Emerging evidence implicates sodium dysregulation as a pivotal mechanism in sodium overload-induced necrosis, a distinct cell death modality characterized by cytotoxic Na⁺ accumulation and mitochondrial membrane collapse. Guided by prior mechanistic studies delineating ion transport pathophysiology, we prioritized seven evolutionarily conserved sodium regulators—NC1, CLT, DHPs, SLC12A2 (NKCC1), SLC8A1 (NCX1), TRPM4, and SLC9A1 (NHE1)—based on their established roles in maintaining transmembrane sodium gradients. To investigate their dysregulation patterns in breast carcinogenesis, we leveraged two orthogonal transcriptomic repositories: The GSE42568 dataset from GEO provided microarray-based expression profiles, while the TCGA-BRCA cohort (n = 1, 226; Illumina HiSeq RNA-seq) furnished paired raw count data and clinicopathological annotations (AJCC Stage I-IV). This dual-cohort approach enabled cross-platform validation of sodium homeostasis-associated transcriptional alterations, with TCGA data subjected to TPM normalization and combat batch correction to minimize technical variability.

Bioinformatic analyses were conducted in R (v4.2.1) using rigorously validated computational pipelines. Raw microarray data from the GSE42568 dataset were retrieved via the GEOquery package (v2.68.0) and subjected to probe-level quality control: Multi-mapped probes were systematically filtered, retaining only the highest intensity probe per gene using the collapseRows algorithm (WGCNA v1.72). Variance-stabilizing transformation (VST) implemented in DESeq2 (v1.38.0) normalized count distributions prior to differential expression analysis, with statistical significance determined by Wald test (|log₂FC| >1, p < 0.05). Parallel processing of TCGA-BRCA RNA-seq data (STAR-aligned, TPM-normalized) incorporated generalized linear models through the stats package (v4.2.1) with empirical Bayes variance moderation. Dimensionality reduction and visualization workflows generated comparative boxplots, principal component analysis (PCA) biplots, and volcano plots using ggplot2 (v3.4.0) with ggrepel (v0.9.3) for label optimization.

Diagnostic performance evaluation incorporated receiver operating characteristic (ROC) curve analysis implemented in the pROC package (v1.18.0). For TRPM4 and SLC9A1, time-dependent AUC values were calculated to assess discriminative capacity between malignant and non-malignant specimens, with statistical significance determined via DeLong’s test for paired ROC curve comparisons. Survival analytics employed Cox proportional hazards regression models (survival v3.3.1) to evaluate sodium dysregulation-associated mortality risk, incorporating molecular subtype stratification and age-adjusted multivariate analyses. Kaplan-Meier curves with log-rank P-values were generated using survminer, featuring time-to-event endpoints censored. Hazard ratios and 95% confidence intervals were computed through restricted mean survival time comparisons between high/low expression cohorts defined by X-tile optimized cutpoints.

TRPM4 and SLC9A1 expression cohorts were stratified to determine optimal survival-discriminatory thresholds. Baseline clinicopathological characteristics between high/low expression groups were compared using multivariate logistic regression models, incorporating Fisher’s exact tests for categorical variables and Welch’s T-tests for continuous parameters (age, tumor size). Missing clinical data were imputed via chained equations with 10 iterations, while inverse probability weighting adjusted for potential selection bias. All statistical tests were two-sided with Bonferroni correction applied to address multiplicity.

Selected patients were divided into two groups based on TRPM4 and SLC9A1 expression, and target genes were extracted by utilizing package “DESeq2” difference analysis with |log₂(FC)|> 0.5 and adjP value < 0.05 as screening criteria. (GO) including biological process (BP), cellular component (CC), molecular function (MF) and (KEGG) pathway enrichment analyses were performed using package “clusterProfiler [4.4.4]” (21), P < 0.05 was considered statistically significant. Gene Set Enrichment Analysis (GSEA) was conducted using the “clusterProfiler” package, and the top 4 enriched terms were displayed in a mountain plot. Te GSEA results were visualized using the “ggplot2” package.

We utilized the miRNet database (https://www.mirnet.ca/) to predict the association between diferentially expressed mRNAs and miRNAs. Subsequently, an mRNA-miRNA regulatory network was constructed based on the obtained results.

