival of patients with gastric cancer following curative resection (6). Despite increasing treatment options that are currently available, outcomes for metastatic gastric cancer are still dismal (7).

ORAI2 belongs to the ORAI family, a gene of emerging interest in metastatic microenvironmental study owing to its association with cell process–regulating ubiquitous second messenger calcium (8, 9). Upon endogenous depletion of Ca²⁺ stores, activated STIM family proteins oligomerize and migrate to endoplasmic reticulum (ER)–plasma membrane junction where they activate and recruit ORAI family proteins through direct interactions (8, 10). Together, they form Ca²⁺ release–activated Ca²⁺ (CRAC) channels and promote store-operated Ca²⁺ entry (SOCE), engaging Ca²⁺ release from the extracellular space (11). Such Ca²⁺ influx with resultant elevated intracellular Ca²⁺ level has long been known as a key in cellular cycles, and altered expressions of ORAIs are believed to regulate tumor metastasis in different cancer types (9–12). Evidences have shown that abnormal activation of SOCE leads to tumor cell migration and cancer metastasis, and local Ca²⁺ influx is essential for building front-to-rear Ca²⁺ gradient, which serves to maintain the morphologic polarity required by migrating cells (13–15). Xia and colleagues firstly identified that increased expression of ORAI1 and STIM1 in gastric cancer facilitated SOCE and thus promoted gastric cancer cell proliferation, metabolism, migration, and invasion through the activation MACC1 (16). In addition to “conventional” interaction between ORAI-STIM that underlies SOCE activation, more complicated “unconventional” mechanisms, which are independent of ER Ca²⁺ stores or STIM sensors, have been identified (14). For instance, binding of ORAI1 to the secretory pathway Ca²⁺-ATPase enables the constitutive Ca²⁺ influx, thus conferring the tumorigenesis in breast cancer (17). Besides, differential expression of SOCE components may be responsible for the cross-talk with other signaling pathways. A switch in the SOCE composition from ORAI1 to ORAI3 was observed particularly in estrogen receptor–positive (ER⁺) breast cancer cells, in contrast to the canonical STIM1/ORAI1 pathway in ER⁻ breast cancer cells (18). Among these studies, ORAI1 is the best-characterized ORAI isoform, whereas current knowledge about ORAI2 and ORAI3 is sparse. Several earlier publications reported that ORAI2 appeared to be ubiquitously expressed in B cells, dendritic cells, macrophages, platelets, and melanocytes, while another study found high expression of ORAI2 in native T cells and thus declared that it was essential in modulating T-cell–mediated immune response (19, 20). Moreover, studies of ERs in parathyroid adenomas revealed an increased expression level of ORAI2 after treatment with diarylpropionitrile and 4-hydroxytamoxifen, thus suggesting a tumor-suppressive role of ER via manipulating ORAI2 expression for parathyroid [Ca²⁺]_i regulation (21). Even though a recent study of promyelocytic cell line HL60 indicated that ORAI2-mediated SOCE resulted in FAK tyrosine phosphorylation and hence HL60 cell migration, the role of ORAI2 in remodeling of cell cytoskeleton, formation of invadopodia, regulation of SOCE oscillations, or cross-talks with other signaling pathways still remain poorly characterized (22).

In this study, we set out to evaluate the expression status of ORAI2 and its clinical relevance in gastric cancer, where we found the presence of ORAI2-positive cells was closely correlated with more aggressive malignancies and poorer survival outcomes. In vitro and in vivo functional assays with ORAI2 ectopic–overexpressing or ORAI2-knockdown gastric cancer cell lines indicated strong abilities of ORAI2 in tumorigenicity and metastasis. We also elucidated the underlying mechanisms by which ORAI2 promotes tumor cell growth and cancer metastasis with particular focus on focal adhesion (FA), an essential process in migrating cells.

All human specimens were obtained from patients with gastric cancer who underwent surgery at Xinyang People's Hospital (Henan, China). The study protocol was approved by the regional Committees for Ethical Review of Research and all the procedures were in accordance with the 1964 Helsinki Declaration. Written informed consent was obtained from all participants in this study. Immortalized gastric cancer cell lines used in this study were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China), cultured in DMEM or RPMI1640 medium, supplemented with 10% FBS (Gibco), 100 U/mL penicillin and 0.1 mg/mL streptomycin (Gibco) under a humidified atmosphere of 5% CO₂ at 37°C. Cell lines at passages greater than 10 were not used in this study. All cell lines were frequently examined to ensure they were free of Mycoplasma contamination.

Migration and invasion assays were performed in 24-well Millicell hanging insert (Millipore) or 24-well BioCoat Matrigel Invasion Chambers (BD Biosciences). For the invasion assay, cells were starved with serum-free medium for 24 hours before the assay. A total of 5 × 10⁴ cells were seeded inside the top chamber with serum-free medium, whereas medium supplemented with 10% FBS was filled into the bottom chamber that served as the chemoattractant. After 48- or 72-hour incubation, cells that migrated or invaded through the membrane (migration) or Matrigel (invasion) were fixed, stained with Crystal Violet, air dried, and mounted on the slides. The number of cells were counted in 10 fields and imaged using SPOT imaging software (Nikon). Data are expressed as the mean ± SD of triplicate wells within the same experiment.

