NCX1 interacts with TRPA1 to promote cell proliferation and tumor growth of colon cancer via disruption of calcium homeostasis.

Introduction
Colon cancer (CC) is prevalent globally and is one of the leading causes of cancer-related death, ranking among the top three in terms of incidence and mortality rates[1]. Once metastases emerge, CC is exceedingly difficult to cure[2]. Therefore, it is essential to clarify its pathophysiology and investigate promising treatment targets to prevent and treat CC. Cytosolic Ca2+ ([Ca2+]cyt), which can result from internal Ca2+ release or extracellular Ca2+ entry via Ca2+-permeable cation channels, is essential for regulating cellular physiological functions and pathological processes, such as proliferation, invasion, migration, immune responses, and drug resistance, etc[3], [4]. Previous studies have indicated that the aberrant [Ca2+]cyt signaling is involved in the progression of digestive cancers, including gastric cancer[5], pancreatic cancer[6], and CC[7]. Disruption of [Ca2+]cyt homeostasis is thought to be an essential driver of cancer onset and progression[8]. Ca2+ channels/transporters on the plasma membrane play an indispensable role in maintaining [Ca2+]cyt homeostasis, and their abnormal expression and function are closely related to digestive tumorigenesis and progression[9], [10]. Therefore, elucidating the molecular mechanisms of how [Ca2+]cyt homeostasis is altered in human cancers, including CC, is crucial, which is still largely unclear.
The Na+/Ca2+ exchanger 1 (NCX1) is a bidirectional transporter protein located in the cell membrane that operates in either Ca2+ exit mode (3Na+ entry and 1 Ca2+ exit) or Ca2+ entry mode (3Na+ exit and 1 Ca2+ entry), depending on the electrochemical gradient of the substrate ions and the membrane potential[11], [12]. Nowadays, the research on NCX1 has mainly focused on the brain[13], heart[14], and kidney[15], and the therapeutic effects of its modulators on the corresponding diseases demonstrated the tremendous therapeutic potential of this target. However, little is currently known about the roles of NCX1 in digestive health and disease. Previous studies showed that NCX1 is expressed in the mouse colon and regulates colonic motility[16]. Further evidence suggests a pathogenic role for NCX1 in hepatocellular carcinomas, breast carcinomas, melanomas, and glioblastomas[17]. Our previous studies have identified a pivotal role for NCX1 in pancreatic cancer[6] and gastric cancer[18], but its expression and function in CC have not been explored to date.
The transient receptor potential ankyrin 1 (TRPA1) channel, as the only member of the TRPA subfamily, is a conserved non-selective Ca2+ permeable channel and is one of the most studied channels in the TRP family[19]. TRPA1 is ubiquitously expressed in the gastrointestinal (GI) tract, including the stomach, duodenum, esophagus, and colon, etc[20]. TRPA1 has been predominantly investigated in the field of GI inflammation and pain [21], [22]. Several studies have indicated that TRPA1 promotes cancer cell survival in diverse tumors by enhancing the antioxidant capacity of cancer cells or upregulating Ca2+-dependent anti-apoptotic pathways[23], [24], [25]. However, very little is known about the role of TRPA1 in the pathogenesis of CC except one recent report that TRPA1 is significantly overexpressed in primary cultures of metastatic CC cells to enhance reactive oxygen species-induced Ca2+ entry into the mitochondria, resulting in its depolarization and caspase-3/7 activated apoptosis[26], which would prevent metastasis of CC cells.
However, this elegant study focusing on cell Ca2+ signaling raises an interesting question: if the expression and function of TRPA1 channels are enhanced in CC cells, leading to significant mitochondrial apoptosis, why does CC still occur? This question prompts us to seek a reasonable answer, as in the real world, CC cells would not express only TRPA1. Moreover, it is currently unclear whether NCX1 or TRPA1 alone or their coupling is critically involved in CC progression. Therefore, in the present study, first, we sought to investigate the pathogenic role of either NCX1 or TRPA1 in CC under in vitro and in vivo conditions. Second, we sought to study whether NCX1 and TRPA1 are simultaneously involved in CC progression. We demonstrate that NCX1/TRPA1 coupling in CC cells prevents excessive increases in [Ca2+]cyt, maintaining a moderate level that promotes cell proliferation and CC growth. This mechanism may explain why, despite TRPA1 upregulation inducing apoptosis, CC persists under real-world conditions.

Material and methods

Animals and ethics statement
The Institutional Animal Care and Use Committee of the Medical College of Qingdao University (QDU-AEC-2024764) granted approval for all animal experiments. All animal care and experimental procedures complied with the NIH Guide for the Care and Use of Laboratory Animals, and animal conduct conformed to the ARRIVE guidelines[27].