Potential protein interactions with TRPM4 and SLC9A1 were collected and integrated through the STRING database (https://string-db.org/) and extracted the relevant genes from these interactions to conduct PPI network analysis.

Paraffin-embedded breast cancer tissues and paired adjacent normal tissues were collected from 90 patients diagnosed with breast cancer via histopathology at Harbin Medical University Cancer Hospital between January 2010 and December 2024. Inclusion criteria: (1) Histopathologically confirmed invasive breast carcinoma; (2) No prior neoadjuvant therapy, chemotherapy, or radiotherapy before surgery; (3) Age range 18–70 years; (4) At least one measurable lesion confirmed by imaging (mammography, ultrasound, or MRI); (5) Normal hematological, hepatic, renal, and cardiac function parameters. For deceased patients, the cause of death was directly attributed to breast cancer progression. Exclusion criteria: (1) Age <18 or >70 years; (2) Severe comorbidities; (3) History of other malignancies; (4) Preoperative anticancer treatments (targeted therapy, immunotherapy); (5) Incomplete clinical follow-up data. Written informed consent was obtained from all participants, and the study protocol was approved by the Ethics Committee of Harbin Medical University Cancer Hospital.

Te specimens were fxed in formalin, dehydrated, embedded, and sectioned into 4-μm continuous slices. Immunohistochemical analysis was performed using the MaxVision method. The TRPM4 and SLC9A1 antibody was purchased from ProteinTech; Te universal immunohistochemistry MaxVision kit and DAB staining solution were purchased from Proteintech. Te working concentration of the TRPM4 antibody was 1:100, and the working concentration of the SLC9A1 antibody was 1:200. Te specifc staining steps were strictly followed according to the instructions provided with the kit.

Protein content was analyzed with a Bicinchoninic Acid Protein Assay Kit (Thermo Fisher Scientific, Waltham, USA). The protein samples were separated with sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and the protein bands were transferred to PVDF membranes (03010040001, Roche). Membranes were probed overnight at 4°C using primary antibodies against the relevant proteins. After primary antibody removal via washing, secondary antibodies were introduced for a 1-hour incubation period at room temperature. Protein band visualization was achieved through chemiluminescent detection with the Abbkine superkine™ ECL kit (China), imaged on the Tanon system (China).

Statistical analysis was performed using one-way ANOVA to compare the means across different groups. All statistical tests were performed with GraphPad Prism software (version 9; GraphPad Software, La Jolla, CA, USA). Data are presented as means ± standard deviations (SDs), as indicated in the figure legends. A P-value of less than 0.05 was considered statistically significant.

Our investigation of NECSO-associated genes in breast carcinogenesis commenced with systematic analysis of the GSE42568 dataset from GEO. Raw microarray data underwent normalization using variance-stabilizing transformation (VST) via DESeq2 (v1.38.0) to mitigate technical variability ( Figure 1A ). Principal component analysis (PCA) revealed distinct clustering patterns between tumor and normal cohorts ( Figure 1B ), confirming substantial transcriptomic divergence. Differential expression analysis identified 4, 996 genes meeting stringent thresholds (|log₂FC| > 0.5, p < 0.05) through DESeq2’s Wald test, visualized as a volcano plot ( Figure 1C ). From seven candidate NECSO regulators previously implicated in sodium homeostasis (NC1, CLT, DHPs, SLC12A2, SLC8A1, TRPM4, SLC9A1), intersection analysis identified two NECSO-associated DEGs - TRPM4 and SLC9A1 - showing consistent tumor-specific upregulation ( Figure 1D ). Cross-validation using TCGA-BRCA RNA-seq data (n = 1, 098 tumors vs. 113 normals) confirmed significant overexpression of both genes in malignant tissues ( Figure 1E ).