All animal procedures were carried out with the approval of the Committee on the Use of Live Animals in Teaching and Research (CULATR) of the University of Hong Kong. For in vivo tumor formation assays, gastric cancer cells (1 × 10⁶) after transfection of ORAI2 or shRNAs and their empty vector controls suspended in PBS were subcutaneously injected into the left and right dorsal flanks of 4- or 5-week-old male BALB/c nude mice. The growth of tumors was measured every week and mice were sacrificed after 5 weeks. The average tumor volume (TV) was calculated according to the equation: TV (mm³) = length × width² × 0.5. For in vivo metastasis assays, 3 × 10⁵ cells in 20 μL PBS were injected into the right hind footpads of 4- or 5-week-old male NOD/SCID mice. Injected cells were luciferase-labeled so that the popliteal lymph node metastasis could be monitored and imaged with a bioluminescence imaging system (PerkinElmer, IVIS Spectrum Imaging System). All of the mice were euthanized 8 weeks after injection. The right-side popliteal lymph nodes were enucleated, measured, paraffin-embedded, and sectioned. To validate the invasion of gastric cancer cells in lymph nodes, serial 4.0-mm sections were obtained and analyzed by IHC using anti-CK7 staining. For experiments with pharmacologic inhibitor SKF96365, 1 × 10⁶ gastric cancer cells were subcutaneously injected into the dorsal flanks of 4- or 5-week-old male BALB/c nude mice. Four days after tumor cells were injected, mice were randomly divided into two groups. The mice in the experimental group were intraperitoneally injected with 100 μL SKF96365 (5 μmol/L) at the dose of 25 mg/kg every other day for 28 days, whereas the control group only received DMSO treatment. After 5 weeks, mice were sacrificed, and tumors were measured.

Cells were scraped and lysed in RIPA-buffer (Thermo Fisher) containing protease inhibitors cocktail. Equal amounts of denatured protein were resolved by 10% SDS-PAGE and transferred to 0.45-μm polyvinylidene fluoride membranes (IPVH00010, Millipore). Membranes were then washed in TBST buffer (TBS containing 0.1% Tween-20) and nonspecific binding sites were blocked by immersing the membranes in blocking buffer containing 5% nonfat milk (Cell Signaling Technology) TBST buffer for 1 hour at room temperature prior to addition of primary antibody. Proteins were detected by chemiluminescence using the Bio-Rad ECL reagent according to the manufacturer's instructions.

Cells (5 × 10⁵) were serum starved overnight for synchronization and allow to resume cell-cycle transition by adding medium supplemented with 10% FBS in the presence or absence of 20 μmol/L SKF96365. Cells were harvested using 0.1% trypsin in 2.5 mmol/L EDTA, washed twice with ice-cold PBS, and then fixed in 1 mL ice-cold 70% ethanol for at least 2 hours. After fixation, cells were washed twice with ice-cold PBS and resuspended in 100 μL RNase A (50 μg/mL; Sigma-Aldrich) in PBS at 37°C for 30 minutes. After incubation with 400 μL propidium iodide (PI; 20 μg/mL; Sigma-Aldrich) at 25°C in the dark for 5 minutes, DNA contents of the stained cells were determined at 560 nm by flow cytometer BD FACSCanto II Analyzer (BD Biosciences) using doublet discrimination gating. All analyses were performed in triplicate and images were analyzed by FlowJo software (Tree Star Inc).

Gastric cancer cells stably expressing ORAI2-shRNAs or empty vectors were seeded on 3.5-cm fibronectin-coated glass-bottom dishes, transfected with plasmid encoding paxillin-pEGFP (23), and experiments were performed 24 hours later at 37°C. Paxillin-pEGFP was a gift from Rick Horwitz (Addgene plasmid # 15233; http://n2t.net/addgene:15233; RRID:Addgene_15233). Fluorescence recovery after photo-bleaching (FRAP) was carried on an inverted microscope with 60× objective (Nikon Eclipse Ti-E). A small area within a single FA was selected (∼200 nm) so that the GFP-paxillin was rapidly bleached using the Micro-Point ablation laser system equipped with a nitrogen-pumped dye laser delivering a wavelength of 488 nm to achieve nearly complete bleaching for each FA. The fluorescence recovery was then recorded at maximum speed by the EMCCD camera of a spinning-disc confocal microscope (PerkinElmer UltraVIEW VoX) until the intensity reached a plateau.

The Fluo-4 NW Calcium Assay Kit was obtained from Invitrogen Life Technologies. Gastric cancer cells were plated in a 96-well plate (2 × 10⁴ cells/well) and grown overnight to allow the cells to adhere to the plate. The following day, the cells were loaded with 100 μL of the 1X Fluo-4-NW dye mix loading solution containing 2 μmol/L Fluo-4 AM and 2.5 mmol/L probenecid, then the plate was incubated at 37°C for 30 minutes and at room temperature for an additional 30 minutes. For measurement of SOCE, 2 μmol/L thapsigargin (TG; Sigma-Aldrich, Inc.) was added to deplete ER calcium stores 30 seconds after incubation in the Hanks' Balanced Salt Solution (HBSS). Ca²⁺ influx was induced by subsequent addition of 2 mmol/L Ca²⁺-containing HBSS after 300 seconds of thapsigargin treatment. The green fluorescence of Fluo-4 was excited by the Argon laser at 485 nm and recorded through a 535 nm channel of a fluorescence microplate reader (Wallac VICTOR3).