Cell culture and reagents
The human normal colonic epithelial cell line (HCoEpiC) and human colon cancer cell lines (Caco-2, SW620, and DLD-1) were acquired from the Chinese Academy of Sciences (Shanghai, China). All cell lines were cultivated in MEM or DMEM-high glucose media (Gibco, USA) with 1 % penicillin/streptomycin (Invitrogen, USA) and 10 % fetal bovine serum (FBS, Gibco, USA) added in a 5 % CO2 incubator at 37 °C. The reagents were presented in Supplementary Table 1.

Preparation and infection of lentiviruses
Lentiviral-based shRNA for TRPA1 and NCX1 were obtained from GenePharma Co., Ltd. (Shanghai, China), and TRPA1-overexpressed lentivirus was provided by OBiO Technology Co., Ltd. (Shanghai, China). All shRNA sequences were shown in Supplementary Table 2 Lentivirus for shTRPA1 was used to silence the TRPA1 expression in SW620 and DLD-1 cells, and lentivirus for shNCX1 was used to silence the NCX1 expression in DLD-1 cells. A lentivirus with the full-length CDS of TRPA1 was designed to enhance its expression in Caco-2 cells. All shRNA groups of TRPA1 and NCX1 that do not have a designed number used shTRPA1-1 and shNCX1-1, respectively. Lentiviruses (MOI 20) were used to infect cells as directed by the manufacturer. In short, 6-well plates were seeded with 1 × 105 cells per well, the lentiviruses were then added to the culture medium, and the medium was replenished after 24 h. Following 72 h of lentivirus infection, the stable cells were screened using puromycin.

Real-time RT-PCR
RT-qPCR was conducted following the protocols outlined in previous studies[18]. β-actin served as the internal standard. 2-ΔΔCt relative quantification was used to quantify the data. The primers were acquired from Sangon Biotech Co., Ltd. (Shanghai, China), as shown in Supplementary Table 2.

Immunofluorescence assay
Immunofluorescence assay was carried out as previously described[18]. After blocked, the cells were incubated with rabbit anti-TRPA1 (1:100; Cat. No. ACC-037, alomone labs, Israel) and murine anti-NCX1 (1:100; Cat. No. MA3-926, Invitrogen, USA) overnight at 4 °C. After being washed with phosphate-buffered saline (PBS) for three times, the cells were cultured with Alexa Fluor 555-labled anti-rabbit (Cat. No. P0179, Beyotime, China) secondary antibody and Alexa Fluor 488-labled anti-mouse (Cat. No. P0188, Beyotime, China) secondary antibody for 1 h. Finally, nuclei were stained with DAPI for 5 min, and images were obtained using a laser scanning confocal microscope (NIKON A1MP, Japan).

Western blot analysis
Western blotting was conducted in accordance with our previous studies[18], [28]. The primary antibodies: anti-TRPA1 (1:250; Cat. No. ACC-037, Alomone Labs, Israel), anti-NCX1 (1:250; Cat. No. ANX-011, alomone labs, Israel), anti-ERK1/2 (1:1000; Cat. No. 9102, Cell Signaling Technology (CST), USA), anti-phospho-ERK1/2 (1:1000; Cat. No. 9101, CST, USA), anti-cyclin D1 (1:1000; Cat. No. MA5-16356, Thermo Fisher Scientific, USA), anti-β-catenin (1:1000; Cat. No. 8480, CST, USA), anti-phospho-β-catenin (1:1000; Cat. No. 9567, CST, USA), anti-PCNA (1:1000; Cat. No. 13110, CST, USA), anti-PI3K (1:2000; Cat. No. ABS1856, Merck, Germany), anti-AKT (1:1000; Cat. No. 4691S, CST, USA), anti-phospho-AKT (1:1000; Cat. No. 4060S, CST, USA), anti-mTOR (1:1000; Cat. No. 2983S, CST, USA), anti-phospho-mTOR (1:1000; Cat. No. 2971S, CST, USA), anti-NF-κB (1:1000; Cat. No. 8242S, CST, USA), anti-phospho-NF-κB (1:1000; Cat. No. 3033S, CST, USA), and anti-GAPDH (1:1000; Cat. No. 3683, CST, USA), the HRP-conjugated secondary antibody (Cat. No. ZB-2301, ZSGB Biotechnology, China).