We employed the miRNet database (v2.0) to predict miRNAs regulating TRPM4 and SLC9A1, identifying 28 shared candidates ( Figure 2A ), including hsa-miR-106a-5p, hsa-miR-18a-5p, hsa-miR-101-3p, hsa-miR-16-5p, hsa-miR-27a-3p, hsa-miR-24-3p, hsa-miR-26a-5p, hsa-miR-17-5p, hsa-miR-19a-3p, hsa-let-7e-5p, hsa-miR-19b-3p, hsa-let-7d-5p, hsa-miR-28-5p, hsa-let-7c-5p, hsa-miR-15b-5p, hsa-let-7i-5p, hsa-let-7b-5p, hsa-let-7g-5p, hsa-miR-34a-5p, hsa-miR-1-3p, hsa-miR-7-5p, hsa-miR-196a-5p, hsa-miR-107, hsa-let-7a-5p, hsa-miR-183-5p, and hsa-miR-15a-5p. Predictions integrated TargetScan (v7.2) and miRDB (v6.0) algorithms with a cumulative interaction weight >0.95. Diagnostic evaluation of these miRNAs in the TCGA-BRCA cohort (n = 1, 098) revealed hsa-let-7c-5p and hsa-miR-183-5p as top discriminators between malignant and normal tissues ( Figures 2B-E ). Literature validation through systematic PubMed mining confirmed experimental evidence for all 28 miRNAs in BC pathogenesis, with hsa-miR-34a-5p, hsa-let-7c-5p, and hsa-miR-183-5p showing direct tumor-suppressive roles.

By integrating TCGA and GTEx data, TRPM4 was found to exhibit the highest expression level in PRAD, while SLC9A1 showed the highest expression in ESCA ( Supplementary Figures 1A, B ). Compared to normal tissues, TRPM4 and SLC9A1 expression levels were significantly upregulated in BRCA samples (P < 0.05), with statistical significance, as shown in Figure 3A . In paired samples, both TRPM4 and SLC9A1 were upregulated in BRCA patients ( Figures 3B, C ). An ROC curve constructed using the BRCA dataset from the TCGA database ( Figure 3D ) demonstrated high diagnostic value for TRPM4. Additionally, TRPM4 and SLC9A1 were associated with tumor staging: their levels were upregulated in T3-T4 stage BRCA patients compared to T1-T2 stage patients ( Figure 3E ). The expression of TRPM4 in BRCA tissues was significantly associated with overall survival (OS, P = 0.01) and disease-specific survival (DSS, P = 0.008), but showed no significant correlation with progression-free interval (PFI, P = 0.052). In contrast, the protein expression levels of SLC9A1 and TRPM4 demonstrated no significant prognostic relevance ( Supplementary Figures 2A-F ). COX proportional hazards regression identified TRPM4, as an independent prognostic biomarker in breast cancer. In univariate analysis, high TRPM4 expression was significantly associated with reduced mortality risk (HR = 0.691; 95% CI: 0.498-0.958; P = 0.027), whereas SLC9A1 expression showed no association with prognosis (HR = 1.097; 95% CI: 0.797-1.509; P = 0.571). After adjusting for clinical covariates, TRPM4 retained its independent protective effect (HR = 0.691; 95% CI: 0.498-0.958; P = 0.027), confirming its robust prognostic value. Notably, SLC9A1 failed to demonstrate significance in multivariate analysis, suggesting its expression is confounded by other clinical factors. These findings highlight TRPM4 as a promising biomarker for risk stratification in breast cancer, while SLC9A1 warrants further investigation to elucidate its biological role ( Supplementary Figure 2G ).

Analysis of genetic alterations in TRPM4 and SLC9A1 across cancers was performed using the cBioPortal platform, encompassing 32 studies and 10, 967 samples ( Figures 4A, B ). TRPM4 harbored 12 mutation sites, including 49 missense mutations, 8 in-frame mutations, with R80Q being the most frequent mutation type. SLC9A1 exhibited 9 mutation sites, predominantly missense mutations, along with truncations, splice variants, structural variants (SVs)/fusions, and A781TV as the most common mutation type. Mutations in TRPM4 and SLC9A1 were observed in 32 high-grade breast cancer tissues ( Figures 4C, D ). TRPM4 mutations may impair calcium ion permeability, leading to intracellular sodium/calcium imbalance and promoting tumor migration. Missense mutations in SLC9A1 could disrupt sodium-hydrogen exchange function, exacerbating tumor microenvironment acidification and enhancing chemotherapy resistance. Although somatic mutations in TRPM4 and SLC9A1 are relatively rare events in breast cancer overall ( Figures 4E, F ), the specific mutations that do occur may offer valuable functional insights. For instance, TRPM4 mutations could potentially impair its calcium ion permeability, which might contribute to intracellular sodium/calcium imbalance—a phenomenon supported by our functional data in knockdown models ( Figures 5G, H ). Similarly, missense mutations in SLC9A1 could disrupt sodium-hydrogen exchange, a mechanism that might exacerbate microenvironment acidification. Interestingly, we observed that these mutations were even less frequent in invasive mixed mucinous carcinoma compared to invasive ductal carcinoma, suggesting a distinct genetic landscape. However, given their low prevalence, their contribution to the overall breast cancer population is likely limited. Therefore, the primary clinical relevance of TRPM4 and SLC9A1 in breast cancer is more robustly underscored by their frequent mRNA/protein overexpression and strong association with poor prognosis, as demonstrated throughout our study, rather than by their mutational status.