To investigate differential gene expression patterns during gastric cancer metastasis, RNA sequencing was applied in 9 pairs of primary tumors (T), their corresponding nontumor (N) tissues and lymph nodes derived from patients with gastric cancer. Notably, 33 upregulated genes were identified in 8 gastric cancer cases (Fig. 1A) and their relative expression profiling was illustrated by heatmap (Fig. 1B). On the basis of literature reviews and functional predictions of these genes, ORAI2 was of particular interest in this study. ORAI2 displayed a generally higher expression in lymph nodes, when compared with the tumor and nontumor tissues (Fig. 1C), as well as compared with its isotypes ORAI1 and ORAI3 (Fig. 1C). Because its isotype ORAI1 was found to promote breast cancer metastasis (15) and there was a recently recognized function of locally confined Ca²⁺ in the programming of directed cell migration (24), we hypothesized that increased expression of the subtype-specific ORAI2 may facilitate tumorigenesis and metastasis by inducing SOCE and signaling activations in gastric cancer.

To see whether ORAI2 exhibits a differential expression pattern in gastric cancer, we analyzed 450 gastric cancer cases from The Cancer Genome Atlas (TCGA) database. It appeared that the expression of ORAI2 in nontumor (n = 35) was much lower than that in gastric tumor samples (n = 415; Fig. 1D). Among the differential ORAI2 expression frequencies of 412 informative gastric cancer cases, the patients with gastric cancer were almost equally divided into low-expression group (≤9.20, n = 205, 49.8%) and high-expression group (>9.20, n = 207, 50.2%). The upregulation of ORAI2 in gastric cancer is closely associated with poor overall survival time (log-rank test, P = 0.1235; Fig. 1E) and disease-free survival time (log-rank test, P = 0.0141; Fig. 1F), by Kaplan–Meier survival analysis. Then, we assessed the clinicopathological characteristics of ORAI2 and found that its upregulation was significantly correlated with poor differentiation (P < 0.0001) and invasion (P = 0.033; Supplementary Table S1). To further explore the clinical relevance of ORAI2 in the metastatic gastric cancer, the expression of ORAI2 was analyzed by IHC on a TMA consisting of 43 paired primary gastric cancer, adjacent nontumor and lymph node specimens. Noninformative samples including lost samples and unrepresentative samples were not included in data compilation. The representative paired clinical samples are shown in Fig. 1G. Semiquantitative IHC scoring standard was applied to evaluate the expression intensity of ORAI2 (Supplementary Fig. S1). It showed that the expression of ORAI2 was particularly high in lymph nodes, when compared with their corresponding nontumor tissues (P = 0.0001; Fig. 1H). In Kaplan–Meier survival analysis of patients with different lymph node ORAI2 expression levels, high lymph node ORAI2 expression (Score = 3) was found significantly associated with poor overall survival (log-rank test, P < 0.0001; Fig. 1I) and poor disease-free survival (log-rank test, P = 0.0047; Fig. 1J), when compared with low lymph node (Score = 1) and median lymph node (Score = 2) ORAI2 expression, suggesting ORAI2 as a prognostic indicator for clinical outcomes of gastric cancer.

To investigate whether ORAI2 possesses tumor-promoting function, two gastric cancer cell lines SGC7901 and NUGC4 with low ORAI2 expression level were stably transduced with lentivirus packaged with either a full-length human ORAI2 or an empty vector (Vec) to generate ORAI2-overexpressing or control cells. Similarly, SUN216 and NCI-N87 cell lines with high ORAI2 expression level were transduced by lentiviral particles encoding either ORAI2-specific short hairpin RNAs (shRNA2 and shRNA3) or a nontarget control (Ctl) to create stable knockdown (shORAI2-2 or shORAI2-3) or control cells. Successful overexpression and knockdown of ORAI2 were then confirmed at both mRNA and protein levels by qPCR, immunofluorescence, and Western blotting, respectively (Fig. 2A–C; Supplementary Fig. S2). Tumorigenic ability of ORAI2 was assessed by in vitro functional assays, including proliferation XTT assay, foci formation assay, and soft agar assay. Compared with control cells, ORAI2-overexpressing cells displayed increased proliferation rate (Fig. 2D), enhanced colony-forming abilities in soft agar (Fig. 2E), and higher foci formation frequencies (Fig. 2F). Conversely, knockdown of ORAI2 expression revealed an opposing effect (Fig. 2D–F). To further assess in vivo tumor-promoting ability of ORAI2, ORAI2-overexpressed SGC7901 or ORAI2-repressed NCI-N87 cells and their respective control cells were subcutaneously injected into the right and left dorsal flank of nude mice. The results showed that overexpression of ORAI2 could significantly shorten the latency and increase the mean tumor volume than those formed by control cells (Fig. 2G), whereas knockdown of ORAI2 exerted opposite effect (Fig. 2H). All these findings indicate the tumor-promoting ability of ORAI2 in gastric cancer, both in vitro and in vivo.