Co-immunoprecipitation (Co-IP) and immunohistochemistry (IHC)
Co-IP and IHC staining were carried out as previously described[18], [29], [30]. The samples were incubated with specific primary antibodies: Anti-TRPA1 (1:100; Cat. No. ACC-037, alomone laboratories, Israel) and anti-NCX1 (1:100; Cat. No. ANX-011, alomone labs, Israel) for Co-IP; Anti-TRPA1 (1:100; Cat. No. ACC-037, alomone laboratories, Israel), anti-NCX1 (1:100; Cat. No. ANX-011, alomone labs, Israel), and anti-Ki67 (1:1000; Cat. No. GB151499-100, Servicebio, China) for IHC. Proteins immunoprecipitated were analyzed by western blotting, and images of all tissue nuclei were obtained using a panoramic tissue cell scanner (Panoramic MIDI, 3DHISTECH Ltd., Hungary). Image J evaluated the optical density.

Cell proliferation and colony formation assay
Cell proliferation was assessed using a CCK8 assay kit (Cat. No. HY-K0301, MedChemExpress, USA). 5000 cells/well was inoculated in 96-well plates; when required, the cells were subsequently exposed to varying concentrations of drugs. After 48 h, 100 μL of 10 % CCK8 regent was added for 1–2 h. The optical density was measured at 450 nm using the Multiskan FC Enzyme Labeler (ThermoFisher Scientific, USA). Colony formation was used to evaluate the long-term survival of cells. Cells (500 cells/well) were inoculated into 6-well plates for 10–12 days. After being washed and fixed, the cells were stained with crystal violet. Image J was utilized to quantify the clone counts.

Cell cycle analysis
Cells were collected, gently resuspended in pre-cooled 75 % ethanol, and subsequently stored at −20°C overnight. After centrifugation and ethanol elimination, the cells were hydrated in PBS for 15 min at room temperature. Then they stained with DNA staining solution (Cat. No. CCS012, MultiSciences, China) at 37 °C for 30 min away from light. Cell cycle analysis employed flow cytometry, which gated samples on living cells with an excitation wavelength of 488 nm and an emission wavelength of 620 nm. The FCS files were subsequently analyzed using FlowJo v10.8.1 software.

Intracellular calcium measurement
Calcium imaging was performed as previously described[18], [28]. Data were analyzed using GraphPad Prism 8.0.

Intracellular sodium determination
Intracellular Na+ was also determined by imaging devices that detected Fura-2 fluorescence changes using Sodium Green™ (Cat. No. S6901, Invitrogen, USA) or SBFI, AM (Cat. No. S1264, Invitrogen, USA) (Sodium Green™, 50 min; SBFI, AM, 2 h). Sodium Green™ was excited at 488 nm, and the fluorescence intensity under 488 nm excitation was recorded by F/F0; the ratio of SBFI fluorescence with excitation at 340 nm or 380 nm (F340/380) was recorded as for Fura-2.

Tumor xenograft in nude mice
Tumor xenograft assay was performed as previously described[18], [28]. Ten four-week-old male nude mice were obtained from Beijing Huafukang Biological Science and Technology Co Ltd (Beijing, China) and randomly divided into two groups. 1 × 106 of the shTRPA1 SW620 cells were injected into the right axilla and negative control (NC) cells into the left axilla of nude mice. Similarly, shNCX1 DLD-1 cells and their respective NC cells were separately injected into the right and left axilla of the nude mice. After thirty days, the mice were executed, and the xenografted tumors were quantified by calculating the tumor volume (mm3) according to the formula 1/2 (length × width2) and weighing the tumors.

Statistical analysis
P < 0.05 was considered statistically significant. Data were presented as means ± SEM and were analyzed using GraphPad Prism 8.0 with one-way ANOVA, the Student-Newman-Keul post hoc test, or Student's t-tests for paired or unpaired samples. All in vitro experiments were repeated using three biological replicates, and the data were obtained at least in triplicate.