The protein-protein interaction (PPI) networks of TRPM4 and SLC9A1 were constructed using a high-confidence threshold (score ≥ 0.7) with the top 30 interacting proteins ( Figures 6A, D ). GO/KEGG enrichment analysis of their interacting proteins revealed that TRPM4 was significantly associated with sodium ion transport, sodium ion transmembrane transport-mediated membrane depolarization, ion channel complexes (including cation/sodium channel complexes), and metal ion transmembrane transporter activity ( Figures 6B, E ), while its KEGG pathway enrichment primarily involved taste transduction. GSEA further demonstrated that TRPM4 was enriched in pathways such as the CDC42 GTPase Cycle, Rab Geranylgeranylation, Developmental Biology, and Keratinization ( Figures 6C, F ). In contrast, SLC9A1’s GSEA results highlighted pathways including the Role of LAT2/NTAL in Calcium Mobilization, Scavenging of Heme from Plasma, CD22-Mediated BCR Regulation, Formation of the Cornified Envelope, and Drug Metabolism by Other Enzymes.

The functional states of TRPM4 and SLC9A1 in breast cancer were investigated using the CancerSEA database ( Figures 7A, B ). TRPM4 exhibited strong positive correlations with angiogenesis ( Figure 7C ), while SLC9A1 showed positive correlations with DNA repair, DNA damage, and cell cycle regulation ( Figure 7D ).

We employed immunohistochemistry to analyze the expression of TRPM4 and SLC9A1 in tumor tissues. The results revealed that the expression levels of TRPM4 and SLC9A1 were significantly upregulated compared to those in the normal tissue group ( Figures 8A, B ). Subsequently, we performed RT-qPCR and Western blot assays to validate the knockdown efficiency of si-TRPM4 and si-SLC9A1 ( Figures 8C, D ). We transfected MDA-MB-468 and BT-549 cells with si-RNAs targeting TRPM4 and SLC9A1, and conducted CCK-8 and colony formation assays to evaluate the proliferative capacity mediated by these genes. The colony formation assay demonstrated that the clonogenic ability was significantly reduced in the TRPM4 and SLC9A1 knockdown groups compared to the negative control group ( Figures 8E, F ).

Similarly, the scratch wound healing assay demonstrated that compared to the NC group, the cell migration ability was significantly reduced in the TRPM4 and SLC9A1 knockdown groups ( Figures 5A-D ). Subsequently, the effect of TRPM4 and SLC9A1 knockdown on cell apoptosis was detected by flow cytometry. It was found that the apoptosis rate was significantly higher in cells after knockdown of these two genes compared to the control group ( Figures 5E, F ). To further verify the correlation between these two genes and sodium channels, we examined the expression of NCX1, a core regulator of intracellular sodium and calcium ions. Western blot analysis showed that genetic knockdown of either TRPM4 or SLC9A1 significantly reduced the protein expression level of NCX1 ( Figures 5G, H ).

These results indicate that knockdown of TRPM4 and SLC9A1 significantly inhibits the proliferation and migration abilities of breast cancer cells, thereby suppressing tumor growth. Mechanistically, TRPM4 and SLC9A1 cooperatively regulate proliferative signaling through ion homeostasis and modulate metastatic capacity by affecting the epithelial-mesenchymal transition pathway, positioning them as novel therapeutic targets. This dual functionality, spanning tumorigenesis, progression, and microenvironmental adaptation, provides a pioneering framework for developing precision therapies against breast cancer, particularly for treatment-resistant subtypes.