As our previous study indicated that upregulation of ORAI2 was significantly associated with tumor invasion and lymph node metastasis in patients with gastric cancer, the effect of ORAI2 on cell motility was therefore studied by in vitro cell migration and invasion transwell assays. When the cells were transfected with ORAI2, enhanced cell migration and invasion abilities were observed, as compared with the vector controls (Fig. 3A). Once endogenous ORAI2 was silenced by shRNAs, the ability of cell migration and invasion was remarkably decreased correspondingly (Fig. 3B). Nevertheless, cells were transiently transfected with lentiviral particles encoding either a human ORAI2 or an empty vector to rescue the phenotypes of SUN216 knockdown (shORAI2-3) cells. In our results, the rescued version of shRNA-SUN216 cells displayed an increased ability of migration and invasion when compared with the control cells, indicating there were no off-target effects associated with the shRNAs. To study the effect of ORAI2 on tumor metastasis in gastric cancer, a lymph node metastasis animal model was adopted by injecting ORAI2-overexpressed or -suppressed cells, as well as their vector controls, into the right hind footpads of NOD/SCID mice. The metastatic tumors were monitored in vivo under bioluminescence imaging system for a period of 8 weeks and the popliteal (sentinel) lymph nodes, which represent the first step of cancer spread, were then examined (25). In the ORAI2-overexpressed group, swollen popliteal lymph nodes were observed in all 5 mice, whereas only 1 in 5 mice developed lymph node metastasis in control group (Fig. 3C). Accordingly, the lymph node size derived from ORAI2-transfected group was significantly larger than those formed by vector cells in the control group (Fig. 3D) and the infiltration of tumor cells was then confirmed by CK7 staining (Fig. 3E). Conversely, silencing of ORAI2 expression resulted in less swollen (1/3) and smaller popliteal lymph nodes, when compared with the controls (Fig. 3F and G). Taken together, our data suggested ORAI2 might promote gastric cancer metastasis by enhancing gastric cancer cell motility.

Given the fact that numerous studies have indicated ORAI1 and STIM1-activated CRAC channel could modulate SOCE in many cancer types, whereas ORAI2 and its SOCE partners have never been identified in gastric cancer, we next investigate whether STIM1 was putative CRAC channel partner of ORAI2 in gastric cancer cells. Intriguingly, when ORAI2 was overexpressed, the mRNA level of STIM1 was not found elevated correspondingly (Fig. 4A, left). However, once the endogenous silencing of ORAI2 was performed, the mRNA levels of STIM1 remarkably reduced, suggesting a differential involvement of STIM1 in gastric cancer (Fig. 4A, right). At the protein level, the expression of STIM1 strikingly increased following the overexpression of ORAI2 and decreased after the depletion of ORAI2, suggesting ORAI2 may affect the translation of STIM1 in gastric cancer (Fig. 4B). Nevertheless, the coexpression of ORAI2 and STIM1 was confirmed by IF (Fig. 4C) and IHC (Fig. 4D). To assess ORAI2-dependent SOCE, Fluo-4 AM probe was adopted to monitor intracellular Ca²⁺ oscillations in gastric cancer cells, and TG, an inhibitor of sarco/ER Ca²⁺-ATPase (SERCA), was applied to deplete ER Ca²⁺ stores. The initiation of Ca²⁺ influx via SOC channels was triggered by addition of 2 mmol/L Ca²⁺ to gastric cancer cells following 5 minutes incubation of TG in Ca²⁺ free solution. Compared with the control cells, notable and sustained elevation of Ca²⁺ influx was observed in ORAI2-transfected cells, where the peak fluorescence reading relating to the intracellular calcium concentration ([Ca²⁺]_i) of SOCE was amplified by approximately two-fold (Fig. 4E; Supplementary Fig. S3). When this assay was done in shORAI2-2 and shORAI2-3 gastric cancer cells, the SOCE-induced Ca²⁺ influx showed a clear decrease, in which the peak fluorescence was reduced by approximately 80% and 60%, respectively (Fig. 4F). Therefore, our results demonstrated the existence of ORAI2-dependent SOCE in gastric cancer.

