Enhanced NCX1 expression in human primary GC tissues
Due to the lack of information on NCX1 expression in the stomach of normal subjects and GC patients, we first collected human primary GC tissues and corresponding adjacent tissues to compare NCX1 expression. By applying Western blotting analysis, total 55 pairs of fresh gastric tissues obtained from GC patients were compared. As shown in Fig. 1, 36 pairs had higher NCX1 protein expression in human GC tissues than in adjacent tissues (Fig. 1A), accounting for 65% (Fig. 1D). In contrast, 13 pairs had lower NCX1 expression in GC tissues (Fig. 1B), accounting for 24% (Fig. 1D). However, 6 pairs had no difference (Fig. 1C), accounting for 11% (Fig. 1D). Therefore, NCX1 protein expression was enhanced in human primary GC tissues.
Second, immunohistochemstry study was applied to human gastric tissues from 80 GC patients. Among these patients, their average age was 64 years old, 76% was male, 56% was diagnosed with advanced-stage (III/IV), and 73% had lymphatic metastasis (Supplementary table 1). As shown in Fig. 1E and F, the protein expression of NCX1 was markedly enhanced in GC tissues compared to their adjacent tissues, but staining was not detected in the negative control, indicating its specific staining to NCX1 proteins. Third, the association between NCX1 expression and clinicopathologic parameters of GC progression was subsequently analyzed. As shown in Fig. 1G-I, the up-regulation of NCX1 expression was correlated with the advanced clinical stage, the large tumor size, lymphatic metastasis. Furthermore, Kaplan-Meier analysis showed that the GC patients with high NCX1 expression had a poor prognosis, but those with low expression had a better prognosis (Fig. 1J). Altogether, the close association between NCX1 expression and clinicopathologic parameters strongly suggests an oncogenic role for NCX1 in human GC.
Co-localization and binding of the enhanced NCX1 and TRPC1 in human GC cells
Since enhanced expression of TRPC1 was closely related to GC prognosis and it exacerbated EMT in GC [24, 25], we first compared the expression of both TRPC1 and NCX1 proteins between normal human gastric epithelial cells (GES1) and five human GC cell lines (MKN45, SGC7901, AGS, BGC823, and SNU216). As shown in Fig. 1K, the expression level of NCX1 proteins was markedly enhanced in all GC cells compared to the GES1 cells. Similarly, the expression level of TRPC1 proteins was also markedly enhanced in all GC cells (Fig. 1L), suggesting both NCX1 and TRPC1 are expressed parallelly. Second, we performed immunofluorescence analysis to further study the expression and localization of NCX1 and TRPC1 proteins in human GC cells. As shown in Fig. 1M, both NCX1 and TRPC1 proteins were confirmed to express parallelly in three GC cells, but non-specific staining was undetected in the negative control without primary antibody. Moreover, both NCX1 and TRPC1 proteins were predominately expressed and co-localized on the plasma membrane of GC cells (Fig. 1M). Finally, our coimmunoprecipitation study clearly showed the binding of NCX1 and TRPC1 in GC cells (Fig. 1N-Q). Therefore, the expression of NCX1 and TRPC1 is not only up-regulated but also co-localized and bound on the plasma membrane of human GC cells.
NCX1 activation promotes proliferation, migration and invasion of human GC cells in vitro
To examine the role of NCX1 in GC, we first determined the cell proliferation of 3 human GC cell lines commonly used in the literature (MKN45, AGS and SGC7901). The varying concentrations of CaCl2 were applied to stimulate the Ca2+ entry mode of NCX1 since no selective activators of NCX1 are commercially available so far [26]. CaCl2 at the concentrations of 0.1-2 mM, dose-dependently promoted proliferation of all GC cells (Fig. 2A, D, G), which was attenuated by KB-R7943, a selective inhibitor for the Ca2+ entry mode of NCX1 (Fig. 2B, E, H). The concentrations of KB-R7943 were chosen in the light of the different sensitivity of GC cell proliferation to the drug (Supplementary Fig. 1A, B, C, D). Similarly, CaCl2 dose-dependently promoted proliferation of CHO cells with NCX1 overexpression (CHO-NCX1) (Fig. 2P), which was attenuated by KB-R7943 (Fig. 2Q). However, CaCl2 could not influence proliferation of CHO cells without NCX1 overexpression (CHO-K1) (Fig. 2R) and GES1 cells without NCX1 expression (Fig. 2T). Therefore, CaCl2 promotes GC cell proliferation most likely via activating the Ca2+ entry mode of NCX1.