Results

Enhanced NCX1 and TRPA1 expression and co-localization in human colon cancer cells
We first compared the expression levels of NCX1 or TRPA1 between three human CC cells (Caco-2, SW620 and DLD-1 cells) and normal human colonic epithelial cells (HCoEpiC). As illustrated in Fig. 1A and B, the expression levels of NCX1 were significantly elevated in SW620 and DLD-1 cells but not Caco-2 cells compared to those in HCoEpiC cells. Similar results were observed for TRPA1 expression levels as depicted in Fig. 1C and D, indicating similar expression trends of NCX1 and TRPA1 in these cells. Moreover, the expressions of both NCX1 and TRPA1 were found consistently at the transcripts and protein levels in these cells. Second, we performed immunofluorescence analysis to observe the expression and localization of NCX1 and TRPA1 proteins in human CC cells. Both NCX1 and TRPA1 proteins were predominantly expressed and co-localized on the plasma membrane of SW620 and DLD-1 cells (Fig. 1E). Third, Co-IP study clearly demonstrates a physical co-localization of NCX1 and TRPA1 in these two distinct human CC cells (Fig. 1F-I). Thus, NCX1 and TRPA1 were parallelly upregulated, co-localized and bound on the plasma membrane of human CC cells.
TRPA1 activation promotes proliferation of human CC cells in vitro.
Since our data are consistent with a previous study showing higher expression of TRPA1 in metastatic CC cells[26], suggesting its potential pro-oncogenic role, we assessed the functional role of TRPA1 in human CC cells, which has not been reported previously. We employed lentiviral infection to obtain TRPA1-knockdown in SW620 and DLD-1 cells as well as Caco-2 cells overexpressing TRPA1. Representative images and summary data demonstrating successful knockdown and overexpression of TRPA1 in these cells at both mRNA and protein levels were presented in Supplementary Fig. S1A-F; the proliferation marker PCNA was increased in TRPA1-overexpressed Caco-2 cells but decreased in TRPA1-knockdown SW620 and DLD-1 cells compared to their respective NC cells. However, cell proliferation and clonogenicity were significantly decreased in TRPA1-knockdown SW620 and DLD-1 cells (Fig. 2A-D), while they were increased in TRPA1-overexpressed Caco-2 cells compared to NC (Fig. 2G and H). We further utilized ASP7663, a selective agonist of TRPA1[31], to activate TRPA1 channels. ASP7663-induced cell proliferation was prevented by TRPA1 knockdown (Fig. 2E and F). Again, ASP7663 dose-dependently increased cell proliferation of SW620 and DLD-1 cells (Supplementary Fig. S2A, B), which was significantly attenuated by HC-030031, a selective blocker of TRPA1 channels (Supplementary Fig. S2D, E). Likewise, HC-030031 diminished ASP7663-induced clonogenicity (Supplementary Fig. S2G, H). Notably, ASP7663 did not affect cell proliferation in normal HCoEpiC cells with low expression of TRPA1 (Supplementary Fig. S2C, F, I). These pharmacological data suggest that TRPA1 exerts a pro-proliferative role in human CC cells.
Given that TRPA1 is a Ca2+-permeable channel, we selected BAPTA-AM, a calcium chelator capable of penetrating cell membranes, to chelate [Ca2+]cyt. The results showed that BAPTA-AM significantly attenuated ASP7663-induced cell proliferation in both SW620 and DLD-1 cells (Fig. 2I, J). These genetic results further support our notion that Ca2+ entry via TRPA1 channels enhances the proliferation of human CC cells.

NCX1 coordinates with TRPA1 to promote the proliferation of human CC cells
Given that NCX1 was also highly expressed in CC cells, we subsequently explored its pathogenic role in CC. To this end, we applied lentiviral infection to successfully generate NCX1-knockdown DLD-1 cells as shown in Supplementary Fig. S1G and H; the proliferation marker PCNA was decreased in NCX1-knockdown DLD-1 cells compared with their corresponding NC. NCX1 knockdown significantly decreased the proliferation of this human CC cell (Fig. 2K). To test if NCX1 knockdown could upregulate other Ca2+ transporters (e.g., SERCA, PMCA) to make compensatory changes, we assessed NCX1 knockdown on the expressions of SERCA and PMCA. Our data showed that NCX1 knockdown did not upregulate SERCA and PMCA (Supplementary Fig. S3).
As yet, there are no selective agonists available for NCX1; therefore, we proceeded to investigate the effect of NCX1 by using its antagonists. We chose three different kinds of NCX1 antagonists according to the literature[32], [33]: 1) SN 6, a relatively selective antagonist targeting Ca2+ entry mode of NCX1; 2) bepridil, a relatively selective antagonist targeting Ca2+ exit mode; and 3) KB-R7943, an antagonist affecting both Ca2+ exit and entry modes. Our data showed that the proliferation of both DLD-1 cells (Fig. 2L-N) and SW620 cells (Supplementary Fig. S4A-C) was similarly inhibited in a dose-dependent manner by bepridil and KB-R7943, while SN 6 did not impact cell proliferation. These findings suggest that Ca2+ exit mode of NCX1 may play a crucial role in the proliferation of human CC cells.
Moreover, to study the coupling of NCX1 and TRPA1 within human CC cells, we combined TRPA1 activator ASP7663 and NCX1 inhibitors and shRNA. As shown in Fig. 2O-R, ASP7663 dose-dependently promoted CC cell proliferation. Moreover, bepridil and shNCX1 significantly attenuated ASP7663-promoted cell proliferation, whereas SN 6 did not influence the proliferation. These results suggest that TRPA1 couples to the Ca2+ exit mode of NCX1 modulate the proliferation of human CC cells.