In an attempt to investigate the mechanism by which ORAI2 enhanced cell proliferation ability, we studied the effects of ORAI2 overexpression on cell-cycle progression by flow cytometry. After synchronization, the cells in G₀–G₁ phase were arrested by Mimosine, an inhibitor that arrests cell-cycle progression; 10% FBS was provided to resume the cell growth. At the 6-hour time point, ORAI2-transfected SGC7901 cells had a conspicuously higher S-phase cell accumulation (S: 59.3%) than that in vector group (S: 40.7%), suggesting ORAI2 could induce faster S-phase entry in gastric cancer (Fig. 5A). Moreover, Western blot analysis revealed a significant rise in the expression levels of phosphorylated Rb (S780 and S807/811) and a series of cyclins (cyclin A, cyclin E, and cyclin D1), as well as a rapid turnover of P-27 in ORAI2-overexpressed SGC7901 cells, when compared with their vector controls (Fig. 5C). We next examined whether cell-cycle–regulatory pathways were upregulated by ORAI2. Immunoblot results showed that the expression level of phosphorylated PI3K (Tyr458) was particularly elevated in ORAI2-overexpressed cells and attenuated in ORAI2-silenced cells, as compared with the controls, of which, class I (catalytic isoforms p110α and p110β) and class III PI3Ks were mostly correlated with the expression changes of ORAI2, suggesting they might participate in the PI3K regulatory pathway in gastric cancer (Fig. 5D). Moreover, aberrant phosphorylation at the specific Ser380 residue of PTEN was identified in ORAI2-transfected cells, indicating an early event of ORAI2-mediated inactivation of PTEN influenced the downstream PI3K/Akt signaling (Fig. 5E). Specifically, the expression of PI3K/Akt pathway–related proteins changed upon manipulation of the ORAI2 expressions in gastric cancer cells, where we found c-Raf at Ser259 residue and Akt at both Ser473 and Thr308 residues were increased in ORAI2-overexpressed cells and decreased in ORAI2-repressed cells, when compared with their respective controls (Fig. 5E). GSK-3β has been known to play an inhibitory role in cell-cycle progression as it can mediate the degradation of both cyclin E and cyclin D1 (26). Our data revealed that ORAI2-transfected cells dramatically increased the phosphorylation level of GSK-3β (inactive form), and therefore resulted in an enhanced expression of cyclin D1 (Fig. 5E). These findings implied that the tumorigenic ability of ORAI2 was achieved by activating the PI3K/Akt pathway and thus accelerated cell-cycle progression in gastric cancer.

Because SOCE is the immediate downstream target of ORAI2, we next investigated whether blockade of SOCE would influence ORAI2-mediated tumorigenesis. After applied the SOCE inhibitor SKF96365 (20 μmol/L), ORAI2-induced cell-cycle progression was analyzed by flow cytometry. The results showed that the percentage of the cell population at S-phase was significantly decreased in both control and ORAI2-overexpressed group, with a concomitant increase of the percentage in G₁ (Fig. 5B). Subsequently, in vitro XTT proliferation assay revealed that treatment of SKF96365 dramatically decreased the rate of cell proliferation both in control and ORAI2-transfected groups (Fig. 5F). In addition, the cell growth rate was found reduced more severely in ORAI2-silenced cells, when compared with the control cells (Fig. 5G). The treatment concentration was based on [IC₅₀]_c when cell viability was reduced to 50% (Supplementary Fig. S4). Moreover, we examined whether the PI3K/Akt pathway could also be influenced by ORAI2-mediated SOC activity. Our Western blot results demonstrated that SKF96365 and FAK autophosphorylation inhibitor Y15 significantly attenuated the phosphorylation levels of PI3K/Akt pathway–related proteins and cell-cycle–related proteins in both ORAI2-overexpressed and ORAI2-suppressed groups, when compared with the respective DMSO-treated groups, suggesting that the treatment of SKF96365 and Y15 could reduce ORAI2-mediated PI3K/Akt signaling in gastric cancer cells (Fig. 6A–C). Then, the effect of SOCE activity was studied in vivo by intraperitoneal injection of SKF96365 or DMSO in mice bearing subcutaneous tumors formed by ORAI2-overexpressed cells and their control cells. In contrast to the DMSO group (Fig. 6D), the treatment of SKF96365 significantly increased the latency and decreased mean tumor volume in mice (Fig. 6E and F). Subsequently, the recovered mice tumors were then subjected to Western blots for investigating the PI3K/Akt activity changes. Our results showed that the treatment of SKF96365 remarkably decreased ORAI2-mediated PI3K/Akt signal transduction and its downstream substrates GSK-3β and cyclin D1, thus exerting negative regulatory effect on cell-cycle progression and tumorigenesis (Fig. 6G). Given these findings, we demonstrated that blockade of SOCE could impair gastric cancer tumor growth both in vitro and in vivo, a process that may function through inhibiting the PI3K/Akt pathway. Nevertheless, to further confirm that ORAI2-induced CRAC current was responsible for tumorigenesis in gastric cancer, another unspecific SOCE inhibitor 2-APB was employed in our study. After addition of 2-APB at 50 μmol/L, the concentration based on [IC₅₀]_c when cell viability was reduced by 50% (Supplementary Fig. S5), cell migration, invasion, and proliferation abilities were then evaluated by transwell and XTT assays. Our results indicated that 2-APB exhibited similar actions as SKF96365 on gastric cancer progression that it could significantly diminished ORAI2-induced gastric cancer cell migration, invasion, and proliferation (Fig. 6H and I), suggesting a potential involvement of ORAI2-mediated SOC activities in gastric cancer progression.