Although NCX1 enhanced migration and invasion of hepatocellular carcinoma [19], its contribution to GC progression is unknown. Second, we examined the role of NCX1 in migration and invasion of human GC cells. Cell scratch test showed that CaCl2 promoted migration of MKN45 and AGS cells, which was attenuated by KB-R7943 (Supplementary Fig. 2A, B). Moreover, transwell assays showed that CaCl2 promoted migration (Supplementary Fig. 2C, D) and invasion (Fig. 2C, F, I) of MKN45, AGS and SGC7901 cells, which were attenuated by KB-R7943 (Supplementary Fig. 2C, D and Fig. 2C, F, I). Finally, after shNCX1 was applied to successfully knock down the protein expression of NCX1 in GC cells (Fig. 3A, B, C), CaCl2-induced cell proliferation (Fig. 3D, F, H), migration (Supplementary Fig. 3A, B, C) and invasion (Fig. 3E, G, I) were all inhibited. Taken together, NCX1 plays a critical role in GC cell proliferation, migration and invasion.
Hp virulence factor promotes GC cell proliferation, migration and invasion via NCX1 activation
Since H. pylori infection is a pivotal risk factor for tumorigenesis of GC and ammonia/ammonium is a major H. pylori virulence factor, NH4Cl was applied to the present study as a well-known ammonia/ammonium [2]. As shown in Fig. 2, like CaCl2, NH4Cl dose-dependently (0.1-2 mM) promotes proliferation of MKN45, AGS and SGC7901 cells (Fig. 2J, L, N), which was attenuated by KB-R7943 (Fig. 2K, M, O). However, NH4Cl did not affect proliferation of CHO-K1 (Fig. 2S) and GES1 cells (Fig. 2U). Similarly, NH4Cl-induced cell proliferation (Fig. 3J, L, N), migration (Supplementary Fig. 3D, E, F) and invasion (Fig. 3K, M, O) were all inhibited by knockdown of NCX1 protein expression in GC cells. Therefore, Hp virulence factor promotes GC cell proliferation, migration and invasion via NCX1 activation.
CaCl 2 , Hp and their virulence factors enhance NCX1 expression in GC cells
After demonstrating the promoted action of CaCl2 and Hp virulence factors on cell proliferation, migration and invasion, we also examined if they affect NCX1 expression of GC cells. Indeed, CaCl2 enhanced NCX1 expression in MKN45, AGS and SGC7901 cells (Fig. 4A, G, M), which was attenuated by either KB-R7943 (Fig. 4B, H, N) or shNCX1 (Fig. 4C, I, O). Moreover, H. pylori virulence factor NH4Cl enhanced NCX1 expression in GC cells (Fig. 4D, J, P), which were attenuated by either KB-R7943 (Fig. 4E, K, Q) or shNCX1 (Fig. 4F, L, R). Similarly, another H. pylori virulence factor lipopolysaccharide (LPS) enhanced NCX1 expression in GC cells (Fig. 4S, T, U). Finally, H. pylori per se co-culture with MKN45 and AGS cells for 24 h also enhanced NCX1 expression (Fig. 4V, W). However, CaCl2, NH4Cl, LPS and H. pylori all did not affect NCX1 expression in GES1 cells as negative control (Supplementary Fig. 4A, B, C, D). Taken together, these data strongly suggest that like CaCl2, H. pylori and their virulence factors promote GC through enhancing NCX1 expression.
NCX1 coordinates with TRPC1 to promote GC cell proliferation and migration
TRPC family is a potential partner for the Ca2+ entry mode of NCX1 [27]. Among them, TRPC1 is highly expressed in human GC to play an oncogenic role in GC progression [24]. We therefore focused on a possible coupling of TRPC1 and NCX1 in GC development. As shown in Fig. 5, CaCl2-induced proliferation and migration of MKN45, AGS and SGC7901 GC cells were attenuated by either KB-R7943 or a TRPC1 blocker SKF96365 (its concentrations were also chosen in the light of the different sensitivity of GC cell proliferation to the drug (Supplementary Fig. 1E, F, G)); however, CaCl2-induced cell proliferation and migration were further attenuated by a combination of the selective inhibitors for both NCX1 and TRPC1 (Fig. 5A-C, G-I). Similarly, CaCl2-induced cell proliferation and migration were further attenuated by a combination of shNCX1 plus SKF96365 (Fig. 5D-F, J-L). Taken together, these data suggest that NCX1 coordinate with TRPC1 to promote GC cell proliferation and migration.