TRPA1 activation enhances colon tumor growth by promoting cell cycle progression
Since the pathogenic role of TRPA1 is currently obscure in CC, we further applied the xenografted CC model of nude mice to validate its role in colon tumor growth. The knockdown of TRPA1 in SW620 cells via shRNA lentiviruses resulted in about 60 % reduction in tumor weight and about 55 % decrease in tumor volume compared to NC group (Fig. 3A, C). IHC analysis revealed that the expression levels of proliferation marker Ki67 were significantly lower in tumors derived from implants pretreated with TRPA1-shRNA lentivirus than those from NC group (Fig. 3B). These results indicate that TRPA1 promotes colon tumor growth in vivo, consistently with its pro-proliferative action on CC cells in vitro described above.
We further elucidated the underlying mechanisms of TRPA1 in pro-proliferative action. Given that cell cycle is an important biological process that regulates cell proliferation, we investigated how TRPA1 influences CC cell cycle by flow cytometry analysis. As illustrated in Fig. 4D, the number of TRPA1 knockdown SW620 cells was increased by around 30–40 % in G1 phase, but was decreased by around 30–40 % in S phase compared to NC cells. Similarly, cells treated with HC-030031 exhibited effects comparable to the cells pre-treated with TRPA1-shRNA lentivirus in terms of cell cycle progression (Supplementary Fig. S2J). Therefore, TRPA1 knockdown impedes G1/S transition to inhibit proliferation of human CC cells in vitro and colon tumor growth in vivo. These data strongly suggest an oncogenic role of TRPA1 channels in the pathogenesis of CC.

NCX1 knockdown retards colon tumor growth by blocking cell cycle at G1 phase
After demonstrating the pro-proliferative role of NCX1 in human CC cells, we further confirmed its involvement in tumor growth in the xenografted CC mode. In this model, NCX1 knockdown (shNCX1) in DLD-1 cells significantly suppressed tumor weights by approximately 29 % and tumor volume by approximately 38 % compared with the NC group, along with a decrease in the expression of Ki67 (Fig. 3E-G), indicating that NCX1 plays a pro-proliferative role in vitro and oncogenic role in vivo in CC.
We further elucidated NCX1-elicited oncogenic mechanisms. Our data revealed that shNCX1 and bepridil led to significant G1 phase arrest of CC cells; however, SN 6 did not affect cell cycle (Fig. 4H-J). Meanwhile, to test whether [Ca2+]cyt contributed to bepridil-induced cell cycle arrest of DLD-1 cells, we used BAPTA-AM to chelate [Ca2+]cyt and found it indeed restored bepridil-elicited G1 phase arrest (Fig. 4J). Therefore, mechanistically, NCX1 operates in Ca2+ exit mode to modulate cell cycle of human CC cells.

Enhanced NCX1 operates in Ca2+exit mode in human CC cells
We applied single-cell Ca2+ imaging to ascertain if NCX1 operates in Ca2+ exit mode to modulate [Ca2+]cyt homeostasis of human colon cells. First, after comparing the basal levels of [Ca2+]cyt among HCoEpiC, SW620 and DLD-1 cells, we found lower basal [Ca2+]cyt in CC cells compared to normal HCoEpiC cells (Fig. 4A), implying the likelihood of enhanced NCX1 operation in Ca2+ exit mode in CC cells.
Second, bepridil induced a marked increase in [Ca2+]cyt in DLD-1 and SW620 cells, which was attenuated by NCX1 knockdown (shNCX1); however, bepridil did not alter [Ca2+]cyt in HCoEpiC (Fig. 4A, B). Moreover, SN 6 did not elicit any increase in [Ca2+]cyt in both DLD-1 and SW620 cells (Fig. 4C, D). Similarly, extracellular 0Na+ that would stimulate NCX1 operation in Ca2+ entry mode failed to trigger a rise in [Ca2+]cyt (Fig. 4E). These data further support the enhancement of NCX1 operation in Ca2+ exit mode rather than in Ca2+ entry mode in human CC cells.
Third, since NCX1 regulates not only [Ca2+]cyt but also [Na+]cyt, we conducted cell Na+ imaging experiments using Na+ indicators, SBFI and Sodium Green™ tetraacetate. We observed the elevated basal [Na+]cyt in DLD-1 and SW620 cells (Fig. 4F), which is opposite to basal [Ca2+]cyt in these cells (Fig. 4A). Moreover, bepridil reduced greater [Na+]cyt in DLD-1 and SW620 cells than that in normal HCoEpiC (Fig. 4F, G); this reduction was restored by NCX1 knockdown (shNCX1) (Fig. 4H). Notably, SN 6 did not impact either [Ca2+]cyt or [Na+]cyt (Fig. 4C-D, 4I). Collectively, these data confirm that the enhanced NCX1 operates in Ca2+ exit mode rather than in Ca2+ entry mode in human CC cells.