In cancer cell migration, the dynamic assembly and disassembly of FA contacts is a crucial determinant for the formation of an aggressive invadopodium (27). Immature FAs, which are small-sized nascent adhesions, are typically observed at the rear-edge of membrane protrusions in rapidly migrating cells. This is contrast to mature FAs, which are larger and more stable adhesions with tethered actin stress fibers that tightly adhere to the extracellular matrix (ECM) underneath the cell body, and cells with these mature FAs are often nonmigratory or move slowly (28, 29). To investigate whether ORAI2 affected FA dynamics and consequently migratory behavior of gastric cancer cells, we performed IF by dual-staining Paxillin and F-actin in ORAI2-knockdown cells. Paxillin, a major component of FAs, serves as a structural adaptor protein, links between ECM and cytoskeleton, and it is highly dynamic and tend to turnover at the base of newly formed protrusions (23). The IF results indicated that, as compared with the small-sized paxillin-containing FAs formed in the control cells, ORAI2-suppressed cells displayed elongated and larger FAs that were predominantly localized in the inner part of the cell body (Fig. 7A). In addition, the defects of FA dynamics in ORAI2-silencing cells were further studied by FRAP. Specifically, cells were transfected with Paxillin-GFP, which allowed the intracellular FA dynamics to be recorded by time-lapse spinning disc confocal microscopy. Then, a selected FA region (0.2–0.5 μm²) was photobleached by a 488-nm Argon laser at 100% intensity. Our results indicated that the deficiency of ORAI2 could cause a significant increase in the fluorescence recovery time after FRAP, when compared with the controls (Fig. 7B; Supplementary Video S1–S3). Together, these data provided a compelling evidence that ORAI2 was responsible for faster FA dynamics in gastric cancer, suggesting ORAI2 may regulate cell migration by inducing FA turnover.

Focal adhesion kinase (FAK), a nonreceptor protein tyrosine kinase located in the FA complex, has been implicated in the interaction that affects cell migration, invasion, and metastasis via crosstalking with Src, integrin, and growth factor receptor signaling pathways (30). To assess the FAK-dependent signaling participated in ORAI2-mediated FA turnover, we examined the impact of ORAI2 on FAK phosphorylation status and its associated downstream partners. Western blot analysis indicated that stimulation of cells overexpressing ORAI2 particularly enhanced the phosphorylation of FAK at Tyr397 site, whereas silencing of ORAI2 expression resulted in a decrease in FAK phosphorylation. On the basis of the fact that Tyr397 is an autophosphorylation site of FAK that creates a docking site for Src-homology 2 (SH2)-containing proteins, which are subsequently recruited to promote FAK phosphorylation at more phosphorylation sites, we then examined the phosphorylation status of FAK at Tyr576/577 and Tyr925 sites, in which we found Tyr925 phosphorylation was the most unaffected (31). Paxillin is a substrate for FAK–Src complex, and the phosphorylation of paxillin at Tyr118 is a pivotal process of coordinating integrin-mediated cell motility by driving FA disassembly (29). This prompted us to explore whether ORAI2 activity could affect the phosphorylation of paxillin. As expected, the Western blot revealed a dramatic increase in the level of phosphorylated paxillin (Tyr118) in cells transfected with ORAI2, and a clear decrease in ORAI2-suppressing cells, when compared with their respective controls (Fig. 7C). Previous literatures reported that FAK-Src-mediated phosphorylation of paxillin (Tyr118) could recruit and subsequently activate ERK, hence triggering the activation of MAPK/ERK cascade (31, 32). Therefore, we next evaluated the expression of signaling molecules in MAPK pathway, where we found that the stimulation of ORAI2 overexpression could enhance the expression level of phosphorylated C-raf, B-raf, MEK1/2, and ERK1/2 (Fig. 7C). To further confirm the effect of ORAI2 on FA was through the FAK/MAPK/ERK signaling, a FAK autophosphorylation inhibitor Y15 was employed in this study. The treatment concentration was based on [IC₅₀]_c when cell viability was reduced by 50% (Supplementary Fig. S6). Our results showed that addition of Y15 and SKF96365 effectively attenuated FAK, ERK, and paxillin phosphorylation in ORAI2-overexpressed cells, and more severely reduced phosphorylation levels of these molecules were observed in ORAI2-silenced cells (Fig. 7D and E). Because it was suggested that the phosphorylated Y397 residue of FAK can create a high-affinity binding site for SH2-containing proteins, such as Src and PI3K, which functions upstream of Src and PI3K-related signaling during cell–ECM migration (33), we hypothesized that FAK may serve as a point of cross-talk between PI3K/Akt and MAPK/ERK pathways. As shown previously in Fig. 6C, inhibition of FAK remarkably alleviated ORAI2-mediated activation of PI3K signaling by decreasing the phosphorylation levels of PI3K, Akt, and c-Raf. Moreover, the cell-cycle–related proteins were reduced correspondingly. Therefore, our data demonstrated that the disruption of formation of FAK complexes with Src and PI3K could result in decreased signal transduction of both PI3K/Akt and MAPK/ERK pathways, as well as suggested a potential cross-talk between these pathways. Subsequently, functional assays revealed that the treatment of Y15 could significantly inhibit ORAI2-induced cell proliferation (Fig. 7F), migration and invasion (Fig. 7H), as compared with the untreated group. In addition, more severe reduction of cell proliferation, migration, and invasion were observed in gastric cancer knockdown cells (Fig. 7G and I). Together, our data implied that ORAI2 may contribute to the transmission of signals from integrins to downstream effectors, and then strengthened the cross-talks between FAK–paxillin axis and MAPK, therefore underlining FA dynamics and ultimately cell migration, invasion, and metastasis in gastric cancer.