Hp virulence factor could stimulate TRPC1 channels to trigger Ca2+ entry mode of NCX1 in GC cells
We next applied single cell live Ca2+ imaging to determine if NCX1 operates in Ca2+ entry mode to induce [Ca2+]cyt increase in GC cells. First, extracellular 0 Na+ that triggers Ca2+ entry mode of NCX1 could significantly induce [Ca2+]cyt signaling in Ca2+-containing solutions but not in Ca2+-free solutions (Fig. 6A). Second, 0 Na+-induced [Ca2+]cyt signaling in Ca2+-containing solutions could be abolished by KB-R7943 (Fig. 6B). Third, 0 Na+ also markedly increased [Ca2+]cyt signaling in CHO-NCX1 cells with NCX1 overexpression (Fig. 6G, I, J), but not in CHO-K1 without NCX1 overexpression (Fig. 6H, I, J). These data strongly support NCX1 operates in Ca2+ entry mode in GC cells like in CHO-NCX1 cells. We further examined if Hp virulence factor NH4Cl and the local acidic micro-environment in Hp infection-induced chronic inflammation and tumorigenesis could stimulate NCX1 activity. Like 0 Na+, NH4Cl and acid (pH4.5) indeed had similar stimulation on Ca2+ entry mode of NCX1 in SGC7901 cells (Fig. 6C, D, E, F).
We further examined TRPC1/NCX1 coupling-mediated Ca2+ signaling in GC cells since NCX1-induced Ca2+ entry requires the electrochemical gradient of the substrate ions and membrane potential, depending on Na+ entry via TRPC channels [27]. After shTRPC1 successfully knocked down TRPC1 expression in GC cells (Fig. 6N, R), both 0 Na+ and NH4Cl-induced Ca2+ signaling was almost abolished (Fig. 6K-M, O-Q). These data verify that Hp virulence factor could stimulate TRPC1/NCX1 coupling to induce Ca2+ signaling in GC cells.
NCX1 couples with TRPC1 to promote GC through AKT pathway
Since NCX1 often coupled with TRPC1 to function [27], we investigated whether TRPC1 channels are involved in NCX1-mediated AKT phosphorylation. Western blotting analysis exhibited that after CaCl2 induced AKT phosphorylation in MKN45 and SGC-7901 cells, either NCX1 inhibitor KB-R7943 or TRPC1 inhibitor SKF96365 significantly attenuated the CaCl2-induced AKT phosphorylation; however, a combination of them further attenuated CaCl2-induced AKT phosphorylation (Fig. 8A, D). Moreover, either a combination of shNCX1 and SKF96365 (Fig. 8B, E) or a combination of shTRPC1 and KB-R7943 (Fig. 8C, F) further attenuated the CaCl2-induced AKT phosphorylation. These data verify a TRPC1 and NCX1 coupling enhances AKT phosphorylation in GC cells.
NCX1 activation enhances GC growth and metastasis in vivo
We applied subcutaneously xenografted GC model of nude mice to further verify the oncogenic role of NCX1 in GC growth in vivo. NCX1 activation by CaCl2 increased tumor weights (Fig. 8G), which was attenuated by KB-R7943 (Fig. 8H). Moreover, the knockdown of NCX1 in SGC-7901 cells by NCX1-shRNA lentiviruses markedly suppressed growth ability of GC cells after their implantation, leading to a significant decrease in tumor weights (Fig. 8I). Immunohistochemical analysis showed that the tumors derived from the implants pre-treated with NCX1-shRNA lentiviruses had lower expression of NCX1 and Ki67 than those pre-treated with control shRNA (Fig. 8K, L, M). Therefore, NCX1 stimulated GC growth in vivo.
We further applied abdominal transplantation tumor model of nude mice to verify the role of NCX1 in promoting GC metastasis in vivo. As shown in Fig. 8J, CaCl2-induced GC cell metastasis was markedly suppressed by pretreatment with NCX1-shRNA lentiviruses. Compared to NC group, tumor numbers in the group pretreatment with NCX1-shRNA was decreased by about 50%. Therefore, NCX1 stimulated GC metastasis in vivo as well.