NCX1 couples with TRPA1 to regulate [Ca2+]cytin human CC cells
Since [Ca2+]cyt rise can derive from either extracellular Ca2+ entry or internal Ca2+ release, we first investigated the source of TRPA1-evoked Ca2+ signaling in human CC cells. ASP7663 enhanced [Ca2+]cyt markedly in SW620 cells but slightly in HCoEpiC (Fig. 5A). TRPA1-evoked Ca2+ signaling in CC cells was abolished by HC-030031, a selective blocker of TRPA1 channels blocker and by TRPA1 knockdown (shTRPA1) (Fig. 5B-C). In the absence of extracellular Ca2+ (0 Ca2+), ASP7663 failed to raise [Ca2+]cyt, whereas reintroduction of Ca2+ (2 mM) immediately elevated [Ca2+]cyt in SW620 cells (Fig. 5D), indicating that activation of TRPA1 channels primarily facilitates Ca2+ entry from external source of CC cells.
We further explored how NCX1/TRPA1 coupling mediates Ca2+ signaling in human CC cells, given that NCX1 operates in Ca2+ exit mode while TRPA1 facilitates extracellular Ca2+ entry in human CC cells. As shown in Fig. 5E, ASP7663-induced [Ca2+]cyt signaling exhibited an initially transient peak (1st phase) followed by a slow decline to the stabilization (2nd phase) in DLD-1 cells, which was likely due to NCX1 operation in Ca2+ exit mode. However, pretreatment with bepridil did not alter the ASP7663-induced 1st phase of [Ca2+]cyt signaling but markedly potentiated the 2nd phase. Similar findings were observed with NCX1-knockdown DLD-1 cells (Fig. 5F), confirming NCX1/TRPA1 coupling in the regulation of [Ca2+]cyt homeostasis in human CC cells.
It is expected that [Ca2+]cyt level would be high in Caco-2 cells overexpressing TRPA1 if NCX1 could not balance it. However, our results clearly showed that Caco-2 cells overexpressing TRPA1 led to moderate rise rather than excessive rise in [Ca2+]cyt level (Supplementary Fig. S5A-C). Since it is well-known that moderate rise in [Ca2+]cyt causes cancer cell proliferation but excessive [Ca2+]cyt rise causes apoptosis, we applied cell proliferation assay by clamping extracellular Ca2+ in human CC cells via Ca2+ chelater EGTA[34]. Indeed, Ca2+ below 4 mM dose-dependently promoted cell proliferation, whereas Ca2+ exceeding 4 mM exerted apoptosis (Supplementary Fig. S6A and B). Taken together, these results strongly indicate that NCX1 reduces [Ca2+]cyt raised by TRPA1 to a moderate [Ca2+]cyt level, finally promoting cancer cell proliferation rather than inducing apoptosis under real-word conditions.

TRPA1 activation promotes proliferation of CC via ERK1/2 and β-catenin pathways
We subsequently elucidated TRPA1/Ca2+-promoted oncogenic mechanisms. Since ERK1/2 and β-catenin pathways play important roles in CC development[35], [36], and aberrant [Ca2+]cyt affects CC progression via these pathways[7]. First, we observed the phosphorylation of either ERK1/2 or β-catenin was significantly increased in TRPA1-overexpressing Caco-2 cells (Fig. 6A and C). Second, ERK1/2 regulates cyclin D1 transcription[37], therefore, we examined cyclin D1 expression. Indeed, TRPA1-overexpression enhanced the expression of cyclin D1 (Fig. 6B). Conversely, ASP7663-induced phosphorylation of ERK1/2 and β-catenin, along with cyclin D1 expression, was attenuated by TRPA1 knockdown (Fig. 6D-F). Third, to further validate the pivotal roles of ERK1/2 and β-catenin in ASP7663-promoted proliferation of CC cells, we employed ERK1/2 inhibitor SCH772984 and β-catenin inhibitor XAV-939, both of which effectively suppressed ASP7663-induced cell proliferation (Fig. 6G-J). Finally, we ruled out the involvements of other signaling pathways since PI3K, phosphorylation of AKT, mTOR and NF-κB was not changed in TRPA1-overexpressing Caco-2 cells (Supplementary Fig. 7A-D). Thus, TRPA1 activation could promote cell proliferation particularly via phosphorylation of both ERK1/2 and β-catenin in human CC cells.