On the basis of our findings that indicated that ORAI2 could increase SOC activity and was responsible for the dynamic regulation of FA, we hypothesized that ORAI2-regulated SOC activity could affect cell migration and invasion. After applying the SOCE inhibitor SKF96365 at previously tested concentration, the migration and invasion ability of gastric cancer cells were evaluated by transwell assays. Our results showed that the number of migrated and invaded cells in the SKF96365-treated group were decreased by approximately 70%–90%, as compared with the DMSO group (Fig. 7H and I). Also, we explored whether the FAK/MAPK/ERK pathway could be also influenced by ORAI2-mediated SOC activity. As previously shown in Fig. 7E, addition of SKF96365 caused an obviously faded phosphorylation of FAK, ERK, and paxillin in both ORAI2-overexpressed and ORAI2-suppressed group, as well as their vector controls. Therefore, these findings suggested that ORAI2-mediated SOC activity was correlated with the abilities of migration and invasion in gastric cancer cells and was dependent on the underlying FAK/MAPK/ERK signaling pathway.

Metastasis is a multi-step process wherein cancer cells break away from its initial site, travel through the blood or lymph system, and form new tumors in other parts of the body (34). To disseminate from primary tumors and build secondary tumors, cancer cells must first acquire the abilities of migration and invasion, which lay a groundwork for the subsequent infiltration into the circulation and the penetration into the capillary endothelium (9, 34). In fact, when cancer cells escape from their primary home, they may not go for a distant site immediately, but rather settle down in adjacent lymph nodes (35). As the presence of lymph node metastasis is closely linked to further spread of cancer cells in distant organs, the regional lymph node metastasis has been widely recognized as a criterion in determining the cancer stage (36).

The effect of intracellular Ca²⁺ concentration homeostasis regulated by other Ca²⁺-permeable channels on cellular processes in gastric cancer have been intensively investigated in previous researches. For instances, upregulation of the transient receptor potential vanilloid 2 and 4 (TRPV2 and TRPV4), members of the transient receptor potential (TRP) superfamily, were demonstrated to predict the prognosis of patients with gastric cancer and facilitate Ca²⁺ signaling axis to potentially promote gastric cancer progression (37–39). On the basis of the fact that Ca²⁺ is a versatile intracellular messenger that links to the activation of many signaling pathways (40, 41) and ORAI family is long known to be associated with SOCE (42), we investigated and then confirmed that ORAI2 and STIM1 cofacilitate SOCE in gastric cancer. Because proliferating cells require Ca²⁺ entry to facilitate cell-cycle transition, it is not surprising that fast proliferating tumor cells remodel their Ca²⁺ signals (43). To explore the underlying mechanism of ORAI2-induced tumorigenesis, we conducted a cell-cycle study and analyzed the expression profiles of a series of cell-cycle–regulating molecules. Previous literatures have indicated that cyclin E is essential to drive G₁–S transition and cyclin A drives the S-phase to mitosis (44). Cyclin E levels are constantly high in early embryos, which allows CDK2 to launch G₁–S transition immediately when mitosis is over (45). Subsequently, CDK4/6-Cyclin D1 and CDK2-Cyclin E complexes phosphorylate Rb, resulting in restrain of its growth suppression properties, in turn promotes the G₁–S cell-cycle transition (46). On the other hand, P27 inhibits G₁–S CDKs by forming heterotrimeric complexes with them and the downregulation of CDK1 activity may be responsible for uncontrolled cell growth in those malignancies with defect P-27 expression (44, 46). Cyclin D1 is an unstable molecule, which is transported by GSK-3β from nucleus to cytoplasm for ubiquitin-mediated degradation (26, 46). However, GSK-3β could be inactivated by PI3K/Akt-induced phosphorylation (46). This is consistent to our data that showed an enhanced PI3K/Akt activity and increased phosphorylation level of GSK-3β in ORAI2-transfected cells. Moreover, we investigated the ORAI2-modulated SOC activity in tumorigenesis by employing a SOCE inhibitor SKF96365, where we found SKF96365 significantly decreased the ORAI2-mediated cell growth both in vitro and in vivo. Together, we concluded that ORAI2-regulated fast cell-cycle transition relied on PI3K/Akt signaling and ORAI2-modulated SOC activity was crucial in regulating it.