NCX1 couples with TRPA1 to promote CC via ERK1/2 and β-catenin pathways
Since NCX1 operates in Ca2+ exit mode to couple with TRPA1 channels in human CC cells, we further investigated whether NCX1 is connected to TRPA1-mediated phosphorylation of ERK1/2 and β-catenin. Western blot analysis showed that ASP7663-induced phosphorylation of both ERK1/2 and β-catenin in CC cells was significantly reduced by either bepridil or NCX1 knockdown (Fig. 7A, C, D, F). Consistently, ASP7663-promoted cyclin D1 expression was also significantly attenuated by bepridil and NCX1 knockdown (shNCX1) (Fig. 7B and E). These data indicate that NCX1/TRPA1 coupling enhances cyclin D1 expression via phosphorylation of ERK1/2 and β-catenin in human CC cells.

Discussion
In the present study, we present the first evidence that NCX1 and TRPA1 channels simultaneously contribute to the development and progression of CC. Multiple lines of evidence indicate that NCX1 promotes CC growth through a novel coupling with TRPA1 channels in human CC cells. First, expression levels of both NCX1 and TRPA1 are in parallel enhanced in most of tested CC cells. Second, NCX1 and TRPA1 proteins are co-expressed, co-localized and bound on the plasma membrane of human CC cells. Third, downregulation of either NCX1 or TRPA1 using pharmacological blockers or genetic shRNA lentiviruses significantly attenuated cell proliferation in vitro and reduced tumor size and cell proliferation in vivo. Fourth, upregulation of NCX1/TRPA1 coupling promoted CC cell proliferation; conversely, downregulation of NCX1 by either bepridil or shRNA lentiviruses suppressed cell proliferation. Fifth, NCX1 operates in Ca2+ exit mode to expel the entered Ca2+ via TRPA1 channels in human CC cells. Sixth, NCX1 couples to TRPA1 channels to promote CC via ERK1/2 and β-catenin pathways.
NCX1 plays a key role in maintaining [Ca2+]cyt homeostasis by operating in either Ca2+ exit mode or Ca2+ entry mode in various types of mammalian cells[11], and its physiological and pathological significances have been extensively studied on cardiovascular[38], nervous[39], urinary[40] and endocrine[41] systems. However, its involvement in tumorigenesis is largely unknown. Our lab has been focusing on the tumorigenesis role of NCX1 in the digestive system, including esophageal[42], pancreatic[6], hepatocellular[43], and gastric cancers[18]. Nowadays, the functional role of NCX1 in CC has not been explored, although its expression was closely correlated to human CC prognosis[44].
In the present study, we demonstrated that both expression and function of NCX1 are enhanced in human CC cells. Interestingly, NCX1 usually functions in Ca2+ entry mode to raise [Ca2+]cyt in most types of digestive cancer except CC described here. Moreover, highly expressed NCX1 operates in Ca2+ exit mode rather than Ca2+ entry mode, maintaining relatively lower [Ca2+]cyt but higher [Na+]cyt in CC cells than in normal colon cells. Importantly, bepridil, a selective blocker of NCX1 in Ca2+ exit mode and NCX1 knockdown, induced G1 phase arrest, which was attenuated by the Ca2+ chelator BAPTA-AM. In addition, upregulated NCX1 expression is correlated positively with enhanced proliferation of human CC cells in vitro and in vivo and larger colon tumor size. These data strongly suggest that NCX1 in Ca2+ exit mode clearly exerts an oncogenic role in CC and can be a potential preventive/therapeutic target for CC.
Previous studies showed that TRPA1 can upregulate Ca2+-dependent anti-apoptotic and pro-survival signaling pathways, thereby inducing oxidative stress tolerance in cancer, especially in TRPA1-enriched breast and lung cancer spheroids[25], [45]. Recently, it was reported that TRPA1 is highly expressed in primary cultures of human metastatic CC cells to induce mitochondrial Ca2+ overload and caspase-3/7-activated apoptosis[26]. However, that elegant study raises an interesting paradox: why does enhanced TRPA1 expression induce mitochondrial Ca2+-dependent apoptosis, which is supposed to cause anti-CC metastasis but pro-CC metastasis? In addition, the previous study mostly focused on cell Ca2+ signaling via TRPA1 channels rather than their pathogenic roles[26]. Therefore, it is necessary to assess the pathogenic roles of TRPA1 channels and to explain the above-described paradox by using human CC cells and animal models.
In the real world, because of the complexity of cell functions, it is impossible for one single channel to work alone. Therefore, we hypothesize NCX1/TRPA1 coupling maintains a moderate [Ca2+]cyt rise in CC cells, ultimately promoting CC cell proliferation and tumor growth. Although TRP channels are usually potential partners of NCX1 in digestive cancers[18], [46], [47], the association of NCX1 with TRPA1 remains unknown in human CC cells. We have provided strong evidence to support NCX1/TRPA1 coupling: 1) parallel upregulated NCX1 and TRPA1 are co-expressed, co-localized and bound on the plasma membrane of human CC cells; 2) NCX1 operates in Ca2+ exit mode to expel the entered Ca2+ via TRPA1 channels; 3) downregulation of either NCX1 or TRPA1 significantly attenuated cell proliferation in vitro and reduced tumor size as well as cell proliferation in vivo; conversely, upregulation promoted CC cell proliferation.
It is acknowledged that [Ca2+]cyt signaling, as a double-edged sword, regulates various cellular functions, such as cell proliferation, apoptosis, and migration[3], [4]. Our findings indicate that NCX1/TRPA1 coupling promotes cell proliferation rather than mitochondrial apoptotic cell death in CC. This aligns with previous reports that [Ca2+]cyt oscillations in cancer cells have long been known to recruit Ca2+-dependent effectors that promote cell proliferation[48], [49], while long-lasting [Ca2+]cyt increase to a certain level can lead to apoptotic cell death[50], [51]. Enhanced TRPA1-induced excessive increase in [Ca2+]cyt in human metastatic CC cells may cause mitochondrial Ca2+-dependent apoptosis[26]; however, if the increase in [Ca2+]cyt via TRPA1 channels is simultaneously expelled by NCX1 operation in Ca2+ exit mode to allow for a moderate elevation in [Ca2+]cyt, which would stimulate cell proliferation and survival, such in the present case. Therefore, we have not only revealed the oncogenic role of TRPA1 channels in CC but also explained the above-described paradox of TRPA1 channels in human metastatic CC cells[26].
It is well known that alterations in [Ca2+]cyt can affect downstream signaling pathways[7], [52]. Our results reveal that the NCX1/TRPA1 coupling induces phosphorylation of ERK1/2 and β-catenin in CC cells. This further corroborates previous reports highlighting the critical roles of ERK1/2[53] and β-catenin[54] in CC development. In terms of the present study, mechanistically, NCX1/TRPA1 coupling-mediated moderate [Ca2+]cyt signaling may exert an oncogenic role by inducing phosphorylation of ERK1/2 and β-catenin in CC cells, such as in some other digestive cancer cells reported previously[55]. Alternatively, it is due to [Na+]cyt caused by NCX1 operation in Ca2+ exit mode in human CC cells. Elevated [Na+]cyt has also been demonstrated in several studies in various tumors such as breast[56], brain[57], and prostate[58], [59]. Similarly, [Na+]cyt may exert an oncogenic role by inducing phosphorylation of ERK1/2[60], which needs further investigation in CC cells.
The present study has several limitations. First, since only a subcutaneous tumor model was used, our findings require further validation in more advanced CC models or genetically engineered mouse models. Second, potential off-target effects in the shRNA knockdown experiments should be taken into account. Third, further validation in human CC samples and upcoming clinical studies is required.

Conclusion
In summary, we demonstrated for the first time that NCX1/TRPA1 coupling promotes CC development. Mechanistically, excessive Ca2+ entry via TRPA1 channels is simultaneously expelled by NCX1 in Ca2+ exit mode to maintain a moderate [Ca2+]cyt rise to promote CC by inducing phosphorylation of ERK1/2 and β-catenin (Fig. 7G). Our mechanistic study may explain why, despite TRPA1 upregulation in CC cells inducing apoptosis, CC persists under real-world conditions. Although research on NCX and TRP channels remains a relatively recent subject in oncology studies, and most investigations are still in their infancy, they have great potential for further discovery and promise clinical breakthroughs in tumor therapy. Importantly, given the critical roles of NCX1 and TRPA1 in modulating both [Ca2+]cyt and [Na+]cyt in human CC cells, the innovative coupling of NCX1 with TRPA1 may offer new strategies for preventing and treating CC.

Author contributions
LYZ: Investigation, data curation, software, formal analysis, methodology, visualization and writing–original draft. GLZ: Investigation, data curation and validation. JHL: Supervision, software and validation. RHG: Data curation, formal analysis, project administration, visualization, writing–review and editing, and funding acquisition. HD: Resources, conceptualization, supervision, writing–review and editing, and funding acquisition. All authors edited and approved the manuscript.

Compliance with Ethics Requirement
All procedures were performed in compliance with relevant laws and institutional guidelines, and all animal experiments were approved by the Institutional Animal Care and Use Committee at the Medical College of Qingdao University (QDU-AEC-2024764). All animal care and experimental procedures were in accordance with the NIH Guide for the Care and Use of Laboratory Animals. Animal conduct conforms to the ARRIVE guidelines.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.