To further dissect the mechanisms of ORAI2-induced cell migration, invasion, and metastasis in gastric cancer, we predominantly focused on FA dynamics, a process believed to actively remodel adhesion contacts and subsequent morphologic change enabling detachment during cell spreading (47). Invasive membrane protrusions or invadopodia, characterized by polarized actin filaments with fast-growing “barbed” ends and slow-growing “pointed” ends, exhibit robust adhesion disassembly at the rear of the cell (28, 48). Efficiency of cell motility is dependent on the stability of FAs (27). In our study, we found large and elongated FAs, as well as deficient FA dynamics in ORAI2-knockdown cells. Because there have been many thorough studies regarding FA turnover in cancers and emerging evidence was found in activation of FAK and ERK signals that was implicated in FA turnover associated with cell migration (28, 29), we hypothesized that FAK and ERK activation was involved in ORAI2-induced cell migration. Previous reports suggested that the cross-talk between growth factor receptor and FAK was often upregulated in cancers and phosphorylation of FAK was a prerequisite for subsequent activation of downstream signaling (31, 49). Autophosphorylation of FAK on Tyr397 facilitates the binding of Src through their SH2 domains, and this FAK–Src complex further mediates the phosphorylation of paxillin (30, 31). Phosphorylation of paxillin on Tyr31 and Tyr118 sites is a significant process during FA disassembly as the expression of mutant paxillin, which lacks Tyr31 and Tyr118 could lead to a dramatic decrease in rate constant of FA disassembly (29). Accordingly, it was also reported that FAK^−/− and Src^−/− cells, in which tyrosine phosphorylation of FA-associated proteins was attenuated, displayed larger and more stable FAs than those formed in normal cells (50). Our results prominently revealed an increase in phosphorylation levels of both FAK (Tyr397) and paxillin (Tyr118) in ORAI2-transfected cells, and a decreased in ORAI2-silenced cells. On the basis of the literatures that suggested that phosphorylation of paxillin at Tyr118 by Src could engage the recruitment and constitutive binding of MAPK/ERK, and hence the activation of MAPK/ERK (31, 32, 51), we next explored the MAPK/ERK signals in our study. As expected, we found the stimulation of ORAI2 expression in cells significantly enhanced the MAPK/ERK activation. This was further confirmed by employing a FAK autophosphorylation inhibitor Y15. Therefore, we concluded that ORAI2 could induce phosphorylation of FAK and paxillin, resulting in FA turnover by enhancing the signal transduction between FAK and MAPK/ERK. In other words, we suggested a considerable cross-talk between adhesion- and growth factor–mediated signaling. Moreover, our research demonstrated that ORAI2-mediated SOC activity was an initiator of metastasis in gastric cancer by activating the FAK–paxillin–MAPK/ERK axis (Supplementary Fig. S7).

In the future, the field needs to advance beyond the bench side and move forward to address personalized therapeutic approaches for patients with gastric cancer, which could be done by identifying more specific and potent drug targets. Our findings have proposed that ORAI2 represents promising targets in gastric cancer management, as its expression and activity are positively correlated with tumorigenesis and metastasis in gastric cancer. Besides, the effective inhibition of ORAI2 activity by SKF96365 in our study has emphasized the potential for the pharmacologic modulation of SOC channels in gastric cancer. In reality, such strategies of applying SOC channel blockers in clinical trials are still far from ideal, mainly due to their relatively low specificity and potency. In addition, blocking of one specific type of SOC activity may not be enough, as it could be compensated by other SOC subtypes or other Ca²⁺ entry methods, such as TRPCs. For these reasons, to study a “pool” of Ca²⁺ homeostasis-related genes as a potential cancer driver of gastric cancer, rather than just looking at isolated genes, will provide us better intervention means with Ca²⁺ influx in cancer treatment (52). Nevertheless, intensive investigations of combined therapies and novel pharmacologic tools targeting CRAC, as well as understanding the specific combination of ORAI/STIM form SOC channels in gastric cancer are future direction of research (52). To put our findings into translational perspectives, the results may enhance patient selection for available target agents, such as PI3K/Akt inhibitors, as well as discovery of combined therapies for anti-metastasis along the ORAI2 signaling pathways. The use of FAK inhibitor in our study exhibited promising ability in attenuating ORAI2-mediated signal transduction, raising the possibility that FAK inhibitors and siRNAs may be valuable in gastric cancer treatment. Although, there were candidates of FAK inhibitors being developed over the past few years and the first phase I clinical trials have been very promising, many questions still remain to be answered and many challenges still lie ahead (53). What are the underlying mechanisms associated with FAK upregulation and which subset of patients will have better response to this therapy (53)? In the case of gastric cancer, the study of ORAI2 may have already provided some answers.

No disclosures were reported.

S. Wu: Data curation, software, formal analysis, investigation, visualization, methodology, writing-original draft, writing-review and editing. M. Chen: Data curation, software, formal analysis, validation, investigation, visualization, methodology, writing-review and editing. J. Huang: Investigation. F. Zhang: Resources, investigation. Z. Lv: Investigation. Y. Jia: Resources. Y.-Z. Cui: Resources, methodology. L.-Z. Sun: Resources, software, investigation. Y. Wang: Investigation. Y. Tang: Investigation. K. Verhoeft: Investigation. Y. Li: Resources, methodology. Y. Qin: Resources. X. Lin: Resources, writing-review and editing. X.-Y. Guan: Conceptualization, resources, data curation, supervision, funding acquisition, validation, visualization, methodology, project administration, writing-review and editing. K.-O. Lam: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, project administration, writing-review and editing.

This work was supported by grants from the Hong Kong Research Grant Council (RGC) grants including Collaborative Research Fund (C7065-18GF), Research Impact Fund (R4017-18), National Natural Science Foundation of China (82073182), the National Key R&D Program of China (2017YFC1309000), Shenzhen Science and Technology program (KQTD 2018041118502879). We also appreciate the Tessy Cheng Grove Donation for Gastric Cancer Research. X.-Y Guan is Sophie YM Chan Professor in Cancer Research.

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.